Split (1)

train · ~9.11M rows (showing the first 93.5k)

identifier stringlengths 11 32	pdf_url stringlengths 17 4.62k ⌀	lang stringclasses 120 values	error stringclasses 1 value	title stringlengths 2 500 ⌀	source_name stringlengths 1 435 ⌀	publication_year float64 1.9k 2.02k	license stringclasses 3 values	word_count int64 0 1.64M	text stringlengths 1 9.75M
https://openalex.org/W2913986178	https://europepmc.org/articles/pmc6450592?pdf=render	English	null	Brain Death and Organ Donation Rates in a City Hospital: A Retrospective Study	Curēus	2,019	cc-by	4,434	Introduction Although organ donation rates have been increasing over the years, the lack of organ donation remains the most important problem in transplantation. By changing strategies, the Cekirge City Hospital in Bursa/Osmangazi has achieved a cadaveric donor rate of 24.9 per one million individuals in 2016; this rate is 21.5 in England, 20.9 in Norway, 14.7 in the Netherlands, and 10.6 in Germany. Brain Death and Organ Donation Rates in a City Hospital: A Retrospective Study Sibel Yılmaz Ferhatoglu , Nihan Yapici 1 1 Sibel Yılmaz Ferhatoglu , Nihan Yapici 1 1 1. Anesthesiology and Reanimation, Siyami Ersek Cardiothoracic Surgery Hospital, Istanbul, TUR  Corresponding author: Sibel Yılmaz Ferhatoglu, sibelyilmazferhatoglu@gmail.com Disclosures can be found in Additional Information at the end of the article Results There were a total of 137 brain death cases. Three of eight cases, five of 12 cases, three of 13 cases, 13 of 25 cases, 16 of 29 cases, and 21 of 50 cases became a donor in 2011, 2012, 2013, 2014, 2015, and 2016, respectively. Methods Brain death cases were retrospectively evaluated between January 1, 2011, and December 31, 2016. © Copyright 2019 Yılmaz Ferhatoglu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License CC-BY 3.0., which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Conclusion Deceased organ donation rates have increased over the years; however, the number of brain dead patients and the acceptance of organ donation by families have been increasing, but the percentage of brain death donations did not increase. We suggest that the reason for this situation is that well-trained and educated physicians diagnose more brain death cases and have a greater desire to treat end-stage organ failure patients, but the tendency of the public to donate has not increased as hoped. Donation and transplantation rates may be increased with a combination of well-trained, educated, and dedicated physicians with public education. Received 01/02/2019 Review began 01/28/2019 Review ended 01/31/2019 Published 02/04/2019 Categories: Anesthesiology, Transplantation Keywords: brain death, determining brain death, transplantation, ancillary test, icu DOI: 10.7759/cureus.4006 Yılmaz Ferhatoglu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License CC-BY 3.0., which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. differ greatly between countries worldwide, and there is no global consensus in diagnostic criteria [2-3]. The American Academy of Neurology published guidelines and practice parameters to help physicians determine brain death. The guidelines of the American Academy of Neurology were published in 1995 and updated in 2010 [4-5]. The brain death determination guidelines of the Turkish Neurological Society were published in 2014 [6]; they are in accordance with the law on the Regulation on Organ and Tissue Transplantation of the Turkish Ministry of Health in 2012 and with the guidelines of the American Academy of Neurology in 2010. Recognizing and determining brain death is important because the number of patients waiting for organ transplantation increases day by day, and the deceased organ donation rates remain insufficient throughout the world. Hesitancy in brain death diagnosis by physicians and the ethical concerns of the families of candidate donor patients are the two largest obstacles in finding the right organ for the receiving patient whose survival depends on the transplantation. Cardiac death during evaluation, donor instability, and death during organ recovery are medical problems that also result in the loss of donor candidate patients. Therefore, teaching and encouraging anesthesiologists, neurologists, and neurosurgeons on how to perform a rapid diagnosis of brain death are crucial. Organ transplantation is the only option for most patients affected by end-stage organ failure, and health policymakers have paid increasing attention to this subject during the last decade. Medical care for end-stage organ failure patients presents a large financial burden on the overall budget of most countries. Over the last decade, the definitive treatment of these patients through organ transplantation has decreased health expenditures; therefore, increasing cadaver donation rates has become more important for health politicians. The universal shortage in cadaver donation has increased living donor rates, particularly for liver and kidney transplantation. A total of 30,973 transplants from 15,064 donors were performed in the United States of America in 2015, and more than 121000 thousand patients affected by end- stage organ failure were waiting for a transplant in 2015 [7]. As the situation is evaluated by the donor candidate patients and their families, age and ethical concerns create an obstacle for organ transplantation. Moreover, less than 1% of all people who die in hospitals can be considered for organ donation because they must die under specific circumstances [8]. Introduction Brain functions were first described by Mollaret et al. in 1959 using a concept similar to the modern definition of brain death. Mollaret and Goulon differentiated "coma dépassé" from other "coma states," such as a vegetative state [1]. Brain death determination criteria were first published in 1968, a year after the first heart transplantation. Attitudes toward brain death How to cite this article Yılmaz Ferhatoglu S, Yapici N (February 04, 2019) Brain Death and Organ Donation Rates in a City Hospital: A Retrospective Study. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 How to cite this article Yılmaz Ferhatoglu S, Yapici N (February 04, 2019) Brain Death and Organ Donation Rates in a City Hospital: A Retrospective Study. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 The aim of this study was to investigate the strategies for determining brain death retrospectively and the statistics of brain death and organ donation in Cekirge State Hospital, Bursa, Turkey. 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 The brain death determining strategy at our hospital In our hospital, we have developed a brain death determining strategy that works in harmony with organ transplantation coordinators, neurologists, and neurosurgeons. The organ transplantation coordination committee consists of two members. One of them has to be a physician and the other one may be a physician or a nurse. We, the anesthesiologists, and the organ transplantation coordinators examine all intensive care unit patients every day, particularly patients in tertiary intensive care units (Glasgow Coma Scale ≤ 7). Patients with head trauma, cerebrovascular disease, and post-cardiopulmonary resuscitation are also followed up closely. When brain death is suspected, first the anesthesiologists and then the neurologists perform a neurologic examination. If the patient is suitable for testing, the anesthesiologists perform the apnea test. If the apnea test is aborted, an ancillary test is performed. The ancillary tests that we can use in our hospital are computed tomography (CT) angiography and magnetic resonance imaging (MRI) angiography. Magnetic resonance imaging 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 2 of 10 angiography is the preferable ancillary test. Even when the apnea test is positive, we perform an ancillary test because it can reduce the time of observation. Finally, a committee consisting of two doctors, one anesthesiologist, and one neurologist or one neurosurgeon confirms the brain death and, then, brain death can be declared to the patient’s family. All procedures are performed according to the Helsinki declaration. Statistical analysis Number Cruncher Statistical System (NCSS; NCSS, LLC., Utah, US) 2007 was used for the statistical analysis. The mean, standard deviation, and percentage values were calculated as descriptive statistics. Qualitative data were compared using Pearson’s chi-square test. A linear- by-linear association test was used to examine the changes over the years. Statistical significance was accepted when p < 0.05. The author of this study, Sibel Yilmaz Ferhatoglu was appointed to Siyami Ersek Cardiothoracic Surgery Hospital from Cekirge State Hospital during the writing process of this article. All patients included in the study were followed-up and treated in the Cekirge State Hospital, Bursa, Turkey, and the assignment of the first author to another hospital did not change the methodology of the study. 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 Number of cases Percentage of cases Male (N) Female (N) Age (mean ± standard deviation) Acute subdural hematoma 10 7.30 10 0 47.30 ± 14.97 Hypoxic encephalopathy 17 12.41 9 8 57.53 ± 18.98 Intracranial tumor 1 0.73 0 1 75 Intracerebral hemorrhage 39 28.47 23 16 57.18 ± 16.31 Cerebrovascular ischemia 21 15.33 9 12 61.67 ± 9.94 Cerebral contusion 6 4.38 6 0 34.50 ± 21.70 Spontaneous subarachnoid hemorrhage 18 13.14 8 10 56.22 ± 15.15 Traumatic subarachnoid hemorrhage 22 16.06 16 6 45.82 ± 18.18 Cerebrospinal fluid fistula/pneumocephalus 1 0.73 0 1 44 Epidural hematoma 1 0.73 1 0 42 Subacute subdural hematoma 1 0.73 0 1 66 TABLE 1: Cause of brain death, age, and sex of the cases TABLE 1: Cause of brain death, age, and sex of the cases TABLE 1: Cause of brain death, age, and sex of the cases Ancillary test In determining brain death, the preference for the use of an ancillary test increased from 2011 to 2016. In 2011, there was no ancillary test confirmation because there was no ancillary test in our hospital. After 2011, ıf there was no contraindication transportation of the patient for an ancillary test, the ancillary test has done. In 2016, 42 cases of brain death were confirmed with ancillary tests (84%). Thirty-one of the 137 cases were determined as brain death only clinically (22.6%), whereas 106 were determined clinically and confirmed with magnetic resonance imaging angiography (77.4%). Apnea test The apnea test was completed and positive in 123 cases (89.9%); in 14 cases (10.1%), the apnea test was aborted primarily because of arrhythmia. Brain death diagnosis time The time from the hospitalization of the patient to the diagnosis of brain death is four days. The mean brain death diagnosis time was 4.40 ± 4.05 days (range 0–26 days). The time is from the hospitalization of the patient to the diagnosis of brain death. Intracranial tumor and cerebrovascular ischemia were the most quickly diagnosed causes. Patient demographics A total of 137 brain death cases were assessed in this study, including 82 male patients (59.9%) and 55 female patients (40.1%). The mean age was 54.23 ± 17.15 years. The mean age of the male patients was 51.04 ± 16.67 years, and the mean age of the female patients was 59.0 ± 16.89 years. All of the patients were over 18 years old. According to blood types, A Rh-positive was the most common blood group (35.8%). None of the brain death cases had an AB Rh-negative blood type. In both genders, intracerebral hemorrhage (28.47%) was the most common cause of brain death. In females, cerebrovascular ischemia (15.33%) was the second most common cause of brain death, and in males, traumatic subarachnoid hemorrhage (16.06%) was the second most common cause of brain death (Table 1). 3 of 10 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 Harvested organs The liver and kidney were the most frequently donated organs, followed by the heart. Over the six-year study period, 61 livers, 100 kidneys, 11 hearts, five corneas, two lungs, and one pancreas were harvested. Of the 137 patients diagnosed with brain death, organs were donated in 61 patients, 26 (42.6%) of whom were female and 35 (57.4%) of whom were male. The acceptance and rejection rates for each year are shown in Figure 1. FIGURE 1: Total number of brain deaths, acceptance, and rejection for each year FIGURE 1: Total number of brain deaths, acceptance, and rejection for each year Donation acceptance and rejection Donation acceptance and rejection 4 of 10 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 Brain death determination rate in our hospital The brain death determination rate in our hospital in 2011 was 0.4%, which increased to 2% in 2016. The brain death determination rate in 2016 was statistically significantly higher than the rates in 2012, 2013, and 2014, with p = 0.001, p < 0.001, and p = 0.007, respectively (Tables 2-3). 2011 2012 2013 2014 2015 2016 Intensive care unit stay 1976 1594 2062 2381 2060 2500 Number of brain death cases 8 12 13 25 29 50 2011 2012 2013 2014 2015 2016 Intensive care unit stay 1976 1594 2062 2381 2060 2500 Number of brain death cases 8 12 13 25 29 50 TABLE 2: The number of patients staying in the intensive care unit and the number of brain death cases per year 019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 5 of 10 TABLE 2: The number of patients staying in the intensive care unit and the number of brain death cases per year 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 5 of 10 Donation Population * Per million people value** 2011 3 789.575 3.7 2012 5 792.219 6.3 2013 3 802.620 3.7 2014 13 813.262 15.9 2015 16 826.742 19.3 2016 21 841.756 24.9 TABLE 4: Organ donation rate / per million people : Population of Bursa, Osmangazi district Donation Population Per million people value** 2011 3 789.575 3.7 2012 5 792.219 6.3 2013 3 802.620 3.7 2014 13 813.262 15.9 2015 16 826.742 19.3 2016 21 841.756 24.9 TABLE 4: Organ donation rate / per million people Years p OR (95% CI) 2011-2012 0.173 1.866 (0.761, 4.576) 2011-2013 0.323 1.561 (0.646, 3.774) 2011-2014 0.019* 2.610 (1.175, 5.800) 2011-2015 0.002 3.513 (1.602, 7.702) 2011-2016 <0.001 5.020 (2.375, 10.614) 2012-2013 0.656 0.836 (0.381, 1.838) 2012-2014 0.339 1.399 (0.701, 2.793) 2012-2015 0.062 1.882 (0.958, 3.701) 2012-2016 0.001** 2.690 (1.428, 5.068) 2013-2014 0.130 1.672 (0.853, 3.278) 2013-2015 0.013* 2.251 (1.167, 4.342) 2013-2016 <0.001 3.217 (1.743, 5.938) 2014-2015 0.278 1.346 (0.786, 2.305) 2014-2016 0.007 1.923 (1.186, 3.119) 2015-2016 0.127 1.429 (0.901, 2.267) TABLE 3: Comparison of organ donation rates by year intervals Pearson's chi-square test p < 0.05 p < 0.01 OR: odds ratio Deceased organ donation rate per million people Deceased organ donation rate increased significantly after 2014. The donor rate in 2016 was 24.9 per million people (Table 4). TABLE 3: Comparison of organ donation rates by year intervals Pearson's chi-square test p < 0.05 *p < 0.01 OR: odds ratio Deceased organ donation rate per million people Deceased organ donation rate increased significantly after 2014. The donor rate in 2016 was 24.9 per million people (Table 4). TABLE 3: Comparison of organ donation rates by year intervals Pearson's chi-square test p < 0.05 *p < 0.01 Deceased organ donation rate per million people Deceased organ donation rate increased significantly after 2014. The donor rate in 2016 was 24.9 per million people (Table 4). 6 of 10 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 : Population of Bursa, Osmangazi district Discussion Organ transplantation is the only solution for most patients affected by end-stage organ failure. Trying to cure these patients by medical treatment is difficult or impossible in most cases and is expensive. An economic analysis of transplantation shows that transplantation is both cheaper and more successful than medical treatment. “Increasing organ transplantation rates by 50% could achieve a further cost saving of two hundred million euros per annum,” as stated in the United Kingdom organ transplantation strategy 2020 review. Increasing organ donation by simple precautions is the main aim of all efforts. The U.K. Organ Donation Taskforce increased rates in after-death organ donation by 50% between 2007/2008 and 2012/2013 by highlighting three key areas: donor identification and reveal, donor coordination, and organ retrieval arrangements [9]. There are different national organ transplantation programs. Spain has developed a successful organ donation program called the Spanish Model of Donation and Transplantation. In this model, organ transplantation coordinators play an important role: they are responsible for enhancing organ donation in their hospital and representing the hospital level of the organizational network [10]. Other successful organ donation models have been established in the United States of America and Australia [10-11]. In our hospital, teamwork between organ donation coordinators, anesthesiologists, neurologists, and neurosurgeons and the development in determining a brain death strategy over time were the key areas contributing to our success. We achieved a cadaveric donor rate of 24.9 per one million individuals in 2016; this rate was 21.5 in England, 20.9 in Norway, 14.7 in the Netherlands, and 10.6 in Germany [12]. We believe that developing special hospital strategies may increase donation rates. In our hospital, 44.5% of brain death cases became donors between 2011 and 2016. Most of the patients diagnosed with brain death were male, similar to other studies [3,13-16]. Trauma is more common in males in some studies in our country, but there is no information about gender in the Turkish Statistical Institute data. The reason for the increased incidence of a brain death diagnosis in males is that most accidents are accompanied by heavy head trauma [17-18]. Intracerebral hemorrhage was the most common cause of brain death in our study, similar to Intracerebral hemorrhage was the most common cause of brain death in our study, similar to 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 7 of 10 In three other studies done in our country, the family acceptance rates for being a donor were 34.2%, 29%, and 8.7%, respectively [14-15,19]. Ethnic and religious aspects are the most important factors affecting the organ donation rate [11,16,20]. In our hospital, deceased organ donation rates have increased over time: in 2011, 37.5% of brain death cases became donors, whereas in 2016, 42% of the cases became donors. Although the number of brain death cases have been increasing, the percentage of organ donation acceptance of families has not been increasing as expected. We hypothesize that this occurs because well-trained and educated physicians experience more brain death cases and have a greater desire to treat end-stage organ failure patients, but the tendency for the public to donate has not increased as hoped. We suggest that both physician awareness and public awareness are important topics for organ donation. According to the Turkish Ministry of Health, there are 24,588 patients currently waiting for organ donation who are on the transplant list in 2016: 21,474 patients for a kidney, 2164 for a liver, 762 for a heart, 272 for a pancreas, 11 for a kidney and a pancreas, three for heart valves, and two for intestines. In 2016, 1999 brain death cases were declared and 564 of these became donors, resulting in 4921 transplantations performed. In 2008, there were two hundred six transplantations; thus, the transplantation rate has increased by 23.8 times in eight years since then [21]. Since there are few studies in the literature that can be used for comparing the data of the present study, the retrospective nature and short follow-up period of the present study are the limitations. On the other hand, we believe that the study gives very important ideas about organ donation and expectations about education related to this topic. the studies by Escudero et al. [13] and Pandey et al. [3]. However, many studies reported different primary causes of brain death [14-15,19]. The difference in the most common cause of brain death in various studies may be a result of using different classifications for the cause of brain death. The brain death determination method differs in each country throughout the world [2-3,13] In our hospital, the clinical determination rate was only 22.6%, whereas the clinical determination and ancillary test confirmation rate was 77.4%. Since there was no contraindication of transportation, we chose to perform an ancillary test to determine brain death in many cases to reduce the observation time. Pandey A. et. al [3] pointed to the variability in diagnosing brain death between 2011 and 2015 in their hospital. In their study, an apnea test was completed in 49.5% of patients, and an ancillary test was completed in 29.8% of patients. Escudero et al. discussed the different practices in brain death diagnoses with a multicenter study that had a clinical determination rate of only 5% and an ancillary test rate of 95% with at least one test [13]. The brain death determination method differs in each country throughout the world [2-3,13] In our hospital, the clinical determination rate was only 22.6%, whereas the clinical determination and ancillary test confirmation rate was 77.4%. Since there was no contraindication of transportation, we chose to perform an ancillary test to determine brain death in many cases to reduce the observation time. Pandey A. et. al [3] pointed to the variability in diagnosing brain death between 2011 and 2015 in their hospital. In their study, an apnea test was completed in 49.5% of patients, and an ancillary test was completed in 29.8% of patients. Escudero et al. discussed the different practices in brain death diagnoses with a multicenter study that had a clinical determination rate of only 5% and an ancillary test rate of 95% with at least one test [13]. In our study, the family acceptance of donation after brain death was 44.5%. In three other studies done in our country, the family acceptance rates for being a donor were 34.2%, 29%, and 8.7%, respectively [14-15,19]. Ethnic and religious aspects are the most important factors affecting the organ donation rate [11,16,20]. In our study, the family acceptance of donation after brain death was 44.5%. 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 Disclosures Human subjects: All authors have confirmed that this study did not involve human participants or tissue. Animal subjects: All authors have confirmed that this study did not involve animal subjects or tissue. Conflicts of interest: In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work. References 1. Machado C, Korein J, Ferrer Y, Portela L, Garcia M, Manero J: The concept of brain death did not evolve to benefit organ transplants. J Med Ethics. 2007, 33:197-200. 10.1136/jme.2006.016931 1. Machado C, Korein J, Ferrer Y, Portela L, Garcia M, Manero J: The concept of brain death did not evolve to benefit organ transplants. J Med Ethics. 2007, 33:197-200. 10.1136/jme.2006.016931 2. Wahlster S, Wijdicks EF, Patel PV, Greer D, Hemphill C, Carone M, Mateen F: Brain death declaration: practices and perceptions worldwide. Neurology. 2015, 84:1870-1879. 10.1212/WNL.0000000000001540 2. Wahlster S, Wijdicks EF, Patel PV, Greer D, Hemphill C, Carone M, Mateen F: Brain death declaration: practices and perceptions worldwide. Neurology. 2015, 84:1870-1879. 10.1212/WNL.0000000000001540 3. Pandey A, Sahota P, Nattanmai P, Newey CR: Variability in diagnosing brain death at an academic medical center. Neuroscience J. 2017, 2017:601795. 10.1155/2017/6017958 4. Scripko PD, Greer DM: An update on brain death criteria: a simple algorithm with complex questions. Neurologist. 2011, 17:237-240. 10.1097/NRL.0b013e318224edfa 5. Wijdicks EFM, Varelas PN, Gronseth GS, Greer DM: Evidence-based guideline update: determining brain death in adults: report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurology. 2010, 74:1911-1918. 10.1212/WNL.0b013e3181e242a8 6. Arsava EM, Demirkaya S, Dora B, et al.: Turkish Neurological Society-Diagnostic guidelines for brain death. Turk J Neurology. 2014, 20:101-104. 7. Girlanda R: Deceased organ donation for transplantation: challenges and opportunities . World J Transplant. 2016, 6:451-459. 10.5500/wjt.v6.i3.451 7. Girlanda R: Deceased organ donation for transplantation: ch World J Transplant. 2016, 6:451-459. 10.5500/wjt.v6.i3.451 8. Australian Organ and Tissue Authority. (2018). Accessed: January 1, 2019: https://donatelife.gov.au/resources/clinical-guidelines-and-protocols/best-practice- guideline-offering-organ-and-tissue. 8. Australian Organ and Tissue Authority. (2018). Accessed: January 1, 2019: https://donatelife.gov.au/resources/clinical-guidelines-and-protocols/best-practice- guideline-offering-organ-and-tissue. 9. Living donor kidney transplantation 2020: a UK strategy . (2017). Accessed: October 27, 2017: https://nhsbtdbe.blob.core.windows.net/umbraco-assets-corp/1434/ldkt_2020_strategy.pdf. 9. Living donor kidney transplantation 2020: a UK strategy . (2017). Accessed: October 27, 2017: https://nhsbtdbe.blob.core.windows.net/umbraco-assets-corp/1434/ldkt_2020_strategy.pdf. 10. Matesanz R, Marazuela R, Dominguez-Gil B, Coll E, Mahillo B, de la Rosa G: The 40 donors per million population plan: an action plan for improvement of organ donation and transplantation in Spain. Transplant Proc. 2009, 41:3453-3456. 10.1016/j.transproceed.2009.09.011 11. Rudge C, Metasanz R, Delmonico FL, Chapman J: International practices of organ donation. Br J Anaesth. 2012, 108:48-55. 10.1093/bja/aer399 12. Global observatory for donation and transplantation. (2017). Accessed: October 27, 2017: http://www.transplant-observatory.org/global/. 13. Escudero D, Valentin MO, Escalante JL, et al.: Brain death diagnosis and organ donation. Anaesthesia. 2015, 70:1130-1139. 10.1111/anae.13065 14. Conclusions Recognizing and determining brain death are two important steps for organ donation. Creating teamwork and harmony between organ donation coordinators, anesthesiologists, neurologists, and neurosurgeons is necessary for the determination of brain death and increasing organ donation rates. Also, the education of health employees and the public is another important issue to achieve better donation rates. Governments and international organizations should work harder on this health problem. We think a national brain death strategy must be identified and governments must pay more attention to this subject. 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 8 of 10 Additional Information 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 References Karasu D, Yılmaz C, Karaduman I, Cınar YS, Buyukkoyuncu PN: Retrospective analysis of patients with brain death. J Med Surg Intensive Care Med. 2015, 6:23-26. 15. Battal M, Horoz A, Karatepe O, Citgez B: Experience of research hospital in determination of brain death. The Medical Bulletin of Sisli Etfal Hospital. 2013, 47:59-62. 16. Palaniswamy V, Sadhasivam S, Selvakumaran S, Priyadharsan J, Ananth SR: Organ donation after brain death in India: a trained intensivist is the key to success. Indian J Crit Care Med. 2016, 20:593-596. 10.4103/0972-5229.192049 17. Karadana GA, Aksu NM, Akkas M, Akman C, Uzumcugil A, Ozmen MM: The epidemiology and 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 9 of 10 2019 Yılmaz Ferhatoglu et al. Cureus 11(2): e4006. DOI 10.7759/cureus.4006 cost analysis of patients presented to emergency department following traffic accidents. Med Sci Moni. 2013, 19:1125-1130. 10.12659/MSM.889539 cost analysis of patients presented to emergency department following traffic accidents. Med Sci Moni. 2013, 19:1125-1130. 10.12659/MSM.889539 18. Dirlik M, Bostancıoglu BC, Elbek T, et al.: Features of the traffic accidents happened in the province of Aydın between 2005 and 2011. Ulus Travma Acil Cerrahi Derg. 2014, 20:353-358. 10.5505/tjtes.2014.18828 19. Uludag O, Sabuncu U, Kusderci H, Kaplan F, Tutak A, Dogukan M: Where are we on organ donation?. J Turk Soc Intens Care. 2016, 14:86-92. 10.4274/tybdd.52824 20. Altınors N, Haberal M: Brain death and transplant in Islamic countries . Exp Clin Transplant. 2016, 3:48-52. 10.6002/ect.tondtdtd2016.P2 21. Turkish Ministry of Health, Chair of Dialysis and Transplantation Report . (2017). Accessed: October 27, 2017: https://organkds.saglik.gov.tr/KamuyaAcikRapor.aspx?q=ORGANNAKLI. 10 of 10
https://openalex.org/W2044637338	https://europepmc.org/articles/pmc2810339?pdf=render	English	null	Allosteric Regulation of HIV-1 Reverse Transcriptase by ATP for Nucleotide Selection	PloS one	2,010	cc-by	9,076	Abstract Background: Human immunodeficiency virus type 1 reverse transcriptase (HIV-1 RT) is a DNA polymerase that converts viral RNA genomes into proviral DNAs. How HIV-1 RT regulates nucleotide selectivity is a central issue for genetics and the nucleoside analog RT inhibitor (NRTI) resistance of HIV-1. Methodology/Principal Findings: Here we show that an ATP molecule at physiological concentrations acts as an allosteric regulator of HIV-1 RT to decrease the Km value of the substrate, decrease the kcat value, and increase the Ki value of NRTIs for RT. Computer-assisted structural analyses and mutagenesis studies suggested the positions of the ATP molecule and NRTI- resistance mutations during a catalytic reaction, which immediately predict possible influences on nucleotide insertion into the catalytic site, the DNA polymerization, and the excision reaction. Conclusions/Significance: These data imply that the ATP molecule and NRTI mutations can modulate nucleotide selectivity by altering the fidelity of the geometric selection of nucleotides and the probability of an excision reaction. Citation: Yokoyama M, Mori H, Sato H (2010) Allosteric Regulation of HIV-1 Reverse Transcriptase by ATP for Nucleotide Selection. PLoS ONE 5(1): e8867. doi:10.1371/journal.pone.0008867 Mori H, Sato H (2010) Allosteric Regulation of HIV-1 Reverse Transcriptase by ATP for Nucleotide Selection. PLoS ONE 5(1): e8867 0008867 Editor: Jean-Pierre Vartanian, Institut Pasteur, France Editor: Jean-Pierre Vartanian, Institut Pasteur, France oyama et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits tion, and reproduction in any medium, provided the original author and source are credited. Copyright: 2010 Yokoyama et al. This is an open-access article distributed under the terms of the Creative Commons Attribut unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was supported by a grant from the Ministry of Health, Labor and Welfare, Japan. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. nding: This work was supported by a grant from the Ministry of Health, Labor and Welfare, Japan. The funders had no role in study desi alysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: yokoyama@nih.go.jp to the excision reaction of the RT to remove the chain terminating NRTIs [8,9,10]. Introduction Human immunodeficiency virus type 1 reverse transcriptase (HIV- 1 RT) is an RNA-dependent DNA polymerase that converts single- stranded viral RNA genomes into double-stranded proviral DNAs after HIV-1 entry into the cells. Active HIV-1 RT is composed of two related chains, termed p51 and p66 [1]. The p66 chain has a catalytic site for DNA polymerization: the fingers, palm, and thumb subdomains form a cavity for the binding of the template, primer, two divalent cations, and dNTPs for DNA synthesis [1], as seen in other DNA polymerases. Although HIV-1 RT exhibits no exonucleolytic proofreading activity, it still retains a relatively high level of fidelity of DNA synthesis, i.e., about 2.5–661024 base substitutions per site [2,3]. Increasing evidence suggests that the high fidelity of DNA synthesis achieved by DNA polymerases—i.e., the discrimination of the correct and incorrect nucleotides for polymer- ization—is primarily due to the geometric selection of nucleotides during nucleotide insertion into the catalytic site [4,5,6]. Enzyme activity is often modulated by an allosteric effector, a small natural compound that binds to the enzyme at a site distinct from the substrate-binding site. In this study, we show by kinetic, structural, and mutagenesis studies that the ATP molecule can act as an allosteric effector of HIV-1 RT to modulate nucleotide selectivity and DNA polymerization. We also show probable three- dimensional (3-D) positions of the bound ATP molecule and NRTI-resistance mutations during a catalytic cycle. The obtained data suggest that the ATP molecule and NRTI mutations can cooperatively modulate physicochemical properties of the p66 catalytic cavity to alter the fidelity of the geometric selection of nucleotides and the probability of an excision reaction. An ATP molecule is a multifunctional nucleotide that exists at a concentration of ,3.2 mM in the cells [7]. Many studies have suggested that the ATP molecule is a cellular factor involved in the drug resistance of HIV-1. Nucleoside analog RT inhibitors (NRTIs) act as chain terminators blocking DNA synthesis, since they lack the 39-OH group required for the phosphodiester bond formation, whereas NRTI-resistant RT catalyzes dinucleoside polyphosphate synthesis in the presence of millimolar concentra- tions of NTP [8]. Thus, the ATP molecule at physiological concentrations in vitro serves as an effective pyrophosphate donor Abstract A previous crystal structure study identified a binding site of ATP in the catalytic cavity of p66 when the RT was free from the template and primer [11]. Although ATP-mediated excision provides a plausible mechanism for the NRTI resistance of HIV-1, some NRTI-resistance mutations are located distantly from the excision site. Therefore, their roles in NRTI resistance are not fully understood [12]. Masaru Yokoyama*, Hiromi Mori, Hironori Sato Pathogen Genomics Center, National Institute of Infectious Diseases, Musashi Murayama-shi, Tokyo, Japan Pathogen Genomics Center, National Institute of Infectious Diseases, Musashi Murayama-shi, Tokyo, Japan Results The ERT-mt6 RT has an 11-amino-acid insertion in the b3-b4 loops of the p66 fingers subdomain and four substitutions–M41L, T69I, L210W, and T215Y–in the polypeptide backbone of 93JP-NH1 [13]. These mutations confer higher levels of resistance of the 93JP-NH1 virus against AZT, d4T, b-L-29,39-dideoxy-3-thiacytidine, 29,39-dideoxyinosine, and 29,39-dideoxycytidine than other mutants in the polypeptide backbone of 93JP-NH1 [13]. Therefore, we used ERT-mt6 RT, which clearly showed that NRTI-resistance mutations enhance the effect of ATP on enzyme kinetics in NRTI-sensitive RT. We next examined whether or not the above ATP effects on NRTI resistance were specific to the ATP molecule by measuring the IC50 of AZTTP in the presence of 5 mM UTP, CTP, GTP, NaPO4, dATP, ADP, cAMP, and AMP-PNP (Figure 2B). While many of the compounds tested increased the IC50 of AZTTP, the magnitude of the fold increase was consistently greater with ERT- mt6 RT than 93JP-NH1 RT. The ATP molecule was most effective at increasing the IC50, yielding an approximately 350-fold increase with the ERT-mt6 RT. The ATP, dATP, and GTP molecules, which have a purine ring, had greater effects on the IC50 increase than the UTP and CTP molecules, which have a pyrimidine ring. NaPO4 and cAMP had little effect on the IC50 in either RT. Notably, AMP-PNP, a non-hydrolyzed analogue of the ATP molecule, also increased the IC50 of AZTTP by approxi- mately 150-fold with the ERT-mt6 RT. These data suggest that nucleotides similar in size to the ATP molecule can assist in the development of NRTI resistance. The initial velocities of dTTP incorporation into poly (rA)?p(dT)12-18 were measured using purified p51/p66 RT heterodimers of the 93JP-NH1 and ERT-mt6 RTs (Figure S1A). In both RTs, the dTTP incorporation followed Michaelis-Menten kinetics (Figure S1B). The Km values for the 93JP-NH1 and ERT- mt6 RTs were 4.060.1 and 13.860.5 mM, respectively, suggesting that the dTTP has a higher Km value for ERT-mt6 RT. The kcat values for 93JP-NH1 and ERT-mt6 RTs were 1.0460.01 and 0.4960.01s21, respectively, suggesting that ERT-mt6 RT has a nucleotide addition reaction with a slower turnover rate. We then used the two RTs to examine whether or not the ATP molecule influences the DNA polymerization kinetics by measur- ing the initial velocity of dTTP incorporation into poly (rA)?p(dT)12-18 in the presence of 0, 1, 2, 3, and 4 mM of ATP. Results The velocity is decreased in both RTs in association with an increase in ATP concentration, indicating that the ATP molecule inhibits the overall catalytic reaction by HIV-1 RT (Figure 1A). Lineweaver-Burk double-reciprocal plots showed that, in both RTs, the straight lines at the different ATP concentrations have different x- and y-intercepts, and that the line slope increases with increasing ATP concentration (Figure 1B). The data suggest that the ATP-mediated inhibition of dTTP incorporation is a mixed noncompetitive inhibition. We further examined whether or not the ATP molecule influences the Ki values of AZTTP and d4TTP to the two RTs in the enzyme reaction. Based on the kinetics data for ATP in Figure 1, the kinetics data for AZTTP in Figure S2, the reported kinetics data for AZTTP inhibition [17,18], and a crystal structure study of the ATP-RT complex [11], a simplified kinetics model in which ATP functions as a mixed noncompetitive inhibitor and nucleoside analogs function as competitive inhibitors was hypoth- esized (Figure 2C). We measured the initial velocities of dTTP incorporation by HIV-1 RTs at various ATP concentrations and various AZTTP or d4TTP concentrations, and calculated the Ki values of these compounds for the 93JP-NH1 and ERT-mt6 RTs using Equations 5 and 6 (see Materials and Methods). Notably, the ATP molecule induced increases in the Ki NRTI value in a dose- dependent manner. The magnitudes of the increases were much greater with ERT-mt6 RT, reaching about 30-fold for the Ki AZTTP value and 8.5-fold for the Ki d4TTP value at 3 mM of ATP compared to those without the ATP molecule (Figure 2D). The magnitude of changes in Ki NRTI values was much smaller with 93JP-NH1 RT, which yielded about 6.8- and 0.6-fold increases in Ki AZTTP and Ki d4TTP values, respectively, at 3 mM of ATP compared to those without the ATP molecule. These data suggest that physiological concentrations of ATP [7] potently increase the Ki value of NRTI when RT has NRTI-resistance mutations. We further examined whether or not the ATP molecule can influence the Km of substrate and kcat values in the enzyme reaction. The substrate-velocity data in Figure 1A were fit to Equations 1 and 2 (see Materials and Methods), and the average Km and kcat values in the presence of 0, 1, 2, 3, and 4 mM of ATP were obtained with six independent experiments. Structural Study on ATP Action A previous study identified a binding site of the ATP molecule in the p66 when the RT was free from the template and primer [11]. To address the binding site of the ATP molecule in RT during DNA synthesis, we conducted a computer-assisted structural analysis. Using the homology modeling method [19], we first constructed 3-D models of 93JP-NH1 and ERT-mt6 RTs at various catalytic stages defined by biochemical and crystallo- graphic data [14,20,21] (Figure 3). In the DNA polymerization processes, each single nucleotide addition cycle was divided into four steps, termed the post-translocation, fingers-open ternary, Results Effects of ATP on HIV-1 RT Reaction Kinetics Effects of ATP on HIV-1 RT Reaction Kinetics Effects of ATP on HIV-1 RT Reaction Kinetics First, we analyzed the effects of ATP on HIV-1 RT reaction kinetics. We began by collecting basic information on the steady- state kinetics of DNA polymerization in the absence of the ATP molecule. We used two HIV-1 RTs for the present study: the January 2010 \| Volume 5 \| Issue 1 \| e8867 January 2010 \| Volume 5 \| Issue 1 \| e8867 1 PLoS ONE \| www.plosone.org Regulation of HIV RT by ATP molecule, nanomole orders of AZTTP effectively inhibited dTTP incorporation by both RTs (Figure 2A, top), suggesting that, in the absence of the ATP molecule, the level of inhibition of DNA polymerization by AZTTP is equivalent between the two RTs. In the presence of 1 mM ATP, the inhibition was more moderate with the ERT-mt6 RT than with the 93JP-NH1 RT, and the difference in the inhibition curve became much greater in the presence of 5 mM ATP (Figure 2A, middle and bottom). These data suggest that the ATP molecule and NRTI mutations cooperatively reduce the NRTI sensitivity of HIV-1 RT in vitro. molecule, nanomole orders of AZTTP effectively inhibited dTTP incorporation by both RTs (Figure 2A, top), suggesting that, in the absence of the ATP molecule, the level of inhibition of DNA polymerization by AZTTP is equivalent between the two RTs. In the presence of 1 mM ATP, the inhibition was more moderate with the ERT-mt6 RT than with the 93JP-NH1 RT, and the difference in the inhibition curve became much greater in the presence of 5 mM ATP (Figure 2A, middle and bottom). These data suggest that the ATP molecule and NRTI mutations cooperatively reduce the NRTI sensitivity of HIV-1 RT in vitro. NRTI-sensitive RT (93JP-NH1) and multi-NRTI-resistant RT (ERT-mt6) [13]. The ERT-mt6 RT has an 11-amino-acid insertion in the b3-b4 loops of the p66 fingers subdomain and four substitutions–M41L, T69I, L210W, and T215Y–in the polypeptide backbone of 93JP-NH1 [13]. These mutations confer higher levels of resistance of the 93JP-NH1 virus against AZT, d4T, b-L-29,39-dideoxy-3-thiacytidine, 29,39-dideoxyinosine, and 29,39-dideoxycytidine than other mutants in the polypeptide backbone of 93JP-NH1 [13]. Therefore, we used ERT-mt6 RT, which clearly showed that NRTI-resistance mutations enhance the effect of ATP on enzyme kinetics in NRTI-sensitive RT. NRTI-sensitive RT (93JP-NH1) and multi-NRTI-resistant RT (ERT-mt6) [13]. Results Notably, both Km and kcat values for the 93JP-NH1 and ERT-mt6 RTs monotonically decreased with increasing ATP concentration (Figure 1C). We also examined the Ki values of ATP to the two RTs. On the basis of information on the RT catalytic cycle [14], we assumed two structures of RT for ATP binding: the RT-template-primer complex (RT complex 1) and the RT-template-primer-dTTP complex (RT complex 2). Using Equations 3 and 4 (see Materials and Methods), we calculated Ki ATP and Ki9 ATP values of ATP to the RT complexes 1 and 2, respectively. The Ki ATP values were 2.961.3 and 2.861.3 mM for the 93JP-NH1 and ERT-mt6 RTs, respectively, suggesting that the ATP molecule binds with the equivalent Ki ATP value to complex 1 of the two RTs. The Ki9 ATP values were 1.260.5 and 1.160.4 mM for the 93JP-NH1 and ERT-mt6 RTs, respectively, suggesting that the ATP molecule also binds with the equivalent Ki9 ATP value to complex 2 of these RTs. These results are consistent with the finding that AZT resistance mutations cause no difference in ATP binding [15,16]. Finally, the Ki ATP value was larger than the Ki9 ATP value in both RTs, suggesting that ATP binds with higher affinity to the dTTP- bound RT than to the substrate-free RT. ATP’s Effects on NRTI Action Next, we examined whether or not the ATP molecule influences the action of NRTI on the two RTs. In the absence of the ATP PLoS ONE \| www.plosone.org January 2010 \| Volume 5 \| Issue 1 \| e8867 2 Regulation of HIV RT by ATP Figure 1. Effects of ATP on HIV-1 RT reaction kinetics. A. The substrate-velocity curves of purified HIV-1 RTs in the presence of ATP. RNA- dependent DNA polymerase activity [13] of the purified RTs was measured using various concentrations of [a-32P]dTTP and poly (rA)?p(dT)12-18 in the presence of ATP. Representative results with 93JP-NH1 RT (left) and ERT-mt6 RT (right) are shown. B. Lineweaver-Burk double-reciprocal plots for ATP-dependent inhibition of dTTP incorporation. Reciprocal values of the initial velocities and substrate concentrations in Figure 1A are plotted. C. Effects of ATP on Km (left) and kcat (right) values in RT reaction. The Km and kcat values were estimated by fitting of the initial velocity of dTTP incorporation to Equations 3 and 4 as described in Materials and Methods. The mean values with variances of the six independent experiments are shown. doi:10.1371/journal.pone.0008867.g001 Figure 1. Effects of ATP on HIV-1 RT reaction kinetics. A. The substrate-velocity curves of purified HIV-1 RTs in the presence of ATP. RNA- dependent DNA polymerase activity [13] of the purified RTs was measured using various concentrations of [a-32P]dTTP and poly (rA)?p(dT)12-18 in the presence of ATP. Representative results with 93JP-NH1 RT (left) and ERT-mt6 RT (right) are shown. B. Lineweaver-Burk double-reciprocal plots for ATP-dependent inhibition of dTTP incorporation. Reciprocal values of the initial velocities and substrate concentrations in Figure 1A are plotted. C. Effects of ATP on Km (left) and kcat (right) values in RT reaction. The Km and kcat values were estimated by fitting of the initial velocity of dTTP incorporation to Equations 3 and 4 as described in Materials and Methods. The mean values with variances of the six independent experiments are shown. doi:10 1371/journal pone 0008867 g001 doi:10.1371/journal.pone.0008867.g001 consistent with the position needed to initiate the base pair formation at the enzyme active center after rotation (,20u) of the b3-b4 loops to form a fingers-closed ternary complex [5] (Figures 3B and 3C). fingers-closed ternary, and pre-translocation complex stages [14] (Figure 3). The 3-D models were used to search for a possible binding site for nucleotides by docking simulations [22]. ATP’s Effects on NRTI Action A previous study suggested that incoming dNTP and NRTIs bind along the p66 fingers subdomain of the fingers-open ternary RT complex at the post-translocation stage [23] (Figures 3A and 3B), as is generally seen in other polymerases [24,25,26,27,28]. The binding position was in agreement with the biochemical mode of NRTI inhibition, competitive inhibition, as shown by our kinetic study of AZTTP (Figure S2) and previous studies [17,18]. This position was also The 3-D models of pre- and post-translocation complexes corresponding to E and ES in the kinetic model (Figure 2C) were used to search for a possible binding site for the ATP molecule by docking simulations [22]. The simulations suggested that the ATP molecule could bind to the 93JP-NH1 and ERT-mt6 RTs at the pre-translocation (Figure 3D) and the post-translocation (Figure 3A) stages. The ATP molecule was predicted to bind PLoS ONE \| www.plosone.org January 2010 \| Volume 5 \| Issue 1 \| e8867 3 Regulation of HIV RT by ATP Figure 2. Effects of ATP on NRTI action. A. Effects of ATP on AZTTP-dependent inhibition of RT activities. The dTMP incorporation into poly (rA)?p(dT)12-18 was measured using [a-32P]dTTP and purified p51/p66 heterodimers in the presence of the indicated concentrations of ATP and AZTTP. Ratios of the dTTP incorporation at given concentrations of AZTTP to that in the absence of AZTTP are shown. B. Effects of nucleotides and related compounds on IC50s of AZTTP. IC50s of AZTTP were determined in the presence of 5 mM of the indicated compounds, and the fold increases in IC50 compared to AZTTP without the compounds are shown. C. A simplified kinetics model of DNA polymerization in the presence of ATP and NRTI. The model was generated on the basis of the kinetics data in Figure 1 and Figure S2, previously reported kinetics data [17,18], and a crystal structure study of the ATP-RT complex [11]. D. Effects of ATP on the Ki values of AZTTP and d4TTP. The Ki AZTTP and Ki d4TTP values were estimated by fitting the initial velocity of dTTP incorporation to Equation 5 as described in Materials and Methods. The mean values with variances are shown for two independent experiments performed with duplicate samples. doi:10.1371/journal.pone.0008867.g002 Figure 2. Effects of ATP on NRTI action. A. Effects of ATP on AZTTP-dependent inhibition of RT activities. ATP’s Effects on NRTI Action The dTMP incorporation into poly (rA)?p(dT)12-18 was measured using [a-32P]dTTP and purified p51/p66 heterodimers in the presence of the indicated concentrations of ATP and AZTTP. Ratios of the dTTP incorporation at given concentrations of AZTTP to that in the absence of AZTTP are shown. B. Effects of nucleotides and related compounds on IC50s of AZTTP. IC50s of AZTTP were determined in the presence of 5 mM of the indicated compounds, and the fold increases in IC50 compared to AZTTP without the compounds are shown. C. A simplified kinetics model of DNA polymerization in the presence of ATP and NRTI. The model was generated on the basis of the kinetics data in Figure 1 and Figure S2, previously reported kinetics data [17,18], and a crystal structure study of the ATP-RT complex [11]. D. Effects of ATP on the Ki values of AZTTP and d4TTP. The Ki AZTTP and Ki d4TTP values were estimated by fitting the initial velocity of dTTP incorporation to Equation 5 as described in Materials and Methods. The mean values with variances are shown for two independent experiments performed with duplicate samples. doi:10.1371/journal.pone.0008867.g002 January 2010 \| Volume 5 \| Issue 1 \| e8867 PLoS ONE \| www.plosone.org 4 Regulation of HIV RT by ATP Figure 3. Structural models of the HIV-1 RT p66 subunit in a DNA polymerization cycle. The 3-D models of the 93JP-NH1 p66-template- primer ternary complex of the fingers-open configuration at post-translocation (A), fingers-open configuration at the stage of dTTP binding (B), fingers-closed configuration after fingers-rotation (C), and fingers-open configuration at pre-translocation stage (D). The models were constructed by homology modeling and docking simulation techniques using two crystal structures [1,14] of the HIV-1 RTs as modeling templates (see Materials and Methods). Catalytic clefts composed of fingers, palm, and thumb subdomains are shown. dTTP, magenta sticks; p66 main chain, grey ribbon; Mg2+ ion, gray spheres; template-primer, grey sticks; b3-b4 loop of the fingers subdomain, blue ribbon. doi:10.1371/journal.pone.0008867.g003 Figure 3. Structural models of the HIV-1 RT p66 subunit in a DNA polymerization cycle. The 3-D models of the 93JP-NH1 p66-template- primer ternary complex of the fingers-open configuration at post-translocation (A), fingers-open configuration at the stage of dTTP binding (B) fingers-closed configuration after fingers-rotation (C), and fingers-open configuration at pre-translocation stage (D). The models were constructed by homology modeling and docking simulation techniques using two crystal structures [1,14] of the HIV-1 RTs as modeling templates (see Materials and Methods). Catalytic clefts composed of fingers, palm, and thumb subdomains are shown. dTTP, magenta sticks; p66 main chain, grey ribbon; Mg2+ ion, gray spheres; template-primer, grey sticks; b3-b4 loop of the fingers subdomain, blue ribbon. doi:10.1371/journal.pone.0008867.g003 Figure 3. Structural models of the HIV-1 RT p66 subunit in a DNA polymerization cycle. The 3-D models of the 93JP-NH1 p66-template- primer ternary complex of the fingers-open configuration at post-translocation (A), fingers-open configuration at the stage of dTTP binding (B), fingers-closed configuration after fingers-rotation (C), and fingers-open configuration at pre-translocation stage (D). The models were constructed by homology modeling and docking simulation techniques using two crystal structures [1,14] of the HIV-1 RTs as modeling templates (see Materials and Methods). Catalytic clefts composed of fingers, palm, and thumb subdomains are shown. dTTP, magenta sticks; p66 main chain, grey ribbon; Mg2+ ion, gray spheres; template-primer, grey sticks; b3-b4 loop of the fingers subdomain, blue ribbon. doi:10.1371/journal.pone.0008867.g003 Figure 3. Structural models of the HIV-1 RT p66 subunit in a DNA polymerization cycle. The 3-D models of the 93JP-NH1 p66-template- primer ternary complex of the fingers-open configuration at post-translocation (A), fingers-open configuration at the stage of dTTP binding (B), fingers-closed configuration after fingers-rotation (C), and fingers-open configuration at pre-translocation stage (D). The models were constructed by homology modeling and docking simulation techniques using two crystal structures [1,14] of the HIV-1 RTs as modeling templates (see Materials and Methods). Catalytic clefts composed of fingers, palm, and thumb subdomains are shown. dTTP, magenta sticks; p66 main chain, grey ribbon; Mg2+ ion, gray spheres; template-primer, grey sticks; b3-b4 loop of the fingers subdomain, blue ribbon. doi:10.1371/journal.pone.0008867.g003 along the highly conserved motif A near the side chains of R72, D110, D113, F116, D185, and K219 residues at the p66 fingers subdomain of both RTs (Figure 4). The ATP-binding position was stabilized through electrostatic and hydrophobic interactions between the ATP molecule and the side chains of surrounding amino acids. The ATP position was similar to the ATP position in the crystal structure of the template-primer-free RT [11] and was indistinguishable between the pre- and post-translocation stages (Figure S3), suggesting that a specific ATP-binding site is preserved in the free-RT and RT-template-primer tertiary complex. The ATP-binding position was distinct from that of dNTP at initial binding [23] and after fingers-domain rotation [1] in the fingers- open and -closed configurations of RT, respectively (Figures 3 and 4), consistent with our kinetic data for allosteric regulation (Figures 1 and 2). The bound ATP molecule was located near the YMDD motif, motif A, and the 39-end of the primer, suggesting that the ATP binding can modulate polymerization and support the excision PLoS ONE \| www.ploson January 2010 \| Volume 5 \| Issue 1 \| e8867 PLoS ONE \| www.plosone.org January 2010 \| Volume 5 \| Issue 1 \| e8867 5 Regulation of HIV RT by ATP Figure 4. Docking simulations of ATP to the HIV-1 RT p66 subunit with NRTI resistance. ATP was docked with the optimized p66- template-primer complex of the ERT-mt6 strain at the pre-translation stage, using the automated ligand docking program ASEDock2005 [22] operated in the Molecular Operating Environment (see Materials and Methods). Catalytic clefts composed of fingers, palm, and thumb subdomains are shown. ATP, red sticks; p66 main chain, grey ribbon; template-primer, grey sticks; motif A, blue ribbon. The side chains of amino acids around ATP are indicated with cyan sticks, and the side chains of amino acids for NRTI resistance (M41L, T69I, L210W, and T215W) with orange sticks. The main chain of an 11-amino-acid insertion at the b3-b4 loops for NRTI resistance is shown in orange. doi:10.1371/journal.pone.0008867.g004 Figure 4. Docking simulations of ATP to the HIV-1 RT p66 subunit with NRTI resistance. ATP was docked with the optimized p66- template-primer complex of the ERT-mt6 strain at the pre-translation stage, using the automated ligand docking program ASEDock2005 [22] operated in the Molecular Operating Environment (see Materials and Methods). Catalytic clefts composed of fingers, palm, and thumb subdomains are shown. ATP, red sticks; p66 main chain, grey ribbon; template-primer, grey sticks; motif A, blue ribbon. The side chains of amino acids around ATP are indicated with cyan sticks, and the side chains of amino acids for NRTI resistance (M41L, T69I, L210W, and T215W) with orange sticks. The main chain of an 11-amino-acid insertion at the b3-b4 loops for NRTI resistance is shown in orange. doi:10.1371/journal.pone.0008867.g004 established [14]. All of the tested substitutions changed the overall DNA polymerization activity of the RT (Figure S4B). The substitutions at positions 72, 110, and 116 (R72A/Q, D110A/N, and F116A/L) resulted in a loss of DNA polymerization activity, suggesting their essential role in this activity. In contrast, those at positions 113 and 219 (D113A/N and K219Q/A) enhanced the incorporation of dTTP (Figure S4B), suggesting their regulatory role in overall DNA polymerization activity. reaction (Figure 4). Site-Directed Mutagenesis Study We further examined how substitutions of amino acids around the predicted ATP-binding site would influence the biochemical properties of the ERT-mt6 RT. Single-amino-acid substitutions were introduced into the p66 chain of ERT-mt6, and their effects on the overall DNA polymerization activity, IC50 of AZTTP, Km value, kcat value, and Ki value of ATP for the ERT-mt6 RT were analyzed. The positions of the substitutions introduced corresponded to positions 72, 110, 113, 116, and 219 of the 93JP-NH1 p66. We did not conduct mutagenesis of D185 in the YMDD loops, because its essential role in the translocation of the template primer has been The bound ATP molecule was positioned between the catalytic site and the b3-b4 loops, suggesting that the ATP binding can modulate the initial binding and translocation of dNTP and NRTIs into the catalytic site (Figures 3 and 4). Taken together, these structural data are well consistent with our kinetic data, biochemical data for excision [8,9,10], and crystal structure data for ATP binding [11]. NRTI-resistance mutations of the ERT-mt6 p66 (Figure 4, orange residues) were located relatively far away from the bound ATP molecule and catalytic center in p66. Thus, it is less likely that these mutations directly influence the ATP-mediated excision. Instead, the M41L, T69I, L210W, and T215Y substitutions augmented the hydrophobicity of the catalytic cavity of p66, which could enhance p66’s ability to exclude water from the catalytic site cleft for a higher fidelity of nucleotide selection [6,29]. This possibility is consistent with our kinetic data. The fingers-domain insertion induced changes in the shape of the b3-b4 loops that could alter the position of the initial binding site of dNTP and NRTI relative to the catalytic site. The active D113A/N and K219Q/A RTs were further examined for changes in the IC50 of AZTTP, and in the Km, kcat, Ki ATP, and Ki9 ATP values. The D113A/N resulted in an approximately 4- to 5-fold reduction in the IC50 of AZTTP (Figure S4C), suggesting that D113 plays an important role in the development of NRTI resistance. Km, kcat, Ki ATP, and Ki9 ATP values were estimated by using the substrate-velocity curves for the D113A/N and K219Q/A RTs (Figure S4D). The D113A/N and K219Q/A substitutions induced changes in the Km, kcat, Ki ATP, and Ki9 ATP values, suggesting that the D113 and K219 residues regulate the Km of substrate, kcat, and Ki values of ATP (Figure 5). The D113A/N substitutions resulted in reductions in Km and kcat values, which paralleled the reductions in Ki ATP and Ki9 ATP values. The K219Q/A substitutions resulted in increases in Km values, which paralleled the increases in Ki ATP values. The kinetics data implied that residues 113 and 219 can regulate the affinity of the substrate and ATP molecule but do not contribute directly to the catalysis of DNA polymerization, suggesting that the ATP binding site would be distinct from the catalytic site for DNA polymerization. Discussion How HIV-1 RT regulates the nucleotide selectivity for DNA synthesis is a central issue for genetics and the NRTI resistance of PLoS ONE \| www.plosone.org January 2010 \| Volume 5 \| Issue 1 \| e8867 6 Regulation of HIV RT by ATP Figure 5. Site-directed mutagenesis study of the HIV-1 RT p66 subunit with NRTI resistance. Single substitutions of amino acids around the predicted ATP-binding site in Figure 3 were introduced into the p66 chain of ERT-mt6. The overall DNA polymerization activity (Figure S4B), IC50 of AZTTP (Figure S4C), and Km, kcat, Ki ATP, and Ki9 ATP values were measured using the [a-32P]dTTP and poly (rA)?p(dT)12-18 system, and fold increases in the Km (A), kcat (B), Ki ATP (C), and Ki9 ATP values (D) compared to those for the ERT-mt6 were calculated. Results for the RT mutants, D113A, D113N, K219Q, and K219A, which retained sufficient polymerization activity for a kinetic study, are shown. doi:10.1371/journal.pone.0008867.g005 Figure 5. Site-directed mutagenesis study of the HIV-1 RT p66 subunit with NRTI resistance. Single substitutions of amino acids around the predicted ATP-binding site in Figure 3 were introduced into the p66 chain of ERT-mt6. The overall DNA polymerization activity (Figure S4B), IC50 of AZTTP (Figure S4C), and Km, kcat, Ki ATP, and Ki9 ATP values were measured using the [a-32P]dTTP and poly (rA)?p(dT)12-18 system, and fold increases in the Km (A), kcat (B), Ki ATP (C), and Ki9 ATP values (D) compared to those for the ERT-mt6 were calculated. Results for the RT mutants, D113A, D113N, K219Q, and K219A, which retained sufficient polymerization activity for a kinetic study, are shown. doi:10.1371/journal.pone.0008867.g005 We next investigated the mechanisms by which the ATP molecule modulates the nucleotide selectivity of HIV-1 RT. Docking simulations predicted that the ATP-binding site would be similar between NRTI-sensitive and -resistant RTs during DNA polymerization. Consistent with the results of the kinetic study, the predicted ATP-binding site was distinct from that of dNTP and NRTIs [23], and single-amino-acid substitutions at positions 113 and 219 around the predicted ATP-binding site indeed induced significant changes in the Ki value of ATP. Importantly, the ATP- binding position suggested possible mechanisms by which ATP could influence DNA polymerization and the excision reactions of RT, as follows: First, interactions between the c-phosphate of the ATP molecule and the side chain of D185 in the YMDD loops could influence the DNA translocation of the primer template [14]. Discussion Second, interactions between charged portions of the ATP molecule and the side chains of D110 and D185 could modulate the Mg2+ position and stability for DNA polymerization [1]. Third, the c-phosphate of the ATP molecule is located near the 59 phosphate of DNA primer terminus and thus could increase the probability of a DNA excision [11] in concert with a reduction in the kcat value of RT. HIV-1. In this study, we showed that the ATP molecule at physiological concentrations acted as an allosteric regulator of HIV-1 RT to modulate nucleotide selectivity. We also showed probable 3-D positions of the bound ATP molecule and NRTI mutations in the catalytic cleft; these positions immediately suggested that a nucleotide-selection mechanism—i.e., an ATP- and RT-mutation-mediated modulation of the geometric selection of nucleotides—played a role in the DNA polymerization and NRTI resistance of HIV-1. HIV-1. In this study, we showed that the ATP molecule at physiological concentrations acted as an allosteric regulator of HIV-1 RT to modulate nucleotide selectivity. We also showed probable 3-D positions of the bound ATP molecule and NRTI mutations in the catalytic cleft; these positions immediately suggested that a nucleotide-selection mechanism—i.e., an ATP- and RT-mutation-mediated modulation of the geometric selection of nucleotides—played a role in the DNA polymerization and NRTI resistance of HIV-1. First, we demonstrated that the ATP molecule modulated the Km and kcat values of the substrate for HIV-1 RT. We showed that the ATP molecule reduced the Km values of dTTP with both NRTI-sensitive and -resistant RTs (Figure 1C). These results suggested that the ATP molecule can decrease the Km value of a natural substrate to HIV-1 RT. We also showed that the ATP molecule reduced the kcat values of these RTs (Figure 1C). These results suggested that the ATP molecule can decrease the rate of DNA polymerization and thereby increase the probability of an excision reaction by HIV-1 RT. Lineweaver-Burk double- reciprocal plots showed that the ATP molecule is a mixed noncompetitive inhibitor of RT, suggesting distinct binding sites for ATP and dNTP. Taken together, these data strongly suggest that the ATP molecule can act as an allosteric regulator to modulate the nucleotide selectivity of HIV-1 RT. Discussion We found no marked increases in the Ki values of AZTTP and d4TTP at the ATP concentrations around the Ki9 ATP value (1.160.4 mM) for ATP binding to the RT-template-primer-dTTP January 2010 \| Volume 5 \| Issue 1 \| e8867 7 PLoS ONE \| www.plosone.org Regulation of HIV RT by ATP the HIV-1 RT. The open reading frames encoding the RT p51 and p66 subunits of the 93JP-NH1 and ERT-mt6 RTs were amplified by PCR and cloned into the BamH1 site of pQE-9 (Qiagen, Germany). The nucleotide sequences of the PCR- amplified fragments and the sequences around the cloning sites were verified with an automated sequencer. Each subunit was expressed individually in XL1-blue by induction with isopropyl-b- D-thiogalactopyranoside, and the cells expressing p51 and p66 were mixed in binding buffer (20 mM sodium phosphate, 500 mM NaCl, 10 mM imidazole, and EDTA-free protease inhibitor mixture (Roche, Germany), lysed with a French press, centrifuged at 10,000 g for 20 min, and filtered (0.45-mm pore size). The p51/p66 heterodimers were purified from the filtered lysates by Ni2+ affinity chromatography (HiTrap Chelating HP; Amersham Biosciences, UK) and size exclusion chromatography (HiLoad 16/60 Superdex 200 pg; Amersham Biosciences, UK). All of the purification processes were carried out at 4uC. About 1.5 (ERT-mt6) and 3 mg (93JP-NH1) of the p51/p66 heterodimers with greater than 95% purity as judged by SDS-polyacrylamide gel electrophoresis (Figure S1) were obtained with a 1-liter culture. The specific activities of the purified RTs were 40,000 and 10,000 units/mg of protein for 93JP-NH1 and ERT-mt6, respectively, wherein one unit is defined as the amount of enzyme required for incorporation of 1.0 nmol of 32P-dTTP into poly(rA)/ poly(dT)12-18 in 10 min at 37uC. complex for the excision reaction (Figures 2C and 2D). The ATP molecule was estimated to bind with equivalent Ki values to the NRTI-sensitive and -resistant RTs, as others have indicated [15,16], suggesting that NRTI-resistance mutations do not necessarily increase ATP-binding affinity. Moreover, some NRTI-resistant mutations, such as M41L and T69I, are located relatively distantly from the ATP-binding site and catalytic site (Figure 4), which makes their direct impact on DNA excision unlikely. Thus, our study suggests that an ATP-mediated DNA excision mechanism alone is insufficient to explain the roles of the NRTI-resistance mutations, as was also noted previously [12]. Therefore, we speculate that NRTI-resistance mutations can decrease the affinity of NRTIs to HIV-1 RT in concert with the ATP molecule. Measurement of RT Activity Measurement of RT Activity The purified RTs were dissolved in the RT stock buffer (50 mM Tris-HCl pH 7.5, 75 mM KCl, 5 mM MgCl2, 2 mM DTT, 0.05% NP40, and 50% glycerol)[32,33] and kept at 230uC until use. RNA-dependent DNA polymerase activity was measured using [a-32P]dTTP and poly(rA)/poly(dT)12-18 as described previously [33]. For the RT reaction in the presence of ATP, RT activities were measured in 100 ml of RT reaction cocktail consisting of 50 mM Tris-HCl pH 7.5, 75 mM KCl, 5 mM MgCl2, 2 mM DTT, 0.05% NP40, and 50% glycerol containing RT (1–10 nM), dTTP (0.2–18 mM), and ATP (0–4 mM). For the RT reaction in the presence of ATP and NRTI, RT activities were measured in 100 ml of the RT reaction cocktail containing RT (1– 10 nM), dTTP (0.2–18 mM), ATP (0–5 mM), and AZTTP or d4TTP (0–1 mM). These experiments were performed in duplicate and repeated two to six times. In conclusion, we demonstrated that the ATP molecule at physiological concentrations acts as an allosteric regulator of HIV-1 RT to decrease the Km value of the substrate, decrease the kca valuet, and increase the Ki value of NRTIs for RT. The effects were independent of NRTI-resistance mutations of RT. The ATP molecule and NRTI mutations could decrease RT’s sensitivity to NRTI of RT in concert with the RT mutation. Our data support the notion that the ATP molecule and NRTI mutations can modulate nucleotide selectivity by altering the fidelity of the geometric selection of nucleotides and the probability of an excision reaction. Materials and Methods ð1Þ Steady-State Kinetic Analysis The averages of the experimental data were fit by a nonlinear regression method using the program Igor Pro (WaveMetrics, USA). The kinetics parameters were determined by the Michaelis- Menten equation: v~ Vmaxapp½S Kmappz½S , ð1Þ Nucleotides where [S] is the substrate concentration; Km app is the apparent Michaelis-Menten constant; and Vmax app is the apparent maximal rate attained when the enzyme active sites are saturated by substrate. Poly(rA)Np(dT)12-18, dNTPs (100 mM, pH 7.5), NTPs (100 mM, pH 7.5), and [a-32P]dTTP were purchased from Pharmacia Biotech Inc. (USA). ADP and AMP were from ICN (USA). Adenosine 59-(b, c-imido) triphosphate (AMP-PNP) was from Sigma Chemical (USA). 39-Azido 39-deoxythymidine 59-triphos- phate (AZTTP), 39-deoxy-29, and 39-didehydrothymidine 59- triphosphate (d4TTP) were from Moravek Biochemicals (USA). Based on the kinetics data in Figure 1 and Figure S2, the previously reported kinetics data [17,18], and a crystal structure study of the ATP-RT complex [11], we assumed mixed noncompetitive inhibition of ATP and competitive inhibition of NRTIs (Figure 2C). Discussion Our data imply that the ATP molecule and NRTI mutations can modulate the nucleotide selectivity of HIV-1 RT by influencing the geometric selection of nucleotides in the catalytic cavity, as follows: First, the presence of bound ATP molecule in the catalytic cavity can sterically influence the initial binding and translocation of dNTP/NRTI into the catalytic site. Second, more hydrophobic side chains of the NRTI-resistance mutations can improve p66’s ability to exclude water from the catalytic cavity and allow more intimate interactions between nucleotides, the primer-template, and amino acids around the catalytic site for distinguishing between correct and incorrect base pairs. In this regard, previous crystal structure analyses have revealed that the active site of a low-fidelity polymerase is more accessible to the solvent than those of more accurate polymerases [30,31]. Third, Ki NRTIs values increased sharply with NRTI-resistant RT (ERT- mt6 RT) at ATP concentrations around the Ki ATP value (2.861.3 mM) for ATP binding to the RT-template-primer complex (Figures 2C and 2D). Fourth, the changes in the Km value of dTTP correlated with those in the Ki value of the ATP molecule to the RT-template-primer complex (Figures 5A and 5C). Taken together, our structural, kinetic and mutagenesis data suggest that the NRTI-resistance mutations and the ATP molecule can cooperatively modulate physicochemical properties of the p66 catalytic cavity to alter the fidelity of the geometric selection of nucleotides and the probability of an excision reaction. PLoS ONE \| www.plosone.org Structural Analysis We constructed the 3-D models of HIV-1 RTs by homology modeling [19] using the Molecular Operating Environment, MOE (Chemical Computing Group, Canada) as previously described [34]. We generated models of the 93JPNH-1 RT and ERT-mt6 RT structures at the pre- and post-translocation stages, which theoretically are competent for the binding of the incoming-ATP. We used two crystal structures of the HIV-1 RTs (PDB code: 1N6Q [14] and 1RTD [1]) as modeling templates. The sequence identities of the 1N6Q and 1RTD with the 93JPNH-1 RT and ERT-mt6 RT are ,90%. We optimized the 3-D structure thermodynamically by energy minimization using MOE and an AMBER94 force field. We further refined the physically unacceptable local structure of the models on the basis of evaluation by the Ramachandran plot using MOE. The optimized models were docked with ATP with the automated ligand docking program ASEDock2005 [22] (Ryoka Systems, Japan) operated in the Molecular Operating Environment. The RT-template-primer-ATP complex structures were thermody- namically and sterically optimized as described above. Figure S2 Lineweaver-Burk double-reciprocal plots of AZTTP- dependent inhibition of dTTP incorporation. A. 93JP-NH1 RT. B. ERT-mt6. The initial velocities of dTMP incorporation into poly (rA)?p(dT)12-18 were measured using [a-32P]dTTP and purified RTs in the presence of AZTTP. Reciprocal values of the initial velocities and substrate concentrations are plotted. Found at: doi:10.1371/journal.pone.0008867.s002 (0.15 MB TIF) Figure S3 Docking simulations of ATP with RT-template-primer ternary complex models. A and C: 93JP-NH1 RT. B and D: ERT- mt6 RT. The 3-D models of the p66-template-primer complexes at the pre-translation stage (A and B) and the post-translation stage (C and D) were constructed by a homology modeling technique and docked with ATP using the ASEDock2005 (see Materials and Methods). Catalytic clefts composed of fingers, palm, and thumb subdomains are shown. ATP, red sticks; p66 main chain, grey ribbon; template-primer, grey sticks; motif A, blue ribbon. Found at: doi:10.1371/journal.pone.0008867.s003 (1.93 MB TIF) Supporting Information Km app~ Km(1z ½IATP KATP i z ½IAZT KAZT i z ½IAZT½IATP K 00ATP i KAZT i ) (1z ½IATP K 0ATP i ) , ð6Þ Figure S1 Data on RTs of 93JP-NH1 and ERT-mt6. A. Electrophoresis of the purified p51/p66 heterodimers of HIV-1 RTs. The purified p51/p66 heterodimers of 93JP-NH1 RT (NH1) and ERT-mt6 RT (mt6) were electrophoresed on an SDS-4/20% polyacrylamide gradient gel. The gel was stained with GelCode Blue Stain Reagent (Pierce, USA). (Lanes 1 and 4) Molecular size markers. B. The substrate-velocity curves of purified HIV-1 RTs. RNA-dependent DNA polymerase activity at the indicated concentrations of [a-32P]dTTP was measured using purified RTs of 93JP-NH1 (1 nM) and ERT-mt6 (10 nM). Found at: doi:10.1371/journal.pone.0008867.s001 (0.29 MB TIF) Figure S1 Data on RTs of 93JP-NH1 and ERT-mt6. A. Electrophoresis of the purified p51/p66 heterodimers of HIV-1 RTs. The purified p51/p66 heterodimers of 93JP-NH1 RT (NH1) and ERT-mt6 RT (mt6) were electrophoresed on an SDS-4/20% polyacrylamide gradient gel. The gel was stained with GelCode Blue Stain Reagent (Pierce, USA). (Lanes 1 and 4) Molecular size markers. B. The substrate-velocity curves of purified HIV-1 RTs. RNA-dependent DNA polymerase activity at the indicated concentrations of [a-32P]dTTP was measured using purified RTs of 93JP-NH1 (1 nM) and ERT-mt6 (10 nM). Found at: doi:10.1371/journal.pone.0008867.s001 (0.29 MB TIF) ð6Þ where Ki ATP is the dissociation constant of ATP; Ki AZTTP is the dissociation constant of AZTTP; [IATP] is the ATP concentration; and [IAZTTP] is the AZTTP concentration. Expression and Purification of HIV-1 RT The kcat app and Km app that can be derived from the model (Figure 2C) are The kcat app and Km app that can be derived from the model (Figure 2C) are kcat app~ kcat (1z ½IATP K 0ATP i ) ð5Þ ð5Þ and and Expression and Purification of HIV-1 RT HIV-1 infectious molecular clones, 93JP-NH1 and ERT-mt6 [13], were used to clone and express the p51 and p66 subunits of Using this model, the enzyme kinetic parameters were calculated using Equations 2–5. PLoS ONE \| www.plosone.org January 2010 \| Volume 5 \| Issue 1 \| e8867 January 2010 \| Volume 5 \| Issue 1 \| e8867 8 Regulation of HIV RT by ATP Vmax app and Km app are defined as Vmax app and Km app are defined as Vmax app and Km app are defined as pQE70 (Qiagen, Germany) containing the coding sequence of the p66 subunit of ERT-mt6 as the template. The positions of the amino acid substitutions corresponded to the positions 72, 110, 113, 116, and 219 of 93JP-NH1. The mutations and oligonucleotides used in the mutagenesis reaction were R72A (59-CGGCCAGCA TTAAATGGgcGAAATTAGTAGATTTCAGAGAG-39), R72Q (59-CGGCCAGCATTAAATGGcaGAAATTAGTAGATTTCA- GAGAG-39), D110A (59-GAAAAAATCAGTAACAGTACTAG- cTGTGGGAGATGCATATTTTTC-39), D110N (59-GAAAAA- ATCAGTAACAGTACTAaATGTGGGAGATGCATATTTT- TC-39), D113A (59-CAGTACTAGATGTGGGAGcTGCATAT- TTTTCAGTTCCTT-39), D113N (59-CAGTACTAGATGTGG- GAaaTGCATATTTTTCAGTTCCTT-39), F116A (59-GGAA- CTGAAgcATATGCATCTCCCACATCTAGTACTG-39), F116L (59-GGAACTGAcAAATATGCATCTCCCACATCTAGTACT- G-39), K219A (59-GGGATTTTATACACCAGACgcAAAGCAT- CAGAAGGAACCTC-39), and K230(219)Q (59-GGGATTTTA- TACACCAGACcAAAAGCATCAGAAGGAACCTC-39), where the introduced mutations appear in lowercase letters. In all cases, the nucleotide sequences of the complete p66 coding region and of cloning sites were verified with an automated sequencer. The mutant p66 subunits were expressed in XL1-blue and used to form the p51/p66 heterodimer using the p51 subunit of 93JP-NH1 in binding buffer, as described above. The p51/p66 heterodimers were purified by Ni2+ affinity chromatography. About 104 to 221 mg of the p51/p66 heterodimers, with about 90% purity as judged by SDS-polyacrylamide gel electrophoresis (Figure S4A), were obtained from a 20 ml culture. The purified RTs were dissolved in the RT stock buffer and kept at 230uC until use. Vmax app~kcat app½E ð2Þ ð2Þ and and Km app~Km (1z ½I Ki ) (1z ½I Ki0 ) , ð3Þ ð3Þ respectively, where Km is the Michaelis-Menten constant; [E] is the enzyme concentration; [I] is the inhibitor concentration; Ki and Ki9 are the inhibition constants for the enzyme and the complex of the enzyme with substrate; and kcat app is the apparent turnover number: kcat app~ kcat (1z ½I Ki0 ) , ð4Þ ð4Þ where kcat is the turnover number. The kcat app and Km app that can be derived from the model (Figure 2C) are where kcat is the turnover number. References q J 24. Li Y, Kong Y, Korolev S, Waksman G (1998) Crystal structures of the Klenow fragment of Thermus aquaticus DNA polymerase I complexed with deoxyr- ibonucleoside triphosphates. Protein Sci 7: 1116–1123. 10. Meyer PR, Matsuura SE, Mian AM, So AG, Scott WA (1999) A mechanism of AZT resistance: an increase in nucleotide-dependent primer unblocking by mutant HIV-1 reverse transcriptase. Mol Cell 4: 35–43. p p 25. Johnson SJ, Taylor JS, Beese LS (2003) Processive DNA synthesis observed in a polymerase crystal suggests a mechanism for the prevention of frameshift mutations. Proc Natl Acad Sci U S A 100: 3895–3900. 11. Das K, Sarafianos SG, Clark AD Jr, Boyer PL, Hughes SH, et al. (2007) Crystal structures of clinically relevant Lys103Asn/Tyr181Cys double mutant HIV-1 reverse transcriptase in complexes with ATP and non-nucleoside inhibitor HBY 097. J Mol Biol 365: 77–89. 26. Yin YW, Steitz TA (2004) The structural mechanism of translocation and helicase activity in T7 RNA polymerase. Cell 116: 393–404. 27. Temiakov D, Patlan V, Anikin M, McAllister WT, Yokoyama S, et al. (2004) Structural basis for substrate selection by t7 RNA polymerase. Cell 116: 381–391. 12. Sarafianos SG, Marchand B, Das K, Himmel DM, Parniak MA, et al. (2009) Structure and function of HIV-1 reverse transcriptase: molecular mechanisms of polymerization and inhibition. J Mol Biol 385: 693–713. 28. Thompson AA, Albertini RA, Peersen OB (2007) Stabilization of poliovirus polymerase by NTP binding and fingers-thumb interactions. J Mol Biol 366: 1459–1474. 13. Sato H, Tomita Y, Ebisawa K, Hachiya A, Shibamura K, et al. (2001) Augmentation of human immunodeficiency virus type 1 subtype E (CRF01_AE) multiple-drug resistance by insertion of a foreign 11-amino-acid fragment into the reverse transcriptase. J Virol 75: 5604–5613. 29. Petruska J, Goodman MF, Boosalis MS, Sowers LC, Cheong C, et al. (1988) Comparison between DNA melting thermodynamics and DNA polymerase fidelity. Proc Natl Acad Sci U S A 85: 6252–6256. 14. Sarafianos SG, Clark AD Jr, Das K, Tuske S, Birktoft JJ, et al. (2002) Structures of HIV-1 reverse transcriptase with pre- and post-translocation AZTMP- terminated DNA. Embo J 21: 6614–6624. y 30. Ling H, Boudsocq F, Woodgate R, Yang W (2001) Crystal structure of a Y- family DNA polymerase in action: a mechanism for error-prone and lesion- bypass replication. Cell 107: 91–102. 15. Ray AS, Murakami E, Basavapathruni A, Vaccaro JA, Ulrich D, et al. References 59-triphosphate with human immunodeficiency virus reverse transcriptase. Proc Natl Acad Sci U S A 83: 8333–8337. 59-triphosphate with human immunodeficiency virus reverse transcriptase. Proc Natl Acad Sci U S A 83: 8333–8337. 1. Huang H, Chopra R, Verdine GL, Harrison SC (1998) Structure of a covalently trapped catalytic complex of HIV-1 reverse transcriptase: implications for drug resistance. Science 282: 1669–1675. 18. Vrang L, Bazin H, Remaud G, Chattopadhyaya J, Oberg B (1987) Inhibition of the reverse transcriptase from HIV by 39-azido-39-deoxythymidine triphosphate and its threo analogue. Antiviral Res 7: 139–149. 2. Preston BD, Poiesz BJ, Loeb LA (1988) Fidelity of HIV-1 reverse transcriptase. Science 242: 1168–1171. g 19. Baker D, Sali A (2001) Protein structure prediction and structural genomics. Science 294: 93–96. 3. Roberts JD, Bebenek K, Kunkel TA (1988) The accuracy of reverse transcriptase from HIV-1. Science 242: 1171–1173. 20. Kati WM, Johnson KA, Jerva LF, Anderson KS (1992) Mechanism and fidelity of HIV reverse transcriptase. J Biol Chem 267: 25988–25997. 4. Goodman MF (1997) Hydrogen bonding revisited: geometric selection as a principal determinant of DNA replication fidelity. Proc Natl Acad Sci U S A 94: 10493–10495. 21. Rittinger K, Divita G, Goody RS (1995) Human immunodeficiency virus reverse transcriptase substrate-induced conformational changes and the mechanism of inhibition by nonnucleoside inhibitors. Proc Natl Acad Sci U S A 92: 8046–8049. 5. Kunkel TA, Bebenek K (2000) DNA replication fidelity. Annu Rev Biochem 69: 497–529. 6. Kunkel TA (2004) DNA replication fidelity. J Biol Chem 279: 16895–16898. 22. Goto J, Kataoka R, Muta H, Hirayama N (2008) ASEDock-docking based on alpha spheres and excluded volumes. J Chem Inf Model 48: 583–590. 7. Traut TW (1994) Physiological concentrations of purines and pyrimidines. Mol Cell Biochem 140: 1–22. 23. Chen R, Yokoyama M, Sato H, Reilly C, Mansky LM (2005) Human immunodeficiency virus mutagenesis during antiviral therapy: impact of drug- resistant reverse transcriptase and nucleoside and nonnucleoside reverse transcriptase inhibitors on human immunodeficiency virus type 1 mutation frequencies. J Virol 79: 12045–12057. 8. Meyer PR, Matsuura SE, So AG, Scott WA (1998) Unblocking of chain- terminated primer by HIV-1 reverse transcriptase through a nucleotide- dependent mechanism. Proc Natl Acad Sci U S A 95: 13471–13476. 9. Arion D, Kaushik N, McCormick S, Borkow G, Parniak MA (1998) Phenotypic mechanism of HIV-1 resistance to 39-azido-39-deoxythymidine (AZT): increased polymerization processivity and enhanced sensitivity to pyrophosphate of the mutant viral reverse transcriptase. Biochemistry 37: 15908–15917. Site-Directed Mutagenesis Figure S4 Data on RT mutants from the ERT-mt6 RT. A. Electrophoresis of the purified RT mutants from the ERT-mt6 RT. B. dTMP incorporations into poly (rA)?p(dT)12-18 by the Site-directed mutagenesis was performed with a QuikChange Multi Site-Directed Mutagenesis Kit (Stratagene, USA), using PLoS ONE \| www.plosone.org January 2010 \| Volume 5 \| Issue 1 \| e8867 9 Regulation of HIV RT by ATP RTs was measured using various concentrations of [a-32P]dTTP and poly (rA)?p(dT)12-18 in the presence of ATP. Representative results with D113A RT (left) and K219A RT (right) are shown. Found at: doi:10.1371/journal.pone.0008867.s004 (0.39 MB TIF) mutant RTs. RNA-dependent DNA polymerase activity of the purified RTs (20 nM) was measured using a [a-32P]dTTP and poly (rA)?p(dT)12-18 system. C. Fold increases in the IC50 of AZTTP by ATP addition. IC50 values of AZTTP with RT mutants were calculated from the amounts of [a-32P]dTTP incorporation in the presence of various concentrations (0– 1 mM) of AZTTP and 5 mM ATP. Fold increases in IC50 compared to the values without ATP are shown. D. The substrate- velocity curves of purified HIV-1 RTs in the presence of ATP. RNA-dependent DNA polymerase activity of the purified mutant mutant RTs. RNA-dependent DNA polymerase activity of the purified RTs (20 nM) was measured using a [a-32P]dTTP and poly (rA)?p(dT)12-18 system. C. Fold increases in the IC50 of AZTTP by ATP addition. IC50 values of AZTTP with RT mutants were calculated from the amounts of [a-32P]dTTP incorporation in the presence of various concentrations (0– 1 mM) of AZTTP and 5 mM ATP. Fold increases in IC50 compared to the values without ATP are shown. D. The substrate- velocity curves of purified HIV-1 RTs in the presence of ATP. RNA-dependent DNA polymerase activity of the purified mutant mutant RTs. RNA-dependent DNA polymerase activity of the purified RTs (20 nM) was measured using a [a-32P]dTTP and poly (rA)?p(dT)12-18 system. C. Fold increases in the IC50 of AZTTP by ATP addition. IC50 values of AZTTP with RT mutants were calculated from the amounts of [a-32P]dTTP incorporation in the presence of various concentrations (0– 1 mM) of AZTTP and 5 mM ATP. Fold increases in IC50 compared to the values without ATP are shown. D. The substrate- velocity curves of purified HIV-1 RTs in the presence of ATP. RNA-dependent DNA polymerase activity of the purified mutant Author Contributions Conceived and designed the experiments: MY HS. Performed the experiments: MY HM. Analyzed the data: MY. Wrote the paper: MY HS. Conceived and designed the experiments: MY HS. Performed the experiments: MY HM. Analyzed the data: MY. Wrote the paper: MY HS. References (2003) Probing the molecular mechanisms of AZT drug resistance mediated by HIV-1 reverse transcriptase using a transient kinetic analysis. Biochemistry 42: 8831–8841. yp p 31. Ling H, Boudsocq F, Plosky BS, Woodgate R, Yang W (2003) Replication of a cis-syn thymine dimer at atomic resolution. Nature 424: 1083–1087. 32. Starnes MC, Cheng YC (1989) Human immunodeficiency virus reverse transcriptase-associated RNase H activity. J Biol Chem 264: 7073–7077. transcriptase-associated RNase H activity. J Biol Chem 264: 7073– 16. Marchand B, White KL, Ly JK, Margot NA, Wang R, et al. (2007) Effects of the translocation status of human immunodeficiency virus type 1 reverse transcriptase on the efficiency of excision of tenofovir. Antimicrob Agents Chemother 51: 2911–2919. 33. Willey RL, Smith DH, Lasky LA, Theodore TS, Earl PL, et al. (1988) In vitro mutagenesis identifies a region within the envelope gene of the human immunodeficiency virus that is critical for infectivity. J Virol 62: 139–147. 34. Shirakawa K, Takaori-Kondo A, Yokoyama M, Izumi T, Matsui M, et al. (2008) Phosphorylation of APOBEC3G by protein kinase A regulates its interaction with HIV-1 Vif. Nat Struct Mol Biol 15: 1184–1191. 17. Furman PA, Fyfe JA, St Clair MH, Weinhold K, Rideout JL, et al. (1986) Phosphorylation of 39-azido-39-deoxythymidine and selective interaction of the PLoS ONE \| www.plosone.org PLoS ONE \| www.plosone.org January 2010 \| Volume 5 \| Issue 1 \| e8867 10
https://openalex.org/W2903163366	https://www.duo.uio.no/bitstream/10852/65868/1/12913_2018_Article_3729.pdf	English	null	Characteristics, course and outcome of patients receiving physiotherapy in primary health care in Norway: design of a longitudinal observational project	BMC health services research	2,018	cc-by	8,190	Characteristics, course and outcome of patients receiving physiotherapy in primar health care in Norway: design of a longitudinal observational project Kari Anne I. Evensen1,2* , Hilde Stendal Robinson3, Ingebrigt Meisingset1,4, Astrid Woodhouse1,5,6, Marit Thielemann3, Wenche S. Bjorbækmo3,7, Gard Myhre4, Anne E. Hansen4, Ottar Vasseljen1 and Nina K. Vøllestad3 Evensen et al. BMC Health Services Research (2018) 18:936 https://doi.org/10.1186/s12913-018-3729-y Evensen et al. BMC Health Services Research (2018) 18:936 https://doi.org/10.1186/s12913-018-3729-y STUDY PROTOCOL Open Access Characteristics, course and outcome of patients receiving physiotherapy in primary health care in Norway: design of a longitudinal observational project Kari Anne I. Evensen1,2* , Hilde Stendal Robinson3, Ingebrigt Meisingset1,4, Astrid Woodhouse1,5,6, Marit Thielemann3, Wenche S. Bjorbækmo3,7, Gard Myhre4, Anne E. Hansen4, Ottar Vasseljen1 and Nina K. Vøllestad3 Open Access * Correspondence: karianne.i.evensen@ntnu.no 1Department of Public Health and Nursing, Norwegian University of Science and Technology (NTNU), Trondheim, Norway 2Department of Clinical and Molecular Medicine, NTNU, Trondheim, Norway Full list of author information is available at the end of the article * Correspondence: karianne.i.evensen@ntnu.no * Correspondence: karianne.i.evensen@ntnu.no 1Department of Public Health and Nursing, Norwegian University of Science and Technology (NTNU), Trondheim, Norway 2Department of Clinical and Molecular Medicine, NTNU, Trondheim, Norway Full list of author information is available at the end of the article Abstract Background: Physiotherapists (PTs) in primary health care manage patients with large variation in medical diagnosis, age, functional status, disability and prognosis. Lack of knowledge and systematically collected data from patients treated by PTs in primary health care has prompted this longitudinal observational physiotherapy project. This paper aims to describe a method for developing a database of patients managed by PTs in primary health care, with the main purpose to study patients’ characteristics, treatment courses and prognostic factors for favourable outcome. Methods: This is a longitudinal observational project, following patients through their physiotherapy treatment periods in primary health care in Norway and until one year after inclusion. The project involves both private practitioners and municipally employed PTs working in primary health care in nine municipalities in Norway. The patients are recruited to three different cohorts depending on age and whether they are referred to a private practitioner or a municipally employed PT. All data are recorded electronically, transferred and stored securely. For all patients we have included extensive questionnaires to obtain information about demographics, disability and function, pain-related variables, psychosocial factors, treatments and evaluation of treatment as well as response to clinical tests. The PTs have access to use their own patients’ data. We have also prepared for linkage to national patient registers and data collected in population-based studies to be able to gather further important data. Discussion: This project will have important implications for physiotherapy services in primary health care. The database contains more than 3000 patients, and data collection is ongoing. Data collected so far suggest that the patients included are representative of the larger population of patients treated by private practitioners or municipally employed PTs in Norway. This large scale prospective physiotherapy project will provide knowledge about the patient groups, applied treatments and short- and long-term outcome of the patients. Trial registration: ClinicalTrials.gov Identifier: NCT03626389. Registered on August 13th 2018 (retrospectively registered). Keywords: Physiotherapy, Primary health care services, Electronic registration, Database, Cohort, FYSIOPRIM, Municipality Trial registration: ClinicalTrials.gov Identifier: NCT03626389. Registered on August 13th 2018 (retrospectively registered). Trial registration: ClinicalTrials.gov Identifier: NCT03626389. Registered on August 13th 2018 (retrospectively registered). Keywords: Physiotherapy, Primary health care services, Electronic registration, Database, Cohort, FYSIOPRIM, Municipality Keywords: Physiotherapy, Primary health care services, Electronic registration, Database, Cohort, FYSIOPRIM, Municipality © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted 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. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Methods/design Aims The primary aim of the project is to build a database by systematically collecting data from baseline throughout the treatment period and beyond, including PROMs, the patients’ and PTs’ main goals and plans for treatment, assessment of goal achievement and outcome. This will enable detailed descriptions of patients receiving physio- therapy services, goal-setting, type of treatments, and how general health, physical function and relevant clin- ical factors change throughout and after the treatment period. Secondly, we want to study associations and interaction effects between clinical characteristics, treatments and outcome, along with health-economic evaluations. We will also examine how physiotherapy practice is affected by being exposed to systematic registration of clinical data. Moreover, we will be able to compare patients receiving physiotherapy services with information from national patient registries. In addition, physical fitness, balance and walking ability will be examined in a group of healthy older adults to study as- sociations between these aspects of physical functioning and health-related quality of life, self-reported physical and psychological functioning as well as response to physical fitness tests. We will compare data from this group with data from patients with musculoskeletal complaints who receive physiotherapy treatment, and with similar data from large population-based studies in Norway. Finally, this project enables methodological studies to validate clinical tests and questionnaires. p p p y In Norway, one large and some smaller local studies of patients treated by PTs in primary health care have been published [2–4]. All are cross-sectional studies and only a limited number of factors are studied. In the largest study, the PTs answered questions about sex, age, diag- nosis, treatment modalities and main goals for treatment for a total of 3196 patients [4]. Patient-reported outcome measures (PROMs) were not used, and none of the studies were designed to describe clinical courses or prognostic factors. An international study used registry data from three different countries (USA, Netherlands and Israel) [1]. Sex, age, diagnoses, affected body regions, duration of complaints as well as number and type of treatment (modality) were registered. The study did, however, not include PROMs as all data were recorded by the PTs. In a study of 7670 patients with musculoskeletal complaints seeking physiotherapy treatment in New Zealand, sex, age, affected body region, patient-reported pain intensity and disability as well as number of treatments were recorded [5]. The study showed clear reduction in pain intensity and disability after treatment. © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted 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. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Page 2 of 10 Page 2 of 10 Evensen et al. BMC Health Services Research (2018) 18:936 Methods/design Aims No other outcome measures were recorded, nor was there any information on effects and course of symptoms after treatment. Follow-up data after treatment and a wider spectrum of baseline prognos- tic factors and outcome measures are needed for com- parative exploration of symptom course and treatment outcome in the different patient groups receiving primary care physiotherapy. Background together a set of standardised methods and tools to enable systematic studies of clinical courses for patients treated by PTs in primary health care in Norway. This paper describes the framework, design and methods used for the systematic data collection, where the purpose is to establish a database for studying patient characteristics, prognostic factors, applied treatments and outcome in various patient groups. The patients include children and adults of all ages receiving primary care physiotherapy services from private practitioners and municipally employed PTs. This large scale pro- spective physiotherapy project will provide knowledge about the patient groups, applied treatments, and short- and long-term outcome of the patients. Background There is a lack of comprehensive and systematically collected data about patients receiving primary care physiotherapy. Physiotherapists (PTs) in primary health care manage patients with large variation in medical diagnosis, age, functional status, disability and prognosis [1]. Among PTs in private practice in Norway, informa- tion about patients and treatments is limited to diagno- ses, number of treatments and costs, which the PTs are obliged to report to the Norwegian Directorate of Health in order to receive reimbursement. Apart from this, little is known about the complaints for which patients seek or receive physiotherapy services. Comparative know- ledge of the different patient groups, applied treatments, clinical courses and outcome measures is missing. Hence, there is a need for robust and comprehensive data of how and to whom the primary care physiother- apy services are delivered, and whether the treatment goals are achieved. This includes systematically collected information about prognostic factors, content and effect of commonly applied treatments. This knowledge will aid health care managerial decision making and policy makers in prioritising among health care services, and to improve rehabilitation of patients in primary health care. In Norway, one large and some smaller local studies of patients treated by PTs in primary health care have been published [2–4]. All are cross-sectional studies and only a limited number of factors are studied. In the largest study, the PTs answered questions about sex, age, diag- nosis, treatment modalities and main goals for treatment for a total of 3196 patients [4]. Patient-reported outcome measures (PROMs) were not used, and none of the studies were designed to describe clinical courses or prognostic factors. An international study used registry data from three different countries (USA, Netherlands and Israel) [1]. Background Sex, age, diagnoses, affected body regions, duration of complaints as well as number and type of treatment (modality) were registered. The study did, however, not include PROMs as all data were recorded by the PTs. In a study of 7670 patients with musculoskeletal complaints seeking physiotherapy treatment in New Zealand, sex, age, affected body region, patient-reported pain intensity and disability as well as number of treatments were recorded [5]. The study showed clear reduction in pain intensity and disability after treatment. No other outcome measures were recorded, nor was there any information on effects and course of symptoms after treatment. Follow-up data after treatment and a wider spectrum of baseline prognos- tic factors and outcome measures are needed for com- parative exploration of symptom course and treatment outcome in the different patient groups receiving primary care physiotherapy. Through the Research Program for Physiotherapy in Primary Health Care, the FYSIOPRIM, we have put g There is a lack of comprehensive and systematically collected data about patients receiving primary care physiotherapy. Physiotherapists (PTs) in primary health care manage patients with large variation in medical diagnosis, age, functional status, disability and prognosis [1]. Among PTs in private practice in Norway, informa- tion about patients and treatments is limited to diagno- ses, number of treatments and costs, which the PTs are obliged to report to the Norwegian Directorate of Health in order to receive reimbursement. Apart from this, little is known about the complaints for which patients seek or receive physiotherapy services. Comparative know- ledge of the different patient groups, applied treatments, clinical courses and outcome measures is missing. Hence, there is a need for robust and comprehensive data of how and to whom the primary care physiother- apy services are delivered, and whether the treatment goals are achieved. This includes systematically collected information about prognostic factors, content and effect of commonly applied treatments. This knowledge will aid health care managerial decision making and policy makers in prioritising among health care services, and to improve rehabilitation of patients in primary health care. Design and setting This is a longitudinal observational project following patients through physiotherapy treatment periods in primary health care in Norway. Data are collected prospectively from the first consultation and until 1 year after baseline. The project started its data collection in June 2015, and will continue through June 2020. The pro- ject involves both private practitioners and municipally Through the Research Program for Physiotherapy in Primary Health Care, the FYSIOPRIM, we have put Evensen et al. BMC Health Services Research (2018) 18:936 Page 3 of 10 Page 3 of 10 collected from 655 adult and older patients receiving physiotherapy services from approximately 55 PTs working in the municipality of Trondheim. The patients may be referred from general practitioners, occupational therapists, health professionals at health care centers, re- habilitation centers or hospitals, or by proxy. The physiotherapy services include “early intervention”, “re- habilitation of activities of daily living (ADL rehabilitation)” or “reablement”, and “regular physiother- apy”, all of which are typically delivered in a home set- ting. Early intervention is a collaboration between health and welfare centers and units for physiotherapy and oc- cupational services. It is offered to patients in need of limited services who are referred to municipality health care services for the first time. ADL rehabilitation/rea- blement is offered to patients receiving home based ser- vices and who are at risk of functional deterioration, while regular physiotherapy is offered to patients in need of more specific physiotherapy services. employed PTs working in primary health care in nine mu- nicipalities of Norway; Oslo, Drammen, Ski, Kongsberg, Stavanger, Trondheim, Orkanger, Rindal, and Alta. All five health regions of Norway are represented. The Norwegian health care system is publicly funded. Most PTs working as private practitioners have a legal agreement to practice within the municipality. They are partly paid by the municipality (as a fixed financial support for practice), partly by a fee for service reimbursement by the Norwegian Health Economics Administration (HELFO), and the patient’s deductible fee (maximum around 2000 NOK per patient per calendar year). Patients seeking physiotherapy services in private practice will normally meet at the PT’s clinic. Municipally employed PTs are on fixed salary with no costs to the patients. They work in an out-patient set- ting, and the patients receive physiotherapy services in their usual daily environment, i.e. for children at home, in kindergarten or at school, and for older patients often in their own home. Design and setting There are no clear guidelines for de- termining which patients should receive physiotherapy from a private practitioner or a municipally employed PT. The choice can be based on previous personal expe- riences with physiotherapy, the possibility to visit a clinic, evaluation of needs and/or benefits of treating the patients in their own setting. Cohort 3: Children receiving physiotherapy services in Trondheim Municipality This part of the project collected data from May 2016 through April 2017. During this period, data were collected from 162 children aged 0–18 years receiving physiotherapy services from approximately 25 PTs work- ing in the municipality of Trondheim. For children, physiotherapy may be initiated on referral from parents, personnel in kindergarten or school, general practi- tioners, occupational therapists, and health professionals at primary health care centers or hospitals. The children typically receive physiotherapy services at primary health care centers, in their own home, in kindergarten, or at school, depending on the child’s usual primary location and condition. Patients in Trondheim are recruited to three different cohorts depending on age and whether they are referred to a private practitioner or to a municipally employed PT. In the rest of the country, patients are recruited from private practitioners only. Description of the cohorts The three cohorts in the project are: Cohort 1: Adult patients seeking physiotherapy services in private practice Data collection This part of the project started its data collection in June 2015, and data collection is planned to continue through June 2020. From June 2015 through December 2017, data were collected from a total of 2754 patients above the age of 18 years seeking physiotherapy treat- ment from 111 PTs in nine municipalities in different parts of Norway, including 536 patients already in treat- ment. Among the 111 PTs, 78 (70%) were general PTs, 20 (18%) were manual therapists and 13 (12%) were PTs with special education in psychomotor physio- therapy. Seventy-seven (69%) of the PTs were working in Trondheim and recruited 2302 (84%) of the patients. A flowchart of the data collection is presented in Fig. 1. Patients are asked to participate on their first encounter with a PT in primary health care. Project information and consent forms are available in Norwegian and English. We collect baseline data in two steps. First, data are registered by the PT and the patient in collaboration. They jointly agree on the main treatment goal and plan for treatment. The PT asks the patient to define and score their most important specific functional problems using the Patient-Specific Functional Scale (PSFS) [6]. The PT registers the patients’ referral and diagnosis, and determines whether the patient should fill in any disease- or region-specific questionnaires. Secondly, the patient completes questionnaires either by using an e-tablet or through a web-link sent by e-mail. Cohort 2: Adult and older patients receiving physio- therapy services in Trondheim Municipality This part of the project collected data from the beginning of May 2016 to the end of May 2018. From January 2016 till the end of September 2017, data were Approximately 1 month after baseline (for patients seeking physiotherapy services from private practitioners), or 2–3 months after baseline (for patients receiving Evensen et al. BMC Health Services Research (2018) 18:936 Page 4 of 10 Fig. 1 Flowchart of data collection and data users in FYSIOPRIM FYSIOPRIM Research Program for Physiotherapy in Primary Health Care, TSD Services for Sensitive Data, PT Physiotherapist ig. Data collection 1 Flowchart of data collection and data users in FYSIOPRIM YSIOPRIM Research Program for Physiotherapy in Primary Health Care, TSD Services for Sensitive Data, PT Physiotherap nd data users in FYSIOPRIM siotherapy in Primary Health Care, TSD Services for Sensitive Data, PT Physiotherapist physiotherapy services from municipally employed PTs), the PT and the patient together evaluate goal achievement and fulfillment of the treatment plan so far. They adjust goals and plans if needed, and the patient recompletes the PSFS. At end of treatment, or maximum 3 months (in pri- vate practice) or 6 months after baseline (in the munici- pality), the PT and the patient again evaluate goal achievement and fulfillment of the treatment plan. The patient then completes the same standardised question- naires as at baseline. The same questionnaires are com- pleted 6 and 12 months after baseline by patients in private practice using a web link sent by sms or e-mail. All patients treated by PTs in private practice receive re- minders once a week, up to three times, by sms and e-mail, if they have not answered the questionnaires. The journal number of the patient is used as the study identifier to enable data flow between the electronic med- ical journal and the secure Infopad server. The data from all patients are copied from the Infopad server to a secure server for research data at the University of Oslo (Services for Sensitive Data) (Fig. 1). Data management is done ac- cording to the quality assurance system of the University of Oslo. We have prepared for linkage between data from FYSIOPRIM and data from national patient registers. This can be done by use of the patient’s journal number to access their unique 11-digit Norwegian personal number. The project is approved by the Regional committee for Medical and Health Research Ethics in Norway (REC no. 2013/2030). All data are collected electronically using an application run on a tablet or through a web-link. The software is provided by Infopad AS (www.infopad.no). Immediately after the completion of the questionnaires all data are transferred to a secure server with in-memory encryption. All PTs have access to their own patients’ data through Infopad’s web site. Depending on journal system, the PTs can import the patients’ data into their electronic medical journal. Most private practitioners have this possibility. Variables An overview of all recorded variables is presented in (Additional file 1: Table S1). For all patients we have in- cluded extensive questionnaires to obtain information about demographics, disability and function, pain-related variables, psychosocial factors, treatments and evaluation of treatment as well as response to clinical tests. For adult patients, we also obtain information about employment/ work status and psychosocial factors. For subgroups of Evensen et al. BMC Health Services Research (2018) 18:936 Page 5 of 10 Page 5 of 10 be used for analyses a range of automated procedures for data preparation and quality checking are performed using standardised written scripts in STATA. Examples are to check i) patterns of missing data, ii) values outside the possible range of a variable (e.g., age > 120 years), iii) conflicting answers (e.g., pregnancy and male), and iv) unexpected frequency distributions for categorical variables. We will explore each variable graphically using histograms, QQ plots and box plots to evaluate the data distribution. Descriptive statistics will be used to describe the patient populations using parametric or non-paramet- ric statistics according to the data distribution. Statistical analyses of associations between variables, prognosis, clin- ical course and outcome will be detailed in pertinent fu- ture publications from the project. patients in private practice, we additionally included disease- or region-specific questionnaires. For children below the age of 1 year, we included a detailed infant history and examination, and for school-aged children we included a questionnaire quantifying physical activity. A list of treatment options is completed by the PTs at follow-up (Additional file 2). Cohort 1: Adult patients seeking physiotherapy services in private practice For adult and older patients receiving physiotherapy services from municipally employed PTs in Trondheim (Cohort 2), we compared data from 68 patients included in FYSIOPRIM with data from Trondheim Municipality for the 72 patients not included in FYSIOPRIM from the end of September to beginning of November 2017. This enabled us to compare sex, age, living conditions, referral entity and cause of referral. Table 1 shows demographics of adult patients seeking physiotherapy services from private practitioners (Co- hort 1). Of the 2754 patients registered in the total FYSIOPRIM database by 31.12.2017, 69.2% were fe- males. Age distribution was similar in the age categories up to 70 years, and relatively few patients were above the age of 80 years. The sex and age distributions were fairly similar between the 2302 patients registered in FYSIO- PRIM Trondheim and data from HELFO Trondheim encompassing all patients treated by private practitioners with reimbursement privilege in Trondheim in 2016– 2017, although patients were somewhat younger in the former (Table 1). Due to a very low number of children registered in FYSIOPRIM after May 2017, we were not able to use the exact same time period for comparison. Thus, for children receiving physiotherapy services from munici- pally employed PTs in Trondheim (Cohort 3), data from 162 children included in FYSIOPRIM from May 2016 to April 2017 were compared with data from Trondheim Municipality for all 72 children receiving physiotherapy services from end of September to begin- ning of November 2017. The distribution of registered diagnostic ICPC code groups in FYSIOPRIM Trondheim was compar- able to the data from HELFO Trondheim, apart from a larger proportion of the FYSIOPRIM patients had osteoarthritis, and slightly fewer had neck and low back pain (Table 2). Organisation of the project Most adult patients seeking physiotherapy services from private practitioners (Cohort 1) were recruited in Trondheim. We compared sex and age distribution for the 2302 patients registered in FYSIOPRIM Trondheim in the period January 2016 to December 2017 with data from HELFO Trondheim in the period July 2016 to June 2017. Diagnosis by the International Classification of Primary Care (ICPC) codes were available for 1725 of the 2302 patients in Trondheim. Of the 577 patients with missing ICPC code, a large proportion had direct access to a manual therapist and did not have their ICPC code registered. g p j The FYSIOPRIM Research Program has three consor- tium partners; the University of Oslo, the Norwegian University of Science and Technology and Trondheim Municipality. The FYSIOPRIM steering committee con- sists of two members from each consortium partner. The committee submits biannual progress reports to the funding body (the Norwegian Fund for Post-Graduate Training in Physiotherapy). Assessment of representativeness In order to assess the representativeness of the patients in the FYSIOPRIM database, we compared sex, age and diagnosis/cause of referral in Cohort 1 with aggregated data from HELFO, which registers all patients treated by private practitioners with a reimbursement privilege. For Cohorts 2 and 3 we used aggregated data from Trondheim Municipality for comparisons. Data preparation and statistical analyses All the data are collected electronically, thereby avoiding the possibility of mistakes when transferring data from paper to electronic data. Data management is performed using STATA 15 (Stata Corp., College Station, TX, USA). The raw data are stored in a secure server for re- search data (Services for Sensitive Data). Before data can Cohort 2: Adult and older patients receiving physiotherapy services in Trondheim municipality Table 3 shows demographics of adult and older patients receiving physiotherapy services from municipally employed PTs in Trondheim (Cohort 2). Of the 655 adult Discussion In this paper, we describe the design and main features of a longitudinal observational project to build a primary health care physiotherapy database (the FYSIOPRIM database). This large scale prospective physiotherapy project in primary health care in Norway will provide knowledge about the patient groups treated, applied treatments and short- and long-term patient outcome. The design, which involves repeated measurements from baseline and up to 1 year after baseline, enables us to examine the patient’s characteristics, symptoms and Cohort 2: Adult and older patients receiving physiotherapy services in Trondheim municipality Most patients (85.6%) were above the age of 70, and 89.7% were living in their own home. In total, 489 (74.7%) re- ceived regular physiotherapy, 112 (17.1%) early intervention and 54 (8.2%) ADL rehabilitation/reablement. The most frequent causes of referral to physiotherapy were geriatrics/ functional deterioration with or without falls (27.5%), orthopedic (14.7%) and neurologic conditions (11.1%). For this cohort, we compared sex, age, living conditions, refer- ral entity and cause of referral for patients included in FYSIOPRIM in a limited time period with data from Trondheim Municipality for patients not included in FYSIOPRIM during the same time period (last two col- umns in Table 3). We found that sex and age distribu- tion was similar, but fewer patients not included in FYSIOPRIM Trondheim were living at home and more of them were referred because of geriatrics/functional deterioration (Table 3). in FYSIOPRIM Trondheim by the end of April 2017, 43.8% were females and 61.1% were below the age of 2 years (Table 4). The most frequent cause of referral to physiotherapy were assessment/guidance related to motor development (32.1%), asymmetrical movement patterns (29.0%) and orthopedic conditions, such as gait and foot alignment (16.0%). To evaluate the representa- tiveness of our child cohort, we compared these propor- tions with data for all children receiving physiotherapy services in Trondheim Municipality during a limited time period. We found that sex and age distributions as well as causes of referral were highly comparable between the samples (Table 4). in FYSIOPRIM Trondheim by the end of April 2017, 43.8% were females and 61.1% were below the age of 2 years (Table 4). The most frequent cause of referral to physiotherapy were assessment/guidance related to motor development (32.1%), asymmetrical movement patterns (29.0%) and orthopedic conditions, such as gait and foot alignment (16.0%). To evaluate the representa- tiveness of our child cohort, we compared these propor- tions with data for all children receiving physiotherapy services in Trondheim Municipality during a limited time period. We found that sex and age distributions as well as causes of referral were highly comparable between the samples (Table 4). and older patients registered in FYSIOPRIM Trondheim by the end of September 2017, 64.1% were females (Table 3). Most patients (85.6%) were above the age of 70, and 89.7% were living in their own home. In total, 489 (74.7%) re- ceived regular physiotherapy, 112 (17.1%) early intervention and 54 (8.2%) ADL rehabilitation/reablement. Cohort 2: Adult and older patients receiving physiotherapy services in Trondheim municipality The most frequent causes of referral to physiotherapy were geriatrics/ functional deterioration with or without falls (27.5%), orthopedic (14.7%) and neurologic conditions (11.1%). For this cohort, we compared sex, age, living conditions, refer- ral entity and cause of referral for patients included in FYSIOPRIM in a limited time period with data from Trondheim Municipality for patients not included in FYSIOPRIM during the same time period (last two col- umns in Table 3). We found that sex and age distribu- tion was similar, but fewer patients not included in FYSIOPRIM Trondheim were living at home and more of them were referred because of geriatrics/functional deterioration (Table 3). Cohort 2: Adult and older patients receiving physiotherapy services in Trondheim municipality Table 3 shows demographics of adult and older patients receiving physiotherapy services from municipally employed PTs in Trondheim (Cohort 2). Of the 655 adult Evensen et al. BMC Health Services Research (2018) 18:936 Page 6 of 10 Table 1 Demographics of adult patients registered in the FYSIOPRIM database and data from HELFO in Trondheim FYSIOPRIM Total 18.06.15–31.12.17 (n = 2754) FYSIOPRIM Trondheima 01.01.16–31.12.17 (n = 2302) HELFO Trondheimb 01.07.16–30.06.17 (n = 19,460) Variables n (%) n (%) n (%) Sex Female 1851 (69.2) 1605 (69.9) 12,863 (66.1) Male 825 (30.8) 691 (30.1) 6590 (33.9) Missing 78 6 7 Age 18–29 years 466 (16.9) 408 (17.7) 2530 (13.9) 30–39 years 46 (16.9) 382 (16.6) 2385 (13.1) 40–49 years 452 (16.4) 373 (16.2) 2930 (16.0) 50–59 years 482 (17.5) 387 (16.8) 3337 (18.3) 60–69 years 476 (17.3) 385 (16.7) 3316 (18.2) 70–79 years 309 (11.2) 265 (11.5) 2624 (14.4) 80+ years 103 (3.7) 102 (4.4) 1143 (6.3) Missing 0 0 1195 Specialistb General PT 1658 (71.7) 1353 (72.7) 12,783 (65.7) Manual therapist 441 (19.1) 310 (16.6) 5575 (28.7) Psychomotor PT 215 (9.3) 199 (10.7) 1099 (5.6) No reimbursement privilege/Missing 440 440 3 FYSIOPRIM Research Program for Physiotherapy in Primary Health Care, HELFO Norwegian Health Economics Administration, PT Physiotherapist aPatients in FYSIOPRIM Trondheim are included in the total FYSIOPRIM database bData from HELFO Trondheim for all adult patients seeking physiotherapy services from private practitioners with a reimbursement privilege Table 1 Demographics of adult patients registered in the FYSIOPRIM database and data from HELFO in Trondheim FYSIOPRIM T t l FYSIOPRIM T dh i a HELFO T d 1 Demographics of adult patients registered in the FYSIOPRIM database and data from HELFO in Trondheim FYSIOPRIM Research Program for Physiotherapy in Primary Health Care, HELFO Norwegian Health Economics Administration, PT Physiotherapist aPatients in FYSIOPRIM Trondheim are included in the total FYSIOPRIM database bData from HELFO Trondheim for all adult patients seeking physiotherapy services from private practitioners with a reimbursement privilege FYSIOPRIM Research Program for Physiotherapy in Primary Health Care, HELFO Norwegian Health Economics Administration, PT Physiotherapist aPatients in FYSIOPRIM Trondheim are included in the total FYSIOPRIM database bData from HELFO Trondheim for all adult patients seeking physiotherapy services from private practitioners with a reimbursement privilege and older patients registered in FYSIOPRIM Trondheim by the end of September 2017, 64.1% were females (Table 3). Cohort 3: Children receiving physiotherapy services in Trondheim municipality Table 4 shows demographics of children receiving physiotherapy services from municipally employed PTs in Trondheim (Cohort 3). Of the 162 children registered Evensen et al. BMC Health Services Research (2018) 18:936 Page 7 of 10 Table 2 Distribution of registered diagnostic ICPC code groups in FYSIOPRIM and HELFO in Trondheim ICPC category ICPC diagnosis ICPC code FYSIOPRIM Trondheim 01.01.16–31.12.17 (n = 1725) HELFO Trondheima 01.07.16–30.06.17 (n = 19,460) n (%) n (%) A General and unspecfied A01 - A99 53 (3.1) 698 (3.6) K Stroke/CVD K90, K91 13 (0.8) 237 (1.2) K Other heart disease Rest of K-categories 4 (0.2) 173 (0.9) L Osteoarthrosis L89, L90, L91 296 (17.2) 2298 (11.8) L Rheumatoid Arthritis L88 66 (3.8) 464 (2.4) L Fract/Sprain/Disloc/Inj L72 - L81, L96 150 (8.7) 1207 (6.2) L Hip pain L13 59 (3.4) 723 (3.7) L Knee pain L15 62 (3.6) 1093 (5.6) L Neck pain L01, L83 123 (7.1) 2125 (10.9) L Low back pain L02, L03, L84, L85, L86 127 (7.4) 2296 (11.8) L Shoulder pain L08, L92 156 (9.0) 1783 (9.2) L Other Musculoskeletal Rest of L-categories 346 (20.1) 3553 (18.3) N Headache N01, N95 21 (1.2) 254 (1.3) N Other Neurology Rest of N-categories 106 (6.1) 803 (4.1) P Psychological P01 - P99 44 (2.6) 393 (2.0) R Respiratory R01 - R99 38 (2.2) 287 (1.5) W Pregnancy, family planning W01 - W99 30 (1.7) 595 (3.1) Other Other ICPC categories All B, D, F, H, S, T, U, X, Y, Z 31 (1.8) 478 (2.5) FYSIOPRIM Research Program for Physiotherapy in Primary Health Care, HELFO Norwegian Health Economics Administration, ICPC International Classification of Primary Care aData from HELFO Trondheim for all adult patients seeking physiotherapy services from private practitioners with a reimbursement privilege FYSIOPRIM Research Program for Physiotherapy in Primary Health Care, HELFO Norwegian Health Economics Administration, ICPC International Classification of Primary Care aData from HELFO Trondheim for all adult patients seeking physiotherapy services from private practitioners with a reimbursement privilege IM Research Program for Physiotherapy in Primary Health Care, HELFO Norwegian Health Economics Administration, ICPC International Classific om HELFO Trondheim for all adult patients seeking physiotherapy services from private practitioners with a reimbursement privilege or Physiotherapy in Primary Health Care, HELFO Norwegian Health Economics Administration, ICPC International Classification of Primary Care for all adult patients seeking physiotherapy services from private practitioners with a reimbursement privilege with the PTs to decrease the burden. Cohort 3: Children receiving physiotherapy services in Trondheim municipality One major constraint for the municipally employed PTs is that their journal system is not able to import patient data from the Infopad system. Most PTs in private practice can incorp- orate data collected through Infopad into their journal system. This enables them to use the information when deciding and evaluating the treatment plan along with their patient. We have also developed summarised reports for individual patient data collected at baseline and end of treatment, providing a quick overview of patient-reported outcomes and changes throughout the treatment period. This information can easily be used in communication with other health care personnel. While it is considered a strength that the patient is involved in the goal-setting process and plans for treatment, some pa- tients may have reported better or different outcomes than they would have done with completely anonymous reporting, which may cause some reporting bias in this project. disabilities, the patients’ trajectories as well as prognostic factors for favourable outcome. Strengths and limitations The strength of the design is the systematically collected data from baseline throughout the treatment period and beyond, including PROMs, the patients’ and PTs’ main goals and plans for treatment, assessment of goal achievement and outcome. Furthermore, this large data- base is supported and supplemented by the possibility to link data to local, regional and national registers. The inclusion of patients is dependent on the cooper- ation of the PTs. The PTs invite the patients to partici- pate, which is time-consuming in terms of recruitment and data management logistics. Most PTs endure high workloads in their ordinary practice, and implementa- tion of additional electronic patient registrations may be demanding. The effect of this intrusion on their practice will be examined more closely in this project. In order to succeed, we have made extensive efforts into making the system easy, feasible and useful for the clinicians. We have had to balance the amount of variables regis- tered against the time constraints of both PTs and pa- tients. Initially, the project started with an extensive number of variables, but was downscaled in cooperation There are different sources of possible selection bias in this project. Firstly, participation by the private practi- tioners is voluntary, and thus they may not be represen- tative of the PTs in their municipalities or in Norway in general. Less than half of the invited PTs chose to par- ticipate. The PTs reported different causes as to why Evensen et al. BMC Health Services Research (2018) 18:936 Page 8 of 10 they accepted or declined participation. Strengths and limitations Common expla- nations for not participating were time-conflicts or tech- comparison showed that a larger proportion of patients was treated by psychomotor PTs and a smaller propor- Table 3 Demographics of adult and older patients registered in FYSIOPRIM Trondheim and data from Trondheim Municipality FYSIOPRIM Trondheim 01.05.16–24.09.17 (n = 655) FYSIOPRIM Trondheim 25.09–03.11.17 (n = 68) Trondheim Municipalitya 25.09–03.11.17 (n = 72) Variables n (%) n (%) n (%) Sex Female 419 (64.1) 42 (61.8) 46 (63.9) Male 235 (35.9) 26 (38.2) 26 (36.1) Missing 1 0 0 Age 18–59 years 35 (5.3) 3 (4.4) 6 (8.5) 60–69 years 59 (9.0) 5 (7.4) 9 (12.7) 70–79 years 152 (23.2) 21 (30.9) 14 (19.7) 80–89 years 291 (44.4) 28 (41.2) 34 (47.9) 90+ years 118 (18.0) 11 (16.2) 8 (11.3) Missing 0 0 1 Living conditions In own home 586 (89.7) 63 (92.7) 49 (72.1) Institution 67 (10.3) 5 (7.4) 19 (27.9) Missing 2 0 4 Referral entity Health and welfare center 150 (22.9) 13 (19.1) 21 (29.2) Patient/proxy 138 (21.1) 14 (20.6) 18 (25.0) General practitioner 120 (18.3) 15 (22.1) 6 (8.3) Home based services 83 (12.7) 12 (17.6) 7 (9.7) Hospital 82 (12.5) 7 (10.3) 8 (11.1) Rehabilitation center 28 (4.3) 2 (2.9) 1 (1.4) Occupational therapist 19 (2.9) 1 (1.5) 4 (5.6) Multidisciplinary team 13 (2.0) 1 (1.5) 0 (0.0) Other/unknown 22 (3.4) 3 (4.4) 7 (9.7) Cause of referral Geriatrics/functional deterioration/fall 180 (27.5) 14 (20.6) 25 (34.7) Orthopedics 96 (14.7) 10 (14.7) 7 (9.7) Neurology 73 (11.1) 11 (16.2) 9 (12.5) Musculoskeletal 65 (9.9) 5 (7.4) 5 (6.9) Heart or lung disease 27 (4.1) 4 (5.9) 2 (2.8) Cancer 21 (3.2) 4 (5.9) 4 (5.6) Syndrome 4 (0.6) 0 (0.0) 2 (2.8) Psychiatry 1 (0.2) 0 (0.0) 2 (2.8) Early intervention 112 (17.1) 9 (13.2) 5 (6.9) ADL rehabilitation/reablement 54 (8.2) 8 (11.8) 3 (4.2) Other/unknown 22 (3.2) 3 (4.4) 8 (11.1) FYSIOPRIM Research Program for Physiotherapy in Primary Health Care aData from Trondheim Municipality for adult and older patients receiving physiotherapy services from municipally employed physiotherapists and who were not included in FYSIOPRIM FYSIOPRIM Research Program for Physiotherapy in Primary Health Care aData from Trondheim Municipality for adult and older patients receiving physiotherapy services from municipally employed physiotherapists and who were not included in FYSIOPRIM comparison showed that a larger proportion of patients was treated by psychomotor PTs and a smaller propor- tion was treated by manual therapists in FYSIOPRIM they accepted or declined participation. Strengths and limitations Common expla- nations for not participating were time-conflicts or tech- nical issues with the electronic registration. Our they accepted or declined participation. Common expla- nations for not participating were time-conflicts or tech- nical issues with the electronic registration. Our Evensen et al. BMC Health Services Research (2018) 18:936 Page 9 of 10 Page 9 of 10 Trondheim compared with data from HELFO Trond- heim (Table 1). Secondly, the PTs did not include all their patients in the project. Even though more than 100 private practi- i f h i d i h Table 4 Demographics of children registered in FYSIOPRIM Trondheim and data from Trondheim Municipality FYSIOPRIM Trondheim 01.05.16–30.04.17 (n = 162) Trondheim Municipalitya 25.09–03.11.17 (n = 72) Variables n (%) n (%) Sex Female 71 (43.8) 28 (38.9) Male 91 (56.2) 44 (61.1) Age 0–1 years 99 (61.1) 41 (56.9) 2–3 years 26 (16.0) 15 (20.8) 4–6 years 12 (7.4) 2 (2.8) 7–9 years 10 (6.2) 7 (9.7) 10–12 years 12 (7.4) 5 (6.9) 13–16 years 3 (1.9) 2 (2.8) 17–18 years 0 (0) 0 (0) Referral entity Primary health care center 82 (50.6) 45 (62.5) Hospital 29 (17.9) 10 (13.9) Personnel in kindergarten 24 (14.8) 5 (6.9) Personnel in school 8 (4.9) 1 (1.4) School health care services 8 (4.9) 4 (5.6) Children’s and family’s services 2 (1.2) 1 (1.4) General practitioner 3 (1.9) 3 (4.2) Occupational therapist 0 (0) 0 (0) Proxy/parents 5 (3.1) 3 (4.2) Other 1 (0.6) 0 (0) Cause of referral Motor development 52 (32.1) 23 (31.9) Asymmetry (0–1 years) 47 (29.0) 22 (30.1) Orthopedics (gait, foot alignment) 26 (16.0) 16 (22.2) Preterm 10 (6.2) 1 (1.4) Diagnosis/syndrome 5 (3.1) 3 (4.2) Advice physical activity 5 (3.1) 1 (1.4) Multidisciplinary assessment 1 (0.6) 1 (1.4) Heart or lung disease 1 (0.6) 0 (0) Other 15 (9.3) 5 (6.9) FYSIOPRIM Research Program for Physiotherapy in Primary Health Care aData from Trondheim Municipality for all children receiving physiotherapy services from municipally employed physiotherapists Table 4 Demographics of children registered in FYSIOPRIM Trondheim and data from Trondheim Municipality number of patients recruited per PT was less than 25 in the 2.5 year period until end of 2017. Municipally employed PTs included less than half of all eligible patients (Table 3). This may be due to time conflicts or that recruitment was simply forgotten in a busy clinical practice. Strengths and limitations It could also be due to the PT’s personal opinion as to whether a patient was eligible. Even though we have compared our population with external data, we cannot completely exclude that selection biases may have influenced the study population. In the period from June 2015 through December 2017, more than 3000 patients participated in the baseline recording. Our descriptive results at baseline showed that patients seek- ing physiotherapy services in private practice were largely similar to those in the register from HELFO. The FYSIOPRIM database contains proportionally a larger number of patients with osteoarthritis in private prac- tice, possibly due to the fact that some of the most dedi- cated PTs performed mostly group-based physiotherapy for this patient group, and thus contributing a large number of patients into the project. A lower proportion of manual therapists in FYSIOPRIM compared with those claiming reimbursement from HELFO may explain the lower rates of neck and low back patients in FYSIO- PRIM Trondheim. Also, a larger proportion of the man- ual therapists did not register their patients’ ICPC code, thus making even fewer patients available for the com- parison with HELFO data. Among the patients of muni- cipally employed PTs, sex and age distribution in children, adult and older patients were similar to the ag- gregated data collected by the municipality in the same time period. A large number of the patients treated by the municipally employed PTs, especially the older pa- tients, were not considered eligible for recruitment for various reasons. As a result, a smaller proportion of pa- tients living in an institution and fewer patients with ge- riatrics/functional deterioration as the cause of referral were included in FYSIOPRIM Thirdly, loss to follow-up and missing data are another concern to the validity of longitudinal time series designs with repeated measurements. This may in particular affect the number of participants with complete trajectories in our project. To evaluate whether differential loss to follow-up occurs, we will compare baseline characteristics of completers with patients lost to follow-up. Usability of a primary health care physiotherapy database This project addresses the lack of comprehensive and systematically collected data for patients receiving physiotherapy services in primary health care in Norway. Several studies are planned, including cross-sectional, longitudinal and methodological studies. The compre- hensive baseline assessments enables description of patients’ demographics, general health and quality of life, Competing interests h h d l h Competing interests The authors declare that they have no competing interests. In this paper we have presented the design and framework of a comprehensive longitudinal observa- tional project in primary care physiotherapy, including characteristics of the participants at baseline. Represen- tativeness was assessed by comparing patients in the FYSIOPRIM database with external populations. These comparisons showed that our patients are largely compar- able to the composition of patients treated by both private practitioners and municipally employed PTs in Norway. Funding 5. Nicholas P, Hefford C, Tumilty S. The use of the patient-specific functional scale to measure rehabilitative progress in a physiotherapy setting. J Man Manip Ther. 2012;20(3):147–52. The FYSIOPRIM Research Program is funded by the Norwegian Fund for Post-Graduate Training in Physiotherapy. The funder had no role in the design of the project and collection, analysis, and interpretation of data or in writing the manuscript. The consortium partners, the University of Oslo, the Norwegian University of Science and Technology and Trondheim Municipality all contribute to the funding by financing researchers and running costs. 6. Stratford P, Gill C, Westaway M, et al. Assessing disability and change on individual patients: a report of a patient specific measure. Physiother Can. 1995;47(4):258–63. Author details 1 1Department of Public Health and Nursing, Norwegian University of Science and Technology (NTNU), Trondheim, Norway. 2Department of Clinical and Molecular Medicine, NTNU, Trondheim, Norway. 3Department of Health Sciences, University of Oslo, Oslo, Norway. 4Unit for Physiotherapy Services, Trondheim Municipality, Trondheim, Norway. 5Department of Circulation and Medical Imaging, NTNU, Trondheim, Norway. 6Norwegian Advisory Unit on Complex Symptom Disorders, St. Olavs hospital, Trondheim University Hospital, Trondheim, Norway. 7Department of Physiotherapy, Oslo Metropolitan University, Oslo, Norway. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Abbreviations FYSIOPRIM R FYSIOPRIM: Research Program for Physiotherapy in Primary Health Care; HELFO: Norwegian Health Economic Administration; PROM: Patient-reported outcome measures; PSFS: Patient-Specific Function Scale; PT: Physiotherapist 2. Paulsen B. Hva dreier det seg om? En analyse av møtet mellom pasienter og behandlere i kommunale fysioterapitjenester. Rapport 4/89. NIS, Sintef: Trondheim; 1989. Additional files Additional file 1: Table S1. Overview of variables in the FYSIOPRIM database. (DOCX 127 kb) Additional file 2: List of treatment options completed by the physiotherapists at follow-up. (DOCX 28 kb) Additional file 1: Table S1. Overview of variables in the FYSIOPRIM database. (DOCX 127 kb) Additional file 1: Table S1. Overview of variables in the FYSIOPRIM database. (DOCX 127 kb) Received: 10 September 2018 Accepted: 16 November 2018 Received: 10 September 2018 Accepted: 16 November 2018 Additional file 2: List of treatment options completed by the physiotherapists at follow-up. (DOCX 28 kb) Additional file 2: List of treatment options completed by the physiotherapists at follow-up. (DOCX 28 kb) Consent for publication l bl h Not applicable as the manuscript does not contain individual person’s data. Not applicable as the manuscript does not contain individual person’s data. Acknowledgements We would like to thank the Norwegian Fund for Post-Graduate Training in Physiotherapy for funding of the FYSIOPRIM Research Program. We are grateful to all physiotherapists and patients participating in FYSIOPRIM. 3. Steiro A, Hauksdottir N. Fysioterapi i Tromsø. Undersøkelse av fysioterapi i Tromsø 1983. Tromsø helseråd: Tromsø; 1987. 4. Vasseljen O, Hansen AE. Pasienter i privat praksis - hvem er de og hva lider de av? Fysioterapeuten. 2002;5(69):13–8. References 1. Swinkels IC, Hart DL, Deutscher D, et al. Comparing patient characteristics and treatment processes in patients receiving physical therapy in the United States, Israel and the Netherlands: cross sectional analyses of data from three clinical databases. BMC Health Serv Res. 2008;8:163. Availability of data and materials The datasets generated and/or analysed during the current study are not publicly available due to permission has not been applied for from neither the participants nor the Ethical Committee but might be available from the corresponding author on reasonable request. Ethics approval and consent to participate The project is conducted according to the Helsinki declaration and approved by the Regional committee for Medical and Health Research Ethics in Norway (REC no. 2013/2030). Project information and consent form were available in Norwegian and English. All patients have provided written informed consent to participate. Parents gave consent when the children were less than 16 years old. Trondheim compared with data from HELFO Trond- heim (Table 1). Trondheim compared with data from HELFO Trond- heim (Table 1). Secondly, the PTs did not include all their patients in the project. Even though more than 100 private practi- tioners so far have recruited patients, the average Page 10 of 10 Evensen et al. BMC Health Services Research (2018) 18:936 Page 10 of 10 diagnoses, psychosocial factors, physical function, pain localisation, pain intensity, as well as working ability and sick leave. Furthermore, it enables studies of associations between patient characteristics and treatments and longi- tudinal assessments of change throughout the treatment course. Moreover, we will be able to identify prognostic factors for improvement for subgroups of patients. Ethics approval and consent to participate Authors’ contributions d h d KAIE participated in the design of the project, contributed to data collection, data analyses and drafted the manuscript. HSR participated in the design of the project, contributed to data collection and helped to draft the manuscript. IM participated in the design of the project, contributed to data collection, data cleaning, data analyses and redrafts. AW and GM participated in the design of the project, contributed to data collection and redrafts. MT contributed to data collection, data cleaning and redrafts. WSB and AEH participated in the design of the project and contributed to redrafts. OV participated in the design of the project, contributed to data analyses and redrafts. NKV participated in the design of the project, supervised data collection and contributed to redrafts. All authors have approved the final version of the manuscript.
https://openalex.org/W4312277360	https://ojs.uajy.ac.id/index.php/jai/article/download/5228/2674	Indonesian	null	Potensi sentra olahan ubi kayu dan mitigasi bencana di Desa Kranggan, Galur, Kulon Progo	Jurnal Atma Inovasia	2,022	cc-by-sa	3,669	Received 16 Juni 2021; Revised - ; Accepted for Publication 28 Juli 2022; Published 28 Juli 2022 Abstract — Kuliah Kerja Nyata (KKN) are part of a mandatory activity to fulfill student obligations, namely community service, KKN in batch 80 is carried out online due to the covid-19 pandemic. Therefore, this research was conducted to fulfill the task given by Atma Jaya University Yogyakarta. The Kranggan Village area is one of the villages located in the Galur sub-district, Kulon Progo Regency which is located close to the southern coast of the island of Java. Therefore, it is necessary to have comprehensive efforts in policy making to reduce the risk caused by natural disasters such as doing mitigation. In addition, this study aims to educate the Kranggan village community in increasing search results from cassava plants, where the Kranggan village area has 91.0 hectares of land for cassava plants in 2018. Data collection techniques in this research are observation, collection of literature studies from the internet and discussions with lecturers. The results of this study were made together with a pocket book or e-book entitled "Potency of Kranggan Village, Processed Cassava" and an e-book or pocket book "Disaster Mitigation". The target of this KKN program is the formation of a home industry that is able to compete in the market and is managed directly by the Kranggan village community and provides education if a natural disaster occurs, so that the community understands what actions to take in certain situations. Therefore, Group 79 chose to make an e-book on village potential on the use of cassava and disaster mitigation counseling. The results of this study are expected to be useful in the future for the people of Kranggan Village in increasing their income from making cassava chips in addition to the work of the majority of the population is farming and is expected to be useful for the people of Kranggan Village to always be responsive and alert in taking action to deal with natural disasters. Target program KKN ini adalah terbentuknya industri rumah tangga yang mampu bersaing di pasaran dan di kelola langsung oleh masyarakat desa kranggan dan memberi edukasi jika nanti terjadi bencana alam, agar masyarakat mengerti tindakan apa yang harus dilakukan dalam situasi-situasi tertentu. Oleh karena itu, Kelompok 79 memilih untuk membuat e-book potensi desa tentang pemanfaatan ubi kayu dan penyuluhan mitigasi bencana. I. PENDAHULUAN Indonesia terletak pada kondisi geografis yang diapit tiga lempeng aktif, yaitu Indo-Australia, Eurasia dan Pasifik yang mengakibatkan negara tersebut memiliki tingkat kerawanan tinggi terhadap bencana geologis dan hidro-klimatologis. Badan Nasional Penanggulangan Bencana (BNPB) mencatat bahwa dalam kurun waktu tahun 2020 telah terjadi 2.925 kejadian bencana alam pada tanah air ini, oleh karena itu tidak sedikit orang yang mengalami kerugian, kerusakan hingga banyak memakan korban jiwa [1]. Hal tersebut dapat menjelaskan bahwa kurangnya kesiapsiagaan terhadap bencana yang terjadi di Indonesia. Menurut [2], bencana adalah peristiwa atau rangkaian peristiwa yang mengancam dan mengganggu kehidupan dan penghidupan masyarakat yang disebabkan, baik oleh faktor alam dan/atau faktor non alam maupun faktor manusia, sehingga mengakibatkan timbulnya korban jiwa manusia, kerusakan lingkungan, kerugian harta benda, dan dampak psikologis. Keywords — village potential, cassava, disaster mitigation, tsunami, flood, counseling, kranggan village, community service program, kulon progo, Abstrak— Kuliah Kerja Nyata (KKN) merupakan bagian kegiatan wajib untuk memenuhi kewajiban mahasiswa yaitu pengabdian masyarakat, KKN pada angkatan 80 ini dilakukan secara daring dikarenakan pandemi covid-19. Oleh karena itu penelitian ini dilakukan untuk memenuhi tugas yang di berikan oleh Universitas Atma Jaya Yogyakarta. Daerah Desa Krangganmerupakan salah satu desa yang terletak di kecamatan Galur Kabupaten Kulon Progo yang kondisinya terletak dekat dengan pesisir pantai selatan pulau Jawa. Oleh karena itu diperlukan adanya upaya-upaya komperhensif dalam pengambilan kebijakan guna mengurangi resiko yang diakibatkan bencana alam seperti melakukan mitigasi. Selain itu, penelitian ini bertujuan untuk mengedukasi masyarakat desa kranggan dalam meningkatkan hasil pencarian dari tanaman ubi kayu, dimana wilayah desa Kranggan memiliki lahan 91,0 Ha untuk tanaman ubi kayu pada tahun 2018. Teknik pengumpulan data dalam dalam penelitian ini adalah dengan melakukan observasi, pengumpulan studi literatur dari internet dan diskusi bersama dosen. Hasil dari penelitian ini dibuat bersama dengan buku saku atau e- book yang berjudul “Potensi Desa Kranggan, Olahan Ubi Kayu" dan e-book atau buku saku "Mitigasi Bencana". Daerah Istimewa Yogyakarta merupakan salah-satu contoh dari banyak daerah yang rawan mengalami kejadian bencana alam, khususnya pada Desa Kranggan yang merupakan salah-satu desa yang terletak di Kecamatan Galur Kabupaten Kulon Progo yang dimana kondisinya terletak dekat dengan pesisir Pantai Selatan Pulau Jawa. Desa Kranggan merupakan daerah dengan iklim tropis dengan tingkat curah hujan yang tinggi [3]. Hal ini yang membuat wilayah tersebut rawan terjadi guncangan gempa bumi yang dapat berdampak pada tsunami mengingat bahwa lokasi wilayah berdekatan dengan pesisir pantai. Received 16 Juni 2021; Revised - ; Accepted for Publication 28 Juli 2022; Published 28 Juli 2022 Hasil penelitian ini diharapkan dapat berguna kelak untuk masyarakat Desa Kranggan dalam menambah penghasilan dari membuat keripik ubi kayu disamping pekerjaan mayoritas penduduk adalah bertani dan diharapkan berguna bagi masyarakat Desa Kranggan untuk selalu tanggap dan waspada dalam mengambil tindakan upaya menghadapi bencana alam. Kata Kunci—potensi desa, ubi kayu, mitigasi bencana, tsunami, banjir, penyuluhan, desa kranggan, kuliah kerja nyata, kulon progo. Jurnal Atma Inovasia (JAI) Vol. 2, No. 4, Juli 2022 Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kranggan, Galur, Kulon Progo Jurnal Atma Inovasia (JAI) Vol. 2, No. 4, Juli 2022 Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kranggan, Galur, Kulon Progo Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kranggan, Galur, Kulon Progo Sergio Pratama, Fatimah Novianti, Cristhopel Gurusinga, Amanda Putri Eriandina Satriani, Maria Btari Larasati Marhaeni,Christina Natalia Kharis, Lo, Dea Katelya, Gisela Radiastasha Yuniarka, Pita Desi Putrianti, Deon Natanael Malau, Brigitta Laksmi Paramita Universitas Atma Jaya Yogyakarta, Jl. Babarsari No 43, Yogyakarta Email: brigitta.laksmi@uajy.ac.id p-ISSN: 2775-9385 e-ISSN: 2775-9113 p-ISSN: 2775-9385 e-ISSN: 2775-9113 Jurnal Atma Inovasia (JAI) Vol. 2, No. 4, Juli 2022 Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kranggan, Galur, Kulon Progo Jurnal Atma Inovasia (JAI) Vol. 2, No. 4, Juli 2022 Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kranggan, Galur, Kulon Progo D. Analisis dan Pembahasan Penelitian ini dilakukan secara deskriptif kualitatif. Analisis dan pembahasan menjadi tolak ukur pada hasil luaran yang telah dibuat. Hasil luaran e-book dan laporan dibuat setelah seluruh data telah lengkap dan terkumpul didukung oleh informasi tambahan yang didapat melalui tinjauan literatur. I. PENDAHULUAN Oleh karena itu, diperlukan adanya upaya-upaya komprehensif dalam pengambilan kebijakan guna mengurangi risiko yang diakibatkan bencana alam seperti melakukan kegiatan mitigasi. Mitigasi merupakan suatu 440 p-ISSN: 2775-9385 e-ISSN: 2775-9113 Jurnal Atma Inovasia (JAI) Vol. 2, No. 4, Juli 2022 Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kranggan, Galur, Kulon Progo Jurnal Atma Inovasia (JAI) Vol. 2, No. 4, Juli 2022 Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kranggan, Galur, Kulon Progo Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kranggan, Galur, Kulon Progo Gambar 1 Kerangka Metode Pengabdian Pengidentifikasian Masalah di Desa Kranggan Studi Pustaka Penentuan Tema mengenai Ubi Kayu dan Mitigasi Bencana Analisis dan Pembahasan Penarikan Kesimpulan tindakan-tindakan guna mengurangi atau meminimalkan potensi dampak negatif dari suatu bencana, sedangkan kegiatan mitigasi merupakan salah satu bagian dari kegiatan penanganan bencana yang difokuskan untuk mengurangi potensi dampak yang mungkin ditimbulkan oleh bencana yang diprediksikan akan terjadi di masa datang. Berdasarkan kejadian peristiwa bencana alam yang sering terjadi pada wilayah daerah Kabupaten Kulonprogo, maka diperlukan adanya upaya mitigasi yang dapat dilakukan dengan cara membuat pemetaan wilayah aman dari dampak gempa bumi dan tsunami. Penarikan Kesimpulan Gambar 1 Kerangka Metode Pengabdian A. Pengidentifikasian Masalah Pengidentifikasian masalah dilakukan di awal dalam rangka mencari tahu permasalahan yang ada di Desa Kranggan. Proses identifikasi masalah dilakukan secara daring melalui situs Desa Kranggan yang menampilkan informasi mengenai kondisi desa. Melalui berbagai informasi yang tersedia, penulis dapat mengetahui permasalahan yang muncul di Desa Kranggan. Permasalahan ini kemudian diproses pada tahap metodologi sebagai upaya penyelesaian masalah C. Penentuan Tema Melimpahnya hasil panen komoditas ubi kayu di Desa Kranggan, membuat peluang bisnis yang menguntungkan bagi masyarakat. Ubi kayu dapat ditingkatkan nilai jualnya dengan pengolahan dan pemasaran yang baik. Ubi Kayu dapat diolah menjadi berbagai jenis olahan yaitu dapat dikategorikan menjadi 3 jenis olahan ubi kayu berupa (1). makanan tradisional seperti tiwul, gogik, gatot, growol, dan tape; (2). makanan pokok seperti liwet singkong dan nasi singkong; (3). makanan jajanan seperti kue kacamata, lemet, getuk, kripik, kerupuk dan lain sebagainya. Pengolahan ubi kayu yang paling sederhana dan tidak memerlukan banyak alat yang sangat cocok untuk masyarakat desa kranggan adalah olahan keripik singkong [5]. Penentuan tema atau topik penelitian dilakukan setelah studi pustaka. Penentuan tema menjadi hal yang penting dalam menentukan arah penelitian. Tema merupakan dasar dari sebuah penelitian, Tema yang menarik dapat menjadi minat bagi pembaca khususnya masyarakat Desa Kranggan. Tema dari penelitian ini adalah mitigasi bencana dan potensi ubi kayu di Desa Kranggan. A. Potensi Desa Kranggan Dari hasil diskusi serta pencarian informasi melalui situs website dan jurnal terkait informasi di Desa Keranggan diperoleh beberapa potensi desa yang dapat dikembangkan masyarakat yang bermanfaat untuk meningkatkan perekonomian desa keranggan, salah satu potensi desa yang dipilih dari diskusi kelompok yaitu pengolahan ubi kayu. Berdasarkan data statistik Badan Pusat Statistik Kulon Progo luas lahan yang digunakan untuk tanaman ubi kayu di tahun 2018 seluas 91,0 Ha, ubi kayu banyak ditanam di desa kranggan kulon progo karena proses penanaman ubi kayu yang mudah dan perawatannya tidak sulit sehingga banyak masyarakat memanfaatkan lahanya selain untuk menanam padi juga menanam ubi kayu [4]. Ubi kayu dapat dimanfaatkan sebagai olahan makanan dan juga dapat dimanfaatkan sebagai pakan ternak. B. Studi Pustaka Studi pustaka dilakukan setelah pengidentifikasian masalah. Data dari penelitian ini bersumber dari data sekunder. Informasi dikumpulkan secara daring seperti jurnal penelitian, artikel, dan situs daring. Studi pustaka dilakukan dengan tujuan untuk mendalami permasalahan yang ada di Desa Kranggan. C. Penentuan Tema Penentuan tema atau topik penelitian dilakukan setelah studi pustaka. Penentuan tema menjadi hal yang penting dalam menentukan arah penelitian. Tema merupakan dasar dari sebuah penelitian, Tema yang menarik dapat menjadi minat bagi pembaca khususnya masyarakat Desa Kranggan. Tema dari penelitian ini adalah mitigasi bencana dan potensi ubi kayu di Desa Kranggan. E. Penarikan Kesimpulan Pengolahan Keripik khas merupakan salah satu alternatif yang dapat digunakan untuk meningkatkan nilai jual Ubi Kayu di kalangan masyarakat. Diperlukan inovasi terhadap Ubi Kayu agar mampu menarik minat pembeli dan meningkatkan pendapatan masyarakat. Untuk mengolah Ubi Kayu menjadi Keripik Khas yang cita rasa dan bernilai, masyarakat perlu memperhatikan Penarikan kesimpulan dilakukan setelah proses analisis dan pembahasan. Kesimpulan diambil berdasarkan identifikasi, pengumpulan data, dan proses- proses sebelumnya. Proses ini akan diketahui manfaat- manfaat yang dapat dihasilkan penelitian ini bagi Desa Kranggan. 441 p-ISSN: 2775-9385 e-ISSN: 2775-9113 Jurnal Atma Inovasia (JAI) Vol. 2, No. 4, Juli 2022 Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kranggan, Galur, Kulon Progo beberapa hal, mulai dari pemilihan Ubi Kayu hingga pengemasan Keripik [6] memasukkan hasil parutan tadi ke dalam karung, kemudian dipres. Hasil parutan yang sudah dipres akan diayak ke dalam wadah dan dicampur dengan bumbu yang terbuat dari bawang putih, ketumbar, garam dan kemiri. Hasil parutan yang sudah dicampur ini akan dikukus menggunakan tungku api, kemudian digiling dengan penggilingan slondok. Adonan dari hasil penggilingan ini dipotong-potong sepanjang 5cm dan dibentuk persegi panjang, lalu diletakkan di wadah untuk dijemur hingga kering. Setelah kering, keripik Slondok siap untuk digoreng dan diberi bumbu pedas manis atau keju sebelum dikemas dengan menarik. p g p Ubi Kayu berkualitas tinggi dipilih yang memiliki ciri-ciri tidak terlalu tua/berkayu dan tidak berwarna biru. Kemudian Ubi Kayu dikupas dan dicuci. Ubi kayu yang sudah dicuci bersih menjadi irisan tipis setebal 1,5-2 mm yang nantinya akan dimasukkan ke dalam larutan kapur sirih (bisa pakai amonium bikarbonat) dengan banyak air dan diaduk hingga larut agar keripik terasa renyah. Kemudian masukkan irisan Ubi Kayu ke dalam larutan kapur sirih dan rendam selama ± 30 menit. Irisan ubi kayu tadi dibilas beberapa kali menggunakan air hingga bersih dan ditiriskan. Persiapkan wajan yang diisi minyak goreng secukupnya. Api dinyalakan kecil untuk mendapatkan tingkat kematangan yang baik, setelah panas masukan irisan Ubi Kayu ke dalam wajan. Setelah keripik ubi kayu mulai matang, besarkan api sedikit dan tunggu hingga keripik ubi kayu berwarna sedikit kecoklatan. Ubi kayu yang sudah matang, diangkat dan ditiriskan. Kemudian Keripik dikeringkan menggunakan mesin spinner agar terbebas dari sisa minyak goreng sehingga masa penyimpanan lebih lama [6] Slondok khas Kranggan yang sudah dikemas akan dipasarkan melalui pasar tradisional dan kios yang berada di kecamatan Galur. Produk ini juga dapat dipasarkan melalui tempat-tempat wisata yang berada di kecamatan Galur. 1. Tsunami Menurut [8], tsunami merupakan suatu peristiwa bencana alam yang dapat terjadi akibat adanya perpindahan badan air yang disebabkan oleh perubahan permukaan laut secara vertikal dalam kondisi tiba-tiba. Upaya penyelamatan tsunami terdiri dari : E. Penarikan Kesimpulan Selain pasar tradisional, Produk ini juga dijual di media sosial agar masyarakat dapat mengenal produk ini dengan lebih mudah dan memesannya. Proses pengolahan ubi kayu menjadi keripik Slondok dapat mempekerjakan sekitar 7-12 orang yang dapat berasal dari kelompok tani desa Kranggan kecamatan Galur. Dengan demikian kelompok tani dapat tetap produktif dan meningkatkan pendapatan dari hasil tani Ubi kayu. Setelah dilakukan proses pembuatan keripik singkong dilanjutkan dengan pengemasan. Pengemasan ini dilakukan dengan menggunakan vacuum. Keripik di masukan kedalam plastik dan ditimbang kurang lebih 75 gram. Setelah di timbang plastik divacuum dan diberi label merek. Setelah itu dimasukan lagi kedalam paperbag. Keripik Khas Kranggan dapat dipasarkan melalui pasar lokal, tempat wisata sekitar desa, dan kios setempat sebagai oleh-oleh khas dari Desa Kranggan. Proses pembuatan Keripik Khas Kranggan ini dapat dilakukan oleh ibu-ibu rumah tangga setempat. Mulai dari pengolahan keripik, pengemasan sampai kepada pemasaran. Ibu-ibu diberikan pelatihan terlebih dahulu sebelumnya. Pelatihan ini diberikan guna meningkatkan ekonomi desa Keranggan dan memberdayakan masyarakat setempat khususnya ibu-ibu agar lebih produktif. D. Mitigasi Bencana di Desa Kranggan Berdasarkan diskusi yang dihasilkan selain potensi desa pengolahan ubi kayu, terdapat mitigasi bencana di Desa Kranggan akibat kurangnya kesiapsiagaan dalam menghadapi bencana alam. Maka hasil buku saku program kerja adalah Mitigasi Bencana Desa Kranggan, yang berisi tentang penjelasan singkat mengenai bencana alam tsunami dan gempa bumi serta upaya mengurangi maupun menghilangkan resiko akibat bencana alam sebagai bentuk pemahaman masyarakat Desa Kranggan. C. Pengolahan Ubi Kayu menjadi Slondok Khas di Desa Kranggan a. Jika guncangan gempa bumi dan disusul air laut dari sekitar pantai, segera lari menuju ke tempat yang tinggi. Banyak inovasi yang telah dilakukan oleh masyarakat khususnya di Indonesia untuk meningkatkan nilai jual Ubi Kayu, salah satunya adalah mengolah Ubi Kayu menjadi kerupuk Slondok. Slondok adalah sejenis jajanan yang dibuat dari bahan baku ubi kayu dengan cita rasa khas yang berbeda dengan keripik ubi kayu pada umumnya, karena cara pembuatannya juga berbeda. Keripik yang berbentuk pipih dengan panjang sekitar 5cm, dan rasa yang gurih dapat dijadikan oleh-oleh khas dengan menambahkan varian rasa baru. b. Jika mendengar berita dari pantai dan sedang berada di perahu atau kapal di tengah laut, jangan mengarahkan perahu atau kapal mendekat ke pesisir pantai. c. Segera mengungsi jika mendapat pemberitahuan umum tentang tsunami dan mengutamakan keselamatan terlebih dahulu. DAFTAR PUSTAKA b. Segera mungkin lari ke luar rumah serta mencari tempat terbuka dan aman. Namun bila berada di dalam rumah, segera berlindung dibawah meja atau kursi yang kokoh. b. Segera mungkin lari ke luar rumah serta mencari tempat terbuka dan aman. Namun bila berada di dalam rumah, segera berlindung dibawah meja atau kursi yang kokoh. [1] BNPB, “Informasi Bencana,” Badan Nasional Penanggulangan Bencana, 2020. https://bnpb.go.id/informasi-bencana. c. Menghubungi posko bantuan seperti Badan Penanggulangan Bencana Alam dan kontak darurat lainnya. c. Menghubungi posko bantuan seperti Badan Penanggulangan Bencana Alam dan kontak darurat lainnya. [2] [2] UU RI, “UNDANG-UNDANG REPUBLIK INDONESIA NOMOR 24 TAHUN 2007 TENTANG PENANGGULANGAN BENCANA,” 2007, [Online]. Available: http://digilib.unila.ac.id/4949/15/BAB II.pdf. Jurnal Atma Inovasia (JAI) Vol. 2, No. 4, Juli 2022 Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kranggan, Galur, Kulon Progo Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di De [6] I. Indardi, “Peningkatan Ekonomi Masyarakat Melalui Pembuatan Keripik Singkong Di Semuluh Kidul, Semanu, Gunung Kidul,” BERDIKARI J. Inov. dan Penerapan Ipteks, vol. 6, no. 1, pp. 53–64, 2018, doi: 10.18196/bdr.6133. [8] K. Rahman, “Mitigasi Bencana Tsunami,” BPBD DIY, 2018. http://www.bpbd.jogjaprov.go.id/berita/mitigasi- bencana-tsunami. Desa kranggan merupakan desa yang berlokasi di dekat pesisir pantai serta memiliki resiko bencana gempa bumi serta tsunami. Kebijakan yang dilakukan dalam mengurangi resiko yang diakibatkan oleh bencana alam dapat dilakukan penyuluhan berupaa kegiatan mitifasi sehingga masyarakat bisa mengetahui upaya – upaya yang dapat dilakukan saat terjadi bencana alam seperti tsunami dan gempa bumi. Penulis juga mengedukasi cara evakuasi setelah terjadi bencana tsunami dan gempa bumi untuk mengurangi resiko korban luka ataupun korban jiwa. Hal ini berguna juga agar memudahkan tim SAR untuk melakukan evakuasi jika sewaktu-waktu terjadi bencana karena masyarakat desa menjadi paham tindakan apa yang harus mereka lakukan. [9] [9] A. M. Nur, “Gempa Bumi, Tsunami Dan Mitigasinya,” J. Geogr., vol. 7, no. 1, 2010, doi: 10.15294/jg.v7i1.92. [10] [10] Humas BNPB, “Mitigasi Bencana bagi Penyandang disabilitas,” BNPB, 2019. https://bnpb.go.id/berita/mitigasi-bencana-bagi- penyandang-disabilitas. Saran yang dapat diberikan oleh penulis adalah di tahap selanjutnya perlu adanya pelatihan berupa materi dunia bisnis untuk membimbing masyarakat Desa Kranggan, agar potensi yang menjadi peluang bisnis ini dapat berjalan berkelanjutan. 3. Bagi Penyandang Disabilitas Berdasarkan [10], penyandang disabilitas menjadi kemungkinan terbesar korban bencana dikarenakan risiko yang lebih tinggi, memperburuk fisik dan mental korban, serta menambah jumlah penyandang disabilitas. Upaya penyelamatan bagi penyandang disabilitas terdiri dari : [3] [3] Pemerintah Kalurahan Kranggan, “Profil Desa Kranggan Kulon Porogo,” 2019. http://kranggan- kulonprogo.desa.id/index.php/artikel/2019/3/5/profil -wilayah-desa. a. Mengelompokkan berdasarkan kondisi maupun keadaan masyarakat, jenis disabilitas, dan setiap wilayahnya. b. Jika terjadi bencana alam, membantu untuk mengarahkan tempat yang lebih aman, menjauh dari tempat tinggi, berbahaya, serta barang yang mudah jatuh dan pecah. [4] Badan Pusat Statistik, “Kapanewon Galur Dalam Angka 2020,” 2020. [5] Indardi, “Peningkatan Ekonomi Masyarakat Dengan Perbaikan Nilai,” pp. 766–776, 2019. p c. Memberikan fasilitas terapi dan konseling psikologi bagi korban atau penyandang disabilitas. [6] 2. Gempa Bumi Menurut [9], gempa Bumi merupakan getaran asli yang berasal dari dalam bumi yang kemudian merambat ke permukaan bumi akibat adanya rekahan bumi pecah dan bergeser dengan keras. Upaya penyelamatan gempa bumi terdiri dari : Menurut [7], Slondok diolah menggunakan cara dan peralatan yang masih tradisional. Ubi kayu yang telah dipanen akan dikumpulkan ke tempat pengolahan dan langsung dikupas. Ubi kayu yang sudah dikupas bersih, akan dicuci kemudian diparut menggunakan mesin parut. Hasil parutan ubi kayu ini masih mengandung banyak air dan zat lain yang harus dipisahkan, sehingga masyarakat a. Tenang dan tidak panik. 442 p-ISSN: 2775-9385 e-ISSN: 2775-9113 p-ISSN: 2775-9385 e-ISSN: 2775-9113 IV. KESIMPULAN Ubi kayu adalah salah satu komoditas di Desa Kranggan yang memiliki potensi peluang bisnis yang baik bagi masyarakat. Peningkatan nilai jual ubi kayu perlu dilakukan untuk meningkatkan peluang bisnis tersebut, yaitu dengan cara mengolah ubi kayu menjadi keripik dan slondok. Kedua jenis makanan tersebut memiliki masa simpan yang relatif lama dan mudah untuk diterima masyarakat, sehingga dapat dijadikan oleh-oleh makanan khas Desa Kranggan. [7] Iswidodo, “Ternyata Begini Cara Bikin Slondok Singkong Khas Magelang,” 2020. https://jateng.tribunnews.com/2020/09/03/ternyata- begini-cara-bikin-slondok-singkong-khas-magelang. [8] UCAPAN TERIMAKASIH UCAPAN TERIMAKASIH Ucapan terima kasih oleh penulis ditujukan kepada Lembaga Penelitian dan Pengabdian Masyarakat (LPPM) Universitas Atma Jaya Yogyakarta, selaku pihak yang menyelenggarakan kegiatan Kuliah Kerja Nyata Society 5.0. 443 p-ISSN: 2775-9385 e-ISSN: 2775-9113 Jurnal Atma Inovasia (JAI) Vol. 2, No. 4, Juli 2022 Vol. 2, No. 4, Juli 2022 Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kra PENULIS Sergio Pratma , prodi Informatika, Fakultas Teknologi Industri, Universitas Atma Jaya Yogyakarta. Fatimah Novianti, prodi Biologi Fakultas Teknobiologi, Universitas Atma Jaya Yogyakarta. Cristhopel Gurusinga, prodi Ekonomi Pembangunan, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. Amanda Putri Eriandina Satriani, prodi Arsitektur, Fakultas Teknik, Universitas Atma Jaya Yogyakarta. Maria Btari Larasati Marhaeni, prodi Manajemen, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. Christina Natalia Kharis, Lo, prodi Informatika, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. - Dea Katelya, prodi Informatika, Fakultas Teknologi Industri, Universitas Atma Jaya Yogyakarta.- Gisela Radiastasha Yuniarka, prodi Akuntansi, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. tra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kranggan, Galur, Kulon Progo Potensi Sentra Olahan Ubi Kayu dan Mitigasi Bencana di Desa Kranggan, Galur, Kulon Progo PENULIS Sergio Pratma , prodi Informatika, Fakultas Teknologi Industri, Universitas Atma Jaya Yogyakarta. Fatimah Novianti, prodi Biologi Fakultas Teknobiologi, Universitas Atma Jaya Yogyakarta. Cristhopel Gurusinga, prodi Ekonomi Pembangunan, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. Pita Desi Putrianti, prodi Ilmu Hukum, Fakultas Hukum, Universitas Atma Jaya Yogyakarta. Deon Natanael Malau, prodi Informatika, Fakultas Teknologi Industri, Universitas Atma Jaya Yogyakarta.- Brigitta Laksmi Paramita, prodi Biologi, Fakultas Teknobiologi, Universitas Atma Jaya Yogyakarta. Pita Desi Putrianti, prodi Ilmu Hukum, Fakultas Hukum, Universitas Atma Jaya Yogyakarta. Maria Btari Larasati Marhaeni, prodi Manajemen, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. Brigitta Laksmi Paramita, prodi Biologi, Fakultas Teknobiologi, Universitas Atma Jaya Yogyakarta. Christina Natalia Kharis, Lo, prodi Informatika, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. - Amanda Putri Eriandina Satriani, prodi Arsitektur, Fakultas Teknik, Universitas Atma Jaya Yogyakarta. Deon Natanael Malau, prodi Informatika, Fakultas Teknologi Industri, Universitas Atma Jaya Yogyakarta.- PENULIS Sergio Pratma , prodi Informatika, Fakultas Teknologi Industri, Universitas Atma Jaya Yogyakarta. Sergio Pratma , prodi Informatika, Fakultas Teknologi Industri, Universitas Atma Jaya Yogyakarta. Deon Natanael Malau, prodi Informatika, Fakultas Teknologi Industri, Universitas Atma Jaya Yogyakarta.- Fatimah Novianti, prodi Biologi Fakultas Teknobiologi, Universitas Atma Jaya Yogyakarta. Fatimah Novianti, prodi Biologi Fakultas Teknobiologi, Universitas Atma Jaya Yogyakarta. Brigitta Laksmi Paramita, prodi Biologi, Fakultas Teknobiologi, Universitas Atma Jaya Yogyakarta. Cristhopel Gurusinga, prodi Ekonomi Pembangunan, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. Cristhopel Gurusinga, prodi Ekonomi Pembangunan, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. Maria Btari Larasati Marhaeni, prodi Manajemen, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. Christina Natalia Kharis, Lo, prodi Informatika, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. - Christina Natalia Kharis, Lo, prodi Informatika, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. - Dea Katelya, prodi Informatika, Fakultas Teknologi Industri, Universitas Atma Jaya Yogyakarta.- Gisela Radiastasha Yuniarka, prodi Akuntansi, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. Gisela Radiastasha Yuniarka, prodi Akuntansi, Fakultas Bisnis dan Ekonomika, Universitas Atma Jaya Yogyakarta. 444
https://openalex.org/W4280521365	https://www.nature.com/articles/s41598-022-11572-3.pdf	English	null	Preprocess dependence of optical properties of ensembles and single siphonaxanthin-containing major antenna from the marine green alga Codium fragile	Scientific reports	2,022	cc-by	13,331	Tatas Hardo Panintingjati Brotosudarmo1,2, Bernd Wittmann1, Soichiro Seki3, Ritsuko Fujii3,4 & Jürgen Köhler1,5,6* Tatas Hardo Panintingjati Brotosudarmo1,2, Bernd Wittmann1, Soichiro Seki3, Ritsuko Fujii3,4 & Jürgen Köhler1,5,6* The siphonaxanthin-siphonein-Chl-a/b-protein (SCP) is the light-harvesting complex of the marine alga Codium fragile. Its structure resembles that of the major light-harvesting complexes of higher plants, LHC II, yet it features a reversed Chl a:Chl b ratio and it accommodates other variants of carotenoids. We have recorded the fluorescence emission spectra and fluorescence lifetimes from ensembles and single SCP complexes for three different scenarios of handling the samples. While the data obtained from ensembles of SCP complexes yield equivalent results, those obtained from single SCP complexes featured significant differences as a function of the sample history. We ascribe this discrepancy to the different excitation intensities that have been used for ensemble and single complex spectroscopy, and conclude that the SCP complexes undergo an aging process during storage. This process is manifested as a lowering of energetic barriers within the protein, enabling thermal activation of conformational changes at room temperature. This in turn leads to the preferential population of a red-shifted state that features a significant decrease of the fluorescence lifetime. Photosynthesis is exploited by plants, bacteria, and algae to convert solar energy into biochemical energy. This is achieved by the absorption of sunlight in specialised light-harvesting complexes (LHCs), and subsequent ultrafast and efficient transport of the excitation energy to a photochemical reaction centre (RC) that acts as transducer driving a transmembrane charge separation1. Since the prerequisites that have to be fulfilled for electron transfer are quite demanding, the structure of the RCs is mainly conserved across the various organisms that perform photosynthesis. In contrast, the rules for energy transfer are rather tolerant resulting in a rich variety for the structures of the light harvesting systems allowing for perfect adaption to the special conditions that apply to the habitat of the organism2. In contrast to the LHCs from higher plants the antennae systems from marine algae have only recently attracted more attention3–7. Codium fragile is a marine alga that occurs in open coasts and tidal pools, but it can also be found under water in a depth of up to about 20 m. Its LHC is the siphonaxanthin- siphonein-Chl-a/b-protein (SCP) that is optimized for using blue-green light, which is the dominating spectral range under water. www.nature.com/scientificreports www.nature.com/scientificreports Preprocess dependence of optical properties of ensembles and single siphonaxanthin‑containing major antenna from the marine green alga Codium fragile OPEN Tatas Hardo Panintingjati Brotosudarmo1,2, Bernd Wittmann1, Soichiro Seki3, Ritsuko Fujii3,4 & Jürgen Köhler1,5,6* Tatas Hardo Panintingjati Brotosudarmo1,2, Bernd Wittmann1, Soichiro Seki3, Ritsuko Fujii3,4 & Jürgen Köhler1,5,6* For SCP the Chl a molecules in position 602, and either those in positions 610 or 612 are exchanged to Chl b as indicated by the cyan colour. The Chl molecules are numbered according to the scheme given in9 (1RWT). The figure has been produced using PyMOL ver. 2.5.0. (Sx + Sn):Neo:Chl a:Chl b = 2.5:1:6:8 was obtained by HPLC analysis for the pigment structure from resonance Raman spectroscopy4. A similar Chl a:Chl b ratio was determined earlier also for the LHCs of PS II from the siphonous green algae Bryopsis maxima and Bryopsis corticulans6,7. In summary, in SCP the two molecules of Lut in LHC II are replaced by Sx/Sn, Vio is not present, and the Chl a to Chl b ratio is reversed with respect to that in LHC II. Both the exchange of the carotenoids and the larger amount of Chl b molecules enhance the light-harvesting efficiency of the complex in the blue-green spectral region around 450 nm, which reflects the adaption of the species to the natural habitat. Since the ligands in the Chl binding pockets are the same in LHC II and SCP7 the reversed Chl a to Chl b ratio hints for some structural flexibility of the SCP complex at these positions. Based on recent cryoEM data10 the Chl a to Chl b exchange sites are assigned to the position 602, and either to positions 610 or 612, see Fig. 1, according to the numbering scheme used in9. g g g Usually, the details in the optical spectra from a macroscopic ensemble of proteins are washed out due to ensemble averaging. In contrast, studying the complexes individually provides information about the distribu- tions of parameters rather than only about their moments, which allows to identify subpopulations that would be obscured otherwise. In the past, single-molecule spectroscopy has been applied extensively to the antennae complexes of purple bacteria11–20 and to the light-harvesting systems associated with plant photosystems I and II21–29 for obtaining information about the energetics and dynamics both at cryogenic temperatures as well as under ambient conditions, recently excellently reviewed in30. At low temperatures thermal motions of the nuclei are frozen out and the electronic transition energies that serve as input for structure-based modelling can be detected with enhanced spectral resolution. Tatas Hardo Panintingjati Brotosudarmo1,2, Bernd Wittmann1, Soichiro Seki3, Ritsuko Fujii3,4 & Jürgen Köhler1,5,6* The structure of this complex is still elusive but given the high sequence identity with the major LHC II complex from green plants7 the idea is that the structure of SCP resembles that of LHC II, which is known with atomic resolution from x-ray crystallography8,9. LHC II is a heterotrimer and each monomer binds 8 Chl a, 6 Chl b, and 4 molecules of carotenoid (Car), namely two non-equivalent Luteins (Lut), one 9’-cis neoxanthin (Neo), and one Violaxanthin (Vio) or Zeaxanthin (Zea)9. For SCP it is known that is forms as well a trimeric complex that accommodates Chl a, Chl b, and that the two Lut molecules are replaced by siphonaxan- thin (Sx) and its esterified variant siphonein (Sn), respectively, and 9’-cis neoxanthin (Neo). Recently, a ratio of 1Spectroscopy of Soft Matter, University of Bayreuth, 95440 Bayreuth, Germany. 2Department of Food Technology, Universitas Ciputra, Citraland CBD Boulevard, Surabaya 60219, Indonesia. 3Graduate School of Science, Osaka City University, 3‑3‑138 Sugimoto, Sumiyoshi‑ku, Osaka 558‑8585, Japan. 4Research Center for Artificial Photosynthesis, Osaka Metropolitan University, 3‑3‑138 Sugimoto, Sumiyoshi‑ku, Osaka 558‑8585, Japan. 5Bavarian Polymer Institute, University of Bayreuth, 95440 Bayreuth, Germany. 6Bayreuther Institut für Makromolekülforschung (BIMF), 95440 Bayreuth, Germany. email: juergen.koehler@uni-bayreuth.de \| https://doi.org/10.1038/s41598-022-11572-3 Scientific Reports \| (2022) 12:8461 www.nature.com/scientificreports/ Figure 1. High-resolution structure of a monomer of LHC II from spinach [PDB:1RWT]. (A, C) Side view of the arrangement of the Chl molecules with (A) and without (C) the protein backbone. (B, D) Top view of the arrangement of the Chl molecules with (B) and without (D) the protein backbone. Chl a molecules are shown in green, Chl b molecules are show in blue. For clarity the carotenoids have been omitted. For SCP the Chl a molecules in position 602, and either those in positions 610 or 612 are exchanged to Chl b as indicated by the cyan colour. The Chl molecules are numbered according to the scheme given in9 (1RWT). The figure has been produced using PyMOL ver. 2.5.0. Figure 1. High-resolution structure of a monomer of LHC II from spinach [PDB:1RWT]. (A, C) Side view of the arrangement of the Chl molecules with (A) and without (C) the protein backbone. (B, D) Top view of the arrangement of the Chl molecules with (B) and without (D) the protein backbone. Chl a molecules are shown in green, Chl b molecules are show in blue. For clarity the carotenoids have been omitted. Results Using τAv = i Aiτi i Ai , yields 3.6 ns for the amplitude averaged lifetime.f i Ai In order to diminish bleaching effects, the experiments on single SCP complexes were conducted under oxygen-free conditions in a vessel that was first flushed with gaseous Argon to remove residual air and then kept under vacuum. In contrast to the experiments reported in29 that were conducted using an electrokinetic trap ensuring a well-defined orientation of the pigment-protein complexes, in our experiments the three-dimensional spatial orientation of the SCP complexes is not controlled. Hence, for each individual complex the projection of the absorbing transition-dipole moment on the electric field vector of the excitation light is different leaving the effective excitation intensity unknown. Because of this ambiguity, the emitted intensity is not a reliable parameter for comparing different individual SCP complexes with each other. Instead, we recorded for each individual SCP complex synchronously the fluorescence lifetime and the emission spectrum as a function of time. For doing so, the spectra were read out consecutively every 5.1 s, and the lifetimes were read out consecutively every 500 ms. From these experiments the spectral peak positions of the emission spectra were extracted by fitting a Gaussian to the main peak (see SI). For having the same statistics for the complexes that were studied under different preparation conditions the observation time for each individual complex was restricted to 20 s. If not stated otherwise the accuracy for the lifetimes is limited by the instrument response which amounts to 250 ps, and the accuracy for the spectral peak positions amounts to 0.5 nm (11 cm−1).h y p p p ( ) The data extracted will be displayed in a two-dimensional "lifetimes versus spectral peak positions" diagram, where the horizontal axis corresponds to the spectral peak position and the vertical axis to the fluorescence life- time, see Fig. 2E. For obtaining these diagrams, periods of both constant peak position and constant lifetime con- tribute one data point as pointed out in Fig. 2D. The example in Fig. 2D shows on the left-hand side the lifetimes and spectral peak positions as a function of time that have been extracted from an experiment on a single SCP complex. All fluorescence decays were compatible with monoexponentials. (For an explanation that addresses the discrepancy between the monoexponential decays of the single complexes and the non-monoexponential decays observed for the ensembles, see SI). Results Samples of unaggregated trimeric SCP complexes dissolved in buffer solution (concentration of 1.45 × 10–7 M) have been prepared from fresh samples defrosted on ice under ambient conditions (protocol 1). The correspond- ing room-temperature absorption spectrum is shown in Fig. 2A by the black line. It features strong peaks in the blue spectral region at 439 nm and 473 nm, and two weaker bands in the red spectral region peaking at 652 nm and 672 nm. The band at 473 nm can be attributed to both the carotenoids and the Chl b Soret band, whereas the band at 439 nm is assigned to the absorptions of the Chl a/b Soret bands4,32. The maxima in the red spectral region correspond to the absorptions of the Qy transitions of Chl b and Chl a, respectively. In order to avoid selective excitation of specific Chl a/b sites in the SCP complexes we have chosen 561 nm as the excitation wavelength, because this is outside the 650 nm spectral range of the excitonically coupled Chl molecules. For the experiments on single SCP complexes these were immobilised in PVA, and the experiments were carried out at an excita- tion intensity of 525 W/cm2 for obtaining a reasonable signal-to-noise ratio. For better comparison the same excitation intensity has been used also for all experiments on ensembles of SCP complexes immobilised in PVA. y p p In Fig. 2B the ensemble emission spectra from a solution of SCP complexes (full black line) and from SCP complexes immobilized in a PVA matrix (dashed black line) are compared. Despite the large difference in the excitation intensity used for recording the two spectra, i.e. 30 μW/cm2 for the SCP complexes in solution versus 525 W/cm2 for those in PVA, both spectra are identical in shape and feature a pronounced peak at 680.0 nm (corresponding to 14,705 cm−1; FWHM 383 cm−1) accompanied by a low energy wing that extends from 700 to 760 nm. The comparison of the two spectra testifies that the spectral profile of the emission spectrum is neither affected by the immobilization of the SCP complexes in PVA nor by the difference in excitation intensity. The observed transient of an ensemble of SCP complexes embedded in PVA, Fig. 2C, is consistent with a biexponen- tial decay featuring lifetime components of 4 ns, and 1.2 ns with amplitudes of 85%, and 15%, respectively. Tatas Hardo Panintingjati Brotosudarmo1,2, Bernd Wittmann1, Soichiro Seki3, Ritsuko Fujii3,4 & Jürgen Köhler1,5,6 On the other hand, experiments under ambient conditions are closer to the natural environment and provide valuable information about the conformational dynamics of the pigment- protein complexes. In particular, a single protein that undergoes conformational transitions is at any time in a Scientific Reports \| (2022) 12:8461 \| https://doi.org/10.1038/s41598-022-11572-3 www.nature.com/scientificreports/ distinct, well-defined state in contrast to an ensemble of proteins due to the lack of synchronization31. To the best of our knowledge results from optical spectroscopy on single SCP complexes have not been published to date. Here we report about spectroscopy on ensembles and single trimeric SCP complexes at room temperature for three different protocols for handling the samples. The first protocol (protocol 1) refers to fresh samples prepared on ice under ambient conditions, the second protocol (protocol 2) refers to SCP complexes from the same batch that have been stored for 7 months at − 80 °C in the dark, and that are prepared under the same conditions as before. The last protocol (protocol 3) refers to experiments where the samples from the 7 months old SCP com- plexes have been prepared in a cold room at 5 °C. While the absorption and emission spectra from ensembles of SCP complexes in solution featured a high degree of similarity for the three protocols, the single complex approach revealed strong differences in the spectral details as a function of the sample preparation history. Results The excitation wavelength is 561 nm and the excitation intensity amounts to 525 W/cm2. (D) Fluorescence lifetimes and spectral peak positions of a single SCP complex. The fluorescence lifetime (scale on the right) was read out every 500 ms and is given by the open dots. The drop out after 4 s presumably results from unresolved blinking. The bin time for recording the spectra was 5.1 s (full line) and the observed spectral peak position of the accumulated spectrum is given by the black dot (scale on the left). The sequence of emission spectra is shown on the right-hand side from top to bottom. For better comparison, the broken vertical line refers to the spectral peak position of the ensemble emission spectrum. (E) Two-dimensional representation of the fluorescence lifetime vs. spectral peak position of individual SCP complexes. Due to fluctuations of both lifetime and/or spectral peak position a single complex can contribute more than one data point to the diagram. The number of individual complexes that contribute to the pattern is given in the top left corner. The histograms correspond to the distributions of the spectral peak position (top) and fluorescence lifetime (right). For ease of comparison an ensemble emission spectrum has been overlaid in the top histogram. The statistical parameters of the two histograms (mean ± sdev) are (14,723 ± 109) cm−1 for the peak positions, and (3.4 ± 0.8) ns for the lifetimes, respectively. Figure 2. Spectroscopy on ensembles and single SCP complexes. The samples for spectroscopy have been prepared under ambient conditions (protocol 1). (A) Room temperature absorption spectrum of an ensemble of SCP complexes dissolved in bulk buffer solution at a concentration of 1.45 × 10–7 M. (B) Peak normalised emission spectra from the bulk buffer solution for an excitation intensity of 30 μW/cm2 (full line) and from an ensemble of SCP complexes embedded in a thin film of PVA (dashed line) for an excitation intensity of 525 W/cm2. The excitation wavelength for recording the emission spectra was 561 nm as indicated by the green arrow in (A). (C) Fluorescence lifetime of an ensemble of SCP complexes immobilized in PVA in semi- logarithmic representation. The full line corresponds to a biexponential fit (see residuals at the bottom) with decay components (amplitudes) of 4.0 ns (85%) and 1.2 ns (15%). The excitation wavelength is 561 nm and the excitation intensity amounts to 525 W/cm2. Results (D) Fluorescence lifetimes and spectral peak positions of a single SCP complex. The fluorescence lifetime (scale on the right) was read out every 500 ms and is given by the open dots. The drop out after 4 s presumably results from unresolved blinking. The bin time for recording the spectra was 5.1 s (full line) and the observed spectral peak position of the accumulated spectrum is given by the black dot (scale on the left). The sequence of emission spectra is shown on the right-hand side from top to bottom. For better comparison, the broken vertical line refers to the spectral peak position of the ensemble emission spectrum. (E) Two-dimensional representation of the fluorescence lifetime vs. spectral peak position of individual SCP complexes. Due to fluctuations of both lifetime and/or spectral peak position a single complex can contribute more than one data point to the diagram. The number of individual complexes that contribute to the pattern is given in the top left corner. The histograms correspond to the distributions of the spectral peak position (top) and fluorescence lifetime (right). For ease of comparison an ensemble emission spectrum has been overlaid in the top histogram. The statistical parameters of the two histograms (mean ± sdev) are (14,723 ± 109) cm−1 for the peak positions, and (3.4 ± 0.8) ns for the lifetimes, respectively. standard deviation of 109 cm−1. The average peak position is in close agreement with the 14,705 cm−1 that was found for the peak of the ensemble spectrum. The distribution of the fluorescence lifetimes is relatively broad, and the statistical parameters amount to 3.4 ns ± 0.8 ns (mean ± sdev), also in good agreement with the average lifetime of 3.6 ns obtained for a large ensemble of SCP complexes. standard deviation of 109 cm−1. The average peak position is in close agreement with the 14,705 cm−1 that was found for the peak of the ensemble spectrum. The distribution of the fluorescence lifetimes is relatively broad, and the statistical parameters amount to 3.4 ns ± 0.8 ns (mean ± sdev), also in good agreement with the average lifetime of 3.6 ns obtained for a large ensemble of SCP complexes. Similar experiments have been conducted on SCP complexes from the same batch that has been stored for 7 months at − 80 °C in the dark, and that were prepared under the same conditions as before (protocol 2). Results For the first 4 s the fluorescence lifetime amounts to about 4.5 ns before it drops to about 3.4 ns for the remaining observation time. The sharp minimum at about 4 s is attributed to blinking processes that are faster than the 5.1 s bin time for reading out the spectra. For this particular SCP complex the first value is slightly longer than the 3.6 ns that have been found for the average ensemble lifetime. On the right-hand side of Fig. 2D the emission spectra that have been accumulated during the four time intervals are shown. For all spectra the spectral peak positions of 14,684 cm−1 are close to the ensemble value and show only little variation as a function of time. For this particular complex we find in total one peak position and two lifetimes, and it therefore contributes two data points to the aforementioned two-dimensional lifetime vs. spectral peak position diagram in Fig. 2E. The two-dimensional diagram with contributions from 31 individual complexes is shown in Fig. 2E, together with the distributions of the spectral peak positions and the fluorescence lifetimes on top of and next to the pattern. The spectral peak positions observed for the individual complexes show only little variation and the corresponding distribution is characterized by a mean of 14,723 cm−1 and a https://doi.org/10.1038/s41598-022-11572-3 Scientific Reports \| (2022) 12:8461 \| www.nature.com/scientificreports/ standard deviation of 109 cm−1. The average peak position is in close agreement with the 14,705 cm−1 that was Figure 2. Spectroscopy on ensembles and single SCP complexes. The samples for spectroscopy have been prepared under ambient conditions (protocol 1). (A) Room temperature absorption spectrum of an ensemble of SCP complexes dissolved in bulk buffer solution at a concentration of 1.45 × 10–7 M. (B) Peak normalised emission spectra from the bulk buffer solution for an excitation intensity of 30 μW/cm2 (full line) and from an ensemble of SCP complexes embedded in a thin film of PVA (dashed line) for an excitation intensity of 525 W/cm2. The excitation wavelength for recording the emission spectra was 561 nm as indicated by the green arrow in (A). (C) Fluorescence lifetime of an ensemble of SCP complexes immobilized in PVA in semi- logarithmic representation. The full line corresponds to a biexponential fit (see residuals at the bottom) with decay components (amplitudes) of 4.0 ns (85%) and 1.2 ns (15%). Results The filled area shows the emission spectrum from a sample of fresh SCP complexes embedded in PVA. Both spectra have been peak normalised for better comparison. The excitation wavelength for recording the emission spectrum was 561 nm as indicated by the green arrow in (A). (C) Fluorescence lifetime of an ensemble of the 7 months old SCP complexes immobilized in PVA in semi- logarithmic representation. The full line corresponds to a biexponential fit (see residuals at the bottom) with decay components (amplitudes) of 4.3 ns (33%) and 1.2 ns (67%). The excitation wavelength is 561 nm, and the excitation intensity amounts to 525 W/cm2. (D) Fluorescence lifetimes and spectral peak positions of a single SCP complex. The fluorescence lifetime (scale on the right) was read out every 500 ms and is given by the open dots. The dropouts after 2.5 s and 14 s presumably result from unresolved blinking. The bin time for recording the spectra was 5.1 s (full line) and the observed spectral peak position of the accumulated spectrum is given by the black dot (scale on the left). The sequence of emission spectra is shown on the right-hand side from top to bottom. For better comparison, the broken vertical line refers to the spectral peak position of the ensemble emission spectrum. (E) Two-dimensional representation of the fluorescence lifetime vs. spectral peak position of individual SCP complexes. Due to fluctuations of both lifetime and/or spectral peak position a single complex can contribute more than one data point to the diagram. The number of individual complexes that contribute to the pattern is given in the top left corner. The histograms correspond to the distributions of the spectral peak position (top) and fluorescence lifetime (right). For ease of comparison an ensemble emission spectrum from the fresh sample embedded in PVA has been overlaid in the top histogram. The statistical parameters of the two histograms (mean ± sdev) are (14,319 ± 178) cm−1 for the peak positions, and (2.0 ± 0.5) ns for the lifetimes, respectively. Figure 3. Spectroscopy on ensembles and single SCP complexes that have been stored for 7 months at − 80 °C in the dark. The samples for spectroscopy have been prepared under ambient conditions (protocol 2). Results (A) Room temperature absorption spectrum of an ensemble of SCP that has been stored for 7 months at − 80 °C in the dark dissolved in bulk buffer solution at a concentration of 1.45 × 10–7 M (black line). For comparison the corresponding absorption spectrum from the fresh sample (cf. Fig. 2A) has been underlaid as filled spectrum. Both spectra have been peak normalised at 476 nm. (B) Emission spectrum from SCP complexes embedded in PVA for an excitation intensity of 525 W/cm2. The filled area shows the emission spectrum from a sample of fresh SCP complexes embedded in PVA. Both spectra have been peak normalised for better comparison. The excitation wavelength for recording the emission spectrum was 561 nm as indicated by the green arrow in (A). (C) Fluorescence lifetime of an ensemble of the 7 months old SCP complexes immobilized in PVA in semi- logarithmic representation. The full line corresponds to a biexponential fit (see residuals at the bottom) with decay components (amplitudes) of 4.3 ns (33%) and 1.2 ns (67%). The excitation wavelength is 561 nm, and the excitation intensity amounts to 525 W/cm2. (D) Fluorescence lifetimes and spectral peak positions of a single SCP complex. The fluorescence lifetime (scale on the right) was read out every 500 ms and is given by the open dots. The dropouts after 2.5 s and 14 s presumably result from unresolved blinking. The bin time for recording the spectra was 5.1 s (full line) and the observed spectral peak position of the accumulated spectrum is given by the black dot (scale on the left). The sequence of emission spectra is shown on the right-hand side from top to bottom. For better comparison, the broken vertical line refers to the spectral peak position of the ensemble emission spectrum. (E) Two-dimensional representation of the fluorescence lifetime vs. spectral peak position of individual SCP complexes. Due to fluctuations of both lifetime and/or spectral peak position a single complex can contribute more than one data point to the diagram. The number of individual complexes that contribute to the pattern is given in the top left corner. The histograms correspond to the distributions of the spectral peak position (top) and fluorescence lifetime (right). For ease of comparison an ensemble emission spectrum from the fresh sample embedded in PVA has been overlaid in the top histogram. Results The room-temperature absorption spectrum from a solution of these complexes is shown in Fig. 3A by the black line. For reference the absorption spectrum obtained from the fresh sample is underlaid as a filled spectrum. Apart from very small changes below 400 nm the absorption spectra obtained for the different protocols are identical. In Fig. 3B the full line corresponds to the emission spectrum from an ensemble of these SCP complexes https://doi.org/10.1038/s41598-022-11572-3 Scientific Reports \| (2022) 12:8461 \| www.nature.com/scientificreports/ scientificreports/ embedded in PVA, which is compared with the emission spectrum obtained from the fresh sample that is again Figure 3. Spectroscopy on ensembles and single SCP complexes that have been stored for 7 months at − 80 °C in the dark. The samples for spectroscopy have been prepared under ambient conditions (protocol 2). (A) Room temperature absorption spectrum of an ensemble of SCP that has been stored for 7 months at − 80 °C in the dark dissolved in bulk buffer solution at a concentration of 1.45 × 10–7 M (black line). For comparison the corresponding absorption spectrum from the fresh sample (cf. Fig. 2A) has been underlaid as filled spectrum. Both spectra have been peak normalised at 476 nm. (B) Emission spectrum from SCP complexes embedded in PVA for an excitation intensity of 525 W/cm2. The filled area shows the emission spectrum from a sample of fresh SCP complexes embedded in PVA. Both spectra have been peak normalised for better comparison. The excitation wavelength for recording the emission spectrum was 561 nm as indicated by the green arrow in (A). (C) Fluorescence lifetime of an ensemble of the 7 months old SCP complexes immobilized in PVA in semi- logarithmic representation. The full line corresponds to a biexponential fit (see residuals at the bottom) with decay components (amplitudes) of 4.3 ns (33%) and 1.2 ns (67%). The excitation wavelength is 561 nm, and the excitation intensity amounts to 525 W/cm2. (D) Fluorescence lifetimes and spectral peak positions of a single SCP complex. The fluorescence lifetime (scale on the right) was read out every 500 ms and is given by the open dots. The dropouts after 2.5 s and 14 s presumably result from unresolved blinking. The bin time for recording the spectra was 5.1 s (full line) and the observed spectral peak position of the accumulated spectrum is given by the black dot (scale on the left). Results The sequence of emission spectra is shown on the right-hand side from top to bottom. For better comparison, the broken vertical line refers to the spectral peak position of the ensemble emission spectrum. (E) Two-dimensional representation of the fluorescence lifetime vs. spectral peak position of individual SCP complexes. Due to fluctuations of both lifetime and/or spectral peak position a single complex can contribute more than one data point to the diagram. The number of individual complexes that contribute to the pattern is given in the top left corner. The histograms correspond to the distributions of the spectral peak position (top) and fluorescence lifetime (right). For ease of comparison an ensemble emission spectrum from the fresh sample embedded in PVA has been overlaid in the top histogram. The statistical parameters of the two histograms (mean ± sdev) are (14,319 ± 178) cm−1 for the peak positions, and (2.0 ± 0.5) ns for the lifetimes, respectively. 5 Vol.:(012345 Scientific Reports \| (2022) 12:8461 \| https://doi.org/10.1038/s41598-022-11572-3 embedded in PVA, which is compared with the emission spectrum obtained from the fresh sample that is again underlaid as a filled spectrum. In contrast, to the absorption spectrum the ensemble emission spectrum from the 7 months old sample is red shifted with respect to the initial spectrum and peaks at 686.6 nm (14,564 cm−1). Moreover, its width (FWHM) is increased to 440 cm−1, and the relative intensity of the low-energy shoulder in the 710 nm to 760 nm spectral range has risen significantly. The decay of the fluorescence from the ensemble is again consistent with a biexponential with lifetimes (amplitudes) of 4.3 ns (33%), and 1.2 ns (67%). While the Figure 3. Spectroscopy on ensembles and single SCP complexes that have been stored for 7 months at − 80 °C in the dark. The samples for spectroscopy have been prepared under ambient conditions (protocol 2). (A) Room temperature absorption spectrum of an ensemble of SCP that has been stored for 7 months at − 80 °C in the dark dissolved in bulk buffer solution at a concentration of 1.45 × 10–7 M (black line). For comparison the corresponding absorption spectrum from the fresh sample (cf. Fig. 2A) has been underlaid as filled spectrum. Both spectra have been peak normalised at 476 nm. (B) Emission spectrum from SCP complexes embedded in PVA for an excitation intensity of 525 W/cm2. www.nature.com/scientificreports/ The room temperature absorption spectrum from a solution, Fig. 4A, resembles closely the one obtained from the fresh sample, also for the spectral range below 400 nm Fig. 4A. Interestingly, this holds true as well for the emission spectrum that features a peak position at 679.6 nm (14,715 cm−1) and a width (FWHM) of 395 cm−1 Fig. 4B, and which reproduces the spectral profile observed for the fresh sample. Similarly for the fluorescence decay with lifetime components (amplitudes) of 4.0 ns (79%), and 1.3 ns (21%). The relative weight of the two decay components is nearly restored to the values observed for the fresh SCP complexes. Accordingly, the averaged lifetime amounts to 3.4 ns. The fluorescence lifetime from the single complex shown in Fig. 4D amounts to 3.2 ns for the first 9 s and then drops to 2.0 ns for the remaining duration of the experiment. During this time the spectral peak position does not show significant variations and is centred around 14,735 cm−1 corresponding to a slight blue shift with respect to the ensemble value. As before such a complex contributes 2 data points to the respective two-dimensional diagram shown in Fig. 4E.h g g The statistical parameters for the full dataset with contributions from 26 individual SCP complexes are 14,537 cm−1 ± 181 cm−1 for the peak positions and 2.8 ns ± 0.9 ns for the fluorescence lifetimes. However, closer inspection of the two-dimensional representation of the data suggests that this diagram represents a superposi- tion of the two patterns observed before for the fresh SCP complexes and for those that have been stored for 7 months in the cold, cf. Figs. 2 and 3. This is corroborated by the separate evaluation of the statistics of the two clusters of data points that are indicated by the blue and red-shaded areas in Fig. 4E. For convenience these will be referred to as the blue and the red cluster in the following. For the data in the blue cluster, the distribution of the peak position is characterised by 14,715 cm−1 ± 60 cm−1 and that for the lifetimes by 3.4 ns ± 0.9 ns. These numbers are both in very close agreement with the parameters found for the fresh sample. The corresponding parameters for the red cluster amount to 14,382 cm−1 ± 77 cm−1 for the peak positions and 2.1 ns ± 0.4 ns for the lifetimes. www.nature.com/scientificreports/ decay times are about the same as those observed before, their relative weight has shifted towards the short-lived component which is reflected in an averaged lifetime of 2.2 ns. A tentative explanation for this observation is to consider two prevailing conformations of the complexes, each associated with one of the lifetime components. Variations of the population ratio of these conformations will directly impact on the corresponding amplitudes of the two decay components, yet without affecting the time constants.hi f An example for a single complex is shown in Fig. 3C. The setup of the figure is similar to the setup of Fig. 2C. For this complex the fluorescence lifetime amounts to 1.6 ns during the whole observation time, which is even shorter than the average lifetime of 2.2 ns found for an ensemble under the same conditions. The spectral peak position shows a significant shift from about 14,430 cm−1 during the first 10 s to about 14,120 cm−1 during the last 20 s, which both are red shifted with respect to the ensemble value. Hence, for this complex we find two peak positions and one lifetime, and this complex contributes two data points to the corresponding two-dimensional lifetime vs. peak position diagram shown in Fig. 3E. For the SCP complexes that have been prepared accord- ing to protocol 2 the "peak position versus lifetime" diagram, and the resulting histograms with contributions from 31 individual complexes are shown in Fig. 3E. The distribution of the spectral peak positions does not show a pronounced maximum but is red shifted and clearly broader than the histogram obtained from the fresh complexes. This is also reflected by the statistical parameters of 14,319 cm−1 ± 178 cm−1. Interestingly, now the histogram for the fluorescence lifetimes is relatively narrow and shifted towards shorter decay times. The statisti- cal parameters for this distribution amount to 2.0 ns ± 0.5 ns, which is in the range of the ensemble average for this preparation protocol. Finally, experiments have been conducted on the 7 months old SCP complexes where all preparation steps including spin coating as well as flushing and evacuation of the sample chamber took place in a cold room at 5 °C (protocol 3). The experiments were carried out immediately after the preparation of the sample. The experimental results are summarized in Fig. 4 which has a similar layout as Figs. 2 and 3. Results The statistical parameters of the two histograms (mean ± sdev) are (14,319 ± 178) cm−1 for the peak positions, and (2.0 ± 0.5) ns for the lifetimes, respectively. embedded in PVA, which is compared with the emission spectrum obtained from the fresh sample that is again underlaid as a filled spectrum. In contrast, to the absorption spectrum the ensemble emission spectrum from the 7 months old sample is red shifted with respect to the initial spectrum and peaks at 686.6 nm (14,564 cm−1). Moreover, its width (FWHM) is increased to 440 cm−1, and the relative intensity of the low-energy shoulder in the 710 nm to 760 nm spectral range has risen significantly. The decay of the fluorescence from the ensemble is again consistent with a biexponential with lifetimes (amplitudes) of 4.3 ns (33%), and 1.2 ns (67%). While the https://doi.org/10.1038/s41598-022-11572-3 Scientific Reports \| (2022) 12:8461 \| www.nature.com/scientificreports/ www.nature.com/scientificreports/ The number of individual complexes that contribute to the pattern is given in the top left corner. The histograms correspond to the distributions of the spectral peak position (top) and fluorescence lifetime (right). For ease of comparison an emission spectrum from a sample of fresh SCP complexes embedded in PVA has been overlaid in the top histogram. The statistical parameters of the two histograms (mean ± sdev) are (14,537 ± 181) cm−1 for the peak positions, and (2.8 ± 0.9) ns for the lifetimes, respectively. Figure 4. Spectroscopy on ensembles and single SCP complexes that have been stored for 7 months at − 80 °C in the dark. The samples for spectroscopy have been prepared in a cold room at 5 °C (protocol 3). (A) Room temperature absorption spectrum of an ensemble of SCP dissolved in bulk buffer solution at a concentration of 1.45 × 10–7 M (black line). For comparison the corresponding absorption spectrum from the fresh sample (cf. Fig. 2A) has been underlaid as filled spectrum. Both spectra have been peak normalised at 476 nm. (B) Emission spectrum from SCP complexes embedded in PVA for an excitation intensity of 525 W/cm2. The filled area shows the emission spectrum from a sample of fresh SCP complexes embedded in PVA. Both spectra have been peak normalised for better comparison. The excitation wavelength for recording the emission spectrum was 561 nm as indicated by the green arrow in (A). (C) Fluorescence lifetime of an ensemble of SCP complexes immobilized in PVA in semi-logarithmic representation. The full line corresponds to a biexponential fit (see residuals at the bottom) with decay components (amplitudes) of 4.0 ns (79%) and 1.3 ns (21%). The excitation wavelength is 561 nm, and the excitation intensity amounts to 525 W/cm2. (D) Fluorescence lifetimes and spectral peak positions of a single SCP complex. The fluorescence lifetime (scale on the right) was read out every 500 ms and is given by the open dots. The bin time for recording the spectra was 5.1 s (full line) and the observed spectral peak position of the accumulated spectrum is given by the black dot (scale on the left). The sequence of emission spectra is shown on the right-hand side from top to bottom. For better comparison, the broken vertical line refers to the spectral peak position of the ensemble emission spectrum. (E) Two- dimensional representation of the fluorescence lifetime vs. www.nature.com/scientificreports/ The mean for the lifetimes is close to the value found for SCP complexes that were prepared according to protocol 2, whereas the mean for the spectral peak positions is more shifted to the red. However, the most intriguing observation in the experiments on these SCP complexes was that the classification whether a result felt into the blue or red cluster correlated with the duration of the total experiment. Starting the experiment on a sample immediately after the preparation in the cold room yielded spectral peak positions and fluorescence lifetimes that belonged to the blue cluster, whereas after about 120 min duration of the experiments the observed combination of spectral peak positions and fluorescence lifetimes felt into the red cluster. Jumps of individual complexes between the two clusters were not observed. The experimental results for the three protocols are summarized in Table 1.h The experiments on single SCP complexes were carried out using an excitation intensity that exceeds by far the (spectrally integrated) intensity of less than 100 μW/cm2 from the sun on a bright day in the natural habitat of the algae. In order to test whether the observed changes of the spectral signatures correlate with the excitation intensity it would be desirable to repeat the experiments using a lower excitation intensity. Due to signal-to-noise restrictions this, however, is only possible for ensembles of SCP complexes. In Fig. 5 the ensemble emission spec- tra for the various protocols are compared as a function of the excitation intensity. At high excitation intensity, Fig. 5A, the spectrum recorded according to protocol 3 initially reproduces the "protocol 1 like" spectrum (dashed black line), yet changes into the "protocol 2 like" spectrum (red dashed line) after a waiting time (without illu- mination) of about 2 h. These observations are consistent with the results obtained on single SCP complexes for the different protocols. At lower excitation intensities, Fig. 5B, the emission spectra for all protocols are identical, and moreover reproduce the emission spectrum obtained for protocol 1 at high excitation intensity. This suggests that the formation of the red-shifted quenched state correlates with the excitation intensity. https://doi.org/10.1038/s41598-022-11572-3 Scientific Reports \| (2022) 12:8461 \| www.nature.com/scientificreports/ ientificreports/ Di i Figure 4. Spectroscopy on ensembles and single SCP complexes that have been stored for 7 months at − 80 °C in the dark. www.nature.com/scientificreports/ The samples for spectroscopy have been prepared in a cold room at 5 °C (protocol 3). (A) Room temperature absorption spectrum of an ensemble of SCP dissolved in bulk buffer solution at a concentration of 1.45 × 10–7 M (black line). For comparison the corresponding absorption spectrum from the fresh sample (cf. Fig. 2A) has been underlaid as filled spectrum. Both spectra have been peak normalised at 476 nm. (B) Emission spectrum from SCP complexes embedded in PVA for an excitation intensity of 525 W/cm2. The filled area shows the emission spectrum from a sample of fresh SCP complexes embedded in PVA. Both spectra have been peak normalised for better comparison. The excitation wavelength for recording the emission spectrum was 561 nm as indicated by the green arrow in (A). (C) Fluorescence lifetime of an ensemble of SCP complexes immobilized in PVA in semi-logarithmic representation. The full line corresponds to a biexponential fit (see residuals at the bottom) with decay components (amplitudes) of 4.0 ns (79%) and 1.3 ns (21%). The excitation wavelength is 561 nm, and the excitation intensity amounts to 525 W/cm2. (D) Fluorescence lifetimes and spectral peak positions of a single SCP complex. The fluorescence lifetime (scale on the right) was read out every 500 ms and is given by the open dots. The bin time for recording the spectra was 5.1 s (full line) and the observed spectral peak position of the accumulated spectrum is given by the black dot (scale on the left). The sequence of emission spectra is shown on the right-hand side from top to bottom. For better comparison, the broken vertical line refers to the spectral peak position of the ensemble emission spectrum. (E) Two- dimensional representation of the fluorescence lifetime vs. spectral peak position of individual SCP complexes. Due to fluctuations of both lifetime and/or spectral peak position a single complex can contribute more than one data point to the diagram. For the distinction between the data points in the red and blue shaded areas see text. The number of individual complexes that contribute to the pattern is given in the top left corner. The histograms correspond to the distributions of the spectral peak position (top) and fluorescence lifetime (right). For ease of comparison an emission spectrum from a sample of fresh SCP complexes embedded in PVA has been overlaid in the top histogram. www.nature.com/scientificreports/ spectral peak position of individual SCP complexes. Due to fluctuations of both lifetime and/or spectral peak position a single complex can contribute more than one data point to the diagram. For the distinction between the data points in the red and blue shaded areas see text. The number of individual complexes that contribute to the pattern is given in the top left corner. The histograms correspond to the distributions of the spectral peak position (top) and fluorescence lifetime (right). For ease of comparison an emission spectrum from a sample of fresh SCP complexes embedded in PVA has been overlaid in the top histogram. The statistical parameters of the two histograms (mean ± sdev) are (14,537 ± 181) cm−1 for the peak positions, and (2.8 ± 0.9) ns for the lifetimes, respectively. www.nature.com/scientificreports/ The statistical parameters of the two histograms (mean ± sdev) are (14,537 ± 181) cm−1 for the peak positions, and (2.8 ± 0.9) ns for the lifetimes, respectively. Figure 4. Spectroscopy on ensembles and single SCP complexes that have been stored for 7 months at − 80 °C in the dark. The samples for spectroscopy have been prepared in a cold room at 5 °C (protocol 3). (A) Room temperature absorption spectrum of an ensemble of SCP dissolved in bulk buffer solution at a concentration of 1.45 × 10–7 M (black line). For comparison the corresponding absorption spectrum from the fresh sample (cf. Fig. 2A) has been underlaid as filled spectrum. Both spectra have been peak normalised at 476 nm. (B) Emission spectrum from SCP complexes embedded in PVA for an excitation intensity of 525 W/cm2. The filled area shows the emission spectrum from a sample of fresh SCP complexes embedded in PVA. Both spectra have been peak normalised for better comparison. The excitation wavelength for recording the emission spectrum was 561 nm as indicated by the green arrow in (A). (C) Fluorescence lifetime of an ensemble of SCP complexes immobilized in PVA in semi-logarithmic representation. The full line corresponds to a biexponential fit (see residuals at the bottom) with decay components (amplitudes) of 4.0 ns (79%) and 1.3 ns (21%). The excitation wavelength is 561 nm, and the excitation intensity amounts to 525 W/cm2. (D) Fluorescence lifetimes and spectral peak positions of a single SCP complex. The fluorescence lifetime (scale on the right) was read out every 500 ms and is given by the open dots. The bin time for recording the spectra was 5.1 s (full line) and the observed spectral peak position of the accumulated spectrum is given by the black dot (scale on the left). The sequence of emission spectra is shown on the right-hand side from top to bottom. For better comparison, the broken vertical line refers to the spectral peak position of the ensemble emission spectrum. (E) Two- dimensional representation of the fluorescence lifetime vs. spectral peak position of individual SCP complexes. Due to fluctuations of both lifetime and/or spectral peak position a single complex can contribute more than one data point to the diagram. For the distinction between the data points in the red and blue shaded areas see text. www.nature.com/scientificreports/ The excitation intensity at 561 nm was always 525 W/cm2. Ensembles Protocol 1 Protocol 2 Protocol 3 Spectral peak position (cm−1) 14,705 cm−1 14,564 cm−1 14,715 cm−1 Spectral peak position (nm) 680.0 686.6 679.6 FWHM (cm−1) 383 440 395 Lifetimes (ns) (amplitudes) 4.0 (85%) 1.2 (15%) 4.3 (33%) 1.2 (67%) 4.0 (79%) 1.3 (21%) Averaged lifetime (ns) 3.6 2.2 3.4 Single complexes Protocol 3 blue cluster Protocol 3 red cluster Spectral peak position mean ± sdev (cm−1) 14,723 ± 109 14,319 ± 178 14,537 ± 180 14,715 ± 60 14,382 ± 77 Mean of spectral peak position (nm) 679.2 698.4 687.9 679.6 695.3 Lifetime mean ± sdev (ns) 3.4 ± 0.8 2.0 ± 0.5 2.8 ± 0.9 3.4 ± 0.9 2.1 ± 0.4 Table 1. Summary of the spectroscopic results from SCP complexes as a function of the sample preparation conditions. For the ensemble data the spectral peak positions and the widths (FWHM) of the emission spectra are provided. For the single complex data, the means and standard deviations of the respective distributions are given. The excitation intensity at 561 nm was always 525 W/cm2. Table 1. Summary of the spectroscopic results from SCP complexes as a function of the sample preparation conditions. For the ensemble data the spectral peak positions and the widths (FWHM) of the emission spectra are provided. For the single complex data, the means and standard deviations of the respective distributions are given. The excitation intensity at 561 nm was always 525 W/cm2. Figure 5. Emission spectra of ensembles of SCP complexes (A) embedded in PVA for excitation intensities of 525 W/cm2, and (B) dissolved in bulk buffer solution at a concentration of 1.45 × 10–7 M, and excitation intensities of 30 µW/cm2. Key: Dashed black lines—emission spectra for protocol 3; dashed red lines—emission spectra for protocol 3 after a waiting time of 2 h. For ease of comparison, the emission spectra for protocols 1 and 2 are shown by the underlaid filled areas, and the full black lines, respectively. Figure 5. Emission spectra of ensembles of SCP complexes (A) embedded in PVA for excitation intensities of 525 W/cm2, and (B) dissolved in bulk buffer solution at a concentration of 1.45 × 10–7 M, and excitation intensities of 30 µW/cm2. Key: Dashed black lines—emission spectra for protocol 3; dashed red lines—emission spectra for protocol 3 after a waiting time of 2 h. www.nature.com/scientificreports/ www.nature.com/scientificreports/ www.nature.com/scientificreports/ might argue that the relatively high excitation intensity of 525 W/cm2 causes some damage to the SCP complexes. However, we note that at 561 nm this intensity produces about the same number of excited states per unit time as an intensity of 250 W/cm2 at 640 nm, which is a common intensity used for spectroscopy of single LHC com- plexes from green plants26,27,33,34. Since such a red-shifted emission has been observed before for LHC II from Table 1. Summary of the spectroscopic results from SCP complexes as a function of the sample preparation conditions. For the ensemble data the spectral peak positions and the widths (FWHM) of the emission spectra are provided. For the single complex data, the means and standard deviations of the respective distributions are given. The excitation intensity at 561 nm was always 525 W/cm2. Ensembles Protocol 1 Protocol 2 Protocol 3 Spectral peak position (cm−1) 14,705 cm−1 14,564 cm−1 14,715 cm−1 Spectral peak position (nm) 680.0 686.6 679.6 FWHM (cm−1) 383 440 395 Lifetimes (ns) (amplitudes) 4.0 (85%) 1.2 (15%) 4.3 (33%) 1.2 (67%) 4.0 (79%) 1.3 (21%) Averaged lifetime (ns) 3.6 2.2 3.4 Single complexes Protocol 3 blue cluster Protocol 3 red cluster Spectral peak position mean ± sdev (cm−1) 14,723 ± 109 14,319 ± 178 14,537 ± 180 14,715 ± 60 14,382 ± 77 Mean of spectral peak position (nm) 679.2 698.4 687.9 679.6 695.3 Lifetime mean ± sdev (ns) 3.4 ± 0.8 2.0 ± 0.5 2.8 ± 0.9 3.4 ± 0.9 2.1 ± 0.4 Figure 5. Emission spectra of ensembles of SCP complexes (A) embedded in PVA for excitation intensities of 525 W/cm2, and (B) dissolved in bulk buffer solution at a concentration of 1.45 × 10–7 M, and excitation intensities of 30 µW/cm2. Key: Dashed black lines—emission spectra for protocol 3; dashed red lines—emission spectra for protocol 3 after a waiting time of 2 h. For ease of comparison, the emission spectra for protocols 1 and 2 are shown by the underlaid filled areas, and the full black lines, respectively. Table 1. Summary of the spectroscopic results from SCP complexes as a function of the sample preparation conditions. For the ensemble data the spectral peak positions and the widths (FWHM) of the emission spectra are provided. For the single complex data, the means and standard deviations of the respective distributions are given. Discussion We have detected the fluorescence lifetimes and the emission spectra from ensembles and single SCP complexes using three different protocols for handling the samples. Taking the data from the fresh sample (protocol 1) as a reference, protocol 2 yields a red shifted ensemble emission spectrum featuring a pronounced low energy shoul- der, and a shortening of the average fluorescence lifetime, whereas these parameters are nearly unchanged for applying protocol 3. Since the results obtained from the fresh samples are reproduced for protocol 3, we conclude that the SCP complexes are still intact immediately after taking them out of the low temperature storage. One https://doi.org/10.1038/s41598-022-11572-3 Scientific Reports \| (2022) 12:8461 \| www.nature.com/scientificreports/ For the 7 months old samples (red line) the barrier height is lowered such that the available thermal energy at room temperature is sufficient to overcome this threshold. Figure 6. Schematic sketch of the highest tier of the protein energy landscape of the SCP protein structure. For the fresh samples (blue line) the states emitting around 680 nm and around 690–700 nm are separated by a high energy barrier. For the 7 months old samples (red line) the barrier height is lowered such that the available thermal energy at room temperature is sufficient to overcome this threshold. Figure 6. Schematic sketch of the highest tier of the protein energy landscape of the SCP protein structure. For the fresh samples (blue line) the states emitting around 680 nm and around 690–700 nm are separated by a high energy barrier. For the 7 months old samples (red line) the barrier height is lowered such that the available thermal energy at room temperature is sufficient to overcome this threshold. Figure 6. Schematic sketch of the highest tier of the protein energy landscape of the SCP protein structure. For the fresh samples (blue line) the states emitting around 680 nm and around 690–700 nm are separated by a high energy barrier. For the 7 months old samples (red line) the barrier height is lowered such that the available thermal energy at room temperature is sufficient to overcome this threshold. However, for a single complex the decay time may fluctuate as a function of time, and for the micro ensemble of individual complexes studied the decay times feature a distribution. Moreover, for protocol 3 the single SCP complexes undergo a change in the course of time from "protocol 1 like" results during the first two hours after sample preparation to "protocol 2 like" results later.l p p p p For LHC II from green plants fluorescence lifetimes between 1.2–1.4 ns and 3.4–3.8 ns with relative contribu- tions between 10–20% for the fast decay time and 80–90% for the slower one, respectively, have been observed before40. The fast component is in very good agreement with the fast component found for SCP, whereas the other component for LHC II is slightly faster with respect to the corresponding lifetime component of SCP. The relative weight of the two contributions, however, closely resembles the partition found for fresh samples of SCP, see results for protocol 1. www.nature.com/scientificreports/ A reduction of the fluorescence lifetime from about 4 ns to 2 ns has been reported for LHC II complexes in thylakoid membranes41 which, however, was not accompanied by a red shift of the emission.t t For single LHC II trimers a reversible switching of the emission between a red-shifted state above 700 nm and a state at 683 nm has been observed before26,28,34. Unfortunately, information about the fluorescence lifetimes of these states is not available. The red-shifted emission was associated with a mixed exciton-charge transfer (CT) state that involves the sites Chl a603, and Chl b609 in close proximity to the lutein 2 molecule, and the appear- ance/disappearance of this band was ascribed to a conformational change within the protein that affects the mutual interactions of these chromophores28,34. Owing to the high degree of homology of the protein structures of SCP and LHC II we hypothesize that in SCP the red-shifted quenched state can be associated as well with structural changes within the protein scaffold that induce the formation of the mixed exciton-charge transfer (CT) state. Given the relatively low fluorescence quantum yield of SCP of 5–10% a large fraction of the average absorbed energy is dissipated by radiationless decay providing a sufficient amount of energy for inducing con- formational fluctuations of the protein backbone. This is consistent with the occurrence of the red-emitting state only for higher excitation intensities, see Fig. 5. However, for using high excitation intensities both ensembles and single SCP complexes prepared according to protocol 3 feature a transition from "protocol 1 like" spectra to "protocol 2 like" spectra after a waiting time of about two hours, which presumably reflects the time that is required for warming up the sample chamber to room temperature. From this we infer that for the 7 months old samples the formation of the red-shifted state requires next to a high excitation intensity in addition some thermal activation, and that the corresponding barrier is in the order of the available thermal energy at room temperature, i.e., about 200 cm−1. Accordingly, the height of this barrier will be significantly higher than this for the fresh sample (protocol 1), because for this scenario the formation of the red-emitting states was not observed despite using high excitation intensities and handling the samples at room temperature prior to the optical experiments. This hypothesis is sketched in Fig. www.nature.com/scientificreports/ For ease of comparison, the emission spectra for protocols 1 and 2 are shown by the underlaid filled areas, and the full black lines, respectively. might argue that the relatively high excitation intensity of 525 W/cm2 causes some damage to the SCP complexes. However, we note that at 561 nm this intensity produces about the same number of excited states per unit time as an intensity of 250 W/cm2 at 640 nm, which is a common intensity used for spectroscopy of single LHC com- plexes from green plants26,27,33,34. Since such a red-shifted emission has been observed before for LHC II from green plants34, and because we do not observe any spectral features that are indicative of free Chlorophyll, namely a fluorescence lifetime of about 5–8 ns35–37 and an emission band around 675 nm38, we exclude that the emission from the red-shifted quenched state is caused by a deterioration of the samples. Moreover, changes in the PVA matrix causing this shift are excluded as well. Firstly, for all three sample preparation scenarios the PVA was added during the last step of the procedure. Hence, any presumed influence of the PVA on the outcome of the experi- ments should be identical for all three scenarios. Secondly, in previous work on single light-harvesting complexes from purple bacteria it was shown that the experimental results obtained from complexes that were embedded in PVA and the results from complexes that were reconstituted into a phospholipid bilayer were equivalent16,39.l For all three protocols the fluorescence transients recorded from ensembles of SCP complexes are consistent with biexponentials with decay times that are identical within experimental accuracy, namely around 4 ns and 1.2 ns. The decrease of the averaged lifetime from 3.6 ns to 2.2 ns for protocol 2 results solely from the variation of the relative contributions of the two lifetime components to the total decay. In contrast, the experiments on single SCP complexes reveal that all fluorescence decays of single complexes were compatible with monoexponentials. https://doi.org/10.1038/s41598-022-11572-3 Scientific Reports \| (2022) 12:8461 \| www.nature.com/scientificreports/ Figure 6. Schematic sketch of the highest tier of the protein energy landscape of the SCP protein structure. For the fresh samples (blue line) the states emitting around 680 nm and around 690–700 nm are separated by a high energy barrier. Materials and methods Cultivation of algae. Cod Subsequently a small droplet (25 μL) of the solution was spin-coated in the dark onto a glass substrate (for 10 s at 500 rpm and 60 s at 2000 rpm) forming a thin polymer film of about 100 nm thickness with embedded SCP complexes. Prior to the spin coating the glass substrate was three times carefully cleaned with acetone and dried with a strong flow of nitrogen gas. The substrate served as window of a vacuum chamber with the sample positioned on the inside which allows to work under oxygen-free conditions. Immediately after mounting the window the vacuum chamber was flushed 3 times with gaseous Argon to remove residual air and then evacuated to 10–3 mbar. Protocol 2 Protocol 2 refers to SCP samples that have been stored at − 80 °C 7–8 months, and that were pre- pared for the spectroscopic experiments following the same procedures as detailed above for protocol 1. g Protocol 3 Protocol 3 refers to SCP samples that have been stored at − 80 °C 7–8 months, but where the whole preparation of the samples for the spectroscopic experiments including spin coating, and mounting of the samples was carried out in a cold room (5 °C) in the dark. After mounting the sample chamber was evacuated, moved to the optical setup, flushed with gaseous Argon and then evacuated again. Optical experiments. For recording the UV/VIS spectra the dissolved SCP complexes at a concentration of 1.45 × 10–7 M were filled into quartz-glass cuvettes (Hellma QS) that was mounted in commercial spectrome- ters (absorption: Perkin Elmer Lambda75, emission: Varian Cary Eclipse spectrometer). The emission spectrum was recorded for an excitation wavelength of 561 nm and an intensity that corresponded to 30 µW/cm2. The single-complex experiments were conducted on a commercial single-molecule optical microscope (MicroTime 200, PicoQuant). The SCP complexes were excited with a laser diode at 561 nm (LHD-D-TA-560B, PicoQuant) operated in pulsed mode with a repetition rate of 20 MHz and an intensity of 525 W/cm2. The output from the laser was coupled into an optical fibre and entered an inverted confocal optical microscope where it was focused with a water immersion objective (60 × UPlanS APO UIS2, NA = 1.2, Olympus) onto the sample. Materials and methods Cultivation of algae. Cod Cultivation of algae. Codium fragile (KU-065, KU-MACC, Kobe, Japan) was cultivated in floating form in Provasoli’s Enriched Sea water (PES) medium as described in42 using artificial sea water (Marine Art SF-1, Tomita Pharmaceutical Co., Ltd, Tokushima, Japan) at 21 °C, and illuminated with a white light LED (AS-010, Fujikura, Osaka, Japan) at 0.5 mW/cm2 with 12 h/12 h light cycles. Protein preparation/purification. Thylakoid membranes were prepared from lyophilized algae accord- ing to45. SCP was purified from the thylakoid membranes as detailed in4 with slight modifications: After solu- bilization, the samples were loaded onto the layer of sucrose density gradient as described in46. The major band was collected and purified by anion exchange chromatography, see4. This step was repeated, and the material was further purified using another sucrose density gradient followed by gel filtration (Sephacryl S-200HR, Merck KGaA, Darmstadt, Germany) with 20 mM Tris–HCl (pH 8.2) containing 0.03% n-dodecyl β-D-maltoside (β-DDM). Purified SCP was flash frozen with liquid nitrogen, stored at − 80 °C, and transported at liquid nitro- gen temperature using Dry Shipper (CX100, Taylor-Wharton, Baytown, Tx, USA). Sample handling. The SCP samples with an optical density of 0.2 at 670 nm were defrosted on an ice bucket and diluted in 25 mM Tris–HCl buffer pH 8.2 containing 0.03% n-dodecyl β-D-maltoside (β-DDM) to give a 1.45 × 10–7 M solution of SCP complexes. For removing oxygen, the buffer was degassed by ultrasonica- tion for 30 min. The solution was immediately divided into aliquots of 10 μL, and either used immediately, or snap-frozen in liquid nitrogen and stored at − 80 °C for later use. All handlings took place in the dark while the samples were kept cold in an ice bucket. p p Protocol 1 Protocol 1 refers to fresh samples. To prepare the sample, the aliquoted SCP solution (1.45 × 10–7 M) was defrosted on an ice bucket. For the experiments on ensembles of SCP complexes embedded in a thin film of polyvinyl alcohol (PVA) about 1 µL from the aliquots was further diluted with the buffer to a volume of 200 µL and 2% of PVA (w/v; mw 124.000–186.000) were added. For the experiments on the single SCP complexes the sample was further diluted with the buffer in three steps to 0.07 pM, and in the last dilution step 2% (w/v) PVA was added. www.nature.com/scientificreports/ 6.hl h The change of the barrier height during storage of the samples could reflect a conformational change that occurs in a part of the protein scaffold such as the formation or the loss of a hydrogen bond, or an isomerization of one of the chromophores or one of the residues in the protein scaffold. As has been shown in42,43 for LHC II and a similar light-harvesting structure very little changes in the liganding of the chromophores will have drastic effects for the photophysical properties of the complexes such as spectral position of the emission and/ or fluorescence lifetimes. Whether these observations are related to the processes that regulate the photoprotec- tion of the photosynthetic machinery44 under strong illumination conditions will need some further attention. Scientific Reports \| (2022) 12:8461 \| https://doi.org/10.1038/s41598-022-11572-3 www.nature.com/scientificreports/ Finally, we note that it is not uncommon to store isolated and purified photosynthetic antennae proteins for some time span at − 80 °C in the dark, and to verify the integrity of the samples by taking UV/VIS spectra. This is usually carried out using low excitation intensities in the order of some 10–100 µW/cm2. However, the results presented here for the light-harvesting complexes of the marine alga Codium fragile show that this might be problematic and that (at least for this particular light-harvesting complex) ensemble UV/VIS spectroscopy might not be sufficient to verify the intactness of the samples. www.nature.com/scientificreports/ transients were deconvoluted with the instrument response function using commercial software (SymPhoTime 64, Picoquant). All experiments have been conducted at room temperature. transients were deconvoluted with the instrument response function using commercial software (SymPhoTime 64, Picoquant). All experiments have been conducted at room temperature. Fluorescence quantum yield measurement. The fluorescence quantum yields were measured and determined according to the procedure described in47. The measurements on SCP were carried out at room temperature using a laser diode (Thorlabs, HL6358MG) operating at 639 nm as excitation light source. For detection a CCD-camera (ANDOR IDUS 420) that was connected to a spectrograph (MS125), which in turn was fibre-coupled to an integrating sphere was used. Received: 1 December 2021; Accepted: 26 April 2022 Received: 1 December 2021; Accepted: 26 April 2022 References g g y 2. Blankenship, R. E. (ed.) Molecular Mechanisms of Photosynthesis (Wiley, 2014). 2. Blankenship, R. E. (ed.) Molecular Mechanisms of Photosynthesis (Wiley, 2014). . Akhtar, P. et al. Spectral tuning of light-harvesting complex II in 3. Akhtar, P. et al. Spectral tuning of light-harvesting complex II in the siphonous alga Bryopsis corticulans and its effect on energy transfer dynamics. Biochim. Biophys. Acta 1861, 148191 (2020). transfer dynamics. Biochim. Biophys. Acta 1861, 148191 (2020) y p y 4. Streckaite, S. et al. Pigment structure in the light-harvesting protein of the siphonous green alga Codium fragile. Biochim. Biophys Acta 1862, 148384 (2021). 5. Nguyen, H. L. et al. An exciton dynamics model of bryopsis corticulans light-harvesting complex II. J. Phys. Chem. B 125, 1134– 1143 (2021). 6. Chen, G. et al. Characterization of chlorophyll-protein complexes isolated from a Siphonous green alga, Bryopsis corticulans Photosynth. Res. 96, 75–81 (2008). y 7. Wang, W. et al. Spectral and functional studies on siphonaxanthin-type light-harvesting complex of photosystem II from Bryopsis corticulans. Photosynth. Res. 117, 267–279 (2013).h y ( ) 8. Kühlbrandt, W. & Wang, D. N. Three-dimensional structure of plant light-harvesting complex determined by electron crystal lography. Nature 350, 130–134 (1991). g p y ( ) 9. Liu, Z. et al. Crystal structure of spinach major light-harvesting complex at 2.72 A resolution. Nature 428, 287–292 (2004) veiling the blue-green light utilization by an oceanic macro green 10. Seki et al. Unveiling the blue-green light utilization by an ocea g g g y g g ( ) 11. Bopp, M. A., Sytnik, A., Howard, T. D., Cogdell, R. J. & Hochstrasser, R. M. The dynamics of structural deformations of immobilized single light-harvesting complexes. Proc. Natl. Acad. Sci. USA 96, 11271–11276 (1999). 11. Bopp, M. A., Sytnik, A., Howard, T. D., Cogdell, R. J. & Hochstrasser, R. M. The dynamics of structur single light-harvesting complexes. Proc. Natl. Acad. Sci. USA 96, 11271–11276 (1999). g g g 2. Hofmann, C., Aartsma, T. J., Michel, H. & Köhler, J. Direct observation of tiers in the energy landscape of a chromoprotein: a single-molecule study. Proc. Natl. Acad. Sci. USA 100, 15534–15538 (2003). 12. Hofmann, C., Aartsma, T. J., Michel, H. & Köhler, J. Direct observation of tiers in the energ single-molecule study. Proc. Natl. Acad. Sci. USA 100, 15534–15538 (2003). 13. Ketelaars, M. et al. Spectroscopy on Individual light-harvesting 1 complexes of Rhodopseudomonas acidophila. Biophys. J. References 83, 1701–1715 (2002).h 14. Cogdell, R. J., Gall, A. & Köhler, J. The architecture and function of the light-harvesting apparatus of purple bacteria: from single molecules to in vivo membranes. Q. Rev. Biophys. 39, 227–324 (2006). Q p y , ( ) 5. Brotosudarmo, T. H. P. et al. Single-molecule spectroscopy reveals that individual low-light LH2 complexes from Rhodopseu- domonas palustris 2.1.6. have a heterogeneous polypeptide composition. Biophys. J. 97, 1491–1500 (2009). 16. Richter, M. F., Baier, J., Cogdell, R. J., Köhler, J. & Oellerich, S. Single-molecule spectroscopic characterization of light-harvesting 2 complexes reconstituted into model membranes. Biophys. J. 93, 183–191 (2007). p y 17. Landsiedel, R. et al. Testing metal-oxide nanomaterials for human safety. Adv. Mater. 22, 2601–2627 (2010). ll l f l h l h h f ll d b l l l 18. Gall, A. et al. Conformational switching in a light-harvesting protein as followed by single-molecule spectroscopy. Biophys. J. 2713–2720 (2015).h 19. Löhner, A. et al. The origin of the split B800 absorption peak in the LH2 complexes from Allochromatium vinosum. Photosynth. Res. 123, 23–31 (2015).il 0. Tubasum, S. et al. Protein configuration landscape fluctuations revealed by exciton transition polarizations in single light harvesting complexes. J. Phys. Chem. B 120, 724–732 (2016). p y 1. Jelezko, F., Tietz, C., Gerken, U., Wrachtrup, J. & Bittl, R. Single-molecule spectroscopy on photosystem I pigment−protein com- plexes. J. Phys. Chem. B 104, 8093–8096 (2000). p y 2. Brecht, M., Radics, V., Nieder, J. B., Studier, H. & Bittl, R. Red antenna states of photosystem I from Synechocystis PCC 6803 Biochemistry-US 47, 5536–5543 (2008). y 3. Brecht, M., Hussels, M., Schlodder, E. & Karapetyan, N. V. Red antenna states of Photosystem I trimers from Arthrospira platensis revealed by single-molecule spectroscopy. Biochim. Biophys. Acta 1817, 445–452 (2012). 4. Brecht, M., Radics, V., Nieder, J. B. & Bittl, R. Protein dynamics-induced variation of excitation energy transfer pathways. Proc Natl. Acad. Sci. USA 106, 11857–11861 (2009). 5. Tietz, C. et al. Single molecule spectroscopy on the light-harvesting complex II of higher plants. Biophys. J. 81, 556–562 (2001). 6 K ü T P J N d hki V I Ili i C & G d ll R Fl l d i f i l LHCII i 25. Tietz, C. et al. Single molecule spectroscopy on the light-harvesting complex II of higher plants. Biophys. J. 81, 556–562 (2001). 26. Krüger, T. P. J., Novoderezhkin, V. Materials and methods Cultivation of algae. Cod In order to find a single complex, the focal spot on the sample could be scanned over a range of 250 µm × 250 µm using a galvo scanner (FLIMbee unit, PoicoQuant). The emission from the sample passed a dichroic mirror (ZT488/561 kpc, AHF/Croma) and suitable optical filters (LP H 560 LPXR, AHF) and was then split in a beamsplitter cube in a ratio of 50/50. One part of the signal was directed towards a spectrometer (Shamrock SR-163, grating 600 lines/ mm blazed at 500 nm) via another fibre, and the emission spectrum was detected using an electron-multiplying charge-coupled device (EMCCD Newton 970, Andor). The spectral resolution was 0.5 nm corresponding to 11 cm−1 at 680 nm. The other half of the signal was focused onto a single-photon counting avalanche diode (SPCM-AQRH-14-TR, Excelitas) and used for time tagged time-resolved (TTTR) fluorescence data collection employing time correlated single photon counting (TCSPC TimeHarp 260 PICO Dual, PicoQuant). The tran- sients were measured with a temporal resolution of 250 ps. In order to avoid pile-up effects in the photon count- ing statistics when measuring ensembles, we ensured that the detection count rate was less than 1% of the laser repetition rate by placing an OD 3 in front of the detector. For obtaining the fluorescence lifetimes the measured https://doi.org/10.1038/s41598-022-11572-3 Scientific Reports \| (2022) 12:8461 \| www.nature.com/scientificreports/ www.nature.com/scientificreports/ Proceedings of the US-Japan conference held at Hakone. Jpn. Soc. Plant Pysiol. 1966, 63–75 (1966). g g J p Jp y , ( ) 43. Wientjes, E., Roest, G. & Croce, R. From red to blue to far-red in Lhca4: how does the protein modulate the spectral properties of the pigments?. Biochim. Biophys. Acta 1817, 711–717 (2012). g g J p Jp y , ( ) 43. Wientjes, E., Roest, G. & Croce, R. From red to blue to far-red in Lhca4: how does the protein modulate the spectral properties of the pigments?. Biochim. Biophys. Acta 1817, 711–717 (2012). 44. Pascal, A. A. et al. Molecular basis of photoprotection and control of photosynthetic light-harvesting. Nature 436, 134–137 (2005). 45. Anderson, J. M. Chlorophyll-protein complexes of a Codium species, including a light-harvesting siphonaxanthin-Chlorophylla ab-protein complex, an evolutionary relic of some Chlorophyta. Biochim. Biophys. Acta 724, 370–380 (1983). 4. Pascal, A. A. et al. Molecular basis of photoprotection and control of photosynthetic light-harvesting. Nature 436, 134–137 (2005) 5. Anderson, J. M. Chlorophyll-protein complexes of a Codium species, including a light-harvesting siphonaxanthin-Chlorophylla ab-protein complex, an evolutionary relic of some Chlorophyta. Biochim. Biophys. Acta 724, 370–380 (1983). 44. Pascal, A. A. et al. Molecular basis of photoprotection and control of photosynthetic light-harvesting. Nature 436, 134–137 (2005). 45. Anderson, J. M. Chlorophyll-protein complexes of a Codium species, including a light-harvesting siphonaxanthin-Chlorophylla ab-protein complex, an evolutionary relic of some Chlorophyta. Biochim. Biophys. Acta 724, 370–380 (1983). p p y p y p y 6. Fujii, R., Yamano, N., Hashimoto, H., Misawa, N. & Ifuku, K. Photoprotection vs. photoinhibition of photosystem II in transplas- tomic lettuce (Lactuca sativa) dominantly accumulating astaxanthin. Plant Cell Physiol. 57, 1518–1529 (2016). 47. de Mello, J. C., Wittmann, H. F. & Friend, R. H. An improved experimental determination of external photoluminescence quantum efficiency. Adv. Mater. 9, 230–232 (1997). 7. de Mello, J. C., Wittmann, H. F. & Friend, R. H. An improved experimental determination of external photoluminescence quantum efficiency. Adv. Mater. 9, 230–232 (1997). 47. de Mello, J. C., Wittmann, H. F. & Friend, R. H. An improved experimental determination of external photoluminescence quantum efficiency. Adv. Mater. 9, 230–232 (1997). Acknowledgements We thank Werner Reichstein for assistance with the MT 200 microscope and Dr. Frank-Julian Kahle for assis- tance with the fluorescence quantum yield measurements. T.H.P.B. thanks the Humboldt Foundation for a Georg-Forster Fellowship. Financial support from the Humboldt Foundation and the State of Bavaria within the initiatives “Solar Technologies go Hybrid" and the Elite Network "Biological Physics" is gratefully acknowledged. Financial support from the grant-in-aid of the Sasakura Enviro-Science Foundation (to S.S.) and the OCU Stra- tegic Research Grant for Basic Researches (to R.F.) are also acknowledged. Author contributions Conceptualization, J.K. and R.F.; methodology, J.K.; software, B.W.; vali-dation, T.H.P.B., B.W., J.K.; formal analy- sis, T.H.P.B.; investigation, T.H.P.B., S.S.; resources, J.K.; data curation, J.K. and R.F.; writing—original draft preparation, J.K.; writing—review and editing, T.H.P.B., S.S., R.F., J.K.; visualization, T.H.P.B.; supervision, J.K.; project administration, J.K.; funding acquisition, J.K. All authors have read and agreed to the published version of the manuscript. References I., Ilioaia, C. & van Grondelle, R. Fluorescence spectral dynamics of single LHCII trimers. Biophys. J. 98, 3093–3101 (2010). 6. Krüger, T. P. J., Novoderezhkin, V. I., Ilioaia, C. & van Grondelle, R. Fluorescence spectral dynamics of single LHCII trimers Biophys. J. 98, 3093–3101 (2010). p y 27. Krüger, T. P. J., Ilioaia, C., Valkunas, L. & van Grondelle, R. Fluorescence intermittency from the main plant light-harvesting complex: sensitivity to the local environment. J. Phys. Chem. B 115, 5083–5095 (2011). y y 28. Krüger, T. P. J., Wientjes, E., Croce, R. & van Grondelle, R. Conformational switching explains the intrinsic multifunctional plant light-harvesting complexes. Proc. Natl. Acad. Sci. USA 108, 13516–13521 (2011). hl h l l l l d fi f h d d h d f h h p g g p 29. Schlau-Cohen, G. S. et al. Single-Molecule Identification of Quenched and Unquenched States of LHCII. J. Phys. Chem. Lett. 6, 860–867 (2015).l 0. Kondo, T., Chen, W. J. & Schlau-Cohen, G. S. Single-molecule fluorescence spectroscopy of photosynthetic systems. Chem. Rev 117, 860–898 (2017). ( ) 31. Köhler, J. & Cogdell, R. J. Studying conformational changes of proteins via single-molecule spectroscopy: Cryogenic temperatures versus room temperature. Adv. Bot. Res. 91, 1–31 (2019).fi p 32. Ilioaia, C., Johnson, M. P., Horton, P. & Ruban, A. V. Induction of efficient energy dissipation in the isolated light-harvesting complex of Photosystem II in the absence of protein aggregation. J. Biol. Chem. 283, 29505–29512 (2008).h p y p gg g 3. Ramanan, C. et al. The role of exciton delocalization in the major photosynthetic light-harvesting antenna of plants. Biophys. J 108, 1047–1056 (2015). ( ) 34. Krüger, T. P. J., Ilioaia, C., Johnson, M. P., Ruban, A. V. & van Grondelle, R. Disentangling the low-energy states of the major light- harvesting complex of plants and their role in photoprotection. Biochim. Biophys. Acta 1837, 1027–1038 (2014). https://doi.org/10.1038/s41598-022-11572-3 Scientific Reports \| (2022) 12:8461 \| Funding g Open Access funding enabled and organized by Projekt DEAL. Competing interests The authors declare no competing interests. www.nature.com/scientificreports/ www.nature.com/scientificreports/ 5. Nordlund, T. M. & Knox, W. H. Lifetime of fluorescence from light-harvesting chlorophyll a/b proteins. Excitation intensity dependence. Biophys. J. 36, 193–201 (1981).l p p y 36. Tredwell, C. J. & Searle, G. F. Picosecond fluorescence from photosynthetic systems in vivo. Ciba F. Symp. 14, 257–281 (197 l p y y ) Picosecond events and their measurement (Academic Press Inc, 1 l 37. Seibert, M. (ed.) Picosecond events and their measurement (Ac 38. van Oort, B., van Hoek, A., Ruban, A. V. & van Amerongen, H. Equilibrium between quenched and nonquenched conformations of the major plant light-harvesting complex studied with high-pressure time-resolved fluorescence. J. Phys. Chem. B 111, 7631–7637 (2007). 39. Böhm, P. S., Kunz, R., Southall, J., Cogdell, R. J. & Köhler, J. Does the reconstitution of RC-LH1 complexes from Rhodopseu- domonas acidophila strain 10050 into a phospholipid bilayer yield the optimum environment for optical spectroscopy?. J. Phys. Chem. B 117, 15004–15013 (2013).l 40. Moya, I., Silvestri, M., Vallon, O., Cinque, G. & Bassi, R. Time-resolved fluorescence analysis of the photosystem II antenna pro in detergent micelles and liposomes. Biochemistry 40, 12552–12561 (2001). 40. Moya, I., Silvestri, M., Vallon, O., Cinque, G. & Bassi, R. Time-resolved fluorescence analysis of the photosystem II antenna proteins in detergent micelles and liposomes. Biochemistry 40, 12552–12561 (2001). 41. Belgio, E., Johnson, M. P., Jurić, S. & Ruban, A. V. Higher plant photosystem II light-harvesting antenna, not the reaction center, determines the excited-state lifetime-both the maximum and the nonphotochemically quenched. Biophys. J. 102, 2761–2771 (2012). 42. L. Provasoli. Media and prospects for the cultivation of marine algae. In: Watanabe A, Hattori A (end) Cultures and collections of algae. Proceedings of the US-Japan conference held at Hakone. Jpn. Soc. Plant Pysiol. 1966, 63–75 (1966). 41. Belgio, E., Johnson, M. P., Jurić, S. & Ruban, A. V. Higher plant photosystem II light-harvesting antenna, not the reaction center, determines the excited-state lifetime-both the maximum and the nonphotochemically quenched. Biophys. J. 102, 2761–2771 (2012). 42. L. Provasoli. Media and prospects for the cultivation of marine algae. In: Watanabe A, Hattori A (end) Cultures and collections of algae. Proceedings of the US-Japan conference held at Hakone. Jpn. Soc. Plant Pysiol. 1966, 63–75 (1966). p y q p y 42. L. Provasoli. Media and prospects for the cultivation of marine algae. In: Watanabe A, Hattori A (end) Cultures and collections of algae. Additional informationh Additional information Supplementary Information The online version contains supplementary material available at https://doi.org/ 10.1038/s41598-022-11572-3. Correspondence and requests for materials should be addressed to J.K. Reprints and permissions information is available at www.nature.com/reprints. © The Author(s) 2022 Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, 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 licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2022 https://doi.org/10.1038/s41598-022-11572-3 Scientific Reports \| (2022) 12:8461 \|
https://openalex.org/W2742681823	https://juniperpublishers.com/jojun/pdf/JOJUN.MS.ID.555588.pdf	English	null	Cancer in Israel: Incidence and Trends	JOJ urology & nephrology	2,017	cc-by	2,499	Introduction reporting of newly diagnosed cancer cases has been mandatory by Israeli law since 1982. The INCR covers the entire Israeli population (more than eight millions citizens) and receives data from several sources including hospitals, pathology and cytology laboratories, death certificates from local health authorities and cause of death information from Israel Central Bureau of Statistics [6]. Data is available online and updated regularly (last update on May 5, 2016). The last year of fully updated data regarding testicular cancer was 2013. All data presented in our study was extracted from the Ministry of Health online database and edited as needed. The incidence of testicular cancer has increased steadily over the past 20 years in the United States and Europe [1-3]. Data from other parts of the world shows a similar trend of increasing incidence [4]. Based on data from GLOBOCAN 2012, Norway has the highest age-standardized incidence rate (ASR) of 12.2 per 100,000 citizens. Other European countries have a high ASR, most of which above 7.0 [5]. Israel has the highest ASR among all Asian countries [5]. There are three large and distinct sub-populations in Israel: Ashkenazi Jews-descendants of Jews from central and eastern Europe; Sephardi Jews who are descendants of Jews from Iberia and Israeli Arabs. Moreover, Israel includes other smaller sub- populations such as Asian Jews and foreign workers. We obtained data on the incidence of testicular cancer in the different ethnic groups in Israel and compared the ASR between them. We also analyzed the trends in the incidence of testicular cancer since 1982. ASR is reported as new cases per 100,000 persons per year and standardized by Segi world standard population. This report presents an updated rates and trends of testicular cancer in Israel based on data from the Israel National Cancer Registry (INCR). Research Asrticle Volume 2 Issue 3 - March 2017 DOI: 10.19080/JOJUN.2017.2.555588 JOJ uro & nephron Copyright © All rights are reserved by Zaher Bahouth Research Asrticle Volume 2 Issue 3 - March 2017 DOI: 10.19080/JOJUN.2017.2.555588 JOJ uro & nephron Copyright © All rights are reserved by Zaher Bahouth Abstract Background: The incidence of testicular cancer is steadily increasing in most parts of the world, with European countries leading the list. In Asia, Israel has the highest age-standardized ratio. However, Israel has several sub-populations with various ethnics and potentially significant differences in incidence. Objectives: To report the incidence and trends of testicular cancer in Israel in the past two decades and to show the differences between the different ethnic groups. Methods: After a legal instruction in 1982, all medical centers in Israel started to report every patient with any cancerous disease to the Israel National Cancer Registry (INCR). Data from the INCR is freely available online. We retrieved data on testicular cancer in Israel and analyzed it. No statistical comparison was done. Results: In Israel, there is a trend of increasing incidence of testicular cancer. Ashkenazi Jewish have the highest risk of developing testicular cancer, with an incidence approaching some European countries. All sub-populations in Israel showed a steady increase in incidence of testicular cancer. Conclusions: Ashkenazi Jews has a 5.7-fold the risk of Sephardi Jews or Israeli Arabs, and 8-fold the risk of Asian Jews. The incidence of testicular cancer in Israel is increasing in all sub-populations. Keywords: Testicular cancer; GLOBOCAN; INCR; Israel Zaher Bahouth* and Ofer Nativ Department of Urology, School of Medicine, Israel Submission: February 03 2017; Published: April Department of Urology, School of Medicine, Israel Submission: February 03, 2017; Published: April 06, 2017 Department of Urology, School of Medicine, Israel Submission: February 03, 2017; Published: April 06, 2017 Submission: February 03, 2017; Published: April 06, 2017 *Corresponding author: Zaher Bahouth, Department of Urology, Bnai-Zion Medical Center, School of Medicine, Technion, Haifa, Israel, Tel: ; Email: Results Results Table 1: ASR and actuarial number of patients diagnosed with testicular cancer in Israel between 1982-2013, in all ethnic groups. Results Year Jewish Arabs Others Overall Asia Sephardi Ashkenazi Israeli 1982 38(2.33) 3(1.02) 4(1.44) 12(4.27) 19(2.01) 4 (2.35) 0 1983 36 (2.06) 0 4 (1.29) 9 (2.4) 23 (2.7) 4 (1.11) 0 1984 37 (2.08) 5 (6.84) 0 10 (1.93) 22 (2.02) 1 (0.43) 0 1985 56 (2.99) 2 (1.11) 1 (0.43) 24 (8.64) 29 (3.41) 1 (0.21) 0 1986 40 (2.28) 3 (0.94) 3 (0.87) 13 (3.31) 21 (2.52) 3 (0.76) 0 1987 46 (2.59) 0 1 (0.39) 21 (7.26) 24 (2.53) 1 (0.25) 1 1988 58 (3.17) 1 (0.25) 3 (2.52) 22 (6.89) 32 (2.49) 2 (0.43) 1 1989 60 (3.15) 2 (1.81) 1 (0.27) 18 (5.5) 39 (4.76) 3 (0.61) 2 1990 53 (2.75) 5 (1.59) 1 (0.98) 16 (3.5) 31 (3.14) 4 (0.98) 0 1991 70 (3.30) 6 (3.15) 3 (4.79) 24 (5.07) 37 (2.82) 3 (1.01) 1 1992 60 (2.83) 3 (0.92) 0 15 (2.61) 42 (4.15) 2 (0.32) 3 1993 70 (3.16) 1 (0.29) 1 (0.52) 21 (3.79) 47 (3.24) 8 (1.36) 4 1994 71 (3.16) 2 (0.56) 2 (0.57) 24 (4.08) 43 (3.3) 4 (0.95) 1 1995 75 (3.35) 2 (1.08) 2 (0.55) 20 (3.17) 51 (3.67) 8 (1.69) 1 1996 72 (3.11) 4 (0.96) 4 (1.11) 25 (4.66) 39 (2.35) 10 (1.95) 4 1997 93 (3.97) 2 (0.94) 1 (0.55) 22 (3.68) 68 (4.43) 8 (1.77) 0 1998 88 (3.61) 3 (2.88) 2 (0.54) 21 (2.72) 61 (4.06) 9 (1.37) 5 1999 101 (4.00) 4 (4.23) 2 (2.47) 27 (4.34) 68 (3.95) 7 (1.10) 5 2000 97 (3.93) 2 (2.39) 4 (1.2) 14 (1.9) 77 (4.65) 11 (1.70) 3 2001 112 (4.37) 1 (0.28) 2 (1.1) 29 (3.64) 80 (4.55) 10 (1.64) 5 2002 122 (4.76) 2 (0.47) 4 (2.71) 29 (4.03) 87 (4.69) 11 (1.89) 4 2003 118 (4.48) 4 (3.76) 1 (0.27) 29 (3.99) 84 (4.25) 12 (1.87) 4 2004 115 (4.40) 1 (0.19) 3 (2.38) 28 (3.56) 83 (4.52) 12 (1.92) 6 2005 126 (4.70) 1 (0.25) 0 24 (3.17) 101 (5.18) 10 (1.47) 2 2006 121 (4.41) 0 3 (3.20) 20 (3.67) 97 (4.73) 16 (2.30) 5 2007 142 (5.11) 0 1 (1.46) 30 (5.44) 111 (5.26) 10 (1.33) 7 2008 115 (4.08) 2 (0.33) 1 (0.53) 20 (3.51) 91 (4.28) 10 (1.43) 4 2009 135 (4.57) 2 (0.19) 0 18 (3.33) 115 (5.57) 16 (2.09) 10 2010 134 (4.68) 1 (0.09) 0 25 (5.29) 108 (5.0) 10 (1.39) 7 2011 131 (4.42) 0 2 (1.86) 41 (7.68) 87 (3.72) 23 (2.98) 3 2012 150 (4.99) 2 (3.13) 1 (1.33) 32 (6.07) 114 (4.89) 10 (1.19) 9 2013 135 (4.53) NA NA NA NA 28 (3.13) 6 Mean 88 (3.63) 2.5 (1.52) 2.2 (1.35) 22 (4.29) 62.3 (3.8) 7.8 (1.35) 4 Median 88 (3.61) 2 (0.95) 2 (1.1) 22 (3.79) 61 (4) 8 (1.37) 4 NA: Not Available. Materials and Methods The INCR has trained registrars whom, for each case, use the available documents to establish site and morphology according to ICD-O-3 coding and record grade, size of tumor, treatment Data on the incidence of testicular cancer patients in Israel was obtained from the INCR. INCR was established in 1960 and 001 JOJ uro & nephron 2(3): JOJUN.MS.ID.555588 (2017) JOJ Urology & Nephrology benign testicular neoplasms, neoplasms of uncertain behavior or testicular lymphomas. provided and stage according to the SEER Summary Staging Manual. It should be emphasized that these data do not include Results Figure 1: ASR of patients diagnosed with testicular cancer in Jewish and Arab males in Israel, from 1982 to 2013 with a trend line for each sub-population. This is mainly an epidemiological study, and as such, has some inherent limitations. No statistical studies can be done to compare the statistical significance of the differences between the ASR. Data in the registry is limited, and some crucial data was not available for us in this study. Moreover, a child of an Ashkenazi parents, who was born in Israel is registered as an Israeli, and another child of two Sephardi parents who was born in Israel is registered as Israeli. Figure 2: Trends in Jews only, divided into Ashkenazi and Sephardi Jews. Figure 2: Trends in Jews only, divided into Ashkenazi and Sephardi Jews. Conclusion The incidence of testicular cancer is still increasing in Israel, as in several other countries around the world. Ashkenazi Jewish males are at the highest risk of developing testicular cancer in Israel, with ASR approaching the highest populations in Scandinavia. Figure 2 shows the trends in Jewish males only, divided into two groups: Ashkenazi and Sephardi Jews. In this figure, we can see an ASR of 4.0-8.0 in Ashkenazi Jews compared to less than 2 in Sephardi Jews. Figure 3 shows the ASR of all the sub- populations in Israel, with the highest rate being in Ashkenazi Jews. Results All numbers are actual number of cases. Numbers in brackets are ASR (1:100.000). No ASR for other groups. We retrieved data from 1982 through 2013. In 2013, with testicular cancer in Israel between 1982-2013, divided into l number of patients diagnosed with testicular cancer in Israel between 1982-2013, in all ethnic groups. NA: Not Available. All numbers are actual number of cases. Numbers in brackets are ASR (1:100.000). No ASR for o We retrieved data from 1982 through 2013. In 2013, testicular cancer was diagnosed in 163 patients in Israel: 135 Jewish and 28 Arabs. This translates into ASR of 4.51 for the Jewish population and 3.13 for the Arabs population. Table 1 shows the ASR and the actuarial number of patients diagnosed with testicular cancer in Israel between 1982-2013, divided into the different sub-populations. Figure 1 shows the ASR of patients diagnosed with testicular cancer in Jewish and Arab males in Israel, from 1982 to 2013 Zaher B, Ofer N. Cancer in Israel: Incidence and Trends. JOJ uro & nephron. 2017; 2(3): 555588. DOI: 10.19080/ How to cite this article: Zaher B, Ofer N. Cancer in Israel: Incidence and Trends. JOJ uro & nephron. 2017; 2(3): 5555 JOJUN.2017.2.555588. 002 JOJ Urology & Nephrology with a trend line for each sub-population. This figure shows a trend of increasing incidence through the past two decades. These factors could also explain the highest rates of testicular cancer in Israel compared to other Asian countries and the high rates in Ashkenazi Jewish-descendants of European countries. trend of increasing incidence through the past two decades. Figure 1: ASR of patients diagnosed with testicular cancer in Jewish and Arab males in Israel, from 1982 to 2013 with a trend line for each sub-population. Figure 2: Trends in Jews only, divided into Ashkenazi and Sephardi Jews. Figure 1: ASR of patients diagnosed with testicular cancer in Jewish and Arab males in Israel, from 1982 to 2013 with a trend line for each sub-population. Significant differences in ASR exist in same countries in different ethnic groups [12]. For example, Horner et al showed that the incidence of testicular cancer in white non-Hispanic Americans is five times higher than Black Americans [7]. In Israel, we showed that the risk in Ashkenazi Jews is, on average, 5.7 times higher than in Sephardi Jews and Israeli Arabs, and 8 times higher than Jewish from Asian countries. References 1. Nigam M, Aschebrook-Kilfoy B, Shikanov S, Eggener S (2014) Increasing incidence of testicular cancer in the United States and Europe between 1992 and 2009. World J Urol 33(5): 623-631. 1. Nigam M, Aschebrook-Kilfoy B, Shikanov S, Eggener S (2014) Increasing incidence of testicular cancer in the United States and Europe between 1992 and 2009. World J Urol 33(5): 623-631. Figure 3: ASR of all the Israeli sub-populations. Figure 3: ASR of all the Israeli sub-populations. Discussion 2. Gilbert D, Rapley E, Shipley J (2011) Testicular germ cell tumours: predisposition genes and the male germ cell niche. Nat Rev Cancer 11(4): 278-288. 3. Dalgaard MD, Weinhold N, Edsgärd D, Silver JD, Pers TH, et al. (2012) A genome-wide association study of men with symptoms of testicular dysgenesis syndrome and its network biology interpretation. J Med Genet 49(1): 58-65. 4. Huyghe E, Matsuda T, Thonneau P (2003) Increasing Incidence of Testicular Cancer Worldwide: A Review. J Urol 170(1): 5-11. Figure 3: ASR of all the Israeli sub-populations. 5. Ferlay J, Soerjomataram I, Ervik M, Dikshit R, Eser S, et al. (2013) GLOBOCAN 2012 v1.0, Cancer Incidence and Mortality Worldwide in 2012 v1.0. IARC CancerBase No. 11. International Agency for Research on Cancer, France. Discussion Discussion Testicular cancer is the most common malignancy among men in their 20-40 years of age [7]. The incidence of testicular cancer appears to be increasing worldwide [8]. In our study, Israel is no different; there is an obvious trend of increasing incidence among all sub-populations in Israel. The trend of increasing incidence is obvious in Jewish and Arab men. 6. Anon. Israel National Cancer Registry. Ministry of Health, State of Israel. 7. Horner M, Ries L, Krapcho M (2009) SEER cancer statistics review, 1975- 2006. Bethesda (MD): National Cancer Institute, 2009. 8. Purdue MP, Devesa SS, Sigurdson AJ, McGlynn KA (2005) International patterns and trends in testis cancer incidence. Int J Cancer 115(5): 822-827. The incidence of testicular cancer varies significantly according to geographic region with rates being highest in Scandinavia and lowest in Asia [9]. However, Israel has the highest rates of testicular cancer among all Asian countries [5]. Genetic and environmental factors could possibly be contributing factors to the increasing incidence of testicular cancer [10,11]. 9. Weijl NI, Rutten MFJ, Zwinderman AH, Keizer HJ, Nooy MA, et al. (2000) Thromboembolic Events During Chemotherapy for Germ Cell Cancer: A Cohort Study and Review of the Literature. J Clin Oncol 18(10): 2169- 2178. her B, Ofer N. Cancer in Israel: Incidence and Trends. JOJ uro & nephron. 2017; 2(3): 555588. DOI: 10.19080/ How to cite this article: Zaher B, Ofer N. Cancer in Israel: Incidence and Trends. JOJ uro & nephron. 2017; 2(3): 555588. DOI: 10.19080/ JOJUN.2017.2.555588. How to cite this article: Zaher B, Ofer N. Cancer in Israel: Incidence and Trends. JOJ uro & nephron. 2017; 2(3): 5555 JOJUN.2017.2.555588. 003 JOJ Urology & Nephrology JOJ Urology & Nephrology (2008) Testicular dysgenesis syndrome and the origin of carcinoma in situ testis. Int J Androl 31(2): 275-287. (2008) Testicular dysgenesis syndrome and the origin of carcinoma in situ testis. Int J Androl 31(2): 275-287. 10. Mai PL, Chen BE, Tucker K, Friedlander M, Phillips KA, et al. (2009) Younger age-at-diagnosis for familial malignant testicular germ cell tumor. Fam Cancer 8(4): 451-456. 12. Wein AJ, Kavoussi LR, Partin AW, Peters CA, (2016). Campbell-Walsh urology (11th edn). Elsevier, Philadelphia, PA, USA. 12. Wein AJ, Kavoussi LR, Partin AW, Peters CA, (2016). Campbell-Walsh urology (11th edn). Elsevier, Philadelphia, PA, USA. 11. Sonne SB, Kristensen DM, Novotny GW, Olesen IA, Nielsen JE, et al. Your next submission with Juniper Publishers will reach you the below assets • Quality Editorial service • Swift Peer Review • Reprints availability • E-prints Service • Manuscript Podcast for convenient understanding • Global attainment for your research • Manuscript accessibility in different formats ( Pdf, E-pub, Full Text, Audio) • Unceasing customer service Track the below URL for one-step submission https://juniperpublishers.com/online-submission.php 004
https://openalex.org/W2054384744	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0037182&type=printable	English	null	No Evidence for a Trade-Off between Reproductive Investment and Immunity in a Rodent	PloS one	2,012	cc-by	10,013	Yan-Chao Xu1,2, Deng-Bao Yang1,2, De-Hua Wang1* 1 State Key Laboratory of Integrated Management of Pest Insects and Rodents, Institute of Zoology, Chinese Academy of Sciences, Beijing, China, 2 Graduate School of the Chinese Academy of Sciences, Beijing, China Abstract This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This study was supported by grants from the Ministry of Science and Technology (2007BC109103), Chinese Academy of Sciences (KSCX2-EW-N-005) and National Natural Science Foundation of China (31071930). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: wangdh@ioz.ac.cn PLoS ONE \| www.plosone.org Abstract Life history theory assumes there are trade-offs between competing functions such as reproduction and immunity. Although well studied in birds, studies of the trade-offs between reproduction and immunity in small mammals are scarce. Here we examined whether reduced immunity is a consequence of reproductive effort in lactating Brandt’s voles (Lasiopodomys brandtii). Specifically, we tested the effects of lactation on immune function (Experiment I). The results showed that food intake and resting metabolic rate (RMR) were higher in lactating voles (6# litter size #8) than that in non- reproductive voles. Contrary to our expectation, lactating voles also had higher levels of serum total Immunoglobulin G (IgG) and anti-keyhole limpet hemocyanin (KLH) IgG and no change in phytohemagglutinin (PHA) response and anti-KLH Immunoglobulin M (IgM) compared with non-reproductive voles, suggesting improved rather than reduced immune function. To further test the effect of differences in reproductive investment on immunity, we compared the responses between natural large (n$8) and small litter size (n#6) (Experiment II) and manipulated large (11–13) and small litter size (2–3) (Experiment III). During peak lactation, acquired immunity (PHA response, anti-KLH IgG and anti-KLH IgM) was not significantly different between voles raising large or small litters in both experiments, despite the measured difference in reproductive investment (greater litter size, litter mass, RMR and food intake in the voles raising larger litters). Total IgG was higher in voles with natural large litter size than those with natural small litter size, but decreased in the enlarged litter size group compared with control and reduced group. Our results showed that immune function is not suppressed to compensate the high energy demands during lactation in Brandt’s voles and contrasting the situation in birds, is unlikely to be an important aspect mediating the trade-off between reproduction and survival. Citation: Xu Y-C, Yang D-B, Wang D-H (2012) No Evidence for a Trade-Off between Reproductive Investment and Immunity in a Rodent. PLoS ONE 7(5): e37182. doi:10.1371/journal.pone.0037182 Edit Sh R Si h N i l C I i U i d S f A i Editor: Shree Ram Singh, National Cancer Institute, United States of America Editor: Shree Ram Singh, National Cancer Institute, United States of America Received January 10, 2012; Accepted April 17, 2012; Published May 23, 2012 Received January 10, 2012; Accepted April 17, 2012; Published May 23, 2012 Copyright: 2012 Xu et al. No Evidence for a Trade-Off between Reproductive Investment and Immunity in a Rodent Yan-Chao Xu1,2, Deng-Bao Yang1,2, De-Hua Wang1* Experiment I Body mass, food intake and resting metabolic rate (RMR). Before pairing, there was no significant difference in body mass between the non-lactating and lactating voles (t = 0.140, df = 10, P = 0.892; Figure 1a). The lactating voles had higher body mass than non-lactating voles during 15 days of lactation (group effect, F1,11 = 6.348, P = 0.028; day effect, F5,55 = 2.222, P = 0.065; interaction group6day, F5,55 = 2.401, P = 0.049; Fig. 1a). , Before pairing, there was no significant difference in dry matter intake between the non-lactating and lactating voles (t = 0.211, df = 10, P = 0.837). The lactating voles had significantly higher dry matter intake than non-lactating voles from day 6 to day 15 during lactation (group effect, F1,10 = 20.956, P,0.001; day effect, F4,40 = 7.631, P,0.001; interaction group6day, F4, 40 = 8.211, P,0.001; Fig. 1b), and dry matter intake was increased by about 100% compared with non-lactating voles on day15. , Before pairing, there was no significant difference in dry matter intake between the non-lactating and lactating voles (t = 0.211, df = 10, P = 0.837). The lactating voles had significantly higher dry matter intake than non-lactating voles from day 6 to day 15 during lactation (group effect, F1,10 = 20.956, P,0.001; day effect, F4,40 = 7.631, P,0.001; interaction group6day, F4, 40 = 8.211, P,0.001; Fig. 1b), and dry matter intake was increased by about 100% compared with non-lactating voles on day15. RMR in lactating voles was also significantly increased by about 100% compared with non-lactating voles (F1, 10 = 13.147, P = 0.05; Fig. 1c) on day 13. Serum hormones. Serum leptin concentrations were signif- icantly decreased in lactating voles compared to non-lactating voles (t = 2.440, df = 10, P = 0.035; table 1). Lactating voles had higher serum prolactin concentrations than non-lactating voles (t = 22.976, df = 11, P = 0.013; table 1). There was no significant difference in serum corticosterone concentrations between lactat- ing voles and non-lactating voles (t = 22.146, df = 10, P = 0.057; table 1). Body composition and organ mass. Thymus mass was reduced in lactating voles compared to non-lactating voles (F1, 10 = 8.438, P = 0.016; Table S1), whereas spleen mass had no significant difference (F1, 10 = 0.01, P = 0.921). Introduction by white adipose tissue, is known to be involved in the regulation of food intake during lactation [13,14]. In addition, prolactin is required for the ongoing maintenance of milk secretion [15] and the regulation of hyperphagia and metabolic process during lactation [16]. These two hormones may also play an important signal driving counterbalance between reproduction and immune function [4]. Elevated corticosterone release may reflect the stress of high energy demand [17], which may suppress immunity [18]. The high cost of lactation requires that energy intake must increase, or that the allocation of energy to other functions reduces [19]. However, sustained energy intake during late lactation might be limited intrinsically by aspects of an animal’s physiology [9,11,12,20]; other physiological functions would be consequently down-regulated. Reproduction and self-maintenance are important for fitness and both require substantial energy investment [1,2,3,4,5,6]. Because animals are frequently constrained by intrinsic physio- logical limitations that govern their capacity to expend energy, they must consequently maintain an optimal allocation of energy between competing physiological functions (e.g. growth, repro- duction and immunity) [7,8]. In small mammals, the costs of reproduction involve higher energy and nutrient demands and energy expenditure [5]. The energy demands of mammalian reproduction increase throughout lactation; particularly late lactation is the energetically critical period of the breeding cycle [9]. The greater expenditure during lactation is related to the mass of nursing young and to the cost of their locomotion and temperature regulation, as well as to the cost of growth [10,11]. Organ remodeling which involves growth of the alimentary tract and other associated metabolic organs (including heart, liver, lung and kidney) and body fat utilization are necessary to achieve the high demands of lactation in many small rodents [10,12]. A number of hormones may play an important role in the energy intake and expenditure during lactation. Leptin, secreted Life-history theory predicts that current reproductive effort gives rise to a fitness cost, which may be observed as reduced survival or future reproduction [21]. To survive, animals must be able to generate immune responses to resist potentially life-threatening diseases. However, mounting an immune response requires substantial energy [1,3,4,22]. Introduction Many studies in lots of species have found support for trade-offs between reproduction and immunity, with immunity being suppressed during energetically reproductive May 2012 \| Volume 7 \| Issue 5 \| e37182 1 May 2012 \| Volume 7 \| Issue 5 \| e37182 PLoS ONE \| www.plosone.org Reproductive Investment and Immunity periods [23,24,25,26,27], although not all studies have showed a suppresive effect of reproductive investment on immune function [28]. In a variety of species, when investment in reproduction increases, there is a concomitant increase in host susceptibility to parasites [29,30]. Figure 1. Changes of body mass (a) and dry matter intake (DMI) (b) over time. RMR (c) on day 13 of lactation in lactating voles. Values are means 6 s.e.m. N represents non-lactating group, and L0 represents lactating group. Significant difference between groups is indicated by an asterisk if P,0.05. doi:10.1371/journal.pone.0037182.g001 The precise reasons for immunosuppression during reproduc- tive period are unclear, but one proposed mechanism considered reproductive effort as the main reason to suppress immune function [31]. Many studies have documented that experimentally increased reproductive effort adversely affected immune function in birds [2,24,25,30,31,32,33,34]. Few studies have focused on the effect of reproductive effort during lactation on the ability to mount immune responses in small mammals [35,36]. We conducted three experiments to test the hypothesis whether the reproductive effort of lactating Brandt’s voles (Lasiopodomys brandtii) negatively affects the immune function. We predicted that i) the immune function would be suppressed in lactating voles, and ii) larger reproductive effort (larger litters) would be associated with greater immunosuppression. Experiment I Body fat mass (F1, 10 = 6.335, P = 0.031) and body fat content (F1, 10 = 6.115, P = 0.043) were decreased in lactating voles compared with non- lactating voles. Liver mass (F1, 10 = 13.856, P = 0.004), gastroin- testinal tract with or without content were increased in lactating voles compared with non-lactating voles except caecum mass (F1, 10 = 3.551, P = 0.089). There were no significant differences in heart (F1, 10 = 0.019, P = 0.893), lungs (F1, 10 = 0.048, P = 0.831), kidneys (F1, 10 = 2.219, P = 0.167) between the two groups. Figure 1. Changes of body mass (a) and dry matter intake (DMI) (b) over time. RMR (c) on day 13 of lactation in lactating voles. Values are means 6 s.e.m. N represents non-lactating group, and L0 represents lactating group. Significant difference between groups is indicated by an asterisk if P,0.05. doi:10.1371/journal.pone.0037182.g001 Innate immunity, PHA response and humoral immunity. Serum total Immunoglobulin G (IgG) concentration was significantly increased in lactating females (t = 28.328, PLoS ONE \| www.plosone.org PLoS ONE \| www.plosone.org May 2012 \| Volume 7 \| Issue 5 \| e37182 2 Reproductive Investment and Immunity Table 1. Effect of lactation on serum hormones in female Brandt’s voles. Table 1. Effect of lactation on serum hormones in female Brandt’s voles. Parameters N L0 T df P Leptin (ng/ml) 4.43460.617a 2.71760.217b 2.440 10 P,0.05 Prolactin (ng/ml) 134.61169.873a 176.673616.305b 22.976 11 P,0.05 Corticosterone (nmol/l) 2.98460.174 3.80160.339 22.146 10 Ns Values are means 6 s.e.m. significant differences are indicated by different superscripts in each row if P,0.05, determined by independent-samples T test. doi:10.1371/journal.pone.0037182.t001 Values are means 6 s.e.m. significant differences are indicated by different superscripts in each row if P,0.05, determined by in doi:10.1371/journal.pone.0037182.t001 df = 11, P,0.001; Fig. 2a) compared with non-lactating voles. No significant difference in PHA response was found between non- lactating and lactating voles (t = 1.174, df = 11, P = 0.265; Fig. 2b). There was no significant difference in anti-keyhole limpet hemocyanin (KLH) Immunoglobulin M (IgM) concentrations between non-lactating and lactating voles (t = 21.341, df = 12, P = 0.205; Fig. 2c). Lactating voles had significantly higher anti- KLH Immunoglobulin G (IgG) concentrations (t = 22.311, df = 12, P = 0.039; Fig. 2d) compared with non-lactating voles. Experiment II E represents enlarged litter size group, C represents non-manipulated litter size group, and R represents reduced litter size group. Changes of dry matter intake (DMI) (c) over time and RMR (d) on day 13 of lactation. Values are means 6 s.e.m. Significant differences in litter mass between L and S group or litter mass and food intake between E, C and R are indicated by an asterisk if P,0.05. A pound sign indicates significant differences in DMI between E and R group, and a double asterisk indicates significant differences between E and R group, and between C and R group if P,0.05, whereas significant differences in RMR between E, C and R are indicated by different letters if P,0.05. doi:10.1371/journal.pone.0037182.g003 voles (t = 0.556, df = 15, P = 0.587; Table S2). There was no significant difference in serum corticosterone concentrations between large litter size voles and small litter size voles (t = 1.459, df = 10, P = 0.175; Table S2). 12.06660.785 g on day 15 of lactation, respectively. The mean pup mass in the large litter size group was smaller than that in the small litter size group (group effect, F1,15 = 16.675, P = 0.001; day effect, F5,75 = 232.576, P,0.001; interaction group6day, F5,75 = 11.239, P,0.001). 12.06660.785 g on day 15 of lactation, respectively. The mean pup mass in the large litter size group was smaller than that in the small litter size group (group effect, F1,15 = 16.675, P = 0.001; day effect, F5,75 = 232.576, P,0.001; interaction group6day, F5,75 = 11.239, P,0.001). 12.06660.785 g on day 15 of lactation, respectively. The mean pup mass in the large litter size group was smaller than that in the small litter size group (group effect, F1,15 = 16.675, P = 0.001; day effect, F5,75 = 232.576, P,0.001; interaction group6day, F5,75 = 11.239, P,0.001). Body composition and organ mass. Thymus (F1, 14 = 0.068, P = 0.798) and spleen mass (F1, 14 = 0.058, P = 0.813) did not differ between large litter size voles and small litter size voles (Table S2). Experiment II Litter size and litter mass. At the end of lactation, the mean numbers of offspring in large and small groups were 9.060.0 and 3.360.5, respectively (Fig. 3a). Females in the large litter size group (L) had a significantly higher total litter mass than those in the small litter size group (S) (group effect, F1,15 = 31.772, P,0.001; day effect, F5,75 = 178.721, P,0.001; interaction group6day, F5,75 = 17.449, P,0.001; Fig. 3a). The mean pup mass in large and small groups were 2.47160.043 g and 2.90960.103 g on day 0 of lactation and 8.12260.507 g and Figure 2. The effects of lactation on serum total IgG (a), PHA response (b), serum anti-KLH IgM (c) and anti-KLH IgG (d) in N and group. Values are means 6 s.e.m. Significant difference between groups is indicated by an asterisk if P,0.05. doi:10.1371/journal.pone.0037182.g002 Figure 2. The effects of lactation on serum total IgG (a), PHA response (b), serum anti-KLH IgM (c) and anti-KLH IgG (d) in N and L0 group. Values are means 6 s.e.m. Significant difference between groups is indicated by an asterisk if P,0.05. doi:10.1371/journal.pone.0037182.g002 May 2012 \| Volume 7 \| Issue 5 \| e37182 PLoS ONE \| www.plosone.org 3 3 Figure 3. Changes of litter mass (a) during lactation. L represents large litter size group, and S represents small litter size group. Change litter mass (b) after manipulation during lactation. E represents enlarged litter size group, C represents non-manipulated litter size group, an represents reduced litter size group. Changes of dry matter intake (DMI) (c) over time and RMR (d) on day 13 of lactation. Values are means 6 s. Significant differences in litter mass between L and S group or litter mass and food intake between E, C and R are indicated by an asterisk if P,0.0 pound sign indicates significant differences in DMI between E and R group, and a double asterisk indicates significant differences between E an group, and between C and R group if P,0.05, whereas significant differences in RMR between E, C and R are indicated by different letters if P,0 doi:10.1371/journal.pone.0037182.g003 Reproductive Investment and Immu Reproductive Investment and Immunity Figure 3. Changes of litter mass (a) during lactation. L represents large litter size group, and S represents small litter size group. Changes of litter mass (b) after manipulation during lactation. Experiment II There were no significant differences in body fat mass (F1, 14 = 1.632, P = 0.222), body fat content (F1, 14 = 1.172, P = 0.297), heart (F1, 14 = 0.407, P = 0.534), liver (F1, 14 = 2.250, P = 0.156), lungs (F1, 14 = 0.411, P = 0.532), kidneys (F1, 14 = 3.806, P = 0.071) and gastrointestinal tract with or without content between the two groups except stomach (F1, 14 = 7.476, P = 0.016) and intestine with content mass (F1, 14 = 9.757, P = 0.007). Maternal body mass. Before pairing, no difference in body mass existed between the Large and small group (t = 2.059, df = 15, P = 0.059). During lactation, change of body mass in large group was not significantly different from that of small group (group effect, F1, 15 = 4.416, P = 0.053; day effect, F5, 75 = 16.941, P,0.001; interaction group6day, F5, 75 = 2.598, P = 0.032; Fig. S2). Serum hormones. Serum leptin concentration had no significant difference in large litter size voles, compared to small litter size voles (t = 0.532, df = 15, P = 0.519; Table S2). An increased trend, but no significant difference in serum prolactin was found in large litter size voles compared to small litter size Innate immunity, PHA response and humoral immunity. Total IgG concentration was greater in large litter size females than small litter size females (t = 4.418, df = 15, PLoS ONE \| www.plosone.org May 2012 \| Volume 7 \| Issue 5 \| e37182 4 Reproductive Investment and Immunity Figure 4. The effect of litter size on serum total IgG (a), PHA response (b), serum anti-KLH IgM (c) and serum anti-KLH IgG (d) in L and S group. Values are means 6 s.e.m. Significant difference between groups is indicated by an asterisk if P,0.05. doi:10.1371/journal.pone.0037182.g004 Figure 4. The effect of litter size on serum total IgG (a), PHA response (b), serum anti-KLH IgM (c) and serum anti-KLH IgG (d) in L and S group Values are means 6 s e m Significant difference between groups is indicated by an asterisk if P,0 05 Figure 4. The effect of litter size on serum total IgG (a), PHA response (b), serum anti-KLH IgM (c) and serum anti-KLH IgG (d) in L and S group. Values are means 6 s.e.m. Significant difference between groups is indicated by an asterisk if P,0.05. Experiment II doi:10.1371/journal.pone.0037182.g004 P,0.001; Fig. 4a).There was no significant difference in the PHA response between the large litter size and small litter size individuals (t = 21.705, df = 15, P = 0.109; Fig. 4b). KLH-IgG (t = 0.278, df = 18, P = 0.784; Fig. 4d) and KLH-IgM (t = 0.218, df = 18, P = 0.830; Fig. 4c) concentrations were also not signifi- cantly different between large litter size and small litter size individuals. 90 = 9.744, P,0.001; interaction group6day, F10, 90 = 2.25, P = 0.022; Fig. S3). Further analysis showed that body mass in reduced group decreased significantly during this period (F5, 35 = 10.660, P,0.001), but not in enlarged group (F5, 30 = 2.251, P = 0.075) and control group (F5, 25 = 2.145, P = 0.093). , Before mating, there was no significant difference in dry matter intake in enlarged, control and reduced groups (group effect, F2,18 = 1.222, P = 0.318; day effect, F2,36 = 0.835, P = 0.442; interaction group6day, F4,36 = 1.083, P = 0.379; Fig. 3c). Signif- icant differences were found among enlarged, control and reduced groups during lactation (group effect, F2,18 = 17.716, P,0.001; day effect, F4,72 = 41.777, P,0.001; interaction group6day, F8,72 = 4.885, P,0.001). From day 3 to day 15 during lactation, food intake of all voles significantly increased (P,0.05). The voles with enlarged litters had significant higher dry matter intake than those of control voles (P = 0.030) except day 15 and the voles with reduced litters throughout lactation (P,0.001), and dry matter intake of the voles with reduced litters was significantly lower than control voles (P = 0.005). PLoS ONE \| www.plosone.org Correlations between RMR and total litter mass Correlations between RMR and total litter mass There was a positive relationship between RMR of manipulated lactating voles (n = 23) and total litter mass (r = 0.868, P,0.001; Fig. S1). Body composition and organ mass. Thymus (F2, 17 = 0.433, P = 0.656) and spleen mass (F2, 17 = 0.368, P = 0.698) did not differ among enlarged litter size voles, control voles and reduced litter size voles (Table S3). The differences were not significant for body fat mass (F2, 17 = 0.012, P = 0.988) and body fat content (F2, 17 = 0.007, P = 0.993). There were no significant differences in heart (F2, 17 = 0.805, P = 0.463), liver (F2, 17 = 1.337, P = 0.289), lungs (F2, 17 = 0.618, P = 0.551), kidney (F2, 17 = 2.017, P = 0.164), gastrointestinal tract with or without content among these three groups except stomach with content (F2, 17 = 3.683, P = 0.047) and stomach (F2, 17 = 5.676, P = 0.013). Experiment III Litter size and litter mass. The mean numbers of offspring in enlarged, control and reduced groups were 11.760.2, 7.560.2 and 2.460.3 at the end of lactation (Fig. 3b). Total litter mass differed significantly among manipulation groups (group effect, F2,18 = 61.373, P,0.001; day effect, F5,90 = 190.617, P,0.001; interaction group6day, F10,90 = 9.180, P,0.001; Fig. 3b). Specif- ically, the enlarged group had a significantly higher total litter mass than both the control and reduced groups (P,0.001), and the reduced group had a significantly lower total litter mass than the control group (P,0.001). Significant difference in RMR was found among enlarged, control and reduced groups (F2,15 = 10.217, P = 0.002; Fig. 3d). Specifically, RMR of the voles raising reduced litters was significant lower than that of control voles (P = 0.015), but there Maternal body mass, food intake and RMR. Before pairing, no difference in body mass existed among enlarged, control and reduced groups (F2, 18 = 1.136, P = 0.343). During lactation, there was no significant difference among these three groups (group effect, F2, 18 = 2.487, P = 0.111; day effect, F5, PLoS ONE \| www.plosone.org May 2012 \| Volume 7 \| Issue 5 \| e37182 5 Reproductive Investment and Immunity was no significant difference between voles with enlarged litters and control voles (P = 0.104). P = 0.643; Fig. 5b). KLH-IgM of voles raising enlarged litters was not significantly different to that of control voles, however, voles with reduced litters had significantly higher anti-KLH IgM than controls (F2,17 = 4.640, P = 0.025; Fig. 5c). KLH-IgG was not significantly different among the three groups (F2, 17 = 0.315, P = 0.734; Fig. 5d). Serum hormones. There were no significant differences in serum leptin concentrations in enlarged voles, compared to control and reduced voles (F2, 18 = 0.321, P = 0.730; Table S3). An increased trend, but no significant difference in serum prolactin was found between reduced group and enlarged group (F2, 18 = 1.001, P = 0.387; Table S3). Serum corticosterone concentra- tions had no significant difference among these three groups (F2, 18 = 0.843, P = 0.447; Table S3). Energetic cost of reproduction Lactation represents the most energetically demanding periods of the life cycle for small female mammals [5,37,38], and energy intake and energy expenditure during lactation are extremely high [5,39,40,41]. A remarkable increase in food intake, RMR and the mass of alimentary tract (see table S1, S2 and S3) assured us that costs of lactation in voles were substantial. Although high energy intake is exported directly to offspring in milk [42], a large proportion is metabolized by females, increasing maternal daily energy expenditure and RMR. The differences observed between lactating and non-lactating animals (e.g. Fig. 1) might actually be the result of some underlying physiological/life-history difference between females which conceived when paired with a male, and those that did not. This is why it is crucial for us to include manipulation of the litter size in Experiment III. Our results showed that RMR increased in lactating voles, especially in the voles with large litter size, and RMR was positively correlated with litter mass (see Fig. S1). These data indicated that the energetic cost of maternal maintenance increased with the increase of energetic investment of lactation. A short-fall in energy intake relative to energy expenditure resulted in declined but non- significant trend of body mass (see Fig. 1a and Fig. S2 and Fig. S3) and mobilization of reserves (body fat) in lactating voles. Actually, some studies have shown that food intake during peak lactation was limited, and did not increase with the increase of litter size [5,10,43,44]. The apparent physiological limit may be imposed by the capacity of alimentary tract to process food into a form for mobilization [5,43], and it may underpin an important life-history trait (the maximum litter size) and an important life- history trade-off. It has been suggested that lifetime reproductive success depends more on parental survival than fecundity [45,46]. Immunity is critical for the survival of small mammals, and it also requires substantial energy [2,3,4,22]. Thus, it is important to note whether the trade-off of resource allocation exists between reproduction and immunity in wild small mammals. In addition, Brandt’s vole was regarded as an income breeder, which compensates for the energetic demands of reproduction by increasing food intake rather than only mobilizing energy stores [7,63]. Our data showed that lactating voles dramatically increased their food intake. Overall, maternal food intake increased with increases in litter size and litter mass. Energetic cost of reproduction The plasticity of digestive tract is necessary for the increased food intake during lactation in several rodents [11,14,64]. The digestive tract (including stomach, intestine, caecum and colon) increased in lactating voles, and large litter size group had a greater digestive tract than small litter size group (see table S1, S2, S3). Therefore, the voles may compensate the high energy requirement during lactation by increasing food intake, but not reducing the energy resources allocated to immunity. However, increased feeding rate during lactation might have some other cost, such as increased exposure to predation and/or disease when foraging. Lactating voles exhibited reduced leptin and elevated prolactin, which may be involved in the regulation of food intake and energy expenditure [14,15,65,66]. Moreover, these two hormones may be important regulators of the reproductive and immune systems and their interactions [4,49,67]. However, we did not find any correlations between these two hormones and immunity in this study. ‘Stress hormone’ corticosterone has been used by ecologists as an indicator of physiological stress in wild vertebrates [68], and it seems that lactation (no change in corticosterone) is not a kind of stress for voles. Although we do not detect any correlations between the hormones and immune function in voles in the Reproductive Investment and Immunity Reproductive Investment and Immunity group compared with control and reduced group; however, these levels were still higher than observed in non-reproductive voles. demands including decreased body mass, fat mass and increased RMR and food intake [14; this study]. Although it is not known why immunity was not suppressed in lactating voles, the vertebrate immune system is very complex. It is likely that different aspects of immunity will respond to energy challenges in different ways [55]. The immune system comprises the acquired arm and the innate arm [56]. Development of the acquired immune system is potentially the largest immunological investment vertebrates make, but the cost of use is modest [4]. Due to this reason, voles might not tend to compromise the acquired arm during lactation which is vital for their long-term survival. Another important component of the immune system is innate immunity (only total IgG measured in the present study). IgG is the most abundant immunoglobulin in circulation, and may represent a state of immunological ‘readiness’ [57]. It is worth noting that total IgG increased in lactating voles, and was also higher in females raising large natural litters. The present result is contrast with that of one previous study in a bird [57]. One possible explanation should be that higher innate immunity may reflect better overall individual condition; thus, those female voles who were in better condition (and had higher IgG levels) could afford to raise larger litters. However, comparing control group (litter size = 6–8) and large group (litter size .8), we found that total IgG was higher in the control group than that in the large group. Thus, the precise mechanism is needed to be clarified in the future study. A potential reason for increased total IgG during lactation may be due to maternal transfer of antibodies which can protect the neonate offspring from infection [58]. In vertebrates, IgG are the primary class of immunoglobulins that transfer via milk [59]. The amount of antibodies transferred to offspring is correlated to the concentration of antibodies in the blood of the females, and mirrors the local disease environment [60,61]. Therefore, mothers with large litter size had to produce more total IgG due to offspring demand. Only acquired immune responses and total IgG were determined in the present study, and single measures of immunity are probably insufficient to charac- terize immunocompetence [4,62]. Discussion Our data indicated that reproduction in Brandt’s vole is physiologically costly, and was reflected by significantly elevated energy intake and expenditure. However, in contrast to our a priori predictions almost all the immune components that we measured were not suppressed by increased reproductive invest- ment. Specifically, PHA response was not affected, and humoral immunity and innate immunity (i.e. total IgG) were significantly enhanced in lactating voles. In addition, differences in reproduc- tive investment in voles raising both natural and manipulated litters did not affect the immune responses except total IgG. Total IgG was higher in voles with natural large litter size than those with natural small litter size, but decreased in enlarged litter size Innate immunity, PHA response and humoral immunity. Voles with enlarged litters had lower total IgG than voles that had their litters reduced and control voles (F2,18 = 3.731, P = 0.042; Fig. 5a). PHA response was not significantly different among enlarged, control and reduced voles (F2,18 = 0.452, Innate immunity, PHA response and humoral immunity. Voles with enlarged litters had lower total IgG than voles that had their litters reduced and control voles (F2,18 = 3.731, P = 0.042; Fig. 5a). PHA response was not significantly different among enlarged, control and reduced voles (F2,18 = 0.452, Figure 5. The effect of litter size manipulation on serum total IgG (a), PHA response (b), serum anti-KLH IgM(c) and serum anti-KLH IgG (d) in E, C and R group. Values are means 6 s.e.m. Significant difference among groups is indicated by an asterisk if P,0.05. doi:10.1371/journal.pone.0037182.g005 Figure 5. The effect of litter size manipulation on serum total IgG (a), PHA response (b), serum anti-KLH IgM(c) and serum anti-KLH IgG (d) in E, C and R group. Values are means 6 s.e.m. Significant difference among groups is indicated by an asterisk if P,0.05. doi:10.1371/journal.pone.0037182.g005 PLoS ONE \| www.plosone.org May 2012 \| Volume 7 \| Issue 5 \| e37182 6 Effect of reproductive effort on immune function during lactation Many studies, mostly in birds, have suggested that trade-offs may exist between reproduction and immune function [22,24,45,47,48,49,50,51]. Many studies of experimentally in- creased clutch [24,25] or brood size [48,52,53] have detected adverse effect of an increased reproductive effort on different components of the immune system in birds. Increased incubation cost could lead to a lower humoral immune responsiveness and a reduction of lymphocyte levels in eiders (Somateria mollissima) [24] and a reduction of survival in great tits (Parus major) [54]. Studies which have experimentally increased brood rearing have docu- mented a reduction of T-cell-mediated immunity (measured as a swelling response to PHA) in enlarged broods in pied flycatchers (Ficedula hypoleuca) [52] and a suppression of humoral and PHA response in enlarged broods in tree swallow (Tachycineta bicolor) [33,53]. In the current study, however, we found that immune responses were not suppressed in lactating voles and some, for example, KLH-IgG and total IgG were even increased. Moreover, the acquired immune responses did not differ between large litter size group and small litter size groups, both in the experiments where litter size was experimentally manipulated or naturally variable. This result was surprising given that lactating voles, especially for the voles with large litters, exhibited typical characteristics suggesting they were experiencing high energy PLoS ON PLoS ONE \| www.plosone.org May 2012 \| Volume 7 \| Issue 5 \| e37182 May 2012 \| Volume 7 \| Issue 5 \| e37182 7 Reproductive Investment and Immunity 8754) solution: KLH group and PHA group. All animals were naı¨ve to KLH and PHA. Animals in KLH group and PHA group were sacrificed on day 18 and day 15 of lactation which are late peak lactation. Each vole was euthanized by CO2 asphyxiation between 0900 and 1100 h, and trunk blood was collected which were allowed to clot for 30 min at 4uC and centrifuged at 4uC for 30 min at 3000 r. p. m. Sera were collected and stored in sealable polypropylene micro-centrifuge tubes at 280uC until assay for total IgG, anti-KLH IgG, leptin, prolactin and corticosterone. present study, fluctuations in these hormones must be helpful for maintaining normal immune function during lactation. Thus, more studies should be conducted to illustrate the roles of these hormones in mediating immune activity in wild animals with a seasonal breeding cycle. Experiment II The second experiment explored the effect of natural large and small litter sizes, which were presumed to reflect different reproductive effort, on immune function in lactating voles. Lactating females whose litter size was more than 8 or less than 6 were selected and defined as the large group (L, n = 13) or the small group (S, n = 24), respectively. Each group was divided into two further groups as in experiment I: the KLH group (LK and SK, n = 6 and n = 14, respectively) and the PHA group (LP and SP, n = 7 and n = 10, respectively). Ethics statement All animal procedures were reviewed and approved by the Institutional Animal Care and Use Committee of the Institute of Zoology, Chinese Academy of Sciences (Permit Number: IOZ11012). All researchers and students had been certified before performing animal studies. Experiment III To further test the relationship between reproductive effort and immune function, in the final experiment we manipulated litter size to examine the effect of increased or decreased reproductive effort on the immune function. Animals whose original litter size at birth was 6–8 were used in this experiment. We manipulated litter size by adding or removing pups on the day of parturition. Litters with same parturition date were mixed together and assigned randomly to females. By adding or removing 5 pups, we assigned pups randomly to three treatment groups: E, enlarged group (initial litter size 6–8, with 5 pups added); C, control group (with the initial litter size unchanged); R, reduced group (initial litter size 6–8, with 5 pups removed). Maternal voles easily accept foreign pups, as the survival of offspring did not differ between original pups and cross-fostered pups. Each group was divided into two groups, KLH group (EK, CK and RK, n = 7, n = 6 and n = 7, respectively) and PHA group (EP, CP and RP, n = 7, n = 6 and n = 8, respectively). Experiment I In the first experiment we examined whether immune function was compromised in lactating voles compared with non-repro- ductive voles. Voles whose litter size was 6–8 were defined as the lactating group (L0, n = 13). Animals that were not pregnant or lactating were defined as the non-reproductive group (N, n = 14). Each group was divided into two further groups depending on the immune function assays performed: the KLH group (L0K and NK, n = 7 and n = 7, respectively) and the PHA group (L0P and NP, n = 6 and n = 7, respectively). In conclusion, our results suggest that immunity is not suppressed in Brandt’s voles during lactation even when repro- ductive investment was experimentally increased. Interestingly, innate immunity was even enhanced during lactation. Although there might be a negligible trade-off between litter size and innate immunity in terms of not enhanced serum total IgG in the experimental enlarged group, overall, these data do not support the idea that a trade-off exists between reproduction and immunity in lactating voles. Taken together, immune function is unlikely to be an important aspect mediating the trade-off between repro- duction and survival in lactating Brandt’s voles. Effect of reproductive effort on immune function during lactation Another possible contributing factor is that Brandt’s voles may have experienced great exposure to pathogens and parasites in the wild, especially during the reproductive season [69], thus they may have evolved a strategy to increase investment in immune function at the cost of resources available for their offspring (e.g. reduced growth rate of the large litter size group) [70]. Given the changes in immunity can have marked consequences for disease resistance and long-term survival costs [48,71], animals with high exposure to pathogens and parasites may be unwilling to down regulate immune function. Study species Brandt’s voles inhabit mainly the grasslands of Inner Mongolia of China, Mongolia, and the Baikal region of Russia. They were non-hibernating herbivores and polygamy. They hoard food in the late fall and living in groups during cold winter. In the wild 90% of female voles commenced breeding in April, and the reproductive season lasted till August. During this period, one female could raise 1–2 litters, the length of gestation was 21 days, and the litter size ranged from 2 to 13. In the wild, Brandt’s vole’s life span is around 14 months, and in the laboratory voles can live for 31 months [69]. Previous studies have shown that energy intake and resting metabolic rate increased, and body fat was depleted during lactation [14]. Animals and housing conditions One hundred and eighteen virgin adult female Brandt’s voles, weighing 40–55 g and aged 120–150 days old, were used in this study. They were the offspring of voles from our laboratory colony. Voles were kept individually in plastic cages (30 cm615 cm620 cm) under a 16 h: 8 h light/dark cycle and room temperature (2161uC). Commercial standard rabbit pellets (Beijing KeAo Feed Co., Beijing, China) and water were provided ad libitum. All the females were randomly paired with males for 1 day and then were immediately separated from the males. The day of parturition was designated as day 0 of lactation [42]. To assess humoral and cellular immunity during lactation, each group was divided into two groups which were immunochallenged with KLH (Sigma 7017) and phytohemagglutinin (PHA; PHA-P, Sigma L- PLoS ONE \| www.plosone.org Statistical analysis Data were analyzed using SPSS 17.0 software (SPSS Inc., Chicago, IL, USA). Prior to all statistical analyses, data were examined for normality of variance using the Kolmogorov- Smirnov test. Differences in body mass, litter mass and mean pup mass were analyzed by one-way repeated-measures ANOVA, while differences in food intake were analyzed by one-way repeated-measures ANCOVA with body mass as covariate Immune response measurements Serum total IgG assay. Serum samples from PHA groups of the three experiments were used to determine serum total IgG. Total IgG is one component of innate immunity which may be particularly important for survival in the wild [57].The concen- tration of total IgG was measured by rat IgG ELISA (enzyme- linked immunosorbent assay) kit (RapidBio Lab, Calabasas, CA, USA). The sensitivity by this assay is 1.0 mg/ml when using 10 ml serum samples. Inter- and intra-assay variations were both ,15%. Serum total IgG assay. Serum samples from PHA groups of the three experiments were used to determine serum total IgG. Total IgG is one component of innate immunity which may be particularly important for survival in the wild [57].The concen- tration of total IgG was measured by rat IgG ELISA (enzyme- linked immunosorbent assay) kit (RapidBio Lab, Calabasas, CA, USA). The sensitivity by this assay is 1.0 mg/ml when using 10 ml serum samples. Inter- and intra-assay variations were both ,15%. Table S1 The effects of lactation on body composition, wet organ mass in female Brandt’s voles. (DOC) Table S2 The effects of litter size on body composition, wet organ mass and hormones in Brandt’s voles. (DOC) Table S3 The effects of manipulation on body compo- sition, wet organ mass and hormones in Brandt’s voles. (DOC) (TIF) Figure S2 Changes of maternal body mass in large and small litter size group before and after lactation. L represents large litter size group, and S represents small litter size group. IgM is the major class of antibody early in a primary antibody response, and IgG is the predominant immunoglobulin class present in the blood following an immune challenge [75]. Enzyme-linked immune-sorbent assay (ELISA) was used to measure serum anti-KLH IgM and IgG concentrations according to [36,75]. The Appendix S1 has displayed further details. Figure S3 Changes of maternal body mass in manipu- lation experiment before and after lactation. E represents enlarged litter size group, C represents non-manipulated litter size group, and R represents reduced litter size group. (TIF) PHA response. To measure delayed-type hypersensitivity responses which are localized antigen-specific responses eliciting swelling and redness at the site of antigen injection in immunized animals, we injected subcutaneously 0.1 mg of PHA (PHA-P, Sigma L-8754) dissolved in 0.03 ml of sterile PBS (pH 7.4) in the middle of the left footpad of PHA group around 0900. Prior to injection, the footpad thickness of left hind foot was measured to the nearest 0.01 mm with a micrometer (Tesa Shopcal, Swiss). Six hours after injection, we measured footpad thickness at the injection site. The PHA response was calculated as the difference between pre- and post-injection measurements divided by initial footpad thickness (PHA response = (post PHA2pre PHA)/pre PHA). Each measurement of PHA response was replicated six times [76,77]. The pre-experiment showed that the maximum PHA response occurs after 6 h of PHA injection (unpublished data). PHA response. To measure delayed-type hypersensitivity responses which are localized antigen-specific responses eliciting swelling and redness at the site of antigen injection in immunized animals, we injected subcutaneously 0.1 mg of PHA (PHA-P, Sigma L-8754) dissolved in 0.03 ml of sterile PBS (pH 7.4) in the middle of the left footpad of PHA group around 0900. Prior to injection, the footpad thickness of left hind foot was measured to the nearest 0.01 mm with a micrometer (Tesa Shopcal, Swiss). Six hours after injection, we measured footpad thickness at the injection site. The PHA response was calculated as the difference between pre- and post-injection measurements divided by initial footpad thickness (PHA response = (post PHA2pre PHA)/pre PHA). Each measurement of PHA response was replicated six times [76,77]. Table S3 The effects of manipulation on body compo- sition, wet organ mass and hormones in Brandt’s voles. (DOC) The methods for measuring body composition, organ mass, blood glucose and serum hormones can be found in Appendix S1. Body mass, food intake and reproductive performance Body mass, food intake and reproductive performance Body mass and food intake were measured at 9:00–11:00 every three days for 9 days before mating and throughout lactation. Dry matter intake was calculated from the difference between the dry food given and the dry food residue [food given6 (1-water content)2food residue6 (1-water content)]. Food samples were taken to determine the water content (5.660.7%, N = 15). Litter size and litter mass were also recorded every three days during lactation. PLoS ONE \| www.plosone.org May 2012 \| Volume 7 \| Issue 5 \| e37182 8 Reproductive Investment and Immunity followed by Tukey’s honestly significant difference post hoc comparisons. Differences in RMR, body compositions and organ mass were analyzed by one-way ANCOVA with body mass as covariate followed by Tukey’s honestly significant difference post hoc comparisons. Serum hormones (leptin, corticosterone and prolactin), immune responses (serum anti-KLH IgM, serum anti- KLH IgG, PHA response) were analyzed by independent-samples T test in the first and second experiment. Group differences in serum hormones and immune responses were analyzed by one- way ANOVA in the third experiment. Finally, Pearson correlation analysis was performed to determine the correlations between RMR and total litter mass. Differences between group means were considered statistically significant at P,0.05. All the variables analyzed (except humoral immunity) are from the PHA group of the three experiments. Resting metabolic rate measurements Resting metabolic rate measurements On day 13 of lactation, RMR of voles from the PHA groups of the three experiments was measured by using an open-flow respirometry system (Sable, FoxBox, USA) at 3060.5uC (within thermal neutral zone 27.5–32.5uC of Brandt’s vole [72], which was controlled with a incubator (Yiheng Model LRH-250, Shanghai, China) as described previously [73]. Further details can be found in the Appendix S1 in supporting information. Acknowledgments We would like to thank Professor John R. Speakman for his constructive suggestions and improvement of the English expression. We are grateful to all the members of Animal Physiological Ecology Group for their assistance. Supporting Information Appendix S1 Supplementary Appendix of methods, including resting metabolic rate, measuring serum anti-KLH IgM and IgG concentrations, measuring body composition and organ mass and serum hormones. (DOC) Appendix S1 Supplementary Appendix of methods, including resting metabolic rate, measuring serum anti-KLH IgM and IgG concentrations, measuring body composition and organ mass and serum hormones. (DOC) Humoral immunity. To assess humoral immunity, animals received a single subcutaneous injection of 100 mg of KLH suspended in 0.1 ml sterile saline on day 8 of peak lactation. KLH is a specific non-replicating antigen which induces a robust antibody response without inducing fever or making the animal sick [74]. Animals in all groups were lightly anesthetized with isoflurane (Shandong LiNuo Pharmaceutical) and bled from retro- orbital sinus 5 days post injection to measure anti-KLH IgM concentrations between 0900 and 1100 h. Humoral immunity. To assess humoral immunity, animals received a single subcutaneous injection of 100 mg of KLH suspended in 0.1 ml sterile saline on day 8 of peak lactation. KLH is a specific non-replicating antigen which induces a robust antibody response without inducing fever or making the animal sick [74]. Animals in all groups were lightly anesthetized with isoflurane (Shandong LiNuo Pharmaceutical) and bled from retro- orbital sinus 5 days post injection to measure anti-KLH IgM concentrations between 0900 and 1100 h. IgM is the major class of antibody early in a primary antibody response, and IgG is the predominant immunoglobulin class present in the blood following an immune challenge [75]. Enzyme-linked immune-sorbent assay (ELISA) was used to measure serum anti-KLH IgM and IgG concentrations according to [36,75]. The Appendix S1 has displayed further details. (TIF) The pre-experiment showed that the maximum PHA response occurs after 6 h of PHA injection (unpublished data). Table S1 The effects of lactation on body composition, wet organ mass in female Brandt’s voles. (DOC) PLoS ONE \| www.plosone.org References Anim Behav 62: 417–425. 55. Drazen DL, Trasy A, Nelson RJ (2003) Photoperiod differentially affects energetics of immunity in pregnant and lactating Siberian hamsters (Phodopus sungorus). Can J Zool 81: 1406–1413. 24. Hanssen SA, Hasselquist D, Folstad I, Erikstad KE (2005) Cost of reproduction in a long-lived bird: incubation effort reduces immune function and future reproduction. Proc R Soc Lond B 272: 1039–1046. 56. Janeway C, Travers P, Walport M, Shlomchik M (2004) Immunobiology. New York: Garland publishing. 25. Knowles SCL, Nakagawa S, Sheldon BC (2009) Elevated reproductive effort increases blood parasitaemia and decreases immune function in birds: a meta- regression approach. Funct Ecol 23: 405–415. 57. Greives T, McGlothlin J, Jawor J, Demas G, Ketterson E (2006) Testosterone and innate immune function inversely covary in a wild population of breeding Dark - Eyed Juncos (Junco hyemalis). Funct Ecol 20: 812–818. ˚ Dark - Eyed Juncos (Junco hyemalis). Funct Ecol 20: 812–818. ˚ 26. French SS, DeNardo DF, Moore MC (2007) Trade-Offs between the reproductive and immune systems: facultative responses to resources or obligate responses to reproduction? Am Nat 170: 79–89. 58. Hasselquist D, Nilsson JA˚ (2009) Maternal transfer of antibodies in vertebrates: trans-generational effects on offspring immunity. Phil Trans R Soc Lond B 364: 51–60. 27. Cox RM, Parker EU, Cheney DM, Liebl AL, Martin LB, et al. (2010) Experimental evidence for physiological costs underlying the trade - off between reproduction and survival. Funct Ecol 24: 1262–1269. 59. Boulinier T, Staszewski V (2008) Maternal transfer of antibodies: raising immuno-ecology issues. Trends Ecol Evol 23: 282–288. 60. Gasparini J, McCoy KD, Haussy C, Tveraa T, Boulinier T (2001) Induced maternal response to the Lyme disease spirochaete Borrelia burgdorferi sensu lato in a colonial seabird, the kittiwake Rissa tridactyla. Proc R Soc Lond B 268: 647–650. 28. Bleu J, Massot M, Haussy C, Meylan S (2011) Experimental litter size reduction reveals costs of gestation and delayed effects on offspring in a viviparous lizard. Proc R Soc Lond B 279: 489–498. 61. Lemke H, Hansen H, Lange H (2003) Non-genetic inheritable potential of maternal antibodies. Vaccine 21: 3428–3431. 29. Richner H, Christe P, Oppliger A (1995) Paternal investment affects prevalence of malaria. Proc Natl Acad Sci U S A 92: 1192–1194. 30. Nordling D, Andersson M, Zohari S, Lars G (1998) Reproductive effort reduces specific immune response and parasite resistance. Proc R Soc Lond B 265: 1291–1298. 62. References Speakman JR, Kro´l E (2010) Maximal heat dissipation capacity and hyperthermia risk: neglected key factors in the ecology of endotherms. J Anim Ecol 79: 726–746. 40. Mattingly DK, McClure PA (1982) Energetics of reproduction in large-littered cotton rats (Sigmodon hispidus). Ecology 63: 183–195. 41. Zhang XY, Li YL, Wang DH (2008) Large litter size increases maternal energy intake but has no effect on UCP1 content and serum leptin concentrations in lactating Brandt’s voles (Lasiopodomys brandtii). J Comp Physiol B Biochem Syst Environ Physiol 178: 637–645. 9. Speakman J, McQueenie J (1996) Limits to sustained metabolic rate: the link between food intake, basal metabolic rate, and morphology in reproducing mice, Mus musculus. Physiol Zool 69: 746–769. 10. Johnson M, Thomson S, Speakman JR (2001) Limits to sustained energy intake. II. Inter-relationships between resting metabolic rate, life-history traits and morphology in Mus musculus. J Exp Biol 204: 1937–1946. y 42. Johnson M, Thomson S, Speakman J (2001) Limits to sustained energy intake. I. Lactation in the laboratory mouse Mus musculus. J Exp Biol 204: 1925–1935. ation in the laboratory mouse Mus musculus. J Exp Biol 204: 1925–193 43. Hammond KA, Diamond J (1997) Maximal sustained energy budgets in humans and animals. Nature 386: 457–462. p gy J p 11. Speakman JR, Kro´l E (2005) Limits to sustained energy intake IX: a review of hypotheses. J Comp Physiol B Biochem Syst Environ Physiol 175: 375–394. 44. Wu SH, Zhang LN, Speakman JR, Wang DH (2009) Limits to sustained energy intake. XI. A test of the heat dissipation limitation hypothesis in lactating Brandt’s voles (Lasiopodomys brandtii). J Exp Biol 212: 3455–3465. 12. Weiner J (1992) Physiological limits to sustainable energy budgets in birds and mammals: ecological implications. Trends Ecol Evol 7: 384–388. mammals: ecological implications. Trends Ecol Evol 7: 384–38 13. Brogan RS, Mitchell SE, Trayhurn P, Smith MS (1999) Suppression of leptin during lactation: contribution of the suckling stimulus versus milk production. Endocrinology 140: 2621–2627. 45. Linden M, Møller AP (1989) Cost of reproduction and covariation of life history traits in birds. Trends Ecol Evol 4: 367–371. 46. Koivula M, Koskela E, Mappes T, Oksanen TA (2003) Cost of reproduction in the wild: Manipulation of reproductive effort in the bank vole. Ecology 84: 398–405. 14. Zhang XY, Wang DH (2007) Thermogenesis, food intake and serum leptin in cold-exposed lactating Brandt’s voles Lasiopodomys brandtii. J Exp Biol 210: 512–521. 47. References Norris K, Evans MR (2000) Ecological immunology: life history trade-offs and immune defense in birds. Behav Ecol 11: 19–26. 15. Knobil E, Neill J (2006) Knobil and Neill’s Physiology of Reproduction. Boston: Academic Press. 48. Ardia DR, Schat KA, Winkler DW (2003) Reproductive effort reduces long- term immune function in breeding tree swallows (Tachycineta bicolor). Proc R Soc Lond B 270: 1679–1683. 16. Grattan DR (2002) Behavioural significance of prolactin signalling in the central nervous system during pregnancy and lactation. Reproduction 123: 497–506. 17. Criscuolo F, Bertile F, Durant JM, Raclot T, Gabrielsen GW, et al. (2006) Body mass and clutch size may modulate prolactin and corticosterone levels in eiders. Physiol Biochem Zool 79: 514–521. 49. Demas GE (2004) The energetics of immunity: a neuroendocrine link between energy balance and immune function. Horm Behav 45: 173–180. 50. Hanssen SA, Hasselquist D, Folstad I, Erikstad KE (2004) Costs of immunity: immune responsiveness reduces survival in a vertebrate. Proc R Soc Lond B 271: 925–930. y 18. Padgett DA, Glaser R (2003) How stress influences the immune response. Trends Immunol 24: 444–448. 19. Gittleman JL, Thompson SD (1988) Energy allocation in mammalian reproduction. Am Zool 28: 863–875. 51. Martin Ii LB, Hasselquist D, Wikelski M (2006) Investment in immune defense is linked to pace of life in house sparrows. Oecologia 147: 565–575. ˚ 20. Hammond KA, Diamond J (1992) An experimental test for a ceiling on sustained metabolic rate in lactating mice. Physiol Zool 65: 952–977. 52. Ilmonen P, Hasselquist D, Langefors A˚ , Wiehn J (2003) Stress, immunocom- petence and leukocyte profiles of pied flycatchers in relation to brood size manipulation. Oecologia 136: 148–154. 21. Williams GC (1966) Natural selection, the costs of reproduction, and a refinement of Lack’s principle. Am Nat 100: 687–690. p g 53. Ardia DR (2005) Tree swallows trade off immune function and reproductive effort differently across their range. Ecology 86: 2040–2046. 22. Sheldon BC, Verhulst S (1996) Ecological immunology: costly parasite defences and trade-offs in evolutionary ecology. Trends Ecol Evol 11: 317–321. and trade-offs in evolutionary ecology. Trends Ecol Evol 11: 317–3 54. Visser M, Lessells C (2001) The costs of egg production and incubation in great tits (Parus major). Proc R Soc Lond B 268: 1271–1277. 23. Adamo SA, Jensen M, Younger M (2001) Changes in lifetime immunocompe- tence in male and female Gryllus texensis (formerly G. integer): trade-offs between immunity and reproduction. References 33. Ardia DR (2005) Individual quality mediates trade-offs between reproductive effort and immune function in tree swallows. J Anim Ecol 74: 517–524. 1. Demas GE, Chefer V, Talan MI, Nelson RJ (1997) Metabolic costs of mounting an antigen-stimulated immune response in adult and aged C57BL/6J mice. Am J Physiol Regul Integr Comp Physiol 273: R1631–1637. 34. Eraud C, Duriez O, Chastel O, Faivre B (2005) The energetic cost of humoral immunity in the Collared Dove, Streptopelia decaocto: is the magnitude sufficient to force energy-based trade-offs? Funct Ecol 19: 110–118. J y g g p y 2. Lochmiller RL, Deerenberg C (2000) Trade-offs in evolutionary immunology: just what is the cost of immunity? Oikos 88: 87–98. 2. Lochmiller RL, Deerenberg C (2000) Trade-offs in just what is the cost of immunity? Oikos 88: 87–98 3. Derting TL, Compton S (2003) Immune response, not immune maintenance, is energetically costly in wild white - footed mice (Peromyscus leucopus). Physiol Biochem Zoo 76: 744–752. 35. Saino N, Canova L, Fasola M, Martinelli R (2000) Reproduction and population density affect humoral immunity in bank voles under field experimental conditions. Oecologia 124: 358–366. 4. Martin LB, Weil ZM, Nelson RJ (2008) Seasonal changes in vertebrate immune activity: mediation by physiological trade-offs. Phil Trans R Soc Lond B 363: 321–339. p g 36. Demas GE, Drazen DL, Nelson RJ (2003) Reductions in total bod humoral immunity. Proc R Soc Lond B 270: 905–911. p g Demas GE, Drazen DL, Nelson RJ (2003) Reductions in total body fa 36. Demas GE, Drazen DL, Nelson RJ (2003) Reductions humoral immunity. Proc R Soc Lond B 270: 905–911. 37. Thompson SD, Nicoll ME (1986) Basal metabolic rate and energetics of reproduction in therian mammals. Nature 321: 690–693. 5. Speakman JR (2008) The physiological costs of reproduction in small mammals. Phil Trans R Soc Lond B 363: 375–398. 38. Speakman JR (2000) The cost of living: field metabolic rates of small mammals. Adv Ecol Res 30: 177–297. 6. McNab BK (2002) The physiological ecology of vertebrates: a view from energetics. Ithaca, NY: Cornell University Press. 39. Innes DGL, Millar JS (1981) Body weight, litter size, and energetics of reproduction in Clethrionomys gapperi and Microtus pennsylvanicus. Can J Zool 59: 785–789. 7. Stearns SC (1992) The evolution of life histories. Oxford: University Press Oxford. 8. Author Contributions Conceived and designed the experiments: YCX DHW. Performed the experiments: YCX DBY. Analyzed the data: YCX. Wrote the paper: YCX DHW. PLoS ONE \| www.plosone.org May 2012 \| Volume 7 \| Issue 5 \| e37182 9 References Matson KD, Cohen AA, Klasing KC, Ricklefs RE, Scheuerlein A (2006) No simple answers for ecological immunology: relationships among immune indices at the individual level break down at the species level in waterfowl. Proc R Soc Lond B 273: 815–822. 31. Deerenberg C, Arpanius V, Daan S, Bos N (1997) Reproductive effort decreases antibody responsiveness. Proc R Soc Lond B 264: 1021–1029. 63. Tuomi J, Hakala T, Haukioja E (1983) Alternative concepts of reproductive effort, costs of reproduction, and selection in life history evolution. Am Zool 23: 25–34. 32. Bonneaud C, Mazuc J, Gonzalez G, Haussy C, Chastel O, et al. (2003) Assessing the cost of mounting an immune response. Am Nat 161: 367–379. PLoS ONE \| www.plosone.org May 2012 \| Volume 7 \| Issue 5 \| e37182 May 2012 \| Volume 7 \| Issue 5 \| e37182 10 Reproductive Investment and Immunity 64. Derting TL, Austin MW (1998) Changes in gut capacity with lactation and cold exposure in a species with low rates of energy use, the pine vole (Microtus pinetorum). Physiol Zool 71: 611–623. 71. Marti A, Marcos A, Martinez J (2001) Obesity and immune function relationships. Obes rev 2: 131–140. 72. Li Q, Huang C (1994) Characteristic of the resting metabolic rate of Brandt’s vole. Acta Theriologica Sinica 14: 217–220. 72. Li Q, Huang C (1994) Characteristic of the vole. Acta Theriologica Sinica 14: 217–220. p ) y 65. Stocker CJ, Wargent E, O’Dowd J, Cornick C, Speakman JR, et al. (2007) Prevention of diet-induced obesity and impaired glucose tolerance in rats following administration of leptin to their mothers. Am J Physiol Regul Integr Comp Physiol 292: R1810–R1818. vole. Acta Theriologica Sinica 14: 217–220. g 73. Chi QS, Wang DH (2011) Thermal physiology and energetics in male desert hamsters (Phodopus roborovskii) during cold acclimation. J Comp Physiol B Biochem Syst Environ Physiol 181: 91–103. y y 74. Dixon FJ, Jacot-Guillarmod H, McConahey PJ (1966) The an p y 66. Naef L, Woodside B (2007) Prolactin/leptin interactions in the control of food intake in rats. Endocrinology 148: 5977–5983. y y 74. Dixon FJ, Jacot-Guillarmod H, McConahey PJ (1966) The antibody responses of rabbits and rats to hemocyanin. J Immunol 97: 350–355. of rabbits and rats to hemocyanin. J Immunol 97: 350–355. 67. Yu-Lee LY (2002) Prolactin modulation of immune and inflammatory responses. Recent Prog Horm Res 57: 435–455. 75. Zysling DA, Demas GE (2007) Metabolic stress suppresses humoral immune function in long-day, but not short-day, Siberian hamsters (Phodopus sungorus). J Comp Physiol B Biochem Syst Environ Physiol 177: 339–347. 68. Romero LM (2004) Physiological stress in ecology: lessons from biomedical research. Trends Ecol Evol 19: 249–255. 76. Smits J, Bortolotti GR, Tella JL (1999) Simplifying the phytohaemagglutinin skin-testing technique in studies of avian immunocompetence. Funct Ecol 13: 567–572. 69. Zhang Z, Wang Z (1998) Ecology and management of rodent pests in agriculture. Beijing: Ocean Press. 77. Bellocq J, Krasnov BR, Khokhlova IS, Pinshow B (2006) Temporal dynamics of a T-cell mediated immune response in desert rodents. Comp Biochem Physiol A Mol Integr Physiol 145: 554–559. 70. Martin II LB, Pless M, Svoboda J, Wikelski M (2004) Immune activity in temperate and tropical house sparrows: a common-garden experiment. Ecology 85: 2323–2331. 77. Bellocq J, Krasnov BR, Khokhlova IS, Pinshow B (2006) Temporal dynamics of a T-cell mediated immune response in desert rodents. Comp Biochem Physiol A Mol Integr Physiol 145: 554–559. Reproductive Investment and Immunity PLoS ONE \| www.plosone.org May 2012 \| Volume 7 \| Issue 5 \| e37182 11
https://openalex.org/W3009942132	https://europepmc.org/articles/pmc7085020?pdf=render	English	null	A Novel Approach to Predict Wrinkling of Aluminum Alloy During Warm/Hot Sheet Hydroforming Based on an Improved Yoshida Buckling Test	Materials	2,020	cc-by	8,787	Received: 18 December 2019; Accepted: 3 March 2020; Published: 5 March 2020 Abstract: In order to predict the wrinkling of sheet metal under the inﬂuence of ﬂuid pressure and temperature during warm/hot hydroforming, a numerical simulation model for sheet wrinkling prediction was established, taking into account through-thickness normal stress induced by ﬂuid pressure. From simulations using linear and quadratic elements, respectively, it was found that the latter gave results that were much closer to experimental data. A novel experimental method based on an improved Yoshida Buckling Test (YBT) was proposed for testing the wrinkling properties of sheets under the through-thickness normal stress. A wrinkling coeﬃcient suitable for predicting wrinkling was also presented. Based on the numerical simulations, an experimental validation of wrinkling performance was conducted. Ridge-height curves measured along the main diagonal tensile direction of the sheet were presented and showed that the wrinkling prediction criterion provided good discrimination. Furthermore, the wrinkling properties of several diﬀerent materials were simulated to evaluate the accuracy of the prediction method, and the results revealed that the improved YBT gave good predictions for wrinkling in the conventional sheet metal forming process, while the prediction results for wrinkling in warm/hot sheet hydroforming were also accurate with the ﬂuid pressure of zero. Keywords: warm sheet hydroforming; hydromechanical deep drawing; wrinkling prediction; through-thickness normal stress Received: 18 December 2019; Accepted: 3 March 2020; Published: 5 March 2020 materials materials materials www.mdpi.com/journal/materials Gaoshen Cai 1, Jubo Fu 1, Dongxing Zhang 2,* , Jinlin Yang 1, Yongfeng Yuan 1, Lihui Lang 3 and Sergei Alexandrov 4 Gaoshen Cai 1, Jubo Fu 1, Dongxing Zhang 2,* , Jinlin Yang 1, Yongfeng Yuan 1, Lihui Lang 3 and Sergei Alexandrov 4 1 Faculty of Mechanical Engineering & Automation, Zhejiang Sci-Tech University, Hangzhou 310018, China; caigaocan@zstu.edu.cn (G.C.); fujubo598@163.com (J.F.); jinlinyang@126.com (J.Y.); yuanyf@zstu.edu.cn (Y.Y.) 1 Faculty of Mechanical Engineering & Automation, Zhejiang Sci-Tech University, Hangzhou 310018, China; caigaocan@zstu.edu.cn (G.C.); fujubo598@163.com (J.F.); jinlinyang@126.com (J.Y.); yuanyf@zstu.edu.cn (Y.Y.) 2 Department of Mechanical & Material Engineering, Western University, London, ON N6A5B9, Canada 3 School of Mechanical Engineering and Automation, Beihang University, Beijing 100191, China; lang@buaa.edu.cn 2 Department of Mechanical & Material Engineering, Western University, London, ON N6A5B9, Canada 3 School of Mechanical Engineering and Automation, Beihang University, Beijing 100191, China; lang@buaa.edu.cn 4 Institute for Problems in Mechanics of the Russian Academy of Sciences, 119526 Moscow, Russia; sergei_alexandrov@yahoo.com * Correspondence: dzhan43@uwo.ca 1. Introduction Warm/hot sheet hydroforming was proposed to enhance the plasticity of lightweight materials [1–4]. This innovative and ﬂexible processing method combines the advantages of cold sheet hydroforming and warm/hot sheet forming [5–7] and is regarded as an advanced technique for improving material formability [8–11]. In contrast to the conventional sheet hot stamping process, ﬂuid pressure provides the through-thickness normal stress and assists the sheet formation process during warm/hot sheet hydroforming [12,13]. Therefore, the stress states under the inﬂuence of ﬂuid pressure are of great signiﬁcance for elucidating the forming mechanism of this technique [14–17] and also play an important role in selection of processing method, parameter optimization, and determination of the Materials 2020, 13, 1165; doi:10.3390/ma13051165 www.mdpi.com/journal/materials Materials 2020, 13, 1165 2 of 19 rules governing deformation during forming [18,19]. Consequently, it is necessary to analyze how the ﬂuid pressure acts on the sheet during the forming process [20,21]. Wrinkling instability is one of the main defects that arise during the sheet metal plastic forming process, and it can directly aﬀect the forming quality and precision of thin-walled parts [22,23]. It can also increase wear on the forming die. Thus, the harmful eﬀects of wrinkling on the forming of parts are obvious, and, in view of this, research on wrinkling prediction and control of sheet metal is a key topic in the ﬁeld of plastic forming. Although warm/hot sheet hydroforming can reduce the tendency to fracture and instability (wrinkling, cracking, etc.) [9,12], wrinkling can still occur during sheet forming. To improve the precision and quality of forming of parts and eﬀectively eliminate wrinkling instability, it is particularly important to obtain accurate predictions of the likelihood of wrinkling in the forming process [24–26]. Therefore, it is necessary to investigate the prediction of wrinkling of sheet metal during warm/hot hydroforming. 2.1. Theoretical Prediction Model for Sheet Plastic Instability To date, research on wrinkling has focused mainly on theoretical and experimental aspects. In terms of plastic instability theory, two basic mathematical models to predict wrinkling have been presented: The bifurcation model and the energy model [27–29]. The bifurcation model is based on a variational approach: When the Hill bifurcation functional has a nonzero solution, the material will wrinkle. The energy model is based on the principle of plastic work and compares the critical wrinkling energy with the plastic deformation work: When certain conditions are met, wrinkling will occur. These theoretical models can predict the wrinkling of typical parts during the forming process, such as the hydromechanical deep drawing of cylindrical cups and the hydraulic bulging of pipes under liquid pressure [30,31]. However, it is diﬃcult to represent the surface shapes of complex thin-walled parts using simple real functions, and the boundary conditions in the forming process are very complicated. As a consequence, it is not possible to obtain analytical solutions of the equations resulting from the models in terms of real functions. In addition, the complicated nature of these equations means that errors may easily occur during calculations. Besides the preferred orientation and the relative thickness (t/D) of sheet metal, wrinkling is related to the stress ratio of deformation, the mechanical properties of the material, the geometric shape of the die, and the contact conditions between workpiece and die [25,26]. However, it is diﬃcult to take account of all of these factors in a theoretical model. The development of ﬁnite element numerical simulation has provided an important approach for predicting wrinkling and developing measures for its control. In view of this, a prediction model for warm/hot sheet hydroforming based on the Yoshida buckling test (YBT) [32] was proposed in this study. 2.2. Prediction Method Based on a Sheet Plastic Instability Test 2.2. Prediction Method Based on a Sheet Plastic Instability Test Before theoretical methods had been established for the prediction of wrinkling in conventional sheet metal forming, experimental approaches, including the YBT proposed by Yoshida [32] in 1983 and the tube wrinkling property test proposed by Reddy, were generally used for qualitative analysis of the wrinkling properties of materials. The YBT, also known as the square sheet diagonal tensile test [33], is an experimental method for conducting a diagonal tensile test on a square sheet to evaluate the resistance to wrinkling due to compressional instability of a sheet materials in an uneven tensile state. Also, it has proved to be an eﬀective simulation test of wrinkling properties [34], allowing qualitative analysis of the wrinkling tendencies of diﬀerent materials. A schematic illustration of the test is shown in Figure 1, in which b is the center span of the basis of measurement (mm), and h is the height of wrinkling (mm). 3 of 19 Materials 2020, 13, 1165 Figure 1. Schematic of the YBT. Figure 1. Schematic of the YBT. Figure 1. Schematic of the YBT. It has been reported that [26,35] wrinkling inevitably accompanies compressive stress. In sheet plastic forming, when the compressive stress in one direction of a shell structure increases to a certain extent, if there is no supporting force in the through-thickness direction or if the supporting force is insuﬃcient for the original equilibrium state to be recovered, then wrinkling will appear. In the deep drawing of a sheet, wrinkling is most likely to occur in the ﬂange region of the blank. To reduce the possibility of wrinkling, the factors aﬀecting wrinkling in the ﬂange region should be considered when simulating the process of sheet metal forming. 2.3. Theoretical Basis of Control of Wrinkling To prevent wrinkling during the forming of sheet metal, it is not suﬃcient just to predict its occurrence: It is also essential to analyze how it can be controlled during a practical machining process. It is necessary to start from the basic theory, study the relevant stability conditions, critical conditions, and basic stress-strain states, and then provide a theoretical basis for controlling wrinkling during sheet forming. A thin shell element according to Donnell-Mushtari-Vlasov (DMV) theory is shown in Figure 2, in which the 2δ is equivalent to h and λ −2ux is equivalent to b in Figure 1. In the plate buckling analysis, the strain deformation is small, based on the assumption that when the ratio of the wrinkling height to the thickness of the sheet is less than 0.2 (δ/t < 0.2), and the loading process before wrinkling is proportional loading, the waveform of the wrinkling is symmetrical. Then, from Timoshenko’s buckling theory, the stability condition of the sheet is as follows: (1) ∆T ≤∆U where ∆U is the plastic buckling energy and ∆T is the work done on the sheet by the external force. This condition indicates that the sheet is in a stable state when the work done by the external force does not exceed the plastic buckling energy, whereas if ∆T > ∆U, then wrinkling of the sheet will occur. 4 of 19 Materials 2020, 13, 1165 Figure 2. Schematic of DMV thin shell element. Figure 2. Schematic of DMV thin shell element. According to Hutchinson’s double-curvature shell element theory, the Lagrangian strain tensor at the neutral layer (x3) of the buckling instantaneous distance can be expressed as: (2) εαβ = Eαβ + x3καβ (2) where α, β ∈{1, 2}. Here, Eαβ and καβ are respectively the tensile and bending strains on the shell elements: 1 where α, β ∈{1, 2}. Here, Eαβ and καβ are respectively the tensile and bending strains on the shell elements: 1 Eαβ = 1 2(uα,β + uβ,α) + bαβw (3) (3) (4) καβ = −w,αβ (4) where uα are the components of the displacement in the x1 and x2 directions, w is the displacement normal to the surface of the sheet, bαβ is the curvature tensor of the neutral layer under wrinkling, and a subscript comma indicates a partial derivative (e.g., uα,β = ∂uα/∂xβ). where uα are the components of the displacement in the x1 and x2 directions, w is the displacement normal to the surface of the sheet, bαβ is the curvature tensor of the neutral layer under wrinkling, and a subscript comma indicates a partial derivative (e.g., uα,β = ∂uα/∂xβ). β β The constitutive equation involving all the variables is σαβ = Lαβκγεκγ. Th i f h l f d b di h h ll l h β β The constitutive equation involving all the variables is σαβ = Lαβκγεκγ. β β The expressions for the resultant force and bending moment on the shell element are then Nαβ = Z t/2 −t/2 σαβ dx3 = tLαβκγ . Eκγ (5) (5) Mαβ = Z t/2 −t/2 σαβx3 dx3 = t3 12LαβκγKκγ (6) (6) According to the assumption of a DMV shell element, the buckling strain capacity of sheet can be expressed as: According to the assumption of a DMV shell element, the buckling strain capacity of sheet can be expressed as: ∆U = Z S Z Mαβdκαβ + Z NαβdEαβ ! 3. Wrinkling Prediction Model for Sheet Metal During Warm Sheet Hydroforming 3.1. Wrinkling Test Method for Wrinkling of Sheet Metal During Warm Hydroforming 3.1. Wrinkling Test Method for Wrinkling of Sheet Metal During Warm Hydroforming As mentioned above, the wrinkling resistance of a material is an important property aﬀecting its behavior during forming. For conventional forming methods at normal temperature, the YBT has been used as the evaluation standard for the wrinkling property of materials [32]. However, there are great technical diﬃculties in testing the wrinkling resistance of materials in a forming environment where the stress normal to the sheet thickness is provided by a warm/hot ﬂuid medium. In addition, appropriate test equipment and evaluation criteria are not available for conducting the corresponding tests. Besides, this method cannot be used to examine the wrinkling performance of sheets under through-thickness normal stress provided by ﬂuid pressure, and it is not always appropriate for testing the wrinkling performance of materials at elevated temperature. To solve this technical problem and investigate the wrinkling performance of sheet metals during hydroforming, a novel approach to test the wrinkling performance of sheet metals under the through-thickness normal stress induced by ﬂuid pressure has been proposed that is based on the YBT. This method can be conducted under high-temperature and high-pressure conditions, the square plate to be tested is placed in a special unit, and the through-thickness normal stress is applied to the sheet by means of ﬂuid pressure. During the test, the elongation λ75 of a 75 mm region in the middle of the sheet is measured. When λ75 = 1%, the height h of the arch (i.e., the height of the wrinkling) is taken as the basic data. The corresponding center span is then about b = 25 mm. The curve of wrinkling height versus ﬂuid pressure is then drawn from the recorded test data. A schematic illustration of this test method is shown in Figure 3. Figure 3. Schematic diagram of the wrinkling test. (a) Principle of the wrinkling test. (b) Basis of measurement. Figure 3. Schematic diagram of the wrinkling test. (a) Principle of the wrinkling test. (b) Basis of measurement. In order to complete this test method, an appropriate experimental device was designed by the present authors. ∆T ≤∆U dS (7) (7) where S is the area where wrinkling occurs. Substituting of the expressions (5) and (6) for Nαβ and Mαβ into (7) then gives Z Z where S is the area where wrinkling occurs. Substituting of the expressions (5) and (6) for Nαβ and Mαβ into (7) then gives Z Z ∆U = t 24 Z S Lαβκγw,κγw,αβdS+ t 2 Z S Lαβκγbαβbκγw2dS (8) (8) Because the external force does work in the neutral layer, Because the external force does work in the neutral layer, Because the external force does work in the neutral layer, ∆T = 1 2 Z S (N11w2 ,1 + N22w2 ,2)dS (9) (9) Wrinkling is considered to occur when ∆T = ∆U. This condition is the critical equation that is used to judge the wrinkling of a sheet, which was implemented in numerical simulation software to simulate and investigate the wrinkling of simple sheet parts to prevent wrinkling. Materials 2020, 13, 1165 5 of 19 3. Wrinkling Prediction Model for Sheet Metal During Warm Sheet Hydroforming 6 of 19 6 of 19 Figure 4. Structural diagram of sheet wrinkling test setup: (a) Front view; (b) cutaway view (A–A). 1, heel block; 2, swash partition; 3, heating unit; 4, insulating layer; 5, master cylinder; 6, master cylinder piston; 7, stretching rod; 8, support block; 9, high-temperature sealing ring; 10, case cover; 11, closed oil cylinder; 12, bolt; 13, nut; 14, displacement sensor; 15, sample chuck; 16, cooling hole; 17, tank body; 18, pressure measuring device; 19, data acquisition device; 20, temperature measuring device; 21, ﬁxed mold; 22, high-pressure chamber; 23, ﬂuid pressure generating device; 24, low-pressure chamber; 25, overﬂow valve. Figure 4. Structural diagram of sheet wrinkling test setup: (a) Front view; (b) cutaway view (A–A). 1, heel block; 2, swash partition; 3, heating unit; 4, insulating layer; 5, master cylinder; 6, master cylinder piston; 7, stretching rod; 8, support block; 9, high-temperature sealing ring; 10, case cover; 11, closed oil cylinder; 12, bolt; 13, nut; 14, displacement sensor; 15, sample chuck; 16, cooling hole; 17, tank body; 18, pressure measuring device; 19, data acquisition device; 20, temperature measuring device; 21, ﬁxed mold; 22, high-pressure chamber; 23, ﬂuid pressure generating device; 24, low-pressure chamber; 25, overﬂow valve. Second, the closed box is ﬁlled with oil by the ﬂuid pressure system, and the heated oil is used to heat the samples to the set temperature, by which is to provide the high-temperature and the through-thickness normal stress conditions. Then two master cylinders are driven to stretch the specimen, and the stretching length is recorded by the displacement sensor. When λ75 reaches 1%, the test is stopped. Second, the closed box is ﬁlled with oil by the ﬂuid pressure system, and the heated oil is used to heat the samples to the set temperature, by which is to provide the high-temperature and the through-thickness normal stress conditions. Then two master cylinders are driven to stretch the specimen, and the stretching length is recorded by the displacement sensor. When λ75 reaches 1%, the test is stopped. Third, the sample is cleaned brieﬂy and data are recorded and processed. The arch height of the center span within b = 25 mm is denoted by h. From the processed data, curves of h versus the pressure in the oil chamber are drawn to provide a preliminary idea of the wrinkling behavior of the material. 6 of 19 Finally, the wrinkling coeﬃcient Lλ = (h/λ75)ep/pm is used as an index of wrinkling resistance: the greater the value of Lλ, the less is the wrinkling resistance of the material, and the easier it is for the material to wrinkle. This coeﬃcient directly indicates the degree of wrinkling of a material, and it provides a better reﬂection of the relationship between wrinkling and other factors than does the wrinkling height H obtained by the conventional YBT. This has important theoretical and practical ramiﬁcations for prediction of sheet metal wrinkling during warm/hot sheet hydroforming. 3.2. Wrinkling Prediction Method for Warm Sheet Hydroforming 3. Wrinkling Prediction Model for Sheet Metal During Warm Sheet Hydroforming It consists of a closed box, a stretching mechanism, a ﬂuid pressure system, a heating system, a cooling system, a pressure measuring system, a temperature measuring device, and a data acquisition and processing device, as shown in Figure 4. acquisition and processing device, as shown in Figure 4. The test method using this device involves the following main steps. The test method using this device involves the following main steps. First, multiple groups of 100 mm ×100 mm square plate samples are prepared, and each of these is then divided into two ﬁxed areas and two clamping areas (see Figure 3): The ﬁxed areas are attached to the ﬁxed mold inside the box, and the clamping areas are attached to the specimen holder, with a clamping width of 40 mm. A measuring reference area of length 75 mm is marked on the tensile axis of the sample. The relative elongation during stretching is denoted by λ75. First, multiple groups of 100 mm ×100 mm square plate samples are prepared, and each of these is then divided into two ﬁxed areas and two clamping areas (see Figure 3): The ﬁxed areas are attached to the ﬁxed mold inside the box, and the clamping areas are attached to the specimen holder, with a clamping width of 40 mm. A measuring reference area of length 75 mm is marked on the tensile axis of the sample. The relative elongation during stretching is denoted by λ75. Materials 2020, 13, 1165 6 of 19 3.3. Numerical Simulation of Sheet Wrinkling Performance Test By ignoring factors with only a slight inﬂuence, the wrinkling behavior of typical structural parts can be analyzed on the basis of an appropriate theoretical model, although for irregular parts with complex structures, it is diﬃcult to establish such models. For practical application, it is necessary to discretize theoretical models with high precision, and in the present study, the ﬁnite element method was adopted to predict the occurrence of wrinkling. 3.2. Wrinkling Prediction Method for Warm Sheet Hydroforming In the test program, two indexes are used to measure the degree of wrinkling of sheet metal: One is the fluid pressure–wrinkling height curve, and the other is the previously mentioned wrinkling coefficient: Lλ = h λ75 ep/pm (p, pm ≥0) (10) (10) where p is the ﬂuid pressure in the test procedure (MPa), Pm is the critical ﬂuid pressure when the sheet is wrinkle-free (MPa), h is the measured wrinkling height (mm), and λ75 is the relative elongation when the sheet is stretched in the reference zone, which is the tensile strain. Equation (10) is the prediction model equation for sheet wrinkling adopted in this study. It shows that if no ﬂuid pressure is applied during the test (p = 0), the wrinkling coeﬃcient becomes Lλ = h/λ75, which is the conventional YBT test method. In contrast, the approach presented here takes the ﬂuid pressure and temperature into account, and is thus suitable for dealing with the test environment Materials 2020, 13, 1165 7 of 19 encountered in warm/hot sheet hydroforming. Therefore, the test data obtained using this method can reﬂect the wrinkling properties of sheets in the real forming process. encountered in warm/hot sheet hydroforming. Therefore, the test data obtained using this method can reﬂect the wrinkling properties of sheets in the real forming process. 3.3.1. Numerical Calculation Method for Shell Elements In ﬁnite element software, shell problems are generally divided into two types: Thick shell problems and thin shell problems. In addition, Abaqus software has available three diﬀerent mathematical descriptions of general three-dimensional shell elements: General shell elements, thin-shell elements, and thick-shell elements (Table 1). These shell elements include linear and quadratic triangular elements and quadrilateral elements. Table 1. Abaqus shell element types. General Shell Elements Thin-Shell Elements Thick-Shell Elements S4R, S3R, SAX1, SAX2, SAX2T STRI3, STRI35, STRI65, S4R5, S8R5, S9R5, SAXA S8R, S8RT Isotropic material shells are considered to be thick shells when the ratio of thickness to span is greater than 1/15 and to be thin shells when this ratio is less than 1/15. In the case of the 100 mm × 100 mm square plate tensile test, the shell is thin. To verify the accuracy of wrinkling prediction using thin-shell elements, S4R and S8R, two element types were used to conduct numerical tests. 3.3.2. Inﬂuence of Diﬀerent Microdisturbances on Simulation Results 3.3.2. Inﬂuence of Diﬀerent Microdisturbances on Simulation Results A model for simulation was established, and the material selected was 2A16 aluminum alloy in the form of a square plate of side length L = 100 mm and thickness 1 mm. The clamping width was 30–50 mm, with a ﬁxed constraint applied to the hydraulic chuck at the lower end of the test piece B, the upper portion and the rigid body A were connected by a tie constraint, the reference body control node RP-1 was assigned to the rigid body, and a displacement constraint along the tensile direction Y was applied to the control node. According to the above analysis, since the thickness of the sheet used in the simulation was much less than the dimensions in the other two directions, the shell element in Abaqus was selected for the simulation. In ﬁnite element simulation, factors such as model size, ﬂatness, surface ﬁnish, and chuck size can be all taken as ideal states, and shell elements could be aﬀected only by the stress along the plane direction of the plate. It is therefore diﬃcult to simulate deformation of the plate along the direction perpendicular to its thickness, and so wrinkling cannot be observed. Therefore, external factors can be considered as disturbances (concentrated forces and surface forces) applied to the plate. (1) Concentrated Force Disturbance (1) Concentrated Force Disturbance (1) Concentrated Force Disturbance In the simulation, a concentrated force (1 N) was applied very briefly to a corner of the non-clamping end of the sheet. The sheet size was again 100 mm × 100 mm, and the tensile displacement was 6 mm. Using a static algorithm, the simulation results of wrinkling height (mm) were obtained, as shown in Figure 5. 8 of 19 Materials 2020, 13, 1165 Figure 5. Simulation results for wrinkling of a sheet disturbed by a concentrated force. Figure 5. Simulation results for wrinkling of a sheet disturbed by a concentrated force. (2) Surface Force Disturbance (2) Surface Force Disturbance By contrast, a very small (10−3 MPa) and uniformly distributed surface force was applied very brieﬂy to the unclamped part of the sheet surface. The sheet size was also 100 mm ×100 mm, and the other settings were the same as before, then the simulation results were shown in Figure 6. Figure 6. Simulation results for wrinkling of a sheet disturbed by a surface force. Figure 6. Simulation results for wrinkling of a sheet disturbed by a surface force. Figures 5 and 6 indicate that when the sheet was disturbed by a concentrated force, the maximum wrinkling height at the center was 4.899 mm, while a disturbance by surface force under the same simulation conditions gave a value of 4.895 mm, a diﬀerence of just 0.004 mm, which shows the consistency of the simulation results under the two disturbance conditions. Both simulation and analytical results show that diﬀerent forms of disturbance have similar eﬀects on the wrinkling state of 2A16 aluminum alloy sheet. 3.3.3. Inﬂuence of Diﬀerent Algorithms on Simulation Results 3.3.3. Inﬂuence of Diﬀerent Algorithms on Simulation Results (1) Static Algorithm Analysis (i) Selection of Element Integral Mode Abaqus software provides linear and seco Abaqus software provides linear and second-order elements for the diﬀerent shell element types. The linear element consumes less computational resources and has better computational stability in the analysis of complex friction contact processes. However, in the sheet buckling problem, the shell element reference surface will be subject to continuous curvature deformation. To accurately describe Abaqus software provides linear and second-order elements for the diﬀerent shell element types. The linear element consumes less computational resources and has better computational stability in the analysis of complex friction contact processes. However, in the sheet buckling problem, the shell element reference surface will be subject to continuous curvature deformation. To accurately describe the consequent changes in shape, it is necessary to make appropriate choices of mesh division and element type. Situation results show that quadratic elements provide a better expression of the deformation of the surface boundary. Besides, the calculation of the grid with the linear element was Materials 2020, 13, 1165 9 of 19 faster than with the quadratic element. The reason for this is the greater number of integral nodes in the quadratic element, which makes the arithmetic and integration processes more complicated. (ii) Choice of Meshing Method The meshing density and partitioning method are related to the accuracy of the simulation results. Therefore, an appropriate choice of mesh density is an important aspect of the simulation. To ensure consistency with the simulation settings described above, the clamping width of the sheet was taken as 30 mm and the stretching stroke was taken as 5 mm in the simulations. Using the S8R element, simulation results for the wrinkle height with diﬀerent grid sizes are shown in Table 2. Table 2. Comparison of calculation results for diﬀerent grid sizes with the S8R element. Grid Size (mm) 2 4 6 Wrinkle height at the center of the sheet (mm) 4.387 4.377 4.336 le 2. Comparison of calculation results for diﬀerent grid sizes with the S8R element. The simulation results show that the calculation speed became faster with the larger grid size divided on the sheet. Conversely, the smaller grid size resulted in the slower calculation speed. 3.3.3. Inﬂuence of Diﬀerent Algorithms on Simulation Results 11 of 19 Materials 2020, 13, 1165 Figure 11. Variation of kinetic energy with time. Figure 12. Variations of internal energy and ratio of kinetic energy to internal energy with time. Figure 11. Variation of kinetic energy with time. Figure 11. Variation of kinetic energy with time. Figure 12. Variations of internal energy and ratio of kinetic energy to internal energy with time. Results show that for a grid size of 2 mm, the proportion of kinetic energy was still below 0.5% during the entire loading process, and this process can again be considered to be quasi-static. Also, when the stretching stroke was 5 mm, the wrinkling height at the center of the sheet was 0.309 mm, which is signiﬁcantly smaller than measured by the test. A comparison with the simulation results with a grid size of 4 mm shows that the simulation results are in accordance. This veriﬁes the above analytical results regarding the selection of grid size. Furthermore, the simulation results are in accordance with those of the theoretical analysis, which proves the reliability and applicability of the simulations. 3.4. Experimental Veriﬁcation of Sheet Wrinkling Properties To verify the accuracy of the above theoretical and numerical simulation results, an experiment was carried out. 3.3.3. Inﬂuence of Diﬀerent Algorithms on Simulation Results In addition, the simulation results with grid sizes of 2 mm and 4 mm were similar, although for the quadratic element, the ﬁner meshing gave rise to the smaller overall stiﬀness of the shell and the greater wrinkle height. (2) Dynamic Algorithm Analysis Prediction of wrinkling with a dynamic algorithm commonly used for the forming process of sheet was studied. The sample size was again 100 mm × 100 mm, the clamping width was 30 mm, the tensile displacement was 5 mm, and the loading time was 1 s. To determine whether the simulation process was quasi-static, it was necessary to check whether the kinetic energy of the analyzed body was less than the internal energy by several percentage points in the post-processing result. As mentioned above, the simulation results for grid sizes of 2 mm and 4 mm were in accordance. Therefore, subsequent simulations were performed for both these sizes. Figure 7 shows the results of the wrinkling height (mm) for a grid size of 4 mm, Figure 8 shows the corresponding variation of kinetic energy with time, and Figure 9 shows the variations of internal energy and kinetic/internal energy ratio with time. Figure 7. Deformation nephogram using the dynamic algorithm with a grid size of 4 mm. Figure 7. Deformation nephogram using the dynamic algorithm with a grid size of 4 mm. 10 of 19 Materials 2020, 13, 1165 , , Figure 8. Variation of kinetic energy with time. Figure 9. Variations of internal energy and of ratio of kinetic energy to internal energy with time. i f k h Figure 8. Variation of kinetic energy with time. Figure 9. Variations of internal energy and of ratio of kinetic energy to internal energy with time. Simulation results indicate that the proportion of kinetic energy in the internal energy remained below 0.5% throughout the loading, and so the whole simulation can be considered to be a quasi-static loading process. When the ﬁnal stretching stroke was 5 mm, the wrinkling height at the center of the sheet was 2.065 mm. For comparison, the simulation was also performed for a grid size of 2 mm, and the simulation results were obtained, as shown in Figures 10–12. Figure 10. Deformation nephogram using the dynamic algorithm with a grid size of 2 mm. Figure 10. Deformation nephogram using the dynamic algorithm with a grid size of 2 mm. 3.4.1. Test Equipment and Material In its unﬁlled state (without ﬂuid pressure), the equipment used for the sheet wrinkling test includes an electronic universal testing machine, chucks (a pair with widths of 50 mm and 80 mm, respectively), and an extensometer, as shown in Figure 13 (WDW-100, Wuxi Huayin Test Instrument CO. LTD, Wuxi, China). At the center of the equipment is the operating part, used for assembling the chuck and samples. The clamping is hydraulic, and the operation is controlled by computer, using TE engineering software. During the test, this software was used to control the rise and fall of the moving beam and input parameters such as the test tensile speed, as well as dynamically displaying the test process and the stress–strain curves. 12 of 19 Materials 2020, 13, 1165 Figure 13. Electronic universal testing machine. Figure 13. Electronic universal testing machine. The material used in this test was 2A16 aluminum alloy sheet. The samples were 100 mm × 100 mm square plates of thickness 1.0 mm. As shown in Figure 14, the diagonal lines drawn on the corners of the plates indicate the places where they were clamped by the chucks. The clamping widths were 30 mm, 40 mm, and 50 mm. Figure 14. Samples before wrinkling. Figure 14. Samples before wrinkling. 3.4.2. Analysis of Experimental Results The wrinkling test was carried out at room temperature, and the wrinkling height was measured as the relative height at the center of the equidistant width L = 25 mm at the intermediate section, using a height measuring instrument. To further verify the trend of variation of the wrinkling height, the measurements obtained were plotted in Figure 15 versus clamping width and tensile length to show the dependences of the wrinkling height on these parameters. Figure 15. Test curves for wrinkling height: (a) Wrinkling height versus tensile length for diﬀerent clamping widths; (b) wrinkling height versus clamping width for diﬀerent tensile lengths. Figure 15. Test curves for wrinkling height: (a) Wrinkling height versus tensile length for diﬀerent clamping widths; (b) wrinkling height versus clamping width for diﬀerent tensile lengths. Materials 2020, 13, 1165 13 of 19 Figure 15a shows that for a given value of the clamping width, the wrinkling height of the samples increased as the tensile length increased from 5 mm to 8 mm to 10 mm. Figure 15b indicates that for a given value of the tensile length, the wrinkling height decreased as the clamping width of the samples increased from 30 mm to 40 mm to 50 mm, which means that the wrinkling becomes less and less obvious. Experimental results indicate that, in general, the wrinkling tendency of warm/hot sheet hydroforming is that the wrinkling height of the sheet decreases with increasing sample clamping width, and it increases with increasing tensile length. Also, the results verify the correction of the numerical simulation of sheet wrinkling performance. 3.4.3. Surface Detection in the Main Diagonal Direction of Test Samples There are insuﬃcient measured data on wrinkling height according to the national standard to fully characterize the wrinkling of diﬀerent samples and determine the wrinkling susceptibility of diﬀerent materials. Therefore, to evaluate the accuracy of the numerical simulation in predicting the wrinkling failure mode of a material, the wrinkling height h was measured at intervals of 10 mm from the center of a square sheet sample in the diagonal direction and was then plotted versus the diagonal ridge line, as shown in Figure 16. Figure 16. Wrinkling height measurement for surface detection: (a) Wrinkling height in the diagonal direction; (b) schematic of measurement procedure. Figure 16. Wrinkling height measurement for surface detection: (a) Wrinkling height in the diagonal direction; (b) schematic of measurement procedure. In these measurements, the imaginary mid-surface of the shell was taken as the measurement surface. According to small-deformation theory, the assumption that the deformation of the sheet was uniform and variations in the thickness of the sheet before and after tension were negligible was applied. The wrinkling height at each measurement point is then: H = h −t (11) (11) where H is wrinkling height of the mid-surface (mm), h is the measured wrinkling height (mm), and t is the sheet thickness (mm). In this test, t was 1 mm. In this test, two group samples with side length L = 100 mm, clamping width B = 30 mm were selected for measurement, with a tensile stroke d = 5 mm. The wrinkling height curves of the two group samples along the tensile diagonal direction under the same loading conditions were obtained, as shown in Figure 17. Figure 17 shows that the results for the two samples were not completely consistent, and there was a deviation of about 2%. There are probably a number of reasons for this, the most likely being that it was due to errors in the test and measurement procedures. To improve the accuracy of experimental data to allow these to serve as a basis for sample comparison, the average value of two tests can be taken, and this was done to verify the subsequent numerical simulation results. 14 of 19 Materials 2020, 13, 1165 Figure 17. Comparison of measurements of wrinkling height of samples in the main diagonal direction. Figure 17. Comparison of measurements of wrinkling height of samples in the main diagonal direction. 3.4.4. Comparison of Calculation Results To evaluate the accuracy of diﬀerent element algorithms in predicting the wrinkling form of sheets, the shape distribution of ridge lines in the main diagonal direction was studied for various simulation conditions, using the surface detection method along the main diagonal described above. Figure 18 shows the results of simulation using the static implicit algorithm with the S8R quadratic element compared with experimental data, and Figure 19 shows the relative error in the results from this algorithm. Figure 18. Comparison of the results of simulation using the S8R quadratic element and experimental data. Figure 18. Comparison of the results of simulation using the S8R quadratic element and experimental data. Figure 19. Relative error in the simulation results using the S8R quadratic element. Figure 19. Relative error in the simulation results using the S8R quadratic element. Materials 2020, 13, 1165 15 of 19 15 of 19 Figures 18 and 19 indicate that for the static simulation algorithm with quadratic elements, the size of the elements has an impact on the simulation results. For an element size of 2 mm, the relative error in the simulation results is about 5%. When the element size is increased to 6 mm, the maximum relative error exceeds 10%. In general, comparison of experimental data and simulation results shows that the static algorithm using quadratic elements has higher accuracy in numerical simulation than the other algorithms, and its results are in good agreement with experimental data. To compare the accuracy of simulation of sheet wrinkling prediction with diﬀerent algorithms, a simulation using the S4R linear element with the static algorithm was also conducted. The simulation results are compared with experimental data in Figure 20, and Figure 21 shows the results of simulation using the dynamic algorithm with the S4R linear element compared with experimental data. Figure 20. Comparison of simulation results using the S4R linear element and the static algorithm with experimental data. Figure 20. Comparison of simulation results using the S4R linear element and the static algorithm with experimental data. Figure 21. Comparison of simulation results using the S4R linear element and the dynamic algorithm with experimental data. Figure 21. Comparison of simulation results using the S4R linear element and the dynamic algorithm with experimental data. Figure 20 shows that the calculation accuracy with the linear element is very poor. For an element size of 2 mm, the error in the highest point is 65.7%. 3.4.4. Comparison of Calculation Results Therefore, it is diﬃcult for an algorithm using the S4R linear element to accurately predict wrinkling during sheet forming process. Figure 21 indicates that the error is even larger than with the static algorithm and that it gets worse with decreasing Materials 2020, 13, 1165 16 of 19 element size. This further conﬁrms that the S4R element is not suitable for use in element predicting wrinkling in the sheet forming process. 3.5. Evaluation of Accuracy of Sheet Wrinkling Test On account of the YBT has long been the basic test to measure the wrinkling properties of sheets, to verify the accuracy of the modiﬁed YBT, tests on diﬀerent materials were conducted. Sheets of diﬀerent materials have diﬀerent wrinkling heights under the same conditions, and so numerical simulations of wrinkling properties and comparisons of wrinkling heights on the ridge line in the main tensile direction of diﬀerent materials were carried out. Wrinkling tests on sheets of 5A06 and 2A16 aluminum alloys, 2198 aluminum-lithium alloy, and TA1 titanium alloy were performed to evaluate the wrinkling performance of these materials. Stress–strain curves at normal temperature are shown in Figure 22. Figure 22. Stress-strain curves of several materials. Figure 22. Stress-strain curves of several materials. The S8R shell element algorithm was again used for the simulations. The samples were 100 mm × 100 mm square sheets of thickness 1.0 mm, and the clamping width was 30 mm. The tensile stroke was 5mm, and the other test conditions were as in the earlier experiments. Curves of ridge line height versus center distance in the main tensile direction are shown in Figure 23, and the maximum wrinkling height distributions are shown in Figure 24. Figure 23. Ridge line height versus center distance in the principal tensile direction. Figure 23. Ridge line height versus center distance in the principal tensile direction. 17 of 19 Materials 2020, 13, 1165 1165 Figure 24. Maximum wrinkling height distributions of diﬀerent materials. Figure 24. Maximum wrinkling height distributions of diﬀerent materials. Figures 23 and 24 show that the behavior of the wrinkling height was similar for the diﬀerent materials, although there were some diﬀerences. The 5A06 aluminum alloy sheet had the poorest resistance to wrinkling, and its maximum wrinkling height was the greatest. The 2198 aluminum-lithium alloy sheet had the greatest resistance to wrinkling, and its wrinkling height distribution was the most ideal. The diﬀerence in maximum wrinkling height among the materials was about 4.39%, so it can be considered that YBT performs well in predicting and evaluating the wrinkling of sheets. These results show that the YBT is of great importance for predicting wrinkling of sheets. They also show that the proposed equation for predicting wrinkling of sheets is suitable for warm/hot sheet hydroforming when the ﬂuid pressure is zero. Author Contributions: Funding acquisition, G.C.; Methodology, G.C., D.Z., Y.Y., and L.L.; Software, Y.Y., and J.F.; Writing—original draft, G.C., and J.F.; Data curation, J.F., and J.Y.; Formal analysis, G.C., and S.A.; Experiment, G.C., J.F., and J.Y.; Project administration, Y.Y.; Writing—review & editing, G.C., D.Z., L.L., and S.A. All authors have read and agreed to the published version of the manuscript. 3.5. Evaluation of Accuracy of Sheet Wrinkling Test This can provide a basis for subsequent study of the prediction of sheet wrinkling under ﬂuid pressure and temperature, as well as providing basic test data for the further study of wrinkling during the forming process. Funding: This research was funded by Zhejiang Provincial Natural Science Foundation of China, grant NO. LQ18E050010; and the Scientiﬁc Research Foundation of Zhejiang Sci-Tech University, grant NO. 17022073-Y. Conﬂicts of Interest: The authors declare no conﬂict of interest. s of Interest: The authors declare no conﬂict of interest. References 1. Hill, R. A theory of the yielding and plastic flow of anisotropic metals. Proc. R. Soc. Lond. Ser. A 1948, 193, 281–297. 2. Koç, M.; Billur, E.; Necati, Ö.N. An experimental study on the comparative assessment of hydraulic bulge test analysis methods. Mater. Des. 2011, 32, 272–281. [CrossRef] 2. Koç, M.; Billur, E.; Necati, Ö.N. An experimental study on the comparative assessment of hydraulic bulge test analysis methods. Mater. Des. 2011, 32, 272–281. [CrossRef] 3. Yang, X.Y.; Lang, L.H.; Liu, K.N.; Liu, B.S. Mechanics analysis of axisymmetric thin-walled part in warm sheet hydroforming. Chin. J. Aeronaut. 2015, 28, 1546–1554. [CrossRef] 3. Yang, X.Y.; Lang, L.H.; Liu, K.N.; Liu, B.S. Mechanics analysis of axisymmetric thin-walled part in warm sheet hydroforming. Chin. J. Aeronaut. 2015, 28, 1546–1554. [CrossRef] 4. Zhou, G.; Wang, Y.N.; Lang, L.H. Accuracy analysis of complex curvature parts based on the rigid-ﬂexible hydroforming. Int. J. Adv. Manuf. Technol. 2018, 99, 247–254. [CrossRef] 5. Mahabunphachai, S.; Koç, M. Investigations on Forming of Aluminum 5052 and 6061 Sheet Alloys at Warm Temperatures. Mater. Des. 2010, 31, 2422–2434. [CrossRef] 6. Palumbo, G.; Piccininni, A.; Guglielmi, P.; Michele, G.D. Warm hydroforming of the heat treatable aluminium alloy AC170PX. J. Manuf. Process. 2015, 20, 24–32. [CrossRef] 7. Liu, K.N.; Lang, L.H.; Cai, G.S.; Yang, X.Y.; Guo, C.; Liu, B.S. A novel approach to determine plastic hardening curves of AA7075 sheet utilizing hydraulic bulging test at elevated temperature. Int. J. Mech. Sci. 2015, 100, 328–338. [CrossRef] 8. Kaya, S.; Altan, T.; Groche, P.; Klöpsch, C. Determination of the Flow Stress of Magnesium Az31-O Sheet at Elevated Temperatures Using the Hydraulic Bulge Test. Int. J. Mach. Tools Manuf. 2008, 48, 550–557. [CrossRef] 9. Liu, S.C.; Lang, L.H.; Guan, S.W. An investigation into the formability and processes of GLARE materials using hydro-bulging test. Int. J. Precis. Eng. Manuf. 2019, 20, 121–128. 10. Cai, G.S.; Wu, C.Y.; Gao, Z.P.; Lang, L.H.; Sergei, A. Research on Al-alloy sheet forming formability during warm/hot sheet hydroforming based on elliptical warm bulging test. AIP Adv. 2018, 8, 055023. [CrossRef] 11. Cai, G.S.; Wu, C.Y.; Gao, Z.P.; Lang, L.H.; Sergei, A. Investigation on the eﬀect of pressure rate on formability of aluminum alloy during warm/hot sheet hydroforming. AIP Adv. 2018, 8, 095313. [CrossRef] 12. Liu, B.S.; Lang, L.H.; Zeng, Y.S.; Lin, J.G. Forming characteristic of sheet hydroforming under the inﬂuence of through-thickness normal stress. J. Mater. Process. Technol. 2012, 212, 1875–1884. 4. Conclusions Using a novel approach based on an improved YBT, wrinkling of sheets in warm/hot sheet hydroforming was predicted using theoretical analysis, numerical simulation, and experimental veriﬁcation. The general conclusions of this study can be summarized as follows: (1) A novel experimental method based on an improved YBT was proposed to test the wrinkling properties of sheets under through-thickness normal stress provided by ﬂuid pressure at diﬀerent temperatures. A wrinkling coeﬃcient Lλ that expresses the ability of a sheet to resist wrinkling during the hydroforming process was also presented. (1) A novel experimental method based on an improved YBT was proposed to test the wrinkling properties of sheets under through-thickness normal stress provided by ﬂuid pressure at diﬀerent temperatures. A wrinkling coeﬃcient Lλ that expresses the ability of a sheet to resist wrinkling during the hydroforming process was also presented. (2) The wrinkling test using the proposed method for sheet wrinkling properties was carried out, and results indicated that the wrinkling height of the sheet decreased with the increasing sample clamping width, and it increased with the increasing tensile length. (2) The wrinkling test using the proposed method for sheet wrinkling properties was carried out, and results indicated that the wrinkling height of the sheet decreased with the increasing sample clamping width, and it increased with the increasing tensile length. (3) Based on numerical simulation, the measurement method that Ridge-height curves measured along the main diagonal tensile direction of sheets was presented. Results indicated that the wrinkling prediction criterion allows good discrimination. (4) The wrinkling properties of several diﬀerent materials were simulated and analyzed, and the results showed that the improved YBT gave good predictions for wrinkling in the conventional sheet metal forming process, while the prediction results for wrinkling in warm/hot sheet hydroforming were also accurate when the ﬂuid pressure was zero. Author Contributions: Funding acquisition, G.C.; Methodology, G.C., D.Z., Y.Y., and L.L.; Software, Y.Y., and J.F.; Writing—original draft, G.C., and J.F.; Data curation, J.F., and J.Y.; Formal analysis, G.C., and S.A.; Experiment, G.C., J.F., and J.Y.; Project administration, Y.Y.; Writing—review & editing, G.C., D.Z., L.L., and S.A. All authors have read and agreed to the published version of the manuscript. Materials 2020, 13, 1165 18 of 19 18 of 19 References [CrossRef] 13. Lang, L.H.; Cai, G.S.; Liu, K.N.; Sergei, A.; Du, P.M.; Zheng, H. Investigation on the eﬀect of through thickness normal stress on forming limit at elevated temperature by using modiﬁed M-K model. Int. J. Mater. Form. 2015, 8, 211–228. [CrossRef] 14. Lang, L.H.; Du, P.M.; Liu, B.S.; Cai, G.S.; Liu, K.N. Pressure rate controlled uniﬁed constitutive equations based on microstructure evolution for warm hydroforming. J. Alloys Compd. 2013, 574, 41–48. [CrossRef] 15. Ji, H.C.; Liu, J.P.; Wang, B.Y.; Tang, X.F.; Lin, J.G.; Huo, Y.M. Microstructure evolution and constitutive equations for the high-temperature deformation of 5Cr21Mn9Ni4N heat-resistant steel. J. Alloy. Compd. 2017, 693, 674–687. [CrossRef] 6. Mei, H.; Lang, L.H.; Liu, K.N.; Yang, X.G. Evaluation study on iterative inverse modeling procedure for determi post-necking hardening behavior of sheet metal at elevated temperature. Metals 2018, 8, 1044. [CrossRef] 17. He, D.G.; Lin, Y.C.; Chen, J.; Chen, D.D.; Huang, J.; Tang, Y.; Chen, M.S. Microstructural evolution and support vector regression model for an aged Ni-based superalloy during two-stage hot forming with stepped strain rates. Mater. Des. 2018, 154, 51–62. [CrossRef] 18. Cai, G.S.; Lang, L.H.; Liu, K.N.; Sergei, A.; Zhang, D.X.; Yang, X.Y.; Guo, C. Research on the eﬀect of ﬂow stress calculation on aluminum alloy sheet deformation behavior based on warm bulging test. Met. Mater. Int. 2015, 21, 365–373. [CrossRef] 19. Lin, Y.C.; Luo, S.C.; Yin, L.X.; Huang, J. Microstructural evolution and high temperature ﬂow behaviors of a homogenized Sr-modiﬁed Al-Si-Mg alloy. J. Alloy. Compd. 2018, 739, 590–599. [CrossRef] 20. Cai, G.S.; Zhou, X.J.; Lang, L.H.; Sergei, A. Research on aluminum alloy sheet thermoplastic deformation behavior based upon warm bulging test. AIP. Adv. 2016, 6, 025023. [CrossRef] 21. Wang, Y.; Yang, L.; Bai, B.; Lang, L.H. Evaluation of limit deformation behavior in hydro-bulging of the double-layer sheet metal using diffuse and localized instability theories. Int. J. Mech. Sci. 2019, 150, 145–153. [CrossRef] 22. Belhassen, L.; Koubaa, S.; Wali, M.; Dammak, F. Numerical prediction of springback and ductile damage in rubber-pad forming process of aluminum sheet metal. Int. J. Mech. Sci. 2016, 117, 218–226. [CrossRef] 23. Chanhee, W.; Hyung, K.; Seokryul, L.; Dongjin, K.; Sunghyuk, P.; Jonghun, Y. Wrinkling prediction for GPa-grade steels in sheet metal forming process. Int. J. Adv. Manuf. Technol. 2019, 102, 9–12. 19 of 19 19 of 19 Materials 2020, 13, 1165 24. Zhou, B.J.; Xu, Y.C. References The eﬀect of upper sheet on wrinkling and thickness distribution of formed sheet part using double-layer sheet hydroforming. Int. J. Adv. Manuf. Technol. 2018, 99, 1175–1182. [CrossRef] 25. Lu, H.; Kearney, M.; Wang, C.; Liu, S.; Meehan, P.A. Part accuracy improvement in two point incremental forming with a partial die using a model predictive control algorithm. Precis. Eng. 2017, 49, 179–188. [CrossRef] 26. Chen, Y.Z.; Liu, W.; Zhang, Z.C.; Xu, Y.C.; Yuan, S.J. Analysis of wrinkling during sheet hydroforming of curved surface shell considering reverse bulging eﬀect. Int. J. Mech. Sci. 2017, 120, 70–80. [CrossRef] 26. Chen, Y.Z.; Liu, W.; Zhang, Z.C.; Xu, Y.C.; Yuan, S.J. Analysis of wrinkling during sheet hydroforming of curved surface shell considering reverse bulging eﬀect. Int. J. Mech. Sci. 2017, 120, 70–80. [CrossRef] 27. Choi, H.; Koc, M.; Ni, J. A study on the analytical modeling for warm hydro-mechanical deep drawing of lightweight materials. Int. J. Mach. Tools. Manuf. 2007, 47, 1752–1766. [CrossRef] 28. Liu, K.N.; Lang, L.H.; Zhang, W.S.; Marai, M.; Liu, B.S. Coupled Eulerian-Lagrangian simulation of granular medium sheet forming process and experimental investigation at elevated temperature. Int. J. Adv. Manuf. Technol. 2017, 88, 2871–2882. [CrossRef] 29. Khosrojerdi, E.; Bakhshi-Jooybari, M.; Gorji, A.; Hosseinipour, S.J. Experimental and numerical analysis of hydrodynamic deep drawing assisted by radial pressure at elevated temperatures. Int. J. Adv. Manuf. Technol. 2017, 88, 185–195. [CrossRef] 30. Rizwan, Z.; Lang, L.H.; Zhang, R.J. Analysis of hydro-mechanical deep drawing and the effects of cavity pressure on quality of simultaneously formed three-layer Al alloy parts. Int. J. Adv. Manuf. Technol. 2015, 80, 2117–2128. 30. Rizwan, Z.; Lang, L.H.; Zhang, R.J. Analysis of hydro-mechanical deep drawing and the effects of cavity pressure on quality of simultaneously formed three-layer Al alloy parts. Int. J. Adv. Manuf. Technol. 2015, 80, 2117–2128. 31. Cai, G.S.; Wu, C.Y.; Zhang, D.X. Investigation on the eﬀect of type of cooling on the properties of aluminum alloy during warm/hot hydromechanical deep drawing. Symmetry 2018, 10, 362. [CrossRef] 31. Cai, G.S.; Wu, C.Y.; Zhang, D.X. Investigation on the eﬀect of type of cooling on the properties of alu alloy during warm/hot hydromechanical deep drawing. Symmetry 2018, 10, 362. [CrossRef] 32. Yoshida, K. Purpose and feature of the Yoshida Buckling Test (YBT). J. Jpn. Soc. Technol. Plast. 1983, 24, 901–908. 32. Yoshida, K. Purpose and feature of the Yoshida Buckling Test (YBT). J. Jpn. Soc. Technol. Plast. 1983, 24, 901–908. 33. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). References Hayashi, H.; Hirata, M.; Yoshida, K. Buckling behavior of a square specimen in the diagonal tensile test and its application. J. Jpn. Soc. Technol. Plast. 1978, 24, 1298–1304. 33. Hayashi, H.; Hirata, M.; Yoshida, K. Buckling behavior of a square specimen in the diagonal tensile test and its application. J. Jpn. Soc. Technol. Plast. 1978, 24, 1298–1304. 34. Kim, J.B.; Yang, D.Y. Finite element analysis of the wrinkling initiation and growth in modiﬁed Yoshida buckling test. Korean J. Met. Mater. 1998, 4, 640–647. [CrossRef] 35. Kim, J.B.; Yoon, J.W.; Yang, D.Y.; Barlat, F. Investigation into wrinkling behavior in the elliptical cup deep drawing process by ﬁnite element analysis using bifurcation theory. J. Mater. Process. Technol. 2001, 111, 170–174. [CrossRef] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
W4387539162.txt	https://journals.wlb-stuttgart.de/ojs/index.php/wfr/article/download/8053/7932	de		Rezension von: Stadt Heubach (Hrsg.), Heubach und die Burg Rosenstein	Württembergisch Franken	2,023	cc-by	458	342 Neue Bücher ster erschlossen, enthalten zahlreiche Informationen, meist stichwortartig formuliert, für die Dürr für die Zeit vor nikalische Aufschriebe ab 1800 die sogenannten »Heilbronner Weinbüchlein«, chro- 1519, sowie verschiedene andere stadtgeschichtliche Arbei- ten benutzte. Für die Bearbeitung des 3. Bandes wurde beim Stadtarchiv eine zusätzliche Planstelle eingerichtet, die mit Werner Föll besetzt wurde. Dieser stellt in einer umfangreichen Einleitung die politische Situation in Heilbronn zur Zeit der Weimarer Republik und die im Vordergrund der Arbeit von Stadtverwaltung und Gemeinderat stehenden Aufgaben dar. Ausführlich geht er auf das wirtschaftliche und das kulturelle Leben in der Stadt ein. Für die Bearbeitung des Chronikteils bedient sich Föll umfangreicher Vorarbeiten von Friedrich und Willy Dürr, Karl Wulle und Helmut Schmolz. Trotz der vom Herausgeber ausführlich dargestellten langen »Entstehungs- und Vorgeschichte« des Buches ist dennoch »der Typus der reinen Chronik, aus und in der Zeit geschrieben, vorrangig erhalten geblieben«, allerdings durch spätere in Forschungen Primärquellen »angereichert«. Die Fülle der nahezu über jeden Tag in den umfangreichen Band aufgenommenen Informationen wird wohltuend unterbrochen durch die über den gesamten Text verteilten 108 Abbildungen, die das sorgfältig dokumentierte Zeitgeschehen beispielhaft illustrieren. Das Stadtarchiv Heilbronn hat mit der Herausgabe dieser nunmehr bis zum Beginn des Dritten Reiches lückenlos vorliegenden dreibändigen Stadtchronik ein großartiges, für ähnliche Vorhaben anderer Städte in allen Teilen mustergültiges Werk geschaffen. M. Heubach und die Burg Rosenstein. Akermann Geschichte, Tradition, Landschaft. Hrsg.: Stadt Heubach. Schwäbisch Gmünd: Einhorn 1984. 412 S., 269 Abb. In Heimatbüchern wird die Heimat meist chronologisch vorgestellt, d. h. also: man be- ginnt mit der Geologie, mit den Zeugen der frühen Erdzeitalter, und man endet in der Gegenwart. Das vorliegende Buch beginnt erst mit der Vorgeschichte von Ort und Umgebung, schildert dann chronologisch das Geschehen über Mittelalter und Industriezeitalter hinweg bis zur Gegenwart. An das nachfolgende schließen sich erst am Ende naturbezogene Themen aus volkskundliche Kapitel der Gegenwart an, in die geologische Pflichtstoff knapp einbezogen wurde. Was an sticht, sind die scheinbar fließenden Übergänge von einem diesem zum Heimatbuch der be- anderen historischen Thema. Berücksichtigt sind hier auch die zwei Partnerstädte in Frankreich und Österreich mit kurzen historischen Abrissen. Ein Heimatbuch mit gut ebenso ausgeführter Gliederung. durchdachter und E. Pastor Stephan Keck; Haucke Schmitt: 500 Jahre Heerbergskirche. Eine Festschrift. Hrsg.: Evangelische Kirchengemeinde Laufen/Kocher. 0.E., 96 S., Abb. Die Jubiläumsschrift enthält einen ausführlichen, urkundlich erarbeiteten Geschichtsabriß der einstigen Wallfahrtskirche. Außerdem eine kunstverständige, bebilderte Be- schreibung des Flügelaltars von Zeytblom, der Figuren von Jörg Syrlin d. J. und der Alabastermadonna und ihrer wechselreichen Unterbringungsgeschichte. Die Kunst- werke selber sind heute nicht mehr in der Kirche, sondern im Alten Schloß Stuttgart, in der Schloßkirche werden dankbar Untergröningen und die Aufsätze lesen, im Kernerhaus darunter auch einen Weinsberg. Kunstfreunde von Justinus Kerner, der damals auf dem Heerberg den Altarmaler Zeytblom für die Kunstgeschichte entdeckte.
https://openalex.org/W2974435767	http://journal.stiem.ac.id/index.php/jureq/article/download/322/246	Indonesian	null	PENGARUH AUDIT OPRASIONAL TERHADAP PERSEDIAAN BARANG DAGANG PADA USAHA DEPOT JAYA MOTOR KOTA PALOPO	Equilibrium/Equilibrium	2,019	cc-by-sa	5,369	PENGARUH AUDIT OPRASIONAL TERHADAP PERSEDIAAN BARANG DAGANG PADA USAHA DEPOT JAYA MOTOR KOTA PALOPO Junaidi1, Huldi1 1Sekolah Tinggi Ilmu Ekonomi Muhammadiyah Palopo Junaidi@stiem.ac.id Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 eISSN 2684-9313 pISSN 2088-7485 Abstract This study aims to determine whether operational audits influence the adequacy of merchandise inventory. Data used in this study are primary data, namely data sourced from observations and questionnaires conducted at Jaya Motor Depot Business, Palopo City. To test data reliability, validity test and reliability test are used. The statistical analysis used is simple linear regression to test the influence of the dependent and independent variables. The test results obtained by Operational Audit t count of 5.518 and t table of 2.131 indicates that the value of tcount> t table with a significance level of 0.000 <0.05 then H0 is rejected and H1 is accepted. This shows that operational audits have a significant influence on the adequacy of merchandise inventory. Keywords: Operational Audit Adequacy of Merchandise Inventory g f f q y f y Keywords: Operational Audit, Adequacy of Merchandise Inventory Abstrak : Penelitian ini bertujuan untuk mengetahui apakah audit operasional berpengaruh terhadap kecukupan persediaan barang dagang.Data yang digunakan dalam penelitian ini merupakan data primer yaitu data yang bersumber dari hasil observasi dan kuesioner yang dilakukan pada Usaha depot Jaya Motor Kota Palopo. Untuk menguji keandalan data maka digunakan uji validitas dan Uji reliabilitas. Analisis statistik yang digunakan yaitu regresi linear sederhana dilakukan untuk menguji pengaruh antara variabel dependen dan independen. Hasil pengujian diperoleh thitung Audit Operasional sebesar 5.518 dan ttabel sebesar 2.131 ini menunjukkan bahwa nilai thitung > ttabel dengan tingkat signifikansi 0.000<0.05 maka H0 ditolak dan H1 diterima. Hal ini menunjukkan bahwa audit operasional memiliki pengaruh yang signifikan terhadap kecukupan persediaan barang dagang. Kata Kunci: Audit Operasional, Kecukupan Persediaan Barang Dagang PENDAHULUAN efisien. Apabila belum dilaksanakan seperti sehrusnya, maka auditor akan memberikan rekomendasi atau saran agar pada masa akan datang menjadi lebih baik, yakni dapat memperbaiki kinerja serta pengendalian intern dan mengangkat efisiensi operasional perusahaan. PENDAHULUAN Semakin berkembangnya perusahaan menuntut pula perkembangan di bidang pemeriksaan. Pemeriksaan yang dilakukan tidak hanya pemeriksaan keuangan tetapi juga pemeriksaan yang menekankan pada penilaian sistematis dan objektif serta berorietasi pada tujuan untuk memperoleh keyakinan tentang keefektifan dan memberikan pendapat atas kewajaran laporan keuangan yang diperiksa. Pimpinan perusahaan memerlukan audit operasional yang menyajikan informasi mengenai aktivitas operasional perusahaan dan tidak terbatas pada informasi keuangan dan akuntansi saja, suatu organisasi yang dikendalikan dengan efektif terletak pada sikap manajemen. Manajemen adalah pihak yang mengelola serta mengendalikan perusahaan. Semakin berkembangnya perusahaan menuntut pula perkembangan di bidang pemeriksaan. Pemeriksaan yang dilakukan tidak hanya pemeriksaan keuangan tetapi juga pemeriksaan yang menekankan pada penilaian sistematis dan objektif serta berorietasi pada tujuan untuk memperoleh keyakinan tentang keefektifan dan memberikan pendapat atas kewajaran laporan keuangan yang diperiksa. Pimpinan perusahaan memerlukan audit operasional yang menyajikan Setiap perusahaan dalam melaksanakan kegiatan pemasaran,tidak terlepas akan kebutuhan barang dagangan yang menjadi faktor utama dalam menunjang jalannya aktivitas pemasaran perusahaan. Dengan terpenuhinya akan barang tepat pada waktunya, maka kegiatan suatu perusahaan akan dapat berjalan dengan lancar sesuai dengan tujuan yang ingin dicapai. Tanpa adanya persediaan barang dagang, perusahaan akan menghadapi risiko dimana pada suatu waktu tidak dapat memenuhi keinginan dari para pelanggannya. informasi mengenai aktivitas operasional perusahaan dan tidak terbatas pada informasi keuangan dan akuntansi saja, suatu organisasi yang dikendalikan dengan efektif terletak pada sikap manajemen. Manajemen adalah pihak yang mengelola serta mengendalikan perusahaan. Audit operasional merupakan evaluasi atas berbagai kegiatan operasional perusahaan, sedangkan sasarannya adalah untuk menilai apakah pelaksanaan kegiatan operasional telah dilaksanakan secara ekonomis, efektif dan Berdasarkan pada penelitian pendahuluan yang dilakukan penulis pada tanggal 26, Januari, 2015 di Usaha Depot Jaya Motor diketahui bahwa: Sebagai perusahaan dagang, Usaha Depot Jaya Motor juga menghadapi masalah- 44 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 eISSN 2684-9313 pISSN 2088-7485 eISSN 2684-9313 pISSN 2088-7485 audit yang sangat terkenal adalah definisi yang berasal dari ASOBAC (A Statement Of Basic Auditing Concepts) yang mendefinisikan auditing sebagai berikut: masalah yang berkaitan dengan penyediaan barang dagang yang diinginkan konsumen. Perusahaan terkadang tidak dapat memenuhi permintaan barang yang diinginkan konsumen sehingga mereka harus menunggu perusahaan untuk memesan barang dari pihak distributor. PENDAHULUAN Namun, sebagai perusahaan yang menjual barang bekas tidaklah mudah untuk mendapatkan barang yang diinginkan konsumen. Perusahaan tersebut harus menunggu barang hasil pelelangan dari Adira Finance. “Auditing adalah suatu proses sistematis untuk menghimpun dan mengevaluasi bukti- bukti secara objektif mengenai asersi-asersi tentang berbagai tindakan dan kejadian ekonomi untuk menentukan tingkat kesesuaian antara asersi-asersi tersebut dengan kriteria yang telah ditentukan dan menyampaikan hasilnya kepada para pemakai yang telah berkepentingan”. Masalah lain yaitu, perusahaan sering mengambil barang dagang tanpa mempertimbangkan bahwa persediaan barang yang ada masih banyak sehingga jika dalam satu bulan persediaan yang ada semua tidak terjual maka perusahaan harus menggunakan modal sendiri untuk membayar uang angsuran kepada Adira Finance. Oleh karena itu, maka penulis tertarik untuk meneliti tentang: Pengaruh Audit Operasional Terhadap kecukupan Persediaan Barang Dagang. Menurut Arens, Elder,Beasley (2004:15): “auditing adalah pengumpulan serta pengevaluasian bukti-bukti atas informasi untuk menentukan dan melaporkan tingkat kesesuaian informasi tersebut dengan kriteria- kriteria yang telah ditetapkan. Audit harus dilaksanakan oleh seseorang yang kompeten dan independen”. Agoes (2012:4) mendefinisikan auditing sebagai: “suatu proses sistematis yang secara objektif mendapatkan dan mengevaluasi bukti mengenai asersi tentang kegiatan-kegiatan dan kejadian-kejadian ekonomi untuk meyakinkan tingkat keterkaitan antara asersi tersebut dan kriteria yang telah ditetapkan dan mengkomunikasikan hasilnya kepada pihak- pihak yang berkepentingan”. Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 eISSN 2684-9313 pISSN 2088-7485 eISSN 2684-9313 pISSN 2088-7485 Persekutuan Firma ( FA ) atau Perusahaan perseorangan. Kondisi tertentu satuan Perseorangan, tetapi dalam kondisi tertentu satuan juga biasa terbentuk divisi atau departemen periode waktu yang diaudit umumnya satu tahun, tetapi ada juga yang satu bulan, satu kuartal atau beberapa tahun 1) Pengumpulan atau pengevaluasian bukti, adanya bukti – bukti yang memadai baik dari segi jumlah maupun dari segi menu sangat diperlukan untuk menentukan kegiatan audit. Bahan bukti dapat terdiri dari bermacam bentuk yang berbeda termasuk peringatan lisan yang berbeda ( Klien ), komunitas dengan pihak ketiga dan hasil pengamatan auditor 2) Auditor harus independen dan kompeten, independen berarti bebas dari pengaruh – pengaruh hingga batas – batas tertentu. Sedangkan kompeten berarti auditor harus mempunyai pengetahuan dan pengalaman yang cukup agar dapat memahami kriteria yang dipergunakan. 3) Pelaporan hasil audit harus mampu memberikan informasi mengenai kesesuaian informasi yang diperiksa dengan kriteria yang telah ditetapkan. komputer, metode produksi, pemasaran dan bidang – bidang lain sesuai dengan keahlian auditor. c) Audit ketaatan ( Compliance Audit ) Audit Ketaatan merupakan proses yang menilai bukti untuk menentukan apakah aktivitas keuangan dalam operasi tertentu dari suatu entitas sesuai dengan kondisi yang ditetapkan dalam peraturan dan kebijakan. Audit ketaatan bertujuan untuk mempertimbangkan apakah auditor atau klien telah mengikuti aturan tertentu yang telah ditetapkan pihak yang memiliki orientasi lebih tinggi. Hasil audit ketaatan biasanya tidak dilaporkan kepada pihak luar tetapi kepada pihak tertentu yang paling berkepentingan atas organisasi adalah pihak yang paling berkepentingan atas terpenuhinya prosedur dan aturan yang telah ditetapkan. Auditing Auditing adalah sebuah proses sistematis yang dilakukan oleh seseorang yang memiliki kompetensi dan independensi, mengenali perolehan dan penilaian atau bukti secara objektif yang dilakukan dengan pengumpulan dan penilaian atas bukti-bukti informasi yang dapat dikuantifikasikan dan terkait dengan entitas suatu ekonomi tertentu, berkenaan dengan pernyataan mengenai tindakan- tindakan dan kejadian-kejadian ekonomi dengan tujuan untuk menentukan tingkat kesesuaian antara pernyataan tersebut dengan kriteria yang telah ditetapkan serta untuk mengkomunikasikan hasil-hasilnya kepada pihak-pihak yang berkepentingan (Purwono,2004:18). Dari definisi-definisi tersebut di atas, dapat disimpulkan bahwa audit adalah menyangkut hal-hal 1) Informasi-informasi yang dapat diukur dan kriteria-kriteria yang telah ditetapkan syarat dalam melakukan pemeriksaan adalah informasi yang terpercaya atau dapat dibuktikan kebenarannya dan kriteria standar yang dapat digunakan oleh auditor sebagai pedoman dalam mengevaluasi informasi-informasi tersebut. 2) Entitas ekonomi, ruang lingkup tanggung jawab auditor yang jelas mengenai penerapan entitas ekonomi dan periode waktu diaudit. Entitas ekonomi merupakan satuan legal misalnya perseroan terbatas (PT), Lembaga Pemerintah Persekutuan komenditar (CV), Koperasi, Menurut Halim (2003 :1) untuk mempelajari auditing dan profesi akuntan secara mendalam, perlu kiranya mengetahui definisi audit. Definisi 45 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 Dalam kenyataannya mengembangkan rekomendasi merupakan salah satu aspek yang paling menantang dalam audit operasional. Tahapan dasar dalam melaksanakan audit operasional menurut Bayangkara (2008), yaitu: a) Audit pendahuluan dilakukan untuk mendapatkan informasi latar belakang terhadap objek yang diaudit, di samping juga dilakukan penelaahaan terhadap kebijakan berkaitan dengan aktifitas yang diaudit, serta menganalisis berbagai informasi yang telah diperoleh untuk mengidentifikasikan hal-hal yang potensial mengandung kelemahan pada perusahaan yang diaudit. b) Review dan pengujian pengendalian manajemen Pada tahapan ini auditor melakukan review dan pengujian terhadap pengendalian manajemen objek audit, dengan tujuan untuk menilai efektifitas pengendalian manajemen dalam mendukung pencapaian tujuan perusahaan. Hasil pengujian pengendalian manajemen ini dapat mendukung tujuan audit sementara menjadi tujuan audit yang sesungguhnya (definitive audit objective) atau mungkin ada beberapa tujuan audit sementara yang gugur. Karena tidak cukup bukti untuk mendukung tujuan audit tersebut. c) Audit terinci pada tahap ini auditor melakukan pengumpulan bukti yang cukup dan kompoten untuk mendukung tujuan audit yang telah ditentukan. Pada tahap ini dilakukan pengembangan temuan untuk mencari keterkaitan antara satu temuan dengan temuan yang lain dalam menguji permasalahan yang berkaitan dengan tujuan audit. Temuan yang cukup relevan, dan kompoten dalam tahap ini disajikan dalam suatu Kertas Kerja Audit (KKA) Untuk mendukung simpulan audit yang dibuat dan rekomendasi yang diberikan. Menurut Arens, Elder, Beasley,(2004) pada dasarnya audit operasional terbagi menjadi 3 jenis yaitu: fungsional,organisasi, penugasan khusus. Ketiga jenis audit operasional itu dapat di uraikan Fungsional Audit fungsional berkaitan dengan sebuah fungsi atau lebih dalam suatu operasional. Ini dapat berhubungan misalnya dengan fungsi penggajian suatu divisi atau untuk perusahaan secara keseluruhan. Keunggulan audit fungsional adalah memungkinkan adanya spesialisasi oleh auditor, kekurangan audit Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 eISSN 2684-9313 pISSN 2088-7485 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 eISSN 2684-9313 pISSN 2088-7485 fungsional adalah tidak dievaluasinya fungsi yang saling berkaitan. keuangan, audit operasional mencakup serangkaian langkah atau prosedur yang terstruktur dan terorganisasi. Aspek ini mencakup perencanaan yang tepat dan juga mendapatkan secara objektif, menilai bukti yang berkaitan dengan aktifitas yang diaudit. a) Penilaian operasi organisasi yang berdasarkan pada suatu kriteria yang ditetapkan atau disetujui. Dalam audit operasional, kriteria sering dinyatakan dalam standar kinerja (performance standars) yang ditetapkan industri. Audit operasional mengukur tingkat hubungan antara kinerja aktual dengan kriteria. b) Tujuan utama dari audit operasional adalah membantu manajemen dari organisasi yang diaudit untuk memperbaiki effectiveness, efficiency, dan economy dari suatu operasi. Dengan demikian, audit operasional memfokuskan prinsip pada masa yang akan datang. Ini berlawanan langsung dengan audit keuangan yang mempunyai fokus historis. c) Penerima yang tepat dari laporan audit operasional adalah manajemen atau individu yang meminta diadakannya audit. Kecuali jika diminta oleh pihak ketiga, pembagian laporan tetap dalam entitas. Dalam kebanyakan hal, dewan komisaris atau panitia audit menerima copy laporan audit operasional. d) Tidak seperti audit laporan keuangan, suatu audit operasional tidak berakhir dengan laporan atas temuan. Audit operasional memperluas dengan memberikan rekomendasi untuk perbaikan. Dalam kenyataannya mengembangkan rekomendasi merupakan salah satu aspek yang paling menantang dalam audit operasional. keuangan, audit operasional mencakup serangkaian langkah atau prosedur yang terstruktur dan terorganisasi. Aspek ini mencakup perencanaan yang tepat dan juga mendapatkan secara objektif, menilai bukti yang berkaitan dengan aktifitas yang diaudit. a) Penilaian operasi organisasi yang berdasarkan pada suatu kriteria yang ditetapkan atau disetujui. Dalam audit operasional, kriteria sering dinyatakan dalam standar kinerja (performance standars) yang ditetapkan industri. Audit operasional mengukur tingkat hubungan antara kinerja aktual dengan kriteria. b) Tujuan utama dari audit operasional adalah membantu manajemen dari organisasi yang diaudit untuk memperbaiki effectiveness, efficiency, dan economy dari suatu operasi. Dengan demikian, audit operasional memfokuskan prinsip pada masa yang akan datang. Ini berlawanan langsung dengan audit keuangan yang mempunyai fokus historis. c) Penerima yang tepat dari laporan audit operasional adalah manajemen atau individu yang meminta diadakannya audit. Kecuali jika diminta oleh pihak ketiga, pembagian laporan tetap dalam entitas. Dalam kebanyakan hal, dewan komisaris atau panitia audit menerima copy laporan audit operasional. d) Tidak seperti audit laporan keuangan, suatu audit operasional tidak berakhir dengan laporan atas temuan. Audit operasional memperluas dengan memberikan rekomendasi untuk perbaikan. Audit Operasional Audit operasional seringkali diartikan sama dengan audit manajemen. Pengertian sederhana dari audit manajemen adalan investigasi dari suatu organisasi dalam semua aspek kegiatan manajemen dari yang paling tinggi sampai kebawah dan pembuatan laporan audit mengeni efektivitasnya atau dari segi probilitas dan efisiensi kegiatan bisnisnya sedangkan pengertian sederhana dari audit operasional adalah uraian efektivitas perusahaan yang sistematis dalam hubungannya dengan tujuan untuk melihat, mengidentifikasi peluang perbaikan, atau mengembangkan rekomendasi untuk perbaikan. Jelas kedua pengertian serupa karena pemeriksaan manajemen dilakukan saat manajemen beroperasi (Boynton, 2003). Jenis – jenis Audit Menurut Boynton, Jonhson, Kell ( 2003 : 490 ) ada tiga jenis audit a) Pelaporan Keuangan ( Financial Audit ) Audit Pelaporan Keuangan merupakan proses yang mencakup memperoleh dan menilai bukti tentang laporan keuangan suatu entitas untuk tujuan menyatakan pendapat apakah informasi keuangan disajikan sesuai kriteria yang ditetapkan. b) Audit Operasional ( Performance Audit) Audit operasional adalah proses mencakup, memperoleh, dan menilai bukti tentang aktivitas operasi suatu entitas berkenaan dengan tujuan khusus yang sering berkaitan baik penilaian kinerja maupun pengambilan keputusan manajemen. Umumnya, pada saat selesainya audit operasional auditor akan memberikan sejumlah saran kepada para manajemen untuk memperbaiki operasional perusahaan. Dalam operasional tujuan dilakukan tidak terbatas pada masalah – masalah akuntansi, tetapi juga meliputi evaluasi terhadap struktur organisasi pemanfaatan j p ( y ) Menurut Agoes (2012: 11) mendefinisikan: j p ( y ) Menurut Agoes (2012: 11) mendefinisikan: 11) j p ( y ) Menurut Agoes (2012: 11) mendefinisikan: “Audit Operasional adalah pemeriksaan terhadap kegiatan perusahaan, termasuk kebijakan akuntansi dan kebijakan operasional yang telah ditentukan tersebut sudah dilakukan secara efektif, efisien, dan ekonomis“. Dari definisi di atas maka diperoleh karakteristik audit operasional a) Audit operasional merupakan suatu proses yang sistematis seperti dalam audit laporan 46 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 eISSN 2684-9313 pISSN 2088-7485 eISSN 2684-9313 pISSN 2088-7485 akan mempengaruhi biaya yang akan dikeluarkan untuk persediaan tersebut. Ruang lingkup audit operasional lebih difokuskan pada fungsi produksi suatu perusahaan, yang berarti melakukan pemeriksaan segi operasional suatu perusahaan. Namun dalam hal ini suatu perusahaan mengalami keterbatasan dalam melaksanakan audit operasional tersebut. Keterbatasan yang terjadi dalam suatu perusahaan dalam melaksaan audit operasional antara lain: a) Waktu. Pemeriksa harus memberikan laporan kepada pihak manajemen sesegera mungkin agar masalah yang timbul dapat segera terselasaikan, sehingga menyebabkan terbatasnya waktu pemeriksaan. Untuk mengatasi keterbatasan waktu ini, audit operasional dapat dilakukan secara teratur untuk menghindari permasalahan agar tidak menjadi berlarut-larut. b) Keahlian kekurangannya pengetahuaan dan penguasaan berbagai disiplin ilmu dan bisnis merupakan salah satu keterbatasan. Tidak mungkin seorang pemeriksa dapat menjadi ahli dalam berbagai disiplin bisnis. c) Biaya yang dapat dihemat dari hasil pemeriksaan haruslah lebih besar dari biaya pemeriksaan itu sendiri. Persediaan merupakan suatu unsur umum yang memegang peranan penting dalam operasi perusahaan, yang secara kontinyu diperoleh atau diproduksi dan dijual mempunyai efek langsung terhadap laba perusahaan. Oleh karena itu para pemimpin perusahaan baik perusahaan dagang maupun industri selalu berusaha untuk mengelola persediaanya sebaik mungkin. Para ahli mendefinisikan persediaan bermacam- macam sebagai berikut: Menurut Prasetyo (2006:65) “persediaan adalah suatu aktiva yang meliputi barang- barang milik perusahaan dengan maksud untuk dijual dalam satu periode usaha yang normal, termasuk yang dalam pengerjaan/ proses produksi menunggu masa penggunaannya pada proses produksi”. Hendriksen (2001:570) menyatakan bahwa: Persediaan meliputi barang perdagangan yang dimaksudkan untuk dijual dalam kondisi usaha normal dan bahan baku serta dalam pembantu yang digunakan dalam proses produksi untuk dijual. Pengertian persediaan menurut Kieso, Weiygand, Warfield, (2007:444), menyatakan bahwa: Audit Atas Persediaan Pada setiap tingkat perusahaan baik perusahaan kecil, menengah, maupun peruasahaan besar. Persediaan sangat penting bagi kelangsungan hidup perusahaan. Perusahaan harus dapat memperkirakan jumlah persediaan yang dimilikinya, persediaan yang dimiliki oleh perusahaan tidak boleh terlalu banyak dan tidak boleh terlalu sedikit karna 47 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 Metode Pengumpulan Data Metode pengumpulan data yang dilakukan dalam penelitian ini adalah penelitian Lapangan (Field Research), yaitu penelitian yang dilakukan dengan mengunjungi secara langsung objek penelitian dengan cara: Observasi, wawancara, kuesioner, Populasi dan Sampel Populasi mengacu pada keseluruhan kelompok orang, kejadian,atau hal minat yang ingin peneliti investigasi. Sedangkan sampel adalah sekelompok atau bagian dari populasi. Dengan mempelajari sampel, peneliti akan mampu menarik kesimpulan yang dapat digeneralisasikan terhadap populasi penelitian (Uma Sekaran 2006:121-123). Adapun yang menjadi populasi sekaligus sampel dalam penelitian ini adalah karyawan dari Usaha Depot Jaya Motor itu sendiri. Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 data yaitu: a) Data primer yaitu data yang bersumber dari hasil observasi dan kuesioner dari pimpinan serta karyawan Usaha Depot Jaya Motor kota Palopo. b) Data sekunder yaitu data yang diperoleh dari perusahaan dalam bentuk yang sudah jadi seperti : struktur organisasi, sejarah berdirinya perusahaan, laporan pembelian, persediaan, dan laporan penjualan. Dalam hubungannya dengan konsep input- proses-output, efisiensi adalah rasio antara output dan input, Seberapa besar output yang dimiliki perusahaan. Metode kerja yang baik akan dapat memandu proses operasi berjalan dengan mengoptimalkan penggunaan sumberdaya yang dimiliki perusahaan. Jadi, efisiensi merupakan ukuran proses yang menghubungkan antara input dan output dalam operasional perusahaan (Bayangkara. 2008:13). Anthony (2005:174) berpendapat bahwa efisiensi adalah output terhadap input, atau jumlah output per unit input. Pusat tanggung jawab A lebih efisien dari pusat tanggung jawab B jika (1) menggunakan jumlah sumberdaya yang lebih sedikit namun memproduksi jumlah output yang sama atau (2) menggunakan jumlah sumberdaya yang sama namun memproduksi jumlah output yang lebih besar. b) Efektifitas bibandingkan dengan efisiensi, yang ditentukan oleh hubungan antara input dan output, efektiftas ditentukan oleh hubungan antara output yang dihasilkan oleh suatu pusat tanggung jawab dengan tujuannya. Semakin besar output yang dikontribusikan terhadap tujuan, maka semakin efektiflah tujuan tersebut. Efektifitas cenderung dinyatakan dalam istilah- istilah subjektif dan nonalitis (Anthony,2005: 174). Efisiensi dan efektifitas berkaitan satu sama lain, setiap pusat tanggung jawab harus efektif dan efisien dimana organisasi harus mencapai tujuannya dengan cara yang optimal. Suatu pusat tanggung jawab yang dijalankan tugasnya dengan konsumsi terendah atas sumberdaya, mungkin akan efisien, tetapi jika output yang dihasilkan gagal dalam memberikan kontribusi yang memadai dalam pencapaian cita-cita organisasi, maka pusat tanggung jawab tersebut tidaklah efektif. ( Anthony: 174-175). Uji validitas dan Reabilitas Sugiyono (2007:137) menjelaskan bahwa uji validitas merupakan alat ukur yang digunakan untuk mendapatkan data itu valid atau dapat digunakan untuk mengukur apa yang seharusnya diukur. Uji validitas dilakukan dengan menggunakan teknik Pearson Correlation, yaitu dengan cara menghitung korelasi antara skor masing-masing butir pernyataan-pernyataan dengan skor total, jika korelasi antara skor masing-masing butir pernyataan dengan total skor mempunyai tingkat signifikansi di bawah 0,05 maka butir pernyataan tersebut dinyatakan valid dan sebaliknya (Ghozali,2009:49). Efisiensi dan Efektivitas Audit operasional dikenal sebagai audit yang berkonsentrasi pada efektivitas dan efisiensi organisasi. Efektivitas mengukur seberapa berhasil suatu organisasi mencapai tujuan dan sasarannya. Efisiensi mengukur seberapa baik suatu entitas menggunakan sumberdaya dalam mencapai tujuannya. Sabagai contoh seorang auditor dapat memeriksa badan federal untuk menentukan apakah badan tersebut telah mencapai tujuannya seperti yang ditetapkan oleh “kongres” (efektivitas) dan menggunakan sumberdaya keuangannya secara benar (efisiensi). Pembahasan mengenai ekonomisasi, efisiensi dan efektivitas akan lebih mudah dipahami jika dibahas dalam kerangka input- proses-output. b) Efisiensi berhubungan dengan bagaimana perusahaan melakukan operasinya. Sehingga dicapai optimalisasi penggunaan sumberdaya yang dimiliki. Efisiensi berhubungan dengan metode kerja (operasi). “ persediaan adalah pos-pos aktiva yang dimiliki perusahaan untuk dijual dalam operasi bisnis normal atau barang yang akan digunakan atau dikonsumsi dalam memproduksi barang yang akan dijual. “ persediaan adalah pos-pos aktiva yang dimiliki perusahaan untuk dijual dalam operasi bisnis normal atau barang yang akan digunakan atau dikonsumsi dalam memproduksi barang yang akan dijual. Menurut Warren Reeve (2009:452).“ persediaan juga didefinisikan sebagai aktiva yang tersedia untuk dijual dalam kegiatan usaha normal dalam proses produksi atau penjualan dalam bentuk bahan atau perlengkapan (supplies) untuk digunakan dalam proses produksi atau pemberian jasa”. Menurut Stice dan Skousen (2009: 571). Persediaan adalah istilah yang diberikan untuk aktiva yang akan dijual dalam kegiatan normal perusahaan atau aktiva yang dimasukkan secara langsung atau tidak langsung ke dalam barang yang akan diproduksi dan kemudian dijual “. Ruang Lingkup Audit Operasional Ruang Lingkup Audit Operasional 48 eISSN 2684-9313 pISSN 2088-7485 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 waktu ke waktu. Suatu konstruk atau variabel dikatakan reliable atau andal jika memberikan nilai cronbach alpha >0,6 (Ghozali, 2005). Penelitian ini akan dilakukan pada usaha Depot Jaya Motor Kota Palopo di Jl.Tandipau No. 21. Penelitian ini akan dilakukan pada usaha Depot Jaya Motor Kota Palopo di Jl.Tandipau No. 21. Uji reliabilitas dilakukan untuk menguji apakah jawaban dari responden konsisten atau stabil. Suatu angket dikatakan andal jika jawaban seseorang terhadap pertanyaan atau pernyataan adalah konsisten atau stabil dari Adapun jenis dan sumber data yang digunakan dalam penelitian ini adalah data kuantatif yaitu data yang diperoleh dari perusahaan dalam bentuk angka-angka. Sumber 49 eISSN 2684-9313 pISSN 2088-7485 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 Uji Regresi Sederhana dan uji hipotesis Analisis regresi dilakukan untuk mengetahui pengaruh antara variabel independen dan variabel dependen. Dalam penelitian ini analisis yang digunakan adalah analisis regresi sederhana. Berikut adalah tabel dari hasil olahan SPSS V 21.0. a. Dependent Variable: Kecukupan Metode Analisis Bertitik tolak dari masalah dan hipotesis yang telah dikemukakan sebelumnya maka dalam membahas dan menganalisis permasalahan sekaligus membuktikan hipotesis maka penulis akan menggunakan metode Regresi Linear Sederhana, menurut Iqbal Hasan (2002:115) regresi linear sederhana adalah regresi linear dimana variabel yang terlibat didalamnya hanya dua, satu varibel terikat Y dan satu variabel bebas X dan perangkat satu. Adapun rumus yang digunakan adalah sebagai berikut : Y = a + bX+e Keterangan : Y = Kecukupan persediaan Barang Dagang X = Audit operasional X = Audit operasional a = intersep b = Koefisien regresi variabel X dan Y menggunakan Skala Likert, dimana indikator tersebut dijadikan sebagai titik tolak untuk menyusun item-item instrument yang dapat berupa pernyataan atau pertanyaan. Sakala Likert didesain untuk menelaah seberapa kuat subjek setuju atau tidak setuju dengan pernyataan pada skala 5 titik dengan respon terhadap sejumlah item yang berkaitan dengan konsep atau variabel tertentu kemudian disajikan kepada tiap responden (Uma Sekaran 2006:31). Agar lebih mudah untuk melihat mengenai variabel penelitian yang dgunakan maka penulis menjabarkannya dalam bentuk operasionalisasi variabel yang dapat dilihat pada tabel berikut ini: 50 Equilibrium Volume 7. No. 2. Tahun 2018 eISSN 2684-9313 Hal. 44 - 54 pISSN 2088-7485 Tabel .1 Operasional variabel X N o Variabel Definisi Indikator Skala 1 Audit operasio nal (X) Audit operasional adalah pemeriksaan terhadap kegiatan perusahaan, termasuk kebijakan akuntansi dan kebijakan operasional yang telah  Indepen si audit operasi onal  Laporan audit  Progra m kinerja auditor dan Tindak lanjut. Ordin al ditentukan manajemen untuk mengetahui apakah kegiatan operasional tersebut sudah dilakukan secara efektif, afisien, dan ekonomis.(A goes 2012:11) Tabel . 2 Operasional variabel Y No Variabel Definisi Indikator Skala 1 persediaan barang dagang (Y) Persediaan adalah suatu aktiva yang meliputi barang-barang milik perusahaan dengan maksud untuk dijual dalam suatu periode usaha yang normal, termasuk yang dalam pengerjaan/proses produksi menunggu masa penggunaannya pada proses produksi (Prasetyo:65).  Efektifitas dan efisiensi operasi.  Keandalan data persediaan.  Pemesanan dan penerimaan barang. Ordinal eISSN 2684-9313 pISSN 2088-7485 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 Tabel . 2 Operasional variabel Y No Variabel Definisi Indikator Skala 1 persediaan barang dagang (Y) Persediaan adalah suatu aktiva yang meliputi barang-barang milik perusahaan dengan maksud untuk dijual dalam suatu periode usaha yang normal, termasuk yang dalam pengerjaan/proses produksi menunggu masa penggunaannya pada proses produksi (Prasetyo:65).  Efektifitas dan efisiensi operasi.  Keandalan data persediaan.  Pemesanan dan penerimaan barang. HASIL DAN PEMBAHASAN Uji Regresi Sederhana dan uji hipotesis Uji Regresi Sederhana dan uji hipotesis Metode Analisis Ordinal TS = Tidak Setuju, diberi skor 2 STS = Sangat Tidak Setuju, diberi skor 1 Masing-masing variabel diatas diukur atas dasar tanggapan responden dan pernyataan yang diberikan pada kuisioner yang diukur dengan menggunakan skala 5 tingkat (Likert) pengukuran yaitu SS = Sangat setuju diberi skor 5 S = Setuju, diberi skor 4 KS = Kurang Setuju, diberi skor 3 Coefficientsa Model Unstandardized Coefficients B Std. Error 1 (Constant) 26.210 3.830 Audit Operasional .473 .086 a. Dependent Variable: Kecukupan Persediaan Barang Dagang Masing-masing variabel diatas diukur atas dasar tanggapan responden dan pernyataan Masing-masing variabel diatas diukur atas dasar tanggapan responden dan pernyataan Masing-masing variabel diatas diukur atas dasar tanggapan responden dan pernyataan dasar tanggapan responden dan pernyataan Coefficientsa Model Unstandardized Coefficients B Std. Error 1 (Constant) 26.210 3.830 Audit Operasional .473 .086 a. Dependent Variable: Kecukupan P di B D Hasil Uji t Hasil dari uji T dengan membandingkan thitung dan ttabel dimana ttabel (df = n-2 atau 17-2=15), hasil diperoleh untuk t tabel sebesar 2,131 dan t hitung sebesar 5.518 dengan tingkat signifikansi sebesar 0.000 . Tabel 3. Korelasi dan Koefisien Determinasi Model Summaryb Model R R Square Adjusted R Square Std. Error of the Estimate 1 .819a .670 .648 2.316 a. Predictors: (Constant), Audit operasional b. Dependent Variable: Kecukupan persediaan barang dagang Sumber: Output SPSS, 21.0, 2015 Dari tabel di atas dapat disimpulkan bahwa: R dalam analisis regresi sederhana menunjukkan korelasi sederhana (korelasi person), yaitu korelasi antara satu variabel Tabel 3. Korelasi dan Koefisien Determinasi Model Summaryb Dimana nilai thitung > ttabel (5.518>2.131) dan dengan nilai signifikansi < 0,05 (0.000<0,05) maka H0 ditolak dan H1 diterima, hal ini menunjukkan bahwa variabel audit operasional berpengaruh secara signifikan terhadap kecukupan persediaan barang dagang. Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 eISSN 2684-9313 pISSN 2088-7485 kecukupan persediaan barang dagang nilainya 26.210. Nilai Koefisien regresi variabel audit operasional adalah 0.473, ini dapat diartikan bahwa setiap peningkatan audit operasional sebesar 1% maka tingkat kecukupan persediaan barang dagang juga akan meningkat sebesar 0.473 atau 47.3%. Uji t (T Test) Uji t digunakan untuk mengetahui apakah variabel independen berpengaruh signifikan atau tidak terhadap variabel dependen, dalam hal ini untuk mengetahui apakah variabel audit operasional berpengaruh secara signifikan atau tidak terhadap kecukupan persediaan barang dagang. Pengujian dilakukan menggunakan tingkat signifikansi 0,05 dengan derajat kebebasan df = n-2. Apabila nilai signifikan lebih kecil dari derajat kebebasan maka hipotesis alternatif diterima, yang menyatakan bahwa suatu variabel independen berpengaruh terhadap variabel dependen. berada di atas 0,5 dan mendekati 1. Dalam penelitian ini, nilai koefisien determinasi yang dipakai adalah Adjusted R Square. Adjusted R Square dianggap lebih baik dari R2 karena Adjusted R Square dapat naik atau turun apabila satu variabel independen ditambahkan ke dalam model regresi. Uji Regresi Linear Sederhana Persediaan Barang Dagang Dari tabel diatas dapat diperoleh rumus regresi linear sederhana sebagai berikut: Y= 26.210+ 0.473X+ e Arti dari persamaan regresi ini Nilai konstanta (a) sebesar 26.210 artinya jika audit operasional nilainya adalah 0, maka Dari tabel diatas dapat diperoleh rumus regresi linear sederhana sebagai berikut: Y= 26.210+ 0.473X+ e Arti dari persamaan regresi ini Nilai konstanta (a) sebesar 26.210 artinya jika audit operasional nilainya adalah 0, maka yang diberikan pada kuisioner yang diukur dengan menggunakan skala 5 tingkat (Likert) pengukuran yaitu Dari tabel diatas dapat diperoleh rumus regresi linear sederhana sebagai berikut: Arti dari persamaan regresi ini Nilai konstanta (a) sebesar 26.210 artinya jika audit operasional nilainya adalah 0, maka 50 eISSN 2684-9313 pISSN 2088-7485 Coefficientsa Model Unstandard ized Coefficients Standardi zed Coefficien ts t Sig B Std. Erro r Beta 1 (Consta nt) 26.2 10 3.83 0 6.8 43 .00 0 Audit operasio nal .473 .086 .819 5.5 18 .00 0 a. Dependent Variable: kecukupan persediaan barang dagang Sumber: Output SPSS, 21.0, 2015 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 Anonimus,2007. PSAK No 14. Pada tabel diatas menunjukkan bahwa nilai koefisien Adjusted R Square sebesar 0.648 atau 64.8%. Maka dapat disimpulkan bahwa audit operasional memiliki pengaruh sebesar 64.8% terhadap kecukupan persediaan barang dagang, sedangkan sisanya 35.2% dipengaruhi oleh variabel lain yang tidak dimasukkan dalam penelitian ini. Arens, Alvin A. Rendal J Elder, Mark S Beasley, 2004. “ Auditing and assurance Service: An Integreted Approach,” Edisi ke-9. PT Indeks kelompok Gramedia, Jakarta. ,2006. “ Auditing dan Pelayanan Verifikasi Pendekatan Terpadu,” Edisi ke-9, Jilid 2, PT Indeks Kelomnpok Gramedia< Jakarta. Standar eror of the Estimate, adalah ukuran kesalahan prediksi, nilainya sebesar 2.316. Artinya kesalahan dalam memprediksi kecukupan persediaan barang dagang sebesar 2.316%. Bayangkara, IBK,2008. “ Audit Manajemen, Prosedur dan Implementasi Management audit”, Salemba Empat, Jakarta. Analisis Koefisien Korelasi dan Koefisien Determinasi (R2) Koefisien korelasi R menunjukkan seberapa besar korelasi atau hubungan antara variabel independen dengan variabel dependen. Koefisien korelasi dikatakan kuat apabila nilai R R dalam analisis regresi sederhana menunjukkan korelasi sederhana (korelasi person), yaitu korelasi antara satu variabel 51 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 eISSN 2684-9313 pISSN 2088-7485 DAFTAR PUSTAKA independen terhadap variabel dependen. Nilai koefisen korelasi (R) pada tabel diatas sebesar 0,819 artinya korelasi atau hubungan antara variabel audit operasional dengan variabel kecukupan persediaan barang dagang terjadi hubungan yang erat karena nilai koefisien korelasi diatas 0,5. Agoes, Sukrisno ,2012. “Auditing Pemeriksaan Akuntansi oleh Kantor Akuntan Publik”, Jlid 1, cetakan ke-4, Lembaga Penerbit salemba Empat, Jakarta. KESIMPULAN Boynton, Johnson, KeIl,2003 . “ Modern auditing, Edisi Ke-7 Jilid 2”. Erlangga, Jakarta. Tujuan dari penelitian ini yaitu untuk mengetahui apakah audit operasional berpengaruh secara signifikan terhadap kecukupan persediaan barang dagang.Berdasarkan pembahasan hasil penelitian, Audit operasional memiliki hubungan yang signifikan terhadap kecukupan persediaan barang dagang.hal ini dapat dilihat dari hasil pengujian thitung Audit Operasional sebesar 5.518 dan ttabel sebesar 2.131 ini menunjukkan bahwa nilai thitung > ttabel dengan tingkat signifikansi 0.000<0.05. Dari nilai koefisien Adjusted R Square sebesar 0.648 atau 64.8%. Maka dapat disimpulkan bahwa audit operasional memiliki pengaruh sebesar 64.8% terhadap kecukupan persediaan barang dagang, sedangkan sisanya 35.2% dipengaruhi oleh variabel lain yang tidak dimasukkan dalam penelitian ini. Ghozali,Imam, 2005. “Aplikasi Analisis Multivariate dengan Program Spss”. Badan penerbit UNDIP,Semarang. Guy M dan C Wayne, Alan J winters ,2003. “Auditing” Edisi Kelima, Erlangga, Jakarta. Hakim, Lukman, 2003 . “ Peranan Sistem dan Prosedur Pencatatan Persediaan Barang pada Perusahaan dalam Melaksanakan Pengendalian Interen”. Jurnal Arthavidya Vol 4, No 1. Halim, Abdul, 2003 . “Auditing”,Edisi ketiga, unit penerbit dan percetakan akademi manajemen perusahaan YKPN, Jakarta. Harnato, 2002 .“Akuntansi Keuangan Menengah”. Cetakan pertama, penerbit BPFE UGM, yokyakarta. 52 Equilibrium Volume 7. No. 2. Tahun 2018 Hal. 44 - 54 Hendriksen,E,S, Breda, M,F ,2001 . “accounting theory” Rivard D Irwin inc, Boston. Iriyadi, 2004. “ Evaluasi Atas Prosedur Pemeriksaan Oprasional dalam Meningkatkan Efektivitas Pengendalian Intern Penjualan”. Jurnal Ilmiah ranggading,Vol 4, No 1. Hlm 15- 20. Iqbal Hasan, 2002. “Pokok-Pokok Materi Metodologi Penelitian dan Aplikasinya”. Penerbit Ghalia Indonesia, Jakarta. Kieso, Waygand Warfield, 2007. “Akuntansi intermediate jilid 1”. Erlangga, Jakarta. Mudricah, 2008 . “sistem Akuntansi Persediaan Bahan Baku Pada PT SINAR LENDOH TERANG AMBARAWA”. Skripsi, Universitas Negeri Semarang. Mulyadi, 2002. “ Auditing”. Edisi ke-6, Jilid 1, Salemba Empat, Jakarta. Prasetyo, 2006. “ Pengembangan Modal Persediaan dengan Mempertimbangkan Waktu Kadaluarssa dan Faktor Unit Diskon, Jurnal Ilmiah Teknik Industri, volume 4, No 3, Universitas Muhammadiyah, Surakarta. Purwono, Edi, 2004. “Aspek-aspek EDP Audit Pengenalan Intern pada Komputerisasi”, ANDI, Yogyakarta. Setiyawan, Agus, 2007. “Pengaruh Sistem Pengendalian Intern dan Pengelolaan Persediaan Terhadap Audit Operasional pada PT PERTAMINA PERSERO”. Skripsi, Universitas Islam eISSN 2684-9313 pISSN 2088-7485 eISSN 2684-9313 pISSN 2088-7485 Sugiyono,2009. “Metode Penelitian Kuantitatif dan Kualitatif”, penerbit Alfabeta,Cv, Bandung. Tunggal,Ak,MBA, 2008 . Audit Manajemen Konteporer.Penerbit Harvarindo. Jakarta Barat. Uma Sekaran, 2006. “Metodologi Penelitian Untuk Bisnis”. Buku 2,Penerbit Salemba Empat,Jakarta. Iqbal Hasan, 2002. “Pokok-Pokok Materi Metodologi Penelitian dan Aplikasinya”. Penerbit Ghalia Indonesia, Jakarta. , 2009. “Metodologi Penelitian Untuk Bisnis”. Buku 1,Penerbit Salemba Empat,Jakarta. KESIMPULAN Kieso, Waygand Warfield, 2007. “Akuntansi intermediate jilid 1”. Erlangga, Jakarta. Warren S carl, James M, Reeve dan Philip E. Fees,2009 . “Pengantar Akuntansi”. Edisi 21, salemba empat, Jakarta. Mudricah, 2008 . “sistem Akuntansi Persediaan Bahan Baku Pada PT SINAR LENDOH TERANG AMBARAWA”. Skripsi, Universitas Negeri Semarang. Mulyadi, 2002. “ Auditing”. Edisi ke-6, Jilid 1, Salemba Empat, Jakarta. Prasetyo, 2006. “ Pengembangan Modal Persediaan dengan Mempertimbangkan Waktu Kadaluarssa dan Faktor Unit Diskon, Jurnal Ilmiah Teknik Industri, volume 4, No 3, Universitas Muhammadiyah, Surakarta. Purwono, Edi, 2004. “Aspek-aspek EDP Audit Pengenalan Intern pada Komputerisasi”, ANDI, Yogyakarta. Setiyawan, Agus, 2007. “Pengaruh Sistem Pengendalian Intern dan Pengelolaan Persediaan Terhadap Audit Operasional pada PT PERTAMINA PERSERO”. Skripsi, Universitas Islam Negeri Syarif Hidayatullah, Jakarta. Stice dan Skousen, 2009. “Akuntansi Intermedite”. Edisi keenambelas, buku 1, Salemba Empat, Jakarta. Stice 53
https://openalex.org/W4386067970	https://aacr.figshare.com/articles/journal_contribution/Supplementary_Figure_S3_from_Natural_Killer_Cell_Activation_by_Ubiquitin-specific_Protease_6_Mediates_Tumor_Suppression_in_Ewing_Sarcoma/24008237/1/files/42102908.pdf	English	null	Supplementary Figure S3 from Natural Killer Cell Activation by Ubiquitin-specific Protease 6 Mediates Tumor Suppression in Ewing Sarcoma	null	2,023	cc-by	3,562	C B D A Primary Mouse NK Cell (LAKs) Killing Assays Primary Human NK cell Killing Assays E 0:1 1:1 2.5:1 5:1 0 25 50 75 100 % CD3+ 7AAD- Live YAC-1 YAC1 killing assay E:T : 0:1 1:1 2.5:1 5:1 % Live cells remaining (CD3+/7AAD-) 0.1 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ p = 0.00049 p = 0.004 p = 0.2 80 % CD107a+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP Hi 7AAD- Compiled_Prim Hu NK + LA673 Killing 125 100 75 50 25 0 p = 0.003 p = 0.003 p = 0.04 E:T % Live cells remaining (GFP+ CD45- 7AAD-) USP6/A673 0:1 0.1:1 0.25:1 0.5:1 125 100 75 50 25 0 E:T % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 0.1:1 0.25:1 0.5:1 USP6/A673 p = 0.02 p = 0.02 p = 0.008 E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP+ 7AAD- Copy of Compiled_Prim Hu NK + LA673 Killing 80 % CD107+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ - Dox + Dox - Dox + Dox - Dox + Dox - Dox + Dox E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- Compiled_LAK + LA673_Killing _ Graph 2 - Dox + Dox - Dox + Dox Supplementary Figure S3: USP6 enhances cytolytic activity and activation of primary mouse NK (LAKs) an primary human NK cells A) Spleen-derived LAKs were co-cultured with mouse YAC1 cells at the indicated E ratios for 24h, and percent live remaining YAC1 cells was quantified to confirm LAK cytolytic activity (n=3). B) Assa for monitoring cytolytic activity and NK cell activation. C) Cytolytic activity of LAKs against USP6/A673 wa determined Right each dot pair represents a distinct LAKs preparation (each comprising spleens from 3 mice USP6/ES (GFP+) (Target (T)) NK cell (Effector (E)) + Quantify live USP6/ES (GFP+/7AAD-) Monitor NK cell activation (CD107a) 24h +/-dox D Primary Human NK cell Killing Assays E 0.1 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ p = 0.00049 p = 0.004 p = 0.2 80 % CD107a+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP Hi 7AAD- Compiled_Prim Hu NK + LA673 Killing 125 100 75 50 25 0 p = 0.003 p = 0.003 p = 0.04 E:T % Live cells remaining (GFP+ CD45- 7AAD-) USP6/A673 0:1 0.1:1 0.25:1 0.5:1 125 100 75 50 25 0 E:T % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 0.1:1 0.25:1 0.5:1 E:T 0:1 1:1 2.5:1 5:1 0:1 1:1 2.5:1 5:1 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP+ 7AAD- Copy of Compiled_Prim Hu NK + LA673 Killing 80 % CD107+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ - Dox + Dox - Dox + Dox - Dox + Dox E:T 0:1 1:1 2.5:1 5:1 - Dox + Dox S l t Fi S3 USP6 h t l ti ti it d ti ti f i NK (LAK ) D Primary Human NK cell Killing Assays E 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP Hi 7AAD- Compiled_Prim Hu NK + LA673 Killing 125 100 75 50 25 0 p = 0.003 p = 0.003 p = 0.04 E:T % Live cells remaining (GFP+ CD45- 7AAD-) USP6/A673 0:1 0.1:1 0.25:1 0.5:1 125 100 75 50 25 0 E:T % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 0.1:1 0.25:1 0.5:1 E:T 0:1 1:1 2.5:1 5:1 0:1 1:1 2.5:1 5:1 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP+ 7AAD- Copy of Compiled_Prim Hu NK + LA673 Killing - Dox + Dox - Dox + Dox E:T 0:1 1:1 2.5:1 5:1 D D E 80 % CD107+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ - Dox + Dox Supplementary Figure S3: USP6 enhances cytolytic activity and activation of primary mouse NK (LAKs) and primary human NK cells A) Spleen-derived LAKs were co-cultured with mouse YAC1 cells at the indicated E:T ratios for 24h, and percent live remaining YAC1 cells was quantified to confirm LAK cytolytic activity (n=3). C B D A Primary Mouse NK Cell (LAKs) Killing Assays Primary Human NK cell Killing Assays E 0:1 1:1 2.5:1 5:1 0 25 50 75 100 % CD3+ 7AAD- Live YAC-1 YAC1 killing assay E:T : 0:1 1:1 2.5:1 5:1 % Live cells remaining (CD3+/7AAD-) 0.1 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ p = 0.00049 p = 0.004 p = 0.2 80 % CD107a+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP Hi 7AAD- Compiled_Prim Hu NK + LA673 Killing 125 100 75 50 25 0 p = 0.003 p = 0.003 p = 0.04 E:T % Live cells remaining (GFP+ CD45- 7AAD-) USP6/A673 0:1 0.1:1 0.25:1 0.5:1 125 100 75 50 25 0 E:T % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 0.1:1 0.25:1 0.5:1 USP6/A673 p = 0.02 p = 0.02 p = 0.008 E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP+ 7AAD- Copy of Compiled_Prim Hu NK + LA673 Killing 80 % CD107+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ - Dox + Dox - Dox + Dox - Dox + Dox - Dox + Dox E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- Compiled_LAK + LA673_Killing _ Graph 2 - Dox + Dox - Dox + Dox Supplementary Figure S3: USP6 enhances cytolytic activity and activation of primary mouse NK (LAKs) an primary human NK cells A) Spleen-derived LAKs were co-cultured with mouse YAC1 cells at the indicated E ratios for 24h, and percent live remaining YAC1 cells was quantified to confirm LAK cytolytic activity (n=3). B) Assa for monitoring cytolytic activity and NK cell activation. C) Cytolytic activity of LAKs against USP6/A673 wa determined Right each dot pair represents a distinct LAKs preparation (each comprising spleens from 3 mice USP6/ES (GFP+) (Target (T)) NK cell (Effector (E)) + Quantify live USP6/ES (GFP+/7AAD-) Monitor NK cell activation (CD107a) 24h +/-dox C B D A Primary Mouse NK Cell (LAKs) Killing Assays Primary Human NK cell Killing Assays E 0:1 1:1 2.5:1 5:1 0 25 50 75 100 % CD3+ 7AAD- Live YAC-1 YAC1 killing assay E:T : 0:1 1:1 2.5:1 5:1 % Live cells remaining (CD3+/7AAD-) 0.1 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ p = 0.00049 p = 0.004 p = 0.2 80 % CD107a+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP Hi 7AAD- Compiled_Prim Hu NK + LA673 Killing 125 100 75 50 25 0 p = 0.003 p = 0.003 p = 0.04 E:T % Live cells remaining (GFP+ CD45- 7AAD-) USP6/A673 0:1 0.1:1 0.25:1 0.5:1 125 100 75 50 25 0 E:T % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 0.1:1 0.25:1 0.5:1 USP6/A673 p = 0.02 p = 0.02 p = 0.008 E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP+ 7AAD- Copy of Compiled_Prim Hu NK + LA673 Killing 80 % CD107+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ - Dox + Dox - Dox + Dox - Dox + Dox - Dox + Dox E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- Compiled_LAK + LA673_Killing _ Graph 2 - Dox + Dox - Dox + Dox Supplementary Figure S3: USP6 enhances cytolytic activity and activation of primary mouse NK (LAKs) an primary human NK cells A) Spleen-derived LAKs were co-cultured with mouse YAC1 cells at the indicated E ratios for 24h, and percent live remaining YAC1 cells was quantified to confirm LAK cytolytic activity (n=3). B) Assa for monitoring cytolytic activity and NK cell activation. C B D A Primary Mouse NK Cell (LAKs) Killing Assays Primary Human NK cell Killing Assays E 0:1 1:1 2.5:1 5:1 0 25 50 75 100 % CD3+ 7AAD- Live YAC-1 YAC1 killing assay E:T : 0:1 1:1 2.5:1 5:1 % Live cells remaining (CD3+/7AAD-) 0.1 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ p = 0.00049 p = 0.004 p = 0.2 80 % CD107a+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP Hi 7AAD- Compiled_Prim Hu NK + LA673 Killing 125 100 75 50 25 0 p = 0.003 p = 0.003 p = 0.04 E:T % Live cells remaining (GFP+ CD45- 7AAD-) USP6/A673 0:1 0.1:1 0.25:1 0.5:1 125 100 75 50 25 0 E:T % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 0.1:1 0.25:1 0.5:1 USP6/A673 p = 0.02 p = 0.02 p = 0.008 E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP+ 7AAD- Copy of Compiled_Prim Hu NK + LA673 Killing 80 % CD107+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ - Dox + Dox - Dox + Dox - Dox + Dox - Dox + Dox E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- Compiled_LAK + LA673_Killing _ Graph 2 - Dox + Dox - Dox + Dox Supplementary Figure S3: USP6 enhances cytolytic activity and activation of primary mouse NK (LAKs) an primary human NK cells A) Spleen-derived LAKs were co-cultured with mouse YAC1 cells at the indicated E ratios for 24h, and percent live remaining YAC1 cells was quantified to confirm LAK cytolytic activity (n=3). B) Assa for monitoring cytolytic activity and NK cell activation. C) Cytolytic activity of LAKs against USP6/A673 wa determined Right each dot pair represents a distinct LAKs preparation (each comprising spleens from 3 mice USP6/ES (GFP+) (Target (T)) NK cell (Effector (E)) + Quantify live USP6/ES (GFP+/7AAD-) Monitor NK cell activation (CD107a) 24h +/-dox B) Assa for monitoring cytolytic activity and NK cell activation. C) Cytolytic activity of LAKs against USP6/A673 wa determined. Right, each dot pair represents a distinct LAKs preparation (each comprising spleens from 3 mice incubated with target cells with or without dox (n=4-6). D/E) Primary human NK cells and USP6/A673 were co-culture with or without dox. Cytolytic activity was measured in D, and degranulation (surface CD107a) in E. Data are from independent human NK donors, with samples assayed in singlicate or duplicate depending on cell availability. C B D A Primary Mouse NK Cell (LAKs) Killing Assays Primary Human NK cell Killing Assays E 0:1 1:1 2.5:1 5:1 0 25 50 75 100 % CD3+ 7AAD- Live YAC-1 YAC1 killing assay E:T : 0:1 1:1 2.5:1 5:1 % Live cells remaining (CD3+/7AAD-) 0.1 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ p = 0.00049 p = 0.004 p = 0.2 80 % CD107a+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP Hi 7AAD- Compiled_Prim Hu NK + LA673 Killing 125 100 75 50 25 0 p = 0.003 p = 0.003 p = 0.04 E:T % Live cells remaining (GFP+ CD45- 7AAD-) USP6/A673 0:1 0.1:1 0.25:1 0.5:1 125 100 75 50 25 0 E:T % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 0.1:1 0.25:1 0.5:1 USP6/A673 p = 0.02 p = 0.02 p = 0.008 E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP+ 7AAD- Copy of Compiled_Prim Hu NK + LA673 Killing 80 % CD107+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ - Dox + Dox - Dox + Dox - Dox + Dox - Dox + Dox E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- Compiled_LAK + LA673_Killing _ Graph 2 - Dox + Dox - Dox + Dox Supplementary Figure S3: USP6 enhances cytolytic activity and activation of primary mouse NK (LAKs) an primary human NK cells A) Spleen-derived LAKs were co-cultured with mouse YAC1 cells at the indicated E ratios for 24h, and percent live remaining YAC1 cells was quantified to confirm LAK cytolytic activity (n=3). B) Assa for monitoring cytolytic activity and NK cell activation. C) Cytolytic activity of LAKs against USP6/A673 wa determined Right each dot pair represents a distinct LAKs preparation (each comprising spleens from 3 mice USP6/ES (GFP+) (Target (T)) NK cell (Effector (E)) + Quantify live USP6/ES (GFP+/7AAD-) Monitor NK cell activation (CD107a) 24h +/-dox C) Cytolytic activity of LAKs against USP6/A673 wa determined Right each dot pair represents a distinct LAKs preparation (each comprising spleens from 3 mice USP6/ES (GFP+) (Target (T)) NK cell (Effector (E)) + Quantify live USP6/ES (GFP+/7AAD-) Monitor NK cell activation (CD107a) 24h +/-dox B A 0:1 1:1 2.5:1 5:1 0 25 50 75 100 % CD3+ 7AAD- Live YAC-1 YAC1 killing assay E:T : 0:1 1:1 2.5:1 5:1 % Live cells remaining (CD3+/7AAD-) USP6/ES (GFP+) (Target (T)) NK cell (Effector (E)) + Quantify live USP6/ES (GFP+/7AAD-) Monitor NK cell activation (CD107a) 24h +/-dox B USP6/ES (GFP+) (Target (T)) NK cell (Effector (E)) + Quantify live USP6/ES (GFP+/7AAD-) Monitor NK cell activation (CD107a) 24h +/-dox A 0:1 1:1 2.5:1 5:1 0 25 50 75 100 % CD3+ 7AAD- Live YAC-1 YAC1 killing assay E:T : 0:1 1:1 2.5:1 5:1 % Live cells remaining (CD3+/7AAD-) B A C Primary Mouse NK Cell (LAKs) Killing Assays USP6/A673 p = 0.02 p = 0.02 p = 0.008 E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- - Dox + Dox E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- Compiled_LAK + LA673_Killing _ Graph 2 - Dox + Dox Primary Mouse NK Cell (LAKs) Killing Assays C D Primary Human NK cell Killing Assays E 0.1 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ p = 0.00049 p = 0.004 p = 0.2 80 % CD107a+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP Hi 7AAD- Compiled_Prim Hu NK + LA673 Killing 125 100 75 50 25 0 p = 0.003 p = 0.003 p = 0.04 E:T % Live cells remaining (GFP+ CD45- 7AAD-) USP6/A673 0:1 0.1:1 0.25:1 0.5:1 125 100 75 50 25 0 E:T % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 0.1:1 0.25:1 0.5:1 p = 0.02 p = 0.02 p = 0.008 E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP+ 7AAD- Copy of Compiled_Prim Hu NK + LA673 Killing 80 % CD107+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ - Dox + Dox - Dox + Dox - Dox + Dox - Dox + Dox E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- - Dox + Dox - Dox + Dox Supplementary Figure S3: USP6 enhances cytolytic activity and activation of primary mouse NK (LAKs) an primary human NK cells A) Spleen-derived LAKs were co-cultured with mouse YAC1 cells at the indicated E ratios for 24h, and percent live remaining YAC1 cells was quantified to confirm LAK cytolytic activity (n=3). C B D A Primary Mouse NK Cell (LAKs) Killing Assays Primary Human NK cell Killing Assays E 0:1 1:1 2.5:1 5:1 0 25 50 75 100 % CD3+ 7AAD- Live YAC-1 YAC1 killing assay E:T : 0:1 1:1 2.5:1 5:1 % Live cells remaining (CD3+/7AAD-) 0.1 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ p = 0.00049 p = 0.004 p = 0.2 80 % CD107a+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP Hi 7AAD- Compiled_Prim Hu NK + LA673 Killing 125 100 75 50 25 0 p = 0.003 p = 0.003 p = 0.04 E:T % Live cells remaining (GFP+ CD45- 7AAD-) USP6/A673 0:1 0.1:1 0.25:1 0.5:1 125 100 75 50 25 0 E:T % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 0.1:1 0.25:1 0.5:1 USP6/A673 p = 0.02 p = 0.02 p = 0.008 E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- 0 0.1 0.25 0.5 0 25 50 75 100 125 E:T ratio % GFP+ 7AAD- Copy of Compiled_Prim Hu NK + LA673 Killing 80 % CD107+ (of CD45+) E:T 60 40 20 0 NK alone 0.1:1 0.25:1 0.5:1 0 0.1 0.25 0.5 0 20 40 60 80 E:T ratio % CD107a+ - Dox + Dox - Dox + Dox - Dox + Dox - Dox + Dox E:T 0:1 1:1 2.5:1 5:1 125 100 75 50 25 0 % Live cells remaining (GFP+ CD45- 7AAD-) 0:1 1:1 2.5:1 5:1 0 25 50 75 100 125 % GFP Hi 7AAD- Compiled_LAK + LA673_Killing _ Graph 2 - Dox + Dox - Dox + Dox Supplementary Figure S3: USP6 enhances cytolytic activity and activation of primary mouse NK (LAKs) an primary human NK cells A) Spleen-derived LAKs were co-cultured with mouse YAC1 cells at the indicated E ratios for 24h, and percent live remaining YAC1 cells was quantified to confirm LAK cytolytic activity (n=3). B) Assa for monitoring cytolytic activity and NK cell activation. C) Cytolytic activity of LAKs against USP6/A673 wa determined Right each dot pair represents a distinct LAKs preparation (each comprising spleens from 3 mice USP6/ES (GFP+) (Target (T)) NK cell (Effector (E)) + Quantify live USP6/ES (GFP+/7AAD-) Monitor NK cell activation (CD107a) 24h +/-dox B) Assay for monitoring cytolytic activity and NK cell activation. C) Cytolytic activity of LAKs against USP6/A673 was determined. Right, each dot pair represents a distinct LAKs preparation (each comprising spleens from 3 mice), incubated with target cells with or without dox (n=4-6). D/E) Primary human NK cells and USP6/A673 were co-cultured with or without dox. Cytolytic activity was measured in D, and degranulation (surface CD107a) in E. Data are from 5 independent human NK donors, with samples assayed in singlicate or duplicate depending on cell availability.
https://openalex.org/W4362680928	https://link.springer.com/content/pdf/10.1007/978-3-658-40496-3_2.pdf	English	null	Long-Term Radioactive Waste Management in the Netherlands: Seeking Guidance for Decision-Making	Energiepolitik und Klimaschutz	2,023	cc-by	8,100	Long-Term Radioactive Waste Management in the Netherlands: Seeking Guidance for Decision-Making Long-Term Radioactive Waste Management in the Netherlands: Seeking Guidance for Decision-Making Romy Dekker, Vincent Lagendijk, Roos Walstock and Rinie van Est Romy Dekker, Vincent Lagendijk, Roos Walstock and Rinie van Est © The Author(s) 2023 M. Arentsen and R. van Est (eds.), The Future of Radioactive Waste Governance, Energiepolitik und Klimaschutz. Energy Policy and Climate Protection, https://doi.org/10.1007/978-3-658-40496-3_2 1 The classification system currently used in the Netherlands resembles the IAEA guideline of 2009 (IAEA Safety Standard, 2009), but consists of four categories instead of six: ‘high- level radioactive waste’ (HLW), ‘low-and medium-level radioactive waste’ (LILW), ‘short- lived waste’ (with a half-time of less than 100 days) and ‘exempt waste’ (COVRA, 2014). A subcategory of LILW consists of waste produced from the use of naturally occurring radio active material (NORM). NORM waste with an activity concentration of up to ten times the exemption threshold, are disposed of as very low-level waste at special licensed dumpsites (Ministry of I&E, 2016, p. 16). In this chapter, we only focus on the long-term manage ment of the waste stored at COVRA. 2 The Rathenau Instituut is an independent technology assessment organization. It has been involved in research and debate about the impact of science, innovation, and technology on society for 35 years. 2.1 Introduction The Netherlands currently stores its radioactive waste above ground at the Cen tral Organization for Radioactive Waste (COVRA). With regard to the long-term management of radioactive waste and spent fuel, the Netherlands pursues a ‘dual strategy’. First, there is a national route in which the government envisions a geo logical disposal facility (GDF) for a part of its radioactive waste and spent fuel to be operational by 2130 (Ministry of I&E, 2016). Nevertheless, the option is left open to deviate from this timeframe, as well as from the currently preferred long- term disposal method (geological disposal), if there is reason to do so. Second, the government pursues an international route with other European Union (EU) Member States for the long-term management of radioactive waste (Ministry of I&E, 2016). Although an approximate timeline has been developed, the concrete decision-making process that should lead to a solution, either nationally or inter nationally, has not yet been established. R. Dekker () · V. Lagendijk · R. van Est  Rathenau Instituut, The Hague, Netherlands e-mail: r.dekker@rathenau.nl V. Lagendijk e-mail: v.lagendijk@rathenau.nl R. van Est e-mail: q.vanest@rathenau.nl R. Walstock  Kirkman Company, Baarn, Netherlands e-mail: r.walstock@kirkmancompany.com R. Dekker () · V. Lagendijk · R. van Est  Rathenau Instituut, The Hague, Netherlands e-mail: r.dekker@rathenau.nl R. Walstock  Kirkman Company, Baarn, Netherlands e-mail: r.walstock@kirkmancompany.com 25 © The Author(s) 2023 M. Arentsen and R. van Est (eds.), The Future of Radioactive Waste Governance, Energiepolitik und Klimaschutz. Energy Policy and Climate Protection, https://doi.org/10.1007/978-3-658-40496-3_2 © The Author(s) 2023 M. Arentsen and R. van Est (eds.), The Future of Radioactive Waste Governance, Energiepolitik und Klimaschutz. Energy Policy and Climate Protection, https://doi.org/10.1007/978-3-658-40496-3_2 26 R. Dekker et al. 2016 3 3 3 2020a 1 2016 2012 2016 2 1 The classification system currently used in the Netherlands resembles the IAEA guideline of 2009 (IAEA Safety Standard, 2009), but consists of four categories instead of six: ‘high- level radioactive waste’ (HLW), ‘low-and medium-level radioactive waste’ (LILW), ‘short- lived waste’ (with a half-time of less than 100 days) and ‘exempt waste’ (COVRA, 2014). A subcategory of LILW consists of waste produced from the use of naturally occurring radio active material (NORM). NORM waste with an activity concentration of up to ten times the exemption threshold, are disposed of as very low-level waste at special licensed dumpsites (Ministry of I&E, 2016, p. 16). 2.2 This section describes how the Netherlands managed radioactive waste from 1945 to 2016, and how decision-making took shape. This timeframe spans the period between the building-up of the nuclear sector to the first National Pro gram for the management of radioactive waste. Since the management of radio active waste is closely linked to its applications, we also take developments in the field of nuclear technologies into account. We show that over the years dif ferent waste management methods have been suggested, discussed, researched, used, regulated, banned, and/or abandoned. For each of these methods, specific decision-making processes took place, in terms of technical viability, and social and political-administrative desirability and legal admissibility. 2.1 Introduction In this chapter, we only focus on the long-term manage ment of the waste stored at COVRA. 2 The Rathenau Instituut is an independent technology assessment organization. It has been involved in research and debate about the impact of science, innovation, and technology on society for 35 years. Long-Term Radioactive Waste Management … 27 2 This chapter investigates various important challenges for the decision-making process regarding the long-term management of radioactive waste in the Neth erlands. To this end, we use the conceptual model of a multi-level governance ecosystem (Kool et al., 2017), as explained in the introductory chapter, which consists of four domains and their interactions: ‘politics and administration’, ‘sci ence and technology’, ‘laws and regulations’, and ‘civil society’. We first describe the historical development of the governance ecosystem in the Netherlands, based on a reading of parliamentary documents, publications from the national waste management organisation (COVRA), governmental organisations, NGOs, news items, previous reviews of the national nuclear sector, as well as literature on the governance of radioactive waste. Based on that overview, we reflect on the devel opments within the separate domains and identify current challenges for decision- making. We end by drawing several conclusions. 2.2.1 Development of the Nuclear Sector and Laws and Regulations for Nuclear Safety and Radiation Protection in the Netherlands After World War II, with the support of the United States, the Dutch government teamed up with scientists to explore the peaceful potential of nuclear technology. 28 R. Dekker et al. To this end, the government set up a knowledge and research infrastructure, developed industrial policy, and provided information about nuclear technol ogy to the general public (Ministerie van Economische Zaken, 1957). The Neth erlands also became a shareholder in Eurochemic, an international company founded in 1957 within the framework of the Organization for European Eco nomic Cooperation (OEEC). Its purpose was to build a factory for reprocessing spent fuel, situated in Belgium. Dutch nuclear reactors came online for research and education in Petten (1960 and 1961), Delft (1963), Wageningen (1963), and Eindhoven (1969). Nuclear power plants (NPP) became operational in Dode waard (1968) and Borssele (1973). There was relatively little attention to radioac tive waste management (RWM) during this build-up phase, which also held true for waste management in most other sectors at that time (cf. IAEA, 2002). Over the years, there was a gradual increase in attention to radioactive waste, driven by an international scientific discussion on how such waste should be handled. In the 1950s and 1960s, new organisations were set up to promote the safe use of nuclear technology, such as the International Atomic Energy Agency (IAEA) in 1957, the European Nuclear Energy Agency (ENEA, later NEA) in 1958, and the European Atomic Energy Community (Euratom) in 1958. The Dutch govern ment implemented legislation and regulations on nuclear safety and radiation protection due to the expected increase in the use of radioactive substances in medical, biological, industrial and agricultural fields (Radioactieve Stoffenbesluit, 1958, p. 7). On the basis of agreements within the Euratom Treaty, the Ionizing Radiation Decree (Radioactieve Stoffenbesluit, 1958) provided guidelines with regard to safety. In addition, the government used existing legislation to create the preconditions for protection against ionizing radiation for employees and the general public. In 1963 the Nuclear Energy Act was passed, that governed nuclear activities and provided regulations for nuclear safety and radiation protection (Kernenergiewet 1963). This law, which still applies, is a so-called framework law, with associated Decrees and Ordinances providing more detailed legislation. i During the first two decades of the nuclear program in the Netherlands, there was no explicit radioactive waste policy, just as there was no regular waste pol icy. 3 This Steering Committee consisted of nine experts from political and scientific circles, and was chaired by Mauk de Brauw, a socialist politician who had been minister of aca demic education and research, and previously worked for various companies including Unilever. 2.2.1 Development of the Nuclear Sector and Laws and Regulations for Nuclear Safety and Radiation Protection in the Netherlands However, a practice of managing radioactive waste did emerge (Berkers et al., 2023). In 1963, the Minister of Social Affairs and Health established a spe cial designated service to collect LILW waste. This radioactive waste was sub sequently stored above ground in Petten, near the Reactor Center Netherlands (RCN). Part of this waste was disposed in the deep sea. This latter practice was supervised by ENEA after 1965. Between 1966 and 1974, spent nuclear fuel was reprocessed at Eurochemic in Belgium, where—according to contract—the leftover HLW remained. After Eurochemic shut down in 1974, new agreements Long-Term Radioactive Waste Management … 29 2 were made for the reprocessing of spent fuel with the United States, United King dom and France. The remaining Dutch HLW would eventually be sent back from the UK and France to the Netherlands. At the same time, social resistance to the dumping of radioactive and other high-toxic waste into the deep sea grew, both within and outside the Netherlands, and an international ban on the dumping of HLW into the sea came into effect in 1972. An international moratorium on dumping LILW was issued in 1983, followed by a complete ban in 1993. 2.2.2 Realizing an Above-Ground Interim Storage Facility for Radioactive Waste RWM became a topic in the Dutch societal arena in the 1970s, as exemplified by the societal resistance to deep sea dumping. RWM also became a central issue in the nuclear energy debate as the anti-nuclear movement arose. Furthermore, the government at the time indicated that expanding nuclear energy was only pos sible after an ‘acceptable solution’ was found for radioactive waste (Ministerie van Economische Zaken, 1974). In line with international scientific debates, the Scientific Council for Nuclear Energy, and the Reactor Center of the Nether lands, among others, advised to examine the possibilities of disposing radioactive waste in deep underground salt domes (cf. WRK, 1972). In 1976, the government wanted to investigate the technological possibilities of disposing HLW in salt formations in the northeastern part of the Netherlands. Regional and local resist ance by citizens, societal organisations, companies and politicians eventually obstructed the proposed in situ test drilling (Berkers et al., 2023). 3 30 R. Dekker et al. 2011 These societal and political developments impacted scientific research as well as the policy process (Schröder, 2012). The Integral National Research Program Nuclear Waste (ILONA) started in 1981. ILONA consisted of various committees that examined different possibilities for the disposal of HLW: storage on land, just above or below the earth’s surface, interim-storage of spent fuel elements and nuclear fission waste, disposal in salt domes in the North Sea, and disposal in geological layers beneath the deep seas (Ministerie van VROM, 1984). The research on North Sea salt domes and deep sea geological disposal was rather quickly abandoned due to higher than expected costs (Berkers et al., 2023). In 1981, the Committee Reconsidering Disposal of Radioactive Waste (Commissie Heroverweging Verwijdering Radioactief Afval, HVRA) was installed to look for alternatives for the deep sea disposal of LILW (Ministerie van VROM, 1984, p. 12). HVRA concluded in March 1983 that it had a preference for disposing LILW in salt layers, e.g. by means of deep salt cavities (Berkers et al., 2023). The first RWM policy was presented in 1984 by the Minister of Housing, Spa tial Planning and the Environment (VROM). It included two goals in the field of radiation protection. First, to comply with the ALARA (As Low As Reasonably Achievable) principle, which had been recommended by the International Com mission on Radiological Protection (ICRP) in 1973 and endorsed by Euratom in 1980. 2.2.2 Realizing an Above-Ground Interim Storage Facility for Radioactive Waste This had to be done by isolating, managing and controlling the waste (in Dutch the so-called IBC-criteria). Second, the sum of the received and expected doses for humans should not exceed the established dose limits. The govern ment also decided to set up an above-ground interim storage facility for LILW and HLW. This should be managed by COVRA, for a period of ‘several decades’, which was later explained during parliamentary debates and related policy docu ments as at least 100 years (Ministry of I&E, 2016). This temporal policy pro vided time to further study options for a final GDF and to explore the possibility of an international disposal facility. A designated committee under the aegis of liberal politician W.J. Geertsema was tasked with finding a suitable and accepta ble location for the intended facility. This eventually led to an above-ground radi oactive waste storage facility in Borsele, near the NPP. Local residents and the anti-nuclear energy movement participated in this decision-making process. By August 1989, COVRA had obtained all necessary permits and was able to start construction of the storage facility. 31 2 Long-Term Radioactive Waste Management … 2 2.2.3 Deep Geological Disposal: Elaboration of Policy and Research In subsequent decades, the government elaborated on its 1984 radioactive waste policy through parliamentary debates, research and public consultation. This included the formulation of an environmental policy framework, informed by a public consultation on the acceptability of geological disposal of (radioactive and highly-toxic) waste. This led to new modified principles for RWM: in line with the IBC-criteria, reversibility of the decision-making process and retrievabil ity of the waste became requirements (Tweede Kamer, 1993). With this in mind, the national scientific Committee on Storage of Radioactive Waste (CORA) was asked by ILONA in 1996 to investigate the feasibility of retrievable disposal of radioactive waste both in salt domes and Boom Clay. In addition to technical aspects, one of its subcommittees focused on ethical and social aspects of long- term RWM in a scoping study amongst environmental organisations (CORA, 2001; Selling, 2002). This was the first time a social scientific angle was included in a national research programme on RWM in the Netherlands. The work of international organisations influenced Dutch RWM policy and research. This concerned, for example, international agreements in the IAEA- framework, international radiation guidelines and standards by the International Commission on Radiological Protection (ICRP) (via Euratom), and the work of the Nuclear Energy Agency (NEA). The NEA developed the concept of a safety case, which ‘comprises the findings of a safety assessment and a statement of confidence in these findings’ (OECD, 1999, p. 22). The safety case methodology was applied in the national Research Program for the Final Storage of Radioac tive Waste (OPERA), which ran from 2011 to 2016. This program was organised by COVRA, and looked in particular into the possibilities of a geological disposal facility in Boom Clay. In 2004, NEA also argued for a stepwise approach to deci sion-making, which was in line with the Dutch principle of reversibility. Further more, NEA stated that it is important that ‘the public, and especially the most affected local public, are meaningfully involved in the planning process’ (OECD, 2004, p. 7). With regard to the legal domain, the Aarhus Convention came into effect in 2001, and grants EU citizens the right to access to information, public participation in decision-making, and access to justice in environmental matters (UN, 1998). Institutional arrangements also changed during this period, with several changes in the division of ministerial tasks and responsibilities in the field of nuclear energy, nuclear safety and radiation protection. 4 The consultancy group consisted of seven members from different backgrounds, such as regional authorities, drinking water boards and universities. 2.2.3 Deep Geological Disposal: Elaboration of Policy and Research For example, since 2010 32 R. Dekker et al. the Ministry of Economic Affairs was responsible for nuclear safety and radia tion protection, nuclear energy, the Nuclear Energy Act and the management of the associated organisational units. The Minister of Social Affairs and Employ ment (SZW) was responsible for the protection of employees, and the Minister of Health, Welfare and Sport (VWS) was responsible for protecting patients against the risks of ionizing radiation (ABDTOPConsult, 2019). Influenced by the House of Representatives and international guidelines from the IAEA, it was decided in 2015 to set up a new independent Authority for Nuclear Safety and Radiation protection under the Ministry of Infrastructure and the Environment (I&E): the ANVS. The new authority could not fall under the Ministry of Economic Affairs, which was responsible for (nuclear) energy policy (Wijzigingswet kernerner giewet, 2016). In 2019, a legal evaluation of the ANVS led to the transfer of policy responsibility for the management of radioactive waste to the Ministry of Infrastructure and Water Management (before I&E). The EU also had an influence on the decision-making process regarding the disposal and management of radioactive waste in the Netherlands. In accordance with the 2011/70/Euratom directive, every EU Member State became obliged to draw up a National Program (European Council, 2011). In preparation for the Dutch National Program, four studies were carried out: 1) an inventory of the cur rent and future volume of radioactive waste by COVRA; 2) an initial study into options for the long-term management of radioactive waste by engineering con sultancy ARCADIS; 3) a study on public participation by the Rathenau Instituut; and 4) a study of the state of affairs concerning international research into dis posal by the Nuclear Research & Consultancy Group (NRG). 2.3 Current Challenges The long-term interim above-ground storage at COVRA provides time to work on a final solution for the long-term management of radioactive waste. While the National Program provides a tentative timeline, the step-by-step process of deci sion-making needs further elaboration. To better understand the challenges that need to be addressed, this section reflects on historic and current development of the four domains of the governance-ecosystem: 1) laws and regulation, 2) pol icy and administration, 3) science and technology, and 4) civil society. Because each domain depends on the others for their functioning, we consider their mutual interactions. Moreover, since various levels of government play a role in dealing with radioactive waste, we also consider the multi-level character of the domains. 2.2.3 Deep Geological Disposal: Elaboration of Policy and Research 2016 2017 4 2 Long-Term Radioactive Waste Management … 33 2021  2.1 2016 The National Program also announced the construction of a consultation group (Klankbordgroep) that would be tasked to focus on issues including: • public participation (‘identifying specific forms of participation’); • siting of radioactive waste disposal facilities (‘potential suitable search areas for the disposal of radioactive waste that can be reserved, and identifying the necessary policy harmonization, given other future functions of the (deep) underground environment at those sites’); • knowledge infrastructure (‘options for maintaining the necessary knowledge infrastructure in the Netherlands for the management of radioactive waste’), and; • knowledge infrastructure (‘options for maintaining the necessary knowledge infrastructure in the Netherlands for the management of radioactive waste’), and; 34 R. Dekker et al. • practical implementation of the principle of retrievability (‘defining the criteria for determining the period of retrievability of radioactive waste from disposal to allow a decision on a period of retrievability supported by society’). (Minis try of I&E, 2016, p. 6) • practical implementation of the principle of retrievability (‘defining the criteria for determining the period of retrievability of radioactive waste from disposal to allow a decision on a period of retrievability supported by society’). (Minis try of I&E, 2016, p. 6) The Ministry of I&W asked Van Soest to explore how such a consultation group process could be organisationally embedded. Van Soest concluded that the mis sion of such a group could be “to think through a possible participatory decision- making process aimed at a social agreement about the disposal of radioactive waste and spent fuel, and to advise relevant parties on this” (Van Soest, 2018, p. 8). As a result of this preliminary investigation, it was decided to task the Rathenau Instituut with issuing advice in 2024 on the decision-making process regarding the long-term management of radioactive waste. 2.3.1 Laws and Regulations 1963 Long-Term Radioactive Waste Management … 35 2 1998 2023 5 5 This study has been conducted as part of the current assignment of the Rathenau Instituut. 6 The OPERA advisory group was asked to advice on the quality of the research and on its social relevance. In addition, they had an advisory role on the communication about the program and the results. 8 The advisory board of the ANVS consists of independent experts from the Netherlands and abroad, its task is to provide the ANVS with solicited and unsolicited advice on matters related to the tasks of the ANVS. 7 The Netherlands Commission for Environmental Assessment (NCEA) advises govern ments on the quality of environmental information in environmental assessment reports. The NCEA does not get involved in decision-making or political considerations. 7 The Netherlands Commission for Environmental Assessment (NCEA) advises gove ments on the quality of environmental information in environmental assessment repo The NCEA does not get involved in decision-making or political considerations. 2.3.2 Policy and Administration  2.2 Dutch RWM policy has been both hailed and criticized. First, the organisation of centralised long-term interim above-ground storage of radioactive waste has been deemed as ‘good governance’ by member states of the Joint Convention on the Safety of Spent Fuel, and on the Safety of Radioactive Waste Management, because the packaging and storage facilities are also designed with a term of 100 R. Dekker et al. 36 to 300 years in mind, which facilitates a high level of safety (Ministry of I&W, 2020). In line with this, the IAEA (2009) considered the incorporation of passive safety features in the design of the packaging and facilities as a ‘good practice’, because it makes the safety of the interim-storage less dependent on maintenance, and the packaged materials can be monitored more easily. Second, the IAEA con cluded that COVRA communicates well with the public, for example through art and by organising open days, which has increased the confidence of the public regarding the activities of COVRA (IAEA, 2009). 2015 6 2017 Long-Term Radioactive Waste Management … 37 2 2015 2015 2015 7   2014 2   2017 2016 2020 8 2022 2016 R. Dekker et al. 38 2017 2017 2016 2016  2.2 2015 Another important development over the last two decades is the redistribution of ministerial tasks and responsibilities. Since the 1950s, responsibility for both nuclear energy and nuclear safety in the Netherlands has rested most of the time with the Ministry of Economic Affairs. In 1994, the IAEA Convention on Nuclear Safety stipulated that each Member State had to ensure a separation between organisations in the fields of nuclear safety and nuclear energy. Currently there is a clear division between the responsibilities for energy policy that lie with the Ministry of Economic Affairs, and on the responsibilities for nuclear safety and radiation protection that lie with the Ministry of I&W. The ANVS, established in 2015, was given both policymaking and supervision responsibilities. In 2020, the responsibility for policymaking was transferred from the ANVS to the Ministry of I&W. This creates a better distinction between the duties of the Ministry of I&W, and ANVS as a supervisor and licensing authority (ABDTopconsult, 2019). In recent years, the nuclear energy discussion has resurfaced, partially because of the climate crisis. In the wake of this discussion, radioactive waste also returned to the political and social agenda. 2.3.2 Policy and Administration In its coalition agreement, the current Long-Term Radioactive Waste Management … 39 2 government intends to keep the nuclear power plant in Borssele open longer, to build two new NPPs, and to provide for the safe disposal of nuclear waste (Kabi netsformatie, 2021). What this will mean for the decision-making process regard ing the long-term management of radioactive waste remains to be seen. 2.3.3 Science and Technology The domain of science and technology continues to play an important role in the development of standards for radiation protection and nuclear safety. The same holds for the investigation of the viability and safety of technical options for (long-term) RWM. It thereby influences the development of legislation, regulations, and policy. Below, we describe how despite increased awareness of the importance of interdisciplinary and transdisciplinary knowledge, an inte gral, participatory and sociotechnical knowledge agenda is still missing in the Netherlands, and it is unclear where the institutional responsibility for such an agenda lies. Moreover, we highlight that the vitality of the science and technology domain would benefit from a long-term vision on knowledge development. Until the 1990s, Dutch RWM research primarily focused on technical aspects, such as the technical feasibility of geological disposal in salt domes. From CORA (which ran from 1996 to 2001) onwards, ethical and social aspects of the long- term management of radioactive waste, and the role of public participation in decision-making, have received attention. Within ILONA (the research program which ran between 1981 and 1993), it was argued that social scientific research should be limited to an inventory of the processes that play a role in decision- making (Tweede Kamer, 2002). OPERA, the subsequent national research pro gram (which ran from 2011 to 2018), set up the previously mentioned OPERA advisory group. The group dealt with the “wider societal issues of disposal, including stakeholder engagement and conditions for an inclusive process for long-term decision-making on disposal” (Verhoef et al., 2017, p. 8). The advi sory group stated that this requires a participatory process and the recognition of emotions and values, which can be used to shape and direct technological devel opment; so-called “value sensitive design” (cf. Correljé et al., 2015). In 2015, the Rathenau Instituut concluded that RWM transcends various academic disci plines (technical, geological, ethical, social, psychological, economic), and that it is therefore important “to retain sight of the issue’s multidisciplinary charac ter and the consequent need for interdisciplinary cooperation” (De Vries et al., 2015, p. 19). Despite intentions to set up a broader research program, the current 40 R. Dekker et al. long-term program at COVRA focuses primarily on technical aspects. A more integrated sociotechnical research agenda is thus still lacking. long-term program at COVRA focuses primarily on technical aspects. A more integrated sociotechnical research agenda is thus still lacking. 9 Although COVRA is fully owned by the Dutch state, it is an independent administrative body (in Dutch: Onafhankelijk Bestuursorgaan) that performs government tasks, but does not fall directly under the authority of a ministry. 10 This is in line with the polluter pays principle. This principle assures that organisations which deliver their radioactive waste to COVRA pay for all costs related its management, including research (COVRA, 2020b). 11 Translation from Dutch by authors. 11 Translation from Dutch by authors. 12 While preparing the national programme, ANVS asked the Rathenau Instituut to formu late a vision on public participation in decision-making about long-term radioactive waste management, to serve as a supporting study for the national programme. 2.3.3 Science and Technology 2000   2023 9 2020b 10 2022 11 2022 2022 2 Long-Term Radioactive Waste Management … 41 2 Long-Term Radioactive Waste Management … 2 2020b 2020b 2000   2015 12 2001 2003 2019   2015 2.3 2020b From the point of view of institutional independence, checks and balances, and spreading knowledge, as well as COVRA, the Ministry of I&W, ANVS and R. Dekker et al. 42 other relevant stakeholders must also continue to have access to sufficient and (independent) knowledge and expertise to continue to perform their institutional duties (cf. Raad van Advies ANVS, 2020). The Ministry of I&W is currently exploring how the knowledge landscape of nuclear technology and radiation can be strengthened, including in the long term (cf. Van de Zande et al., 2020). 2.3.4 Civil Society Local residents were also consulted by COVRA in the search for a suitable location for the interim aboveground stor age of radioactive waste. consultation regarding the desirability of geological disposal of radioactive and highly toxicological waste in the early 1990s. Local residents were also consulted by COVRA in the search for a suitable location for the interim aboveground stor age of radioactive waste. Scientific studies show that public participation can lead to higher levels of trust, but that this is not a causal relationship (cf. Wang & Wan Wart, 2007; OECD, 2017; Liu et al. 2019). The broad societal discussion on energy policy (BMD) in the Netherlands, for example, has not led to more confidence in the government among opponents of nuclear energy. The government saw the BMD as a public consultation and had clearly indicated in advance that it would not automatically follow the outcome of the BMD. At the start of the BMD, many opponents of nuclear energy had low faith in the government and saw the BMD as fake participation. After the BMD in 1984, they felt reinforced in that opinion when the government ignored the outcome of the BMD, that there should be no further expansion of nuclear energy. This experience still seems to play a role in current debates on RWM (Berkers et al, 2023). In a 2015 study (De Vries et al. 2015), the Rathenau Instituut concluded that there is “limited trust in the gov ernment with regard to this specific policy issue” (p. 7), and that there is a lack of urgency to participate because of the absence of actual decisions. The Rath enau Instituut proposed that the best way of bolstering trust and willingness to participate, is “to develop a public participation model whose subject matter and procedural design enjoys widespread support” (De Vries et al., 2015, p. 15), and concluded that decision-making on long-term RWM requires a more extensive, long-term and systematic process of public participation. To this end, it proposed the following: 1) the development of a shared plan for public participation, 2) the tailoring of issue-based participation clusters, and 3) periodic reflection to deviate from the plan if necessary. Public participation has in fact become obligatory following the 1998 Aarhus Convention (UN 1998) and through the 2011/70/Euratom directive (European Council, 2011) and Dutch legislation (Akerboom, 2023). 2.3.4 Civil Society Rather than serving as blueprints for the participatory process, these guidelines and regulations should therefore be seen as ‘minimum standards’. 2.3.4 Civil Society Over the last few decades, the societal domain has been involved in decision- making on the long-term management of radioactive waste in several ways. In some cases, involvement was initiated by the government, and in others by civil society actors themselves. We show here that this had various outcomes, and took multiple forms: from protest and resistance to informing and consulting. Over the years, awareness of the importance of public involvement in radioactive waste decision-making has increased. It even became legally required by international and national guidelines, treaties and laws. Part of the assignment to the Rathenau Instituut is to develop an advice on a (possibly participatory) decision-making process for the final disposal of radioactive waste on behalf of the Ministry of I&W. This has been lacking until now, partly due to the absence of concrete deci sion-making steps, which influences willingness to participate. Civil society has influenced both policy and research. In the 1950s, companies worked closely with scientific institutions to erect a nuclear industry in the Neth erlands. In the early 1970s, governmental policy to expand the nuclear sector was criticized by an emerging anti-nuclear energy movement. Dealing with radioac tive waste became a central issue in the nuclear energy debate after the govern ment indicated that expanding nuclear energy was only possible if an ‘acceptable solution’ had been found for radioactive waste. At the same time, the existing practice of disposing radioactive waste in the deep sea was also met with increas ing public protest, resulting in a ban for first HLW (1975), and subsequently LILW (temporary moratorium in 1983, legal ban in 1993). The announcement of in situ test drillings in salt layers in the northeastern part of the Netherlands to find potentially suitable places for a GDF, led to regional political-administrative and social resistance among both proponents and opponents of nuclear energy. To this day, conducting research in the subsurface for the geological disposal of radioactive waste remains a sensitive issue (De Vries et al., 2015). The government has also consulted the public on policy development at vari ous times—often under pressure from parliament, local and regional authorities, and society. This was the case during the BMD in the early 1980s, and the broad Long-Term Radioactive Waste Management … 43 2 consultation regarding the desirability of geological disposal of radioactive and highly toxicological waste in the early 1990s. 2.4 Conclusions Since the materialization of NPPs in the 1960s and 1970s, political and societal debates on radioactive waste and nuclear technology, in particular nuclear energy, have become entangled. This has made long-term RWM a sensitive issue. While R. Dekker et al. 44 the government opted for the further development of nuclear power since the opening of the nuclear power plant in Borsele in 1972, the 1986 Chernobyl acci dent put a stop to such plans. Recently, the political discussion on nuclear energy has resurfaced. This might influence the long-term management of radioactive waste. The entanglement is also visible from an institutional point of view. For example, the Nuclear Energy Act regulates the licensing of nuclear installations and the safe management of radioactive materials, including radioactive waste. And the division of ministerial responsibilities has shifted because of a perceived conflict of interest between the fields of nuclear energy, and nuclear safety and radiation protection, including RWM. Over the years, societal resistance against various waste management options, such as disposal into the deep sea and exploratory drillings for geological dis posal in salt layers, has influenced the political decision-making process regard ing long-term RWM. Since 1984, the radioactive waste policy has resulted in the establishment of an above-ground storage facility, after which geological disposal is foreseen either nationally or internationally. The National Program in 2016 stated that a decision on a final GDF should be taken around 2100. The possibil ity is left open to deviate from this scenario—both in terms of the timeframe and disposal method. In this way, future generations should not be left with an unrea sonable burden. In addition, various policy principles and requirements have been formulated over the years, and a legal framework has been developed that pro vides guidelines for the decision-making process that lies ahead. Moreover, over the past two decades the ministerial tasks and responsibilities regarding nuclear energy, nuclear safety and radiation protection are split between the Ministries of Economic Affairs and I&W, which may contribute to the checks and balances in the field of RWM. Although there is a National Program and a suggested timeline for RWM, the road to a final solution is only partly worked out. The Dutch policy of long-term above-ground storage, followed by a GDF, is not only hailed but also criticized. 2.4 Conclusions It is seen as good governance and as a good practice in terms of safety and commu nication with the public. On the other hand, parties such as the European Com mission, societal organisations and the Rathenau Instituut are concerned that the century-long period for decision-making leads to a decrease in political urgency and willingness to participate and act on the subject. Various parties, such as the NCEA and the Council of the ANVS, therefore recommend using the interim period meaningfully and to concretize and possibly bring forward the decision- making process. The National Program named several issues that should be clari fied as part of the decision-making process. 2 Long-Term Radioactive Waste Management … 45 1) Define the optimal period of retrievability, 1) Define the optimal period of retrievability, 1) Define the optimal period of retrievability, 2) Set up criteria to reserve potential search locations for a GDF, 3) Clarify options for maintaining the necessary knowledge landscape, and 4) Concretize the role of public participation within research and various national and decentralised political decision-making processes. 4) Concretize the role of public participation within research and various national and decentralised political decision-making processes. The reflections in this chapter on the separate domains bring to light four addi tional cross-domain issues to be addressed in the decision-making process: 5) Further elaborate the requirement of reversibility of the decision-making pro cess to clarify to what extent decisions should be reversible and how this will be assessed. 5) Further elaborate the requirement of reversibility of the decision-making pro cess to clarify to what extent decisions should be reversible and how this will be assessed. 6) Develop a long-term, integral and participatory knowledge agenda to support the decision-making process and to keep the science and technology domain vital, now and in the future. 7) Spread knowledge over various (public) institutions, so that there can be an institutionally sound knowledge landscape with sufficient checks and bal ances. And lastly, 8) Develop a participatory decision-making process for the final disposal of radi oactive waste that enjoys broad public and political support in terms of content and procedure to bolster trust and willingness to participate. Laws and regula tions in the field of participation can serve as minimum standards. 2.4 Conclusions Acknowledgements The authors would like to thank Dhoya Snijders for his contribution to earlier versions of this chapter and the participants of the societal review from the societal, scientific and public administration domains for their highly appreciated feedback. ABDTOPConsult. (2019). Wettelijke evaluatie van het zbo ANVS. Algemene Bestuursdi enst. Den Haag: Ministerie van Binnenlandse Zaken en Koninkrijksrelaties. Akerboom, S. (forthcoming 2023). Rapport rechtendomein lange termijn beheer radioac tief afval. Een studie naar het juridisch en governance-kader omtrent opslag en berging van radioactief afval in Nederland. Den Haag: Rathenau Instituut. Berenschot. (2022). Evaluatie Radioactief afval. Een evaluatie van de Nota Radioactief afval en het Nationaal Programma Radioactief afval. Utrecht: Berenschot. Adviesgroep OPERA. (2017). Van Afval naar Berging. Tijd voor Verantwoordelijkheid. Nieuwdorp: COVRA. ABDTOPConsult. (2019). Wettelijke evaluatie van het zbo ANVS. Algemene Bestuursdi enst. Den Haag: Ministerie van Binnenlandse Zaken en Koninkrijksrelaties. Adviesgroep OPERA. (2017). Van Afval naar Berging. Tijd voor Verantwoordelijkheid. Nieuwdorp: COVRA. Akerboom, S. (forthcoming 2023). Rapport rechtendomein lange termijn beheer radioac tief afval. Een studie naar het juridisch en governance-kader omtrent opslag en berging van radioactief afval in Nederland. Den Haag: Rathenau Instituut. Berenschot. (2022). Evaluatie Radioactief afval. Een evaluatie van de Nota Radioactief afval en het Nationaal Programma Radioactief afval. Utrecht: Berenschot. References ABDTOPConsult. (2019). Wettelijke evaluatie van het zbo ANVS. Algemene Bestuursdi enst. Den Haag: Ministerie van Binnenlandse Zaken en Koninkrijksrelaties. Adviesgroep OPERA. (2017). Van Afval naar Berging. Tijd voor Verantwoordelijkheid. Nieuwdorp: COVRA. Akerboom, S. (forthcoming 2023). Rapport rechtendomein lange termijn beheer radioac tief afval. Een studie naar het juridisch en governance-kader omtrent opslag en berging van radioactief afval in Nederland. Den Haag: Rathenau Instituut. Berenschot. (2022). Evaluatie Radioactief afval. Een evaluatie van de Nota Radioactief afval en het Nationaal Programma Radioactief afval. Utrecht: Berenschot. R. Dekker et al. 46 Berkers, E., Lagendijk, V., Dekker, R., Snijders, D., en R. van Est (2023). Een kwestie van tijd: Besluitvorming over radioactief afval in Nederland van 1945 tot 2016. Den Haag: Rathenau Instituut. Brunnengräber, A. (2019). The wicked problem of long term radioactive waste governance. In A. Brunnengräber, & M.R. Di Nucci (Eds.), Conflicts, Participation and Acceptabil ity in Nuclear Waste Governance (pp. 335–355). Wiesbaden: Springer VS. Correljé, A., Cuppen, E., Dignum, M., Pesch, U., & Taebi, B. (2015). Responsible Innova tion in Energy Projects: Values in the Design of Technologies, Institutions and Stake holder Interactions. In B.-J. Koops, I. Oosterlaken, H. Romijn, T. Swierstra, & J. van den Hoven (Eds.), Responsible Innovation 2 (pp. 183–200). Springer International Pub lishing, Cham. Commissie m.e.r. (2015). Nationaal uitvoeringsprogramma voor het langetermijnbeheer van radioactief afval en verbruikte splijtstoffen. Utrecht: Commissie voor de milieuef fectenrapportage. CORA. (2001). Terugneembare berging, een begaanbaar pad? Onderzoek naar de mogeli jkheden van terugneembare berging van radioactief afval in Nederland. Den Haag: Economische Zaken. COVRA (2014). Inventaris Radioactief Afval in Nederland. https://www.covra.nl/app/ uploads/2019/08/Inventaris-radioactief-afval-in-Nederland.pdf. COVRA. (2020a). Jaarrapport 2020. Nieuwdorp: COVRA. COVRA. (2020b). Long-Term Research Programme for Geological Disposal of Radioac tive Waste. Overall research program and work program for 2020–2025. Nieuwdorp: COVRA and Technopolis. COVRA. (2021). Vereniging voor Multinationale Samenwerking Radioactief Afval opger icht in Nieuwdorp: ERDO. https://www.covra.nl/nl/organisatie/nieuws/vereniging-voor- multinationale-samenwerking-radioactief-afval-opgericht-in-nieuwdorp-erdo/. De Vries, A., van Waes, A., van Est, R., van der Meulen, B., & Brom, F. (2015). Enabling participation: A vision on public participation in decision-making about long term radioactive waste management. Den Haag: Rathenau Instituut. European Council. (2011). Council Directive 2011/70/Euratom of 19 July 2011 establish ing a Community framework for the responsible and safe management of spent fuel and radioactive waste. OJ L 199/48, 2 August 2011. Gibbons, M., & Nowotny, H. (2001). The potential of transdisciplinarity. Transdisciplinar ity: joint problem solving among science, technology, and society, 67–80. IAEA Safety Standards. (2009). References Regulations for the Safe Transport of Radioactive Mate rial. Vienna: International Atomic Energy Agency. IAEA Safety Standards. (2009). Regulations for the Safe Tran rial. Vienna: International Atomic Energy Agency. IAEA (2000) Legal and governmental infrastructure for nuclear, radiation, radioactive waste and transport safety. Vienna: International Atomic Energy Agency Specific Safety Requirements, GS-R-1. IAEA. (2002). Radioactive Waste Management Profiles. Vienna: International Atomic Energy Agency. IAEA. (2009). Peer Review of the Radioactive Waste Management Activities of COVRA, Netherlands. International Atomic Energy Agency: Vienna, IAEA Safety Standards Applications Series, 8. pp Kabinetsformatie. (2021). Omzien naar elkaar, vooruitkijken naar de toekomst. Den Haag: Kabinetsformatie. (2021). Omzien naar elkaar, vooruitkijken naar de toekomst. Den Haag: Bureau Woordvoering Kabinetsformatie Kabinetsformatie. (2021). Omzien naar elkaar, vooruitkijken naar de toekomst. De Bureau Woordvoering Kabinetsformatie. Kernenergiewet. (1963) Staatsblad van het Koninkrijk der Nederlanden, 82. Bureau Woordvoering Kabinetsformatie. Kernenergiewet. (1963) Staatsblad van het Koninkrijk der Nederlanden, 82. g Kernenergiewet. (1963) Staatsblad van het Koninkrijk der Nederlanden, 82. 2 Long-Term Radioactive Waste Management … 47 2 Kool, L., Timmer, J., Royakkers, R., & van Est, R. (2017). Urgent upgrade: Protect public values in our digitized society. Den Haag: Rathenau Instituut. LAKA. (2015, September 8). Pleitnotitie—Uitbreiding. COVRA. LAKA. https://www.laka. org/info/uitbreiding-covra/2015-09-08_pleidooi_uitbreiding_covra_anoniem.pdf. Liu, L., Bouman, T., Perlaviciute, G., & Steg, L. (2019). Effects of trust and public par ticipation on acceptability of renewable energy projects in the Netherlands and China. Energy Research & Social Science, 53, 137–144. Ministerie van Economische Zaken. (1957). Nota inzake de Kernenergie. Den Haag: Minis terie van Economische Zaken. Ministerie van Economische Zaken. (1974). Energienota. Den Haag: SDU. Ministry of I&E. (2016). The national programme for the management of radioactive waste and spent fuel. Den Haag: Ministry of Infrastructure and the Environment (I&E). Ministry of I&W. (2020). Joint convention on the safety of spent fuel management and on the safety of radioactive waste management. Den Haag: Ministry of Infrastructure and Water Management (I&W). Ministerie van VROM. (1984) Radioactief afval. Den Haag: Ministerie van Volkshuisvest ing, Ruimtelijke Ordening en Milieubeheer. Nowotny, H. (2003). Democratising expertise and socially robust knowledge. Science and Public Policy, 30(3), 151–156. OECD. (1999). Confidence in the Long-term Safety of Deep Geological Repositories. Its Development and Communication. Paris: Nuclear Energy Agency / Organisation for Economic Co-operation and Development. OECD. (2000). Geological disposal of radioactive waste: Review of developments in the last decade. Paris: Nuclear Energy Agency / Organisation for Economic Co-operation and Development. OECD. (2004). References Stepwise approach to decision making for long-term radioactive waste management: Experience, issues and guiding principles. Paris: Organisation for Eco nomic Co-operation and Development, Nuclear Energy Agency. OECD. (2017). Trust and Public Policy: How Better Governance Can Help Rebuild Public Trust. OECD Public Governance Reviews. Paris: Organisation for Economic Co-opera tion and Development. Raad van Advies ANVS. (2020). Advies van Raad van Advies over de rol van ANVS in relatie tot eindberging van radioactief afval. Den Haag: Autoriteit Nucleaire Veiligheid en Stralingsbescherming. g g Radioactieve stoffenbesluit. (1958). Staatsblad van Het Koninkrijk der Nederlanden 317, 7 Schröder, J. (2012). Identifying remaining socio-technical challenges at the na the Netherlands. SOTEC working paper. Antwerp: University of Antwerp. Selling, H.A. (2002). Policy of radioactive waste disposal in the Netherlands. https://inis. iaea.org/collection/NCLCollectionStore/_Public/33/037/33037868.pdf. Smink, M., van den Broek, J., Metze, T., Cuppen, E., & van Est, R. (2017). In E. van de Grift, & A. van Waes (eds.), Samen kennis aanboren: Verkenning van kennis en opvat tingen over ultradiepe geothermie. Den Haag: Rathenau Instituut. Tweede Kamer. (1993). Opbergen van afval in de diepe ondergrond. Den Haag: SDU, 1992–1993, 23163–1. Tweede Kamer. (2002). Radioactief afvalbeleid. Den Haag: SDU, 2002–2003, 28674–1. Tweede Kamer. (2017). Opwerking van radioactief materiaal. Den Haag: SDU, 2016– 2017, 25422–204. R. Dekker et al. 48 UN. (1998). Convention on Access to Information, Public Participation in Decision-Mak ing and Access to Justice in Environmental Matters. Geneva: The United Nations Eco nomic Commission for Europe (UNECE). Van der Zande, A., Wolterbeek, B., & C. Leijen. (2020). Naar een Agenda en Platform Nucleaire Technologie en Straling. Den Haag: Autoriteit Nucleaire Veiligheid en Stral ingsbescherming. Van Soest, J.P. (2018). Diepgravende Dialogen,Bouwen aan Vertrouwen.Eindrapportage Kwartiermaker Eindberging radioactief afvalen verbruikte splijtstoffen. Deventer: De Gemeynt. Van Waes A., de Vries, A., van Est, R., & F. Brom. (2014). Broadening the debate on shale gas: Guidelines for decision making based on the Dutch experience. Rathenau Instituut. Verhoef, E., Bartol, J., Neeft, E., Chapman, N., & McCombie, C. (2017). Summary OPERA Safety Case. Nieuwdorp: COVRA. Vijftig jaar HFR. (2011). Elst: Kernvisie 6(6), 1–5. Wang, X., & Wan Wart, M. (2007). When public participation in administration leads to trust: An empirical assessment of managers’ perceptions. Public administration review, 67(2), 265–278. Wijzigingswet Kernenergiewet. (2016) Staatsblad van Het Koninkrijk der Nederlanden. WRK. (1972). Vestigingsplaatsen van energiereactoren en de opslag van radio-actief afval. WRK. (1972). Vestigingsplaatsen van energiereactoren en de opslag van radio-actief afval. References Den Haag: Wetenschappelijke Raad voor de Kernenergie Den Haag: Wetenschappelijke Raad voor de Kernenergie. Den Haag: Wetenschappelijke Raad voor de Kernenergie. Romy Dekker is a senior researcher at the Rathenau Instituut, the Dutch parliamentary technology assessment (TA) and science systems assessment (SciSA) organisation in The Hague. She has a degree in cultural anthropology and development sociology. Her research focuses on topics related to the governance of sustainability. She is a part-time PhD-candi date in the democratic governance of science and technology at Eindhoven University of Technology. Vincent Lagendijk is a senior researcher at the Rathenau Instituut, the Dutch parliamen tary technology assessment (TA) and science systems assessment (SciSA) organisation in The Hague. He was trained as a historian and in Science and Technology Studies, and has a keen interest in infrastructures. He also is a part-time Assistant Professor at Maastricht University’s Faculty of Arts and Social Sciences. Roos Walstock works at Kirkman Company, a consultancy that helps organisations iden tify opportunities for truly tackling societal issues (‘wicked problems’) in an innovative way. In 2021, she first worked as a research intern, and from September to December as a junior researcher at the Rathenau Instituut, the Dutch Parliamentary organisation for Tech nology Assessment. Roos studied strategic innovation management, and philosophy and society at the University of Groningen. As part of her Masters in philosophy she researched different ways of decision-making under deep uncertainty and their connection with values. 2 Long-Term Radioactive Waste Management … 49 2 Rinie van Est is a research coordinator at the Rathenau Instituut, the Dutch parliamentary technology assessment (TA) and science systems assessment (SciSA) organisation in The Hague. He has a degree in applied physics and public administration, and a PhD in political science. He is a global expert in the field of TA, politics of innovation and public participa tion. For more than twenty-five years he has been involved in the energy and digital transi tions, with special attention to the role of emerging technologies, such as robotics and AI. He is part-time Professor of Technology Assessment and Governance at Eindhoven Univer sity of Technology. References Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, 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 Crea tive Commons license and indicate if changes were made. The images or other third party material in this chapter are included in the chapter’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
W4304690628.txt	https://www.businessperspectives.org/images/pdf/applications/publishing/templates/article/assets/17163/KPM_2022_01_Hamed.pdf	en		Adoption of big data analytics in medium-large supply chain firms in Saudi Arabia	Knowledge & performance management (Online)/Knowledge & performance management	2,022	cc-by	7,998	“Adoption of big data analytics in medium-large supply chain firms in Saudi Arabia” AUTHORS Adel Hamed Abdul Manaf Bohari ARTICLE INFO Adel Hamed and Abdul Manaf Bohari (2022). Adoption of big data analytics in medium-large supply chain firms in Saudi Arabia. Knowledge and Performance Management, 6(1), 62-74. doi:10.21511/kpm.06(1).2022.06 DOI http://dx.doi.org/10.21511/kpm.06(1).2022.06 RELEASED ON Wednesday, 12 October 2022 RECEIVED ON Friday, 17 June 2022 ACCEPTED ON Monday, 05 September 2022 LICENSE This work is licensed under a Creative Commons Attribution 4.0 International License JOURNAL "Knowledge and Performance Management" ISSN PRINT 2543-5507 ISSN ONLINE 2616-3829 PUBLISHER LLC “Consulting Publishing Company “Business Perspectives” FOUNDER Sp. z o.o. Kozmenko Science Publishing NUMBER OF REFERENCES NUMBER OF FIGURES NUMBER OF TABLES 59 0 2 © The author(s) 2022. This publication is an open access article. businessperspectives.org Knowledge and Performance Management, Volume 6, 2022 Adel Hamed (Malaysia), Abdul Manaf Bohari (Malaysia) BUSINESS PERSPECTIVES LLC “СPС “Business Perspectives” Hryhorii Skovoroda lane, 10, Sumy, 40022, Ukraine www.businessperspectives.org Adoption of big data analytics in medium-large supply chain firms in Saudi Arabia Abstract Received on: 17th of June, 2022 Accepted on: 5th of September, 2022 Published on: 12th of October, 2022 © Adel Hamed, Abdul Manaf Bohari, 2022 Adel Hamed, Ph.D. Candidate, Lecturer, School of Business Management (SBM), College of Business (COB), University Utara Malaysia (UUM), Sintok, Kedah Darul Aman, Malaysia. (Corresponding author) Abdul Manaf Bohari, Ph.D. in Geographical Information Systems, Associate Professor, School of Business Management (SBM), College of Business (COB), University Utara Malaysia (UUM), Sintok, Kedah Darul Aman, Malaysia. This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International license, which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited. Conflict of interest statement: Author(s) reported no conflict of interest 62 Big Data Analytics (BDA) is one of the most digital innovations for supporting supply chain firms’ activities. Empirically, multiple benefits of BDA in Supply Chain Management (SCM) have been demonstrated. The study aimed to investigate the relationship between technical, organizational, and environmental factors and supply chain firms’ performance using the Technology-Organization-Environment (TOE) framework and the Diffusion of Innovation (DOI) theory. This study was conducted at medium-large supply chain firms in Saudi Arabia, the sample size reached 700 firms recognized by Saudi Arabia’s Ministry of Commerce and Industry in different domains. In this study, a questionnaire was used to collect primary data. The collected data are analyzed using SPSS version 26.0. SPSS is used to describe respondents’ demographic profiles. The percentage of respondents to the questionnaire reached 57%. In addition, to test hypotheses and accomplish research goals, PLS-SEM version 3.0 is used to examine the relationship between independent and dependent variables. From the PLS results, the study reported that complexity (β = 0.097, t = 2.817), security (β = 0.222, t = 3.486), IT expertise (β = 0.108, t = 1.993), and external support (β = 0.211, t = 3.468) were positively related to firm’s performance; in contrast, relative advantage (β = –0.006, t = 0.200), compatibility (β = –0.020, t = 0.314), top management support (β = –0.046, t = 0.386), organizational resources (β = –0.065, t = 1.179), competitive pressure (β = –0.011, t = 0.199), and privacy (β = –0.05, t = 0.872) were negatively related to firm’s performance. Keywords supply chain management, performance, diffusion of innovation, technology-organization-environment, Saudi Arabia INTRODUCTION Saudi Arabia has established goals for 2020 and a strategic vision for 2030, focusing on logistics and supply chain management. One of the primary goals of these strategies is to diversify the country’s economic streams away from oil. In Saudi Arabia, BDA technologies will be one of the primary facilitators for the success of Vision 2030, since data will power the Kingdom’s blueprint for the future, creating new opportunities, expanding the economy, and transforming society, which are key objectives of the Saudi Vision 2030. Another goal is to increase investments and lead the digital economy through technology. The Saudi government is also looking for private-sector collaborations to develop information technology infrastructure and telecommunications (Thamir et al., 2020). Implementation of BDA may be problematic due to the high cost, liabilities, complexities, and applicability associated with new technological innovation. These factors might contribute to ambiguity in the choice to implement BDA in SCM operations. Saudi Arabia, continuously shifting environmental patterns, mostly as a result of the Kingdom’s Vision 2030 implementation and other un- http://dx.doi.org/10.21511/kpm.06(1).2022.06 Knowledge and Performance Management, Volume 6, 2022 expected situations like as the COVID-19 epidemic, may cause Saudi firms to re-evaluate their choices, potentially leading them to adopt BDA in order to sustain, or indeed obtain, a market leader (Thamir et al., 2020). Additionally, big data helps firms improve their supplier evaluations and govern procurement activities in any circumstance. Firms may also simulate their supply networks using big data. While numerous studies have exhibited the advantages of utilizing BDA in SCM, research in developing economies is limited. Most of the previous research focused on rich industrialized countries. In contrast, few prior published studies have evaluated Saudi Arabia’s drive to adopt and apply big data analytics technology to SC firms (Thamir et al., 2020). Furthermore, Choi et al. (2018) indicate increasing uses of BD technologies and strategies in many areas of SCM, such as prediction, income management, and risk assessment. However, many CEOs still do not use big data analytics in their decision-making activities. The expected sample size reaches near 700 firms recognized by the Ministry of Commerce and Industry in Saudi Arabia in different domains. Therefore, the findings of this study will provide generalization knowledge and determine the main factors facing firms to adopt big data through testing the adoption of a big data model in firms. Furthermore, this study is important as it contributes to expanding the literature on BDA and SCM by conducting in Saudi Arabia. Moreover, it might used as a reference for future academics seeking to expand their study focus on factors affecting the adoption of big data technologies in middle east countries. 1. LITERATURE REVIEW velocity, veracity, and value) required to provide actionable information for sustained delivery, perA supply chain is a network of suppliers, manufac- formance assessment, and competitive advantage turers, transportation companies, warehouses, re- (Emani et al, 2015). tailers, and customers. Supply chain management aims to control the movement of finances, infor- Generally, big data adoption has an impact on mation, and goods within a supply chain in order SCM performance in many aspects, and it is to maintain a high degree of product availability shown in many studies such as Thamir et al. (2020), and service to the customer at the lowest feasible Govindan et al. (2020), Verhoef et al. (2019), Kamble cost. Nowadays, there is a plethora of records gen- and Gunasekaran (2019), Feki (2019), Agrawal et erated as a result of transactions between suppliers al. (2015), and many more. However, Mikalef et al. and purchasers. The use of big data methods and (2019) observed that recent research asserted that tactics in a variety of supply chain management a sizable proportion of firms are incapable of levthemes, including forecasting, revenue manage- eraging value from the prospects that BDA may ment, and risk analysis, using case studies from provide for their businesses. Some academics have leading brands. However, despite the rising use of even rejected those businesses may improve their big data in supply chain management, some man- performance through BDA. As a result, there may agers continue to avoid incorporating big data an- be a lack of knowledge and conflicting opinions regarding how firms might profit from BDA funds alytics into their decision-making processes. (Wamba et al., 2017). Additionally, the promises of 1.1. Big Data analytics adoption BDA are not thoroughly investigated by businesses such as SCM companies (Mikalef et al., 2019). The term “big data” refers to a massive dataset that cannot be controlled or analyzed using standard BDA’s applications have contributed to the imdatabases (Lin et al., 2019; Mishra et al., 2017). In provement and development of various business comparison, other experts saw big data as a busi- models, including Supply Chain Management ness strategy that enables organizations to exam- (SCM) (Kamble & Gunasekaran, 2019; and ine merely a massive volume of data (Wright et al., Chehbi-Gamoura et al., 2020). Countless prior 2019; Malaka & Brown, 2015). In addition, BDA is studies and publications had shown BDA adopa holistic strategy for managing, processing and tion benefits, including improved risk reduction, analyzing data of various sizes of (volume, variety, demand forecasting, batch size optimization, in- http://dx.doi.org/10.21511/kpm.06(1).2022.06 63 Knowledge and Performance Management, Volume 6, 2022 ventory reduction, and the development of more innovative solutions to boost the satisfaction of clients (Queiroz & Pereira, 2019; Tahiduzzaman et al., 2017; Galea-Pace, 2020; Nguyen et al., 2018; and Nambisan et al., 2019). Despite these benefits, other researchers claim that many organizations are hesitant to accept and use BDA technology (Arunachalam et al., 2018), necessitating an empirical investigation to ascertain the variables that impact and modify enterprises’ motivation to adopt and use BDA in SCM. analytics is a generic phrase that is a term that refers to advanced big data analytics used in managing the supply chain. These analyses might be classified as predictive, descriptive, or prescriptive (Wang et al., 2016). Numerous benefits of BDA in SCM are supported by empirical research, including cost reductions, enhanced SC flexibility, and satisfied customers (Ramanathan et al., 2017). As a result, there is growing interest in defining a distinct skill set for SCM data scientists (Waller & Fawcett, 2013). Big data analytics is rapidly gaining traction among academics and is becoming a 1.2. Big Data analytics and supply priority for businesses to deploy. Daily, incomprehensible amounts of data are transported and colchain performance lected as a result of the widespread use and growth Many companies in various industries use big da- of big data-enabling tools such as social networks, ta analytics technology to manage risks, reduce mobile devices and identification technologies operating costs, and improve supply chain visi- that give consent to the ‘Internet of Things.’ Given bility and traceability, Feki (2019). According to that increased data equates to increased knowlQueiroz and Pereira (2019), BD is a powerful tool edge, firms are increasingly leveraging these techto help organizations conduct analyses. According nologies to establish and preserve competitive adto Abawajy (2015), availability of large amounts of vantage (Wamba et al., 2017). data may play an essential role in creating insights into decision-making processes. Zhu et al. (2018) 1.3. Underpinning theory believe that big data technology improves the performance of supply chain companies. Several pre- A theoretical model is a collection of pre-existing, vious studies conducted on the applications of big approved hypotheses taken from the academic litdata analytics technology showed the possibility of erature. The TOE (technology organization enviapplying the technology in various sectors (Grover ronment) framework (Tornatzky et al., 1990) em& Kar, 2017; Palanisamy & Thirunavukarasu, ployed as the study’s main underpinning theory and 2017; Raut et al., 2019). Despite the hype around Diffusion of innovations (DOI) (Rogers, 1995) are the adoption of big data, none of the studies has some of the most often used Information systems (IS) been investigated in-depth, particularly in a me- adoption theories for analyzing individual or organdium-sized enterprise. However, the relationship izational IT adoption decisions. Many scholars have between the most emphasized factors for Big Data confirmed and tested these hypotheses in various adoption evaluated in a specific situation, Baig et adoption situations, including e-commerce, e-learnal. (2019) recommended using large samples and ing, healthcare, and tourism (Yadegaridehkordi et focusing on mixed-method methods. To develop al., 2018; Chandra & Kumar, 2018; Fan et al., 2018). and validate current theoretical models and en- They play an important role in considering the assure accuracy for future studies, an in-depth and pects that affect any technology acceptance decision by generating a blueprint (Oliveira & Martins, 2011); detailed study is needed. theoretical models must be considered to solve big Supply chain analytics is not a new notion. data acceptance and execution issues at the organiHistorically, supply chain management has de- zational or individual level. Several factors influence pended on statistics and operations research to big data adoption, including IT infrastructure, useoptimize supply and demand matching objectives. fulness, and complexity. With the support of an information system, business analytics has a strong association with supply In conclusion, a survey of the literature, adopting chain performance. Nevertheless, the emergence big data analytics is the process through which an of big data applications in supply chain manage- innovation changes the architecture of an organment does open up new possibilities. Supply chain ization. Big data adoption includes improved in- 64 http://dx.doi.org/10.21511/kpm.06(1).2022.06 Knowledge and Performance Management, Volume 6, 2022 formation processing methods and technological H3: advancements that support decision-making. It gives organizations new opportunities to use information and gain a competitive advantage. The use of big data enhances risk prediction, boosts H4: production, and efficiently satisfies customers. In addition, big data usage helps businesses and sectors to outperform their rivals. Therefore, big data H5: use could be time-consuming and expensive, but the long-term advantages might lead to success. Compatibility is positively related to supply chain management performance at Saudi firms. While big data have been frequently employed in predictive research, there are relatively few studies measuring prediction error in large data. More exactly, beyond the raw data’s quality, the accuracy of big data analysis is strongly influenced by the model used to analyses the data. We still have a long way to go in terms of generating metrics that can be used to assess the accuracy of a method for analyzing large data. The majority of current research on big data applications in supply chain management is theoretical and conceptual, with a distinct dearth of research on analytical models. Additionally, present analytical models focus primarily on the use of big data in modelling sustainability. As a result, there is still a gap in the use of big data to supply chain optimization. Even though empirical research has examined factors influencing BD adoption in a variety of areas, the prior literature demonstrates a dearth of empirical evidence for big data adaptation in Saudi Arabia. As a result, this study will focus on medium-large supply chain enterprises in Saudi Arabia in order to address this absence from the literature H6: IT expertise is positively related to supply chain management performance at Saudi firms. H7: Organizational resources are positively related to supply chain management performance at Saudi firms. H8: Competitive pressure is positively related to supply chain management performance at Saudi firms. H9: External support is positively related to supply chain management performance at Saudi firms. Security is positively related to supply chain management performance at Saudi firms. Top Management Support is positively related to supply chain management performance at Saudi firms. H10: Privacy is positively related to supply chain management performance at Saudi firms. 3. METHOD This study was conducted at medium-large supply chain firms in Saudi Arabia, the sample size reached to 700 firms recognized by Saudi 2. AIMS AND HYPOTHESES Arabia’s Ministry of Commerce and Industry in different domains (such as Airlines, Banks, This study aims to examine the relationship be- Broadcasting & entertainment, Building matween technical, organizational, and environ- terials & fixtures, Business support services, mental factors, and SCM performance at medi- Exploration & Production, Food products, Food um-large supply chain firms in Saudi Arabia retailers & wholesalers, General mining, Heavy construction, Insurance, Investment Services, Ten (10) hypotheses were formulated as follows: Marine transportation, Publishing, Railroad, Real estate holding & development, Restaurants, H1: Relative advantage is positively related Specialty chemicals, telecommunications) and to supply chain management performance different locations (such as Riyadh, Dammam, at Saudi firms. Dhahran, Jeddah, Jubail, Khobar). The study was targeted employees at various levels within their H2: Complexity is positively related to supply chain organizations, including owners, senior managemanagement performance at Saudi firms. ment, middle management, information tech- http://dx.doi.org/10.21511/kpm.06(1).2022.06 65 Knowledge and Performance Management, Volume 6, 2022 nology expertise, and staff members. The study sample also included all departments at supply chain firms such as procurement, inventory, buying, expediting, import/export, operations, and logistics. The study population was compiled from two primary sources. A commercial directory published by Saudi Arabia’s Ministry of Commerce and Industry is the first source. The second source population for the study was derived from a Saudi Arabian commercial directory (Daleeli), which provided firm names, phone numbers, fax numbers, and web page links to broaden the target population. To achieve the study’s objectives, a quantitative methodology was employed. In this study, a questionnaire was used to collect primary data, the respondents were asked to answer the items presented in a questionnaire. The researchers conducted both online and hard-copy forms to distribute these questionnaires to the respondents within three months in the central cities of Saudi Arabia. The survey identified the demographic profile and included five variables, current position of employees, company field, company size by annual revenue, company size by employees, and company experience in the field of business. According to the survey’s findings, the percentage of employees who responded to the questionnaire were workers in CIO/IT director/Technology director, and their percentage reached about 77.1%. In addition, most of the companies that responded to the questionnaire were companies working in the field of Transport/ logistics/post, as their percentage reached 31.3%. Moreover, the percentage of responding companies in terms of annual revenues (85-150 million SAR) reached 52.1%. In the end, the responding companies met the employee size range (50-100 employees) (63.2%). The questionnaire was divided into two sections, factors of adoption and firm performance. The Factors of Adoption consist of three main variables; Technical Contexts consist of four sub-variables (Relative advantage, Complexity, Compatibility, and Security). Organizational Contexts consist of three sub-variables (Top Management Support, IT expertise, and Organizational). Environmental context consists of three sub-variables (Competitive Pressure, External Support, and 66 Privacy). Measurement items of factors of adoption formation (37 question items) were adapted from Davis (1989), Boonsiritomachai (2014). The firm performance (Score) consists of five variables, which are Plan, Source, Make, Deliver, and Return. The measurement items for firm performance (17question items) were adapted from Santos and Leite (2018). The questionnaire’s measuring items were all scored on a Likert scale with a range of 1 (Strongly Disagree) to 5 (Strongly Agree). The pretest was done utilizing expert judgments, as the questionnaire in this study was composed of measurements from various situations. As a result, the questionnaire was revised based on the advice of three experts to assure the scales’ reliability and validity. The expert’s identity has been withheld to adhere to the expert’s confidentiality obligations. a pilot test with 100 responders from one of the SCM firms included in this research is also done. All indicators have a Cronbach’s Alpha value greater than 0.70. Technical Contexts α = 0.862; Organizational Contexts α = 0.881; Environmental Contexts α = 0.910; And Firm Performance α = .915) indicating adequate internal consistency. The collected data is analyzed using Statistical Package of Social Science (SPSS) techniques via SPSS software version 26.0. SPSS is used to describe respondents’ demographic profiles, report descriptive statistics, and Partial Least Squares Structural Equation Modelling (PLS-SEM) techniques via SmartPLS software version 3.0. PLSSEM is applied to investigate the connection between independent and dependent variables to verify hypotheses and achieve research objectives. 4. RESULTS The measurement model was assessed by evaluating outer loadings, composite reliability and validity, Cronbach’s alpha (α), average variance extracted (AVE), and factor loadings, as shown in Table 1. All the constructs had Cronbach’s alpha (α) and CR greater than 0.70, meeting construct reliability criteria. All average variance extracted (AVE) values were more significant than 0.50, meeting convergent validity criteria. And the factor loadings, greater than 0.70. http://dx.doi.org/10.21511/kpm.06(1).2022.06 Knowledge and Performance Management, Volume 6, 2022 Table 1. Results of Outer Loading, Cronbach’s Alpha, CR, and AVE for composite variables Latent variables Relative advantage Complexity Technical Contexts Compatibility Security Top Management Support Organizational Contexts IT expertise Organizational Resources Competitive Pressure Environmental Contexts External Support Privacy Plan Source Firm Performance Make Delivery Return Indicators Factor loadings TRA1 TRA2 TRA3 TRA4 TCOM1 TCOM2 TCOM3 TCOM4 TCOP1 TCOP2 TCOP3 TCOP4 TSEC1 TSEC2 TSEC3 TSEC4 TSEC5 OTMS1 OTMS2 OTMS3 OITE1 OITE2 OITE3 OOR1 OOR2 OOR3 ECP1 ECP2 ECP3 ECP4 EES1 EES2 EES3 EPR1 EPR2 EPR3 EPR4 FPP1 FPP2 FPP3 FPS1 FPS2 FPS3 FPS4 FPS5 FPM1 FPM2 FPM3 FPM4 FPM5 FPD1 FPD2 FPD3 FPR1 FPR2 FPR3 FPR4 0.870 0.820 0.856 0.836 0.779 0.750 0.714 0.807 0.793 0.796 0.813 0.743 0.843 0.767 0.850 0.736 0.861 0.827 0.778 0.845 0.863 0.821 0.803 0.858 0.788 0.755 0.859 0.793 0.831 0.813 0.857 0.791 0.864 0.732 0.876 0.835 0.877 0.848 0.842 0.716 0.786 0.755 0.831 0.711 0.805 0.825 0.791 0.805 0.704 0.819 0.768 0.818 0.789 0.723 0.860 0.868 0.804 Cronbach’s alpha CR AVE 0.868 0.912 0.716 0.762 0.848 0.583 0.796 0.867 0.619 0.871 0.907 0.661 0.751 0.857 0.668 0.772 0.868 0.687 0.721 0.843 0.642 0.843 0.895 0.68 0.787 0.876 0.702 0.851 0.904 0.693 0.726 0.845 0.647 0.838 0.885 0.606 0.848 0.892 0.624 0.703 0.835 0.628 0.832 0.888 0.666 Note: Factor Loadings above 0.708; Cronbach’s Alpha above 0.708; Composite Reliability (CR) above 0.708; Average Variance Extracted (AVE) above 0.50. http://dx.doi.org/10.21511/kpm.06(1).2022.06 67 Knowledge and Performance Management, Volume 6, 2022 The structural model was assessed by evaluating collinearity statistics (VIF), coefficient of determination (R 2), path coefficients (β), effect size (f2), and predictive relevance (Q2). The VIF values were less than five, confirming collinearity. Furthermore, the path coefficients and total effects were analyzed and supported four out of ten hypotheses. The estimated model in this research study could explain 73.8% of the variance in the plan, 81.5% of the variance in source, 77.8% of the variance in the making, 67.7% of the variance in delivery, and 70.8% of the variance in return. Predictive relevance (Q2) values were more outstanding than zero, indicating that the estimated model was predictive. Table 1 shows the results, indicating that the importance of the path coefficient is proven using t statistics and p-value evaluation at the .05, .01, and .001 confidence interval levels. Table 2 shows the structural model estimates of the hypothesized relationships, both direct and indirect effects. The direct effect of relative advantage on firm performance is statistically and negatively insignificant (β = –0.006, p < 0.842, t = 0.200), not supporting Hypothesis H1. The direct effect of complexity on firm performance is statistically significant (β = 0.097, p < 0.005, t = 2.817), supporting H2. The direct effect of compatibility on firm performance is statistically insignificant (β = –0.020, p < 0.754, t = 0.314), not supporting H3. Security directly affected firm performance that is statistically significant (β = 0.222, p < 0.001, t = 3.486), supporting H4. The direct effect of top management support on firm performance was statistically insignificant (β = 0.046, p < 0.868, t = 0.386), not supporting H5. The direct effect of IT expertise on firm performance is sta- tistically significant (β = 0.108, p < 0.047, t = 1.993), supporting H6. Organizational resources on firm performance were statistically insignificant (β = 0.065, p < 0.239, t = 1.179), not supporting H7. The direct effect of competitive pressure on firm performance is statistically negative insignificant (β = –0.011, p < 0.842, t = 0.199), not supporting H8. The direct effect of external support on firm performance was statistically significant (β = 0.211, p < 0.001, t = 3.468), supporting H9. Finally, the direct effect of privacy on firm performance is statistically insignificant (β = 0.05, p < 0.384, t = 0.872), not supporting H10. 5. DISCUSSION 5.1. Discussion of the relationship between technical contexts and SCM performance This section discusses how to evaluate hypotheses H1, H2, H3, and H4. H1: Relative advantage is positively related to supply chain management performance at Saudi firms. The data analysis results did not support H1, implying that there is no association between relative advantage and business performance. The study’s findings on the link between relative advantage and business performance are less significant in the setting of Big Data adoption. Agrawal (2015) and Nam et al. (2015) found no evidence for relative advantage in studies of Big Data adoption, hinting that relative advantage may be redundant in such investigations. However, the lack of awareness of how big data analytics technology may be used to improve Table 2. Path coefficient of variables H H1 H2 H3 H4 H5 H6 H7 H8 H9 H10 68 Paths Relative advantage → Firm Performance Complexity → Firm Performance Compatibility → Firm Performance Security → Firm Performance Top Management Support → Firm Performance IT expertise → Firm Performance Organizational Resources → Firm Performance Competitive Pressure → Firm Performance External Support → Firm Performance Privacy → Firm Performance β T statistics P values Support –0.006 0.097 –0.020 0.222 0.046 0.108 0.065 –0.011 0.211 0.05 0.200 2.817 0.314 3.486 0.868 1.993 1.179 0.199 3.468 0.872 0.842 0.005 0.754 0.001 0.386 0.047 0.239 0.842 0.001 0.384 No Yes No Yes No Yes No No Yes No http://dx.doi.org/10.21511/kpm.06(1).2022.06 Knowledge and Performance Management, Volume 6, 2022 an organization’s performance can negatively affect its adoption. In comparison to the first hypothesis. The employee of a business feels that big data analytics technology offers several benefits, including improving the firm’s performance. As a result, the benefits of using big data analytics technologies are obvious to business decision-makers and staff. H2: Complexity is positively related to supply chain management performance at Saudi firms. H4: Security is positively related to supply chain management performance at Saudi firms. H4 is corroborated by the path coefficient values. As a result, the security factor and SCM performance have a favorable correlation. This result suggested that security was associated with company performance. In contrast, earlier research has established a negative correlation between security and the use of Big Data technologies (Salleh & Janczewski, 2016; Nguyen & Petersen, 2017). The survey resulted in a high mean security factor (4.40), indicating that businesses possess the necessary security capabilities to adopt this technology, possess the necessary skills to ensure data security when utilizing this BDA technology, can easily integrate security policies for technology, have adequate tools and mechanisms in place to ensure effective data protection when utilizing this BDA technology, and possess the necessary security capabilities to adopt this technology. As a consequence, businesses can protect their customers’ data and prevent security concerns. The customers should have no concerns about security. The findings on the influence of complexity on business performance indicated a positive link, corroborating H2. This study suggested that the complexity of the firm had a positive correlation with its success. While earlier research indicates that complexity has a detrimental effect on BDA acceptance (Verma & Chaurasia, 2019; Agrawal, 2015), a few further studies indicate that complexity does not affect BDA (Lai et al., 2018) or other IT innovation adoption (Maduku et al., 2016). In comparison, the degree of complexity is mostly determined by the users who will interact with technology, as the descriptive demographics revealed that the majority of respondents (77.1 per- 5.2. Discussion of correlation cent) specialize in information technology. As between organizational contexts a result, users will have no difficulty interacting and SCM performance with this technology, since it is simple to grasp, engaging with it is clear and intelligible, and it is sim- This section explains the methodology for evaluating hypotheses H5, H6, and H7. ple to become proficient at utilizing it. H3: Compatibility is positively related to supply chain management performance at Saudi firms. Analyses of the data indicate a negative correlation between compatibility and company performance. This means that the infrastructure compatibility for adopting big data analytics technology negatively affects the firm performance. Hence, H3 is not supported. Previous studies have shown an insignificant effect of compatibility on BDA adoption (Maroufkhani et al., 2020). Furthermore, managers’ decisions on BDA adoption are influenced by compatibility. Organizational decision-makers will be more receptive to adopting and using BDA technology in various aspects of supply chain operations if they believe it is consistent with the data acquired, current operating procedures, IT infrastructure, and company values and beliefs. http://dx.doi.org/10.21511/kpm.06(1).2022.06 H5: Top management support is positively related to supply chain management performance at Saudi firms. The data analysis results did not support H5, indicating that there is a direct influence of top management support on business performance. This finding suggested that TMS had no bearing on company performance. On the other hand, the outcome contradicts earlier research indicating that TMS is a recurring and critical factor in organizations’ intentions to utilize big data (Maroufkhani, et al., 2020; Thamir et al., 2020; Jang et al., 2018; Sun et al., 2019; Lai et al., 2018; Verma & Chaurasia, 2019; Nguyen and Petersen, 2017). However, senior management support is crucial in fostering a climate conducive to company adoption of new technologies (Asiaei & Rahim, 2019; Scupola, 2009; Maduku et al., 2016). Top ex- 69 Knowledge and Performance Management, Volume 6, 2022 ecutives are catalysts for organizational change, communicating and promoting values through a well-articulated vision for the business (CruzJesus et al., 2019; Kandil et al., 2018). Furthermore, assistance from senior management can aid in the learning and spread of technology throughout the organization (Asiaei & Rahim, 2019). As a result, they are critical throughout the BDA adoption process. However, the study’s findings indicate that this factor has a negligible effect on top management’s adoption of big data technology, owing to the existence of other factors that may have a greater impact on technology adoption, such as security and privacy concerns, the technology uses, and IT expertise in big data analytics technology. H6: IT expertise is positively related to supply chain management performance at Saudi firms. 2009; Thamir et al., 2020). Additionally, this study supports Gangwar (2018) finding that organizational resources may not have an impact on big data adoption at the original intention stage but rather only during the post-adoption phases, which reflects the degree of implementation and utilization. In addition, many firms seek to acquire benefits from using cloud computing technologies, where cloud computing can provide all the technical needs of firms, thus firms can save the costs of acquiring infrastructure and save time to manage it. 5.3. Discussion of correlation between environmental contexts and SCM performance This section discusses hypotheses H8, H9, and H10. H8: The effect of IT expertise on firm performance revealed a positive relationship, therefore, supporting H6. On the other hand, these results implied that there is a positive relationship between IT expertise and firm performance. Therefore, empirical evidence indicates that IT competence has a beneficial effect on IT adoption across a range of technologies, including Big Data adoption (Nam et al., 2015). Expertise in information technology is regarded as one of the company’s most valuable assets. Companies can provide all facilities to adopt big data analytics technology such as IT infrastructure, but they are facing big problems in providing big data technology specialists. Moreover, especially in the Middle East countries, the existence of IT expertise in BDA is very little, and it is not easy to attract them from other countries, so the presence of this expertise will greatly help the company’s decisions in adopting big data technology to raise the performance of the company. The data analysis results indicate that H8 is incorrect, implying that there is a negative link between competitive pressure and company performance. By contrast, the findings contradict prior research on competitive pressure as a predictor of BDA uptake (Lai et al., 2018; Jang et al., 2018; Gangwar, 2018; Verma & Chaurasia, 2019; Schüll & Maslan, 2018; Thamir et al., 2020). Competitive pressure is a larger motivator when it comes to deciding to use Big Data technologies. This means that firms may feel obligated to adopt Big Data technologies ahead of time to prevent a future competitive disadvantage owing to competition pressure. As a result, if enterprises in Saudi Arabia wish to preserve their market position, the adoption of Big Data technology may be considered a strategic requirement. H9: H7: Organizational resources are positively related to supply chain management performance at Saudi firms. H7 is not supported based on the results of the path coefficient. As a result, the association between organizational resources and SCM performance is minimal. The findings corroborate prior studies (Alsaad et al., 2019; Wang & Ahmed, 70 Competitive pressure is positively related to supply chain management performance at Saudi firms. External support is positively related to supply chain management performance at Saudi firms. The outcomes of the effect of external support on firm performance revealed a positive relationship, therefore, supporting H9. In other words, these results imply that there is a positive relationship between external support and firm performance. Furthermore, the research http://dx.doi.org/10.21511/kpm.06(1).2022.06 Knowledge and Performance Management, Volume 6, 2022 finding consistent with previous research has provided a correlation between external support and new technology adoption (Al-Isma’ili et al., 2016; Premkumar & Roberts, 1999; Nguyen & Petersen, 2017; Hung et al., 2016). In light of the hypothesis’s findings, the existence of vendors and agents in the Saudi market working to provide support and training for companies will motivate firms to adopt big data analytics technology, which will reflect positively on the firm’s performance. As a result, the more external help available, the more probable it is that Big Data technologies will be embraced. and security are paramount (Priyadarshinee et al., 2017). Fear of losing control of personal information and the danger of information being leaked to competitors may hinder BDA adoption. Additionally, Nam et al. (2015) uncovered evidence for a link between privacy regulations and Big Data usage in a prior study on the issue. Privacy issues are among the most important issues in Saudi society. Due to the nature of the conservative society, it is not easy to obtain all the information about customers because there is a belief among customers that it is possible to use this information for other things that go beyond marketing studies. In addition, there is H10: Privacy is positively related to supply chain a prevailing belief in the society of Saudi Arabia that their data is vulnerable to penetration management performance at Saudi firms. or distortion or distribution to other parties. H10 is not supported based on the results of Despite the existence of regulations and legislathe path coefficient. Therefore, there is an in- tion that punishes anyone who infringes the prisignificant relationship between privacy and vacy of others with deterrent punishments, soSCM Performance. This finding suggested that ciety still has a fear of sharing data with others. security was related to the firm’s performance. As a result, the presumed challenges associated According to previous studies, the major con- with complying with privacy requirements decerns of business owners are when it comes to crease the likelihood that Big Data technologies deploying data-related technologies, privacy will be implemented. CONCLUSION The purpose of this study was to investigate the relationship between technical, organizational, and environmental adoption factors, and SCM performance at medium-large supply chain firms in Saudi Arabia. Two crucial implications may be drawn from the PLS results. First, complexity, security, IT expertise, and external support factors demonstrate their positive effect on BDA adoption, which improves firm performance. In contrast, relative advantage, compatibility, top management support, organizational resources, competitive pressure, and privacy factors have been shown to negatively affect the adoption of BDA, thus decreasing firm performance. Based on the findings, the study recommended that firms prioritize spreading and growing awareness inside the organization about the key advantages of BDA technology to all stakeholders. Additionally, firms must think about the challenges of implementing BDA technology as well as additional elements that might affect acceptance, such resistance to change, data quality and integration, organizational learning culture, and decision-making culture. Therefore, firms are required to develop appropriate plans on how to overcome these challenges during the planning stages for the deployment of technology. Future studies should incorporate the suggested paradigm for big data analytics technology adoption with other theories, such as RBV, TTF, and TAM. The framework’s ability to explain previously unrecognized features of big data analytics technology adoption and deployment may be enhanced by including more theory-based context. In the end, the study was limited to medium-sized Saudi Arabian supply chain companies. To generalize the study’s results to the Saudi market, a framework that can be evaluated at small- to medium-sized businesses in Saudi Arabia would be needed. http://dx.doi.org/10.21511/kpm.06(1).2022.06 71 Knowledge and Performance Management, Volume 6, 2022 AUTHOR CONTRIBUTIONS Conceptualization: Adel Hamed, Abdul Manaf Bohari. Data curation: Adel Hamed, Abdul Manaf Bohari. Methodology: Adel Hamed. Project administration: Adel Hamed, Abdul Manaf Bohari. Writing – original draft: Adel Hamed. Writing – review & editing: Abdul Manaf Bohari. REFERENCES 1. 2. 3. 4. 5. 6. 72 Abawajy, J. (2015). Comprehensive analysis of big data variety landscape. International Journal of Parallel, Emergent and Distributed Systems, 30(1), 5-14. https://doi.org /10.1080/17445760.2014.925548 7. Agrawal, K. P. (2015). Investigating the determinants of big data analytics (BDA) adoption in emerging economies. Academy of Management Proceedings, 1, 1121011290. https://doi.org/10.5465/ ambpp.2015.11290abstract Alsaad, A., Mohamad, R., & Ismail, N. A. (2019). The contingent role of dependency in predicting the intention to adopt B2B e-commerce. Information Technology for Development, 25(4), 686-714. https://doi.org/10.1080/02681102.2 018.1476830 Arunachalam, D., Kumar, N., & Kawalek, J. P. (2018). Understanding Big Data Analytics Capabilities in Supply Chain Management: Unravelling the Issues, Challenges and Implications for Practice. Transportation Research Part E: Logistics and Transportation Review, 114, 416436. https://doi.org/10.1016/j. tre.2017.04.001 Asiaei, A., & Rahim, N. Z. A. (2019). A multifaceted framework for adoption of cloud computing in Malaysian SMEs. Journal of Science and Technology Policy Management, 10(3), 708-750. https://doi. org/10.1108/JSTPM-05-2018-0053 Baig, M. I., Shuib, L., & Yadegaridehkordi, E. (2019). Big data adoption: State of the art and research challenges. Information Processing & Management, 56(6), 102095. https://doi.org/10.1016/j. ipm.2019.102095 8. 9. Boonsiritomachai, W., McGrath, M., & Burgess, S. (2014). A research framework for the adoption of Business Intelligence by Small and Medium-sized enterprises. In Proceedings of the 27th Annual Conference on Small Enterprise Association of Australia & New Zealand (pp. 1-22). SEAANZ. Retrieved from https:// silo.tips/download/a-researchframework-for-the-adoption-ofbusiness-intelligence-by-smalland-medi Chandra, S., & Kumar, K. N. (2018). Exploring factors influencing organizational adoption of augmented reality in e-commerce: Empirical analysis using technology organization environment model. Journal of Electronic Commerce Research, 19(3), 237-265. Retrieved from http://www.jecr.org/sites/default/ files/2018vol19no3_paper3.pdf Chehbi-Gamoura, S., Derrouiche, R., Damand, D., & Barth, M. (2020). Insights from Big Data Analytics in Supply Chain Management: An All-Inclusive Literature Review Using the SCOR Model. Production Planning and Control, 31(5), 355-382. https:// doi.org/10.1080/09537287.2019.1 639839 framework. Computers in Industry, 109, 1-13. https://doi.org/10.1016/j. compind.2019.03.007 12. Davis, F. D. (1989). Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology. MIS Quarterly: Management Information Systems, 13(3), 319-339. https://doi. org/10.2307/249008 13. Emani, C. K., Cullot, N., & Nicolle, C. (2015). Understandable big data: a survey. Computer Science Review, 17, 70-81. https://doi. org/10.1016/j.cosrev.2015.05.002 14. Fan, W., Liu, J., Zhu, S., & Pardalos, P. M. (2018). Investigating the impacting factors for the healthcare professionals to adopt artificial intelligence-based medical diagnosis support system (AIMDSS). Annals of Operations Research, 294, 567-592. https://doi. org/10.1007/s10479-018-2818-y 15. Feki, M. (2019). Big data analytics-driven supply chain transformation. In K. Mezghani & W. Aloulou (Eds.), Business transformations in the era of digitalization (pp. 106124). IGI Global. https://doi. org/10.4018/978-1-5225-7262-6. ch007 10. Choi, T. M., Wallace, S. W., & Wang, Y. (2018). Big data analytics in operations management. Production and Operations Management, 27(10), 18681883. https://doi.org/10.1111/ poms.12838 16. Galea-Pace, S. (2020). How Is Big Data Transforming the Supply Chain? Supply Chain Digital. Retrieved from https://supplychaindigital.com/technology/howbig-data-transforming-supplychain 11. Cruz-Jesus, F., Pinheiro, A., & Oliveira, T. (2019). Understanding CRM adoption stages: empirical analysis building on the TOE 17. Gangwar, H. (2018). Understanding the determinants of big data adoption in India: An analysis of the manufacturing and services http://dx.doi.org/10.21511/kpm.06(1).2022.06 Knowledge and Performance Management, Volume 6, 2022 sectors. Information Resources Management Journal (IRMJ), 31(4), 1-22. Retrieved from https://www. igi-global.com/article/understanding-the-determinants-of-big-dataanalytics-adoption/230368 puting adoption in Egypt. The Business & Management Review, 9(4), 113-123. Retrieved from https://cberuk.com/cdn/conference_proceedings/2019-07-12-2127-47-PM.pdf 18. Govindan, K., Rajeev, A., Padhi, S. S., & Pati, R. K. (2020). Supply Chain Sustainability and Performance of Firms: A Meta-Analysis of the Literature. Transportation Research Part E: Logistics and Transportation Review, 137, 101923. https://doi. org/10.1016/j.tre.2020.101923 25. Lai, Y., Sun, H., & Ren, J. (2018). Understanding the determinants of big data analytics (BDA) adoption in logistics and supply chain management: An empirical investigation. International Journal of Logistics Management, 29(2), 676-703. https://doi. org/10.1108/IJLM-06-2017-0153 19. Grover, P., & Kar, A. K. (2017). Big data analytics: A review on theoretical contributions and tools used in literature. Global Journal of Flexible, Systems Management, 18(3), 203-229. https://doi. org/10.1007/s40171-017-0159-3 26. Maduku, D., Mpinganjira, M., & Duh, H. (2016). Understanding mobile marketing adoption intention by South African SMEs: A multi-perspective framework. International Journal of Information Management, 36(5), 711-723. https://doi.org/10.1016/j. ijinfomgt.2016.04.018 20. Hung, S. Y., Huang, Y. W., Lin, C. C., Chen, K. C., & Tarn, J. M. (2016). Factors influencing business intelligence systems implementation success in the enterprises. PACIS 2016 Proceedings, 297. Retrieved from https://aisel.aisnet. org/pacis2016/297/ 21. Isma’ili, A., Li, M., Shen, J., & He, Q. (2016). Cloud computing adoption determinants: an analysis of Australian SMEs. PACIS 2016 Proceedings, 209. Retrieved from https://aisel.aisnet.org/pacis2016/209 22. Jang, W. J., Kim, S. S., Jung, S. W., & Gim, G. Y. (2018, July). A study on the factors affecting intention to introduce Big data from smart factory perspective. In 3rd IEEE/ ACIS International Conference on Big Data, Cloud Computing, and Data Science Engineering (pp. 129156). Springer, Cham. https://doi. org/10.1007/978-3-319-96803-2_11 23. Kamble, S. S., & Gunasekaran, A. (2019). Big data-driven supply chain performance measurement system: a review and framework for implementation. International Journal of Production Research, 58(1), 65-86. https://doi.org/10.108 0/00207543.2019.1630770 24. Kandil, A. M. N. A., Ragheb, M. A., Ragab, A. A., & Farouk, M. (2018). Examining the effect of TOE model on cloud com- http://dx.doi.org/10.21511/kpm.06(1).2022.06 27. Malaka, I., & Brown, I. (2015). Challenges to the Organisational Adoption of Big Data Analytics: A Case Study in the South African Telecommunications Industry. SAICSIT 15: The 2015 Annual Research Conference of the South African Institute of Computer Scientists and Information Technologists (pp. 1-27). https:// doi.org/10.1145/2815782.2815793 28. Maroufkhani, P., Tseng, M. L., Iranmanesh, M., Ismail, W. K. W., & Khalid, H. (2020). Big data analytics adoption: Determinants and performances among small to medium-sized enterprises. International Journal of Information Management, 54, 102190. https://doi.org/10.1016/j.ijinfomgt.2020.102190 29. Mikalef, P., Boura, M., Lekakos, G., & Krogstie, J. (2019). Big data analytics and firm performance: Findings from a mixed-method approach. Journal of Business Research, 98, 261-276. https://doi. org/10.1016/j.jbusres.2019.01.044 30. Mishra, D., Gunasekaran, A., Papadopoulos, T., & Childe, S. J. (2018). Big Data and Supply Chain A., Management: A Review and Bibliometric Analysis. Annals of Operations Research, 270(1-2), 313-336. https://doi.org/10.1007/ s10479-016-2236-y 31. Nam, D. W., Kang, D. W., & Kim, S. (2015). Process of big data analysis adoption: Defining big data as a new IS innovation and examining factors affecting the process. In 2015 48th Hawaii International Conference on System Sciences (pp. 4792-4801). IEEE. Retrieved from https://ieeexplore.ieee.org/document/7070389 32. Nambisan, S., Wright, M., & Feldman, M. (2019). The Digital Transformation of Innovation and Entrepreneurship: Progress, Challenges and Key Themes. Research Policy, 48(8), 103773. https://doi.org/10.1016/j.respol.2019.03.018 33. Nguyen, T., & Petersen, T. E. (2017). Technology adoption in Norway: organizational assimilation of big data (Master’s Thesis). Retrieved from https://openaccess. nhh.no/nhh-xmlui/bitstream/ handle/11250/2455449/masterthesis.PDF 34. Nguyen, T., Zhou, L., Spiegler, V., Ieromonachou, P., & Lin, Y. (2018). Big Data Analytics in Supply Chain Management: A State-of-the-Art Literature Review. Computers and Operations Research, 98, 254-264. https://doi. org/10.1016/j.cor.2017.07.004 35. Oliveira, T., & Martins, M. F. (2011). Literature review of information technology adoption models at the firm level. The Electronic Journal Information Systems Evaluation, 14(1), 110-121. Retrieved from https://academicpublishing.org/index.php/ejise/ article/view/389/352 36. Palanisamy, V., & Thirunavukarasu, R. (2017). Implications of big data analytics in developing healthcare frameworks – A review. Journal of King Saud University Computer and Information Sciences, 31(4), 415-425. https://doi.org/10.1016/j. jksuci.2017.12.007 37. Premkumar, G., & Roberts, M. (1999). Adoption of new information technologies in rural small businesses. Omega, 27(4), 467-484. https://doi.org/10.1016/S03050483(98)00071-1 38. Priyadarshinee, P., Raut, R. D., Jha, M. K., & Kamble, S. S. (2017). A 73 Knowledge and Performance Management, Volume 6, 2022 cloud computing adoption in Indian SMEs: Scale development and validation approach. The Journal of High Technology Management Research, 28(2), 221-245. https://doi. org/10.1016/j.hitech.2017.10.010 39. Queiroz, M. M., & Pereira, S. C. F. (2019). Intention to adopt big data in supply chain management: A Brazilian perspective. Revista de Administração de Empresas, 59(6), 389-401. https://doi.org/10.1590/ S0034-759020190605 40. Ramanathan, U., Subramanian, N., & Parrott, G. (2017). Role of Social Media in Retail Network Operations and Marketing to Enhance Customer Satisfaction. International Journal of Operations and Production Management, 37(1), 105-123. https://doi.org/10.1108/ IJOPM-03-2015-0153 41. Raut, R. D., Mangla, S. K., Narwane, V. S., Gardas, B. B., Priyadarshinee, P., & Narkhede, B. E. (2019). Linking big data analytics and operational sustainability practices for sustainable business management. Journal of Cleaner Production, 224, 10-24. https://doi.org/10.1016/j. jclepro.2019.03.181 42. Rogers, E. M. (1995). Lessons for guidelines from the Diffusion of innovations. Joint Commission Journal on Quality and Patient Safety, 21(7), 324-328. https://doi. org/10.1016/s1070-3241(16)301559 43. Salleh, K. A., & Janczewski, L. (2016). Adoption of Big Data Solutions: A study on its security determinants using Sec-TOE Framework. CONF-IRM 2016 Proceedings, 66. Retrieved from http://aisel.aisnet.org/confirm2016/66 44. Santos, T. F., & Leite, M. S. A. (2018). Performance Measurement System Based on Supply Chain Operations Reference Model: Review and Proposal. In G. P. Moynihan (Ed.), Contemporary Issues and Research in Operations Management. https://doi. org/10.5772/INTECHOPEN.76307 45. Schüll, A., & Maslan, N. (2018). On the Adoption of Big Data Analytics: Interdependencies of Con- 74 textual Factors. In Proceedings of the 20th International Conference on Enterprise Information Systems (ICEIS), 1, 425-431. https://doi. org/10.5220/0006759904250431 46. Scupola, A. (2009). SMEs’e‐commerce adoption: perspectives from Denmark and Australia. Journal of Enterprise Information Management, 22(1-2), 152-166. https://doi. org/10.1108/17410390910932803 47. Sun, S., Cegielski, C. G., Jia, L., & Hall, D. J. (2018). Understanding the factors affecting the organizational adoption of big data. Journal of Computer Information Systems, 58(3), 193-203. https:// doi.org/10.1080/08874417.2016.1 222891 48. Tahiduzzaman, Md., Rahman, M., Dey, S. K., Rahman, Md S., & Akash, S. M. (2017). Big Data and Its Impact on Digitized Supply Chain Management. IJRDOJournal of Business Management, 3(9), 196-208. 49. Thamir, H. A., Mezghani, K., & Alsadi, A. K. (2020). Examining the adoption of Big data analytics in supply chain management under competitive pressure: evidence from Saudi Arabia. Journal of Decision Systems, 30(23), 300-320. https://doi.org/10.108 0/12460125.2020.1859714 50. Tornatzky, L. G., Fleischer, M., & Chakrabarti, A. K. (1990). Processes of technological innovation. Lexington Books. 51. Verhoef, P. C., Thijs, B., Yakov, B., Abhi, B., John, Q., Nicolai, F., & Michael, H. (2019). Digital transformation: A multidisciplinary reflection and research agenda. Journal of Business Research, 122, 889901. https://doi.org/10.1016/j. jbusres.2019.09.022 52. Verma, S., & Chaurasia, S. (2019). Understanding the determinants of big data analytics adoption. Information Resources Management Journal (IRMJ), 32(3), 1-26. https://doi.org/10.4018/ IRMJ.2019070101 53. Waller, M. A., & Fawcett, S. E. (2013). Data Science, Predictive Analytics, and Big Data: A Revolution That Will Transform Supply Chain Design and Management. Journal of Business Logistics, 34(2), 77-84. https://doi. org/10.1111/jbl.12010 54. Wamba, S. F., Ngai, E. W. T., Riggins, F., & Akter, S. (2017). Guest editorial: Transforming operations and production management using big data and business analytics: future research directions. International Journal of Operations & Production Management, 37(1), 2-9. https://doi.org/10.1108/ IJOPM-07-2016-0414 55. Wang, C., Li, X., Zhou, X., Wang, A., & Nedjah, N. (2016). Soft Computing in Big Data Intelligent Transportation Systems. Applied Soft Computing Journal, 38, 10991108. https://doi.org/10.1016/j. asoc.2015.06.006 56. Wang, Y., & Ahmed, P. K. (2009). The moderating effect of the business strategic orientation on eCommerce adoption: Evidence from UK family run SMEs. The Journal of Strategic Information Systems, 18(1), 16-30. https://doi. org/10.1016/j.jsis.2008.11.001 57. Wright, L. T., Robin, R., Stone, M., & Aravopoulou, E. (2019). Adoption of Big Data Technology for Innovation in B2B Marketing. Journal of Business-to-Business Marketing, 26(3-4), 281-93. https://doi.org/10.1080/105171 2X.2019.1611082 58. Yadegaridehkordi, E., Nilashi, M., Nasir, M. H. N. B. M., & Ibrahim, O. (2018). Predicting determinants of hotel success and development using structural equation modelling (SEM)ANFIS method. Tourism Management, 137(1), 199-210. https://doi. org/10.1016/j.techfore.2018.07.043 59. Zhu, S., Song, J., Hazen, B.T., Lee, K., & Cegielski, C. (2018). How supply chain analytics enables operational supply chain transparency: An organisational information processing theory perspective. International Journal of Physical Distribution & Logistics Management, 48(1), 47-68. https://doi.org/10.1108/IJPDLM-11-2017-0341 http://dx.doi.org/10.21511/kpm.06(1).2022.06
https://openalex.org/W3199760333	https://opus.bibliothek.uni-augsburg.de/opus4/files/91138/Thummerer_2021_J_Phys _Conf_Ser.pdf	English	null	Hybrid modeling of the human cardiovascular system using NeuralFMUs	Journal of physics. Conference series	2,021	cc-by	5,985	PAPER • OPEN ACCESS PAPER • OPEN ACCESS You may also like Interaction between runaway electrons and internal kink in a post-disruption plasma Yueqiang Liu, L. Li, C.C. Kim et al. - HYDRO-PACIAE, a hydrodynamic and transport hybrid model for ultra-relativistic heavy ion collisions Yu-Liang Yan, Yun Cheng, Dai-Mei Zhou et al. - LITHIUM DEPLETION IS A STRONG TEST OF CORE-ENVELOPE RECOUPLING Garrett Somers and Marc H. Pinsonneault - Journal of Physics: Conference Series Journal of Physics: Conference Series You may also like Interaction between runaway electrons and internal kink in a post-disruption plasma Yueqiang Liu, L. Li, C.C. Kim et al. - HYDRO-PACIAE, a hydrodynamic and transport hybrid model for ultra-relativistic heavy ion collisions Yu-Liang Yan, Yun Cheng, Dai-Mei Zhou et al. - LITHIUM DEPLETION IS A STRONG TEST OF CORE-ENVELOPE RECOUPLING Garrett Somers and Marc H. Pinsonneault - To cite this article: Tobias Thummerer et al 2021 J. Phys.: Conf. Ser. 2090 012155 View the article online for updates and enhancements. This content was downloaded from IP address 137.250.100.44 on 06/12/2021 at 08:17 This content was downloaded from IP address 137.250.100.44 on 06/12/2021 at 08:17 This content was downloaded from IP address 137.250.100.44 on 06/12/2021 at 08:17 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 IC-MSQUARE 2021 Journal of Physics: Conference Series 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 Hybrid modeling of the human cardiovascular system using NeuralFMUs Tobias Thummerer, Johannes Tintenherr and Lars Mikelsons Chair of Mechatronics, Augsburg University, Germany E-mail: {tobias.thummerer@informatik, johannes.tintenherr@student, lars.mikelsons@informatik}.uni-augsburg.de Abstract. Hybrid modeling, the combination of ﬁrst principle and machine learning models, is an emerging research ﬁeld that gathers more and more attention. Even if hybrid models produce formidable results for academic examples, there are still diﬀerent technical challenges that hinder the use of hybrid modeling in real-world applications. By presenting NeuralFMUs, the fusion of a Functional Mock-up Unit (FMU), a numerical ODE solver and an artiﬁcal neural network, we are paving the way for the use of a variety of ﬁrst principle models from diﬀerent modeling tools as parts of hybrid models. This contribution handles the hybrid modeling of a complex, real-world example: Starting with a simpliﬁed 1D-ﬂuid model of the human cardiovascular system (arterial side), the aim is to learn neglected physical eﬀects like arterial elasticity from data. We will show that the hybrid modeling process is more comfortable, needs less system knowledge and is therefore less error-prone compared to modeling solely based on ﬁrst principle. Further, the resulting hybrid model has improved in computation performance, compared to a pure ﬁrst principle white-box model, while still fulﬁlling the requirements regarding accuracy of the considered hemodynamic quantities. The use of the presented techniques is explained in a general manner and the considered use-case can serve as example for other modeling and simulation applications in and beyond the medical domain. Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. Published under licence by IOP Publishing Ltd 1 2 https://github.com/ThummeTo/FMIFlux.jl 1 https://github.com/ThummeTo/FMI.jl 1 https://github.com/ThummeTo/FMI.jl 2 https://github.com/ThummeTo/FMIFlux.jl 1. Introduction h l The structural integration of physical white-box models into artiﬁcal neural networks (ANNs) to retrieve a hybrid model is a growing research ﬁeld, see [1] or [2]. One milestone was the use of state-of-the-art numerical solvers for ordinary diﬀerential equations (ODEs) inside of ANNs instead of residual net structures to reproduce numerical integration and learn dynamic system behavior, like in [3]. The resulting combination of an ANN and an ODE solver - so called NeuralODEs - lead to improvements in model accuracy while also enhancing computation and memory cost. y Based on the idea of NeuralODEs we presented NeuralFMUs in [4], the structural integration of a ﬁrst-principle model in form of a FMU, a modern numerical solver like Tsit5 [5] or CVODE [6] and a feed-forward ANN. Further, we provided the open-source frameworks FMI.jl1 and FMIFlux.jl 2 to allow for the setup and training of NeuralFMUs just like a convenient ANN in the Julia programming language. In this contribution, the considered technologies shall be tested with a model of the human cardiovascular system (s. Sec. 2.1). 1 https://github.com/ThummeTo/FMI.jl 2 https://github.com/ThummeTo/FMIFlux.jl 1 IC-MSQUARE 2021 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 Journal of Physics: Conference Series One of the core challenges in hybrid modeling, the fusion of data driven and ﬁrst principle models, is to provide an eﬃcient training process for the resulting heterogeneous structure. For training of ANNs in general, the gradient of the loss function according to the net parameters is needed. This belongs also to hybrid models: The gradient must be determined along the ANN, the numerical solver and the model of the physical system. The diﬀerent jacobians must be determined by diﬀerent methods, because of the availability (or non-availability) of information and interfaces between the NeuralFMU components. For example, the mathematical operations inside of ANNs are known, therefore many methods to retrieve partial derivatives, like automatic diﬀerentiation, are possible to use. Models exported from modeling tools on the other hand need diﬀerent techniques, because they do not necessarily provide the symbolic structure of the model or often actively hide this information because it may contain sensible company knowledge. For FMUs, we suggest the use of the optional built-in function fmi2GetDirectionalDerivative or - if not available - sampling of additional simulation points and partial derivative approximation via ﬁnite (central) diﬀerences. Both approaches are implemented as part of FMIFlux.jl. Finally, if all jacobians over every component are retrieved, the needed diﬀerentiation chains must be deployed to the machine learning framework. For an overview and more detailed insight into the necessary technical steps, see [4]. In the following, some short style explanations about involved tools and standards are given. 1.1. Julia programming language The Julia Programming Language (hereinafter: Julia) is a dynamic typing language, being developed at the Massachusetts Institute of Technology since 2009 and ﬁrst published in 2012 [7]. Julia provides the ability to run code platform-independent and with performance benchmarks similar to native C-implementations. Regardless of the fact that Julia is relatively young (in terms of programming languages), the community is growing rapidly and many other research facilities joined the development of the language and language extensions. 1.2. Modelica The physical modeling was performed in the object-orientated modeling language Modelica (https://modelica.org/). Modelica allows acausal modeling, meaning causalization of the system of equations is handled automatically by the compiler at compile time. This allows for the building of large models, while keeping them human understandable by using an intuitive topology and sub-model enclosures. The two most common tools for graphical supported programming with Modelica are Dymola (by Dassault Syst`emes ®) and OMEdit (open-source), both tools allow for model export as FMU (s. next subsection). 2.1. Reference model f In [8], a model to simulate arterial pulse waves and their changes over aging of healthy patients is introduced. Together with the model itself, simulation results for arterial blood pressure, volume ﬂow, cross section (luminal area) and photoplethysmography (PPG) for 4,374 model patients of ages between 25 and 75 years are published. The model and simulation data was validated with in vivo data from diﬀerent sources. The ﬂuid simulation is implemented as laminar, incompressible, newtonian 1D-ﬂow, extended by a viscoelastic term to model diameter changes in arterial cross sections. The simulation setup and parameterization was done in Matlab (by MathWorks ®), the 1D-ﬂuid simulation was performed with Nektar1D (http://haemod.uk/nektar) using the hemodynamic model from [9]. The arterial system is closed by boundary conditions: The heart is abstracted as predetermined mass ﬂow into the Aorta, the vascular beds (the remaining boundary conditions of the ﬂuid simulation) where modeled as three-element windkessels. Please note, that only the arterial side of the cardiovascular system is modeled: Starting at the heart, the blood ﬂow passes multiple arteries and ﬁnally reaches the vascular beds. 0.6 0.8 [s] Arterial Network Arterial Head Legs Arm Ra Art ow P1D Q1D Q1D , P1D Aorta Inle Diame R 60/HR Sup. Temporal Brachial Digital Femoral Anterior Tibial Sup. Mid. Cerebal Aortic Root Thoracic Aorta Abdominal Aorta Iliac Bifurcation Common Iliac Figure 1: The simpliﬁed structure of the arterial side of the human cardiovascular system (ﬁgure adapted from [8]). The hemodynamics at the 10 selected artery segments (red) will be learned from data. 0.6 0.8 [s] Arterial Network Arterial Head Legs Arm Ra Art ow P1D Q1D Q1D , P1D Aorta Inle Diame R 60/HR Sup. Temporal Brachial Digital Femoral Anterior Tibial Sup. Mid. Cerebal Aortic Root Thoracic Aorta Abdominal Aorta Iliac Bifurcation Common Iliac Sup. Temporal Aortic Root Thoracic Aorta Abdominal Aorta Common Iliac Figure 1: The simpliﬁed structure of the arterial side of the human cardiovascular system (ﬁgure adapted from [8]). The hemodynamics at the 10 selected artery segments (red) will be learned from data. Figure 1: The simpliﬁed structure of the arterial side of the human cardiovascular system (ﬁgure adapted from [8]). The hemodynamics at the 10 selected artery segments (red) will be learned from data. The simulation database of the introduced contribution, further referenced as Pulse Wave Database (PWDB) (https://zenodo.org/record/3275625), serves as an ideal starting point for the presented machine learning task. 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 Journal of Physics: Conference Series Journal of Physics: Conference Series This paper further divides into three sections: The presentation of the cardiovascular system model, modeling assumptions and the extension to a hybrid model (s. Sec. 2), followed by training and validation of the hybrid model (s. Sec. 3) and ﬁnally a short conclusion, including current and future work (s. Sec. 4). 1.3. Functional Mock-up Interface p f Functional Mock-up Interface (FMI) allows for the simulation and parameterization of models outside of the original modeling environment in a standardized and platform-independent format. It is possible to generate standard-compliant models, so called FMUs, in two diﬀerent modes: model exchange (ME) and co-simulation (CS). ME-FMUs oﬀer an interface for the system dynamics, meaning the FMU computes a system state derivative for a given system state. In a subsequent step outside of the FMU, the next system state can be derived by numerically integrating the state derivative. CS-FMUs already include a numerical ODE solver, which allows for an even easier simulation. On the other hand, this inclusion prevents manipulation of the system dynamics before the numerical integration and consequently makes this mode unattractive for the aim of this contribution. 2 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 IC-MSQUARE 2021 IC-MSQUARE 2021 Journal of Physics: Conference Series • Thoracic Aorta, end of segment 18 • Abdominal Aorta, start of segment 39 • Iliac Bifurcation, end of segment 41 • left Superﬁcial Temporal Artery, end of segment 87 • left Superior Middle Cerebral Artery, end of segment 72 • left Brachial Artery, three quarters along segment 21 • left Digital Arteries, end of segment 112 • left Iliac Artery, half way of segment 44 • left Femoral Artery, half way of segment 46 • left Tibial Artery, end of segment 49 2.1. Reference model The PWDB includes ready-to-use data for 12 arterial segments, the following 10 segments of this set will be used for the later training process: 3 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 Journal of Physics: Conference Series Journal of Physics: Conference Series • The diameters of the arterial segments are assumed constant, meaning the arterial cross section does not vary dependent on the system state (pressure). • The diameters of the arterial segments are assumed constant, meaning the arterial cross section does not vary dependent on the system state (pressure). • Finally, the arteries are assumed cylindrical, so the inlet and outlet radius are the same size. To parameterize the pipes with parameter data from the reference model, which assumes a pipe with constant diameter change over length, the mean value of inlet and outlet diameter was used. • Finally, the arteries are assumed cylindrical, so the inlet and outlet radius are the same size. To parameterize the pipes with parameter data from the reference model, which assumes a pipe with constant diameter change over length, the mean value of inlet and outlet diameter was used. (a) Thoracic Aorta (#18) (b) Digital Arteries (#112) Figure 3: Deviation between the Modelica model (green/dashed) and the reference model (red/dotted) for the simulation of a single pressure pulse wave of subject #1, observed at two diﬀerent segments. (a) Thoracic Aorta (#18) (b) Digital Arteries (#112) (b) Digital Arteries (#112) (a) Thoracic Aorta (#18) Figure 3: Deviation between the Modelica model (green/dashed) and the reference model (red/dotted) for the simulation of a single pressure pulse wave of subject #1, observed at two diﬀerent segments. As to expect, these modeling simpliﬁcations lead to a signiﬁcantly diﬀerent simulation result compared to the reference model (s. Fig 3). However, the pulse wave characteristic is roughly visible despite the simpliﬁcations. The main reason for the diﬀerence between reference and simpliﬁed Modelica model is the lack of arterial dynamics: The system pressure inside the simpliﬁed model is the same at any location inside the arterial network, but varies over time with the heart pulse wave. On the other hand, the simpliﬁed model has a much better computational performance. The motivation for hybrid modeling is to retrieve model accuracy close to the reference model, but providing better computational performance at the same time. , p g p p To conclude, the only remaining system dynamics are the pressures over the 46 terminal windkessels, meaning the system state xwk (windkessels) can be uniquely described using a 46 entries state vector. 2.2. First principle model (Modelica) We start by building up a ﬁrst principle model in form of an intuitive, object-orientated Modelica model, using default components from the Modelica Standard Library (MSL) as far as possible. However, the presented procedure is not limited to Modelica models, of course. m m_flow Input 87 41 39 44 46 49 112 72 21 18 Figure 2: The physical model of the human cardiovascular system (arterial side), consisting of 116 static pipes (arterial segments, red/blue) and 46 three-element windkessels (vascular beds, green), in Dymola. From the set of pipes, 10 segments were picked for the machine learning task (red). The mass ﬂow from heart into the Aorta m ﬂow is given as model input. The model was build with components of the MSL. Figure 2: The physical model of the human cardiovascular system (arterial side), consisting of 116 static pipes (arterial segments, red/blue) and 46 three-element windkessels (vascular beds, green), in Dymola. From the set of pipes, 10 segments were picked for the machine learning task (red). The mass ﬂow from heart into the Aorta m ﬂow is given as model input. The model was build with components of the MSL. The ﬁrst principle model of the cardiovascular system is a simpliﬁcation of the reference model concerning the following aspects: The ﬁrst principle model of the cardiovascular system is a simpliﬁcation of the reference model concerning the following aspects: • The arteries are modeled as static pipes (MSL 3.2.3: Modelica.Fluid.Pipes.StaticPipe), meaning mass, momentum and energy balances are assumed steady-state. As a result, the pipe itself does not store any mass or energy, but physical conservation principles are satisﬁed. • The arteries are modeled as static pipes (MSL 3.2.3: Modelica.Fluid.Pipes.StaticPipe), meaning mass, momentum and energy balances are assumed steady-state. As a result, the pipe itself does not store any mass or energy, but physical conservation principles are satisﬁed. 4 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 IC-MSQUARE 2021 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 To retrieve a more exact model that describes a dynamic pressure drop over the arterial segments dependent on arterial cross section change by elasticity, more states are needed. To setup the Modelica model for machine learning, placeholder states xart (arteries) were inserted into the system at the considered artery locations, resulting in the concatenated system state vector x = xwk\|xart. This state space expansion occurs by simply inserting components into the object-orientated model at the desired locations. Two placeholders types are examined: The insertion of ﬂuid capacities with small capacitance (hereinafter: C- placeholders, one state) and the insertion of parallel circuits consisting of a capacitor and an inductance (hereinafter: LC-placeholders, two states). This reﬂects a straightforward model enhancement process: Adding dynamic placeholders to a model at the locations needed, while the correct physical term is learned from data. Note, that the use of a variety of placeholders is possible, like RLC-circuits for example. After this ﬁnal step, both models (one with C- and one with LC-placeholders) were exported as FMUs. 5 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 IC-MSQUARE 2021 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 IC-MSQUARE 2021 Journal of Physics: Conference Series 2.3. Hybrid model 2.3. Hybrid model 2 3 ybrid model Hybrid modeling is a technique picked often, if higher modeling accuracy is needed, but conventional ﬁrst principle modeling does not increase precision, is not economically practicable anymore or is simply not possible because of lack of system knowledge. On the other hand, almost any development process in medicine, mechanics or electronics generates data, that can be used to achieve better (more precise or faster computing) hybrid models. During hybrid modeling, we will neglect knowing the reference system and view the problem from a diﬀerent perspective: Starting with the simple Modelica model, we want to improve model accuracy without knowing the physical principles needed for this step. Therefore, the improvement process is done solely on basis of the reference model simulation data, not on physical system knowledge. This mirrors a typical use case: Model accuracy shall be improved, further ﬁrst principle modeling is not an option, but measurement data of a more accurate (or the real) system is available and can be used. ) To start with hybrid modeling, the FMU models exported from Dymola are imported into Julia using the library FMI.jl. With the library extension FMIFlux.jl, we can setup and train NeuralFMUs on basis of the Modelica model FMUs. For the considered use-case, two NeuralFMUs as in Fig. 4 (one with C- and one with LC-placeholders) with layer dimensions as in Tab. 1 were examined. Table 1: Topologies of the considered NeuralFMUs. Layer Type C-placeholders LC-placeholders Activation Inputs Outputs Inputs Outputs #1 state vector separation 56 10 \| 46 66 20 \| 46 none #2 data pre-processing 10 10 20 20 none #3 bias 10 10 20 20 none #4 data post-processing 10 10 20 20 none #5 state vector merging 10 \| 46 56 20 \| 46 66 none #6 FMU 56 56 66 66 none #7 derivative vector separation 56 10 \| 46 66 20 \| 46 none #8 data pre-processing 10 10 20 20 none #9 dense 10 30 20 30 tanh #10 dense 30 10 30 20 none #11 data post-processing 10 10 20 20 none #12 derivative vector merging 10 \| 46 56 20 \| 46 66 none 3. Training & Validation During training, three pulse wave cycles for patient #1 are simulated (2.466 s, varies between subjects because of diﬀerent heart rates). 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 Journal of Physics: Conference Series 2.3. Hybrid model The ﬁrst pulse wave is ignored in the training process and allows the system to retrieve a stationary (periodically repeating) state. The training takes place on two cycles to promote learning a time-invariant system behavior, meaning a periodically repeating input (heart) should generate a periodically repeating output (arterial pressure curves) with same period length. Even if the PWDB includes data observed at 500 Hz, the training process is sampled down to 40 Hz because of training performance optimization. To conclude, training takes place on only two pulse waves of one patient, resulting in 66 data points `a 10 Table 1: Topologies of the considered NeuralFMUs. Layer Type C-placeholders LC-placeholders Activation Inputs Outputs Inputs Outputs #1 state vector separation 56 10 \| 46 66 20 \| 46 none #2 data pre-processing 10 10 20 20 none #3 bias 10 10 20 20 none #4 data post-processing 10 10 20 20 none #5 state vector merging 10 \| 46 56 20 \| 46 66 none #6 FMU 56 56 66 66 none #7 derivative vector separation 56 10 \| 46 66 20 \| 46 none #8 data pre-processing 10 10 20 20 none #9 dense 10 30 20 30 tanh #10 dense 30 10 30 20 none #11 data post-processing 10 10 20 20 none #12 derivative vector merging 10 \| 46 56 20 \| 46 66 none Table 1: Topologies of the considered NeuralFMUs. 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 arterial pressure values. Note, that this is a very small training base in the ﬁeld of machine learning. arterial pressure values. Note, that this is a very small training base in the ﬁeld of machine learning. The training results, comparing the original FMU pressure (from the Modelica model), the improved NeuralFMU pressures and the target pressure (reference system data) can be found in Fig. 5. As loss function, a simple mean-squared-error with increasing time horizon between the reference system and NeuralFMU arterial blood pressures was deployed. The loss function rates the deviation only for a random sub-set of three artery segments, while changing the sub-set every training cycle (Batching). The training itself was controlled via parameter freezing: Until reaching a signiﬁcant small error (loss), parameters of the derivative ANN (s. Fig. 4: bottom) were locked to force compensation of state oﬀsets by the state ANN (s. Fig. 4: top) and prevent early corruption of the system dynamics. After reaching a loss threshold, all net parameters were unlocked and trained in parallel. For the C-placeholders, the static pressure wave from the FMU model is vertically scaled and shifted to ﬁt the reference system data as good as possible. The LC-placeholders provide a better ﬁt, because the LC-circuit additionally allows for phase-shifting the original pulse waves. The NeuralFMUs produce less data (only 10 dynamic locations instead of 116) and the output result does only approximate the reference model output, because of the relative small ANN layout and a lack of states to interfere with the system. On the other hand, there is a signiﬁcant gain in performance, which motivates the use of NeuralFMUs for use-cases, that focus on less measurement locations and/or more on fast computations than maximum precision, like embedded hardware or wearable computing. The reference model has a simulation time3 of ≈756 s for 10 pulse waves, resulting in an average ≈75.6 s per pulse waves. The NeuralFMUs on the other hand, have a simulation time4 of only ≈0.20 s per 10 pulse waves, meaning an average of ≈0.02 s per pulse wave5. Even if both simulations were performed on diﬀerent systems, a signiﬁcant performance gain by a factor of ≈3750 is clearly visible. This scope can be used to improve accuracy through deeper and/or wider network topologies or the injection of more and/or more complex state placeholders if necessary. 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 Testing the NeuralFMUs with unobserved data from patient #2 leads to similar ﬁtting pressure curves (s. Fig. 6), however the remaining average error is slightly bigger. Even if the derivative ANN (bottom) did learn a simpliﬁed abstraction of the pressure dynamics, the simpliﬁcation itself was understood in a generalized manner. Please note at this point, that the heart rate diﬀers between the two patients (≈73 bpm for #1 and ≈77 bpm for #2). The state ANN (top) was only able to learn static oﬀsets between the states of the reference system and the ﬁrst principle model parameterized for patient #1, but the oﬀsets in data diﬀer for patient #2. It is visible in the plots, that a small vertical oﬀset between training and testing results exists. These oﬀsets could be reduced by training the state ANN on data of patient #2. A more detailed result validation is omitted at this point, but will be part of a pursuing contribution. 3 AMD RyzenTM 9 3900X on Ubuntu 20.04.2 LTS 4 Intel® CoreTM i7-8565U on Windows 10 Enterprise 20H2 5 Even if the number of states and mathematical complexity of the FMUs diﬀer between C- and LC-placeholders, the diﬀerence in simulation time of the resulting NeuralFMUs was marginal. 6 https://github.com/ThummeTo/FMI.jl 3 AMD RyzenTM 9 3900X on Ubuntu 20.04.2 LTS 4 Intel® CoreTM i7-8565U on Windows 10 Enterprise 20H2 3. Training & Validation 3. Training & Validation 3. Training & Validation During training, three pulse wave cycles for patient #1 are simulated (2.466 s, varies between subjects because of diﬀerent heart rates). The ﬁrst pulse wave is ignored in the training process and allows the system to retrieve a stationary (periodically repeating) state. The training takes place on two cycles to promote learning a time-invariant system behavior, meaning a periodically repeating input (heart) should generate a periodically repeating output (arterial pressure curves) with same period length. Even if the PWDB includes data observed at 500 Hz, the training process is sampled down to 40 Hz because of training performance optimization. To conclude, training takes place on only two pulse waves of one patient, resulting in 66 data points `a 10 g During training, three pulse wave cycles for patient #1 are simulated (2.466 s, varies between subjects because of diﬀerent heart rates). The ﬁrst pulse wave is ignored in the training process and allows the system to retrieve a stationary (periodically repeating) state. The training takes place on two cycles to promote learning a time-invariant system behavior, meaning a periodically repeating input (heart) should generate a periodically repeating output (arterial pressure curves) with same period length. Even if the PWDB includes data observed at 500 Hz, the training process is sampled down to 40 Hz because of training performance optimization. To conclude, training takes place on only two pulse waves of one patient, resulting in 66 data points `a 10 6 6 IC-MSQUARE 2021 Journal of Physics: Conference Series IC-MSQUARE 2021 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 6 https://github.com/ThummeTo/FMI.jl Acknowledgments This work has been partly supported by the ITEA 3 cluster programme for the project UPSIM - Unleash Potentials in Simulation. 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 the library extension FMIFlux.jl 7. Finally, the simulation data was compared to the target values from the PWDB. The resulting NeuralFMU produces, dependent on the requirements of the underlying application, suﬃciently accurate results. Model precision can be further improved by using more or other circuits for the state placeholders or wider and deeper ANN structures. Further, the use of more than one patient during training will improve the NeuralFMU prediction quality on unknown patients. In terms of computational eﬃciency, a signiﬁcant performance gain is visible for the considered example, making the presented approach interesting for performance critical applications. NeuralFMUs have further advantages, e.g. even though they might include black-box models, they are fully diﬀerentiable. This allows for the use of other machine learning techniques or eﬃcient model examination methods, like gradient based algorithms. NeuralFMUs open up to many new interesting use-cases, but bring up many new challenges, too. For example, most physical systems are stable by default, meaning they converge against a physical equilibrium if not disturbed. By manipulating the system dynamics via an ANN, this natural stability is not guaranteed anymore, which may result in a destabilized training process. Best practices for network initialization and stabilization methods during training must be con- sidered and will be part of a pursuing contribution. The corresponding sources for the presented example will be published soon as part of the FMIFlux.jl library repository. 7 https://github.com/ThummeTo/FMIFlux.jl 4. Conclusion & Future Work We highlighted a workﬂow to improve a ﬁrst principle model (at the example of a simple, object-orientated model) using a NeuralFMU without knowing the underlying physical equations necessary. The enhancements were learned solely based on a small set of data, in the presented example generated by a more accurate reference system. Starting with the FMI model export from the modeling tool, the resulting FMU was imported into the Julia programming language using the open-source library FMI.jl6. Inside Julia, a NeuralFMU was set up and trained, using 7 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 IC-MSQUARE 2021 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 IC MSQUARE 2021 Journal of Physics: Conference Series References [1] Willard J, Jia X, Xu S, Steinbach M and Kumar V 2020 Integrating physics-based modeling with machine learning: A survey [1] Willard J, Jia X, Xu S, Steinbach M and Kumar V 2020 Integrating physics-based modeling with machine learning: A survey [2] Rai R and Sahu C K 2020 Driven by Data or Derived Through Physics? A Review of Hybrid Physics Guided Machine Learning Techniques With Cyber-Physical System (CPS) Focus vol 8 pp 71050–71073 [2] Rai R and Sahu C K 2020 Driven by Data or Derived Through Physics? A Review of Hybrid Physics Guided Machine Learning Techniques With Cyber-Physical System (CPS) Focus vol 8 pp 71050–71073 [3] Chen T Q, Rubanova Y, Bettencourt J and Duvenaud D 2018 N 1806.07366) URL http://arxiv.org/abs/1806.07366 ( p , p ) [5] Tsitouras C 2011 Runge–Kutta pairs of order 5(4) satisfying only the ﬁrst column simplifying assumption vol 62 pp 770–775 URL https://www.sciencedirect.com/science/article/pii/S0898122111004706 [5] Tsitouras C 2011 Runge–Kutta pairs of order 5(4) satisfying only the ﬁrst column simplifying assumption vol 62 pp 770–775 URL https://www.sciencedirect.com/science/article/pii/S0898122111004706 [6] Hindmarsh A C, Serban R, Balos C J, Gardner D J, Reynolds D R and Woodward C S 2021 User documentation for cvode v5.7.0 (sundials v5.7.0) Tech. rep. URL https://computing.llnl.gov/sites/ default/files/cv_guide-5.7.0.pdf [6] Hindmarsh A C, Serban R, Balos C J, Gardner D J, Reynolds D R and Woodward C S 2021 User documentation for cvode v5.7.0 (sundials v5.7.0) Tech. rep. References URL https://computing.llnl.gov/sites/ default/files/cv_guide-5.7.0.pdf g p [7] Bezanson J, Karpinsky S, Shah V B and Edelman A 2012 Julia: A fast dynamic language for technical computing (Preprint 1209.5145) URL http://arxiv.org/abs/1209.5145 [8] Charlton P H, Mariscal Harana J, Vennin S, Li Y, Chowienczyk P and Alastruey J 2019 Modeling arterial pulse waves in healthy aging: a database for in silico evaluation of hemodynamics and pulse wave indexes vol 317 pp H1062–H1085 pMID: 31442381 (Preprint https://doi.org/10.1152/ajpheart.00218.2019) URL https://doi.org/10.1152/ajpheart.00218.2019 [9] Alastruey J, Parker K H and Sherwin S J 2012 Arterial pulse wave haemodynamics (Virtual PiE Led Group) pp 401–443 ISBN 9781855981331 7 https://github.com/ThummeTo/FMIFlux.jl 8 8 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 2090 (2021) 012155 IOP Publishing doi:10.1088/1742-6596/2090/1/012155 IC-MSQUARE 2021 Journal of Physics: Conference Series 𝒙 STATE ANN ME-FMU ሶ𝒙𝑓𝑚𝑢 𝑢𝑓𝑚𝑢 𝑡 ෝ𝒙𝑓𝑚𝑢 DATA PRE-PROCESSING DATA POST-PROCESSING STATE VECTOR SEPARATION STATE VECTOR MERGING 𝒙𝑎𝑟𝑡 ෝ𝒙𝑎𝑟𝑡 DERIVATIVE VECTOR SEPARATION 𝒙𝑤𝑘 DATA PRE-PROCESSING ሶ𝒙𝑎𝑟𝑡 𝒙 STATE ANN ME-FMU ሶ𝒙𝑓𝑚𝑢 𝑢𝑓𝑚𝑢 𝑡 ෝ𝒙𝑓𝑚𝑢 DERIVATIVE ANN DATA PRE-PROCESSING DATA POST-PROCESSING STATE VECTOR SEPARATION STATE VECTOR MERGING 𝒙𝑎𝑟𝑡 ෝ𝒙𝑎𝑟𝑡 ሶ𝒙𝑤𝑘 DERIVATIVE VECTOR SEPARATION ෡ሶ𝒙𝑎𝑟𝑡 NUMERICAL ODE SOLVER න DERIVATIVE VECTOR MERGING ෡ሶ𝒙 𝒙𝑤𝑘 DATA PRE-PROCESSING DATA POST-PROCESSING ሶ𝒙𝑎𝑟𝑡 𝒙(𝑡+ ℎ) ෝ𝒙𝑡+ ℎ Figure 4: The hybrid model structure. From top to bottom: Because not all states of the NeuralFMU shall be part of the training process, the state vector x is separated into the two sub-vectors xwk (the windkessels’ states) and xart (the arterial pressure states). Before being fed into the upper ANN, the arterial state values xart are pre-processed to ensure being inside an ANN-compliant range and are transformed back after the ANN pass. The state ANN at the top of the topology is able to learn and compensate state dependent modeling failures, like static oﬀsets in measurement data in this special case. The new arterial state estimate ˆxart is merged together with the original windkessel states xwk into the modiﬁed system state vector ˆxfmu, that is further passed, together with the system input ufmu and the current simulation time t, to the ME-FMU. Inside the FMU, the current state derivatives ˙xfmu are computed based on the given system state ˆxfmu. DERIVATIVE ANN ሶ𝒙𝑤𝑘 ෡ሶ𝒙𝑎𝑟𝑡 NUMERICAL ODE SOLVER න DERIVATIVE VECTOR MERGING ෡ሶ𝒙 DATA PRE-PROCESSING DATA POST-PROCESSING ሶ𝒙𝑎𝑟𝑡 𝒙(𝑡+ ℎ) ෝ𝒙𝑡+ ℎ f As for the system state, the system state derivative vector is separated into the arterial ˙xart and windkessel state derivative vector ˙xwk. References The arterial pressure dynamics are pre-processed, fed to the ANN, post-processed and the resulting ˆ˙xart is merged back together with ˙xwk into the system state derivative estimate ˆ˙x. The derivative ANN is therefore able to manipulate the system dynamics to retrieve a diﬀerent system behavior. In the considered use case, the pressure drop over the arterial segments can be learned based on data. Because the ANN is learning on basis of state derivatives, generalized learning of the correct physical laws is promoted. ˆ ෡ሶ𝒙 Finally, the system dynamics estimate ˆ˙x is numerically integrated with step size h into the next system state estimate ˆx(t+h) or the next system state x(t+h) respectively. 9 IC-MSQUARE 2021 IOP Publishing 2090 (2021) 012155 g doi:10.1088/1742-6596/2090/1/012155 Journal of Physics: Conference Series (a) Thoracic Aorta (#18) (b) Abdominal Aorta (#39) (c) Iliac Bifurcation (#41) (d) Sup. Temp. Artery (#87) (e) Mid. Cerebral Artery (#72) (f) Brachial Artery (#21) (g) Digital Arteries (#112) (h) Com. Iliac Artery (#44) (i) Femoral Artery (#46) (j) Ant. Tibial Artery (#49) Figure 5: Training results of the 10 considered arterial segments, based on data of subject #1 of the PWDB from [8]. The horizontal axis labels the simulation time t in seconds, the vertical axis the arterial pressure in Pascals. The plots show the pressure of the Modelica model (black/dotted, clipped), the target reference system pressure (red/dashed), the pressure learned by the NeuralFMU with C-placeholders (blue/solid) and with LC-placeholders (green/solid). (a) Thoracic Aorta (#18) (b) Abdominal Aorta (#39) (c) Iliac Bifurcation (#41) (d) Sup. Temp. Artery (#87) (e) Mid. Cerebral Artery (#72) (f) Brachial Artery (#21) (f) Brachial Artery (#21) (e) Mid. Cerebral Artery (#72) (h) Com. Iliac Artery (#44) (g) Digital Arteries (#112) (i) Femoral Artery (#46) (g) Digital Arteries (#112) (j) Ant. Tibial Artery (#49) (j) Ant. Tibial Artery (#49) Figure 5: Training results of the 10 considered arterial segments, based on data of subject #1 of the PWDB from [8]. The horizontal axis labels the simulation time t in seconds, the vertical axis the arterial pressure in Pascals. The plots show the pressure of the Modelica model (black/dotted, clipped), the target reference system pressure (red/dashed), the pressure learned by the NeuralFMU with C-placeholders (blue/solid) and with LC-placeholders (green/solid). References 10 IC-MSQUARE 2021 IOP Publishing 2090 (2021) 012155 g doi:10.1088/1742-6596/2090/1/012155 Journal of Physics: Conference Series Journal of Physics: Conference Series 2090 (2021) 012155 doi:10.1088/1742-6596/2090/1/012155 (a) Thoracic Aorta (#18) (b) Abdominal Aorta (#39) (c) Iliac Bifurcation (#41) (d) Sup. Temp. Artery (#87) (e) Mid. Cerebral Artery (#72) (f) Brachial Artery (#21) (g) Digital Arteries (#112) (h) Com. Iliac Artery (#44) (i) Femoral Artery (#46) (j) Ant. Tibial Artery (#49) Figure 6: Testing results on the 10 considered arterial segments. Tests where performed against data of (unknown) subject #2 of the PWDB from [8]. The horizontal axis labels the simulation time t in seconds, the vertical axis the arterial pressure in Pascals. The plots show the pressure of the Modelica model (black/dotted, clipped), the target reference system pressure (red/dashed), the pressure learned by the NeuralFMU with C-placeholders (blue/solid) and with LC-placeholders (green/solid). (a) Thoracic Aorta (#18) (b) Abdominal Aorta (#39) (c) Iliac Bifurcation (#41) ( ) ( ) (d) Sup. Temp. Artery (#87) (e) Mid. Cerebral Artery (#72) (f) Brachial Artery (#21) (f) Brachial Artery (#21) (g) Digital Arteries (#112) (i) Femoral Artery (#46) (h) Com. Iliac Artery (#44) (j) Ant. Tibial Artery (#49) (j) Ant. Tibial Artery (#49) Figure 6: Testing results on the 10 considered arterial segments. Tests where performed against data of (unknown) subject #2 of the PWDB from [8]. The horizontal axis labels the simulation time t in seconds, the vertical axis the arterial pressure in Pascals. The plots show the pressure of the Modelica model (black/dotted, clipped), the target reference system pressure (red/dashed), the pressure learned by the NeuralFMU with C-placeholders (blue/solid) and with LC-placeholders (green/solid). 11
https://openalex.org/W3097508014	https://biblio.ugent.be/publication/8702036/file/8702816.pdf	English	null	An Unbiased Approach to Mapping the Signaling Network of the Pseudorabies Virus US3 Protein	Pathogens	2,020	cc-by	9,869	Received: 15 October 2020; Accepted: 3 November 2020; Published: 5 November 2020 Abstract: The US3 serine/threonine protein kinase is conserved among the alphaherpesvirus family and represents an important virulence factor. US3 plays a role in viral nuclear egress, induces dramatic alterations of the cytoskeleton, represses apoptosis, enhances gene expression and modulates the immune response. Although several substrates of US3 have been identiﬁed, an unbiased screen to identify US3 phosphorylation targets has not yet been described. Here, we perform a shotgun and phosphoproteomics analysis of cells expressing the US3 protein of pseudorabies virus (PRV) to identify US3 phosphorylation targets in an unbiased way. We identiﬁed several cellular proteins that are diﬀerentially phosphorylated upon US3 expression and validated the phosphorylation of lamin A/C at serine 404, both in US3-transfected and PRV-infected cells. These results provide new insights into the signaling network of the US3 protein kinase and may serve as a basis for future research into the role of the US3 protein in the viral replication cycle. ywords: US3; kinase; PRV;pseudorabiesvirus; alphaherpesvirus; phosphoproteome; massspectrometry Pathogens 2020, 9, 916; doi:10.3390/pathogens9110916 pathogens pathogens pathogens An Unbiased Approach to Mapping the Signaling Network of the Pseudorabies Virus US3 Protein Robert J. J. Jansens 1, Sandra Marmiroli 2 and Herman W. Favoreel 1,* 1 Department of Virology, Parasitology and Immunology, Faculty of Veterinary Medicine, Ghen 9820 Merelbeke, Belgium; robert.jansens@ugent.be 2 Cellular Signaling Laboratory, Department of Surgery, Medicine, Dentistry, and Morphology, University of Modena & Reggio Emilia, 41121 Modena, Italy; sandra.marmiroli@unimore.it * Correspondence: Herman.Favoreel@UGent.be 1 Department of Virology, Parasitology and Immunology, Faculty of Veterinary Medicine, Ghent 9820 Merelbeke, Belgium; robert.jansens@ugent.be 2 Cellular Signaling Laboratory, Department of Surgery, Medicine, Dentistry, and Morphology, University of Modena & Reggio Emilia, 41121 Modena, Italy; sandra.marmiroli@unimore.it * Correspondence: Herman.Favoreel@UGent.be Received: 15 October 2020; Accepted: 3 November 2020; Published: 5 November 2020 www.mdpi.com/journal/pathogens 1. Introduction Herpesviruses are among the most successful pathogens worldwide, establishing lifelong latent infections in their natural host. Their success correlates with the tight control that herpesviruses exert on diﬀerent cellular pathways, sometimes by mimicking cellular proteins. One strategy to gain control over the host cell employed by herpesviruses is the expression of viral protein kinases that phosphorylate a wide range of both viral and cellular proteins [1,2]. All herpesviruses encode at least one protein kinase, while alphaherpesviruses encode two. The herpes simplex virus (HSV) and pseudorabies virus (PRV) homologues of the protein kinase conserved in all herpesviruses are called UL13. The conserved alphaherpesvirus protein kinase homologues in HSV and PRV are called US3. Although neither of these viral kinases are required for replication in cell culture, mutant viruses deﬁcient in either kinase are severely attenuated in in vivo infection models [3–6]. The US3 protein is a multifunctional serine/threonine protein kinase. US3 expression modulates a wide array of cellular processes, including virus nuclear egress, inhibition of apoptosis, reorganization of the cytoskeleton and several immune modulators [2,7,8]. Some of these eﬀects have been investigated at a mechanistic level, and several phosphorylation targets of US3 have been identiﬁed [2,9]. The US3 protein of HSV1 has been shown to directly phosphorylate Lamin A/C, IRF3, TSC2, Beclin1, Bad, PKA and KIF3A, in addition to several viral proteins [10–17]. The US3 protein of PRV, on the other hand, has been shown to directly phosphorylate PAK1 and PAK2 [18]. These phosphorylated proteins were identiﬁed in hypothesis-driven experiments and likely do not encompass the entire spectrum of US3 phosphorylation targets. In addition to directly phosphorylating diﬀerent targets, US3 may Pathogens 2020, 9, 916; doi:10.3390/pathogens9110916 www.mdpi.com/journal/pathogens 2 of 13 Pathogens 2020, 9, 916 also modulate the activity of several cellular kinases and/or phosphatases, an aspect missed when identifying direct phosphorylation targets [19,20]. A better understanding of the signaling networks aﬀected by the US3 protein would enhance our understanding of the functions of the US3 protein, and the underlying mechanisms. Here, we used an unbiased mass spectrometry (MS) approach to identify pathways modulated by the PRV US3 protein. Mass spectrometry-based proteomics has previously been used in several studies to gain a better overview of diﬀerentially regulated or phosphorylated proteins during alphaherpesvirus infections [21–24]. By using transfected cells, we focused on proteins diﬀerentially phosphorylated by expression of US3 without the complicating background of a viral infection. 1. Introduction Indeed, viral infection could confound ﬁndings, among others due to the expression of another viral kinase, UL13, and other viral proteins aﬀecting the phosphorylation state of cellular proteins [5]. This approach allows us to generate a wide overview of cellular pathways aﬀected by US3. Several diﬀerentially phosphorylated proteins involved in pathways known to be aﬀected by US3 were identiﬁed. In addition, we were able to identify and conﬁrm a speciﬁc serine residue in lamin A/C that is phosphorylated via US3, leading to new insights into lamin phosphorylation in herpesvirus infections [11]. Surprisingly, we also found several proteins involved in RNA processing to be diﬀerentially phosphorylated, hinting to a hitherto unknown function of US3 in mRNA processing. 2.1. The Phosphoproteome of US3-Transfected Cells In order to discover phosphorylation targets of the PRV US3 protein, we transfected swine testicle (ST) cells with a plasmid encoding the full length US3 protein, or a kinase inactive version of this plasmid (US3 KD) [8]. Cells were lysed at 16 hpt because the activity of US3 (anti-apoptosis and intercellular transport by TNTs) causes a higher apparent transfection eﬃciency compared to kinase negative US3 at later time points [25,26]. The average transfection eﬃciencies of US3 and US3 KD at 16 hpt were 27% and 26%, respectively (Figure S1). The transfected cell lysates were used both for shotgun proteomics, determining total protein levels, and for phosphoproteomics. Since PRV US3 functions as a kinase, no major changes in total protein expression were expected. In line with this, only three proteins were found to be differentially regulated in the shotgun proteomics experiment (Figure 1A). Even though the transfection efficiencies were similar, the wild type, active version of PRV US3 showed higher expression compared to the inactive form of US3, which is in line with previous observations in our lab that kinase-active US3 appears to be more stable than its kinase negative version (unpublished results). Additionally, a component of the mitoribosome and a DNA topoimerase were found to be downregulated in cells expressing wild type US3 (Table S1). The phosphoproteome experiment yielded a larger amount of diﬀerentially phosphorylated peptides (Figure 1B). In total, 14 cellular peptides were signiﬁcantly dephosphorylated in wild type US3 transfected cells compared to kinase inactive US3 transfected cells (Tables 1 and S1), while 64 cellular peptides showed signiﬁcantly increased phosphorylation upon wild type US3 transfection compared to kinase inactive US3 transfection (Tables 2 and S1). The list of signiﬁcantly phosphorylated sites includes four sites in lamin A and C, a component of the nuclear lamina previously shown to be phosphorylated by the HSV1 homologue of US3 [11]. However, several other proteins known to be diﬀerentially phosphorylated by HSV or PRV US3, including group A PAKs, Bad, coﬁlin and RhoA, were not identiﬁed [8,18,20]. Our results do not indicate that these phosphorylation sites were not diﬀerentially phosphorylated, merely that these sites were not identiﬁed using our MS approach. This shows the value of using alternative methodologies to identify phosphorylation sites and indicates that the list of phosphorylated sites upon US3 expression is not exhaustive. 3 of 13 Pathogens 2020, 9, 916 Figure 1. Proteome and phosphoproteome of US3-transfected cells. 2.1. The Phosphoproteome of US3-Transfected Cells However, several other proteins known to be differentially phosphorylated by HSV or PRV US3, including group A PAKs, Bad, cofilin and RhoA, were not identified [8,18,20]. Our results do not indicate that these phosphorylation sites were not differentially phosphorylated, merely that these sites were not identified using our MS approach. This shows the value of using alternative methodologies to identify phosphorylation sites and indicates that the list of phosphorylated sites upon US3 expression is not exhaustive. Table 1. Top five dephosphorylated proteins in wild type US3 transfected cells compared to kinase inactive US3 transfected cells. Gene Name Protein Name Site Log2(KD/WT) -Log2(p-Value) CDS2 Phosphatidate cytidylyltransferase S33 6.38 4.56 BCKDHA 2-oxoisovalerate dehydrogenase alpha S313 5.83 3.48 BCKDHA 2-oxoisovalerate dehydrogenase alpha S303 5.54 3.25 PPP6R1 S/T-protein phosphatase 6 subunit 1 S531 4.39 2.92 WDR20 WD repeat-containing protein 20 S348 4.16 3.83 Table 1. Top ﬁve dephosphorylated proteins in wild type US3 transfected cells compared to kinase inactive US3 transfected cells. Gene Name Protein Name Site Log2(KD/WT) -Log2(p-Value) CDS2 Phosphatidate cytidylyltransferase S33 6.38 4.56 BCKDHA 2-oxoisovalerate dehydrogenase alpha S313 5.83 3.48 BCKDHA 2-oxoisovalerate dehydrogenase alpha S303 5.54 3.25 PPP6R1 S/T-protein phosphatase 6 subunit 1 S531 4.39 2.92 WDR20 WD repeat-containing protein 20 S348 4.16 3.83 Table 2. Top 10 Phosphorylated proteins in US3 transfected cells compared to kinase negative US3 transfected cells. Gene Name Protein Name Site Log2(KD/WT) -Log2(p-Value) TSSC1 EARP and GARP complex interacting protein 1 S320 −6.50 4.65 RAB11FIP5 Rab11 family-interacting protein 5 T162 −6.24 3.53 RAB11FIP5 Rab11 family-interacting protein 5 S164 −6.09 3.98 TOMM70 Mitochondrial import receptor subunit TOM70 S97 −6.01 4.52 TESK2 Dual speciﬁcity testis-speciﬁc protein kinase 2 S8 −5.85 6.37 SZRD1 SUZ domain-containing protein 1 S17 −5.48 2.28 DDX17 Probable ATP-dependent RNA helicase DDX17 S575 −5.23 3.85 LMNA Prelamin-A/C S12 −5.21 1.56 SZRD1 SUZ domain-containing protein 1 S19 −5.09 5.62 PALMD Palmdelphin T255 −4.78 3.29 identified using our MS approach. This shows the value of using alternative methodologies to identify phosphorylation sites and indicates that the list of phosphorylated sites upon US3 expression is not exhaustive. Table 1. Top five dephosphorylated proteins in wild type US3 transfected cells compared to kinase inactive US3 transfected cells. 2.1. The Phosphoproteome of US3-Transfected Cells (A) Volcano plot of differential expression of proteins between ST cells expressing active US3 or kinase inactive US3. Proteins upregulated in wild type US3-expressing cells are shown in red, proteins downregulated in wild type US3-expressing cells are shown in green. (B) Volcano plot of differential phosphorylation of proteins between ST cells expressing active US3 or kinase negative US3. Proteins that show increased phosphorylation in wild type US3-expressing cells are shown in red, proteins that show reduced phosphorylation in wild type US3-expressing cells are shown in green. Figure 1. Proteome and phosphoproteome of US3-transfected cells. (A) Volcano plot of differential expression of proteins between ST cells expressing active US3 or kinase inactive US3. Proteins upregulated in wild type US3-expressing cells are shown in red, proteins downregulated in wild type US3-expressing cells are shown in green. (B) Volcano plot of differential phosphorylation of proteins between ST cells expressing active US3 or kinase negative US3. Proteins that show increased phosphorylation in wild type US3-expressing cells are shown in red, proteins that show reduced phosphorylation in wild type US3-expressing cells are shown in green. Figure 1. Proteome and phosphoproteome of US3-transfected cells. (A) Volcano plot of differential expression of proteins between ST cells expressing active US3 or kinase inactive US3. Proteins upregulated in wild type US3-expressing cells are shown in red, proteins downregulated in wild type US3-expressing cells are shown in green. (B) Volcano plot of differential phosphorylation of proteins between ST cells expressing active US3 or kinase negative US3. Proteins that show increased phosphorylation in wild type US3-expressing cells are shown in red, proteins that show reduced phosphorylation in wild type US3-expressing cells are shown in green. Figure 1. Proteome and phosphoproteome of US3-transfected cells. (A) Volcano plot of differential expression of proteins between ST cells expressing active US3 or kinase inactive US3. Proteins upregulated in wild type US3-expressing cells are shown in red, proteins downregulated in wild type US3-expressing cells are shown in green. (B) Volcano plot of differential phosphorylation of proteins between ST cells expressing active US3 or kinase negative US3. Proteins that show increased phosphorylation in wild type US3-expressing cells are shown in red, proteins that show reduced phosphorylation in wild type US3-expressing cells are shown in green. Figure 1. Proteome and phosphoproteome of US3-transfected cells. (A) Volcano plot of differential Figure 1. Proteome and phosphoproteome of US3-transfected cells. 2.1. The Phosphoproteome of US3-Transfected Cells (A) Volcano plot of differential The phosphoproteome experiment yielded a larger amount of differentially phosphorylated eptides (Figure 1B). In total, 14 cellular peptides were significantly dephosphorylated in wild type S3 transfected cells compared to kinase inactive US3 transfected cells (Tables 1 and S1), while 64 Table 1. Top ﬁve dephosphorylated proteins in wild type US3 transfected cells compared to kinase inactive US3 transfected cells. peptides (Figure 1B). In total, 14 cellular peptides were significantly dephosphorylated in wild type US3 transfected cells compared to kinase inactive US3 transfected cells (Tables 1 and S1), while 64 cellular peptides showed significantly increased phosphorylation upon wild type US3 transfection compared to kinase inactive US3 transfection (Tables 2 and S1). The list of significantly phosphorylated sites includes four sites in lamin A and C, a component of the nuclear lamina previously shown to be phosphorylated by the HSV1 homologue of US3 [11]. However, several other proteins known to be differentially phosphorylated by HSV or PRV US3, including group A PAKs, Bad, cofilin and RhoA, were not identified [8,18,20]. Our results do not indicate that these phosphorylation sites were not differentially phosphorylated, merely that these sites were not identified using our MS approach. This shows the value of using alternative methodologies to d f h h l d d h h l f h h l d US Table 1. Top ﬁve dephosphorylated proteins in wild type US3 transfected cells compared to kinase inactive US3 transfected cells. Gene Name Protein Name Site Log2(KD/WT) -Log2(p-Value) CDS2 Phosphatidate cytidylyltransferase S33 6.38 4.56 BCKDHA 2-oxoisovalerate dehydrogenase alpha S313 5.83 3.48 BCKDHA 2-oxoisovalerate dehydrogenase alpha S303 5.54 3.25 PPP6R1 S/T-protein phosphatase 6 subunit 1 S531 4.39 2.92 WDR20 WD repeat-containing protein 20 S348 4.16 3.83 Table 2. Top 10 Phosphorylated proteins in US3 transfected cells compared to kinase negative US3 peptides (Figure 1B). In total, 14 cellular peptides were significantly dephosphorylated in wild type US3 transfected cells compared to kinase inactive US3 transfected cells (Tables 1 and S1), while 64 cellular peptides showed significantly increased phosphorylation upon wild type US3 transfection compared to kinase inactive US3 transfection (Tables 2 and S1). The list of significantly phosphorylated sites includes four sites in lamin A and C, a component of the nuclear lamina previously shown to be phosphorylated by the HSV1 homologue of US3 [11]. 2.2. Validation of Identiﬁed S404 Phosphosite in Lamin A/C 2.2. Validation of Identiﬁed S404 Phosphosite in Lamin A/C The majority of all sites identiﬁed as diﬀerentially phosphorylated have not been functionally described before, and consequently do not have phospho-speciﬁc tools available. One exception is the S404 site of lamin A/C. This phosphosite shows an approximately 10-fold increase in phosphorylation in US3-transfected cells compared to cells transfected with kinase-inactive US3 (Table S1). Lamin A/C S404 has been previously described as a phosphorylation target of Akt and S404 phosphorylation was shown to target lamin A/C for degradation [27–29]. We used a S404 phospho-speciﬁc antibody to validate the mass spectrometry results. Transfection of active US3 indeed leads to an increase in p-S404 lamin A/C signal in Western blotting assays (Figure 2A). To investigate whether these results bear relevance in an infection context, Western blotting was performed in ST cells that were mock-infected or infected with wild type, US3null or a rescue virus (Figure 2B). A clear increase in p-S404 lamin A/C signal was observed in cells infected with wild type or rescue virus, while this signal was absent in cells infected with a virus that does not express US3. Total lamin A/C blotting also revealed an increase in lamin A/C levels during infection, but this eﬀect was independent of US3 expression. An eﬀect of PRV infection on total lamin A/C levels also appeared to be cell type dependent as lamin A/C levels were reduced upon infection in rabbit or swine kidney cells (RK-13 and SK cells) (data not shown). S404 of lamin A/C was initially described as a phosphorylation target of Akt [27–29]. In order to ﬁnd out whether the US3 triggered phosphorylation of lamin A/C occurs through Akt, we infected cells with PRV in the absence or presence of 5 µM of the Akt inhibitor MK-2206. MK-2206 treatment did not aﬀect phosphorylation of S404 in Lamin A/C, in line with a previous report of HSV1 US3 acting as an Akt mimic, rather than as an Akt-triggering viral protein (Figure 2C) [12]. Pathogens 2020, 9, x FOR PEER REVIEW 5 of 14 Figure 2. US3 triggers S404 phosphorylation of lamin A/C. (A) Swine testicle (ST) cells transfected for 48 h with eukaryotic expression vectors encoding GFP (pTrip), wild type US3 (pKG1) or kinase inactive US3 (pHF61). (B) ST cells were either mock infected, or infected with wild type (WT), US3null or US3null rescue PRV (16hpi). 2.2. Validation of Identiﬁed S404 Phosphosite in Lamin A/C (C) ST cells were mock infected, or infected with wild type PRV. At 2 hpi, the virus inoculum was washed away and cells were treated with 5 µM of the Akt inhibitor MK- 2206 for an additional 14 h. 2 3 G E i h A l i Figure 2. US3 triggers S404 phosphorylation of lamin A/C. (A) Swine testicle (ST) cells transfected for 48 h with eukaryotic expression vectors encoding GFP (pTrip), wild type US3 (pKG1) or kinase inactive US3 (pHF61). (B) ST cells were either mock infected, or infected with wild type (WT), US3null or US3null rescue PRV (16hpi). (C) ST cells were mock infected, or infected with wild type PRV. At 2 hpi, the virus inoculum was washed away and cells were treated with 5 µM of the Akt inhibitor MK-2206 for an additional 14 h. Figure 2. US3 triggers S404 phosphorylation of lamin A/C. (A) Swine testicle (ST) cells transfected for 48 h with eukaryotic expression vectors encoding GFP (pTrip), wild type US3 (pKG1) or kinase inactive US3 (pHF61). (B) ST cells were either mock infected, or infected with wild type (WT), US3null or US3null rescue PRV (16hpi). (C) ST cells were mock infected, or infected with wild type PRV. At 2 hpi, the virus inoculum was washed away and cells were treated with 5 µM of the Akt inhibitor MK- 2206 for an additional 14 h. Figure 2. US3 triggers S404 phosphorylation of lamin A/C. (A) Swine testicle (ST) cells transfected for 48 h with eukaryotic expression vectors encoding GFP (pTrip), wild type US3 (pKG1) or kinase inactive US3 (pHF61). (B) ST cells were either mock infected, or infected with wild type (WT), US3null or US3null rescue PRV (16hpi). (C) ST cells were mock infected, or infected with wild type PRV. At 2 hpi, the virus inoculum was washed away and cells were treated with 5 µM of the Akt inhibitor MK-2206 for an additional 14 h. 2.1. The Phosphoproteome of US3-Transfected Cells Gene Name Protein Name Site Log2(KD/WT) -Log2(p-Value) CDS2 Phosphatidate cytidylyltransferase S33 6.38 4.56 BCKDHA 2-oxoisovalerate dehydrogenase alpha S313 5.83 3.48 BCKDHA 2-oxoisovalerate dehydrogenase alpha S303 5.54 3.25 PPP6R1 S/T-protein phosphatase 6 subunit 1 S531 4.39 2.92 WDR20 WD repeat-containing protein 20 S348 4.16 3.83 Table 2. Top 10 Phosphorylated proteins in US3 transfected cells compared to kinase negative US3 transfected cells. Gene Name Protein Name Site Log2(KD/WT) -Log2(p-Value) TSSC1 EARP and GARP complex interacting protein 1 S320 −6.50 4.65 RAB11FIP5 Rab11 family-interacting protein 5 T162 −6.24 3.53 RAB11FIP5 Rab11 family-interacting protein 5 S164 −6.09 3.98 TOMM70 Mitochondrial import receptor subunit TOM70 S97 −6.01 4.52 TESK2 Dual speciﬁcity testis-speciﬁc protein kinase 2 S8 −5.85 6.37 SZRD1 SUZ domain-containing protein 1 S17 −5.48 2.28 DDX17 Probable ATP-dependent RNA helicase DDX17 S575 −5.23 3.85 LMNA Prelamin-A/C S12 −5.21 1.56 SZRD1 SUZ domain-containing protein 1 S19 −5.09 5.62 PALMD Palmdelphin T255 −4.78 3.29 ified using our MS approach. This shows the value of using alternative methodologies to ify phosphorylation sites and indicates that the list of phosphorylated sites upon US3 expression ble 2. Top 10 Phosphorylated proteins in US3 transfected cells compared to kinase negative US3 entified using our MS approach. This shows the value of using alternative methodologies to entify phosphorylation sites and indicates that the list of phosphorylated sites upon US3 expression not exhausti e Table 2. Top 10 Phosphorylated proteins in US3 transfected cells compared to kinase negative US3 transfected cells. Pathogens 2020, 9, 916 4 of 13 3 Discussion 3. Discussion 3. Discussion The US3 protein is an important virulence factor in alphaherpesviruses and fulfills a wide array of functions in virus-host interactions of both HSV and PRV. Here, we used a mass spectrometry approach to identify proteins differentially phosphorylated upon expression of the US3 protein versus expression of a kinase-inactive US3 mutant. In total we identified 78 differentially phosphorylated phosphosites in cells transfected with wild type US3 versus cells transfected with kinase inactive US3, of which 64 were phosphorylated and 14 were dephosphorylated. These may include proteins directly phosphorylated by US3 and/or phosphosites targeted by cellular kinases or phosphatases that are activated or inactivated by US3 expression. We identified and confirmed S404 as a phosphorylation site in lamin A/C in both US3-transfected and PRV-infected cells. S404 of lamin A/C was previously shown to be phosphorylated by the cellular kinase Akt, further pointing towards US3 as a functional mimic of Akt [28,30]. US3-mediated phosphorylation of S404 lamin A/C in PRV- infected cells does not depend on Akt, as phosphorylation still occurs in the presence of an Akt inhibitor. This is in line with a study on HSV1 US3, where it was shown that the US3-dependent phosphorylation of the Akt target 4E-BP1 does not require Akt activity [12]. Akt phosphorylation of lamin A/C induces lamin degradation and an alternative morphology of the nuclear lamina, resembling the phenotype observed in Emery Dreifuss muscular dystrophy cells [27 29] Although The US3 protein is an important virulence factor in alphaherpesviruses and fulﬁlls a wide array of functions in virus-host interactions of both HSV and PRV. Here, we used a mass spectrometry approach to identify proteins diﬀerentially phosphorylated upon expression of the US3 protein versus expression of a kinase-inactive US3 mutant. In total we identiﬁed 78 diﬀerentially phosphorylated phosphosites in cells transfected with wild type US3 versus cells transfected with kinase inactive US3, of which 64 were phosphorylated and 14 were dephosphorylated. These may include proteins directly phosphorylated by US3 and/or phosphosites targeted by cellular kinases or phosphatases that are activated or inactivated by US3 expression. We identiﬁed and conﬁrmed S404 as a phosphorylation site in lamin A/C in both US3-transfected and PRV-infected cells. S404 of lamin A/C was previously shown to be phosphorylated by the cellular kinase Akt, further pointing towards US3 as a functional mimic of Akt [28,30]. 2.3. Gene Enrichment Analysi I d t id tif 2.3. Gene Enrichment Analysis In order to identify specific pathways affected by US3 expression, we performed gene ontology (GO) enrichment analysis of differentially phosphorylated proteins (Figure 3 and Table S2). Surprisingly, we found biological processes related to RNA processing to be substantially enriched. Apart from promoting transcriptional activity by phosphorylation of HDAC proteins, US3 has no known functions in RNA processing. Biological processes related to “nucleocytoplasmic transport” l t d i diff ti ll h h l ti it (T bl S2) It i th f ibl In order to identify specific pathways affected by US3 expression, we performed gene ontology (GO) enrichment analysis of differentially phosphorylated proteins (Figure 3 and Table S2). Surprisingly, we found biological processes related to RNA processing to be substantially enriched. Apart from promoting transcriptional activity by phosphorylation of HDAC proteins, US3 has no known functions in 5 of 13 Pathogens 2020, 9, 916 RNA processing. Biological processes related to “nucleocytoplasmic transport” were also overrepresented in differentially phosphorylation sites (Table S2). It is therefore possible that some of the functions of the US3 protein are performed by shuttling proteins between the nucleus and the cytoplasm. “Negative regulation of cardiac muscle adaptation” and “positive regulation of histone H3-K9 trimethylation” were both found to also be enriched, however these ontologies are poorly annotated and each contain only two or three annotated proteins in the porcine database (Table S2). It can therefore not be excluded at this point that these ontologies may possibly represent false positives. Pathogens 2020, 9, x FOR PEER REVIEW 6 of 14 Figure 3. Overview of significantly enriched gene ontologies in differentially phosphorylated proteins. Each circle represents a significantly enriched gene ontology. The number between brackets is the number of proteins present in this ontology. Figure created via BioRender.com. Figure 3. Overview of signiﬁcantly enriched gene ontologies in diﬀerentially phosphorylated proteins. Each circle represents a signiﬁcantly enriched gene ontology. The number between brackets is the number of proteins present in this ontology. Figure created via BioRender.com. Figure 3. Overview of significantly enriched gene ontologies in differentially phosphorylated proteins. Each circle represents a significantly enriched gene ontology. The number between brackets is the number of proteins present in this ontology. Figure created via BioRender.com. Figure 3. Overview of signiﬁcantly enriched gene ontologies in diﬀerentially phosphorylated proteins. Each circle represents a signiﬁcantly enriched gene ontology. The number between brackets is the number of proteins present in this ontology. 2.3. Gene Enrichment Analysi I d t id tif 2.3. Gene Enrichment Analysis Figure created via BioRender.com. 3 Discussion 3. Discussion US3-mediated phosphorylation of S404 lamin A/C in PRV-infected cells does not depend on Akt, as phosphorylation still occurs in the presence of an Akt inhibitor. This is in line with a study on HSV1 US3, where it was shown that the US3-dependent phosphorylation of the Akt target 4E-BP1 does not require Akt activity [12]. Akt phosphorylation of lamin A/C induces lamin degradation and an alternative morphology of the nuclear lamina, resembling the phenotype observed in Emery–Dreifuss muscular dystrophy cells [27–29]. Although no reduction in total lamin A/C levels Pathogens 2020, 9, 916 6 of 13 was observed upon PRV infection in ST cells, a substantial reorganization of the nuclear lamina upon herpesvirus infections has been previously described [11,31,32]. Apart from lamin A/C, several promising novel phosphorylation targets were identiﬁed. Although these phosphosites have not been functionally characterized before, they might represent new research avenues, leading to a better understanding of the functions of US3. One of the most striking eﬀects of the US3 protein is the dramatic reorganization of the actin cytoskeleton, involving the breakdown of actin stress ﬁbers and the formation of tunneling nanotubes [7,26]. These changes were shown to depend on dephosphorylation of coﬁlin at position S3 [33]. Interestingly, one of the signiﬁcantly phosphorylated proteins identiﬁed in our screen was testis-speciﬁc protein kinase 2 (TESK2). TESK2 belongs to the TESK/LIMK family of protein kinases, and has been shown to phosphorylate coﬁlin at position S3 [34,35]. Although regulation of TESK2 by phosphorylation has not been described, the activity of the related LIMK protein is regulated by phosphorylation [36]. Although speculative at this point, this may indicate that US3-induced phosphorylation of TESK2 may aﬀect TESK2 activity, thereby modulating S3 phosphorylation of coﬁlin. Another diﬀerentially phosphorylated protein potentially involved in the cytoskeletal rearrangements induced by the US3 protein is ENAH. ENAH, a member of the Ena/VASP family of actin regulators, promotes the formation of ﬁlopodia [37]. Although the activity of ENAH has been previously shown to be regulated by phosphorylation [38,39], the eﬀect of the phosphosite identiﬁed in the current screen is unknown. p p y p p Tunneling nanotubes induced by US3 are remarkably stable compared to endogenous TNTs described in literature [26]. We previously showed that they contain microtubules with post-translational modiﬁcations associated with stability, and that the contact area with connected cells is enriched in cadherins and beta-catenin [26]. 3 Discussion 3. Discussion Some of the proteins identiﬁed in the current phosphoproteomics screen might shed light on the mechanism underlying this TNT stabilization. Two proteins involved in the regulation of microtubule stability were found to be diﬀerentially phosphorylated. The microtubule-associated protein 1B (MAP1B) regulates microtubule dynamics and is regulated by phosphorylation [40–42]. Interestingly, the phosphosite that was found to be dephosphorylated in the current screen was previously found to be dephosphorylated in a phosphoproteomics screen investigating the eﬀects of a combined inhibition of ROCK1 and epidermal growth factor [43]. ROCK1 inhibition leads to the formation of cell protrusions resembling US3-induced TNTs [7], pointing towards a possible function of S891 MAP1B dephosphorylation in TNT formation. CLASP1, another microtubule regulator, was found to be signiﬁcantly phosphorylated and has previously been described to be important in US3-mediated microtubule stabilization [44]. Although the phosphosite identiﬁed in the current screen has not been characterised, it was previously identiﬁed in a screen identifying phosphorylation targets of AMP-activated protein kinase (AMPK), together with several other proteins involved in cell motility and invasion [45]. AMPK activation has previously been shown to induce stress ﬁber disassembly, a phenotype also observed upon US3 expression [46,47]. Finally, phosphorylation of cadherin-6, also known as K-cadherin, could also be involved in the stability of US3-induced TNTs. K-cadherin is phosphorylated in its C-terminal intracellular domain. Phosphorylation of E-cadherin in its intracellular domain enhances its association with beta-catenin and strengthens cell–cell adhesion [48,49]. Although K-cadherin phosphorylation has not been functionally described before, it is possible that K-cadherin is regulated similarly and that this may contribute to stabilization of US3-induced TNTs. One of the most surprising ﬁndings of the current phosphoproteome screen was the large number of proteins involved in RNA processing. Gene ontology enrichment analysis showed a marked enrichment of phosphorylated proteins involved in mRNA processing. US3 has previously been shown to phosphorylate histone deacetylases (HDAC), leading to increased reporter gene expression [50,51], but the gene ontologies enriched in proteins phosphorylated by US3 are mostly related to RNA splicing. US3 is not known to aﬀect splicing, indicating that this might be a novel function of US3. Despite the large amount of promising new phosphosites identiﬁed in the current screen, most of the proteins that were demonstrated before to be phosphorylated by US3 were not identiﬁed. 7 of 13 Pathogens 2020, 9, 916 The absence of these phosphosites does not contradict previous ﬁndings. 3 Discussion 3. Discussion Rather, it shows that the list of US3 phosphorylation targets is not complete. While most known proteins phosphorylated by US3 were identiﬁed via hypothesis-driven experiments, our approach does not rely on assumptions or prior knowledge. Although this allows the identiﬁcation of uncharacterized phosphosites, there is no selection for biologically functional sites. In order to draw conclusions on the eﬀect of these sites, extensive validation will be required. The absence of these phosphosites does not contradict previous ﬁndings. Rather, it shows that the list of US3 phosphorylation targets is not complete. While most known proteins phosphorylated by US3 were identiﬁed via hypothesis-driven experiments, our approach does not rely on assumptions or prior knowledge. Although this allows the identiﬁcation of uncharacterized phosphosites, there is no selection for biologically functional sites. In order to draw conclusions on the eﬀect of these sites, extensive validation will be required. In conclusion, we identiﬁed a wide range of novel phosphosites aﬀected by PRV US3 expression. These include several phosphorylated residues of lamin A/C, a known US3 substrate, of which S404 phosphorylation was validated using a phospho-speciﬁc antibody. Enrichment analysis of signiﬁcantly phosphorylated proteins revealed a remarkable number of proteins involved in mRNA processing. Our results provide several research avenues that can lead to a better understanding of the functions of the US3 protein in the PRV replication cycle, as well as of the regulation of cellular proteins via phosphorylation. 4.1. Cells, Inhibitors and Viruses Swine testicle (ST) cells were cultured in MEM (Gibco) supplemented with 10% fetal calf serum (FCS), 1mMsodiumpyruvate, 105 U/Lpenicillin, 100mg/Lstreptomycinand50mg/Lgentamycin. Confluentcells were infected at an MOI of 10 and lysed at 16 hpi. Wild type NIA3, an isogenic US3null mutant and the corresponding rescue virus were previously described and were kindly donated by the ID-DLO (The Netherlands) [52,53]. The Akt inhibitor MK-2206 was purchased from Selleckchem. Cells were infected for 2 h after which the initial inoculum was washed away and 5 µM of MK-2206 in complete medium was added, in order to prevent the inhibitor from interfering with viral entry. 4.2. Transfection Cells were transfected with 2250 ng of plasmid DNA for each well of a 6 well plate using JetPrime (poly-plus) according to the manufacturer’s instructions. The plasmids encoding full length wild type NIA3 US3 protein (pKG1) and full length kinase inactive NIA3 US3 protein with a K138Q mutation (pHF61) were described previously [8,25]. These plasmids allow the expression of both the short and long isoform of the US3 proteins, as is the case during PRV infection. The pTrip plasmid encoding eGFP was a kind gift of B. Verhasselt (Ghent University, Ghent, Belgium). Cells intended for mass spectrometry were lysed at 16 h post transfection (hpt) to prevent large diﬀerences in US3 expression between the active and the kinase inactive US3 plasmid. Three replicates were performed. Cells transfected in parallel were used to determine the transfection eﬃciency by immunoﬂuorescence followed by ﬂow cytometry. Transfection eﬃciencies of both samples were similar in each of the replicates. 4.4. Mass Spectrometry Sample Preparation Cells were lysed in a urea lysis buﬀer containing 9 M urea, 20 mM HEPES pH 8.0 and PhosSTOP phosphatase inhibitor cocktail (Roche, Basel, Switzerland), 1 tablet/10 mL buﬀer). The samples were sonicated with 3 pulses of 15 s at an amplitude of 20% using a 3 mm probe, with incubation on ice for 1 min between pulses. After centrifugation for 15 min at 20,000× g at room temperature to remove insoluble components, proteins were reduced by addition of 5 mM DTT and incubation for 30 min at 55 ◦C and then alkylated by addition of 10 mM iodoacetamide and incubation for 15 min at room temperature in the dark. The protein concentration was measured using a Bradford assay (Bio-rad) and from each sample 2 mg protein was used to continue the protocol. Samples were further diluted with 20 mM HEPES pH 8.0 to a ﬁnal urea concentration of 4 M and proteins were digested with 8 µg LysC (Wako, Osaka, Japan) (1/250, w/w) for 4 h at 37 ◦C. Samples were again diluted to 2 M urea and digested with 20 µg trypsin (Promega) (1/100, w/w) overnight at 37 ◦C. The resulting peptide mixture was acidiﬁed by addition of 1% triﬂuoroacetic acid (TFA) and after 15 min incubation on ice, samples were centrifuged for 15 min at 1780× g at room temperature to remove insoluble components. Next, peptides were puriﬁed on SampliQ SPE C18 cartridges (500 mg, Agilent, Santa Clara, USA). Columns were ﬁrst washed with 5 mL 100% acetonitrile (ACN) and pre-equilibrated with 15 mL of solvent A (0.1% TFA in water/ACN (98:2, v/v)) before samples were loaded on the column. After peptide binding, the column was washed again with 5 mL of solvent A and peptides were eluted twice with 700 µL elution buﬀer (0.1% TFA in water/ACN (20:80, v/v)). The eluted peptides were divided in two parts: 100 µL was dried completely in a speedvac vacuum concentrator for shotgun analysis, while the remainder was used for phosphopeptide enrichment. Phosphopeptides were enriched with MagReSyn® Ti-IMAC beads following the protocol according to the manufacturer’s instructions with slight modiﬁcations. Brieﬂy, 200 µl MagReSyn® Ti-IMAC beads (per sample) were washed twice with 70% EtOH, once with 1% NH4OH and three times with a mixture of water/ACN/TFA (14:80:6, v/v/v). 4.3. Flow Cytometry Cells transfected with US3 or kinase negative US3 were collected by trypsinization. After washing the cells with PBS, the cells were ﬁxed using 3% paraformaldehyde and permeabilized using 0.1% saponin (Sigma Aldrich, St. Louis, MO, USA). Cells were incubated with a 1/100 dilution of a mouse anti-US3 antibody in PBS with 0.1% saponin. The anti-US3 antibody was a kind gift of Leigh Anne Olsen and Lynn Enquist (Princeton University, Princeton, NJ, USA). Following three washing steps in PBS, cells were incubated with a 1/200 dilution of goat anti-mouse AF647 antibody (ThermoFisher, Waltham, MA, USA). After three additional washing steps the cells were analyzed using a BD FACSAria III Cell Sorter (BD Biosciences, Franklin Lakes, NJ, USA). 8 of 13 Pathogens 2020, 9, 916 4.4. Mass Spectrometry Sample Preparation Next, the digested sample was incubated with the washed beads for 30 min at room temperature, the beads were washed once with a mixture of water/ACN/TFA (14:80:6, v/v/v) and three times with a mixture of water/ACN/TFA (19:80:1, v/v/v). Phosphopeptides were eluted from the beads by adding three times 80 µl 1% NH4OH. In total, 60 µL 10% formic acid (FA) was added to the combined eluate and the samples were dried completely in a speedvac vacuum concentrator. 4.6. Data Analysis Analysis of the shotgun and phosphoproteomics data was performed in two separate MaxQuant searches (version 1.5.6.5) with identical, mainly default search settings including a false discovery rate set at 1% on PSM, peptide and protein level. Spectra were searched against the Sus scrofa proteins in the Uniprot/Swiss-Prot database (database release version of November 2016 containing 26,101 pig protein sequences, downloaded from http://www.uniprot.org) supplemented with the sequences of the kinase active/inactive US3 Pseudorabies virus (NIA3 strain). The mass tolerance for precursor and fragment ions was set to 4.5 and 20 ppm, respectively, during the main search. Enzyme speciﬁcity was set as C-terminal to arginine and lysine, also allowing cleavage at proline bonds with a maximum of two missed cleavages. Variable modiﬁcations were set to oxidation of methionine residues, acetylation of protein N-termini and phosphorylation of serine, threonine or tyrosine residues, while carbamidomethylation of cysteine residues was set as ﬁxed modiﬁcation. Matching between runs was enabled with a matching time window of 1 min and an alignment time window of 20 min. Only proteins with at least one unique or razor peptide were retained leading to the identiﬁcation of 4743 proteins in the shotgun samples and 5583 phosphorylation sites in the phosphopeptide enriched samples. Proteins identiﬁed in the shotgun samples were quantiﬁed by the MaxLFQ algorithm integrated in the MaxQuant software [56]. A minimum ratio count of two unique or razor peptides was required for quantiﬁcation. Further data analysis was performed with the Perseus software (version 1.5.5.3) after loading the proteingroups ﬁle from MaxQuant. Reverse database hits and proteins that were only identiﬁed by site were removed, LFQ intensities were log2 transformed and replicate samples were grouped. Proteins with less than three valid values in at least one group were removed and missing values were imputed from a normal distribution around the detection limit leading to a list of 3109 quantiﬁed proteins that was used for further data analysis. Then, a t-test was performed (FDR = 0.05 and s0 = 1) to compare wild type and mutant samples and a volcano plot was generated. Three proteins were found to be signiﬁcantly regulated. For further analysis of the phosphoproteomics data, the phospho(STY)sites ﬁle was loaded in the Perseus software (version 1.5.5.3). Reverse hits were removed, the site table was expanded and the intensity values were log2 transformed. 4.5. LC-MS/MS Analysis Puriﬁed peptides for shotgun analysis were re-dissolved in 100 µL solvent A and the peptide concentration was determined on a Lunatic spectrophotometer (Unchained Labs) [54]. In total, 2 µg of each sample was injected for LC-MS/MS analysis on an Ultimate 3000 RSLCnano system (Thermo) in-line connected to a Q Exactive HF mass spectrometer (Thermo) equipped with a Nanospray Flex Ion source (Thermo). Peptides resulting from phosphopeptide enrichment were re-dissolved in 20 µl solvent A of which 10 µL was injected for LC-MS/MS analysis on the same system. Trapping of peptides was performed at 10 µL/min for 4 min in solvent A on a 20 mm trapping column (made in-house, 100 µm internal diameter (I.D.), 5 µm beads, C18 Reprosil-HD, Dr. Maisch, Ammerbuch, Germany) and the sample was loaded on an analytical column packed in the needle (made in-house, 75 µm I.D. × 400 mm, 1.9 µm beads C18 Reprosil-HD, Dr. Maisch). Peptides were eluted by a non-linear increase from 2 to 56% MS solvent B (0.1% FA in water/ACN (2:8, v/v)) over 145 min at a constant ﬂow rate of 250 nl/min and a constant temperature of 50 ◦C (CoControl 3.3.05, Sonation), followed by a 15-min wash reaching 99% MS solvent B and re-equilibration with MS solvent A (0.1% FA in water/ACN (2:8, v/v)). The mass spectrometer was operated in data-dependent mode, automatically switching between MS and MS/MS acquisition for the 16 most abundant ion peaks per MS spectrum. Full-scan MS spectra (375–1500 m/z) were acquired at a resolution of 60,000 in the orbitrap analyser after accumulation to a target value of 3,000,000. The 16 most intense ions above a threshold value of 22,000 (shotgun) or 13,000 (phospho) were isolated (window of 1.5 Th) for fragmentation at a normalized collision energy of 28% after ﬁlling the trap at a target value of 100,000 for maximum 45 ms (shotgun) or 9 of 13 Pathogens 2020, 9, 916 80 ms (phospho). MS/MS spectra (200–2000 m/z) were acquired at a resolution of 15,000 in the orbitrap analyser. The S-lens RF level was set at 55 and we excluded precursor ions with single, unassigned and >6 charge states from fragmentation selection. QCloud was used to control instrument longitudinal performance during the project [55]. 4.6. Data Analysis Replicate samples were grouped, phosphosites with less than three valid values in at least one group were removed and missing values were imputed from a normal distribution around the detection limit leading to a list of 2971 quantiﬁed phosphopeptides that was used for further data analysis. Then, a t-test was performed (FDR = 0.05 and s0 = 1) to compare control and KO samples and a volcano plot was generated. In total, 81 phosphopeptides were signiﬁcantly regulated. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE [57] partner repository with the dataset identiﬁer PXD021751. 4.8. Gene Ontology Enrichment Analysis Gene Ontology Enrichment Analysis The enrichment of “biological process” gene ontologies of the diﬀerentially phosphorylated proteins was analyzed using the PANTHER overrepresentation test against the Sus Scrofa database (GO Ontology database DOI: 10.5281/zenodo.3873405 Released 2020-06-01). Fisher’s Exact test was used with a 0.05% false discovery rate correction. Supplementary Materials: The following are available online at http://www.mdpi.com/2076-0817/9/11/916/s1, Table S1: Overview of shotgun and phospho- mass spectrometry data, Table S2: signiﬁcantly enriched GO terms. Figure S1: Flow cytometry analysis of transfection eﬃciency of samples used for mass spectrometry. Author Contributions: Conceptualization, R.J.J.J. and H.F.; methodology, R.J.J.J. and H.W.F.; investigation, R.J.J.J.; resources, S.M.; writing—original draft preparation, R.J.J.J.; writing—review and editing, H.W.F.; supervision, H.W.F.; project administration, H.W.F.; funding acquisition, H.W.F. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by F.W.O.-Vlaanderen (research grants G019617N and G060119N) and the Special Research Fund of Ghent University (concerted research grant GOA013-17 and basic infrastructure grant BAS003-18). Acknowledgments: We would like to thank CliﬀVan Waesberghe for the excellent technical assistance provided, the VIB Proteomics Core for the analysis of the proteomics data, Leigh Anne Olsen and Lynn Enquist (Princeton University, Princeton, NJ, USA) for their kind gift of the anti-PRV US3 monoclonal antibodies, Bruno Verhasselt (Ghent University, Ghent, Belgium) for the pTrip plasmid and the ID-DLO (The Netherlands) for the NIA3 PRV strains. Conﬂicts of Interest: The 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. 4.7. Western Blotting Cells were lysed at 48 hpt or 16 hpi in RIPA buffer (Abcam, Cambridge, UK) with cOmplete mini EDTA free protease inhibitor cocktail (Roche, Basel, Switzerland ) and PhosStop (Roche). Cell lysates were separated on a 10% polyacrylamide gel, followed by blotting on PVDF membrane (GE healthcare, Piscataway, NJ, USA). Regular blots were blocked in 5% nonfat milk diluted in 0.1% Tween-20 in PBS for 1 h at room temperature. When using phosphorylated protein-specific antibodies, 5% bovine serum albumin (MP Biomedicals, Santa Ana, USA) diluted in 0.1% Tween-20 in PBS was used for blocking. Primary antibodies were incubated overnight at 4 ◦C (Table 3). Following 3 consecutive 5 min washes in PBS-T, the membranes were incubated with the secondary antibody for 1h at room temperature. Following 3 more 5 min washes, the blots were detected using Pierce enhanced chemiluminescence (ECL) 10 of 13 Pathogens 2020, 9, 916 substrate (ThermoFisher, Waltham, MA, USA), ECL Plus substrate (GE Healthcare, Piscataway, NJ, USA), or SuperSignal West Femto maximum sensitivity substrate (ThermoFisher, Waltham, MA, USA) on a ChemiDoc MP imaging device (Bio-Rad, Hercules, CA, USA). Table 3. Primary antibodies used for Western blotting. Table 3. Primary antibodies used for Western blotting. Target Concentration Source Alpha-tubulin 1/1000 ab40742, Abcam LMNA 1/500 ab208798, Abcam p-S404 LMNA 1/1000 [27] US3 1/100 Leigh Anne Olsen and Lynn Enquist, [4] Princeton University, Princeton, USA 4.8. Gene Ontology Enrichment Analysis References 1. Jacob, T.; Broeke, C.V.D.; Favoreel, H.W. Viral Serine/Threonine Protein Kinases. J. Virol. 2010, 85, 1158–1173. [CrossRef] [PubMed] 1. Jacob, T.; Broeke, C.V.D.; Favoreel, H.W. Viral Serine/Threonine Protein Kinases. J. Virol. 2010, 85, 1158–1173. [CrossRef] [PubMed] 2. Deruelle, M.J.; Favoreel, H.W. Keep it in the subfamily: The conserved alphaherpesvirus US3 protein kinase. J. Gen. Virol. 2011, 92, 18–30. [CrossRef] [PubMed] 2. Deruelle, M.J.; Favoreel, H.W. Keep it in the subfamily: The conserved alphaherpesvirus US3 protein kinase. J. Gen. Virol. 2011, 92, 18–30. [CrossRef] [PubMed] 3. Kimman, T.G.; De Wind, N.; De Bruin, T.; De Visser, Y.; Voermans, J. Inactivation of Glycoprotein gE and Thymidine Kinase or the US3-Encoded Protein Kinase Synergistically Decreases In Vivo Replication of Pseudorabies Virus and the Induction of Protective Immunity. Virology 1994, 205, 511–518. [CrossRef] [PubMed] 4. Olsen, L.M.; Ch’Ng, T.H.; Card, J.P.; Enquist, L.W. Role of Pseudorabies Virus Us3 Protein Kinase during Neuronal Infection. J. Virol. 2006, 80, 6387–6398. [CrossRef] [PubMed] 4. Olsen, L.M.; Ch’Ng, T.H.; Card, J.P.; Enquist, L.W. Role of Pseudorabies Virus Us3 Protein Kinase during Neuronal Infection. J. Virol. 2006, 80, 6387–6398. [CrossRef] [PubMed] 5. De Wind, N.; Domen, J.; Berns, A. Herpesviruses encode an unusual protein-serine/threonine kinase which is nonessential for growth in cultured cells. J. Virol. 1992, 66, 5200–5209. [CrossRef] 5. De Wind, N.; Domen, J.; Berns, A. Herpesviruses encode an unusual protein-serine/threonine kinase which is nonessential for growth in cultured cells. J. Virol. 1992, 66, 5200–5209. [CrossRef] 6. Koyanagi, N.; Kato, A.; Takeshima, K.; Maruzuru, Y.; Kozuka-Hata, H.; Oyama, M.; Arii, J.; Kawaguchi, Y. Regulation of Herpes Simplex Virus 2 Protein Kinase UL13 by Phosphorylation and Its Role in Viral Pathogenesis. J. Virol. 2018, 92, 1–19. [CrossRef] 6. Koyanagi, N.; Kato, A.; Takeshima, K.; Maruzuru, Y.; Kozuka-Hata, H.; Oyama, M.; Arii, J.; Kawaguchi, Y. Regulation of Herpes Simplex Virus 2 Protein Kinase UL13 by Phosphorylation and Its Role in Viral Pathogenesis. J. Virol. 2018, 92, 1–19. [CrossRef] 11 of 13 Pathogens 2020, 9, 916 7. Favoreel, H.W.; van Minnebruggen, G.; Adriaensen, D.; Nauwynck, H.J. Cytoskeletal rearrangements and cell extensions induced by the US3 kinase of an alphaherpesvirus are associated with enhanced spread. Proc. Natl. Acad. Sci. USA 2005, 102, 8990–8995. [CrossRef] 8. Deruelle, M.; Geenen, K.; Nauwynck, H.J.; Favoreel, H.W. A point mutation in the putative ATP binding site of the pseudorabies virus US3 protein kinase prevents Bad phosphorylation and cell survival following apoptosis induction. Virus Res. References 2007, 128, 65–70. [CrossRef] 9. Kato, A.; Kawaguchi, Y. Us3 Protein Kinase Encoded by HSV: The Precise Function and Mechanism on Viral Life Cycle. Adv. Exp. Med. Biol. 2018, 1045, 45–62. [CrossRef] 10. Wang, S.; Wang, K.; Lin, R.; Zheng, C. Herpes Simplex Virus 1 Serine/Threonine Kinase US3 Hyperphosphorylates IRF3 and Inhibits Beta Interferon Production. J. Virol. 2013, 87, 12814–12827. [CrossRef] 11. Mou, F.; Forest, T.; Baines, J.D. US3 of Herpes Simplex Virus Type 1 Encodes a Promiscuous Protein Kinase That Phosphorylates and Alters Localization of Lamin A/C in Infected Cells. J. Virol. 2007, 81, 6459–6470. [CrossRef] [PubMed] 12. Chuluunbaatar, U.; Roller, R.; Feldman, M.E.; Brown, S.; Shokat, K.M.; Mohr, I. Constitutive mTORC1 activation by a herpesvirus Akt surrogate stimulates mRNA translation and viral replication. Genes Dev. 2010, 24, 2627–2639. [CrossRef] 13. Cartier, A.; Komai, T.; Masucci, M.G. The Us3 protein kinase of herpes simplex virus 1 blocks apoptosis and induces phosporylation of the Bcl-2 family member Bad. Exp. Cell Res. 2003, 291, 242–250. [CrossRef] 14. Kato, A.; Yamamoto, M.; Ohno, T.; Kodaira, H.; Nishiyama, Y.; Kawaguchi, Y. Identiﬁcation of Proteins Phosphorylated Directly by the Us3 Protein Kinase Encoded by Herpes Simplex Virus 1. J. Virol. 2005, 79, 9325–9331. [CrossRef] 15. Xiong, R.; Rao, P.; Kim, S.; Li, M.; Wen, X.; Yuan, W. Herpes Simplex Virus 1 US3 Phosphorylates Cellular KIF3A To Downregulate CD1d Expression. J. Virol. 2015, 89, 6646–6655. [CrossRef] [PubMed] 16. Benetti, L.; Roizman, B. Herpes simplex virus protein kinase US3 activates and functionally overlaps protein kinase A to block apoptosis. Proc. Natl. Acad. Sci. USA 2004, 101, 9411–9416. [CrossRef] [PubMed] 17. Rubio, R.M.; Mohr, I. Inhibition of ULK1 and Beclin1 by an α-herpesvirus Akt-like Ser/Thr kinase limits autophagy to stimulate virus replication. Proc. Natl. Acad. Sci. USA 2019, 116, 26941–26950. [CrossRef] 18. Broeke, C.V.D.; Radu, M.; Deruelle, M.; Nauwynck, H.; Hofmann, C.; Jaﬀer, Z.M.; Chernoﬀ, J.; Favoreel, H.W. Alphaherpesvirus US3-mediated reorganization of the actin cytoskeleton is mediated by group A p21-activated kinases. Proc. Natl. Acad. Sci. USA 2009, 106, 8707–8712. [CrossRef] 19. Benetti, L.; Roizman, B. Protein kinase B/Akt is present in activated form throughout the entire replicative cycle of deltaU(S)3 mutant virus but only at early times after infection with wild-type herpes simplex virus 1. J. Virol. 2006, 80, 3341–3348. [CrossRef] 20. Jacob, T.; Broeke, C.V.D.; Van Waesberghe, C.; Van Troys, L.; Favoreel, H.W. Pseudorabies virus US3 triggers RhoA phosphorylation to reorganize the actin cytoskeleton. J. References Gen. Virol. 2015, 96, 2328–2335. [CrossRef] 20. Jacob, T.; Broeke, C.V.D.; Van Waesberghe, C.; Van Troys, L.; Favoreel, H.W. Pseudorabies virus US3 triggers RhoA phosphorylation to reorganize the actin cytoskeleton. J. Gen. Virol. 2015, 96, 2328–2335. [CrossRef] 21. Yang, S.; Pei, Y.; Zhao, A. iTRAQ-based Proteomic Analysis of Porcine Kidney Epithelial PK15 cells Infected ith P d bi i S i R 2017 7 45922 [C R f] 21. Yang, S.; Pei, Y.; Zhao, A. iTRAQ-based Proteomic Analysis of Porcine Kidney Epithelial PK15 cells Infected with Pseudorabies virus. Sci. Rep. 2017, 7, 45922. [CrossRef] 21. Yang, S.; Pei, Y.; Zhao, A. iTRAQ-based Proteomic Analysis of with Pseudorabies virus. Sci. Rep. 2017, 7, 45922. [CrossRef] 22. Soh, T.K.; Davies, C.T.; Muenzner, J.; Hunter, L.M.; Barrow, H.G.; Connor, V.; Bouton, C.R.; Smith, C.; Emmott, E.; Antrobus, R.; et al. Temporal Proteomic Analysis of Herpes Simplex Virus 1 Infection Reveals Cell-Surface Remodeling via pUL56-Mediated GOPC Degradation. Cell Rep. 2020, 33, 108235. [CrossRef] 23. Kulej, K.; Avgousti, D.C.; Sidoli, S.; Herrmann, C.; Della Fera, A.N.; Kim, E.T.; Garcia, B.A.; Weitzman, M.D. Time-resolved Global and Chromatin Proteomics during Herpes Simplex Virus Type 1 (HSV-1) Infection. Mol. Cell. Proteom. 2017, 16, S92–S107. [CrossRef] 24. Magalhães-Junior, M.J.; Baracat-Pereira, M.C.; Pereira, L.K.J.; Vital, C.E.; Santos, M.R.; Cunha, P.S.; Fernandes, K.M.; Bressan, G.C.; Fietto, J.L.R.; Silva-Júnior, A.; et al. Proteomic and phosphoproteomic analyses reveal several events involved in the early stages of bovine herpesvirus 1 infection. Arch. Virol. 2019, 165, 69–85. [CrossRef] 25. Geenen, K.; Favoreel, H.W.; Olsen, L.; Enquist, L.W.; Nauwynck, H.J. The pseudorabies virus US3 protein kinase possesses anti-apoptotic activity that protects cells from apoptosis during infection and after treatment with sorbitol or staurosporine. Virolology 2005, 331, 144–150. [CrossRef] 12 of 13 Pathogens 2020, 9, 916 26. Jansens, R.J.J.; Broeck, W.V.D.; De Pelsmaeker, S.; Lamote, J.A.S.; Van Waesberghe, C.; Couck, L.; Favoreel, H.W. Pseudorabies Virus US3-Induced Tunneling Nanotubes Contain Stabilized Microtubules, Interact with Neighboring Cells via Cadherins, and Allow Intercellular Molecular Communication. J. Virol. 2017, 91, e00749-17. [CrossRef] 27. Cenni, V.; Bertacchini, J.; Beretti, F.; Lattanzi, G.; Bavelloni, A.; Riccio, M.; Ruzzene, M.; Marin, O.; Arrigoni, G.; Parnaik, V.; et al. Lamin A Ser404 Is a Nuclear Target of Akt Phosphorylation in C2C12 Cells. J. Proteome Res. 2008, 7, 4727–4735. [CrossRef] 28. Bertacchini, J.; Beretti, F.; Cenni, V.; Guida, M.; Gibellini, F.; Mediani, L.; Marin, O.; Maraldi, N.M.; De Pol, A.; Lattanzi, G.; et al. References The protein kinase Akt/PKB regulates both prelamin A degradation and Lmna gene expression. FASEB J. 2013, 27, 2145–2155. [CrossRef] 29. Naeem, A.S.; Zhu, Y.; Di, W.L.; Marmiroli, S.; O’Shaughnessy, R.F.L. AKT1-mediated Lamin A/C degradation is required for nuclear degradation and normal epidermal terminal diﬀerentiation. Cell Death Diﬀer. 2015, 22, 2123–2132. [CrossRef] [PubMed] 30. Chuluunbaatar, U.; Mohr, I. A herpesvirus kinase that masquerades as Akt. Cell Cycle 2011, 10, 2064–2068. [CrossRef] 31. Reynolds, A.E.; Liang, L.; Baines, J.D. Conformational Changes in the Nuclear Lamina Induced by Herpes Simplex Virus Type 1 Require Genes U L 31 and U L 34. J. Virol. 2004, 78, 5564–5575. [CrossRef] 32. Leach, N.R.; Roller, R.J. Signiﬁcance of host cell kinases in herpes simplex virus type 1 egress and lamin-associated protein disassembly from the nuclear lamina. Virology 2010, 406, 127–137. [CrossRef] 33. Jacob, T.; Broeke, C.V.D.; Van Troys, M.; Waterschoot, D.; Ampe, C.; Favoreel, H.W. Alphaherpesviral US3 Kinase Induces Coﬁlin Dephosphorylation To Reorganize the Actin Cytoskeleton. J. Virol. 2013, 87, 4121–4126. [CrossRef] [PubMed] 34. Toshima, J. Coﬁlin Phosphorylation and Actin Reorganization Activities of Testicular Protein Kinase 2 and Its Predominant Expression in Testicular Sertoli Cells. J. Biol. Chem. 2001, 276, 31449–31458. [CrossRef] 35. Røsok, Ø.; Pédeutour, F.; Ree, A.H.; Aasheim, H.-C. Identiﬁcation and Characterization of TESK2, a Novel Member of the LIMK/TESK Family of Protein Kinases, Predominantly Expressed in Testis. Genomics 1999, 61, 44–54. [CrossRef] [PubMed] 6. Edwards, D.C.; Sanders, L.C.; Bokoch, G.M.; Gill, G.N. Activation of LIM-kinase by Pak1 couples Rac/Cd GTPase signalling to actin cytoskeletal dynamics. Nat. Cell Biol. 1999, 1, 253–259. [CrossRef] GTPase signalling to actin cytoskeletal dynamics. Nat. Cell Biol. 1999, 1, 253–259. [CrossRef] 37. Krugmann, S.; Jordens, I.; Gevaert, K.; Driessens, M.; Vandekerckhove, J.; Hall, A. Cdc42 induces ﬁlopodia by promoting the formation of an IRSp53: Mena complex Curr Biol 2001 11 1645–1655 [CrossRef] g g y y 37. Krugmann, S.; Jordens, I.; Gevaert, K.; Driessens, M.; Vandekerckhove, J.; Hall, A. Cdc42 induces ﬁlopodia by promoting the formation of an IRSp53: Mena complex. Curr. Biol. 2001, 11, 1645–1655. [CrossRef] 38. Loureiro, J.J.; Rubinson, D.A.; Bear, J.E.; Baltus, G.A.; Kwiatkowski, A.V.; Gertler, F.B. Critical Roles of Phosphorylation and Actin Binding Motifs, but Not the Central Proline-rich Region, for Ena/Vasodilator-stimulated Phosphoprotein (VASP) Function during Cell Migration. Mol. Biol. Cell 2002, 13, 2533–2546. [CrossRef] [PubMed] 39. Lambrechts, A.; Kwiatkowski, A.V.; Lanier, L.M.; Bear, J.E.; Vandekerckhove, J.; Ampe, C.; Gertler, F.B. References cAMP-dependent Protein Kinase Phosphorylation of EVL, a Mena/VASP Relative, Regulates Its Interaction with Actin and SH3 Domains. J. Biol. Chem. 2000, 275, 36143–36151. [CrossRef] 40. Kawauchi, T.; Chihama, K.; Nishimura, Y.V.; Nabeshima, Y.I.; Hoshino, M. MAP1B phosphorylation is differentially regulated by Cdk5/p35, Cdk5/p25, and JNK. Biochem. Biophys. Res. Commun. 2005, 331, 50–55. [CrossRef] 41. Trivedi, N.; Marsh, P.; Goold, R.G.; Wood-Kaczmar, A.; Gordon-Weeks, P.R. Glycogen synthase kinase-3β phosphorylation of MAP1B at Ser1260 and Thr1265 is spatially restricted to growing axons. J. Cell Sci. 2005, 118, 993–1005. [CrossRef] 42. Fischer, I.; Romano-Clarke, G. Changes in Microtubule-Associated Protein MAP1B Phosphorylation During Rat Brain Development. J. Neurochem. 1990, 55, 328–333. [CrossRef] 43. Rontogianni, S.; Iskit, S.; Van Doorn, S.; Peeper, D.S.; Altelaar, A.F.M. Combined EGFR and ROCK Inhibition in Triple-negative Breast Cancer Leads to Cell Death Via Impaired Autophagic Flux. Mol. Cell. Proteom. 2019, 19, 261–277. [CrossRef] [PubMed] 44. Naghavi, M.H.; Gundersen, G.G.; Walsh, D. Plus-end tracking proteins, CLASPs, and a viral Akt mimic regulate herpesvirus-induced stable microtubule formation and virus spread. Proc. Natl. Acad. Sci. USA 2013, 110, 18268–18273. [CrossRef] 13 of 13 13 of 13 Pathogens 2020, 9, 916 45. Schaffer, B.E.; Levin, R.S.; Hertz, N.T.; Maures, T.J.; Schoof, M.L.; Hollstein, P.E.; Benayoun, B.A.; Banko, M.R.; Shaw, R.J.; Shokat, K.M.; et al. Identification of AMPK Phosphorylation Sites Reveals a Network of Proteins Involved in Cell Invasion and Facilitates Large-Scale Substrate Prediction. Cell Metab. 2015, 22, 907–921. [CrossRef] 46. Miranda, L.M.; Carpentier, S.; Platek, A.; Hussain, N.; Gueuning, M.-A.; Vertommen, D.; Ozkan, Y.; Sid, B.; Hue, L.; Courtoy, P.J.; et al. AMP-activated protein kinase induces actin cytoskeleton reorganization in epithelial cells. Biochem. Biophys. Res. Commun. 2010, 396, 656–661. [CrossRef] 47. Van Minnebruggen, G.; Favoreel, H.W.; Jacobs, L.; Nauwynck, H.J.; Minnebruggen, G.V. Pseudorabies Virus US3 Protein Kinase Mediates Actin Stress Fiber Breakdown. J. Virol. 2003, 77, 9074–9080. [CrossRef] 48. McEwen, A.E.; Maher, M.T.; Mo, R.; Gottardi, C.J. E-cadherin phosphorylation occurs during its biosynthesis to promote its cell surface stability and adhesion. Mol. Biol. Cell 2014, 25, 2365–2374. [CrossRef] 49. Lickert, H.; Bauer, A.; Kemler, R.; Stappert, J. Casein Kinase II Phosphorylation of E-cadherin Increases E-cadherin/β-Catenin Interaction and Strengthens Cell-Cell Adhesion. J. Biol. Chem. 2000, 275, 5090–5095. [CrossRef] 50. Walters, M.S.; Kinchington, P.R.; Banﬁeld, B.W.; Silverstein, S. Hyperphosphorylation of Histone Deacetylase 2 by Alphaherpesvirus US3 Kinases. J. Virol. 2010, 84, 9666–9676. [CrossRef] 51. Poon, A.P.W.; Gu, H.; Roizman, B. References ICP0 and the US3 protein kinase of herpes simplex virus 1 independently block histone deacetylation to enable gene expression. Proc. Natl. Acad. Sci. USA 2006, 103, 9993–9998. [CrossRef] 52. De Wind, N.; Zijderveld, A.; Glazenburg, K.; Gielkens, A.; Berns, A. Linker insertion mutagenesis of herpesviruses: Inactivation of single genes within the Us region of pseudorabies virus. J. Virol. 1990, 64, 4691–4696. [CrossRef] 53. Kimman, T.G.; De Wind, N.; Oei-Lie, N.; Pol, J.M.A.; Berns, A.J.M.; Gielkens, A.L.J. Contribution of single genes within the unique short region of Aujeszky’s disease virus (suid herpesvirus type 1) to virulence, pathogenesis and immunogenicity. J. Gen. Virol. 1992, 73, 243–251. [CrossRef] 54. Maia, T.M.; Staes, A.; Plasman, K.; Pauwels, J.; Boucher, K.; Argentini, A.; Martens, L.; Montoye, T.; Gevaert, K.; Impens, F. Simple Peptide Quantiﬁcation Approach for MS-Based Proteomics Quality Control. ACS Omega 2020, 5, 6754–6762. [CrossRef] 55. Chiva, C.; Olivella, R.; Borràs, E.; Espadas, G.; Pastor, O.; Solé, A.; Sabidó, E. QCloud: A cloud-based quality control system for mass spectrometry-based proteomics laboratories. PLoS ONE 2018, 13, e0189209. [CrossRef] 56. Cox, J.; Hein, M.Y.; Luber, C.A.; Paron, I.; Nagaraj, N.; Mann, M. Accurate Proteome-wide Label-free Quantiﬁcation by Delayed Normalization and Maximal Peptide Ratio Extraction, Termed MaxLFQ. Mol. Cell. 55. Chiva, C.; Olivella, R.; Borràs, E.; Espadas, G.; Pastor, O.; Solé, A.; Sabidó, E. QCloud: A cloud-based quality control system for mass spectrometry-based proteomics laboratories. PLoS ONE 2018, 13, e0189209. [CrossRef] , ; , ; , ; p , ; , ; , ; , Q q y control system for mass spectrometry-based proteomics laboratories. PLoS ONE 2018, 13, e0189209. [CrossRef] 56. Cox, J.; Hein, M.Y.; Luber, C.A.; Paron, I.; Nagaraj, N.; Mann, M. Accurate Proteome-wide Label-free Quantiﬁcation by Delayed Normalization and Maximal Peptide Ratio Extraction, Termed MaxLFQ. Mol. Cell. Proteom. 2014, 13, 2513–2526. [CrossRef] [PubMed] 56. Cox, J.; Hein, M.Y.; Luber, C.A.; Paron, I.; Nagaraj, N.; Mann, M. Accurate Proteome-wide Label-free Quantiﬁcation by Delayed Normalization and Maximal Peptide Ratio Extraction, Termed MaxLFQ. Mol. Cell. Proteom. 2014, 13, 2513–2526. [CrossRef] [PubMed] 57. Perez-Riverol, Y.; Csordas, A.; Bai, J.; Bernal-Llinares, M.; Hewapathirana, S.; Kundu, D.J.; Inuganti, A.; Griss, J.; Mayer, G.; Eisenacher, M.; et al. The PRIDE database and related tools and resources in 2019: Improving support for quantiﬁcation data. Nucleic Acids Res. 2018, 47, D442–D450. [CrossRef] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional aﬃliations. © 2020 by the authors. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). References Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W2001711125	https://bmcgenomics.biomedcentral.com/counter/pdf/10.1186/1471-2164-15-927	English	null	Transcriptional and epigenetic responses to mating and aging in Drosophila melanogaster	BMC genomics	2,014	cc-by	10,756	* Correspondence: anholt@ncsu.edu Department of Biological Sciences, W. M. Keck Center for Behavioral Biology and Program in Genetics, Box 7614, North Carolina State University, Raleigh, NC 27695-7617, USA RESEARCH ARTICLE Open Access Transcriptional and epigenetic responses to mating and aging in Drosophila melanogaster Transcriptional and epigenetic responses to mating and aging in Drosophila melanogaster Zhou et al. Transcriptional and epigenetic responses to mating and aging in Drosophila melanogaster Zhou et al. Zhou et al. Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Abstract Background: Phenotypic plasticity allows organisms to respond rapidly to changing environmental circumstances, and understanding its genomic basis can yield insights regarding the underlying genes and genetic networks affecting complex phenotypes. Female Drosophila melanogaster undergo dramatic physiological changes mediated by seminal fluid components transferred upon mating, including decreased longevity. Their physiological and behavioral effects have been well characterized, but little is known about resulting changes in regulation of gene expression or the extent to which mating-induced changes in gene expression are the same as those occurring during aging. Results: We assessed genome-wide mRNA, microRNA, and three common histone modifications implicated in gene activation for young and aged virgin and mated female D. melanogaster in a factorial design. We identified phenotypically plastic transcripts and epigenetic modifications associated with mating and aging. We used these data to derive phenotypically plastic regulatory networks associated with mating of young flies, and aging of virgin and mated flies. Many of the mRNAs, microRNAs and epigenetic modifications associated with mating of young flies also occur with age in virgin flies, which may reflect mating-induced accelerated aging. We functionally tested the plastic regulatory networks by overexpressing environmentally sensitive microRNAs. Overexpression resulted in altered expression of ~70% of candidate target genes, and in all cases affected oviposition. Conclusions: Our results implicate microRNAs as mediators of phenotypic plasticity associated with mating and provide a comprehensive documentation of the genomic and epigenomic changes that accompany mating- and aging-induced physiological changes in female D. melanogaster. Keywords: microRNA, Histone modification, ChipSeq, Environmental plasticity, Systems genetics Most previous studies of genome wide transcriptional responses to environmental perturbations have utilized microarray technology, which has a lower dynamic range than direct sequencing of mRNAs and can be influenced by DNA polymorphisms between the target genotype and the reference genotype used to design the array. We used whole genome RNA sequencing to alleviate these limitations and included a comprehensive analysis of microRNAs (miRNAs). MicroRNAs bind to the 3′ un- translated regions of mRNA, preventing its translation, polyadenylation or stability, and play essential roles in signaling pathways that regulate development and differ- entiation [6,7]. Although miRNAs have been implicated in lifespan determination in Caenorhabditis elegans and Drosophila [8,9], little is known regarding their role in phenotypic plasticity. Finally, epigenetic modifications have been proposed as an important aspect of phenotypic Transcriptional and epigenetic responses to mating and aging in Drosophila melanogaster Shanshan Zhou, Trudy FC Mackay and Robert RH Anholt* © 2014 Zhou et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. mRNA plasticity We obtained mRNA expression profiles for each of the four physiological conditions by RNA sequencing. An average 95.7% of sequences had quality scores higher than 37.65, with over 65× coverage of the estimated 30 Mb Drosophila transcriptome. We detected 26,151 mRNA sequences, including 15,610 annotated transcripts and 10,541 previously unannotated transcripts. Some of the unannotated transcripts could be artifacts of mapping short reads corresponding to premature unspliced mRNAs back to the genome; others represent novel intergenic long non-coding RNAs. We identified 647 unique genes that were differentially expressed post-mating or as a function of age. These include 344 genes with predicted transcripts of unknown function, 23 annotated non-coding tran- scripts and two novel (i.e., previously unannotated) genes. The remaining 278 genes included multiple members of gene families, such as the IM, Jonah, Cp, Cpr, Twdl, Osi, Vm, Mal and Obp families (Additional file 1). Among the differentially expressed genes, 474 showed altered tran- script levels in young flies after mating; 160 showed al- tered transcript levels between young and old virgin flies; and 279 genes showed altered expression between young and old mated flies (Additional file 1). There was modest overlap among these categories (Additional file 2: Figure S1). Only 18 differentially expressed tran- scripts were identified between aged virgin and mated flies. This suggests that the largest post-mating differ- ences in transcript profiles are observed in young females and these differences fade as flies age. Here, we utilized next generation sequencing technology to assess the plastic responses of mRNAs, miRNAs and three major histone modifications (H3K4me1, H3K4me3, and H3K9ac) to mating and aging in D. melanogaster females, in a factorial design. Mating and aging have previously been shown to elicit transcriptional re- sponses in Drosophila [4,9,13-20]. During mating, male D. melanogaster transfer a complex mixture of accessory gland derived peptides and proteins together with sperm into the female reproductive tract. These peptides exert profound physiological and behavioral changes in the fe- male which are thought to enhance the male’s competitive reproductive success [21-23]. They activate oocyte devel- opment and accelerate egg laying, reduce receptivity of the female to additional matings, and enhance the female’s immune competence [21-23]. In addition, mated females have a reduced lifespan compared to virgin females [4,24]. The effects of several accessory gland peptides have been well characterized, including ovulin [25] and the sex pep- tide [26]. Background Phenotypic plasticity is the phenomenon whereby organ- isms with the same genotype can achieve different phenotypic endpoints in response to changes in external or internal environments. Phenotypic plasticity thus al- lows organisms to respond rapidly to changing environ- mental circumstances. Understanding the genomic basis of phenotypic plasticity can yield insights regarding the underlying genes and genetic networks affecting com- plex phenotypes, since mutations in genes whose expres- sion changes in different environments also affect the phenotype [1-5]. Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Page 2 of 15 molecular response profiles that would otherwise occur later in life. plasticity through their regulation of gene expression [10-12]. However, we know little of the complex relation- ships between genome-wide transcriptional responses to environmental perturbations and their regulation by miR- NAs and epigenetic modifications. mRNA plasticity The latter exerts its effect by binding to recep- tors that are expressed in the reproductive tract and in the central nervous system [26]. While the abundances of many hundreds of transcripts are known to change with age [2,8,9,14-16,20,27], it is not known to what extent these patterns vary between virgin and mated flies and whether the reduced longevity of mated females reflects an acceleration of normal senescence. We used our data to construct interaction networks between phenotypically plastic miRNAs, their target genes and associated histone 3 markers. We functionally validated the effects of miRNAs on oviposition in mated females to demon- strate that miRNAs are instrumental in facilitating post-mating physiological and behavioral changes in female D. melanogaster. We performed Gene Ontology (GO) enrichment ana- lyses for the two treatments with the largest numbers of environmentally responsive transcripts: the plastic re- sponse to mating for young flies (Figure 1a) and the plastic response to aging of virgin flies (Figure 1b). Tran- scripts with altered expression levels following mating show strong enrichment of GO terms associated with egg development and proteolysis, in agreement with a previous study [13]; as well as GO terms associated with bacterial defense and immune responses (Figure 1a). GO categories associated with immune/defense response, egg shell formation and metabolism were enriched in the comparison between young and old virgin flies (Figure 1b), also in agreement with previous studies [14-16]. There is partial concordance between GO categories between the transcript profiles of young flies before and after mating and young and aged virgin flies, indicating that some bio- logical processes undergo specific post-mating or aging- related changes in transcript abundance levels, whereas others reflect post-mating changes that would occur later in the lifespan in the absence of mating. Results We assessed changes in genome wide mRNA, miRNA and epigenetic modifications for 3–5 day old and 4-week old virgin Canton S (B) [28] D. melanogaster females and females which were maintained in the continuous presence of males with multiple opportunities for mat- ing. We refer to the latter as “mated” females. We evalu- ated overall differences in transcriptional and epigenetic responses to mating and to aging, as well as differences in the molecular signatures of plasticity between young and old virgin and mated flies, and mated and virgin young and old flies. This design enables us to assess whether or to what extent mating induces changes in Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Zhou et al. BMC Genomics 2014, 15:927 Page 3 of 15 http://www.biomedcentral.com/1471-2164/15/927 Figure 1 Gene ontology enrichment of transcripts with altered expression as a result of mating or aging. Gene ontology enrichment was assessed for transcripts that changed expression levels in young flies before and after mating (a) and in young vs aged virgin flies (b) Light blue bars exceed the Benjamini corrected threshold of P <0.05. Figure 1 Gene ontology enrichment of transcripts with altered expression as a result of mating or aging. Gene ontology enrichment was assessed for transcripts that changed expression levels in young flies before and after mating (a) and in young vs aged virgin flies (b) Light blue bars exceed the Benjamini corrected threshold of P <0.05. maternal to zygotic transition following fertilization [29,30]. We mapped the 15-18 bp RNA fragments across all samples to the genome sequence and identified 147 genes that give rise to degradation fragments exclusively in young flies after mating (Additional file 3: Table S1). Among these genes, 64 encode predicted transcripts of unknown function. These genes also comprise genes as- sociated with developmental processes and, intriguingly, four genes encoding chemoreceptors (Gr39b, Gr93d, Or13a, Or45a). miRNA plasticity We extracted and sequenced small RNAs (<200 nt, in- cluding miRNAs) from the same samples used to assess genomic responses of mRNAs to mating and aging. Examination of the distribution profiles of small RNAs showed prominent peaks at 22 bp and 30 bp, corre- sponding to miRNAs and rasiRNAs (the Drosophila equivalent of piRNAs), respectively (Figure 2). A prom- inent peak of small RNA fragments around 15 bp is evi- dent in replicate samples of flies following mating. These fragments are transient and are not observed for older mated flies, nor are they observed in young or old virgin flies. These small mRNA fragments are likely due to degradation of maternal RNA in embryos during the We detected 149 mature miRNAs from 105 precur- sors, of which 41 miRNAs from 39 precursors with 28 conserved seeds showed significant differences across the four conditions (Additional file 4). Similar numbers Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Zhou et al. BMC Genomics 2014, 15:927 Page 4 of 15 Page 4 of 15 http://www.biomedcentral.com/1471-2164/15/927 Figure 2 Size distributions of small RNAs in young (blue) and aged (red) virgin and mated flies. Black and open arrowheads indicate the regions corresponding to miRNAs and rasiRNAs, respectively. The arrows point at small fragments apparent only in young flies after mating that likely reflect mRNA degradation products generated during the maternal to zygotic transition. Figure 2 Size distributions of small RNAs in young (blue) and aged (red) virgin and mated flies. Black and open arrowheads indicate the regions corresponding to miRNAs and rasiRNAs, respectively. The arrows point at small fragments apparent only in young flies after mating that Figure 2 Size distributions of small RNAs in young (blue) and aged (red) virgin and mated flies. Black and open arrowheads indicate the regions corresponding to miRNAs and rasiRNAs, respectively. The arrows point at small fragments apparent only in young flies after mating that likely reflect mRNA degradation products generated during the maternal to zygotic transition. Figure 2 Size distributions of small RNAs in young (blue) and aged (red) virgin and mated flies. Black and open arrowheads indicate the regions corresponding to miRNAs and rasiRNAs, respectively. The arrows point at small fragments apparent only in young flies after mating that likely reflect mRNA degradation products generated during the maternal to zygotic transition. loop configurations with similar conserved seed regions [31,32]. miRNA plasticity Zygotic expression of these miRNAs has been associated with degradation of maternal mRNA in em- bryos [29,32]. Indeed, these miRNAs are only expressed in mated flies and are virtually absent in virgins (Figure 3). Among the 147 genes that are uniquely im- plicated in maternal degradation, 45 contain 21 con- served binding sites for miRNAs that are differentially expressed after mating, and 18 genes are targets of the miRNA 309 cluster. of miRNAs show plastic responses to the four treat- ments. We find 27 differentially regulated miRNAs after mating in young flies and 22 in old flies; and 17 miR- NAs differentially regulated with aging in virgin flies and 28 in mated flies, with considerable overlap in plas- tic responses of miRNAs among the four physiological conditions (Additional file 5: Figure S2). Of particular interest is the miRNA 309 cluster, which, in addition to mir-309, comprises mir-286, mir-3, mir-4, mir-5 and mir-6. The latter gives rise to three alternative stem- Page 5 of 15 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Page 5 of 15 Zhou et al. BMC Genomics 2014, 15:927 Figure 3 Changes in mir-309 cluster expression after mating in young (blue) and old (red) flies. induce further chromatin remodeling [11,12,34,35]. We examined three common histone 3 modifications (H3K4me1, H3K4me3, H3K9ac) previously implicated in gene activation [36]. We used immuno-coprecipitation to identify DNA segments differentially associated with each of the his- tone 3 modifications across the genome. We identified 3,484, 5,467, and 4,986 peaks of co-precipitated DNA fragments for H3K4me1, H3K4me3 and H3K9ac, re- spectively; of which 710 (H3K4me1), 1,513 (H3K4me3) and 1,819 (H3K9ac) co-precipitated fragments were differentially expressed among the four physiological conditions (Additional file 6: Figure S3). We found similar numbers of up- and down-regulated genes for each variable histone modification (Table 1; Additional files 7, 8 and 9). The number of histone modifications associated with aging was larger than those associated with post-mating changes, especially for H3K4me3 (Table 1). We identified candidate histone regulated genes associated with co-precipitated fragments by applying a 2 kb window downstream and upstream of each gene. On average, 3.3 candidate genes were asso- ciated with each histone peak (Table 1), of which ~24% are novel transcripts. miRNA plasticity About 10% of differentially modified fragments after mating in young flies were shared between H3K4me1 and H3K4me3, and between H3K4me1 and H3K9ac, while about 15% were shared among H3K4me3 and H3K9ac, and only about 1% of these fragments were asso- ciated with all three modified histone marks (Additional file 10: Figure S4). Similarly, about 17% of differentially modified fragments identified as a function of aging of vir- gin females were shared between H3K4me1 and H3K4me3, and between H3K4me1 and H3K9ac, while as much as 36% of precipitated DNA sequences were shared among H3K4me3 and H3K9ac, and about 5% of se- quences were associated with all three modified his- tone marks (Additional file 10: Figure S4). Figure 3 Changes in mir-309 cluster expression after mating in young (blue) and old (red) flies. We asked to what extent histone modifications in- duced by mating are permanent throughout the lifespan of reproductively active females. We found that 18%, 9% and 11% of H3K4me1, H3K4me3 and H3K9ac modifica- tions, respectively, that occurred post-mating in young flies persisted at 4 weeks of age when females were maintained in the presence of males. Next, we asked how many of these histone modifications also occur as a result of aging in the absence of mating. We found that 65%, 73% and 77% of these H3K4me1, H3K4me3 and H3K9ac modifications, respectively, ultimately occur in virgin flies, but only at later age. These markers, there- fore, represent epigenetic mating-induced accelerated aging changes and are associated with up to 539 unique candidate genes, of which 421 are annotated (Additional file 11). These include genes associated Epigenetic plasticity Epigenetic regulation in D. melanogaster occurs primarily through association of modified histones with target DNA sequences [33]. Numerous post-translational modifications of histone proteins have been identified, including methyla- tion, acetylation, phosphorylation, ubiquitylation, SUMOy- lation, biotinylation and ADP-ribosylation [11,34,35]. Some histone modifications act in cis to change chromatin structure directly, whereas others act in trans to affect the recruitment of a protein complex that regulates gene expression [36]. Modified histone protein residues can serve as docking sites for transcription factors and other chromatin-modifying enzymes to regulate transcription or Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Page 6 of 15 Table 1 Number of genes associated with changes in histone modifications* Histone modification Post-mating changes in young flies Post-mating changes in aged flies Changes associated with aging in virgin flies Changes associated with aging in mated flies Histone peaks Genes Histone peaks Genes Histone peaks Genes Histone peaks Genes H3K4me1 Up-regulated 178 536 259 847 253 750 284 911 Down-regulated 205 677 249 772 291 971 299 1002 H3K4me3 Up-regulated 281 959 530 1899 611 2000 703 2482 Down-regulated 383 1293 482 1708 693 2382 710 2534 H3K9ac Up-regulated 512 1618 427 1427 816 2376 806 2598 Down-regulated 536 1680 458 1428 823 2553 834 2755 Differences between conditions were estimated by subtracting histone peak amplitudes in virgin flies from those in mated flies and in young flies from those in aged flies. Only histone modification peaks that are significantly variable across the four conditions (FDR <0.05) with differences that are larger than one pooled standard deviation are counted. Table 1 Number of genes associated with changes in histone modifications Differences between conditions were estimated by subtracting histone peak amplitudes in virgin flies from those in mated flies and in young flies from those in aged flies. Only histone modification peaks that are significantly variable across the four conditions (FDR <0.05) with differences that are larger than one pooled standard deviation are counted. with oocyte differentiation, including stonewall, 14-3-3zeta, bazooka, missing oocyte, and egalitarian. Histone modifica- tions that affect these candidate genes may reflect persistent post-mating stimulation of oogenesis. of mating or aging, they illustrate that different transcripts interact with histone 3 upon mating and as a function of age. Epigenetic plasticity Among genes with altered transcriptional regulation that interact with modified histone 3, only a few are asso- ciated with more than one histone mark, indicating target specific interactions for each histone modification with these genes (Table 2). To assess to what extent environmentally plastic tran- scripts might be causally associated with epigenetic regulation we asked which putative histone regulated genes showed altered transcriptional regulation across the four experimental conditions (Table 2). The majority of genes with altered transcriptional regulation corre- sponded with candidate target genes for the three his- tone marks when young virgin flies are compared to young mated flies and when young mated flies are com- pared to aged mated flies (Table 2). Many fewer genes with altered transcriptional regulation corresponded with candidate target genes when young virgin flies were compared to aged virgins and when old virgin flies were compared to old mated flies (Table 2). Although these proportions roughly parallel differences in the numbers of transcripts with altered abundance levels as a consequence Numbers in the table represent genes associated with histone marks. Up-regulated and Down-regulated refer to the histone marks detected by immuno co-precipitation. Differences between conditions were estimated by subtracting histone peak amplitudes in virgin flies from those in mated flies and in young flies from those in aged flies Phenotypically plastic regulatory networks themselves potential targets for H3K4me3. It is reason- able to infer from the network structure that regula- tory modifications at a limited number of focal genes can affect the expression of a suite of interacting partners. The network associated with post-mating changes in young flies comprised 46 target genes for 20 miRNAs, and these target genes have 92 interacting partners (Figure 4; Additional file 12). In addition to genes that encode predicted transcripts of unknown function, de- velopmental genes feature prominently among the latter. The network reveals several notable features. First, there are shared targets for multiple miRNAs. Second, these target genes interact with multiple partners and each interacting partner interacts with multiple miRNA tar- gets, thus forming a collage of interacting ensembles that are embedded in the network. Third, 13% of miRNA target genes and 37% of their interacting partners are targets for regulation through histone modifications. In addition, two miRNAs, mir-193 and mir-1000, are A similar analysis of transcriptional changes between young and old virgin flies gives rise to a smaller network, comprised of only seven miRNAs, five target genes, and 14 genes interacting with these targets (Figure 5; Additional file 12). Of the latter, 10 are potential targets for regulation by modified histone 3, as are two of the miRNA target genes. Among the seven miRNAs, five overlap with the network associated with post-mating changes, including mir-193. It should be noted that in this case the compos- ition of differentially expressed miRNA target genes re- stricts connectivity, which limits the size of the network. Nevertheless, it is clear that the genetic architectures that underlie changes in gene expression are distinct with Figure 4 A network of miRNAs, their target genes, and potential interacting partners associated with post-mating changes. miRNAs that change expression in young flies after mating are shown in the inner circle in yellow boxes and are connected to their target genes, shown in the middle circle, with red lines. The miRNA target genes are connected with known interaction partners, shown in the outer circle, with grey lines. Genes that are targets for H3K4me1, H3K4me3 and H3K9ac, are shown in orange, green and red boxes, respectively. Targets for two histone 3 marks are indicated with an additional border color. Phenotypically plastic regulatory networks We identified target genes that undergo changes in ex- pression for all miRNAs with altered transcriptional regu- lation, and recruited genes known to interact with these target genes from the Drosophila Interactions Database [37] to obtain integrative networks that provide a compre- hensive visualization of post-mating or aging-associated changes in whole genome transcriptional profiles, and superimposed histone marks on the networks. We only in- cluded genes with more than one interaction with miRNA target genes. Since very few transcripts were altered in the comparison of old virgin and mated flies, we derived these Table 2 Numbers of candidate histone regulated genes with altered expression after mating or upon aging* Comparison H3K4me1 H3K4me3 H3K9ac Up-regulated Down-regulated Up-regulated Down-regulated Up-regulated Down-regulated Young flies Virgin vs. Mated 13 13 12 15 19 20 Virgin flies Young vs. Aged 4 3 4 5 2 4 Mated flies Young vs. Aged 5 6 24 18 16 11 Aged flies Virgin vs. Mated 1 0 0 0 0 0 Numbers in the table represent genes associated with histone marks. Up-regulated and Down-regulated refer to the histone marks detected by immuno co-precipitation. Differences between conditions were estimated by subtracting histone peak amplitudes in virgin flies from those in mated flies and in young flies from those in aged flies. Table 2 Numbers of candidate histone regulated genes with altered expression after mating o Comparison H3K4me1 H3K4me3 f candidate histone regulated genes with altered expression after mating or upon aging Numbers of candidate histone regulated genes with altered expression after mating or upon aging on H3K4me1 H3K4me3 H3K9ac Numbers in the table represent genes associated with histone marks. Up-regulated and Down-regulated refer to the histone marks detected by immuno co-precipitation. Differences between conditions were estimated by subtracting histone peak amplitudes in virgin flies from those in mated flies and in young flies from those in aged flies. Numbers in the table represent genes associated with histone marks. Up-regulated and Down-regulated refer to the histone marks detected by immuno co-precipitation. Differences between conditions were estimated by subtracting histone peak amplitudes in virgin flies from those in mated flies and in young flies from those in aged flies. Page 7 of 15 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Page 7 of 15 plastic regulatory networks for three conditions: young virgin and mated flies, young and old mated flies, and young and old virgin flies. Phenotypically plastic regulatory networks Genes that are targets for H3K4me1, H3K4me3 and H3K9ac, are shown in orange, green and red boxes, respectively. Myb is a target for both H3K4me3 and H3K9ac, as indicated by the red border around the green box. Note that only miRNAs and target genes that undergo altered transcriptional regulation are included in the network, and only genes interacting with more than one miRNA target are represented in the outer circle. regard to post-mating changes and aging, as might be expected. that no viable offspring could be obtained when mir-281-1 and mir-310 were overexpressed. The remaining overex- pression lines generated F1 progeny, in which we used quantitative RT-PCR to monitor expression of target genes in virgin and mated flies contemporaneously with F1 progeny obtained by crossing the control progenitor strain to the ubiquitin-GAL4 driver line. We assessed whether changes in their corresponding target genes that were differentially expressed between young virgin and mated females were significantly different after mating in control lines compared to the miRNA over- expression lines (Student’s t-test, P <0.05). Despite differences in genetic backgrounds between the GAL4- UAS offspring and the original Canton S (B) line used to derive the networks, we observed significant effects for 13 of the 20 target genes tested (70%; Table 3). Thus, perturbation of miRNAs results in disruption of target gene expression that occurs during changes in physiological state and directly implicates miRNAs in The network associated with aging in mated flies comprises 15 miRNAs, 27 miRNA target genes and 64 interacting partners (Figure 6; Additional file 12). Two miRNAs, mir-193 and mir-279 are potential targets of H3K4me3 and H3K4me1, respectively. Eight miRNA tar- get genes (~29%) and 39 (~60%) interacting partners are potentially regulated by histone modifications. Phenotypically plastic regulatory networks Note that only miRNAs and target genes that undergo altered transcriptional regulation are included in the network, and only genes interacting with more than one miRNA target are represented in the outer circle. Figure 4 A network of miRNAs, their target genes, and potential interacting partners associated with post-mating changes. miRNAs that change expression in young flies after mating are shown in the inner circle in yellow boxes and are connected to their target genes, shown in the middle circle, with red lines. The miRNA target genes are connected with known interaction partners, shown in the outer circle, with grey lines. Genes that are targets for H3K4me1, H3K4me3 and H3K9ac, are shown in orange, green and red boxes, respectively. Targets for two histone 3 marks are indicated with an additional border color. Note that only miRNAs and target genes that undergo altered transcriptional regulation are included in the network, and only genes interacting with more than one miRNA target are represented in the outer circle. Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Zhou et al. BMC Genomics 2014, 15:927 Page 8 of 15 http://www.biomedcentral.com/1471-2164/15/927 Figure 5 A network of miRNAs, their target genes, and potential interacting partners associated with aging. miRNAs that change expression upon aging of virgin flies are shown in the inner circle in yellow boxes and are connected to their target genes, shown in the middle circle, with red lines. The miRNA target genes are connected with known interaction partners, shown in the outer circle, with grey lines. Genes that are targets for H3K4me1, H3K4me3 and H3K9ac, are shown in orange, green and red boxes, respectively. Myb is a target for both H3K4me3 and H3K9ac, as indicated by the red border around the green box. Note that only miRNAs and target genes that undergo altered transcriptional regulation are included in the network, and only genes interacting with more than one miRNA target are represented in the outer circle. Figure 5 A network of miRNAs, their target genes, and potential interacting partners associated with aging. miRNAs that change expression upon aging of virgin flies are shown in the inner circle in yellow boxes and are connected to their target genes, shown in the middle circle, with red lines. The miRNA target genes are connected with known interaction partners, shown in the outer circle, with grey lines. Effects of micro-RNA overexpression on target gene expression To test the connectivity of the computational networks (Figures 4, 5 and 6) and evaluate causality, we overex- pressed six miRNAs for which overexpression lines were available (mir-281-1, mir-286, mir-34, mir-92b, mir-310 and mir-988) using the binary GAL4-UAS expression system [38] under a universal ubiquitin promoter. Evidence that the driver was effective comes from the observation Zhou et al. BMC Genomics 2014, 15:927 http://www biomedcentral com/1471-216 Zhou et al. BMC Genomics 2014, 15:927 Page 9 of 15 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 http://www.biomedcentral.com/1471-2164/15/927 Figure 6 miRNAs, their target genes, and potential interacting partners associated with aging and mating. miRNAs that change expression after aging are shown in the inner circle in yellow boxes and are connected to their target genes, shown in the middle circle, with red lines. The miRNA target genes are connected with known interaction partners, shown in the outer circle, with grey lines. Genes that are targets for H3K4me1, H3K4me3 and H3K9ac, are shown in orange, green and red boxes, respectively. Targets for two histone 3 marks are indicated with an additional border color. Note that only miRNAs and target genes that undergo altered transcriptional regulation are included in the network, and only genes interacting with more than one miRNA target are represented in the outer circle. Figure 6 miRNAs, their target genes, and potential interacting partners associated with aging and mating. miRNAs that change expression after aging are shown in the inner circle in yellow boxes and are connected to their target genes, shown in the middle circle, with red lines. The miRNA target genes are connected with known interaction partners, shown in the outer circle, with grey lines. Genes that are targets for H3K4me1, H3K4me3 and H3K9ac, are shown in orange, green and red boxes, respectively. Targets for two histone 3 marks are indicated with an additional border color. Note that only miRNAs and target genes that undergo altered transcriptional regulation are included in the network, and only genes interacting with more than one miRNA target are represented in the outer circle. expression levels are causally associated with physiological changes. the regulation of post-mating and aging-dependent gene expression. Discussion Overexpression of a miRNA has an effect on the target gene when differential expression of the target gene is only significant (P <0.05) in control females, but not in the miRNA overexpressing females, or vice versa. Figure 7 Effects of miRNA overexpression on egg laying by mated females. Number of eggs laid by five young mated females over 18 hours between control and mir-286, mir-34, mir-92b and mir-988 overexpression lines. Two stars indicate P <0.01 and three stars indicate P <0.0001. Error bar shows standard error. Table 3 Real time PCR confirmation of interactions between miRNAs and target genes* Figure 7 Effects of miRNA overexpression on egg laying by Figure 7 Effects of miRNA overexpression on egg laying by mated females. Number of eggs laid by five young mated females over 18 hours between control and mir-286, mir-34, mir-92b and mir-988 overexpression lines. Two stars indicate P <0.01 and three stars indicate P <0.0001. Error bar shows standard error. transcript profiles occur in young females after mating. Although the genetic architectures that underlie changes in gene expression are distinct with regard to post-mating changes and aging, we found epigenetic modifications that occur post-mating in young flies and at later age in virgin flies, consistent with mating-induced accelerated aging and consequent shortening of the lifespan of mated females [24] (Additional file 3). Student’s t-tests were used to calculate P-values of significant differences between expression in virgin and mated flies. Overexpression of a miRNA has an effect on the target gene when differential expression of the target gene is only significant (P <0.05) in control females, but not in the miRNA overexpressing females, or vice versa. Student’s t-tests were used to calculate P-values of significant differences between expression in virgin and mated flies. Overexpression of a miRNA has an effect on the target gene when differential expression of the target gene is only significant (P <0.05) in control females, but not in the miRNA overexpressing females, or vice versa. The power of our analysis is limited by the small num- ber of replicate samples and histone modifications ana- lyzed, and possible differences in genomic responses in different genetic backgrounds. Thus, the complement of genes that undergo altered regulation as a consequence of physiological state and epigenetic modifications that could regulate their expression provides only a partial view of the genetic architectures that regulate physiological state-induced regulation of gene expression. Discussion Effects of micro-RNA overexpression on egg laying Stimulation of oogenesis and oviposition is the most profound change induced in females as a consequence of mating [21]. To assess causality between post-mating modulation of miRNA expression and oviposition, we measured egg laying after mating by control females and females overexpressing mir-286, mir-34, mir-92b, or mir-988. Mating causes a reduction in expression of mir-34, mir-92b and mir-988 and an increase in expres- sion of mir-286 (Figure 3; Additional file 4). Females that overexpress mir-34 and mir-92b lay 34% and 37% fewer eggs after mating, whereas females that overexpress mir-286 and mir-988 lay 16% and 33% more eggs than controls (Figure 7). Thus, post-mating changes in miRNA We identified genomic and epigenomic changes in gene expression that are associated with mating and aging in female D. melanogaster. We found that 26 miRNAs change expression levels after mating, including transi- ent expression of members of the mir-309 cluster associ- ated with the maternal to zygotic transition that results in maternal mRNA breakdown [29,30]. The mir-309 cluster is analogous to mir-430, which mediates the maternal to zygotic transition in zebrafish [39]. Up- regulation of the mir-309 cluster after mating is accompanied by the transient accumulation of small RNA fragments, likely representing breakdown prod- ucts of degraded maternal RNAs in embryos. In D. melanogaster, loci required for the destabilization of Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Page 10 of 15 Table 3 Real time PCR confirmation of interactions between miRNAs and target genes* miRNA Target gene Expression in virgin vs. mated female flies (P-Value) Control cross miRNA overexpression cross Effect on target gene expression mir-286 CG15212 0.498 0.037 + CG4000 0.033 0.064 + dyl 0.241 0.018 + Osi7 0.380 0.006 + Pkcdelta 0.044 0.364 + Vha68-1 0.013 0.348 + mir-34 CG16857 0.034 0.434 + CG34040 0.046 0.036 - dan 0.204 0.367 - Gbeta76c 0.277 0.010 + prc 0.240 0.410 - prm 0.060 0.441 - mir-92b CG31869 0.351 0.008 + CG8303 0.446 0.026 + fln 0.378 0.034 + Tusp 0.075 0.032 + mir-988 CG15571 0.038 0.226 + CG31086 0.036 0.417 + Ela 0.240 0.139 - vkg 0.143 0.058 - *Student’s t-tests were used to calculate P-values of significant differences between expression in virgin and mated flies. Discussion Furthermore, genes associated with histone marks are commonly assigned based on chromosomal location [36,41,42]. Dem- onstrating causal relationships with candidate genes within such chromosomal regions remains challenging, and it is difficult to unambiguously identify false positives or ex- clude long-distance effects of modified histones on gene expression. We alleviated this concern to some extent by focusing on histone modifications associated with co- localized genes that undergo altered expression. We also note that our analyses were performed on whole flies and we cannot exclude that some changes in miRNA expres- sion or histone modifications may be compartmentalized to certain tissues. distinct subsets of maternal mRNAs during the mater- nal to zygotic transition have been mapped to different chromosome arms [40], indicating that different mech- anisms may mediate maternal RNA degeneration in the zygote. The appearance of RNA fragments only in young flies after mating could be due to subsequent changes that may include epigenetic modifications, which increase the efficiency of the removal of such fragments during repeated matings. distinct subsets of maternal mRNAs during the mater- nal to zygotic transition have been mapped to different chromosome arms [40], indicating that different mech- anisms may mediate maternal RNA degeneration in the zygote. The appearance of RNA fragments only in young flies after mating could be due to subsequent changes that may include epigenetic modifications, which increase the efficiency of the removal of such fragments during repeated matings. We identified target genes for miRNAs that change transcript abundance levels after mating. We examined similar changes in genome-wide expression levels of miRNAs and their associated target genes that occur as females age. In addition, we analyzed epigenetic modifi- cations mediated via histone 3 that are associated with mating and aging. We were able to construct integrative genetic networks from a subset of genes with altered expression due to mating or aging to visualize part of the genetic architectures that underlie these physio- logical changes. We found that the largest differences in Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Page 11 of 15 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Previous studies have implicated steroid hormones in regulating expression of miRNAs [44,45]. Odorant bind- ing proteins expressed in the male accessory gland could conceivably serve as carriers for such ligands [19]. How- ever, the precise mechanisms by which signals transmit- ted during mating regulate gene expression remain to be explored. Discussion Nevertheless, this study presents to the best of our knowledge the first demonstration that miRNAs play a critical role in mediating post-mating changes in female flies and a comprehensive documentation of the genomic and epigenomic changes that accompany mating-induced physiological changes as well as aging in female D. melanogaster. A large proportion of genes with changes in transcript abundance are members of gene families. Members of the Jonah, Cp, Twdl, and Mal gene families show antagon- istic changes in transcript abundances after mating and as a result of aging, whereas Vm family members show changes in the same direction. Similarly, some genes with altered expression that belong to gene ontology categories shared between mating and aging conditions also show antagonistic effects. We speculate that these antagonistic changes may be part of the molecular machinery that me- diates trade-offs between reproduction and lifespan [24]. In addition to restricted expression in a limited number of cells, antagonistic regulation in different tissues could also be a limiting factor in detecting genes that undergo altered transcriptional regulation (e.g. we did not detect altered transcript abundance of smaug, which triggers the mater- nal to zygotic degradation [30,43]). Conclusions To gain insight regarding the genes and regulatory net- works underpinning plastic changes after mating and during aging, we identified mRNAs, microRNAs and epigenetic modifications with significantly different ex- pression after mating or as a function of aging. We used these data to derive phenotypically plastic regulatory networks centered on environmentally sensitive micro- RNAs associated with aging and mating, and identified several biomarkers of mating-induced accelerated aging. Overexpression of several plastic microRNAs resulted in altered expression of candidate target genes and affected oviposition. MicroRNAs are thus critical in mediating post-mating changes in gene expression. These data pro- vide comprehensive documentation of the genomic and epigenomic changes that accompany mating- and aging- induced physiological changes in female D. melanogaster. We observed many histone modifications after mating or upon aging that are not accompanied by changes in corresponding transcript abundances. Here, the target gene might be remote from the histone-DNA interaction site. Alternatively, interaction between the histone and the target gene might prevent activation of this gene at a later stage or inactivation of a gene that is actively expressed. Furthermore, the networks centered on miR- NAs and their target genes (Figures 4, 5 and 6) do not include all transcripts with altered expression levels, nor do they include all potential target genes for epigenetic regulation. Nevertheless, the complex structure of these networks raises the possibility that epigenetic regulation of target genes for histone 3 could also indirectly affect the expression of interconnected genes. Drosophila stocks C S Isogenic Canton S (B) flies [28] were reared on cornmeal- molasses-yeast medium at 25°C under a 12 hour light–dark cycle. Adults were collected immediately after eclosion and subjected to two treatments. In the first treatment, adults were allowed to mate at a dens- ity of 25 females and 25 males per vial. In the second treatment, 50 females were placed in each vial and maintained as virgins. Flies were transferred to new vials every 2 days. Female flies were collected at 3–5 days (young flies) or at 4 weeks (aged flies) at the same time of day (1 pm-3 pm) and flash frozen on dry ice. Overexpression of miRNAs confirmed the biological relevance of the derived environmentally plastic regula- tory networks. Despite differences in genetic back- grounds, altered expression of miRNAs correlated with altered expression of ~70% of their candidate target genes, and in all cases miRNA overexpression affected post-mating oviposition. Although we analyzed only a subset of miRNAs for which viable overexpression lines could be obtained, it is reasonable to assume that these observations would extend to the larger ensemble of miRNAs. Methods In addition to epigenetic modifications that uniquely occur after mating, we identified identical epigenetic markers that occur also in virgin females at advanced age. These mark post-mating changes that relate to the shortening of female lifespan after mating, possibly due to the increased energy expenditure required for egg production [21,24]. mRNA and miRNA sequencing Two independent samples of 25 female flies were used for each condition (young virgin flies, young mated flies, aged virgin flies and aged mated flies). RNA was ex- tracted and separated into a small RNA (<200 nts) and large RNA fraction (>200 nts) using the Qiagen miR- Neasy kit. rRNA was depleted from the large RNA frac- tion using Ribo-Zero rRNA Removal Kits (Epicentre, Since this study focused on whole flies, it remains to be determined to what extent the genomic and epige- nomic changes we have observed are compartmentalized to reproductive organs or the central nervous system. Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Page 12 of 15 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Inc.) to enrich for mRNA. We used enriched mRNA to prepare bar-coded libraries. Four mRNA samples were pooled in equal molarity from each replicate and se- quenced in one 50 bp single read Illumina HiSeq2000 lane. Small RNA libraries were prepared using the NEX- flex Small RNA Sequencing kit (Bioo, Inc.) and enriched for miRNA, siRNA and piRNA following gel electro- phoresis. All 8 bar-coded samples were pooled in equal molarities and sequenced in one 50 bp single read Illu- mina HiSeq2000 lane. (FPKM), corrected for background using median FPKM of non-peak intervals in each sample and followed by quan- tile normalization across samples. We used background corrected normalized values of FPKM for each peak for differential modification analysis, as described above, using permutation with one-way ANOVA and FDR to correct for multiple-testing under dependency [51]. We used one pooled standard deviation for post hoc pairwise compari- sons. Raw data can be accessed in the National Center for Biotechnology Information Sequence Read Archive (SRA) under accession number SRP048404. mRNA sequences were aligned and assembled using Bowtie2-2.0.6, Tophat-2.0.7 and Cufflinks-2.0.2 [46,47]. We used Cuffdiff to analyze pairwise differential expres- sion between the conditions of transcripts, genes, spli- cing and promoter uses [48]. We first trimmed small RNA sequences to remove the adapter sequence before aligning them to the Drosophila miRBase [49] using miRExpress [50]. We then used quantile normalized counts for differential analysis implemented by the R package (lmPerm) using permutation with a one-way ANOVA model, Y = μ + C + ε, where μ is the overall mean, C designates condition and ε the error term. We used FDR to correct for multiple-testing under depend- ency [51]. Gene ontology analysis Gene ontology enrichment analyses for transcripts that changed expression after mating or aging was done using DAVID [56] with a cut-off for nominal statistical significance of P < 0.01. mRNA and miRNA sequencing One pooled standard deviation was used for post hoc pairwise comparisons. We identified putative miRNA targets using the TargetScan Fly [52] and MinoTar data bases [53]. Raw data can be accessed in the National Center for Biotechnology Information Sequence Read Archive (SRA) under accession number SRP048388. Confirmation of interactions between miRNAs and target genes We obtained UAS-mir-281-1, UAS-mir-286, UAS-mir-34, UAS-mir-92b, UAS-mir-988, mir-310 and the co-isogenic control lines from the Bloomington Stock Center. Females from each homozygous UAS-mir line and their progenitor control were crossed with males from a ubiquitin-GAL4 driver line (w1118: Ubi-Gal4). We did not obtain viable off- spring when mir-281-1 or mir-310 was overexpressed under the ubiquitin promoter. For the other crosses 4–6 day old virgin and mated F1 females were flash frozen for RNA extraction. We extracted total RNA from 10 female flies of each replicate with three biological replicates of each cross. RNAs were converted to cDNA using the iScript cDNA Synthesis Kit (Bio-Rad, Inc.). We used real time quantitative PCR on an ABI 7900HT qRT-PCR ther- mal cycler with 3 technical replicates for each sample and ABsolute qPCR SYBR Green Mixes (Thermo Scientific, Inc.) to evaluate the difference in expression of miRNA target genes between virgin and mated F1 females in con- trols and between virgin and mated F1 females that over- express miRNAs (Table 3) using the following primers: CCCTGCTCCTTCTCCTTCTT and TGAACTTGAACC CCACCTTG for Mlc1, GGTCATCGTCTGTCCGATCT and CTAACTTCCTGGGCGACAAC for dyl, CACCC CAATTCACTCTGGTT and CCTTCAGCCGAAGATT GAAA for Egfr, GGAGACACAGCTCCGACAAT and ATTCTGGCTACCCGGCTTAT for Vha68-1, TTTCGGT GTCTTGTGTCTGG and AGGCAGTCGCTGTAGATG CT for Osi7, GTCCCTGATCGGCGATAATA and AACC- CAGGAGCTTCTTCAACT for CG15212, TCCGAGTA CTCTCCCTCCAG and TCTTGCTAGTCCAGCCACCT for CG4000, CAGTAGTGGAGCAGGCAGTG and ATG ATAGGCCTCCTCCTTGG for Pkcdelta, GGGATAGG CTGGATTGGATT and CGTCGAGAGTAGCTCCGATT for CG31869, AGTAGTTGCGATGCCAATCC and AG Chromatin immunoprecipitation and ChIP sequencing Two independent samples of 400 female flies were used for each condition. Flies were ground in liquid nitrogen. Samples were then cross-linked with 1% formaldehyde for 10 min and lysed, and chromatin was digested into mono- and di-nucleosomes with micrococcal nuclease. Aliquots of digested chromatin samples were then immunoprecipitated with antibodies against H3K4me1, H3K4me3, and H3K9ac. Antibodies of which the speci- ficities have been ascertained by the modENCODE pro- ject [36,54] were obtained from Abcam, Inc., and Active Motif, Inc. Competing interests Th h d l h The authors declare that they have no competing interests. References 1. Lakowski B, Hekimi S: The genetics of caloric restriction in Caenorhabditis elegans. Proc Natl Acad Sci U S A 1998, 95(22):13091–13096. 1. Lakowski B, Hekimi S: The genetics of caloric restriction in Caenorhabditis elegans. Proc Natl Acad Sci U S A 1998, 95(22):13091–13096. 2. Magwire MM, Yamamoto A, Carbone MA, Roshina NV, Symonenko AV, Pasyukova EG, Morozova TV, Mackay TFC: Quantitative and molecular genetic analyses of mutations increasing Drosophila life span. PLoS Genet 2010, 6(7):e1001037. Additional file 3: Table S3. Genes that generate degraded fragments after mating. Additional file 3: Table S3. Genes that generate degraded fragments after mating. Additional file 4: Changes in miRNA expression levels upon mating and aging. Authors’ contributions SZ, TFCM and RRHA designed the experiments, SZ performed the experiments and analyzed the data, SZ, TFCM and RRHA wrote the manuscript. All authors read and approved the final manuscript. Chromatin immunoprecipitation and ChIP sequencing Genes associated with H3K9ac modification changes. Additional file 11: Histone modifications and associated genes that are response for mating-induced accelerated aging. Table S10. Genes associated with H3K4me1 modification changes. Table S11. Genes associated with H3K4me3 modification changes. Table S12. Genes associated with H3K9ac modification changes. Additional file 12: Genes identified from the DroID database that can interact with miRNA target genes which change expression after mating or upon aging. Table S13. Genes that can interact with miRNA target genes which change expression in young flies after mating. Table S14. Genes that can interact with miRNA target genes which Additional file 12: Genes identified from the DroID database that can interact with miRNA target genes which change expression after mating or upon aging. Table S13. Genes that can interact with miRNA target genes which change expression in young flies after mating. Table S14. Genes that can interact with miRNA target genes which change expression in virgin flies upon aging. Table S15. Genes that can interact with miRNA target genes which change expression in mated flies upon aging. change expression in virgin flies upon aging. Table S15. Genes that can interact with miRNA target genes which change expression in mated flies upon aging. Oviposition We collected 5–7 day old-mated F1 control females and females overexpressing miRNAs under the ubiquitin-GAL4 promoter and counted the number of eggs laid by 5 females during 18 hours (4:00 pm-11:00 am). Ten replicates of 5 F1 females were tested from each cross. We used Dunnett’s test to assess variation of egg laying between the control and miRNA overexpression crosses. Received: 24 July 2014 Accepted: 13 October 2014 Published: 23 October 2014 Additional file 2: Figure S1. Four-way Venn diagram that shows overlap among transcripts that change in abundance level after mating of young flies (green circle) and aged flies (red circle), and as a result of aging in virgin flies (blue circle) and mated flies (yellow circle). Chromatin immunoprecipitation and ChIP sequencing Enriched chromatin extracts were used to prepare bar- coded sequencing libraries. ChIP libraries of the four con- ditions from each replicate were pooled and sequenced in one 50 bp single read Illumina HiSeq2000 lane. ChIP se- quences were aligned to the reference genome using Bowtie2-2.0.6 followed by peak calling using MACS1.4.2 (Model-based Analysis for ChIP-Seq) [47,55]. Peak calls were merged across samples if the peak was present in more than one sample and the peak width bigger than 146 bp. Sequence counts were then summed for each peak, and normalized across samples by calculating Fragment Per Kilobase per Million reads mapped Page 13 of 15 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 GCGCTATAATTTCGATGC for CG8303, TGATCTTG GCAGTGGACTTG and AAGATTCTGCGCAAGAAG GA for Prm, CACTGGCATACCTTTGGTTG and TGG CAGTGACATTGATTCGT for fln, AGAATTGCCTTT TGCCACA and CCGACTACGATTCGAAAAGC for Tusp, ATTGCAGTGTGGTTGTTCCA and CCAGTTT GGTTGGAGACGAT for CG34040, GCGAAAGAGAG TCCGAGAGA and GGTTGTGTGCTGGTCAGTTG for dan, CATCGGCATCAGCATCTATC and GCATAGATG GATGGGTGGAC for CG16867, GTAATGACGGTGG CCTTTTC and CTTTGACCATCTTGGCCACT for Gbeta76C, AGGTCTTGGGCCTAGGTGTT and GAGG CCGATCCTGATGAATA for prc, CAAATCCTGCGCC TCATAGT and CACATCGAGTGCCTTGGATA for CG 15571, GGTGGAGCTGGAGAACTACG and AGGTCT CTCCGTCGACATTC for CG31086, TCACCAGATTC CCGCTTATC and GTAACCACCGGATGATGAGG for Ela and ACTGAGAGCCAGGACTGGAA and CTTCC GGTTTCACAGATGGT for vkg. Gapdh expression (Primers: AGGCGTTTGTGACTTCTGGAA and TCT GGCCGTTGAGCATTTC) in each sample was used to normalize gene expression by mean normalization. flies. Table S5. Changes in H3K4me1 modification between virgin and mated young female flies. Additional file 8: Changes in H3K4me3 modification. Table S6. Changes in H3K4me3 modification between young and aged female flies. Table S7. Changes in H3K4me3 modification between virgin and mated female flies. Additional file 9: Changes in H3K9ac modification. Table S8. (a) Changes in H3K9ac modification between young and aged virgin female flies. (b) Changes in H3K9ac modification between young and aged mated female flies. Table S9. (a) Changes in H3K9ac modification between virgin and mated young female flies. (b) Changes in H3K9ac modification between virgin and mated aged female flies. Additional file 10: Figure S4. Three-way Venn diagrams that show overlap among H3K4me1 (blue circle), H3K4me3 (green circle) and H3K9ac (yellow circle) modifications that change in modification intensity after mating of young flies (a), and aged flies (b), and as a result of aging in virgin flies (c) and mated flies (d). Additional file 11: Histone modifications and associated genes that are response for mating-induced accelerated aging. Table S10. Genes associated with H3K4me1 modification changes. Table S11. Genes associated with H3K4me3 modification changes. Table S12. Additional files We thank Rachel Lyman, Sarah McAdam and Genevieve Evans for technical assistance and Dr. Wen Huang for valuable discussions. Additional file 1: Differences in gene expression levels. Table S1. Differences in gene expression levels between young and aged female flies. Table S2. Differences in gene expression levels between virgin and mated female flies. Received: 24 July 2014 Accepted: 13 October 2014 Published: 23 October 2014 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 9. Liu N, Landreh M, Cao K, Abe M, Hendriks GJ, Kennerdell JR, Zhu Y, Wang LS, Bonini NM: The microRNA miR-34 modulates ageing and neurodegeneration in Drosophila. Nature 2012, 482(7386):519–523. 33. Lyko F, Ramsahoye BH, Jaenisch R: DNA methylation in Drosophila melanogaster. Nature 2000, 408(6812):538–540. 34. Allis CD, Berger SL, Cote J, Dent S, Jenuwien T, Kouzarides T, Pillus L, Reinberg D, Shi Y, Shiekhattar R, Shilatifard A, Workman J, Zhang Y: New nomenclature for chromatin-modifying enzymes. Cell 2007, 131(4):633–636. 10. Duncan EJ, Gluckman PD, Dearden PK: Epigenetics, plasticity, and evolution: How do we link epigenetic change to phenotype? J Exp Zool B Mol Dev Evol 2014, 322(4):208–220. 11. Gardner KE, Allis CD, Strahl BD: Operating on chromatin, a colorful language where context matters. J Mol Biol 2011, 409(1):36–46. 35. Kouzarides T: SnapShot: Histone-modifying enzymes. Cell 2007, 131(4):822. 12. Spencer VA, Davie JR: Role of covalent modifications of histones in regulating gene expression. Gene 1999, 240(1):1–12. 36. Nègre N, Brown CD, Ma L, Bristow CA, Miller SW, Wagner U, Kheradpour P, Eaton ML, Loriaux P, Sealfon R, Li Z, Ishii H, Spokony RF, Chen J, Hwang L, Cheng C, Auburn RP, Davis MB, Domanus M, Shah PK, Morrison CA, Zieba J, Suchy S, Senderowicz L, Victorsen A, Bild NA, Grundstad AJ, Hanley D, MacAlpine DM, Mannervik M, et al: A cis-regulatory map of the Drosophila genome. Nature 2011, 471(7339):527–531. 13. McGraw LA, Clark AG, Wolfner MF: Post-mating gene expression profiles of female Drosophila melanogaster in response to time and to four male accessory gland proteins. Genetics 2008, 179(3):1395–1408. 14. Pletcher SD, Macdonald SJ, Marguerie R, Certa U, Stearns SC, Goldstein DB, Partridge L: Genome-wide transcript profiles in aging and calorically restricted Drosophila melanogaster. Curr Biol 2002, 12(9):712–723. 37. Murali T, Pacifico S, Yu J, Guest S, Roberts GG 3rd, Finley RL Jr: DroID 2011: a comprehensive, integrated resource for protein, transcription factor, RNA and gene interactions for Drosophila. Nucleic Acids Res 2011, 39(Database issue):D736–D743. 15. McCarroll SA, Murphy CT, Zou S, Pletcher SD, Chin CS, Jan YN, Kenyon C, Bargmann CI, Li H: Comparing genomic expression patterns across species identifies shared transcriptional profile in aging. Nat Genet 2004, 36(2):197–204. 38. Brand AH, Perrimon N: Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 1993, 118(2):401–415. 16. Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Lai CQ, Parnell LD, Lyman RF, Ordovas JM, Mackay TFC: Candidate genes affecting Drosophila life span identified by integrating microarray gene expression analysis and QTL mapping. Mech Ageing Dev 2007, 128(3):237–249. 39. Giraldez AJ, Mishima Y, Rihel J, Grocock RJ, Van Dongen S, Inoue K, Enright AJ, Schier AF: Zebrafish MiR-430 promotes deadenylation and clearance of maternal mRNAs. Science 2006, 312(5770):75–79. 40. De Renzis S, Elemento O, Tavazoie S, Wieschaus EF: Unmasking activation of the zygotic genome using chromosomal deletions in the Drosophila embryo. PLoS Biol 2007, 5(5):e117. 17. Smith G, Fang Y, Liu X, Kenny J, Cossins AR, de Oliveira CC, Etges WJ, Ritchie MG: Transcriptome-wide expression variation associated with environmental plasticity and mating success in cactophilic Drosophila mojavensis. Evolution 2013, 67(7):1950–1963. 41. Yin H, Sweeney S, Raha D, Snyder M, Lin H: A high-resolution whole- genome map of key chromatin modifications in the adult Drosophila melanogaster. PLoS Genet 2011, 7(12):e1002380. 18. Dalton JE, Kacheria TS, Knott SR, Lebo MS, Nishitani A, Sanders LE, Stirling EJ, Winbush A, Arbeitman MN: Dynamic, mating-induced gene expression changes in female head and brain tissues of Drosophila melanogaster. BMC Genomics 2010, 11:541. 42. Nagarajan M, Veyrieras JB, de Dieuleveult M, Bottin H, Fehrmann S, Abraham AL, Croze S, Steinmetz LM, Gidrol X, Yvert G: Natural single-nucleosome epi-polymorphisms in yeast. PLoS Genet 2010, 6(4):e1000913. 19. McGraw LA, Gibson G, Clark AG, Wolfner MF: Genes regulated by mating, sperm, or seminal proteins in mated female Drosophila melanogaster. Curr Biol 2004, 14(16):1509–1514. 43. Benoit B, He CH, Zhang F, Votruba SM, Tadros W, Westwood JT, Smibert CA, Lipshitz HD, Theurkauf WE: An essential role for the RNA-binding protein Smaug during the Drosophila maternal-to-zygotic transition. Development 2009, 136(6):923–932. 20. Lai CQ, Leips J, Zou W, Roberts JF, Wollenberg KR, Parnell LD, Zeng ZB, Ordovas JM, Mackay TFC: Speed-mapping quantitative trait loci using microarrays. Nat Methods 2007, 4(10):839–841. 44. Jin H, Kim VN, Hyun S: Conserved microRNA miR-8 controls body size in response to steroid signaling in Drosophila. Genes Dev 2012, 26(13):1427–1432. 21. Wolfner MF: Battle and ballet: molecular interactions between the sexes in Drosophila. J Hered 2009, 100(4):399–410. 22. Chapman T: Seminal fluid-mediated fitness traits in Drosophila. Heredity (Edinb) 2001, 87(Pt 5):511–521. 45. Kucherenko MM, Barth J, Fiala A, Shcherbata HR: Steroid-induced microRNA let-7 acts as a spatio-temporal code for neuronal cell fate in the developing Drosophila brain. EMBO J 2012, 31(24):4511–4523. 23. Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Wolfner MF: The gifts that keep on giving: physiological functions and evolutionary dynamics of male seminal proteins in Drosophila. Heredity (Edinb) 2002, 88(2):85–93. 46. Kim D, Pertea G, Trapnell C, Pimentel H, Kelley R, Salzberg SL: TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions. Genome Biol 2013, 14(4):R36. 24. Chapman T, Liddle LF, Kalb JM, Wolfner MF, Partridge L: Cost of mating in Drosophila melanogaster females is mediated by male accessory gland products. Nature 1995, 373(6511):241–244. 47. Langmead B, Salzberg SL: Fast gapped-read alignment with Bowtie 2. Nat Methods 2012, 9(4):357–359. 25. Rubinstein CD, Wolfner MF: Drosophila seminal protein ovulin mediates ovulation through female octopamine neuronal signaling. Proc Natl Acad Sci U S A 2013, 110(43):17420–17425. 48. Trapnell C, Roberts A, Goff L, Pertea G, Kim D, Kelley DR, Pimentel H, Salzberg SL, Rinn JL, Pachter L: Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nat Protoc 2012, 7(3):562–578. 26. Yapici N, Kim YJ, Ribeiro C, Dickson BJ: A receptor that mediates the post-mating switch in Drosophila reproductive behaviour. Nature 2008, 451(7174):33–37. 49. Griffiths-Jones S, Grocock RJ, van Dongen S, Bateman A, Enright AJ: miRBase: microRNA sequences, targets and gene nomenclature. Nucleic Acids Res 2006, 34:D140–D144. 27. Wilson RH, Lai CQ, Lyman RF, Mackay TFC: Genomic response to selection for postponed senescence in Drosophila. Mech Ageing Dev 2013, 134(3–4):79–88. 50. Wang WC, Lin FM, Chang WC, Lin KY, Huang HD, Lin NS: miRExpress: Analyzing high-throughput sequencing data for profiling microRNA expression. BMC Bioinformatics 2009, 10:328. 28. Norga KK, Gurganus MC, Dilda CL, Yamamoto A, Lyman RF, Patel PH, Rubin GM, Hoskins RA, Mackay TFC, Bellen HJ: Quantitative analysis of bristle number in Drosophila mutants identifies genes involved in neural development. Curr Biol 2003, 13(16):1388–1396. 51. Benjamini Y, Drai D, Elmer G, Kafkafi N, Golani I: Controlling the false discovery rate in behavior genetics research. Behav Brain Res 2001, 125(1–2):279–284. 29. Barckmann B, Simonelig M: Control of maternal mRNA stability in germ cells and early embryos. BBA-Gene Regul Mech 2013, 1829(6–7):714–724. 52. Kheradpour P, Stark A, Roy S, Kellis M: Reliable prediction of regulator targets using 12 Drosophila genomes. Genome Res 2007, 17(12):1919–1931. 30. Tadros W, Lipshitz HD: The maternal-to-zygotic transition: a play in two acts. Development 2009, 136(18):3033–3042. 53. Additional file 4: Changes in miRNA expression levels upon mating and aging. 3. Weindruch R, Kayo T, Lee CK, Prolla TA: Microarray profiling of gene expression in aging and its alteration by caloric restriction in mice. J Nutr 2001, 131(3):918S–923S. Additional file 5: Figure S2. Four-way Venn diagram that shows overlap among miRNAs that change in abundance level after mating of young flies (green circle) and aged flies (red circle), and as a result of aging in virgin flies (blue circle) and mated flies (yellow circle). Additional file 5: Figure S2. Four-way Venn diagram that shows overlap among miRNAs that change in abundance level after mating of young flies (green circle) and aged flies (red circle), and as a result of aging in virgin flies (blue circle) and mated flies (yellow circle). 4. Zhou S, Campbell TG, Stone EA, Mackay TFC, Anholt RRH: Phenotypic plasticity of the Drosophila transcriptome. PLoS Genet 2012, 8(3):e1002593. 4. Zhou S, Campbell TG, Stone EA, Mackay TFC, Anholt RRH: Phenotypic plasticity of the Drosophila transcriptome. PLoS Genet 2012, 8(3):e1002593. 5. Sambandan D, Carbone MA, Anholt RRH, Mackay TFC: Phenotypic plasticity and genotype by environment interaction for olfactory behavior in 4. Zhou S, Campbell TG, Stone EA, Mackay TFC, Anholt RRH: Phenotypic plasticity of the Drosophila transcriptome. PLoS Genet 2012, 8(3):e1002593. 5. Sambandan D, Carbone MA, Anholt RRH, Mackay TFC: Phenotypic plasticity and genotype by environment interaction for olfactory behavior in Drosophila melanogaster. Genetics 2008, 179(2):1079–1088. p y p p 5. Sambandan D, Carbone MA, Anholt RRH, Mackay TFC: Phenotypic plasticity and genotype by environment interaction for olfactory behavior in Drosophila melanogaster. Genetics 2008, 179(2):1079–1088. Additional file 6: Figure S3. Four-way Venn diagrams that show overlap among histone modifications that change in modification intensity after mating of young flies (green circle) and aged flies (red circle), and as a result of aging in virgin flies (blue circle) and mated flies (yellow circle). (a) H3K4me1 modification. (b) H3K4me3 modification. (c) H3K9ac modification. g 6. Bourc’his D, Voinnet O: A small-RNA perspective on gametogenesis, fertilization, and early zygotic development. Science 2010, 330(6004):617–622. 7. Bushati N, Cohen SM: MicroRNA functions. Annu Rev Cell Dev Biol 2007, 23:175–205. Additional file 7: Changes in H3K4me1 modification. Table S4. Changes in H3K4me1 modification between young and aged female Additional file 7: Changes in H3K4me1 modification. Table S4. Changes in H3K4me1 modification between young and aged female 8. Kato M, Slack FJ: Ageing and the small, non-coding RNA world. Ageing Res Rev 2013, 12(1):429–435. Additional file 4: Changes in miRNA expression levels upon mating and aging. Page 14 of 15 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Schnall-Levin M, Zhao Y, Perrimon N, Berger B: Conserved microRNA targeting in Drosophila is as widespread in coding regions as in 3ʹUTRs. Proc Natl Acad Sci U S A 2010, 107(36):15751–15756. 31. Ruby JG, Stark A, Johnston WK, Kellis M, Bartel DP, Lai EC: Evolution, biogenesis, expression, and target predictions of a substantially expanded set of Drosophila microRNAs. Genome Res 2007, 17(12):1850–1864. 54. Egelhofer TA, Minoda A, Klugman S, Lee K, Kolasinska-Zwierz P, Alekseyenko AA, Cheung MS, Day DS, Gadel S, Gorchakov AA, Gu T, Kharchenko PV, Kuan S, Latorre I, Linder-Basso D, Luu Y, Ngo Q, Perry M, Rechtsteiner A, Riddle NC, Schwartz YB, Shanower GA, Vielle A, Ahringer J, Elgin SC, Kuroda MI, 32. Bushati N, Stark A, Brennecke J, Cohen SM: Temporal reciprocity of miRNAs and their targets during the maternal-to-zygotic transition in Drosophila. Curr Biol 2008, 18(7):501–506. Page 15 of 15 Page 15 of 15 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Zhou et al. BMC Genomics 2014, 15:927 http://www.biomedcentral.com/1471-2164/15/927 Pirrotta V, Ren B, Strome S, Park PJ, et al: An assessment of histone-modification antibody quality. Nat Struct Mol Biol 2011, 18(1):91–93. 55. Feng J, Liu T, Qin B, Zhang Y, Liu XS: Identifying ChIP-seq enrichment using MACS. Nat Protoc 2012, 7(9):1728–1740. 56. da Huang W, Sherman BT, Lempicki RA: Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 2009, 4(1):44–57. doi:10.1186/1471-2164-15-927 Cite this article as: Zhou et al.: Transcriptional and epigenetic responses to mating and aging in Drosophila melanogaster. BMC Genomics 2014 15:927. doi:10.1186/1471-2164-15-927 Cite this article as: Zhou et al.: Transcriptional and epigenetic responses to mating and aging in Drosophila melanogaster. BMC Genomics 2014 15:927. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit
https://openalex.org/W2950618289	https://www.biorxiv.org/content/biorxiv/early/2017/05/05/134650.full.pdf	English	null	How to normalize metatranscriptomic count data for differential expression analysis	bioRxiv (Cold Spring Harbor Laboratory)	2,017	cc-by	15,626	BACKGROUND Differential expression analysis on the basis of RNA-Seq count data has become a standard tool in transcriptomics. Several studies have shown that prior normalization of the data is crucial for a reliable detection of transcriptional differences. Until now it is not clear whether and how the transcriptomic approach can be used for differential expression analysis in metatranscriptomics. The potential side effects that may result from direct application of transcriptomic tools to metatranscriptomic count data have not been studied so far. How to normalize metatranscriptomic count 1 data for differential expression analysis 2 Heiner Klingenberg1 and Peter Meinicke1 3 1Abteilung f¨ur Bioinformatik Institut f¨ur Mikrobiologie und Genetik Universit¨at Corresponding author: 6 Peter Meinicke1 7 Email address: peter@gobics.de 8 METHODS We propose a model for differential expression in metatranscriptomics that explicitly accounts for variations in the taxonomic composition of transcripts across different samples. As a main consequence the correct normalization of metatranscriptomic count data requires the taxonomic separation of the data into organism-speciﬁc bins. Then the taxon-speciﬁc scaling of organism proﬁles yields a valid normalization and allows to recombine the scaled proﬁles into a metatranscriptomic count matrix. This matrix can then be analyzed with statistical tools for transcriptomic count data. For taxon-speciﬁc scaling and recombination of scaled counts we provide a simple R script. RESULTS When applying transcriptomic tools for differential expression analysis directly to metatranscriptomic data the organism-independent (global) scaling of counts implies a high risk of falsely predicted functional differences. In simulation studies we show that incorrect normalization not only tends to loose signiﬁcant differences but especially can produce a large number of false positives. In contrast, taxon-speciﬁc scaling can equalize the variation of relative library sizes from different organisms and therefore shows a reliable detection of signiﬁcant differences in all simulations. On real metatranscriptomic data the results from taxon-speciﬁc and global scaling can largely differ. In our study, global scaling shows a high number of extra predictions which are not supported by single transcriptome analyses. Inspection of the scaling error suggests that these extra predictions may actually correspond to artifacts of an incorrect normalization. . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint How to normalize metatranscriptomic count 1 data for differential expression analysis 2 Heiner Klingenberg1 and Peter Meinicke1 3 1Abteilung f¨ur Bioinformatik,Institut f¨ur Mikrobiologie und Genetik,Universit¨at 4 G¨ottingen, Goldschmidtstr. 1, 37077 G¨ottingen, Germany 5 Corresponding author: 6 Peter Meinicke1 7 Email address: peter@gobics.de 8 BACKGROUND 42 Metagenome analysis can provide a comprehensive view on the metabolic potential of a microbial 43 community (Eisen, 2007; Simon and Daniel, 2009). In addition to the static functional proﬁle of the 44 metagenome, metatranscriptomic RNA sequencing (RNA-Seq) can highlight the multi-organism dynamics 45 in terms of the corresponding expression proﬁles (Poretsky et al., 2005; Frias-Lopez et al., 2008; Gilbert 46 et al., 2008; Urich et al., 2008). In particular, metatranscriptomics makes it possible to investigate the 47 functional response of the community to environmental changes (Gilbert et al., 2008; Poretsky et al., 48 2009). 49 In single organism transciptome studies, differential expression analysis based on RNA-Seq data 50 has become an established tool (Marioni et al., 2008; Trapnell et al., 2012). For the analysis, ﬁrst, 51 quality-checked sequence reads are mapped to the organisms genome for transcript identiﬁcation. Then 52 the transcript counts are compared between different experimental conditions to identify statistically 53 signiﬁcant differences. Several studies have shown that read count normalization has a great impact on 54 the detection of signiﬁcant differences (Bullard et al., 2010; Dillies et al., 2013; Lin et al., 2016). The 55 aim of the count normalization is to make the expression levels comparable across different samples 56 and conditions. This is an essential prerequisite for distinguishing condition-dependent differences from 57 spurious variation of expression levels. 58 In metatranscriptomics, already the transcript identiﬁcation step can be challenging. In many cases, 59 RNA-Seq faces a mixture of organisms for which no reference genome sequence is available. Several 60 strategies have been suggested: de novo transcriptome assembly combined with successive homology- 61 based annotation (Celaj et al., 2014), the direct functional annotation of reads by classiﬁcation according 62 to some protein database (Huson et al., 2011; Nacke et al., 2014; Hesse et al., 2015) or parallel sequencing 63 of the corresponding metagenome with successive mapping of RNA-Seq reads to assembled and annotated 64 contigs (Mason et al., 2012; Franzosa et al., 2014; Ye and Tang, 2016). For the subsequent comparison 65 of counts between different conditions no standard protocol exists for differential expression analysis 66 on metatranscriptomic data. Several studies and tools apply methods that have been developed for 67 differential expression analysis in transcriptomics to metatranscriptomic count data (McNulty et al., 2013; 68 Martinez et al., 2016; Macklaim et al., 2013). CONCLUSIONS As in transcriptomics, a proper normalization of count data is also essential for differential expression analysis in metatranscriptomics. Our model implies a taxon-speciﬁc scaling of counts for normalization of the data. The application of taxon-speciﬁc scaling consequently removes taxonomic composition variations from functional proﬁles and therefore effectively prevents the risk of false predictions due to incorrect normalization. 7 8 9 0 1 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint BACKGROUND 42 However, the question under which conditions established 69 models from single organism transcriptomics also apply to organism communities has not been addressed 70 sufﬁciently so far. 71 Here we present an extended statistical model for count data from metatranscriptomic RNA-Seq 72 experiments. Theoretical considerations as well as studies on simulated and real count data show that 73 correct normalization of the data is crucial and in general requires an organism-speciﬁc rescaling of 74 expression proﬁles. This implies that a valid differential analysis should only include data that can be 75 attributed to single organisms. The application of differential expression analysis to mixed-species data 76 without prior separation can be found in several metatranscriptomic studies (Nacke et al., 2014; McNulty 77 et al., 2013; De Filippis et al., 2016) as well as in dedicated pipelines for metatranscriptome analysis 78 (Martinez et al., 2016; Westreich et al., 2016). Our results suggest that inadequate normalization of 79 metatranscriptomic count data always bares the risk of serious errors in differential expression analysis 80 and should be avoided consequently. 81 NORMALIZATION 82 95 where the relative feature abundance is equal for all samples and all size factors have been comprised in 92 the row vector s. Thus, for NDE features the size factors are proportional to the expected counts. If we 93 knew which features are actually NDE, we would be able to estimate the required size (scaling) factors 94 for normalization from the corresponding counts. 95 p g With the common choice of a mean scaling factor of 1 the scaling factors can be estimated b 96 With the common choice of a mean scaling factor of 1 the scaling factors can be estimated b 96 ˆs = Y (NDE) i• 1 n ∑jY (NDE) ij (3) ˆs = Y (NDE) i• 1 n ∑jY (NDE) ij (3) where n is the number of samples. The denominator in the above equation corresponds to the arithmetic 97 mean of the counts for feature i. If the library sizes of the samples strongly diverge, possibly by orders of 98 magnitude, the geometric mean is more suitable (Anders and Huber, 2010). 99 To make the expected counts of different samples comparable, i.e. in order to compare the feature 100 concentrations, the columns of the data matrix are divided by the sample-speciﬁc scaling factors prior to 101 the testing for signiﬁcant differences. As above, it is common to choose an average scaling factor 1. Thus, 102 if we would actually know an NDE feature beforehand, in principle, we could use it to estimate the scaling 103 factors. Usually this is not the case and we need to make some assumptions. A common assumption that 104 is used in current tools is that most of the features are NDE. Then it is possible to estimate the scaling 105 factors by some robust statistics. In DESeq for each sample the putative scaling factors from all features 106 are calculated and then the median of all these values is used as an estimator of the sample-speciﬁc scaling 107 factor (Anders and Huber, 2010). The median is highly robust, with a breakdown point of 50% and 108 therefore the estimator can be used if at least half of the data corresponds to NDE features. Without any 109 distinction between DE and NDE features the scaling factors have also been estimated from the count 110 sums of all samples. NORMALIZATION 82 To clarify our arguments for an alternative normalization of metatranscriptomic data we need to explain 83 the statistical nature of the normalization problem. We ﬁrst follow the approach of Anders and Huber 84 (Anders and Huber, 2010) for single organism RNA-Seq count data and start with a basic model for the 85 mean of the observed counts. The expected (mean) count E[Yi j] for gene (feature) i and sample j arises 86 from a product of the per-gene quantity λicj under condition cj and a size factor sj: 87 (1) E[Yij] = λicjsj The factor λic j is proportional to the mean concentration of feature i under condition cj. The size factor sj 88 represents the sampling depth or library size. Usually, both factors are unknown. If we assume the i-th 89 feature to be non-differentially expressed (NDE) we can represent the corresponding row of the count 90 matrix by 91 2/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint E h Y (NDE) i• i = λis (2) E h Y (NDE) i• i = λis (2) where the relative feature abundance is equal for all samples and all size factors have been comprised in 92 the row vector s. Thus, for NDE features the size factors are proportional to the expected counts. If we 93 knew which features are actually NDE, we would be able to estimate the required size (scaling) factors 94 for normalization from the corresponding counts. . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint Taxon-speciﬁc scaling and global scaling 136 We propose a method to prepare metatranscriptomic data for differential expression analysis. The 137 method is referred to as taxon-speciﬁc scaling. As an essential prerequisite our approach requires that 138 the data is ﬁrst partitioned according to the contributing organisms. Then the count data matrix from 139 each partition is normalized separately. Here, established tools from transcriptomics can be used to 140 estimate the corresponding scaling factors. Finally, the normalized count data matrices are summed 141 up to provide normalized metatranscriptomic count data which can be analyzed in terms of differential 142 expression (Fig. 1). Here all statistical models and tools for count-based differential expression analysis 143 in transcriptomics can in principle be used to identify differentially expressed features. 144 p p p y y p If we denote the original count matrix for organism k as Yk and the associated vector of estimated 145 scaling factors as ˆsk the normalized metatranscriptomic count matrix is computed by 146 ˜Y = ∑ k Yk diag−1(ˆsk) (8) Here, the diag−1 operator transforms the scaling vector to a diagonal matrix with inverse scaling factors 147 on the diagonal and zeros everywhere else. We provide an R script where we use DESeq2 for scaling 148 factor estimation and identiﬁcation of signiﬁcant differences (see Additional File 1). 149 In principle, our method is computationally simple and the hard work has to be done beforehand in 150 order to provide the partitioned data in terms of the organism-speciﬁc count matrices. This is the realm of 151 binning methods and, in addition, may require sequence assembly tools to achieve a sufﬁcient sequence 152 length for reliable separation. 153 At this point, the question may arise why to get back to metatranscriptomic data when differential 154 expression analysis could be performed for separate organisms or speciﬁc taxa. There are several reasons 155 why the analysis of the recombined metatranscriptome data can be useful: ﬁrst of all, the statistical power 156 of organism-speciﬁc tests may be low due to decreased counts. If several organisms show the same slight 157 difference, this difference may only become statistically signiﬁcant when accumulating their normalized 158 counts. Or a feature may show differences for single organisms but these differences may cancel out 159 when correctly summarized. In this case the corresponding feature is not indicative for the experimental 160 condition with regard to the whole community. NORMALIZATION 82 It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint (7) Sλ i = sαTλ i = ˜λis where ˜λi results from the dot product of the organism and the feature rates. This corresponds to equation 126 (2) and allows to apply DESeq or other tools for single organism differential expression analysis to the 127 metatranscriptomic count matrix. However, the underlying assumption that S has column rank 1, i.e. all 128 column vectors are collinear, would be hard to justify in practice. Implicitly we would assume that the 129 relative contributions of all organisms are constant accross all samples. In general, this assumption is not 130 met, because for a real metatranscriptome, the organism composition of transcripts cannot be controlled 131 and will be different for different samples. In our approach to normalization of metatranscriptomic 132 counts we preprocess the data according to an organism-speciﬁc rescaling of seperated counts so that the 133 recombined count data actually meet the former assumption. 134 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint NORMALIZATION 82 However, the potential shortcomings of this total count normalization have widely 111 been discussed (Anders and Huber, 2010; Robinson and Oshlack, 2010; Soneson and Delorenzi, 2013). 112 In metatranscriptomics the situation is more complicated because for each organism we can have a 113 different scaling factor. So we have to extend the above sampling model to an N-organism mixture that 114 includes a matrix S of organism-speciﬁc scaling factors s jk: 115 E[Yij] = N ∑ k=1 λijks jk E[Yij] = N ∑ k=1 λijks jk (4 (4) where i,j,k are the feature, sample and organism indices, repectively. We omitted the condition dependency 116 (cj) for convenience. 117 I l t ti (2) f NDE f t h th f ll i d l f f t i f th where i,j,k are the feature, sample and organism indices, repectively. We omitted the condition dependency 116 (cj) for convenience. 117 In analogy to equation (2) for NDE features we have the following model for a feature row i of the In analogy to equation (2) for NDE features we have the following model for a feature row i of th nt matrix: E h Y (NDE) i• i = λ T i ST (5 (5) where the column vector λ i contains all organism-speciﬁc rates for feature i and λ T i indicates transposition 120 of this vector. 121 where the column vector λ i contains all organism-speciﬁc rates for feature i and λ T i indicates transposition 120 of this vector. 121 Application of the above single-organism scheme for estimation of scaling factors is only valid if the 122 matrix of scaling factors has the following form: 123 S = [α1s,α2s,...,αKs] = sαT (6) S = [α1s,α2s,...,αKs] = sαT (6 (6) where α is a column vector of organism-speciﬁc abundances and s contains the sample-speciﬁc scaling 124 factors, now in a column vector, which is equal for all organisms. Then we can write 125 where α is a column vector of organism-speciﬁc abundances and s contains the sample-speciﬁc scaling 124 factors, now in a column vector, which is equal for all organisms. Then we can write 125 3/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. Taxon-speciﬁc scaling and global scaling 136 Therefore the analysis of separate organism transcriptomes 161 and the analysis of the rectiﬁed metatranscriptome data should be combined to provide a complete picture 162 of the community response. 163 In our study and in the supplied R script we use DESeq2 to compute scaling factors and to identify 164 signiﬁcant differences on the basis of the normalized count matrics. We decided for DESeq2 for several 165 reasons. It is an established tool in transcriptomics which has shown a good performance in compar- 166 ative studies (Soneson and Delorenzi, 2013; Dillies et al., 2013) and which has already been used for 167 metatranscriptome analysis (McNulty et al., 2013; Martinez et al., 2016; De Filippis et al., 2016). In 168 particular, the estimation of scaling factors is robust and can be performed as a separate prior step apart 169 from the computation of signiﬁcant differences. The latter aspect is important for taxon-speciﬁc scaling 170 which requires to apply the normalization independently. However, we would like to emphasize that our 171 4/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint arguments for the taxon-speciﬁc scaling approach do not depend on a particular statistical tool and in fact 172 the main ﬁndings of our study can be reproduced with other tools, such as edgeR (Robinson et al., 2010), 173 SAMseq(Li and Tibshirani, 2013) or limma(Ritchie et al., 2015). In some experiments we also used 174 edgeR and total count (TC) normalization to study the impact of different transcriptomic scaling methods. 175 In contrast to taxon-speciﬁc scaling, global scaling performs the normalization of metatranscriptomic 176 data without prior separation, i.e. sample-speciﬁc scaling factors are estimated from the original metatran- 177 scriptomic counts. In general, taxon-speciﬁc and global scaling will result in distinct normalized count 178 matrices which in turn can lead to largely differing results in differential expression analysis. Synthetic data generation and analysis tools 181 The tool compcodeR (Soneson, 2014) was used to generate all simulated data. The tool generates count 182 data based on a negative binomial distribution model with parameters estimated from real transcriptome 183 data (Pickrell et al., 2010; Cheung et al., 2010). If not explicitly speciﬁed, the compcodeR parameters in 184 the R function “generate.org.mat” are used (see Additional File 1). All analyses were performed with R 185 version 3.3.0 and DESeq2 version 1.8.2, edgeR version 3.10.5. 186 Simulated metatranscriptome 187 209 TPk = \|{i : ˆLi = DE∧Li = DE}\| (9) FPk = \|{i : ˆLi = DE∧Li = NDE}\| (10) TPk = \|{i : ˆLi = DE∧Li = DE}\| (9) FPk = \|{i : ˆLi = DE∧Li = NDE}\| (10) TPk = \|{i : ˆLi = DE∧Li = DE}\| FPk = \|{i : ˆLi = DE∧Li = NDE}\| TPk = \|{i : ˆLi = DE∧Li = DE}\| (9) (10) Taxon-speciﬁc scaling and global scaling 136 We tried to 179 show this on simulated and real count data as described in the following. 180 Simulated metatranscriptome 187 A metatranscriptome arises from a mixture of various organisms, each with individual features. As a 188 result, a metatranscriptome can include features covered by all taxa as well as features occurring only 189 in few or a single organism. Generally, the count contributions from different organisms are not equal 190 and vary across samples. We refer to this as the variation of the library size. Therefore, compcodeR 191 was used to generate multiple data sets with different total count numbers to simulate the variation of 192 organism-speciﬁc library sizes. Thereby, each generated data set mimics the contribution of a single 193 organism. The data sets were then combined to simulate a metatranscriptomic count matrix. 194 As with all simulations, the data can only provide a coarse approximation of real metatranscriptomic 195 counts which depends on particular parameters. Therefore, settings for the number of features and the 196 number of total counts inﬂuence the results. Each organism is simulated with 100 differentially expressed 197 features (DEF), 50 of them upregulated, and with 900 features that were non-differentially expressed 198 (NDE). 199 Each data set consists of two conditions, A and B, with six samples (replicates) per condition and 200 ﬁve organisms (Org1 to Org5) per sample. In the ﬁrst three simulations, the different organism proﬁles 201 are stacked, to exclude any interference between features from different organisms in the combined data. 202 Accordingly, the ﬁnal count matrix has 12 columns and 5000 rows that correspond to samples and features, 203 respectively. The data generation process provides the necessary information to calculate the number 204 of true positives (TP) and false positives (FP). The label Li is DE or NDE according to feature i being 205 differentially expressed or non-differentially expressed. The statistical test used to detect DEF, provides a 206 p-value for each feature. The predicted label ˆLi is DE if the adjusted p-value (Benjamini and Hochberg, 207 1995) is below a threshold of 0.05 for feature i. The TP and FP counts are calculated for each organism k 208 individually. Simulation IV: “Mixed feature effects” 233 In the previous simulations I-III, the generated count matrices from different organisms are stacked in 234 the combined count matrix. By stacking the organism proﬁles, the provided feature labels from the data 235 generation process can be used to validate the predictions. Now we want to analyze which effects can 236 be observed if each feature can accumulate counts from different organisms. This case is common for 237 real metatranscriptomic counts which arise from a mixture of organisms. Of particular interest are two 238 effects that we refer to as ”cancellation” and ”boosting”. To simplify the analysis of these effects we 239 restricted our simulation to a mixture of two organisms, which had the same base counts as Org1 and 240 Org2 in Simulation I. The cancellation effect is observed when DEF that are signiﬁcant in one or both 241 organism datasets loose signiﬁcance in the mixture. In contrast, the boosting effect is observed for DEF 242 which are only signiﬁcant in the combined dataset. We generated data for three samples per condition, to 243 limit the variability of the superimposed count data. The ﬁrst 100 features of the organism proﬁles were 244 DE while the remaining 900 were NDE. The simulation is divided into part A, where we superimposed 245 features with the same DE direction and part B, where the corresponding DE features of Org1 and Org2 246 had opposite directions. 247 Note that the aim of Simulation IV was not to compare the different normalization approaches but 248 instead to demonstrate the possible effects that may result from mixed organism count data. However, 249 the simulation cannot be used to draw conclusions about the frequencies of the effects for real data. In 250 particular, we expect the boosting effect to be much stronger for real data where organisms with a similar 251 response may provide correlated features that can emphasize trends or differences between conditions 252 when superimposing their counts. 253 Part A For this part of the simulation, we superimposed equally directed features of the two organisms. 254 With 100 features selected as DE, the ﬁrst 50 are “upregulated” followed by 50 “downregulated” features 255 and 900 NDE features. This simulation was expected to show the boosting effect as well as the cancellation 256 effect. 257 Part B In part B we tried to further increase the frequency of the cancellation effect. Simulation III: “Condition dependent variation” 226 In the third simulation, we investigated to what extent a condition dependent variation of LS 227 In the third simulation, we investigated to what extent a condition dependent variation of LS can affect 227 the normalization results. Under condition A we increase LS of Org1 by a random factor between 1.5 and 228 2 while under condition B we decrease the LS by a random factor between 0.5 and 0.667. For Org2 the 229 direction of change is reversed, with a random decrease under condition A and an increase for condition 230 B. For Org1 and Org2 the same base count as for Simulation I and II is used and for Org3-5 all parameters 231 from Simulation II are used. 232 Simulation IV: “Mixed feature effects” 233 An important 258 aspect of identifying DEF is the difference between the mean count values of the two conditions. To 259 bring the mean count values of the mixture for the two conditions closer together, we added the sorted 260 ”upregulated” features of Org1 and the sorted “downregulated” features of Org2 and vice versa. The true 261 mean values available from the data generation process are the basis for the sorting. For the generated 262 DEF the sorting ensures that high count values in condition A from one organism are balanced with high 263 count values in condition B from the other organism. 264 Simulation II: “With library size variation” 220 Simulation II: “With library size variation” 220 In the second experiment we simulated a more realistic situation, with varying LS for all included 221 organisms. 222 Organism base counts are identical to the ﬁrst simulation but the LS is randomly increased or reduced 223 according to a random factor between 0.5 and 2. Due to the different library sizes of the samples, a prior 224 normalization is required. 225 Simulation I: “Without library size variation” 210 In the ﬁrst experiment we simulate the case where the libray size (LS) for each of the ﬁve organisms does 211 not vary accross different samples. Although this is an unrealistic case we performed this simulation 212 to verify that both normalization approaches work equally well under ideal conditions. In addition, we 213 wanted to investigate how different organism abundances affect the identiﬁcation of DEF. Each organism 214 was assigned a ﬁxed total count number across all samples, without variation in library size. We simulated 215 organism Org1 with a base count of 1e7 followed by organism Org2 with 5e6, 1e6 for organism Org3 216 and organism Org4, Org5 with 5e5,1e5 respectively. Because data is generated without variation for the 217 number of counts per sample, no normalization is required, i.e. the correct scaling factors for all samples 218 and all organisms are the same (= 1). 219 5/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint Error calculation of the scaling factor 276 The scaling error Ek was estimated from the difference between the sample-speciﬁc scaling factors ˆs jk and the actual ”true” scaling factors sjk for each organism k as provided by the simulation parameters. In both cases the factors are scaled to provide a unit mean across samples. To obtain a scaling error between 0 and 1, we compute the error by: Ek = ∑j \|ˆsjk −sjk\| 2n (11) Ek = ∑j \|ˆsjk −sjk\| 2n (11) where n is the number of samples. In addition, we used the logarithmic measure log2 ˆs jk/sjk to represent 277 the directed error. 278 Simulation V: “False positive control” 265 The results of the differential expression analysis 274 for global and taxon-speciﬁc scaling on the superimposed data were compared 275 The LS ranges were identical to those in Simulation III. The results of the differential expression analysis 274 for global and taxon-speciﬁc scaling on the superimposed data were compared. 275 Metatranscriptome data 279 For a real data study, we chose a metatranscriptome dataset from mice gut(McNulty et al., 2013). The 280 experiment includes 12 different species (see Additional File 2 : Tab. 1) representing an artiﬁcial human 281 gut microbial community which was inserted into germ free mice. In the original study the diet for 282 the mice was changed at different time points. Metatranscriptomic data is available for 6 time points 283 which provide the conditions for our analysis. The available processed count data was obtained from the 284 European Bioinformatics Institute [http://www.ebi.ac.uk/, ArrayExpress, E-GEOD-48993] and contains 285 gene names and the associated numbers. 286 g Because the gene to Pfam (Finn et al., 2014) mapping is available for most organisms, we selected 287 Pfam protein domains as features for the differential expression analysis. Each Pfam domain family is a 288 feature in the resulting vector, including only Pfams observed at least once. We transformed the available 289 RPKM values for the genes back to raw counts. For genes with multiple Pfam annotations, we add the 290 raw count values of the gene to all associated Pfam features. From the available data, we constructed 291 a count matrix for each condition and organism (Additional File 3). Here, each column constitutes a 292 different sample and each row represents a particular feature. Because the count data from Bacteroides 293 cellulosilyticus WH2 did not map to gene names, all related counts are excluded from the analysis. 294 A differential expression analysis for all pairwise combinations of distinguished conditions was 295 performed to compare the results of global and taxon-speciﬁc scaling. We calculated the number of DEF 296 predicted a) with both methods with the same fold change direction, b) with both methods but with an 297 opposite fold change direction and c) with only one scaling method. In addition, we investigated the 298 overlap between the single organism transcriptome analyses and the differential expression analysis for 299 the mixture. We applied a signiﬁcance threshold on the adjusted p-value of 0.05 for the prediction of DEF. 300 RESULTS & DISCUSSION 301 In the ﬁrst part of our evaluation we examined the performance of taxon-speciﬁc and global scaling 302 methods on simulated data. Because simulations I to III had been designed to provide a clear ground truth 303 we were able to distinguish true positive predictions of DEF from falsely classiﬁed features. In the second 304 part we show results on real metatranscriptomic count data. Here the ground truth is not known and 305 therefore we restrict the analysis on the comparison of the results from the two normalization approaches. 306 Because it is impossible to verify the correctness of predictions we focus on analyzing the agreement or 307 disagreement on DEF detection in this case. 308 Simulation V: “False positive control” 265 In the ﬁnal simulation, the aim was to investigate the effect of an LS shift within mixed count data 266 without any DE features. Here we particularly wanted to measure the impact of the normalization on the 267 false positive rate. This kind of analysis has also been proposed for transcriptomics to validate the false 268 discovery control in differential expression analysis tools (Soneson and Delorenzi, 2013). 269 We used the parameters from Simulation III, but instead of six samples per condition, compcodeR 270 generated 12 samples per condition where we only used the samples for the ﬁrst condition. For Org3-5 271 6/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint we use the LS variation as used in Simulation III. For Org1 and Org2 the LS shift between different 272 conditions in Simulation III is now applied between samples 1 6 (condition A) and 7 12 (condition B) 273 we use the LS variation as used in Simulation III. For Org1 and Org2 the LS shift between different 272 we use the LS variation as used in Simulation III. For Org1 and Org2 the LS shift between different 272 conditions in Simulation III is now applied between samples 1-6 (condition A) and 7-12 (condition B). 273 The LS ranges were identical to those in Simulation III. The results of the differential expression analysis 274 for global and taxon-speciﬁc scaling on the superimposed data were compared. 275 conditions in Simulation III is now applied between samples 1-6 (condition A) and 7-12 (condition B). 273 The LS ranges were identical to those in Simulation III. The results of the differential expression analysis 274 for global and taxon-speciﬁc scaling on the superimposed data were compared. 275 pp p The LS ranges were identical to those in Simulation III. Simulation II 323 When introducing organism-speciﬁc LS variation across samples the picture changes. For the global 324 scaling approach the results show a decrease in the average TP rate for all organisms (see Fig. 3). This 325 trend is also visible when edgeR and TC normalization are used for differential expression analysis (see 326 Additional File 2 : Fig 2). On the other hand, with taxon-speciﬁc scaling the results are very similar to 327 results from Simulation I (see Fig. 2). With this method more DEF are correctly identiﬁed than with 328 global scaling. The difference in the number of correctly identiﬁed DEF for global scaling is dependent 329 on the parameter settings for the LS variation. For a lower amplitude of the LS variation, the TP rate for 330 global scaling increases (Additional File 2 : Fig. 3). The range of TP for the most abundant organisms 331 Org1 and Org2 is broader with global scaling (see Fig. 3) which also shows a higher scaling error (SE) for 332 organisms with a lower sequencing depth (see Additional File 2 : Fig. 4). 333 The receiver operating characteristics (ROC) shows a higher area under curve (AUC) value for taxon- 334 speciﬁc scaling (0.8776) than for global scaling (0.8282). The curve for global scaling also shows a higher 335 degree of variation across different simulation runs (Fig. 4 dotted lines). 336 Simulation III 337 With the inclusion of a condition dependent variation of the LS this simulation experiment can be viewed 338 as a worst case study. For global scaling, the observed number of true positives is higher for all data sets 339 compared to Simulation II (see Fig. 5).However, the number of false positive predictions explodes and 340 even exceeds the total number of DEF (500) resulting in average TP and FP numbers of 228 (± 11) and 341 1523 (± 78), i.e. ∼35 % of all features are predicted to be DEF. 342 In particular, the biggest portion of FP accumulates in features from Org1 and Org2 (see Fig. 5). 343 Inspecting the log2 fold changes (see Fig. 6) a shift from the correct center of 0 upwards and downwards 344 can be observed for Org1 and Org2, respectively. As a result, many DEF are identiﬁed with a wrong 345 (opposite) direction and most of the false positive detections just reﬂect the direction of this shift. This 346 situation implies a total loss of control over the false discovery rate. The results with edgeR and TC 347 normalization show a similarly high FP rate (see Additional File 2 : Fig. 5). 348 As a consequence, the ROC curve collapses for global scaling (see Fig. 7) with an AUC of 0.6396. 349 In contrast, taxon-speciﬁc scaling does not suffer from condition-dependent LS variation and the results 350 compare well with those of Simulation I & II showing a similar shape of the ROC curve (AUC: 0.8785). 351 With taxon-speciﬁc scaling, the average TP across all species is 237 which corresponds to a sensitivity of 352 ∼47 %. For global scaling the total number of predicted DEF (TP + FP) is dependent on the amplitude of 353 the condition dependent LS shift and increases for bigger shifts. 354 Simulation I 309 In this experiment, we measured the ability to detect DEF in a metatranscriptome without variation of 310 organism-speciﬁc libray sizes accross different samples. This situation, in principle, does not require any 311 normalization and therefore we expected taxon-speciﬁc (”tax”) and global (”glo”) scaling to yield similar 312 results. This is conﬁrmed by the resulting true positive (TP) predictions of DEF for the included organisms 313 (Fig. 2). For both approaches the number of true positives is higher for more abundant organisms due to 314 7/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a ified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: oRxiv preprint an increased statistical power of the corresponding tests. The ﬁnal proﬁle includes 100 DE and 900 NDE 315 features for each organism, resulting in 5000 features in total. 316 We repeated the analysis with edgeR and TC normalizations to estimate the scaling factors (see 317 Additional File 2 : Fig. 1). For edgeR, the number of correctly identiﬁed DEF is lower for all organisms 318 (see Additional File 2 : Fig. 1). For this particular data set, the library size (LS) was correctly adjusted by 319 DESeq2 with scaling factors close to 1 for all samples. Again both normalization approaches performed 320 equally well. A similar picture can be expected for a varying total LS of the metatranscriptome samples 321 as long as the relative LS of the organisms does not vary across different samples. 322 Simulation IV 355 With this simulation we wanted to analyze the effects that result from the superposition of counts from 356 different species. In this case we do not compare the two normalization approaches. DEF that can not be 357 detected for each organism separately may be identiﬁed as DE in the mixture (boosting effect) and DEF 358 that can be detected for single organisms may disappear in the mixture (cancellation effect). Here we 359 analyzed the frequencies of the different effects for the features that were labeled DE according to the 360 data generation process. The presented numbers result from averaging over 100 iterations and for the 361 observed effects these numbers sum up to the total number of DE-labeled features (100). 362 Part A First, we just added the data matrices of the two simulated organisms, i.e. the upregulated and 363 downregulated feature counts are summed up. In the results, boosting as well as cancellation effects can 364 be observed. The chance to observe the boosting effect is low because the features of the two organisms 365 and conditions are not correlated. Furthermore, when counts with different orders of magnitude are added, 366 8/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a tified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint only slight changes in the mean counts can be observed. As a result, the boosting and cancellation effects 367 only show median frequencies of 5 and 9, respectively (see Fig. 8). Simulation V 384 In the ﬁnal simulation, we wanted to investigate the effect of the scaling on the false discovery rate 385 for mixed organism count data. Therefore the data matrices (12 samples and 1000 features) for this 386 experiment were generated without any DEF. 387 For taxon-speciﬁc scaling, the number of signiﬁcant features is negligible with an average number of 388 1.6, which is well within the range of the estimated false discovery rate (FDR). In contrast, for global 389 scaling the average number of predicted DEF is 628 ± 22. While with taxon-speciﬁc scaling there is 390 little if any signiﬁcant difference, global scaling predicted ∼63 % of all features to be DE. Only a small 391 number of predicted DEF results from analysis of the organism transcriptomes with a mean of 1.2, again 392 well within the FDR range for an adjusted p-value threshold of 0.05. The results from Simulation IV 393 show that the prevalent effect of the superposition of uncorrelated data is the cancellation effect and the 394 boosting effect could only be observed in a few cases. However, in Simulation IV we focused on DEF 395 that were marked as differentially expressed by the data generation process. Therefore we can exclude 396 that the high number of DEF predicted by global scaling, results from the boosting effect. 397 Simulation IV 355 The main portion of features identiﬁed 368 as DE in the mixture is also identiﬁed as DE for at least one of the organisms, followed by the fraction of 369 features that are insigniﬁcant in all cases (see Fig. 8). Increasing the number of samples per condition 370 from 3 to 6 further reduces the median number of boosted features (3) while also increasing the overall 371 ability to correctly identify DE features (see Additional File 2 : Fig. 6). 372 Part B In the second part of the simulation, we changed the order of the features to intensify the 373 cancellation effect. This was achieved by adding sorted upregulated features of one organism to sorted 374 downregulated features of the other organism. The median number of features identiﬁed as DE for both 375 cases, mixture and single organism analysis, is reduced to 7 and the boosting effect almost completely 376 disappears (see Fig. 8). Due to our simulation setup, features identiﬁed as DE in one of the organisms but 377 not identiﬁed in the mixture were the second most frequent (see Fig. 8) while features predicted as NDE 378 for both organisms and for the mixture were most frequent. 379 Increasing the number of samples per condition from three to six resulted in a stronger cancellation 380 effect and a higher number of features identiﬁed as DE for one of the organisms and in the mixture (see 381 Additional File 2 : Fig. 6). Again, the increased number of samples improves the overall detection of 382 DEF. 383 Analysis of “day 13” vs “day 27” 423 In this category, taxon-speciﬁc scaling and global scaling predicted 5 additional DEF uniquely. 433 GO terms with predicted DEF from taxon-speciﬁc scaling alone included magnesium ion binding 434 and fucose metabolic process with 3 predicted DEF each. For global scaling alone, we found DNA 435 modiﬁcation, molybdopterin cofactor biosynthetic process, and RNA modiﬁcation with 3, 2 and 2 436 predicted DEF respectively (see Additional File 4 for a complete list). 437 scaling. In this category, taxon-speciﬁc scaling and global scaling predicted 5 additional DEF uniquely. 433 GO terms with predicted DEF from taxon-speciﬁc scaling alone included magnesium ion binding 434 and fucose metabolic process with 3 predicted DEF each. For global scaling alone, we found DNA 435 modiﬁcation, molybdopterin cofactor biosynthetic process, and RNA modiﬁcation with 3, 2 and 2 436 predicted DEF respectively (see Additional File 4 for a complete list). 437 Extra predictions In the condition comparison “day 13” vs “day 27”, both normalization approaches 438 shared 376 features predicted as DE. With taxon-speciﬁc scaling, 136 extra predictions were observable 439 while global scaling resulted in 380 extra predictions. When comparing the results of both scaling methods 440 to the single organism transcriptome analyses, global scaling and taxon-speciﬁc scaling resulted in 252 441 and 69 predictions that were insigniﬁcant in all single analyses (see Fig. 9). Both methods shared an 442 overlap of 53 DEF that were not detected in any of the transcriptome analyses. Thus, global scaling would 443 suggest a boosting effect for ∼13 % of the features that were insigniﬁcant in all transcriptome analyses. In 444 contrast, with taxon-speciﬁc scaling, a putative boosting effect for only ∼4% of the transcriptomic NDE 445 features can be observed. In the single organism transcriptomes, a total of 1302 features were predicted to 446 be DEF at least once. Both methods lead to a similar cancellation rate of ∼66% for taxon-speciﬁc scaling 447 and ∼61% for global scaling. 448 The fraction of shared DEF predictions between the two scaling methods is lower if the DEF are 449 supported by a smaller number of transcriptome analyses. For features supported by one transcriptome the 450 agreement was ∼48%, increasing to ∼56% for two transcriptomes. In the range of three to six supporting 451 transcriptomes, the agreement increases to ∼59%, ∼80%, 100% and 100% respectively. Real metatranscriptome data 398 It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a ertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint When taking into account the direction of the differential expression, the number of DEF predicted by 420 both methods but with contrary regulation direction is low. Here comparison “day 16” vs “day 30” shows 421 the highest number of signiﬁcant features with an opposite direction (5). 422 When taking into account the direction of the differential expression, the number of DEF predicted by 420 both methods but with contrary regulation direction is low. Here comparison “day 16” vs “day 30” shows 421 the highest number of signiﬁcant features with an opposite direction (5). 422 both methods but with contrary regulation direction is low. Here comparison “day 16” vs “day 30” shows 421 the highest number of signiﬁcant features with an opposite direction (5). 422 Real metatranscriptome data 398 While the objective of the simulation studies was to evaluate and compare the two normalization ap- 399 proaches in terms of correctly identiﬁed DEF, we do not have a ground truth for the analysis of the real 400 data. Therefore we focused on an analysis of the (dis-)agreement in results between both approaches 401 without assessing the actual detection performance. The analyzed data comprises Pfam counts from 11 402 organisms, 6 conditions according to different time points and 4 replicates per condition. An overview on 403 predicted DEF in all pairwise condition comparisons is shown in Fig. 9. 404 p p p g For both approaches, the number of DEF peaked at “day 13” vs. “day 27” with 512 and 756 signiﬁcant 405 features for taxon-speciﬁc and global scaling. The number of DEF was low when conditions close together 406 on the time line were compared (“day 15” vs. “day 16” or “day 29” vs. “day 30”). 407 For global scaling, the number of predicted DEF was generally higher than for taxon-speciﬁc scaling. 408 For some of the comparisons, the number of extra predictions under global scaling was even higher 409 than the number of shared predictions (see Fig. 9). The high number of extra predictions observed with 410 global scaling is especially prevalent for the comparisons “day 15” vs “day 16” with 16 times more extra 411 predictions than predictions shared with taxon-speciﬁc scaling and “day 13” vs “day 16” with 3 times 412 more extra predictions than shared predictions. 413 We also compared the results of the mixture analysis for global scaling and taxon-speciﬁc scaling 414 to the transcriptome analyses of the individual organisms. For 10 of the 15 comparisons, the majority 415 of features predicted as DE with global scaling were not predicted as DE in any transcriptome (see 416 Additional File 2 : Fig. 7). In contrast, with taxon-speciﬁc scaling only the comparisons “day 29” vs “day 417 30” and “day 15” vs “day 16” show a higher fraction of signiﬁcant features not predicted as DE in the 418 transcriptomes. These two comparisons are also the ones with the smallest total number of predicted DEF. 419 9/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. Analysis of “day 13” vs “day 27” 423 As described in the original study, “day 13” and “day 27” each correspond to the ﬁnal day of a particular 424 diet. Because the number of predicted DEF for both scaling methods (global and taxon-speciﬁc) was the 425 highest here, we analyze the results for this comparison in more detail. 426 We mapped the Pfam-annotated features which were predicted as DE for global scaling and taxon- 427 specitic scaling to Gene Ontology (GO) terms and compared the results. With taxon-speciﬁc scaling 250 428 GO terms with at least one DEF mapping were identiﬁed, while global scaling resulted in 311 GO terms. 429 GO terms associated to biological processes with a high agreement between the two methods were for 430 example cellular amino acid metabolic process where both methods identiﬁed 7 of the 9 associated Pfams 431 as DE and carbohydrate metabolic process with 11 DEF shared between taxon-speciﬁc scaling and global 432 scaling. In this category, taxon-speciﬁc scaling and global scaling predicted 5 additional DEF uniquely. 433 We mapped the Pfam-annotated features which were predicted as DE for global scaling and taxon- 427 specitic scaling to Gene Ontology (GO) terms and compared the results. With taxon-speciﬁc scaling 250 428 GO terms with at least one DEF mapping were identiﬁed, while global scaling resulted in 311 GO terms. 429 GO terms associated to biological processes with a high agreement between the two methods were for 430 example cellular amino acid metabolic process where both methods identiﬁed 7 of the 9 associated Pfams 431 as DE and carbohydrate metabolic process with 11 DEF shared between taxon-speciﬁc scaling and global 432 scaling. In this category, taxon-speciﬁc scaling and global scaling predicted 5 additional DEF uniquely. 433 GO terms with predicted DEF from taxon-speciﬁc scaling alone included magnesium ion binding 434 and fucose metabolic process with 3 predicted DEF each. For global scaling alone, we found DNA 435 modiﬁcation, molybdopterin cofactor biosynthetic process, and RNA modiﬁcation with 3, 2 and 2 436 di d DEF i l ( Addi i l Fil 4 f l li ) GO terms associated to biological processes with a high agreement between the two methods were for 430 example cellular amino acid metabolic process where both methods identiﬁed 7 of the 9 associated Pfams 431 as DE and carbohydrate metabolic process with 11 DEF shared between taxon-speciﬁc scaling and global 432 scaling. CONCLUSIONS 492 Differential expression analysis in metatranscriptomics is challenging. Metatranscriptomic count data 493 from RNA-Seq experiments show two main modes of biological variation. The functional composition 494 of transcripts reﬂects the activity of organisms and systematic changes might indicate a metabolic 495 response to experimental conditions. The taxonomic composition of transcripts can change as well and a 496 change may not necessarily be explainable in terms of controlled experimental conditions. In contrast to 497 metagenomics, in metatranscriptomics the questions ”who is there?” and ”what are they doing” are not 498 necessarily connected and need to be answered separately. If the two questions are not separated, there is 499 a considerable risk, to interpret variations in the taxonomic composition as functional changes. This may 500 even happen if the functional proﬁles of all organisms stay the same under different conditions. 501 pp p g y Normalization of metatranscriptomic data must have the goal to eliminate the inﬂuence of taxonomic 502 variations from functional analysis. We argue that for a correct normalization the metatranscriptome needs 503 to be decomposed to normalize the organism proﬁles independently. Then the metatranscriptomic count 504 data may be recombined from the normalized proﬁles to look for any global trends in the superimposed 505 count data. If differential expression tools are directly applied to the metatranscriptomic count matrix a 506 high risk of erroneous results is encountered. Our simulations indicate that the main risk is not to miss 507 some of the true differences but the real danger is to detect a large number of false functional differences 508 which arise from taxonomic abundance variations across samples. In particular, if these variations are 509 condition dependent the false positive rate can explode, circumventing all statistical control mechanisms 510 for bounding the false discovery rate. 511 We would like to point out that our ﬁndings do not affect metatransciptome studies that just aim 512 to analyze the functional repertoire from RNA-Seq data. The question which functions or genes are 513 expressed is much easier to answer than the question what is the functional response to a change of 514 experimental conditions. However, it is important to note that our results do not only apply to the classic 515 two conditions setup that we used throughout our study. COMPETING INTERESTS 519 The authors declare that they have no competing interests. 520 The authors declare that they have no competing interests. 520 Analysis of “day 13” vs “day 27” 423 CC-BY 4.0 International license a ed by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: Rxiv preprint Single feature analysis For the comparison “day 13” vs “day 27” we now show several examples for 474 features that reﬂect particular mixed organism effects or the different behavior of the scaling methods. 475 With regard to the scaling error discussed above, several features actually collect counts from mainly 476 one organism (see Fig. 11, 12 and Additional File 2 : Fig. 8). In addition, we found several features 477 that show an observable boosting effect with global as well as taxon-speciﬁc scaling (e.g. Fig. 13 or 478 Additional File 2 : Fig. 9). For these features, signiﬁcant differences were only observed in the combined 479 metatranscriptome. Some of the extra predictions obtained with taxon-speciﬁc scaling are results of 480 a putative boosting effect (see Additional File 2 : Fig. 10 and Fig. 11). In contrast, some of the extra 481 predictions resulting in a putative boosting effect were observable only with global scaling (see Fig. 11 482 and Additional File 2 : Fig 8). In both cases, the incorrect scaling factors resulted in the detection of DE 483 for mainly one organism. For other features, the combination of multiple incorrect scaling factors also 484 predicted DE when global scaling was used (see Additional File 2 : Fig. 12). 485 The prevalent effect for both normalization methods was cancellation. Often, the DEF from multiple 486 transcriptomes cancel each other out (see Fig. 14 and Additional File 2 : Fig. 13) and in some cases an 487 organism switch could be observed (see Additional File 2 : Fig. 14 and 15). 488 An example for a contradicting expression direction is shown in Fig 15. For this feature, the incorrect 489 scaling factors obtained by global scaling for Org2 and Org4 suggest a higher expression of this feature in 490 “day 13”, while taxon-speciﬁc scaling predicted the expression to be higher in the “day 27” condition. 491 CONCLUSIONS 492 Also for multiple conditions and time series a 516 correct normalization is essential to separate functional from taxonomic trends in the metatranscriptomic 517 composition variations. 518 Analysis of “day 13” vs “day 27” 423 While the relative 452 agreement between taxon-speciﬁc and global scaling increases, the total number of features supported by 453 multiple transcriptomes decreases (Additional File 2 : Fig. 7). 454 Scaling error In the simulations, the differences in the estimated scaling factors for the single organism 455 proﬁles in comparison to the actual scaling factors were low with a scaling error of ∼0.01. In Simulation 456 II we found the scaling error to be high in general with global scaling (Additional File 2 : Fig. 3) and 457 in Simulation III we showed the drastic increase of features falsely identiﬁed as DE with global scaling 458 when two organisms with condition-dependent abundance shifts were combined. 459 To further investigate the increased number of DEF predicted by global scaling, we compared the 460 scaling factors estimated for the single organism proﬁles with the estimated scaling factors for the global 461 normalization in the comparison “day 13” vs “day 27”. For several organisms, a pattern emerged which 462 showed a condition speciﬁc scaling error (Fig. 10). While the scaling factors for one condition are 463 too small, the scaling factors in the other condition are too high. As a result, global scaling leads to 464 condition-dependent errors which may cause extra predictions of DEF because an artiﬁcial shift between 465 the two conditions is introduced. 466 Incorrect scaling is especially problematic for features, which mainly comprise counts from one 467 organism or when counts from mixed organisms with the same scaling shift are analyzed. For a quantiﬁ- 468 cation we determined the number of features, for which counts from a single organism (or the mixture of 469 organisms with the same scaling error direction) exceed 80% of the normalized counts for that feature. 470 For extra predictions obtained only with global scaling without evidence from the transcriptome analyses, 471 the counts from a single organism are the main contribution for 82 of 199 features. Additional 43 features 472 are predicted from the summed counts from organisms with the same scaling shift. 473 10/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . REFERENCES 528 Anders, S. and Huber, W. (2010). Differential expression analysis for sequence count data. Genome 529 Biology, 11(10):R106. 530 Benjamini, Y. and Hochberg, Y. (1995). Controlling the False Discovery Rate: A Practical and Powerful 531 Approach to Multiple Testing. Journal of the Royal Statistical Society. Series B (Methodological), 532 57(1):289–300. 533 Bullard, J. H., Purdom, E., Hansen, K. D., and Dudoit, S. (2010). Evaluation of statistical methods for 534 normalization and differential expression in mRNA-Seq experiments. BMC Bioinformatics, 11:94. 535 Celaj, A., Markle, J., Danska, J., and Parkinson, J. (2014). Comparison of assembly algorithms for 536 improving rate of metatranscriptomic functional annotation. Microbiome, 2:39. 537 Cheung, V. G., Nayak, R. R., Wang, I. X., Elwyn, S., Cousins, S. M., Morley, M., and Spielman, R. S. 538 (2010). Polymorphic Cis– and Trans–Regulation of Human Gene Expression. PLoS Biology, 8(9):1–14. 539 De Filippis, F., Genovese, A., Ferranti, P., Gilbert, J. A., and Ercolini, D. (2016). Metatranscriptomics re- 540 veals temperature-driven functional changes in microbiome impacting cheese maturation rate. Scientiﬁc 541 Reports, 6:21871. 542 Dillies, M. A., Rau, A., Aubert, J., Hennequet-Antier, C., Jeanmougin, M., Servant, N., Keime, C., 543 Marot, G., Castel, D., Estelle, J., Guernec, G., Jagla, B., Jouneau, L., Laloe, D., Le Gall, C., Schaeffer, 544 B., Le Crom, S., Guedj, M., and Jaffrezic, F. (2013). A comprehensive evaluation of normalization 545 methods for Illumina high-throughput RNA sequencing data analysis. Brieﬁngs in Bioinformatics, 546 B., Le Crom, S., Guedj, M., and Jaffrezic, F. (2013). A comprehensive evaluation of normalization 545 B., Le Crom, S., Guedj, M., and Jaffrezic, F. (2013). A comprehensive evaluation of normalization 545 methods for Illumina high-throughput RNA sequencing data analysis. Brieﬁngs in Bioinformatics, 546 14(6):671–683. 547 en, J. A. (2007). Environmental shotgun sequencing: its potential and challenges for studying th hidden world of microbes. PLoS Biology, 5(3):e82. Finn, R. D., Bateman, A., Clements, J., Coggill, P., Eberhardt, R. Y., Eddy, S. R., Heger, A., Hetherington, 550 K., Holm, L., Mistry, J., Sonnhammer, E. L., Tate, J., and Punta, M. (2014). Pfam: the protein families 551 database. Nucleic Acids Res., 42(Database issue):D222–230. 552 Franzosa, E. A., Morgan, X. C., Segata, N., Waldron, L., Reyes, J., Earl, A. M., Giannoukos, G., Boylan, 553 M. R., Ciulla, D., Gevers, D., Izard, J., Garrett, W. S., Chan, A. T., and Huttenhower, C. (2014). 554 Relating the metatranscriptome and metagenome of the human gut. FUNDING 525 This work was partly funded by DFG (Project: “Computational models for metatranscriptome analysis”, 526 Me3138). 527 AUTHOR’S CONTRIBUTIONS 521 HK implemented the method and performed all experiments. PM designed the model and wrote the ”Nor- 522 malization” section of the manuscript. HK & PM designed the simulations, performed data interpretation 523 and wrote the manuscript. All authors read and approved the ﬁnal manuscript. 524 11/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint REFERENCES 528 Genome Research, 18(9):1509– 577 1517. 578 Martinez, X., Pozuelo, M., Pascal, V., Campos, D., Gut, I., Gut, M., Azpiroz, F., Guarner, F., and 579 Manichanh, C. (2016). MetaTrans: an open-source pipeline for metatranscriptomics. Scientiﬁc Reports, 580 6:26447. 581 Mason, O. U., Hazen, T. C., Borglin, S., Chain, P. S., Dubinsky, E. A., Fortney, J. L., Han, J., Holman, 582 H. Y., Hultman, J., Lamendella, R., Mackelprang, R., Malfatti, S., Tom, L. M., Tringe, S. G., Woyke, 583 T., Zhou, J., Rubin, E. M., and Jansson, J. K. (2012). Metagenome, metatranscriptome and single-cell 584 Mason, O. U., Hazen, T. C., Borglin, S., Chain, P. S., Dubinsky, E. A., Fortney, J. L., Han, J., Holman, 582 H. Y., Hultman, J., Lamendella, R., Mackelprang, R., Malfatti, S., Tom, L. M., Tringe, S. G., Woyke, 583 , , , , , p g, , , , , , g , , y , T., Zhou, J., Rubin, E. M., and Jansson, J. K. (2012). Metagenome, metatranscriptome and single-cell 584 sequencing reveal microbial response to Deepwater Horizon oil spill. ISME J, 6(9):1715–1727. 585 McNulty, N. P., Wu, M., Erickson, A. R., Pan, C., Erickson, B. K., Martens, E. C., Pudlo, N. A., Muegge, 586 B. D., Henrissat, B., Hettich, R. L., and Gordon, J. I. (2013). Effects of Diet on Resource Utilization by 587 a Model Human Gut Microbiota Containing Bacteroides cellulosilyticus WH2, a Symbiont with an 588 Extensive Glycobiome. PLoS Biology, 11(8):1–20. 589 B. D., Henrissat, B., Hettich, R. L., and Gordon, J. I. (2013). Effects of Diet on Resource Utilization by 587 a Model Human Gut Microbiota Containing Bacteroides cellulosilyticus WH2, a Symbiont with an 588 Extensive Glycobiome. PLoS Biology, 11(8):1–20. 589 Nacke, H., Fischer, C., Th¨urmer, A., Meinicke, P., and Daniel, R. (2014). Land use type signiﬁcantly 590 affects microbial gene transcription in soil. Microbial Ecology, 67(4):919–930. 591 Pickrell, J. K., Marioni, J. C., Pai, A. A., Degner, J. F., Engelhardt, B. E., Nkadori, E., Veyrieras, J.-B., 592 Stephens, M., Gilad, Y., and Pritchard, J. K. (2010). Understanding mechanisms underlying human 593 gene expression variation with RNA sequencing. Nature. 594 Poretsky, R. S., Bano, N., Buchan, A., LeCleir, G., Kleikemper, J., Pickering, M., Pate, W. M., Moran, 595 M. A., and Hollibaugh, J. T. (2005). Analysis of microbial gene transcripts in environmental samples. 596 Applied and Environmental Microbiology, 71(7):4121–4126. 597 Poretsky, R. S., Hewson, I., Sun, S., Allen, A. E., Zehr, J. REFERENCES 528 Proceedings of the National 555 Academy of Sciences of the United States of America, 111(22):E2329–2338. 556 Frias-Lopez, J., Shi, Y., Tyson, G. W., Coleman, M. L., Schuster, S. C., Chisholm, S. W., and Delong, E. F. 557 (2008). Microbial community gene expression in ocean surface waters. Proceedings of the National 558 Academy of Sciences of the United States of America, 105(10):3805–3810. 559 Gilbert, J. A., Field, D., Huang, Y., Edwards, R., Li, W., Gilna, P., and Joint, I. (2008). Detection of Large 560 Numbers of Novel Sequences in the Metatranscriptomes of Complex Marine Microbial Communities. 561 PLoS ONE, 3(8):e3042. 562 Hesse, C. N., Mueller, R. C., Vuyisich, M., Gallegos-Graves, L. V., Gleasner, C. D., Zak, D. R., and Kuske, 563 C. R. (2015). Forest ﬂoor community metatranscriptomes identify fungal and bacterial responses to N 564 deposition in two maple forests. Front Microbiol, 6:337. 565 Huson, D. H., Mitra, S., Ruscheweyh, H. J., Weber, N., and Schuster, S. C. (2011). Integrative analysis of 566 environmental sequences using MEGAN4. Genome Res., 21(9):1552–1560. 567 Li, J. and Tibshirani, R. (2013). Finding consistent patterns: a nonparametric approach for identifying 568 differential expression in RNA-Seq data. Stat Methods Med Res, 22(5):519–536. 569 Lin, Y., Golovnina, K., Chen, Z. X., Lee, H. N., Negron, Y. L., Sultana, H., Oliver, B., and Harbison, S. T. 570 (2016). Comparison of normalization and differential expression analyses using RNA-Seq data from 571 726 individual Drosophila melanogaster. BMC Genomics, 17:28. 572 p g Macklaim, J. M., Fernandes, A. D., Di Bella, J. M., Hammond, J. A., Reid, G., and Gloor, G. B. (2013). 573 Comparative meta-RNA-seq of the vaginal microbiota and differential expression by Lactobacillus 574 iners in health and dysbiosis. Microbiome, 1(1):12. 575 y , ( ) Marioni, J. C., Mason, C. E., Mane, S. M., Stephens, M., and Gilad, Y. (2008). RNA-seq: an assessment of 576 Marioni, J. C., Mason, C. E., Mane, S. M., Stephens, M., and Gilad, Y. (2008). RNA-seq: an assessment of 576 12/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint technical reproducibility and comparison with gene expression arrays. Genome Research, 18(9):1509– 577 1517. 578 technical reproducibility and comparison with gene expression arrays. REFERENCES 528 P., and Moran, M. A. (2009). Comparative 598 day/night metatranscriptomic analysis of microbial communities in the North Paciﬁc subtropical gyre. 599 Environmental Microbiology, 11(6):1358–1375. 600 Ritchie, M. E., Phipson, B., Wu, D., Hu, Y., Law, C. W., Shi, W., and Smyth, G. K. (2015). limma powers 601 differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Research, 602 43(7):e47. 603 Robinson, M. D., McCarthy, D. J., and Smyth, G. K. (2010). edgeR: a Bioconductor package for 604 differential expression analysis of digital gene expression data. Bioinformatics, 26(1):139–140. 605 Robinson, M. D. and Oshlack, A. (2010). A scaling normalization method for differential 606 analysis of RNA-seq data. Genome Biology, 11(3):R25. 607 y q gy, ( ) Simon, C. and Daniel, R. (2009). Achievements and new knowledge unraveled by metagenomic ap- 608 proaches. Applied Microbiology and Biotechnology, 85(2):265–276. 609 Soneson, C. (2014). compcodeR – an R package for benchmarking differential expression methods for 610 RNA-seq data. Bioinformatics, 30(17):2517–2518. 611 Soneson, C. and Delorenzi, M. (2013). A comparison of methods for differential expression analysis of 612 RNA-seq data. BMC Bioinformatics, 14:91. 613 Trapnell, C., Roberts, A., Goff, L., Pertea, G., Kim, D., Kelley, D. R., Pimentel, H., Salzberg, S. L., 614 Rinn, J. L., and Pachter, L. (2012). Differential gene and transcript expression analysis of RNA-seq 615 experiments with TopHat and Cufﬂinks. Nature Protocols, 7(3):562–578. 616 Urich, T., Lanzen, A., Qi, J., Huson, D. H., Schleper, C., and Schuster, S. C. (2008). Simultaneous assess- 617 ment of soil microbial community structure and function through analysis of the meta-transcriptome. 618 PLoS ONE, 3(6):e2527. 619 Westreich, S. T., Korf, I., Mills, D. A., and Lemay, D. G. (2016). SAMSA: a comprehensive metatran- 620 scriptome analysis pipeline. BMC Bioinformatics, 17(1):399. 621 Ye, Y. and Tang, H. (2016). Utilizing de Bruijn graph of metagenome assembly for metatranscriptome 622 analysis. Bioinformatics, 32(7):1001–1008. 623 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint FIGURES FIGURES 624 13/27 a) c) d) e) f) NDE NDE NDE DE DE Feature counts for each sample b) Feature proﬁles for each taxon Proﬁle normalization for each taxon Combining the normalized proﬁles Diﬀerential expression analysis Samples for two conditions Figure 1. Workﬂow for taxon-speciﬁc normalization. a) Sequence samples from conditions A (white) and B (light gray). Assign each sequence read to taxonomic and feature categories. b) Compute feature proﬁles from the assignment counts. c) Obtain count matrix from taxon-speciﬁc feature proﬁles. d) Normalize feature proﬁles of each taxon-speciﬁc count matrix separately. e) Recombine normalized feature proﬁles of all taxa into a metatranscriptomic proﬁle. f) Perform differential expression analysis on metatranscriptomic count matrix. a) c) Feature proﬁles for each taxon Samples for two conditions d) e) Combining the normalized proﬁles Figure 1. Workﬂow for taxon-speciﬁc normalization. a) Sequence samples from conditions A (white) and B (light gray). Assign each sequence read to taxonomic and feature categories. b) Compute feature proﬁles from the assignment counts. c) Obtain count matrix from taxon-speciﬁc feature proﬁles. d) Normalize feature proﬁles of each taxon-speciﬁc count matrix separately. e) Recombine normalized feature proﬁles of all taxa into a metatranscriptomic proﬁle. f) Perform differential expression analysis on metatranscriptomic count matrix. 14/27 . CC-BY 4.0 International license a ertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint #features Figure 2. Simulation I. Number of true positive (TP) and false positive (FP) features identiﬁed with DESeq2 for global (glo) and taxon-speciﬁc (tax) scaling: Boxplots represent variation over 100 runs of the simulation. #features Figure 2. Simulation I. Number of true positive (TP) and false positive (FP) features identiﬁed with DESeq2 for global (glo) and taxon-speciﬁc (tax) scaling: Boxplots represent variation over 100 runs of the simulation. #features Figure 3. Simulation II. Number of true positive (TP) and false positive (FP) features identiﬁed with DESeq2 for global (glo) and taxon-speciﬁc (tax) scaling. . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint FIGURES FP boxplots appear compressed due to outliers. Organism order for FP is the same as for TP boxplots. Figure 3. Simulation II. Number of true positive (TP) and false positive (FP) features identiﬁed with DESeq2 for global (glo) and taxon-speciﬁc (tax) scaling. FP boxplots appear compressed due to outliers. Organism order for FP is the same as for TP boxplots. 15/27 FPR Figure 4. ROC curves for Simulation II. Average curve for taxon-speciﬁc scaling (blue) vs. average curve for global scaling (red) with false positive rate (FPR) on x-axis and true positive rate (TPR) on y-axis. Dotted lines above and below indicate the standard deviation for each method. The average area under curve (AUC) is 0.8776 for taxon-speciﬁc scaling and 0.8282 for global scaling. FPR Figure 4. ROC curves for Simulation II. Average curve for taxon-speciﬁc scaling (blue) vs. average curve for global scaling (red) with false positive rate (FPR) on x-axis and true positive rate (TPR) on y-axis. Dotted lines above and below indicate the standard deviation for each method. The average area under curve (AUC) is 0.8776 for taxon-speciﬁc scaling and 0.8282 for global scaling. 16/27 #features Figure 5. Simulation III. Number of true positive (TP) and false positive (FP) features identiﬁed with DESeq2 for global (glo) and taxon-speciﬁc (tax) scaling. Boxplots represent variation over 100 runs of the simulation. Figure 5. Simulation III. Number of true positive (TP) and false positive (FP) features identiﬁed with DESeq2 for global (glo) and taxon-speciﬁc (tax) scaling. Boxplots represent variation over 100 runs of the simulation. 17/27 Figure 6. Simulation III: Log2 fold changes. Log2 fold changes of features for global normalization on one example data set. Along x-axis, features (dots) are ordered according to ﬁve stacked organism proﬁles, each with 1000 features of which the ﬁrst 50 features are “upregulated”, and the next 50 features are “downregulated”. Gray dots correspond to correctly detected NDE features, light green dots to downregulated features which are missed and dark green dots to correctly identiﬁed downregulated DEF. Light blue dots correspond to missed upregulated DEF and dark blue dots to correctly identiﬁed upregulated DEF. Red dots mark DEF where global scaling leads to an incorrect direction. Purple dots correspond to NDE features which are incorrectly identiﬁed as signiﬁcant features. Figure 6. Simulation III: Log2 fold changes. Log2 fold changes of features for global normalization on one example data set. . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint FIGURES Along x-axis, features (dots) are ordered according to ﬁve stacked organism proﬁles, each with 1000 features of which the ﬁrst 50 features are “upregulated”, and the next 50 features are “downregulated”. Gray dots correspond to correctly detected NDE features, light green dots to downregulated features which are missed and dark green dots to correctly identiﬁed downregulated DEF. Light blue dots correspond to missed upregulated DEF and dark blue dots to correctly identiﬁed upregulated DEF. Red dots mark DEF where global scaling leads to an incorrect direction. Purple dots correspond to NDE features which are incorrectly identiﬁed as signiﬁcant features. 18/27 FPR Figure 7. ROC curves for Simulation III. Average curve for taxon-speciﬁc scaling (blue) vs. average curve for global scaling (red) with false positive rate (FPR) on x-axis and true positive rate (TPR) on y-axis. Dotted lines above and below indicate the standard deviation for each method. The average area under curve (AUC) is 0.8785 for taxon-speciﬁc scaling and 0.6369 for global scaling. FPR Figure 7. ROC curves for Simulation III. Average curve for taxon-speciﬁc scaling (blue) vs. average curve for global scaling (red) with false positive rate (FPR) on x-axis and true positive rate (TPR) on y-axis. Dotted lines above and below indicate the standard deviation for each method. The average area under curve (AUC) is 0.8785 for taxon-speciﬁc scaling and 0.6369 for global scaling. 19/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint a (boost) b (cancel) c d #features a (boost) b (cancel) c d Figure 8. Simulation IV. Number of DEF that show a particular mixed organism effect. Effects are distinguished according to a) features not detected as DE in a single transcriptome but predicted as DE in the mixture (boosting), b) features detected as DE in at least one transcriptome but predicted as NDE in the mixture (cancellation), c) features identiﬁed as DE in at least one transcriptome and also predicted as DE in the mixture and d) features not detected as DE for both, transcriptomes and mixture. Boxplots represent variation over 100 runs of the simulation. a (boost) b (cancel) c d #features a (boost) b (cancel) c d Figure 8. . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint FIGURES Simulation IV. Number of DEF that show a particular mixed organism effect. Effects are distinguished according to a) features not detected as DE in a single transcriptome but predicted as DE in the mixture (boosting), b) features detected as DE in at least one transcriptome but predicted as NDE in the mixture (cancellation), c) features identiﬁed as DE in at least one transcriptome and also predicted as DE in the mixture and d) features not detected as DE for both, transcriptomes and mixture. Boxplots represent variation over 100 runs of the simulation. 20/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint opposite direction shared unique #features #features Figure 9. Predicted DEF for real data. Number of signiﬁcant features from taxon-speciﬁc scaling (“tax”, right bar) and global scaling (“glo”, left bar) for different condition comparisons. Colors indicate shared signiﬁcant features with same direction of difference (grey), shared signiﬁcant features with opposite direction (red) and mutually exclusive features (purple) that are only found to be signiﬁcant for one scaling method. Smaller ﬁgure: histogram for predicted DEF according to the number of single organism analyses that show a signiﬁcant difference (x-axis). Upper part shows results for taxon-speciﬁc scaling and lower part for global scaling. For example, a high bar at “0” means that many features are found to be signiﬁcant for the metatranscriptome which are not signiﬁcant for any of the single transcriptome analyses opposite direction shared unique #features #features Figure 9. Predicted DEF for real data. Number of signiﬁcant features from taxon-speciﬁc scaling (“tax”, right bar) and global scaling (“glo”, left bar) for different condition comparisons. Colors indicate shared signiﬁcant features with same direction of difference (grey), shared signiﬁcant features with opposite direction (red) and mutually exclusive features (purple) that are only found to be signiﬁcant for one scaling method. Smaller ﬁgure: histogram for predicted DEF according to the number of single organism analyses that show a signiﬁcant difference (x-axis). Upper part shows results for taxon-speciﬁc scaling and lower part for global scaling. . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint FIGURES For example, a high bar at “0” means that many features are found to be signiﬁcant for the metatranscriptome which are not signiﬁcant for any of the single transcriptome analyses 21/27 . CC-BY 4.0 International license a ertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint log2 ratio sample organism Figure 10. Global scaling condition bias. Direction of global scaling “error”” in terms of the log2-ratio of scaling factors from transcriptomic and global scaling. Results for different organisms in the comparison of “day 13” vs. “day 27”. For symmetry of the color range the negative log2-ratio was capped at -1.25, with error scores below that threshold showing the same color (blue). Samples 1-4 are from condition A and samples 5-8 from condition B. For the species name abbreviations see Additional File 2 : Tab. 1. D. longicatena DSM 13814 was not observed in that particular condition comparison. log2 ratio sample organism Figure 10. Global scaling condition bias. Direction of global scaling “error”” in terms of the log2-ratio of scaling factors from transcriptomic and global scaling. Results for different organisms in the comparison of “day 13” vs. “day 27”. For symmetry of the color range the negative log2-ratio was capped at -1.25, with error scores below that threshold showing the same color (blue). Samples 1-4 are from condition A and samples 5-8 from condition B. For the species name abbreviations see Additional File 2 : Tab. 1. D. longicatena DSM 13814 was not observed in that particular condition comparison. 22/27 raw tax glo 0 10000 20000 30000 40000 2 4 6 8 2 4 6 8 2 4 6 8 sample count organism 7 8 PF07554 Figure 11. Single feature analysis (PF07554). Stacked bars in three parts (x-axis) show organism-speciﬁc counts before scaling (left), after taxon-speciﬁc scaling (middle) and after Taxon-speciﬁc and global scaling result in adjusted p-values 0.98 and 7.23−56, respectively. . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint FIGURES raw tax glo 0 10000 20000 30000 40000 2 4 6 8 2 4 6 8 2 4 6 8 sample count organism 7 8 PF07554 Figure 11. Single feature analysis (PF07554). Stacked bars in three parts (x-axis) show organism-speciﬁc counts before scaling (left), after taxon-speciﬁc scaling (middle) and after Taxon-speciﬁc and global scaling result in adjusted p-values 0.98 and 7.23−56, respectively. 23/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint global scaling (right). Extra prediction of DEF by (mis)scaling of one organism with global scaling. raw tax glo 0 25000 50000 75000 2 4 6 8 2 4 6 8 2 4 6 8 sample count organism 1 2 3 5 6 7 11 PF07881 Figure 12. Single feature analysis (PF07881). Stacked bars in three parts (x-axis) show organism-speciﬁc counts before scaling (left), after taxon-speciﬁc scaling (middle) and after global scaling (right). Loss of signiﬁcance due to (mis)scaling of proﬁles from mainly one organism. Feature is signiﬁcant for organisms Org2, Org5 and Org11 in transcriptome analysis. Taxon-speciﬁc and global scaling result in adjusted p-values 1.79e−3 and 0.66, respectively. raw tax glo 0 25000 50000 75000 2 4 6 8 2 4 6 8 2 4 6 8 sample count organism 1 2 3 5 6 7 11 PF07881 Figure 12. Single feature analysis (PF07881). Stacked bars in three parts (x-axis) show organism-speciﬁc counts before scaling (left), after taxon-speciﬁc scaling (middle) and after global scaling (right). Loss of signiﬁcance due to (mis)scaling of proﬁles from mainly one organism. Feature is signiﬁcant for organisms Org2, Org5 and Org11 in transcriptome analysis. Taxon-speciﬁc and global scaling result in adjusted p-values 1.79e−3 and 0.66, respectively. 24/27 raw tax glo 0 10000 20000 30000 40000 50000 2 4 6 8 2 4 6 8 2 4 6 8 sample count organism 1 2 3 4 5 6 7 9 11 PF00204 Figure 13. Single feature analysis (PF00204). Stacked bars in three parts (x-axis) show organism-speciﬁc counts before scaling (left), after taxon-speciﬁc scaling (middle) and after global scaling (right). With boosting effect observable for both methods, i.e. feature is insigniﬁcant in transcriptome analyses and signiﬁcant for metatranscriptomic counts. . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint FIGURES For Org1 and Org6 the adjusted p-values are close to 0.05 for single analysis, with taxon-speciﬁc scaling an adjusted p-value p = 0.005 is achieved. raw tax glo 0 10000 20000 30000 40000 50000 2 4 6 8 2 4 6 8 2 4 6 8 sample count organism 1 2 3 4 5 6 7 9 11 PF00204 Figure 13. Single feature analysis (PF00204). Stacked bars in three parts (x-axis) show organism-speciﬁc counts before scaling (left), after taxon-speciﬁc scaling (middle) and after global scaling (right). With boosting effect observable for both methods, i.e. feature is insigniﬁcant in transcriptome analyses and signiﬁcant for metatranscriptomic counts. For Org1 and Org6 the adjusted p-values are close to 0.05 for single analysis, with taxon-speciﬁc scaling an adjusted p-value p = 0.005 is achieved. 25/27 . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint raw tax glo 0 10000 20000 30000 2 4 6 8 2 4 6 8 2 4 6 8 sample count organism 1 2 3 4 5 6 7 8 9 11 PF01979 Figure 14. Single feature analysis (PF01979). Stacked bars in three parts (x-axis) show organism-speciﬁc counts before scaling (left), after taxon-speciﬁc scaling (middle) and after global scaling (right). Feature is signiﬁcant for transcriptome analysis but becomes insigniﬁcant for metatranscriptomic counts (cancellation effect), i.e. the adjusted p-value with both scaling methods is above 0.05. raw tax glo 0 10000 20000 30000 2 4 6 8 2 4 6 8 2 4 6 8 sample count organism 1 2 3 4 5 6 7 8 9 11 PF01979 Figure 14. Single feature analysis (PF01979). Stacked bars in three parts (x-axis) show organism-speciﬁc counts before scaling (left), after taxon-speciﬁc scaling (middle) and after global scaling (right). Feature is signiﬁcant for transcriptome analysis but becomes insigniﬁcant for metatranscriptomic counts (cancellation effect), i.e. the adjusted p-value with both scaling methods is above 0.05. 26/27 raw tax glo 0 20000 40000 60000 80000 2 4 6 8 2 4 6 8 2 4 6 8 sample count organism 1 2 3 4 5 6 PF12667 Figure 15. Single feature analysis (PF12667). . CC-BY 4.0 International license a certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under The copyright holder for this preprint (which was not this version posted May 5, 2017. ; https://doi.org/10.1101/134650 doi: bioRxiv preprint FIGURES Stacked bars in three parts (x-axis) show organism-speciﬁc counts before scaling (left), after taxon-speciﬁc scaling (middle) and after global scaling (right). Signiﬁcant feature with opposite direction for the two scaling methods. Taxon-speciﬁc and global scaling result in adjusted p-values 6.91e−5 and 1.33e−5, respectively. The log2 fold change for condition A in comparison to condition B is 0.70 for taxon-speciﬁc scaling and -0.82 for global scaling. raw tax glo 0 20000 40000 60000 80000 2 4 6 8 2 4 6 8 2 4 6 8 sample count organism 1 2 3 4 5 6 PF12667 Figure 15. Single feature analysis (PF12667). Stacked bars in three parts (x-axis) show organism-speciﬁc counts before scaling (left), after taxon-speciﬁc scaling (middle) and after global scaling (right). Signiﬁcant feature with opposite direction for the two scaling methods. Taxon-speciﬁc and global scaling result in adjusted p-values 6.91e−5 and 1.33e−5, respectively. The log2 fold change for condition A in comparison to condition B is 0.70 for taxon-speciﬁc scaling and -0.82 for global scaling. 27/27
https://openalex.org/W2218580750	https://openresearch-repository.anu.edu.au/bitstream/1885/220052/1/01_Alexander_THE_CONCISE_GUIDE_TO_2015.pdf	English	null	SLC6 neurotransmitter transporter family	null	2,013	cc-by	9,478	THE CONCISE GUIDE TO PHARMACOLOGY 2015/16: Overview Stephen PH Alexander1, Eamonn Kelly2, Neil Marrion2, John A Peters3, Helen E Benson4, Elena Faccenda4, Adam J Pawson4, Joanna L Sharman4, Christopher Southan4, O Peter Buneman5, William A Catterall6, John A Cidlowski7, Anthony P Davenport8, Doriano Fabbro9, Grace Fan10, John C McGrath11, Michael Spedding12, Jamie A Davies4 and CGTP Collaborators 1, Eamonn Kelly2, Neil Marrion2, John A Peters3, Helen E Benson4, Elena Faccenda4, N L , y , , J , , , na L Sharman4, Christopher Southan4, O Peter Buneman5, William A Catterall6, 1School of Biomedical Sciences, University of Nottingham Medical School, Nottingham, NG7 2UH, UK School of Biomedical Sciences, University of Nottingham Medical School, Nottingham, NG7 2UH, UK 2School of Physiology and Pharmacology, University of Bristol, Bristol, BS8 1TD, UK gy, y f g , , , tute of Environmental Health Sciences, National Institutes of Health, Department of Health and Human Services, le Park, NC 27709, USA 8Clinical Pharmacology Unit, University of Cambridge, Cambridge, CB2 0QQ, UK Q p , , 10The Agnes Irwin School, Rosemont, Pennsylvania, USA g , , y , 11School of Life Sciences, University of Glasgow, Glasgow, G12 8QQ, UK 12Spedding Research Solutions SARL, Le Vésinet 78110, France S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmaco S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmaco se Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 Abstract The Concise Guide to PHARMACOLOGY 2015/16 provides concise overviews of the key properties of over 1750 human drug targets with their pharmacology, plus links to an open access knowledgebase of drug targets and their ligands (www.guidetopharmacology.org), which provides more detailed views of target and ligand properties. The full contents can be found at http://onlinelibrary.wiley.com/doi/10. 1111/bph.13347/full. This compilation of the major pharmacological targets is divided into eight areas of focus: G protein-coupled receptors, ligand-gated ion channels, voltage-gated ion channels, other ion channels, nuclear hormone receptors, catalytic receptors, enzymes and transporters. These are presented with nomenclature guidance and summary information on the best available pharmacological tools, alongside key references and suggestions for further reading. The Concise Guide is published in landscape format in order to facilitate comparison of related targets. It is a condensed version of material contemporary to late 2015, which is presented in greater detail and constantly updated on the website www.guidetopharmacology.org, superseding data presented in the previous Guides to Receptors & Channels and the Concise Guide to PHARMACOLOGY 2013/14. It is produced in conjunction with NC-IUPHAR and provides the ofﬁcial IUPHAR classiﬁcation and nomenclature for human drug targets, where appropriate. It consolidates information previously curated and displayed separately in IUPHAR-DB and GRAC and provides a permanent, citable, point-in-time record that will survive database updates. Table of contents Overview 5729 Table of contents 5729 Overview 5734 Other Protein Targets 5734 Adiponectin receptors 5735 Blood coagulation components 5735 Non-enzymatic BRD containing proteins 5736 Carrier proteins 5737 CD molecules 5738 Methyllysine reader proteins 5739 Cytokines and growth factors 5739 Fatty acid-binding proteins 5741 Sigma receptors 5742 Tubulins 5744 G protein-coupled receptors 5746 Orphan and other 7TM receptors 5746 Class A Orphans 5756 Class C Orphans 5756 Taste 1 receptors 5757 Taste 2 receptors 5758 Other 7TM proteins 5759 5-Hydroxytryptamine receptors 5764 Acetylcholine receptors (muscarinic) 5766 Adenosine receptors 5768 Adhesion Class GPCRs 5770 Adrenoceptors 5774 Angiotensin receptors 5775 Apelin receptor Searchable database: http://www.guidetopharmacology.org/index.jsp Overview 5729 Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full 5729 Overview 5734 Other Protein Targets 5734 Adiponectin receptors 5735 Blood coagulation components 5735 Non-enzymatic BRD containing proteins 5736 Carrier proteins 5737 CD molecules 5738 Methyllysine reader proteins 5739 Cytokines and growth factors 5739 Fatty acid-binding proteins 5741 Sigma receptors 5742 Tubulins 5744 G protein-coupled receptors 5746 Orphan and other 7TM receptors 5746 Class A Orphans 5756 Class C Orphans 5756 Taste 1 receptors 5757 Taste 2 receptors 5758 Other 7TM proteins 5759 5-Hydroxytryptami 5764 Acetylcholine recep 5766 Adenosine receptor 5768 Adhesion Class GPC 5770 Adrenoceptors 5774 Angiotensin recepto 5775 Apelin receptor Overview 5729 Overview 5729 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharm S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. Table of contents Vitamin D receptor-like receptors 5965 2A. Hepatocyte nuclear factor-4 receptors 5966 2B. Retinoid X receptors 5967 2C. Testicular receptors 5968 2E. Tailless-like receptors 5969 2F. COUP-TF-like receptors 5970 3B. Estrogen-related receptors 5971 4A. Nerve growth factor IB-like receptors 5972 5A. Fushi tarazu F1-like receptors 5973 6A. Germ cell nuclear factor receptors 5974 0B. DAX-like receptors 5975 Steroid hormone receptors 5975 3A. Estrogen receptors 5976 3C. 3-Ketosteroid receptors 5979 Catalytic receptors 5981 Cytokine receptor family 5981 IL-2 receptor family 5983 IL-3 receptor family 5983 IL-6 receptor family 5985 IL-12 receptor family 5985 Prolactin receptor family p 5958 1A. Thyroid hormone receptors y p 5852 Thyrotropin-releasing hormone receptors 5852 Trace amine receptor p 5854 Urotensin receptor 5854 Vasopressin and oxytocin receptors 5967 2C. Testicular receptors p y 5856 VIP and PACAP receptors p 5968 2E. Tailless-like receptors p 5969 2F. COUP-TF-like receptors 5870 Ligand-Gated Ion Channels p 5970 3B. Estrogen-related receptors g 5871 5-HT3 receptors g p 5971 4A. Nerve growth factor IB-like receptors 5873 Acid-sensing (proton-gated) ion channels (ASICs) g 5972 5A. Fushi tarazu F1-like receptors g p g 5875 Epithelial sodium channels (ENaC) 5973 6A. Germ cell nuclear factor receptors p 5877 GABAA receptors 5974 0B. DAX-like receptors 5882 Glycine receptors p 5975 Steroid hormone receptors y p 5885 Ionotropic glutamate receptors p 5975 3A. Estrogen receptors 5976 3C. 3-Ketosteroid receptors 5891 Nicotinic acetylcholine receptors 5979 Catalytic receptors 5896 P2X receptors y p 5981 Cytokine receptor family y p 5981 IL-2 receptor family 5983 IL-3 receptor family 5904 Voltage-gated ion channels 5983 IL-6 receptor family p y 5985 IL-12 receptor family 5905 CatSper and Two-Pore channels p y 5985 Prolactin receptor family 5907 Cyclic nucleotide-regulated channels 5986 Interferon receptor family 5909 Potassium channels 5987 IL-10 receptor family 5910 Calcium-activated potassium channels 5988 Immunoglobulin-like family of IL-1 receptors 5912 Inwardly rectifying potassium channels 5989 IL-17 receptor family 5990 GDNF receptor family 5991 Integrins Table of contents British Journal of Pharmacology (2015) 172, 5729–5743 5844 Proteinase-activated receptors 5846 QRFP receptor 5846 Relaxin family peptide receptors 5848 Somatostatin receptors 5850 Succinate receptor 5850 Tachykinin receptors 5852 Thyrotropin-releasing hormone receptors 5852 Trace amine receptor 5854 Urotensin receptor 5854 Vasopressin and oxytocin receptors 5856 VIP and PACAP receptors 5870 Ligand-Gated Ion Channels 5871 5-HT3 receptors 5873 Acid-sensing (proton-gated) ion channels (ASICs) 5875 Epithelial sodium channels (ENaC) 5877 GABAA receptors 5882 Glycine receptors 5885 Ionotropic glutamate receptors 5891 IP3 receptor 5891 Nicotinic acetylcholine receptors 5896 P2X receptors 5898 Ryanodine receptor 5900 ZAC 5904 Voltage-gated ion channels 5905 CatSper and Two-Pore channels 5907 Cyclic nucleotide-regulated channels 5909 Potassium channels 5910 Calcium-activated potassium channels 5912 Inwardly rectifying potassium channels 5915 Two-P potassium channels 5917 Voltage-gated potassium channels 5920 Transient Receptor Potential channels 5934 Voltage-gated calcium channels 5936 Voltage-gated proton channel 5937 Voltage-gated sodium channels 5942 Other ion channels 5943 Aquaporins 5944 Chloride channels 5944 ClC family 5947 CFTR 5948 Calcium activated chloride channel 5949 Maxi chloride channel 5950 Volume regulated chloride channels 5952 Connexins and Pannexins 5954 Sodium leak channel, non-selective y org/index jsp 5777 Bile acid receptor 5778 Bombesin receptors 5780 Bradykinin receptors 5781 Calcitonin receptors 5783 Calcium-sensing receptors 5784 Cannabinoid receptors 5785 Chemerin receptor 5785 Chemokine receptors 5791 Cholecystokinin receptors 5792 Class Frizzled GPCRs 5793 Complement peptide receptors 5795 Corticotropin-releasing factor receptors 5796 Dopamine receptors 5798 Endothelin receptors 5799 G protein-coupled estrogen receptor 5800 Formylpeptide receptors 5801 Free fatty acid receptors 5803 GABAB receptors 5805 Galanin receptors 5806 Ghrelin receptor 5807 Glucagon receptor family 5809 Glycoprotein hormone receptors 5810 Gonadotrophin-releasing hormone receptors 5811 GPR18, GPR55 and GPR119 5812 Histamine receptors 5814 Hydroxycarboxylic acid receptors 5815 Kisspeptin receptor 5816 Leukotriene receptors 5818 Lysophospholipid (LPA) receptors 5819 Lysophospholipid (S1P) receptors 5820 Melanin-concentrating hormone receptors 5821 Melanocortin receptors 5822 Melatonin receptors 5823 Metabotropic glutamate receptors 5826 Motilin receptor 5827 Neuromedin U receptors 5828 Neuropeptide FF/neuropeptide AF receptors 5829 Neuropeptide S receptor 5828 Neuropeptide W/neuropeptide B receptors 5830 Neuropeptide Y receptors 5832 Neurotensin receptors 5833 Opioid receptors 5835 Orexin receptors 5836 Oxoglutarate receptor 5836 P2Y receptors 5838 Parathyroid hormone receptors 5839 Platelet-activating factor receptor 5840 Prokineticin receptors 5841 Prolactin-releasing peptide receptor 5842 Prostanoid receptors 5956 Nuclear hormone receptors 5958 1A. Thyroid hormone receptors 5959 1B. Retinoic acid receptors 5960 1C. Peroxisome proliferator-activated receptors 5961 1D. Rev-Erb receptors 5962 1F. Retinoic acid-related orphans 5963 1H. Liver X receptor-like receptors 5964 1I. 5991 Integrins 5991 Integrins g g p 5920 Transient Receptor Potential channels 5994 Natriuretic peptide receptor family 5934 Voltage-gated calcium channels 5996 Pattern recognition receptors g 5996 Toll-like receptor family 5996 Toll-like receptor family 5997 NOD-like receptor family p y 5999 Receptor serine/threonine kinase (RSTK) family p 6000 Type I receptor serine/threonine kinases 6001 Type II receptor serine/threonine kinases yp p 6001 Type III receptor serine/threonine kinases 6002 RSTK functional heteromers 6003 Receptor tyrosine kinases p y 6004 Type I RTKs: ErbB (epidermal growth factor) receptor family 6005 Type II RTKs: Insulin receptor family 6005 Type III RTKs: PDGFR, CSFR, Kit, FLT3 receptor family yp p y 6007 Type IV RTKs: VEGF (vascular endothelial growth factor) Overview 5730 Overview 5730 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Overview S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. receptor family p y 6008 Type V RTKs: FGF (ﬁbroblast growth factor) receptor family p yg 6080 Leukotriene and lipoxin metabolism yp ( g ) p y 6008 Type VI RTKs: PTK7/CCK4 yp ( 6008 Type VI RTKs: PTK7/CCK4 6081 GABA turnover yp 6009 Type VII RTKs: Neurotrophin receptor/Trk fami yp p 6010 Type VIII RTKs: ROR family p y 6083 1-phosphatidylinositol 4-kinase family yp 6010 Type IX RTKs: MuSK p p y y 6083 Phosphatidylinositol-4-phosphate 3-kinase family yp 6010 Type X RTKs: HGF (hepatocyte growth factor) receptor family p y p p 6084 Phosphatidylinositol 3-kinase family p y y 084 Phosphatidylinositol-4,5-bisphosphate 3-kinase fam y 6011 Type XI RTKs: TAM (TYRO3-, AXL- and MER-TK) receptor family 6012 Type XII RTKs: TIE family of angiopoietin recepto yp y g p 6012 Type XIII RTKs: Ephrin receptor family yp p 6013 Type XIV RTKs: RET yp 6014 Type XV RTKs: RYK yp 6014 Type XVI RTKs: DDR (collagen receptor) family yp g 6015 Type XVII RTKs: ROS receptors p p 6089 Phosphatidylcholine-speciﬁc phospholipase D yp p 6015 Type XVIII RTKs: LMR family yp y 6016 Type XIX RTKs: Leukocyte tyrosine kinase (LTK) receptor family y 6016 Type XX RTKs: STYK1 y 6016 Type XX RTKs: STYK1 yp 6017 Receptor tyrosine phosphatases (RTP) p y p p 6018 Tumour necrosis factor (TNF) receptor family 6018 Tumour necrosis factor (TNF) receptor family 6024 Enzymes 6028 Protein Kinases (EC 2.7.x.x) 6028 Rho kinase 6029 Protein kinase C (PKC) 6029 Alpha subfamily 6029 Delta subfamily 6030 Eta subfamily 6030 FRAP subfamily 6031 CDK4 subfamily 6031 GSK subfamily 6032 Polo-like kinase (PLK) family 6032 STE7 family 6033 Abl family 6033 Ack family 6034 Janus kinase (JakA) family 6034 Src family 6035 Tec family 6035 RAF family 6036 Peptidases and proteinases 6036 A1: Pepsin 6037 A22: Presenilin 6037 C14: Caspase 6037 M1: Aminopeptidase N 6038 M2: Angiotensin-converting (ACE and ACE2) 6038 M10: Matrix metallopeptidase 6039 M12: Astacin/Adamalysin 6039 M28: Aminopeptidase Y 5991 Integrins British Journal of Pharm 6079 Lipoxygenases 6080 Leukotriene and lipoxin metabolism 6081 GABA turnover 6082 Glycerophospholipid turnover 6082 Phosphatidylinositol kinases 6083 1-phosphatidylinositol 4-kinase family 6083 Phosphatidylinositol-4-phosphate 3-kinase family 6084 Phosphatidylinositol 3-kinase family 6084 Phosphatidylinositol-4,5-bisphosphate 3-kinase family 6085 1-phosphatidylinositol-3-phosphate 5-kinase family 6085 Type I PIP kinases (1-phosphatidylinositol-4-phosphate 5-kinase family) 6086 Type II PIP kinases (1-phosphatidylinositol-5-phosphate 4-kinase family) 6087 Phosphoinositide-speciﬁc phospholipase C 6088 Phospholipase A2 6089 Phosphatidylcholine-speciﬁc phospholipase D 6090 Lipid phosphate phosphatases 6091 Haem oxygenase 6092 Hydrogen sulphide synthesis 6093 Hydrolases 6093 Inositol phosphate turnover 6094 Inositol 1,4,5-trisphosphate 3-kinases 6094 Inositol polyphosphate phosphatases 6094 Inositol monophosphatase 6095 Lanosterol biosynthesis pathway 6097 Nucleoside synthesis and metabolism 6099 Sphingosine 1-phosphate turnover 6100 Sphingosine kinase 6100 Sphingosine 1-phosphate phosphatase 6101 Sphingosine 1-phosphate lyase 6101 Thyroid hormone turnover 6103 1.14.11.29 2-oxoglutarate oxygenases 6103 2.4.2.30 poly(ADP-ribose)polymerases 6104 2.5.1.58 Protein farnesyltransferase 6104 3.5.3.15 Peptidyl arginine deiminases (PADI) 6104 RAS subfamily 6105 4.2.1.1 Carbonate dehydratases 6105 5.99.1.2 DNA Topoisomerases 6110 Transporters 6113 ATP-binding cassette transporter family 6113 ABCA subfamily 6115 ABCB subfamily 6116 ABCC subfamily 6117 ABCD subfamily of peroxisomal ABC transporters 6118 ABCG subfamily 6119 F-type and V-type ATPases 6119 F-type ATPase 6120 V-type ATPase 6040 M19: Membrane dipeptidase 6040 S1: Chymotrypsin 6041 T1: Proteasome 6042 S8: Subtilisin 6042 S9: Prolyl oligopeptidase 6042 Acetylcholine turnover 6044 Adenosine turnover 6045 Amino acid hydroxylases 6046 L-Arginine turnover 6047 Arginase 6047 Arginine:glycine amidinotransferase 6047 Dimethylarginine dimethylaminohydrolases 6048 Nitric oxide synthases 6048 Carboxylases and decarboxylases 6049 Carboxylases 6050 Decarboxylases 6052 Catecholamine turnover 6055 Ceramide turnover 6055 Serine palmitoyltransferase 6056 Ceramide synthase 6057 Sphingolipid 14-desaturase 6058 Sphingomyelin synthase 6058 Sphingomyelin phosphodiesterase 6059 Neutral sphingomyelinase coupling factors 6059 Ceramide glucosyltransferase 6060 Acid ceramidase 6060 Neutral ceramidases 6061 Alkaline ceramidases 6061 Ceramide kinase 6062 Chromatin modifying enzymes 6062 2.1.1.- Protein arginine N-methyltransferases 6062 3.5.1.- Histone deacetylases (HDACs) 6063 Cyclic nucleotide turnover 6063 Adenylyl cyclases 6064 Soluble guanylyl cyclase 6065 Exchange protein activated by cyclic AMP (Epac) 6066 Phosphodiesterases, 3’,5’-cyclic nucleotide 6069 Cytochrome P450 6069 CYP1 family 6070 CYP2 family 6070 CYP3 family 6071 CYP4 family 6072 CYP5, CYP7 and CYP8 families 6072 CYP11, CYP17, CYP19, CYP20 and CYP21 families 6073 CYP24, CYP26 and CYP27 families 6074 CYP39, CYP46 and CYP51 families 6075 Endocannabinoid turnover 6076 Eicosanoid turnover 6077 Cyclooxygenase 6077 Prostaglandin synthases 6024 Enzymes p p 6094 Inositol polyphosphate phosphatases y 6028 Protein Kinases (EC 2.7.x.x) 6094 Inositol monophosphatase 6028 Rho kinase 6029 Protein kinase C (PKC) y 6099 Sphingosine 1-phosphate turnover p g 6100 Sphingosine 1-phosphate phosphatase 6101 Sphingosine 1-phosphate lyase p g p p y 6101 Thyroid hormone turnover Overview 5731 Overview 5731 Overview 5731 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharm S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 se Guide to PHARMACOLOGY 2015/16: Overview. transporters (HATs) p 6130 SLC3 family y 6130 SLC7 family y 6131 SLC4 family of bicarbonate transporters y 6132 Anion exchangers g 6132 Sodium-dependent HCO 3 transporters 3 6133 SLC5 family of sodium-dependent glucose transp 6134 Hexose transporter family p 6137 Sodium myo-inositol cotransporter transporters y p p 6138 SLC6 neurotransmitter transporter family 6024 Enzymes British Journal of Pharmacology (2015) 172, 5729–5743 6120 P-type ATPases 6121 Na C/K C-ATPases 6121 Ca2C-ATPases 6122 H C/K C-ATPases 6122 Cu C-ATPases 6122 Phospholipid-transporting ATPases 6123 Major facilitator superfamily (MFS) of transporters 6123 SLC superfamily of solute carriers 6124 SLC1 family of amino acid transporters 6124 Glutamate transporter subfamily 6126 Alanine/serine/cysteine transporter subfamily 6127 SLC2 family of hexose and sugar alcohol 6127 Class I transporters 6128 Class II transporters 6129 Proton-coupled inositol transporter 6129 SLC3 and SLC7 families of heteromeric amino acid transporters (HATs) 6130 SLC3 family 6130 SLC7 family 6131 SLC4 family of bicarbonate transporters 6132 Anion exchangers 6132 Sodium-dependent HCO 3 transporters 6133 SLC5 family of sodium-dependent glucose transporters 6134 Hexose transporter family 6135 Choline transporter 6136 Sodium iodide symporter, sodium-dependent multivita- min transporter and sodium-coupled monocarboxylate transporters 6137 Sodium myo-inositol cotransporter transporters 6138 SLC6 neurotransmitter transporter family 6138 Monoamine transporter subfamily 6139 GABA transporter subfamily 6141 Glycine transporter subfamily 6142 Neutral amino acid transporter subfamily 6144 SLC8 family of sodium/calcium exchangers 6145 SLC9 family of sodium/hydrogen exchangers 6145 SLC10 family of sodium-bile acid co-transporters 6147 SLC11 family of proton-coupled metal ion transporters 6148 SLC12 family of cation-coupled chloride transporters 6149 SLC13 family of sodium-dependent sulphate/carboxylate transporters 6150 SLC14 family of facilitative urea transporters 6151 SLC15 family of peptide transporters 6152 SLC16 family of monocarboxylate transporters 6154 SLC17 phosphate and organic anion transporter family 6154 Type I sodium-phosphate co-transporters 6155 Sialic acid transporter 6155 Vesicular glutamate transporters (VGLUTs) 6156 Vesicular nucleotide transporter 6156 SLC18 family of vesicular amine transporters 6158 SLC19 family of vitamin transporters 6159 SLC20 family of sodium-dependent phosphate transporters 6160 SLC22 family of organic cation and anion transporters 6160 Organic cation transporters (OCT) 6161 Organic zwitterions/cation transporters (OCTN) 6162 Organic anion transporters (OATs) 6163 Urate transporter 6163 SLC23 family of ascorbic acid transporters 6164 SLC24 family of sodium/potassium/calcium exchangers 6165 SLC25 family of mitochondrial transporters 6165 Mitochondrial di- and tri-carboxylic acid transporter subfamily 6166 Mitochondrial amino acid transporter subfamily 6167 Mitochondrial phosphate transporters 6167 Mitochondrial nucleotide transporter subfamily 6168 Mitochondrial uncoupling proteins 6169 Miscellaneous SLC25 mitochondrial transporters 6170 SLC26 family of anion exchangers 6170 Selective sulphate transporters 6170 Chloride/bicarbonate exchangers 6171 Anion channels 6171 Other SLC26 anion exchangers 6172 SLC27 family of fatty acid transporters 6173 SLC28 and SLC29 families of nucleoside transporters 6173 SLC28 family 6174 SLC29 family 6176 SLC30 zinc transporter family 6176 SLC31 family of copper transporters 6177 SLC32 vesicular inhibitory amino acid transporter 6178 SLC33 acetylCoA transporter 6179 SLC34 family of sodium phosphate co-transporters 6180 SLC35 family of nucleotide sugar transporters 6181 SLC36 family of proton-coupled amino acid transporters 6182 SLC37 family of phosphosugar/phosphate exchangers 6182 SLC38 family of sodium-dependent neutral amino acid transporters 6183 System A-like transporters 6183 System N-like transporters 6184 Orphan SLC38 transporters 6185 SLC39 family of metal ion transporters 6186 SLC40 iron transporter 6187 SLC41 family of divalent cation transporters 6187 SLC42 family of Rhesus glycoprotein ammonium transporters 6188 SLC43 family of large neutral amino acid transporters 6189 SLC44 choline transporter-like family 6190 SLC45 family of putative sugar transporters 6191 SLC46 family of folate transporters 6192 SLC47 family of multidrug and toxin extrusion transporters 6192 SLC48 heme transporter 6193 SLC49 family of FLVCR-related heme transporters 6194 SLC50 sugar transporter 6195 SLC51 family of steroid-derived molecule transporters 6195 SLC52 family of riboﬂavin transporters 6196 SLCO family of organic anion transporting polypeptides 6199 Patched family yp 6121 Na C/K C-ATPases 6121 Ca2C-ATPases / 6122 Cu C-ATPases 6122 Phospholipid-transporting ATPases p p p g 6123 Major facilitator superfamily (MFS) of transporte j p y ( 6123 SLC superfamily of solute carriers p y 6124 SLC1 family of amino acid transporters y p 6124 Glutamate transporter subfamily p y 6126 Alanine/serine/cysteine transporter subfamily y p y 6127 SLC2 family of hexose and sugar alcohol y 6127 Class I transporters p 6128 Class II transporters p 6129 Proton-coupled inositol transporter Introduction Phar- macological tools listed are prioritised on the basis of selectivity and availability. That is, agents (agonists, antagonists, inhibitors, activators, etc.) are included where they are both available (by donation or from commercial sources, now or in the near future) AND the most selective. This edition of the Concise Guide is di- vided into nine sections, which comprise pharmacological targets of similar structure/function. These are G protein-coupled recep- tors, ligand-gated ion channels, voltage-gated ion channels, other ion channels, catalytic receptors, nuclear hormone receptors, en- zymes, transporters and other protein targets. A new aspect of the Concise Guide 2015/16 is that each of these sections contains a cise Guide, although structural information is available on the on- line database. An expansion in the current version of the Concise Guide is the increased inclusion of approved drugs, which reﬂects the aim of the online database to reﬂect the clinical exploitation of human molecular targets. Although many of these agents are much less selective than the tool compounds listed to deﬁne in- dividual targets or groups of targets, we have included them for the signiﬁcant interest associated with their use and mechanisms of action. The emphasis on approved drugs means that the online database has been expanded to include 8024 ligands (as of August 2015), meaning that additional records now appear in the Con- cise Guide, primarily in the enzymes section. The organisation of the data is tabular (where appropriate) with a standardised format, where possible on a single page, intended to aid understanding of and comparison within a particular target group. The Concise Guide is intended as an initial resource, with links to additional complete listing of the families available for inspection on the on- line database, identifying those families reported in the Concise Guide by their page numbers. We hope that the Concise Guide will provide for researchers, teachers and students a state-of-the- art source of accurate, curated information on the background to their work that they will use in the Introductions to their Research Papers or Reviews, or in supporting their teaching and studies. We recommend that any citations to information in the Concise Guide are presented in the following format: Alexander SPH et al. (2015). The Concise Guide to PHARMACOL- OGY 2015/16: Overview. Br J Pharmacol XXX. Introduction eral guidelines from NC-IUPHAR. This current edition, the Con- cise Guide to PHARMACOLOGY 2015/16, is the latest snapshot of the database in print form, following on from the Concise Guide to PHARMACOLOGY 2013/14. It contains data drawn from the online database as a rapid overview of the major pharmacolog- ical targets. Thus, there are fewer targets presented in the Con- cise Guide (1761) compared to the online database (2761, as of August 2015). The priority for inclusion in the Concise Guide is the presence of quantitative pharmacological data. This means that often orphan family members are not presented in the Con- In order to allow clarity and consistency in pharmacology, there is a need for a comprehensive organisation and presentation of the targets of drugs. This is the philosophy of the IUPHAR/BPS Guide to PHARMACOLOGY presented on the online free access database (http://www.guidetopharmacology.org/). This database is supported by the British Pharmacological Society (BPS), the In- ternational Union of Basic and Clinical Pharmacology (IUPHAR), the Wellcome Trust and the University of Edinburgh. Data in- cluded in the Guide to PHARMACOLOGY are derived in large part from interactions with the subcommittees of the Nomenclature Committee of the International Union of Basic and Clinical Phar- macology (NC-IUPHAR). The Editors of the Concise Guide have compiled the individual records, in concert with the team of Cu- rators, drawing on the expert knowledge of these latter subcom- mittees. The tables allow an indication of the status of the nomen- clature for the group of targets listed, usually previously published in Pharmacological Reviews. In the absence of an established sub- committee, advice from several prominent, independent experts has generally been obtained to produce an authoritative consen- sus on nomenclature, which attempts to ﬁt in within the gen- Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Overview 5732 S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmaco S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/1 reviews and resources for greater depth and information. Phar- macological and structural data focus primarily on human gene products, wherever possible, with links to HGNC gene nomen- clature and UniProt IDs. In a few cases, where data from human proteins are limited, data from other species are indicated. Introduction In this overview are listed protein targets of pharmacological inter- est, which are not G protein-coupled receptors, ligand-gated ion channels, voltage-gated ion channels, ion channels, nuclear hor- mone receptors, catalytic receptors, transporters or enzymes. A dedication bers mentioned above, Tony established the online database at http://www.guidetopharmacology.org/ as the exceptional re- source it is today. This Edition of the Concise Guide to PHARMACOLOGY is ded- icated to Tony Harmar (1951-2014). Tony was a friend and colleague, who was involved with IUPHAR for over 15 years and worked on the IUPHAR database for over a decade at Edinburgh, working hard to establish the curators as a team of highly informed and informative individuals imbued with Tony’s passion and dogged determination to focus on high- quality data input, ensuring high-quality data output. With time and the resources of the BPS and Wellcome Trust, com- bined with the expertise of the NC-IUPHAR committee mem- Acknowledgements We are extremely grateful for the ﬁnancial contributions from the British Pharmacological Society, the International Union of Basic and Clinical Pharmacology, the Wellcome Trust (099156/Z/12/Z]), which support the website and the University of Edinburgh, who host the guidetopharmacology.org website. We are also tremendously grateful to the long list of collaborators from NC-IUPHAR subcommittees and beyond, who have assisted in the construction of the Concise Guide to PHARMACOLOGY 2015/16 and the online database www.GuideToPHARMACOLOGY.org Conﬂict of interest The authors state that there are no conﬂicts of interest to disclose. c 2015 The Authors. British Journal of Pharmacology published by John Wiley & Sons Ltd on behalf of The British Pharmacological Society. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided th g y , which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Overview 5733 Overview 5733 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full se Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 Other Protein Targets 5739 Fatty acid-binding proteins – Heat shock proteins – Immunoglobulins – Inhibitors of apoptosis (IAP) protein family – Kelch-like proteins – Kinesins – Mitochondrial-associated proteins – Notch receptors – Pentaxins – Serum pentaxins – Regulators of G protein signaling (RGS) proteins – RZ family – R4 family – R7 family – R12 family – Reticulons – Ribosomal factors 5741 Sigma receptors 5742 Tubulins – Tumour-associated p – WD repeat-containin 5734 Adiponectin receptors – B-cell lymphoma 2 (Bcl-2) protein family 5735 Blood coagulation components – Bromodomain-containing proteins 5735 Non-enzymatic BRD containing proteins 5736 Carrier proteins 5737 CD molecules – Chromatin-interacting transcriptional repressors 5738 Methyllysine reader proteins – Circadian clock proteins 5739 Cytokines and growth factors – EF-hand domain containing 5739 Fatty acid-binding proteins – Heat shock proteins – Immunoglobulins – Inhibitors of apoptosis (IAP) protein family – Kelch-like proteins – Kinesins – Mitochondrial-associated proteins – Notch receptors – Pentaxins – Serum pentaxins – Regulators of G protein signaling (RGS) proteins – RZ family – R4 family – R7 family – R12 family – Reticulons – Ribosomal factors 5741 Sigma receptors 5742 Tubulins – Tumour-associated protein – WD repeat-containing prot 5734 Adiponectin receptors – B-cell lymphoma 2 (Bcl-2) protein family 5735 Blood coagulation components – Bromodomain-containing proteins 5735 Non-enzymatic BRD containing proteins 5736 Carrier proteins 5737 CD molecules – Chromatin-interacting transcriptional repressors 5738 Methyllysine reader proteins – Circadian clock proteins 5739 Cytokines and growth factors – EF-hand domain containing Adiponectin receptors Other protein targets ! Adiponectin receptors Other protein targets ! Adiponectin receptors teraction, PH domain and leucine zipper containing 1, Q9UKG1 [52]). The adiponectin receptors are a class of proteins (along with membrane progestin receptors), which contain seven sequences of aliphatic amino acids reminiscent of GPCRs, but which are struc- turally and functionally distinct from that class of receptor. Overview: Adiponectin receptors (provisional nomen- clature, ENSFM00500000270960) respond to the 30 kDa complement-related protein hormone adiponectin (also known as ADIPOQ: adipocyte, C1q and collagen domain-containing protein; ACRP30, adipose most abundant gene transcript 1; apM-1; gelatin-binding protein: Q15848) originally cloned from adipocytes [49]. Although sequence data suggest 7TM domains, immunological evidence indicates that, contrary to typical 7TM topology, the carboxyl terminus is extracellular, while the amino terminus is intracellular [86]. Signalling through these recep- tors appears to avoid G proteins. Adiponectin receptors appear rather to stimulate protein phosphorylation via AMP-activated protein kinase and MAP kinase pathways [86], possibly through the protein partner APPL1 (adaptor protein, phosphotyrosine in- Nomenclature Adipo1 receptor Adipo2 receptor HGNC, UniProt ADIPOR1, Q96A54 ADIPOR2, Q86V24 Rank order of potency globular adiponectin ( ADIPOQ, Q15848) > adiponectin ( ADIPOQ, Q15848) globular adiponectin ( ADIPOQ, Q15848) = adiponectin ( ADIPOQ, Q15848) Comments: T-Cadherin (CDH13, P55290) has also been suggested to be a receptor for (hexameric) adiponectin [35]. Searchable database: http://www.guidetopharmacology.org/index.jsp Adiponectin receptors 5734 Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Adipo1 receptor ADIPOR1, Q96A54 globular adiponectin ( ADIPOQ, Q15848) > adiponectin ( ADIPOQ, Q15848) Adipo2 receptor ADIPOR2, Q86V24 Adipo2 receptor ADIPOR2, Q86V24 Adipo1 receptor ADIPOR1, Q96A54 Adipo1 receptor ADIPOR1, Q96A54 mments: T-Cadherin (CDH13, P55290) has also been suggested to be a receptor for (hexameric) adiponectin [35] Adiponectin receptors 5734 Adiponectin receptors 5734 Adiponectin receptors 5734 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 Further Reading Further Reading Astermark J (2015) FVIII inhibitors: pathogenesis and avoidance. Blood 125: 2045-2051 [PMID:25712994] Further Reading Shetty S et al. (2009) Adiponectin in health and disease: evaluation of adiponectin-targeted drug development strategies. Trends Pharmacol. Sci. 30: 234-9 [PMID:19359049] Buechler C et al. (2010) Adiponectin receptor binding proteins–recent advances in elucidating adiponectin signalling pathways. FEBS Lett. 584: 4280-6 [PMID:20875820] Buechler C et al. (2010) Adiponectin receptor binding proteins–recent advances in elucidating adiponectin signalling pathways. FEBS Lett. 584: 4280-6 [PMID:20875820] p g g p y [ ] Dalamaga M et al. (2012) The role of adiponectin in cancer: a review of current evidence. Endocr. Rev. 33: 547-94 [PMID:22547160] [ ] Goldstein BJ et al. (2009) Protective vascular and myocardial effects of adiponectin. Nat Clin Pract Cardiovasc Med 6: 27-35 [PMID:19029992] [ ] Juhl C et al. (2012) Molecular tools to characterize adiponectin activity. Vitam. Horm. 90: 31-56 [PMID:23017711] a ( ) [PMID:23017711] Blood coagulation components Other protein targets ! Blood coagulation components Other protein targets ! Blood coagulation components Other protein targets ! Blood coagulation components components of the coagulation cascade targetted by agents in cur- rent clinical usage. components of the coagulation cascade targetted by agents in cur- rent clinical usage. lation and aggregation of platelets, as well as proteins circulating in the plasma. The coagulation cascade involves multiple proteins being converted to more active forms from less active precursors, typically through proteolysis (see Proteases). Listed here are the Overview: Coagulation as a patho/physiological process is inter- preted as a mechanism for reducing excessive blood loss through the generation of a gel-like clot local to the site of injury. The process involves the activation, adhesion (see Integrins), degranu- serpin peptidase inhibitor, clade C (antithrombin), member 1 SERPINC1, P01008 heparin (pKd 7.8) [25], fondaparinux (pKd 7.5) [65], dalteparin [34], danaparoid [15, 58], enoxaparin [17], tinzaparin [19] – Nomenclature coagulation factor V (proaccelerin, labile factor) coagulation factor VIII, procoagulant component serpin peptidase inhibitor, clade C (antithrombin), member 1 HGNC, UniProt F5, P12259 F8, P00451 SERPINC1, P01008 Selective activators – – heparin (pKd 7.8) [25], fondaparinux (pKd 7.5) [65], dalteparin [34], danaparoid [15, 58], enoxaparin [17], tinzaparin [19] Selective antagonists drotrecogin alfa (Inhibition) [40, 41] drotrecogin alfa (Inhibition) [40, 41] – Further Reading Astermark J (2015) FVIII inhibitors: pathogenesis and avoidance. Blood 125: 2045-2051 [PMID:25712994] Non-enzymatic BRD containing proteins Other protein targets ! Bromodomain-containing proteins ! Non-enzymatic BRD containing proteins Overview: bromodomains bind proteins with acetylated lysine residues, such as histones, to regulate gene transcription. Listed herein are examples of bromodomain-containing proteins for which sufﬁcient pharmacology exists. Searchable database: http://www.guidetopharmacology.org/index.jsp Non-enzymatic BRD containing proteins 5735 Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full coagulation factor VIII, procoagulant component F8, P00451 serpin peptidase inhibitor, clade C (antithrombin), member 1 SERPINC1, P01008 drotrecogin alfa (Inhibition) [40, 41] drotrecogin alfa (Inhibition) [40, 41] Shetty S et al. (2009) Adiponectin in health and disease: evaluation of adiponectin-targeted drug development strategies. Trends Pharmacol. Sci. 30: 234-9 [PMID:19359049] Further Reading S Gallenkamp D et al. (2014) Bromodomains and their pharmacological inhibitors. ChemMedChem 9: 438-464 [PMID:24497428] Brand M et al. (2015) Small molecule inhibitors of bromodomain-acetyl-lysine interactions. ACS Chem Biol 10:22-39 [PMID:25549280] [ ] Filippakopoulos P and Knapp S (2014) Targeting bromodomains: epigenetic readers of lysine acety- lation. Nat Rev Drug Discov 13: 337-356 [PMID:24751816] S Non-enzymatic BRD containing proteins Other protein targets ! Bromodomain-containing proteins ! Non-enzymatic BRD containing proteins Overview: bromodomains bind proteins with acetylated lysine residues, such as histones, to regulate gene tra sufﬁcient pharmacology exists. Non-enzymatic BRD containing proteins 5735 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 Nomenclature bromodomain adjacent to zinc ﬁnger domain, 2A bromodomain adjacent to zinc ﬁnger domain, 2B CREB binding protein polybromo 1 SWI/SNF related, matrix associ- ated, actin dependent regula- tor of chromatin, subfamily a, member 4 HGNC, UniProt BAZ2A, Q9UIF9 BAZ2B, Q9UIF8 CREBBP, Q92793 PBRM1, Q86U86 SMARCA4, P51532 Selective inhibitors GSK2801 (pKd 6.6) [73] GSK2801 (Binding) (pKd 6.9) [73] I-CBP112 (pKd 6.8) [72] PFI-3 (Binding) (pKd 7.3) [79] PFI-3 (Binding) (pKd 7.1) [79] Further Reading Brand M et al. (2015) Small molecule inhibitors of bromodomain-acetyl-lysine interactions. ACS Chem Biol 10:22-39 [PMID:25549280] Filippakopoulos P and Knapp S (2014) Targeting bromodomains: epigenetic readers of lysine acety- lation. Nat Rev Drug Discov 13: 337-356 [PMID:24751816] Gallenkamp D et al. (2014) Bromodomains and their pharmacological inhibitors. ChemMedChem 9: 438-464 [PMID:24497428] Sanchez R et al. (2014) The bromodomain: from epigenome reader to druggable target. Biochim Biophys Acta 1839: 676-685 [PMID:24686119] Nomenclature bromodomain adjacent to zinc ﬁnger domain, 2A bromodomain adjacent to zinc ﬁnger domain, 2B CREB binding protein polybromo 1 SWI/SNF related, matrix associ- ated, actin dependent regula- tor of chromatin, subfamily a, member 4 HGNC, UniProt BAZ2A, Q9UIF9 BAZ2B, Q9UIF8 CREBBP, Q92793 PBRM1, Q86U86 SMARCA4, P51532 Selective inhibitors GSK2801 (pKd 6.6) [73] GSK2801 (Binding) (pKd 6.9) [73] I-CBP112 (pKd 6.8) [72] PFI-3 (Binding) (pKd 7.3) [79] PFI-3 (Binding) (pKd 7.1) [79] domain-acetyl-lysine interactions. ACS ains: epigenetic readers of lysine acety- 6] Gallenkamp D et al. (2014) Bromodomains and their pharmacological inhibitors. ChemMedChem 9: 438-464 [PMID:24497428] Sanchez R et al. (2014) The bromodomain: from epigenome reader to druggable target. Biochim Biophys Acta 1839: 676-685 [PMID:24686119] Carrier proteins Other protein targets ! Carrier proteins can also form pathological amyloid ﬁbrils [85]. Pharmacological intervention to reduce or prevent TTR dissociation is being pur- sued as a theapeutic strategy. To date one small molecule kinetic stabilising molecule (tafamidis) has been approved for FAP, and is being evaluated in clinical trials for other TTR amyloidoses. amyloid ﬁbril formation [66]. These amyloidogenic mutants are linked to the development of pathological amyloidoses, including familial amyloid polyneuropathy (FAP) [1, 13], familial amyloid cardiomyopathy (FAC) [37], amyloidotic vitreous opacities, carpal tunnel syndrome [57] and others. In old age, non-mutated TTR Overview: TTR is a homo-tetrameric protein which transports thyroxine in the plasma and cerebrospinal ﬂuid and retinol (vi- tamin A) in the plasma. Many disease causing mutations in the protein have been reported, many of which cause complex dissoci- ation and protein mis-assembly and deposition of toxic aggregates Nomenclature Common abreviation HGNC, UniProt Carrier proteins 5736 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 CD molecules Other protein targets ! CD molecules PDCD1, Q15116 pembrolizumab (Binding) (pKd 10) [9], nivolumab (Binding) (pKd 9.1) [29, 42, 43] The endogenous ligands for human PD-1 are programmed cell death 1 ligand 1 (PD-L1 aka CD274 ( CD274, Q9NZQ7)) and programmed cell death 1 ligand 2 (PD-L2; PDCD1LG2). These ligands are cell surface peptides, normally involved in immune system regulation. Many types of cancer cells evolve mechanisms to evade control and elimination by the immune system. Such mechanisms can include inhibition of so-called ’immune checkpoints’, which would normally be involved in the maintenance of immune homeostasis. An increasingly important area of clinical oncology research is the development of new agents which impede these evasion techniques, thereby switching immune vigilance back on, and effecting immune destruction of cancer cells. Three molecular targets of checkpoint inhibitors which are being extensively pursued are cytotoxic T-lymphocyte antigen 4 (CTLA4), programmed cell death 1 (PD-1), and programmed cell death ligand 1 (PD-L1). Using antibody-based therapies targeting these pathways, clinical responses have been reported in various tumour types, including melanoma, renal cell carcinoma [64] and non-small cell lung cancer [39, 51]. pembrolizumab is the ﬁrst-in-class, anti-PD-1 antibody to be approved by the US FDA, with ongoing clinical trials for nivolumab (e.g. NCT01673867, NCT01721746) and pidilizumab (NCT02077959, NCT01952769). Methyllysine reader proteins Other protein targets ! Chromatin-interacting transcriptional repressors ! Methyllysine reader proteins Overview: Methyllysine reader proteins bind to methylated proteins, such as histones, allowing regulation of g llysine reader proteins bind to methylated proteins, such as histones, allowing regulation of gene expression. l(3)mbt-like 3 (Drosophila) L3MBTL3, Q96JM7 UNC1215 (pKd 6.9) [38] Other protein targets ! CD molecules not possible in the Guide to PHARMACOLOGY; listed herein are selected members of the family targetted for therapeutic gain. not possible in the Guide to PHARMACOLOGY; listed herein are selected members of the family targetted for therapeutic gain. see CD73 ecto-5’-nucleotidase) or receptors (for example, see CD41 integrin, alpha 2b subunit). Many CDs are targetted for therapeutic gain using antibodies for the treatment of prolifera- tive disorders. A full listing of all the Clusters of Differentiation is Overview: Cluster of differentiation refers to an attempt to catalogue systematically a series of over 300 cell-surface proteins associated with immunotyping. Many members of the group have identiﬁed functions as enzymes (for example, Nomenclature CD2 CD3e molecule, epsilon (CD3-TCR complex) CD20 (membrane- spanning 4-domains, subfamily A, member 1) CD33 CD52 CD80 CD86 cytotoxic T-lymphocyte- associated protein 4 (CD152) Common abreviation – – – – – – – CTLA-4 HGNC, UniProt CD2, P06729 CD3E, P07766 MS4A1, P11836 CD33, P20138 CD52, P31358 CD80, P33681 CD86, P42081 CTLA4, P16410 Selective inhibitors – – – – – abatacept [84], belatacept [16] abatacept [84], belatacept [16] – Selective antagonists alefacept (Inhibition) [56, 89] – – – – – – – Antibodies – catumaxomab (Binding) [46], muromonab-CD3 (Binding) [24], otelixizumab (Binding) [7] ofatumumab (Binding) (pKd 9.9) [47], rituximab (Binding) (pKd 8.5) [78], ibritumomab tiuxetan (Binding), obinutuzumab (Binding) [2, 68], tositumomab (Binding) lintuzumab (Binding) (pKd 10) [8], gemtuzumab ozogamicin (Binding) [6] alemtuzumab (Binding) [22] – – ipilimumab (Binding) (pKd >9) [28], tremelimumab (Binding) (pKd 8.9) [30] CD molecules 5737 CD molecules 5737 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full se Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 Nomenclature Common abreviation HGNC, UniProt Antibodies Comments programmed cell death 1 (CD279) PD-1 PDCD1, Q15116 PDCD1, Q15116 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Methyllysine reader proteins 5738 Further Reading nd DNA h A t Thinnes CC et al. (2014) Targeting histone lysine demethylases - progress, challenges, and the future. Biochim Biophys Acta 1839: 1416-1432 [PMID:24859458] overy Part 2: Histone demethylation and DNA D:25857453] ng methyllysine readout Biochim Biophys Acta Thinnes CC et al. (2014) Targeting histone lysine demethylases - progress, challenges, and the future. Biochim Biophys Acta 1839: 1416-1432 [PMID:24859458] Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Methyllysine reader proteins 5738 Methyllysine reader proteins 5738 Methyllysine reader proteins 5738 Methyllysine reader proteins 5738 S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 Other protein targets ! Cytokines and growth factors Other protein targets ! Cytokines and growth factors Overview: cytokines and growth factors are a group of small proteins released from cells, which act upon the same cell or neighbouring cells, often with a role in immune regulation and/or proliferation. Listed herein are examples of cytokines and growth factors targetted for therapeutic beneﬁt. Overview: cytokines and growth factors are a group of small proteins released from cells, which act upon the sam Listed herein are examples of cytokines and growth factors targetted for therapeutic beneﬁt. Nomenclature interleukin 1, beta tumor necrosis factor vascular endothelial growth factor A HGNC, UniProt IL1B, P01584 TNF, P01375 VEGFA, P15692 Antagonists – – aﬂibercept (Inhibition) [10, 11, 82] Selective antagonists – etanercept (Inhibition) [18, 23] pegaptanib (Inhibition) [26, 61] Antibodies gevokizumab (Binding) (pKd 12.5) [36, 53, 71], canakinumab (Binding) (pKd 10.5) [27], rilonacept (Binding) [32, 55] golimumab (Inhibition) (pIC50 10.7) [77], inﬂiximab (Inhibition) (pKd 8.7) [44], adalimumab (Inhibition) (pKd >8) [75], certolizumab pegol (Inhibition) [60] ranibizumab (Inhibition) (pKd 9.8) [3], bevacizumab (Inhibition) (pIC50 8–8.3) [3] gevokizumab (Binding) (pKd 12.5) [36, 53, 71], canakinumab (Binding) (pKd 10.5) [27], rilonacept (Binding) [32, 55] Fatty acid-binding proteins Other protein targets ! Fatty acid-binding proteins Other protein targets ! Fatty acid-binding proteins [76]) or for interaction with metabolic enzymes. Although se- quence homology is limited, crystallographic studies suggest con- served 3D structures across the group of binding proteins. Overview: Fatty acid-binding proteins are low molecular weight (100-130 aa) chaperones for long chain fatty acids, fatty acyl CoA esters, eicosanoids, retinols, retinoic acids and related metabolites and are usually regarded as being responsible for allowing the oth- erwise hydrophobic ligands to be mobile in aqueous media. These binding proteins may perform functions extracellularly (e.g. in plasma) or transport these agents; to the nucleus to interact with nuclear receptors (principally PPARs and retinoic acid receptors fatty acid binding protein 1, liver fatty acid binding protein 2, intestinal Crystal structure of the human FABP5 [33]. A broader substrate speciﬁcity than other FABPs, binding two fatty acids per protein [83]. Crystal structure of the human FABP3 [87]. Fatty acid-binding proteins 5739 Fatty acid-binding proteins 5739 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 retinol binding protein 3, interstitial retinol binding protein 4, plasma RBP4, P02753 retinol binding protein 5, cellular RBP5, P82980 retinol binding protein 3, interstitial retinol binding protein 7, cellular retinaldehyde binding protein 1 cellular retinoic acid binding protein 1 cellular retinoic acid binding protein 2 ot RBP7, Q96R05 RLBP1, P12271 CRABP1, P29762 CRABP2, P29373 potency – 11-cis-retinal, 11-cis-retinol > 9-cis-retinal, 13-cis-retinal, 13-cis-retinol, all-trans-retinal, retinol [14] tretinoin > alitretinoin stearic acid > palmitic acid, oleic acid, linoleic acid, α-linolenic acid, arachidonic acid [70] – lthough not tested at all FABPs, BMS309403 exhibits high afﬁnity for FABP4 (pIC50 8.8) compared to FABP3 or FABP5 (pIC50 <6.6) [20, 81]. HTS01037 is reported to interfere with FABP4 ltiple pseudogenes for the FABPs have been identiﬁed in the human genome. cellular retinoic acid binding protein 2 CRABP2, P29373 retinol binding protein 7, cellular RBP7, Q96R05 PMP2, P02689 In silico modelling suggests that FABP8 can bind both fatty acids and cholesterol [50]. retinol binding protein 3, interstitial RBP3, P10745 peripheral myelin protein 2 fatty acid binding protein 9, testis fatty acid binding protein 7, brain FABP7, O15540 Crystal structure of the human FABP7 [4]. Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Further Reading nuclear receptors and gene transcription. Lipids 43: 1-17 [PMID:17882463] Chmurzy«ska A. (2006) The multigene family of fatty acid-binding proteins (FABPs): function, struc- ture and polymorphism. J. Appl. Genet. 47: 39-48 [PMID:16424607] Furuhashi M et al. (2008) Fatty acid-binding proteins: role in metabolic diseases and potential as drug targets. Nat Rev Drug Discov 7: 489-503 [PMID:18511927] Kralisch S et al. (2013) Adipocyte fatty acid binding protein: a novel adipokine involved in the pathogenesis of metabolic and vascular disease? Diabetologia 56: 10-21 [PMID:23052058] Schroeder F et al. (2008) Role of fatty acid binding proteins and long chain fatty acids in modulating Storch J et al. (2010) Tissue-speciﬁc functions in the fatty acid-binding protein family. J. Biol. Chem. 285: 32679-83 [PMID:20716527] Yamamoto T et al. (2009) Classiﬁcation of FABP isoforms and tissues based on quantitative evalua- tion of transcript levels of these isoforms in various rat tissues. Biotechnol. Lett. 31: 1695-701 [PMID:19565192] Storch J et al. (2010) Tissue-speciﬁc functions in the fatty acid-binding protein family. J. Biol. Chem. 285: 32679-83 [PMID:20716527] Yamamoto T et al. (2009) Classiﬁcation of FABP isoforms and tissues based on quantitative evalua- tion of transcript levels of these isoforms in various rat tissues. Biotechnol. Lett. 31: 1695-701 [PMID:19565192] Fatty acid-binding proteins 5740 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Fatty acid-binding proteins 5740 S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 Other protein targets ! Sigma receptors Overview: Although termed ‘receptors’, the evidence for coupling through conventional signalling pathways is lacking. Initially described as a subtype of opioid receptors, there is only a modest pharmacological overlap and no structural convergence with the G protein-coupled receptors. A wide range of compounds, ranging from psychoactive agents to antihistamines, have been observed to bind to these sites, which appear to be intracellular. d ‘receptors’, the evidence for coupling through conventional signalling pathways is lacking. Initially described as a subtype of opioid receptors, there is only a modest no structural convergence with the G protein-coupled receptors. A wide range of compounds, ranging from psychoactive agents to antihistamines, have been observed to t b i t ll l Overview: Although termed ‘receptors’, the evidence for coupling through conventional signalling pathways is lacking. Initially described as a subty pharmacological overlap and no structural convergence with the G protein-coupled receptors. A wide range of compounds, ranging from psychoactive a bind to these sites, which appear to be intracellular. 2 – PB-28 (pKi 8.3) [5], 1,3-ditolylguanidine (pKi 7.4) [45] – Guinea pig – – SM 21 (pIC50 7.2) [48] – [3H]-di-o-tolylguanidine (Agonist) There is no molecular correlate of the 2 receptor. Nomenclature sigma non-opioid intracellular receptor 1 2 HGNC, UniProt SIGMAR1, Q99720 – Agonists – PB-28 (pKi 8.3) [5], 1,3-ditolylguanidine (pKi 7.4) [45] – Guinea pig (Sub)family-selective agonists (RS)-PPCC (pKi 8.8) [67] – Selective agonists PRE-084 (pIC50 7.4) [80], (+)-SK&F10047 – Antagonists (-)-pentazocine SM 21 (pIC50 7.2) [48] Selective antagonists NE-100 (pIC50 8.4) [62], BD-1047 (pIC50 7.4) [54] – Labelled ligands [3H]pentazocine (Agonist) [3H]-di-o-tolylguanidine (Agonist) Comments – There is no molecular correlate of the 2 receptor. Comments: (-)-pentazocine also shows activity at opioid receptors. Sigma receptors Other protein targets ! Sigma receptors Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Further Reading Dubrovsky B. (2006) Neurosteroids, neuroactive steroids, and symptoms of affective disorders. Phar- macol. Biochem. Behav. 84: 644-55 [PMID:16962651] Guitart X et al. (2004) Sigma receptors: biology and therapeutic potential. Psychopharmacology (Berl.) 174: 301-19 [PMID:15197533] Matsumoto RR et al. (2003) Sigma receptors: potential medications development target for anti- cocaine agents. Eur. J. Pharmacol. 469: 1-12 [PMID:12782179] de Medina P et al. (2011) Importance of cholesterol and oxysterols metabolism in the pharmacology of tamoxifen and other AEBS ligands. Chem. Phys. Lipids 164: 432-7 [PMID:21641337] Sigma receptors 5741 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 Other protein targets ! Tubulins Other protein targets ! Tubulins Overview: Tubulins are a family of intracellular proteins most commonly associated with microtubules, part of the cytoskeleton. They are exploited for therapeutic gain in cancer chemotherapy as targets for agents derived from a variety of natural products: taxanes, colchicine and vinca alkaloids. These are thought to act primarily through β-tubulin, thereby interfering with the normal processes of tubulin polymer formation and disassembly. Nomenclature tubulin, alpha 1a tubulin, alpha 4a tubulin, beta class I tubulin, beta 3 class III tubulin, beta 4B class IVb tubulin, beta 8 class VIII HGNC, UniProt TUBA1A, Q71U36 TUBA4A, P68366 TUBB, P07437 TUBB3, Q13509 TUBB4B, P68371 TUBB8, Q3ZCM7 Inhibitors – – vinblastine (pIC50 9), vincristine – – – (Sub)family-selective inhibitors – – eribulin (pIC50 8.2) [59], paclitaxel (Mitotic cell cycle arrest in A431 cells) (pEC50 8.1) [63], colchicine (pIC50 8) [12], cabazitaxel, docetaxel, ixabepilone – – – Further Reading Kaur R et al. (2014) Recent developments in tubulin polymerization inhibitors: An overview. Eur J Med Chem 87: 89-124 [PMID:25240869] Lu Y et al. (2012) An overview of tubulin inhibitors that interact with the colchicine binding site. Pharm. Res. 29: 2943-71 [PMID:22814904] Perdiz D et al. (2011) The ins and outs of tubulin acetylation: more than just a post-translational modiﬁcation? Cell. Signal. 23: 763-71 [PMID:20940043] Schappi JM et al. (2014) Tubulin, actin and heterotrimeric G proteins: coordination of signaling and structure. Biochim. Biophys. Acta 1838: 674-81 [PMID:24071592] Song Y et al. (2015) Post-translational modiﬁcations of tubulin: pathways to functional diversity of microtubules. Trends Cell Biol. 25: 125-36 [PMID:25468068] Yu I et al. (2015) Writing and Reading the Tubulin Code. J. Biol. Chem. 290: 17163-72 [PMID:25957412] Nomenclature tubulin, alpha 1a tubulin, alpha 4a tubulin, beta class I tubulin, beta 3 class III tubulin, beta 4B class IVb tubulin, beta 8 class VIII HGNC, UniProt TUBA1A, Q71U36 TUBA4A, P68366 TUBB, P07437 TUBB3, Q13509 TUBB4B, P68371 TUBB8, Q3ZCM7 Inhibitors – – vinblastine (pIC50 9), vincristine – – – (Sub)family-selective inhibitors – – eribulin (pIC50 8.2) [59], paclitaxel (Mitotic cell cycle arrest in A431 cells) (pEC50 8.1) [63], colchicine (pIC50 8) [12], cabazitaxel, docetaxel, ixabepilone – – – Kaur R et al. (2014) Recent developments in tubulin polymerization inhibitors: An overview. Eur J Med Chem 87: 89-124 [PMID:25240869] Lu Y et al. (2012) An overview of tubulin inhibitors that interact with the colchicine binding site. Pharm. Res. 29: 2943-71 [PMID:22814904] Perdiz D et al. (2011) The ins and outs of tubulin acetylation: more than just a post-translational modiﬁcation? Cell. Signal. 23: 763-71 [PMID:20940043] Schappi JM et al. (2014) Tubulin, actin and heterotrimeric G proteins: coordination of signaling and structure. Biochim. Biophys. Acta 1838: 674-81 [PMID:24071592] Song Y et al. (2015) Post-translational modiﬁcations of tubulin: pathways to functional diversity of microtubules. Trends Cell Biol. 25: 125-36 [PMID:25468068] Yu I et al. (2015) Writing and Reading the Tubulin Code. J. Biol. Chem. 290: 17163-72 [PMID:25957412] Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full Further Reading Tubulins 5742 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full S.P.H. Alexander et al. The Concise Guide to PHARMACOLOGY 2015/16: Overview. British Journal of Pharmacology (2015) 172, 5729–5743 ( ) [ 22. Ginaldi L et al. (1998) [9593475] g 83. Thompson J et al. (1997) [9054409] y 60. Nesbitt A et al. (2007) [17636564] ( ) [ ] 23. Goldenberg MM. (1999) [10090426] 61. Nimjee SM et al. (2005) [15660527] 84. Vicente Rabaneda EF et al. (2013) [23899231] g 24. Goldstein G. (1987) [3105134] g ( ) [ 11. Chu QS. (2009) [19236257] 53. Masat L et al. (2009) IL-1β binding antibod- ies and fragments thereof. Patent number: 53. Masat L et al. (2009) IL-1β binding antibod- ies and fragments thereof. Patent number: US7531166. Assignee: Xoma Technology, Ltd. Priority date: 27/02/2015. Publication date: 12/03/2010. ( ) [ ] 12. Cifuentes M et al. (2006) [16504507] g ( ) 36. Issafras H et al. (2014) [24194526] 76. Schroeder F et al. (2008) [17882463] ( ) [ 13. Coelho T. (1996) [8894411] ( ) [ 13. Coelho T. (1996) [8894411] ( ) 37. Jacobson DR et al. (1997) [9017939] 77. Shealy DJ et al. (2010) [20519961] ( ) 14. Crabb JW et al. (1998) [9541407] 38. James LI et al. (2013) [23292653] y 78. Stein R et al. (2004) [15102696] J ( ) [ ] 15. Cziraky MJ et al. (1993) [8137606] ( ) [ ] 39. Johnson DB et al. (2014) [25096781] 79. Structural Genomics Consortium. PFI- 3: Selective chemical probe for SMARCA bromodomains. Accessed on 10/11/2014. http://www.thesgc.org. y J ( ) [ ] 16. El-Charabaty E et al. (2012) [22992146] J ( ) [ 40. Kanji S et al. (2001) [11714212] ( ) [ ] 19. Friedel HA et al. (1994) [7528134] ( ) [ ] 19. Friedel HA et al. (1994) [7528134] 43. Korman AJ et al. (2006) Human mon- oclonal antibodies to programmed death 1(pd-1) and methods for treating cancer us- ing anti-pd-1 antibodies alone or in com- bination with other immunotherapeutics. Patent number: WO2006121168. Assignee: Ono Pharmaceutical Co. Priority date: 09/05/2005. Publication date: 02/03/2015. ( ) [ 35. Hug C et al. (2004) [15210937] ( ) [ 35. Hug C et al. (2004) [15210937] 24. Goldstein G. (1987) [3105134] 85. Westermark P et al. (1990) [2320592] 62. Okuyama S et al. (1993) [7901723] p g g 80. Su TP et al. (1991) [1658302] 57. Murakami K et al. (1999) [10403814] 80. Su TP et al. (1991) [1658302] ( ) [ ] 20. Furuhashi M et al. (2007) [17554340] ( ) [ ] 81. Sulsky R et al. (2007) [17502136] 58. Nakase J et al. (2009) [19398784] ( ) [ ] 21. Garcia-Calvo M et al. (2005) [15928087] y 82. Tang PA et al. (2013) [24179482] 59. Narayan S et al. (2011) [21324687] ( ) [ 22. Ginaldi L et al. (1998) [9593475] p ( ) [ ] 42. Kline J et al. (2010) [21154117] 18. Feldman M et al. (1998) [9865320] p 42. Kline J et al. (2010) [21154117] ( ) [ 56. Mitchell P. (2002) [12089534] , y Publication date: 15/10/1998. 45. Lever JR et al. (2006) [16463398] 71. Roell MK et al. (2010) [20410301] 4. Balendiran GK et al. (2000) [10854433] 46. Linke R et al. (2010) [20190561] ( ) 72. SGC. I-CBP112 - a CREBBP/EP300-selective chemical probe. Accessed on 03/03/2015. thesgc.org. 29. Hall RD et al. (2013) [23302904] ( ) [ 5. Berardi F et al. (1996) [8568804] 47. Liu Q. (2013) Fully human antibodies against human cd20. Patent number: WO2013007052. Assignee: Qingfa Liu. Priority date: 13/07/2011. Publication date: 17/01/2013. ( ) 6. Bernstein ID. (2000) [10720144] 30. Hanson DC et al. (2004) Human mono- clonal antibodies to CTLA-4. Patent num- ber: US6682736 B1. Assignee: Abgenix, Inc., Pﬁzer Inc.. Priority date: 22/12/1998. Publi- cation date: 27/01/2004. y cation date: 27/01/2004. 48. Mach RH et al. (1999) [10096443] g g 74. Sacchettini JC et al. (1989) [2671390] 31. Hertzel AV et al. (2009) [19754198] 49. Maeda K et al. (1996) [8619847] 75. Salfeld JG et al. (2001) Human antibod- ies that bind human TNFα. Patent num- ber: US6258562. Assignee: Basf Aktienge- sellschaft. Priority date: 09/02/1996. Pub- lication date: 10/07/2001. 32. Hoffman HM et al. (2008) [18668535] 50. Majava V et al. (2010) [20421974] ( ) [ 33. Hohoff C et al. (1999) [10493790] 51. Malas S et al. (2014) [24969320] ( ) 34. Holmer E et al. (1986) [3744129] 10. Chang AA et al. (2013) [24144450] 52. Mao X et al. (2006) [16622416] g ( ) [ 11. Chu QS. (2009) [19236257] ( ) [ 7. Bolt S et al. (1993) [8436176] ( ) 7. Bolt S et al. (1993) [8436176] g g 73. SGC. GSK2801: A Selective Chemical Probe for BAZ2B/A bromodomains. Accessed on 03/03/2015. thesgc.org. ( ) 40. Kanji S et al. (2001) [11714212] 54. Matsumoto RR et al. (1995) [8566098] y ( ) [ 17. Eriksson BI et al. (1995) [7667822] j ( ) [ ] 41. Kapur S et al. (2001) [11463021] 55. McDermott MF. (2009) [19649332] References Novartis Ag. Priority date: 22/08/2000. Pub- lication date: 04/11/2008. 66. Penchala SC et al. (2013) [23716704] anti-TNF antibodies. Patent number: US5656272. Assignee: New York Univer- sity Medical Center, Centocor, Inc.. Pri- ority date: 18/03/1991. Publication date: 12/08/1997. 1. ANDRADE C. (1952) [12978172] 1. ANDRADE C. (1952) [12978172] ( ) [ ] 2. Alduaij W et al. (2011) [21378274] ( ) [ ] 67. Prezzavento O et al. (2007) [17328523] 28. Halk EL et al. (2001) Human ctla-4 an- tibodies and their uses. Patent number: WO2001014424. Assignee: Medarex Inc. Priority date: 24/08/1999. Publication date: 01/03/2001. j ( ) [ ] 3. Baca M et al. (1998) Anti-vegf antibodies. Patent number: WO1998045331. Assignee: Genentech Inc,. Priority date: 07/04/1997. Publication date: 15/10/1998. ( ) 68. Reslan L et al. (2013) [23537278] 69. Richieri GV et al. (1994) [7929039] 70. Richieri GV et al. (2000) [10852718] ( ) [ ] 8. Caron PC et al. (1992) [1458463] ( ) [ ] 8. Caron PC et al. (1992) [1458463] ( ) [ ] 9. Carven GJ et al. (2010) Antibodies to human programmed death receptor PD-1. Patent number: US20100266617. Assignee: Organon NV. Priority date: 13/06/2008. Publication date: 21/10/2010. ( ) [ ] 25. Gotti R et al. (2013) [23598032] 25. Gotti R et al. (2013) [23598032] 86. Yamauchi T et al. (2003) [12802337] 63. Ouyang X et al. (2006) [16377187] ( ) [ ] 26. Gragoudas ES et al. (2004) [15625332] 26. Gragoudas ES et al. (2004) [15625332] ( ) [ ] 87. Young AC et al. (1994) [7922029] y g ( ) 64. Pal SK et al. (2014) [24892254] 44. Le J et al. (1997) Methods of treating TNF- α-mediated Crohn’s disease using chimeric g ( ) [ ] 27. Gram H et al. (2008) Antibodies to human IL- ( ) [ ] 65. Paolucci F et al. (2002) [12383040] ( ) 1β. Patent number: US7446175. Assignee: g 88. Zwicker BL et al. (2013) [23 65. Paolucci F et al. (2002) [12 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full ( ) 1β. Patent number: US7446175. References 5743 References 5743 References 5743 Searchable database: http://www.guidetopharmacology.org/index.jsp Full Contents of ConciseGuide: http://onlinelibrary.wiley.com/doi/10.1111/bph.13347/full
https://openalex.org/W2073630173	https://europepmc.org/articles/pmc3068169?pdf=render	English	null	Environmental Salinity Determines the Specificity and Need for Tat-Dependent Secretion of the YwbN Protein in Bacillus subtilis	PloS one	2,011	cc-by	10,137	Abstract This is an open-access article distributed under the terms unrestricted use, distribution, and reproduction in any medium, provided the original author and Funding: This research was supported by the Commission of the European Union (CEU) projects LSHG-CT-2004-503468, LSHG-CT-2004-005257, LSHM-CT-2006- 019064, LSHG-CT-2006-037469 and PITN-GA-2008-215524, the transnational Systems Biology of Microorganisms (SysMO) initiative through projects BACELL SysMO1 and 2, the European Science Foundation under the EUROCORES Programme EuroSCOPE, and grant 04-EScope 01-011 from the Research Council for Earth and Life Sciences of the Netherlands Organization for Scientific Research, the "Deutsche Forschungsgemeinschaft", the "Fonds der Chemischen Industrie" and the "Bundesministerium fu¨r Bildung und Forschung" (03ZIK012). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: j.m.van.dijl@med.umcg.nl Environmental Salinity Determines the Specificity and Need for Tat-Dependent Secretion of the YwbN Protein in Bacillus subtilis Rene´ van der Ploeg1, Ulrike Ma¨der2, Georg Homuth2, Marc Schaffer2, Emma L. Denham1, Carmine G. Monteferrante1, Marcus Miethke3, Mohamed A. Marahiel3, Colin R. Harwood4, Theresa Winter5, Michael Hecker5, Haike Antelmann5, Jan Maarten van Dijl1* 1 Department of Medical Microbiology, University Medical Center Groningen and University of Groningen, Groningen, The Netherlands, 2 Interfaculty Institute for Genetics and Functional Genomics, Department for Functional Genomics, Ernst-Moritz-Arndt-University Greifswald, Greifswald, Germany, 3 Department of Chemistry/ Biochemistry, Philipps-University Marburg, Marburg, Germany, 4 Centre for Bacterial Cell Biology, Institute of Cell and Molecular Biosciences, Newcastle University, Newcastle upon Tyne, United Kingdom, 5 Institut fu¨r Mikrobiologie und Molekularbiologie, Ernst-Moritz-Arndt-Universita¨t Greifswald, Greifswald, Germany PLoS ONE \| www.plosone.org Abstract Twin-arginine protein translocation (Tat) pathways are required for transport of folded proteins across bacterial, archaeal and chloroplast membranes. Recent studies indicate that Tat has evolved into a mainstream pathway for protein secretion in certain halophilic archaea, which thrive in highly saline environments. Here, we investigated the effects of environmental salinity on Tat-dependent protein secretion by the Gram-positive soil bacterium Bacillus subtilis, which encounters widely differing salt concentrations in its natural habitats. The results show that environmental salinity determines the specificity and need for Tat-dependent secretion of the Dyp-type peroxidase YwbN in B. subtilis. Under high salinity growth conditions, at least three Tat translocase subunits, namely TatAd, TatAy and TatCy, are involved in the secretion of YwbN. Yet, a significant level of Tat-independent YwbN secretion is also observed under these conditions. When B. subtilis is grown in medium with 1% NaCl or without NaCl, the secretion of YwbN depends strictly on the previously described ‘‘minimal Tat translocase’’ consisting of the TatAy and TatCy subunits. Notably, in medium without NaCl, both tatAyCy and ywbN mutants display significantly reduced exponential growth rates and severe cell lysis. This is due to a critical role of secreted YwbN in the acquisition of iron under these conditions. Taken together, our findings show that environmental conditions, such as salinity, can determine the specificity and need for the secretion of a bacterial Tat substrate. Citation: van der Ploeg R, Ma¨der U, Homuth G, Schaffer M, Denham EL, et al. (2011) Environmental Salinity Determines the Specificity and Need for Tat- Dependent Secretion of the YwbN Protein in Bacillus subtilis. PLoS ONE 6(3): e18140. doi:10.1371/journal.pone.0018140 Editor: Pierre Cornelis, Vrije Universiteit Brussel, Belgium Editor: Pierre Cornelis, Vrije Universiteit Brussel, Belgium Editor: Pierre Cornelis, Vrije Universiteit Brussel, Belgium Received November 9, 2010; Accepted February 24, 2011; Published March 30, 2011 Received November 9, 2010; Accepted February 24, 2011; Published March 30, 2011 der Ploeg et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits tion, and reproduction in any medium, provided the original author and source are credited. Copyright: 2011 van der Ploeg et al. This is an open-access article distributed under the terms of the Creative Commons Attribu unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Copyright: 2011 van der Ploeg et al. Results TatAyCy-independent secretion of YwbN at high salinity To investigate whether high salinity might impact on Tat- dependent secretion, we investigated the secretion of YwbN in LB medium with a final salt concentration of 6%, which is comparable to that of highly saline environments or brine. It should be noted that the ‘standard’ LB medium, as originally defined by Luria and Bertani and used in all our previously published secretion studies, already contains 1% NaCl. Thus, its salinity resembles that of brackish water or the cytoplasm. LB with no added salt would reflect the situation encountered in fresh water. To facilitate YwbN detection, this protein was provided with a C-terminal Myc-epitope, and the corresponding gene construct was expressed ectopically from the xylose-inducible xylA promoter. The cassette encoding the xylA-ywbN-myc fusion was named X-YwbN (Table 1). When the cells reached an OD600 of 2, expression of ywbN was induced and growth was continued for 3 hours. Cellular and growth medium fractions were collected and used for SDS-PAGE and Western blotting to monitor YwbN secretion. The results revealed a significant impact of high salinity on the specificity of YwbN secretion (Figure 1, upper panels). Clearly, YwbN secretion was no longer strictly dependent on the tatAyCy genes when cells were grown in medium with 6% salt. Although deletion of tatAyCy resulted in a strong reduction of the extracellular level of YwbN, the tatAyCy mutant nevertheless secreted substantial amounts of this protein. This contrasts markedly with the secretion of YwbN by cells grown in LB with 1% or no salt, which is strictly TatAyCy-dependent (Figure 1, middle and lower panels). Remarkably, at 6% NaCl the secretion of YwbN was also reduced in a tatAdCd mutant. This suggested for the first time an involvement of TatAd and/or TatCd in YwbN secretion in a strain in which tat gene expression was not genetically engineered. Furthermore, this observation was intrigu- ing because a function for tatAdCd had only been demonstrated so far under phosphate starvation conditions [18]. Consistent with the finding that tatAyCy or tatAdCd mutant strains secreted reduced amounts of mature YwbN, the secretion of YwbN was also strongly reduced in the total-tat2 mutant strain (Figure 1), which lacks all B. subtilis tat genes (Table 1). These data show that YwbN secretion was partially Tat-independent when cells were grown in the presence of 6% salt. Results To verify that this Tat-independent secretion of YwbN was not due to cell lysis, the locations of the Sec-dependently secreted LipA protein and the cytoplasmic marker protein TrxA were verified by Western blotting. No extracellular TrxA was detected, indicating that the observed Tat- independent secretion of YwbN was not due to cell lysis (Figure S1). Furthermore, secreted LipA was readily detectable (Figure 1), and Sypro Ruby-staining of the respective gels did not reveal any major differences in the total amounts of protein secreted by the parental and tat mutant strains grown in LB with 6% salt (Figure S2). Taken together, these findings show that the TatAyCy translocase is not exclusively involved in YwbN secretion under high salinity growth conditions, and that TatAd and/or TatCd are l i l d i thi p p [ ] In silico analyses suggested that 69 proteins of B. subtilis have signal peptides with potential RR-motifs [19]. However, proteo- mics and molecular biological analyses revealed that only two of these - YwbN and PhoD – were strictly dependent on a functional Tat translocase for secretion. The signal peptides of the B. subtilis QcrA and YkuE proteins were shown to direct Tat-dependent protein secretion in Streptomyces [28], but the Tat-dependence of these proteins in B. subtilis remains to be shown. Recently, we demonstrated that the esterase LipA can be secreted via both the Sec and Tat pathways of B. subtilis [29]. This became evident under conditions of LipA hyperproduction, indicating that an overflow mechanism exists to re-direct LipA from the normally used Sec-dependent export route into the Tat-dependent route. Importantly, if LipA is produced at wild-type levels this protein is secreted Sec-dependently and, in accordance with this view, no Tat-dependent secretion of the LipA protein can be demonstrated in the sequenced B. subtilis strain 168 [19,29]. These observations on LipA secretion suggest that seemingly Sec-dependent proteins can be targeted to the Tat pathway depending on intracellular and perhaps even extracellular conditions. In this context, it is noteworthy that the Tat machinery is extensively used for protein secretion by certain halophilic archaea, such as Haloferax volcanii and Haloarcula hispanica, indicating that Tat has evolved into a mainstream secretion pathway for some organisms that grow in highly saline milieus [30,31,32,33,34,35]. This raised the question as to what extent salinity can impact on Tat-dependent protein secretion by microorganisms, like the soil bacterium B. Introduction folding of Sec-dependent proteins occurs post-translocationally [7]. The transport of proteins across biological membranes and their subsequent secretion into external milieus are vital processes for all known microorganisms. These processes depend on the activity of dedicated molecular machines. A first critical step in protein secretion is the passage of transported proteins through the cytoplasmic membrane, which can occur either in an unfolded state via the general secretion (Sec) machinery, or in a folded state via the twin-arginine translocation (Tat) machinery [1,2,3,4,5]. Accordingly, the Sec and Tat machines work independently of each other, using distinct mechanisms for protein translocation. Acceptance of a protein by the Sec or Tat complexes is dictated by the presence of an N-terminal signal peptide with or without a selective recognition motif for Tat (i.e. the twin-arginine motif) [6], and the folding state of the transported protein [7,8]. In general, Tat-dependent proteins fold prior to translocation, while the Early studies have defined the twin-arginine (RR-) motif in the N-region of signal peptides as S/T-R-R-x-F-L-K [6,9,10,11,12]. However, this motif is not always strictly conserved in Tat- dependently exported proteins. The observed natural variations, as well as site-directed mutagenesis studies, have defined a more general RR-consensus sequence as R/K-R-x-#-#, where # is a hydrophobic residue [13,14,15]. This RR-motif is recognized by the membrane-embedded Tat machinery of which two general types are known. Gram-negative bacteria, like Escherichia coli, contain a TatABC-type machinery consisting of three Tat proteins (TatA, TatB and TatC) that are indispensable for translocation [16,17]. In contrast, most Gram-positive bacteria contain a ‘‘minimal’’ TatAC machinery that lacks the TatB protein [18,19]. In these minimal TatAC translocases, the role of TatB is fulfilled by a bifunctional TatA protein [20,21]. The precise PLoS ONE \| www.plosone.org 1 March 2011 \| Volume 6 \| Issue 3 \| e18140 PLoS ONE \| www.plosone.org March 2011 \| Volume 6 \| Issue 3 \| e18140 Salinity-Determined Usage of Tat in Bacillus Salinity-Determined Usage of Tat in Bacillus mechanism of Tat-dependent protein translocation is currently unknown, but studies in E. coli and the thylakoids of plant chloroplasts have shown that RR-signal peptide recognition involves a TatB-TatC complex [22,23,24]. Subsequently, TatBC-precursor complexes merge with TatA sub-complexes to facilitate the translocation process in such a way that folded proteins of varying sizes and complexity can pass across the membrane [24,25]. Introduction dependent secretion of YwbN at high salinity does not only involve the TatAy and TatCy subunits as previously shown, but also the TatAd subunit. If no NaCl is included in the LB medium, only TatAy and TatCy are required for YwbN secretion. Under these conditions TatAy and TatCy are of major importance for growth and cell viability, which relates to an important role of YwbN in iron acquisition. [ ] Notably, some organisms contain multiple Tat translocases [26]. This was shown for the Gram-positive bacterium Bacillus subtilis, which contains two minimal TatAC translocases named TatAdCd and TatAyCy that can function independently of each other [20]. The tatAd-tatCd and tatAy-tatCy genes coding for these translocases are organised in operons at separate genomic loci [18]. In what follows these operons are referred to as tatAdCd and tatAyCy, respectively. B. subtilis contains a third tatA gene (tatAc) with no demonstrable role in Tat-dependent protein transport [19,20]. The specificities of the TatAdCd and TatAyCy translocase complexes are non-identical, but overlapping. Only two B. subtilis proteins, YwbN and PhoD, have been shown to be secreted in a strictly Tat-dependent manner [19,20]. YwbN is a member of the family of Dyp-type peroxidases that contain an iron-heme co- factor [27]. This protein is a preferred substrate for the constitutively expressed TatAyCy translocase when cells are grown in standard Luria-Bertani (LB) medium. In contrast, secretion of the phosphodiesterase PhoD is strictly dependent on the TatAdCd translocase. Notably, the phoD gene is located upstream of the tatAd-tatCd genes, and all three genes are induced under phosphate starvation conditions [18]. Results subtilis, that live in ecological niches where the salinity can fluctuate markedly. As a first approach to address this question, we studied the influence of different NaCl concentrations in the growth medium of B. subtilis on the Tat-dependent secretion of the YwbN protein. The results show that the NaCl content of the growth medium has a strong impact on the Tat-dependency of YwbN secretion. Remarkably, when cells were grown at high salinity some Tat- independent secretion of YwbN was observed. Moreover, the Tat- March 2011 \| Volume 6 \| Issue 3 \| e18140 March 2011 \| Volume 6 \| Issue 3 \| e18140 2 PLoS ONE \| www.plosone.org Salinity-Determined Usage of Tat in Bacillus Table 1. Strains and Plasmids used in this study. y Plasmids Relevant properties Reference pGDL48 Contains multiple cloning site to place genes under the control of the erythromycin promoter; 6.8 kb; ApR; KmR [54] pCAd pGDL48 containing the tatAd gene; 7.0 kb; ApR; KmR [36] pCCd pGDL48 containing the tatCd gene; 7.5 kb; ApR; KmR [36] pCACd pGDL48 containing the tatAd-tatCd operon; 7.7 kb; ApR; KmR [36] pCAy pGDL48 containing the tatAy gene; 7.0 kb; ApR; KmR [20] pCCy pGDL48 containing the tatCy gene; 7.5 kb; ApR; KmR [20] pCACy pGDL48 containing the tatAy-tatCy operon; 7.7 kb; ApR; KmR [20] pXTc Expression vector with the xylose-inducible xylA promoter; ApR; TcR [55] pXTc-ywbN-myc pXTc with the ywbN-myc gene; results in the production of YwbN with a C-terminal Myc tag This study B. subtilis 168 trpC2 [56] tatAy trpC2; tatAy::Em; EmR [20] tatCy trpC2; tatCy::Sp; SpR [18] tatAyCy trpC2; tatAy-tatCy::Sp; SpR [19] tatCd trpC2; tatCd::Km; KmR [18] tatAdCd trpC2; tatAd-tatCd::Km; KmR; [20] tatAy-tatAd trpC2; tatAy::Em; EmR; tatAd::Km; KmR [20] tatCy-tatCd trpC2; tatCy::Sp; SpR; tatCd::Km; KmR total-tat trpC2, tatAc::Em; EmR; tatAy-tatCy::Sp; SpR; tatAd-tatCd::Cm; CmR [19] total-tat2 trpC2; tatAd-tatCd::Km; KmR; tatAy-tatCy::Sp; SpR; tatAc::Em; EmR [20] 168 X-ywbN trpC2; amyE::xylA-ywbN-myc; CmR [19] tatAyCy X-ywbN trpC2; tatAy-tatCy::Sp; SpR; amyE::xylA-ywbN-myc; CmR [19] tatAdCd X-ywbN trpC2; tatAd-tatCd::Km; KmR; amyE::xylA-ywbN-myc; CmR [19] tatAdCd X-ywbN2 trpC2; tatAd-tatCd::Cm; CmR; amyE::xylA-ywbN-myc; TcR This study total-tat2 X-ywbN trpC2; tatAd-tatCd::Km; KmR; tatAy-tatCy::Sp; SpR; tatAc::Em; EmR; amyE::xylA-ywbN-myc; CmR [19] ywbL trpC2, ywbL::pMutin2::Em; EmR (BFA3211) [57] ywbM trpC2, ywbM::pMutin2::Em; EmR (BFA3212) [57] ywbN trpC2, ywbN::pMutin2::Em; EmR (BFA3213) [57] ywbN X-ywbN trpC2, ywbN::pMutin2::Em; EmR amyE::xylA-ywbN-myc: CmR This study doi:10.1371/journal.pone.0018140.t001 transcript that was detectable at comparable amounts in cells grown in the presence of 1% NaCl or without NaCl (Figure 2). Results When cells were grown in LB with 6% NaCl, the level of this tatAd transcript was elevated. Most likely, it corresponds to the complete tatAd gene and part of the tatCd gene (data not shown). Conversely, the levels of a tatAyCy-specific transcript of ,1.2 kb showed a clear decrease in cells grown in medium with 6% salt compared to cells grown in medium with 1% salt or no salt. The same is true for some tatAyCy-specific read-through transcripts that accumulate in the zone of the rRNA bands, where the electrophoretic separation of mRNAs is somewhat impaired. The ,1.2 kb transcript and the read-through transcripts originate from a promoter immediately upstream of tatAy ([37]; data not shown). The levels of a less abundant ,2.3 kb transcript corresponding to moaC, rex, tatAy and tatCy were not affected by growth medium salinity (Figure 2). These findings imply that the relative levels of TatAd and TatAyCy production can differ in cells grown in LB media of high or low salinity, which may explain the altered Tat component specificity observed for YwbN secretion in medium with 6% salt as shown in Figure 1. Notably, the finding that there was no tatCd transcript detectable when cells were grown in LB with 6% salt It should be noted that the cellular amounts of YwbN-Myc decreased when the salt concentration in the growth medium was reduced (Figure 1). It is currently not known why this occurs, but it may relate to the xylose-induced expression of the xylA-ywbN-myc cassette. Furthermore, we also observed that the total yields of secreted LipA became lower when the NaCl concentration in the medium was reduced (Figure 1). This decrease seems specific for LipA as the total amounts of secreted proteins were not substantially affected by differing salt concentrations (Figure S2). Again, this may relate to variation in gene expression under the different conditions, but it is also conceivable that salt has an impact on certain stages in the LipA secretion process. PLoS ONE \| www.plosone.org TatAd, TatAy and TatCy are required for efficient YwbN secretion at high salinity Taken together, these observations show that TatAd, TatAy and TatCy are involved in the Tat-dependent secretion of YwbN by cells growing in LB with 6% NaCl. This is reminiscent of the situation in Gram-negative bacteria that contain a three-component TatABC translocase. Furthermore, the results indicate that a separate expression of tatAy or tatCy can strongly interfere with YwbN secretion by cells grown at high salinity. Consistent with the view that TatAd has a role in YwbN secretion when cells are grown in LB medium with 6% NaCl, the plasmid-directed expression of tatAd was sufficient for full restoration of the secretion of YwbN by tatAdCd mutant cells under these conditions (Figure 3, compare the first two lanes for the tatAdCd mutant strain with the control). Furthermore, YwbN secretion under these conditions was also fully restored by the expression of plasmid-borne tatAdCd or tatAyCy genes. Conversely, no restoration of YwbN secretion was observed in the tatAdCd mutant upon expression of a plasmid-borne tatCd gene alone, and YwbN secretion by the tatAdCd mutant was even completely suppressed by the presence of a plasmid-borne copy of tatCy. The reduced secretion of YwbN by a tatAyCy mutant grown in LB with 6% NaCl was complemented by plasmid-borne tatAyCy, but it was completely inhibited by the presence of plasmid-borne copies of either tatAy or tatCy alone (Figure 3, rightside panels). Lastly, TatAd, TatAy and TatCy are required for efficient YwbN secretion at high salinity Recent studies by Eijlander et al. [36] indicated that engineered tatAdCd expression was able to compensate for the absence of the tatAyCy genes in the secretion of YwbN. We therefore investigated the levels of tatAdCd- and tatAyCy- specific transcripts in cells grown in LB media with varying NaCl concentrations by Northern blotting. Interestingly, we identified a 0.8 kb tatAd-specific PLoS ONE \| www.plosone.org March 2011 \| Volume 6 \| Issue 3 \| e18140 3 Salinity-Determined Usage of Tat in Bacillus Figure 1. Tat-dependence of YwbN secretion in media with differing salinity. Cell and growth medium fractions of B. subtilis tat mutant strains and the parental strain 168 were separated by centrifugation and used for SDS-PAGE, Western blotting and immunodetection of YwbN-Myc and LipA with specific antibodies. From top to bottom the panels show results obtained for cells grown in LB with 6%, 1% or no added NaCl. Protein loading was corrected for OD600. The YwbN-Myc (YwbN and preYwbN) and LipA proteins, and Mw markers are indicated. A slight ‘smiling effect’ as observed for the YwbN and LipA bands in the growth medium sample of the total-tat2 mutant grown in LB with 6% NaCl is due to some residual salt in the sample (see also Figure S2). doi:10.1371/journal.pone.0018140.g001 Figure 1. Tat-dependence of YwbN secretion in media with differing salinity. Cell and growth medium fractions of B. subtilis tat mutant strains and the parental strain 168 were separated by centrifugation and used for SDS-PAGE, Western blotting and immunodetection of YwbN-Myc and LipA with specific antibodies. From top to bottom the panels show results obtained for cells grown in LB with 6%, 1% or no added NaCl. Protein loading was corrected for OD600. The YwbN-Myc (YwbN and preYwbN) and LipA proteins, and Mw markers are indicated. A slight ‘smiling effect’ as observed for the YwbN and LipA bands in the growth medium sample of the total-tat2 mutant grown in LB with 6% NaCl is due to some residual salt in the sample (see also Figure S2). doi:10.1371/journal.pone.0018140.g001 indicated that TatCd has no, or only a very minor role in YwbN secretion during growth at high salinity. expression of plasmid-borne tatAdCd genes in the tatAyCy mutant did not restore secretion of YwbN, which contrasts with the afore- mentioned findings by Eijlander et al. [36]. PLoS ONE \| www.plosone.org Salinity-Determined Usage of Tat in Bacillus Salinity-Determined Usage of Tat in Bacillus Figure 2. Northern blotting analysis of tatAd and tatAy transcription in cells grown in LB media of differing salinity. RNA isolation and Northern blotting were performed as described in the Materials and Methods section. For each sample, 5 mg of RNA per lane were loaded as indicated. Molecular markers and the positions of the 16S and 23S rRNA are indicated. doi:10.1371/journal.pone.0018140.g002 mutants exhibited a lower growth yield as compared to the tatAyCy proficient strains and they were even unable to enter directly into a stationary phase of growth. Instead, mutants lacking tatAyCy displayed a severe lysis phenotype as reflected by a significant drop in the OD600 of the culture. Interestingly, the surviving tatAyCy mutant cells resumed (or continued) growth about 1.5 hours after the onset of cell lysis, but at a lower rate than during the initial exponential growth. This growth phenotype could be reversed by complementation of the chromosomal tatAyCy mutations with plasmid-borne copies of the tatAyCy genes (Figure S3, panel B). Specifically, the growth defect of cells lacking tatAy and tatCy at low salinity was only reversed when both the tatAy and tatCy genes were present on the complementing plasmid. Since tatAyCy mutant cells that had recovered from the lysis phase showed the same phenotype upon re-cultivation under low salinity conditions, it seems that the cells that resumed growth had adapted to these conditions rather than expressing a suppressor mutation. Impor- tantly, the observed phenotype revealed that B. subtilis needs to have an active TatAyCy translocase for optimal growth and entry into stationary phase at low salinity. Figure 2. Northern blotting analysis of tatAd and tatAy transcription in cells grown in LB media of differing salinity. RNA isolation and Northern blotting were performed as described in the Materials and Methods section. For each sample, 5 mg of RNA per lane were loaded as indicated. Molecular markers and the positions of the 16S and 23S rRNA are indicated. doi:10.1371/journal.pone.0018140.g002 TatAyCy is required for optimal growth of B. subtilis and entry into stationary phase at low salinity The experiments to monitor Tat-dependent secretion of YwbN by the tatAdCd, tatAyCy or total-tat mutant strains cultivated in LB medium without added salt revealed an unexpectedly strong growth phenotype for mutant strains lacking the tatAy and/or tatCy genes (Figure 4A and Figure S3). Compared to the parental strain PLoS ONE \| www.plosone.org March 2011 \| Volume 6 \| Issue 3 \| e18140 4 TatAyCy-dependent secretion of YwbN has an important role in iron acquisition The results in Figure 4C show that, indeed, the growth rates of total-tat mutants in BOC medium were significantly lower than that of the parental strain 168. This growth defect was largely relieved by the addition of 10 mM Fe2+ to the medium (Figure 4C), showing that the B. subtilis Tat pathway is required for optimal growth under iron- limited conditions. TatAyCy-dependent secretion of YwbN has an important role in iron acquisition Importantly, the addition of FeSO4 to tatAyCy or total-tat mutant cells growing exponentially in LB without salt completely prevented the occurrence of the lysis phase (Figure S4). This implies that the resumed growth after the lysis phase is most likely due to the release of iron by the lysed cells that can then be used by the surviving cells to resume or continue growth. The addition of 10 mM Fe3+ to the low salinity medium had only a moderately stimulating effect on the growth of the tatAyCy, ywbM or ywbN mutant strains, while Fe3+ had barely any growth-promoting effect on the ywbL mutant strain (Figure S3, panel D). Limited rescue of the mutants by addition of Fe3+ is consistent with the previously reported lack of Fe3+-specific bacillibactin-dependent iron acqui- sition in the B. subtilis 168 strain [40], while excess of Fe2+ is directed towards low affinity uptake by divalent cation importers [41,42]. The results imply that YwbL is the decisive Fe3+ permease for elemental iron uptake as previously suggested [43], and that the B. subtilis TatAyCy translocase and its substrate YwbN are of pivotal importance for the acquisition of iron during growth at low salinity. Since this suggests a possible general role for the B. subtilis Tat pathway in growth under iron-limited conditions, we tested the growth of two independently constructed B. subtilis total-tat mutants in synthetic iron-limited BOC medium. The results in Figure 4C show that, indeed, the growth rates of total-tat mutants in BOC medium were significantly lower than that of the parental strain 168. This growth defect was largely relieved by the addition of 10 mM Fe2+ to the medium (Figure 4C), showing that the B. subtilis Tat pathway is required for optimal growth under iron- limited conditions. B; Figure S3, panel C), while its growth phenotype was normal in the presence of 1% salt (Figure S3, panel F). Furthermore, the growth phenotype of the ywbN mutant strain was fully comple- mented by xylose-induced ectopic expression of a copy of ywbN- Myc that was integrated into the chromosomal amyE locus (Figure S3, panel C). Consistent with these findings, the secretion of Myc- tagged YwbN was strictly dependent on TatAyCy under conditions of low salinity (Figure 1, lower panels). TatAyCy-dependent secretion of YwbN has an important role in iron acquisition Since ywbN is part of a well-conserved cluster of three genes of which the two other genes, ywbL and ywbM, have been implicated in iron uptake [20,38,39], we also investigated whether ywbL or ywbM mutant strains have a growth defect at low salinity. Indeed, this was the case (Figure 4B; Figure S3, panel C), suggesting that YwbL (a homologue of known iron permeases) and YwbM (a lipoprotein), as well as TatAyCy-dependently secreted YwbN could be required for the uptake of sufficient amounts of iron to sustain growth under conditions of low salinity. To test this idea, all growth experiments at low salinity were repeated in the presence of 10 mM FeCl3 (Figure S3, panel D) or FeSO4 (Figure S3, panel E). The results showed that addition of 10 mM Fe2+ to the low salinity LB medium strongly stimulated the growth of the tatAyCy, ywbL, ywbM, or ywbN mutants to levels that were indistinguishable from growth in the presence of 1% salt (Figure S3, compare panels E and F). Importantly, the addition of FeSO4 to tatAyCy or total-tat mutant cells growing exponentially in LB without salt completely prevented the occurrence of the lysis phase (Figure S4). This implies that the resumed growth after the lysis phase is most likely due to the release of iron by the lysed cells that can then be used by the surviving cells to resume or continue growth. The addition of 10 mM Fe3+ to the low salinity medium had only a moderately stimulating effect on the growth of the tatAyCy, ywbM or ywbN mutant strains, while Fe3+ had barely any growth-promoting effect on the ywbL mutant strain (Figure S3, panel D). Limited rescue of the mutants by addition of Fe3+ is consistent with the previously reported lack of Fe3+-specific bacillibactin-dependent iron acqui- sition in the B. subtilis 168 strain [40], while excess of Fe2+ is directed towards low affinity uptake by divalent cation importers [41,42]. The results imply that YwbL is the decisive Fe3+ permease for elemental iron uptake as previously suggested [43], and that the B. subtilis TatAyCy translocase and its substrate YwbN are of pivotal importance for the acquisition of iron during growth at low salinity. Since this suggests a possible general role for the B. subtilis Tat pathway in growth under iron-limited conditions, we tested the growth of two independently constructed B. subtilis total-tat mutants in synthetic iron-limited BOC medium. TatAyCy-dependent secretion of YwbN has an important role in iron acquisition doi:10.1371/journal.pone.0018140.g003 March 2011 \| Volume 6 \| Issue 3 \| e18140 PLoS ONE \| www.plosone.org 5 Salinity-Determined Usage of Tat in Bacillus Salinity-Determined Usage of Tat in Bacillus diamonds), tatAyCy (open triangles), and 168 (open squares) in LB medium without NaCl. C. Growth of the B. subtilis strains total-tat (triangles), total-tat2 (circles) and 168 (squares) in BOC medium: open symbols, no added iron; closed symbols, addition of 10 mM FeSO4. doi:10.1371/journal.pone.0018140.g004 Figure 4. Growth phenotypes of tatAyCy and ywbLMN mutant strains at low salinity. A. Growth of the B. subtilis strains tatAdCd (filled rectangles), tatAyCy (open triangles), and 168 (open squares) in LB medium without NaCl. B. Growth of B. subtilis strains ywbN (open circles), ywbN XywbN plus xylose (closed rectangles), ywbL (closed PLoS ONE \| www plosone org 6 B; Figure S3, panel C), while its growth phenotype was normal in the presence of 1% salt (Figure S3, panel F). Furthermore, the growth phenotype of the ywbN mutant strain was fully comple- mented by xylose-induced ectopic expression of a copy of ywbN- Myc that was integrated into the chromosomal amyE locus (Figure S3, panel C). Consistent with these findings, the secretion of Myc- tagged YwbN was strictly dependent on TatAyCy under conditions of low salinity (Figure 1, lower panels). Since ywbN is part of a well-conserved cluster of three genes of which the two other genes, ywbL and ywbM, have been implicated in iron uptake [20,38,39], we also investigated whether ywbL or ywbM mutant strains have a growth defect at low salinity. Indeed, this was the case (Figure 4B; Figure S3, panel C), suggesting that YwbL (a homologue of known iron permeases) and YwbM (a lipoprotein), as well as TatAyCy-dependently secreted YwbN could be required for the uptake of sufficient amounts of iron to sustain growth under conditions of low salinity. To test this idea, all growth experiments at low salinity were repeated in the presence of 10 mM FeCl3 (Figure S3, panel D) or FeSO4 (Figure S3, panel E). The results showed that addition of 10 mM Fe2+ to the low salinity LB medium strongly stimulated the growth of the tatAyCy, ywbL, ywbM, or ywbN mutants to levels that were indistinguishable from growth in the presence of 1% salt (Figure S3, compare panels E and F). TatAyCy-dependent secretion of YwbN has an important role in iron acquisition To test whether the known TatAyCy substrate YwbN was implicated in the growth kinetics observed at low salinity, experiments were performed with a ywbN mutant. Indeed, the ywbN mutant strain displayed a similar growth phenotype as strains lacking tatAyCy in the absence of added salt (Figure 4, panels A and 168 and tatAdCd mutant strains, the tatAyCy and total-tat mutant strains showed significantly reduced growth rates during the exponential growth phase. Most strikingly, the tatAyCy and total-tat Figure 3. Complementation analysis of tatAdCd or tatAyCy mutant strains grown at high salinity. Growth medium and cell fractions of B. subtilis tat mutant strains and the parental strain 168 grown in LB with 6% added NaCl were separated and used for SDS-PAGE, Western blotting and immunodetection of YwbN-Myc and LipA using specific antibodies. The tatAdCd or tatAyCy mutant strains (marked at the bottom of the Figure) were complemented with plasmid-borne tat genes as indicated at the top of the Figure. The plasmids used for this purpose were pCAd, pCCd, pCACd, pCAy, pCCy, pCACy (see Table 1). Note that the tatAdCd mutant strains contained an X-ywbN cassette in which the original CmR marker had been replaced by a TcR marker (tatAdCd X-ywbN2), whereas the 168 strain and the tatAyCy mutant strains contained the previously published X-ywbN cassette (see Table 1). The YwbN-Myc (YwbN and preYwbN) and LipA proteins, and Mw markers are indicated. doi:10.1371/journal.pone.0018140.g003 Figure 3. Complementation analysis of tatAdCd or tatAyCy mutant strains grown at high salinity. Growth medium and cell fractions of B. subtilis tat mutant strains and the parental strain 168 grown in LB with 6% added NaCl were separated and used for SDS-PAGE, Western blotting and immunodetection of YwbN-Myc and LipA using specific antibodies. The tatAdCd or tatAyCy mutant strains (marked at the bottom of the Figure) were complemented with plasmid-borne tat genes as indicated at the top of the Figure. The plasmids used for this purpose were pCAd, pCCd, pCACd, pCAy, pCCy, pCACy (see Table 1). Note that the tatAdCd mutant strains contained an X-ywbN cassette in which the original CmR marker had been replaced by a TcR marker (tatAdCd X-ywbN2), whereas the 168 strain and the tatAyCy mutant strains contained the previously published X-ywbN cassette (see Table 1). The YwbN-Myc (YwbN and preYwbN) and LipA proteins, and Mw markers are indicated. Salinity-Determined Usage of Tat in Bacillus The observation that, at high salinity, TatAd was involved in YwbN secretion to a similar extent as TatAyCy implies that the TatAd, TatAy and TatCy subunits of B. subtilis can cooperate under these conditions. This contrasts strongly with the previously demonstrated exclusive dependence of YwbN secretion on the TatAyCy translocase, when cells were grown in LB with 1% salt [20]. It was recently reported that YwbN can be secreted in a TatAdCd-dependent manner by cells growing in LB medium with 1% NaCl. However, this only occurs when the TatAdCd translo- case is produced ectopically from plasmid-borne tatAdCd genes [36]. In this somewhat artificial situation, TatAdCd was able to sustain YwbN secretion in the absence of TatAyCy, which is clearly not the case for the parental 168 strain growing at high salinity (Figure 3, last lane), where cooperation between TatAd, TatAy and TatCy appears to be necessary for optimal YwbN secretion via Tat. Interestingly, our Northern blotting analyses indicate that B. subtilis is able to fine-tune the levels of TatAd, TatAy and TatCy in cells grown in high salinity medium through an as yet unidentified transcriptional regulatory mechanism. The present findings thus show that the TatAd subunit is not only involved in Tat-dependent protein secretion upon phosphate starvation, but also when high levels of salt are present in the growth medium. To date, it is not known whether the observed cooperativity between TatAd, TatAy and TatCy is triggered directly by high levels of salt (e.g. high ionic strength), or by the altered expression levels of tatAd and tatAyCy. Cells grown at high salinity secreted substantial amounts of YwbN in a Tat-independent manner. Most likely, this YwbN secretion takes place via the B. subtilis Sec pathway since this pathway has been shown to accept proteins with RR-signal peptides [44]. We attempted to confirm the involvement of Sec in the Tat-independent secretion of YwbN by growing cells in the presence of the SecA inhibitor sodium azide. Unfortunately however, sodium azide interfered significantly with the growth of B. subtilis in LB with 6% salt making it impossible to carefully verify a role for the Sec pathway in YwbN secretion. Furthermore, it is presently difficult to say whether the observed Tat- independent secretion of YwbN under high salinity growth conditions relates to indirect effects of these growth conditions on the pre-translocational folding or assembly of YwbN. Salinity-Determined Usage of Tat in Bacillus Since the internal NaCl concentration is unlikely to change in a significant way when cells are grown in LB medium with 6% salt, any possible folding or assembly defects of YwbN could perhaps relate to changes in the intracellular concentrations of as yet unidentified salt stress-dependent proteins, compatible solutes (osmolytes) or ions [45,46]. Clearly, impaired folding of pre-YwbN would make this protein a potential substrate for the Sec pathway. Taken together, our present results show for the first time that environmental salinity is a critical determinant for Tat-dependent protein secretion in B. subtilis. Depending on the salinity levels, the YwbN protein is seemingly directed from the Tat pathway into the Sec pathway. This observation is reminiscent of our previous observation that the B. subtilis esterase LipA can be redirected from the Sec to the Tat pathway under conditions of hyperproduction via an overflow mechanism [29]. This suggests that there may be additional, as yet unidentified factors that can impact on the choice between Sec or Tat pathway usage. A parameter that can impact strongly on the rates of protein (un-)folding is temperature. Nevertheless, growth temperature by itself does not seem to be a factor determining secretion pathway dependency, since we did not detect any obvious differences in Tat-dependence of YwbN secretion when cells were grown at 15, 30, 37 or 48uC (unpublished observations). However, being an organism that lives in the soil, B. subtilis can be exposed to a plethora of environmental insults that may either directly influence the activity and specificity of the Tat translocases, or that may indirectly impact on the pre-translocational folding of secretory precursor proteins. We are therefore convinced that more proteins will use the Sec and/or Tat pathways of B. subtilis in a growth condition- dependent manner. Identification of such conditions will be of interest not only from a fundamental scientific point of view, but also for the biotechnological application of B. subtilis as a cell factory for the production of high-value proteins. The present experiments revealed an essential role of the TatAyCy-secreted YwbN in the acquisition of iron during growth in LB at low salinity and in iron-limiting synthetic BOC medium. At low salinity, the growth defects and post-exponential growth lysis phenomenon of tatAyCy or ywbN mutants could be fully suppressed by the addition of Fe2+, and to a lesser extent Fe3+, to the growth medium. Salinity-Determined Usage of Tat in Bacillus Salinity-Determined Usage of Tat in Bacillus a ywbL mutant in the presence of Fe3+, but not in the presence of Fe2+, indicating that B. subtilis YwbL is specific for Fe3+ uptake like Ftr1p. Taken together, our findings show that severe iron limitation is the reason why cells lacking YwbLMN or TatAyCy grow at reduced rates in low salinity LB medium, and start to lyse instead of entering the stationary growth phase. The finding that such mutants resume growth after the lysis phase indicates that cell lysis results in the liberation of iron, which can be reused by the surviving cells. This view is confirmed by the observation that the lysis phenomenon can be largely prevented if iron is added to the culture during exponential growth. The precise role of YwbN in making iron available for the cells is not entirely clear but, being a homologue of the Dyp-type iron-dependent peroxidase EfeB, it seems most likely that this enzyme is involved in oxidation of Fe2+ for subsequent uptake of Fe3+ by YwbL, which is analogous to the yeast Fet3p-Ftr1p high-affinity iron uptake system [27,37]. Furthermore, recent studies have shown that EfeB of E. coli can extract iron from heme without affecting the protoporphyrin ring [50], which suggests a role for B. subtilis YwbN in the liberation of complexed iron. A function of YwbN in iron metabolism in B. subtilis is fully consistent with the finding by Helmann and co- workers that the ywbN gene is part of the Fur regulon [43]. By contrast, tatAyCy does not seem to be controlled by Fur, which is consistent with a ‘‘mainstream’’ role in Tat-dependent protein translocation, as documented in the present studies. It is remarkable to note that an iron starvation response was previously documented for B. subtilis cells grown at high salinity [51,52]. This apparent iron limitation at high salinity growth conditions did not result in a requirement for the TatAyCy translocase or YwbLMN as observed under low salinity growth conditions. However, it should be noted that tatAyCy mutants do secrete YwbN when grown in LB medium with 6% NaCl, which might suffice for iron acquisition. salinity had a strong impact on the specificity of the Tat pathway of B. subtilis, as shown by the role of the TatAd component in YwbN secretion. Salinity-Determined Usage of Tat in Bacillus This implies that cells lacking TatAyCy and/or YwbN are iron-starved at low salinity. The same turned out to be true for cells lacking the lipoprotein YwbM and the integral membrane protein YwbL. Interestingly, the E. coli homologue of YwbM, named EfeO, is a periplasmic protein that has been implicated in high-affinity iron-binding [37]. YwbL is a homologue of the EfeU iron permease in the E. coli inner membrane [37,47], and the high-affinity iron permease Ftr1p of fungi and yeast [48]. While Ftr1p is an Fe3+ permease [48,49], EfeU was shown to permeate Fe2+ in vitro into proteoliposomes [47]. The present study shows impaired growth of Discussion Figure 4. Growth phenotypes of tatAyCy and ywbLMN mutant strains at low salinity. A. Growth of the B. subtilis strains tatAdCd (filled rectangles), tatAyCy (open triangles), and 168 (open squares) in LB medium without NaCl. B. Growth of B. subtilis strains ywbN (open circles), ywbN XywbN plus xylose (closed rectangles), ywbL (closed In the present studies we have addressed the question as to whether the salinity of the growth medium is a determinant in Tat- dependent secretion of the YwbN protein by B. subtilis. The results show that this is indeed the case. Intriguingly, high growth medium PLoS ONE \| www.plosone.org March 2011 \| Volume 6 \| Issue 3 \| e18140 March 2011 \| Volume 6 \| Issue 3 \| e18140 6 Materials and Methods Plasmids, bacterial strains, media and growth conditions The plasmids and bacterial strains used in this study are listed in Table 1. Strains were grown with agitation at 37uC in March 2011 \| Volume 6 \| Issue 3 \| e18140 PLoS ONE \| www.plosone.org 7 Salinity-Determined Usage of Tat in Bacillus Luria Bertani (LB) medium consisting of 1% tryptone, 0.5% yeast extract and 0%, 1% or 6% added NaCl, pH 7.4. Belitsky minimal medium without citrate (BOC medium) was prepared as described [53]. If appropriate, media for were supplemented with the following antibiotics: erythromycin (Em), 2 mg ml21; chloramphenicol (Cm), 5 mg ml21; tetracycline (Tc), 10 mg ml21; spectinomycin (Sp), 100 mg ml21; kanamycin (Km), 20 mg ml21. SDS-PAGE and Western blotting Cells were separated from the growth medium by centrifuga- tion. Cellular or secreted proteins were separated by SDS-PAGE using pre-cast Bis-Tris NuPAGE gels (Invitrogen). Separated proteins were stained with SYPRO Ruby protein gel stain (Molecular Probes Inc.). The presence of YwbN-Myc, LipA, or TrxA in cellular or growth medium fractions was detected by Western blotting. For this purpose, proteins separated by SDS- PAGE were semi-dry blotted (75 min at 1 mA/Cm2) onto a nitrocellulose membrane (ProtranH, Schleicher & Schuell). Subse- quently, the LipA and TrxA proteins were detected with specific polyclonal antibodies raised in rabbits. YwbN-Myc was detected with monoclonal antibodies against the Myc-tag (Gentaur). Visualisation of bound antibodies was performed with fluorescent IgG secondary antibodies (IRDye 800 CW goat anti-rabbit or goat anti-mouse from LiCor Biosciences) in combination with the Odyssey Infrared Imaging System (LiCor Biosciences). Fluores- cence was recorded at 800 nm. Quantification of the recorded data was done using the ImageJ software package (http://rsbweb. nih.gov/ij/). Figure S3 Growth phenotypes of tatAyCy and ywbLMN mutant strains at low salinity. B. subtilis tat mutant strains or the parental strain 168 were grown for 7.5 to 14 hours in LB medium without NaCl (panels A-E), or 1% NaCl (panel F). For the experiments in panel D, LB medium was supplemented with 10 mM FeCl3, and for the experiments in panel F with 10 mM FeSO4 (panel E). A. Growth of tat mutant strains: tatAd tatAy (+), tatCd (filled diamonds), tatCy (filled triangles), tatCd tatCy (filled circles), tatAdCd (filled rectangles), tatAyCy (open triangles), total-tat2 (filled squares). Parental strain 168 (open squares). B. Growth of the tatAyCy mutant strain complemented with tatAy (pCAy; open triangles), tatCy (pCCy; X), or tatAyCy (pCACy; filled circles). Controls: tatAyCy mutant with empty vector pGDL48 (closed squares), parental strain 168 (open squares). C-F. Growth of mutant strains: tatAyCy (open triangles), ywbL (closed diamonds), ywbM (X), ywbN (open circles), ywbN XywbN (no xylose; +), ywbN XywbN (plus xylose; closed rectangles). Control: parental strain 168 (open squares). Supporting Information Figure S1 TrxA control for cell lysis in media with differing salinity. Cell and growth medium fractions of B. subtilis tat mutant strains and the parental strain 168 were separated by centrifugation and used for SDS-PAGE, Western blotting and immunodetection of the cytoplasmic marker protein TrxA with specific antibodies. The panels show results obtained for cells grown in LB with 1% or 6% NaCl. Protein loading was corrected for OD600. The positions of TrxA and Mw markers are indicated. The samples correspond to those of the experiment depicted in Figure 1. No TrxA can be detected in the growth medium fractions indicating that cell lysis was negligible in this experiment. p (TIF) Figure S2 Tat-dependence of YwbN secretion in media with differing salinity. Cell and growth medium fractions of B. subtilis tat mutant strains and the parental strain 168 were separated by centrifugation and used for SDS-PAGE and Sypro Ruby staining (left panels) or SDS-PAGE, Western blotting and immunodetection of YwbN-Myc and LipA with specific antibodies (right panels). From top to bottom the panels show results obtained for cells grown in LB with 6%, 1% or no added NaCl. Protein loading was corrected for OD600. The YwbN-Myc (YwbN) and LipA proteins, and Mw markers are indicated. A slight ’smiling effect’ as observed for the YwbN and LipA bands in the growth medium sample of the total-tat2 mutant grown in LB with 6% NaCl is due to some residual salt in the sample (compare also left and right panels). The samples correspond to those of the experiment depicted in Figure 1. (TIF) Growth experiments Strains were pre-cultured in LB medium containing 1% NaCl and subsequently diluted in LB medium without salt or with 1% or 6% NaCl to an optical density at 600 nm (OD600) of ,0.01. Growth was continued in triplicate wells of a 96-well microtiter plate (Greiner) that was incubated in a Biotek Synergy 2 plate reader (37uC, variable shaking). OD600 readings were recorded for 8 or 14 hours. References Panahandeh S, Maurer C, Moser M, Delisa MP, Muller M (2008) Following the path of a twin-arginine precursor along the TatABC translocase of Escherichia coli. J Biol Chem 283: 33267–33275. 32. Thomas JR, Bolhuis A (2006) The tatC gene cluster is essential for viability in halophilic archaea. FEMS Microbiol Lett 256: 44–49. 9. Wexler M, Bogsch EG, Klosgen RB, Palmer T, Robinson C, et al. (1998) Targeting signals for a bacterial Sec-independent export system direct plant thylakoid import by the delta pH pathway. FEBS Lett 431: 339–342. 33. Hutcheon GW, Bolhuis A (2003) The archaeal twin-arginine translocation pathway. Biochem Soc Trans 31: 686–689. 10. Berks BC (1996) A common export pathway for proteins binding complex redox cofactors? Mol Microbiol 22: 393–404. 34. Dilks K, Gimenez MI, Pohlschroder M (2005) Genetic and biochemical analysis of the twin-arginine translocation pathway in halophilic archaea. J Bacteriol 187: 8104–8113. 11. Chaddock AM, Mant A, Karnauchov I, Brink S, Herrmann RG, et al. (1995) A new type of signal peptide: central role of a twin-arginine motif in transfer signals for the delta pH-dependent thylakoidal protein translocase. Embo J 14: 2715–2722. 35. Gimenez MI, Dilks K, Pohlschroder M (2007) Haloferax volcanii twin-arginine translocation substates include secreted soluble, C-terminally anchored and lipoproteins. Mol Microbiol 66: 1597–1606. 12. Brink S, Bogsch EG, Edwards WR, Hynds PJ, Robinson C (1998) Targeting of thylakoid proteins by the delta pH-driven twin-arginine translocation pathway requires a specific signal in the hydrophobic domain in conjunction with the twin-arginine motif. FEBS Lett 434: 425–430. p p 36. Eijlander RT, Jongbloed JD, Kuipers OP (2009) Relaxed specificity of the Bacillus subtilis TatAdCd translocase in Tat-dependent protein secretion. J Bacteriol 191: 196–202. J 37. Rasmussen S, Nielsen HB, Jarmer H (2009) The transcriptionally active regions in the genome of Bacillus subtilis. Mol Microbiol 73: 1043–1057. 13. Stanley NR, Palmer T, Berks BC (2000) The twin arginine consensus motif of Tat signal peptides is involved in Sec-independent protein targeting in Escherichia coli. J Biol Chem 275: 11591–11596. 38. Cao J, Woodhall MR, Alvarez J, Cartron ML, Andrews SC (2007) EfeUOB (YcdNOB) is a tripartite, acid-induced and CpxAR-regulated, low-pH Fe2+ transporter that is cryptic in Escherichia coli K-12 but functional in E. coli O157:H7. Mol Microbiol 65: 857–875. 14. Cristobal S, de Gier JW, Nielsen H, von Heijne G (1999) Competition between Sec- and TAT-dependent protein translocation in Escherichia coli. Embo J 18: 2982–2990. 15. References 26. Yen MR, Tseng YH, Nguyen EH, Wu LF, Saier MH, Jr. (2002) Sequence and phylogenetic analyses of the twin-arginine targeting (Tat) protein export system. Arch Microbiol 177: 441–450. 1. Robinson C, Bolhuis A (2004) Tat-dependent protein targeting in prokaryotes and chloroplasts. Biochim Biophys Acta 1694: 135–147. 1. Robinson C, Bolhuis A (2004) Tat-dependent protein targeting in prokaryotes and chloroplasts. Biochim Biophys Acta 1694: 135–147. 2. Palmer T, Sargent F, Berks BC (2005) Export of complex cofactor-containing proteins by the bacterial Tat pathway. Trends Microbiol 13: 175–180. 2. Palmer T, Sargent F, Berks BC (2005) Export of complex cofactor-containing proteins by the bacterial Tat pathway. Trends Microbiol 13: 175–180. 27. Sturm A, Schierhorn A, Lindenstrauss U, Lilie H, Bruser T (2006) YcdB from Escherichia coli reveals a novel class of Tat-dependently translocated hemoproteins. J Biol Chem 281: 13972–13978. 3. Driessen AJ, Nouwen N (2008) Protein translocation across the bacterial cytoplasmic membrane. Annu Rev Biochem 77: 643–667. cytoplasmic membrane. Annu Rev Biochem 77: 643–667. 28. Widdick DA, Eijlander RT, van Dijl JM, Kuipers OP, Palmer T (2007) A Facile Reporter System for the Experimental Identification of Twin-Arginine Translocation (Tat) Signal Peptides from All Kingdoms of Life. J Mol Biol. 4. Natale P, Bruser T, Driessen AJ (2008) Sec- and Tat-mediated protein secretion across the bacterial cytoplasmic membrane—distinct translocases and mecha- nisms. Biochim Biophys Acta 1778: 1735–1756. p y 5. Sargent F (2007) The twin-arginine transport system: moving folded proteins across membranes. Biochem Soc Trans 35: 835–847. 29. Kouwen TR, van der Ploeg R, Antelmann H, Hecker M, Homuth G, et al. (2009) Overflow of a hyper-produced secretory protein from the Bacillus Sec pathway into the Tat pathway for protein secretion as revealed by proteogenomics. Proteomics. 5. Sargent F (2007) The twin-arginine transport system 6. Berks BC, Sargent F, Palmer T (2000) The Tat protein export pathway. Mol Microbiol 35: 260–274. 30. Rose RW, Bruser T, Kissinger JC, Pohlschroder M (2002) Adaptation of protein secretion to extremely high-salt conditions by extensive use of the twin-arginine translocation pathway. Mol Microbiol 45: 943–950. 7. DeLisa MP, Tullman D, Georgiou G (2003) Folding quality control in the export of proteins by the bacterial twin-arginine translocation pathway. Proc Natl Acad Sci U S A 100: 6115–6120. 31. Bolhuis A (2002) Protein transport in the halophilic archaeon Halobacterium sp. NRC-1: a major role for the twin-arginine translocation pathway? Microbiology 148: 3335–3346. 8. Northern blotting analyses For Northern blotting, cells from overnight cultures in LB medium with 1% NaCl were used to inoculate LB medium without NaCl or with 1% or 6% NaCl. Growth was continued till an OD600 of ,2 and then the cells were harvested for RNA extraction and blotting. Preparation of total RNA and Northern blot analysis using specific RNA probes were performed as described previously [29]. The RNA preparations (5 or 10 mg per lane) were separated electrophoretically in a 1.2% agarose gel. After hybridization of the blotted RNA with gene-specific probes, chemiluminescence was detected using a Lumi-Imager (Roche Diagnostics). Transcript sizes were determined by comparison with an RNA size marker (Fermentas RiboRuler High Range RNA Ladder). Digoxygenin-labeled specific RNA probes were synthesized by in vitro transcription using T7 RNA polymerase and a specific PCR product as template. Synthesis of the DNA template was performed by PCR using the following pairs of oligonucleotides: tatAd-for (59-ATGTTTTCAAACATTGGAAT- 39) and tatAd-rev (59-CTAATACGACTCACTATAGGGA- GAGCCCGCGTTTTTGTCCTGCT-39), tatAy-for (59- ATGC- CGATCGGTCCTGGAAG-39) and tatAy-rev (59-CTAATAC- GACTCACTATAGGGAGACTGATCTTCTTTCTTTTTTT- 39). The underlined sequence indicates the T7 promoter region. Figure S4 Iron additions can prevent the growth defects of tatAyCy and total-tat mutant B. subtilis strains in LB medium without salt. B. subtilis tat mutant strains or the parental strain 168 were grown for 11 hours in LB medium without NaCl. Growth was monitored by OD600 readings. The cultures were supplemented with 100 mM FeCl3 or 100 mM FeSO4 when cells had reached the mid-exponential growth phase after 190 min of cultivation (T1), or when cells had entered the transition phase between the exponential and post-exponential growth phases after 290 min of cultivation (T2). A. parental B. subtilis strain 168, B. tatAyCy mutant strain, and C. total-tat mutant strain. T1 and T2 are marked with arrows. Open squares, no addition to the culture; open triangles, FeCl3 was added at T1; PLoS ONE \| www.plosone.org March 2011 \| Volume 6 \| Issue 3 \| e18140 8 Salinity-Determined Usage of Tat in Bacillus Salinity-Determined Usage of Tat in Bacillus filled circles, FeSO4 was added at T1; filled squares, FeCl3 was added at T2; crosses, FeSO4 was added at T2. filled circles, FeSO4 was added at T1; filled squares, FeCl3 was added at T2; crosses, FeSO4 was added at T2. filled circles, FeSO4 was added at T1; filled squares, FeCl3 was added at T2; crosses, FeSO4 was added at T2. (TIF) Author Contributions Conceived and designed the experiments: RvdP UM GH ELD MM JMvD. Performed the experiments: RvdP MS ELD CGM MM TW HA. Analyzed the data: RvdP UM GH HA JMvD. Contributed reagents/ materials/analysis tools: GH MAM CRH MH JMvD. Wrote the paper: RvdP UM MM HA JMvD. References DeLisa MP, Samuelson P, Palmer T, Georgiou G (2002) Genetic analysis of the twin arginine translocator secretion pathway in bacteria. J Biol Chem 277: 29825–29831. 39. Biswas L, Biswas R, Nerz C, Ohlsen K, Schlag M, et al. (2009) Role of the twin- arginine translocation pathway in Staphylococcus. J Bacteriol 191: 5921–5929. 40. May JJ, Wendrich TM, Marahiel MA (2001) The dhb operon of Bacillus subtilis encodes the biosynthetic template for the catecholic siderophore 2,3-dihydrox- ybenzoate-glycine-threonine trimeric ester bacillibactin. J Biol Chem 276: 7209–7217. 16. Sargent F, Bogsch EG, Stanley NR, Wexler M, Robinson C, et al. (1998) Overlapping functions of components of a bacterial Sec-independent protein export pathway. Embo J 17: 3640–3650. 41. Hantke K (1997) Ferrous iron uptake by a magnesium transport system is toxic for Escherichia coli and Salmonella typhimurium. J Bacteriol 179: 6201–6204. p p y J 17. Bogsch EG, Sargent F, Stanley NR, Berks BC, Robinson C, et al. (1998) An essential component of a novel bacterial protein export system with homologues in plastids and mitochondria. J Biol Chem 273: 18003–18006. 42. Moore CM, Helmann JD (2005) Metal ion homeostasis in Bacillus subtilis. Curr Opin Microbiol 8: 188–195. 18. Jongbloed JD, Martin U, Antelmann H, Hecker M, Tjalsma H, et al. (2000) TatC is a specificity determinant for protein secretion via the twin-arginine translocation pathway. J Biol Chem 275: 41350–41357. 43. Ollinger J, Song KB, Antelmann H, Hecker M, Helmann JD (2006) Role of the Fur regulon in iron transport in Bacillus subtilis. J Bacteriol 188: 3664–3673. 19. Jongbloed JD, Antelmann H, Hecker M, Nijland R, Bron S, et al. (2002) Selective contribution of the twin-arginine translocation pathway to protein secretion in Bacillus subtilis. J Biol Chem 277: 44068–44078. 44. Kolkman MA, van der Ploeg R, Bertels M, van Dijk M, van der Laan J, et al. (2008) The twin-arginine signal peptide of Bacillus subtilis YwbN can direct either Tat- or Sec-dependent secretion of different cargo proteins: secretion of active subtilisin via the B. subtilis Tat pathway. Appl Environ Microbiol 74: 7507–7513. J 20. Jongbloed JD, Grieger U, Antelmann H, Hecker M, Nijland R, et al. (2004) Two minimal Tat translocases in Bacillus. Mol Microbiol 54: 1319–1325. 45. Hoper D, Bernhardt J, Hecker M (2006) Salt stress adaptation of Bacillus subtilis: a physiological proteomics approach. Proteomics 6: 1550–1562. 21. Acknowledgments RvdP UM MM HA JMvD. We thank Dirk Ho¨per for early stage contributions to the present studies, Robyn Eijlander and Oscar Kuipers for plasmids and strains, Jette Anklam for technical assistance, and Thijs Kouwen, Albert Bolhuis and Erhard Bremer for helpful discussions. Bremer for helpful discussions. 57. Kobayashi K, Ehrlich SD, Albertini A, Amati G, Andersen KK, et al. (2003) Essential Bacillus subtilis genes. Proc Natl Acad Sci U S A 100: 4678–4683. y g p p 55. Darmon E, Dorenbos R, Meens J, Freudl R, Antelmann H, et al. (2006) A disulfide bond-containing alkaline phosphatase triggers a BdbC-dependent secretion stress response in Bacillus subtilis. Appl Environ Microbiol 72: 6876–6885. Bacillus subtilis: identification of a eubacterial homolog of archaeal and eukaryotic signal peptidases. Genes Dev 12: 2318–2331. 51. Hoffmann T, Schutz A, Brosius M, Volker A, Volker U, et al. (2002) High- salinity-induced iron limitation in Bacillus subtilis. J Bacteriol 184: 718–727. 56. Kunst F, Ogasawara N, Moszer I, Albertini AM, Alloni G, et al. (1997) The complete genome sequence of the gram-positive bacterium Bacillus subtilis. Nature 390: 249–256. y 52. Steil L, Hoffmann T, Budde I, Volker U, Bremer E (2003) Genome-wide transcriptional profiling analysis of adaptation of Bacillus subtilis to high salinity. J Bacteriol 185: 6358–6370. 54. Tjalsma H, Bolhuis A, van Roosmalen ML, Wiegert T, Schumann W, et al. (1998) Functional analysis of the secretory precursor processing machinery of 53. Miethke M, Klotz O, Linne U, May JJ, Beckering CL, et al. (2006) Ferri- bacillibactin uptake and hydrolysis in Bacillus subtilis. Mol Microbiol 61: 1413–1427. 50. Letoffe S, Heuck G, Delepelaire P, Lange N, Wandersman C (2009) Bacteria capture iron from heme by keeping tetrapyrrol skeleton intact. Proc Natl Acad Sci U S A 106: 11719–11724. References Jongbloed JD, van der Ploeg R, van Dijl JM (2006) Bifunctional TatA subunits in minimal Tat protein translocases. Trends Microbiol 14: 2–4. 22. Bolhuis A, Mathers JE, Thomas JD, Barrett CM, Robinson C (2001) TatB and TatC form a functional and structural unit of the twin-arginine translocase from Escherichia coli. J Biol Chem 276: 20213–20219. 46. Bursy J, Pierik AJ, Pica N, Bremer E (2007) Osmotically induced synthesis of the compatible solute hydroxyectoine is mediated by an evolutionarily conserved ectoine hydroxylase. J Biol Chem 282: 31147–31155. 23. Cline K, Mori H (2001) Thylakoid DeltapH-dependent precursor proteins bind to a cpTatC-Hcf106 complex before Tha4-dependent transport. J Cell Biol 154: 719–729. 47. Grosse C, Scherer J, Koch D, Otto M, Taudte N, et al. (2006) A new ferrous iron-uptake transporter, EfeU (YcdN), from Escherichia coli. Mol Microbiol 62: 120–131. 24. Alami M, Luke I, Deitermann S, Eisner G, Koch HG, et al. (2003) Differential interactions between a twin-arginine signal peptide and its translocase in Escherichia coli. Mol Cell 12: 937–946. 48. Kosman DJ (2003) Molecular mechanisms of iron uptake in fungi. Mol Microbiol 47: 1185–1197. 25. Mori H, Cline K (2002) A twin arginine signal peptide and the pH gradient trigger reversible assembly of the thylakoid [Delta]pH/Tat translocase. J Cell Biol 157: 205–210. 49. Askwith C, Kaplan J (1997) An oxidase-permease-based iron transport system in Schizosaccharomyces pombe and its expression in Saccharomyces cerevisiae. J Biol Chem 272: 401–405. PLoS ONE \| www.plosone.org March 2011 \| Volume 6 \| Issue 3 \| e18140 March 2011 \| Volume 6 \| Issue 3 \| e18140 9 Salinity-Determined Usage of Tat in Bacillus Salinity-Determined Usage of Tat in Bacillus Salinity-Determined Usage of Tat in Bacillus 50. Letoffe S, Heuck G, Delepelaire P, Lange N, Wandersman C (2009) Bacteria capture iron from heme by keeping tetrapyrrol skeleton intact. Proc Natl Acad Sci U S A 106: 11719–11724. 55. Darmon E, Dorenbos R, Meens J, Freudl R, Antelmann H, et al. (2006) A disulfide bond-containing alkaline phosphatase triggers a BdbC-dependent secretion stress response in Bacillus subtilis. Appl Environ Microbiol 72: 6876–6885. 51. Hoffmann T, Schutz A, Brosius M, Volker A, Volker U, et al. (2002) High- salinity-induced iron limitation in Bacillus subtilis. J Bacteriol 184: 718–727. 52. Steil L, Hoffmann T, Budde I, Volker U, Bremer E (2003) Genome-wide transcriptional profiling analysis of adaptation of Bacillus subtilis to high salinity. J Bacteriol 185: 6358–6370. 56. Kunst F, Ogasawara N, Moszer I, Albertini AM, Alloni G, et al. (1997) The complete genome sequence of the gram-positive bacterium Bacillus subtilis. Nature 390: 249–256. 53. Miethke M, Klotz O, Linne U, May JJ, Beckering CL, et al. (2006) Ferri- bacillibactin uptake and hydrolysis in Bacillus subtilis. Mol Microbiol 61: 1413–1427. 57. Kobayashi K, Ehrlich SD, Albertini A, Amati G, Andersen KK, et al. (2003) Essential Bacillus subtilis genes. Proc Natl Acad Sci U S A 100: 4678–4683. 54. Tjalsma H, Bolhuis A, van Roosmalen ML, Wiegert T, Schumann W, et al. (1998) Functional analysis of the secretory precursor processing machinery of PLoS ONE \| www.plosone.org March 2011 \| Volume 6 \| Issue 3 \| e18140 10
https://openalex.org/W2472582738	https://dash.harvard.edu/bitstream/1/30371100/1/4965288.pdf	English	null	Nanoscale imaging of RNA with expansion microscopy	Nature methods	2,016	cc-by	9,626	Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#LAA Permanent link http://nrs.harvard.edu/urn-3:HUL.InstRepos:30371100 Published Version doi:10.1038/nmeth.3899 Published Version doi:10.1038/nmeth.3899 Citation Chen, F., A. T. Wassie, A. J. Cote, A. Sinha, S. Alon, S. Asano, E. R. Daugharthy, et al. 2016. “Nanoscale Imaging of RNA with Expansion Microscopy.” Nature methods 13 (8): 679-684. doi:10.1038/nmeth.3899. http://dx.doi.org/10.1038/nmeth.3899. Author Contributions F.C., A.T.W., E.R.D., A.M., G.M.C. and E.S.B. conceived of RNA tethering strategies to the ExM gel. F.C. and A.T.W. conceived and developed the LabelX reagent. F.C., A.T.W., J.-B. C., and S. Alon developed ExM gel stabilization by re-embedding. F.C., A.T.W., and E.R.D. conceived and developed reversible HCR strategies. F.C., A.T.W., and E.S.B. designed and F.C. and A.T.W. performed experiments. A.C. and A.R. provided FISH reagents and guidance on usage, and A.C. performed experiments. A.S. performed data analysis. S. Asano performed lightsheet imaging and analysis. E.S.B. supervised the project. F.C., A.T.W., A.S., and E.S.B. wrote the paper, and all authors contributed edits and revisions. Nanoscale Imaging of RNA with Expansion Microscopy Fei Chen1,2,3,10, Asmamaw T. Wassie1,2,3,10, Allison J. Cote4, Anubhav Sinha5, Shahar Alon2,3, Shoh Asano2,3, Evan R. Daugharthy6,7, Jae-Byum Chang2,3, Adam Marblestone2,3, George M. Church6,8, Arjun Raj4, and Edward S. Boyden1,2,3,9 1Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA Fei Chen1,2,3,10, Asmamaw T. Wassie1,2,3,10, Allison J. Cote4, Anubhav Sinha5, Shahar Alon2,3, Shoh Asano2,3, Evan R. Daugharthy6,7, Jae-Byum Chang2,3, Adam Marblestone2,3, George M. Church6,8, Arjun Raj4, and Edward S. Boyden1,2,3,9 1Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA Author Manuscript 2Media Lab, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA 3McGovern Institute, Massachusetts Institute of Technology, Cambridge, Massachusetts, epartment of Bioengineering, University of Pennsylvania, Philadelphia, Pennsylvania, USA 5Division of Health Sciences and Technology, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA 5Division of Health Sciences and Technology, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA 6Wyss Institute for Biologically Inspired Engineering, Boston, Massachusetts, USA 7Department of Systems Biology, Harvard Medical School, Boston, Massachusetts, USA 8Department of Genetics, Harvard Medical School, Boston, Massachusetts, USA 6Wyss Institute for Biologically Inspired Engineering, Boston, Massachusetts, USA 7Department of Systems Biology, Harvard Medical School, Boston, Massachusetts, USA 8Department of Genetics, Harvard Medical School, Boston, Massachusetts, USA 7Department of Systems Biology, Harvard Medical School, Boston, Massachusetts, USA 8Department of Genetics, Harvard Medical School, Boston, Massachusetts, USA 9Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA Author Manuscript p g F.C., A.T.W., S. Alon, E.R.D., J.-B. C., A.M., G.M.C., and E.S.B. are inventors on one or more patents or patent applications related to the technologies here discussed. E.S.B. is co-founder of Expansion Technologies, a company whose goal is to facilitate access to expansion microscopy technologies by the scientific community. HHS Public Access Author manuscript Nat Methods. Author manuscript; available in PMC 2017 January 04. Author Manuscript Published in final edited form as: Nat Methods. 2016 August ; 13(8): 679–684. doi:10.1038/nmeth.3899. Abstract The ability to image RNA identity and location with nanoscale precision in intact tissues is of great interest for defining cell types and states in normal and pathological biological settings. Here, we present a strategy for expansion microscopy (ExM) of RNA. We developed a small molecule linker that enables RNA to be covalently attached to a swellable polyelectrolyte gel synthesized throughout a biological specimen. Then, post-expansion, fluorescent in situ hybridization (FISH) imaging of RNA can be performed with high yield and specificity, with single molecule precision, in both cultured cells and intact brain tissue. Expansion FISH (ExFISH) Users may view, print, copy, and download text and data-mine the content in such documents, for the purposes of academic research, subject always to the full Conditions of use: http://www.nature.com/authors/editorial_policies/license.html#terms Author Manuscript Share Your Story The Harvard community has made this article openly available. Please share how this access benefits you. Submit a story . Accessibility Correspondence should be addressed to E.S.B (esb@media.mit.edu). 10These authors contributed equally to this work. rrespondence should be addressed to E.S.B (esb@media.mit.edu). These authors contributed equally to this work. Competing Financial Interests Co pet g a c a te ests F.C., A.T.W., S. Alon, E.R.D., J.-B. C., A.M., G.M.C., and E.S.B. are inventors on one or more patents or patent applications related to the technologies here discussed. E.S.B. is co-founder of Expansion Technologies, a company whose goal is to facilitate access to expansion microscopy technologies by the scientific community. Page 2 Chen et al. de-crowds RNAs and supports amplification of single molecule signals (i.e., via hybridization chain reaction (HCR)) as well as multiplexed RNA FISH readout. ExFISH thus enables super- resolution imaging of RNA structure and location with diffraction-limited microscopes in thick specimens, such as intact brain tissue and other tissues of importance to biology and medicine. Author Manuscript Nat Methods. Author manuscript; available in PMC 2017 January 04. Introduction Nanoscale-resolution imaging of RNA throughout cells, tissues, and organs is key for an understanding of local RNA processing, mapping structural roles of RNA, and defining cell types and states. However, it has remained difficult to image RNA in intact tissues with the nanoscale precision required to pinpoint associations with cellular compartments or proteins important for RNA function. Recently we developed an approach to physically magnify tissues, expansion microscopy (ExM)1. ExM isotropically magnifies tissues, enabling super- resolution imaging on conventional diffraction-limited microscopes. For example, ~4× linear expansion yields ~70 nm resolution using a ~300 nm diffraction-limited objective lens. In our original protocol, fluorophore tags were first targeted to proteins of interest via antibodies, and then anchored to a swellable polyelectrolyte gel synthesized in situ. Isotropic expansion was subsequently enabled by proteolytic treatment to homogenize specimen mechanical properties followed by osmotic swelling of the specimen-gel composite. Author Manuscript Here, we have developed a small molecule linker that enables RNA to be covalently attached to the ExM gel. We show that this procedure, which we call ExFISH, enables RNA fluorescent in situ hybridization (FISH), which enables identification of transcripts in situ with single molecule precision. In RNA FISH, a set of fluorescent probes complementary to a target strand of mRNA are delivered2,3. Single molecule FISH (smFISH) can be performed with multiple fluorophores delivered to a single mRNA via oligonucleotide probes4. In intact tissues, amplification strategies, such as hybridization chain reaction (HCR)5,6 and branched DNA amplification7,8, can enable a large number of fluorophores to be targeted to a single mRNA. We show that ExFISH can support smFISH in cell culture, and HCR-amplified FISH in intact mouse brain tissues. We demonstrate the power of ExFISH for revealing nanoscale structures of long non-coding RNAs (lncRNAs), as well as for localizing neural mRNAs to individual dendritic spines. ExFISH will be useful for a diversity of questions relating the structure and location of RNA to biological functions. Author Manuscript ExFISH: Design and Validation of RNA Anchoring Chemistry We first determined a strategy for covalently linking RNAs directly to the ExM gel. Although transcripts are crosslinked to proteins during fixation, the strong proteolysis of ExM precludes a reliance on proteins for RNA retention (Supplementary Fig. 1). We thus reasoned that covalently securing RNA molecules directly to the ExM gel via a small molecule linker would enable the interrogation of these molecules post-expansion. To achieve this aim, we synthesized a reagent from two building blocks: a molecule containing both an amine as well as an alkylating group that primarily reacts to the N7 of guanine, and a molecule that contains an amine-reactive succinamide ester and a polymerizable Author Manuscript Chen et al. Page 3 acrylamide moiety. Commercially available reagents exist that satisfy each of these two profiles, such as Label-IT Amine (MirusBio) and 6-((Acryloyl)amino)hexanoic acid (Acryloyl-X SE, here abbreviated AcX, Life Technologies; all reagents are listed in Supplementary Table 1). We named this molecule, which enables RNA to be covalently functionalized with a free radical polymerizable group, LabelX (Fig. 1a). We verified that LabelX does not impede smFISH readout (Supplementary Fig. 2). We then designed a procedure where a sample could be treated with LabelX to make its RNAs gel-anchorable, followed by gel formation, proteolysis, and osmotic swelling as performed in the original ExM protocol. Once a sample was thus expanded, the RNAs could then be interrogated through FISH (Fig. 1b). Author Manuscript To quantify RNA transcript anchoring yield after expansion, we used smFISH probes targeting mRNAs of varying copy number (7 targets, with copy number ranging from ~10 to ~10,000 per cell, n = 59 cells across all 7 targets). smFISH images, taken with probes delivered before (Fig. 1c) and after (Fig. 1d) expansion, to the same cells, showed no loss of transcript detectability with expansion for both low- and high-copy number transcripts (Fig. 1e). The ratio of transcripts detected was near unity at low transcript counts (e.g., in the 10’s), however, more transcripts were detected after expansion for highly expressed mRNAs (e.g., in the 1,000’s) (Supplementary Fig. 3, Supplementary Table 2). This difference arises from the high density of smFISH spots for these targets in the un-expanded state, with the expansion process de-crowding spots that previously were indistinguishable. For example, for smFISH against ACTB, we were able to resolve individual ACTB mRNA puncta post- expansion even within transcriptional foci in the nucleus (Fig. ExFISH: Design and Validation of RNA Anchoring Chemistry 1c, versus 1d), which can be dense with mRNA due to transcriptional bursting. Thus, ExFISH is capable of supporting single molecule RNA readout in the expanded state. Since Label-IT also reacts to DNA, the ExFISH process enables uniform expansion of the nucleus (Supplementary Fig. 4). The isotropy of ExFISH (Supplementary Fig. 5) was numerically similar to that observed when protein targets were labeled and expanded in the original ExM protocol1. In recent ExM protocols in which proteins are anchored to the same hydrogel as used in ExFISH, with a similar linker9,10, the distortion is small (a few percent distortion, in cells and tissues). These earlier results, since they were obtained with similar polymer chemistry, serve to bound the ExFISH distortion. The expansion factor is slightly lower than in our original ExM paper (i.e., ~3.3× versus ~4×, expansion factors can be found in Figure Legends of this manuscript) due to the salt required to support hybridization of probes. Author Manuscript Author Manuscript Nat Methods. Author manuscript; available in PMC 2017 January 04. Nanoscale Imaging of lncRNA with ExFISH Nanoscale Imaging of lncRNA with ExFISH We imaged long non-coding RNAs (lncRNAs) known to serve structural roles in cell biology. We imaged the lncRNA XIST, whose role in inactivating the X chromosome may depend on initial association with specific chromatin subregions through a process which is still being revealed11. The pre-expansion image (Fig. 1f) shows two bright globular fluorescent regions, presumably corresponding to the X chromosomes of HEK cells undergoing inactivation11–13, but post-expansion, individual puncta were apparent both within the globular regions as well as nearby (Fig. 1g). We additionally used ExFISH to examine the previously described14 ring-shaped morphology of ensembles of NEAT1 lncRNAs (Fig. 1h), which has been hypothesized to play an important role in gene Author Manuscript Chen et al. Chen et al. Page 4 Page 4 expression regulation and nuclear mRNA retention15. Before expansion, NEAT1 presents in the form of bright, diffraction-limited puncta (Fig. 1h, Fig. 1i), but after expansion, the ring- shaped morphology becomes clear (Fig. 1h, Fig. 1i). Given the complex 3-D structure of the genome16, mapping lncRNAs may be useful in defining key chromatin regulatory complexes and their spatial configurations. Author Manuscript Super-resolved, Multiplexed Imaging of RNA with ExFISH The combination of covalent RNA anchoring to the ExM gel, and the de-crowding of the local environment that results from expansion, could facilitate strategies that have been proposed for multiplexed RNA readout17–19 based upon sequential hybridization with multiple probe sets. In order to facilitate multiple cycles of FISH, we re-embedded expanded specimens in charge-neutral polyacrylamide. This process allowed expanded gels to be immobilized for multi-round imaging, and additionally stabilized the expanded specimen throughout salt concentration changes in the protocol. Such re-embedded samples exhibited similar expansion factors as non-re-embedded samples (i.e., ~3×), and were robust to multiple wash-stain cycles as assessed by repeated application of the same probe set (Fig. 2a, Supplementary Fig. 6, showing 5 rounds of smFISH staining against GAPDH on cultured cells). This stability was observed even under stringent wash conditions designed to minimize cycle-to-cycle crosstalk (e.g., 100% formamide). Across the 5 rounds, there was no distortion of the locations of individual RNA spots from round to round (Fig. 2b), nor variance in detection efficiency or signal-to-noise ratio (Fig. 2c, 2d). Having validated the cycle-to-cycle consistency, we next demonstrated the capability of multiplexed ExFISH by applying probes for GAPDH, UBC, NEAT1, USF2, ACTB, and EEF2 in series, enabling 6 individual RNAs to be identified and localized in the same cell (Fig. 2e, Supplementary Fig. 6). Thus, serial FISH is applicable to samples expanded after securing RNA to the swellable polymer as here described, making it straightforward to apply probe sets computationally designed to yield more information per FISH cycle, e.g. MERFISH18–20. Author Manuscript Author Manuscript Nat Methods. Author manuscript; available in PMC 2017 January 04. 3D Nanoscale Imaging of RNA in Mouse Brain Tissue ExM allows for facile super-resolution imaging of thick 3-D specimens such as brain tissue on conventional microscopy hardware1. We applied ExFISH to samples of Thy1-YFP mouse brain tissue21, using the YFP protein to delineate neural morphology (Fig. 3a, 3b). Endogenous YFP protein was anchored to the polyacrylate gel via AcX using the proExM protocol9, and RNA anchored via LabelX. Since smFISH yields signals too dim to visualize in intact tissues using confocal imaging, we applied the previously described technique of hybridization chain reaction (HCR)5, in particular the next-generation DNA HCR amplifier architecture6 (schematic in Supplementary Fig. 7). In samples containing mouse cortical and hippocampal regions, mRNAs for YFP (Fig. 3c) and glutamic acid decarboxylase 1 Gad1 (Fig. 3d) were easily visualized using a widefield microscope, with YFP mRNA well localized to YFP-fluorescing cells (Fig. 3e), and Gad1 mRNA localized to a population of cells with characteristic arrangement throughout specific layers of the cortex and hippocampus22. Examining brain specimens at high magnification using a confocal spinning disk microscope revealed that individual transcripts could be distinguished due to the physical magnification of ExM (Fig. 3f, with YFP and Gad1 mRNA highlighted), with even highly overexpressed transcripts (e.g., YFP) cleanly resolved into individual puncta (Fig. 3f). Author Manuscript Chen et al. Page 5 Page 5 When FISH probes were omitted, minimal background HCR amplification was observed (Supplementary Fig. 8). Given that ExM enables super-resolution imaging on diffraction limited microscopes, which can be scaled to very fast imaging speeds23, we used a commercially available lightsheet microscope on a Thy1-YFP brain slice to enable visualization of multiple transcripts, with single molecule precision, throughout a volume of ~575 μm × 575 μm × 160 μm thick in just 3 hours (~6×1010 voxels in 3 colors; Supplementary Fig. 9, Supplementary Video 1). Author Manuscript Author Author Manuscript HCR amplifies a target binding event into a bright fluorescent signal (Supplementary Fig. 7). A stringent method for assessing detection accuracy is to label individual RNAs with different probe sets bearing different colors24,25, which shows that 50–80% of mRNAs thus targeted will be doubly labeled, when assessed in cell culture; a 50% co-localization is interpreted as detection efficiency (assuming probe independence); this is a lower bound as it excludes false positives. Nat Methods. Author manuscript; available in PMC 2017 January 04. Discussion We present a novel reagent, easily synthesized from commercial precursors, that enables RNA to be covalently anchored for expansion microscopy. The resulting procedure, ExFISH, enables RNAs to be probed through single-molecule FISH labeling as well as hybridization chain reaction (HCR) amplification. We validated RNA retention before versus after expansion, finding excellent yield, and de-crowding of RNAs for more accurate RNA counts and localization. This enabled us to visualize, with nanoscale precision and single molecule resolution, RNA structures such as XIST and NEAT1, long non-coding RNAs whose emergent structure has direct implications for their biological roles. The anchoring was robust enough to support serial smFISH, including repeated washing and probe hybridization steps, and multiplexed readout of RNA identity and location, implying that using probes designed according to specific coding strategies17–19 would support combinatorial multiplexing, in which each additional cycle yields exponentially more transcript information. The covalent anchoring of RNA to the ExM gel may also support enzymatic reactions to be performed in expanded samples – such as reverse transcription, rolling circle amplification (RCA), fluorescent in situ sequencing (FISSEQ)27, and other strategies for transcriptomic readout or SNP detection28, within intact samples. Author Manuscript ExM, being a physical form of magnification, enables nanoscale resolution even on conventional diffraction limited microscopes. Expanding samples makes them transparent and homogeneous in index of refraction, in part because of the volumetric dilution, and in part because of washout of non-anchored components1. Thus, strategies combining ExM with fast diffraction limited methods like lightsheet microscopy23 may result in “best of both worlds” performance metrics: the voxel sizes of classical super-resolution methods, but the voxel acquisition rates of increasingly fast diffraction limited microscopes1. The de- crowding of RNAs enables another key advantage: reducing the effective size of the self- assembled amplification product of HCR. An HCR amplicon of size 500 nm in the post- expanded sample would, because of the greater distance between RNAs, have an effective size of 500 / 3.5 = ~150 nm. The lower packing density of amplicons facilitates the imaging of more transcripts per experiment19 with nanoscale precision. Other methods of achieving brighter signals may be possible. For example, brighter fluorophores such as quantum dots29 or bottlebrush fluorophores30 could obviate the need for signal amplification, in principle. The expanded state may enable better delivery of these and other bulky fluorophores into samples. 3D Nanoscale Imaging of RNA in Mouse Brain Tissue Page 6 ability to map mRNAs at synapses throughout neuronal arbors may be useful for a diversity of questions in neuroscience ranging from plasticity to development to degeneration. Author Manuscript 3D Nanoscale Imaging of RNA in Mouse Brain Tissue In order to assess the false positive and negative rates for single molecule visualization in expanded tissues, we delivered pairs of probe sets targeting the same transcript with different initiators. This scheme results in amplified fluorescent signals of two different colors from the same target (Supplementary Fig 10), giving a measure of the hybridization efficiency. Delivering probe sets against a nonexistent transcript also gives a measure of false positive rate. We delivered a probe set against a missense probe (Dlg4 reversed, Fig. 3g) as well as a nonexistent transcript (mCherry, Supplementary Table. 3), using Thy1-YFP mouse brain samples, and found a low but nonzero spatial density of dim, yet amplified, puncta (1 per 61 μm3 in unexpanded coordinates, Dlg4 reversed; 1 per 48 μm3, mCherry). Essentially zero of these puncta exhibited co-localization (0/1,209 spots, Dlg4 reversed; 4/1,540 spots mCherry). In contrast, when a transcript was present (Actb), a large fraction of the puncta exhibited co-localization (an average of 58% of probes in one color co-localized with other color, 15,866/27,504 spots, Fig. 3h, Supplementary Table 3), indicative of a 75% detection efficiency, comparable to the non-amplified single molecule studies described above. Author Manuscript We used two-color HCR ExFISH against mRNAs to image their position within cellular compartments such as dendritic spines, which require nanoscale resolution for accurate identification or segmentation. We probed the Dlg4 mRNA, which encodes the prominent postsynaptic scaffolding protein PSD-95, and which is known to be dendritically enriched7. We obtained a degree of co-localization (53%, 5,174/9,795 spots) suggesting a high detection efficiency, 73% (Fig. 3i). We also probed the mRNA for Camk2a, finding a detection efficiency of 78% (co-localization, 61%, 8,799/14,440 spots, Supplementary Fig 10). We focused on puncta which were co-localized, thus suppressing false positive errors, and giving a lower-bound on transcript detection (Supplementary Fig 10). Focusing on individual dendrites in these expanded samples revealed that individual Dlg4 (Fig. 3j) and Camk2a (Fig. 3k) mRNAs could indeed be detected in a sparse subset of dendritic spines as well as fine dendritic processes. To facilitate multiplexed HCR readout, we developed modified HCR hairpins that can be disassembled using toe-hold mediated strand displacement26 (Supplementary Fig. 11). These modified HCR amplifiers enable multiple cycles of HCR by disassembling the HCR polymer between subsequent cycles. Given that neurons can have tens of thousands of synapses, and mRNAs can be low copy number, the Author Manuscript Chen et al. Nat Methods. Author manuscript; available in PMC 2017 January 04. Online Methods A table of all reagents and chemicals with part numbers and suppliers can be found in Supplementary Table 1. Preparation of LabelX Preparation of LabelX Acryloyl-X, SE (6-((acryloyl)amino)hexanoic acid, succinimidyl ester, here abbreviated AcX; Thermo-Fisher) was resuspended in anhydrous DMSO at a concentration of 10 mg/mL, aliquoted and stored frozen in a desiccated environment. Label-IT ® Amine Modifying Reagent (Mirus Bio, LLC) was resuspended in the provided Mirus Reconstitution Solution at 1mg/ml and stored frozen in a desiccated environment. To prepare LabelX, 10 μL of AcX (10 mg/mL) was reacted with 100 μL of Label-IT ® Amine Modifying Reagent (1 mg/mL) overnight at room temperature with shaking. LabelX was subsequently stored frozen (−20 °C) in a desiccated environment until use. Author Manuscript Discussion Other amplification strategies may be possible as well, including enzymatic (e.g., RCA28, tyramide amplification22, HRP amplification) as well as nonenzymatic (e.g., branched DNA) methods, although reaction efficiency and diffusion of reagents into the sample must be considered. Author Manuscript Author Manuscript ExFISH may find many uses in neuroscience and other biological fields. In the brain, for example, RNA is known to be trafficked to specific synapses as a function of local synaptic activity31 and intron content32, and locally translated7,33,34, and the presence and translation of axonal RNAs remains under investigation35. We anticipate that, coupled to Chen et al. Page 7 straightforward multiplexed coding schemes, this method can be used for transcriptomic profiling of neuronal cell-types in situ, as well as for the super-resolved characterization of neuronal connectivity and synaptic organization in intact brain circuits, key for an integrative understanding of the mechanisms underlying neural circuit function and dysfunction. More broadly, visualizing RNAs within cells, and their relationship with RNA processing and trafficking machinery, may reveal new insights throughout biology and medicine. Author Manuscript Online Methods Nat Methods. Author manuscript; available in PMC 2017 January 04. Cell Culture and Fixation HeLa (ATCC CCL-2) cells and HEK293-FT cells (Invitrogen) were cultured on Nunc Lab- Tek II Chambered Coverglass (Thermo Scientific) in D10 medium (Cellgro) supplemented with 10% FBS (Invitrogen), 1% penicillin/streptomycin (Cellgro), and 1% sodium pyruvate (BioWhittaker). Cells were authenticated by the manufacturer and tested for mycoplasma contamination to their standard levels of stringency, and were here used because they are common cell lines for testing new tools. Cultured cells were washed once with DPBS (Cellgro), fixed with 10% formalin for 10 mins, and washed twice with 1× PBS. Fixed cells were then stored in 70% Ethanol at 4°C until use. Author Manuscript LabelX Treatment of Cultured Cells and Brain Slices Author Manuscript Fixed cells were washed twice with 1× PBS, once with 20 mM MOPS pH 7.7, and incubated with LabelX diluted to a desired final concentration in MOPS buffer (20 mM MOPS pH 7.7) at 37 °C overnight followed by two washes with 1× PBS. For cells, ranges of LabelX were used that resulted in a Label-IT ® Amine concentration of 0.006–0.02 mg/mL; higher concentrations resulted in somewhat dimmer smFISH staining (Supplementary Fig. 12), but otherwise no difference in staining quality was observed with Label-IT ® Amine concentrations in this range. For Fig. 1e, Supplementary Fig 1, Supplementary Fig 2, and Supplementary Fig 3 fixed cells were incubated with LabelX diluted to a final Label-IT ® Amine concentration of 0.02 mg/mL. For all other experiments in cells, fixed cells were treated with LabelX diluted to a final Label-IT ® Amine concentration of 0.006 mg/mL. Brain slices, as prepared above, were incubated with 20mM MOPS pH 7.7 for 30 mins and subsequently incubated with LabelX diluted to a final Label-IT ® Amine concentration of 0.1 mg/mL (due to their increased thickness and increased fragmentation from formaldehyde postfixation) in MOPS buffer (20 mM MOPS pH 7.7) at 37°C overnight. For YFP retention, slices were treated with 0.05 mg/mL AcX in PBS for >6 hours @ RT. Author Manuscript Mouse perfusion All methods for animal care and use were approved by the Massachusetts Institute of Technology Committee on Animal Care and were in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All solutions below were made up in 1× phosphate buffered saline (PBS) prepared from nuclease free reagents. Mice were anesthetized with isoflurane and perfused transcardially with ice cold 4% paraformaldehyde. Brains were dissected out, left in 4% paraformaldehyde at 4°C for one day, before moving to PBS containing 100 mM glycine. Slices (50 μm and 200 μm) were sliced on a vibratome (Leica VT1000S) and stored at 4 °C in PBS until use. The mouse used in Fig. 3 and related analyses was a Thy1-YFP (Tg(Thy1-YFP)16Jrs) male mouse in the age range 6–8 weeks. No sample size estimate was performed, since the goal was to demonstrate a technology. No exclusion, randomization or blinding of samples was performed. Author Manuscript Nat Methods. Author manuscript; available in PMC 2017 January 04. Page 8 Page 8 Chen et al. smFISH in Fixed Cultured Cells Before Expansion Fixed cells were briefly washed once with wash buffer (10% formamide, 2× SSC) and hybridized with RNA FISH probes in hybridization buffer (10% formamide, 10% dextran sulfate, 2× SSC) overnight at 37 °C. Following hybridization, samples were washed twice with wash buffer, 30mins per wash, and washed once with 1× PBS. Imaging was performed in 1× PBS. smFISH probe sets targeting the human transcripts for TFRC, ACTB, GAPDH, XIST, and 5′ portion of NEAT1 were ordered from Stellaris with Quasar 570 dye. Probe sets against UBC, EEF2, USF2, TOP2A and full length NEAT1 were synthesized, conjugated to fluorophores, and subsequently purified by HPLC as described previously36. Oligonucleotide sequences for probe sets and accession numbers can be found in Supplementary Table 4. Author Manuscript Nat Methods. Author manuscript; available in PMC 2017 January 04. Gelation, Digestion and Expansion Monomer solution (1× PBS, 2 M NaCl, 8.625% (w/w) sodium acrylate, 2.5% (w/w) acrylamide, 0.15% (w/w) N,N′-methylenebisacrylamide) was mixed, frozen in aliquots, and thawed before use. Monomer solution was cooled to 4°C before use. For gelling cultured cells treated with LabelX, a concentrated stock of VA-044 (25% w/w, chosen instead of the Ammonium persulfate (APS)/Tetramethylethylenediamine (TEMED) of the original ExM protocol1 because APS/TEMED resulted in autofluorescence that was small in magnitude but appreciable in the context of smFISH) was added to the monomer solution to a final concentration of 0.5% (w/w) and degassed in 200 μl aliquots for 15 mins. Cells were briefly incubated with the monomer solution plus VA-044 and transferred to a humidified chamber. Subsequently, the humidified chamber was purged with nitrogen gas. To initiate gelation, the humidified chamber was transferred to a 60 °C incubator for two hours. For gelling brain slices treated with LabelX, gelation was performed as in the original ExM protocol (since, Author Manuscript Chen et al. Chen et al. Page 9 Page 9 with HCR amplification, the slight autofluorescence of APS/TEMED was negligible). Gelled cultured cells and brain slices were digested with Proteinase K (New England Biolabs) diluted 1:100 to 8 units/mL in digestion buffer (50 mM Tris (pH 8), 1 mM EDTA, 0.5% Triton X-100, 500 mM NaCl) and digestion was carried out overnight at 37 °C. The gels expand slightly in the high osmolarity digestion buffer (~1.5×). After digestion, gels were stored in 1× PBS until use and expansion was carried out as previously described. with HCR amplification, the slight autofluorescence of APS/TEMED was negligible). Gelled cultured cells and brain slices were digested with Proteinase K (New England Biolabs) diluted 1:100 to 8 units/mL in digestion buffer (50 mM Tris (pH 8), 1 mM EDTA, 0.5% Triton X-100, 500 mM NaCl) and digestion was carried out overnight at 37 °C. The gels expand slightly in the high osmolarity digestion buffer (~1.5×). After digestion, gels were stored in 1× PBS until use and expansion was carried out as previously described. Author Manuscript smFISH Staining After Expansion Expanded gels were incubated with wash buffer (10% formamide, 2× SSC) for 30 mins at room temperature and hybridized with RNA FISH probes in hybridization buffer (10% formamide, 10% dextran sulfate, 2× SSC) overnight at 37 °C. Analysis of Expansion Isotropy smFISH images before and after expansion of TOP2A was rigidly aligned via two control points using the FIJI plugin Turboreg38. Spots were localized and counted via a custom spot counting Matlab code developed by the Raj lab (complete source code and instructions can be found at https://bitbucket.org/arjunrajlaboratory/rajlabimagetools/wiki/Home). Length measurements were performed among all pairs of points before expansion and the corresponding pairs of points after expansion via a custom Matlab script. Measurement error was defined as the absolute difference between the before and after expansion length measurements (Supplementary Fig. 5c). Author Manuscript Image Processing and Analysis of smFISH performed on Cultured Cells Image Processing and Analysis of smFISH performed on Cultured Cells Widefield images of smFISH staining performed before or after expansion were first processed using a rolling-ball background subtraction algorithm (FIJI)37 with a 200 pixel radius. Subsequently, maximum intensity Z-projections of these images were generated. Spots were then localized and counted using a code developed by the Raj lab and available online (http://rajlab.seas.upenn.edu/StarSearch/launch.html). This image analysis was performed for Fig. 1c–e, Fig.2a–c, Supplementary Fig. 2–4, 6, 8. Widefield images of smFISH staining performed before or after expansion were first processed using a rolling-ball background subtraction algorithm (FIJI)37 with a 200 pixel radius. Subsequently, maximum intensity Z-projections of these images were generated. Spots were then localized and counted using a code developed by the Raj lab and available online (http://rajlab.seas.upenn.edu/StarSearch/launch.html). This image analysis was performed for Fig. 1c–e, Fig.2a–c, Supplementary Fig. 2–4, 6, 8. smFISH Staining After Expansion Author Manuscript Nat Methods. Author manuscript; available in PMC 2017 January 04. Gelation, Digestion and Expansion Following hybridization, samples were washed twice with wash buffer, 30 minutes per wash, and washed once with 1× PBS for another 30 mins. Imaging was performed in 1× PBS. Staining of Re-embedded Gels Staining of Re-embedded Gels Re-embeded staining of gels were performed with exact conditions as described above for expanded gels, except post-hybridization washes were changed to twice with wash buffer (10% formamide), 60 minutes per wash. Probes were removed for multiple rounds of hybridization via treatment with DNAse I or 100% formamide. For DNAse I, samples were treated with DNAse I at 0.5 U/μL for 6 hours at RT. For formamide stripping, samples were treated with 100% formamide at 6 hours at 37C. Author Manuscript Re-embeded staining of gels were performed with exact conditions as described above for expanded gels, except post-hybridization washes were changed to twice with wash buffer (10% formamide), 60 minutes per wash. Probes were removed for multiple rounds of hybridization via treatment with DNAse I or 100% formamide. For DNAse I, samples were treated with DNAse I at 0.5 U/μL for 6 hours at RT. For formamide stripping, samples were treated with 100% formamide at 6 hours at 37C. Probes were removed for multiple rounds of hybridization via treatment with DNAse I or 100% formamide. For DNAse I, samples were treated with DNAse I at 0.5 U/μL for 6 hours at RT. For formamide stripping, samples were treated with 100% formamide at 6 hours at 37C. Probe Design for HCR-FISH Probe sequences and accession numbers for mRNA targets can be found in Supplementary Table 4. Probes were designed for HCR-FISH by tiling the CDS of mRNA targets with 22- mer oligos spaced by 3–7 bases. HCR initiators were appended to tiled sequences via a 2 base spacer (AA). For 2 color probe-sets, even and odd tiled probes were assigned different HCR-initiators to allow for amplification in different color channel. Author Manuscript Re-embedding of Expanded Gels in Acrylamide Matrix For serial staining in cells, expanded gels were re-embeded in acrylamide to stabilize the gels in the expanded state. Briefly: gels were expanded in water and cut manually to ~1 mm thickness with a stainless steel blade. Cut gels were incubated in 3% acrylamide, 0.15% N,N ′-Methylenebisacrylamide with 0.05% APS, 0.05% TEMED and 5 mM Tris ph 10.5 for 20 minutes on a shaker. There is a ~30% reduction in gel size during this step. Excess solution is removed from the gels and the gels are dried with light wicking from a laboratory wipe. Gels are placed on top of a bind-silane treated (see below) coverslip or glass bottom plate with a coverslip placed on top of the gels before moving into a container and purged with nitrogen. The container is moved to a 37 °C incubator for gelation for 1.5 hours. Author Manuscript Page 10 Page 10 Chen et al. Staining of Re-embedded Gels Bind-silane Treatment of Coverslips Coverslips and glass bottom 24 well plates were treated with Bind-Silane, a silanization reagent which incorporates acryloyl groups onto the surface of glass to perform in free radical polymerization. Briefly, 5 μL of Bind-Silane reagent was diluted into 8 mL of ethanol, 1.8 mL of ddH2O and 200 μL of acetic acid. Coverslips and glass bottom 24 well plates were washed with ddH2O followed by 100% ethanol, followed by the diluted Bind- Silane reagent. After a brief wash with the diluted Bind-Silane reagent, the cover-slip was dried, then washed with 100% ethanol, and then dried again. Coverslips were prepared immediately before use. Author Manuscript Probe Design for HCR-FISH Probe sequences and accession numbers for mRNA targets can be found in Supplementary Table 4. Probes were designed for HCR-FISH by tiling the CDS of mRNA targets with 22- mer oligos spaced by 3–7 bases. HCR initiators were appended to tiled sequences via a 2 base spacer (AA). For 2 color probe-sets, even and odd tiled probes were assigned different HCR-initiators to allow for amplification in different color channel. RNA FISH with Hybridization Chain Reaction (HCR) Amplification Gelled samples were incubated with wash buffer (20% formamide, 2× SSC) for 30mins at room temperature and hybridized with HCR initiator tagged FISH probes in hybridization buffer (20% formamide, 10% dextran sulfate, 2× SSC) overnight at 37 °C. Following hybridization, samples were washed twice with wash buffer, 30mins per wash, and incubated with 1× PBS for 2hrs at 37°C. Subsequently, samples were incubated with 1× PBS for at least 6hrs at room temperature. Before HCR amplification, hybridized samples were pre- incubated with amplification buffer (10% dextran sulfate, 5× SSC, 0.1% Tween 20) for 30 mins. To initiate amplification, HCR hairpin stocks (Alexa 456 and Alexa 647 fluorophores) at 3 μM were snap-cooled by heating to 95°C for 90 seconds, and leaving to cool at room temperature for 30 mins. Gelled samples were then incubated with HCR hairpins diluted to 60 nM in amplification buffer for 3hrs at room temperature. After amplification, gels were washed with 5× SSCT (5× SSC, 0.1% Tween 20) twice with one hour per wash. Author Manuscript Imaging of Expanded Brain Slices For epifluorescence imaging of brain sections before and after expansion (Fig. 3a–e) and to quantify expansion factors of tissue slices specimens were imaged on a Nikon Ti-E epifluorescence microscope with a 4× 0.2 NA air objective, a SPECTRA X light engine (Lumencor), and a 5.5 Zyla sCMOS camera (Andor), controlled by NIS-Elements AR software. Author Manuscript Post-expansion confocal imaging of expanded brain tissue was performed on an Andor spinning disk (CSU-X1 Yokogawa) confocal system with a 40× 1.15 NA water objective (Fig. 3f–k, Supplementary Fig. 10) on a Nikon TI-E microscope body. GFP was excited with a 488 nm laser, with 525/40 emission filter. Alexa 546 HCR amplicons were excited with a 561 nm laser with 607/36 emission filter. Alexa 647 amplicons were excited with a 640 nm laser with 685/40 emission filter. Gels were expanded in with 3 washes, 15 minutes each of 0.05× SSC. The expansion factor can be controlled with the salt concentration, we found that 0.05× SSC gives 3× expansion, while still giving enough salt for hybridization stability. To stabilize the gels against drift during imaging following expansion, gels were placed in glass bottom 6 well plates with all excess liquid removed. If needed, liquid low melt agarose (2% w/w) was pipetted around the gel and allowed to solidify, to encase the gels before imaging. Lightsheet imaging was performed on a Zeiss Z.1 lightsheet microscope. Briefly, the sample was fixed on a custom-made plastic holder using super glue and mounted on the freely rotating stage of the Z.1 lightsheet. Lightsheets were generated by two illumination objectives (5×, NA 0.1), and the fluorescence signal detected by a 20× water immersion objective (NA 1.0). Both lightsheets were used for data collection. The image volume dimensions of a single tile were 1400×1400×1057 pixels, with a voxel size of 227 nm laterally and 469 nm axially. The laserlines used for excitation were 488 nm, 561 nm and 638 nm. The individual laser transmissions were set to 5%, with the maximum output of 50 mW (488 nm and 561 nm) and 75 mW (638 nm). Optical filters used to separate and clean the fluorescence response included a Chroma T560lpxr as a dichroic, and a Chroma 59001m for GFP and 59007m for Alexa 546 and Alexa 647. Two PCO.Edge 5.5m sCMOS cameras were used to capture two fluorescence channels simultaneously. Imaging of Cultured Cells using ExFISH Both cultured cells as well as LabelX treated and expanded cultured cells were imaged on a Nikon Ti-E epifluorescence microscope with a SPECTRA X light engine (Lumencor), and a Nat Methods. Author manuscript; available in PMC 2017 January 04. Chen et al. Page 11 Page 11 5.5 Zyla sCMOS camera (Andor), controlled by NIS-Elements AR software. For Fig. 1c, 1d, and Supplementary Fig 3–5 a 40× 1.15 NA water immersion objective was used. For all other experiments with cultured cells, a 60× 1.4 NA oil immersion objective was used. 5.5 Zyla sCMOS camera (Andor), controlled by NIS-Elements AR software. For Fig. 1c, 1d, and Supplementary Fig 3–5 a 40× 1.15 NA water immersion objective was used. For all other experiments with cultured cells, a 60× 1.4 NA oil immersion objective was used. 5.5 Zyla sCMOS camera (Andor), controlled by NIS-Elements AR software. For Fig. 1c, 1d, and Supplementary Fig 3–5 a 40× 1.15 NA water immersion objective was used. For all other experiments with cultured cells, a 60× 1.4 NA oil immersion objective was used. Author Manuscript For imaging smFISH probes labeled with fluorophores, the following filter cubes (Semrock, Rochester, NY) were used: Alexa 488, GFP-1828A-NTE-ZERO; Quasar 570, LF561-B-000; Alexa 594, FITC/TXRED-2X-B-NTE; Atto 647N, Cy5-4040C-000. For imaging smFISH probes labeled with fluorophores, the following filter cubes (Semrock, Rochester, NY) were used: Alexa 488, GFP-1828A-NTE-ZERO; Quasar 570, LF561-B-000; Alexa 594, FITC/TXRED-2X-B-NTE; Atto 647N, Cy5-4040C-000. Nat Methods. Author manuscript; available in PMC 2017 January 04. Two Color Analysis in Slices Author Manuscript A sliding window averaging (or minimization) scheme in Z (3 optical sections) was used to suppress movement artifacts before spot detection processing. RNA puncta were detected via a custom 3D spot counting Matlab code developed by the Raj lab; complete source code and instructions can be found at https://bitbucket.org/arjunrajlaboratory/rajlabimagetools/ wiki/Home. Spot centroids were extracted from both color channels, and spots were determined to be co- localized if their centroids were within a 3 pixel radius in the x,y dimensions and a 2 pixel radius in the z dimension. Imaging of Expanded Brain Slices Tiled datasets were taken with the Zeiss ZEN Software, and subsequently merged and processed with FIJI, Arivis Vision4D and Bitplane Imaris. Author Manuscript Page 12 Page 12 Chen et al. Supplementary Material Refer to Web version on PubMed Central for supplementary material. Nat Methods. Author manuscript; available in PMC 2017 January 04. HCR Reversal via Toe-Hold Mediated Strand Displacement HCR amplification commences upon the addition of two HCR metastable amplifier hairpins. We designed a pair of HCR amplifiers, B2H1T and B2H2 (see below for sequence), where B2H1T bears a 6bp toe-hold for strand displacement. To initiate HCR amplification, aliquots of these amplifiers at 3 μM were snap-cooled by heating to 95 °C for 90 seconds, and leaving to cool at room temperature for 30 mins. Gelled samples were then incubated with HCR hairpins diluted to 60 nM in amplification buffer for 3hrs at room temperature. After amplification, gels were washed with 5× SSCT (5× SSC, 0.1% Tween 20) twice with one hour per wash. Subsequently, HCR reversal was initiated by the addition of a displacement strand (see below for sequence) at 200 nM in 5× SSCT. Author Manuscript B2H1T: ggCggTTTACTggATgATTgATgAggATTTACgAggAgCTCAgTCCATCCTCg TAAATCCTCA TCAATCATCAAATAG B2H2: /5′-Alexa546-C12/ CCTCgTAAATCCTCATCAATCATCCAgTAAACCgCCgATgATTgATgAggA TTTACgAggA TggACTgAgCT Displacement Strand: CTATTTGATGATTGATGAGGATTTAcGAGGATGGAcTGAGcT B2H1T: ggCggTTTACTggATgATTgATgAggATTTACgAggAgCTCAgTCCATCCTCg TAAATCCTCA TCAATCATCAAATAG B2H2: /5′-Alexa546-C12/ CCTCgTAAATCCTCATCAATCATCCAgTAAACCgCCgATgATTgATgAggA TTTACgAggA TggACTgAgCT Displacement Strand: CTATTTGATGATTGATGAGGATTTAcGAGGATGGAcTGAGcT B2H1T: ggCggTTTACTggATgATTgATgAggATTTACgAggAgCTCAgTCCATCCTCg TAAATCCTCA TCAATCATCAAATAG B2H2: /5′-Alexa546-C12/ CCTCgTAAATCCTCATCAATCATCCAgTAAACCgCCgATgATTgATgAggA TTTACgAggA TggACTgAgCT Displacement Strand: CTATTTGATGATTGATGAGGATTTAcGAGGATGGAcTGAGcT B2H1T: Author Manuscript Displacement Strand: CTATTTGATGATTGATGAGGATTTAcGAGGATGGAcTGAGcT References Author Manuscript 1. Chen F, Tillberg PW, Boyden ES. Expansion microscopy. Science (80-). 2015; 347:543–548. 2. Femino AM, Fay F, Fogarty K, Singer R. Visualization of Single RNA Transcripts in Situ. Science (80-). 1998; 280:585–590. 3. Levsky JM, Singer RH. Fluorescence in situ hybridization: past, present and future. J Cell Sci. 2003; 116:2833–2838. [PubMed: 12808017] 4. Raj A, van den Bogaard P, Rifkin SA, van Oudenaarden A, Tyagi S. Imaging individual mRNA molecules using multiple singly labeled probes. Nat Methods. 2008; 5:877–9. [PubMed: 18806792] 5. Choi HMT, et al. Programmable in situ amplification for multiplexed imaging of mRNA expression. Nat Biotechnol. 2010; 28:1208–12. [PubMed: 21037591] 6. Choi HMT, Beck VA, Pierce NA. Next-Generation in Situ Hybridization Chain Reaction: Higher Gain, Lower Cost, Greater Durability. ACS Nano. 2014; 8:4284–4294. [PubMed: 24712299] 7. Cajigas IJ, et al. The local transcriptome in the synaptic neuropil revealed by deep sequencing and high-resolution imaging. Neuron. 2012; 74:453–66. [PubMed: 22578497] Author Manuscript 8. Wang F, et al. RNAscope: A novel in situ RNA analysis platform for formalin-fixed, paraffin- embedded tissues. J Mol Diagnostics. 2012; 14:22–29. 9. Tillberg PW, Chen F, et al. Expansion Microscopy of Biological Specimens with Protein Retention. Nature Biotechnology. In Press. 10. Chozinski TJ, et al. Expansion microscopy with conventional antibodies and fluorescent proteins. Nat Methods. 2016; doi: 10.1038/nmeth.3833 11. Engreitz JM, et al. The Xist lncRNA exploits three-dimensional genome architecture to spread across the X chromosome. Science. 2013; 341:1237973. [PubMed: 23828888] 12. Panning B, Dausman J, Jaenisch R. X chromosome inactivation is mediated by Xist RNA stabilization. Cell. 1997; 90:907–16. [PubMed: 9298902] 13. Plath K, Mlynarczyk-Evans S, Nusinow DA, Panning B. Xist RNA and the mechanism of X chromosome inactivation. Annu Rev Genet. 2002; 36:233–78. [PubMed: 12429693] 14. Mito M, Kawaguchi T, Hirose T, Nakagawa S. Simultaneous multicolor detection of RNA and proteins using super-resolution microscopy. Methods. 2015; doi: 10.1016/j.ymeth.2015.11.007 Author Manuscript 15. Clemson CM, et al. An architectural role for a nuclear noncoding RNA: NEAT1 RNA is essential for the structure of paraspeckles. Mol Cell. 2009; 33:717–26. [PubMed: 19217333] 16. Lieberman-Aiden E, et al. Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science. 2009; 326:289–93. [PubMed: 19815776] 17. Lubeck E, Cai L. Single-cell systems biology by super-resolution imaging and combinatorial labeling. Nat Methods. 2012; 9:743–8. [PubMed: 22660740] 18. Lubeck E, Coskun AF, Zhiyentayev T, Ahmad M, Cai L. Single-cell in situ RNA profiling by sequential hybridization. Acknowledgments Lightsheet imaging was performed in the W.M. Keck Facility for Biological Imaging at Whitehead Institute for Biomedical Research. We would like to acknowledge W. Salmon for assistance with the Zeiss Z.1 lightsheet and S. Olenych from Carl Zeiss Microscopy for providing the microscopy filters, and H. T. Choi and N. Pierce for advice and consultation on HCR. E.R.D. is supported by NIH CEGS grant P50 HG005550, NIH CEGS grant 1 RM1 HG008525, and NSF GRF grant DGE1144152. A.T.W. acknowledges the Hertz Foundation Fellowship. F.C. acknowledges the NSF Fellowship and Poitras Fellowship. E.S.B. acknowledges support by the New York Stem Cell Foundation-Robertson Award, NSF CBET 1053233 (E.S.B.), MIT Media Lab Consortium, the MIT Synthetic Intelligence Project, NIH Director’s Pioneer Award 1DP1NS087724 (E.S.B.), NIH 2R01DA029639 (E.S.B.), NIH Director’s Transformative Award 1R01MH103910, NIH 1R24MH106075, IARPA D16PC00008 (G.M.C.), and Jeremy and Joyce Wertheimer. J.-B. C. was supported by a Simons Postdoctoral Fellowship. Author Manuscript Nat Methods. Author manuscript; available in PMC 2017 January 04. Page 13 Chen et al. References Nat Methods. 2014; 11:360–1. [PubMed: 24681720] 19. Chen KH, Boettiger AN, Moffitt JR, Wang S, Zhuang X. Spatially resolved, highly multiplexed RNA profiling in single cells. Science (80-). 2015; 348:aaa6090–aaa6090. 20. Beliveau BJ, et al. Versatile design and synthesis platform for visualizing genomes with Oligopaint FISH probes. Proc Natl Acad Sci U S A. 2012; 109:21301–6. [PubMed: 23236188] 21. Feng G, et al. Imaging neuronal subsets in transgenic mice expressing multiple spectral variants of GFP. Neuron. 2000; 28:41–51. [PubMed: 11086982] Author Manuscript 22. Lein ES, et al. Genome-wide atlas of gene expression in the adult mouse brain. Nature. 2007; 445:168–76. [PubMed: 17151600] 23. Huisken J, Swoger J, Del Bene F, Wittbrodt J, Stelzer EHK. Optical Sectioning Deep Inside Live Embryos by Selective Plane Illumination Microscopy. Science. 2004; 305:1007–1009. [PubMed: 15310904] 24. Batish M, van den Bogaard P, Kramer FR, Tyagi S. Neuronal mRNAs travel singly into dendrites. Proc Natl Acad Sci. 2012; 109:4645–4650. [PubMed: 22392993] 25. Cabili MN, et al. Localization and abundance analysis of human lncRNAs at single-cell and single- molecule resolution. Genome Biol. 2015; 16:20. [PubMed: 25630241] Nat Methods. Author manuscript; available in PMC 2017 January 04. Chen et al. Page 14 Page 14 26. Zhang DY, Seelig G. Dynamic DNA nanotechnology using strand-displacement reactions. Nat Chem. 2011; 3:103–113. [PubMed: 21258382] Author Manuscript 27. Lee JH, et al. Highly Multiplexed Subcellular RNA Sequencing in Situ. Science (80-). 2014; 343:1360–1363. 28. Ke R, et al. In situ sequencing for RNA analysis in preserved tissue and cells. Nat Methods. 2013; 10:857–60. [PubMed: 23852452] 29. Bruchez M, et al. Semiconductor nanocrystals as fluorescent biological labels. Science. 1998; 281:2013–6. [PubMed: 9748157] 30. Fouz MF, et al. Bright Fluorescent Nanotags from Bottlebrush Polymers with DNA-Tipped Bristles. ACS Cent Sci. 2015; 1:431–438. [PubMed: 27163005] 31. Steward O, Wallace CS, Lyford GL, Worley PF. Synaptic activation causes the mRNA for the leg Arc to localize selectively near activated postsynaptic sites on dendrites. Neuron. 1998; 21:741– 751. [PubMed: 9808461] 32. Buckley PT, et al. Cytoplasmic Intron Sequence-Retaining Transcripts Can Be Dendritically Targeted via ID Element Retrotransposons. Neuron. 2011; 69:877–884. [PubMed: 21382548] Author Manuscript 33. Steward O, Schuman EM. Compartmentalized synthesis and degradation of proteins in neurons. Neuron. 2003; 40:347–359. [PubMed: 14556713] 34. Buxbaum AR, Wu B, Singer RH. Single -Actin mRNA Detection in Neurons Reveals a Mechanism for Regulating Its Translatability. Science (80-). 2014; 343:419–422. 35. Jung H, Yoon BC, Holt CE. Nat Methods. Author manuscript; available in PMC 2017 January 04. References Axonal mRNA localization and local protein synthesis in nervous system assembly, maintenance and repair. Nat Rev Neurosci. 2012; 13:308–24. [PubMed: 22498899] 36. Raj, A.; Tyagi, S. Methods in enzymology. Elsevier Inc; 2010. Detection of individual endogenous RNA transcripts in situ using multiple singly labeled probes; p. 472 37. Schindelin J, et al. Fiji: an open-source platform for biological-image analysis. Nat Methods. 2012; 9:676–82. [PubMed: 22743772] 38. Thévenaz P, Ruttimann UE, Unser M. A pyramid approach to subpixel registration based on intensity. IEEE Trans Image Process. 1998; 7:27–41. [PubMed: 18267377] Author Manuscript Author Manuscript Page 15 Chen et al. Figure 1. Design and validation of ExFISH chemistry. (a) Acryloyl-X SE (top left) is reacted to Label- IT® amine (top right) via NHS-ester chemistry to form LabelX (middle), which serves to make RNA gel-anchorable by alkylating its bases (e.g., the N7 position of guanines) (bottom). (b) Workflow for ExFISH: biological specimens are treated with LabelX (left), which enables RNA to be anchored to the ExM gel (middle). Anchored RNA can be probed via hybridization (right), after gelation, digestion, and expansion. (c) smFISH image of ACTB before expansion. Inset shows zoomed-in region, highlighting transcription sites in nucleus. (d) As in (c), using ExFISH. (e) smFISH counts before versus after expansion for seven different transcripts (n = 59 cells; each symbol represents one cell). (f) smFISH image of XIST long non-coding RNA (lncRNA) in the nucleus of a HEK293 cell before expansion (white line denotes nuclear envelope in f–h). (g) As in (f), using ExFISH. (h) smFISH image before expansion (top), and using ExFISH (bottom), of NEAT1 lncRNA in the nucleus of a HeLa cell. Magenta and green indicate probesets binding to different parts of the 5′ (1–3756 nts) of NEAT1 (see Methods). (i) Insets showing a NEAT1 cluster (boxed region of (h)) with Author Manuscript Author Manuscript Author Manuscript Author Manuscript Figure 1. Figure 1. Design and validation of ExFISH chemistry. (a) Acryloyl-X SE (top left) is reacted to Label- IT® amine (top right) via NHS-ester chemistry to form LabelX (middle), which serves to make RNA gel-anchorable by alkylating its bases (e.g., the N7 position of guanines) (bottom). (b) Workflow for ExFISH: biological specimens are treated with LabelX (left), which enables RNA to be anchored to the ExM gel (middle). Anchored RNA can be probed via hybridization (right), after gelation, digestion, and expansion. (c) smFISH image of ACTB before expansion. Nat Methods. Author manuscript; available in PMC 2017 January 04. Nat Methods. Author manuscript; available in PMC 2017 January 04. smFISH (left) and ExFISH (right). Scale bars (white, in pre-expansion units; blue scale bars are divided by the expansion factor noted))): (c, d) 10 μm (expansion factor, 3.3×), inset 2 μm; (f, g) 2 μm (3.3×), Z scale represented by color coding in pre-expansion units; (h) 2 μm (3.3×); (i) 200 nm (3.3×). References Inset shows zoomed-in region, highlighting transcription sites in nucleus. (d) As in (c), using ExFISH. (e) smFISH counts before versus after expansion for seven different transcripts (n = 59 cells; each symbol represents one cell). (f) smFISH image of XIST long non-coding RNA (lncRNA) in the nucleus of a HEK293 cell before expansion (white line denotes nuclear envelope in f–h). (g) As in (f), using ExFISH. (h) smFISH image before expansion (top), and using ExFISH (bottom), of NEAT1 lncRNA in the nucleus of a HeLa cell. Magenta and green indicate probesets binding to different parts of the 5′ (1–3756 nts) of NEAT1 (see Methods). (i) Insets showing a NEAT1 cluster (boxed region of (h)) with Author Manuscript smFISH (left) and ExFISH (right). Scale bars (white, in pre-expansion units; blue scale bars are divided by the expansion factor noted))): (c, d) 10 μm (expansion factor, 3.3×), inset 2 μm; (f, g) 2 μm (3.3×), Z scale represented by color coding in pre-expansion units; (h) 2 μm (3.3×); (i) 200 nm (3.3×). Page 16 Chen et al. Page 16 smFISH (left) and ExFISH (right). Scale bars (white, in pre-expansion units; blue scale bars are divided by the expansion factor noted))): (c, d) 10 μm (expansion factor, 3.3×), inset 2 μm; (f, g) 2 μm (3.3×), Z scale represented by color coding in pre-expansion units; (h) 2 μm (3.3×); (i) 200 nm (3.3×). smFISH (left) and ExFISH (right). Scale bars (white, in pre-expansion units; blue scale bars are divided by the expansion factor noted))): (c, d) 10 μm (expansion factor, 3.3×), inset 2 μm; (f, g) 2 μm (3.3×), Z scale represented by color coding in pre-expansion units; (h) 2 μm (3.3×); (i) 200 nm (3.3×). Author Manuscript Author Manuscript A Author Manuscript Author Manuscript Page 17 Page 17 Chen et al. Figure 2. Serially hybridized and multiplexed ExFISH. (a) Widefield fluorescence image of ExFISH targeting GAPDH. (b) Boxed region of (a), showing 5 repeated re-stainings following probe removal (see Methods); lower right panel, overlay of the 5 images (with each a different color, red, green, blue, magenta, yellow), showing co-localization. (c) ExFISH RNA counts for each round, normalized to the round 1 count; plotted is mean ± standard error; n = 3 regions of (a). Nat Methods. Author manuscript; available in PMC 2017 January 04. References (d) Signal-to-noise ratio (SNR) of ExFISH across the five rounds of staining of (a), computed as the mean puncta brightness divided by the standard deviation of the background. (e) Composite image showing ExFISH with serially delivered probes against six RNA targets in a cultured HeLa cell (raw images in Supplementary Fig. 6); colors are as follows: NEAT1, blue; EEF2, orange; GAPDH, yellow; ACTB, purple; UBC, green; USF2, light blue. Scale bars (expanded coordinates): (a) 20 μm; (b) 10 μm; (e) 20 μm. Author Manuscript Author Manuscript Author Ma Author Manuscript Figure 2. Inset, one plane of the boxed region (red, YFP protein; cyan, YFP mRNA; magenta, Gad1 mRNA). (g) Confocal image (i) and processed image (ii) of HCR-ExFISH using a missense Dlg4 probe, in Thy1-YFP mouse tissue (green, YFP protein). The raw image (i) uses alternating probes in two colors (red, Dlg4 missense even; blue, Dlg4 missense odd). The processed image (ii) shows zero co-localized spots (magenta). (h) As in (g), but for HCR-ExFISH targeting Actb in Thy1-YFP mouse brain (green, YFP protein; red, Actb even, and blue, Actb odd in (i); co-localized spots in magenta (ii)). (i) Confocal image of hippocampal tissue showing co-localized Dlg4 puncta (magenta) overlaid on YFP (green). (j) Dendrites with Dlg4 mRNA localized to spines (arrows). (i), (ii), two representative examples. (k) As in (j), but with HCR-ExFISH of Camk2a mRNA showing transcripts in dendritic spines and processes. Scale bars (white, in pre-expansion units; blue scale bars are divided by the expansion factor noted): (a) 500 μm; (b–e) 500 μm (expansion factor 2.9×); (f) 50 μm (2.9×), inset 10 μm; (g–i) 10 μm (3×); (j,k) 2 μm (3×). (e,i) maximum-intensity projection (MIP) 27 μm thick (pre-expanded units); (g,h,j,k) MIPs ~1.6 μm thick. Author Manuscript Author Manuscript Figure 3. Figure 3. Nanoscale imaging of RNA in mammalian brain. (a) Widefield fluorescence image of Thy1- YFP mouse brain. (b) Post-expansion widefield image of (a). (c) Widefield fluorescence showing HCR-ExFISH of YFP mRNA in the sample of (b). (d) As in (c), but for Gad1 mRNA. (e) Composite of (b–d), highlighting distribution of Gad1 versus Thy1-YFP mRNAs. (f) Confocal image of mouse hippocampal tissue from (e) showing single RNA puncta. Inset, one plane of the boxed region (red, YFP protein; cyan, YFP mRNA; magenta, Gad1 mRNA). (g) Confocal image (i) and processed image (ii) of HCR-ExFISH using a missense Dlg4 probe, in Thy1-YFP mouse tissue (green, YFP protein). The raw image (i) uses alternating probes in two colors (red, Dlg4 missense even; blue, Dlg4 missense odd). The processed image (ii) shows zero co-localized spots (magenta). (h) As in (g), but for HCR-ExFISH targeting Actb in Thy1-YFP mouse brain (green, YFP protein; red, Actb even, and blue, Actb odd in (i); co-localized spots in magenta (ii)). (i) Confocal image of hippocampal tissue showing co-localized Dlg4 puncta (magenta) overlaid on YFP (green). (j) Dendrites with Dlg4 mRNA localized to spines (arrows). (i), (ii), two representative examples. Figure 2. Figure 2. Serially hybridized and multiplexed ExFISH. (a) Widefield fluorescence image of ExFISH targeting GAPDH. (b) Boxed region of (a), showing 5 repeated re-stainings following probe removal (see Methods); lower right panel, overlay of the 5 images (with each a different color, red, green, blue, magenta, yellow), showing co-localization. (c) ExFISH RNA counts for each round, normalized to the round 1 count; plotted is mean ± standard error; n = 3 regions of (a). (d) Signal-to-noise ratio (SNR) of ExFISH across the five rounds of staining of (a), computed as the mean puncta brightness divided by the standard deviation of the background. (e) Composite image showing ExFISH with serially delivered probes against six RNA targets in a cultured HeLa cell (raw images in Supplementary Fig. 6); colors are as follows: NEAT1, blue; EEF2, orange; GAPDH, yellow; ACTB, purple; UBC, green; USF2, light blue. Scale bars (expanded coordinates): (a) 20 μm; (b) 10 μm; (e) 20 μm. g Serially hybridized and multiplexed ExFISH. (a) Widefield fluorescence image of ExFISH targeting GAPDH. (b) Boxed region of (a), showing 5 repeated re-stainings following probe removal (see Methods); lower right panel, overlay of the 5 images (with each a different color, red, green, blue, magenta, yellow), showing co-localization. (c) ExFISH RNA counts for each round, normalized to the round 1 count; plotted is mean ± standard error; n = 3 regions of (a). (d) Signal-to-noise ratio (SNR) of ExFISH across the five rounds of staining of (a), computed as the mean puncta brightness divided by the standard deviation of the background. (e) Composite image showing ExFISH with serially delivered probes against six RNA targets in a cultured HeLa cell (raw images in Supplementary Fig. 6); colors are as follows: NEAT1, blue; EEF2, orange; GAPDH, yellow; ACTB, purple; UBC, green; USF2, light blue. Scale bars (expanded coordinates): (a) 20 μm; (b) 10 μm; (e) 20 μm. Author Manuscript Author Manuscript Page 18 Chen et al. Figure 3. Nanoscale imaging of RNA in mammalian brain. (a) Widefield fluorescence image of Thy1- YFP mouse brain. (b) Post-expansion widefield image of (a). (c) Widefield fluorescence showing HCR-ExFISH of YFP mRNA in the sample of (b). (d) As in (c), but for Gad1 mRNA. (e) Composite of (b–d), highlighting distribution of Gad1 versus Thy1-YFP mRNAs. (f) Confocal image of mouse hippocampal tissue from (e) showing single RNA puncta. Nat Methods. Author manuscript; available in PMC 2017 January 04. Figure 2. (k) As in (j), but with HCR-ExFISH of Camk2a mRNA showing transcripts in dendritic spines and processes. Scale bars (white, in pre-expansion units; blue scale bars are divided by the expansion factor noted): (a) 500 μm; (b–e) 500 μm (expansion factor 2.9×); (f) 50 μm (2.9×), inset 10 μm; (g–i) 10 μm (3×); (j,k) 2 μm (3×). (e,i) maximum-intensity projection (MIP) 27 μm thick (pre-expanded units); (g,h,j,k) MIPs ~1.6 μm thick. Author Manuscript Author Manuscript
https://openalex.org/W4394890968	https://dl.acm.org/doi/pdf/10.1145/3625468.3647619	English	null	QV4	null	2,024	cc-by	10,740	ABSTRACT Italy. ACM, New York, NY, USA, 11 pages. https://doi.org/10.1145/3625468. 3647619 Volumetric videos allow six degrees of freedom (6DoF) movement for viewers, enabling numerous applications in domains such as en- tertainment, healthcare, and education. MPEG’s Video-based Point Cloud Compression (V-PCC) is a recent new standard for volumet- ric video compression that achieves a considerable compression rate while maintaining the quality of the point cloud sequence. However, V-PCC is hard to fit into existing tiling-based volumetric video streaming framework due to the lack of proper user viewing adaptive techniques. In this paper, we propose QV4 , a Quality-of- Experience (QoE) based streaming pipeline for viewpoint-aware V-PCC-encoded volumetric video. Specifically, we leverage the in- termediate results produced by the V-PCC encoder to achieve effec- tive and efficient viewpoint-aware tiling for V-PCC. We then build a QoE model and a 6DoF movement model based on real-world user data, to predict the users’ viewing experience and behaviors, respectively. The proposed QoE model and 6DoF movement model are combined with viewpoint-aware V-PCC tiling to maximize the visual quality of volumetric videos. Extensive simulations show that by enabling viewpoint-aware adaptation and optimization for V-PCC-encoded volumetric videos, QV4 can achieve up to 14.67% improvement in structural similarity index (SSIM) and 7.39% im- provement in video multi-method assessment fusion (VMAF) over highly dynamic viewing behaviors in a network with limited and fluctuating bandwidth. QV4: QoE-based Viewpoint-Aware V-PCC-encoded Volumetric Video Streaming Yuang Shi School of Computing National University of Singapore Singapore yuangshi@u.nus.edu Wei Tsang Ooi School of Computing National University of Singapore Singapore ooiwt@comp.nus.edu.sg 1https://github.com/google/draco 1 INTRODUCTION Volumetric video is an emerging and highly interactive form of media that immerses users in a dynamic 3D experience [11]. These videos employ a data representation called point cloud, which con- sists of a set of 3D points, each with coordinates and color. How- ever, achieving high-quality 3D object representation in volumetric videos demands point clouds composed of millions, if not billions, of points, posing a challenge for transmission-friendly codecs, adap- tive transmission methods, and accurate prediction of viewing be- havior. The codecs for point cloud compression can be categorized into projection-based codecs, geometry-based codecs, and neural-based codecs. MPEG standardizes two compression technologies for point cloud compression (PCC): video-based PCC (V-PCC) and geometry- based PCC (G-PCC). V-PCC is a projection-based approach that projects each frame of 3D point clouds into three 2D images. The generated 2D images can then be encoded using state-of-the-art video codecs (e.g. HEVC). G-PCC uses octree or triangle soup (trisoup) data structures and performs arithmetical encoding on different attributes. Similarly, Google’s Draco 1 uses the kd-tree data structure to compress grid data and point clouds. PU-GAN [20] and MPU [38] models resort to deep learning techniques, to extract the features of the point cloud surface by neural networks and upsample the point cloud. CCS CONCEPTS • Information systems →Multimedia streaming; • Comput- ing methodologies →Animation. V-PCC, in particular, takes advantage of successful 2D video compression techniques, to achieve up to 100× better compres- sion ratio compared with geometry-based PCC techniques (e.g., Google’s Draco) and neural-based PCC techniques (e.g., MPU [38] or PU-GAN [20]), showing great potential for the applications of volumetric videos. Moreover, as a projection-based method, V-PCC is well suited for dense point cloud sequences since V-PCC can generate continuous and smooth surfaces of 2D projections from dense point clouds [3]. As a result, however, V-PCC cannot be fully utilized by tiling-based streaming techniques, that need to segment the video into small and sparse tiles for adaptive optimization. KEYWORDS Volumetric video, media streaming, video-based point cloud com- pression 2.1 Point Cloud Compression Specifically, to take advantage of V-PCC and decompose the point clouds for viewpoint-aware optimization, we leverage the intermediate results produced by the V-PCC encoder as suggested by Rudolph et al. [30]. To take the viewer’s subjective perception and quality of the viewing experience into account, a QoE model is built based on 7,680 ratings of V-PCC-encoded volumetric videos from 120 users by Cox et al.[6]. Our QoE model can accurately predict the subjective quality of volumetric videos encoded by V- PCC with 0.06 root-mean-square error (RMSE). Additionally, to enable effective and efficient viewpoint adaptation for better user perception quality under diverse and dynamic user viewpoints in 6DoF, we develop a viewpoint prediction model for 6DoF viewpoint prediction based on real-world viewpoint data collected from 26 users by Subramanyam et al. [33] and emulated dynamic viewpoint data, which achieves 0.02 mean absolute error (MAE) and 0.07 MAE on average for normalized translational movement and rotational movement, respectively. We combine the proposed QoE model and viewpoint prediction model with the state-of-the-art bitrate alloca- tion algorithm to maximize the visual quality of volumetric video under dynamic network conditions and diverse user viewpoints. We perform extensive simulations under complex and diverse net- work profiles and viewpoint trajectories to verify the feasibility and performance of QV4 . The results show that QV4 achieves up to 14.67% improvement in SSIM and 7.39% improvement in VMAF. of effective quality metrics, bitrate allocation, and accurate view- ing prediction. Because of the aforementioned limitations, most of the works of tiling-based volumetric video streaming resort to geometry-based compression and deep-learning-based compres- sion [13, 16, 18, 20, 38, 40], which achieve low compression ratios and fail to work under limited and fluctuating network conditions. l g In this paper, we present QV4 , a QoE-based streaming pipeline for Viewpoint-aware V-PCC-encoded Volumetric Video. To our best knowledge, QV4 is the first tile-based volumetric video streaming system that exploits user viewing adaptations on V-PCC-encoded content. Specifically, to take advantage of V-PCC and decompose the point clouds for viewpoint-aware optimization, we leverage the intermediate results produced by the V-PCC encoder as suggested by Rudolph et al. [30]. To take the viewer’s subjective perception and quality of the viewing experience into account, a QoE model is built based on 7,680 ratings of V-PCC-encoded volumetric videos from 120 users by Cox et al.[6]. ACM Reference Format: Yuang Shi, Bennett Clement, and Wei Tsang Ooi. 2024. QV4: QoE-based Viewpoint-Aware V-PCC-encoded Volumetric Video Streaming. In ACM Multimedia Systems Conference 2024 (MMSys ’24), April 15–18, 2024, Bari, Recently, Rudolph et al. [30] propose a point cloud decomposition method called NoVA-PCC. They propose to utilize the 2D-projection mechanism of V-PCC for point cloud decomposition, taking the first step to enable user-adaptive optimization techniques on V- PCC-encoded volumetric videos. Nevertheless, V-PCC-encoded vol- umetric video streaming is still very challenging due to the lack This work is licensed under a Creative Commons Attribution International 4.0 License. MMSys ’24, April 15–18, 2024, Bari, Italy © 2024 Copyright held by the owner/author(s). ACM ISBN 979-8-4007-0412-3/24/04. https://doi.org/10.1145/3625468.3647619 This work is licensed under a Creative Commons Attribution International 4.0 License. MMSys ’24, April 15–18, 2024, Bari, Italy © 2024 Copyright held by the owner/author(s). ACM ISBN 979-8-4007-0412-3/24/04. https://doi.org/10.1145/3625468.3647619 144 MMSys ’24, April 15–18, 2024, Bari, Italy Yuang Shi, Bennett Clement, and Wei Tsang Ooi The rest of this paper is organized as follows. Section 2 presents the related work on volumetric video compression and streaming, and introduces the background of V-PCC. Section 3 introduces the proposed methods to achieve QoE-based viewpoint-aware volumet- ric video streaming encoded by V-PCC, including viewpoint-aware V-PCC tiling, 6DoF viewpoint prediction, QoE estimation, and QoE- based bitrate adaptation. We show the experiments and results in Section 4. Finally, we discuss and conclude this paper in Section 5. Occupancy Map Attribute Image Geometry Image Segmented Point Cloud Original Point Cloud Figure 1: Example of patch generation and packing of V-PCC. From left to right: The original point cloud. The segmented point cloud with different colors represents different patches. The occupancy map, attribute image, and geometry image are generated from these patches. 2.1 Point Cloud Compression Our QoE model can accurately predict the subjective quality of volumetric videos encoded by V- PCC with 0.06 root-mean-square error (RMSE). Additionally, to enable effective and efficient viewpoint adaptation for better user perception quality under diverse and dynamic user viewpoints in 6DoF, we develop a viewpoint prediction model for 6DoF viewpoint prediction based on real-world viewpoint data collected from 26 users by Subramanyam et al. [33] and emulated dynamic viewpoint data, which achieves 0.02 mean absolute error (MAE) and 0.07 MAE on average for normalized translational movement and rotational movement, respectively. We combine the proposed QoE model and viewpoint prediction model with the state-of-the-art bitrate alloca- tion algorithm to maximize the visual quality of volumetric video under dynamic network conditions and diverse user viewpoints. We perform extensive simulations under complex and diverse net- work profiles and viewpoint trajectories to verify the feasibility and performance of QV4 . The results show that QV4 achieves up to 14.67% improvement in SSIM and 7.39% improvement in VMAF. 2.1 Point Cloud Compression Segmented Point Cloud Original Point Cloud Original Point Cloud The exploration of point cloud compression has an extensive back- ground, originating from previous studies conducted by Devillers and Gandoin [8] as well as Peng and Kuo [22]. These early works focused on compressing the geometry of point clouds by employ- ing tree-based data structures for representation. Recently, MPEG defines two PCC standards: V-PCC for dynamic point clouds and G-PCC for scenes and objects. By adopting the representation of 2D projection structure, V-PCC projects the original 3D point cloud into the 2D space, and compresses them with 2D video compres- sion methods. 3D data structures (e.g., kd-tree [4], octree [21] or trisoup [11]) are used in representative algorithms including G-PCC and Google’s Draco to exploit 3D correlations. The field of neural point cloud compression has recently gained attention. For example, Quach et al. have contributed to this area with their series of works, employing neural networks to compress point cloud geometry [26]. They frame the decoding process as a binary classification problem. Another notable contribution comes from Yan et al., who propose a specialized autoencoder architecture for compressing point cloud geometry [36]. Que et al. introduce a deep-learning framework that utilizes voxel context to compress octree-structured data in both static and dynamic point cloud compression scenarios [27]. Sheng et al. present an approach using neural networks for point cloud at- tribute compression, incorporating second-order point convolution [31]. Furthermore, Quach et al. provide a comprehensive survey summarizing recent advancements in neural point cloud compres- sion [25]. While neural compression techniques show promise, the current research emphasizes static point clouds. Figure 1: Example of patch generation and packing of V-PCC. From left to right: The original point cloud. The segmented point cloud with different colors represents different patches. The occupancy map, attribute image, and geometry image are generated from these patches. of effective quality metrics, bitrate allocation, and accurate view- ing prediction. Because of the aforementioned limitations, most of the works of tiling-based volumetric video streaming resort to geometry-based compression and deep-learning-based compres- sion [13, 16, 18, 20, 38, 40], which achieve low compression ratios and fail to work under limited and fluctuating network conditions. In this paper, we present QV4 , a QoE-based streaming pipeline for Viewpoint-aware V-PCC-encoded Volumetric Video. To our best knowledge, QV4 is the first tile-based volumetric video streaming system that exploits user viewing adaptations on V-PCC-encoded content. 2.2 Video-based Point Cloud Compression V-PCC’s main idea is to project each frame of 3D point clouds into three kinds of 2D images. The generated 2D images can then be encoded using state-of-the-art video codecs (e.g. HEVC).i On the encoder side, patch generation first projects the 3D point cloud into 2D space with different projection angles. Specifically, normal estimation is applied to each point, and the points are as- signed to a projection plane based on its normal direction. By de- fault, the six faces of a bounding box are used as projection planes. By projecting the points to projection planes, the point cloud is segmented into patches. These patches are further refined to ensure connectedness in the projection. Figure 1 shows an example of a segmented point cloud with different colors representing different patches. Patch packing then sequentially places the patches into 2D video frames while maintaining the close position and orientation of each patch across frames. The objective of patch packing is to 145 QV4: QoE-based Viewpoint-Aware V-PCC-encoded Volumetric Video Streaming MMSys ’24, April 15–18, 2024, Bari, Italy Viewpoint-aware V-PCC Encoding V-PCC Encoded Tiles Bitrate Adaptation Network Condition Viewpoint Prediction QoE Prediction Bandwidth Estimation V-PCC Decoder & Render X, Y, Z, FoV Server Client Visible Part Invisible Part Figure 2: The pipeline of our proposed streaming system. On the server side, V-PCC tiling is performed to partition the point cloud into view-dependent representations of varying qualities. On the client side, different adaptation techniques described in §3.2, §3.3, §3.4, and §3.5 are leveraged to optimize the quality of the visible parts, resulting in higher visual quality of visible parts while lower quality of invisible parts. Viewpoint-aware V-PCC Encoding V-PCC Encoded Tiles Server Viewpoint Prediction Client Server Figure 2: The pipeline of our proposed streaming system. On the server side, V-PCC tiling is performed to partition the point cloud into view-dependent representations of varying qualities. On the client side, different adaptation techniques described in §3.2, §3.3, §3.4, and §3.5 are leveraged to optimize the quality of the visible parts, resulting in higher visual quality of visible parts while lower quality of invisible parts. point clouds. AITransfer [16] proposed a neural adaptive trans- mission scheme, but it does not address color attributes in point clouds. Transmission systems such as ViVo [13] and Groot [18] combined compression and optimization methods to achieve paral- lel acceleration based on visual characteristics. 2.2 Video-based Point Cloud Compression Nonetheless, these approaches adopt geometry-based compression techniques for tile- based transmission, thus suffering from limitations in compression ratios and bandwidth requirements. Both ViVo and Groot achieve a low compression ratio (less than 8), thus failing to work well under complex network conditions [37]. fit all the patches onto 2D frames which can be compressed effec- tively by existing video coding standards. The process consists of applying a packing method followed by a global patch allocation (GPA). V-PCC produces three artifacts after patch packing: the oc- cupancy map which contains information if a pixel in the resulting projection plane is occupied, the geometry image which contains information on the distance between points from the projection plane (i.e., depth), and the attribute image which contains informa- tion on attributes, e.g. colors, of the projected points. The example of these three artifacts is shown in Figure 1. Finally, the generated 2D video is encoded with a 2D video codec, such as the HEVC. As a result, V-PCC compression rates are controlled by the geometry and texture quantization parameter (QP) of the 2D video codec. In contrast, we capitalize on the utilization of V-PCC and viewing- aware optimization to tile-based volumetric video streaming system. These unique components enable our approach to achieve a sig- nificantly higher average compression ratio of 610, while keeping satisfactory quality even under limited network conditions and complex viewing behavior. Further details of our experiments and analyses are contained in Section 4. On the decoder side, the 2D video decoder decodes and gen- erates the occupancy map, geometry map, and attribute map, re- spectively. The geometry and attribute of the point cloud are then reconstructed. Post-processing (e.g., duplicate pruning, smoothing) is then applied to the reconstructed point cloud. Please refer to [11] for more details. 3.2 Viewpoint-Aware V-PCC Tiling In viewpoint-aware point cloud segmentation, a point cloud is expected to be divided into 𝑘groups corresponding to 𝑘view di- rections, denote D = {𝑑1,𝑑2, . . . ,𝑑𝑘} as the collection of predefined view directions. As the performance of V-PCC is highly dependent on the performance of 2D video codec, naïve point cloud segmenta- tion can destroy the spatial and temporal redundancy in projected 2D images, making view-dependent V-PCC challenging [41].i (b) Synthesized trajectories. (a) Real user trajectories from [33]. As discussed in Section 2.2, the V-PCC encoder first generates patches which are the 2D projections of 3D point clouds. The pro- jections are then packed and compressed by sophisticated 2D com- pression techniques. Rudolph et al. [30] first observe that during patch generation, the assignment of projections can be interpreted as the view directions. For example, in Figure 1, the points of the point cloud are projected into patches based on their directions of normals which also indicate the view directions. Therefore, we can fully utilize the assignments of 2D projections produced within V-PCC encoding for point cloud segmentation, thus incorporating the high compression efficiency brought by V-PCC.i Figure 3: Examples of viewpoint trajectories. The centroid of the point cloud object is located at (0,0,0). and bitrate allocation, as discussed in Section 3.4 and Section 3.5, respectively. One-hot encoding is applied to represent the tiles. Specifically, we denote the tiles at 𝑇+ 𝑝as t𝑇+𝑝∈{0, 1}𝑀where the 𝑖-th entry is 1 if and only if the 𝑖-th tile falls into the user view, and is 0 otherwise. Viewpoint Trajectory. To better understand how real users move when watching volumetric videos, we adopt real viewpoint trajectories from [33]. The viewpoint trajectories are collected from 26 participants. Participants wore an Oculus Rift HMD to watch volumetric videos rendered from 8i Dataset [9]. They were free to make translational movements (pitch and yaw), move backward or forward (Z dimension), leftward or rightward (X dimension), and upward or downward (Y dimension). Figure 3(a) shows an example of the real viewpoint trajectories. We characterize the viewing behaviors and make two interesting observations. First, the horizontal positions X and Z range from -0.13 m to 1.36 m, and 0.77 m to 2.75 m, respectively, indicating that the participants spent all the time looking toward the frontal body of the avatars in volumetric videos and never moved to the back. 3.2 Viewpoint-Aware V-PCC Tiling Second, the median and 90th percentile of vertical position Y are 0.02 m and 0.09 m, respectively, meaning that the participants hardly moved vertically. Based on the above observations, we can find that the participants spent most of their time in a certain area looking toward the frontal body of the avatar in volumetric videos, which is in agreement with previous studies [2, 29, 39]. g pfi y g y Specifically, the V-PCC encoder generates patches in the follow- ing steps. First, a normal per point is estimated according to Hoppe et al. [14]. A tangent plane and its normal are defined per point, based on the nearest neighbors within a predefined search distance. To make the directions more uniform, each point’s normal is en- forced to point in similar directions to its neighbors’ normals. Then, each point is associated with the plane that has the closest nor- mal. As a result, the point cloud is segmented into a set of patches. Several techniques are further used to refine the segmentation to en- sure the inter-smoothness and intra-connectedness of the patches. Instead of packing all the patches into one 2D image, we pack them into distinct tiles. Each tile corresponds to a direction of projection that indicates the view direction. As a result, for the pre-defined view directions 𝑑1,𝑑2, . . . ,𝑑𝑘, we will create 𝑘attribute, geometry, and occupancy videos. In the end, 𝑘bitstreams are produced by the V-PCC encoder. By default, 𝑘is set to 6 since six faces of a bounding box are used by V-PCC as projection planes. Because the aforementioned bitstreams are independent and non-overlapping, we can allocate different bitrates to point cloud segments based on the viewpoint of users. Therefore, to better evaluate the generalization ability of our method, we additionally emulate more dynamic user movement and generate synthesized viewpoint trajectories. In the synthesized trajectories, the viewers are kept moving around the object hori- zontally and vertically, as shown in Figure 3(b). The corresponding positional values for X and Z axes span from -2.92 m to 3.16 m and -3.08 m to 3.21 m, respectively, while the Y axis ranges from -3.08 m to 3.21 m, resulting in a wide spatial coverage. The standard deviations for X, Y, and Z coordinates amount to 1.58 m, 0.64 m, and 1.39 m, respectively, indicating notable dispersion tendencies within the dataset. 2.3 Volumetric Video Streaming QV4 is, to the best of our knowledge, the first tile-based streaming system for V-PCC-encoded volumetric video that considers user viewing experience and behaviors. It is designed based on dynamic adaptive streaming over HTTP (DASH) standards. Figure 2 gives an overview of QV4 . Several volumetric video streaming systems have been developed, demonstrating the feasibility of volumetric video transmission [12, 13, 16–19, 24, 33]. YuZu [40] pioneered the use of point cloud super-resolution techniques, enhancing the user experience with models like PU-GAN [20] and MPU [38]. However, YuZu’s experi- ments lacked a comparison with conventional point cloud codecs and were limited to small datasets. Subramanyam et al. [33] pro- posed a view-adaptive streaming approach that segments individ- ual point cloud objects into non-overlapping tiles. The tiles are transmitted with different bitrates based on the position and orien- tation of the users’ viewport. Li et al. [19] introduced an innovative method that incorporates human saliency to optimize 3D tiling schemes for dynamic point cloud streaming, improving efficiency. Shi et al. [32] observed that V-PCC suffers a great quality drop at low bit-rate. To address this problem, they proposed to exploit the redundancy of 3D point clouds by simple 3D sub-sampling and achieved considerable quality improvement of V-PCC encoded QV4 streams video-on-demand volumetric content stored on an Internet server to client hosts. On the server side, the volumetric videos are segmented into view-aware tiles which are indepen- dent and non-overlapping (§3.2). Then each tile is independently encoded with different quality levels by V-PCC. We include tile metadata for each tile in the Media Presentation Description (MPD) document. During playback, the client continuously predicts the user’s viewpoint (§3.3) and estimates the QoE of tiles (§3.4). The tiles with optimal quality are chosen based on the bitrate of tiles, predicted viewpoint, estimated QoE, network condition, and buffer status (§3.5). Then, the client sends requests to the server for the selected tiles. After receiving the encoded tiles from the server, the V-PCC decoder decodes the tiles and stitches them together to 146 Yuang Shi, Bennett Clement, and Wei Tsang Ooi MMSys ’24, April 15–18, 2024, Bari, Italy form a complete Group of Frames (GoF) of the point cloud. The buffer finally stores the GoF and sends the GoFs to the renderer sequentially for the user. (a) Real user trajectories from [33]. (b) Synthesized trajectories. Figure 3: Examples of viewpoint trajectories. 2.3 Volumetric Video Streaming The centroid of the point cloud object is located at (0,0,0). (b) Synthesized trajectories. (a) Real user trajectories from [33]. 3.2 Viewpoint-Aware V-PCC Tiling This augmented dataset provides a more realistic representation of dynamic user movement, enabling a more robust evaluation of the proposed method’s effectiveness in handling di- verse motion patterns. 3.3 6DoF Viewpoint Prediction 6DoF Viewpoint prediction is a critical component of real-time volumetric video streaming, as it facilitates a smooth and seamless experience, where the user’s movements are tracked in real-time and the appropriate content is prefetched, thus minimizing delay. In this section, we propose our model to achieve accurate viewpoint prediction.ii Problem Definition. For current time𝑇, our first objective is to predict the position of viewpoint at time 𝑇+ 𝑝based on the history viewpoints of frames from 𝑇−ℎto 𝑇, where 𝑝is the prediction window and ℎis the history window. Given the predicted view- point v𝑇+𝑝, we then infer the in-the-view tiles for QoE prediction 147 MMSys ’24, April 15–18, 2024, Bari, Italy QV4: QoE-based Viewpoint-Aware V-PCC-encoded Volumetric Video Streaming (a) The predictions of LR. (b) The predictions of MLP. (c) The predictions of SVR. (d) The predictions of GRU. Figure 4: Visualization of the predictions of four models and the ground truth when ℎ= 60 and 𝑝= 120. (d) The predictions of GRU. (b) The predictions of MLP. (c) The predictions of SVR. (a) The predictions of LR. (b) The predictions of MLP. (a) The predictions of LR. (c) The predictions of SVR. (d) The predictions of GRU. 4: Visualization of the predictions of four models and the ground truth when ℎ= 60 and 𝑝= 120. (a) Translational movement. (b) Rotational movement. Figure 5: CDF of MAE of translational and rotational move- ment prediction when ℎ= 60 and 𝑝= 120. (b) Rotational movement. (a) Translational movement. visualize the predicted trajectory of the viewpoint of these four models in Figure 4, which shows that GRU is capable of accurately predicting the viewpoint trajectory, compared to other models. After predicting the viewpoint, we then infer the tiles that fall into the view based on the predicted viewpoint. Three metrics are utilized to evaluate the performance of in-the-view tile prediction: exact-match accuracy, Hamming score, and weighted-average F1 score. We also measure the total inference time of viewpoint pre- diction and in-the-view tile prediction on a computer with Apple M1 Pro with 32 GB RAM running on macOS 13.2. (a) Translational movement. (b) Rotational movement. Figure 5: CDF of MAE of translational and rotational move- ment prediction when ℎ= 60 and 𝑝= 120. As shown in Table 1, GRU achieves the best performance with 0.6440 exact-match accuracy, 0.8686 Hamming score, and 0.8742 weighted-average F1 score. Table 1: Performance of in-the-view tile prediction. Table 1: Performance of in-the-view tile prediction. Model Exact-match Hamming Weighted-average Inference Accuracy ↑ Score ↑ F1 Score ↑ time (ms) ↓ LR 0.2243 0.6669 0.6698 0.0014 MLP 0.3900 0.7484 0.7545 0.0049 SVR 0.1587 0.6069 0.5834 0.4308 GRU 0.6440 0.8686 0.8742 0.0485 3.3 6DoF Viewpoint Prediction In terms of time efficiency, the inference of these four models can all be completed within 1 ms for five dimensions. We thus adopt GRU as our viewpoint prediction model based on the above results. Prediction Model. Volumetric videos involve multiple dimen- sions of movement, which can make it challenging to predict the user’s viewport in real-time. As suggested by Han et al. [13], predict- ing each dimension separately and then combining the results to derive the predicted viewport allows for more accurate predictions and better real-time performance, making it suitable for real-world applications. We thus consider four learning-based models to pre- dict each dimension of X, Y, Z, yaw, and pitch: Linear Regression (LR), Multi-Layer Perceptron (MLP), Support Vector Regression (SVR), and Gated Recurrent Unit RNN (GRU). Specifically, the MLP is configured with 2 hidden layers, with 64 neurons for the first hidden layer and 32 neurons for the second hidden layer. Hyper- bolic tangent is employed as the activation function and Adam is used for optimization. Default settings are applied to LR and SVR. As for the GRU model, 2 GRU layers are stacked together, with 64 neurons for each layer. Adam is adopted as the optimizer. 3.4 QoE Prediction Model We then explain the proposed QoE prediction model. The model considers two factors: (i) parameters associated with the encoder (i.e., V-PCC compression rate), and (ii) the viewpoint of the user. Evaluation. We consider several history windows from 30 frames to 60 frames and prediction windows from 30 frames to 120 frames at a step of 30 frames for an FPS of 30. The X, Y, Z, yaw, and pitch are normalized using mean normalization. Mean absolute error (MAE) is utilized to evaluate the performance of the models. Figure 5 plots the MAEs of translational (X, Y, Z) and rotational (yaw, pitch) movement prediction, with ℎ= 60 (the upper bound of inference time) and 𝑝= 120 (the lower bound of prediction accuracy). As shown, GRU outperforms the other models in both translational and rotational movement prediction, with the lowest MAE. To bet- ter illustrate the prediction performance of the models, we also We start by discussing the trivial version of our QoE prediction model, which only considers the effect of V-PCC compression rate. Given a tile 𝑡𝑗, we want to estimate its quality with respect to its representation level 𝑗, where the representation level is controlled by geometry QP and texture QP as mentioned in Section 2.2. To accurately predict the quality, we resort to supervised ma- chine learning algorithms. The volumetric video quality assessment dataset (VOLVQAD) [6] is adopted to train and evaluate our model. VOLVQAD consists of 376 video sequences and 7,680 ratings from 120 users. The volumetric video sequences are encoded with MPEG 148 Yuang Shi, Bennett Clement, and Wei Tsang Ooi MMSys ’24, April 15–18, 2024, Bari, Italy V-PCC using 4 different avatar models from the 8i Dataset [9] and 16 varying quality, where the quality is controlled by geometry QP and texture QP. The volumetric video sequences are then rendered into test videos for subjective quality assessment. The participants were asked to compare the QoE of the rendered videos by providing mean opinion scores (MOS), ranging from 1 to 5. Each participant viewed 64 pairs of videos. The order of the videos was chosen randomly for each participant.i x z 𝜃! 𝜃" 𝜃# x z Figure 6: An example of the relationship between viewpoint and tiles. 3.5 QoE-based Tile Rate Adaptation Three metrics are used to evaluate the ML models: RMSE, MAE, and 𝑅2. The prediction performance of the five models is reported in Table 2. We find that PR achieves the best performance with 0.06 RMSE, 0.20 MAE, and 0.52 𝑅2, compared to other methods. We thus adopt PR to predict the quality of tiles with respect to geometry QP and texture QP, given its effectiveness and efficiency. For a tile 𝑡𝑗, we denote its geometry QP as 𝑄𝑃𝑔 𝑗and texture QP as 𝑄𝑃𝑡 𝑗, the quality is then denoted as 𝑞(𝑄𝑃𝑔 𝑗,𝑄𝑃𝑡 𝑗). We denote the quality as 𝑞𝑗for simplification. Then, we predict its quality 𝑞𝑗with: Given the tiles of point clouds segmented based on the method described in Section 3.2, the user viewpoint predicted by the model discussed in Section 3.3, and the QoE of tiles estimated by the QoE model described in Section 3.4, the next step is allocating high bitrate to the portion of the point cloud contents that fall within the viewing frustum while transmitting the rest tiles with low bitrate. Our algorithm involves two steps. First, the client estimates the optimal bitrate budget of the next point cloud contents to download based on the network condition and buffer state. The estimated bitrate budget is then allocated to the tiles of the volumetric video, given the user viewpoint, the quality and bitrate of each represen- tation level of tiles, and the buffer states.i 𝑞𝑗= ®𝑋𝑇· ®𝛽+ 𝜀, (1) (1) where ®𝑋𝑇= 𝑄𝑃𝑔 𝑗,𝑄𝑃𝑡 𝑗, 𝑄𝑃𝑔 𝑗 2 ,𝑄𝑃𝑔 𝑗· 𝑄𝑃𝑡 𝑗, 𝑄𝑃𝑡 𝑗 2 , ®𝛽= [−0.002, 0.208, −0.005, 0.006, −0.007]𝑇, 𝜀= 2.29. (2) In the first step, we determine the ideal bitrate by using a state- of-the-art bitrate allocation algorithm QUETRA [35], which is an ef- fective yet simple rate adaptation algorithm based on an 𝑀/𝐷/1/𝐾 queuing model of DASH. Note that other rate adaptation algorithms, such as ELASTIC [5] or BBA [15], can also be used in our pipeline. (2) We then consider the effect of the viewpoint of users. Ideally, we want to attach more importance to the tiles facing the users’ viewpoint and attach less importance to the occluded tiles. To depict the position relationship between tiles and viewpoint, we calculate the cosine of the angle between the line of sight and the normal of tiles. Table 2: Comparison of the QoE models. Model RMSE ↓ MAE ↓ 𝑅2 ↑ PR 0.06 0.20 0.52 SVR 0.53 0.47 -2.07 RF 0.92 0.77 -3.48 MLP 0.40 0.34 -0.68 K-NN 0.49 0.41 0.00 g Therefore, the proposed QoE prediction model is: 𝑄(𝑡𝑗, 𝑣) = −cos(𝑡𝑗, 𝑣) · 𝑞𝑗. (3) (3) where 𝑞𝑗is the predicted quality of 𝑡𝑗obtained by Eq. 1. where 𝑞𝑗is the predicted quality of 𝑡𝑗obtained by Eq. 1. 3.4 QoE Prediction Model We consider five lightweight machine learning (ML) models to predict the quality: (i) Polynomial Regression (PR), (ii) Support Vector Regression (SVR), (iii) Random Forest (RF), (iv) Multi-Layer Perceptron (MLP), and (v) K-Nearest Neighbour (K-NN). During the training stage, we perform 5-fold cross-validation for hyper- parameter tuning by using the grid search algorithm where the Coefficient of Determination (𝑅2) is used as an objective criterion. The following hyperparameters are selected after hyperparameter tuning: (i) PR has two degrees, (ii) SVR has linear kernel; (iii) RF has 100 estimators with 5 maximum depth; (iv) MLP has 0.001 learning rate with a hidden layer with 100 neurons; (v) KNN has two neigh- bours with uniform weights. Other hyperparameters are kept in default. Figure 6: An example of the relationship between viewpoint and tiles. tile 𝑡𝑗, their cos(𝑡𝑗, 𝑣) decreases. Figure 6 gives an example of the viewpoint-tile relationship, where the point cloud is divided into six tiles. In this example, let’s denote the angles between the line of sight and the normals of frontal, right, and back tile as 𝜃1, 𝜃2, and 𝜃3, respectively. We can find that cos𝜃1 is the smallest compared to others because the frontal tile is closer to the viewpoint while the right and back tiles are almost occluded. 3.5 QoE-based Tile Rate Adaptation We denote the cosine of the angle between the line of sight and the normal of tile 𝑡𝑗as cos(𝑡𝑗, 𝑣) ∈[−1, 1], where 𝑣is the position of the viewpoint. With the viewpoint getting closer to the front of the We then allocate the bitrate budget estimated from QUETRA to the tiles. The tile rate allocation problem can be modeled as a Multi- class Knapsack problem, which is a special case of 0-1 Knapsack Problem with a disjoint multi-class constraint. A frame is segmented into 𝑀tile groups, each group has 𝑁representation levels. Let 𝑡𝑖,𝑗 be the tile of group 𝑖and representation level 𝑗, denote its estimated bitrate of representation level as 𝑏𝑖,𝑗and its rendered quality as 149 MMSys ’24, April 15–18, 2024, Bari, Italy QV4: QoE-based Viewpoint-Aware V-PCC-encoded Volumetric Video Streaming 𝑄(𝑡𝑖,𝑗, 𝑣), where 𝑣is the position of user viewpoint. The rendered quality 𝑄(𝑡𝑖,𝑗, 𝑣) is obtained by Eq. 3, and the viewpoint position 𝑣 is predicted by our viewpoint prediction model described in Section 3.3.f 26 Mbps, varying every second. The throughput variation does not follow any regular pattern, making the rate adaptation challenging. Figure 7 shows their throughput variations over 60 seconds. Figure 7: Visualization of P1, P2, P3, and P4. We wish to choose tiles from different representation levels to maximize the expected rendered quality, while the total rate is kept within the given bitrate budget 𝐵estimated from QUETRA. Let X = 𝑥𝑖,𝑗: 1 ≤𝑖≤𝑀, 1 ≤𝑗≤𝑁 , where 𝑥𝑖,𝑗is 1 if and only if tile 𝑡𝑖,𝑗is selected to be downloaded for rendering, and is 0 otherwise. We can formulate the problem as:︁︁ argmax X 𝑀 ∑︁ 𝑖=1 𝑁 ∑︁ 𝑗=1 𝑄(𝑡𝑖,𝑗, 𝑣)𝑥𝑖,𝑗, subject to 𝑀 ∑︁ 𝑖=1 𝑁 ∑︁ 𝑗=1 𝑏𝑖,𝑗𝑥𝑖,𝑗≤𝐵, 𝑁 ∑︁ 𝑗=1 𝑥𝑖,𝑗= 1, 𝑖= 1, . . . , 𝑀, 𝑥𝑖,𝑗∈{0, 1}, 𝑖= 1, . . . , 𝑀, 𝑗= 1, . . . , 𝑁. (4) Figure 7: Visualization of P1, P2, P3, and P4. (4) Viewpoint Trajectory. We use two viewpoint trajectories (V1 and V2) for evaluation. As discussed in Section 3.3, V1 is the set of real viewpoint trajectories collected from 26 participants [33]. V2 is the synthesized trajectory generated from the emulation of more dynamic user movement. ︁ 𝑥𝑖,𝑗∈{0, 1}, 𝑖= 1, . . . , 𝑀, 𝑗= 1, . . . , 𝑁. 3.5 QoE-based Tile Rate Adaptation To address the multi-class knapsack problem, we adopted the solution based on Pisinger’s greedy heuristic [23]. This heuristic begins by sorting representation levels of each tile by their bitrate and selecting the lightest one from each tile to cover all tiles. Next, it sorts representations in each tile group based on their quality-to- bitrate slope 𝑄(𝑡𝑖,𝑗, 𝑣) −𝑄(𝑡𝑖,𝑗−1, 𝑣) /𝑏𝑖,𝑗−𝑏𝑖,𝑗−1 , which quan- tifies the improvement in quality for selecting a higher representa- tion level from a sorted list. The algorithm then replaces previously- selected levels with the next level from the sorted sequence within the same tile until the total bitrate reaches the budget bitrate. y Encoder Settings. V-PCC reference software version 15 2 is used as our point cloud codec. We use the “lossy geometry lossy attribute random access" configurations defined in the V-PCC com- mon test condition (CTC) [10]. We divide the point cloud into six tiles with the decomposition method described in Section 3.2. V- PCC compression rates are controlled by the geometry and texture quantization parameter (QP). There are five compression rates de- fined in the CTC, labeled as R5 to R1. R5 has the highest quality (lowest compression) and R1 has the lowest quality (highest com- pression). We also define two additional rate points with higher distortion levels than that of R1. We denote those two additional levels as R0 and R-1. The geometry quantization parameter is set to 36 and 40 respectively; The attribute quantization parameter is set to 47 and 52 respectively. Therefore, a total of seven representation levels are given for each tile. Table 4 summarizes the CTC encoder settings as well as our additional setting. 4.1 Experiment Setup Volumetric Video Dataset. We use the four dynamic point cloud sequences from the 8i Dataset [9]: RedAndBlack, Loot, Soldier, and LongDress, for our evaluation. Each sequence has 300 frames with a frame rate of 30 fps. The average number of points per frame and corresponding bitrates (in Gbps) of the uncompressed volumetric videos are summarized in Table 3. Table 4: Settings for MPEG V-PCC Reference Encoder Table 4: Settings for MPEG V-PCC Reference Encoder Rate Level Occupancy Q Factor Q Factor Resolution (Geometry Map) (Attribute Map) 5 2 16 22 4 4 20 27 3 4 24 32 2 4 28 37 1 4 32 42 0 4 36 47 -1 4 40 52 2https://github.com/MPEGGroup/mpeg-pcc-tmc2/releases/tag/release-v15.0 Table 3: Dynamic 3D Point Cloud Dataset RedAndBlack Loot Soldier LongDress Points (×106) 0.7 0.8 1.5 0.8 Bitrate (Gbps) 3.6 3.9 5.5 4.3 Network Profile. Four network profiles (P1, P2, P3, and P4) are used for evaluation. We follow the DASH Industry Forum Guide- lines [7] to generate Profile P1 and P2. Specifically, P1 follows a high-low-high pattern and P2 follows a low-high-low pattern. The throughput of P1 and P2 varies regularly at the interval of 3 seconds in five levels with rate {3, 4, 6, 8, 10} Mbps. Profile P3 and P4 are real 4G/LTE network traces collected from a moving bus and car [28], respectively. The throughput of P3 and P4 ranges from 0 to Decoding and Rendering Setting. To enable real-time V-PCC decoding, we reconstructed the V-PCC reference decoder in Rust with various optimizations, achieving an average of 15 times faster 150 MMSys ’24, April 15–18, 2024, Bari, Italy Yuang Shi, Bennett Clement, and Wei Tsang Ooi (a) RedAndBlack. (b) Loot. (c) Soldier. (d) LongDress. Figure 8: SSIM. Each point represents the average quality per second with the corresponding bitrate. (a) RedAndBlack. (b) Loot. (c) Soldier. (d) LongDress. Figure 9: VMAF. Each point represents the average quality per second with the corresponding bitrate. (a) RedAndBlack. (b) Loot. (c) Soldier. (d) LongDress. Figure 8: SSIM. Each point represents the average quality per second with the corresponding bitrate. (c) Soldier. (d) LongDress. (a) RedAndBlack. (b) Loot. (b) Loot. gure 8: SSIM. Each point represents the average quality per second with the corresponding bitrate. (a) RedAndBlack. (b) Loot. (c) Soldier. (d) LongDress. Figure 9: VMAF. Each point represents the average quality per second with the corresponding bitrate. (b) Loot. (c) Soldier. (a) RedAndBlack. (d) LongDress. (d) LongDress. Figure 9: VMAF. Each point represents the average quality per second with the corresponding bitrate. (Base). Instead, it adopts QUETRA [35] for bitrate estimation, which is the same as ours when estimating the bitrate budget.i than the reference decoder. The code is made public to support further research 3. We also work with a toolkit called VVTk 4, which provides an extendable video player for real-time dynamic point cloud rendering. The streaming simulation and rendering in our experiments are all conducted online and in real-time through our V- PCC decoder and VVTk, running on a computer with Apple M1 Pro with 32 GB RAM running on macOS 13.2. 3https://github.com/benclmnt/tmc2-rs 4https://github.com/nus-vv-streams/vvtk Table 3: Dynamic 3D Point Cloud Dataset The width and height of the rendered videos are set to 900 and 1600, respectively. The point size is set to 1, and the background color of all the rendered videos is set to black. The playback was looped in 60 seconds for each point cloud sequence, network profiles, and viewpoint trajectories, resulting in 32 combinations. Visual Quality. We first show the quality of the four point cloud sequences versus the encoded bitrate in Figure 8 and Figure 9. The figures plot the results under all the viewpoint trajectories and network profiles. Furthermore, curves are fitted based on the data points for better comparison. As observed, QV4 achieves bet- ter quality at the same bitrate compared with V-PCC (Base). Our method outperforms V-PCC (Base) by at most 24.3% in SSIM and 28.6 in VMAF on RedAndBlack; 36.9% in SSIM and 24.4 in VMAF on Loot; 27.2% in SSIM and 30.3 in VMAF on Soldier; and 33.6% in SSIM and 31.9 in VMAF on LongDress. Evaluation Metrics. For evaluation, PCC Arena [34] is adopted to compute the 2D quality of rendered volumetric videos. Two metrics are used to quantify the quality of rendered volumetric videos: VMAF [1] and SSIM. To avoid the disturbance from the background which is irrelevant to the quality of 3D point clouds, we generate a mask map to exclude the background pixels of each rendered 2D image when using SSIM to evaluate the quality of rendered point cloud frames. Table 5: Overall quality of our proposed streaming method and V-PCC (Base) on dynamic viewpoint trajectories (V1 and V2) and network conditions (P1, P2, P3, and P4). Network V1 V2 Profile Method SSIM ↑ VMAF ↑ SSIM ↑ VMAF ↑ P1 V-PCC (Base) 0.73 69.93 0.74 59.02 Proposed 0.85 75.10 0.77 64.47 P2 V-PCC (Base) 0.72 68.64 0.73 64.70 Proposed 0.80 71.90 0.75 66.61 P3 V-PCC (Base) 0.75 71.02 0.75 66.03 Proposed 0.86 75.30 0.78 70.11 P4 V-PCC (Base) 0.71 67.85 0.72 58.43 Proposed 0.79 72.32 0.76 61.32 4.2 Experimental Resultsf To demonstrate the effectiveness of our QoE model and viewing adaptations, we first evaluate the performance of QV4 and compare it with the baseline streaming pipeline denoted as V-PCC (Base). V-PCC (Base) is selected for encoding and decoding the whole point cloud by V-PCC without tiling. The proposed QoE-based tile rate allocation and viewpoint prediction are also disabled for V-PCC We also list the average SSIM and VMAF for the proposed method and V-PCC (Base) over four dynamic point cloud sequences in Table 151 QV4: QoE-based Viewpoint-Aware V-PCC-encoded Volumetric Video Streaming MMSys ’24, April 15–18, 2024, Bari, Italy (a) Bitrate of in-the-view tiles. (b) Bitrate of out-of-the-view tiles. Figure 10: The distributions of bitrate of tiles in the view (left) and out of the view (right) on viewpoint trajectories (V1 and V2) and network conditions (P1, P2, P3, and P4). The quartiles are shown for better comparison. (a) Bitrate of in-the-view tiles. (a) Bitrate of in-the-view tiles. (b) Bitrate of out-of-the-view tiles. Figure 10: The distributions of bitrate of tiles in the view (left) and out of the view (right) on view and network conditions (P1, P2, P3, and P4). The quartiles are shown for better comparison. Figure 10: The distributions of bitrate of tiles in the view (left) and out of the view (right) on viewpoint trajectories (V1 and V2) and network conditions (P1, P2, P3, and P4). The quartiles are shown for better comparison. Table 6: Performance of the proposed method with and with- out viewpoint prediction. V1 V2 Method SSIM ↑ VMAF ↑ SSIM ↑ VMAF ↑ QV4 (NVP) 0.79 71.88 0.71 59.92 Proposed 0.83 73.66 0.77 65.63 (a) Proposed. (b) QV4 (NVP). Figure 11: Sample rendered point clouds of LongDress using the proposed method with and without viewpoint prediction. Table 6: Performance of the proposed method with and with- out viewpoint prediction. 5. We can make several observations from the results. First, we can observe that our proposed method consistently achieves the best quality with up to 14.67% improvement in SSIM and 7.39% improvement in VMAF on average, compared with V-PCC (Base). Besides, we can find that both QV4 and V-PCC (Base) achieve better performance on V1 than V2. As discussed in Section 3.3, users in V1 spent most of the time in a certain area looking at the frontal body, while the movement in V2 was more dynamic and complex. 4.2 Experimental Resultsf Under a dynamic viewpoint trajectory, the viewpoint predictions are less accurate. Wrong-predicted viewpoints lead to wrong bitrate allocation, thus degrading the video quality. Our method still achieves 3% improvement in SSIM and 3.58 in VMAF on average over four network profiles, compared with the baseline. The above results demonstrate the superiority of the proposed method compared with the method without our proposed viewing optimizations in terms of 2D visual quality. To sum up, the overall quality reported in Table 5, and the results of quality versus encoded bitrate shown in Figure 8 and Figure 9 indicate that our method can effectively and robustly improve the visual quality of volumetric videos, under diverse network conditions and complex viewing behaviors.fi (a) Proposed. Figure 11: Sample rendered point clouds of LongDress using the proposed method with and without viewpoint prediction. Bandwidth Efficiency. From Figure 8 and Figure 9, we can observe that when comparing the bit-rate of QV4 and V-PCC (Base) at the same quality, our method generally consumes much less bandwidth compared with V-PCC (Base) at the same quality in terms of SSIM and VMAF. Taking LongDress as an example, QV4 saves up to 49.6% and 41.1% bitrate at the same quality in terms of SSIM and VMAF, respectively, compared to V-PCC (Base). In summary, the results plotted in Figure 8 and Figure 9 demonstrate the bandwidth efficiency of our method compared to V-PCC (Base).f A main observation is that QV4 consistently assigns higher quality to the tiles within the view and decreases the bitrate of the tiles out of the view, under all four network profiles and two viewpoint trajectories. Overall, QV4 allocates representation levels with 45.5% bigger bitrate to the visible tiles and 62.5% smaller bitrate to the tiles that are not visible, compared to QV4 (NVA). By adapting to the viewpoint of users, QV4 decreases the bitrate of invisible tiles and increases the bitrate of in-the-view tiles, thus achieving high-quality volumetric video streaming under limited and complex network conditions.f fi Effect of Viewpoint Adaptation. The main reason that QV4 can save considerable bandwidth while keeping high visual quality is that QV4 considers users’ viewing behaviors and assigns higher quality to the tiles in the view and lower quality to the tiles out of the view. 5 DISCUSSION AND CONCLUSION We present QV4 , a volumetric video streaming pipeline empowered by QoE-based viewpoint-aware optimizations. With the proposed QoE model and viewpoint-aware optimizations, QV4 allows to adap- tively choose the optimal tiles under complex user viewing behav- iors and network conditions, thus maximizing the user’s viewing experience. Extensive simulations are conducted. The experiments show that our pipeline can improve the volumetric video quality by up to 14.67% in SSIM and 7.39% in VMAF over dynamic viewpoint trajectories and limited and fluctuating network conditions. jl g There are several limitations to our work that we can improve upon. First, the point cloud is partitioned into non-overlapping tiles based on the predefined view directions. The tiles are then encoded independently. However, the main issue incurred by segmentation is that smaller tiles contain less inter-spatial redundancy, making the overall compression efficiency worse than that of unsegmented video. Especially, V-PCC is sensitive to the density and complexity of the volumetric videos. In this paper, we only validate the feasi- bility and effectiveness of six tiles, which is the default setting of V-PCC. Therefore, exploring the trade-off between the segmenta- tion size and overhead from V-PCC remains a potential research direction. Secondly, visual saliency serves as an important feature in tile rate allocation. By allocating higher bitrate to the tiles with higher visual saliency and lower bitrate to the unattended tiles, we can significantly reduce the utilized bandwidth and client-side decoding overhead, and improve the users’ viewing experience. We hope to propose a saliency-detection model to segment the tiles based on their visual importance, thus allowing saliency-aware tile rate allocation. Thirdly, while the results of visual quality through SSIM and VMAF provide valuable insights into the performance of our proposed method, user studies will provide a more com- prehensive understanding of the effectiveness of QV4 in terms of the viewing experience. Therefore, future work should involve subjective evaluations to further assess of our method. f QoE. Moreover, we calculate the QoE value of each frame for the proposed and baseline streaming method. Figure 12 presents the distribution of QoE of four dynamic point clouds. The mean values are also plotted for better comparison.l The results indicate that when the bandwidth is limited and fluc- tuating, QV4 always provides considerably better QoE on average. 4.2 Experimental Resultsf As 152 Yuang Shi, Bennett Clement, and Wei Tsang Ooi MMSys ’24, April 15–18, 2024, Bari, Italy Figure 12: Average QoE of four dynamic point clouds over different viewpoint trajectories (V1 and V2) and network fil ( d ) h l d d LongDress RedAndBlack Loot Soldier 0 10 20 30 40 50 60 Overall Frame Rate (fps) Figure 13: Overall frame rate of QV4 . LongDress RedAndBlack Loot Soldier 0 10 20 30 40 50 60 Overall Frame Rate (fps) Figure 13: Overall frame rate of QV4 . Figure 12: Average QoE of four dynamic point clouds over different viewpoint trajectories (V1 and V2) and network profiles (P1, P2, P3, and P4). The mean values are indicated by white dots. Figure 13: Overall frame rate of QV4 . Figure 12: Average QoE of four dynamic point clouds over different viewpoint trajectories (V1 and V2) and network profiles (P1, P2, P3, and P4). The mean values are indicated by white dots. volumetric video streaming [18, 30, 32]. Notably, the latency anal- ysis excludes the inference time of viewpoint prediction and QoE prediction, as their durations are negligible (within 0.01 ms). shown, compared with the proposed method, the overall quality of the volumetric videos produced from QV4 (NVP) degrades from 4% to 6% in SSIM and from 1.78 to 5.71 in VMAF. To visualize the dif- ference, Figure 11 provides an example of the rendered point clouds with and without viewpoint prediction. One could find that QV4 (NVP) suffers great quality degradation, compared to the proposed method. By predicting the user viewpoint, QV4 can prefetch the 3D contents with higher quality and reduce the delay. In contrast, QV4 (NVP) cannot detect the user behavior, thus failing to fetch the accurate 3D contents in advance. As a result, QV4 (NVP) has to lower the quality of point clouds for low-latency volumetric video streaming. In conclusion, Table 6 and Figure 11 validate the effectiveness of the proposed viewpoint prediction. 5 DISCUSSION AND CONCLUSION Specifically, compared to V-PCC (Base), QV4 improves the QoE of LongDress, RedAndBlack, Loot, and Soldier by 45.45%, 29.65%, 12.98%, and 14.70% on average, respectively. Besides, the MOS of 52.34% of point cloud frames streamed by QV4 are predicted as 4 (good) or 5 (excellent) whereas only 26.92% of frames’ MOS streamed by V-PCC (Base) are predicted as good or excellent. The improvement on QoE is because the viewpoint-aware optimiza- tions of QV4 reduce the bitrate of the tiles outside the viewport and assign a higher quality to the tiles that are more visually important. System Performance. We also conduct a comprehensive per- formance evaluation of QV4 under varying viewpoint trajectories (V1 and V2) and network conditions (P1, P2, P3, and P4). Figure 13 shows the overall frame update rate, measured considering decod- ing and rendering time. Remarkably, our findings demonstrate that QV4 consistently achieves a frame rate exceeding 30 frames per second (fps), reaching up to 60 fps, across all four dynamic point clouds. This outstanding performance is maintained in the face of complex and dynamic network conditions and viewing trajectories. Furthermore, we assessed the average per-frame decoding and ren- dering latency of QV4 , yielding 16.87 ms and 8.11 ms, respectively. The result reveals that QV4 effectively keeps decoding latency be- low 30 ms, making V-PCC decoding no longer the bottleneck for 4.2 Experimental Resultsf To better illustrate the effect of viewpoint adaptation of QV4 , we present Figure 10 which depicts the bitrate distributions of in-the-view and out-of-the-view tiles for all video frames over two viewpoint trajectories and four network profiles. For comparison, we disable the QoE-based tile rate adaptation of QV4 as described in Section 3.5 so that each tile is assigned the same bitrate, and denote it as QV4 (NVA), i.e. No Viewpoint Adaptation. Effect of Viewpoint Adaptation. The main reason that QV4 can save considerable bandwidth while keeping high visual quality is that QV4 considers users’ viewing behaviors and assigns higher quality to the tiles in the view and lower quality to the tiles out of the view. To better illustrate the effect of viewpoint adaptation of QV4 , we present Figure 10 which depicts the bitrate distributions of in-the-view and out-of-the-view tiles for all video frames over two viewpoint trajectories and four network profiles. For comparison, we disable the QoE-based tile rate adaptation of QV4 as described in Section 3.5 so that each tile is assigned the same bitrate, and denote it as QV4 (NVA), i.e. No Viewpoint Adaptation. Effect of Viewpoint Adaptation. The main reason that QV4 can save considerable bandwidth while keeping high visual quality is that QV4 considers users’ viewing behaviors and assigns higher quality to the tiles in the view and lower quality to the tiles out of the view. To better illustrate the effect of viewpoint adaptation of QV4 , we present Figure 10 which depicts the bitrate distributions of in-the-view and out-of-the-view tiles for all video frames over two viewpoint trajectories and four network profiles. For comparison, we disable the QoE-based tile rate adaptation of QV4 as described in Section 3.5 so that each tile is assigned the same bitrate, and denote it as QV4 (NVA), i.e. No Viewpoint Adaptation. Effect of Viewpoint Prediction. We further conduct compar- ison experiments to demonstrate the feasibility and effectiveness of the proposed viewpoint prediction. We remove the viewpoint prediction model from QV4 and denote it as QV4 (NVP), i.e. No Viewpoint Prediction. We conducted experiments under four net- work profiles and two viewpoint trajectories. The average SSIM and VMAF over four network profiles are shown in Table 6. REFERENCES [21] Donald Meagher. Geometric modeling using octree encoding. Computer Graphics and Image Processing, 19(2):129–147, Jun 1982. [1] Anne Aaron, Zhi Li, Megha Manohara, Joe Yuchieh Lin, Eddy Chi-Hao Wu, and C.-C. Jay Kuo. Challenges in cloud based ingest and encoding for high quality streaming media. In Proceedings of the 2015 IEEE International Conference on Image Processing, ICIP 2015, Quebec City, QC, Canada, September 27-30, 2015, pages 1732–1736. IEEE, 2015. g g ( ) [22] Jingliang Peng and Chung J. Kuo. Octree-based progressive geometry encoder. In Internet Multimedia Management Systems IV, volume 5242, pages 301–311. SPIE, Nov 2003. 23] David Pisinger. A minimal algorithm for the multiple-choice knap [23] David Pisinger. A minimal algorithm for the multiple-choice knapsack prob European Journal of Operational Research, 83(2):394–410, Jan 1995. European Journal of Operational Research, 83(2):394–410, Jan 1995. [2] Evangelos Alexiou and Touradj Ebrahimi. Exploiting user interactivity in quality assessment of point cloud imaging. In Proceedings of the 11th International Conference on Quality of Multimedia Experience QoMEX 2019, Berlin, Germany, June 5-7, 2019, pages 1–6. IEEE, 2019. [24] Feng Qian, Bo Han, Jarrell Pair, and Vijay Gopalakrishnan. Toward practical volumetric video streaming on commodity smartphones. In Proceedings of the 20th International Workshop on Mobile Computing Systems and Applications, HotMobile 2019, Santa Cruz, CA, USA, February 27-28, 2019, pages 135–140. ACM, 2019. p p g y pp 2019, Santa Cruz, CA, USA, February 27-28, 2019, pages 135–140. ACM, [3] Evangelos Alexiou, Irene Viola, Tomás M Borges, Tiago A Fonseca, Ricardo L De Queiroz, and Touradj Ebrahimi. A comprehensive study of the rate-distortion performance in mpeg point cloud compression. APSIPA Transactions on Signal and Information Processing, 8:e27, Nov 2019. [25] Maurice Quach, Jiahao Pang, Dong Tian, Giuseppe Valenzise, and Frederic Dufaux. Survey on deep learning-based point cloud compression. Frontiers in Signal Processing, 2, Feb 2022. [26] Maurice Quach, Giuseppe Valenzise, and Frédéric Dufaux. Learning convolutional transforms for lossy point cloud geometry compression. In Proceedings of the 2019 IEEE International Conference on Image Processing, ICIP 2019, Taipei, Taiwan, September 22-25, 2019, pages 4320–4324. IEEE, 2019. f g [4] Jon Louis Bentley. Multidimensional binary search trees used for associative searching. Commun. ACM, 18(9):509–517, Sept 1975. [5] Luca De Cicco, Vito Caldaralo, Vittorio Palmisano, and Saverio Mascolo. ELASTIC: A client-side controller for dynamic adaptive streaming over HTTP (DASH). In Proceedings of the 20th International Packet Video Workshop, PV 2013, San Jose, CA, USA, December 12-13, 2013, pages 1–8. REFERENCES An overview of ongoing point cloud compression standardization activities: Video-based (V-PCC) and geometry-based (G-PCC). APSIPA Transactions on Signal and Information Processing, 9:e13, Apr 2020. [32] Yuang Shi, Pranav Venkatram, Yifan Ding, and Wei Tsang Ooi. Enabling low bit- rate MPEG V-PCC-encoded volumetric video streaming with 3D sub-sampling. In Proceedings of the 14th Conference on ACM Multimedia Systems, MMSys 2023, Vancouver, BC, Canada, June 7-10, 2023, pages 108–118. ACM, 2023. APSIPA Transactions on Signal and Information Processing, 9:e13, Apr [12] Serhan Gül, Dimitri Podborski, Thomas Buchholz, Thomas Schierl, and Cornelius Hellge. Low-latency cloud-based volumetric video streaming using head motion prediction. In Proceedings of the 30th ACM Workshop on Network and Operating Systems Support for Digital Audio and Video, NOSSDAV 2020, Istanbul, Turkey, June 10-11, 2020, pages 27–33. ACM, 2020. [33] Shishir Subramanyam, Irene Viola, Alan Hanjalic, and Pablo César. User centered adaptive streaming of dynamic point clouds with low complexity tiling. In Proceedings of the 28th ACM International Conference on Multimedia, MM 2020, Virtual Event / Seattle, WA, USA, October 12-16, 2020, pages 3669–3677. ACM, 2020. p g [13] Bo Han, Yu Liu, and Feng Qian. ViVo: visibility-aware mobile volumetric video streaming. In Proceedings of the 26th Annual International Conference on Mobile Computing and Networking, MobiCom 2020, London, United Kingdom, September 21-25, 2020, pages 11:1–11:13. ACM, 2020. [34] Cheng-Hao Wu, Chih-Fan Hsu, Tzu-Kuan Hung, Carsten Griwodz, Wei Tsang Ooi, and Cheng-Hsin Hsu. Quantitative comparison of point cloud compression algorithms with PCC arena. IEEE Transactions on Multimedia, 25:3073–3088, Feb 2023.i p g [14] Hugues Hoppe, Tony DeRose, Tom Duchamp, John McDonald, and Werner Stuetzle. Surface reconstruction from unorganized points. In Proceedings of the 19th Annual Conference on Computer Graphics and Interactive Techniques, SIGGRAPH 1992, Chicago, IL, USA, July 27-31, 1992, pages 71–78. ACM, 1992. [35] Praveen Kumar Yadav, Arash Shafiei, and Wei Tsang Ooi. QUETRA: A queuing theory approach to DASH rate adaptation. In Proceedings of the 2017 ACM on Multimedia Conference, MM 2017, Mountain View, CA, USA, October 23-27, 2017, pages 1130–1138. ACM, 2017. [15] Te-Yuan Huang, Ramesh Johari, Nick McKeown, Matthew Trunnell, and Mark Watson. A buffer-based approach to rate adaptation: evidence from a large video streaming service. In Proceedings of the 2014 ACM conference on SIGCOMM, SIGCOMM 2014, Chicago, IL, USA, August 17-22, 2014, pages 187–198. ACM, 2014. p g [36] Wei Yan, Shan Liu, Thomas H Li, Zhu Li, Ge Li, et al. REFERENCES IEEE, 2013. p p g [27] Zizheng Que, Guo Lu, and Dong Xu. VoxelContext-Net: An Octree based frame- work for point cloud compression. In Proceedings of the 2021 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2021, virtual, June 19-25, 2021, pages 6042–6051. Computer Vision Foundation / IEEE, 2021. [6] Samuel Rhys Cox, May Lim, and Wei Tsang Ooi. VOLVQAD: an MPEG V-PCC volumetric video quality assessment dataset. In Proceedings of the 14th Conference on ACM Multimedia Systems, MMSys 2023, Vancouver, BC, Canada, June 7-10, 2023, pages 357–362. ACM, 2023. [28] Darijo Raca, Jason J. Quinlan, Ahmed H. Zahran, and Cormac J. Sreenan. Beyond throughput: a 4g LTE dataset with channel and context metrics. In Proceedings of the 9th ACM Multimedia Systems Conference, MMSys 2018, Amsterdam, The Netherlands, June 12-15, 2018, pages 460–465. ACM, 2018. p g [7] DASH Industry Forum. Guidelines for implementation: DASH-AVC/264 test cases and vectors. http://dashif.org/guidelines/, 2014. Netherlands, June 12-15, 2018, pages 460–465. ACM, 2018. [29] Silvia Rossi, Irene Viola, and Pablo César. Behavioural analysis in a 6-DoF VR system: Influence of content, quality and user disposition. In Proceedings of the 1st Workshop on Interactive eXtended Reality, IXR@MM 2022, Lisbon, Portugal, 14 October 2022, pages 3–10. ACM, 2022. [8] Olivier Devillers and Pierre-Marie Gandoin. Geometric compression for interac- tive transmission. In Proceedings of the 11th IEEE Visualization Conference, IEEE Vis 2000, Salt Lake City, UT, USA, October 8-13, 2000, Proceedings, pages 319–326. IEEE Computer Society and ACM, 2000. [30] Michael Rudolph and Amr Rizk. View-adaptive streaming of point cloud scenes through combined decomposition and video-based coding. In Proceedings of the 1st International Workshop on Advances in Point Cloud Compression, Processing and Analysis, APCCPA 2022, October 28 – November 1, Lisbon, Portugal, APCCPA ’22, page 41–49, New York, NY, USA, 2022. Association for Computing Machinery. p y [9] Eugene d’Eon, Bob Harrison, Taos Myers, and Philip A Chou. 8i voxelized full bodies-a voxelized point cloud dataset. ISO/IEC JTC1/SC29 Joint WG11/WG1 (MPEG/JPEG) input document WG11M40059/WG1M74006, 7(8):11, 2017. [10] Eugene d’Eon, Bob Harrison, Taos Myers, and Philip A Chou. Common test conditions for point cloud compression. ISO/IEC JTC1/SC29/WG11 w17766, 2018. [31] Xihua Sheng, Li Li, Dong Liu, Zhiwei Xiong, Zhu Li, and Feng Wu. Deep-PCAC: An end-to-end deep lossy compression framework for point cloud attributes. IEEE Transactions on Multimedia, 24:2617–2632, Jun 2022. [11] Danillo Graziosi, Ohji Nakagami, Satoru Kuma, Alexandre Zaghetto, Teruhiko Suzuki, and Ali Tabatabai. REFERENCES Deep autoencoder-based lossy geometry compression for point clouds. arXiv preprint arXiv:1905.03691, 2019. g g p g [16] Yakun Huang, Yuanwei Zhu, Xiuquan Qiao, Zhijie Tan, and Boyuan Bai. AITrans- fer: Progressive AI-powered transmission for real-time point cloud video stream- ing. In Proceedings of the 29th ACM Multimedia Conference, MM 2021, Virtual Event, China, October 20 - 24, 2021, pages 3989–3997. ACM, 2021. [37] Mengyu Yang, Zhenxiao Luo, Miao Hu, Min Chen, and Di Wu. A compara- tive measurement study of point cloud-based volumetric video codecs. IEEE Transactions on Broadcasting, 69(3):715–726, May 2023. [17] Gazi Karam Illahi, Ashutosh Vaishnav, Teemu Kämäräinen, Matti Siekkinen, and Mario Di Francesco. Learning to predict head pose in remotely-rendered virtual reality. In Proceedings of the 14th Conference on ACM Multimedia Systems, MMSys 2023, Vancouver, BC, Canada, June 7-10, 2023, pages 27–38. ACM, 2023. [38] Wang Yifan, Shihao Wu, Hui Huang, Daniel Cohen-Or, and Olga Sorkine- Hornung. Patch-based progressive 3D point set upsampling. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5958–5967, 2019. 2023, Vancouver, BC, Canada, June 7-10, 2023, pages 27–38. ACM, 202 [18] Kyungjin Lee, Juheon Yi, Youngki Lee, Sunghyun Choi, and Young Min Kim. GROOT: a real-time streaming system of high-fidelity volumetric videos. In Proceedings of the 26th Annual International Conference on Mobile Computing and Networking, MobiCom 2020, London, United Kingdom, September 21-25, 2020, pages 57:1–57:14. ACM, 2020. [39] Emin Zerman, Radhika Kulkarni, and Aljosa Smolic. User behaviour analysis of volumetric video in augmented reality. In Proceedings of the 13th International Conference on Quality of Multimedia Experience, QoMEX 2021, Montreal, Canada, June 14-17, 2021, pages 129–132. IEEE, 2021. [40] Anlan Zhang, Chendong Wang, Bo Han, and Feng Qian. YuZu: Neural-enhanced volumetric video streaming. In Proceedings of the 19th USENIX Symposium on Networked Systems Design and Implementation, NSDI 2022, Renton, WA, USA, April 4-6, 2022, pages 137–154. USENIX Association, 2022. [19] Jie Li, Cong Zhang, Zhi Liu, Richang Hong, and Han Hu. Optimal volumetric video streaming with hybrid saliency based tiling. IEEE Transactions on Multimedia, 25:2939–2953, Feb 2023. [20] Ruihui Li, Xianzhi Li, Chi-Wing Fu, Daniel Cohen-Or, and Pheng-Ann Heng. PU-GAN: A point cloud upsampling adversarial network. In Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision, ICCV 2019, Seoul, Korea (South), October 27 - November 2, 2019, pages 7202–7211. IEEE, 2019. [41] Wenjie Zhu, Zhan Ma, Yiling Xu, Li Li, and Zhu Li. View-dependent dynamic point cloud compression. ACKNOWLEDGMENTS This work was supported by the Singapore Ministry of Education Academic Research Fund Tier 1 (T1 251RES2038). 153 MMSys ’24, April 15–18, 2024, Bari, Italy QV4: QoE-based Viewpoint-Aware V-PCC-encoded Volumetric Video Streaming REFERENCES IEEE Transactions on Circuits and Systems for Video Technology, 31(2):765–781, Feb 2021. 154
https://openalex.org/W4389384755	https://www.researchsquare.com/article/rs-3689383/latest.pdf	English	null	The mediating role of resilience and interaction anxiousness in the effects of physical activity on mobile phone addiction among college students	Research Square (Research Square)	2,023	cc-by	6,286	The mediating role of resilience and interaction anxiousness in the effects of physical activity on mobile phone addiction among college students Jinlong Wu Southwest University Wen Xiao Shenzhen University Bowen Liu Shenzhen University Jingxuan Yu Shenzhen University Kangyong Zheng The Hong Kong Polytechnic University Qiuqiong Shi (  gaby-qiuqiong.shi@polyu.edu.hk ) Laboratory for Artificial Intelligence in Design Zhanbing Ren Shenzhen University Conclusion It is essential to improve resilience and reduce interaction anxiousness to reduce MPA problems through regular engagement in PA among college students. Results PA were negatively associated with MPA behavior. Resilience and interaction anxiousness moderated the relationship between PA and MPA. More importantly, PA could also influence MPA through the chain- mediating effects of resilience and interaction anxiousness. Methods The participants were 590 college students (272 males; mean age = 19.67) who completed a psychosocial battery, including the international physical activity questionnaire—short form (IPAQ-SF), the connor - davidson resilience scale (CD-RISC), the interaction anxiousness scale (IAS), and the mobile phone addiction index (MPAI). Correlations of variables were computed using Pearson’s test. Mediation models were tested using SPSSS PROCESS macro with the regression bootstrapping method. Background Recent studies have shown that regular physical activity (PA) can positively influence mobile phone addiction (MPA) behaviors in college students. However, it remains unknown whether this effect is mediated by other factors. Evidence suggests that resilience and interaction anxiousness may be candidate mediators that partly explain the positive effect of PA on MPA. This study aims to explore the impact of PA on MPA through a mediation model, and the role of resilience and interaction anxiousness in this relationship. Research Article Posted Date: December 6th, 2023 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Additional Declarations: No competing interests reported. Additional Declarations: No competing interests reported. Page 1/16 1. Introduction Mobile phone addiction, also known as MPA, is a type of addictive behavior that is linked to excessive usage of mobile phones [1,2]. It is characterized by individuals having a strong desire for and dependency on their mobile devices, causing them to partake in activities that involve the overuse of these devices [3,4]. The harmful effects of MPA are particularly prominent among college students, who may experience a variety of symptoms such as headaches, dizziness, body aches, numbness in the extremities, dry eyes, blurred vision, and mental health disorders like depression, social anxiety, stress, and insomnia [5–9]. Experts specializing in mental health believe that MPA will become one of the most prevalent forms of Page 2/16 technology addiction in the 21st century [10]. Therefore, it is crucial to identify the risk and protective factors that contribute to MPA in order to prevent and manage this issue effectively. Research has indicated that the development of MPA behavior is associated with various adverse consequences, including detrimental mental and physical well-being, academic underachievement, and interpersonal difficulties [11,12]. Specifically, individuals at the college level often encounter significant challenges in their daily lives, such as interpersonal and academic hurdles. Due to their limited self- control capabilities, they experience negative emotions when confronted with these challenges [13–15]. In the presence of negative moods, college students tend to resort to engaging in MPA behavior as a means of regulating their emotions and obtaining immediate rewards. The use of mobile phones, being one of the most convenient and accessible electronic devices, provides an instant solution to fulfill multiple needs and serves as an easily accessible method for stress relief, thereby contributing to the emergence of MPA behavior among college students [16,17]. The recommended amount of physical activity (PA) by the World Health Organization is at least 150 minutes per week for moderate-to-intense PA. Alternatively, engaging in vigorous to intense PA for more than 75 minutes each week is also advised. This is to prevent chronic diseases, as regular PA offers significant benefits for both physical and mental health [18]. Several studies have found negative correlations between PA and MPA [19,20]. Recent research has indicated that PA can effectively predict MPA in a negative manner. Additionally, the potential mediating role of self-control between PA and MPA has been taken into consideration [21]. 1. Introduction These findings underscore the potential advantages of PA, which could serve as an important protective mechanism against MPA. Resilience, defined by the American Psychological Association (2014), refers to "the ability to adapt well in the face of hardship, trauma, adversity, significant sources of stress, or even tragedy" [22]. In simpler terms, resilience allows individuals to return to a previous state of normalcy or health following a trauma, accident, tragedy, or illness, which is vital for both mental and physical well-being [23]. It is worth mentioning that engaging in regular physical activity (PA) has been scientifically proven to effectively enhance resilience among college students. This improvement in resilience subsequently boosts their problem-solving skills, self-confidence, and ability to regulate emotions [24] [25]. According to the ego- depletion model of self-regulation, individuals who struggle to successfully adapt to social issues may experience reduced self-control when it comes to using their phones [26]. Consequently, it can be inferred that college students with low levels of resilience may excessively rely on the internet as a coping mechanism when faced with challenges that they are unable to resolve, particularly social problems [27]. Building upon this existing evidence, resilience can be considered an internal mechanism that mediates the relationship between PA and MPA. Interaction anxiousness, also known as social anxiety, is a psychological condition characterized by an individual's persistent fear and emotional discomfort in social or performance situations. Individuals experiencing interaction anxiousness often seek to escape or avoid association with other people [28]. Embodied cognition theories suggest a close connection between our mental states and the physical Page 3/16 Page 3/16 situations we find ourselves in [29]. Research has consistently shown that regular physical activity (PA) can lead to higher emotional stability and reduced anxiety [30–33]. Additionally, the compensatory Internet use theory proposes that individuals who face psychological challenges in the real world may turn to the Internet or smartphones as a means of escape and pain reduction [34]. Notably, individuals with higher levels of interaction anxiety tend to seek support from the Internet, leading to excessive use of mobile phones [35]. Taking into account this scientific evidence, it is plausible to consider interaction anxiousness as a significant mediating factor in the relationship between PA and mobile phone addiction (MPA) among college students. 1. Introduction According to the dynamic model of psychological resilience, resilience can be considered a protective factor against negative emotions (e.g., interaction anxiousness) [36], which can be improved by regular physical exercise [37,38]. In other words, individuals with a higher level of resilience could more easily restore their mental balance while experiencing psychological distress, accompanied by a reduction in anxiety levels. Although we hypothesized that resilience and interaction anxiousness may both mediate the effect of PA on MPA, sequential mediation models may be based on the relationship between resilience and interaction anxiousness. Based on the theories and literature, it appears to be a negative association between PA and MPA among college students in daily life. In sporadic research that has explored the underlying mechanisms between PA and MPA, it is suggested that resilience and interaction anxiousness play a mediating role during the process. The present study was guided by the following hypotheses: Based on the theories and literature, it appears to be a negative association between PA and MPA among college students in daily life. In sporadic research that has explored the underlying mechanisms between PA and MPA, it is suggested that resilience and interaction anxiousness play a mediating role during the process. The present study was guided by the following hypotheses: (1) Resilience and interaction anxiousness mediate the relationship between PA and MPA in college students (H1→H5; H4→H3). (2) Resilience and interaction anxiousness play a chain-mediating role in the relationship between PA and MPA among college students (H1→H2→H3). This study followed these assumptions to investigate the relationships between the four key variables using the research hypothesis model shown in Fig. 1. This study followed these assumptions to investigate the relationships between the four key variables using the research hypothesis model shown in Fig. 1. 2.1. Participants A convenience sampling approach was used to recruit targeted college students aged between 18 and 25 years in three university in Guangdong province, southern China. Eligible participants were described in the study and were asked whether they were willing to participate. After providing written informed consent, the participants were asked to complete a comprehensive questionnaire to collect demographic information and other subsequent procedures. Questionnaires were collected from March 2022 to May 2022, and the initial sample comprised 647 college students. After the questionnaires were collected, we conducted repeated inspections and screenings to eliminate questionnaires that took less than 2 min to Page 4/16 Page 4/16 complete, had irregular answers, and were logically contradictory. The final sample of valid responses was obtained from 590 participants (272 males and 318 females; Mage = 19.67, SD = 1.48). The recruitment and data collection procedures were approved by the Ethics Committee of Shenzhen University. complete, had irregular answers, and were logically contradictory. The final sample of valid responses was obtained from 590 participants (272 males and 318 females; Mage = 19.67, SD = 1.48). The recruitment and data collection procedures were approved by the Ethics Committee of Shenzhen University. Interaction Anxiousness In order to evaluate the levels of interaction anxiousness, the Chinese adaptation of the interaction anxiousness scale (IAS) was employed [41]. Consisting of 15 items, this version requires participants to rate their responses on a 5-point Likert scale. Respondents rate the items on a scale ranging from 1 (never) to 5 (always). Enhanced scores signify elevated levels of interaction anxiety. The reliability of the IAS, as measured by Cronbach's alpha, was found to be 0.89 in the current study. Physical Activity The levels of PA among college students were evaluated through the utilization of the validated Chinese version[39] of the international questionnaire on physical activity—abbreviated form (IPAQ-SF)[40]. IPAQ- SF determines the PA levels in the past 7 days by gathering and dividing the time spent on vigorous activity, moderate activity, walking, and sitting. The total PA level (represented by metabolic equivalent [MET]) was calculated based on the quantities of vigorous, moderate, and walking activities. In this study, Cronbach’s alpha for the IPAQ-SF was determined to be 0.87. Resilience To measure resilience over the last month, we utilized the Chinese edition of the Connor-Davidson resilience scale (CD-RISC) [35]. This scale, consisting of 25 items, incorporates three main aspects: tenacity, strength, and optimism. Participants rated each item on a 5-point Likert scale, ranging from 1 (hardly ever) to 5 (almost always). A greater score indicates a higher degree of resilience. Our study observed a high level of internal consistency (Cronbach’s α = 0.93). 2.2. Measure Physical Activity 3.1. Common Method Bias Common method bias was tested using Harman’s single-factor method. We first conducted exploratory factor analyses for all items and each scale. The results indicated that 11 factors had eigenvalues higher than 1.0; the first factor explained only 20.74% of the total variance, which is lower than 50%. Thus, the results indicated that the common method bias was not large enough to distort the results. Mobile Phone Addiction To evaluate mobile phone addiction, researchers utilized the mobile phone addiction index (MPAI)[42]. This index comprises 17 items that gauge four different aspects of smartphone addiction: lack of control over cravings, anxiety levels, disorientation and withdrawal, and diminished productivity. Survey respondents provided answers to these items using a 5-point scale, ranging from 1 (never) to 5 (always). Prior investigations have attested to the MPAI's reliability and validity among Chinese adolescents and young adults[43]. In our current study, the measure displayed strong internal consistency, as indicated by Cronbach's α coefficient of 0.88. 2.3. Data Analysis Data were analyzed using SPSS 21.0 software and the PROCESS macro[44]. Specifically, the total score of each scale was first calculated according to the corresponding formula and rules. Second, Harman’s single-factor test was used to assess common method bias, which indicated a concern if one factor explained more than 50% of the total variance [45,46]. Third, gender differences in all tested variables were tested by independent sample t-test. Pearson’s correlations between each of the two dependent variables (PA, resilience, interaction anxiousness, and MPA) were tested to support the research hypotheses. Finally, according to our hypotheses, we constructed three models to test for validity using SPSS 21.0 (Model 6, the PROCESS macro). The bias-corrected percentile bootstrap method was used to test the mediation models (mediation model of resilience; mediation model of interaction anxiousness) and the chain mediation model of resilience and interaction anxiousness. The mediation model was implemented with 5,000 bootstrap samples and 95% corrected confidence intervals (CIs) [47,48]. Note. : P < 0.05, : P < 0.01 Note. : P < 0.05, *: P < 0.01 3.2. Descriptive Statistics Descriptive statistics showed that the scores of self-reported levels of PA, resilience, interaction anxiousness, and MPA were moderate (Table 1). Moreover, resilience and interaction anxiousness showed significant gender differences; specifically, male students demonstrated higher resilience scores than female students, and female students had higher interaction anxiousness than male students. Pearson’s correlation analysis found that PA had a significantly positive correlation with resilience, and it was negatively correlated with interaction anxiety and MPA. Resilience was positively associated with interaction anxiousness and MPA, and interaction anxiousness had a significantly positive association with MPA. Correlation coefficients are displayed in Table 2. Page 6/16 Page 6/16 Table 1 Gender difference in all tested variables. Note. : P < 0.05, : P < 0.01 Variables Total (n = 590) Male (n = 272) Female (n = 318) M ± SD M ± SD M ± SD T P Age 19.67 ± 1.48 19.95 ± 1.70 19.44 ± 1.22 4.16 0.00 Physical Activity (MET) 3696.46 ± 2295.19 3683.254 ± 2281.37 3707.76 ± 2310.48 -0.13 0.90 Resilience 77.25 ± 20.35 79.23 ± 19.823 75.55 ± 20.67 2.197* 0.03 Interaction Anxiousness 42.47 ± 10.21 42.56 ± 10.26 45.1 ± 10.04 -3.039 0.00 Mobile Phone Addiction 42.47 ± 10.56 42.19 ± 9.75 42.71 ± 11.22 -0.59 0.55 Table 2 Correlations of all tested variables. - Physical Activity (MET) Resilience Interaction Anxiousness Mobile Phone Addiction Physical Activity (MET) - - - - Resilience − .30 - - - Interaction Anxiousness .23 − .31 - - Mobile Phone Addiction − .21 .31 .22** - Correlations of all tested variables. 3.3. Chained Mediating Analyses According to the model 6 in the Process program, a chain mediation model was established with PA as an independent variable, resilience and interaction anxiousness as mediating variables, and MPA as a dependent variable. As shown in Fig. 2, PA significantly and positively predicted resilience (H1 = 0.23, P < 0.01), resilience significantly negatively predicted interaction anxiousness (H2 = -0.18, P < 0.01), and interaction anxiousness significantly and positively predicted MPA (H3 = 0.22 P < 0.01). Table 3 showed the overall path coefficients of the mediation analysis. Specifically, in the total effect model, the direct predictive effect of PA on MPA in this study revealed that PA was able to predict MPA significantly and negatively (H7 = -0.30, 95%CIs [-0.002, -0.001]). When resilience and interaction anxiousness as mediating variables were added to construct a chain mediation model, PA still played a significant role in predicting MPA (H6 = -0.20, 95%Cls [-0.002, -0.001]). The indirect effect of perceived social support through resilience and interaction anxiousness was also significant (H1→H2→H3 = -0.01, Page 7/16 Page 7/16 95% CIs [-0.016, -0.003]), which confirmed the research hypotheses. Each mediation path had a significant effect on MPA (H1→H5, H4→H3= -0.05, -0.04, 95% CIs [-0.075, -0.029], [-0.063, -0.018], respectively), which also confirmed the research hypotheses. Table 3 Chain mediation analysis results. Standardized Effect Standard Error 95% Confidence Interval Lower Upper H7 Total effect -0.30 0.000 -0.002 -0.001 H6 Direct effect -0.20 0.000 -0.002 -0.001 H1→H5 Path 1 -0.05 0.012 -0.075 -0.029 H4→H3 Path 2 -0.04 0.012 -0.063 -0.018 H1→H2→H3 Path 3 -0.01 0.003 -0.016 -0.004 Table 3 Chain mediation analysis results. 4. Discussion The purpose of this study was to analyze the effect of PA on MPA among college students and the mediating role of resilience and interaction anxiousness in this relationship, which further extends previous studies [46] on the mechanism of PA on MPA and helps to gain a better understanding of the potential benefit of PA concerning MPA. Three mediating paths were found in the mechanism underlying the relationship between PA and MPA among Chinese college students. Specifically, resilience and interaction anxiousness moderated the relationship between PA and MPA. More importantly, PA could also influence MPA through the chain-mediating effects of resilience and interaction anxiousness. Physical exercise is a positive and effective means of health promotion, and an increase in exercise can improve the brain’s emotional processing ability to relieve anxiety and reduce phone addiction. Thus, based on the abovementioned mechanisms, physical exercise could be an intervention to treat college students with MPA. First, resilience appears to be an important mediator in the effects of PA on MPA among college students. College students with higher levels of PA had higher levels of resilience, which significantly and positively predicted MPA among them. Our study found that PA has a significant effect on resilience, similar to the findings of previous studies [49,50], because it can enhance emotion regulation by protecting neurons in regions of the brain, such as the striatum and hippocampus [25]. According to the dynamic model of psychological resilience, when college students face adverse events such as bad interpersonal relationships and academic pressure, individuals with high resilience levels regulate their negative emotions and avoid being frustrated by real problems and indulging in the online world [51]. The university stage is a transition period into social life, where many stressors and challenges are Page 8/16 Page 8/16 experienced in daily life and studies. Participation in PA may be an appropriate way to regulate emotions by enhancing psychological resilience, resulting in reduced MPA. Second, it appears that interaction anxiousness mediated the association between PA and MPA among college students. Due to the mature entertainment and social functions of cell phones, as well as their universality and accessibility, their use has become the main way for college students to satisfy the need for interaction while reducing the anxiety of realistic interactions. The final result is that college students indulge more in the virtual environment built by cell phones, thus increasing their MPA level [52]. 4. Discussion The interaction of the person-affect-cognition-execution (I-PACE) model suggests that individuals’ characteristics (e.g., personality, spirit) may influence their cognitive function and behaviors; thus, reducing negative emotions through regular PA is a potential mechanism for reducing MPA [53]. As known, exercise stimulates the release of endorphins, a kind of polypeptide beneficial for reducing negative emotions [46]. Third, this study found that resilience and interaction anxiousness played a chain-mediating role in the association between PA and MPA among college students, revealing the mechanism by which PA affects MPA through the combined effect of resilience and interaction anxiousness. Our findings suggest close connections between resilience and interaction anxiousness, which echoes a previous study of negative emotions [54]. Based on our findings, we speculate that the higher the PA level of college students, the higher the level of resilience, and the more the benefits of flexibly dealing with the negative effects of changes in the external environment, such as reducing social avoidance and distress in social situations, which can reduce the tendency toward cell phone addiction [55]. Third, this study found that resilience and interaction anxio association between PA and MPA among college students, MPA through the combined effect of resilience and interact connections between resilience and interaction anxiousnes emotions [54]. Based on our findings, we speculate that the higher the level of resilience, and the more the benefits of f changes in the external environment, such as reducing soci which can reduce the tendency toward cell phone addiction 5. Practical implications and limitations In summary, the results support the hypotheses of our study, and this study introduced two variables – resilience and interaction anxiousness – which expands on the existing research into the impact of PA on MPA. Following the PA Guidelines for World Health Organization explanation, we suggest the actual amount of PA can higher than the PA Guidelines for World Health Organization recommend because it can bring more mental health benefits. In particular, during the COVID-19 epidemic, periods of confinement can lead to physical dysfunction and mental distress (e.g., MPA), partially attributed to reductions in habitual physical activity. Our findings indicate that college students should actively perform PA to improve resilience and reduce interaction anxiousness, which is conducive to reducing the negative consequences of MPA. Our finding have certain guiding significance for the prevention of MPA behavior among college students. Therefore, college students need to closely monitor their own the level of PA during the COVID-19 epidemic, especially for those who want to regulate their own mobile phone use behaviors. However, some limitations of this study should be mentioned when interpreting its results. First, the data collected were cross-sectional and could only reveal a correlation—and not a causal relationship— between PA and MPA among college students. Further research and experiments are needed to verify this Page 9/16 Page 9/16 causality. Second, all subjects in this study were from one district of Guangdong province, China, which limited the representativeness of the sample. Extending the survey to a national sample is necessary for future studies. Third, all data in this study were collected through self-reported scales, and bias may exist because of the social desirability effect and/or memory errors. Finally, from the perspective of data analysis, gender differences were observed in resilience and interaction anxiety. More research is thus needed to uncover the mechanisms underlying the association between PA and MPA. Acknowledgements The authors would like to thank all students who participated in this study. 6. Conclusion In this study, we found that PA was significantly related to MPA among college students and that this relationship was mediated by a series of associations between resilience and interaction anxiety. It is thus recommended to improve resilience and reduce interaction anxiousness to curb MPA problems through regular engagement in PA among college students. Availability of data and materials The data and materials are available upon request from the corresponding author. Funding Research Foundation for Young Teacher of Shenzhen University[grant number QNJS0274]；High-level Scientific Research Foundation for the Introduction of Talent of Shenzhen University[grant number RC00228]; Natural Science Featured Innovation Projects in Ordinary Universities in Guangdong Province(grant number 2021KTSCX297) Authors’ contributions Conception and design of the study (JLW, ZBR); acquisition of data (JLW,WX); analysis and reporting of data (WX,BWL); drafting of the manuscript (JLW, WX). Review and editing manuscript (KYZ, BWL,QQS, JXY and ZBR). All authors have read and agreed to the published version of the manuscript. Ethics approval and consent to participate Ethical approval for the study was obtained from the Ethics Committee of Shenzhen University. Informed consent to participate was obtained for all participants before survey initiation. All methods were per‑formed in accordance with the relevant guidelines and regulations. Consent for publication Consent for publication Page 10/16 Not applicable. Competing interests The authors declare no conflict of interest with respect to the research, authorship, and/or publication of this article. References 1. Yang, X.J.; Liu, Q.Q.; Lian, S.L.; Zhou, Z.K. Are bored minds more likely to be addicted? The relationship between boredom proneness and problematic mobile phone use. Addictive Behaviors. 2020, 108, 106426. 2. Zhang, G.; Yang, X.; Tu, X.; Ding, N.; Lau, J.T.F. Prospective relationships between mobile phone dependence and mental health status among Chinese undergraduate students with college adjustment as a mediator. Journal of Affective Disorders. 2020, 260, 498-505. 3. Zou, Z.; Wang, H.; d’Oleire Uquillas, F.; Wang, X.; Ding, J.; Chen, H. Definition of substance and non- substance addiction. Substance and Non-substance Addiction. 2017, 21-41. 3. Zou, Z.; Wang, H.; d’Oleire Uquillas, F.; Wang, X.; Ding, J.; Chen, H. Definition of substance and non- substance addiction. Substance and Non-substance Addiction. 2017, 21-41. 4. Billieux, J.; Maurage, P.; Lopez-Fernandez, O.; Kuss, D.J.; Griffiths, M.D. Can disordered mobile phone use be considered a behavioral addiction? An update on current evidence and a comprehensive model for future research. Current Addiction Reports. 2015, 2, 156-162. 4. Billieux, J.; Maurage, P.; Lopez-Fernandez, O.; Kuss, D.J.; Griffiths, M.D. Can disordered mobile phone use be considered a behavioral addiction? An update on current evidence and a comprehensive model for future research. Current Addiction Reports. 2015, 2, 156-162. 5. Hong, F.Y.; Chiu, S.I.; Huang, D.-H. A model of the relationship between psychological characteristics, mobile phone addiction and use of mobile phones by Taiwanese university female students. Computers in human behavior. 2012, 28, 2152-2159. 5. Hong, F.Y.; Chiu, S.I.; Huang, D.-H. A model of the relationship between psychological characteristics, mobile phone addiction and use of mobile phones by Taiwanese university female students. Computers in human behavior. 2012, 28, 2152-2159. 6. Višnjić, A.; Veličković, V.; Sokolović, D.; Stanković, M.; Mijatović, K.; Stojanović, M.; Milošević, Z.; Radulović, O. Relationship between the manner of mobile phone use and depression, anxiety, and stress in university students. International journal of environmental research and public health. 2018, 15, 697. 7. Ivanova, A.; Gorbaniuk, O.; Błachnio, A.; Przepiórka, A.; Mraka, N.; Polishchuk, V.; Gorbaniuk, J. Mobile phone addiction, phubbing, and depression among men and women: a moderated mediation analysis. Psychiatric Quarterly. 2020, 91, 655-668. 8. Zhang, T.; Gong, N.; Jia, R.; Li, H.; Ni, X. Stroop effect in smartphone addiction among college students. Medicine. 2021, 100. 8. Zhang, T.; Gong, N.; Jia, R.; Li, H.; Ni, X. Stroop effect in smartphone addiction among college students. Medicine. 2021, 100. 9. References Kaya, F.; Bostanci Daştan, N.; Durar, E. Smart phone usage, sleep quality and depression in university students. International Journal of Social Psychiatry. 2021, 67, 407-414. 9. Kaya, F.; Bostanci Daştan, N.; Durar, E. Smart phone usage, sleep quality and depression in university students. International Journal of Social Psychiatry. 2021, 67, 407-414. 10. Lin, Y.H.; Lin, S.H.; Yang, C.C.H.; Kuo, T.B.J. Psychopathology of everyday life in the 21st century: smartphone addiction. In Internet addiction; Springer: 2017; pp. 339-358. 10. Lin, Y.H.; Lin, S.H.; Yang, C.C.H.; Kuo, T.B.J. Psychopathology of everyday life in the 21st century: smartphone addiction. In Internet addiction; Springer: 2017; pp. 339-358. 11. Li, L.; Xu, D.D.; Chai, J.X.; Wang, D.; Li, L.; Zhang, L.; Lu, L.; Ng, C.H.; Ungvari, G.S.; Mei, S.L. Prevalence of Internet addiction disorder in Chinese university students: A comprehensive meta-analysis of observational studies. Journal of behavioral addictions. 2018, 7, 610-623. Page 11/16 12. Deci, E.L.; Ryan, R.M. The" what" and" why" of goal pursuits: Human needs and the self-determination of behavior. Psychological inquiry. 2000, 11, 227-268. 13. Augner, C.; Hacker, G.W. Associations between problematic mobile phone use and psychological parameters in young adults. International journal of public health. 2012, 57, 437-441. 14. Kuang-Tsan, C.; Fu-Yuan, H. Study on relationship among university students’ life stress, smart mobile phone addiction, and life satisfaction. Journal of Adult Development. 2017, 24, 109-118. 15. Lepp, A.; Barkley, J.E.; Karpinski, A.C. The relationship between cell phone use, academic performance, anxiety, and satisfaction with life in college students. Computers in human behavior. 2014, 31, 343-350. 16. Lf, A.; Pw, A.; Meng, Z.B.; Xin, X.C.; Ye, C.D.; Jia, N.A.; Li, L.A.J.C.; Review, Y.S. Can emotion regulation difficulty lead to adolescent problematic smartphone use? A moderated mediation model of depression and perceived social support - ScienceDirect. 2020, 108. 17. Southwick, S.M.; Ozbay, F.; Mayes, L.C. Psychological and biological factors associated with resilience to stress and trauma. 2008. 17. Southwick, S.M.; Ozbay, F.; Mayes, L.C. Psychological and biological factors associated with resilience to stress and trauma. 2008. 18. World Health, O. Public health implications of excessive use of the internet, computers, smartphones and similar electronic devices: Meeting report, Main Meeting Hall, Foundation for Promotion of Cancer Research, National Cancer Research Centre, Tokyo, Japan, 27-29 August 2014; World Health Organization: 2015. 19. Kim, S.-E.; Kim, J.-W.; Jee, Y.-S. Relationship between smartphone addiction and physical activity in Chinese international students in Korea. Journal of behavioral addictions. References 2015, 4, 200-205. 20. Zagalaz-Sánchez, M.L.; Cachón-Zagalaz, J.; Sánchez-Zafra, M.; Lara-Sánchez, A. Mini review of the use of the mobile phone and its repercussion in the deficit of physical activity. Frontiers in psychology. 2019, 10, 1307. 21. Guo, K.l.; Ma, Q.s.; Yao, S.j.; Liu, C.; Hui, Z.; Jiang, J.; Lin, X. The Relationship Between Physical Exercise and Mobile Phone Addiction Tendency of University Students in China: A Moderated Mediation Model. Frontiers in Psychology. 2022, 13. 21. Guo, K.l.; Ma, Q.s.; Yao, S.j.; Liu, C.; Hui, Z.; Jiang, J.; Lin, X. The Relationship Between Physical Exercise and Mobile Phone Addiction Tendency of University Students in China: A Moderated Mediation Model. Frontiers in Psychology. 2022, 13. 22. Cousins, J.N.; Fernández, G. The impact of sleep deprivation on declarative memory. Progress in brain research. 2019, 246, 27-53. doi:10.1016/bs.pbr.2019.01.007. 22. Cousins, J.N.; Fernández, G. The impact of sleep deprivation on declarative memory. Progress in brain research. 2019, 246, 27-53. doi:10.1016/bs.pbr.2019.01.007. 23. Babić, R.; Babić, M.; Rastović, P.; Ćurlin, M.; Šimić, J.; Mandić, K.; Pavlović, K. Resilience in health and illness. Psychiatria Danubina. 2020, 32, 226-232. 23. Babić, R.; Babić, M.; Rastović, P.; Ćurlin, M.; Šimić, J.; Mandić, K.; Pavlović, K. Resilience in health and illness. Psychiatria Danubina. 2020, 32, 226-232. 24. Xia, B.; Ma, Z.; Hu, Y. Research on the relationship between physical exercise, psychological flexibility and positive emotion of college students based on computer mathematical model. 2020; p. 012009. 24. Xia, B.; Ma, Z.; Hu, Y. Research on the relationship between physical exercise, psychological flexibility and positive emotion of college students based on computer mathematical model. 2020; p. 012009. 25. Xu, S.; Liu, Z.; Tian, S.; Ma, Z.; Jia, C.; Sun, G. Physical activity and resilience among college students: The mediating effects of basic psychological needs. International Journal of Environmental Research and Public Health. 2021, 18, 3722. 25. Xu, S.; Liu, Z.; Tian, S.; Ma, Z.; Jia, C.; Sun, G. Physical activity and resilience among college studen The mediating effects of basic psychological needs. International Journal of Environmental Research and Public Health. 2021, 18, 3722. 26. Tokunaga, R.S.; Rains, S.A. An evaluation of two characterizations of the relationships between problematic Internet use, time spent using the Internet, and psychosocial problems. Human Communication Research. 2010, 36, 512-545. 26. Tokunaga, R.S.; Rains, S.A. An evaluation of two characterizations of the relationships between problematic Internet use, time spent using the Internet, and psychosocial problems. 12. Deci, E.L.; Ryan, R.M. The" what" and" why" of goal pursuits: Human needs and the self-determination of behavior. Psychological inquiry. 2000, 11, 227-268. References Human Communication Research. 2010, 36, 512-545. Page 12/16 Page 12/16 27. Peng, W.; Liu, M. Online gaming dependency: a preliminary study in China. Cyberpsychology, behavior, and social networking. 2010, 13, 329-333. 28. Buyukbayraktar, C.G. Predictive relationships among smartphone addiction, fear of missing out and interaction anxiousness. European Journal of Educational Sciences. 2020, 7, 1-16. 29. Egger, S.W.; Remington, E.D.; Chang, C.J.; Jazayeri, M. Internal models of sensorimotor integration regulate cortical dynamics. Nature neuroscience. 2019, 22, 1871-1882. 30. Qian, H.B.; Wang, Z.Y. Literature Review of Physical Exercise and Mental Health Study. China School Physical Education (Higher Education): Nanjing, China. 2017. 31. Di Benedetto, M.; Len Burns, G.; Lindner, H.; Kent, S. A biopsychosocial model for depressive symptoms following acute coronary syndromes. Psychology and Health. 2010, 25, 1061-1075. 32. Vasey, M.W.; Bosmans, G.; Ollendick, T.H. The developmental psychopathology of anxiety. In Handbook of developmental psychopathology; Springer: 2014; pp. 543-560. 33. Schuch, F.B.; Bulzing, R.A.; Meyer, J.; Vancampfort, D.; Firth, J.; Stubbs, B.; Grabovac, I.; Willeit, P.; Tavares, V.D.O.; Calegaro, V.C. Associations of moderate to vigorous physical activity and sedentary behavior with depressive and anxiety symptoms in self-isolating people during the COVID-19 pandemic: A cross-sectional survey in Brazil. Psychiatry research. 2020, 292, 113339. 34. Kardefelt-Winther, D. A conceptual and methodological critique of internet addiction research: Towards a model of compensatory internet use. Computers in human behavior. 2014, 31, 351-354. 35. Bian, M.; Leung, L. Linking loneliness, shyness, smartphone addiction symptoms, and patterns of smartphone use to social capital. Social science computer review. 2015, 33, 61-79. 36. Li, X.; Yu, H.; Yang, N. The mediating role of resilience in the effects of physical exercise on college students’ negative emotions during the COVID-19 epidemic. Scientific reports. 2021, 11, 1-8. 37. Spies, G.; Seedat, S. Depression and resilience in women with HIV and early life stress: does trauma play a mediating role? A cross-sectional study. BMJ open. 2014, 4, e004200. 38. Richardson, G.E. The metatheory of resilience and resiliency. Journal of clinical psychology. 2002, 58, 307-321. 39. Wang, C.; Chen, P.; Zhuang, J. Validity and reliability of international physical activity questionnaire– short form in Chinese youth. Research Quarterly for Exercise and Sport. 2013, 84, S80-S86. 40. Meeus, M.; Van Eupen, I.; Willems, J.; Kos, D.; Nijs, J.O. Is the International Physical Activity Questionnaire-short form (IPAQ-SF) valid for assessing physical activity in Chronic Fatigue Syndrome? Disability and rehabilitation. 2011, 33, 9-16. 41. Leary, M.R. References Social anxiousness: The construct and its measurement. Journal of personality assessment. 1983, 47, 66-75. 41. Leary, M.R. Social anxiousness: The construct and its measurement. Journal of personality assessment. 1983, 47, 66-75. 42. Leung, L. Leisure boredom, sensation seeking, self-esteem, and addiction: Symptoms and patterns of cell phone use. In Mediated interpersonal communication; Routledge: 2008; pp. 373-396. 42. Leung, L. Leisure boredom, sensation seeking, self-esteem, and addiction: Symptoms and patterns of cell phone use. In Mediated interpersonal communication; Routledge: 2008; pp. 373-396. 43. Lian, L.; You, X.; Huang, J.; Yang, R. Who overuses smartphones? Roles of virtues and parenting style in smartphone addiction among Chinese college students. Computers in Human Behavior. 2016, 65, 43. Lian, L.; You, X.; Huang, J.; Yang, R. Who overuses smartphones? Roles of virtues and parenting style in smartphone addiction among Chinese college students. Computers in Human Behavior. 2016, 65, Page 13/16 Page 13/16 92-99. 44. Hayes, A.F. Introduction to mediation, moderation, and conditional process analysis: A regression- based approach; Guilford publications: 2017. 45. Podsakoff, P.M.; MacKenzie, S.B.; Lee, J.-Y.; Podsakoff, N.P. Common method biases in behavioral research: a critical review of the literature and recommended remedies. Journal of applied psychology. 2003, 88, 879. 46. Akdoğan, R.; Çimşir, E. Linking inferiority feelings to subjective happiness: Self-concealment and loneliness as serial mediators. Personality and Individual Differences. 2019, 149, 14-20. 47. Hayes, A.F. PROCESS: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling. 2012. 48. Bolin, J.H. Introduction to mediation, moderation, and conditional process analysis: a regression- based approach. 2014. 49. Zschucke, E.; Renneberg, B.; Dimeo, F.; Wüstenberg, T.; Ströhle, A. The stress-buffering effect of acute exercise: Evidence for HPA axis negative feedback. Psychoneuroendocrinology. 2015, 51, 414-425. 50. Yoshikawa, E.; Nishi, D.; Matsuoka, Y.J. Association between regular physical exercise and depressive symptoms mediated through social support and resilience in Japanese company workers: a cross- sectional study. BMC Public Health. 2016, 16, 1-8. 51. Furlong, M.J.; Ritchey, K.M.; O’Brennan, L.M. Developing norms for the California Resilience Youth Development Module: Internal assets and school resources subscales. The California school psychologist. 2009, 14, 35-46. 51. Furlong, M.J.; Ritchey, K.M.; O’Brennan, L.M. Developing norms for the California Resilience Youth Development Module: Internal assets and school resources subscales. The California school psychologist. 2009, 14, 35-46. 52. Mimi, M.Y.; Tang, S.K.; Wan, V.T.C.; Vong, S.K.S. References The effectiveness of physical exercise training in pain, mobility, and psychological well-being of older persons living in nursing homes. Pain Management Nursing. 2014, 15, 778-788. 52. Mimi, M.Y.; Tang, S.K.; Wan, V.T.C.; Vong, S.K.S. The effectiveness of physical exercise training in pain, mobility, and psychological well-being of older persons living in nursing homes. Pain Management Nursing. 2014, 15, 778-788. 53. Brand, M.; Young, K.S.; Laier, C.; Wölfling, K.; Potenza, M.N. Integrating psychological and neurobiological considerations regarding the development and maintenance of specific Internet-use disorders: An Interaction of Person-Affect-Cognition-Execution (I-PACE) model. Neuroscience & Biobehavioral Reviews. 2016, 71, 252-266. 53. Brand, M.; Young, K.S.; Laier, C.; Wölfling, K.; Potenza, M.N. Integrating psychological and neurobiological considerations regarding the development and maintenance of specific Internet-use disorders: An Interaction of Person-Affect-Cognition-Execution (I-PACE) model. Neuroscience & Biobehavioral Reviews. 2016, 71, 252-266. 54. Fbs, A.; Rab, A.; Jm, B.; Dv, C.; Jfd, E.; Bsf, G.; Ig, H.; Pwi, J.; Vdot, K.; Vcc, L.J.P.R. Associations of moderate to vigorous physical activity and sedentary behavior with depressive and anxiety symptoms in self-isolating people during the COVID-19 pandemic: A cross-sectional survey in Brazil. 2020, 292. 54. Fbs, A.; Rab, A.; Jm, B.; Dv, C.; Jfd, E.; Bsf, G.; Ig, H.; Pwi, J.; Vdot, K.; Vcc, L.J.P.R. Associations of moderate to vigorous physical activity and sedentary behavior with depressive and anxiety symptoms in self-isolating people during the COVID-19 pandemic: A cross-sectional survey in Brazil. 2020, 292. 55. Ma, X.; Wang, Y.; Hu, H.; Tao, X.G.; Zhang, Y.; Shi, H. The impact of resilience on prenatal anxiety and depression among pregnant women in Shanghai. Journal of affective disorders. 2019, 250, 57-64. 55. Ma, X.; Wang, Y.; Hu, H.; Tao, X.G.; Zhang, Y.; Shi, H. The impact of resilience on prenatal anxiety and depression among pregnant women in Shanghai. Journal of affective disorders. 2019, 250, 57-64. 55. Ma, X.; Wang, Y.; Hu, H.; Tao, X.G.; Zhang, Y.; Shi, H. The impact of resilience on prenatal anxiety and depression among pregnant women in Shanghai. Journal of affective disorders. 2019, 250, 57-64. Figures Page 14/16 Figure 1 Research model. egression analysis of the chain mediating model. : P < 0.05, : P < 0.01 Figure 1 Research model. Figure 2 Regression analysis of the chain mediating model. : P < 0.05, *: P < 0.01 Regression analysis of the chain mediating model. : P < 0.05, **: P < 0.01 Page 16/16
https://openalex.org/W1975400386	https://zenodo.org/records/1957946/files/article.pdf	German	null	Ueber die Umwandlung der Isodialursäure in Dialursäure;	Justus Liebigs Annalen der Chemie	1,901	public-domain	4,277	Diese Bnnaleii 251; 24e. (Einyelaufen am 8. Januar 1901.) (Einyelaufen am 8. Januar 1901.) 246 246 KO ech, Ueber die Uwawa.ndlung der Isodiialurs&wre Die Saure ist in "armem Wasser sehr leicht lijslich und krystallisirt in glanzenden , weissen Schuppchen. Sehr wenig lost sie sich in kochendem Benzol, Schwefelkohlenstoff und Chloroform, sehr leicht hingegen in Aceton. Aus der heissen Acetonlosung wird sie auf Benzolzusatz in dunnen, verwachsenen Nadelchen ausgeschieden, die untcr den? Polarisationsmikroskope als lange Saulen mit theils gerader, theils schiefer Ausl6schung erscheinen ; sie gehoren daher dem monoklinen System an. Die einzelnen Flachen waren nicht bestimmbar. Silbernitrat fallt aus der mit Ammoniak neutralisirten, wassrigen Losung der Saure einen flockigen, weissen Nieder- schlag des Silbersalzes. Es ist in der Falluugsfliissiglreit zieni- lich bestandig, schwarzt sich aber nach dem Filtriren, auch in einem braunen Exsiccator, in kurzer Zeit. die Urnwandlung der Isodialurs&ure in Dialurssure ; von Paul Koech. die Urnwandlung der Isodialurs&ure in Dialurssure ; die Urnwandlung der Isodialurs&ure in Dialurssure ; die Urnwandlung der Isodialurs&ure in Dialurssure ; von Paul Koech. (Mitgetheilt von Robert Behreml.) (Mitgetheilt von Robert Behreml.) Mittheilungen aux dem organixch - chemixchen Laboratorium der Technischen Hochxchule Hannover, Mittheilungen aux dem organixch - chemixchen Laboratorium der Technischen Hochxchule Hannove (Einyelaufen am 8. Januar 1901.) ’) lac. cit. 244. (Mitgetheilt von Robert Behreml.) Bei ihren Versuchen zur Synthose der I-Iarnsaure haben B e h r e n d und R o o s e n l) eine der Dialursaure NH-CO I I I 1 CO CHOH NIT-CO isomere Verbindung, die Isodialursaure, erhalteu , welcher sic die Formel 247 in Dialursaure. 247 in Dialursaure. NH-CHOH KH-CHOH 1 I CO CO i 1 f 2H20 oder CO ,<OH f H,O 1 OH ! I NH-CO NH-CO eilen. NH-CHOH KH-CHOH 1 I CO CO i 1 f 2H20 oder CO ,<OH f H,O 1 OH ! I NH-CO NH-CO ertheilen. zuertheilen. centrirter Schwefelsaure leicht mit Harnstoff zu Harnsaure. Die Isodialursaure condensirt sich bei Gegenwart von con- Versuche , eine analoge Condensation mit Guanidin behufs Gewinnung einer Iminoharnsaure durchzufuhren, lieferten keine fassbaren Producte. Verwendet man aber statt der Schwefel- saure Eisessig, so erhalt man statt des gewunschten Conden- sationsproductes in reichlicher Menge Alloxantin. Die Bildung dieses Korpers erscheint auf den ersten Blick recht auffallig, da das Alloxantin nur durch einen Oxydationsvorgang ent- stehen kann, 2C,H4N,04 + 0 == C,H,K40, f 2H,O, die Isodialursaure aber im Gegensatze zu der ausserst leicht oxydirbaren Dialursaure in saurer Losung gegen Oxydations- mittel recht bestandig ist, wie schon aus ihrer Entstehung durch Oxydation der Isobarbitursaure vermittelst uberschussigen Brom- wassers hervorgeht. In der That verlauft der Vorgang auch etwas verwickelter; die Isodialursaure geht zunachst in Dialursaure iiber, welche dann in bekannter Weise durch den Luftsauerstoff zu Alloxantin oxydirt wird. Erhitzt man die Isodialursaure im Kohlensaure- strome mit Eisessig und Guanidin, so erhdt man anstatt des Alloxantins Dialursaure. Das wirksame Agens ist nicht der Eisessig, mit welchem die Isodialursaure ohne wesentlich ver- andert zu werden, langere Zeit erhitzt werden kann, sondern das Guanidin und zwar wirkt dasselbe als Base. B e h r e n d und R o o s e n 2, haben bereits beobachtet, dass die Isodialursanre zwar Kaliumacetat oder Carbonat nicht zer- setzt, mit freier Kalilauge aber ein Salz bildet, welches die K O ech, Ueber die Urnwandlzclzg der Isodialursazcre 248 grosste Aehnlichkeit mit dialursaureni Kali zeigt. Eine ge- nauere Untersuchung hat denn auch ergeben, dass dieses Salz thatsachlich dialursaures Kali ist. Dieselbe Umwandlung, welche in der Kiilte nur freie Alkalien bewirken, wird aber in der Hitze auch durch Acetate hervorgerufen, selbst bei Gegenwart von uberschussigem Eisessig. (Mitgetheilt von Robert Behreml.) Kaliuni- und Natriumacetat liefern dabei in glatter Reaction die dialursaureii Salze, Guanidinacetat giebt vermuthlich zunachst ebenfalls das Dialurat , doch kann man dasselbe nicht rein erhalten, da es schon wahrend der Reaction, vermuthlich in Folge theilweiser Dissociation zu Allo- santin oxydirt wird. Dialursaures Ammoniak kann man eben- falls auf diesem Wege nicht erhalten. Der ausserordentlich leicht und glatt verlaufende Ueber- gang der Isodialursiiure in Dialursaure SH-CHOH SH--CO I I I I ?O CO i CO CHOH NH-CO 1 1 I I NH 4 0 I uiiter dem Eiuflusse von Basen hat zunachst etwas Ueberraschendes. Berucksichtigt man aber, dass Ion den beiden Molekeln Wasser, welche die Isodialnrsaure enthalt, die eine schon bei l l O o , die zweite aber erst bei 140° unter theilweiser Zersetzung der Saure entweicht , also wohl als Constitutionswasser aufzufassen ist, so wird der Vorgang verstandlicher: I. 1 I. 111. SH-PHOH SH-C OH NH- CO CO C/OH = CO ( ' OH f H,O = CO CHOI-I f H,O. 'OH NH-CO ?u'€I-CO S H CO Die Isodialursaure wurde demnach zunachst unter Ab- spaltung von Wasser in eine Verbiiidung yon der Formel TI dbergehen. Die weitere Umlagerung wurde nach allen bis- herigen Erfahrungen nichts Auffalliges bieten. Nicht ausgeschlossen erscheint es ubrigens, dass die Dialur- saure gar nicht als Tartronylharnstoff aufzufassen ist , sondern Dialursawe. ommt. Dafur wurde die a sprechen, welche man e ilz Dialursawe. 249 dass ihr die Formel I1 zukommt. Dafur wurde die ausnahms- weise grosse Oxydirbarkeit sprechen, welche man eher von einem mehrwerthigen Phenole, als von einem Ureid der Tartron- saure erwarten konnte. Experimenteller Theil. 2 g feingepulverte Isodialursaure wurden mit 8 ccm Eis- essig iibergossen, l,2 g Guanidincarbonat (berechnet fur gleiclie Nolelieln 1 g) zagesetzt und mit Steigrohr im Wasserbade erhitzt. Nachdem vorubergehend vollige Losung eingetreten war, schieden sich nach einer bis zwei Minuten Krystalle aus, nach funf Minuten war bereits alles zu einem dicken Brei er- starrt. Nach 30 Minuten wurde mit 15 ccm Wasser verdunnt, nach dem Erkalten abgesaugt und mit Wasser, Alkohol und Aether gewaschen. Das Product stellte eine schwach rothlich gefarbte, krystallinische Masse dar , die Ausbeute betrug bei verschiedenen Versuchen 1,5-1,7 g. Der Korper konnte niclit in reinem Zustande erhalten werden. In heissem Wasser lost er sicli schwierig auf, beim Er- kalten scheidet sich niir eine feine Ilaut an1 Boden des Ge- fasses ab. Die Losung reagirt schwach sauer, giebt mit Baryt- wasser , Silbernitrat, sowie Eisenchlorur und Ammoniak die Alloxantinreactionen, und niit Pikrinsaure eine Fallung von Guanidinpikrat. Kalilauge liefert in der Kiilte geringe Mengen von Ammoniak. PrBparate verschiedener Darstellung enthielten 21,75-24,25 pC. Stickstoff, wahrend Alloxantin 17,40 pC., Dialursaure 1749 pC., dialursaures Guanidin 34,48 pC., dialur- saures Ammon 26,08 pC. verlangen. Wahrscheinlich liegt ein Gemisch von Alloxantin mit Dialursaure und dereu Salzen vor. In heisser verdunnter Salzsaure lost sich das Product bis auf geringe Mengen eines gelben Pulvers leicht auf, beim Er- kalten krystallisirt Alloxantin aus. Ausbeute etwa 30 pC. des Rohproductes. Zuweilen sind den kleinen derben Krystallchen des Alloxantins geringe Menge feiner Nadelchen , vielleicht Ko ech, Ueber die Umwawdlurzg der Isodialursiiure 250 Uramil, beigemengt. Das Alloxantin giebt die sammtlichen oben erwahnten, fur diesen Korper charakteristischen Reactionen. Die Analyse des lufttrocknen Praparates ergab folgende Zahlen : Zahlen : 1. 0,2237 g gaben 0,2418 CO, und 0,0705 H,O. 0,1766 g Druck. ,, 26,l ccm feuchtes Stickgas bei 11" und 759,7 mm 11. 0,1936 g gaben 0,2103 CO, iind 0,0604 H,O. 0,2161 g ,, 32,4 ccm feuchtes Stickgas bei 12,5" und 741 mm Druck. 11. 0,1936 g gaben 0,2103 CO, iind 0,0604 H,O. 0,2161 g ,, 32,4 ccm feuchtes Stickgas bei 12,5" und 741 mm Druck. Berechnet fur Gefunden - C,H,N,O, f 3 H,O I. n. C 29,82 29,47 29,62 H 3,10 3,50 3,47 N 17,40 17,60 17$4 Im Vacuum neben Schwefelsaure verlor das Alloxantin iiicht an Gewicht, bei 1200 entwichen ewei Mol. Wasser, wobei schwacbe Rothfarbung eintrat. Experimenteller Theil. Beim Erhitzen auf 140° fand keine weitere Gewichtsabnahrne statt, bei 170° entwich die dritte Molekel Wasser; doch liess sich bei dieser Temperatur nicht vollige Gewichtsconstanz erzielen, da bereits geringe Zer- setzung eintrat, welche oberhalb 170° schnell fortschritt. 0,7885 g vcrloren bei 120' 0,0913 g, entsprechend 11,68 pC. 0,7885 g ,, ,, 170O 0,1463 g, ,, 18,56 pC. 18,56 pC. g zwei Mol. H,O 11,18 pC., fur drei Mo!. H,O l6,77 pC. Aus der Formel CsH,N,O, + 3H,O berechnen sich fur zwei Mol. H,O 11,18 pC., fur drei Mo!. H,O l6,77 pC. Aus der Formel CsH,N,O, + 3H,O berechnen sich fur zwei Mol. H,O 11,18 pC., fur drei Mo!. H,O l6,77 pC. I. 0,1503 g gaben 0,1844 CO, und 0,0352 H,O. 0,3610 g ,, 58 ccm feuchtes Stickgas bei 17" m d 763 mm Drnck. I. 0,1503 g gaben 0,1844 CO, und 0,0352 H,O. 0,3610 g ,, 58 ccm feuchtes Stickgas bei 17" m d 763 mm Drnck. 11. 0,2185 g gaber? 0,2651 CO, und 0,0509 H,O. 0,1980 g ,, 33,l ccm feuclites Stickgas bei 1 7 O uiid 763 mm Druck. , g g , , , , 0,1980 g ,, 33,l ccm feuclites Stickgas bei 1 7 O uiid 763 mm Druck. 0,1980 g ,, 33,l ccm feuclites Stickgas bei 1 7 O uiid 763 mm Druck. Berechnet fur Gefunden - C,H,NIO, f H,O I. n. C 33,57 33,46 33,09 H 2,09 2,60 2,59 Is 19,58 19,21 19,49 251 in DialursMure. Ueber das Verhalten des Alloxantins beim Erhitzen finden sich Angaben, welche nicht ganz mit dem obigen Befunde uber- einstimmen. L i e b i g und Wohler3)sagen: AlloxantinCsH,,N,O,o verliert bei der Siedehitze der Schwefelsiiure 15,4 pC. Wasser, was ziemlich genau drei Atomen entspricht. R. Otto4) giebt an, dass gewassertes Alloxantin bei 150° bis 12,6 pC., nicht mehr bei 180° verliert und sich bei 200° zersetzt. E s wurde daher Alloxantin aus Harnsaure bereitet j das- selbe verhielt sich genau, wie das aus Isodialursaure gewonnene Product. 0,6821 g verloren bri 120' 0,0808 g, eiitvpreehend 11,86 pC. 0,6821 g ,, ,, 170' 0,1213 g, ,, 17,77 pC. 0,6821 g verloren bri 120' 0,0808 g, eiitvpreehend 11,86 pC. 0,6821 g ,, ,, 170' 0,1213 g, ,, 17,77 pC. Das bei 1 70° getrocknete Product nimmt beim Umkrystalli- siren aus Wasser nicht wieder Wasser auf, hat also bei der Entwiisserung eine tiefergehende Umwandlung erlitten. ") Gmelin, Handbuch Suppl. 1 , 2, 872. a) Diese Annalen 26, 282. Experimenteller Theil. Jeden- falls besitzt die eine Molekel Wasser eine andere Function als die beiden iibrigen und es ist daher das Alloxantin wohl richtig als ein Additionsproduct von Alloxan und Dialrirsaure zu betrachten und C,H,N,O, + 2H,O zu schreiben. Die krystallographische Untersuchung ergab die Identitat der verschiedenen Praparate j allerdings konnten wegen der geringen Dimensionen der Krystalle nur annahernde Messungen ausgefiihrt werden. Danach stellten die Krystalle rhombische Tafelchen dar, deren sanimtliche Flachen schiefwinklig waren. Der stumpfe Winkel betrag auf allen Flachen etwas mehr als looo, der spitze etwas weniger als 80°. Allc Flachen zeigten schiefe Ausloschung im parallelen linearpolarisirten Lichte. Bis- weilen fanden sich Zwillinge nach einer der schmalen Seiten- flachen. Eine rohe Bestimmung der Loslichkeit fiihrte ebenfalls bei den verschiedenen Praparaten zu dem gleichen Ergebniss. Von 1 g Alloxantin aus Harnsiiure, welches in 22 ccm heissen K O e c h , Ueber die L~wwia.ndlung der Isodialursiiure 252 mit zwei Tropfen Salzsaure angesauerten Wassers gelost waren, lwystallisirten bei Zimmertemperatur 0,82 g wieder Bus, VOR 1 g Alloxantin aus Isodialursaure 0,75 g. Dass bei der Einwirkung von Eisessig und Guanidin auf Isodialursaure nicht sofort Alloxantin, sondern zunachst Dialur- saure entsteht, lasst sich nachweissen, wenn man die Reaction bei Luftausschluss vor sich gehen lasst. 3 g feingepulverte Isodialursaure wurden mit 12 ccm Eis- essig und 1,8 g Guanidincarbonat Stunde im Wasserbade erhitzt und zugleich ein Stroni trockner Kohlensaure durch das Kolbchen geleitet. Die zu einem dicken Rrei erstarrte Masse wurde nun sofort rnit 20 ccm zehnprocentiger Salzsaure versetzt, im Kohlensaurestrome abgekiihlt, die ausgeschiedenen sehr feinen Krystallchen unter miigliclistem Luftausschluss abgesaugt, mit. ausgekochtem Wasser, Alliohol und Aether gewaschen und im Vacuum iiber Schwefelslure getrocknet. Die Ausbeute betrug 1,22 g ; entsprechend der grosseren Loslichkeit der Dialursaure war sie geringer, als wenn man bei Luftzntritt arbeitete, also statt Dialursaure Alloxantin erhielt. Die Mutterlaugen schieden allmahlich noch 0,98 g Alloxantin ab. Die so erhaltene Di- alarsdure gab mit Eiserichloriir, Silbernitrat und Barytwasser die Reactionen des Alloxantins, unterschied sich aber von diesem Korper dadurch, dass sie Alkalicarbonate unter starker Kohlen- saureentwickelung zerlegte , niit Kaliumacetat is der Kiilte die sehr schwerloslichen , lanzettformigen Krystallchen von Ealium- dialurat gab und im Exsiccator an Gewicht verlor. Von einer Analyse der Dialursaure als solcher wurde Abstand genommen, da die Zusammensctzung so nahe mit der des Alloxantins zu- sammenfallt, dass einc Unterscheidung nicht moglich ist. ') Diese dniialeii 127, 12. Experimenteller Theil. Da eine Analyse der Dialnrsaure iiberhaupt noch nicht veroffentlicht zu sein scheint, so wurde ein Praparat nach den dngaben yon B a e y c r fi) aus Harnsaure dargestellt, unter mog- lichstem Ausschluss von Luft isolirt , nach dem Suswaschen 253 ir, Dialzcrsazcre. init Alkohol und Aether eine Stunde im Vacuum uber Schwefel- saure getrocknet und in einer Kohlensaureatmosphare aufbewahrt. init Alkohol und Aether eine Stunde im Vacuum uber Schwefel- saure getrocknet und in einer Kohlensaureatmosphare aufbewahrt. init Alkohol und Aether eine Stunde im Vacuum uber Schwefel- saure getrocknet und in einer Kohlensaureatmosphare aufbewahrt. Die Analyse gab folgende Zahlen: 0,2219 g gaben 0,2379 CO, uud 0,0806 H,O. 0,2212 g 32,4 ccin fenchtes Stickgas bei 16,5O nnd 7%,5 inm 3,5957 g verloren im Vacnum uber Schwefelsiure in zwei Monaten 1,5318 g verloren im Kohlenslarestrome bei 120-12b0 0,1724 g, ,, Druck. 0,3902 g, entsprechend 11,27 pC. entsprechend 10,85 pC. Berechnet fur Gefundeii , C,H,N,O, f H,O C&N,O, + 2H20 (Alloxantin) C 29,63 29,82 29,24 H 3,70 3,lO 4,04 N 17,25 17,40 l6,99 H,O 11,11 11,18 (2H,O) 11,27 i0,85 Um also Dialursaure yon Alloxantin sicher zu unterscheiden, g g 0,2212 g 32,4 ccin fenchtes Stickgas bei 16,5O nnd 7%,5 inm ,, Druck. p p Berechnet fur Gefundeii , C,H,N,O, f H,O C&N,O, + 2H20 (Alloxantin) C 29,63 29,82 29,24 H 3,70 3,lO 4,04 N 17,25 17,40 l6,99 H,O 11,11 11,18 (2H,O) 11,27 i0,85 Um also Dialursaure yon Alloxantin sicher zu unterscheiden, Um also Dialursaure yon Alloxantin sicher zu unterscheiden, muss man sich der oben angefuhrten Reactiouen bedienen. muss man sich der oben angefuhrten Reactiouen bedienen. 1 g der aus Isodialursiiure erhaltenen Substanz wurde in schwach erwarmtem Wasser gelost und eine Losung von Kalium- hydroxyd in verdiinnter Essigsiiure zugesetzt. Alsbald schied sich dialursaures Kali in den charakteristischen Lauzetten ab. Die Menge betrug 0,98 g; sie ist also etwa gleich derjenigen der angewendeten Saure und gleich 88 pC. der berechneten 1,i2 g. 0 2402 g gaben 0 2224 CO und 0 0502 HO 0,2108 g ,, 29,9 ccm fenchtes Stickgns bei 16,5' nnd 733 mm Druck. ,2024 g gaben 0,0942 K,S04. Berechnet fur Gefunden C,€I,N,04K C 26,36 26,38 II 1,66 232 N 15,RS 16,35 I< 21,43 20,86 Das Salz loste sich leicht in verdunnter Kalilauge und wurde durch Essigsaure unverandert gefallt. Ganz ebenso ver- halt sich dialursaures Kali aus Harnsaure. Siehe weiter unten. 0,1725 g gabeu 0,1905 CO, und 0,0635 H,O. 0,1816 g 0,6167 g verloren bei 120" 0,0710 H,O. 0,6167 g ,, ., 170° 0,1155 H,O. ,, 27,8 ccm fenchtes Stickgas bei 15' und 757 mm Druck. Experimenteller Theil. 254 Koech , lrereber die Umwalzdlzcng der Isodialursaure Um iibrigens die Dialursaure von Alloxantin durch die Rildung des Kaliumdialurates zu unterscheiden , muss man in der Kaltc oder bei massiger Warme arbeiten und die Ausbeuten berucksichtigen. In der Siedehitze giebt namlich auch Allo- xantin mit Kaliumacetat dialursaures Kali, aber begreiflicherweise nur halb so vie1 wie Dialursaure. 3 g Alloxantin wurden in 75 ccm siedenden Wassers gelost nnd mit 1,2 g Aetzkali, in verdunnter Essigsaure gelost, versetzt. Beim Erkalten schieden sich 1,76 g Kaliumdialarat (= 95 pC. der berechneten 1,86 g) mit den charakteristischen Eigen- schaften ab. In der Losung konnten hochstens 0,2 g Kalium- salz geblieben sein. Siehe \veitcr unten. 3 g Alloxantin wurden in 75 ccm siedenden Wassers gelost nnd mit 1,2 g Aetzkali, in verdunnter Essigsaure gelost, versetzt. Beim Erkalten schieden sich 1,76 g Kaliumdialarat (= 95 pC. der berechneten 1,86 g) mit den charakteristischen Eigen- schaften ab. In der Losung konnten hochstens 0,2 g Kalium- salz geblieben sein. Siehe \veitcr unten. 0.2024 g guben 0,1956 CO, und 0,0443 H20. 0,2012 g ,, 29,l ccm feuclites Stickgas bei 15' und 7563 mm Druck. 0.2024 g guben 0,1956 CO, und 0,0443 H20. 0,2012 g ,, 29,l ccm feuclites Stickgas bei 15' und 7563 mm Druck. 0.2024 g guben 0,1956 CO, und 0,0443 H20. 0,2012 g ,, 29,l ccm feuclites Stickgas bei 15' und 7563 mm Druck. 0,2300 g gaben 0,1084 K,SO,. Berechnet fur Gefiinden C,TI,3?J,0,K (7 2636 26,35 H 1,65 2,43 N 15,3s 16,4S K 21,43 21,13 Bus den beschriebenen Versuchen geht hervor, dass das erste Product der Einwirkung von Eisessig und Guanidin auf Isodialursaure Dialursaure oder ein Salz derselben ist, welches weiter zu Alloxantin oxydirt wird. In der That wird Dialur- saure, wenn man sie genau wie die Isodialursaure mit Eisessig und Guanidin behandelt, in Alloxantin verwandelt. Aus 1 g Dialursaure wurden 0,77 g eines blaugriin ge- farbten Pulvers gewonnen. Dieses lieferte aus verdunnter Salz- sLure krystallisirt 0,62 g Alloxantin, welches Alkalicarbonate und Acetate in der Kalte nicht zersetzte. 0,1725 g gabeu 0,1905 CO, und 0,0635 H,O. 0,1816 g 0,6167 g verloren bei 120" 0,0710 H,O. 0,6167 g ,, ., 170° 0,1155 H,O. ,, 27,8 ccm fenchtes Stickgas bei 15' und 757 mm Druck. 255 ir, Dialursaure. 0,3142 g gaben 0,3106 CO, und 0,0616 H,O. 0,2447 g ,, 0,2059 g ,, 0,0992 K,SO,. 31,7 ccm feuchtes Stickgas bei 14" und 762 mm Druck. Experimenteller Theil. Berechnet fur Gefuiiden C,H,N;O, f 2 H,O C 29,82 30,12 N 17,40 l7,85 3R,O 16,77 18,73 H 3,lO 4,09 2H,O 11,18 11,52 Um zu erfahren ob der Eisessig bei der Umlagerung der Isodialursaure eine Rolle spiele, wurde 1 g der Saure mit 4 ccm Eisessig li2 Stunde im Wasserbade erwarmt. Beim Verdunnen mit Wasser blieben nur 0,05 g eines anscheinend amorphen Polvers ) welches die Alloxantinreactionen nicht gab, ungelost. Aus der stark eingeengten Losung wurden 0,54 g reine Iso- dialursaure in schiinen Krystallen zuriickgewonnen) die Mutter- lauge gab noch weitere Mengen. Der Eisessig ist mithin wirkungslos; wie in der Einleitung erwahnt, bewirkt vielmehr das Guanidin als Basis die Um- lagerung. Bereits in der Kalte geht dieselbe unter dem Ein- flusse freier Aetzalkalien vor sich. 3,33 g wasserhaltige Isodialursaure wurden in 148 ccm Wasser gelost und mit einer Losung von 1,8 g Aetzkali in 9 ccm Wasser versetzt. Sofort schieden sich die mikroskopischen, lan- zettformigen Krystalle des dialursauren Kaliums aus, die Losung wurde nun schwach mit Essigsaure angesauert, die Krystalle ab- gesaugt und mit Wasser bis zum Eintritt schwach alkalischer Reaction gewaschen. Ausbeute an Kaliumsalz 2,6 g statt 3,36 g. Zum Vergleich wurden 3 g Dialursaure (entsprechend 3,33 g Isodialursaure) in 148 g Wasser geliist und mit einer schwach mit Essigsaure angesauerten Losung von 1,s g Aetzkali in 9 ccrn Wasser versetzt. Erhalten wurden 2,76 g Kaliumsalz. Die Los- lichkeit der beiden Salze waren also gleich. Ebenzo zeigten die anderen Eigenschaften keine Verschiedenheit. Beide Salze losten sich leicht in iiberschussiger Kalilauge und wurden durch Essigsaure unverandert gefallt. Beim Auswaschen mit Wasser nahm dieses schaach alkalische Reaction an. Infolge dessen fand man den Kaliumgehalt des Salzes leicht zu niedrig. Koech, Ueber die Urnwalzdlung der Isodialursuzcre 256 Die Krystlllchen stellten vierseitige Prismen mit meist ge- krummten Flachen dar , wodurch der lanzettformige Habitus entsteht. Im parallelcn, linearpolarisirten Lichte zeigte sich auf allen Flachen Ausloschung parallel und senkrecht zu den Begrenzungslinien. I. Analyse des Salzes aus Isodialursaure. 0,2799 g gaben 0,2691 CO, und 0,0471 H,O. 0,2616 g ,, Druck. 0,5079 g gaben 0,2431 K,SO,. 34,3 ccin feuchtes Stickgas bei 14O und 767 mm 0,2616 g ,, Druck. 34,3 ccin feuchtes Stickgas bei 14O und 767 mm 11. Analyse des Salzes aus DialursSinre. 11. Analyse des Salzes aus DialursSinre. 0,2255 g gaben 0,2276 CO, und 0,0443 H,O. 0,2946 g Druck. ,, 393 ccm feuchtes Stickgm bei 14" und 768 mm 11. Experimenteller Theil. Analyse des Salzes aus DialursSinre. 0,2255 g gaben 0,2276 CO, und 0,0443 H,O. 0,2946 g Druck. ,, 393 ccm feuchtes Stickgm bei 14" und 768 mm 0,5006 g gaben 0,2331 K,SO,. Bereclinet fur Gefunden /- C,H,N,O4K I. 11. c; 26,36 26,22 27,52 H 1,65 1:87 2,18 N 15,38 16,58 1.5,82 K 21,43 2146 20,87 Nachdem so nachgewiesen war, dass Kalilauge die Iso- dialursaure in Dialursaure verwandelt, andererseits durch einen besonderen Versuch die Angabe von B e h r e n d und R o o s e n bestatigt war, dass Isodialursaure durch Kaliumacetat in der Kalte nicht verandert werde, blieb noch zu prufen, ob die Um- Iagerung durch Kaliurnacetat in der Hitze erfolgt. Dies ist in der That der Fall. 1 g Isodialursaure wurde mit 4 ccm Eisessig, in welchem 0,6 g Kaliumacetat gelost waren, l/z Stunde im Wasserbade er- hitzt. Die nach vorubergehender Losung zum Brei erstarrte Masse wurde mit 15 ccm Wasser verdunnt und die Krystalle isolirt. Das Product (0,9 g) zeigte alle Eigenschaften des Kaliumdialurates. 257 in. Dialursazcre. Berechnet Gefunden C 26,36 26,96 H 1,65 2,18 N 15,38 1539 K 21,43 21,60 Ganz analog dem Kaliumacetat wirkt auch Natriumacetat. Ganz analog dem Kaliumacetat wirkt auch Natriumacetat. Aus 2 g Isodialursaure wurden 1,3 g dialursaures Natrium er- halten. Das Salz verliert bereits uber Schwefelsaure Wasser, analysirt wurde das lufttrockne Salz. 0,2261 g gaben 0,2230 C 0 2 und 0,0611 H,O. 0,2103 g ,, 0,2138 g gaben 0,0843 Na,SO,. 0,2389 g verloren bei 130" 0,0118 H20. 0,2389 g ,, ,, 160' 0,0250 H,O. 29,75 ccm feuchtes Stickgas bei 15" nnd 758,5 Druck. Bereclinet fur Gefunden C4H,N,0,Na + HSO C 26,08 26,89 €I 2,72 3,OO N 15,22 16,54 Na 12,50 12,77 'i2H20 4,89 4,94 (bei 130') 1H20 9,78 10,46 (bei 160') 0,2261 g gaben 0,2230 C 0 2 und 0,0611 H,O. 0,2103 g ,, 29,75 ccm feuchtes Stickgas bei 15" nnd 758,5 mm Druck. Bereclinet fur Gefunden C4H,N,0,Na + HSO C 26,08 26,89 €I 2,72 3,OO N 15,22 16,54 Na 12,50 12,77 'i2H20 4,89 4,94 (bei 130') 1H20 9,78 10,46 (bei 160') Die Versuche, auf gleichem Wege dialursaures Ammonium darzustellen, schlugen fehl; auch wenn der Versuch unter Luft- abschluss ausgefuhrt wurde, bildete sich ein Korper von den Eigenschaften des Murexids. ') Diese Annalen 182, 70. Annalen der Cbemie 315. Bd. Experimenteller Theil. Stets wurde nur das Salz C711sKa2N401,, gebildct, dessen Umwaiitllung in C,II,NaN,O, trotz vielfacher Versuche nicht gelang. Gegenwart von Lllausiiurc. Stets wurde nur das Salz C711sKa2N401,, gebildct, dessen Umwaiitllung in C,II,NaN,O, trotz vielfacher Versuche nicht gelang. Wir habeii, wie oben angcgcbci~ , aus Isodialursiiure und Natriumacctat nur das normalc Salz C4H8KaXs0,1 erhaltcn. ljh scheint daiiach fast, als ob die nach dcn von Men- s c h u t k i n angeweiidetcn BIetlioclen dargestclltcn Sake gar nicht mit den aus freier Dialursiiure bczw. Isodialursiiure gebildeten identisch sind und deninacli gar nicht die cigcntliclien Salze diescr Siiure darstellen, Jedenfalls verdieiieii diese Verhiiltiiisse und namentlich aucli die merkwiirdigen, von Xcii schutkiii bcobachteten Umwand- lungeii dcr Salzc cine geiiauere Untcrsuchung. I) Sielie die vorangehonde Xittlieil~ulg. roil Ciirt c. Vogel. roil Ciirt c. Vogel. Experimenteller Theil. Ein auffallender Mange1 an Uebereinstimmung unserer Beobachtungen an den dialursauren Salzen mit den Angaben N e n s c h u t k i n ' ~ ~ ) muss hier noch hervorgehoben werden. M e n s c h u t k i n findet, dass die dialursauren Salze in zwei ver- schiedenen Formen existiren von den allgemeinen Formeln C,H,Me,N,O,, und C,H,MeN,O,, von denen die ersteren die bestandigeren sind. Das Ammoniumsalz, C,H,(NH4),N40io , wurde von M en- s c h u t k i n nach L i e b i g und W o h l e r aus Alloxan durch Re- 17 KO ech, Ueber die LTmwandlung der Isodialursuure etc. 258 duction mit Schwefelwasserstofl und Schwefelammonium bezw. Schwefelwasserstoff und Ammoniumcarbonat, sowie nach S tr e c k e r aus Alloxan durch Einwirkung yon Ammoniak bei Gegenwrt von Blausaure erhalten. Durch wiederholte Krystallisation aus sieden- dem Wasser bei Gegenwart von Ammoniumcarbonat ging es in das Salz C,H,(RII,)N,O, iiber , welches seinerseits durch Kry- stallisation aus heissem, reinen Wasser sofort wieder in das erstere Salz verwandelt wnrde. AUS den beiden Amnioniumsalzen erhielt M e n s c h u t k i n die entsprechenden Kalinmsalze durch Unisetzung mit Kalium- acetat. Die wechselseitige Umwandlung der beiden Sake in- einander verlief unter denselben Bedingnngen , wie diejenigeii der Ammoniumsalze. Bei unseren T’ersnchen haben wir nur das Salz C,H,KN,O, beobachtet. Der Versuch, dasselbe in das Salz C,HsK,N,O,, umzuwandeln, schlug fehl. 3,06 g dialursaures Kali aus der freien Saure und Kalinni- acetat warden in 250 ccm siedenden Wassers geliist, wobei noch eiii geringer Ruckstand blieb. Kach 12 stundigem Stehen der erkalteten Liisung hatten sich 1,45 g Krystalle \on den Eigenschaften des Kaliumdialurates abgeschieden. 0,2210 g gaben 0,2107 CO, m d 0,0427 H,O. 0,1893 g ,, 0,0890 B,HO,. iierechnet fur Befundrn C4H,I<N,0, C,H,K,N40,, ( I %,36 21.75 2539 H 1,63 2,07 2,14 li 21,43 20,24 21,05 Die Mutterlmgen gaben beini Einengen im Vacuum tiber Schwefelsaure eine klare, zahe Masse , aus der liein weiteres krystallinisches S a l ~ gewonnen werden konnte. Das dialursaure Xatrium hat M e n s c h u t k i n nach ver- schiedenen Methoden dargestellt : 1) aus dem Ammoniumsalz und Natriumacetat, 2 ) aus Alloxantin, Schwefelwasserstoff und Xatriumcarbonat , 3) aus Alloxan uiid Natriumcarbonat bei 259 u. Tog e l , Ueber die Codensation con Isoclialursiiure etc. Gegenwart von Lllausiiurc. (Mitgethcilt von Robert Helwend.) I’ a ~i 1 K o e c h l) hat gefundeii, dass Guanidin sich nicht gleich dem IIarnstoff mit Isodialursiiure bci Gegcnvart von concentrirter Schwefelsaure condciisircn 18sst. Ebcnsowenig giebt Tliioharnstoff unter tliescn Cnistanden ein fassbares Con- densationsproduct. Dagegen erhalt man ein solches, uenn man Eiscssig als Condensationsmittel bcniitzt. Das Product besitzt die complicirtc Zusammcnsctzung C,,II,6Sl,0,LSa ; cine eiufache Entstehuiigsglcichung lasst sich nicht aufstrllcn ; aucli war es bishcr niclit nibglich , die Constitution der Verbindung zu er- mitteln. Beim Iiochen niit Xatronlauge spaltet tier Iiorper allcn Schwefel und ciiicn Theil dcs Stickstofk ab und licfcrt cin in Wasser fast unliisliches Natriuinsalz 7 on der Zusainmensctzung 1 7 %
https://openalex.org/W3201329318	https://ieeexplore.ieee.org/ielx7/16/9583611/09532000.pdf	English	null	Technology of GaN-Based Large Area CAVETs With Co-Integrated HEMTs	I.E.E.E. transactions on electron devices/IEEE transactions on electron devices	2,021	cc-by	6,765	Technology of GaN-Based Large Area CAVETs With Co-Integrated HEMTs Philipp Döring , Rachid Driad, Richard Reiner , Patrick Waltereit, Stefano Leone , Michael Mikulla, and Oliver Ambacher devices offer better area efﬁciency in the high voltage region, as the drift region can be scaled independently of the chip area. In contrast to other reported vertical GaN transistor designs [5]–[9], the current aperture vertical electron transistor (CAVET) combines the advantage of the vertical depletion- and drift-region with the proven gate–source module of the lateral HEMT. This combination opens the possibility for monolithic integration of gate drivers as well as protection and sensor systems [10] of the low-voltage HEMT-technology and simultaneously avoids the requirement of an intricate ﬁeld-plate design since the high electric ﬁeld is moved into the bulk material. In addition, the current blocking layer (CBL), which serves as a potential barrier between the source and drain in CAVET technology, can be used to shield the low-voltage logic of the HEMT-technology from the high drain/substrate potential of the vertical transistor, when con- nected to the source contact. Even though CAVETs were demonstrated by various fabrication strategies, the intrinsic device design, the challenges in the fabrication process, and their impact on the performance of the fabricated devices are rarely explored [11]–[20]. The use of Mg-implantation to form the CBL seems to be an advantageous process to allow for the direct implementation of the HEMT surface process technology. In contrast to CAVETs demonstrated by Shibata et al. [11] or Nie et al. [12], trench etching in the active area can be avoided and a planar regrowth can be used to ﬁnalize the CAVET structure, as the Mg-implantation process leaves the surface morphology unaffected. Promising results of CAVETs were presented [17] with Mg-implanted CBL’s fab- ricated by molecular beam epitaxy (MBE) regrowth to prevent Mg-diffusion from the CBL, which compensates the 2-D elec- tron gas (2DEG) at the AlGaN/GaN interface [14]. However, these demonstrated devices were limited to the mA-range and large periphery devices could not be demonstrated. We have recently presented the ﬁrst results on large area CAVETs completely grown by metalorganic chemical vapor deposition (MOCVD) [21], which circumvents the need for an indus- trially unsuitable and expensive MBE regrowth step [17]. In this work, recent progress on the technology of large area CAVETs is presented with lateral HEMTs co-integrated on-chip fabricated by Mg-implantation and MOCVD regrowth under standard growth conditions. 5547 IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 68, NO. 11, NOVEMBER 2021 Manuscript received May 3, 2021; revised June 22, 2021 and August 18, 2021; accepted August 31, 2021. Date of publication Sep- tember 9, 2021; date of current version October 22, 2021. This work was supported in part by the Deutsche Forschungsgemeinschaft (DFG)under Grant AM 105/47-1 and in part by Fraunhofer Society under VERTIGO Grant. The review of this article was arranged by Editor S. Chowdhury. (Corresponding author: Philipp Döring.) Philipp Döring is with the Department of Power Electronics, Albert-Ludwigs-Universität Freiburg, 79108 Freiburg im Breisgau, Germany (e-mail: philipp.doering@inatech.uni-freiburg.de). Rachid Driad, Richard Reiner, Patrick Waltereit, Stefano Leone, and Michael Mikulla are with Fraunhofer Institute for Applied Solid State Physics (Fraunhofer IAF), 79108 Freiburg im Breisgau, Germany. Oliver Ambacher is with the Department of Power Electronics, Albert-Ludwigs-Universität Freiburg, 79108 Freiburg im Breisgau, Germany, and also with Fraunhofer Institute for Applied Solid State Physics (Fraunhofer IAF), 79108 Freiburg im Breisgau, Germany. Color versions of one or more ﬁgures in this article are available at https://doi.org/10.1109/TED.2021.3109840. Digital Object Identiﬁer 10 1109/TED 2021 3109840 Technology of GaN-Based Large Area CAVETs With Co-Integrated HEMTs An analysis of the intrinsic CAVET design and its impact on the ON-state and transfer characteristics of the fabricated devices is given. An optimized Abstract— In this work, multi-ﬁnger current aperture vertical electron transistors (CAVETs) are fabricated with co-integrated high electron mobility transistors (HEMTs). The devices are realized by Mg-ion implantation and met- alorganic chemical vapor deposition (MOCVD) regrowth. The intrinsic CAVET design is optimized for robust device performance and applied on multi-ﬁnger devices having a total gate periphery of WG = 13.5 mm and WG = 77 mm. Mappings of the transfer characteristics revealed reliable turn-off behavior demonstrating the suitability of the intrin- sic device layout. The largest CAVETs revealed a total ON-state resistance of RON = 1.67 and a maximum drain current of ID,MAX = 20.3 A at VGS = 3 V. A pulse robustness of PPULS = 976 W at VDS = 50 V and a pulsewidth of 500 μs is shown without thermal destruction. Additionally, HEMTs are co-integrated on-chip. This combination of HEMTs and reliable large area CAVETs enables the design of high- performance, monolithically integrated GaN power circuits (GaN power ICs) based on the CAVET technology. Index Terms— Current aperture vertical electron transis- tor (CAVET), GaN, power electronics, vertical transistor. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ Rachid Driad, Richard Reiner, Patrick Waltereit, Stefano Leone, and Michael Mikulla are with Fraunhofer Institute for Applied Solid State Physics (Fraunhofer IAF), 79108 Freiburg im Breisgau, Germany. s work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativec Philipp Döring is with the Department of Power Electronics, Albert-Ludwigs-Universität Freiburg, 79108 Freiburg im Breisgau, Germany (e-mail: philipp.doering@inatech.uni-freiburg.de). p g Digital Object Identiﬁer 10.1109/TED.2021.3109840 I. INTRODUCTION Measured data of three single transistors (WG = 100 µm) are given in green. Fig. 2. TCAD simulations of the area-speciﬁc ON-state resistance RON × A versus the aperture length LAP in dependence of the donor concentration in the aperture region. A free donor concentration of ND = 2 × 1016 cm−3 is assumed in the drift region. Measured data of three single transistors (WG = 100 µm) are given in green. Fig. 1. Schematic overview of the CAVET fabrication process and the ﬁnal device layout. (a) CBL-formation by Mg-ion implantation. (b) MOCVD AlGaN/GaN regrowth. (c) Metallization and passivation. (d) Deﬁnition of intrinsic design parameters. (e) On-chip HEMTs with CBL in vicinity to the AlGaN-barrier. (f) CAVET schematic. Fig. 2. TCAD simulations of the area-speciﬁc ON-state resistance RON × A versus the aperture length LAP in dependence of the donor concentration in the aperture region. A free donor concentration of ND = 2 × 1016 cm−3 is assumed in the drift region. Measured data of three single transistors (WG = 100 µm) are given in green. revealed a root mean square of rms = 0.19 nm (10 × 10 μm2 scan). The device processing [Fig. 1(c)] is based on standard III–V technology [3]. Alignment markers were etched and protected by SiOX prior to regrowth to orientate the surface processing with respect to the aperture region. A Ti/Al/Ni/Au metal stack is deposited and alloyed at around 825 ◦C in N2 ambient to form the ohmic contacts. Afterward, the devices are passivated with a SiXNY dielectric deposited by plasma-enhanced chemical vapor deposition (PECVD) and then isolated by Ar-implantation. Afterward, the pas- sivation layers are opened by inductively coupled plasma etching (ICP-RIE) and Ni-Au-based Schottky gate contacts are deposited. The surface process is ﬁnalized by an additional SiXNY passivation layer and a TiPtAu-based interconnection metallization. The drain contact in the vertical devices is realized on the backside of the GaN substrate by the previously mentioned alloyed ohmic and interconnection metallization. Fig. 1. Schematic overview of the CAVET fabrication process and the ﬁnal device layout. (a) CBL-formation by Mg-ion implantation. (b) MOCVD AlGaN/GaN regrowth. (c) Metallization and passivation. (d) Deﬁnition of intrinsic design parameters. (e) On-chip HEMTs with CBL in vicinity to the AlGaN-barrier. (f) CAVET schematic. III. RESULTS AND DISCUSSION The initial testing of the fabricated CAVETs comprises the characterization of standard Hall- and transfer length method (TLM)-structures. Hall measurements on isolated areas reveal sheet carrier density of nS = 8.7 × 1012 cm−2, electron mobility of μ = 1320 cm2/V·s, and sheet resistance of RS = 538 /□. Contact resistance mapping by TLM reveals an RC = 0.3 – 0.35 ·mm over the complete wafer, which is slightly above literature values [20]. One of the critical design parameters of the CAVET is the aperture length LAP [Fig. 1(d)] [15], [21], [22]. To avoid current choking, the aperture resistivity should be low, which can be achieved either by intentional n-type doping of the aperture region or by increasing the aperture length. In order to investigate the impact of the aperture resistivity on the total area-speciﬁc ON-state resistance RON×A, TCAD simulations are carried out using a self-consistent Poisson-solver (Silvaco Atlas) (Fig. 2). It is visible, that for small LAP, the RON × A is signiﬁcantly increasing because of the dominating RAP, particularly for low donor concentrations in the aperture. However, the impact I. INTRODUCTION G G aN-BASED devices offer great potential for high power switching applications due to their superior physical properties, compared to Si and SiC [1], [2]. The most investi- gated AlGaN/GaN high electron mobility transistor (HEMT) has already entered the high-power and high-frequency elec- tronics market [3], [4]. However, for high power applications, the lateral technology suffers from high gate-to-drain spac- ing to sustain high voltage operation. In contrast, vertical Manuscript received May 3, 2021; revised June 22, 2021 and August 18, 2021; accepted August 31, 2021. Date of publication Sep- tember 9, 2021; date of current version October 22, 2021. This work was supported in part by the Deutsche Forschungsgemeinschaft (DFG)under Grant AM 105/47-1 and in part by Fraunhofer Society under VERTIGO Grant. The review of this article was arranged by Editor S. Chowdhury. (Corresponding author: Philipp Döring.) p g Digital Object Identiﬁer 10.1109/TED.2021.3109840 Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4 IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 68, NO. 11, NOVEMBER 2021 5548 Fig. 1. Schematic overview of the CAVET fabrication process and the ﬁnal device layout. (a) CBL-formation by Mg-ion implantation. (b) MOCVD AlGaN/GaN regrowth. (c) Metallization and passivation. (d) Deﬁnition of intrinsic design parameters. (e) On-chip HEMTs with CBL in vicinity to the AlGaN-barrier. (f) CAVET schematic. Fig. 2. TCAD simulations of the area-speciﬁc ON-state resistance RON × A versus the aperture length LAP in dependence of the donor concentration in the aperture region. A free donor concentration of ND = 2 × 1016 cm−3 is assumed in the drift region. Measured data of three single transistors (WG = 100 µm) are given in green. revealed a root mean square of rms = 0.19 nm (10 × 10 μm2 scan). The device processing [Fig. 1(c)] is based on standard III–V technology [3]. Alignment markers were etched and protected by SiOX prior to regrowth to orientate the surface processing with respect to the aperture region. A Ti/Al/Ni/Au metal stack is deposited and alloyed at around 825 ◦C in N2 ambient to form the ohmic contacts. Afterward, the devices are passivated with a SiXNY dielectric deposited Fig. 2. TCAD simulations of the area-speciﬁc ON-state resistance RON × A versus the aperture length LAP in dependence of the donor concentration in the aperture region. A free donor concentration of ND = 2 × 1016 cm−3 is assumed in the drift region. I. INTRODUCTION device layout is derived and applied on large area multi-ﬁnger CAVETs with a chip size of 1 × 1 mm2 and 2 × 2 mm2, and their ON-state, transfer, OFF-state characteristics are analyzed as well as their pulse power stability. Vertical and lateral transistors are realized on-chip to compare the vertical device performance directly as a function of the total gate width and to demonstrate the possibility of a direct integration of the AlGaN/GaN-HEMT. The combination of CAVETs and on-chip integrated HEMTs enables the introduction of several new design opportunities in the AlGaN/GaN-technology for high-functional, monolithic integrated GaN power ICs. II. EXPERIMENTAL SECTION Transfer characteristics at VDS = 10 V of small gate width CAVETs (WG = 100 µm) with a gate aperture overlap of LGAP = 2 µm and varying aperture length (black). In addition, the gate aperture overlap of CAVETs with LAP = 10 µm is reduced to LGAP = 1 µm (red) and LGAP = 0 µm (blue), showing increasing loss of the gate control. The measured data reveal higher resistances as expected from the TCAD simulations. It is assumed that the effec- tive donor concentration in the aperture and drift layer is lower due to slight carrier compensation by carbon, which causes a slightly higher resistance. Carbon is a deep acceptor that may be introduced by auto-doping during the epitaxial growth, because of the carbon atoms present in the MOCVD precursors. Another potential contributor could be related to surface contamination during the aperture deﬁnition process. Nevertheless, the measured RON × A for CAVETs with an aperture length of LAP = 10 μm is close to the expected values as the active device area becomes larger and the impact of the low carrier concentration is decreasing. A second critical parameter in the design of the CAVETs is the gate aperture overlap LGAP [13], [17] shown in Fig. 1(d). The gate aperture overlap can be deﬁned as the difference between the length of the gate LG and the aperture length LAP. LGAP determines the gate control by suppressing source-drain leakage through the GaN-channel below the gate. The required overlap needs to be adjusted with respect to the channel thickness between the CBL and source contact in the fabricated device structure. To determine the required dimension of the gate-aperture overlap, small gate width CAVETs were fabricated with LGAP ranging between 0 and 2 μm and a gate width of WG = 100 μm. The corresponding transfer characteristics at VDS = 10 V and varying VGS = −5 to 1 V are shown in Fig. 3. The device with an aperture length of LAP = 10 μm and LGAP = 2 μm presented a threshold voltage of VTH = −2.82 V. Similar VTH is obtained for on-chip lateral HEMTs test structures. The device with an aperture length of LAP = 10 μm and LGAP = 2 μm presented a threshold voltage of VTH = −2.82 V. Similar VTH is obtained for on-chip lateral HEMTs test structures. II. EXPERIMENTAL SECTION The fabrication of the devices started with the growth of a 2 μm n−-GaN layer by MOCVD on 2-in bulk n+-GaN substrates. The HVPE grown substrate has a resistivity of ρ < 0.5 ·cm and a dislocation density of DD < 5 × 106 cm−2. The CBL was formed by Mg-ion implantation [Fig. 1(a)] with an energy of 100 keV and a dose of 2 × 1014 cm−2, which gives a peak Mg-concentration of 1 × 1019 cm−3. This Mg-concentration allows for the regrowth of the following epitaxial layers under standard MOCVD growth conditions [22]. The aperture region was protected by a ∼3-μm-thick photoresist during implantation. After the pho- toresist was removed, the CAVET was ﬁnalized by MOCVD regrowth of a 250 nm uid-GaN channel, a 24 nm Al0.24Ga0.76N barrier, and a 4 nm GaN-cap layer [Fig. 1(b)]. The regrowth surface temperature of about 1050 ◦C also served as in situ annealing for the Mg-implanted CBL without the require- ment of a capping layer. Atomic force microscopy (AFM) DÖRING et al.: TECHNOLOGY OF GaN-BASED LARGE AREA CAVETs WITH CO-INTEGRATED HEMTs 5549 Fig. 3. Transfer characteristics at VDS = 10 V of small gate width CAVETs (WG = 100 µm) with a gate aperture overlap of LGAP = 2 µm and varying aperture length (black). In addition, the gate aperture overlap of CAVETs with LAP = 10 µm is reduced to LGAP = 1 µm (red) and LGAP = 0 µm (blue), showing increasing loss of the gate control. of RAP successively vanishes with increasing LAP, almost independent of the free donor concentration in the aperture. Furthermore, the lowest minimum RON×A is shifting to higher LAP with decreasing ND. Thus, even though low RON × A is achievable with small LAP, higher aperture length (>4 μm) still seems beneﬁcial when addressing high voltage stability due to the higher critical ﬁelds at low donor concentrations. However, for high aperture length (>10 μm), the lateral spreading of the transistor increases the total RON × A which makes very large apertures unsuitable. Small gate width tran- sistors (WG = 100 μm) are fabricated on-chip to measure RON × A empirically. The length of the active area LACT is deﬁned between the centers of the two source contacts of each transistor (gate–source distance of LGS = 4 μm and length of the ohmic contacts of LOHM = 5 μm). Fig. 3. II. EXPERIMENTAL SECTION Thus, the turn-off behavior is determined by the depletion of the 2DEG and unmodulated electrons in the GaN-channel seem to play a minor role in the OFF-state at VDS = 10 V. is visible, as the sub-threshold swing increases drastically to SS-TH = 404.1 mV/decade and the threshold voltage shifts to VTH = −3.79 V. Devices without any LGAP did not reveal any turn-off behavior and the gate control is completely lost. To investigate the suitability of the intrinsic device layout on the large area cell topology, two different dimensioned comb structures are fabricated with a chip size of 1 × 1 mm2 and 2 × 2 mm2. The two different multi-ﬁnger transistors feature an aperture length of LAP = 10 μm, a gate-ﬁnger length of LG = 14 μm, and a gate-aperture overlap of LGAP = 2 μm. The 1 × 1 mm2-chip device has a total gate width of WTOTAL = 26 × 0.53 mm = 13.78 mm (number of ﬁngers × ﬁnger width = N × WG) and an active ﬁnger area of AACT = 0.53 mm × 0.85 mm (WG × LACT). The 2 × 2 mm2-chip device features a total gate periphery of WTOTAL = 58 × 1.32 mm = 77 mm and an active ﬁnger area of AACT = 1.32 mm × 1.85 mm. Both comb structures are realized on the same chip with the previously shown small transistors (WG = 100 μm) to allow a direct comparison. Transfer characteristics of both large-area devices are mapped over the complete wafer (Fig. 4). A threshold volt- age shift due to Mg-diffusion, observed for MOCVD-regrown CAVETs [14], is not observable, demonstrating the suitability of our MOCVD-based regrowth process even for much larger devices. In order to analyze the OFF-state behavior of the large area CAVETs, IV-characteristics at VGS = −5 V are measured until device breakdown (Fig. 5). The small gate width transistors reveal OFF-state drain currents in the range of ID,OFF = 8 × 10−6 A/cm2 below VDS = 200 V. Breakdown is measured in the range of VBR = 206 −222 V (at room temperature) independent of the aperture length and also for the transistor without an aperture. Thus, it is assumed that breakdown is caused by electrons punch through the CBL. IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 68, NO. 11, NOVEMBER 2021 The devices feature an aperture length of LAP = 2 µm, a gate aperture overlap of LGAP = 2 µm. Transfer curves are shown for 1 × 1 mm2-chip devices (red) and 2 × 2 mm2-chip devices (black) having a total gate width of WTOTAL = 13.5 mm and WTOTAL = 77 mm, respectively. The current compliance was set to 1 A. of small gate width CAVETs in [17] (1.72 MV/cm, WG = 75 μm). The critical ﬁeld strength of the 1 × 1 mm2 (ECRIT = 2.3 MV/cm) is even exceeding these values, which demon- strates the suitability of the present in situ annealing/regrowth process for large multi-ﬁnger CAVETs as reported in this work. However, a higher breakdown voltage can be expected by increasing the drift layer thickness, as demonstrated in other vertical GaN-designs [5]–[9], [11], [12], in view to increase the area-efﬁciency with respect to lateral GaN-based power transistors. Fig. 5. OFF-state characteristics at VGS = −5 V of the large area devices with a chip size of 1 × 1 mm2 (red), 2 × 2 mm2 (blue). Small gate width CAVETs on the same chip with different aperture lengths are given in black. In addition, a CAVET without an aperture was measured (orange) to force the current through the CBL. To demonstrate the behavior of the devices in terms of current handling capability and pulse power stability, pulsed output characteristics of the fabricated multi-ﬁnger CAVETs are measured (Fig. 6). A pulsewidth of 100 μs and a sample time of 50 μs is used with a duty cycle far below 1%. The CAVET with a chip size of 1 × 1 mm2 revealed a maximum absolute drain current of ID,MAX = 4.97 A (normalized to the active area ID,MAX = 1.08 kA/cm2) and a total differ- ential ON-state resistance of RON = 3.56 (RON × A = 16.2 m/cm2) at VGS = 1 V. The knee-voltage is found at VK = 25 V, which corresponds with a pulse power of PPULS = 121 W or a pulse power density of 26.86 kW/cm2. The larger multi-ﬁnger CAVET with a chip-size of 2 × 2 mm2 revealed a maximum absolute drain current of ID,MAX = 14.76 A and a total differential ON-state resistance of RON = 1.81 (RON × A = 40.9 m/cm2) at VGS = 1 V. IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 68, NO. 11, NOVEMBER 2021 The CAVET with a chip size of 1 × 1 mm2 revealed a maximum absolute drain current of ID,MAX = 4.97 A (normalized to the active area ID,MAX = 1.08 kA/cm2) and a total differ- ential ON-state resistance of RON = 3.56 (RON × A = 16.2 m/cm2) at VGS = 1 V. The knee-voltage is found at VK = 25 V, which corresponds with a pulse power of PPULS = 121 W or a pulse power density of 26.86 kW/cm2. The larger multi-ﬁnger CAVET with a chip-size of 2 × 2 mm2 revealed a maximum absolute drain current of ID,MAX = 14.76 A and a total differential ON-state resistance of RON = 1.81 (RON × A = 40.9 m/cm2) at VGS = 1 V. The Fig. 4. Mapping of the transfer characteristics at VDS = 10 V of large area CAVETs. The devices feature an aperture length of LAP = 2 µm, a gate aperture overlap of LGAP = 2 µm. Transfer curves are shown for 1 × 1 mm2-chip devices (red) and 2 × 2 mm2-chip devices (black) having a total gate width of WTOTAL = 13 5 mm and WTOTAL = 77 mm Fig. 6. Pulsed output characteristics of the multi-ﬁnger CAVETs with a chip size of 1 × 1 (blue) and 2 × 2 mm2 (black). A pulsewidth of 100 µs and a sample time after 50 µs are used, with a duty cycle below 1%. Fig. 6. Pulsed output characteristics of the multi-ﬁnger CAVETs with a chip size of 1 × 1 (blue) and 2 × 2 mm2 (black). A pulsewidth of 100 µs and a sample time after 50 µs are used, with a duty cycle below 1%. Fig. 6. Pulsed output characteristics of the multi-ﬁnger CAVETs with a chip size of 1 × 1 (blue) and 2 × 2 mm2 (black). A pulsewidth of 100 µs and a sample time after 50 µs are used, with a duty cycle below 1%. Fig. 6. Pulsed output characteristics of the multi-ﬁnger CAVETs with a chip size of 1 × 1 (blue) and 2 × 2 mm2 (black). A pulsewidth of 100 µs and a sample time after 50 µs are used, with a duty cycle below 1%. Fig. 4. Mapping of the transfer characteristics at VDS = 10 V of large area CAVETs. II. EXPERIMENTAL SECTION Besides, a sub-threshold slope of SS-TH = 87.6 mV/decade and an ION/IOFF ratio of 1.18 × 105 are derived from the transfer characteristics. A signiﬁcant change is not observable for CAVETs with the same gate aperture overlap but different aperture lengths between 2 and 30 μm. Thus, the ratio of the gate aperture overlap and the aperture was found to be not relevant. However, when decreasing the gate-aperture overlap to 1 μm, a signiﬁcant reduction of the gate control Further improvement is expected by connecting the CBL with the source contact. In this case, the formation of a body diode improves reverse leakage behavior and self-biasing can be avoided by keeping the CBL at the source potential. The breakdown voltage of the 1 × 1 mm2 device is found within IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 68, NO. 11, NOVEMBER 2021 The maximum drain current normalized to the active device area is reduced to ID,MAX = 0.61 kA/cm2, when compared to the smaller multi-ﬁnger CAVET. The knee voltage is found at VK = 34 V, corresponding to a total pulse power of PPULS = 501.8 W (20.4 kW/cm2). As Hall- and TLM-measurements revealed typical values, the high ON-state resistance and the high knee voltage VK in the fabricated devices are assumed to be the result of reduced effective carrier concentration in the nominally undoped aperture and drift region as mentioned previously. Additional testing of the aperture and drift region further indicated this assumption (not shown here). Fig. 5. OFF-state characteristics at VGS = −5 V of the large area devices with a chip size of 1 × 1 mm2 (red), 2 × 2 mm2 (blue). Small gate width CAVETs on the same chip with different aperture lengths are given in black. In addition, a CAVET without an aperture was measured (orange) to force the current through the CBL. the same range at VBR = 201 V. However, the drain leakage current is signiﬁcantly increased, especially in the range of VDS > 100 V. It is assumed that the higher leakage currents are the cause of a lower yield in the upscaled device but further data are needed to verify this assumption. A further increase in the drain leakage current of about two orders of magnitude is observable in the 2 × 2 mm2 device, and the breakdown voltage is reduced to VBR = 122 V, which is in agreement with the assumption of the loss in yield by further upscaling of the device area. Another reason for the higher leakage and reduced breakdown voltage could also result from dislocations/defects originating from the GaN-substrate. the same range at VBR = 201 V. However, the drain leakage current is signiﬁcantly increased, especially in the range of VDS > 100 V. It is assumed that the higher leakage currents are the cause of a lower yield in the upscaled device but further data are needed to verify this assumption. A further increase in the drain leakage current of about two orders of magnitude is observable in the 2 × 2 mm2 device, and the breakdown voltage is reduced to VBR = 122 V, which is in agreement with the assumption of the loss in yield by further upscaling of the device area. IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 68, NO. 11, NOVEMBER 2021 IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 68, NO. 11, NOVEMBER 2021 5550 Fig. 4. Mapping of the transfer characteristics at VDS = 10 V of large area CAVETs. The devices feature an aperture length of LAP = 2 µm, a gate aperture overlap of LGAP = 2 µm. Transfer curves are shown for 1 × 1 mm2-chip devices (red) and 2 × 2 mm2-chip devices (black) having a total gate width of WTOTAL = 13.5 mm and WTOTAL = 77 mm, respectively. The current compliance was set to 1 A. Fig. 5. OFF-state characteristics at VGS = −5 V of the large area devices with a chip size of 1 × 1 mm2 (red), 2 × 2 mm2 (blue). Small gate width CAVETs on the same chip with different aperture lengths are given in black. In addition, a CAVET without an aperture was measured (orange) to force the current through the CBL. the same range at VBR = 201 V. However, the drain leakage current is signiﬁcantly increased, especially in the range of VDS > 100 V. It is assumed that the higher leakage currents are the cause of a lower yield in the upscaled device but further Fig. 6. Pulsed output characteristics of the multi-ﬁnger CAVETs with a chip size of 1 × 1 (blue) and 2 × 2 mm2 (black). A pulsewidth of 100 µs and a sample time after 50 µs are used, with a duty cycle below 1%. of small gate width CAVETs in [17] (1.72 MV/cm, WG = 75 μm). The critical ﬁeld strength of the 1 × 1 mm2 (ECRIT = 2.3 MV/cm) is even exceeding these values, which demon- strates the suitability of the present in situ annealing/regrowth process for large multi-ﬁnger CAVETs as reported in this work. However, a higher breakdown voltage can be expected by increasing the drift layer thickness, as demonstrated in other vertical GaN-designs [5]–[9], [11], [12], in view to increase the area-efﬁciency with respect to lateral GaN-based power transistors. To demonstrate the behavior of the devices in terms of current handling capability and pulse power stability, pulsed output characteristics of the fabricated multi-ﬁnger CAVETs are measured (Fig. 6). A pulsewidth of 100 μs and a sample time of 50 μs is used with a duty cycle far below 1%. DÖRING et al.: TECHNOLOGY OF GaN-BASED LARGE AREA CAVETs WITH CO-INTEGRATED HEMTs DÖRING et al.: TECHNOLOGY OF GaN-BASED LARGE AREA CAVETs WITH CO-INTEGRATED HEMTs 5551 Fig. 7. Pulsed output characteristics of the multi-ﬁnger CAVETs with a chip size 2 × 2 mm2. VDS is measured up to 50 V and VGS is varied from −5 to 3 V. A pulsewidth of 500 µs and a sample time after 50 µs are used. A pulse power of PPULS = 968 W was measured at VDS = 50 V which is close to the predicted device failure of PFAIL = 976 W (using EFAIL = 200 mJ/mm2). Fig. 8. Output characteristic of the lateral on-chip HEMT with varying gate–source voltage of VGS = −6 to 0 V for VDS = 0 to 40 V. The device features a gate width of WG = 50 µm, a length of the active area of LACT = 25.5 µm and a GaN-channel thickness of tCHA = 250 nm between the Mg-implanted CBL and the AlGaN barrier. Fig. 8. Output characteristic of the lateral on-chip HEMT with varying gate–source voltage of VGS = −6 to 0 V for VDS = 0 to 40 V. The device features a gate width of WG = 50 µm, a length of the active area of LACT = 25.5 µm and a GaN-channel thickness of tCHA = 250 nm between the Mg-implanted CBL and the AlGaN barrier. Fig. 8. Output characteristic of the lateral on-chip HEMT with varying gate–source voltage of VGS = −6 to 0 V for VDS = 0 to 40 V. The device features a gate width of WG = 50 µm, a length of the active area of LACT = 25.5 µm and a GaN-channel thickness of tCHA = 250 nm between the Mg-implanted CBL and the AlGaN barrier. Fig. 7. Pulsed output characteristics of the multi-ﬁnger CAVETs with a chip size 2 × 2 mm2. VDS is measured up to 50 V and VGS is varied from −5 to 3 V. A pulsewidth of 500 µs and a sample time after 50 µs are used. A pulse power of PPULS = 968 W was measured at VDS = 50 V which is close to the predicted device failure of PFAIL = 976 W (using EFAIL = 200 mJ/mm2). vertical transistors. On the other hand, this potential barrier is expected to protect the low-voltage HEMT-logic from the high drain potential. DÖRING et al.: TECHNOLOGY OF GaN-BASED LARGE AREA CAVETs WITH CO-INTEGRATED HEMTs However, connecting the CBL to the source contact of the CAVET will be beneﬁcial to avoid any impact of the drain potential on the static and dynamic behavior of the lateral HEMTs. of EFAIL = 200 mJ/mm2 at VDS = 50 V is assumed, which was experimentally derived by Mocanu et al. [24]. Using a pulse power PPULS = 500 W and an active area of AACT = 2.16 mm2, a destruction power of PFAIL = 976 W is calculated. Pulsed output characteristics of the 2 × 2 mm2-chip CAVET are measured for VDS = 0 to 50 V and VGS = −5 to 3 V (Fig. 7). The multi-ﬁnger CAVET exhibited an absolute drain current of 20.3 A at VGS = 3 V and a total differential ON-state resistance of RON = 1.67 . A pulse power of PPULS = 968 W was measured at VDS = 50 V, close to the predicted destruction power of a lateral HEMT. Self-heating is visible but device failure in terms of sudden drain current increase or physical breakdown was not observed [24]. IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 68, NO. 11, NOVEMBER 2021 Another reason for the higher leakage and reduced breakdown voltage could also result from dislocations/defects originating from the GaN-substrate. The measured breakdown voltage of the 2 × 2 mm2 device corresponds to a critical ﬁeld strength of ECRIT = 1.4 MV/cm assuming depletion over the complete drift layer distance (according to [2]). It is also worth noting that the calculated ﬁeld strength is still close to published values To compare the pulse power stability with published data for lateral Schottky-gate HEMTs, a speciﬁc destruction energy IV. CONCLUSION In this article, recent progress in the CAVET technology is presented using low-dose Mg-implantation and MOCVD regrowth. The intrinsic CAVET design was optimized with respect to device performance and gate control. Larger aperture lengths were chosen to avoid current choking for low donor concentration. The fabricated multi-ﬁnger CAVETs presented in this work exhibit high absolute current capability and high pulse power stability. Lateral HEMTs were co-integrated on- chip, with the 2DEG in the vicinity of the Mg-implanted CBL. This combination of a large area CAVET and an on-chip integrated HEMT demonstrates a pathway to allow monolithically integrated GaN power converters with power switches based on the CAVET technology. Future works will focus on the optimization of the aperture and drift region conductivity, as well as the increase in breakdown voltage to enable higher area efﬁciency in the vertical design with respect to its lateral counterpart. Finally, lateral transistors are processed on-chip, without an aperture below the gate. The fabricated HEMTs have a gate width of WG = 50 μm and an active area length of LACT = 25.5 μm based on our 600 V HEMT-technology [23] but features the Mg-implanted CBL in the vicinity to the 2DEG [Fig. 1(e)]. Transfer characteristics (not shown here) at VDS = 10 V and VGS = −6 to 0 V revealed a sub-threshold swing of SS-TH = 93.4 mV/decade and a threshold voltage of VTH = −2.81 V. The measured gate leakage current was lower than 1 × 10−4 A/mm2. Output characteristics are measured for VDS = 0 to 40 V in 0.1 V steps (Fig. 8). p g The lateral FET reveals a maximum drain current of ID,MAX = 0.4 A/mm or ID,MAX = 1.59 kA/cm2 at VGS = 0 V and an area-speciﬁc ON-state resistance of RON × A = 2.98 m·cm2. The resulting output characteristics of the lateral FETs demonstrate the suitability of the used fabrication process. However, in comparison to the vertical FETs, a signif- icantly improved current density and ON-state conductivity can be observed. This is an additional indication that the aperture and drift region conductivity is reduced. The measurements of IV-characteristics of the on-chip integrated lateral transistors reveal the potential of future monolithic integration of the conventional HEMT-technology. On one hand, the CBL serves as a potential barrier to suppress source-drain leakage in the REFERENCES REFERENCES [1] A. Q. Huang, “Power semiconductor devices for smart grid and renewable energy systems,” Proc. IEEE, vol. 105, no. 11, pp. 2019–2047, Nov. 2017, doi: 10.1109/JPROC.2017.2687701. [2] B. J. Baliga, “Power semiconductor device ﬁgure of merit for high- frequency applications,” IEEE Electron Device Lett., vol. 10, no. 10, pp. 455–457, Oct. 1989, doi: 10.1109/55.43098. [3] P. Waltereit et al., “GaN-based high voltage transistors for efﬁcient power switching,” Phys. Status Solidi C, vol. 10, no. 5, pp. 831–834, 2013, doi: 10.1002/pssc.201200563. [4] S. Krause, P. Bruckner, M. Dammann, and R. Quay, “High-power- density AlGaN/GaN technology for 100-V operation at L-band frequencies,” in IEDM Tech. Dig., San Francisco, CA, USA, Dec. 2019, pp. 1–17. IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 68, NO. 11, NOVEMBER 2021 5552 [5] Y. Zhang et al., “Large-area 1.2-kV GaN vertical power FinFETs with a record switching ﬁgure of merit,” IEEE Electron Device Lett., vol. 40, no. 1, pp. 75–78, Jan. 2019, doi: 10.1109/LED.2018. 2880306. [15] D. Ji, A. Agarwal, W. Li, S. Keller, and S. Chowdhury, “Demonstration of GaN current aperture vertical electron transistors with aperture region formed by ion implantation,” IEEE Trans. Electron Devices, vol. 65, no. 2, pp. 483–487, Feb. 2018, doi: 10.1109/TED.2017.2786141. y p no. 2, pp. 483–487, Feb. 2018, doi: 10.1109/TED.2017.27 [6] T. Oka, Y. Ueno, T. Ina, and K. Hasegawa, “Vertical GaN-based trench metal oxide semiconductor ﬁeld-effect transistors on a free-standing GaN substrate with blocking voltage of 1.6 kV,” Appl. Phys. Exp., vol. 7, no. 2, p. 21002, 2014, doi: 10.7567/APEX.7.021002. [16] R. Yeluri et al., “Design, fabrication, and performance analysis of GaN vertical electron transistors with a buried p/n junction,” Appl. Phys. Lett., vol. 106, pp. 183502-1–183502-5, May 2015, doi: 10.1063/1.4919866. [17] S. Chowdhury, M. H. Wong, B. L. Swenson, and U. K. Mishra, “CAVET on bulk GaN substrates achieved with MBE-regrown AlGaN/GaN layers to suppress dispersion,” IEEE Electron Device Lett., vol. 33, no. 1, pp. 41–43, Jan. 2012, doi: 10.1109/LED.2011.2173456. [7] D. Ji et al., “Large-area in situ oxide, GaN interlayer-based vertical trench MOSFET (OG-FET),” IEEE Electron Device Lett., vol. 39, no. 5, pp. 711–714, May 2018, doi: 10.1109/LED.2018.2813312. [8] M. Sun, Y. Zhang, X. Gao, and T. Palacios, “High-performance GaN vertical ﬁn power transistors on bulk GaN substrates,” IEEE Electron Device Lett., vol. 38, no. 4, pp. 509–512, Apr. 2016, doi: 10.1109/LED.2017.2670925. [18] D. Ji, A. Agarwal, H. Li, W. Li, S. Keller, and S. REFERENCES Chowdhury, “880 V/2.7 m·cm2 MIS gate trench CAVET on bulk GaN substrates,” IEEE Electron Device Lett., vol. 39, no. 6, pp. 863–865, Jun. 2018, doi: 10.1109/LED.2018.2828844. [9] T. Oka, T. Ina, Y. Ueno, and J. Nishii, “1.8 m·cm2 vertical GaN-based trench metal-oxide-semiconductor ﬁeld-effect transistors on a free- standing GaN substrate for 1.2-kV-class operation,” Appl. Phys. Exp., vol. 8, no. 5, p. 54101, 2015, doi: 10.7567/APEX.8.054101. [19] S. Mandal et al., “Dispersion free 450-V p GaN-gated CAVETs with Mg-ion implanted blocking layer,” IEEE Electron Device Lett., vol. 38, no. 7, pp. 933–936, Jul. 2017, doi: 10.1109/LED.2017.2709940. [20] S. Rajabi et al., “A demonstration of nitrogen polar gallium nitride current aperture vertical electron transistor,” IEEE Electron Device Lett., vol. 40, no. 6, pp. 885–888, Jun. 2019, doi: 10.1109/LED.2019.2914026. [10] S. Moench et al., “Monolithic integrated AlGaN/GaN power con- verter topologies on high-voltage AlN/GaN superlattice buffer,” Phys. Status Solidi A, vol. 218, no. 3, p. 2000404, 2021, doi: 10.1002/pssa.202000404. [21] P. Doering, R. Driad, R. Reiner, P. Waltereit, M. Mikulla, and O. Ambacher, “Metal organic chemical vapour deposition regrown large area GaN-on-GaN current aperture vertical electron transistors with high current capability,” Electron. Lett., vol. 57, no. 3, pp. 145–147, Feb. 2021, doi: 10.1049/ell2.12068. [11] D. Shibata et al., “1.7 kV/1.0 m·cm2 normally-off vertical GaN transis- tor on GaN substrate with regrown p-GaN/AlGaN/GaN semipolar gate structure,” in IEDM Tech. Dig., San Francisco, CA, USA, Dec. 2016, pp. 10.1.1–10.1.4. [22] P. Doering et al., “Growth and fabrication of quasivertical current aperture vertical electron transistor structures,” Phys. Status Solidi A, vol. 218, no. 3, Feb. 2021, Art. no. 2000379, doi: 10.1002/pssa.202000379. pp [12] H. Nie et al., “1.5-kV and 2.2-m-cm2 vertical GaN transistors on bulk GaN substrates,” IEEE Electron Device Lett., vol. 35, no. 9, pp. 939–941, Sep. 2014, doi: 10.1109/LED.2014.2339197. p [23] P. Waltereit et al., “Impact of GaN cap thickness on optical, electrical, and device properties in AlGaN/GaN high electron mobility transis- tor structures,” J. Appl. Phys., vol. 106, no. 2, p. 23535, 2009, doi: 10.1063/1.3184348. pp p [13] I. Ben-Yaacov, Y.-K. Seck, U. K. Mishra, and S. P. DenBaars, “AlGaN/GaN current aperture vertical electron transistor with regrown channels,” J. Appl. Phys., vol. 95, no. 4, p. 2073, 2004, doi: 10.1063/1.1641520. [14] S. Chowdhury, B. L. Swenson, and U. K. Mishra, “Enhancement and depletion mode AlGaN/GaN CAVET with Mg-ion-implanted GaN as current blocking layer,” IEEE Electron Device Lett., vol. 29, no. 6, pp. REFERENCES 543–545, Jun. 2008, doi: 10.1109/LED.2008.922982. [24] M. Mocanu, C. Unger, M. Pfost, P. Waltereit, and R. Reiner, “Thermal stability and failure mechanism of Schottky gate AlGaN/GaN HEMTs,” IEEE Trans. Electron Devices, vol. 64, no. 3, pp. 848–855, Mar. 2017, doi: 10.1109/TED.2016.2633725.
https://openalex.org/W2605602162	https://www.scielo.br/j/coluna/a/DbR43TrhWfpxyvkJCKvBHCp/?lang=en&format=pdf	English	null	EFFECTIVENESS OF AN INTERDISCIPLINARY PROGRAM IN PATIENTS WITH FAILED BACK SURGERY SYNDROME	Coluna/Columna	2,017	cc-by	4,749	EFICACIA DE UN PROGRAMA INTERDISCIPLINARIO EN PACIENTES CON SÍNDROME POSTLAMINECTOMÍA José Eduardo Nogueira Forni1, Ana Marcia Rodrigues Cunha2, Carlos Eduardo D’Aglio Rocha3, Lilian Chessa Dias2, Marco Henrique D’All Aglio Foss3, Randolfo dos Santos Junior2, Gerardo Maria de Araujo Filho4, Marielza Regina Ismael Martins3 1. Faculdade de Medicina de São José do Rio Preto, Departamento de Ortopedia e Traumatologia, São José do Rio Preto, SP, Brazil. 2. Faculdade de Medicina de Rio Preto, Hospital de Base de São José do Rio Preto, São José do Rio Preto, SP, Brazil. 1. Faculdade de Medicina de São José do Rio Preto, Departamento de Ortopedia e Traumatologia, São José do Rio Preto, SP, Brazil. 2. Faculdade de Medicina de Rio Preto, Hospital de Base de São José do Rio Preto, São José do Rio Preto, SP, Brazil. 3. Faculdade de Medicina de São José do Rio Preto, Departamento de Ciências Neurológicas, São José do Rio Preto, SP, Brazil. 4. Faculdade de Medicina de São José do Rio Preto, Departamento de Psiquiatria e Psicologia Médica, São José do Rio Preto, SP, Brazil. 3. Faculdade de Medicina de São José do Rio Preto, Departamento de Ciências Neurológicas, São José do Rio Preto, SP, Brazil. 4. Faculdade de Medicina de São José do Rio Preto, Departamento de Psiquiatria e Psicologia Médica, São José do Rio Preto, SP, Brazil. Original Article/Artigo Original/Artículo Original Original Article/Artigo Original/Artículo Original RESUMO Objetivo: Avaliar os resultados de um programa interdisciplinar aplicado a pacientes com síndrome pós-laminectomia, visando melhora funcional, modulação da dor, redução de sintomas de ansiedade e depressão e aumento da qualidade de vida. Método: Trata-se de um estudo prospectivo não randomizado, com amostra constituída por um grupo de pacientes com síndrome dolorosa pós-laminectomia diagnosticados com dor persistente ou recorrente após cirurgias da coluna lombar (laminectomia e artrodese) encaminhados para interconsultas na Clínica da Dor (n= 26). Os instrumentos utilizados foram: Inventário Breve de Dor, Questionário Roland-Morris e Inventários Beck de ansiedade e depressão. Para avaliar a qualidade de vida, foi utilizado Questionário Genérico Whoqol-bref13 e o medo de se movimentar foi avaliado pela Escala Tampa para Cinesiofobia. Resultados: Houve predomínio de mulheres, a média de idade foi 42,3 ± 5,8 anos, 43% eram casados e a média de escolaridade foi de 7 ± 4,5 anos. O tempo médio de dor relatado foi de 8 ± 6,8 meses, além de níveis altos de ansiedade, depressão e cinesiofobia. Após a intervenção houve melhora significativa na percepção da qualidade de vida e de todos os parâmetros avaliados (p < 0,05) observando-se ganhos funcionais, assim como a diminuição do limiar de dor. Conclusão: A intervenção interdisciplinar nos pacientes com síndrome pós-laminectomia proporciona melhor desempenho funcional, diminui a intensidade da dor, os sintomas de ansiedade e depressão, melhorando a qualidade de vida. A inclusão desta intervenção associada ao tratamento farmacológico pode desenvolver um estilo de vida ativo e independente. Descritores: Dor lombar; Pesquisa interdisciplinar; Reoperação. ABSTRACT Objective: To evaluate the results of an interdisciplinary program administered to patients with failed back surgery syndrome, aiming at functional improvement, modulation of pain, reduction of anxiety symptoms and depression, and improvement of quality of life. Method: This is a non-randomized prospective study with a sample of patients with failed back surgery pain syndrome diagnosed with persistent or recurrent pain after surgery to the lumbar spine (laminectomy and arthrodesis) referred to liaison in the Pain Clinic (n= 26). The instruments used were Brief Pain Inventory, Roland-Morris Questionnaire and Beck Anxiety and Depression Inventories. The generic WHOQOL-bref13 questionnaire was used to evaluate the quality of life and the fear of moving was assessed by the Tampa Scale for Kinesiophobia. Results: There was a predominance of females, the mean age was 42.3 ± 5.8 years, 43% were married and average schooling was 7 ± 4.5 years. The mean time of pain reported was 8 ± 6.8 months in addition to high levels of anxiety, depression and kinesiophobia. After the intervention, there was a significant improvement in the perception of quality of life and of all parameters evaluated (p<0.05), with functional gains as well as decreased pain threshold. Conclusion: The interdisciplinary intervention in patients with failed back surgery syndrome provides better functional performance, decreases the intensity of pain, anxiety and depression symptoms, and improves quality of life. The inclusion of this intervention associated with drug therapy may the patient develop an active and independent lifestyle. Keywords: Low back pain; Interdisciplinary research; Reoperation. Study conducted at the Faculdade de Medicina de Rio Preto, Hospital de Base de São José do Rio Preto, São José do Rio Preto, SP, Brazil. Correspondence: Marielza R. I. Martins. Departamento de Ciências Neurológicas Faculdade de Medicina de São Jose do Rio Preto. Avenida Brigadeiro Faria Lima 5416, São Jose do Rio Preto, SP, Brasil. 15090-000. marielzamartins@famerp.br EFICÁCIA DE UM PROGRAMA INTERDISCIPLINAR EM PACIENTES COM SÍNDROME PÓS-LAMINECTOMIA EFICACIA DE UN PROGRAMA INTERDISCIPLINARIO EN PACIENTES CON SÍNDROME POSTLAMINECTOMÍA Received on 01/25/2016, accepted on 07/20/2016. INTRODUCTION were conducted by professionals in the areas of physical therapy, occupational therapy, psychology and social service. Each session was conducted by a professional, with the others as collaborators (Table 1). In the 12th session, the patients were reassessed by the same criteria of efficiency, then the results were compared. All patients continued their pharmacological treatment. Failed back surgery syndrome (FBSS) is characterized by the persistent maintenance of lumbar and / or lower limb pain complaints in patients previously submitted to lumbar vertebral surgery, with possible changes in physical examination and imaging1 resulting in a cascade of clinical and surgical events causing loss of function, reduced quality of life and increased costs to society.2 The evaluation of patients was performed using a semi-struc- tured interview with sociodemographic data: age, sex, marital status, education, employment status and socioeconomic classification (Brazilian Association of Market Research Institutes/ABIPEME),16 weight and height. For measuring pain and its impact on some areas of life, we applied the Brief Pain Inventory (BPI),17 which is an instrument comprised of 15 items divided into two parts: the first part evaluates the intensity of pain and the second, the interference of the pain in aspects of life (general activities, mood, ability to walk, sleep, work, relationships with others and enjoyment of life). The intensity and interference of pain are scored from 0 (no pain) to 10 (worst possible pain). The degree of disability was assessed by the Roland-Morris Questionnaire (RM)18 which consists of 24 questions, selected to cover a range of aspects related to activities of daily living, pain and function. The questions are objective and simple, with a score of “1” for each statement the patient agrees with, and “0” for each statement the patient does not agree with. The score is the sum of the values, with a minimum of “0” and a maximum of “24. The closer the score is to “24”, the more severe the individual’s disability with chronic low back pain is. This questionnaire has a cutoff score of “14”, in other words, individuals with a score higher than 14 have disabilities. The validity and reproducibility of the Portuguese version, translated and adapted, are well established in the literature. To evaluate the quality of life, we used the generic WHOQOL-bref19 Questionnaire, which comprises 26 questions covering a broad area and four specific domains (physical, psy- chological, social relationships and environment). INTRODUCTION The quality of life scores of the WHOQOL-bref domains are from zero to 100; the higher the value for each domain, the better the quality of life. Emotional factors such as anxiety and depression were assessed by the Beck The incidence of FBSS varies considerably in different studies, and different factors involved in its pathophysiology are mentioned.1,3 Factors such as incorrect or incomplete diagnosis before surgery, inappropriate surgery, failure to consider the patient’s psychosocial profile (anxiety, depression, substance abuse, alexithymia) and the presence of inflammation at the surgical site all contribute to this outcome, ranging from 10% to 40% in accordance with the number of spine surgeries performed.4 The results of failed back surgery syndrome are extremely variable, ranging from complete relief to a life of pain. Reports indicate that when the initial surgery does not obtain good results, the success rate decreases with each additional surgery.5,6 Multiple surgeries in elderly patients also have poorer results.7 In this context, rehabilitation intervention is extremely important in this group of patients8 by the resource of different therapeutic means.8,9 The aim of rehabilitation is to maximize the functional abilities of the individual because, with this syndrome a chronic condition, a comprehensive evaluation of an interdisciplinary pain clinical is indicated for a temporal concept, focusing on the future and not the past. The aim of this study was to evaluate the impact of a group service based on rehabilitation strategies, guidance for patients, and changes in attitudes and beliefs in relation to pain. Coluna/Columna. 2017;16(1):48–51 EFFECTIVENESS OF AN INTERDISCIPLINARY PROGRAM IN PATIENTS WITH FAILED BACK SURGERY SYNDROME EFFECTIVENESS OF AN INTERDISCIPLINARY PROGRAM IN PATIENTS WITH FAILED BACK SURGERY SYNDROME 49 de altos niveles de ansiedad, depresión y kinesiofobia. Después de la intervención se observó una mejoría significativa en la percepción de la calidad de vida y de todos los parámetros (p < 0,05), siguiendo las ganancias funcionales, así como la disminución del umbral del dolor. Conclusión: La intervención interdisciplinaria en pacientes con síndrome postlaminectomía mejora la funcionalidad, disminuye la intensidad del dolor, los síntomas de ansiedad y depresión y mejora la calidad de vida. La inclusión de esta intervención asociada con el tratamiento farmacológico puede proporcionar una vida activa e independiente. Descriptores: Dolor de la región lumbar; Investigación interdisciplinaria; Reoperación. RESUMEN Objetivo: Evaluar los resultados de un programa interdisciplinario subministrado a pacientes con síndrome postlaminectomía que busca la mejora funcional, la modulación del dolor, la reducción de los síntomas de ansiedad y depresión y calidad de vida. Método: Se trata de un estudio prospectivo no aleatorio de una muestra compuesta por un grupo de pacientes con síndrome doloroso postlaminectomía diagnosticados con dolor persistente o recurrente después de la cirugía de la columna lumbar (laminectomía y fusión) referidos a interconsultas en la Clínica del Dolor (n= 26). Los instrumentos utilizados fueron: Cuestionario breve del dolor, Cuestionario Roland-Morris e inventarios de ansiedad y depresión de Beck. Para evaluar la calidad de vida se utilizó el cuestionario genérico WHOQOL-BREF13 y el miedo de moverse fue evaluado por la Escala de Tampa de Kinesiofobia. Resultados: Hubo un predominio de mujeres, la edad promedio fue de 42,3 ± 5,8 años, 43% eran casados y la educación promedio fue de 7 ± 4,5 años. Se reportó una duración promedio del dolor de 8 ± 6,8 meses, además Study conducted at the Faculdade de Medicina de Rio Preto, Hospital de Base de São José do Rio Preto, São José do Rio Preto, SP, Brazil. Correspondence: Marielza R. I. Martins. Departamento de Ciências Neurológicas Faculdade de Medicina de São Jose do Rio Preto. Avenida 15090-000. marielzamartins@famerp.br http://dx.doi.org/10.1590/S1808-185120171601158955 Received on 01/25/2016, accepted on 07/20/2016. Coluna/Columna. 2017;16(1):48–51 MATERIAL AND METHODS This is a prospective non-randomized study conducted with patients with failed back surgery syndrome in an outpatient group of the Pain Clinic, Hospital de Base in São José do Rio Preto/SP/ Brazil. This purpose of this study was to gather knowledge on the effect of a group intervention in these individuals, based on specific protocols and systematic research carried out from August to November 2014, after gaining the approval of the Research Ethics Committee under no. 384/2012. Table 1. Description of group activities. 1ª e 12ª sessions Initial and final evaluation 2ª session Educational and awareness activities as the anatomy of the spine and the physiological aspects of the low back pain Flexibility and stretching exercises 3ª session Awareness of the possibilities for social participation Ergonomic aspects related to activities of daily living, work and leisure 4ª session Group dynamics for psychosocial approach 5ª session Aerobic and paravertebral muscles strengthening exercises 6ª session Guidance on food education 7ª session Educational and awareness activities Aspects related to activities of daily living 8ª session Exercises with gradual increase of performance 9ª session Psychosocial approach 10ª session Awareness of the possibilities for social participation 11ª session Group dynamics Table 1. Description of group activities. 1ª e 12ª Table 1. Description of group activities. The population consisted of individuals of both sexes, with a sufficient cognitive level to understand the procedures and follow the guidelines, and who agreed to participate by signing an informed consent form. A total of 26 patients with painful syndrome were included in the study, having been diagnosed with persistent or recurrent pain after lumbar surgery (laminectomy and arthrodesis). Patients with psychiatric illness, lack of clinical follow-up in the Pain Clinic, metabolic (osteoporosis), inflammatory (arthritis, spondylitis) and oncological diseases, were excluded. g From the 2nd to the 11th sessions (once a week, lasting 60 minutes), patients performed flexibility and stretching exercises, with a gradual improvement in performance10 guidance on rest positions,11,12 aerobic and paravertebral muscle strengthening exercises,8 educational and awareness activities relating to the anatomy of the spine and the physiological aspects of the low back pain,8,13 encourage opportunities for participation social 16,17 guidance on food education,13 ergonomic aspects related to activities of daily living, work and leisure8,13,14 and group dynamics for the psychosocial approach15. All the sessions 50 Table 3. Mean values according to the Tampa Scale of Kinesiophobia (TSK) and Brief Pain Inventory (BPI) pre-and-post-group intervention (n= 26). MATERIAL AND METHODS Instruments Initial evaluation average and SD ( ± ) Average reevaluation and SD ( ± ) P-Value BPI Pain Intensity 9.04 ± 2.3 6.30 ± 3.45 0.03* Pain interference with general activity Mood 8.35 ± 3.04 6.50 ± 2.04 0.04* Ability to walk 7.32 ± 2.50 6.60 ± 3.67 0.06 Sleep 9.35 ± 1.98 5.20 ± 2.72 0.03* Work 8.50 ± 2.35 6.12 ± 4.09 0.05* Personal relationship 5.65 ± 3.25 2.93 ± 1.99 0.04* Enjoyment of life 6.50 ± 3.50 3.81 ± 1.86 0.06 TSK 49.5 ± 6.73 26.4 ± 5.82 0.03* Table Caption: BPI - Brief Pain Inventory; TSK - Tampa Scale of Kinesiophobia.* P ≤0.05 - Wilcoxon non-parametric test Table 3. Mean values according to the Tampa Scale of Kinesiophobia (TSK) and Brief Pain Inventory (BPI) pre-and-post-group intervention (n= 26). Scale. In this research we used the Beck Anxiety Inventory (BAI) and the Beck Depression Inventory (BDI). The BAI20 was proposed to measure the common symptoms of anxiety. It consists of 21 symptoms listed, with four alternatives for each one, in ascending order of level of anxiety. The scale classifies anxiety as minimum (0-9 points); mild (10-16 points); moderate (17-29 points); and severe (30-63 points). The BDI21 comprises 21 categories of symptoms and activities, containing four alternatives for each one, in ascending order by level of depression. The patient is asked to choose the answer that best fits their symptoms in the last week. The sum of the scores identifies the level of depression. The following result is proposed for the degree of depression: minimum (from 0-11 points); mild (12-19 points); moderate (20-35 points); and severe (36-63 points). Excessive fear of movement and physical activity, which results in feelings of vulnerability to pain or fear of recurrence of the lesion, was observed by the Tampa Scale of Kinesiophobia (TSK),22 which consists of 17 statements about pain. the patient must decide whether they agree or disagree with each statement, using a four-point scale. The final score is a minimum of 17 and a maximum of 68 points. The higher the score, the greater the degree of kinesiophobia, indicating that the individual is afraid to move because of the low back pain. Table Caption: BPI - Brief Pain Inventory; TSK - Tampa Scale of Kinesiophobia.* P ≤0.05 - Wilcoxon non-parametric test Figure 1. MATERIAL AND METHODS Distribution (%) of the domains (physical, psychological, social relations, environment) of quality of life, pre- and post-intervention. Physical; 65.47 Physical; 46.49 Pre Post Psychological; 49.97 Social relations; 51.34 Social relations; 63.64 Environment; 70.65 Environment; 49.5 Psychological; 72.54 As criteria for improvement and worsening, rates of score changes were calculated, for the questionnaires applied before and after the group intervention. We performed a descriptive analysis for all the study variables. The analyses were performed by ANOVA to determine differences. To see which variables differed, the non-parametric Wilcoxon and Mann Whitney tests were applied. The significance level used for the tests was 0.05. RESULTS Variable N Mean and standard deviation ( ± ) % Education 26 7 ± 4.5 years Weight 26 50-60kg 38.4 61-70kg 34.6 >70kg 27 Height 26 1.62 ± 10.8 m Pain Duration 26 8 ± 6.8 meses Number of Surgeries 2.2 Employment Status 26 Formal Working 7.9 Disability Insurance 46 Retired 26.9 Informal Working 19.2 Table 2. Characterization of sociodemographic and clinical profile of the study group (n= 26). Figure 2. Levels of anxiety and depression assessed by the Beck Scale (BAI–anxiety and BDI–depression) pre and post interdisciplinary intervention (n= 26). The group initially had greater functional disabilities (17.05 ± 3.48) assessed by the Roland Morris Questionnaire. We observed statistically significant difference in post-intervention functional capacity (6.78 ± 2.35, p= 0.02), and decreasing physical limitations of referred pain in the lumbar spine (Mann Whitney test). RESULTS Our sample consisted mainly of females (67%), married (43%), with a mean age of 42.3 ± 5.8 years. In relation to socioeconom- ic classification, there was a predominance of the category low (½ to 2 minimum wages). More than half of the sample belonged to this socioeconomic level (55.7%). In relation to current occupation, 46% (n= 12) of the patients were inactive (disability insurance). The mean pain time of these patients before joining the group was 8 ± 2.5 months. No patients were lost to follow-up. Figure 1. Distribution (%) of the domains (physical, psychological, social relations, environment) of quality of life, pre- and post-intervention. Figure 2. Levels of anxiety and depression assessed by the Beck Scale (BAI–anxiety and BDI–depression) pre and post interdisciplinary intervention (n= 26). Pre 45 40 35 30 25 20 15 10 5 0 Post anxiety GT anxiety GC depression GT depression GC p p Other data regarding the sociodemographic and clinical profile are shown in Table 2. Regarding pain and fear of movement, the table below (Table 3) shows a comparison between the pre and post-intervention periods. Regarding pain and fear of movement, the table below (Table 3) shows a comparison between the pre and post-intervention periods. The relationship between perceived quality of life in the pre- and post-intervention periods is shown in the figure below (Figure 1). The levels of anxiety and depression decreased after the interdisciplinary intervention (Figure 2) showing statistical difference The relationship between perceived quality of life in the pre- and post-intervention periods is shown in the figure below (Figure 1). p p g ( g ) The levels of anxiety and depression decreased after the interdisciplinary intervention, (Figure 2) showing statistical difference (p <0.05, ANOVA test). Table 2. Characterization of sociodemographic and clinical profile of the study group (n= 26). Variable N Mean and standard deviation ( ± ) % Education 26 7 ± 4.5 years Weight 26 50-60kg 38.4 61-70kg 34.6 >70kg 27 Height 26 1.62 ± 10.8 m Pain Duration 26 8 ± 6.8 meses Number of Surgeries 2.2 Employment Status 26 Formal Working 7.9 Disability Insurance 46 Retired 26.9 Informal Working 19.2 Table 2. Characterization of sociodemographic and clinical profile of the study group (n= 26). Coluna/Columna. 2017;16(1):48–51 REFERENCES atividade física? Acta Fisiatr. 2011;18(1):27-31. atividade física? Acta Fisiatr. 2011;18(1):27-31. 1. Slawsky A, Hennemann SA. Análise retrospectiva das causas de falha no tratamento cirúrgico das doenças da coluna vertebral. Coluna/Columna. 2008;7(3):235-40. 1. Slawsky A, Hennemann SA. Análise retrospectiva das causas de 15. Boonstra AM, Reneman MF, Waaksma BR, Schiphorst Preuper HR, Stewart RE. Predictors of multidisciplinary treatment outcome in patients with chronic musculoskeletal pain. Disabil Rehabil. 2015;37(14):1242-50. 2. Manca A, Eldabe S, Buchser E, Kumar K, Taylor RS. Relationship between health-related quality of life, pain, and functional disability in neuropathic pain patients with failed back surgery syndrome. Value Health. 2010;3(1):95-102. 16. Associação Brasileira de Institutos de Pesquisa de Mercado (Abipeme). Classificação socioeconômica - critério Abipeme. Anexo 3. [acesso em 2015 jan 28]. Disponível em: http://www.ufrn.br/sites/fonprace/perfil_anexo3.doc 3. Matias AC, Antunes F. Qualidade de vida na síndrome de cirurgia lombar falhada. Rev Soc Port Med FIs Reab. 2012;21(1):32-6. 4. Chan CW, Peng P. Failed back surgery syndrome. Pain Med. 2011;1 17. Cleeland C, Ryan K. Pain assessment:global use of the Brief Inventory. Ann Academy Med Singapore.1994;23(2):129-38. 5. Thomson S, Jacques L. Demographic characteristics of patients with severe neuropathic pain secondary to failed back surgery syndrome. Pain Pract. 2009;9(3):206-15. 18. Nusbaum L, Natour J, Ferraz MB, Goldenberg J. Translation, adaptation and Validation of the Roland-Morris questionnaire – Brazil Roland Morris. Braz J Med Biol Res. 2001;34(2):203-10. 6. Sciubba DM, Gokaslan ZL. Spinal cord stimulation for failed back surgery syndrome-- does it work and is it cost-effective? Nat Clin Pract Neurol. 2008;4(9):472-3. 19. Fleck MP, Louzada S, Xavier M, Chachamovich E, Vieira G, Santos L, et al. Application of the Portuguese version of the abbreviated instrument of quality life WHOQOL-bref. Rev Saude Publica. 2000;30(2):178-83. 7. Archer KR, Seebach CL, Mathis SL, Riley LH, Wegener ST. Early postoperative fear of movement predicts pain, disability, and physical health six months after spinal surgery for degenerative conditions. J Spine. 2014;13(5):759-67. 8. Miller B, Gatchel RJ, Lou L, Stowell A, Robinson R, Polatin PB. Interdisciplinary treatment of failed back surgery syndrome (FBSS): a comparison of FBSS and non-FBSS patients. Pain Pract. 2005;5(3):190-202. 20. Cunha JA. Inventário Beck de Ansiedade (BAI). Manual da versão em português das escalas Beck. São Paulo: Casa do Psicólogo; 2001. 21. Cunha JA. Inventário Beck de Depressão (BDI). Manual da versão em português das escalas Beck. São Paulo: Casa do Psicólogo; 2001. 9. Martins MRI, Polvero LO, Rocha CE, Foss MH, Santos Junior R. CONCLUSION Self-perception of improved quality of life and decreased pain intensity after the interdisciplinary program were significant, corroborating the study by Al Kaysi et al.25 who propose a care pathway for patients with failed back surgery syndrome with temporal concept, focusing on the future and not the past. This study defines an optimized care pathway, with a multidisciplinary program combining intensive physical rehabilitation with cognitive and behavioral treatment for pain management. Based on the data of the present study, we conclude that the interdisciplinary intervention in patients with failed back surgery syndrome provided better functional performance, decreased the severity of pain, anxiety and symptoms of depression, and improved their quality of life. The inclusion of this intervention, combined with pharmacological treatment, can systematically develop self-regulation skills needed for the rehabilitation transition to maintain an active and independent lifestyle, adding specific activities to their daily or weekly planning, and the regular follow-up should stimulate long- term maintenance of behavioral change. g p g This study also demonstrated that patients after surgery showed an improvement in symptoms of anxiety and depression. The data are similar to those previously reported in the literature for patients with chronic pain.19,24 Meinhart and McCaffery26 also point out that anxiety is a feeling often associated with acute pain, and is recognized as being responsible for increased pain perception. Anxiety is related to the intensity of pain and the possible meanings that the pain has for the patient. Another factor demonstrated by McGregor and Hughes27 is that patients undergoing surgery for lumbar stenosis have All the authors declare that there are no potential conflicts of interest regarding this article. CONTRIBUTIONS OF THE AUTHORS: Each author made an individual and significant contribution to the development of the manuscript. JENF, AMRC, CEDR and GMAF collected and evaluated the clinical data, MRIM, MHAAF, LCD and RSJ evaluated the data from the statistical analysis, performed the literature review, revised the manuscript, and contributed to the intellectual concept of the study. DISCUSSION The objective of this study was to evaluate the response of patients with failed back surgery syndrome to participating in an interdisciplinary program, in terms of pain, functional limitations, quality of life and emotional disturbances. Coluna/Columna. 2017;16(1):48–51 EFFECTIVENESS OF AN INTERDISCIPLINARY PROGRAM IN PATIENTS WITH FAILED BACK SURGERY SYNDROME 51 unrealistic expectations of the treatment, and this is associated with poorer satisfaction scores in the postoperative period. In this regard, the detailed preoperative evaluation and clarification are extremely important for patients and their families, because the underlying disease and the surgery can affect their activities of daily living. The instruments used in this study were important to verify the effectiveness of treatment and the patients’ responses to it, as well as determining the individual needs, even though the service was performed in a group setting. The mean age of the study population was 42.3 years, and there was a high incidence of incapacity for work (68%) with disability insurance. This corroborating the work by Steenstra et al.23 Regarding sex, the majority of the patients were female. In the study by Heyer et al.,12 these levels were similar, but the work by Teixeira et al found different levels.24 g y y g Regarding fear of movement, this is a risk factor for increased pain, disability and reduced physical function in patients after back surgery.17 In this study, performing exercises, stretching, ergonomic counseling and cognitive-behavioral approach led to the groups achieving better muscular and aerobic conditioning, and encouraged the patients to perform regular exercises, improving their functional performance. In the study by Archer et al.,13 the researchers conducted a multicenter pilot study, using a similar approach in postoperative groups with pain, obtaining satisfactory results. Given this context, we noted that the use of targeted pro- grams to improve the quality of life and functionality has practical implications for the integral management of these patients. In the present study, body weight was above the anthropometric parameters. Several studies have identified increased weight as an important risk factor for increasing pain before or after surgery.20,21 In relation to socioeconomic situation, the most prevalent in this study is consistent with the work by Martinez et al.14 The results report that educational work should be undertaken in this social group, since it provides organizational behavior changes (structural) able to benefit the health of this population. CONTRIBUTIONS OF THE AUTHORS: Each author made an individual and significant contribution to the development of the manuscript. JENF, AMRC, CEDR and GMAF collected and evaluated the clinical data, MRIM, MHAAF, LCD and RSJ evaluated the data from the statistical analysis, performed the literature review, revised the manuscript, and contributed to the intellectual concept of the study. Coluna/Columna. 2017;16(1):48–51 REFERENCES Uso de questionários para avaliar a multidimensionalidade e a qualidade de vida do fibromiálgico. Rev Bras Reumatol. 2012;52(1):21-6. 22. Siqueira FB; Teixeira- Salmela LF,Magalhães LC. Análise das propriedades psicométricas da versão brasileira da escala tampa de cinesiofobia. Acta Ortop. Bras. 2007;15(1)19-24. 10. Hussain A, Erdek M. Interventional pain management for failed back surgery syndrome. Pain Pract. 2014;14(1):64-78. 23. Steenstra IA, Franche RL, Furlan AD, Amick B 3rd, Hogg-Johnson S. The Added Value of Collecting Information on Pain Experience When Predicting Time on Benefits for Injured Workers with Back Pain. J Occup Rehabil. 2016;26(2):117-24. g p g Workers with Back Pain. J Occup Rehabil. 2016;26(2):117-24. 11. Nijs J, Meeus M, Cagnie B, Roussel NA, Dolphens M, Van Oosterwijck J, et al. A modern neuroscience approach to chronic spinal pain: combining pain neuroscience education with cognition-targeted motor control training. Phys Ther. 2014;94(5):730-8. 24. Teixeira MJ, Atento LT, Garcia OG, Fonoff ET, Paica WS. Síndrome dolorosa pós- laminectomia: estudo descritivo da abordagem terapêutica em 56 pacientes. Rev Assoc Med Bras. 2011;57(3):286-9. 12. Heyer EJ, Sharma R, Winfree CJ, Mocco J, McMahon DJ, McCormick PA, et al. Severe pain confounds neuropsychological test performance. J Clin Exp Neuropsychol. 2008;22(5):633–9. 25. Al Kaisy A, Pang D, Desai MJ, Pries P, North R, Taylor RS. Failed back surgery syndrome: who has failed? Neurochirurgie. 2015:61(1);S6-S14. 26. Meinhart NT, McCaffery M. Pain: a nursing approach to assessment and analysis. New York: Appleton-Century; 1983. 13. Archer KR, Coronado RA, Haug CM, Vanston SW, Devin CJ, Fonnesbeck CJ, et al. A comparative effectiveness trial of postoperative management for lumbar spine surgery: changing behavior through physical therapy (CBPT) study protocol. BMC Muscul Disord. 2014;15:325. 27. McGregor AH, Hughes SP. The evaluation of the surgical management of nerve root compression in patients with low back pain: Part 2: patient expectations and satisfaction. Spine (Phila Pa 1976). 2002;27(13):1471-6. 27. McGregor AH, Hughes SP. The evaluation of the surgical management of nerve root compression in patients with low back pain: Part 2: patient expectations and satisfaction. Spine (Phila Pa 1976). 2002;27(13):1471-6. ; 14. Martinez AP, Martinez JE, Lanza LB. Há correlação entre classe social e a prática de Coluna/Columna. 2017;16(1):48–51
https://openalex.org/W4212987229	https://hal.archives-ouvertes.fr/hal-03607931/file/deVerneilBiogeosciences2022.pdf	English	null	Evaluating the Arabian Sea as a regional source of atmospheric CO<sub>2</sub>: seasonal variability and drivers	Biogeosciences	2,022	cc-by	21,427	Evaluating the Arabian Sea as a regional source of atmospheric CO2: seasonal variability and drivers Alain de Verneil, Zouhair Lachkar, Shafer Smith, Marina Lévy To cite this version: Alain de Verneil, Zouhair Lachkar, Shafer Smith, Marina Lévy. Evaluating the Arabian Sea as a regional source of atmospheric CO2: seasonal variability and drivers. Biogeosciences, 2022, 19, pp.907 - 929. 10.5194/bg-19-907-2022. hal-03607931v2 To cite this version: Alain de Verneil, Zouhair Lachkar, Shafer Smith, Marina Lévy. Evaluating the Arabian Sea as a regional source of atmospheric CO2: seasonal variability and drivers. Biogeosciences, 2022, 19, pp.907 - 929. 10.5194/bg-19-907-2022. hal-03607931v2 Evaluating the Arabian Sea as a regional source of atmospheric CO2: seasonal variability and drivers Alain de Verneil, Zouhair Lachkar, Shafer Smith, Marina Lévy Correspondence: Alain de Verneil (ajd11@nyu.edu) Correspondence: Alain de Verneil (ajd11@nyu.edu) Received: 1 February 2021 – Discussion started: 17 February 2021 Revised: 24 November 2021 – Accepted: 8 December 2021 – Published: 14 February 2022 Received: 1 February 2021 – Discussion started: 17 February 2021 Revised: 24 November 2021 – Accepted: 8 December 2021 – Published: 14 February 2022 Abstract. The Arabian Sea (AS) was conﬁrmed to be a net emitter of CO2 to the atmosphere during the interna- tional Joint Global Ocean Flux Study program of the 1990s, but since then few in situ data have been collected, leav- ing data-based methods to calculate air–sea exchange with fewer and potentially out-of-date data. Additionally, coarse- resolution models underestimate CO2 ﬂux compared to other approaches. To address these shortcomings, we employ a high-resolution (1/24◦) regional model to quantify the sea- sonal cycle of air–sea CO2 exchange in the AS by focusing on two main contributing factors, pCO2 and winds. We com- pare the model to available in situ pCO2 data and ﬁnd that uncertainties in dissolved inorganic carbon (DIC) and total alkalinity (TA) lead to the greatest discrepancies. Neverthe- less, the model is more successful than neural network ap- proaches in replicating the large variability in summertime pCO2 because it captures the AS’s intense monsoon dynam- ics. In the seasonal pCO2 cycle, temperature plays the ma- jor role in determining surface pCO2 except where DIC de- livery is important in summer upwelling areas. Since sea- sonal temperature forcing is relatively uniform, pCO2 dif- ferences between the AS’s subregions are mostly caused by geographic DIC gradients. We ﬁnd that primary productiv- ity during both summer and winter monsoon blooms, but also generally, is insufﬁcient to offset the physical delivery of DIC to the surface, resulting in limited biological con- trol of CO2 release. The most intense air–sea CO2 exchange occurs during the summer monsoon when outgassing rates reach ∼6 molCm−2 yr−1 in the upwelling regions of Oman and Somalia, but the entire AS contributes CO2 to the atmo- sphere. Despite a regional spring maximum of pCO2 driven by surface heating, CO2 exchange rates peak in summer due to winds, which account for ∼90 % of the summer CO2 ﬂux variability vs. 6 % for pCO2. In comparison with other es- timates, we ﬁnd that the AS emits ∼160 TgCyr−1, slightly higher than previously reported. Altogether, there is 2× vari- ability in annual ﬂux magnitude across methodologies con- sidered. Correspondence: Alain de Verneil (ajd11@nyu.edu) Future attempts to reduce the variability in estimates will likely require more in situ carbon data. Since summer monsoon winds are critical in determining ﬂux both directly and indirectly through temperature, DIC, TA, mixing, and primary production effects on pCO2, studies looking to pre- dict CO2 emissions in the AS with ongoing climate change will need to correctly resolve their timing, strength, and up- welling dynamics. 1 Introduction The global ocean represents a major reservoir of inorganic carbon on the planet’s surface (40× atmosphere) and up to the present has on average acted to uptake ∼23 % of the 11 Gt excess anthropogenic carbon (Friedlingstein et al., 2020; Ciais et al., 2013; Khatiwala et al., 2009). The Ara- bian Sea (AS) is a region of the ocean that has been found to naturally release CO2 to the atmosphere (∼90 MtCyr−1; Sarma et al., 1998), mitigating the ocean’s role in moderat- ing atmospheric CO2 accumulation. While the AS as a re- gional basin is considered too small to greatly impact global budgets of air–sea CO2 exchange (Naqvi et al., 2005), it at- Evaluating the Arabian Sea as a regional source of atmospheric CO2: seasonal variability and drivers Alain de Verneil1,4, Zouhair Lachkar1,4, Shafer Smith2,4, and Marina Lévy3 1Center for Prototype Climate Modeling, New York University Abu Dhabi, Abu Dhabi, UAE 2Courant Institute of Mathematical Sciences, New York University, New York, USA 3Sorbonne Université (CNRS/IRD/MNHN), LOCEAN-IPSL, Paris, France 4Arabian Center for Climate and Environmental Sciences, New York University Abu Dhabi, Abu Dhabi, UAE Alain de Verneil1,4, Zouhair Lachkar1,4, Shafer Smith2,4, and Marina Lévy3 1Center for Prototype Climate Modeling, New York University Abu Dhabi, Abu Dhabi, UAE 2Courant Institute of Mathematical Sciences, New York University, New York, USA 3Sorbonne Université (CNRS/IRD/MNHN), LOCEAN-IPSL, Paris, France 4Arabian Center for Climate and Environmental Sciences, New York University Abu Dhabi, Abu Dhabi, UAE HAL Id: hal-03607931 https://hal.science/hal-03607931v2 Submitted on 14 Mar 2022 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Research article Biogeosciences, 19, 907–929, 2022 https://doi.org/10.5194/bg-19-907-2022 © Author(s) 2022. This work is distributed under the Creative Commons Attribution 4.0 License. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux with pCO2 reported in the Surface Ocean Carbon Atlas (SO- CAT; Bakker et al., 2016; Pfeil et al., 2013). with pCO2 reported in the Surface Ocean Carbon Atlas (SO- CAT; Bakker et al., 2016; Pfeil et al., 2013). tracts attention because high rates of air–sea CO2 ﬂux 7– 33 molCm−2 yr−1 and values > 700 µatm of partial pressure of CO2, or pCO2, have been observed there, in addition to unique features such as the world’s thickest oxygen mini- mum zone (OMZ) (Morrison et al., 1999; Acharya and Pani- grahi, 2016; Lachkar et al., 2016) and corresponding carbon maximum zone (CMZ) (Paulmier et al., 2011). Where data are sparse in the AS, numerical circulation models have been used to complement the lack of spatiotem- poral coverage. These models ﬁll the domain with their own estimates of carbon variables, such as pCO2, while also pro- viding detailed information on the factors affecting them (for example, DIC, temperature, biological productivity, etc.). For example, in the wake of the JGOFS expeditions, the synthesis study of Sarma et al. (2003) used a numerical model to ex- amine biological and chemical aspects of the annual carbon budget in the central and eastern AS. Further studies focus on other aspects over different timescales, such as intrasea- sonal pCO2 variability due to temperature vs. DIC (Valsala and Murtugudde, 2015) or decadal trends in pH (Sreeush et al., 2019a). These approaches, without more in situ data, are the best estimates we have of the current AS carbon sys- tem’s behavior. Therefore, it is incumbent that these mod- els are vigorously validated against what precious few data exist. The need to reduce uncertainty is further emphasized when modeled carbon chemistry quantities are utilized as a proxy for other things. For example, a recent modeling study in the AS found that pCO2 could be used to indicate com- munity compensation depth, which reﬂects the complicated balance between primary production and respiration in the water column (Sreeush et al., 2019b). As a result, the possi- bility exists to propagate uncertainties beyond carbon chem- istry. However, these AS modeling studies compare output to established climatologies, such as TK09, which are coarse in spatial resolution and smooth out unique features of the AS such as coastal upwelling, although some studies have begun using ARGO ﬂoat proﬁles for model validation (Chakraborty et al., 2018). A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux ( ) ( , ) The role of the AS as a region of net CO2 emission, while suspected for decades (Keeling, 1968; Naqvi et al., 1993), was more ﬁrmly established with observations con- ducted under the international collaborative efforts of the Joint Global Ocean Flux Study (JGOFS) program during the 1990s (Sarma et al., 1998; Millero et al., 1998a; Goyet et al., 1998b; Naqvi et al., 2005); see Smith (2005) and the accom- panying Progress in Oceanography special issue for greater context. Conducted over several years, a major focus was to sample over the particularly strong seasonal monsoon cy- cle present in the AS, complete with surface current rever- sals, coastal upwelling, and intense phytoplankton blooms (Schott and McCreary Jr, 2001; Kumar et al., 2001; Lévy et al., 2007). JGOFS carbon data were ﬁrst used to create linear statistical models, which were then extrapolated over a greater region of the AS to produce larger-scale estimates of seasonal CO2 ﬂux showing emission to the atmosphere (Sabine et al., 2000; Sarma, 2003; Bates et al., 2006). JGOFS data still represent the greatest source of data for current de facto standard global products, such as Takahashi et al. (2009) (hereafter TK09), who produced a global climatol- ogy of pCO2 and CO2 ﬂux gridded onto a 4◦× 5◦grid us- ing a horizontal advection–diffusion scheme. In recent years, neural networks have been applied instead of simpler statisti- cal models to likewise produce global climatologies, such as Landschützer et al. (2015) (hereafter L15) on an increased- resolution 1◦× 1◦grid. All these different methodologies, al- though of differing sophistication, still rely on the availability of in situ data. Despite the wealth of information that models provide, they have their own weaknesses. In a review of CO2 ﬂux esti- mates from various independent methodologies, Sarma et al. (2013) found that coupled ocean biogeochemical models un- derestimated the air–sea CO2 ﬂux in the AS. The underesti- mate was attributed to poor resolution of monsoonal currents, speciﬁcally near the coasts of Oman and Somalia. The need for sufﬁcient resolution of monsoon and upwelling currents is underscored by the roles that small-scale horizontal (Ma- hadevan et al., 2004) and vertical (Mahadevan et al., 2011; Resplandy et al., 2019) currents can play in advecting carbon. Additionally, Sarma et al. Published by Copernicus Publications on behalf of the European Geosciences Union. Published by Copernicus Publications on behalf of the European Geosciences Union. 908 https://doi.org/10.5194/bg-19-907-2022 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux For all the comparison datasets, air–sea CO2 ﬂux is calculated from monthly values. The 1pCO2 values are calculated using Keeling curve data (downloaded from https://gml.noaa.gov/ccgg/trends/gl_data. html, downloaded 1 February 2022) of atmospheric xCO2 for the respective calibrated year of each dataset. The same cli- matological winds as used in the model (Sect. 2.2) are ap- plied to the pCO2 products. The gridded product TK09 is chosen because previous modeling studies in the AS use it as validation (see “Introduction”). The L15 climatology, while based upon the same in situ data mentioned above, represents different processing methodologies and, as a high-resolution, global pCO2 dataset, also serves to provide independent con- text to the model validation. pCO2 is also calculated from DIC and TA provided by the statistical ﬁts to JGOFS data by Sarma (2003) and to the gridded GLODAP climatolog- ical product. The statistical ﬁts of Sarma (2003) are used twice, ﬁrst using model SST, SSS, and chl a, and second with World Ocean Atlas (WOA) 2009 SST and SSS with SeaW- iFS chl a. GLODAP-derived pCO2 also uses WOA 2009 SST and SSS applied to the annual DIC and TA values. Calcula- tions of pCO2 are performed using the CO2SYS software package (Van Heuven et al., 2011). Since all calculations are conducted at the near surface, differences between this soft- ware suite and Orr and Epitalon (2015) are minimal. Fur- thermore, for air–sea CO2 ﬂux intercomparison purposes, all pCO2 values except for TK09 are interpolated to the same 1◦× 1◦grid already shared by GLODAP, WOA, and L15. Due to the model’s higher resolution, the re-gridding pro- cess reduces the area covered, consequently lowering the to- tal model CO2 ﬂux quoted in later sections of this study. For this study, we choose to focus on the seasonal cycle due to the strength of the monsoon in the AS and because it is resolved by the in situ data, although models suggest in- terannual (Valsala and Maksyutov, 2013; Valsala et al., 2020) and intraseasonal (Valsala and Murtugudde, 2015) variability exists. The study begins with a description of pCO2 datasets used, along with the model conﬁguration and methods of analysis in Sect. 2. Following this in Sect. 3 is a descrip- tion of the model validation and results, with discussion in Sect. 4. We conclude in Sect. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux role of the AS as a CO2 source by quantifying air–sea CO2 ﬂux with a targeted approach. First, by employing a higher- resolution regional numerical model of the AS carbon sys- tem, monsoonal and upwelling currents will be sufﬁciently resolved. Furthermore, model validation will use raw data, not a smoothed climatological product, to evaluate the model air–sea CO2 ﬂux. Quantiﬁcation of seasonal air–sea CO2 ﬂux will focus on the contributing factors of 1pCO2, the differ- ence in seawater and atmospheric pCO2, and wind. In par- ticular, the role of sea surface temperature (SST), sea sur- face salinity (SSS), DIC, and TA in determining the seasonal cycle of pCO2 will be investigated for the entire domain of the AS, as well as its spatial heterogeneity within the AS. A further budget analysis of surface DIC compares the physi- cal and biological mechanisms governing carbon sources and sinks, such as advection and mixing vs. biological production and respiration, among others. The relative impacts of pCO2 and winds upon the seasonal cycle of CO2 ﬂux are also com- pared, culminating in a meta-analysis of the model’s CO2 ﬂux estimates relative to alternative approaches. SOCAT data are unreported. SOCAT fugacity (fCO2) values are converted to pCO2 and mole fraction (xCO2) using re- ported SST and SSS data included in the products using rou- tines from the CO2SYS software package (Van Heuven et al., 2011). The anthropogenic effect of increasing surface pCO2 is removed by calculating a ﬁt linear trend of 2 µatm yr−1, slightly higher than ≈1.5 seen in Tjiputra et al. (2014). The pCO2 values are calibrated to the year 2005, the representa- tive year used for the model’s atmospheric xCO2. The year 2005 is chosen for the model’s xCO2 concentration because it is the end of the historical period for the Intergovernmen- tal Panel of Climate Change (IPCC) models used in its ﬁfth report published 2014. The earliest SOCAT data comes from 1962, and different databases used in this study stem from similarly different time spans. As a result, we assume there is a baseline seasonal cycle of pCO2 and air–sea CO2 ﬂux which has held stable over the past decades. p Alternative pCO2 products are used for comparison pur- poses. A complete list of these datasets and their char- acteristics is provided in Table 1. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux 5 with perspectives and rec- ommendations regarding future studies of pCO2 and air–sea CO2 ﬂux in the AS. 2.1 The pCO2 data In this study, sea surface pCO2 is used as the primary in situ data for model validation. Whereas models favor DIC and TA (Wolf-Gladrow et al., 2007), shipboard pCO2 can be mea- sured underway, and hence there are more observations avail- able. Additionally, since model pCO2 is calculated from DIC and TA (see Sect. 2.2), pCO2 measurements act as an inde- pendent dataset. Here, pCO2 validation stems from in situ un- gridded data merged from SOCAT v. 2019 (downloaded from https://www.socat.info/index.php/version-2019/, last access: 3 September 2019) and the Lamont–Doherty Earth Obser- vatory (LDEO) surface pCO2 database (Takahashi et al., 2019). Both databases aggregate all available in situ sur- face pCO2 data, including JGOFS. SOCAT and LDEO con- tain > 180 000 and ∼90 000 data points on the AS, respec- tively. SOCAT has more data because it includes multiple methodologies. As a result, SOCAT data are preferred, and LDEO observations are included for the years 1980–81 when A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux (2013) found that the peak of air– sea CO2 ﬂux observed in boreal summer occurred slightly out of phase, with models leading observations by over a month in the AS. Finally, the modeled pCO2 in the AS found a springtime maximum not seen in the observations based on the data from TK09. Clearly, an effort must be made to establish whether these discrepancies are residual effects of low resolution, endemic to models generally, or indicative of a real pattern that suggests future concerted in situ sampling. The wealth of information provided by the JGOFS expe- ditions has been invaluable for understanding the AS, but there has been little subsequent in situ sampling in the re- gion, as has been previously remarked (Hood et al., 2016). For example, in the Global Ocean Data Analysis Project v2 (GLODAP; Olsen et al., 2019) database, there are no re- ported observations in the AS of two important carbon vari- ables, dissolved inorganic carbon (DIC) and total alkalinity (TA), more recent than 1998, with a similar > 98 % of data predating 2000 for pCO2. Thus, the global products of TK09 and L15 are based upon conditions in the AS from 20 years ago. Since quantities like surface pCO2 concurrently trend with rising atmospheric CO2 concentration (Tjiputra et al., 2014), the dearth of recent sampling means that uncertainty in the AS’s carbon system will only grow with time. The gap in data collection also means that the AS is proportionally underrepresented in global datasets: whereas the AS is 2 % of the ocean surface, DIC and TA measurements in the AS are < 1 % of the GLODAP ensemble, which is also the case Considering the challenges speciﬁc to studying the AS carbon cycle, in this paper we aim to put into context the Biogeosciences, 19, 907–929, 2022 https://doi.org/10.5194/bg-19-907-2022 909 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Table 1. Summary of pCO2 datasets used in this study. Included is whether the product is gridded and, if so, its spatial and temporal resolution. Reference year (Ref. year) indicates the year from which Keeling atmospheric xCO2 values are used to calculate CO2 ﬂux. Purpose designates use case within the article. pCO2 calculated indicates whether product provides pCO2 (no) or whether pCO2 was calculated using DIC, TA, temperature, salinity, and possibly chl a (yes). Table 1. Summary of pCO2 datasets used in this study. Included is whether the product is gridded and, if so, its spatial and temporal resolution. Reference year (Ref. year) indicates the year from which Keeling atmospheric xCO2 values are used to calculate CO2 ﬂux. Purpose designates use case within the article. pCO2 calculated indicates whether product provides pCO2 (no) or whether pCO2 was calculated using DIC, TA, temperature, salinity, and possibly chl a (yes). Dataset Gridded (yes/no), resolution Ref. year (xCO2) Domain Purpose pCO2 calculated Reference Surface Ocean Carbon Atlas (SOCAT) No, NA 2005 Global Model pCO2 validation No Bakker et al. (2016) Lamont–Doherty Earth Obser- vatory pCO2 database (LDEO) No, NA 2005 Global Model pCO2 validation No Takahashi et al. (2019) Takahashi 2009 (TK09) Yes, 4◦× 5◦, monthly 2005 Global Air–sea CO2 ﬂux estimate No Takahashi et al. (2009) Landschützer 2015 (L15) Yes, 1◦× 1◦, monthly 2001 Global pCO2 comparison and air–sea CO2 ﬂux estimate No Landschützer et al. (2015) Sarma statistical model, T /S/chl a from model (ROMS) Yes, 1/24◦ (interpolated to 1◦), seasonal 1995 AS north of 10◦N Air–sea CO2 ﬂux estimate Yes Sarma (2003) Sarma statistical model, World Ocean Atlas T/S, SeaWiFS chl a Yes, 1◦× 1◦, seasonal 1995 AS north of 10◦N Air–sea CO2 ﬂux estimate Yes Sarma (2003) GLODAP DIC/TA, World Ocean Atlas T/S Yes, 1◦× 1◦, annual 2002 Global Air–sea CO2 ﬂux estimate Yes Olsen et al. (2019) NA: not available Figure 1. Vertically integrated net primary production in the Ara- bian Sea (gCm−2 yr−1) from the VGPM algorithm (Behrenfeld and Falkowski, 1997) for SeaWiFS data (years 1997–2010) (a, c) and model output (b, d) for summer (JJAS, a, b) and winter (DJFM, c, d) monsoons. White boxes in (b, d) denote regions of analysis in the paper. order upstream biased algorithm to reduce spurious mixing (Marchesiello et al., 2009). The K-proﬁle parameterization (KPP; Large et al., 1994) for vertical mixing is used. 2.2 Model details and setup The model we use is the Regional Ocean Modeling Sys- tem Adaptive Grid Reﬁnement In Fortran (ROMS-AGRIF) https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 910 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Table 2. List of parameters and their values used in the biogeochemical model. Table 2. List of parameters and their values used in the biogeochemical model. ameters and their values used in the biogeochemical model. Parameter Value Kw, seawater light attenuation 0.04 m−1 KChl, chl a light attenuation 0.024 m−1 Palpha, initial slope of P -I curve 1.0 Wm−2 d−1 C : NP, carbon-to-nitrogen ratio of phytoplankton 6.625 molCmolN−1 C : NZ, carbon-to-nitrogen ratio of zooplankton 6.625 molCmolN−1 O2 : NO3, oxygen-to-nitrogen ratio for nitrate uptake 9.375 molO2 molNO3−1 O2 : NH4, oxygen-to-nitrogen ratio for ammonium uptake 7.375 molO2 molNO3−1 N : Cden, nitrate-to-DIC ratio for denitriﬁcation 0.8 molNO3 molDIC−1 O2 den, oxygen threshold for denitriﬁcation 4.0 mmolO2 m−3 RCaCO3, ratio of calcium carbonate precipitation to production 0.07 molCaCO3 molC−1 2m, maximum chl a to carbon ratio 1.3538 mgchl amgC−1 KNO3, half-saturation rate for nitrate uptake 0.75 mmolNm−3 KNH4, half-saturation rate for ammonium uptake 0.5 mmolNm−3 µnitr, nitriﬁcation rate 0.05 d−1 µP, phytoplankton mortality rate 0.072 d−1 Gmax, maximum zooplankton growth rate 0.6 d−1 β, zooplankton assimilation efﬁciency 0.75 Kphy, half-saturation rate for zooplankton ingestion 1.0 d−1 µexc, zooplankton excretion rate 0.1 d−1 µZmor, zooplankton mortality rate 0.025 d−1 Zgam, fraction of sloppy feeding to fecal pellets 0.33 µSD, small detritus breakdown rate to ammonium 0.03 d−1 µagg, speciﬁc aggregation rate of small detritus and phytoplankton 0.005 mmolN−1 d−1 µLD, large detritus breakdown 0.01 d−1 Tdissol, water column dissolution rate of calcium carbonate 0.0057 d−1 Tsedremin, remineralization rate in sediments 0.003 d−1 wP, phytoplankton sinking velocity 0.5 md−1 wSD, small detritus sinking velocity 1.0 md−1 wLD, large detritus sinking velocity 10.0 md−1 wCaCO3, vertical sinking speed of calcium carbonate 20 md−1 Heat ﬂux, evaporation and precipitation, and restoring SSS are provided by the Comprehensive Ocean-Atmosphere Data Set (COADS; da Silva et al., 1994). SST forcing is provided by a monthly climatology of Pathﬁnder data from 1985 to 1997 (Casey and Cornillon, 1999). Wind stress is produced using the QuikSCAT/SCOW monthly climatology from 1999 to 2009 (Risien and Chelton, 2008). Tracer values for the ini- tial conditions and the boundaries are given by WOA 2009 for temperature, salinity, NO3, and oxygen. Horizontal veloc- ities u,v for initial and boundary conditions derive from the Simple Ocean Data Assimilation (SODA) analysis (Carton and Giese, 2008). A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux The model domain spans from 5.3◦S to 30.5◦N and from 33 to 78.1◦E (Fig. 1). For the sake of comparison with Sarma et al. (2013), we will present the region north of the Equator and exclude the Red Sea and Arabian Gulf. The model’s horizon- tal resolution is 1/24◦, resulting in ∼5 km horizontal grid spacing. p g Coupled to the hydrodynamic model is a nitrogen-based biogeochemical model with two components for nutrients, nitrate and ammonium, with one phytoplankton, one zoo- plankton, and two detrital pools (Gruber et al., 2006). Bio- logical parameters for the model are the same as those used in Gruber et al. (2011). A carbon module is also applied to the model with the state variables of DIC, TA, and calcium carbonate (CaCO3) (Gruber et al., 2012; Hauri et al., 2013; Lachkar and Gruber, 2013). In addition to the usual phys- ical transport and mixing, CaCO3 is allowed to vertically sink at 20 md−1. The chosen sinking rate is a simpliﬁca- tion in that it does not include the faster rates observed for foraminifera shells (Curry et al., 1992), which as a biological group are not resolved by the biological model due to numer- ical constraints. Organic carbon is linked to organic nitrogen through the Redﬁeld ratio 106 : 16. DIC is altered by air–sea CO2 ﬂux, primary production, respiration, remineralization, and dissolution/precipitation of CaCO3. TA changes with the removal and creation of nitrate (NO3), including nitriﬁca- tion and denitriﬁcation, as well as dissolution/precipitation of CaCO3. The amount of CaCO3 precipitation is linked to primary production through a constant ratio of 0.07, meaning Figure 1. Vertically integrated net primary production in the Ara- bian Sea (gCm−2 yr−1) from the VGPM algorithm (Behrenfeld and Falkowski, 1997) for SeaWiFS data (years 1997–2010) (a, c) and model output (b, d) for summer (JJAS, a, b) and winter (DJFM, c, d) monsoons. White boxes in (b, d) denote regions of analysis in the paper. version 3.1.1. (Shchepetkin and McWilliams, 2005). Previ- ously used in the AS by Lachkar et al. (2016), the model is a free-surface primitive equation model, with a sigma and curvilinear grid for the vertical and horizontal dimen- sions, respectively. ROMS implements a forward–backward time-stepping algorithm with split baroclinic and barotropic modes. The advection of tracers uses a rotated-split third- https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 911 https://doi.org/10.5194/bg-19-907-2022 The air–sea ﬂux in the model is calculated using FCO2 = K0α pCOsea 2 −pCOair 2 = K0α1pCO2, (1) FCO2 = K0α pCOsea 2 −pCOair 2 1pCO2 ≈∂pCO2 ∂DIC 1DIC + ∂pCO2 ∂TA 1TA + ∂pCO2 ∂SST 1SST + ∂pCO2 ∂SSS 1SSS, (3) (1) = K0α1pCO2, where K0 is the solubility determined by temperature and salinity (Weiss, 1974), α is the CO2 piston velocity with a quadratic wind speed dependence (Wanninkhof, 1992), and the difference in ocean and atmosphere pCO2, 1pCO2, is ar- ranged so that the ﬂux convention is positive outward from the ocean. The choice of Wanninkhof (1992) for the solubil- ity parameterization is for direct comparison with previous modeling studies (see “Introduction”), despite the fact that more recent formulations are available, such as Wanninkhof (2014). The objective being to characterize seasonal anoma- lies of air–sea CO2 ﬂux, here we use a Reynolds decompo- sition. Brieﬂy, a Reynolds decomposition takes a time series and divides it into a temporal mean and ﬂuctuating compo- nent. When applied correctly, multiple terms can be produced in isolation showing their ﬂuctuating contribution to the to- tal. Noting that temperature effects upon solubility (K0) and piston velocity (α) approximately cancel, meaning that their product mostly reﬂects wind forcing, we have the following where K0 is the solubility determined by temperature and salinity (Weiss, 1974), α is the CO2 piston velocity with a quadratic wind speed dependence (Wanninkhof, 1992), and the difference in ocean and atmosphere pCO2, 1pCO2, is ar- ranged so that the ﬂux convention is positive outward from the ocean. The choice of Wanninkhof (1992) for the solubil- ity parameterization is for direct comparison with previous modeling studies (see “Introduction”), despite the fact that more recent formulations are available, such as Wanninkhof (2014). The objective being to characterize seasonal anoma- lies of air–sea CO2 ﬂux, here we use a Reynolds decompo- sition. Brieﬂy, a Reynolds decomposition takes a time series and divides it into a temporal mean and ﬂuctuating compo- nent. When applied correctly, multiple terms can be produced in isolation showing their ﬂuctuating contribution to the to- tal. 2.4 Analysis of air–sea CO2 ﬂux, pCO2, and DIC variability The proximate variables that affect pCO2 change in the model are DIC, TA, SST, and SSS. Following previous stud- ies (Lovenduski et al., 2007; Turi et al., 2014), we use a ﬁrst- order Taylor expansion to decompose pCO2 into contribu- tions from these four, neglecting contributions from nutrients (phosphate and silicate). Initially, the decomposition would follow the form 2.3 Domains of analysis In this study we focus on six distinct regions (Fig. 1). The ﬁrst, the entire analysis domain, is the AS north of the Equa- tor. The upwelling regions of the Omani and Somalian coasts are included separately to focus on the summer monsoon im- pact of enhanced DIC but also enhanced biological produc- tivity (Schott and McCreary Jr, 2001). The Omani region be- gins at the coast and extends 300 km outward. The Soma- lian region begins near 3.8◦N and extends north to the tip of the Horn of Africa, with an eastern extension to 58.6◦E so as to encompass the region known as the Great Whirl (Vic et al., 2014), shown to be important for air–sea exchange in previous studies (Valsala and Murtugudde, 2015). The north region is deﬁned by a rectangle from 21◦N, 59.4◦E to 26.5◦N, 69.5◦E, encompassing the northern part of the AS where the winter monsoon’s primary productivity is most in- tense (Kumar et al., 2001). An oligotrophic region represent- ing the central AS, which has less productivity and chloro- phyll a on average (Fig. 1), is deﬁned by a rectangle from 3.3◦N, 61.31◦E to 17◦N, 70.8◦E. The last region, covering the western coast of India, extends from the coastline 100 km offshore. (2) cross terms where ′ indicates an anomaly and x is a 5-year average of variable x, which are calculated at each grid point. The 5- year average is necessary for exact closure in the Reynolds decomposition. F ′ CO2 is the seasonal ﬂux anomaly, with groupings based on wind anomalies (K0α)′, 1pCO2′ anoma- lies, and cross-terms involving both. The winds in this study are prescribed, so uncertainty in air–sea ﬂux stems from pCO2. The SOCAT protocol assigns a minimum uncertainty of 2 µatm to observations. Using the average SST and SSS from the SOCAT observations, the solubility change is 2.68 × 10−2 mmolCm−3 µatm−1. Wind speeds of 1, 5, and 10 ms−1 will then produce a shift of 0.0018, 0.0443, and 0.177 molCm−2 yr−1, respectively. The model presents a median value of 1.28 molCm−2 yr−1 with median winds of 5 ms−1, so therefore the baseline uncer- tainty in air–sea CO2 is ∼3.5 %. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux xCO2 values are set to 380 ppm, equivalent to 2005 levels, with an annual sinusoidal perturbation of 2.9 ppm. arrangement for the decomposition of ﬂux anomalies (Doney et al., 2009b): F ′ CO2 = (K0α)′1pCO2 \| {z } wind + (K0α)1pCO2′ \| {z } pCO2 + (K0α)′(1pCO2)′ −(K0α)′1pCO2′ \| {z } cross terms , (2) A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Initial and boundary conditions for DIC and TA come from GLODAP from 300 m down to the bot- tom. Surface TA was calculated using the relations from Lee et al. (2006), and the corresponding DIC was calculated us- ing WOA phosphate, silicate, T , and S values along with L15 pCO2. DIC and TA values between the surface and 300 m are calculated using density weighting. The model is spun up for 30 years, with 5 additional years for analysis. Atmospheric 0.07 moles of CaCO3 are produced for each mole of organic carbon. The dissolution rate is a constant 0.0057 d−1 in the water column and 0.002 d−1 in the sediments. Surface ﬂuxes of DIC and TA due to evaporation, precipitation, and river in- put are included as virtual ﬂuxes proportional to SSS forcing. Inside the module, surface carbon chemistry is calculated us- ing routines from the Ocean Carbon-Cycle Model Intercom- parison Project (OCMIP) carbonate chemistry routines (http: //ocmip5.ipsl.jussieu.fr/OCMIP/phase3/simulations/, last ac- cess: 1 February 2022). Carbon chemistry coefﬁcients used here include K1 and K2 CO2 dissociation from Millero (1995) and original data from Mehrbach et al. (1973) which was reﬁt by Dickson and Millero (1987). A summary of the biological parameters used in the biogeochemical model is provided in Table 2. The model is run with 360 d years and interpolated, cli- matologically averaged monthly forcing. The different cli- matological products derive from datasets spanning slightly different periods, and so here we assume that the dynamics represented within them have not changed in the time since. https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 912 2.4.1 Air–sea CO2 variability The air–sea ﬂux in the model is calculated using https://doi.org/10.5194/bg-19-907-2022 2.4.3 DIC budget ∂pCO2 ∂DIC 1DIC = ∂pCO2 ∂ SSS/S0DICs1 SSS/S0DICs = DICs S0 ∂pCO2 ∂DIC 1SSS + S S0 pCO2 ∂DIC1DICs. (5) Whereas the state variables of DIC, TA, SST, and SSS pro- vide the chemical context which determines carbon availabil- ity to potential air–sea ﬂux via pCO2, tracking the overall in- ventory of inorganic carbon (i.e., DIC) allows for the parsing of numerous source and sink processes governing the total amount of carbon reaching the surface. Beyond the biological processes impacting DIC as outlined in Sect. 2.2, the phys- ical processes impacting DIC are air–sea CO2 ﬂux, surface evaporation and precipitation, horizontal and vertical advec- tion, and horizontal and vertical mixing. In order to diagnose the relative importance of these terms (i.e., to weigh com- petition between upwelling circulation source and biological drawdown sink), we calculate the budget IDIC in a 3D vol- ume by integrating (5) Collectively, the 1SSS term in Eq. (5) and its counterpart in TA can be added to the original 1SSS term in Eq. (3) to represent all salinity effects in a “freshwater” term so that we now have the following (Turi et al., 2014): 1pCO2 ≈∂pCO2 ∂DICS 1DICS \| {z } 1pCO2DICs + ∂pCO2 ∂TAS 1TAS \| {z } 1pCO2TAs + ∂pCO2 ∂T 1T \| {z } 1pCO2T + ∂pCO2 ∂SSS 1SSS \| {z } 1pCO2SSS . (6) IDIC = ZZ A η Z −z(σ) J(x,y,z)dAdz (8) (8) (6) 1pCO2SSS with For the remainder of this paper, when discussing the re- sults of the Taylor series decomposition method, it will be understood that DIC and TA refer to DICs and TAs, and SSS will refer to the combined term. J = −PPNew+Reg −CaCO3prec-remin + Zooresp + Detremin \| {z } Biology −FAS \|{z} Air–Sea + Advx + Advy + Mixx + Mixy \| {z } Horz. Circ + Advz + Mixz \| {z } Vert. Circ + Evap-Precip \| {z } Forc , (9) J = −PPNew+Reg −CaCO3prec-remin + Zooresp + Detremin \| } Biology −FAS \|{z} Air–Sea + Advx + Advy + Mixx + Mixy \| {z } Horz. Circ + Advz + Mixz \| {z } Vert. Circ + Evap-Precip \| {z } Forc , (9) The contributions of DIC, TA, SST, and SSS to pCO2 vari- ability are used to construct maps and time series of pCO2 anomalies. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux et al., 2007): et al., 2007): 2.4.3 DIC budget In order to calculate the anomaly, 1pCO2 re- quires calculating both the 1 deviations of DIC, TA, T, and SSS, as well as partial derivatives. In this study, we calcu- late both temporal and spatial anomalies. To consider spatial variability, starting with annual means of pCO2, DIC, TA, SST, and SSS, an average value for the whole domain is cal- culated and removed from each grid point’s annual mean to get a 1 perturbation or anomaly. Similarly, for temporal vari- ability, with the monthly values of pCO2, DIC, TA, SST, and SSS at each grid point, the annual average at that grid point is removed to produce the monthly 1 perturbation/anomaly. Partial derivatives are approximated via centered differences. These are obtained by calculating pCO2 with slight devia- tions of DIC, TA, SST, and SSS from the mean value. Both positive and negative deviations are used to construct cen- tered differences, with deviation magnitude determined by Orr et al. (2018). For example to calculate the monthly pCO2 anomaly due to SST for a grid point with annual mean pCO2 of 430 µatm, annual mean SST of 24 ◦C, and monthly SST of 26 ◦C, the following equation is used: (9) which is the volume-speciﬁc ﬂux J of DIC in a given grid cell. PPNew+Reg is net community primary production scaled by the Redﬁeld ratio, CaCO3remin-prec is net CaCO3 precip- itation and remineralization, Zooresp is zooplankton respira- tion, and Detremin is remineralization of both detrital pools. All these terms are grouped together into “Biology” because they represent all biological processes. FAS is air–sea ﬂux, with a sign convention of positive outward. Advx is advective ﬂux in the x direction, with corresponding y and z compo- nents. Mixx is the x component of mixing ﬂux, again with y and z components. All x and y components of both advec- tive and mixing DIC ﬂuxes are grouped into horizontal cir- culation, with a similar grouping for vertical circulation in the z direction. Evap-Precip is the forced virtual ﬂux from evaporation and precipitation at the surface. A is the two- dimensional horizontal area to be considered, which in our study includes the entire domain but also the subregions of analysis. The bottom boundary of integration, −z(σ), is the sigma-layer depth at which integration starts, moving up to the free-moving surface η. The air–sea ﬂux in the model is calculated using Noting that temperature effects upon solubility (K0) and piston velocity (α) approximately cancel, meaning that their product mostly reﬂects wind forcing, we have the following (3) where 1pCO2 is the perturbation of pCO2 from a mean value, and the 1 terms for DIC, TA, SST, and SSS likewise express deviations from a prescribed value depending on whether the deviations are spatial or temporal in nature (see below). The coefﬁcients of the 1 terms are partial derivatives of pCO2 with respect to these variables, namely DIC, TA, SST, and SSS, and are calculated via centered differences described below. However, in order to control for salinity ef- fects on DIC and TA (Keeling et al., 2004), we normalize DIC and TA by the salinity S0 = 35 psu to create the vari- ables DICs = S0 DIC SSS and TAs = S0 TA SSS. (4) (4) Substituting these terms into Eq. (3), we can expand to produce, for example with DIC, the following (Lovenduski Substituting these terms into Eq. (3), we can expand to produce, for example with DIC, the following (Lovenduski Biogeosciences, 19, 907–929, 2022 https://doi.org/10.5194/bg-19-907-2022 913 2.4.3 DIC budget We chose to integrate the top ﬁve sigma layers of the model, corresponding to ∼20 m depth. This level was chosen because below this depth, annual cy- cles of IDIC begin to deviate from the surface DIC, which is our focus in this study of air–sea CO2 ﬂux. 1pCO2 ≈∂pCO2 ∂SST 1SST + ... ≈pCO2 24 + 1 × 10−4,... −pCO2 24 −1 × 10−4 2 · 1 × 10−4 · (26–24) + ..., (7) where 1 × 10−4 is the recommended SST deviation. where 1 × 10−4 is the recommended SST deviation. where 1 × 10−4 is the recommended SST deviation. https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 914 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Figure 3. Seasonal surface pCO2 (µatm) from data (left column, a, d, g, j) and the model (middle, b, e, h, k), as well as their differ- ences (right, c, f, i, l). Plots are arranged by season: winter monsoon (DJFM, a–c), spring intermonsoon (AM, d–f), summer monsoon (JJAS, g–i), and fall intermonsoon (ON, j–l). Figure 2. (a) Average surface in situ 1pCO2 (ppm), with proba- bility density function of all 1pCO2 values inset. 1pCO2 data are calculated in comparison to Keeling atmospheric pCO2 and then binned into a 1◦× 1◦grid. (b) Monthly distribution of in situ data sampling times, color-coded by sampling year. Figure 2. (a) Average surface in situ 1pCO2 (ppm), with proba- bility density function of all 1pCO2 values inset. 1pCO2 data are calculated in comparison to Keeling atmospheric pCO2 and then binned into a 1◦× 1◦grid. (b) Monthly distribution of in situ data sampling times, color-coded by sampling year. (a) Average surface in situ 1pCO2 (ppm), with proba Figure 2. (a) Average surface in situ 1pCO2 (ppm), with proba- bility density function of all 1pCO2 values inset. 1pCO2 data are calculated in comparison to Keeling atmospheric pCO2 and then binned into a 1◦× 1◦grid. (b) Monthly distribution of in situ data sampling times, color-coded by sampling year. 3.1 Model validation and pCO2 data-model comparisons The implementation of ROMS-AGRIF presented here has been used in previous studies of the AS (Lachkar et al., 2016). Model output of net primary productiv- ity (NPP) captures the summer monsoon highs near the upwelling regions of Oman and Somalia (model > 400 vs. data > 500 gCm−2 yr−1), with enhanced NPP in the north during the winter monsoon (model ∼300 vs. data > 400 gCm−2 yr−1) (Fig. 1). The model also captures the vertical distributions of temperature and salinity (Figs. S1 and S2 in the Supplement) with deviations from WOA around 1 oC and 0.2 psu. Depth proﬁles of nitrate, oxygen, DIC, and TA are similarly conserved (Figs. S3–S6 in the Sup- plement). Nitrate, DIC, and TA all show their usual nutrient- like proﬁles, while oxygen is its minimum within the OMZ. The deviations seen between in situ data and model output are greatest at depths less than 500 m. Deviations in near- surface NO3 (Fig. S3) can be large for intermediate values (5–20 µM) but overall do not show a systematic bias. DIC (Fig. S5) also has large deviations (∼50 µM) in the top 500 m and with a slight positive bias. It is in TA (Fig. S6) that de- viations, while similarly ∼50 µMeq, show a consistent near- surface underestimation. The surface currents in the model also demonstrate the monsoonal shifts and reversals seen in the AS (Fig. S7 in the Supplement). Figure 3. Seasonal surface pCO2 (µatm) from data (left column, a, d, g, j) and the model (middle, b, e, h, k), as well as their differ- ences (right, c, f, i, l). Plots are arranged by season: winter monsoon (DJFM, a–c), spring intermonsoon (AM, d–f), summer monsoon (JJAS, g–i), and fall intermonsoon (ON, j–l). Table 3. Mean and standard deviation (in parentheses) of annual and seasonal surface pCO2 (µatm) in both the merged dataset and model. Data Model Annual 426 (68) 428 (32) Winter (DJFM) 389 (14) 418 (30) Spring (AM) 398 (13) 439 (26) Summer (JJAS) 439 (77) 433 (36) Fall (ON) 393 (12) 427 (27) spring intermonsoon (Fig. 3d) ﬁnds pCO2 values similar to the winter (range: 354–451 µatm), with data coverage im- proving in the western AS. The summer monsoon, with the best data coverage (Fig. 3g), has pCO2 peaking at 773 µatm. In contrast, the fall intermonsoon (Fig. 3.1 Model validation and pCO2 data-model comparisons 3j) has very little data coverage, with pCO2 ranging from 311 to 485 µatm. Similar to the data, model pCO2 (Fig. 3b) is at its lowest during the winter. However, in the spring (Fig. 3e) open-ocean pCO2 ﬁnds its peak with a domain average of 439 µatm, which is not reﬂected in the in situ dataset (Fig. 3d and e). Maximum model pCO2 is found in the summer monsoon near upwelling regions (Fig. 3h), with values attaining > 800 µatm in Oman. Fall model pCO2 (Fig. 3k) still has elevated values averaging Regarding pCO2, in situ data from the merged SOCAT– LDEO database show that ∼90 % of 1pCO2 values in the AS are positive (Fig. 2a, inset), indicating a positive ﬂux to the atmosphere that is applicable geographically (Fig. 2a). Sampling dates for pCO2 (Fig. 2b) show that ∼70 % are from the summer monsoon months (June–September, JJAS). Most observations similarly date from the 1990s, with 1995 and 1997 alone accounting for 96 %. Seasonal pCO2 distributions from both data and the model are shown in Fig. 3. During the winter monsoon, pCO2 val- ues are at their lowest (range: 348–455 µatm; Fig. 3a). The Biogeosciences, 19, 907–929, 2022 https://doi.org/10.5194/bg-19-907-2022 915 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Figure 4. Taylor diagram of modeled vs. observed surface pCO2, both annually and seasonally. Data are from merged SOCAT and LDEO databases, corrected to the year 2005. Distance from ori- gin (concentric solid lines) is normalized model standard deviation. Angle from vertical axis is Pearson correlation coefﬁcient. Distance from observation point (black dot) is root-mean square deviation (dashed blue lines). Color of each point denotes model bias; i.e., positive values are overestimates. model overestimates variability (1.1 and 1.6, respectively). For all periods apart from summer, model pCO2 has a posi- tive bias (9.1, 24.6, 48.4, and 33.7 µatm for the annual, win- ter, spring, and fall, respectively). During the summer, the model has a negative bias of −3.1 µatm. g µ The source of bias in pCO2 is linked to the four state vari- ables SST, SSS, DIC, and TA. Comparisons with the model are made with SST and SSS from the merged SOCAT–LDEO database, while DIC and TA come from the ungridded GLO- DAP product (Fig. S8 in the Supplement). In this case, model SST and SSS (Fig. 3.1 Model validation and pCO2 data-model comparisons Most differences show that model output is higher in value than the data, averaging 24.6, 48.4, and 33.7 µatm higher for the winter, spring, and fall seasons, respectively. Ungridded DIC and TA data from GLODAP, though more sparse (n = 334 data points with both DIC and TA at depths ≤50 m), show more deviation from the 1 : 1 line (Fig. S8c and d) with overall negative biases of −15.8 µmolkg−1 and −30.0 µmoleqkg−1 for DIC and TA. These biases result in pCO2 perturbations of −33.8 and +45.7 µatm, respectively, when accounted for individually. Since the buffering capacity of seawater is related to the ratio of TA and DIC, when both biases are considered, average pCO2 shifts +16.7 µatm. As a result, while the DIC model bias lowers pCO2, the stronger bias in TA is the most likely cause for the model’s overall positive pCO2 bias, which may in part be due to the unre- solved fast sinking rates of foraminifera in the model. A Taylor diagram (Taylor, 2001) comparing in situ pCO2 data with model output shows the model’s relative perfor- mance (Fig. 4). The distance from the origin is model vari- ability normalized by standard deviation of the in situ data. The angle created from the y axis is the Pearson correlation coefﬁcient between the model and in situ data. If the model were to perfectly reproduce the data, it would appear at the position (1,0), equivalent to a normalized standard deviation of 1 and correlation coefﬁcient of 1. For the entire dataset, as well as for the spring and summer seasons, the model’s cor- relation with data is ∼0.5. Winter and fall have lower values at 0.2 and 0.06, respectively. Variability expressed as normal- ized standard deviation shows that overall, and during spring and summer periods, the model underestimates data variabil- ity (∼0.5 µatm). During the winter and fall, however, the Direct comparisons between the in situ and model output demonstrate the positive bias and middling correlations of the model with respect to the data, as well as the model’s tendency to underrepresent variability. As a result, it is neces- sary to investigate how these shortcomings compare with al- ternative pCO2 estimates in the AS. Figure 5 shows monthly comparisons of the pCO2 probability distribution functions from in situ data, model output, and L15. For most of the year, the data (Fig. 3.1 Model validation and pCO2 data-model comparisons S8a and b) largely overlap with a 1 : 1 relationship but with slight positive biases of ∼0.4 oC and 0.3 psu. Removing these biases from the model results in a pCO2 shift of −6.8 and −3.5 µatm for SST and SSS, re- spectively. These deviations are close in magnitude to the best-case measurement error of ∼2 µatm. Taylor diagrams for SST and SSS (Fig. S9 in the Supplement) further show the seasonal performance of these two variables. The model performs best for SST (Fig. S9a) during the winter, with a correlation of 0.93 and a normalized standard deviation of 0.97. The other seasons have lower correlations (0.74–0.81) and reduced standard deviations (0.63–0.8) except for the fall with a standard deviation of 1. SSS (Fig. S9b) has lower cor- relations and standard deviations than SST, with all seasons demonstrating a positive bias (0.02–0.39 psu). Correlation is best in the winter at 0.89 and worst in the fall at 0.46. Model variability in SSS is also less than the data, with standard de- viations ranging from 0.33 to 0.72. Lower variability is most likely due to the raw nature of the in situ data used here, in opposition to the monthly averaged climatological forcing and initial conditions of the model. Figure 4. Taylor diagram of modeled vs. observed surface pCO2, both annually and seasonally. Data are from merged SOCAT and LDEO databases, corrected to the year 2005. Distance from ori- gin (concentric solid lines) is normalized model standard deviation. Angle from vertical axis is Pearson correlation coefﬁcient. Distance from observation point (black dot) is root-mean square deviation (dashed blue lines). Color of each point denotes model bias; i.e., positive values are overestimates. 427 µatm but less than the summer period. Certain regions in the model show persistent maxima in pCO2, such as the Gulf of Oman and the Strait of Hormuz, which are not reﬂected in the few data collected there. Model pCO2 values in the Gulf of Aden increase during spring and then peak during the summer, a pattern which is unclear from the data. Annual and seasonal pCO2 means, with standard deviations in paren- theses, are displayed in Table 3 for both the data and model. Differences from interpolated model output and in situ data are shown on the right column of Fig. 3 (Fig. 3c, f, i, and l). 3.2 Air–sea CO2 ﬂux, drivers of seasonal variability, and ﬂux intercomparison Figure 5. Monthly probability density distributions of surface pCO2 (µatm) in (a) merged SOCAT–LDEO in situ data, (b) modeled pCO2, and (c) L15 pCO2 climatology. Modeled annual mean atmospheric ﬂux of CO2 (Fig. 6a) shows outgassing (positive, red) throughout the entire do- main, producing an average annual CO2 ﬂux density rate of 1.9 molCm−2 yr−1 and a total of 162.6 Tg C yr−1. Simi- lar to pCO2, several hotspots appear in the geographic distri- bution. Near the coast of Oman, the average ﬂux density is 2.7, with 3.2 in Somalia and 2.4 along the coast of India, pro- ducing a ﬂux of 11.4, 32.9, and 4.9 Tg C yr−1, respectively. The other regions, the north AS and oligotrophic central AS, have average densities of 2.0 and 1.5 molCm−2 yr−1, with total ﬂuxes of 10.5 and 28.6 Tg C yr−1. The seasonal air–sea ﬂux (Fig. 6b–e) has minima during fall and winter, with an increase in spring and a strong maximum during the sum- mer monsoon. Omani and Somalian ﬂux densities during the summer monsoon are 5.8 and 5.9 molCm−2 yr−1, respec- tively. The distribution of enhanced summer air–sea CO2 ﬂux coincides with the southwest monsoon winds (Fig. S10 in the Supplement), as well as the band of cooler temperatures im- pacting spatial pCO2 anomalies (see Sect. 3.3.1). The entire domain ﬂuxes of 32.0, 26.6, 90.9, and 13.1 Tg C yr−1 for the winter, spring, summer, and fall periods, respectively, each contribute 19.7, 16.3, 55.9, and 8.1 % of the annual total. peak. In the model (Fig. 5b), pCO2 is almost entirely above 400 µatm, with the median value increasing during the spring intermonsoon and peaking in June (453 µatm). Similar to the data, the upper-bound variability in pCO2 peaks in August. L15 (Fig. 5c), by contrast, has a tighter envelope of variabil- ity, with 5–95 percentile values never going beyond the range of 368–434 µatm. Median pCO2 in L15 peaks in the summer like the data at 402 µatm, but there is no large increase in upper-bound variability, with the 95 % upper bound in L15 reaching 434 µatm in September. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux 916 Figure 5. Monthly probability density distributions of surface pCO2 (µatm) in (a) merged SOCAT–LDEO in situ data, (b) modeled pCO2, and (c) L15 pCO2 climatology. Figure 5. Monthly probability density distributions of surface pCO2 (µatm) in (a) merged SOCAT–LDEO in situ data, (b) modeled pCO2, and (c) L15 pCO2 climatology. Figure 6. (a) Modeled annual mean air–sea CO2 ﬂux density (molCm−2 yr−1). (b–e) Seasonal ﬂux density for winter (DJFM), spring (AM), summer (JJAS), and fall (ON), respectively. Positive is ﬂux out of the ocean. Figure 6. (a) Modeled annual mean air–sea CO2 ﬂux density (molCm−2 yr−1). (b–e) Seasonal ﬂux density for winter (DJFM), spring (AM), summer (JJAS), and fall (ON), respectively. Positive is ﬂux out of the ocean. 3.1 Model validation and pCO2 data-model comparisons 5a) stays within a relatively narrow range (375–425 µatm) except for the summer monsoon when values can exceed 500 µatm and the median value has its https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 3.2 Air–sea CO2 ﬂux, drivers of seasonal variability, and ﬂux intercomparison Towards this end, we use multiple pCO2 products to calculate CO2 ﬂux with the same wind and parameterization as the model (Fig. 9). As summarized in Table 1, pCO2 from TK09, L15, GLODAP data, and Sarma (2003), interpolated to the WOA 1◦× 1◦grid, is used in these calculations (except for TK09 in which the coarse resolution reduced coverage). The original applicability of the Sarma (2003) model is north of 10◦N, and so ﬂux is calculated for this region, as well. Figure 7. (a) Anomaly of air–sea CO2 ﬂux during the summer mon- soon (JJAS; molCm−2 yr−1). Summer ﬂux anomaly contributions due to (b) wind, (c) pCO2, and (d) cross-terms in Eq. (2). as described in Eq. (2). The temporal anomalies for the sum- mer monsoon, the period with the strongest CO2 ﬂux signal, are presented in Fig. 7. Most of the domain has positive but variable strength anomalies in air–sea ﬂux (Fig. 7a), aver- aging 1.3 molCm−2 yr−1 with a standard deviation of 1.35. The wind contribution to ﬂux variability, κα (Fig. 7b), is also positive in most of the domain except the Gulf of Aden and the southeastern corner of the domain. The wind anomaly’s magnitude and distribution closely match the total anomaly in Fig. 7a, with a mean ﬂux anomaly of 1.18 molCm−2 yr−1 and 0.96 standard deviation. The 1pCO2 contribution to sea- sonal ﬂux anomaly (Fig. 7c) has a lower-magnitude effect overall (mean ﬂux anomaly 0.1, deviation 0.5, maximum 6.2 molCm−2 yr−1), with positive values north of 10◦N and slightly negative to the south. The maxima approaching 6.2 molCm−2 yr−1 are in the upwelling centers of Oman, So- malia, and the Indian coast. Second-order cross-term values (Fig. 7d) are almost all positive, with maxima also occurring near upwelling centers similar to the 1pCO2 term but weaker in magnitude with an average of 0.04 molCm−2 yr−1. All calculations have their peak CO2 ﬂux sometime in the summer, conﬁrming the role of winds in CO2 ﬂux timing. This study’s model consistently produces one of the higher estimates with 120 TgCyr−1 (reduced from 162.6 due to re- gridding) and 57 TgCyr−1 north of 10◦N. The only estimate higher than the model is GLODAP data in the region north of 10◦N with 65 TgCyr−1 possibly driven by summer mon- soon sampling bias. The high model estimate is perhaps un- surprising, considering the pCO2 bias. 3.2 Air–sea CO2 ﬂux, drivers of seasonal variability, and ﬂux intercomparison In summary, the survey of available data and compar- ing it to the model output produces a few distinct features: (1) available in situ data show that the majority of observa- tions are skewed towards the summer monsoon during the years 1995 and 1997; (2) most in situ data show CO2 out- gassing in the AS; (3) the model has a net positive bias in surface pCO2, driven by a joint DIC-TA bias which is slightly stronger in TA; and (4) the model captures the high summer monsoon pCO2 values better than the alternative L15 clima- tology. The variability in air–sea CO2 ﬂux can be attributed to the contributions of winds, 1pCO2, and interacting cross-terms, Biogeosciences, 19, 907–929, 2022 https://doi.org/10.5194/bg-19-907-2022 917 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux effect (Somalian May pCO2 anomaly of 1.1 molCm−2 yr−1, wind anomaly of 0.1). Summer monsoon winds represent the majority contribution to CO2 ﬂux variability, with a min- imum 64.7 % contribution relative to the total anomaly in India, a maximum of 112.8 % in the oligotrophic AS, and 90.8 % for the whole domain. By contrast, summer pCO2 and cross-terms contribute 6.0 % and 3.1 % to the domain’s anomaly, respectively. Fall intermonsoon months resemble the winter monsoon, with negative wind anomalies contribut- ing most with small or negative pCO2 contributions. In most scenarios, pCO2 contributes in the same direction as the winds or little at all, with the notable exceptions of Oman, oligotrophic AS, Somalia, and the domain during spring in- termonsoon. Figure 7. (a) Anomaly of air–sea CO2 ﬂux during the summer mon- soon (JJAS; molCm−2 yr−1). Summer ﬂux anomaly contributions due to (b) wind, (c) pCO2, and (d) cross-terms in Eq. (2). While strong monsoon winds dominate the timing of air– sea CO2 ﬂux, and the AS is always a source of CO2 due to positive 1pCO2, differences in pCO2 between independent sources can still result in a wide range of overall magnitudes. In the AS, CO2 outgassing estimates vary from 7 TgCyr−1 (Goyet et al., 1998b) to > 90 TgCyr−1 (Sarma, 2003) and everything in between (Somasundar et al., 1990), with each study using their own pCO2 data and wind parameterizations. Considering the important seasonal role of winds, the best way to investigate the role of pCO2 variability is to keep winds (and their ﬂux parameterization) constant. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux 10c) contributes to- ward negative pCO2 anomalies in a southwest-to-northeast band along the coasts of east Africa and the Arabian penin- sula, up to the coasts of Pakistan and the northern coast of India near Gujarat. The cold SST structure contributes a −20 µatm effect on pCO2 and largely overlaps the stronger summer monsoon winds (Fig. S10). The opposite trend is found in the central oligotrophic and Indian regions, where the average temperature contribution to pCO2 is 20 µatm de- spite upwelling along the southern Indian coast. The distri- A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Figure 8. Monthly CO2 air–sea ﬂux anomaly (molCm−2 yr−1) for (a) the domain, (b) Oman, (c) north AS, (d) oligotrophic central AS, (e) Somali coast, and (f) Indian coast. Contributors to the ﬂux are solubility and winds (kα, blue), pCO2 (red), and cross-terms (orange). Gray regions indicate winter and summer monsoons. Figure 8. Monthly CO2 air–sea ﬂux anomaly (molCm−2 yr−1) for (a) the domain, (b) Oman, (c) north AS, (d) oligotrophic central AS, (e) Somali coast, and (f) Indian coast. Contributors to the ﬂux are solubility and winds (kα, blue), pCO2 (red), and cross-terms (orange). Gray regions indicate winter and summer monsoons. Figure 9. (a) Monthly CO2 ﬂux (TgCmonth−1) from the AS as calculated using pCO2 from TK09 (cyan), L15 (blue), model (black), and GLODAP (red). (b) Monthly CO2 ﬂux from 10◦N and north using pCO2 from L15 (blue), model (black), GLODAP (red), Sarma using model output (purple), and Sarma using WOA data (orange). Dashed line in (b) is the zero ﬂux axis, and gray regions denote winter and summer monsoons. Positive ﬂux is out from the ocean surface. Figure 9. (a) Monthly CO2 ﬂux (TgCmonth−1) from the AS as calculated using pCO2 from TK09 (cyan), L15 (blue), model (black), and GLODAP (red). (b) Monthly CO2 ﬂux from 10◦N and north using pCO2 from L15 (blue), model (black), GLODAP (red), Sarma using model output (purple), and Sarma using WOA data (orange). Dashed line in (b) is the zero ﬂux axis, and gray regions denote winter and summer monsoons. Positive ﬂux is out from the ocean surface. pCO2 bias, as previously mentioned the GLODAP estimate supersedes it in the region north of 10◦N, as does the orig- inal Sarma (2003) estimate of 70 TgCyr−1. Thus, while we may think the model overestimates ﬂux, it is still within the range of previous studies in the AS. clear that Oman, the Indian coast, and the north AS host en- hanced pCO2, with average positive anomalies of 8.6, 21.5, and 49 µatm, respectively. In contrast, both the oligotrophic central AS and Somalian regions have negative pCO2 anoma- lies (−13.7 and −2.9 µatm, respectively). The contributing factors to these pCO2 anomalies, SST, DIC, TA, and SSS, display differing distributions. SST (Fig. 3.2 Air–sea CO2 ﬂux, drivers of seasonal variability, and ﬂux intercomparison The range in estimates of total CO2 ﬂux is 57–120 TgCyr−1, resulting in a ratio of 2.1× variability. In the reduced domain of the AS north of 10◦N, estimates range from 12.3 to 65.6, resulting in 5.3× variability. The 5.3× ratio is quite high and is in part driven by the low estimates from the Sarma (2003) model, which are 12.3 and 17.6 using tracer data from WOA and ROMS, respectively. Indeed, the Sarma (2003) model estimates have negative CO2 ﬂux for some months, which is not observed in the original publication, and the total ﬂuxes are quite smaller than the 70 TgCyr−1 reported. If the two lower estimates are removed, the range in air–sea CO2 ﬂux in the domain north of 10◦N is 41–65 TgCyr−1, providing a ratio of 1.6 similar to 2.1 for the whole domain. Even considering the model’s The seasonal ﬂux anomalies for all regions are displayed in Fig. 8. The summer monsoon ﬂux is so strong that it makes the anomalies (black lines) for all the other seasons nega- tive except for May in the spring. During the winter months (DJFM), both wind and pCO2 terms produce negative ﬂux anomalies (ranging to −0.78 and −0.38 in the domain for wind and pCO2, respectively; Fig. 8a), indicating the relative lack of winds and minimum pCO2 values. In winter, while the negative wind term is universally strongest, within the upwelling regions the pCO2 term is 58 % (Fig. 8b) of the wind term’s magnitude, and 49 % for the entire domain. The spring intermonsoon, where many regions such as Somalia and the central oligotrophic AS (Fig. 8d and e) experience their pCO2 maximum, shows a positive pCO2 effect on ﬂux anomaly that is as large as or larger than the negative wind https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 918 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Figure 10. (a) Spatial anomaly of time-averaged surface pCO2 (µatm). Black boxes represent regions of analysis used in (b) to show averaged contributions of four parameters to pCO2 variability. The changes in pCO2 due to these variables are shown for (c) tem- perature, (d) DIC, (e) TA, and (f) SSS. Both TA and SSS effects are negative for the ﬁrst half of the year before becoming slightly positive in the second half, never reaching 10 µatm in magnitude. Different pCO2 anomaly cycles can be found in the up- welling regions of Oman, Somalia, and India (Fig. 11b, e, and f). Here, a positive temperature peak appears in the spring (27–45 µatm), which is then supplanted by a positive DIC peak during the summer monsoon (41–81 µatm). In both Oman and India, the summertime DIC peak is strong enough to contribute to the annual pCO2 peak despite cooler tem- peratures. In Somalia, the summertime DIC peak is not suf- ﬁciently stronger than temperature (41 vs. −34 µatm) such that in sum with the other terms maximum pCO2 is found in the spring, not the summer, similar to the whole domain and oligotrophic regions. Both TA and SSS effects in these three regions are lower in magnitude (never exceeding 18.4 and 7.3 µatm for TA and SSS, respectively) and generally run counter to DIC. A completely different regime occurs in the north AS (Fig. 11c). Here, while temperature effects (49 µatm in June) create a similar spring–summertime peak in pCO2 (15.9 µatm) somewhat counteracted by DIC (−40 µatm), dur- ing the winter monsoon temperature and DIC effects are both maximal and in opposing amplitudes (−49.5 and 51.4 µatm for SST and DIC, respectively). This occurs due to the con- vective mixing that occurs during winter in the north AS, where cooling temperatures lower pCO2, but subsurface wa- ter introduces more DIC, resulting in a near-balance. Figure 10. (a) Spatial anomaly of time-averaged surface pCO2 (µatm). Black boxes represent regions of analysis used in (b) to show averaged contributions of four parameters to pCO2 variability. The changes in pCO2 due to these variables are shown for (c) tem- perature, (d) DIC, (e) TA, and (f) SSS. The oligotrophic central region (Fig. 11d), the largest in area, has similar pCO2 and temperature impacts as the whole domain, with the two largely overlapping. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux DIC, TA, and SSS impacts also follow similar patterns but have slightly higher magnitudes in the central AS, with DIC reaching 32 µatm. bution of DIC-induced anomalies (Fig. 10d) shows a posi- tive inﬂuence near coastal regions and the western AS off the coast of Somalia (+25 µatm), whereas a strong minimum is found in an oval region encompassing the central, open- ocean AS (−36.6 µatm). TA effects (Fig. 10e) show a north– south gradient similar to SSS, with positive contributions to pCO2 of +20 µatm occurring in the north and −20 µatm to- wards the south, resulting in magnitudes similar to SST con- tributions. SSS contributions (Fig. 10f) show a similar distri- bution as TA but are weaker in magnitude (± 10 µatm). 3.4 Near-surface DIC budgets and cycling SST’s effect on pCO2 reﬂects physical processes like surface heating and cooling, mixing, and advection. DIC, by con- trast, reﬂects both physical and biological processes because in addition it is also impacted by photosynthesis, CaCO3 shell formation and dissolution, zooplankton respiration, de- tritus remineralization (bacterial respiration), and air–sea ex- change. Budgets of DIC ﬂuxes in the upper 20 m (Fig. 12; see Fig. S11 in the Supplement for a volume-speciﬁc DIC ﬂux) show that two major processes dominate: vertical cir- culation (light blue lines) and net biological processes (ma- genta lines). In the entire domain and all subregions, and for all months, vertical circulation (advection and mixing) acts as a source of DIC, with the sum of all biological processes acting as a sink (NB the top 20 m does not constitute the entire euphotic zone, so respiration and remineralization at depth is not included). Maximum magnitudes of both ver- tical circulation and biological ﬂux occur during the sum- mer monsoon for all regions except for the north AS where they occur during the winter monsoon bloom (Fig. 12c). 3.3.1 Spatial pCO2 distribution Spatial pCO2 anomalies calculated from the annual mean highlight the geographic hotspots of pCO2 inside the do- main (Fig. 10a). The pCO2 anomalies range from −89 to +415 µatm, indicative of a positive skew in the distribution. Within the regions of analysis prescribed in this study, it is https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 919 3.3.2 Seasonal pCO2 cycle 12d) oscillates from being ei- ther positive or negative four times during the year, with mag- nitudes rivaling air–sea ﬂux at times (5 × 10−2 PgCyr−1). Horizontal advection (dark blue lines) is negative on average for the whole domain (−0.2 PgCyr−1), denoting net export (Fig. 12a). The same pattern occurs for all subregions except India with net horizontal import of surface DIC (Fig. 12f; 2.9 × 10−3 PgCyr−1). The Omani upwelling region and the oligotrophic region experience positive peaks of horizontal import during the summer monsoon (27 and 56 TgCyr−1 for Omani and oligotrophic regions, respectively), though for Somalia this period is the maximum DIC export, peaking at 220 TgCyr−1 in July. The maximum DIC ﬂux in the domain due to vertical cir- culation is 1.76 PgCyr−1, whereas biological ﬂux peaks at −1.0 PgCyr−1. Biological ﬂuxes are nearly phase-matched with vertical circulation, though peaks in summer biological ﬂux lag vertical circulation by a month (Fig. 12d, e, and f). Comparing the two ﬂux terms, after normalizing biological ﬂux by vertical circulation ﬂux, the relative strength of bio- logical processes vs. vertical sources of DIC becomes appar- ent. In the whole domain, biological ﬂux ranges from −90 % to −34.5 % of vertical ﬂux, similar to Rixen et al. (2005). As a result, biological ﬁxation of carbon is generally weaker than physical vertical delivery of DIC. The maximum DIC ﬂux in the domain due to vertical cir- culation is 1.76 PgCyr−1, whereas biological ﬂux peaks at −1.0 PgCyr−1. Biological ﬂuxes are nearly phase-matched with vertical circulation, though peaks in summer biological ﬂux lag vertical circulation by a month (Fig. 12d, e, and f). Comparing the two ﬂux terms, after normalizing biological ﬂux by vertical circulation ﬂux, the relative strength of bio- logical processes vs. vertical sources of DIC becomes appar- ent. In the whole domain, biological ﬂux ranges from −90 % to −34.5 % of vertical ﬂux, similar to Rixen et al. (2005). As a result, biological ﬁxation of carbon is generally weaker than physical vertical delivery of DIC. Air–sea ﬂux (red lines) is always negative due to the high pCO2 values, peaking during the summer monsoon. DIC ﬂux due to atmospheric escape, while reaching its maxi- mum magnitude of ∼0.32 PgCyr−1 in June and July for the whole domain (Fig. 3.3.2 Seasonal pCO2 cycle The previous section outlines the geographic regions within the AS that have overall high or low pCO2 values, but in order to investigate the strong seasonal monsoon cycle in the AS, the decomposition of variables affecting monthly pCO2 values is calculated at each model grid point and av- eraged into each analysis region (Fig. 11). Regarding the whole domain (Fig. 11a), pCO2 variability is similar to that seen in Fig. 5b, with a spring pCO2 anomaly peak (20 µatm) and minimum during fall and winter (−9.4 µatm). Tempera- ture effects largely mirror the overall pCO2 cycle (May peak 30 µatm, January minimum −17 µatm). Change in pCO2 as- sociated with DIC acts in opposition to temperature but with lower magnitude (16 µatm in February, −8 µatm in June). https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 920 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Figure 11. Time series of pCO2 anomalies (µatm) (black lines) for (a) the entire domain, (b) Oman, (c) north AS, (d) oligotrophic central AS, (e) Somalia, and (f) India. Dashed gray lines indicate horizontal axis. Gray shading shows summer and winter monsoons. Additional lines show change in pCO2 due to temperature (blue), DIC (red), TA (orange), and SSS (magenta). Figure 11 Time series of pCO2 anomalies (µatm) (black lines) for (a) the entire domain (b) Oman (c) north AS (d) oligotrophic central Figure 11. Time series of pCO2 anomalies (µatm) (black lines) for (a) the entire domain, (b) Oman, (c) north AS, (d) oligotrophic central AS, (e) Somalia, and (f) India. Dashed gray lines indicate horizontal axis. Gray shading shows summer and winter monsoons. Additional lines show change in pCO2 due to temperature (blue), DIC (red), TA (orange), and SSS (magenta). Figure 11. Time series of pCO2 anomalies (µatm) (black lines) for (a) the entire domain, (b) Oman, (c) north AS, (d) oligotrophic central AS, (e) Somalia, and (f) India. Dashed gray lines indicate horizontal axis. Gray shading shows summer and winter monsoons. Additional lines show change in pCO2 due to temperature (blue), DIC (red), TA (orange), and SSS (magenta). net evaporation, averaging 0.07 PgCyr−1 in the domain) ex- cept India where it is negative (net precipitation, averaging −4.8 × 10−3 PgCyr−1). The oligotrophic region’s evapora- tion and precipitation ﬂux (Fig. 4.1 Model pCO2 vs. data The search for the model bias source is hindered by the lack of in situ data in the region. As already noted, GLO- DAP has 334 locations with DIC and TA in the top 50 m. The few available in situ data that do exist in the AS have a number of deﬁciencies for the purpose of validating model output. First, the data available are both old and concentrated around the years 1995 and 1997. While the JGOFS stud- ies were quintessential in diagnosing the seasonal cycle of pCO2, they preclude being able to decipher the secular trend in surface pCO2 due to increasing atmospheric CO2 con- centrations. In our analysis, we estimated a +2 µatm yr−1 trend, close to that of Tjiputra et al. (2014), though ﬁnding an interannual linear trend requires more data at regular in- tervals. Second, due to the nature of strong upwelling in the AS, previous cruise sampling also biases not only the sum- mer months (≈70 % of data) but also places in the vicinity of the Omani coast (Fig. 3g). As a result, it is difﬁcult to deter- mine to what extent the data are representative of the entire AS. Consider that in the model, ﬂux intensities are lower in The pCO2 output from the model has a positive bias with re- spect to the in situ data, as is clear from Figs. 3 to 5. The question becomes whether the model bias precludes its use in acquiring a reasonable air–sea CO2 ﬂux estimate. Regard- ing the direction of CO2 ﬂux (positive outgassing or negative uptake), since most in situ 1pCO2 data are already positive (Fig. 2), an additional positive bias will not impact ﬂux direc- tion, reafﬁrming the previous ﬁndings of Sarma et al. (1998) and subsequent work demonstrating that the AS is a source of CO2 to the atmosphere. A positive model bias in pCO2 has been noted in previous modeling studies. For instance, in the global data assimilation study of Valsala and Maksyu- tov (2010), they found an overall positive bias in the north- ern Indian Ocean, ∼+5–15 µatm above TK09 (compared to our −3.1 to +48.4 µatm with respect to in situ data). Ad- ditionally, that study found a similar underestimate near the upwelling regions (summer negative bias in the model) of the AS and overestimate elsewhere (their Figs. 3 and 4). In Sreeush et al. 4 Discussion tive pCO2 bias, whereas the ofﬂine Ocean Transport Tracer Model (OTTM) produced negative bias in pCO2 in compari- son to TK09. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux 921 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Figure 12. Time series of DIC ﬂuxes (PgCyr−1) in the top 20 m for (a) the domain, (b) Oman, (c) north AS, (d) oligotrophic central AS, (e) Somalia, and (f) India. Dashed gray line shows x = 0 axis. Gray shading denotes summer and winter monsoons. Figure 12. Time series of DIC ﬂuxes (PgCyr−1) in the top 20 m for (a) the domain, (b) Oman, (c) north AS, (d) oligotrophic central AS, (e) Somalia, and (f) India. Dashed gray line shows x = 0 axis. Gray shading denotes summer and winter monsoons. 3.3.2 Seasonal pCO2 cycle 12a), only surpasses biological ﬂux in May, when 0.23 PgCyr−1 is released to the atmosphere com- pared to 0.15 PgCyr−1 in biological processes. Evaporation and precipitation (brown lines) result in higher DIC for most of the year in the entire domain and upwelling regions (i.e., https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 4.2 Spatial distribution of air–sea CO2 ﬂux and pCO2 Apart from the aforementioned bias leading to heightened absolute values (though Bates et al., 2006, have > 400 µatm for large parts of the AS), the relatively enhanced pCO2 values near Oman, along the west coast of India, and in the Gulf of Aden have already been observed (Sabine et al., 2000; Bates et al., 2006; Sarma et al., 2000; Körtzinger et al., 1997). These same studies, however, note a minimum of pCO2 outside of the summer monsoon near the southwest coast of India due to freshwater inﬂux, which is not replicated well in the model. Additionally, elevated pCO2 near the Equator is not observed (Sabine et al., 2000; Bates et al., 2006), although it can ap- pear in other models (Valsala and Murtugudde, 2015). The model’s seasonal pCO2 minimum during the winter monsoon is also not reﬂective of results found elsewhere (Goyet et al., 1998a, b; Bates et al., 2006; though many studies highlight the north AS, where minimum model pCO2 occurs during the spring). Instead, these papers state pCO2 is minimal during the fall intermonsoon. Likewise, the large-scale spring max- imum of pCO2 seen in the model is not found in these stud- ies, except for in Louanchi et al. (1996), though this result is somewhat anomalous since that study showed a pCO2 min- imum during the summer monsoon. Thus, while the model agrees with previous work insofar as the coastal regions im- pacted by upwelling show enhanced pCO2, mismatches do appear in the seasonal timing of maxima and minima, espe- cially within certain subregions. p 2, The spatial decomposition of factors inﬂuencing pCO2 (Fig. 10) highlights how geographically DIC can be the strongest factor, with SST and TA taking secondary roles and SSS being a weak contributor. Since DIC and TA can co-vary with salinity, when they are not normalized, their distribution in the AS mirrors the north–south salinity gra- dient (see Figs. 2 and 3 in Bates et al., 2006). Once corrected for salinity, it is clear that the upwelling region of Oman still has elevated DIC, whereas the central, oligotrophic AS shows a DIC deﬁcit. By contrast, the onshore–offshore gra- dient in TA is weaker. 4.2 Spatial distribution of air–sea CO2 ﬂux and pCO2 the central, oligotrophic region (Fig. 6), but due to its surface area the total ﬂux (28.6 TgCyr−1) was close to that of Soma- lia (32.9 TgCyr−1), an observation also made by Lendt et al. (2003). Determining to what extent the model over- or under- estimates CO2 ﬂux due to pCO2 bias would require more in situ sampling, which would need to be designed around solv- ing the problems of areal coverage (outside of Oman and up- welling zones) and temporal coverage (off-summer months and recurrent over multiple years). The model results both afﬁrm the conclusions of previous studies in terms of CO2 ﬂux direction and seasonality and yet ﬁnd difference in magnitudes. As previously stated, the AS is an atmospheric CO2 source, with most ﬂux occurring (56 %) during the summer monsoon (Fig. 6). In our results, however, there is no region during any of the seasons where CO2 uptake takes place. While somewhat expected, this is still in disagreement with some of the other pCO2 datasets previously considered, such as in Sarma (2003), in which negative 1pCO2 values appear, such as during the winter monsoon near the south coast of India. The model’s posi- tive pCO2 bias may be to blame for this, making it so that no negative 1pCO2 appears. Despite the positive pCO2 bias, a few other patterns are clear in comparison to other CO2 ﬂux estimates. Sabine et al. (2000) and Sarma (2003) both ﬁnd the maximum ﬂux occurring during the summer monsoon centered around the upwelling regions, which is also quite visible in the model results (Fig. 6d). However, Bates et al. (2006) found that a secondary maximum of ﬂux occurs dur- ing the winter monsoon, though due to the color scale in their Fig. 6 it is difﬁcult to ascertain much beyond CO2 outgassing from the AS during all months of the year. Their secondary max in ﬂux may be partly attributable to higher wintertime pCO2, as well. The distribution of model pCO2 is both similar to and dif- ferent from previous data-based and modeling studies. 4.1 Model pCO2 vs. data (2019a), ROMS resulted in a systematic posi- Biogeosciences, 19, 907–929, 2022 https://doi.org/10.5194/bg-19-907-2022 922 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Biogeosciences, 19, 907–929, 2022 https://doi.org/10.5194/bg-19-907-2022 4.3.2 The pCO2 seasonality Decomposition of seasonal pCO2 anomalies within regions portrays a slightly different picture where temperature is the dominant force, with DIC countervailing in the upwelling re- gions. Not only is this seasonal cycle more akin to that seen in the California Current (Turi et al., 2014), the dueling role of these two forces is also reﬂected in a similar analysis by Sreeush et al. (2019a) for pH instead of pCO2 in the AS. In- terestingly, in that study both ROMS and OTTM were com- pared side-by-side, and in OTTM, TA played a larger role than in ROMS. Similarly, in Valsala and Maksyutov (2013), TA played an important role in regulating interannual pCO2 variability in the AS. A preliminary TA budget of the model (Fig. S12 in the Supplement) shows that while vertical cir- culation and biological processes dominate the seasonal cy- cle of near-surface DIC, TA has multiple forces inﬂuencing its time evolution. However, the magnitude of the ﬂuxes are ∼1 5 those of DIC, indicating that TA is less seasonally vari- able than DIC (reﬂected also in Fig. 11). These results, from another model, as well as the low variability in this model’s TA, raise the possibility that TA’s importance is underesti- mated in the current study. The summer ﬂux signal is such that in nearly all the re- gions outside of summer, the anomaly is negative. Further- more, the contribution of winds in particular is so strong that it is the largest factor all year except for the spring intermon- soon, when peak pCO2 is important relative to the effects of wind (or lack thereof) in the central oligotrophic AS, So- malia, and the averaged domain. This suggests that, on ﬁrst order, winds are the most important factor in determining the seasonal air–sea ﬂux cycle in the AS. We should keep in mind, however, that these results conﬂict with the analysis of Roobaert et al. (2019). In their global study of coastal waters, while seasonal CO2 ﬂux variability in the AS is relatively high compared to other regions (their Fig. 6), the largest con- tributions come from 1pCO2 and cross-terms (their Fig. 7), especially near the Horn of Africa. As a result, further work should be conducted to reduce uncertainty in sea surface pCO2 values to determine whether winds, 1pCO2, or cross- terms are signiﬁcant drivers of air–sea ﬂux. 4.3.1 Air–sea CO2 Flux The fact that model CO2 ﬂux for the entire domain peaks in summer despite a spring peak in pCO2 for the domain as a whole (along with the Somalian and oligotrophic regions) is the ﬁrst sign that perhaps pCO2 is not the primary driver in determining ﬂux timing. The Reynolds decomposition of CO2 ﬂux terms (Fig. 8) clearly shows that a large proportion of the summer ﬂux is due to the arrival of the strong south- west summer monsoon winds. The positive contributions due to pCO2 occur in the usual upwelling regions, though their contribution in magnitude is relatively muted and negative in the southern portion of the AS. Cross-terms, while non- zero, are inconsequential in determining the overall anomaly in summer ﬂux intensity, as has been seen elsewhere (Doney et al., 2009b). Indeed, in a scenario in which the cross-term contribution is at its maximum amplitude, the Omani up- welling region during summer, the cross-term is not strong enough to sway the direction of the ﬂux anomaly. 4.2 Spatial distribution of air–sea CO2 ﬂux and pCO2 Differences between coastal and off- shore normalized DIC and TA in the AS have been previ- ously observed (Millero et al., 1998b; Lendt et al., 2003), but the stronger relative absence of DIC in the central AS and its role in affecting pCO2 has not been emphasized. A simi- lar analysis in the California Current upwelling system (Turi et al., 2014) indicates near-compensation of DIC and temper- ature in opposing directions, nearly overlapping each other. In that scenario, DIC overpowers temperature at the coast, with TA and SSS being secondary. For the AS, while the up- welling regions of Oman and Somalia show temperature and DIC working against each other, they are not as well com- pensated for. Furthermore, the gradients of positive/negative pCO2 contributions from temperature and DIC do not over- lap, leading to the curious scenario in which temperature and DIC both contribute positively to the pCO2 anomaly along the Indian coast. The positioning of these gradients and the surprising negative inﬂuence of DIC away from upwelling regions perhaps underscores how the AS is rather unique, Despite the model’s limitations, its advantages are also clear. Beyond the obvious increase in spatiotemporal cov- erage, capturing the monsoon’s strong seasonal dynamics helps the model where other approaches fall short. This is especially illustrated in Fig. 5. Since upwelling regions are limited in geographic extent near the coast, capturing their high pCO2 values can be difﬁcult for other approaches, such as TK09 with its coarse grid. Even the L15 product, with its ﬁner grid, is unable to produce the higher pCO2 values seen during the summer. Judging from these comparisons, the trade-off appears to be that the model currently may pro- duce less accurate pCO2 values outside of summer, but the explicit resolving of upwelling allows for enhanced pCO2 values during the summer monsoon, the peak of CO2 ﬂux. Biogeosciences, 19, 907–929, 2022 https://doi.org/10.5194/bg-19-907-2022 923 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux where strong seasonal upwelling winds mingle with strong tropical heating and there is the inﬂuence of outﬂows from marginal seas (Prasad et al., 2001; l’Hégaret et al., 2015). where strong seasonal upwelling winds mingle with strong tropical heating and there is the inﬂuence of outﬂows from marginal seas (Prasad et al., 2001; l’Hégaret et al., 2015). mate of CO2 ﬂux, and (3) this leaves reality somewhere in between. The only way to rectify these differences and ar- rive at a more accurate estimate will be to conduct sufﬁcient in situ sampling of DIC, TA, and pCO2 in more regions than the upwelling zones, as well as preferably outside of the sum- mer and over the course of multiple years. With the advent of ARGO ﬂoats with pH sensors, and the advancement of tech- nology for other variables such as TA, the possibility emerges of using autonomous sampling platforms to expand beyond the limitations of shipboard measurements to ﬁll the data gap in the AS carbon system. 4.3.3 DIC seasonality The potential for biological control in setting pCO2 has been found in Sri Lanka near the AS (Chakraborty et al., 2018). In this study, it was found that the source water in Sri Lanka was sufﬁciently low in DIC relative to inorganic nutrients that upwelling actually reduced surface pCO2. In a similar vein, Takahashi et al. (2002) found, using a metric comparing tem- perature and “biological” effects (i.e., everything else), that the AS’s pCO2 is reduced more by biological production than temperature effects. Conducting this analysis on the model output (Fig. S13 in the Supplement), it appears that “bio- logical” control appears dominant over the upwelling areas (Omani coast, coast of Somalia, India) and near the Equa- tor east of 60◦E, but for the majority of the AS temperature dominates. This cursory analysis aside, as is evident in the results of Chakraborty et al. (2018), the more useful compar- ison is in determining whether biological production is sufﬁ- cient to outweigh DIC enhancement from subsurface water. The majority of ﬂux occurs during the summer as opposed to a modeled spring pCO2 maximum due to the inﬂuence of winds. A Reynolds decomposition of both pCO2 and wind variability shows that the intense winds of the summer mon- soon contribute 90 % of that season’s ﬂux anomaly. In fact, winds play a more important role than the increase in pCO2 in the upwelling regions. Even though winds represent such a major variable in determining AS CO2 ﬂux timing, the vari- ability in total ﬂux due to different pCO2 products leads to a 2× range in magnitude. These results suggest that in addition to the expected increase in surface ocean pCO2 due to an- thropogenic climate change, possible changes in the timing, location, and magnitude of monsoon winds (Lachkar et al., 2018; Praveen et al., 2020) will have downstream impacts on seasonal air–sea ﬂux. g In summary, the results in Fig. 12 indicate that for the en- tire AS, DIC enhancement by vertical circulation (both ad- vection and mixing) brings more DIC into the near-surface than is removed by net biological processes, and so no bio- logically induced decrease in pCO2 occurs in the ﬁnal pCO2 signal. 5 Conclusions In this study, we used a regional circulation model coupled with a biogeochemical model to investigate the annual mag- nitude, seasonal cycle, and drivers of air–sea CO2 ﬂux in the AS, primarily winds and 1pCO2. This effort was made to complement previous ﬂux estimates, for which limited data or insufﬁcient model resolutions have produced contrasting results. Consistent with previous work, we ﬁnd that the AS is a source of CO2 to the atmosphere for the entire year, with the bulk occurring during the summer monsoon. Our esti- mate of ﬂux, ∼160 TgCyr−1, with concentrated ﬂux den- sities up to 6 molCm−2 yr−1 in the upwelling regions, is larger than most previous reports but not inconsistent with the range of other ﬁndings (Sarma, 2003; Naqvi et al., 2005; Sarma et al., 2013) . Since the AS lacks carbon data, here we subjected the model to validation with raw data instead of smoothed climatologies. The model is shown to have a posi- tive bias in pCO2, attributed to TA and DIC, with TA bias be- ing stronger. Despite this, pCO2 variability compares favor- ably to alternative products in the region. The bias results in strongly positive 1pCO2 throughout the domain year-round. While positive 1pCO2 values have been observed before in the AS, we likely overestimate CO2 ﬂux outside of the sum- mer monsoon. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux offshore normalized DIC gradient) from the upwelling re- gions does not signiﬁcantly contribute to the central AS or other regions. and S9), and the forcing datasets for SST and heat ﬂux corre- spond to data that predate or include the pCO2 sampling pe- riod (i.e., before 2000), so a climate change bias is unlikely. What might be more likely, then, is a sampling bias towards summertime Oman, one of the few areas in the AS with a summertime instead of springtime pCO2 max. Such a bias could possibly obscure what is happening in the rest of the AS. Regardless, the discrepancy between models and obser- vations during the spring period can be added as yet another reason to conduct more in situ sampling to either conﬁrm or disavow whether the model results are spurious. 4.3.2 The pCO2 seasonality Additionally, when considering the inconsistencies of models in estimat- ing air–sea CO2 ﬂux (Sarma et al., 2013), uncertainties from incomplete representation of winds and the various parame- terizations of piston velocity must be considered in addition to pCO2, especially in light of recent work in the ﬁeld (Ho et al., 2006; Wanninkhof, 2014; Roobaert et al., 2018). Zooming out from the upwelling regions and looking at the whole AS, the dominance of temperature on the seasonal pCO2 cycle is clear. In the domain average, temperature ef- fects nearly overlap with the overall pCO2 anomaly. This result brings back into focus the seasonal timing of pCO2 minima and maxima in the model vis à vis previous work. In the earlier studies, which either use data directly or build statistical models from those data, there is no spring inter- monsoon pCO2 maximum driven by heating. Indeed, Sabine et al. (2000) noted that pCO2 in the spring was much lower than would be expected given the SST but attributed this to drawdown due to biological production. The model, how- ever, indicates that this is precisely the season when biologi- cal production is at its lowest. The presence of these spring- time maxima can be seen in other models, as is visible in the results of Valsala and Maksyutov (2010) and a synthesis by Sarma et al. (2013). Since the model indicates temperature is producing the maxima, it reduces the concern that erro- neous DIC or TA values in the model are driving this signal. The model SST matches well with the in situ data (Figs. S8 Wind parameterizations notwithstanding, once winds are controlled in our metaanalysis (Fig. 9) it appears that on bal- ance (1) gridded data-based pCO2 products will underesti- mate the upwelling zone maxima of pCO2 and CO2 ﬂux dur- ing the summer, (2) the model overestimates pCO2 the rest of the year, eventually contributing to a possible overesti- https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 924 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux https://doi.org/10.5194/bg-19-907-2022 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux (last access: 1 February 2022). The Comprehensive Ocean-Atmosphere Data Set (COADS) can be down- loaded from http://iridl.ldeo.columbia.edu/SOURCES/ .DASILVA/.SMD94/.halfbyhalf/.climatology/ (last ac- cess: 1 February 2022). The Scatterometer Climatol- ogy of Ocean Winds (SCOW) can be downloaded from https://chapman.ceoas.oregonstate.edu/scow/index.html (last access: 1 February 2022). AVHRR SST data can be down- loaded from https://www.ncei.noaa.gov/data/oceans/pathﬁnder/ Version5.0_Climatologies/1982_2008/Monthly/ (last access: 1 February 2022). NPP data can be downloaded from https: //sites.science.oregonstate.edu/ocean.productivity/custom.php (last access: 1 February 2022). (last access: 1 February 2022). The Comprehensive Ocean-Atmosphere Data Set (COADS) can be down- loaded from http://iridl.ldeo.columbia.edu/SOURCES/ .DASILVA/.SMD94/.halfbyhalf/.climatology/ (last ac- cess: 1 February 2022). The Scatterometer Climatol- ogy of Ocean Winds (SCOW) can be downloaded from https://chapman.ceoas.oregonstate.edu/scow/index.html (last access: 1 February 2022). AVHRR SST data can be down- loaded from https://www.ncei.noaa.gov/data/oceans/pathﬁnder/ Version5.0_Climatologies/1982_2008/Monthly/ (last access: 1 February 2022). NPP data can be downloaded from https: //sites.science.oregonstate.edu/ocean.productivity/custom.php (last access: 1 February 2022). ocean. Instead, temperature follows a southwest–northeast monsoon wind pattern, whereas DIC is enhanced nearest to the coasts. The resulting apparent deﬁcit of normalized DIC in the central, oligotrophic AS has not been emphasized pre- viously. Finally, we ﬁnd that despite the intense biological activity in the AS, primary production by phytoplankton is insufﬁcient to counter the increased carbon supply provided by vertical circulation during bloom periods. y g p Models can be used to expand spatiotemporal coverage when data are scarce. However, models’ limitations of- ten manifest when there are no new data to test their ﬁ- delity. Limitations in the spatiotemporal coverage of exist- ing datasets stem from biases in sampling during the sum- mer monsoon, sampling close to the Omani upwelling re- gion, and sampling limited in scope to the years of JGOFS expeditions of the 1990s. In order to fully characterize the pCO2 cycle outside of summer in the rest of the AS, as well as to determine the secular trend of surface pCO2 due to an- thropogenic carbon additions to the atmosphere, more in situ data of the carbon system (e.g., DIC, TA, pCO2), from ship- board measurements or autonomous sampling platforms, are sorely needed. Finally since 1pCO2 is generally positive in the AS, the direction of air–sea CO2 exchange examined here is robust to model error, whereas other important indicators such as pH and aragonite saturation, a, which at important thresholds of low values have deleterious impacts for vari- ous biological taxa (Doney et al., 2009a; Bednaršek et al., 2019, 2021) will be less so. Code availability. ROMS-AGRIF is provided by https://www. croco-ocean.org (last access: 1 February 2022). Financial support. This research has been supported by Tamkeen under the NYU Abu Dhabi Research Institute (grant nos. G110 and CG009). Data availability. Model output used in this study can be accessed and cited from the following: https://doi.org/10.5281/zenodo.5937512 (de Verneil, 2022). The Surface Ocean CO2 Atlas (SOCAT; Bakker et al., 2016) can be downloaded from https://www.socat.info/index.php/version-2019/ (last access: 1 February 2022). The Lamont–Doherty Earth Observatory pCO2 database can be downloaded from https://www.ncei.noaa.gov/access/ocean-carbon-data-system/ oceans/LDEO_Underway_Database/ (last access: 1 Febru- ary 2022). The Simple Ocean Data Assimilation (SODA) reanalysis data can be downloaded from http://apdrc.soest.hawaii. edu/datadoc/soda_2.2.4.php (last access: 1 February 2022). The World Ocean Atlas 2009 can be downloaded from https://www.ncei.noaa.gov/data/oceans/woa/WOA09/DATA/ Review statement. This paper was edited by Peter Landschützer and reviewed by two anonymous referees. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux These data are thus critical for resolving the possible responses of the carbon system in the AS to ongoing climate change, whether from changes in tim- ing or magnitude of monsoon wind forcing, the impact of increased surface heating on stratiﬁcation and vertical circu- lation, or changing levels of primary and ﬁsheries produc- tivity with altered carbonate solubility. Without this baseline information, it will be difﬁcult to predict what the future has in store for the AS carbon system. Supplement. The supplement related to this article is available on- line at: https://doi.org/10.5194/bg-19-907-2022-supplement. Author contributions. AdV, ZL, and ML conceived the study, AdV and ZL ran the model, AdV, ZL, SS, and ML conducted analysis, AdV generated ﬁgures and text, and ZL, SS, and ML revised ﬁgures and text. Competing interests. The contact author has declared that neither they nor their co-authors have any competing interests. Disclaimer. Publisher’s note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. Acknowledgements. Support for this research comes from the Cen- ter for Prototype Climate Modeling (CPCM), the New York Univer- sity Abu Dhabi (NYUAD) Research Institute. Computations were conducted at the High Performance Cluster (HPC) at NYUAD, Dalma. We deeply thank both Benoit Merchand and Muataz Al Bar- wani for their technical support. We are also grateful for the work of two anonymous reviewers and associate editor Peter Landschützer, who greatly improved the manuscript. Code availability. ROMS-AGRIF is provided by https://www. croco-ocean.org (last access: 1 February 2022). 4.3.3 DIC seasonality The timing of biological drawdown, occurring at the same time or lagging vertical circulation, is consistent with the general phenology of blooms and similar to previous ﬁnd- ings (Louanchi et al., 1996; Rixen et al., 2006; Sharada et al., 2008). The result that biological cycling of carbon is much larger than the air–sea ﬂux of CO2 also corroborates the re- sults of Lendt et al. (2003), who found net community pro- duction to be ∼3.6 times larger than CO2 emission. The relatively low impact of horizontal advection is an interest- ing detail to consider; in other upwelling systems, signiﬁcant proportions of water and biological production are advected offshore (Nagai et al., 2015). Lendt et al. (2003) suggest up- welled nitrate is assimilated and does not arrive in the central AS, while Resplandy et al. (2011) show that a large fraction of total nutrients in the central AS come from the upwelling zones. Thus, although water may be advected offshore, the relevant timescale for DIC cycling processes (i.e., air–sea emission, biological uptake) may be short enough so that hor- izontal export of enhanced DIC (keep in mind the onshore– An important result of this modeling study is that temper- ature drives a springtime maximum of pCO2 in the AS. This maximum has been observed in lower-resolution models but is not found in the in situ data. Due to the fact that tempera- ture is not sensitive to biological processes like DIC and TA, this discrepancy suggests that more sampling is necessary to determine whether it is an artifact of spotty sampling or an in- herent problem in models unrelated to resolving coastal up- welling. Additionally, we ﬁnd that spatial gradients of DIC and temperature do not overlap as they do elsewhere in the https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 925 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux References Acharya, S. S. and Panigrahi, M. K.: Eastward shift and mainte- nance of Arabian Sea oxygen minimum zone: Understanding the paradox, Deep-Sea Res. Pt. I, 115, 240–252, 2016. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux K., Lefèvre, N., Lo Monaco, C., Manke, A M hi J T M li L Mill F J M i P M S to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, edited by: Stocker, T. F., Qin, D., Plattner, G. K., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P. M., Cambridge University Press, Cam- bridge, United Kingdom and New York, NY, USA, 2013. Curry, W., Ostermann, D., Guptha, M., and Ittekkot, V.: Foraminiferal production and monsoonal upwelling in the Ara- bian Sea: evidence from sediment traps, Geological Society, Lon- don, Special Publications, 64, 93–106, 1992. da Silva, A. M., Young, C. C., and Levitus, S.: Atlas of surface ma- rine data 1994, Vol. 4: Anomalies of fresh water ﬂuxes, NOAA Atlas, NESDIS, U.S. Department of Commerce, NOAA, NES- DIS, 9, 1994. A., Mathis, J. T., Merlivat, L., Millero, F. J., Monteiro, P. M. S., de Verneil, A.: Evaluating the Arabian Sea as a regional source of atmospheric CO2: seasonal variability and drivers, Zenodo [Data set], https://doi.org/10.5281/zenodo.5937512, 2022. A., Mathis, J. T., Merlivat, L., Millero, F. J., Monteiro, P. M. S., Munro, D. R., Murata, A., Newberger, T., Omar, A. M., Ono, T., Paterson, K., Pearce, D., Pierrot, D., Robbins, L. L., Saito, S., Munro, D. R., Murata, A., Newberger, T., Omar, A. M., Ono, T., Paterson, K., Pearce, D., Pierrot, D., Robbins, L. L., Saito, S., Dickson, A. and Millero, F. J.: A comparison of the equilibrium constants for the dissociation of carbonic acid in seawater media, Deep-Sea Res. Pt. I, 34, 1733–1743, 1987. Salisbury, J., Schlitzer, R., Schneider, B., Schweitzer, R., Sieger, Salisbury, J., Schlitzer, R., Schneider, B., Schweitzer, R., Sieger, R., Skjelvan, I., Sullivan, K. F., Sutherland, S. C., Sutton, A. J., Tadokoro, K., Telszewski, M., Tuma, M., van Heuven, S. M. A. C., Vandemark, D., Ward, B., Watson, A. J., and Xu, S.: A multi- decade record of high-quality fCO2 data in version 3 of the Sur- face Ocean CO2 Atlas (SOCAT), Earth Syst. Sci. Data, 8, 383– 413, https://doi.org/10.5194/essd-8-383-2016, 2016. R., Skjelvan, I., Sullivan, K. F., Sutherland, S. C., Sutton, A. J., Doney, S. C., Fabry, V. J., Feely, R. A., and Kley- pas, J. A.: Ocean Acidiﬁcation: The Other CO2 Problem, Annu. Rev. Mar. Sci., 1, 169–192, https://doi.org/10.1146/annurev.marine.010908.163834, pMID: 21141034, 2009a. Bates, N. R., Pequignet, A. C., and Sabine, C. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Bakker, D. C. E., Pfeil, B., Landa, C. S., Metzl, N., O’Brien, K. M., Olsen, A., Smith, K., Cosca, C., Harasawa, S., Jones, S. D., Nakaoka, S., Nojiri, Y., Schuster, U., Steinhoff, T., Sweeney, C., Takahashi, T., Tilbrook, B., Wada, C., Wanninkhof, R., Alin, S. R., Balestrini, C. F., Barbero, L., Bates, N. R., Bianchi, A. A., Bonou, F., Boutin, J., Bozec, Y., Burger, E. F., Cai, W.-J., Castle, R. D., Chen, L., Chierici, M., Currie, K., Evans, W., Feather- stone, C., Feely, R. A., Fransson, A., Goyet, C., Greenwood, N., Gregor, L., Hankin, S., Hardman-Mountford, N. J., Harlay, J., Hauck, J., Hoppema, M., Humphreys, M. P., Hunt, C. W., Huss, B., Ibánhez, J. S. P., Johannessen, T., Keeling, R., Kitidis, V., Körtzinger, A., Kozyr, A., Krasakopoulou, E., Kuwata, A., Land- schützer, P., Lauvset, S. K., Lefèvre, N., Lo Monaco, C., Manke, A., Mathis, J. T., Merlivat, L., Millero, F. J., Monteiro, P. M. S., Munro, D. R., Murata, A., Newberger, T., Omar, A. M., Ono, T., Paterson, K., Pearce, D., Pierrot, D., Robbins, L. L., Saito, S., Salisbury, J., Schlitzer, R., Schneider, B., Schweitzer, R., Sieger, R., Skjelvan, I., Sullivan, K. F., Sutherland, S. C., Sutton, A. J., Tadokoro, K., Telszewski, M., Tuma, M., van Heuven, S. M. A. C., Vandemark, D., Ward, B., Watson, A. J., and Xu, S.: A multi- decade record of high-quality fCO2 data in version 3 of the Sur- face Ocean CO2 Atlas (SOCAT), Earth Syst. Sci. Data, 8, 383– Bakker, D. C. E., Pfeil, B., Landa, C. S., Metzl, N., O’Brien, K. M., Olsen, A., Smith, K., Cosca, C., Harasawa, S., Jones, S. D., Nakaoka, S., Nojiri, Y., Schuster, U., Steinhoff, T., Sweeney, C., Takahashi, T., Tilbrook, B., Wada, C., Wanninkhof, R., Alin, S. R., Balestrini, C. F., Barbero, L., Bates, N. R., Bianchi, A. A., Bonou, F., Boutin, J., Bozec, Y., Burger, E. F., Cai, W.-J., Castle, R. D., Chen, L., Chierici, M., Currie, K., Evans, W., Feather- stone, C., Feely, R. A., Fransson, A., Goyet, C., Greenwood, N., Gregor, L., Hankin, S., Hardman-Mountford, N. J., Harlay, J., Hauck, J., Hoppema, M., Humphreys, M. P., Hunt, C. W., Huss, B., Ibánhez, J. S. P., Johannessen, T., Keeling, R., Kitidis, V., Körtzinger, A., Kozyr, A., Krasakopoulou, E., Kuwata, A., Land- schützer, P., Lauvset, S. https://doi.org/10.5194/bg-19-907-2022 https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 Biogeosciences, 19, 907–929, 2022 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux 926 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux L.: Ocean carbon cycling in the Indian Ocean: 1. Spatiotempo- ral variability of inorganic carbon and air–sea CO2 gas exchange, Global Biogeochem. Cy., 20, GB3020, https://doi.org/10.1029/2005GB002491, 2006. Doney, S. C., Lima, I., Feely, R. A., Glover, D. M., Lindsay, K., Mahowald, N., Moore, J. K., and Wanninkhof, R.: Mechanisms governing interannual variability in upper-ocean inorganic car- bon system and air–sea CO2 ﬂuxes: Physical climate and atmo- spheric dust, Deep-Sea Res. Pt. II, 56, 640–655, 2009b. Bednaršek, N., Feely, R. A., Howes, E. L., Hunt, B. P. V., Kessouri, F., León, P., Lischka, S., Maas, A. E., McLaughlin, K., Ne- zlin, N. P., Sutula, M., and Weisberg, S. B.: Systematic Re- view and Meta-Analysis Toward Synthesis of Thresholds of Ocean Acidiﬁcation Impacts on Calcifying Pteropods and In- teractions With Warming, Frontiers in Marine Science, 6, 227, https://doi.org/10.3389/fmars.2019.00227, 2019. Friedlingstein, P., O’Sullivan, M., Jones, M. W., Andrew, R. M., Hauck, J., Olsen, A., Peters, G. P., Peters, W., Pongratz, J., Sitch, S., Le Quéré, C., Canadell, J. G., Ciais, P., Jackson, R. B., Alin, S., Aragão, L. E. O. C., Arneth, A., Arora, V., Bates, N. R., Becker, M., Benoit-Cattin, A., Bittig, H. C., Bopp, L., Bultan, S., Chandra, N., Chevallier, F., Chini, L. P., Evans, W., Florentie, L., Forster, P. M., Gasser, T., Gehlen, M., Gilﬁllan, D., Gkritza- lis, T., Gregor, L., Gruber, N., Harris, I., Hartung, K., Haverd, V., Houghton, R. A., Ilyina, T., Jain, A. K., Joetzjer, E., Kadono, K., Kato, E., Kitidis, V., Korsbakken, J. I., Landschützer, P., Lefèvre, N., Lenton, A., Lienert, S., Liu, Z., Lombardozzi, D., Marland, G., Metzl, N., Munro, D. R., Nabel, J. E. M. S., Nakaoka, S.-I., Niwa, Y., O’Brien, K., Ono, T., Palmer, P. I., Pierrot, D., Poul- ter, B., Resplandy, L., Robertson, E., Rödenbeck, C., Schwinger, J., Séférian, R., Skjelvan, I., Smith, A. J. P., Sutton, A. J., Tan- hua, T., Tans, P. P., Tian, H., Tilbrook, B., van der Werf, G., Vuichard, N., Walker, A. P., Wanninkhof, R., Watson, A. J., Willis, D., Wiltshire, A. J., Yuan, W., Yue, X., and Zaehle, S.: Global Carbon Budget 2020, Earth Syst. Sci. Data, 12, 3269– 3340, https://doi.org/10.5194/essd-12-3269-2020, 2020. Quéré, C., Canadell, J. G., Ciais, P., Jackson, R. B., Alin, S., Aragão, L. E. O. C., Arneth, A., Arora, V., Bates, N. R., Becker, M., Benoit-Cattin, A., Bittig, H. C., Bopp, L., Bultan, S., Chandra, N., Chevallier, F., Chini, L. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux 927 ecosystem dynamics in the California Current System, Deep-Sea Res. Pt. I, 53, 1483–1516, 2006. ecosystem dynamics in the California Current System, Deep-Sea Res. Pt. I, 53, 1483–1516, 2006. Large, W. G., McWilliams, J. C., and Doney, S. C.: Oceanic vertical mixing: A review and a model with a nonlocal boundary layer parameterization, Rev. Geophys., 32, 363–403, 1994. Gruber, N., Lachkar, Z., Frenzel, H., Marchesiello, P., Münnich, M., McWilliams, J. C., Nagai, T., and Plattner, G.-K.: Eddy-induced reduction of biological production in eastern boundary upwelling systems, Nat. Geosci., 4, 787–792, 2011. Lee, K., Tong, L. T., Millero, F. J., Sabine, C. L., Dickson, A. G., Goyet, C., Park, G.-H., Wanninkhof, R., Feely, R. A., and Key, R. M.: Global relationships of total alkalinity with salinity and temperature in surface waters of the world’s oceans, Geophys. Res. Lett., 33, L19605, https://doi.org/10.1029/2006GL027207, 2006. Gruber, N., Hauri, C., Lachkar, Z., Loher, D., Frölicher, T. L., and Plattner, G.-K.: Rapid progression of ocean acidiﬁcation in the California Current System, Science, 337, 220–223, 2012. Lendt, R., Thomas, H., Hupe, A., and Ittekkot, V.: Re- sponse of the near-surface carbonate system of the north- western Arabian Sea to the southwest monsoon and re- lated biological forcing, J. Geophys. Res.-Oceans, 108, 3222, https://doi.org/10.1029/2000JC000771, 2003. Hauri, C., Gruber, N., Vogt, M., Doney, S. C., Feely, R. A., Lachkar, Z., Leinweber, A., McDonnell, A. M. P., Munnich, M., and Plat- tner, G.-K.: Spatiotemporal variability and long-term trends of ocean acidiﬁcation in the California Current System, Biogeo- sciences, 10, 193–216, https://doi.org/10.5194/bg-10-193-2013, 2013. Lévy, M., Shankar, D., André, J.-M., Shenoi, S., Durand, F., and de Boyer Montégut, C.: Basin-wide seasonal evolution of the In- dian Ocean’s phytoplankton blooms, J. Geophys. Res.-Oceans, 112, C12014, https://doi.org/10.1029/2007JC004090, 2007. Ho, D. T., Law, C. S., Smith, M. J., Schlosser, P., Harvey, M., and Hill, P.: Measurements of air-sea gas exchange at high wind speeds in the Southern Ocean: Implications for global parameterizations, Geophys. Res. Lett., 33, L16611, https://doi.org/10.1029/2006GL026817, 2006. L’Hégaret, P., Duarte, R., Carton, X., Vic, C., Ciani, D., Baraille, R., and Corréard, S.: Mesoscale variability in the Arabian Sea from HYCOM model results and observations: impact on the Persian Gulf Water path, Ocean Sci., 11, 667–693, https://doi.org/10.5194/os-11-667-2015, 2015. Hood, R. R., Urban, E. R., McPhaden, M. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux J., Su, D., and Raes, E.: The 2nd International Indian Ocean Expedition (IIOE-2): Moti- vating New Exploration in a Poorly Understood Basin, Limnol- ogy and Oceanography Bulletin, 25, 117–124, 2016. Louanchi, F., Metzl, N., and Poisson, A.: Modelling the monthly sea surface fCO2 ﬁelds in the Indian Ocean, Mar. Chem., 55, 265– 279, 1996. Keeling, C. D.: Carbon dioxide in surface ocean waters: 4. Global distribution, J. Geophys. Res., 73, 4543–4553, https://doi.org/10.1029/JB073i014p04543, 1968. Lovenduski, N. S., Gruber, N., Doney, S. C., and Lima, I. D.: En- hanced CO2 outgassing in the Southern Ocean from a positive phase of the Southern Annular Mode, Global Biogeochem. Cy., 21, GB2026, https://doi.org/10.1029/2006GB002900, 2007. Keeling, C. D., Brix, H., and Gruber, N.: Seasonal and long- term dynamics of the upper ocean carbon cycle at Station ALOHA near Hawaii, Global Biogeochem. Cy., 18, GB4006, https://doi.org/10.1029/2004GB002227, 2004. Mahadevan, A., Lévy, M., and Mémery, L.: Mesoscale variability of sea surface pCO2: What does it respond to?, Global Biogeochem. Cy., 18, GB1017, https://doi.org/10.1029/2003GB002102, 2004. Khatiwala, S., Primeau, F., and Hall, T.: Reconstruction of the his- tory of anthropogenic CO2 concentrations in the ocean, Nature, 462, 346–349, 2009. Mahadevan, A., Tagliabue, A., Bopp, L., Lenton, A., Mémery, L., and Lévy, M.: Impact of episodic vertical ﬂuxes on sea surface pCO2, Philos. T. R. Soc. A, 369, 2009–2025, https://doi.org/10.1098/rsta.2010.0340, 2011. Körtzinger, A., Duinker, J. C., and Mintrop, L.: Strong CO2 emis- sions from the Arabian Sea during south-west monsoon, Geo- phys. Res. Lett., 24, 1763–1766, 1997. Marchesiello, P., Debreu, L., and Couvelard, X.: Spurious diapycnal mixing in terrain-following coordinate models: The problem and a solution, Ocean Model., 26, 156–169, 2009. Kumar, S. P., Ramaiah, N., Gauns, M., Sarma, V., Muraleedharan, P., Raghukumar, S., Kumar, M. D., and Madhupratap, M.: Physi- cal forcing of biological productivity in the Northern Arabian Sea during the Northeast Monsoon, Deep-Sea Res. Pt. II, 48, 1115– 1126, 2001. Mehrbach, C., Culberson, C., Hawley, J., and Pytkowicx, R.: Mea- surement of the apparent dissociation constants of carbonic acid in seawater at atmospheric pressure 1, Limnol. Oceanogr., 18, 897–907, 1973. Lachkar, Z. and Gruber, N.: Response of biological production and air–sea CO2 ﬂuxes to upwelling intensiﬁcation in the Califor- nia and Canary Current Systems, J. Marine Syst., 109, 149–160, 2013. Millero, F. J.: Thermodynamics of the carbon dioxide system in the oceans, Geochim. Cosmochim. Acta, 59, 661–677, 1995. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux P., Evans, W., Florentie, Bednaršek, N., Calosi, P., Feely, R. A., Ambrose, R., Byrne, M., Chan, K. Y. K., Dupont, S., Padilla-Gamiño, J. L., Spicer, J. I., Kessouri, F., Roethler, M., Sutula, M., and Weisberg, S. B.: Synthesis of Thresholds of Ocean Acidiﬁcation Im- pacts on Echinoderms, Frontiers in Marine Science, 8, 261, https://doi.org/10.3389/fmars.2021.602601, 2021. L., Forster, P. M., Gasser, T., Gehlen, M., Gilﬁllan, D., Gkritza- lis, T., Gregor, L., Gruber, N., Harris, I., Hartung, K., Haverd, V., Houghton, R. A., Ilyina, T., Jain, A. K., Joetzjer, E., Kadono, K., Kato E Kitidis V Korsbakken J I Landschützer P Lefèvre Kato, E., Kitidis, V., Korsbakken, J. I., Landschützer, P., Lefèvre, i i b d i l d Lenton, A., Lienert, S., Liu, Z., Lombardozzi, D., Mar ., Metzl, N., Munro, D. R., Nabel, J. E. M. S., Nakaoka, S.-I., Behrenfeld, M. J. and Falkowski, P. G.: Photosynthetic rates de- rived from satellite-based chlorophyll concentration, Limnol. Oceanogr., 42, 1–20, https://doi.org/10.4319/lo.1997.42.1.0001, 1997. Carton, J. A. and Giese, B. S.: A reanalysis of ocean climate using Simple Ocean Data Assimilation (SODA), Mon. Weather Rev., 136, 2999–3017, 2008. Casey, K. S. and Cornillon, P.: A comparison of satellite and in situ–based sea surface temperature climatologies, J. Climate, 12, 1848–1863, 1999. Goyet, C., Metzl, N., Millero, F., Eischeid, G., O’sullivan, D., and Poisson, A.: Temporal variation of the sea surface CO2/carbonate properties in the Arabian Sea, Mar. Chem., 63, 69–79, 1998a. Chakraborty, K., Valsala, V., Gupta, G., and Sarma, V.: Dominant biological control over upwelling on pCO2 in sea east of Sri Lanka, J. Geophys. Res.-Biogeo., 123, 3250–3261, 2018. Goyet, C., Millero, F. J., O’Sullivan, D., Eischeid, G., McCue, S., and Bellerby, R.: Temporal variations of pCO2 in surface seawa- ter of the Arabian Sea in 1995, Deep-Sea Res. Pt. I, 45, 609–623, 1998b. Ciais, P., Sabine, C., Bala, G., Bopp, L., Brovkin, V., Canadell, J., Chhabra, A., DeFries, R., Galloway, J., Heimann, M., Jones, C., Le Quéré, C., Myneni, R. B., Piao, S., and Thornton, P.: Carbon and Other Biogeochemical Cycles, in: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I Gruber, N., Frenzel, H., Doney, S. C., Marchesiello, P., McWilliams, J. C., Moisan, J. R., Oram, J. J., Plattner, G.-K., and Stolzenbach, K. D.: Eddy-resolving simulation of plankton Biogeosciences, 19, 907–929, 2022 https://doi.org/10.5194/bg-19-907-2022 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux 928 offshore transport of carbon and nutrients in the California Cur- rent System, J. Geophys. Res.-Oceans, 120, 5318–5341, 2015. offshore transport of carbon and nutrients in the California Cur- rent System, J. Geophys. Res.-Oceans, 120, 5318–5341, 2015. Resplandy, L., Lévy, M., and McGillicuddy Jr., D. J.: Ef- fects of Eddy-Driven Subduction on Ocean Biological Carbon Pump, Global Biogeochem. Cy., 33, 1071–1084, https://doi.org/10.1029/2018GB006125, 2019. Naqvi, S. W. A., Gupta, R. S., and Kumar, M. D.: Carbon dioxide and nitrous oxide in the Arabian Sea, Geophysical Monograph Series, American Geophysical Union, Washington, DC, 75, 85– 92, https://doi.org/10.1029/GM075p0085, 1993. Risien, C. M. and Chelton, D. B.: A global climatology of surface wind and wind stress ﬁelds from eight years of QuikSCAT scat- terometer data, J. Phys. Oceanogr., 38, 2379–2413, 2008. Naqvi, S. W. A., Bange, H. W., Gibb, S. W., Goyet, C., Hat- ton, A. D., and Upstill-Goddard, R. C.: Biogeochemical ocean– atmosphere transfers in the Arabian Sea, Prog. Oceanogr., 65, 116–144, 2005. Rixen, T., Guptha, M., and Ittekkot, V.: Deep ocean ﬂuxes and their link to surface ocean processes and the biological pump, Prog. Oceanogr., 65, 240–259, 2005. Rixen, T., Goyet, C., and Ittekkot, V.: Diatoms and their inﬂu- ence on the biologically mediated uptake of atmospheric CO2 in the Arabian Sea upwelling system, Biogeosciences, 3, 1–13, https://doi.org/10.5194/bg-3-1-2006, 2006. Olsen, A., Lange, N., Key, R. M., Tanhua, T., Álvarez, M., Becker, S., Bittig, H. C., Carter, B. R., Cotrim da Cunha, L., Feely, R. A., van Heuven, S., Hoppema, M., Ishii, M., Jeansson, E., Jones, S. D., Jutterström, S., Karlsen, M. K., Kozyr, A., Lau- vset, S. K., Lo Monaco, C., Murata, A., Pérez, F. F., Pfeil, B., Schirnick, C., Steinfeldt, R., Suzuki, T., Telszewski, M., Tilbrook, B., Velo, A., and Wanninkhof, R.: GLODAPv2.2019 – an update of GLODAPv2, Earth Syst. Sci. Data, 11, 1437–1461, https://doi.org/10.5194/essd-11-1437-2019, 2019. Roobaert, A., Laruelle, G. G., Landschützer, P., and Regnier, P.: Uncertainty in the global oceanic CO2 uptake induced by wind forcing: quantiﬁcation and spatial analysis, Biogeosciences, 15, 1701–1720, https://doi.org/10.5194/bg-15-1701-2018, 2018. Roobaert, A., Laruelle, G. G., Landschützer, P., Gruber, N., Chou, L., and Regnier, P.: The spatiotemporal dynamics of the sources and sinks of CO2 in the global coastal ocean, Global Bio- geochem. Cy., 33, 1693–1714, 2019. Orr, J. C. and Epitalon, J.-M.: Improved routines to model the ocean carbonate system: mocsy 2.0, Geosci. Model Dev., 8, 485–499, https://doi.org/10.5194/gmd-8-485-2015, 2015. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux S., Schlitzer, R., Schneider, B., Schuster, U., Sieger, R., Skjel- van, I., Steinhoff, T., Suzuki, T., Takahashi, T., Tedesco, K., Tel- szewski, M., Thomas, H., Tilbrook, B., Tjiputra, J., Vandemark, D., Veness, T., Wanninkhof, R., Watson, A. J., Weiss, R., Wong, C. S., and Yoshikawa-Inoue, H.: A uniform, quality controlled Surface Ocean CO2 Atlas (SOCAT), Earth Syst. Sci. Data, 5, 125–143, https://doi.org/10.5194/essd-5-125-2013, 2013. J., Brown, P. J., Cai, W.-J., Chavez, F. P., Chen, A., Cosca, C., Sarma, V., Swathi, P., Kumar, M. D., Prasannakumar, S., Bhat- tathiri, P., Madhupratap, M., Ramaswamy, V., Sarin, M., Gauns, M., Ramaiah, N., Sardessai, S., and de Sousa, S. N.: Carbon budget in the eastern and central Arabian Sea: An Indian JGOFS synthesis, Global Biogeochem. Cy., 17, 1102,https://doi.org/10.1029/2002GB001978, 2003. , , , , , , , , , Hoppema, M., Hunt, C. W., Hydes, D., Ishii, M., Johannessen, T., Jones, S. D., Key, R. M., Körtzinger, A., Landschützer, P., Lauvset, S. K., Lefèvre, N., Lenton, A., Lourantou, A., Merlivat, Sarma, V. V. S. S., Dileep, K. M., Gauns, M., and Mad- hupratap, M.: Seasonal controls on surface pCO2 in the central and eastern Arabian Sea, Academy Proceed- ings in Earth and Planetary Sciences, 109, 471–479, available at: http://proxy.library.nyu.edu/login?url=https: //www.proquest.com/docview/214115753?accountid=12768 (last access: 1 February 2022), copyright – Indian Academy of Sciences 2000; Last updated: 11 August 2010; Subject- sTermNotLitGenreText – Arabian Sea, 2000. L., Midorikawa, T., Mintrop, L., Miyazaki, C., Murata, A., Naka- date, A., Nakano, Y., Nakaoka, S., Nojiri, Y., Omar, A. M., Padin, k i i A., Ríos, A. F., Santana-Casiano, J. M., Salisbury, J., Sarma, V. V. S. S., Schlitzer, R., Schneider, B., Schuster, U., Sieger, R., Skjel- van, I., Steinhoff, T., Suzuki, T., Takahashi, T., Tedesco, K., Tel- szewski, M., Thomas, H., Tilbrook, B., Tjiputra, J., Vandemark, D., Veness, T., Wanninkhof, R., Watson, A. J., Weiss, R., Wong, C. S., and Yoshikawa-Inoue, H.: A uniform, quality controlled Surface Ocean CO2 Atlas (SOCAT), Earth Syst. Sci. Data, 5, 125–143, https://doi.org/10.5194/essd-5-125-2013, 2013. A., Ríos, A. F., Santana-Casiano, J. M., Salisbury, J., Sarma, V. V. Sarma, V. V. S. S., Lenton, A., Law, R. M., Metzl, N., Pa- tra, P. K., Doney, S., Lima, I. D., Dlugokencky, E., Ra- monet, M., and Valsala, V.: Sea–air CO2 ﬂuxes in the Indian Ocean between 1990 and 2009, Biogeosciences, 10, 7035–7052, https://doi.org/10.5194/bg-10-7035-2013, 2013. Prasad, T., Ikeda, M., and Kumar, S. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Sabine, C., Wanninkhof, R., Key, R., Goyet, C., and Millero, F. J.: Seasonal CO2 ﬂuxes in the tropical and subtropical Indian Ocean, Mar. Chem., 72, 33–53, 2000. Orr, J. C., Epitalon, J.-M., Dickson, A. G., and Gattuso, J.-P.: Rou- tine uncertainty propagation for the marine carbon dioxide sys- tem, Mar. Chem., 207, 84–107, 2018. Sarma, V.: Monthly variability in surface pCO2 and net air–sea CO2 ﬂux in the Arabian Sea, J. Geophys. Res.-Oceans, 108, 3255, https://doi.org/10.1029/2001JC001062, 2003. Paulmier, A., Ruiz-Pino, D., and Garçon, V.: CO2 maximum in the oxygen minimum zone (OMZ), Biogeosciences, 8, 239–252, https://doi.org/10.5194/bg-8-239-2011, 2011. Pfeil, B., Olsen, A., Bakker, D. C. E., Hankin, S., Koyuk, H., Kozyr, A., Malczyk, J., Manke, A., Metzl, N., Sabine, C. L., Akl, J., Alin, S. R., Bates, N., Bellerby, R. G. J., Borges, A., Boutin, J., Brown, P. J., Cai, W.-J., Chavez, F. P., Chen, A., Cosca, C., Fassbender, A. J., Feely, R. A., González-Dávila, M., Goyet, C., Hales, B., Hardman-Mountford, N., Heinze, C., Hood, M., Hoppema, M., Hunt, C. W., Hydes, D., Ishii, M., Johannessen, T., Jones, S. D., Key, R. M., Körtzinger, A., Landschützer, P., Lauvset, S. K., Lefèvre, N., Lenton, A., Lourantou, A., Merlivat, L., Midorikawa, T., Mintrop, L., Miyazaki, C., Murata, A., Naka- Sarma, V., Kumar, M. D., and George, M.: The central and eastern Arabian Sea as a perennial source of atmospheric carbon dioxide, Tellus B, 50, 179–184, 1998. Pfeil, B., Olsen, A., Bakker, D. C. E., Hankin, S., Koyuk, H., Kozyr, A., Malczyk, J., Manke, A., Metzl, N., Sabine, C. L., Akl, J., Alin, S. R., Bates, N., Bellerby, R. G. J., Borges, A., Boutin, J., Brown, P. J., Cai, W.-J., Chavez, F. P., Chen, A., Cosca, C., Fassbender, A. J., Feely, R. A., González-Dávila, M., Goyet, C., Hales, B., Hardman-Mountford, N., Heinze, C., Hood, M., Hoppema, M., Hunt, C. W., Hydes, D., Ishii, M., Johannessen, T., Jones, S. D., Key, R. M., Körtzinger, A., Landschützer, P., Lauvset, S. K., Lefèvre, N., Lenton, A., Lourantou, A., Merlivat, L., Midorikawa, T., Mintrop, L., Miyazaki, C., Murata, A., Naka- date, A., Nakano, Y., Nakaoka, S., Nojiri, Y., Omar, A. M., Padin, X. A., Park, G.-H., Paterson, K., Perez, F. F., Pierrot, D., Poisson, A., Ríos, A. F., Santana-Casiano, J. M., Salisbury, J., Sarma, V. V. S. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux Lachkar, Z., Smith, S., Lévy, M., and Pauluis, O.: Eddies reduce denitriﬁcation and compress habitats in the Arabian Sea, Geo- phys. Res. Lett., 43, 9148–9156, 2016. Millero, F. J., Degler, E. A., O’Sullivan, D. W., Goyet, C., and Eis- cheid, G.: The carbon dioxide system in the Arabian Sea, Deep- Sea Res. Pt. II, 45, 2225–2252, 1998a. Millero, F. J., Lee, K., and Roche, M.: Distribution of alkalinity in the surface waters of the major oceans, Mar. Chem., 60, 111–130, https://doi.org/10.1016/S0304-4203(97)00084-4, 1998b. Lachkar, Z., Lévy, M., and Smith, S.: Intensiﬁcation and deepening of the Arabian Sea oxygen minimum zone in response to increase in Indian monsoon wind intensity, Biogeosciences, 15, 159–186, https://doi.org/10.5194/bg-15-159-2018, 2018. Morrison, J. M., Codispoti, L., Smith, S. L., Wishner, K., Flagg, C., Gardner, W. D., Gaurin, S., Naqvi, S., Manghnani, V., Prosperie, L., Prosperie, L., and Gundersen, J. S.: The oxygen minimum zone in the Arabian Sea during 1995, Deep-Sea Res. Pt. II, 46, 1903–1931, 1999. Landschützer, P., Gruber, N., Haumann, F. A., Rödenbeck, C., Bakker, D. C., Van Heuven, S., Hoppema, M., Metzl, N., Sweeney, C., Takahashi, T., Tilbrook, B., and Wanninkhof, R.: The reinvigoration of the Southern Ocean carbon sink, Science, 349, 1221–1224, 2015. Nagai, T., Gruber, N., Frenzel, H., Lachkar, Z., McWilliams, J. C., and Plattner, G.-K.: Dominant role of eddies and ﬁlaments in the https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux topography-following-coordinate oceanic model, Ocean Model., 9, 347–404, 2005. Turi, G., Lachkar, Z., and Gruber, N.: Spatiotemporal variability and drivers of pCO2 and air–sea CO2 ﬂuxes in the California Current System: an eddy-resolving modeling study, Biogeosciences, 11, 671–690, https://doi.org/10.5194/bg-11-671-2014, 2014. topography-following-coordinate oceanic model, Ocean Model., 9, 347–404, 2005. Smith, S. L.: The Arabian Sea of the 1990s: New biogeochemical understanding, Prog. Oceanogr., 2, 113–115, 2005. g g g Somasundar, K., Rajendran, A., Kumar, M. D., and Gupta, R. S.: Valsala, V. and Maksyutov, S.: Simulation and assimilation of global ocean pCO2 and air–sea CO2 ﬂuxes using ship obser- vations of surface ocean pCO2 in a simpliﬁed biogeochemical ofﬂine model, Tellus B, 62, 821–840, 2010. Somasundar, K., Rajendran, A., Kumar, M. D., and Gupta, R. S.: Carbon and nitrogen budgets of the Arabian Sea, Mar. Chem., 30, 363–377, 1990. Sreeush, M. G., Rajendran, S., Valsala, V., Pentakota, S., Prasad, K., and Murtugudde, R.: Variability, trend and controlling factors of Ocean acidiﬁcation over Western Arabian Sea upwelling region, Mar. Chem., 209, 14–24, 2019a. Valsala, V. and Maksyutov, S.: Interannual variability of the air–sea CO2 ﬂux in the north Indian Ocean, Ocean Dynam., 63, 165– 178, 2013. Sreeush, M. G., Valsala, V., Santanu, H., Pentakota, S., Prasad, K., Naidu, C., and Murtugudde, R.: Biological production in the In- dian Ocean upwelling zones – Part 2: Data based estimates of variable compensation depth for ocean carbon models via cyclo- stationary Bayesian Inversion, Deep-Sea Res. Pt. II, 179, 104619, https://doi.org/10.1016/j.dsr2.2019.07.007, 2019b. Valsala, V. and Murtugudde, R.: Mesoscale and intraseasonal air– sea CO2 exchanges in the western Arabian Sea during boreal summer, Deep-Sea Res. Pt. I, 103, 101–113, 2015. Valsala, V., Sreeush, M. G., and Chakraborty, K.: The IOD Impacts on the Indian Ocean Carbon Cycle, J. Geophys. Res.-Oceans, 125, e2020JC016485, https://doi.org/10.1029/2020JC016485, 2020. Takahashi, T., Sutherland, S. C., Sweeney, C., Poisson, A., Metzl, N., Tilbrook, B., Bates, N., Wanninkhof, R., Feely, R. A., Sabine, C., Olafsson, J., and Nojiri, Y.: Global sea–air CO2 ﬂux based on climatological surface ocean pCO2, and seasonal biological and temperature effects, Deep-Sea Res. Pt. II, 49, 1601–1622, 2002. Van Heuven, S., Pierrot, D., Rae, J., Lewis, E., and Wallace, D.: MATLAB program developed for CO2 system calcula- tions, ORNL/CDIAC-105b, Carbon Dioxide Information Anal- ysis Center, Oak Ridge National Laboratory, US Department of Energy, Oak Ridge, Tennessee, 530, 2011. Takahashi, T., Sutherland, S. C., Wanninkhof, R., Sweeney, C., Feely, R. A., Chipman, D. A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux P.: Seasonal spreading of the Persian Gulf Water mass in the Arabian Sea, J. Geophys. Res.- Oceans, 106, 17059–17071, 2001. Praveen, V., Valsala, V., Ajayamohan, R., and Balasubramanian, S.: Oceanic Mixing over the Northern Arabian Sea in a Warming Scenario: Tug of War between Wind and Buoyancy Forces, J. Phys. Oceanogr., 50, 945–964, 2020. Schott, F. A. and McCreary Jr., J. P.: The monsoon circulation of the Indian Ocean, Prog. Oceanogr., 51, 1–123, 2001. Sharada, M., Swathi, P., Yajnik, K., and Devasena, C. K.: Role of bi- ology in the air-sea carbon ﬂux in the Bay of Bengal and Arabian Sea, J. Earth Syst. Sci., 117, 429–447, 2008. Resplandy, L., Lévy, M., Madec, G., Pous, S., Aumont, O., and Ku- mar, D.: Contribution of mesoscale processes to nutrient bud- gets in the Arabian Sea, J. Geophys. Res.-Oceans, 116, C11, https://doi.org/10.1029/2011JC007006, 2011. Shchepetkin, A. F. and McWilliams, J. C.: The regional oceanic modeling system (ROMS): a split-explicit, free-surface, https://doi.org/10.5194/bg-19-907-2022 Biogeosciences, 19, 907–929, 2022 929 https://doi.org/10.5194/bg-19-907-2022 A. de Verneil et al.: Arabian Sea Air–Sea CO2 ﬂux W., Hales, B., Friederich, G., Chavez, F., Sabine, C., Watson, A., Bakker, D.C.E., Schuster, U., Metzl, N., Yoshikawa-Inoue, H., Ishii, M., Midorikawa, T., Nojiri, Y., Körtzinger, A., Steinhoff, T., Hoppema, M., Olafsson, J., Arnar- son, T. S., Tilbrook, B., Johannessen, T., Olsen, A., Bellerby, R., Wong, C. S., Delille, B., Bates, N. R., and de Baar, H. J. W.: Cli- matological mean and decadal change in surface ocean pCO2, and net sea–air CO2 ﬂux over the global oceans, Deep-Sea Res. Pt. II, 56, 554–577, 2009. Vic, C., Roullet, G., Carton, X., and Capet, X.: Mesoscale dynamics in the Arabian Sea and a focus on the Great Whirl life cycle: A numerical investigation using ROMS, J. Geophys. Res.-Oceans, 119, 6422–6443, https://doi.org/10.1002/2014JC009857, 2014. Wanninkhof, R.: Relationship between wind speed and gas ex- change over the ocean, J. Geophys. Res.-Oceans, 97, 7373–7382, 1992. Wanninkhof, R.: Relationship between wind speed and gas ex- change over the ocean revisited, Limnol. Oceanogr.-Meth., 12, 351–362, 2014. Takahashi, T., Sutherland, S. C., and Kozyr, A.: Global Ocean Sur- face Water Partial Pressure of CO2 Database: Measurements Performed During 1957–2018 (LDEO Database Version 2018) (NCEI Accession 0160492), NOAA National Centers for Envi- ronmental Information, 2019. Weiss, R.: Carbon dioxide in water and seawater: the solubility of a non-ideal gas, Mar. Chem., 2, 203–215, 1974. Wolf-Gladrow, D. A., Zeebe, R. E., Klaas, C., Körtzinger, A., and Dickson, A. G.: Total alkalinity: The explicit conservative ex- pression and its application to biogeochemical processes, Mar. Chem., 106, 287–300, 2007. Taylor, K. E.: Summarizing multiple aspects of model performance in a single diagram, J. Geophys. Res.-Atmos., 106, 7183–7192, 2001. Tjiputra, J. F., Olsen, A., Bopp, L., Lenton, A., Pfeil, B., Roy, T., Segschneider, J., Totterdell, I., and Heinze, C.: Long-term surface pCO2 trends from observations and models, Tellus B, 66, 23083, https://doi.org/10.3402/tellusb.v66.23083, 2014. Biogeosciences, 19, 907–929, 2022 Biogeosciences, 19, 907–929, 2022 https://doi.org/10.5194/bg-19-907-2022
https://openalex.org/W4213443430	https://www.frontiersin.org/articles/10.3389/fimmu.2022.834595/pdf	English	null	Matrix Remodeling-Associated Protein 8 as a Novel Indicator Contributing to Glioma Immune Response by Regulating Ferroptosis	Frontiers in immunology	2,022	cc-by	9,963	ORIGINAL RESEARCH published: 24 February 2022 doi: 10.3389/fimmu.2022.834595 Matrix Remodeling-Associated Protein 8 as a Novel Indicator Contributing to Glioma Immune Response by Regulating Ferroptosis Zhijie Xu 1†, Xi Chen 2,3†, Liying Song 4, Fang Yuan 5 and Yuanliang Yan 2,3* 1 Department of Pathology, Xiangya Hospital, Central South University, Changsha, China, 2 Department of Pharmacy, Xiangya Hospital, Central South University, Changsha, China, 3 National Clinical Research Center for Geriatric Disorders, Xiangya Hospital, Central South University, Changsha, China, 4 Department of Pharmacy, The Third Xiangya Hospital, Central South University, Changsha, China, 5 Department of Plastic Surgery, The Third Xiangya Hospital, Central South University, Changsha, China The co-cultured models of glioma cells and M2 macrophages showed MXRA8 knockdown glioma cells alleviated the inﬁltration of M2 macrophage, while the reduced Edited by: Junxia Zhang, Nanjing Medical University, China Edited by: Junxia Zhang, Nanjing Medical University, China Reviewed by: Ying Zhao, Zhejiang University, China Zhiwang Song, The First Afﬁliated Hospital of Nanchang University, China Correspondence: Yuanliang Yan yanyuanliang@csu.edu.cn †These authors have contributed equally to this work Reviewed by: Ying Zhao, Zhejiang University, China Zhiwang Song, The First Afﬁliated Hospital of Nanchang University, China Correspondence: Yuanliang Yan yanyuanliang@csu.edu.cn †These authors have contributed equally to this work Correspondence: Yuanliang Yan yanyuanliang@csu.edu.cn †These authors have contributed equally to this work Specialty section: This article was submitted to Cancer Immunity and Immunotherapy, a section of the journal Frontiers in Immunology Specialty section: This article was submitted to Cancer Immunity and Immunotherapy, a section of the journal Frontiers in Immunology Received: 13 December 2021 Accepted: 31 January 2022 Published: 24 February 2022 Received: 13 December 2021 Accepted: 31 January 2022 Published: 24 February 2022 Glioma is a highly malignant brain tumor with a poor survival rate. Novel biomarkers that act as prompt indicators of glioma are urgently needed. In this study, we identiﬁed and validated prognosis-related differentially expressed genes by datasets of glioma in the GEO and TCGA databases. Ferroptosis is a newly recognized process of cell death playing a vital role in cancer biology. Pearson correlation coefﬁcient were used to discovery the prognosis-related genes which have the highest correlation with ferroptosis. Matrix remodeling-associated protein 8 (MXRA8) was identiﬁed as a novel prognosis indicator which may be involved in ferroptosis. The expression of MXRA8 was signiﬁcantly higher in glioma compared with normal brain tissue, and increased expression of MXRA8 was associated with unfavorable survivals. Furthermore, in vitro analysis showed that knockdown of MXRA8 inhibited the cell viability in T98G and U251 cells and increased the sensitivity of glioma cells to temozolomide. We further observed that downregulation of MXRA8 elevated the levels of intracellular ferrous iron and lipid peroxidation, accompanied by upregulation of NCOA4 and suppression of FTH1. Moreover, co-expression analyses showed that GO term and KEGG pathways were mainly enriched in immunity-related pathways, such as neutrophil-related immunity, adaptive immune response, and cytokine binding. Through ssGSEA algorithm and TISIDB database, immunological analyses showed that MXRA8 was signiﬁcantly correlated with various immune inﬁltration cells including NK cells, macrophages, and neutrophils. Meanwhile, MXRA8 was also associated with chemokines and multiple immunoinhibitory molecules, such as TGF-b1, IL-10, PD-L1, and CTLA4. We also found that MXRA8 was positively associated with immune inﬁltration score, and patients with higher immune score underwent worse overall survivals. Moreover, IHC staining indicated a highly positive correlation of MXRA8 with a macrophage marker CSF1R. INTRODUCTION of the relationship between ferroptosis activation and immune system regulation. Glioma is one of the most common primary brain tumor, with an annual incidence of 3–8 per 100,000 population (1). As the most malignant subtype of glioma, glioblastoma multiforme (GBM) have an around 14 months of median survival (2, 3). Despite the availability of conventional treatments, such as surgical resection, adjuvant radiotherapy, and chemotherapy, the effects of prolonged survivals in GBM patients were still limited (4). Hence, exploring novel prognostic biomarkers for glioma therapy is crucial. In the present study, we thus explored the differential expression of ferroptosis-related genes in glioma samples through the GEO datasets and TCGA database. We identiﬁed that Matrix remodeling-associated protein 8 (MXRA8) is highly expressed in glioma sample and is associated with poor prognosis of glioma. MXRA8 is reported to confer debilitating acute and chronic musculoskeletal disease, since it serves as an entry mediator for multiple arthritogenic alphaviruses (15). In terms of tumor progression, several studies have reported that high expressions of MXRA8 had poor overall survival (OS) in kidney renal clear cell carcinoma patients and is correlated with tumor microenvironment in esophageal squamous cell carcinoma (16). However, the underlying functions of MXRA8 in gliomas have not been elucidated. To further understand the role of MXRA8 in glioma, in vitro analysis is conducted and shows that knockdown of MXRA8 inhibited cell proliferations, increased the TMZ sensitivity, and induced ferroptosis in glioma. Further bioinformatics analysis revealed that MXRA8 function could be enriched in immune responses, and MXRA8 expression was positively associated with immunoinhibitory chemokines and checkpoints. Our data suggest that MXRA8 might act as a potential prognostic marker for glioma involving ferroptosis and immunity mediation. Ferroptosis, a newly discovered form of regulated cell death, involves an iron-dependent accumulation of lethal lipid peroxidation (5). Dysregulation of iron metabolism were thought to be a hallmark for ferroptosis and triggers intracellular reactive oxygen species (ROS) production by the Fenton reaction (6). Furthermore, disruption of glutathione metabolism system could also cause overload of lipid peroxidation, leading to extremely harmful cell damages in the process offerroptosis (7). Over the past years, the essential role of ferroptosis in regulation of diverse physiological conditions and pathological progression has been identiﬁed in humans, and it has been gradually recognized as an adaptive process to render cancer cells vulnerable to death (8). Citation: Xu Z, Chen X, Song L, Yuan F and Yan Y (2022) Matrix Remodeling- Associated Protein 8 as a Novel Indicator Contributing to Glioma Immune Response by Regulating Ferroptosis. Front. Immunol. 13:834595. doi: 10.3389/fimmu.2022.834595 February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org Xu et al. MXRA8 as a Glioma Indicator M2 macrophage inﬁltration generated by MXRA8 could be rescued by Fer-1 treatment. These results suggest that MXRA8 promotes glioma progression and highlight the pivotal role of MXRA8 in ferroptosis and immune microenvironment of glioma. Therefore, MXRA8 may serve as a novel prognostic marker and therapeutic target for glioma. Keywords: MXRA8, immune response, ferroptosis, glioma, prognosis MATERIALS AND METHODS The DEGs in Glioma Screened From the GEO and TCGA Datasets INTRODUCTION Some studies have shown that tumor suppressor gene, TP53, could repress the synthesis of glutathione leading to ferroptosis promotion (9). In glioma, it has been reported that p53-dependent inhibition of SLC7A11, a unit of the glutamate-cystine antiporter required for ferroptosis induction, could deplete glutathione and elevate the levels of intracellular ferrous iron (Fe2+), H2O2, and lipid peroxidation (10). Furthermore, ferroptosis in glioma could be triggered by a battery of inducers, such as RSL3 and erastin, accompanied by suppression of proliferation and angiogenesis (11). Hence, induction of ferroptosis is emerging to be an effective strategy to eliminate glioma development, and the role of ferroptosis in glioma still needs to be further elucidated. Frontiers in Immunology \| www.frontiersin.org The DEGs in Glioma Screened From the GEO and TCGA Datasets The GEO (http://www.ncbi.nlm.nih.gov/geo) is an open-source platform for the storage of gene expression, chips, and microarray data. Two expression proﬁling datasets (GSE147352 and GSE59612) were, respectively, downloaded from the GEO database. The GSE147352 dataset includes 103 glioma and 15 normal brain tissues. The GSE59612 dataset includes 39 glioma and 17 nonneoplastic brain tissues. According to the setting cutoff criteria (fold change ≥2, p < 0.001), the differentially expressed genes (DEGs) between glioma and normal brain were identiﬁed. LASSO Cox regression was applied to predict the prognosis- related genes in glioma. For LASSO analysis, RNA-seq data and the latest clinical follow-up information were downloaded from the TCGA using GDC API (https://portal.gdc.cancer.gov/ repository), including 606 RNA-seq data samples. The gene expression proﬁles were normalized using the scale method provided in the R package “limma”. To analyze the gene Emerging studies have established that immunotherapy could become an effective method by manipulating the immune system to recognize and attack cancer cells (12). Recently, the crosstalk between ferroptosis and antitumor immunity has been increasingly recognized. Study showed that ferroptosis could be integrated with immunotherapy through activated CD8+ T cells, by enhancing lipid peroxidation in cancers (13). Additionally, tumor cells treated by ferroptosis inhibit increased the resistance of a PD-L1 inhibitor (13). CD8+ T cell-derived IFN-g cooperated with radiotherapy-activated ataxia-telangiectasia mutation (ATM) could induce ferroptosis in human ﬁbrosarcoma cells and melanoma cells (14). These studies provoked a new insight February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 2 Xu et al. MXRA8 as a Glioma Indicator demographic characteristics of MXRA8 expression and cancer immunology research in GBM, and then TISIDB was employed to cross-validate the roles of MXRA8 in immune-associated signaling pathways, such as tumor-inﬁltrating immune cells, and immunomodulators. signature that contains the most helpful biomarkers for prognosis, R package “glmnet” was used and then the risk score of each sample in all the datasets was calculated through the signature (17). For survival analysis, samples were divided into high- and low-risk groups relying on their own median risk score by using R package “survival.” To verify the prognostic value of the MXRA8- based classiﬁer, time-dependent (3-year, 5-year, and 10-year) receiver operating characteristics (ROC) were analyzed using the R package “timeROC” in the TCGA datasets (18). The GO and KEGG Pathway Analyses LinkedOmics is an open-access portal that provides multiomics data across various cancer types with three analytical modules: LinkFinder, LinkInterpreter, and LinkCompare (23). The heatmaps of positively or negatively correlated genes with MXRA8 were analyzed with the LinkFinder module. Correlation analyses were performed using Spearman’s correlation test. The LinkInterpreter module was used to analyze the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway. CCK-8 Assays and IC50 To explore the relationship between ferroptosis and our prognostic genes, we ﬁrst identiﬁed 60 ferroptosis-associated genes from the GeneCards (https://www.genecards.org/) (24), which contains up-to-date list of human ferroptosis-related genes. After merging samples, 142 glioma cases from the GEO database (GSE147352 and GSE59612) were qualiﬁed for subsequent analysis. Pearson’s test was performed to examine the correlation between ferroptosis-related DEGs and our prognostic gene expression. T98G and U251 cells were seeded in 96-well plates (2 × 103 cells/ well) after transfection with siRNAs for 24 h. Various concentrations of TMZ (0, 0.39, 1.58, 6.25, 25, 100, and 400 mM) were added into the medium for 48 h. On the day of analysis, CCK-8 assays were conducted to examine cell viability following the manufacturer’s protocols. In Brief, 10 ml CCK-8 test solution (B34304, Bimake, USA) was added into each well and incubated for 1 h. Optical density (OD) of the samples was detected at 450 nm using a VICTOR X2 microplate reader (PerkinElmer, USA). Variable slope was conducted to calculate IC50 values using the Graphpad Prism8.0 software [nonlinear regression; dose-response inhibition; log (inhibitor) vs. response- variable slope (4 parameters)]. The DEGs in Glioma Screened From the GEO and TCGA Datasets GEPIA2 is an interactive web server to analyze the RNA sequencing expression data of different cancer types and normal samples from the TCGA database and the GTEx projects, using a standard processing pipeline (26). Here, we computed the correlation between MXRA8 and immune monitoring sites in GBM and the corresponding normal tissues. Cell Cultures, Reagents, siRNA Transfections, and Macrophage Polarization T98G and U251 human glioma cell lines were obtained from the Cancer Research Institute, Central South University, China, as described in our previous study (27); glioma cells were incubated in DMEM (C11995500, HyClone, Logan, UT, USA) supplemented with 10% fetal bovine serum (04-001-1A, BI, Israel) and 1% penicillin and streptomycin (10378016, Gibco, Waltham, MA, USA) at 37°C with 5% CO2. Temozolomide (TMZ) was purchased from Selleckchem (S1237, Berlin, Germany). Ras-selective lethal small molecule 3 (RSL3) and ferrostatin-1 (Fer-1) were purchased from APExBIO (B6095 for RSL3, A4371 for Fer-1, Houston, TX, USA). RSL3 and FER-1 were dissolved in DMSO under sterile conditions to storage concentration of 10 mM. Small interfering RNAs targeting MXRA8 and negative control siRNAs were synthesized by RiboBio (Guangzhou, China). As previously described (28), siRNAs were transfected into T98G and U251 cells for 48 h using Lipofectamine 3000 (L300015, Thermo Fisher Scientiﬁc, USA) according to the manufacturer’s protocol. To polarize M0 macrophages to M1, THP-1 cells (2.5 × 105) were plated into 6-well plate and incubated with 320 nM phorbol 12- myristate 13-acetate (PMA; HY-18739, MedChemExpress, Monmouth Junction, NJ, USA) at 37°C for 6 h. To polarize M2 macrophages, cells were then treated with M2-polarizing reagents IL-4 (20 ng/ml; HY-P70445, MedChemExpress, USA) plus IL-13 (20 ng/ml; 200-13-10UG, Peprotech, USA) and incubated at 37°C for 72 h. Clinical Feature Analysis GEPIA2.0 (http://gepia.cancer-pku.cn/) (19) and UALCAN (http:// ualcan.path.uab.edu/analysis.html) (20) were used to access mRNA data of glioma and normal specimens from the TCGA database. Furthermore, the clinical and survival analyses were obtained from the CGGA database (http://www.cgga.org.cn). In total, 325 samples and 693 samples of RNA sequencing data from the CGGA database 2 (http://www.cgga.org) were used for subtype analysis of the grade, IDH type, methylation status, age, and survivals. Clinical specimens were collected from glioma patients admitted for operation in Beijing Tiantan Hospital (21). Xiantao tool (https://www.xiantao. love/products) is a comprehensive bioinformatics analysis portal which can perform differential expression, enrichment analysis, interaction network, and clinical signiﬁcance across pan-cancer types (22). Univariate regression analysis was performed to analyze the demographic characteristics of MXRA8 expression in glioma using the Xiantao tool. Cell Cultures, Reagents, siRNA Transfections, and Macrophage Polarization Colony Formation Assay y y The protocols used for the colony formation assay have been described previously (28). In brief, cells were transfected with siRNAs for 24 h and seeded in 6-well plates at a density of 1,000 cells per well. After incubating for 48 h, the cells were treated with different doses of TMZ. After incubation for approximately 14 days, colonies were stained with 0.3% w/v crystal violet/ methanol for 20 min at room temperature. Iron Assay Ferrous iron concentration was analyzed in T98G and U251 cells using an iron colorimetric assay kit (ab83366, Abcam, Cambridge, MA, UUSA) according to the manufacturer’s protocol. Brieﬂy, cells were digested by trypsin without EDTA and were lysed in acid assay buffer at a density of 1 × 106 cells per 1.5 ml tube. The samples were centrifuged at 14,000×g for 15 min to remove insoluble material. Ferene S was added to react with Fe2+, and a stable-colored complex was generated following the detection of absorbance at 593 nm by VICTOR X2 microplate reader (PerkinElmer, Waltham, MA, USA). The relative Fe2+ level was calculated as the ratio of concentration of all groups to concentration of siNC. Statistical Analysis All experiments were conducted and repeated at least three times and were reported as the means ± SD. The differences between which were analyzed for signiﬁcance using Student’s t-test for pairwise comparisons or ANOVA for multivariate analysis. Kaplan–Meier survival curves were estimated using the log- rank test to assess survival differences between groups. Data analysis was performed using GraphPad Prism 5 and SPSS 23.0. Differences were considered signiﬁcant at p < 0.05, p < 0.01, and p < 0.001 for all tests. Identiﬁcation of Prognostic-Related Gene MXRA8 The present study’s ﬂow diagram is shown in Figure 1A. Two datasets from the GEO database (GSE147352 and GSE59612) were selected which have more than 50 samples. A total of 2,113 and 7,010 DEGs were identiﬁed between tumor and normal tissues, and 820 codifference genes were qualiﬁed for subsequent analysis (Figure 1B). Combined with the prognostic information, LASSON regression was then implemented to screen prognostic genes from the DEGs in the TCGA datasets. Finally, we found that 45 DEGs were signiﬁcantly correlated with the OS of glioma patients (Figure 1C; Supplemental Table S1). Ferroptosis is an iron-dependent cell death with distinct Malondialdehyde Determination Malondialdehyde (MDA) lipid peroxidation assay (MAK085, Sigma, St. Louis, MO, USA) was used according to manufacturers’ instructions. In brief, cells were collected by trypsin and lysed at a density of 1 × 106 cells per 1.5 ml tube. Furthermore, lysis solution was reacted with thiobarbituric acid. Colorimetric assays were used to measure the absorbance at 532 nm by VICTOR X2 microplate reader (PerkinElmer, USA). The relative MDA level was calculated as the ratio of concentration of all groups to concentration of siNC. Immunohistochemistry The glioma and normal brain tissues were obtained from the Department of Pathology, Xiangya Hospital. The tissues were parafﬁn embedded and the immunohistochemistry (IHC) was performed as previously described (29, 30). In brief, IHC was conducted using a Histomouse SP Kit (959551; Invitrogen, Waltham, MA, USA). Parafﬁn sections were immunostained through a streptavidin peroxidase procedure after microwave antigen retrieval. The signal was detected using a 3,3′- diaminobenzidine solution. The antibodies against MXRA8 (1:100, ab185444) and CSF1R (1:100, ab183316) were all purchased from Abcam. Images of the sections were independently examined and differentially quantiﬁed by two pathologists. IHC intensity score was scored as 0 (negative), 1 (weak brown), 2 (moderate brown), or 3 (strong brown). The extent of staining was scored as 0 (≤10%), 1 (11%–25%), 2 (26%– 50%), 3 (51%–75%), or 4 (>75%). The ﬁnal staining score was determined by multiplication of intensity scores and extent score and was classiﬁed as weakly positive (1–3), positive (4–6), and strongly positive (7–12). All parafﬁn-embedded specimens were collected following the ethical standards of the human experimental committee. Western Blot Analysis Cells were collected and lysed with RIPA lysis buffer with protease inhibitor cocktails (B14012, Bimake, Houston, TX, USA) at a ratio of 100:1 (v/v). The 50-µg protein extracts were equally loaded on 10% or 12% SDS-PAGE and then transferred to PVDF membranes (0.22 µm: ISEQ00010; 0.45 µm: IPVH00010). Following blocking with 5% skimmed milk for 1 h at room temperature, then membranes were incubated with primary antibodies diluted in 5% bovine serum albumin (D620272, Sangon Biotech, Shanghai, China) overnight at 4°C. Primary antibodies including MXRA8 antibody (1:1,000; ab185444, Abcam), NCOA4 antibody (1:1,000; ab86707, Abcam), FTH1 antibody (1:1,000; 4393, Cell Signaling Technology, Danvers, MA, USA), and GAPDH (1:10,000; 60004-1-Ig, Proteintech, Rosemont, IL, USA). For detection, the membranes were visualized by Immobilon Western chemiluminescent reagents (WBKLS0500, Millipore, Burlington, MA, USA). Immunological Analysis Interactions between tumor and immune system were performed with the TISIDB database (25) and Xiantao tool (https://www. xiantao.love/products). We used the Xiantao tool to identify the February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 3 MXRA8 as a Glioma Indicator Xu et al. a Transwell plate (size 8 mm, Corning, Corning, NY, USA), while 2.5 × 105 T98G and U251 glioma cells were incubated with10% FBS in bottom plate. After coincubation at 37°C for 24 h, the cells in the upper chamber were ﬁxed in 4% formalin and stained with 0.3% crystal violet. The inﬁltrated M2 macrophage cells were counted in three randomly selected ﬁelds from each membrane. M2 Macrophage Inﬁltration Assays The M2 macrophage inﬁltration assays were performed as previous described (21). In brief, 2.5 × 105 M2 macrophage cells were seeded without serum for 12 h in the upper chamber of February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 4 MXRA8 as a Glioma Indicator Xu et al. Xu et al. A B C D E F G H I J GURE 1 \| MXRA8 was identiﬁed as a novel prognostic-related gene in glioma. (A) Flowchart presenting the process of identifying the gene and prognostic mogram of glioma in this study. (B) Venn diagram to identify differentially expressed genes between glioma and normal brain tissue from the GEO database. 4 overlapping genes were all upregulated in tumor tissue. The 416 overlapping genes were all downregulated in tumor tissue. (C) Prognostic genes were ide ing the least absolute shrinkage and selection operator Cox regression model (LASSO). The generated coefﬁcient distribution plots for the logarithmic (lambd quence for the selection of the best parameter (lambda). (D) Heatmap of the correlations between 35 upregulated prognostic genes and the 14 ferroptosis-r EGs. (E, F) The expression of MXRA8 was compared between normal and GBM tissue in the GSE59612 (E) and GSE147352 (F) cohorts. (G) AUC of time- pendent ROC curves veriﬁed the prognostic performance of the risk score based on MARA8 in the TCGA cohort. Three-year survival (H), 5-year survival (I), -year survival (J) curves for the OS of patients in the high-risk and low-risk groups in the TCGA cohort. A B B D C D D C G F G E E F G F G F E I J I H I J H FIGURE 1 \| MXRA8 was identiﬁed as a novel prognostic-related gene in glioma. (A) Flowchart presenting the process of identifying the gene and prognostic nomogram of glioma in this study. (B) Venn diagram to identify differentially expressed genes between glioma and normal brain tissue from the GEO database. The 404 overlapping genes were all upregulated in tumor tissue. The 416 overlapping genes were all downregulated in tumor tissue. (C) Prognostic genes were identiﬁed using the least absolute shrinkage and selection operator Cox regression model (LASSO). The generated coefﬁcient distribution plots for the logarithmic (lambda) sequence for the selection of the best parameter (lambda). (D) Heatmap of the correlations between 35 upregulated prognostic genes and the 14 ferroptosis-related DEGs. Validation of the Expression Level and Clinical Signiﬁcance of MXRA8 y To further examine the biological effect of MXRA8 in glioma, we knocked down MXRA8 with two small interfering RNAs in T98G and U251 cell lines (Figure 4A). Cell growth was signiﬁcantly decreased in siMXRA8 compared with siNC (Figure 4B). Likewise, in colony formation assays, cell proliferations were signiﬁcantly inhibited after transfection of siMXRA8 in T98G and U251 cells (Figure 4C). Overcoming the resistance of TMZ has been a vital problem in glioma treatment (33). To conﬁrm whether MXRA8 could inﬂuence TMZ sensitivity in glioma, T98G and U251 cells were transfected with MXRA8 siRNA or siNC, followed by temozolomide incubation. Upon knockdown of MXRA8, T98G and U251 cells both showed increased sensitivity to TMZ, which manifested as reduced cell proliferation rates and about a 6- fold inhibition in IC50 (Figure 4D). Similarly, MXRA8 reduction markedly decreased the colony formation rates of T98G and U251 cells after treatment of 200 and 100 mM TMZ, respectively (Figure 4E). Collectively, these results suggest that downregulation of MXRA8 improved the sensitivity of glioma to TMZ. To further validate the role of MXRA8 in the malignant progression of glioma, we analyzed its expression levels in the TCGA datasets from the GEPIA2.0 and UALCAN platform, respectively. As shown in Figure 2A, MXRA8 expression levels were higher in glioma than those in normal specimens. We also detected the expression of MXRA8 in different grades of glioma from two CGGA datasets (RNAseq_325 and RNAseq_693) and observed that MXRA8 was obviously elevated in high-grade glioma (Figure 2B). IDH mutation and 1p/19q codeletion are distinctive indicators in glioma, with a tendency of favorable prognosis (32). We then explored the association between MXRA8 expression and the status of IDH and 1p/19q. MXRA8 expression signiﬁcantly increased in patients with wild-type IDH and noncodel 1p/19q (Figures 2C, D). Additionally, patients over 42 years old showed higher MXRA8 expression (Figure 2E). The lower expression of MXRA8 was correlated with better prognosis of gliomas (Figure 2F). Furthermore, the demographic characteristics of MXRA8 from the Xiantao tool were in satisfactory concordance with the above results (Table 1). M2 Macrophage Inﬁltration Assays (E, F) The expression of MXRA8 was compared between normal and GBM tissue in the GSE59612 (E) and GSE147352 (F) cohorts. (G) AUC of time- dependent ROC curves veriﬁed the prognostic performance of the risk score based on MARA8 in the TCGA cohort. Three-year survival (H), 5-year survival (I), and 10-year survival (J) curves for the OS of patients in the high-risk and low-risk groups in the TCGA cohort. February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org Xu et al. MXRA8 as a Glioma Indicator properties and recognizing functions involved in cancers (31). The correlations between the 45 prognostic genes and 14 ferroptosis-related DEGs in the GEO dataset were analyzed in Figure 1D and Supplementary Figure S1. Interestingly, a receptor for multiple arthritogenic alphaviruses, MXRA8, which showed the highest correlation with ferroptosis may be involved in glioma progression and prognosis. GSE59612 and GSE147352 datasets respectively showed that expression of MXRA8 was signiﬁcantly higher in gliomas than those in normal tissues (Figures 1E, F). Furthermore, to determine the prognostic effect of MXRA8, time-dependent ROC was applied to determine the prognostic values of the MXRA8-based risk scores. The AUCs for 3-, 5-, and 10-year overall survival predictions for the risk scores were 0.780, 0.772, and 0.754, respectively (Figure 1G). Patients from the TCGA dataset were stratiﬁed into two (high-risk and low-risk) groups. The survival curves revealed signiﬁcantly favorable overall survival in all groups with lower-risk scores (p = 0.000) both in 3, 5, and 10 years (Figures 1H–J) . expression genes of MXRA8 mainly participated in extracellular matrix structural constituent, peptidase regulator activity, cytokine binding, antigen binding, immunoglobulin binding, etc. (Figure 3C). GO-CC analysis showed that these genes were signiﬁcantly enriched in vesicle lumen, endoplasmic reticulum lumen, primary lysosome, ﬁcolin-1-rich granule, vacuolar lumen, and so on (Figure 3D). GO-BP enrichment analysis suggested that these genes were signiﬁcantly related to neutrophil-mediated immunity, granulocyte activation, interferon-gamma production, response to interferon-gamma, adaptive immune response, etc. (Figure 3E). KEGG pathway analysis indicated these genes were signiﬁcantly related to inﬂammation and immune-associated signaling pathways, such as Staphylococcus aureus infection, autoimmune thyroid disease, leishmaniasis, intestinal immune network for IgA production, allograft rejection, etc. (Figure 3F). These ﬁndings suggested that MXRA8 might be involved in network of immunity-related functions in glioma. The Effects of MXRA8 on M2 Macrophage Inﬁltration level of NCOA4 was increased in MXRA8 knockdown glioma cells, while FTH1 was reduced (Figures 5A, B). Ferroptosis is a process characterized by increased intracellular Fe2+ and lipid peroxidation, and the ﬁnal product of lipid peroxidation is MDA. After MXRA8 knockdown, the intracellular Fe2+ and MDA levels were signiﬁcantly increased by approximately 2-fold in T98G and U251 cells compared with controls (Figures 5C, D). To further understand the role of MXRA8 in cell proliferation and ferroptosis, Fer-1, a small molecule that serves as a ferroptosis inhibitor (34), was treated after MXRA8 knockdown. The results showed that Fer-1 could rescue the effects of cell growth inhibition induced by MXRA8 knockdown (Figure 5E). Moreover, Fer-1 reversed the increase of intracellular Fe2+ and MDA induced by downregulated MXRA8 (Figures 5F, G). Also, the protein expression of MXRA8 was reduced after treatment of ferroptosis agonist RSL3 (Figure 5H). Taken together, these data indicated that knockdown of MXRA8 could suppress cell survival and enable ferroptosis in glioma cells. The above results conﬁrmed positive correlation between MXRA8 and macrophage inﬁltrations (Figures 6A, B) and cytokines (such as IL-10 and TGF-b1) that are involved in tumor-associated macrophage inﬁltration were highly consistent with expression of MXRA8. Through a correlation coefﬁcient analysis performed in the GEO datasets (GSE59612 and GSE147352), we investigated the relationship between the expression of MXRA8 and the typical phenotype macrophage markers, including M (AIF1), M1 (IL12A, TNF, NOS2, PTGS2), and M2 (IL10, CCL163, TGFB1, CSF1R) (21) (Figure 7A; Table 2). The M2 macrophage markers CSF1R showed signiﬁcantly positive correlation (cor_spearman >0.3, p < 0.05) with MXRA8 in both datasets. Using 10 parafﬁn- embedded specimens, IHC staining indicated the intensity of MXRA8 and CSF1R was stronger in GBM than normal. The staining intensity of macrophage M2 marker CSF1R was positively correlated with MXRA8 (Pearson’s r = 0.697, p = 0.025) (Figure 7B). To further investigate MXRA8 effect in ferroptosis and macrophage inﬁltration, THP-1 cells were differentiated into M2 macrophage, and U251 and T98G cells were treated with Fer-1 after knockdown of MXRA8. The results showed that downregulation of MXRA8 decreased the inﬁltration of M2 macrophage, while the reduced M2 macrophage inﬁltration generated by MXRA8 knockdown was rescued by Fer-1 treatment (Figure 7C). The Coexpression Network of MXRA8 in Glioma g The correlations between the 45 prognostic genes and 14 ferroptosis-related DEGs in the GEO dataset are shown in Figure 1D and Supplementary Figure S1. Pearson’s correlation coefﬁcient indicated ferroptosis-related genes including DPP4, GCLM, HSPB1, NFE2L2, SAT1, STEAP3, and TFRC may be involved in MXRA8 function (Pearson’s r >0.5). To elucidate the molecular mechanism of MXRA8, the correlation between MXRA8 expression and ferroptosis markers, NCOA4 and FTH1 were performed using the two CGGA datasets (mRNA_seq 323 and 693). The expression of MXRA8 was positively associated with levels of NCOA4 and was negatively related to levels of FTH1. Also, the protein expression In order to explore the biological functions of MXRA8 in glioma, we performed the coexpression pattern of MXRA8 screened from the TCGA-GBMLGG cohort by the LinkFinder module of LinkedOmics. Figures 3A, B shows the heatmaps of top 50 genes positively and negatively correlated with MXRA8, respectively. Additionally, we further performed GO and KEGG functional enrichment analyses to elaborate the potential biological roles of MXRA8 in GBM through the LinkInterpreter module. GO enrichment analysis mainly includes three parts: molecular function (MF), cellular component (CC), and biological process (BP). GO-MF enrichment analysis revealed that co- February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 6 Xu et al. Xu et al. MXRA8 as a Glioma Indicator A B C D E F tion of the expression level and clinical signiﬁcance of MXRA8. (A) Expression comparisons of MXRA8 between glioma and normal control CAN platforms. The “asterisk” represents p < 0.01. (B–F) The correlation of MXRA8 expression level with WHO grade (B), IDH status (C), D), ages (E), and survivals (F) from the CGGA database. A B C B C C D E D E F FIGURE 2 \| Validation of the expression level and clinical signiﬁcance of MXRA8. (A) Expression comparisons of MXRA8 between glioma and normal control from GEPIA2.0 and UALCAN platforms. The “asterisk” represents p < 0.01. (B–F) The correlation of MXRA8 expression level with WHO grade (B), IDH status (C), 1p/19q codeletion status (D), ages (E), and survivals (F) from the CGGA database. February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 7 MXRA8 as a Glioma Indicator Xu et al. TABLE 1 \| The demographic characteristics of MXRA8 expression in glioma. Characteristics Total (N) Odds ratio (OR) p-value WHO grade (G4 vs. The Role of MXRA8 in Immune Mircoenviroment of Glioma Increasing evidence has revealed the association between ferroptosis and immune microenvironment (35). Therefore, we further explored the role of MXRA8 in the immune microenvironment of glioma. The correlations between MXRA8 and various immune inﬁltration cells were analyzed by Xiantao tool using ssGSEA algorithm. Natural killer cells (NK), macrophages, and neutrophils were found to be positively correlated with MXRA8 in glioma (Figures 6A–D). Using TISIDB database, we further analyzed the associations between MXRA8 with immunoinhibitors. As shown in Figures 6E, F, indicated immunoinhibitors were found to have signiﬁcant associations with the expression of MXRA8, while transforming growth factor beta 1 (TGFB1) showed the highest correlation coefﬁcient. Moreover, Figures G–O shows the positive associations between MXRA8 expression levels and multiple immune checkpoint molecules, including PD-L1 and B7-H3. In addition, Figures 6P–R exhibits a positive correlation between MXRA8 and immune inﬁltration score in GSE59612, GSE147352, and TCGA glioma datasets. We further investigated the prognostic value of the immune inﬁltration score in the TCGA glioma dataset, and the results showed that patients with higher immune inﬁltration score had a poorer OS (Figure 6S). These results together indicated that MXRA8 might inﬂuence the prognosis of glioma patients through regulating the immune response. The Coexpression Network of MXRA8 in Glioma G2 and G3) 635 10.482 (6.656–17.150) <0.001 1p/19q codeletion (noncodel vs. codel) 689 4.732 (3.207–7.117) <0.001 Primary therapy outcome (PD vs. CR and PR and SD) 462 1.682 (1.091–2.592) 0.018 IDH status (WT vs. Mut) 686 10.629 (7.273–15.836) <0.001 Histological type (glioblastoma vs. astrocytoma and oligoastrocytoma and oligodendroglioma) 696 9.529 (6.084–15.516) <0.001 Gender (male vs. female) 696 1.098 (0.813–1.484) 0.540 Age (>60 vs. ≤60) 696 3.237 (2.182–4.882) <0.001 Race (White vs. Asian and Black or African American) 683 0.920 (0.502–1.676) 0.784 TABLE 1 \| The demographic characteristics of MXRA8 expression in glioma. DISCUSSION In this work, we aimed to investigate critical and novel indicators for prognosis of glioma. Through GEO datasets and TCGA database, 45 DEGs were screened, as these genes were strongly associated with OS of glioma patients. Intriguingly, among the 45 DEGs, MXRA8 was screened as an indicator of poor prognosis with malignancy clinical characteristics of malignancy including IDH wild type and 1p/19q noncodeletion. Furthermore, we demonstrated that knockdown of MXRA8 inhibited cell proliferations in vitro and contributed to increasing sensitivity of glioma cells to TMZ. These data ﬁrstly provide a novel perspective of MXRA8 which could serve as a prognostic indicator in glioma. Emerging evidence showed that ferroptosis could be triggered by several signal pathways and/or small molecular compounds in glioma. For instance, dihydroartemisinin was identiﬁed to trigger ferroptosis accompanied with ROS generation and lipid February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 8 MXRA8 as a Glioma Indicator Xu et al. B D F (A, B) Heatmaps of top 50 genes positively and negatively related to MXRA8. Red represents enes. (C–F) GO terms for molecular functions, cellular component, biological process, and KEGG A B C D E F FIGURE 3 \| The co-expression network of MXRA8 in glioma. (A, B) Heatmaps of top 50 genes positively and negatively related to MXRA8. Red represents positively linked genes and blue represents negatively linked genes. (C–F) GO terms for molecular functions, cellular component, biological process, and KEGG pathways of MXRA8 by GSEA analyses. B A A B D F of top 50 genes positively and negatively related to MXRA8. Red represents rms for molecular functions, cellular component, biological process, and KEGG C D C D E E F FIGURE 3 \| The co-expression network of MXRA8 in glioma. (A, B) Heatmaps of top 50 genes positively and negatively related to MXRA8. Red represents positively linked genes and blue represents negatively linked genes. (C–F) GO terms for molecular functions, cellular component, biological process, and KEGG pathways of MXRA8 by GSEA analyses. February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 9 MXRA8 as a Glioma Indicator Xu et al. Xu et al. A B C E D FIGURE 4 \| Knockdown of MXRA8 inhibits proliferation and increased the sensitivity of TMZ of glioma. DISCUSSION (A) After U251 and T98G cells were transfected with siMXRA8 or siNC for 48 h, the protein expression of MXRA8 was analyzed by Western blot assay. (B) T98G and U251 cell proliferation was examined after treating with siMXRA8 or siNC for 5 days. (C) The colony formation was analyzed in T98G and U251 cells after treating with siMXRA8 or siNC. (D) Cell viability assay was performed in U251 and T98G cells to evaluate cytotoxicity of TMZ after transfection of siMXRA8. (E) The colony formation assay was conducted to examine after incubation with TMZ in U251 and T98G cells transfected with siMXRA8 or siNC. p < 0.01; p < 0.001. A B C E D FIGURE 4 \| Knockdown of MXRA8 inhibits proliferation and increased the sensitivity of TMZ of glioma. (A) After U251 and T98G cells were transfected with siMXRA8 or siNC for 48 h, the protein expression of MXRA8 was analyzed by Western blot assay. (B) T98G and U251 cell proliferation was examined after treating A B D B A C E D C E E FIGURE 4 \| Knockdown of MXRA8 inhibits proliferation and increased the sensitivity of TMZ of glioma. (A) After U251 and T98G cells were transfected with siMXRA8 or siNC for 48 h, the protein expression of MXRA8 was analyzed by Western blot assay. (B) T98G and U251 cell proliferation was examined after treating with siMXRA8 or siNC for 5 days. (C) The colony formation was analyzed in T98G and U251 cells after treating with siMXRA8 or siNC. (D) Cell viability assay was performed in U251 and T98G cells to evaluate cytotoxicity of TMZ after transfection of siMXRA8. (E) The colony formation assay was conducted to examine after incubation with TMZ in U251 and T98G cells transfected with siMXRA8 or siNC. p < 0.01; *p < 0.001. peroxidation (36). Furthermore, erastin could sensitize GBM cells to temozolomide by restraining SLC7A11 and cystathionine-g- lyase function (37). This evidence suggests that targeting the pathway of ferroptosis could bring out the new strategies for glioma treatment. In our studies, we ﬁrstly found that downregulation of MXRA8 could induce ferroptosis by elevating the Fe2+ and MDA levels in glioma cells. Furthermore, we observed a signiﬁcant correlation between MXRA8 expression and ferroptosis-related marker NCOA4 and FTH1 according to the CCGA datasets. Frontiers in Immunology \| www.frontiersin.org DISCUSSION NCOA4 is a selective cargo receptor for the selective autophagic turnover of ferritin in ferroptosis, while ferritin heavy polypeptide 1 (FTH1) is a subunit of ferritin. Since degradation of ferritin could elevate iron levels and generate cellular oxidative stress by the Fenton reaction (38), it has been proven that NCOA4 could bind up to FTH1 and thus deliver ferritin to the lysosome for degradation (39). In glioma, a study reported that NCOA4 caused degradation of ferritin and a subsequent increase in levels of Fe2+and ultimately ferroptosis (40). Likewise, our study also demonstrated that inhibition of MXRA8 increased the protein expression of NCOA4 and decreased FTH1 protein levels in glioma cells. These above results could help us further understand the potential role of MXRA8 in mediating ferroptosis. Multiple studies have revealed that immunosuppressive microenvironment was a nature of glioma, which further inhibits antitumor immune responses (41). An amount of immune cells, chemokines, and checkpoints were recognized as biomarkers of immune evasion and tumor progression. Glioma- associated macrophages could secrete cytokines such as IL-10 and TGF-b, which decrease the activities of immune cells and promote tumorigenesis (42, 43). Glioma-associated neutrophil inﬁltration which promoted tumor growth and resistance to antiangiogenic therapies was considered an indicator of poor prognosis (44, 45). Moreover, glioma cells express increased levels of immunosuppressive factors such as PD-L1, TIGIT, CD96, and CTLA-4, which negatively regulate presentation of antigens and T-cell responses (46–48). Especially, expression of PD-L1 could respond to the presence of T-cell-generating February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 10 MXRA8 as a Glioma Indicator Xu et al. Xu et al. A B C F D G H E 5 \| MXRA8 regulates ferroptosis in glioma. (A) The correlation between expression of MXRA8 and ferroptosis-related proteins including NCOA4 and FTH1. A4 and FTH1 were examined by Western blot in U251 and T98G cells transfected with siNC or siMXRA8. (C, D) Fe2+ and MDA levels were detected in d T98G cells transfected with siNC or siMXRA8. (E) Cell viability was measured in siNC- or siMXRA8-transfected cells after pretreatment with RSL3 (5 mM) sequent Fer-1 (20 mM). (F, G) Fe2+ and MDA levels detected in siNC- or siMXRA8-transfected cells pretreated with RSL3 (5 mM) and Fer-1 (20 mM) for 24 h, vely. DISCUSSION (H) After treatment of 5 mM of RSL3 in U251 and T98G cells, Western blot assay were performed to examine the protein expression of MXRA8. 5; p < 0.01. A C B C B B F D F D D G H E E G FIGURE 5 \| MXRA8 regulates ferroptosis in glioma. (A) The correlation between expression of MXRA8 and ferroptosis-related proteins including NCOA4 and FTH1. (B) NCOA4 and FTH1 were examined by Western blot in U251 and T98G cells transfected with siNC or siMXRA8. (C, D) Fe2+ and MDA levels were detected in U251 and T98G cells transfected with siNC or siMXRA8. (E) Cell viability was measured in siNC- or siMXRA8-transfected cells after pretreatment with RSL3 (5 mM) and subsequent Fer-1 (20 mM). (F, G) Fe2+ and MDA levels detected in siNC- or siMXRA8-transfected cells pretreated with RSL3 (5 mM) and Fer-1 (20 mM) for 24 h, successively. (H) After treatment of 5 mM of RSL3 in U251 and T98G cells, Western blot assay were performed to examine the protein expression of MXRA8. p < 0.05; p < 0.01. February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org Xu et al. Xu et al. MXRA8 as a Glioma Indicator A C D B E G H F I K L J M O P N Q S R GURE 6 \| The role of MXRA8 in immune microenvironment of glioma. (A) Immune inﬁltration cells associated with MXRA8 using ssGESA algorithm from the TCGA atabase. (B–D) Validations of NK cells, macrophages, and neutrophils by the TISIDB database. (E–O) Associations of the MXRA8 expression with chemokines E, F) and multiple immunoinhibitory molecules (G–O) from the TISIDB database. (P–R) The association between MXRA8 with immune score calculated by ESTIMAT gorithm, based on GSE59612, GSE147352, and TCGA glioma datasets. (S) The Kaplan–Meier curve of immune score on overall survival of TCGA glioma datasets B C D B C D A B C D G F H E F E G H J I K L I L J K P O M N P M N O S R Q S S R R Q FIGURE 6 \| The role of MXRA8 in immune microenvironment of glioma. (A) Immune inﬁltration cells associated with MXRA8 using ssGESA algorithm from the TCGA database. (B–D) Validations of NK cells, macrophages, and neutrophils by the TISIDB database. DISCUSSION (E–O) Associations of the MXRA8 expression with chemokines (E, F) and multiple immunoinhibitory molecules (G–O) from the TISIDB database. (P–R) The association between MXRA8 with immune score calculated by ESTIMATE algorithm, based on GSE59612, GSE147352, and TCGA glioma datasets. (S) The Kaplan–Meier curve of immune score on overall survival of TCGA glioma datasets. February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 12 Xu et al. MXRA8 as a Glioma Indicator Xu et al. ki i f d i d i f T ll killi l l f h i i hibi f i l d i A B C GURE 7 \| The effects of MXRA8 on M2 macrophage inﬁltration in vitro by mediating ferroptosis. (A) The correlation coefﬁcient between MXRA8 and classical enotype markers of macrophages in the GEO datasets. (B) The expression level of MXRA8 and a M2 macrophage phenotype marker (CSF1R) in GBM and mal brain specimens determined by IHC analysis. (C) Inﬁltration of M2 macrophages in transfected siNC or siMXRA8 cells (U251 and T98G) after pretreatment h RSL3 (5 mM) and subsequent Fer-1 (20 mM). p < 0.001. A B B B C RE 7 \| The effects of MXRA8 on M2 macrophage inﬁltration in vitro by mediating ferroptosis. (A) The correlation coefﬁcient between MXRA8 and classical type markers of macrophages in the GEO datasets. (B) The expression level of MXRA8 and a M2 macrophage phenotype marker (CSF1R) in GBM and brain specimens determined by IHC analysis. (C) Inﬁltration of M2 macrophages in transfected siNC or siMXRA8 cells (U251 and T98G) after pretreatment SL3 (5 mM) and subsequent Fer-1 (20 mM). p < 0.001. C C C FIGURE 7 \| The effects of MXRA8 on M2 macrophage inﬁltration in vitro by mediating ferroptosis. (A) The correlation coefﬁcient between MXRA8 and classical phenotype markers of macrophages in the GEO datasets. (B) The expression level of MXRA8 and a M2 macrophage phenotype marker (CSF1R) in GBM and normal brain specimens determined by IHC analysis. (C) Inﬁltration of M2 macrophages in transfected siNC or siMXRA8 cells (U251 and T98G) after pretreatment with RSL3 (5 mM) and subsequent Fer-1 (20 mM). **p < 0.001. levels of these immunoinhibitory factors involved in microenvironment of glioma. Additionally, immune inﬁltration score in the TCGA glioma datasets showed that patients with high immune inﬁltration had a tendency with upregulation of MXRA8, which couples with worse OS. DISCUSSION These results altogether imply that the expression of MXRA8 might be parallel to cytokine interferon gamma and trigger reduction of T-cell killing capacity, which further caused adaptive immune resistance (49, 50). In the present study, the results showed MXRA8 might participate in neutrophil-related immunity, adaptive immune response, and cytokine binding. Furthermore, our studies revealed that MXRA8 expression was positively associated with February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 13 MXRA8 as a Glioma Indicator Xu et al. TABLE 2 \| The correlation coefﬁcient analysis between MXRA8 and phenotype macrophage markers in the GEO datasets. GSE147352 GSE147352 Gene name Gene name cor_spearman p-value Gene name Gene name cor_spearman p-value MXRA8 IL12A −0.422 0.007 MXRA8 IL12A −0.190 0.002 NOS2 −0.011 0.948 TNF −0.041 <0.001 IL10 0.163 0.007 CSF1R 0.331 0.025 PTGS2 0.343 0.322 PTGS2 0.350 <0.001 CSF1R 0.412 0.016 IL10 0.471 <0.001 TGFB1 0.491 0.038 AIF1 0.557 <0.001 TABLE 2 \| The correlation coefﬁcient analysis between MXRA8 and phenotype macrophage markers in the GEO datasets. SUPPLEMENTARY MATERIAL The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/ﬁmmu.2022.834595/ full#supplementary-material ETHICS STATEMENT immunosuppression and developed as a therapeutic target of immunotherapy in glioma. The studies involving human participants were reviewed and approved by the ethics committee of Xiangya Hospital, Central South University. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. The role of macrophages in tumor growth was complex in dependence on various factors and might be polarized to antitumor (M1) or protumor (M2) phenotypes (51). Targeting M2 macrophages in the glioma microenvironment could become an alternative therapeutic strategy. CSF1R-regulated signaling is crucial for the differentiation of M2 macrophage polarization in various cancers, and blockade of CSF1R could convert macrophage polarization and inhibit glioma progression (52, 53). In our study, we found that CSF1R was positively correlated with MXRA8 expression in GBM, which suggested that functional activity of MXRA8 might be related to regulation of M2 macrophages. Furthermore, several studies have reported that ferroptosis-related gene signature promotes tumor progression by inﬂuencing macrophage inﬁltration in various cancers (54, 55), but the biological roles of ferroptosis in glioma-associated macrophages were still unclear. Our present study ﬁrstly demonstrated that knockdown of MXRA8 could inhibit M2 macrophage migration to gliomacellsthroughregulationofferroptosis.Theseﬁndingsindicate thattargetingMXRA8-relatedferroptosismightbecomeapromising strategy for immunotherapy. FUNDING This study is supported by grants from the China Postdoctoral Science Foundation (2021T140754, 2020M672521), the Science and Technology Innovation Program of Hunan Province (2021RC3029), the Natural Science Foundation of Hunan Province (2020JJ5934), and the Postdoctoral Science Foundation of Central South University (248485). From now on, the integrationbetween ferroptosis and immunity environment remains unclear, especially in glioma. Our study uncovers the essential roles of MXRA8 in glioma progression and highlights the importance of MXRA8 in regulating ferroptosis and immune microenvironments. We identiﬁed that MXRA8 supersession could enable ferroptosis activation, and a hint of MXRA8 in regulation of immune microenvironment was provided. Therefore, our ﬁndings suggest that MXRA8 could be a promising prognostic biomarker in patients with glioma. DATA AVAILABILITY STATEMENT Supplementary Table 1 \| 45 DEGs were signiﬁcantly correlated with the OS of glioma patients. Supplementary Table 1 \| 45 DEGs were signiﬁcantly correlated with the OS of glioma patients. The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material. Supplementary Figure 1 \| Heatmap of the correlations between 10 downregulated prognostic genes and the 14 ferroptosis related DEGs. AUTHOR CONTRIBUTIONS ZX and YY contributed to the conception and design of the study. YY and XC contributed to the writing, review, and/or revision of the manuscript. XC, YF, and LS provided administrative, technical, or material support. All authors contributed to the article and approved the submitted version. 2. Bai J, Varghese J, Jain R. Adult Glioma WHO Classiﬁcation Update, Genomics, and Imaging: What the Radiologists Need to Know. Top Magn Reson Imaging (2020) 29(2):71–82. doi: 10.1097/RMR.0000000000000234 3. Chen X, Yan Y, Zhou J, Huo L, Qian L, Zeng S, et al. Clinical Prognostic Value of Isocitrate Dehydrogenase Mutation, O-6-Methylguanine-DNA 1. Gusyatiner O, Hegi ME. Glioma Epigenetics: From Subclassiﬁcation to Novel Treatment Options. Semin Cancer Biol (2018) 51:50–8. doi: 10.1016/ j.semcancer.2017.11.010 REFERENCES Chen X, Yan YL, Zeng SS, Gong ZC, Xu ZJ. Caveolin-1 Promotes Radioresistance via IRGM-Regulated Autophagy in Lung Cancer. Ann Transl Med (2021) 9(1):47. doi: 10.21037/atm-20-3293 8. Liang C, Zhang X, Yang M, Dong X. Recent Progress in Ferroptosis Inducers for Cancer Therapy. Adv Mater (2019) 31(51):e1904197. doi: 10.1002/ adma.201904197 9. Tarangelo A, Magtanong L, Bieging-Rolett KT, Li Y, Ye J, Attardi LD, et al. P53 Suppresses Metabolic Stress-Induced Ferroptosis in Cancer Cells. Cell Rep (2018) 22(3):569–75. doi: 10.1016/j.celrep.2017.12.077 29. Li J, Yao M, Zhu X, Li Q, He J, Chen L, et al. YAP-Induced Endothelial- Mesenchymal Transition in Oral Submucous Fibrosis. J Dent Res (2019) 98 (8):920–9. doi: 10.1177/0022034519851804 30. Yan Y, Xu Z, Qian L, Zeng S, Zhou Y, Chen X, et al. Identiﬁcation of CAV1 and DCN as Potential Predictive Biomarkers for Lung Adenocarcinoma. Am J Physiol Lung Cell Mol Physiol (2019) 316(4):L630–L43. doi: 10.1152/ ajplung.00364.2018 10. Wang Z, Ding Y, Wang X, Lu S, Wang C, He C, et al. Pseudolaric Acid B Triggers Ferroptosis in Glioma Cells via Activation of Nox4 and Inhibition of xCT. Cancer Lett (2018) 428:21–33. doi: 10.1016/j.canlet.2018.04.021 11. Hu Z, Mi Y, Qian H, Guo N, Yan A, Zhang Y, et al. A Potential Mechanism of Temozolomide Resistance in Glioma-Ferroptosis. Front Oncol (2020) 10:897. doi: 10.3389/fonc.2020.00897 31. Mou Y, Wang J, Wu J, He D, Zhang C, Duan C, et al. Ferroptosis, a New Form of Cell Death: Opportunities and Challenges in Cancer. J Hematol Oncol (2019) 12(1):34. doi: 10.1186/s13045-019-0720-y 12. Xu S, Tang L, Li X, Fan F, Liu Z. Immunotherapy for Glioma: Current Management and Future Application. Cancer Lett (2020) 476:1–12. doi: 10.1016/j.canlet.2020.02.002 32. Ludwig K, Kornblum HI. Molecular Markers in Glioma. J Neurooncol (2017) 134(3):505–12. doi: 10.1007/s11060-017-2379-y 13. Wang W, Green M, Choi JE, Gijon M, Kennedy PD, Johnson JK, et al. CD8(+) T Cells Regulate Tumour Ferroptosis During Cancer Immunotherapy. Nature (2019) 569(7755):270–4. doi: 10.1038/s41586-019-1170-y 33. Yan Y, Xu Z, Chen X, Wang X, Zeng S, Zhao Z, et al. Novel Function of lncRNA ADAMTS9-AS2 in Promoting Temozolomide Resistance in Glioblastoma via Upregulating the FUS/MDM2 Ubiquitination Axis. Front Cell Dev Biol (2019) 7:217. doi: 10.3389/fcell.2019.00217 14. Lang X, Green MD, Wang W, Yu J, Choi JE, Jiang L, et al. Radiotherapy and Immunotherapy Promote Tumoral Lipid Oxidation and Ferroptosis via Synergistic Repression of SLC7A11. Cancer Discov (2019) 9(12):1673–85. doi: 10.1158/2159-8290.CD-19-0338 34. REFERENCES Miotto G, Rossetto M, Di Paolo ML, Orian L, Venerando R, Roveri A, et al. Insight Into the Mechanism of Ferroptosis Inhibition by Ferrostatin-1. Redox Biol (2020) 28:101328. doi: 10.1016/j.redox.2019.101328 35. Tang R, Xu J, Zhang B, Liu J, Liang C, Hua J, et al. Ferroptosis, Necroptosis, and Pyroptosis in Anticancer Immunity. J Hematol Oncol (2020) 13(1):110. doi: 10.1186/s13045-020-00946-7 15. Zhang R, Kim AS, Fox JM, Nair S, Basore K, Klimstra WB, et al. Mxra8 Is a Receptor for Multiple Arthritogenic Alphaviruses. Nature (2018) 557 (7706):570–4. doi: 10.1038/s41586-018-0121-3 36. Chen Y, Mi Y, Zhang X, Ma Q, Song Y, Zhang L, et al. Dihydroartemisinin- Induced Unfolded Protein Response Feedback Attenuates Ferroptosis via PERK/ATF4/HSPA5 Pathway in Glioma Cells. J Exp Clin Cancer Res (2019) 38(1):402. doi: 10.1186/s13046-019-1413-7 16. Zhang D, Qian C, Wei H, Qian X. Identiﬁcation of the Prognostic Value of Tumor Microenvironment-Related Genes in Esophageal Squamous Cell Carcinoma. Front Mol Biosci (2020) 7:599475. doi: 10.3389/fmolb. 2020.599475 17. Friedman J, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models via Coordinate Descent. J Stat Softw (2010) 33(1):1–22. doi: 10.18637/jss.v033.i01 37. Chen L, Li X, Liu L, Yu B, Xue Y, Liu Y. Erastin Sensitizes Glioblastoma Cells to Temozolomide by Restraining xCT and Cystathionine-Gamma-Lyase Function. Oncol Rep (2015) 33(3):1465–74. doi: 10.3892/or.2015.3712 18. Blanche P, Dartigues JF, Jacqmin-Gadda H. Estimating and Comparing Time- Dependent Areas Under Receiver Operating Characteristic Curves for Censored Event Times With Competing Risks. Stat Med (2013) 32 (30):5381–97. doi: 10.1002/sim.5958 38. Santana-Codina N, Mancias JD. The Role of NCOA4-Mediated Ferritinophagy in Health and Disease. Pharmaceuticals (Basel) (2018) 11 (4):114. doi: 10.3390/ph11040114 39. Hou W, Xie Y, Song X, Sun X, Lotze MT, Zeh HJ3rd, et al. Autophagy Promotes Ferroptosis by Degradation of Ferritin. Autophagy (2016) 12 (8):1425–8. doi: 10.1080/15548627.2016.1187366 19. Wang F, Yuan C, Wu HZ, Liu B, Yang YF. Bioinformatics, Molecular Docking and Experiments In Vitro Analyze the Prognostic Value of CXC Chemokines in Breast Cancer. Front Oncol (2021) 11:665080. doi: 10.3389/ fonc.2021.665080 40. Zhang Y, Kong Y, Ma Y, Ni S, Wikerholmen T, Xi K, et al. Loss of COPZ1 Induces NCOA4 Mediated Autophagy and Ferroptosis in Glioblastoma Cell Lines. Oncogene (2021) 40(8):1425–39. doi: 10.1038/s41388-020-01622-3 20. Chandrashekar DS, Bashel B, Balasubramanya SAH, Creighton CJ, Ponce- Rodriguez I, Chakravarthi B, et al. UALCAN: A Portal for Facilitating Tumor Subgroup Gene Expression and Survival Analyses. Neoplasia (2017) 19 (8):649–58. doi: 10.1016/j.neo.2017.05.002 41. REFERENCES 2. Bai J, Varghese J, Jain R. Adult Glioma WHO Classiﬁcation Update, Genomics, and Imaging: What the Radiologists Need to Know. Top Magn Reson Imaging (2020) 29(2):71–82. doi: 10.1097/RMR.0000000000000234 3. Chen X, Yan Y, Zhou J, Huo L, Qian L, Zeng S, et al. Clinical Prognostic Value of Isocitrate Dehydrogenase Mutation, O-6-Methylguanine-DNA 2. Bai J, Varghese J, Jain R. Adult Glioma WHO Classiﬁcation Update, Genomics, and Imaging: What the Radiologists Need to Know. Top Magn Reson Imaging (2020) 29(2):71–82. doi: 10.1097/RMR.0000000000000234 1. Gusyatiner O, Hegi ME. Glioma Epigenetics: From Subclassiﬁcation to Novel Treatment Options. Semin Cancer Biol (2018) 51:50–8. doi: 10.1016/ j.semcancer.2017.11.010 3. Chen X, Yan Y, Zhou J, Huo L, Qian L, Zeng S, et al. Clinical Prognostic Value of Isocitrate Dehydrogenase Mutation, O-6-Methylguanine-DNA February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 14 Xu et al. MXRA8 as a Glioma Indicator Methyltransferase Promoter Methylation, and 1p19q Co-Deletion in Glioma Patients. Ann Transl Med (2019) 7(20):541. doi: 10.21037/atm.2019.09.126 24. Liang JY, Wang DS, Lin HC, Chen XX, Yang H, Zheng Y, et al. A Novel Ferroptosis-Related Gene Signature for Overall Survival Prediction in Patients With Hepatocellular Carcinoma. Int J Biol Sci (2020) 16(13):2430–41. doi: 10.7150/ijbs.45050 Methyltransferase Promoter Methylation, and 1p19q Co-Deletion in Glioma Patients. Ann Transl Med (2019) 7(20):541. doi: 10.21037/atm.2019.09.126 4. Jiapaer S, Furuta T, Tanaka S, Kitabayashi T, Nakada M. Potential Strategies Overcoming the Temozolomide Resistance for Glioblastoma. Neurol Med Chir (Tokyo) (2018) 58(10):405–21. doi: 10.2176/nmc.ra.2018-0141 25. Ru B, Wong CN, Tong Y, Zhong JY, Zhong SSW, Wu WC, et al. TISIDB: An Integrated Repository Portal for Tumor-Immune System Interactions. Bioinformatics (2019) 35(20):4200–2. doi: 10.1093/bioinformatics/btz210 5. Stockwell BR, Jiang X, Gu W. Emerging Mechanisms and Disease Relevance of Ferroptosis. Trends Cell Biol (2020) 30(6):478–90. doi: 10.1016/ j.tcb.2020.02.009 26. Tang Z, Li C, Kang B, Gao G, Li C, Zhang Z. GEPIA: A Web Server for Cancer and Normal Gene Expression Proﬁling and Interactive Analyses. Nucleic Acids Res (2017) 45(W1):W98–W102. doi: 10.1093/nar/gkx247 6. Hirschhorn T, Stockwell BR. The Development of the Concept of Ferroptosis. Free Radic Biol Med (2019) 133:130–43. doi: 10.1016/j.freeradbiomed. 2018.09.043 27. Dai S, Yan Y, Xu Z, Zeng S, Qian L, Huo L, et al. SCD1 Confers Temozolomide Resistance to Human Glioma Cells via the Akt/GSK3beta/beta-Catenin Signaling Axis. Front Pharmacol (2017) 8:960. doi: 10.3389/fphar.2017.00960 7. Zheng J, Conrad M. The Metabolic Underpinnings of Ferroptosis. Cell Metab (2020) 32(6):920–37. doi: 10.1016/j.cmet.2020.10.011 28. REFERENCES Litak J, Mazurek M, Grochowski C, Kamieniak P, Rolinski J. PD-L1/PD-1 Axis in Glioblastoma Multiforme. Int J Mol Sci (2019) 20(21):5347. doi: 10.3390/ijms20215347 21. Ma W, Zhang K, Bao Z, Jiang T, Zhang Y. SAMD9 Is Relating With M2 Macrophage and Remarkable Malignancy Characters in Low-Grade Glioma. Front Immunol (2021) 12:659659. doi: 10.3389/ﬁmmu.2021.659659 42. Qi L, Yu H, Zhang Y, Zhao D, Lv P, Zhong Y, et al. IL-10 Secreted by M2 Macrophage Promoted Tumorigenesis Through Interaction With JAK2 in Glioma. Oncotarget (2016) 7(44):71673–85. doi: 10.18632/oncotarget.12317 43. Cui X, Morales RT, Qian W, Wang H, Gagner JP, Dolgalev I, et al. Hacking Macrophage-Associated Immunosuppression for Regulating Glioblastoma Angiogenesis. Biomaterials (2018) 161:164–78. doi: 10.1016/j.biomaterials. 2018.01.053 22. Zhang Z, Qiu X, Yan Y, Liang Q, Cai Y, Peng B, et al. Evaluation of Ferroptosis-Related Gene AKR1C1 as a Novel Biomarker Associated With the Immune Microenvironment and Prognosis in Breast Cancer. Int J Gen Med (2021) 14:6189–200. doi: 10.2147/IJGM.S329031 44. Lin YJ, Wei KC, Chen PY, Lim M, Hwang TL. Roles of Neutrophils in Glioma and Brain Metastases. Front Immunol (2021) 12:701383. doi: 10.3389/ ﬁmmu.2021.701383 23. Vasaikar SV, Straub P, Wang J, Zhang B. LinkedOmics: Analyzing Multi- Omics Data Within and Across 32 Cancer Types. Nucleic Acids Res (2018) 46 (D1):D956–D63. doi: 10.1093/nar/gkx1090 February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 15 Xu et al. MXRA8 as a Glioma Indicator Glioma. Cancers (Basel) (2021) 13(10):2400. doi: 10.3390/cancers 13102400 45. Massara M, Persico P, Bonavita O, Mollica Poeta V, Locati M, Simonelli M, et al. Neutrophils in Gliomas. Front Immunol (2017) 8:1349. doi: 10.3389/ ﬁmmu.2017.01349 54. Yan Y, Cai J, Huang Z, Cao X, Tang P, Wang Z, et al. A Novel Ferroptosis- Related Prognostic Signature Reveals Macrophage Inﬁltration and EMT Status in Bladder Cancer. Front Cell Dev Biol (2021) 9:712230. doi: 10.3389/ fcell.2021.712230 46. Qian J, Wang C, Wang B, Yang J, Wang Y, Luo F, et al. The IFN-Gamma/PD- L1 Axis Between T Cells and Tumor Microenvironment: Hints for Glioma Anti-PD-1/PD-L1 Therapy. J Neuroinﬂamm (2018) 15(1):290. doi: 10.1186/ s12974-018-1330-2 55. Qiu CJ, Wang XB, Zheng ZR, Yang CZ, Lin K, Zhang K, et al. Development and Validation of a Ferroptosis-Related Prognostic Model in Pancreatic Cancer. Invest New Drugs (2021) 39(6):1507–22. doi: 10.1007/s10637-021- 01114-5 47. Hung AL, Maxwell R, Theodros D, Belcaid Z, Mathios D, Luksik AS, et al. February 2022 \| Volume 13 \| Article 834595 REFERENCES TIGIT and PD-1 Dual Checkpoint Blockade Enhances Antitumor Immunity and Survival in GBM. Oncoimmunology (2018) 7(8):e1466769. doi: 10.1080/ 2162402X.2018.1466769 Conﬂict of Interest: 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. 48. Liu F, Huang J, Liu X, Cheng Q, Luo C, Liu Z. CTLA-4 Correlates With Immune and Clinical Characteristics of Glioma. Cancer Cell Int (2020) 20:7. doi: 10.1186/s12935-019-1085-6 49. Tumeh PC, Harview CL, Yearley JH, Shintaku IP, Taylor EJ, Robert L, et al. PD-1 Blockade Induces Responses by Inhibiting Adaptive Immune Resistance. Nature (2014) 515(7528):568–71. doi: 10.1038/nature13954 Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. 50. von Roemeling CA, Wang Y, Qie Y, Yuan H, Zhao H, Liu X, et al. Therapeutic Modulation of Phagocytosis in Glioblastoma can Activate Both Innate and Adaptive Antitumour Immunity. Nat Commun (2020) 11(1):1508. doi: 10.1038/s41467-020-15129-8 Copyright © 2022 Xu, Chen, Song, Yuan and Yan. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 51. Yunna C, Mengru H, Lei W, Weidong C. Macrophage M1/M2 Polarization. Eur J Pharmacol (2020) 877:173090. doi: 10.1016/j.ejphar.2020.173090 52. Cannarile MA, Weisser M, Jacob W, Jegg AM, Ries CH, Ruttinger D. Colony- Stimulating Factor 1 Receptor (CSF1R) Inhibitors in Cancer Therapy. J Immunother Cancer (2017) 5(1):53. doi: 10.1186/s40425-017-0257-y 53. Przystal JM, Becker H, Canjuga D, Tsiami F, Anderle N, Keller AL, et al. Targeting CSF1R Alone or in Combination With PD1 in Experimental February 2022 \| Volume 13 \| Article 834595 Frontiers in Immunology \| www.frontiersin.org 16
W4229561155.txt	https://www.mdpi.com/2411-9660/5/1/21/pdf?version=1616386057	en		Retraction: Amathes, P.; et al. Interpreting Locked Photographic Data: The Case of Apollo 17 Photo GPN-2000-00113. Designs 2021, 5, 8	Designs	2,021	cc-by	239	Retraction Retraction: Amathes, P.; et al. Interpreting Locked Photographic Data: The Case of Apollo 17 Photo GPN-2000-00113. Designs 2021, 5, 8 Designs Editorial Office MDPI, St. Alban-Anlage 66, 4052 Basel, Switzerland; designs@mdpi.com The journal Designs retracts the 8th February, 2021 article “Interpreting Locked Photographic Data: The Case of Apollo 17 Photo GPN-2000-00113” [1]. Following publication, serious concerns were brought to the attention of the publisher. The article was re-examined, revealing that the complaints were valid and that the article does not meet the standards of editorial and scientific soundness for Designs. The article [1] is therefore retracted. This retraction was approved by the Editor in Chief of the journal, the authors did not agree to this retraction. Reference 1. Amathes, P.; Christodoulides, P. Interpreting Locked Photographic Data: The Case of Apollo 17 Photo GPN-2000-00113. Designs 2021, 5, 8. [CrossRef] Citation: Designs Editorial Office. Retraction: Amathes, P.; et al. Interpreting Locked Photographic Data: The Case of Apollo 17 Photo GPN-2000-00113. Designs 2021, 5, 8. Designs 2021, 5, 21. https://doi.org/ 10.3390/designs5010021 Received: 15 March 2021 Accepted: 17 March 2021 Published: 22 March 2021 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Copyright: © 2021 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). Designs 2021, 5, 21. https://doi.org/10.3390/designs5010021 https://www.mdpi.com/journal/designs
https://openalex.org/W3035815283	https://bib-pubdb1.desy.de/record/450147/files/PhysRevLett.125.131804.pdf	English	null	Solar Axions Cannot Explain the XENON1T Excess	Physical review letters	2,020	cc-by	4,941	DOI: 10.1103/PhysRevLett.125.131804 and because the location of the peak around 2–3 keV corresponds roughly to the maximum of the axion energy spectrum for the ABC processes, the Primakoff and 57Fe components are both allowed to be absent as long as there is a nonzero ABC component. This selects gae as the crucial coupling to attempt to explain the data in terms of the QCD axion [5–9]. Taken at face value, the strength of the XENON1T excess requires gae ≳10−12, corresponding to an axion decay constant fa ≲108 GeV, and in turn to an axion mass ma ≳0.06 eV. However, astrophysical con- siderations indicate that such a large value of gae is not tenable, as stellar evolution would be drastically affected by the exceedingly large energy losses via axion emission. The strategy that we will follow consists of assuming that gae and gaγ lie in the 90% C.L. regions resulting from the XENON1T fit [1]. We will then estimate the effects of extra energy losses on a set of astrophysical observables related to red giants branch (RGB) and horizontal branch (HB) stars and to white dwarfs (WDs). Introduction.—The XENON1T collaboration [1] has reported an excess in low-energy electronic recoil data below 7 keV and peaking around 2–3 keV. The collabora- tion cautions that the excess could be due to an un- accounted background from β decays due to a trace amount of tritium, but they also explore the possibility that the signal is due to different types of new physics. The most intriguing interpretation, which also provides the best fit to the data, is given in terms of solar axions, favored over the background-only hypothesis at the 3.5σ level. Three production mechanisms contribute to the solar axion flux: (i) atomic recombination and deexcitation, bremsstrahlung, and Compton (ABC) interactions [2] that are controlled by the axion-electron coupling gae, (ii) Primakoff conversion of photons into axions [3] induced by the axion-photon coupling gaγ, and (iii) axion emission in the M1 nuclear transition of 57Fe [4] that produces monoenergetic 14.4 keV axions and is controlled by and effective axion-nucleon coupling geff an. Since this last process cannot play any role in producing events below 10 keV, we will not include in our analysis astrophysical observables sensitive to geff an. Conversely, axions produced through (i) and (ii) feature a wide spectrum peaking around a few keV. Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article’s title, journal citation, and DOI. Funded by SCOAP3. PHYSICAL REVIEW LETTERS 125, 131804 (2020) PHYSICAL REVIEW LETTERS 125, 131804 (2020) Editors' Suggestion Solar Axions Cannot Explain the XENON1T Excess Luca Di Luzio ,1,* Marco Fedele ,2,† Maurizio Giannotti ,3,‡ Federico Mescia ,2,§ and Enrico Nardi 4,∥ 1Deutsches Elektronen-Synchrotron DESY, Notkestraße 85, D-22607 Hamburg, Germany 2Department de Física Qu`antica i Astrofísica, Institut de Ci`encies del Cosmos (ICCUB), Universitat de Barcelona, Martí i Franqu`es 1, E-08028 Barcelona, Spain 3Physical Sciences, Barry University, 11300 NE 2nd Avenue, Miami Shores, Florida 33161, USA 4INFN, Laboratori Nazionali di Frascati, C.P. 13, 100044 Frascati, Italy (Received 3 July 2020; revised 23 July 2020; accepted 30 July 2020; published 24 September 2020) d 3 July 2020; revised 23 July 2020; accepted 30 July 2020; published 24 September 2020) We argue that the interpretation in terms of solar axions of the recent XENON1T excess is not tenable when confronted with astrophysical observations of stellar evolution. We discuss the reasons why the emission of a flux of solar axions sufficiently intense to explain the anomalous data would radically alter the distribution of certain type of stars in the color-magnitude diagram in the first place and would also clash with a certain number of other astrophysical observables. Quantitatively, the significance of the discrepancy ranges from 3.3σ for the rate of period change of pulsating white dwarfs and exceeds 19σ for the R parameter and for MI;TRGB. DOI: 10.1103/PhysRevLett.125.131804 (2), this expression is derived from an analytic fit to evolutionary track points calculated with stellar evolutionary codes modified to account for axion emission [19,28]; thus, it is quantitatively reliable up to values of gae not much larger than those corresponding to the last point fitted (for definiteness ge13 ∼9). Thus, we will not input into these expressions the much larger XENON1T values ge13 ∼30. Rather, very conservatively, we will limit ourselves to estimate the tension between the observed values of MI;TRGB and R, and the values resulting from Eqs. (2) and (3) when evaluated at ge13 ∼9 (gaγ ≈0). As regards values of gaγ too large to be used in Eq. (3), they can be directly constrained from the lifetime of HB stars which, in the presence of extra cooling, scales as ∼L0=ðL0 þ LaÞ with L0 (La) the standard (axion) core luminosity [15]. Hence, for gγ10 ≳1, rather than Eq. (3), we will use where δMc is the change in the He-core mass, and Y ≃ 0.255 0.002 is the primordial He abundance. The relative errors on δMc, which represents the main theoretical uncertainty from astrophysics. and the one on Y, are of the same order. Hence, due to the larger coefficient multiplying Y, the uncertainty from δMc can be neglected. Similar to Eq. (2), this expression is derived from an analytic fit to evolutionary track points calculated with stellar evolutionary codes modified to account for axion emission [19,28]; thus, it is quantitatively reliable up to values of gae not much larger than those corresponding to the last point fitted (for definiteness ge13 ∼9). Thus, we will not input into these expressions the much larger XENON1T values ge13 ∼30. Rather, very conservatively, we will limit ourselves to estimate the tension between the observed values of MI;TRGB and R, and the values resulting from Eqs. (2) and (3) when evaluated at ge13 ∼9 (gaγ ≈0). As regards values of gaγ too large to be used in Eq. (3), they can be directly constrained from the lifetime of HB stars which, in the presence of extra cooling, scales as ∼L0=ðL0 þ LaÞ with L0 (La) the standard (axion) core luminosity [15]. Hence, for gγ10 ≳1, rather than Eq. DOI: 10.1103/PhysRevLett.125.131804 The production rates are independent of the axion mass for ma ≲100 eV. As regards detection, electron recoils occur via the axio- electric effect, which is controlled by gae. Because of this, Astrophysical observables and axion couplings.—The axion interactions with photons and electrons read Lint ¼ 1 4 gaγaFμν ˜Fμν þ gae ∂μa 2me ¯eγμγ5e; ð1Þ ð1Þ where the couplings can be related to model-dependent dimensionless coefficients as gaγ ¼ ðα=2πÞðCaγ=faÞ and gae ¼ Caeðme=faÞ. In benchmark axion models Caγ and Cae are typically of Oð1Þ, although strong enhancements and/or suppressions are possible in specific cases [10–14]. In the following, we will adopt the notation gγ10 ≡gaγ × ð1010 GeVÞ and ge13 ≡gae × 1013. Axions with couplings as large as ge13 ∼10, gγ10 ∼1 would be abundantly 131804-1 Published by the American Physical Society 0031-9007=20=125(13)=131804(6) PHYSICAL REVIEW LETTERS 125, 131804 (2020) processes both directly and indirectly. If gae is sufficiently large, axion emission would directly produce extra cooling of the He core. The star self-regulates by slightly con- tracting and the core temperature increases, speeding up the He burning rate. Once the He fuel is exhausted, the star turns into a WD. The indirect effect is related to the growth of the degenerate He core during the RGB phase previously discussed. HB stars would unavoidably inherit a more massive core from the parent RGs, resulting in an increased He burning rate to contrast the larger gravitational pull and shortening further the duration of the HB phase. Note that the indirect effect of gae is so important that, for ge13 ∼15, it would suffice to depopulate almost completely the HB in the color-magnitude diagram (CMD) (R ≈0). Cooling of HB stars can also proceed via the Primakoff effect ∝g2aγ, which is particularly efficient at the typical temperatures and densities of HB cores (T ∼108 K, ρ ¼ 104 g cm−3). For sufficiently large values of gaγ, R can still decrease well below the observed values even when gae ≈0. Hence, an accurate determination of this observable allows us to probe the axion coupling to both photons and electrons. In terms of gae and gaγ, the R parameter can be written as [19,28] produced in several types of stars without being trapped, and thus would efficiently drain energy from the star cores. produced in several types of stars without being trapped, and thus would efficiently drain energy from the star cores. DOI: 10.1103/PhysRevLett.125.131804 Astrophysical considerations have been systematically used to place severe bounds on light, weakly interacting particles, such as neutrinos and axions [15]. Noticeably, a set of anomalous observations have recently led to speculations that new physics is at play [14,16,17], and the axion case appears especially compelling [18,19]. The most effective observables to constrain gae and gaγ are described below. Tip of RGB stars in globular cluster: We denote by MI;TRGB the luminosity of the tip of the RGB in globular clusters. RG stars are characterized by a He core and a burning H shell. During the RGB evolution, the ashes from the burning shell increase the He-core mass, while the star luminosity (determined by equilibrium at the surface of the He core between thermal pressure supporting the non- degenerate envelope against the gravity pull from the core) keeps growing. The process continues until the core reaches sufficiently large temperatures and densities (T ∼108 K, ρ ¼ 106 g cm−3) to ignite He, an event known as the He flash. At this stage, the star has reached its maximum luminosity MI;TRGB, after which it shrinks and moves to the HB. If an additional core-cooling mechanism were at play, He ignition would be delayed, the core would accrete a larger mass, and the star would reach higher luminosities. Therefore, measurements of MI;TRGB allow us to test the rate of cooling during the RGB phase. The method is particularly effective for constraining gae since, in RG cores, axions can be efficiently produced via electron bremsstrahlung. The most recent analyses [20–22] have derived comparable constraints. Here we adopt the result of the analysis of the Large Magellanic Cloud in Refs. [23,24], which provides the most conservative bound MI;TRGB ¼ −4.047 0.045 mag. In terms of gae, this observable can be written as [20,25] Rtheo ¼ 7.33Y −0.095 ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 21.86 þ 21.08gγ10 q þ 0.02 −1.61δMc −0.005g2 e13; δMc ¼ 0.024ð ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ g2 e13 þ 1.232 q −1.23 −0.138g1.5 e13Þ; ð3Þ ð3Þ where δMc is the change in the He-core mass, and Y ≃ 0.255 0.002 is the primordial He abundance. The relative errors on δMc, which represents the main theoretical uncertainty from astrophysics. and the one on Y, are of the same order. Hence, due to the larger coefficient multiplying Y, the uncertainty from δMc can be neglected. Similar to Eq. DOI: 10.1103/PhysRevLett.125.131804 To a very good approximation _Π=Π is proportional to the cooling rate _T=T; hence. a measurement of _Π allows us to constrain possible sources of extra cooling (see Ref. [31] for a review). Here we consider four WDVs: G117-B15A [32], R548 [33], L19-2 [34] (for two pulsation modes), and PG1351 þ 489 [35]. We list in Table I the corresponding measured values of _Π=Π. Theoretically, the rate of change in the WD pulsating period as a function of ge13 can be parametrized as [18] _Πtheo WDi ¼ ai þ big2 e13, where ai and bi are parameters specific for each WD. with a ¼ ð6.26Y −0.12Þ and b ¼ 0.41 [27]. Note that Eq. (4) neglects both direct and indirect effects of gae on HB and RGB stars, and hence it would also yield conservative limits. White dwarf luminosity function: The third observable we consider is the distribution of WDs as a function of their luminosity (WDLF). The WDLF measures the WD cooling efficiency, and thus allows us to place strong bounds on new exotic cooling processes, including axion emission (see Ref. [29] for a review). WDs are compact objects whose hydrostatic equilibrium is supported by electron degeneracy pressure; hence, axion emission from WDs would dominantly depend on gae. Here we will use the bound gWDLF e13 ≤2.8 obtained in Ref. [30]. i i p p XENON1T vs astrophysics.—Figure 1 shows contours of the axion energy-loss rates per unit mass in a temperature vs density plane, for a pure He plasma. Contour isolines for energy loss due to Compton (dashed blue) and bremsstrahlung (solid red) processes, which are controlled by gae, are also shown. For reference, we have fixed ge13 ¼ 4.3, which corresponds to the RGB bound from M5 [20]. Energy-loss rates for different values of gae can be easily obtained, recalling that they scale as g2ae. The labeled disks in the figure show the position of the RGB tip and of a typical HB star (of mass 0.8 M⊙) and a range of WDs with luminosities varying from 5 × 10−4 to 5 × 10−1L⊙(dashed gray rectangle). The blue disk indicates the temperature and density of a typical WD variable [31]. The location of the Sun is marked with a yellow disk on top of the broken gray line, which locates main sequence (MS) stars of different masses. DOI: 10.1103/PhysRevLett.125.131804 (3), we will use Mtheo I;TRGB ¼ −4.08 −0.25ð ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ g2 e13 þ 0.93 q −0.96 −0.17g1.5 e13Þ; ð2Þ which results from an analytic fit to ten evolutionary track points reaching close to the RGB tip obtained from numerical simulations [26] and corresponding to values of ge13 up to 9 [20]. The associated theoretical uncertainty is σ2 ¼ 0.0392 þ ð0.046 þ 0.012ge13Þ2 [20]. R parameter: After He ignition, the RG core expands and its density decreases by 2 orders of magnitude. The star migrates to the HB and remains supported by He burning in a nondegenerate core. The ratio R ¼ NHB=NRGB between the number of stars in globular clusters in the HB and in the upper portion of the RGB directly measures the duration of He burning in the HB phase. The value R ¼ 1.39 0.03 was obtained in Ref. [27] from the analysis of 39 clusters. The duration of the HB phase can be affected by gae-related Rtheo ≈ a2 a þ bg2 γ10 ; ð4Þ Rtheo ≈ a2 a þ bg2 γ10 ; ð4Þ 131804-2 131804-2 PHYSICAL REVIEW LETTERS 125, 131804 (2020) TABLE I. Measured values of astrophysic observables and expected ranges, for gae, gaγ falling within the 1σ region of the XENON1T fit (¯ge13 ∈½28; 35). _ΠWDi are in units of ½10−15 s=s. For R and MI;TRGB the expected regions and tensions correspond to ge13 ¼ ¯ge13ðgaγ ¼ 0Þ ≥9 (see text). TABLE I. Measured values of astrophysic observables and expected ranges, for gae, gaγ falling within the 1σ region of the XENON1T fit (¯ge13 ∈½28; 35). _ΠWDi are in units of ½10−15 s=s. For R and MI;TRGB the expected regions and tensions correspond to ge13 ¼ ¯ge13ðgaγ ¼ 0Þ ≥9 (see text). Observable Measured Expected Tension R parameter 1.39 0.03 ≤0.83ðge13 ¼ 9Þ 19σ⋆ MLMC I;TRGBðmagÞ −4.047 0.045 ≤−4.92ðge13 ¼ 9Þ 19σ⋆ gWDLF e13 ≤2.8ð3σÞ 29.7 4.8 5.6σ _Πð113Þ L19−2 3.0 0.6 57 16 3.4σ _Πð192Þ L19−2 3.0 0.6 95 27 3.4σ _ΠPG1351þ489 200 90 19620 5730 3.4σ _ΠG117−B15A 4.2 0.7 113 33 3.3σ _ΠR548 3.3 1.1 87 25 3.3σ with a period Π ranging from a few to several minutes. Because the oscillation period depends on the luminosity, a secular change of the period _Π tracks the rate of decrease of the star luminosity. DOI: 10.1103/PhysRevLett.125.131804 Note that, since MS stars, including the Sun, are supported by H burning cores, their position with respect to the energy-loss isolines for the He plasma is approximate and slightly shifted toward larger rates. The picture shows clearly that the Sun is a relatively faint axion emitter with respect to other stellar objects, so that values ge13 ≳10 as required to account for the XENON1T excess would unavoidably turn other stars into bright “axion light- houses.” The RGB would extend to higher luminosities e13 Rate of period change of WD variables: WD variables (WDVs) are WDs whose luminosity varies periodically, FIG. 1. Contours of the axion energy-loss rates per unit mass, in erg g−1 s−1, for a pure He plasma and ge13 ¼ 4.3. FIG. 1. Contours of the axion energy-loss rates per unit mass, in erg g−1 s−1, for a pure He plasma and ge13 ¼ 4.3. 131804-3 131804-3 PHYSICAL REVIEW LETTERS 125, 131804 (2020) inferred from helioseismology observations [40]. This is about a factor of 2 stronger than the bound labeled “solar ν” in the upper panel of Fig. 8 in Ref. [1], which is taken from Ref. [41] (see [42]). To assess quantitatively the dis- crepancy between the values of gae and gaγ needed to reproduce the XENON1T excess, we proceed as follows: we first extract the allowed ranges from the 90% C.L. region of Ref. [1] not excluded by solar data (the blue area in Fig. 2). This region can be parametrized by means of an effective coupling [14] than the ones observed, and the decreased duration of the He burning phase would depopulate the HB, to the point that, for smaller clusters with relatively few stars, already for ge13 ∼15 we would expect R ≈0. In short, regardless of other details, a value ge13 ∼30 would definitely destroy the agreement between stellar evolution models and the observed CMD. Quantifying the tension.—The projections of the XENON1T 90% C.L. best fit region onto the ðgae; gaγÞ, ðgae; geff anÞ, and ðgaegaγ; gaegeff anÞ planes are given in Fig. 8 of Ref. [1]. Since only gae and gaγ can be responsible for the anomalous XENON1T data below 7 keV, we focus on the best fit region for these two couplings, which corresponds the blue band in Fig. 2. DOI: 10.1103/PhysRevLett.125.131804 In the figure we also show the 2σ limits on gae, gaγ obtained from each single astrophysical observable, as well as the result of a global fit to the entire set of stellar cooling data. The curve depicting the CAST [36] limit in the ðgae; gaγÞ plane in Ref. [1] was taken from Ref. [37]. We update this bound with the most recent CAST results [38] which, in the gae ≃0 limit, and for ma ≲20 meV (ma ≲0.7 eV), corres- pond to gaγ < 0.66ð2.0Þ × 10−10 GeV−1. These limits are represented in Fig. 2 by the two green lines, in which we have folded in the effects of a nonzero gae that would increase the production of solar axions and strengthen the bounds. The vertical dashed line is the LUX limit [39]. The gray horizontal line at gγ10 ¼ 4.1 corresponds to the limit from a global fit to solar data, which includes the measured flux of 8B and 7Be neutrinos as well as additional data ¯g4 e13 ¼ g2 e13ðg2 e13 þ 200g2 γ10Þ: ð5Þ ð5Þ The 90% C.L. (68% C.L.) region of XENON1T is then well represented by the range ¯ge13 ∈½26; 37 (¯ge13 ∈½28; 35). Varying gae and gaγ with the constraint that ¯ge13 remains within this range, we estimate the range of values for the astrophysical observables implied by the XENON1T data, and we confront them with the measured values. Our results are collected in Table I. For each observable, the tension given in the fourth column is evaluated by dividing the difference between the value implied by the XENON1T data and the astrophysical determination by the total uncertainty. Given that the statistical distributions are at best only approximately known, these results are only indicative and have no rigorous Gaussian meaning. It is apparent that the large gae required to fit the XENON1T excess are in strong conflict with all the astrophysical observables. The discrepancy is at the level of ∼3.4σ for the WDVs cooling rates (last five rows in the Table) and reaches ∼6σ for the WDLF in the third row. As regards the first two rows, the expected values of Rtheo and of Mtheo I;TRGB reported in the table are obtained, respectively, from Eqs. (3) and (2) by setting ge13 ¼ 9, rather than by inserting the much larger values ge13 ∼30 needed to account for the XENON1T data. PHYSICAL REVIEW LETTERS 125, 131804 (2020) [15] G. Raffelt, Stars as Laboratories for Fundamental Physics: The Astrophysics of Neutrinos, Axions, and Other Weakly Interacting Particles (University of Chicago Press, Chicago, 1996). If it will be eventually found that the tritium background or other systematic effects [45,46] are not responsible for the excess, other mechanisms involving either absorption or scattering of new particles of nonsolar origin off target electrons [47–51], although less compelling than the QCD axion, might still provide viable explanations for the XENON1T data. [16] S. Hoof, F. Kahlhoefer, P. Scott, C. Weniger, and M. White, J. High Energy Phys. 03 (2019) 191; 11 (2019) 099(E). [17] P. Di Vecchia, M. Giannotti, M. Lattanzi, and A. Lindner, Proc. Sci., Confinement2018 (2019) 034. We thank Axel Lindner for useful comments. L. D. L. is supported by the Marie Skłodowska-Curie Individual Fellowship Grant AXIONRUSH (GA 840791). M. F. and F. M. are supported by MINECO Grant No. FPA2016- 76005-C2-1-P, Maria de Maetzu program Grants No. MDM-2014-0367 of ICCUB and No. 2017 SGR 929. E. N. is supported by the INFN Iniziativa Specifica, Theoretical Astroparticle Physics (TAsP). We thank Axel Lindner for useful comments. L. D. L. is supported by the Marie Skłodowska-Curie Individual Fellowship Grant AXIONRUSH (GA 840791). M. F. and F. M. are supported by MINECO Grant No. FPA2016- 76005-C2-1-P, Maria de Maetzu program Grants No. MDM-2014-0367 of ICCUB and No. 2017 SGR 929. E. N. is supported by the INFN Iniziativa Specifica, Theoretical Astroparticle Physics (TAsP). [18] M. Giannotti, I. G. Irastorza, J. Redondo, A. Ringwald, and K. Saikawa, J. Cosmol. Astropart. Phys. 10 (2017) 010. [19] M. Giannotti, I. Irastorza, J. Redondo, and A. Ringwald, J. Cosmol. Astropart. Phys. 05 (2016) 057. [20] N. Viaux, M. Catelan, P. B. Stetson, G. G. Raffelt, J. Redondo, A. A. R. Valcarce, and A. Weiss, Phys. Rev. Lett. 111, 231301 (2013). [21] O. Straniero, I. Dominguez, M. Giannotti, and A. Mirizzi, in 13th Patras Workshop on Axions, WIMPs and WISPs (2018), pp. 172–176 [arXiv:1802.10357]. Note added.—Recently, Refs. [52,53] appeared claiming that, besides the axioelectric effect, also the inverse Primakoff process can contribute to the detection of solar axions by XENON1T. This would relax the best fit region toward lower values of gae at the cost of increasing gaγ. This can relax the tension with astrophysical bounds; however, using the results of Ref. PHYSICAL REVIEW LETTERS 125, 131804 (2020) [52,53] we have verified that the discrepancy with the R parameter remains at least at the level of 8σ. [22] S. A. Díaz, K.-P. Schröder, K. Zuber, D. Jack, and E. E. B. Barrios, arXiv:1910.10568. , [23] W. L. Freedman et al., Astrophys. J. 882, 34 (2019). [24] W. L. Freedman, B. F. Madore, T. Hoyt, I. S. Jang, R. Beaton, M. G. Lee, A. Monson, J. Neeley, and J. Rich, Astrophys. J. 891, 57 (2020). [25] F. Capozzi and G. Raffelt, arXiv:2007.03694. [26] N. Viaux, M. Catelan, P. B. Stetson, G. Raffelt, J. Redondo, A. A. R. Valcarce, and A. Weiss, Astron. Astrophys. 558, A12 (2013). [27] A. Ayala, I. Dominguez, M. Giannotti, A. Mirizzi, and O. Straniero, Phys. Rev. Lett. 113, 191302 (2014). [28] O. Straniero, A. Ayala, M. Giannotti, A. Mirizzi, and I. Dominguez, in 11th Patras Workshop on Axions, WIMPs and WISPs (2015), pp. 77–81, https://dx.doi.org/10.3204/ DESY-PROC-2015-02/straniero_oscar. [29] J. Isern, in IAU Symposium 357: White Dwarfs as Probes of Fundamental Physics and Tracers of Planetary, Stellar & Galactic Evolution (2020) [arXiv:2002.08069]. [1] E. Aprile et al. (XENON Collaboration), arXiv:2006.09721. [2] J. Redondo, J. Cosmol. Astropart. Phys. 12 (2013) 008. [30] M. M. M. Bertolami, B. E. Melendez, L. G. Althaus, and J. Isern, J. Cosmol. Astropart. Phys. 10 (2014) 069. [3] H. Primakoff, Phys. Rev. 81, 899 (1951). [3] H. Primakoff, Phys. Rev. 81, 899 (1951). [4] S. Moriyama, Phys. Rev. Lett. 75, 3222 (1995). [31] A. H. Córsico, L. G. Althaus, M. M. M. Bertolami, and S. Kepler, Astron. Astrophys. Rev. 27, 7 (2019). [5] R. D. Peccei and H. R. Quinn, Phys. Rev. Lett. 38, 1440 (1977). [32] A. H. Corsico, L. G. Althaus, M. M. Bertolami, A. D. Romero, E. Garcia-Berro, J. Isern, and S. Kepler, Mon. Not. R. Astron. Soc. 424, 2792 (2012). [6] R. D. Peccei and H. R. Quinn, Phys. Rev. D 16, 1791 (1977). [7] S. Weinberg, Phys. Rev. Lett. 40, 223 (1978). [33] A. Corsico, L. Althaus, A. Romero, A. Mukadam, E. Garcia-Berro, J. Isern, S. Kepler, and M. Corti, J. Cosmol. Astropart. Phys. 12 (2012) 010. [8] F. Wilczek, Phys. Rev. Lett. 40, 279 (1978). [9] Our results apply also to explanations based on generic axionlike particles, for which there is no theoretical relation between ma and the coupling strengths, and that are unrelated to the strong CP problem. [34] A. H. Córsico, A. D. Romero, L. DOI: 10.1103/PhysRevLett.125.131804 This is a precautionary procedure that we have adopted to avoid extrapolating Eqs. (2) and (3) to values of gae, for which the quantitative accuracy of these parametrizations cannot be easily assessed. We have then marked with a ⋆the corresponding tensions. We expect that values of the observables in agreement with the XENON1T solar axion fit would result in much larger tensions. For example, already for ge13 ≈15 Eq. (3) would yield R ≈0, corresponding to a complete depopulation of the HB, and 46σ away from observations. FIG. 2. XENON1T 90% C.L. fit (blue region). 3σ exclusion limit from solar data (gray hatched region). 2σ LUX limit (gray dashed line) and CAST limits for ma < 20 meV and ma < 0.7 eV (green lines). Individual 2σ limits from R para- meter, TRGB, WDLF, WDVs (gray lines), and 2σ global bound from astrophysics (red region). Conclusions.—In this Letter, we have explained why astrophysical observations firmly exclude that solar axions could account for the XENON1T excess. Other explana- tions based on solar production of new light particles or on modifications of neutrino properties (such as a neutrino magnetic moment) are also prone to severe astrophysical constraints, and as long as the corresponding new physics processes would also occur in RG, HB, and WD stellar cores, they can likewise be excluded [43]. FIG. 2. XENON1T 90% C.L. fit (blue region). 3σ exclusion limit from solar data (gray hatched region). 2σ LUX limit (gray dashed line) and CAST limits for ma < 20 meV and ma < 0.7 eV (green lines). Individual 2σ limits from R para- meter, TRGB, WDLF, WDVs (gray lines), and 2σ global bound from astrophysics (red region). 131804-4 PHYSICAL REVIEW LETTERS 125, 131804 (2020) PHYSICAL REVIEW LETTERS 125, 131804 (2020) G. Althaus, E. García- Berro, J. Isern, S. Kepler, M. M. M. Bertolami, D. J. Sullivan, and P. Chote, J. Cosmol. Astropart. Phys. 07 (2016) 036. [10] L. Di Luzio, F. Mescia, and E. Nardi, Phys. Rev. Lett. 118, 031801 (2017). [35] T. Battich, A. H. Córsico, L. G. Althaus, M. M. M. Bertolami, and M. Bertolami, J. Cosmol. Astropart. Phys. 08 (2016) 062. [11] L. Di Luzio, F. Mescia, and E. Nardi, Phys. Rev. D 96, 075003 (2017). [12] L. Di Luzio, F. Mescia, E. Nardi, P. Panci, and R. Ziegler, Phys. Rev. Lett. 120, 261803 (2018). [36] S. Andriamonje et al. (CAST Collaboration), J. Cosmol. Astropart. Phys. 12 (2009) 002. [13] F. Bjorkeroth, L. Di Luzio, F. Mescia, E. Nardi, P. Panci, and R. Ziegler, Phys. Rev. D 101, 035027 (2020). [37] K. Barth et al., J. Cosmol. Astropart. Phys. 05 (2013) 010. [14] L. Di Luzio, M. Giannotti, E. Nardi, and L. Visinelli, arXiv:2003.01100. [38] V. Anastassopoulos et al. (CAST Collaboration), Nat. Phys. 13, 584 (2017). 131804-5 131804-5 PHYSICAL REVIEW LETTERS 125, 131804 (2020) PHYSICAL REVIEW LETTERS 125, 131804 (2020) [39] D. Akerib et al. (LUX Collaboration), Phys. Rev. Lett. 118, 261301 (2017). [45] C. Dessert, J. W. Foster, Y. Kahn, and B. R. Safdi, arXiv:2006.16220. [46] M. Szydagis, C. Levy, G. Blockinger, A. Kamaha, N. Parveen, and G. Rischbieter, arXiv:2007.00528. [40] N. Vinyoles, A. Serenelli, F. L. Villante, S. Basu, J. Redondo, and J. Isern, J. Cosmol. Astropart. Phys. 10 (2015) 015. [47] F. Takahashi, M. Yamada, and W. Yin, arXiv:2006.10035. [41] P. Gondolo and G. G. Raffelt, Phys. Rev. D 79, 107301 (2009). [48] K. Kannike, M. Raidal, H. Veermäe, A. Strumia, and D. Teresi, arXiv:2006.10735. [42] For values of the couplings allowed by astrophysics the solar axion luminosity La is a negligible fraction of the total luminosity, for example, La ≈1.85 × 10−3g2 γ10L⊙for Primakoff emission [41]. Hence, effects on the Sun lifetime are also negligible. [49] G. Alonso-Álvarez, F. Ertas, J. Jaeckel, F. Kahlhoefer, and L. Thormaehlen, arXiv:2006.11243. [50] C. Boehm, D. G. Cerdeno, M. Fairbairn, P. A. Machado, and A. C. Vincent, arXiv:2006.11250. [51] B. Fornal, P. Sandick, J. Shu, M. Su, and Y. Zhao, arXiv:2006.11264. [43] Astrophysical constraints could only be evaded in exotic models in which the couplings strongly depend on the stellar environment, like the core density and temperature, see, e.g., [44]. [52] C. Gao, J. Liu, L.-T. Wang, X.-P. Wang, W. Xue, and Y.-M. Zhong, this issue, Phys. Rev. Lett. 125, 131806 (2020). [53] J. B. Dent, B. Dutta, J. L. Newstead, and A. Thompson, following Letter, Phys. Rev. Lett. 125, 131805 (2020). [44] J. Redondo, Can the PVLAS particle be compatible with the astrophysical bounds? Ph. D. thesis, Barcelona, Autonoma University, 2007. 131804-6 131804-6
https://openalex.org/W2131853364	https://etsmjournal.biomedcentral.com/track/pdf/10.1186/2040-7378-4-1	English	null	An experimental protocol for mimicking pathomechanisms of traumatic brain injury in mice	Experimental & translational stroke medicine	2,012	cc-by	4,297	© 2012 Albert-Weißenberger et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Traumatic brain injury (TBI) is a result of an outside force causing immediate mechanical disruption of brain tissue and delayed pathogenic events. In order to examine injury processes associated with TBI, a number of rodent models to induce brain trauma have been described. However, none of these models covers the entire spectrum of events that might occur in TBI. Here we provide a thorough methodological description of a straightforward closed head weight drop mouse model to assess brain injuries close to the clinical conditions of human TBI. Keywords: closed head injury, traumatic brain injury, neurobehavioural deficits, astrocyte, microglia, neurons 1 Introduction axonal swelling, axoplasmic ovoid retraction balls and expression of amyloid beta peptides [17]. The original weight drop model in rats by Feeney [18,19] was optimi- mized to produce an open-head brain injury whereas the model developed by Shohami [20-22] produces a closed-head brain injury with both focal and diffuse injury and enables its adaptation for mice. We provide here a methodological description of our modification of the Shohami model [14,21] to induce closed head weight drop injury in mice. Traumatic brain injury (TBI) is a result of an outside force causing immediate mechanical disruption of brain tissue and delayed pathogenic events that can exacerbate the injury (reviewed by [1]). It represents a leading cause of death and disability in the industrialized countries [2,3] and a growing health problem in the developing countries [4-7]. To better understand the pathological mechanisms underlying TBI and to develop strategies and interventions to limit the secondary damage, the use of rodent models is essential. A number of rodent models to induce brain trauma have been described; however, none of them covers the entire spectrum of events that might occur in TBI [8]. As an example, the cortical cryolesion model is particularly suited for inves- tigating TBI-associated focal lesions with blood-brain barrier leakage and vasogenic brain edema [9-13] but contre coup and diffuse axonal injuries that typically complicate human head injuries [1] are missing. This pathophysiology is present in the weight drop models of TBI which use the gravitational forces of a free falling weight to produce a mix of focal and diffuse brain injury [14-16]. One main characteristic of TBI associated dif- fuse axonal injury is the axonal disruption caused by shearing forces. Typical pathological changes include An experimental protocol for mimicking pathomechanisms of traumatic brain injury in mice Christiane Albert-Weißenberger1, Csanád Várrallyay2, Furat Raslan3, Christoph Kleinschnitz1 an 2 Methods All i All experiments require an appropriate animal experi- mentation facility and need to be conducted in accor- dance with the laws and regulations of the regulatory authorities for animal care. The animal experiments pre- sented here were approved by and conducted in accor- dance with the laws and regulations of the regulatory authorities for animal care and use in Lower Franconia. A total of 20 male C57BL/mice (10-16 weeks) with a body weight of > 20 g mice were used. Correspondence: Albert_C2@klinik.uni-wuerzburg.de 1University of Würzburg, Department of Neurology, Würzburg, Germany Full list of author information is available at the end of the article Albert-Weißenberger et al. Experimental & Translational Stroke Medicine 2012, 4:1 http://www.etsmjournal.com/content/4/1/1 Albert-Weißenberger et al. Experimental & Translational Stroke Medicine 2012, 4:1 http://www.etsmjournal.com/content/4/1/1 Albert-Weißenberger et al. Experimental & Translational Stroke Medicine 2012, 4:1 http://www.etsmjournal.com/content/4/1/1 Open Access 2.1 Weight drop model - Surgery After induction of anaesthesia with 4% isoflurane, a mouse was placed onto the platform directly under the weight of the weight drop device (Figure 1). Anaesthesia was maintained by mask inhalation of isoflurane vapor- ized at concentrations of 1.5% during the surgical * Correspondence: Albert_C2@klinik.uni-wuerzburg.de 1University of Würzburg, Department of Neurology, Würzburg, Germany Full list of author information is available at the end of the article Page 2 of 5 Albert-Weißenberger et al. Experimental & Translational Stroke Medicine 2012, 4:1 http://www.etsmjournal.com/content/4/1/1 Albert-Weißenberger et al. Experimental & Translational Stroke Medicine 2012, 4:1 http://www.etsmjournal.com/content/4/1/1 Figure 1 Weight drop injury device. The falling height of the free-falling weight with a silicon-covered blunt tip of 2 mm diameter can be carefully determined by a locking mechanism. Releasing a weight of 75 g from a height of 10 cm will cause a final impact of 0.01 J. Figure 2 Target area for rod impact. The correct area to be impacted should lie on the right hemisphere and is identified by the asterisk. Figure 1 Weight drop injury device. The falling height of the free-falling weight with a silicon-covered blunt tip of 2 mm diameter can be carefully determined by a locking mechanism. Releasing a weight of 75 g from a height of 10 cm will cause a final impact of 0.01 J. Figure 2 Target area for rod impact. The correct area to be impacted should lie on the right hemisphere and is identified by the asterisk. associated with physical and cognitive deficits (for a recent review see [23]). In parallel to human TBI, the brain injury caused by the weight drop model presented here is also associated with neurobehavioral deficits [21]. The neurobehavioral status of mice was obtained by the neurological severity score (NSS) (adapted from [14]). It consists of 10 individual clinical parameters, including tasks on motor function, alertness and physio- logical behaviour. Each parameter is described below. procedures and depth of anaesthesia was monitored by toe pinch using tweezers. A midline longitudinal scalp incision was made and the skull exposed. After identifi- cation of the right anterior frontal area (1.5 mm lateral to the midline in the mid-coronal plane) as impact area (Figure 2), the weight was released and dropped with a final impact of 0.01 J onto the skull. As the impact can result in trauma-induced respira- tory depression and death, posttraumatic oxygen was immediately applied. 2.1 Weight drop model - Surgery Then the scalp wound was closed by standard suture material (3-0 Ethilon) and the wound area was treated with lidocain cream. Mice were returned to cages immediately at the end of the surgical procedures where access to water and food is freely available. A) Exit circle The mouse was placed in the middle of a platform (Figure 3d) and monitored how long it takes the mouse to exit the platform. Owing to their intrinsic seeking behaviour, healthy mice usually exited the plat- form within 2 min (0 points). If the mouse failed to exit the platform within 2 min there was given 1 point. B) Seeking behavior The mouse was placed on the plat- form and the exploring of the environment and sniffing behavior was monitored. While healthy mice explored the environment (0 points), injured mice do not display this physiological behaviour (1 point). 2.2 Outcome Key readout parameters include the assessment of neu- robehavioral outcome, activation of resident cells, neu- rodegeneration, and morphological changes. C) Monoparesis/hemiparesis Monoparesis or hemipar- esis was represented by failure to use one (monoparesis) or two (hemiparesis) paws. When being picked up by 2.2.1 Neurobehavioral outcome During the last years, an extensive number of studies have been accomplished showing that human TBI are Page 3 of 5 Albert-Weißenberger et al. Experimental & Translational Stroke Medicine 2012, 4:1 http://www.etsmjournal.com/content/4/1/1 Figure 3 Depiction of the equipment used for assessment of the NSS. (a) Beam walk test. The healthy mouse will spontaneously cross 3-, 2- and 1-cm beams. The beam tests should start with the 3-cm beam and increase the level of difficulty by gradually reducing the beam width to 2 and 1 cm. (b) Beam for the beam balancing test. (c) The round stick balance test represents the most difficult test for mice. (d) The platform is used to evaluate the exploring behavior. Owing to their intrinsic inquisitiveness, a healthy mouse exits the platform within 2 min. Figure 4 Round stick balance test evaluation. Figure 3 Depiction of the equipment used for assessment of the NSS. (a) Beam walk test. The healthy mouse will spontaneously cross 3-, 2- and 1-cm beams. The beam tests should start with the 3-cm beam and increase the level of difficulty by gradually reducing the beam width to 2 and 1 cm. (b) Beam for the beam balancing test. (c) The round stick balance test represents the most difficult test for mice. (d) The platform is used to evaluate the exploring behavior. Owing to their intrinsic inquisitiveness, a healthy mouse exits the platform within 2 min. Figure 3 Depiction of the equipment used for assessment of the NSS. (a) Beam walk test. The healthy mouse will spontaneously cross 3-, 2- and 1-cm beams. The beam tests should start with the 3-cm beam and increase the level of difficulty by gradually reducing the beam width to 2 and 1 cm. (b) Beam for the beam balancing test. (c) The round stick balance test represents the most difficult test for mice. (d) The platform is used to evaluate the exploring behavior. Owing to their intrinsic inquisitiveness, a healthy mouse exits the platform within 2 min. Figure 3 Depiction of the equipment used for assessment of Figure 4 Round stick balance test evaluation. (Figure 3c and 4). 2.2 Outcome If a mouse failed to hang onto string by at least two paws 1 point was given. The points of each task were summed up to obtain the NSS. Initial severity of the trauma was assessed 1 h after trauma. All animals in this study exhibited neuro- logical deficits 1 h after trauma, with a group mean of 4.90 +- 0.62 (n = 10) (Figure 5). Evaluation of the NSS values at later time points revealed that the neurological status improved over time and compared to the initial score value 1 h after trauma, the NSS values were signif- icantly decreased at day 3 and 7 (n = 10, P < 0.05). These results show that the NSS is an excellent tool to estimate the initial severity of and recovery from brain injury induced by weight drop. Sham operated mice possessed no neurobehavioral deficits, reflected by a score of 0. In order to avoid bias in interpretation of the results evaluation of task performance were carried out in a blinded manner. the tail, a healthy mouse intrinsically gripped a small forceps touching its paws and holds on to it (0 points). If the mouse failed to grip there was given 1 point. D) Straight walk The mouse was placed on an even surface and alertness, initiative and motor ability to walk straight was assessed. If a mouse failed to walk straight due to missing initiative or dragging of one or several paws 1 point was given. E) Startle reflex A healthy mouse reacted with a bounce and/or wince to a sound stimulus (hand clap- ping) (score of 0). If the mouse failed to respond 1 point was given. 2.2.2 Immunhistochemical characterization F) Beam balancing The mouse was placed on a beam of 7 mm × 7 mm (Figure 3b). Healthy mice are able to balance on the beam for at least 10 s. If a mouse failed to balance, 1 point was given. Studies characterising the posttraumatic response of resi- dent brain cells in humans, show that TBI leads to Figure 5 The assessment of the NSS after trauma allows evaluation of the recovery. Initial severity of the trauma was assessed 1 h after trauma (group mean of 4.90 +- 0.62, n = 10). Compared to the initial score value 1 h after trauma, the NSS values were significantly decreased at day 3 and 7 (P < 0.05). 2.2 Outcome The 10-μm thin brain cryosections mounted on slides were immersed in a solution contain- ing 1% sodium hydroxide in 80% alcohol followed by 70% alcohol and water. They were then transferred to a solution of 0.06% potassium permanganate for 10 min- utes and rinsed in water. After 20 min staining with a concentration of 0.0004% Fluoro Jade B, the slides were rinsed with water and then dried at approximately 50°C. One week after weight-drop injury we observed numerous activated microglia (cell bodies with numer- ous branching processes) (Figure 6) and astrocytes (a process that is accompanied by hypertrophy and increased expression of the glial-specific intermediate filament GFAP). At the same time, also degenerating neurons were detected after trauma (Figure 7). Figure 7 Coronal T1 and T2 weighted magnetic resonance scans obtained on 1 day demonstrates a contusion in the ipsilateral hemisphere and on day 14 after injury enlarged ventricles indicating brain atrophy. 2.2.3 Magnetic resonance imaging h l l h d Morphological changes and progression of brain damage induced by weight drop can be assessed by magnetic resonance imaging (MRI) [29]. The representative T1- and T2-weighted MR-images in Figure 7, were obtained with a 3.0- Tesla magnetic resonance apparatus (Vision, Siemens) using a brain coil for rodents 1 day and 14 days after induction of injury. We detected a contusion in the ipsilateral hemisphere the 1st day after injury, and on the 14th day, enlarged ventricles were observed, indi- cating brain atrophy. 2.2 Outcome G) Beam walk Motor coordination and balance was also evaluated by the ability of a mouse to transverse a graded series of beams (Figure 3a). If a mouse failed to cross the 3-cm-wide and 30-cm-long beam, 3 points were scored and the test stopped. If a mouse managed to cross the 3- cm beam, then the test was repeated using a beam with a width of 2 cm. For a mouse failing to cross the 2-cm- beam, 2 points were scored and the test stopped. If a mouse managed to cross the 3-cm and 2-cm beam, but failed to cross the 1-cm beam, 1 point was given. Healthy mice managed to cross all beams (0 points). Figure 5 The assessment of the NSS after trauma allows evaluation of the recovery. Initial severity of the trauma was assessed 1 h after trauma (group mean of 4.90 +- 0.62, n = 10). Compared to the initial score value 1 h after trauma, the NSS values were significantly decreased at day 3 and 7 (P < 0.05). H) Round stick balancing Balance and grip strength are required to hold on a round stick of 3 mm diameter Albert-Weißenberger et al. Experimental & Translational Stroke Medicine 2012, 4:1 http://www.etsmjournal.com/content/4/1/1 Page 4 of 5 activation of astrocytes, microglia, and neurodegeneration [24-27]. Similarly, brain injury induced by the weight drop model presented here initiates pathological processes lead- ing to cellular changes in glia and neurons. Figure 7 Coronal T1 and T2 weighted magnetic resonance scans obtained on 1 day demonstrates a contusion in the ipsilateral hemisphere and on day 14 after injury enlarged ventricles indicating brain atrophy. g g g To detect activation of glia immunohistochemical stainings were performed as previously described [28]. Briefly, 10-μm thin brain cryosections were prepared using a cryostat (Leica, Bensheim, Germany). The immunohistochemical staining with antibodies for the detection of glial fibrillary acidic protein (GFAP)-expres- sing astrocytes (rabbit, 1:500, anti-GFAP; Ab7260; Abcam, Cambridge, UK) and activated microglia/macro- phages (rat, 1:100, anti-CD11b; MCA7; Serotec, Raleigh, USA) was performed following standard methods using an avidin-biotin system (Vector Laboratories, Burlin- game, USA) and 0.02% diaminobenzidine as chromogen (Kem-En-Tec Diagnostics, Taastrup, Denmark). Nega- tive controls included omission of either the primary or secondary antibody and gave no signals (not shown). To detect degenerating neurons, Fluoro Jade B (Millipore AG310) was used. 1. Gaetz M: The neurophysiology of brain injury. Clin Neurophysiol 2004, 115:4-18. 24. Castejon OJ: Morphological astrocytic changes in complicated human brain trauma. A light and electron microscopic study. Brain Inj 1998, 12:409-427, discussion 407. 24. Castejon OJ: Morphological astrocytic changes in complicated human brain trauma. A light and electron microscopic study. Brain Inj 1998, 12:409-427, discussion 407. 2. Faul M, Xu L, Wald MM, Coronado VG: Traumatic Brain Injury in the United States: Emergency Department Visits, Hospitalizations, and Deaths, 2002-2006. Book Traumatic Brain Injury in the United States: Emergency Department Visits, Hospitalizations, and Deaths, 2002-2006 City; 2010. 2. Faul M, Xu L, Wald MM, Coronado VG: Traumatic Brain Injury in the United States: Emergency Department Visits, Hospitalizations, and Deaths, 2002-2006. Book Traumatic Brain Injury in the United States: Emergency Department Visits, Hospitalizations, and Deaths, 2002-2006 City; 2010. 25. Castejon OJ: Electron microscopic analysis of cortical biopsies in patients with traumatic brain injuries and dysfunction of neurobehavioural system. J Submicrosc Cytol Pathol 1998, 30:145-156. 25. Castejon OJ: Electron microscopic analysis of cortical biopsies in patients with traumatic brain injuries and dysfunction of neurobehavioural system. J Submicrosc Cytol Pathol 1998, 30:145-156. 26. Oehmichen M, Meissner C: Routine techniques in forensic neuropathology as demonstrated by gunshot injury to the head. Legal Medicine 2009, 11(Supplement 1):S50-S53. 26. Oehmichen M, Meissner C: Routine techniques in forensic neuropathology as demonstrated by gunshot injury to the head. Legal Medicine 2009, 11(Supplement 1):S50-S53. 3. Maegele M, Engel D, Bouillon B, Lefering R, Fach H, Raum M, Buchheister B, Schaefer U, Klug N, Neugebauer E: Incidence and outcome of traumatic brain injury in an urban area in Western Europe over 10 years. Eur Surg Res 2007, 39:372-379. 3. Maegele M, Engel D, Bouillon B, Lefering R, Fach H, Raum M, Buchheister B, Schaefer U, Klug N, Neugebauer E: Incidence and outcome of traumatic brain injury in an urban area in Western Europe over 10 years. Eur Surg Res 2007, 39:372-379. 27. Oehmichen M, Walter T, Meissner C, Friedrich HJ: Time course of cortical hemorrhages after closed traumatic brain injury: statistical analysis of posttraumatic histomorphological alterations. J Neurotrauma 2003, 20:87-103. 4. Mock C, Joshipura M, Goosen J, Lormand JD, Maier R: Strengthening trauma systems globally: the Essential Trauma Care Project. J Trauma 2005, 59:1243-1246. 4. Mock C, Joshipura M, Goosen J, Lormand JD, Maier R: Strengthening trauma systems globally: the Essential Trauma Care Project. J Trauma 2005, 59:1243-1246. 28. Author details 1 1University of Würzburg, Department of Neurology, Würzburg, Germany. 2University of Würzburg, Department of Neuroradiology Würzburg, Germany. 3University of Würzburg, Department of Neurosurgery, Würzburg, Germany. 1University of Würzburg, Department of Neurology, Würzburg, Germany. 2University of Würzburg, Department of Neuroradiology Würzburg, Germany. 3University of Würzburg, Department of Neurosurgery, Würzburg, Germany. 21. Chen Y, Constantini S, Trembovler V, Weinstock M, Shohami E: An experimental model of closed head injury in mice: pathophysiology, histopathology, and cognitive deficits. J Neurotrauma 1996, 13:557-568. Received: 30 January 2012 Accepted: 2 February 2012 Published: 2 February 2012 Received: 30 January 2012 Accepted: 2 February 2012 Published: 2 February 2012 22. Shapira Y, Shohami E: Experimental studies on brain oedema after blunt head injury: experimental approaches from animal experimentation to actual or possible clinical application. Eur J Anaesthesiol 1993, 10:155-173. Acknowledgements h bl f g This publication was funded by the German Research Foundation (DFG) and the University of Wuerzburg in the funding programme Open Access Publishing. 19. Dail WG, Feeney DM, Murray HM, Linn RT, Boyeson MG: Responses to cortical injury: II. Widespread depression of the activity of an enzyme in cortex remote from a focal injury. Brain Res 1981, 211:79-89. 20. Panikashvili D, Simeonidou C, Ben-Shabat S, Hanus L, Breuer A, Mechoulam R, Shohami E: An endogenous cannabinoid (2-AG) is neuroprotective after brain injury. Nature 2001, 413:527-531. 3 Outlook In this article we provide a thorough methodological description of a straightforward weight-drop mouse model reproducing brain injuries similar to those Figure 6 Weight drop injury leads to activation of microglia (A.), astrocytes (B.) and neurodegeneration (C.) as shown by immunhistological stainings. Figure 6 Weight drop injury leads to activation of microglia (A.), astrocytes (B.) and neurodegeneration (C.) as shown by immunhistological stainings. Page 5 of 5 Page 5 of 5 Albert-Weißenberger et al. Experimental & Translational Stroke Medicine 2012, 4:1 http://www.etsmjournal.com/content/4/1/1 Albert-Weißenberger et al. Experimental & Translational Stroke Medicine 2012, 4:1 http://www.etsmjournal.com/content/4/1/1 14. Flierl MA, Stahel PF, Beauchamp KM, Morgan SJ, Smith WR, Shohami E: Mouse closed head injury model induced by a weight-drop device. Nat Protoc 2009, 4:1328-1337. experienced in the human with a mix of diffuse injury pattern and focal contusion. A major advantage of this model is that it allows neurological scoring immediately after injury and in the sequel. Thus, although the inflicted brain injury varies in its severity, clinically rele- vant randomization of mice into the various treatment groups is possible and this animal model might facilitate the discovery of improved drug treatments for TBI. 15. Foda MA, Marmarou A: A new model of diffuse brain injury in rats. Part II: Morphological characterization. J Neurosurg 1994, 80:301-313. 15. Foda MA, Marmarou A: A new model of diffuse brain injury in rats. Part II: Morphological characterization. J Neurosurg 1994, 80:301-313. 16. Marmarou A, Foda MA, van den Brink W, Campbell J, Kita H, Demetriadou K: A new model of diffuse brain injury in rats. Part I: Pathophysiology and biomechanics. J Neurosurg 1994, 80:291-300. 16. Marmarou A, Foda MA, van den Brink W, Campbell J, Kita H, Demetriadou K: A new model of diffuse brain injury in rats. Part I: Pathophysiology and biomechanics. J Neurosurg 1994, 80:291-300. 17. Loane DJ, Pocivavsek A, Moussa CE, Thompson R, Matsuoka Y, Faden AI, Rebeck GW, Burns MP: Amyloid precursor protein secretases as therapeutic targets for traumatic brain injury. Nat Med 2009, 15:377-379. 18. Feeney DM, Boyeson MG, Linn RT, Murray HM, Dail WG: Responses to cortical injury: I. Methodology and local effects of contusions in the rat. Brain Res 1981, 211:67-77. 1. Gaetz M: The neurophysiology of brain injury. Clin Neurophysiol 2004, 115:4-18. Kleinschnitz C, Stoll G, Bendszus M, Schuh K, Pauer HU, Burfeind P, Renne C, Gailani D, Nieswandt B, Renne T: Targeting coagulation factor XII provides protection from pathological thrombosis in cerebral ischemia without interfering with hemostasis. J Exp Med 2006, 203:513-518. 5. Yattoo G, Tabish A: The profile of head injuries and traumatic brain injury deaths in Kashmir. J Trauma Manag Outcomes 2008, 2:5. 5. Yattoo G, Tabish A: The profile of head injuries and traumatic brain injury deaths in Kashmir. J Trauma Manag Outcomes 2008, 2:5. 6. Chiu WT, Huang SJ, Tsai SH, Lin JW, Tsai MD, Lin TJ, Huang WC: The impact of time, legislation, and geography on the epidemiology of traumatic brain injury. J Clin Neurosci 2007, 14:930-935. 29. Tsenter J, Beni-Adani L, Assaf Y, Alexandrovich AG, Trembovler V, Shohami E: Dynamic changes in the recovery after traumatic brain injury in mice: effect of injury severity on T2-weighted MRI abnormalities, and motor and cognitive functions. J Neurotrauma 2008, 25:324-333. 29. Tsenter J, Beni-Adani L, Assaf Y, Alexandrovich AG, Trembovler V, Shohami E: Dynamic changes in the recovery after traumatic brain injury in mice: effect of injury severity on T2-weighted MRI abnormalities, and motor and cognitive functions. J Neurotrauma 2008, 25:324-333. j y 7. Martins ET, Linhares MN, Sousa DS, Schroeder HK, Meinerz J, Rigo LA, Bertotti MM, Gullo J, Hohl A, Dal-Pizzol F, Walz R: Mortality in severe traumatic brain injury: a multivariated analysis of 748 Brazilian patients from Florianopolis City. J Trauma 2009, 67:85-90. doi:10.1186/2040-7378-4-1 Cite this article as: Albert-Weißenberger et al.: An experimental protocol for mimicking pathomechanisms of traumatic brain injury in mice. Experimental & Translational Stroke Medicine 2012 4:1. 8. Albert-Weissenberger C, Siren AL: Experimental traumatic brain injury. Exp Transl Stroke Med 2010, 2:16. 9. Rakos G, Kis Z, Nagy D, Lur G, Farkas T, Hortobagyi T, Vecsei L, Toldi J: Evans Blue fluorescence permits the rapid visualization of non-intact cells in the perilesional rim of cold-injured rat brain. Acta Neurobiol Exp (Wars) 2007, 67:149-154. 10. Raslan F, Schwarz T, Meuth SG, Austinat M, Bader M, Renne T, Roosen K, Stoll G, Siren AL, Kleinschnitz C: Inhibition of bradykinin receptor B1 protects mice from focal brain injury by reducing blood-brain barrier leakage and inflammation. J Cereb Blood Flow Metab 2010, 30:1477-1486. doi:10.1186/2040-7378-4-1 Cite this article as: Albert-Weißenberger et al.: An experimental protocol for mimicking pathomechanisms of traumatic brain injury in mice. Experimental & Translational Stroke Medicine 2012 4:1. References 23. Shukla D, Devi BI, Agrawal A: Outcome measures for traumatic brain injury. Clinical Neurology and Neurosurgery 2011, 113:435-441. 1. Gaetz M: The neurophysiology of brain injury. Clin Neurophysiol 2004, 115:4-18. 1. Gaetz M: The neurophysiology of brain injury. Clin Neurophysiol 2004, 115:4-18. 1. Gaetz M: The neurophysiology of brain injury. Clin Neurophysiol 2004, 115:4-18. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: Albert-Weißenberger et al. Experimental & Translational Stroke Medicine 2012, 4:1 http://www.etsmjournal.com/content/4/1/1 Submit your next manuscript to BioMed Central and take full advantage of: 11. Stoffel M, Blau C, Reinl H, Breidt J, Gersonde K, Baethmann A, Plesnila N: Identification of brain tissue necrosis by MRI: validation by histomorphometry. J Neurotrauma 2004, 21:733-740. • Convenient online submission 12. Eriskat J, Furst M, Stoffel M, Baethmann A: Correlation of lesion volume and brain swelling from a focal brain trauma. Acta Neurochir Suppl 2003, 86:265-266. 12. Eriskat J, Furst M, Stoffel M, Baethmann A: Correlation of lesion volume and brain swelling from a focal brain trauma. Acta Neurochir Suppl 2003, 86:265-266. 13. Siren AL, Radyushkin K, Boretius S, Kammer D, Riechers CC, Natt O, Sargin D, Watanabe T, Sperling S, Michaelis T, et al: Global brain atrophy after unilateral parietal lesion and its prevention by erythropoietin. Brain 2006, 129:480-489.
https://openalex.org/W2887425698	https://www.nature.com/articles/s41598-018-30584-6.pdf	English	null	Advancing micro-scale cooling by utilizing liquid-liquid phase separation	Scientific reports	2,018	cc-by	8,290	Advancing micro-scale cooling by utilizing liquid-liquid phase separation OPEN Received: 31 January 2018 Accepted: 2 August 2018 Published: xx xx xxxx Received: 31 January 2018 Accepted: 2 August 2018 Published: xx xx xxxx Wei Xing1, Amos Ullmann2, Neima Brauner2, Joel Plawsky3 & Yoav Peles4 Achieving effective cooling within limited space is one of the key challenges for miniaturized product design. State-of-the-art micro-scale cooling enhancement techniques incorporate flow disturbances and boiling to reach high performance. However, these methods face the inherent issues of extra pressure drop, flow instability and dry-out that limits heat flux. Here we demonstrate that substantial cooling capability enhancement, up to 2.5 times, is realized by introducing the phase separation of a triethylamine (TEA)/water mixture at the micro-scale. Our experiments show that the enhancement behavior is closely related to the system’s initial composition, temperature, and flow conditions. Moreover, the mixture system exhibits reduced pressure drop after separation, which makes it more promising in serving practical applications. The results reveal new possibilities for liquid coolant selection and provide the experimental foundation for further research in this area. Effective thermal management — the ability to control the temperature of a system by leveraging the science of thermodynamics and heat transfer — has been a key technological pursuit of modern society. Decades of extensive research and development aimed at enabling the highly efficient operation of equipment, such as gas turbines, medical devices, and arrays of electronic equipment (e.g., electronic cabinets, computer chips, laser diodes, power amplifiers, batteries, motors etc.). In recent years, a new set of challenges has been posted by the miniaturization of powered machines and the development of high performance semiconductor chips, which require advanced micro-scale cooling methods to ensure operational safety. Traditional convective heat trans- fer enhancement methods include building extended surfaces for increased solid/liquid contact area, stirring to induce flow mixing, using the latent heat of phase change, and so forth1–6. Among these strategies, boiling and condensation two-phase flow heat transfer systems are generally regarded to be most effective. However, phase change heat transfer systems are vulnerable to some inherent issues, such as flow instability and critical heat flux, especially at the micro-scale7,8. Despite extensive research, these issues often hinder the wide spread use of phase change processes in practical systems.h The phase separation of a partially miscible liquid system is characterized by the formation and movement of fluid domains9–12. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Advancing micro-scale cooling by utilizing liquid-liquid phase separation OPEN In several recent publications, it has been shown that this self-propelled motion acts as a stirring mechanism that positively affects thermal transport13,14. Poesio et al. pioneered the phase separation heat transfer research by confining the water/toluene system in a closed cell. They observed faster temperature response when phase separation was induced13. Molin and Mauri used the one-parameter Margules correlation and numerically modeled phase separation process for a quiescent system. Their results indicated that the Nusselt number increases with the mass transfer Peclet number14. For convection systems, enhanced thermal transport was also observed15,16. Most phase separation processes are triggered by reducing the system temperature, and thus, they are suita- ble for heating applications. However, when strong polar interactions are present (such as hydrogen bonding), some liquid mixtures are only miscible below certain temperatures, termed lower critical solution temperatures (LCSTs). An LCST system undergoes phase separation when the temperature exceeds a certain threshold value, and is potentially suitable for cooling applications (See Supplementary Note 1 for thermodynamic point of view 1Department of Mechanical, Aerospace and Nuclear Engineering, Rensselaer Polytechnic Institute, 110 8th Street, Troy, NY, 12180, USA. 2School of Mechanical Engineering, Faculty of Engineering, Tel-Aviv University, Tel-Aviv, Israel. 3Howard P. Isermann Department of Chemical and Biological Engineering, Rensselaer Polytechnic Institute, 110 8th Street, Troy, NY, 12180, USA. 4Department of Mechanical and Aerospace Engineering, University of Central Florida, 12760 Pegasus Blvd, Orlando, FL, 32816, USA. Correspondence and requests for materials should be addressed to W.X. (email: xingw@rpi.edu) or Y.P. (email: Yoav.Peles@ucf.edu) Scientific RePorTs \| (2018) 8:12093 \| DOI:10.1038/s41598-018-30584-6 1 www.nature.com/scientificreports/ of solution behavior). Ullmann et al. pioneered the study for phase separation convection cooling with an LCST system, and successfully demonstrated enhanced heat transfer performance17. Xing et al. first introduced phase separation cooling at micro-scale, and provided experimental evidence for improved heat transfer coefficient18. Based on the existing studies, more work needs to be conducted to assist more in-depth understanding of the micro-scale phase separation heat transfer, i.e. geometrical effects, concentration effects etc. p p gff Herein we employ the triethylamine (TEA)/water system, which has an LCST of 18 °C and critical composi- tion of 32.1% TEA mass fraction, to examine its phase separation heat transfer characteristics and multi-phase flow behavior at the micro-scale (See Supplementary Note 2 for detailed information of the binodal behavior of TEA/water mixture; the system’s physical/thermal properties are shown in Supplementary Fig. 1). Resultsh The convection heat transfer coefficient, h, quantifies the effectiveness of a thermal transport process. It measures the amount of heat transferred per unit area per unit temperature difference between the solid surface and the fluid at a given flow condition: = − h Q A T T ( ) (1) w f (1) Here, we apply h to the phase separating flow. The calculation formulas of h are slightly different, depending on the specific experimental conditions. Here, we apply h to the phase separating flow. The calculation formulas of h are slightly different, depending on the specific experimental conditions. Average thermal transport at the critical composition. The average rate of thermal transport was evaluated using a rectangular microchannel structure. The channel was 22 mm long, 2 mm wide, and 0.4 mm deep. Two resistance heaters of different lengths were deposited at the channel bottom. The long heater was 10 mm long and 1 mm wide, and the short heater was 3 mm long with the same width (Fig. 1(a) and Supplementary Fig. 2). The liquid solution entered the heated channel at its critical composition (i.e., at 32.1% TEA mass fraction) and below its critical temperature. When electrical power was applied to the heater, the mixture’s temperature first rose to the critical value (Tcritical) while the mixture remained single phase. When phase separation occurred, the mixture absorbed heat to break its hydrogen bonds, and separated into liquid phases of different compositions. The equilibrium concentration of each phase was determined by consulting the phase diagram17,19 (Fig. 1(b)). At the critical composition, phase separation occurs via spinodal decomposition. b f h l h fl l ( (b)) h h l h fl p p p p p It can be seen from the long heater flow visualization images (Fig. 1(b)) that the multi-phase flow pattern changed with the applied heat flux. At low heat flux (e.g. Q″ = 4 W/cm2), the flow exhibited a mist flow pattern at the downstream section of the heater, where the fluid temperature was just above the critical value (Tcritical). With the increasing heat flux (e.g. Q″ = 18 W/cm2), the onset of phase separation propagated upstream, and the phase separating flow covered a larger heater area. Since the fluid was heated above Tcritical, the system was far removed from its initial equilibrium state. The fluid domains are distinguishable with an elongated droplet flow pattern. Resultsh As the heat flux continued to increase (e.g. Q″ = 74 W/cm2), the onset of phase separation remained unchanged, but the region of phase separation penetrated deeper into the fluid bulk. The fluid domains are more distinct, and the flow pattern becomes annular-like (note that no phase separation occurred at the top layer of the flow passage). The flow field on the short heater resembles the first 1/3 of the long heater.fii hlii g Figure 1(c and d) depict the average heat transfer coefficient, havg, for both the long and the short heaters at fixed mass fluxes (m″). havg,long is characterized by three zones (Fig. 1(c)): Zone I, at a heat flux of Q″ ~ 2 W/cm2, was a region where the mixture remained single phase; in Zone II, havg,long increased rapidly with heat flux (Q″ ~ 4–18 W/cm2); and in Zone III, havg,long stabilized at perhaps an asymptotic value (Q″ > 18 W/cm2). Because of the smaller area, the total heat transferred to the fluid from the short heater was much less than that transferred from the long heater when both operated at the same heat flux. Therefore, havg,short increased, but the three-zone behavior was less apparent. The flow images provide insight into the three-zone behavior. Zone II corresponds to the expansion of the phase separated region at the heater/liquid interface, and thus, a larger area of the heated solid surface is experiencing the heat transfer benefit caused by phase separation. This leads to a rapid increase of havg,long. Once the phase separation region ceases to propagate upstream and extends into the fluid bulk perpendicular to the heated wall, the enhancement of the thermal transport slows down. Thus, havg,long increases at a diminishing rate in Zone III. The phase separation flow exhibits greater heat transfer coefficient, compared to the case where water is used as the coolant at a similar mass flux (Fig. 1(c)).hfll l g The effect of fluid inlet temperature (Tin) is also examined in Fig. 1(c and d). At low heat fluxes, a mixture with higher Tin is closer to the critical temperature, and thus shows higher havg. However, at intermediate to high heat fluxes, the effect of Tin yields no significant difference in havg. Due to the shape of the binodal curve, after the onset of phase separation, the compositions of the separated phases are nearly independent of the fluid temperature (Fig. 1(b)). Advancing micro-scale cooling by utilizing liquid-liquid phase separation OPEN Our results experimentally demonstrate the feasibility of using an LCST mixture for micro-scale cooling applications, and the effects of initial composition, inlet temperature and heat flux are studied. We find that, at the micro-scale, phase separation at the critical TEA/water composition can enhance heat transfer up to 2.5 times with reduced pressure drop. Two new parameters, an equivalent mass quality, xw, and an augmentation factor, AF, were defined to rep- resent the amount of fluid that has separated into two phases and the effectiveness of heat transfer for the phase separating flow over single phase flow. The phase separating flow morphologies, which indicate the concentration field, were observed and characterized against the specific flow conditions. Future research directions and chal- lenges within this subject matter are identified and discussed. Resultsh Therefore, the thermal transport performance is nearly independent of the fluid temperature (as well as the heat flux), once the system temperature is much higher than Tcritical. The viscosities of the separated phases Scientific RePorTs \| (2018) 8:12093 \| DOI:10.1038/s41598-018-30584-6 2 www.nature.com/scientificreports/ Figure 1. Flow visualization and heat transfer results of critical composition experiments. (a) Schematics of the three-layer testing environment. (b) Phase diagram and flow visualization images. Photos on the left column are the short heater visualizations (m″ = 284 kg/m2s, Tin = 14.1 °C), long heater visualizations are on the right column (m″ = 208 kg/m2s, Tin = 11.4 °C). The system’s location on the phase diagram is approximated using the average wall temperature. For the long heater case with Q″ = 73.8 W/cm2, Tw = 79.3 °C, its location is marked at the topmost location on the phase diagram. (c) Long heater average heat transfer coefficient results at m″ = 208 kg/m2s, m″water = 190 kg/m2s. (d) Short heater average heat transfer coefficients results at m″ = 284 kg/m2s, m″water = 274 kg/m2s. (e) Pressure drop results for the long heater at m″ = 208 kg/m2s. Full quality flow visualization images can be found in Supplementary Fig. 7. Figure 1. Flow visualization and heat transfer results of critical composition experiments. (a) Schematics of the three-layer testing environment. (b) Phase diagram and flow visualization images. Photos on the left column are the short heater visualizations (m″ = 284 kg/m2s, Tin = 14.1 °C), long heater visualizations are on the right column (m″ = 208 kg/m2s, Tin = 11.4 °C). The system’s location on the phase diagram is approximated using the average wall temperature. For the long heater case with Q″ = 73.8 W/cm2, Tw = 79.3 °C, its location is marked at the topmost location on the phase diagram. (c) Long heater average heat transfer coefficient results at m″ = 208 kg/m2s, m″water = 190 kg/m2s. (d) Short heater average heat transfer coefficients results at m″ = 284 kg/m2s, m″water = 274 kg/m2s. (e) Pressure drop results for the long heater at m″ = 208 kg/m2s. Full quality flow visualization images can be found in Supplementary Fig. 7. are both less than that of the mixture (Supplementary Fig. 1(b)), and so the pressure drop of the system decreases after phase separation (Fig. 1(e)). Local thermal transport at the critical composition. Resultsh The critical composition local thermal transport process was examined with localized measurements using resistance temperature detectors (RTDs). The RTDs were located along the flow direction at several locations (Supplementary Fig. 2). The three-zone behavior is indistinct for hlocal (Fig. 2(a)), since the total heat input was less from the short heater. Because RTD2 locates at a more upstream location than RTD3 and RTD4 (Supplementary Fig. 2), the entry length effect results in the great- est local heat transfer coefficient among all RTDs. fi g To further analyze the thermal transport characteristics, the data is presented in a non-dimensional form. Borrowing the concept from flow boiling, an equivalent mass quality, xw, was defined based on the wall tem- perature. The flow quality presents the percentage of fluid that has separated into two phases assuming that the enthalpy of mixing is dependent upon the wall temperature20–22. An augmentation factor, AF, was defined as the ratio of h between the phase separated flow and the single phase flow of the same mixture at the same flow rate and inlet conditions. Figure 2(b) shows the local AF at various locations along the heater. As the mixture moves downstream and receives more energy, AFlocal increases due to the enthalpy of phase separation and the flow mixing effect. The maximum AFlocal is observed at RTD4, where the mass quality peaks. With the aid of the dimensionless parameters, all experimental data at the critical composition can be superimposed onto a single plot (Fig. 2(c)). AF and xw clearly show an asymptotic behavior, and this behavior holds for different heating lengths, inlet temperatures and flow rates at both the average and local levels. The maximum AF observed during experiment is about 2.5. The rapid increase for havg occurs at mass qualities less than 0.1, indicating that AFavg is mainly boosted by phase separation in the vicinity of the wall, not in the bulk of the fluid. Similarly, AFlocal also experienced a rapid increase in this range of quality. It is proposed that the maximum AF occurs when the flow quality reaches unity, indicating that all the mixture undergoes phase separation. Thermal transport at off-critical compositions. Two off-critical concentrations, 15% and 50% TEA mass fraction, were examined. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 2. Heat transfer coefficients and AF data for the critical composition. (a) Local heat transfer coefficients on the short heater, m″ = 208 kg/m2s. The fluid flows from RTD2 to RTD4 in local measurement experiments. (b) AF by locations of RTDs, m″ = 284 kg/m2s, Tin = 10.2 °C. (c) AF vs. mass quality at various experimental conditions. For clarity, not all uncertainties are plotted. Figure 2. Heat transfer coefficients and AF data for the critical composition. (a) Local heat transfer coefficients on the short heater, m″ = 208 kg/m2s. The fluid flows from RTD2 to RTD4 in local measurement experiments. (b) AF by locations of RTDs, m″ = 284 kg/m2s, Tin = 10.2 °C. (c) AF vs. mass quality at various experimental conditions. For clarity, not all uncertainties are plotted. separation morphology clearly exhibits a nucleation pattern (Fig. 3(a)). As more heat is delivered to the fluid, the string shape regions coalesce to form thicker strings. Under bulk advection and the shear force, the resulting flow pattern resembles a droplet/string flow. Flow visualization on the short heater confirms the results from the long heater and provides a closer look at the drop-like domains. The heat transfer coefficient using a 15% mixture increased almost linearly (Fig. 3(b and c)) with applied heat flux, both on average and on local basis. However, the enhancement ratio was weaker than the case of critical composition.l p Similar to the 15% mixture, the 50% mixture exhibited a droplet/string flow pattern, but the sizes of the droplets and strings were smaller (Fig. 4(a)). At this composition, h first increased slightly and then dropped to values below the corresponding single phase flow h. According to the lever rule, after phase separation, the 15% mixture has more content of the thermally favorable phase (water-rich) than the 50% mixture (Table 1). As a result, the majority of the solid/liquid interface was covered by the favorable phase for the 15% mixture. On the other hand, the 50% TEA mixture benefits the most from heat of mixing effects (Table 1). With the onset of phase separation, latent heat and flow mixing boost the thermal energy transport. However, the solid/liquid interface was gradually covered by the unfavorable phase (TEA-rich), due to its greater wettability on the device surface (Supplementary Fig. 3). Resultsh At off-critical compositions, phase separation proceeds by nucleation and domain growth, which is a localized and less rigorous process, compared to spinodal decomposition23–25 (See supplemen- tary Note 3 for more information about nucleation and spinodal decomposition). For the 15% mixture, the phase Scientific RePorTs \| (2018) 8:12093 \| DOI:10.1038/s41598-018-30584-6 3 Scientific RePorTs \| (2018) 8:12093 \| DOI:10.1038/s41598-018-30584-6 www.nature.com/scientificreports/ Photos on the left column are the short heater visualizations (m″ = 215 kg/m2s, Tin = 14.6 °C), long heater visualizations are on the right column (m″ = 241 kg/m2s, Tin = 12.8 °C). The system’s location on the phase diagram is approximated using the average wall temperature. (b) Average heat transfer coefficient results at m″ = 402 kg/m2s, Tin = 12.8 °C. (c) Short heater local heat transfer coefficient results at m″ = 358 kg/m2s, Tin = 14.0 °C. Full quality flow visualization images can be found in Supplementary Fig. 8. Figure 3. Flow visualization and heat transfer results summary of 15% TEA mass fraction experiments. (a) Phase diagram and flow visualization images. Photos on the left column are the short heater visualizations (m″ = 215 kg/m2s, Tin = 14.6 °C), long heater visualizations are on the right column (m″ = 241 kg/m2s, Tin = 12.8 °C). The system’s location on the phase diagram is approximated using the average wall temperature. (b) Average heat transfer coefficient results at m″ = 402 kg/m2s, Tin = 12.8 °C. (c) Short heater local heat transfer coefficient results at m″ = 358 kg/m2s, Tin = 14.0 °C. Full quality flow visualization images can be found in Supplementary Fig. 8. immediately separates into different phases. As the separated fluids flow downstream, heat is transferred to the separated phases that are already at concentrations determined by the binodal curve. That is to say, the fluid mix- ture always stays within the gap between the binodal and spinodal curve. y y g p p Observation of the 15% mixture also reveals that the two separated phases (i.e., water-rich and TEA-rich) end up with larger droplets than the 50% mixture. The size of the droplets is determined by the principle of minimiz- ing the Gibb’s free energy of the system. Therefore, the dependence of Gibb’s free energy on the system’s concen- tration might be different in both cases. The different droplet sizes might also be linked to the difference in the systems’ viscosities (Supplementary Fig. 1(b)) and the hydrodynamics of the flow field. The viscous force tends to tear the fluid domains into smaller droplets, and thus, the 50% mixture (greater viscosity) tends to separate into domains with smaller sizes. Flow visualization and image analysis at the critical composition. www.nature.com/scientificreports/ As a result, the heat transfer process from the wall to the bulk of the fluid is hindered, i.e., a layer of low thermal conductivity material inhibits thermal energy transport. Consequently, the overall heat transfer performance deteriorates compared to the corresponding single phase flow at the same mass flux. It is shown that the 50% mixture yields reduced cooling performance, while the 15% mixture shows good heat transfer enhancement. However, the heat transfer coefficient for the 15% mixture varies considerably with input heat flux with no asymptotic behavior, such that h is sensitive to heat flux. For practical applications, the actual performance is difficult to predict. As a short conclusion, the critical composition mixture seems to exhibit desir- able performance augmentation. p g Similar to spinodal decomposition, the nucleated droplets coalesce to form bigger domains. Interestingly, at the downstream section of the heater, where the fluid near the wall reaches higher temperatures, the flow pattern seems to be string flow, which is different from the spinodal decomposition pattern. Generally, as a mixture reaches higher temperatures, it transitions from a nucleation region into a spinodal decomposition region (as shown in the phase diagram). However, this was not observed during our experiments. Two scenarios are possibly involved here. The spinodal curve is narrow, such that spinodal decomposition only occurs at compositions very close to the critical composition. Alternatively, as nucleation occurs in the heater’s upstream section, the mixture Scientific RePorTs \| (2018) 8:12093 \| DOI:10.1038/s41598-018-30584-6 4 www.nature.com/scientificreports/ Figure 3. Flow visualization and heat transfer results summary of 15% TEA mass fraction experiments. (a) Phase diagram and flow visualization images. Photos on the left column are the short heater visualizations (m″ = 215 kg/m2s, Tin = 14.6 °C), long heater visualizations are on the right column (m″ = 241 kg/m2s, Tin = 12.8 °C). The system’s location on the phase diagram is approximated using the average wall temperature (b) Average heat transfer coefficient results at m″ = 402 kg/m2s, Tin = 12.8 °C. (c) Short heater local heat transf coefficient results at m″ = 358 kg/m2s, Tin = 14.0 °C. Full quality flow visualization images can be found in Supplementary Fig. 8. Figure 3. Flow visualization and heat transfer results summary of 15% TEA mass fraction experiments. (a) Phase diagram and flow visualization images. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Composition xh (water-rich phase) xl (TEA-rich phase) Δhmix (kJ/kg) 32.1% (critical) 64.8% 35.2% 46.2 15% 86.5% 13.5% 29.5 50% 49.1% 50.9% 50.6 Table 1. Comparison of water-rich phase percentage (xh), TEA-rich phase percentage (xl) and enthalpy of mixing between critical composition, 15% and 50% TEA mass fraction, assuming the system’s temperature is 50 °C. Composition xh (water-rich phase) xl (TEA-rich phase) Δhmix (kJ/kg) 32.1% (critical) 64.8% 35.2% 46.2 15% 86.5% 13.5% 29.5 50% 49.1% 50.9% 50.6 Table 1. Comparison of water-rich phase percentage (xh), TEA-rich phase percentage (xl) and enthalpy of mixing between critical composition, 15% and 50% TEA mass fraction, assuming the system’s temperature is 50°C. Table 1. Comparison of water-rich phase percentage (xh), TEA-rich phase percentage (xl) and enthalpy of mixing between critical composition, 15% and 50% TEA mass fraction, assuming the system’s temperature is 50 °C. i l h i t d th t i h h l i di t f TEA t ti Figure 4. Flow visualization and heat transfer results summary of 50% TEA mass fraction experiments. (a) Phase diagram and flow visualization images. Photos on the left column are the short heater visualizations (m″ = 358 kg/m2s, Tin = 14.5 °C), long heater visualizations are on the right column (m″ = 178 kg/m2s, Tin = 12.2 °C). The system’s location on the phase diagram is approximated using the average wall temperature. (b) Average heat transfer coefficient results at m″ = 358 kg/m2s, Tin = 11.8 °C. (c) Short heater local heat transfer coefficient at m″ = 358 kg/m2s, Tin = 13.8 °C. Full quality flow visualization images can be found in Supplementary Fig. 9. Figure 4. Flow visualization and heat transfer results summary of 50% TEA mass fraction experiments. ( ) Ph d d fl l Ph h l ft l h h h l Figure 4. Flow visualization and heat transfer results summary of 50% TEA mass fraction experiments. (a) Phase diagram and flow visualization images. Photos on the left column are the short heater visualizations (m″ = 358 kg/m2s, Tin = 14.5 °C), long heater visualizations are on the right column (m″ = 178 kg/m2s, Tin = 12.2 °C). The system’s location on the phase diagram is approximated using the average wall temperature. (b) Average heat transfer coefficient results at m″ = 358 kg/m2s, Tin = 11.8 °C. www.nature.com/scientificreports/ For spinodal decomposition of quiescent systems, the morphology of the domains has been shown to be bi-continuous and dendritic11,26,27. However, flow visualization in the current study shows a different result: the morphology during spinodal decom- position changes with forced advection28,29. At a given flow rate, the overall concentration field evolves with increasing heat flux and the fluid temperature (Fig. 1(b)). For the same average fluid outlet temperature, the flow pattern varies at different flow rates (Fig. 5(a)). It is postulated that the flow morphology also depends on the flu- id’s residence time in the heated channel. Because the TEA-rich phase appears darker (greater turbidity) than the Scientific RePorTs \| (2018) 8:12093 \| DOI:10.1038/s41598-018-30584-6 5 www.nature.com/scientificreports/ (d) separation layer thickness estimation at m″ = 567 kg/m2s. presence of darker shades resulting from phase separation. The gray value reached an asymptote at a heat flux of about 18 W/cm2. The asymptotic characteristics of gm is consistent with the trend observed for havg (Fig. 1(c)). Figure 5(a) also presents a comparison between different flow rates with similar outlet mean fluid temperature. Higher flow rates correspond to shorter residence times, which leads to a higher mean gray value in the last 10% length of the heater. g The separated phases reside in the sub-layer of the thermal boundary layer where the local temperature exceeds Tcritical. The thickness of this phase separation boundary layer, denoted as δs, is linked to the local mass quality and substantially affects the observed gray value. The local gray value (gm,local) analysis was conducted while assuming that the temperature of the fluid close to the heated wall could be approximately measured by the wall RTDs. The same gray value acquisition technique was applied at the RTD locations. Due to the absence of a non-isothermal phase separation flow model, a single phase conjugate heat transfer model was used to evaluate the separation boundary layer thickness. To match the simulated wall temperatures with measured values, both the mixture specific heat (Supplementary Fig. 1(a)) and an effective thermal conductivity (keff = 0.68 W/mK) is used, noting that the value of keff is about twice of the thermal conductivity of the single phase mixture, and is greater than that of the water (for details of the model see Supplementary Note 4). This indicates that both the latent heat effect and lateral mixing induced by phase separation are responsible for the heat transfer enhancement.fl Figure 5(c) depicts the mean gray value on top of RTD3 and RTD4 at different fluid inlet temperatures. It can be seen that the inlet temperature has very little effect on gm,local. However, for gray values at RTD3 and RTD4 at the same inlet temperature and wall temperature, major differences are observed due to different residence times and separation boundary layer thicknesses (Fig. 5(d)). For RTD3 with Tin = 15.0 °C and RTD4 with Tin = 12.0 °C, δs is almost identical. Thus, the difference in residence time accounts for the difference in gm,local. www.nature.com/scientificreports/ (c) Short heater local heat transfer coefficient at m″ = 358 kg/m2s, Tin = 13.8 °C. Full quality flow visualization images can be found in Supplementary Fig. 9. Figure 4. Flow visualization and heat transfer results summary of 50% TEA mass fraction experiments. (a) Phase diagram and flow visualization images. Photos on the left column are the short heater visualizations (m″ = 358 kg/m2s, Tin = 14.5 °C), long heater visualizations are on the right column (m″ = 178 kg/m2s, Tin = 12.2 °C). The system’s location on the phase diagram is approximated using the average wall temperature. (b) Average heat transfer coefficient results at m″ = 358 kg/m2s, Tin = 11.8 °C. (c) Short heater local heat transfer coefficient at m″ = 358 kg/m2s, Tin = 13.8 °C. Full quality flow visualization images can be found in Supplementary Fig. 9. ingle phase mixture and the water-rich phase, a gray value, gm, can serve as an indicator for TEA concentration n the separated phases.hih The gray value image analysis technique was first applied to the long heater. The average gray values of the middle 10% and the last 10% length of the heater were calculated (Fig. 5(b)). The sample images were extracted from 20 different frames from the flow visualization videos. The first comparison was made between different heat fluxes under the same flow rate. With increasing heat flux, the mean gray value declined, indicating the Scientific RePorTs \| (2018) 8:12093 \| DOI:10.1038/s41598-018-30584-6 6 www.nature.com/scientificreports/ Figure 5. Critical composition flow visualization analysis. (a) direct comparison of flow pattern of similar fluid outlet temperatures but different flow rates. gm,10% is the average gray value of the last 10% length of the long heater. (b) average gray value at the middle and last 10% of the long heater at m″ = 208 kg/m2s, Tin = 11.4 °C. (c) RTD local gray value at m″ = 567 kg/m2s. (d) separation layer thickness estimation at m″ = 567 kg/m2s. Figure 5. Critical composition flow visualization analysis. (a) direct comparison of flow pattern of similar fluid outlet temperatures but different flow rates. gm,10% is the average gray value of the last 10% length of the long heater. (b) average gray value at the middle and last 10% of the long heater at m″ = 208 kg/m2s, Tin = 11.4 °C. (c) RTD local gray value at m″ = 567 kg/m2s. Discussionh The current work demonstrates that phase separation of a TEA/water mixture can enhance thermal energy trans- port at the micro-scale by up to 2.5 times compared to a single phase flow of the same mixture. Significant thermal transport enhancement over water is also observed. At the critical composition, both the local and average heat transfer coefficients of the mixture increase asymptotically with heat flux. Data from the critical composition mixture flow experiments over a range of flow rates, inlet temperatures, heat fluxes, and average and local meas- urements, suggest a unified heat transfer characteristic. A rapid increase in AF occurs at low xw values (less than 0.1), demonstrating that enhanced heat transfer is mainly the result of phase separation at the vicinity of the wall, not in the fluid bulk. The mass quality analysis implies that the theoretical maximum AF is reached when xw = 1. However, due to the non-uniform temperature distribution in the flow, the maximum xw cannot be reached. The pressure drop decreases after phase separation due to a reduction in the system’s viscosity, which is an additional benefit. The water-rich (15%) composition shows some heat transfer enhancement, but not as much as the critical composition. The TEA-rich (50%) mixture yields unfavorable heat transfer performance and is not suitable for any cooling applications. Critical composition phase separation flow shows a mist flow, elongated droplet flow, and annular-like flow pattern, while off-critical compositions show a droplet/string flow pattern. It has been shown in the flow visualization analysis that both thermodynamics and hydrodynamics affect the evolution of flow pattern and the concentration field.h l pi The asymptotic thermal transport behavior is mainly caused by the behavior of the binodal curve for the TEA/water system and might vary depending on the mixture’s content. The domain morphology for forced flow deviates from the quiescent case due to the presence of inertial and shear forces. Since the flow is non-isothermal, the temperature distribution determines the concentration field. As the local temperature varies, so do the local compositions of the separated phases. On the other hand, the fluid away from the heated surface might not reach the critical temperature and remain single phase. The resulting flow consists of multiple phases with concentra- tions determined by the local temperature.lhf y p Beside the TEA/water mixture, other possible fluid systems need to be studied. www.nature.com/scientificreports/ However, comparing gm,RTD3, Tin = 12.0 °C and gm,RTD4, Tin = 15.0 °C, it can be concluded that both the residence time and the separation layer thickness affect gm,local. When TRTD = 40 °C, gm,RTD reaches an asymptotic value, which is in agreement with the shape of the binodal curve.l p Another important observation through the flow visualization images (Fig. 5(a)) is that the dark and light shades are not completely separated apart from one another. It is clearly to see that some dark shades are inside of the lighter shade domains. As the flow is three-dimensional and the camera was set to provide a top view, the observation might be due to the overlap of separated domains. Apart from this experiment-induced uncertainty, the complex flow conditions, i.e. presence of bulk advection, sharp temperature gradient, preferable wetting of one component, cause a mismatch in time response between the geometrical coarsening induced by hydrody- namic effect and mass diffusion driven by thermodynamic effect30–32. Scientific RePorTs \| (2018) 8:12093 \| DOI:10.1038/s41598-018-30584-6 7 7 www.nature.com/scientificreports/ Discussionh The different thermal/physical properties as well as phase separation characteristics can lead to different thermal transport and multi-phase flow behavior. While studying various fluid mixtures, the mixture’s physical/thermal properties need to be carefully measured. It has been proven that the critical solution temperature can be adjusted by adding miniscule amount of a third component. Therefore, the mixture’s critical temperature can be adjusted to satisfy an application’s specific needs. Once the desirable liquid/liquid mixture is configured, it can be readily applied without substan- tial modification to the cooling equipment. This makes the application of such partially miscible fluid systems promising. The physics of the process requires deeper understanding, as the time scales of the mass, momentum and thermal transport are interlinked during the process. The transient interactions of the transport phenomenon are of great interest, and more detailed studies are needed. Meanwhile, the liquid/liquid interface deserves more attention as strong chemical potential gradient presents. Methods Experimental setup for measuring heat transfer coefficient. The TEA/water solution is stored in a 316 Stainless Steel tank, which is immersed in a cooling water bath. A magnetic stir bar constantly mixes the fluid solution, which is maintained below its critical solution temperature. The solution temperature in the tank is monitored by a thermocouple. The fluid is delivered to the test section using a micro pump, and a 60-micron pore size filter is placed between the storage tank and the pump to protect the gear pump from unexpected con- taminants. Two pressure transducers are placed at the inlet and outlet of the test section. A flowmeter is placed after the test section where the fluid mixture reaches the ambient temperature (Supplementary Fig. 4). The test section is a rectangular micro channel whose cross section is 2 mm wide and 0.4 mm high. Two resistance heaters of length 10 mm and 3 mm are deposited on the channel bottom, and five RTDs are also deposited on the chan- nel bottom. The RTD layer is separated from the heater layer by a 1 µm thick layer of silicon dioxide. The bot- tom Pyrex substrate, one piece of 0.4 mm thick silicone rubber, and the top transparent Pyrex cover are pressed together to form the microchannel. The test section is enclosed in a device package to allow for fluid flow and electrical connections (Supplementary Figs 2 and 4). Heat transfer coefficient calculation. The average heat transfer coefficient calculation for critical com- position is performed according to: Heat transfer coefficient calculation. The average heat transfer coefficient calculation for critical com- osition is performed according to: = ″ − h Q T T (2) avg w avg f , (2) where Q″ is the effective heat flux after subtracting heat loss; Tw,avg is the average wall temperature measured by the heater resistance, note that the higher Tw,avg achieved in experiments at a given flow was about 80 °C; Tf is the average fluid temperature inside the channel, and it is calculated as half of the sum of fluid inlet temperature, Tin, and mean fluid outlet temperature, Tout. Tin is measured by RTD0 assuming the local wall temperature is in equi- librium with the fluid bulk. Tout is calculated via a numerical integration using energy balance and specific heat data of the mixture at critical composition (ṁ is the mass flow rate of the mixture), noting that the specific heat data (Supplementary Fig. 1(a)) takes the latent heat into consideration. ∫ =  Q m c dT (3) T T p in out (3) The local heat transfer coefficient is calculated according to: Scientific RePorTs \| (2018) 8:12093 \| DOI:10.1038/s41598-018-30584-6 8 www.nature.com/scientificreports/ = ″ − h Q T T ( local w local f local , , = ″ − h Q T T local w local f local , , (4) where Tw,local is the RTD temperature measurement, Tf,local is the mean fluid temperature at the corresponding RTD location, which is calculated by the inlet temperature and energy balance similar to Tout. For all off-critical compositions, due to the absence of specific heat data, Tout cannot be calculated. Alternatively, only the fluid inlet temperature is used in the denominators. Temperature measurements. Besides the RTDs, the heaters are also used as temperature sensors. During experiments, two Digital Multi-Meters (DMMs) measure the current and voltage across the heater, thus the heater resistance is obtained. The RTD resistance is measured by an excitation module, which provides a con- stant DC current of 100 μA and a sensing module, which measures the voltage across a RTD. The heater or RTD resistance is a function of its temperature. The resistance-temperature relation is obtained through a calibration process, which is achieved by placing the device in a temperature controlled oven and simultaneously recording the resistances and temperatures of the device. Typical resistance-temperature relations of heaters and RTDs are shown in Supplementary Fig. 5. References References 1. Steinke, M. E. & Kandlikar, S. G. Single-phase heat transfer enchancement techniques in microchannel and microchannels flows. In Microchannels and Minichannels (2004).f Microchannels and Minichannels (2004). 2. Cho, H. J., Mizerak, J. P. & Wang, E. N. Turning bubbles on and off during boiling using charged surfactants. Nat. Commun. 6, 8599 (2015). ( ) 2. Cho, H. J., Mizerak, J. P. & Wang, E. N. Turning bubbles on and off during boiling using charged surfactants. Nat. Commun. 6, 8599 (2015).l 3. Woodcock, C., Yu, X., Plawsky, J. & Peles, Y. Piranha Pin Fin (PPF) - Advanced flow boiling microstructures with low surface tension dielectric fluids. Int. J. Heat Mass Transf. 90, 591–604 (2015). l f 4. Wang, Y., Houshmand, F., Elcock, D. & Peles, Y. Convective heat transfer and mixing enhancement in a microchannel with a pillar Int. J. Heat Mass Transf. 62, 553–561 (2013).li J f , ( ) 5. Kosar, A. & Peles, Y. Convective flow of refrigerant (R-123) across a bank of micro pin fins. Int. J. Heat Mass Transf. 49, 3142–3155 (2006). Nd S L H J P l Y & J M K H f h f fi b d I J H 6. Ndao, S., Lee, H. J., Peles, Y. & Jensen, M. K. Heat transfer enhancement from micro pin fins subjected to an impinging jet. Int. J. Hea Mass Transf. 55, 413–421 (2012).lh J J pi j p g g j J Mass Transf. 55, 413–421 (2012). 7. Kandlikar, S. G. Fundamental issues related to flow boiling in minichannels and microchannels. Exp. Therm. Fluid Sci. 26, 389–407 (2002) Mass Transf. 55, 413–421 (2012). 7. Kandlikar, S. G. Fundamental issues related to flow boiling in minichannels and microchannels. Exp. Therm. Fluid Sci. 26, 389–407 (2002) f ( ) 7. Kandlikar, S. G. Fundamental issues related to flow boiling in minichannels and microchannels. Exp. Therm. Fluid Sci. 26, 389–407 (2002).l ( ) 8. Chang, K. H. & Pan, C. Two-phase flow instability for boiling in a microchannel heat sink. Int. J. Heat Mass Transf. 50, 2078–2088 (2007).l 9. Tanaka, H. & Sigehuzi, T. Spinodal decomposition of a symmetric binary fluid mixture in quasi two dimensions: Local orientationa ordering of fluid tubes. Phys. Rev. E 52 (1995). l 10. Tanaka, H. New Coarsening Mechanisms for Spinodal Decomposition Having Droplet Pattern in Binary Fluid Mixture: Collision- Induced Collisions. Phys. Rev. Lett. Heat transfer coefficient calculation. The average heat transfer coefficient calculation for critical com- position is performed according to: Mass quality calculation. At critical composition, an equivalent mass quality, which estimates the fraction f initial mixture experienced phase separation, is defined as: = − ∆  x Q Q m h T ( ) (5) w tp sp mix w = − ∆  x Q Q m h T ( ) (5) w tp sp mix w = − Q h A T T ( ) (6) sp sp heater w f = + − − yh y h h (1 ) (7) mix l mix h mix critical , , ,ll = − ∆  x Q Q m h T ( ) w tp sp mix w (5) = − Q h A T T ( ) (6) sp sp heater w f = − Q h A T T ( ) (6) sp sp heater w f (6) ∆ = + − − h T yh y h h ( ) (1 ) (7) mix w mix l mix h mix critical , , , (7) where Qtp is the two-phase flow total energy input to the fluid mixture, Qsp is the single phase flow total energy input based on single phase heat transfer coefficient at the same temperature gradient as in the two-phase flow; y is the mass percentage of the lighter phase in the separated mixture of the two phases, and is calculated via the lever rule using the wall temperature; Δhmix is the difference in heat of mixing according to the wall temperature. where Qtp is the two-phase flow total energy input to the fluid mixture, Qsp is the single phase flow total energy input based on single phase heat transfer coefficient at the same temperature gradient as in the two-phase flow; y is the mass percentage of the lighter phase in the separated mixture of the two phases, and is calculated via the lever rule using the wall temperature; Δhmix is the difference in heat of mixing according to the wall temperature. Flow visualization. The flow visualization videos are taken using a high-speed camera (phantom vision research v4.2) and a microscope (Leica TYPE020). The mixture preparation. The mixture of a certain concentration is prepared using mass scale of accuracy of 0.1 g. The mixture preparation. The mixture of a certain concentration is prepared using mass scale of accuracy of 0.1 g. Data availability. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Data availability. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. References Phys. Fluids 17, 34107 (2005).f 24. Lamorgese, A. G. & Mauri, R. Diffuse-interface modeling of phase segregation in liquid mixtures. Int. J. Multiph. Flow 34, 9 (2008). 25. Mauri, R., Califano, F., Calvi, E., Gupta, R. & Shinnar, R. Convection-driven phase segregation of deeply quenched liquid m J. Chem. Phys. 118, 8841 (2003). y 26. Leptoukh, G. & Strickland, B. Phase separation in Two-Dimensional Fluid Mixtures. Phys. Rev. Lett. 74, 18–21 (1995). 27. Vladimirova, N., Malagoli, A. & Mauri, R. Two-dimensional model of phase segregation in liquid binary mixtures. Phys. Rev Phys. Plasmas. Fluids. Relat. Interdiscip. Topics 60, 6968–77 (1999).ll 28. Hashimoto, T., Matsuzaka, K., Moses, E. & Onuki, A. String phase in phase-separating fluids under shear flow. Phys. Rev. Le 126–129 (1995). ( ) 29. Berthier, L., Barrat, J. & Kurchan, J. Phase Separation in a Chaotic Flow. Phys. Rev. Lett. 86, 3–6 (2001). 30. Tanaka, H. & Araki, T. Spontaneous double phase separation induced by rapid hydrodynamic coarsening in two-dimensional mixtures. Phys. Rev. Lett. 81, 389–392 (1998).l y ( ) 1. Araki, T. & Tanaka, H. Hydrodynamic delocalization of phase separation in a locally cooled fluid mixture. Europhys. Lett. 65 214–220 (2004).l 2. Tanaka, H. Interplay between wetting and phase separation in binary fluid mixtures: Roles of hydrodynamics. J. Phys. Condens Matter 13, 4637–4674 (2001). 32. Tanaka, H. Interplay between wetting and phase separation in binary fluid mixtures: Roles of hydrodynamics. J. Phys. Condens. Matter 13, 4637–4674 (2001). Author Contributions W.X. designed and performed the experiments, and analyzed the data. A.U., N.B., J.P. and Y.P. supervised the research. W.X. and Y.P. wrote the paper. Acknowledgementsh g This study was supported through generous funds provided by the Israel Ministry of Defense (IMOD) under contract number 4440715854 and the Office of Naval Research (ONR) under grants number N00014-15-1-2071 and N62909-15-1-2032. The authors also thank Dr. Yingying Wang, Dr. Xiangfei Yu, Dr. Corey Woodcock, Dr. Ashwin Vutha and Dr. Sameer Rao for insightful discussions and suggestions. References 72 (1994).f y ( ) 11. Anders, D. & Weinberg, K. Numerical simulation of diffusion induced phase separation and coarsening in binary alloys. Comput. Mater. Sci. 50, 1359–1364 (2011). ( ) 2. Santonicola, G., Mauri, R. & Shinnar, R. Phase Separation of Initially Inhomogeneous Liquid Mixtures. Ind. Eng. Chem. Res. 40 2004–2010 (2001). 13. Poesio, P., Lezzi, A. & Beretta, G. Evidence of convective heat transfer enhancement induced by spinodal decomposition. Phys. Rev. E 75, 66306 (2007). 14. Molin, D. & Mauri, R. Enhanced heat transport during phase separation of liquid binary mixtures, 1–11, https://doi. org/10.1063/1.2749810 (2007). 9 Scientific RePorTs \| (2018) 8:12093 \| DOI:10.1038/s41598-018-30584-6 9 www.nature.com/scientificreports/ 5. Di Fede, F., Poesio, P. & Beretta, G. P. Heat transfer enhancement in a small pipe by spinodal decomposition of a low viscosity, liquid liquid, strongly non-regular mixture. Int. J. Heat Mass Transf. 55, 897–906 (2012). q g y g f 6. Gat, S., Brauner, N. & Ullmann, A. Heat transfer enhancement via liquid-liquid phase separation. Int. J. Heat Mass Transf. 52 1385–1399 (2009). ( ) 17. Ullmann, A., Lipstein, I. & Brauner, N. Applying phase separation of a solvent system with a lower critical solution temperature for enhancement of cooling rates by forced and free convection. Proc. 15th Int. Heat Transf. Conf., 1–15 (2014). g y Liquid/liquid phase separation heat transfer at the microscale. Int. g q q p p f 19. Kohler, F. & Rice, O. K. Coexistence curve of the triethylamine-water system. J. Chem. Phys. 26, 1614–1618 (1957). y y y 0. Bertrand, G. L., Larson, J. W., Hepler, L. G., Larson, J. W. & Heplerlb, L. G. Thermochemical Investigations of the Water Triethylamine System. J. Phys. Chem. 72, 4194–4197 (1968).hh Triethylamine System. J. Phys. Chem. 72, 4194–4197 (1968). y y y 1. Chand, A., Handa, Y. P. & Fenby, D. Excess volumes of triethylamine + chloroform at 298.15 and 308.15 K The. J. Chem. Thermodyn 7, 401–402 (1975). 1. Chand, A., Handa, Y. P. & Fenby, D. Excess volumes of triethylamine + chloroform at 298.15 and 308.15 K The. J. Chem. Thermodyn 7, 401–402 (1975). 2. Kartzmark, E. M. System triethylamine–water: the equilibrium diagram and some physical properties. Can. J. Chem. 45, 1089–1091 2. Kartzmark, E. M. System triethylamine–water: the equilibrium diagram and some physical properties. Can. J. Chem. 45, 1089–109 (1967). ( ) 23. Lamorgese, A. G. & Mauri, R. Nucleation and spinodal decomposition of liquid mixtures. Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-018-30584-6. Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-018-30584- Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-018- Competing Interests: The authors declare no competing interests. Competing Interests: The authors declare no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps an institutional affiliations. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, 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 Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, 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 Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2018 Scientific RePorTs \| (2018) 8:12093 \| DOI:10.1038/s41598-018-30584-6 10
https://openalex.org/W2096207854	https://stemcellres.biomedcentral.com/counter/pdf/10.1186/scrt371	English	null	Design considerations for an integrated microphysiological muscle tissue for drug and tissue toxicity testing	Stem cell research & therapy	2,013	cc-by	4,697	Introduction In vitro microphysiological multiorgan systems using human cells provide a novel method to identify promising drug candidates [1]. Skeletal muscle is an important component of such systems because skeletal muscle Abstract Microphysiological systems provide a tool to simulate normal and pathological function of organs for prolonged periods. These systems must incorporate the key functions of the individual organs and enable interactions among the corresponding microphysiological units. The relative size of diff erent microphysiological organs and their fl ow rates are scaled in proportion to in vivo values. We have developed a microphysiological three-dimensional engineered human skeletal muscle system connected to a circulatory system that consists of a tissue- engineered blood vessel as part of a high-pressure arterial system. The engineered human skeletal muscle tissue reproduces key mechanical behaviors of skeletal muscle in vivo. Pulsatile fl ow is produced using a novel computer-controlled magnetically activated ferrogel. The system is versatile and the muscle unit can be integrated with other organ systems. Periodic monitoring of biomechanical function provides a non- invasive assessment of the health of the tissue and a way to measure the response to drugs and toxins. Sarcopenia, the age-associated loss of skeletal muscle mass, is a major consequence of aging and leads to lower metabolic adaptation, slowed immunological responses to disease, weakness, poor function and disability with advancing age [3]. Th e onset and progression of sarco- penia is a strong predictor of mortality and prevalence increases with advancing age [4]. Type 2 diabetes or low- grade inﬂ ammation is more common in individuals who have sarcopenia [5]. Other major diseases that involve muscle include muscular dystrophy and neurodegenera- tive diseases. Skeletal muscle is both a potential target for drug therapy and a site of drug toxicity. Treatments for type 2 diabetes aﬀ ect and even target skeletal muscle metabo- lism as therapeutic mechanisms. Several candidate drugs are under investigation to treat sarcopenia, although special considerations are needed to deﬁ ne eﬃ cacy [6] and to ensure that side eﬀ ects are limited [7]. Keywords: microphysiological system, skeletal muscle myoblasts, tissue-engineered blood vessel, contractile force, oxygen gradients, diff erentiation Mitochondrial toxicity is a major reason why a drug may fail in clinical trials or is recalled after approval [8]. Like the liver and heart, skeletal muscle is aﬀ ected by mitochondrial drug toxicity. Certain statins, used to treat dyslipidemia, can cause severe myopathy. Drugs known to cause mitochondrial toxicity, such as cerivastatin [9] and thiazolidinediones [8] (used to treat type 2 diabetes), also aﬀ ect muscle function. Correspondence: gtruskey@duke.edu 1Department of Biomedical Engineering, Duke University, 136 Hudson Hall, CB 90281, Durham, NC 27708-0281, USA Full list of author information is available at the end of the article Truskey et al. Stem Cell Research & Therapy 2013, 4(Suppl 1):S10 http://stemcellres.com/content/4/S1/S10 Design considerations for an integrated microphysiological muscle tissue for drug and tissue toxicity testing George A Truskey1, Hardean E Achneck2,3, Nenad Bursac1, Hon Fai Chan1, Cindy S Cheng1, Cristina Fernandez1, Sungmin Hong4, Youngmee Jung1, Tim Koves5, William E Kraus5, Kam Leong1, Lauran Madden1, William M Reichert1 and Xuanhe Zhao4 accounts for 40% of the body’s mass and about 30% of the resting metabolic rate [2]. Under resting conditions about 20% of cardiac output passes through muscle, and during heavy exercise muscle requires as much as 80% of cardiac output. Muscle is responsible for about 75% of whole- body insulin-stimulated glucose uptake. Engineered three-dimensional human skeletal muscle tissue Th e two engineered tissues are inte grated into a perfusion system that enables monitoring of function. Major challenges for in vitro culture of human skeletal muscle are the lack of eﬃ cient methods to diﬀ erentiate large numbers of human muscle cells from induced pluripotent stem cells and the slower doubling time and rate of fusion for primary human skeletal muscle cells compared with mouse or rat myoblasts. Further, rat and mouse myotubes undergo spontaneous contractions in two-dimensional and three-dimensional cultures, where- as contractile human muscle ﬁ bers in vitro have only been obtained following co-culture with rat or human motoneurons [11]. To further enhance diﬀ erentiation, normal contractile function of skeletal muscle under various physical demands can be simulated by applying cyclic stretch or electrical stimulation (Figure 1, inset). Electrical stimula- tion is applied via two platinum electrodes and can be synchronized with the mechanical stimulation [14]. Th e electrical stimulation waveform is precisely bipolar and the amplitude of electrical stimulation is adjusted to accommodate potential changes in the excitation threshold expected to occur with the muscle maturation. We utilized our prior ﬁ ndings with engineered rodent muscles [12-18] to fabricate three-dimensional primary human skeletal muscle tissue bundles. Optimization of the dimensions of the muscle bundles, the media and hydrogel composition, the cell density, and the diﬀ eren- tiation procedure (Figure 1) resulted in three-dimensional human muscle bundles with highly aligned, cross-striated myoﬁ bers containing myogenin-positive nuclei and acetylcholine receptor clusters (Figure 2b1). Functional properties of such engineered muscle were tested using standard force test protocols [17]. In response to single electrical stimuli, the bundles produced twitch contrac- tions; at higher stimu lation frequencies, the twitch responses fused into a more forceful tetanic contraction. Amplitudes of twitch and tetanus-speciﬁ c force were signiﬁ cantly lower than those of native muscle, probably due to a lower muscle ﬁ ber density and a smaller ﬁ ber diameter [19], as well as the presence of immature forms of muscle proteins. Similar to native muscle, engineered human muscle bundles exhibited a positive Starling-like force–length relation ship. After the necessary optimization, the method for engineering functional human muscle tissues is robust and reproducible and tissue bundles remain functional for 4 weeks. Engineered three-dimensional human skeletal muscle tissue *Correspondence: gtruskey@duke.edu 1Department of Biomedical Engineering, Duke University, 136 Hudson Hall, CB 90281, Durham, NC 27708-0281, USA Full list of author information is available at the end of the article To evaluate functional changes to human skeletal muscle after exposure to drugs, we developed a micro physio- logical system consisting of highly aligned, functional, and metabolically active engineered human muscle and © 2010 BioMed Central Ltd © 2013 BioMed Central Ltd Truskey et al. Stem Cell Research & Therapy 2013, 4(Suppl 1):S10 http://stemcellres.com/content/4/S1/S10 Truskey et al. Stem Cell Research & Therapy 2013, 4(Suppl 1):S10 http://stemcellres.com/content/4/S1/S10 Page 2 of 5 develop mental processes: miR-133 a blocks myoblast prolifera tion by inhibiting serum response factor [20], and miR-696 inhibits mitochondrial biogenesis and oxidative metabolism by blocking the metabolic transcriptional co-activator PGC-1α [21]. Engineered human skeletal muscle myoblasts were transfected with anti-miR-133a, with anti-miR-696, or with both anti- miRNAs. Th e miRNA transfection produced longer and more aligned and cross-striated myoﬁ bers when compared with myoblasts that received a scrambled RNA sequence. Myoblasts transfected with both anti-miR- 133a and anti-miR-696 exhibited considerable slow myosin heavy chain, indicative of type I muscle ﬁ bers, and generated higher speciﬁ c contractile forces com- pared with bundles pre pared from myoblasts transfected with only a single anti-miRNA. human tissue-engineered blood vessel (TEBV) (Figure 1). Myoblasts were obtained from biopsies of the vastus lateralis of healthy middle-aged volunteers, endothelial cells were cultured from blood-derived late outgrowth endothelial progenitor cells in human umbilical cord blood [10], and vessel wall medial cells were either human dermal ﬁ broblasts or mesenchymal stem cells. Th ese individual units are cultured separately for about 2 weeks to enable the skeletal myoblasts to fuse and diﬀ erentiate and to allow the blood vessels to develop suﬃ cient mechanical strength. Th e two engineered tissues are inte grated into a perfusion system that enables monitoring of function. human tissue-engineered blood vessel (TEBV) (Figure 1). Myoblasts were obtained from biopsies of the vastus lateralis of healthy middle-aged volunteers, endothelial cells were cultured from blood-derived late outgrowth endothelial progenitor cells in human umbilical cord blood [10], and vessel wall medial cells were either human dermal ﬁ broblasts or mesenchymal stem cells. Th ese individual units are cultured separately for about 2 weeks to enable the skeletal myoblasts to fuse and diﬀ erentiate and to allow the blood vessels to develop suﬃ cient mechanical strength. Designing a microperfusion system for engineered skeletal muscle Engineered human muscle systems are used either as a dynamically conditioned standalone culture system to enable muscle maturation or as an integrated module with other microphysiological units to examine metabolic and drug interactions (Figure 2). Th e microphysiological system includes a tissue-engineered blood vessel, 750 to 1,000 μm in diameter, in parallel with the muscle tissue. Th e blood vessel consists of a conﬂ uent layer of human endothelial cells and a contractile medial layer of human dermal ﬁ broblasts. Th e medial layer of the TEBV was fabricated either: from aligned human mesenchymal stem cell sheets rolled into a tubular structure, cultured in a rotating wall bioreactor, followed by maturation in a perfusion bioreactor; or by preparing a tubular structure of dense collagen containing human neonatal dermal ﬁ broblasts that is then cultured in a perfusion chamber for 1 week [22]. TEBV mechanical strength is enhanced using oscillating pressure and ﬂ ow, similar to approaches used for larger diameter vessels [23-25]. Physiological pulsatile ﬂ ow is generated using a novel magnetoactive porous ferrogel [26,27] that acts as a valve, changing its hydraulic conductivity under applied magnetic ﬁ elds to tune the ﬂ ow rate in the microphysiological system on demand. Other microphysiological organs (for example, myo cardium, liver) can be added in parallel to the muscle (Figure 2a). To promote diﬀ erentiation and the production of adult forms of muscle contractile proteins, and thereby increase contractile force [20], we have transfected myo- blasts with miRNAs prior to formation of the engineered human muscle bundles. miR-133a and miR-696 are of particular interest because they aﬀ ect important Truskey et al. Stem Cell Research & Therapy 2013, 4(Suppl 1):S10 http://stemcellres.com/content/4/S1/S10 Truskey et al. Stem Cell Research & Therapy 2013, 4(Suppl 1):S10 http://stemcellres.com/content/4/S1/S10 Page 3 of 5 Figure 1. Schematic of the regimen for optimization and validation of engineered three-dimensional human skeletal muscle cultures. 3D, three-dimensional; ECs, endothelial cells. Designing a microperfusion system for engineered skeletal muscle Cell Sourcing Muscle - Optimize isolation, culture and differentiation of primary cells Endothelium - Optimize co-culture medium composition - Alter cell density, construct geometry, biomaterials - Transient expression of microRNAs - Test force parameters, Ca2+ handling, cell and fiber morphology 3D Human Muscle 3-D Test Bed - Vary flow-rate, cell density - Optimize muscle function and viability during four weeks of perfusion - Design for desired flowrates - Real-time imaging - Mechanical - Electrical Perfusion System Stimulation Camera Sampling Port Anchor Elastic Anchor + _ Electrodes ECs muscle Electromagnet - Alter cell density, construct geometry, biomaterials - Transient expression of microRNAs - Test force parameters, Ca2+ handling, cell and fiber morphology 3D Human Muscle Camera Sampling Port Anchor Elastic Anchor + _ Electrodes ECs muscle Electromagnet Sampling Port Elastic Anchor Camera ng Port Elastic Anchor Elastic Anchor igure 1. Schematic of the regimen for optimization and validation of engineered three-dimensional human skeletal D, three-dimensional; ECs, endothelial cells. Functional evaluation in the system involves monitor- ing the mechanical behavior of engineered muscle and blood vessels. Th e diameter of the engineered blood vessel is measured in response to changes in pressure, from which the incremental elastic modulus and the ultimate tensile strength prior to failure are determined [24]. Vessel dilation in response to changes in ﬂ ow and nitric oxide release are used to monitor the function of the endothelium. For the muscle tissue, oxygen uptake and the contractile force provide an assessment of muscle function.h dissolved in culture medium, the ﬂ ow rate to muscle tissue under resting conditions needs to be at least 2.54 μl/second. Since 21% of the cardiac output ﬂ ows through skeletal muscle, the total ﬂ ow rate in the system is 12 μl/second. Adjusting the culture medium viscosity to the blood viscosity, the time-averaged shear stress acting on the endothelium in the TEBV is 0.43 Pa for a 1,000 μm diameter vessel and is 0.84 Pa for an 800 μm diameter vessel, well within the range of values reported in vivo. Th e ﬂ ow is laminar, with Reynolds number around 5 for the TEBV, and is quasi-static, similar to conditions in arterioles [30]. Designing a microperfusion system for engineered skeletal muscle Th e design of the system with human cells requires that the shear stress on endothelium in the blood vessel ranges from 0.4 to 2.0 Pa [28]; that the rate of oxygen delivery to the muscle equals or exceeds the rate of oxygen uptake by muscle; and that materials used should not bind drugs. Additional requirements for multiorgan systems are that the relative size of diﬀ erent micro physio- logical organs and their ﬂ ow rates should be scaled in proportion to in vivo values [1]; that a common media is used for all of the diﬀ erent engineered tissues; and that the system must operate for at least 4 weeks. Since skeletal muscle is more abundant than other tissues included in an integrated microphysiological system, this ﬂ ow rate should meet metabolic demands of other tissues – including the liver, which has a high metabolic rate but has 20 times less mass than muscle. Control of ﬂ ow distribution with magnetoactive valves allows simple interfacing with lower-ﬂ ow microﬂ uidic platforms and permits adjustment of the ﬂ ow rate distribution in response to diﬀ erent physiological stimuli. Continuous monitoring of the oxygen concentration in the perfusate combined with a feedback loop can serve to tune ﬂ ow rate in a physiological-like fashion. Under resting conditions, oxygen uptake in humans in vivo is approximately 1 × 10–8 moles O2/second/cm3 muscle tissue and can increase 50-fold during exercise [29], similar to our estimate for the oxygen uptake levels of murine skeletal muscle ﬁ bers in three-dimensional cultures [13]. Consequently, we assume that the in vivo uptake rates are representative of muscle ﬁ bers in vitro. For a muscle cell density of 1 × 108 cells/ml and oxygen Validation and testing of microphysiological systems Validation of the microphysiological system involves measuring vessel relaxation and constriction, skeletal muscle contractile force and metabolism over a 4-week Truskey et al. Stem Cell Research & Therapy 2013, 4(Suppl 1):S10 http://stemcellres.com/content/4/S1/S10 Page 4 of 5 pump Lung Muscle Tissue Arterial TEBV Q QM Tissue 1 Tissue 2 Tissue 3 QT1 QT2 QT3 Ferrogel valve (a) Figure 2. Microphysiological system with engineered human muscle. (a) Schematic of the perfusion system showing the relation among the ferrogel pump, human tissue-engineered blood vessel (TEBV), muscle tissue and other microphysical tissue or organ systems. Q, fl ow rate; QM, fl ow rate to engineered muscle; QT, fl ow rate to other microphysiological tissues. (b1) Aligned three-dimensional human skeletal muscle tissue showing actin striations indicative of sarcomeres. (b2) Human endothelial cells covering a monolayer of diff erentiated human skeletal muscle in two- dimensional cultures. (c1) Schematic of fl uidic magnetically activated ferrogel valve. (c2) Eff ect of magnetic fi eld on magnetically activated ferrogel. (c3) Velocity of microparticles at 1 Hz and 5 μl/minute in fl uidic channel. pump Lung Muscle Tissue Arterial TEBV Q QM Tissue 1 Tissue 2 Tissue 3 QT1 QT2 QT3 Ferrogel valve 125 μm OFF ON (a) (b1) (b2) 25 μm Time, s Velocity, μm/min 0 1 2 3 0 125 250 (c1) (c2) (c3) 125 μm (b2) (b1) 25 μm (a) (b1) (b2) Time, s Velocity, μm/min 0 1 2 3 0 125 250 (c3) g (c1) (c1) g OFF ON c1) (c2) OFF ON Figure 2. Microphysiological system with engineered human muscle. (a) Schematic of the perfusion system showing the relation among the ferrogel pump, human tissue-engineered blood vessel (TEBV), muscle tissue and other microphysical tissue or organ systems. Q, fl ow rate; QM, fl ow rate to engineered muscle; QT, fl ow rate to other microphysiological tissues. (b1) Aligned three-dimensional human skeletal muscle tissue showing actin striations indicative of sarcomeres. (b2) Human endothelial cells covering a monolayer of diff erentiated human skeletal muscle in two- dimensional cultures. (c1) Schematic of fl uidic magnetically activated ferrogel valve. (c2) Eff ect of magnetic fi eld on magnetically activated ferrogel. (c3) Velocity of microparticles at 1 Hz and 5 μl/minute in fl uidic channel. period. Th e robustness of the system can be assessed by evaluating perturbations from normal physiology, including acute simulated exercise and acute exposure to TNFα. Authors’ contributions GAT designed experiments, discussed interpretation of the results, and wrote the paper. HEA analyzed data, suggested experiments and edited the paper. NB designed and analyzed experiments, wrote portions of the paper, and edited the paper. HFC performed experiments and analyzed data. CSC designed and performed experiments and analyzed data. CF performed experiments and analyzed data. SH designed and performed experiments and analyzed data. YJ designed and performed experiments and analyzed data. TK analyzed results and edited paper. WEK analyzed results and edited paper. KL designed and analyzed experiments, wrote portions of the paper, and edited the paper. LM designed and analyzed experiments. WMR designed and analyzed experiments, wrote portions of the paper, and edited the paper. XZ designed and analyzed experiments, wrote portions of the paper, and edited the paper. All authors read and approved the fi nal manuscript. In summary, we have developed the components of a microphysiological system that uses functional measures of blood vessel and skeletal muscle to assess the eﬀ ect of drugs and toxins. By adjusting cell numbers and ﬂ ow rates, the system is ﬂ exible enough to integrate with other Validation and testing of microphysiological systems microﬂ uidic and perfusion systems to examine the response of a number of organs and tissues to drugs. Abbreviations Abbreviations miRNA, microRNA; TEBV, tissue-engineered blood vessel; TNF, tumor necrosis factor. Once the function of the tissues in a microphysiological system is validated, a set of candidate drugs known to aﬀ ect skeletal muscle and blood vessels, as well as other tissues, can be tested. Compounds to be examined include those that test normal function (phenylephrine), drugs that are safe and eﬀ ective (lovastatin and met- formin [31]), drugs that are eﬀ ective and unsafe (ceriva- statin [32,33]), and compounds with known toxicity to skeletal muscle or blood vessels (for example, antimycin [34], rotenone [34] and doxorubicin [35,36]). Initial results indicate that, based upon contractile force genera- tion, engineered human skeletal muscle bundles are more sensitive to doxorubicin than rodent skeletal muscle ﬁ bers [35]. Competing Interests The authors declare that they have no competing interests. Author details 22. Ghezzi CE, Risse PA, Marelli B, Muja N, Barralet JE, Martin JG, Nazhat SN: An airway smooth muscle cell niche under physiological pulsatile fl ow culture using a tubular dense collagen construct. Biomaterials 2013, 34:1954-1966. 1Department of Biomedical Engineering, Duke University, 136 Hudson Hall, CB 90281, Durham, NC 27708-0281, USA. 2Department of Surgery, Duke University Medical Center, Durham, NC 27710, USA. 3Department of Pathology, Duke University Medical Center, Durham, NC 27710, USA. 4Department of Mechanical Engineering and Materials Science, Duke University, Durham, NC 27708-0287, USA. 5Department of Medicine, Duke University Medical Center, Durham, NC 27710, USA. 23. Kumar VA, Brewste r LP, Caves JM, Chaikof EL: Tissue engineering of blood vessels: functional requirements, progress, and future challenges. Cardiovasc Eng Technol 2011, 2:137-148. 24. Syedain ZH, Meier LA, Bjork JW, Lee A, Tranquillo RT: Implantable arterial grafts from human fi broblasts and fi brin using a multi-graft pulsed fl ow-stretch bioreactor with noninvasive strength monitoring. Biomaterials 2011, 32:714-722. References 1. Esch MB, King TL, Shuler ML: The ro le of body-on-a-chip devices in drug and toxicity studies. Ann Rev Biomed Eng 2011, 13:55-72. 1. Esch MB, King TL, Shuler ML: The ro le of body-on-a-chip devices in drug and toxicity studies. Ann Rev Biomed Eng 2011, 13:55-72. 2. Zurlo F, Larson K, Bogardus C, Ravu ssin E: Skeletal muscle metabolism is a major determinant of resting energy expenditure. J Clin Invest 1990, 86:1423-1427. 2. Zurlo F, Larson K, Bogardus C, Ravu ssin E: Skeletal muscle metabolism is a major determinant of resting energy expenditure. J Clin Invest 1990, 86:1423-1427. 26. Zang J, Ryu J, Pu gno N, Wang Q, Tu Q, Buehler MJ, Zhao X: Multifunctionality and control of the crumpling and unfolding of large-area graphene. Nat Mater 2012, 12:321-325. 3. Evans WJ, Paolisso G, Abbatecola AM , Corsonello A, Bustacchini S, Strollo F, Lattanzio F: Frailty and muscle metabolism dysregulation in the elderly. Biogerontology 2010, 11:527-536. 3. Evans WJ, Paolisso G, Abbatecola AM , Corsonello A, Bustacchini S, Strollo F, Lattanzio F: Frailty and muscle metabolism dysregulation in the elderly. Biogerontology 2010, 11:527-536. 27. Zhao XH, Kim J, C ezar CA, Huebsch N, Lee K, Bouhadir K, Mooney DJ: Active scaff olds for on-demand drug and cell delivery. Proc Natl Acad Sci U S A 2011, 108:67-72. 4. Iannuzzi-Sucich M, Prestwood KM, Ke nny AM: Prevalence of sarcopenia and predictors of skeletal muscle mass in healthy, older men and women. J Gerontol A Biol Sci Med Sci 2002, 57:M772-M777. 28. Chiu JJ, Chien S: Eff ects of disturbed fl ow on vascular endothelium: pathophysiological basis and clinical perspectives. Physiol Rev 2011, 91:327-387. 29. Blomstrand E, Rad egran G, Saltin B: Maximum rate of oxygen uptake by human skeletal muscle in relation to maximal activities of enzymes in the Krebs cycle. J Physiol 1997, 501:455-460. p g q 6. Brass EP, Sietsema KE: Consideratio ns in the development of drugs to treat sarcopenia. J Am Geriatr Soc 2011, 59:530-535. 30. Truskey GA, Yuan F, Katz DF: Transport Phenomenon in Biological Systems. 2nd edition. Upper Saddle River, NJ: Pearson/Prentice Hall; 2009. 7. Evans WJ: Drug discovery and develo pment for ageing: opportunities and challenges. Philos Trans R Soc B Biol Sci 2011, 366:113-119. 31. Sarabia V, Lam L, Burdett E, Leiter LA, Klip A: Glucose transport in human skeletal muscle cells in culture. Stimulation by insulin and metformin. Published: 20 December 2013 25. Zaucha MT, Raykin J, Wan W, Gauvin R, Auger FA, Germain L, Michaels TE, Gleason RL: A novel cylindrical biaxial computer-controlled bioreactor and biomechanical testing device for vascular tissue engineering. Tissue Eng Part A 2009, 15:3331-3340. Acknowledgements The work was supported by grants R01AR055226 (to NB), R21AR055195 (to GAT) and UH2TR000505 and the National Institutes of Health Common Fund for the Microphysiological Systems Initiative and Darwin Prockop (NIH grant P40 RR017447) for providing the adult human mesenchymal stem cells. The publication costs of this article were funded by UH2TR000505. Page 5 of 5 Truskey et al. Stem Cell Research & Therapy 2013, 4(Suppl 1):S10 http://stemcellres.com/content/4/S1/S10 Truskey et al. Stem Cell Research & Therapy 2013, 4(Suppl 1):S10 http://stemcellres.com/content/4/S1/S10 Declarations 18. Bian W, Juhas M, Pfeiler TW, Bursac N: Local tissue geometry determines contractile force generation of engineered muscle networks. Tissue Eng Part A 2012, 18:957-967. Publication of this supplement has not been supported by sponsorship. Articles have undergone the journal’s standard review process. The Editors declare that they have no competing interests. 19. Trappe S, Williamson D, Godard M, P orter D, Rowden G, Costill D: Eff ect of resistance training on single muscle fi ber contractile function in older men. J Appl Physiol 2000, 89:143-152. This article has been published as part of Stem Cell Research & Therapy Volume 4 Supplement 1, 2013: Stem cells on bioengineered microphysiological platforms for disease modeling and drug testing. The full contents of the supplement are available online at http://www.stemcellres.com/supplements/4/S1. pp y 20. Rhim C, Cheng CS, Kraus WE, Truskey GA: Eff ect of microRNA modulation on bioartifi cial muscle function. Tissue Eng Part A 2010, 16:3589-3597. 21. Aoi W, Naito Y, Mizushima K, Takana mi Y, Kawai Y, Ichikawa H, Yoshikawa T: The microRNA miR-696 regulates PGC-1α in mouse skeletal muscle in response to physical activity. Am J Physiol Endocrinol Metab 2010, 298:E799-E806. to physical activity. Am J Physiol Endocrinol Metab 2010, 298:E799-E806 References J Clin Invest 1992, 90:1386-1395. 8. Dykens JA, Will Y: The signifi cance of mitochondrial toxicity testing in drug development. Drug Discov Today 2007, 12:777-785. 9. Torok M, Zahno A, Waldhauser KM, Br echt K, S. Kraahenbuhl S: Toxicity of statins on rat skeletal muscle mitochondria. Cell Mol Life Sci 2006, 63:2415-2425. 32. Vandenburgh H, Sh ansky J, Benesch-Lee F, Barbata V, Reid J, Thorrez L, Valentini R, Crawford G: Drug-screening platform based on the contractility of tissue- engineered muscle. Muscle Nerve 2008, 37:438-447. 10. Brown MA, Wallace CS, Angelos M, T ruskey GA: Characterization of umbilical cord blood-derived late-outgrowth endothelial progenitor cells exposed to laminar shear stress. Tissue Eng Part A 2009, 15:3575-3587. 10. Brown MA, Wallace CS, Angelos M, T ruskey GA: Characterization of umbilical cord blood-derived late-outgrowth endothelial progenitor cells exposed to laminar shear stress. Tissue Eng Part A 2009, 15:3575-3587. 33. Kaufmann P, Török M, Zahno A, Waldhauser KM, Brecht K, Krahenbuhl S: Toxicity of statins on rat skeletal muscle mitochondria. Cell Mol Life Sci 2006, 63:2415-2425. 33. Kaufmann P, Török M, Zahno A, Waldhauser KM, Brecht K, Krahenbuhl S: Toxicity of statins on rat skeletal muscle mitochondria. Cell Mol Life Sci 2006, 63:2415-2425. 11. Guo X, Gonzalez M, Stancescua M, V andenburgh HH, Hickman JJ: Neuromuscular junction formation between human stem cell-derived motoneurons and human skeletal muscle in a defi ned system. Biomaterials 2011, 32:9602-9611.f 11. Guo X, Gonzalez M, Stancescua M, V andenburgh HH, Hickman JJ: Neuromuscular junction formation between human stem cell-derived motoneurons and human skeletal muscle in a defi ned system. Biomaterials 2011, 32:9602-9611. 34. Shaham O, Slatea NG , Goldbergera O, Xud Q, Ramanathanb A, Souzab AL, Clishb CB, Simsa KB, Moothaa VK: A plasma signature of human mitochondrial disease revealed through metabolic profi ling of spent media from cultured muscle cells. Proc Natl Acad Sci U S A 2010, 107:1571-1575. 34. Shaham O, Slatea NG , Goldbergera O, Xud Q, Ramanathanb A, Souzab AL, Clishb CB, Simsa KB, Moothaa VK: A plasma signature of human mitochondrial disease revealed through metabolic profi ling of spent media from cultured muscle cells. Proc Natl Acad Sci U S A 2010, 107:1571-1575. 12. Zhang SJ, Truskey GA, Kraus WE: Eff ect of cyclic stretch on β1D-integrin expression and activation of FAK and RhoA. Am J Physiol Cell Physiol 2007, 292:C2057-C2069. 12. doi:10.1186/scrt371 Cite this article as: Truskey GA, et al.: Design considerations for an integrated microphysiological muscle tissue for drug and tissue toxicity testing. Stem Cell Research & Therapy 2013, 4(Suppl 1):S10. References Zhang SJ, Truskey GA, Kraus WE: Eff ect of cyclic stretch on β1D-integrin expression and activation of FAK and RhoA. Am J Physiol Cell Physiol 2007, 292:C2057-C2069. 35. van Norren K, van H elvoort A, Argiles JM, van Tuijl S, Arts K, Gorselink M, Laviano A, Kegler D, Haagsman HP, van der Beek EM: Direct eff ects of doxorubicin on skeletal muscle contribute to fatigue. Br J Cancer 2009, 100:311-314. 35. van Norren K, van H elvoort A, Argiles JM, van Tuijl S, Arts K, Gorselink M, Laviano A, Kegler D, Haagsman HP, van der Beek EM: Direct eff ects of doxorubicin on skeletal muscle contribute to fatigue. Br J Cancer 2009, 100:311-314. 13. Rhim C, Lowell DA, Reedy MC, Slentz DH, Zhang SJ, Kraus WE, Truskey GA: Morphology and ultrastructure of diff erentiating three-dimensional mammalian skeletal muscle in a collagen gel. Muscle Nerve 2007, 36:71-80.f 13. Rhim C, Lowell DA, Reedy MC, Slentz DH, Zhang SJ, Kraus WE, Truskey GA: Morphology and ultrastructure of diff erentiating three-dimensional mammalian skeletal muscle in a collagen gel. Muscle Nerve 2007, 36:71-80.f 36. Gilliam LA, St Clai r DK: Chemotherapy-induced weakness and fatigue in skeletal muscle: the role of oxidative stress. Antioxid Redox Signal 2011, 15:2543-2563. 14. Liao IC, Liu JB, Bursac N, Leong KW : Eff ect of electromechanical stimulation on the maturation of myotubes on aligned electrospun fi bers. Cell Mol Bioeng 2008, 1:133-145. g 15. Bian W, Bursac N: Engineered skelet al muscle tissue networks with controllable architecture. Biomaterials 2009, 30:1401-1412. 16. Bian W, Bursac N: Soluble miniagrin enhances contractile function of engineered skeletal muscle. Faseb J 2011, 26:955-965. 17. Hinds S, Bian W, Dennis RG, Bursac N: The role of extracellular matrix composition in structure and function of bioengineered skeletal muscle. Biomaterials 2011, 32:3575-3583. 17. Hinds S, Bian W, Dennis RG, Bursac N: The role of extracellular matrix composition in structure and function of bioengineered skeletal muscle. Biomaterials 2011, 32:3575-3583.
https://openalex.org/W2606287847	https://www.ajol.info/index.php/wsa/article/download/155171/144787	English	null	Determinant of farmers’ ability to pay for improved irrigation water supply in rural KwaZulu-Natal, South Africa	Water S.A./Water SA	2,017	cc-by	7,905	INTRODUCTION farmers will gradually decrease as they will increasingly have to pay for their water use (DWAF, 2004). It is in this context that studies investigating willingness to pay (WTP) are more widely available (Adepoju and Omonona, 2009; Akter, 2006; Alhassan, 2012; Bogale and Urgessa, 2012; Futija et al., 2005; Kanyoka et al., 2008; Moffat et al., 2011; Ndetewio et al., 2013). Ability to pay (ATP) studies are not common and yet are important, since, whilst irrigators may be willing to pay for improved water services, the question of their ability to do so still remains. While WTP is a maximum amount which water-users are willing to pay for a given service, it is difficult to directly use this as basis of setting tariffs as there are variations in characteristics that determine this willingness (Ndetewio et al., 2013). Moreover, it is necessary to set the irrigation water fee at a price level that the majority of irrigators can actually afford to pay, especially where most of these farmers depend on irrigation for their livelihoods. For that purpose, ATP is frequently referred to and used as it reflects an affordable price. h Smallholder irrigation is an important rural development factor, creating employment opportunities, generating income and enhancing food security in Africa in general, and in South Africa (SA) in particular (Bacha et al., 2011; Van Averbeke et al., 2011; Denison and Manona, 2007; Speelman, 2009). As a result, South Africa has invested in the sector by rehabilitating existing irrigation schemes (Perret and Geyser, 2007). In 2012 the government budgeted R15 million for rehabilitation of smallholder irrigation schemes (DAFF, 2012). On the other hand, the agriculture sector, being the largest water user, is under pressure to release water to other sectors (Kanyoka et al., 2008). The growing water scarcity continues to put pressure on farmers to use water more efficiently. A growing area of interest in SA is that of finding balanced and effective resource management strategies for allocating water among the key sectors (Speelman, 2009; Yokwe, 2006).f lf The ATP is considered as the amount which irrigators can pay for water for their cropping needs and is calculated with reference to gross margins (Fujita et al., 2005). Charging farmers for irrigation water should be done carefully because if prices are set too low, revenues may not be sufficient to cover the full costs of supplying water. ABSTRACT The aim of this study was to determine smallholder farmers’ ability to pay (ATP) for improved irrigation water supply, using their gross margins, in rural KwaZulu-Natal, South Africa. The analysis was conducted on a sample of 161 irrigators. Production data were collected from the irrigators and the residual imputation method (RIM) was used to calculate the gross margins that the farmers realised. An ordinary least squares regression was used to investigate factors affecting ATP. Results indicate that factors such as labour, training, household assets and road conditions positively influence ATP. The study highlights the importance of support to farmers and their institutions. The study also concludes that farmers are making profits from their irrigated crops, especially tomatoes, and therefore recommends that farmers start paying for the water used for their crops. Keywords: smallholder irrigation, ability to pay, residual imputation method, gross margins, OLS regression Determinant of farmers’ ability to pay for improved irrigation water supply in rural KwaZulu-Natal, South Africa Sinenhlanhla Njoko1 and Maxwell Mudhara1* Sinenhlanhla Njoko and Maxwell Mudhara 1School of Agricultural, Earth and Environmental Sciences, University of KwaZulu-Natal, Private Bag X01, Scottsville 3209, Pietermaritzburg, South Africa 1School of Agricultural, Earth and Environmental Sciences, University of KwaZulu-Natal, Priva South Africa http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence INTRODUCTION On the other hand, if water prices are set too high, irrigators may not be able to afford the new improved irrigation water supply. Therefore, to set the required water price, information on the ability of irrigators to pay for such services is essential. Since pricing of water is a key component of an appropriate incentive for efficiency, sustainability and accountability, there is a need to research the demand for the service in order to understand the fundamental value that irrigators place on the improved water service, so that the price that reflects the ability of the irrigators to pay for the improved water services can be established (Alebel, 2002). This study, therefore, aimed to estimate the amount farmers are able to pay for irrigation services using farmers’ gross margins. Furthermore, the study investigates the factors that determine farmers’ ATP for irrigation in Msinga Local Municipality. The study was part of a project initiated, managed and funded by Effective water resource management requires that water be treated as an economic good. As such, making rational decisions about water resource management requires reliable estimates of the economic value of water (Speelman, 2009; Hellegers and Perry, 2006). Knowledge of this value contributes significantly in designing fair, informed and rational pricing systems, providing incentives to irrigators to use water sparingly and efficiently, and allowing recovery operations and maintenance (Perret and Geyser, 2007). Moreover, understanding water values plays an important role when making investment decisions in water resources development, and policy decisions on sustainable water use and water allocations. In South Africa, the issue of water valuation among small- scale irrigation schemes is topical, following the new water policy released by the Minister of Water Affairs and Forestry in April 1997. According to the new policy, water subsidies paid to To whom all correspondence should be addressed. Tel: +27 (33) 260-5518; e-mail: Mudhara@ukzn.ac.za Received 17 November 2014; accepted in revised form 28 February 2017 http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence 229 the Water Research Commission (WRC) (K5/2176) entitled ‘Empowerment of women in rural areas through water use security and agricultural skills training for gender equity and poverty reduction in KwaZulu-Natal (KZN) and North West Province’. STUDY AREA Primary data for the study were collected in Msinga Local Municipality, Mzinyathi District, South Africa. Over 70% of this land (1 725 km2) is under traditional authority while the remaining 30% is commercial farm land (Dearlove, 2007). Much of the terrain is located in deep gorges of the Tugela and Buffalo rivers. This effectively isolates the municipal area from the immediate surrounding municipal areas. The Tugela Ferry Irrigation Scheme (TFIS) is located on both banks of the Tugela river, which supplies the scheme with water. The scheme was planned and constructed by the Natal Native Trust between 1898 and 1902 and has been operational ever since (Cousins, 2012). The scheme is made up of 7 blocks of irrigable land covering 837 ha (Cousins, 2012; Fanadzo, 2012). A total of about 1 500 irrigators participate in the scheme, growing various crops. According to Cousins (2012), the Tugela Ferry irrigators comprise about 15% of all smallholder irrigation farmers in the KZN Province (Sinyolo, 2013). g Msinga has very limited employment opportunities. Agriculture is one of the most important economic sectors in Msinga, with most households engaged in smallholder farming (Cousins, 2012; Sinyolo, 2013). Rain-fed crop production, however, is very challenging in Msinga because the area is both hot and dry. The area is characterized by frequent droughts, making irrigation the main mode of household food production (Cousins, 2012). One opportunity that exists for some of the rural people to increase incomes and participate in the local economy is provided by irrigation farming, specifically in the Tugela Ferry and Mooi River Irrigation Schemes. The two irrigation schemes play an important role in the local economy of Msinga as sources of food, employment and market for agricultural inputs (Cousins, 2012; Gomo, 2012; Sinyolo, 2013) Figure 4.1 shows the location of Msinga Local Municipality in SA. Farmers in the irrigation scheme were initially allocated 2 plots each of 0.1 ha in size. Over time, some farmers acquired more plots through leasing or borrowing. The main access to land is through the traditional authorities who allocate land to households. Selling of land is not permissible under the current traditional land tenure system. Initially, all of the blocks were to receive water from the main gravity-fed canal; however, over time water shortages have persisted, leaving only 4 blocks to benefit from the canal while other blocks use motorized pumps. INTRODUCTION The distinctive features of irrigation farming in the Msinga schemes are very similar to those found in other low-cost, gravity-fed schemes in South Africa. They are similar in plot sizes, which are small. In irrigation schemes like Msinga plot sizes are 0.1 ha, although farmers can accumulate a number of plots and end up with an average of 0.24 ha each (Sinyolo et al., 2014). The systems of production are highly labour-intensive and common cash crops grown include green maize, tomatoes, cabbage, potatoes, and leafy green vegetables. The production of specialized types of fresh produce like garlic, beetroot, etc., for niche markets is absent or very limited (Denison and Manona, 2007; Van Averbeke and Khosa, 2011; Cousins, 2012).h http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence METHODS It is assumed that the opportunity costs of non-water inputs are given by their market prices. Therefore, the shadow price of water can be calculated as TVP less the costs of all non-water inputs to production and returns to land. The residual, obtained by subtracting the non-water input (including land) costs from total annual crop revenue, equals the gross margin (GM). The GM can be interpreted as the maximum amount the farmer could pay for water and variable costs. STUDY AREA The 4 blocks receiving Figure 1 Location of the Tugela Ferry and Mooi River Irrigation Schemes in Msinga Local Municipality, South Africa Figure 1 Figure 1 Figure 1 Location of the Tugela Ferry and Mooi River Irrigation Schemes in Msinga Local Municipality, South Africa Location of the Tugela Ferry and Mooi River Irrigation Schemes in Msinga Local Municipality, South Africa 230 http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence 230 water under the gravity system use a canal that is 31 km long to draw the water. Another block uses a diesel pump whilst the remaining 2 blocks use electric pumps. TABLE 1 Distribution of sampled farmers in each study area Irrigation method Total number of households in scheme Number of respondents sampled in each scheme Gravity-fed 473 100 Pump-fed 403 61 Total 876 161 In the Mooi River Irrigation Scheme (MRIS), a diversion constructed across the Mooi River abstracts water into a canal, which runs for a distance of 20.8 km from the abstraction point to the end of the scheme (DAEA, 2011). The scheme has 15 blocks of different sizes. The exact year that the scheme was established is not known; however, it is speculated that it may have been early in the 20th century. Most of the farmers currently in MRIS and TFIS, and even some non-members, grew up with their parents participating in the scheme. Essentially, the purpose of the scheme is to improve the livelihoods of those in the surrounding areas through food production and job creation (Gomo, 2012). The scheme covers a total area of 600 ha, divided into plots which are 0.1 ha. The scheme is made up of approximately 824 farmers, each occupying at least 1 plot; however, some farmers occupy more than one (DAEA, 2011). Irrigators do not have other land outside the irrigation scheme, although dryland farming is prevalent in other parts of the area.h Residual imputation method The residual imputation method (RIM) was used to measure the return to water out of the gross margin obtained from all the production inputs employed. The residual imputation method is the most frequently used approach to applied shadow pricing of producers’ goods, particularly for irrigation water (Young, 1996; Speelman, 2009; Yokwe, 2006). For the RIM, the incremental contribution of each input in the production process is determined. If appropriate prices can be assigned by market forces to all inputs but one, the remainder of total value of product is attributed to the remaining or residual input, which in this case is water (Young, 1996; Lange and Hassan, 2007).h The scheme is managed through block committees, which, amongst other things, see to the distribution of water. As a whole, the scheme is managed by the Irrigation Management Committee which ensures equitable water distribution and resolves conflicts (Gomo, 2012). The first 11 blocks draw water under gravity and the last 4 blocks receive water that is diesel pumped. Initially, all the blocks received water from the main canal, but water shortages due to leakages and multiple uses, and increased numbers of participants, has meant that only the first 11 blocks benefit from the gravity-fed canal (Gomo, 2012). The residual valuation assumes that if all markets are competitive, except the one for water, the Total Value Product (TVP) equals exactly the opportunity costs (OC) of all the inputs: TVP = ∑ VMPiQi +VMPwQw where: TVP = total value of the commodity produced; VMPi = value of marginal product of input i, VMPw = value of marginal product of water, Qi = quantity of input i used in production and Qw= quantity of water used in production. http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence Data collection A structured questionnaire was used to interview the farmers who were household heads. The data collected include household characteristics, land, crop systems, market, sources of off-farm income, credit, water management aspects, and problems associated with agricultural practices in general. The questionnaire was pre-tested and modified accordingly to improve its reliability and validity. Field data were collected in November 2013 over a period of 3 weeks from 2 irrigation schemes, i.e., MRIS and TFIS. However, the gross margin would also include the replacement costs or depreciation, for the scheme to be sustainable. In other words, the realistic ability to pay is less than the gross margin by the adjustment for depreciation. h The technique is based on two principles, as discussed in Young (1996) and Speelman (2009): Households in the two study sites were stratified into gravity- fed and pump-fed irrigators (Table 1). Gravity-fed irrigators in Tugela Ferry were not included in the study as the scheme was under rehabilitation at the time of the survey. The farmers had not been producing crops for a year, which made it difficult for irrigators in that part of the scheme to meaningfully participate in the survey. The proportional random sampling method was used to select the women used for the study, in order to equally represent the categories of gravity-fed and pump-fed farmers. A random sampling procedure was employed to select a total of 161 respondents, where 131 were from the MRIS and 30 were from TFIS. Most blocks in TFIS where not functional during the time of the survey. Hence, a sample of 30 farmers was a fair representation of the operational units in the scheme. • The prices of all resources should equal returns at the margin. This is a well-known condition for competitive equilibrium, i.e., as would occur if perfectly competitive markets were to exist for all agricultural inputs. • The total value of production (TVP) can be divided into shares; in such a way that each resource is paid according to the value of its marginal productivity (VMP) and the total product is completely exhausted (Young, 1996; Lange and Hassan, 2007). The RIM has the advantage of being relatively easy to implement. However, it is sensitive to small variations in the specification of the production function and assumptions http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. RESULTS Where: GM = gross margin and Pi = price of input i. TVP = [PcQc+ PmQm+ PtQt+ PsQs+ PpQp+ PbQb+ PbuQbu] ∑PiQi = PfQf+ PpeQpe+ PtiQti+ PfuQfu+ PlQl+ PseQse Qc = Quantity of cabbage heads harvested per hectare Qm = Quantity of maize harvested per hectare Qt = Quantity of tomatoes harvested per hectare Qs = Quantity of spinach bunches harvested per hectare Qp = Quantity of potato bags harvested per hectare Qb = Quantity of bean bags harvested Qbu = Quantity of butternut bags harvested per hectare Qf = Quantity of fertilizer used per hectare Qpe = Quantity of pesticide used per hectare Qti = Number of times tillage done per hectare Qfu = Number of litres used per hectare Ql = Number of times labour employed per hectare Qse = Number of seed packets used per hectare Pc = Price per cabbage head (R5) Pm = Price per ton of maize (R3 000) Pt = Price per tomato crate (R100) Ps = Price per bunch of spinach (R10) Pp = Price per bag of potatoes (R40) Pb = Price per 5kg bag of beans (R40) Pbu = Price per bag of butternuts (R30) Pf = Price per bag of 50kg fertilizer (R600) Ppe = Price per bag of 50kg pesticide (R540) Pti = Price for tillage per ha (R500) Pfu = Price of fuel per litre (R12.48) Pl = Price of labour per day (R20) Pse = Price per packet of seeds (R11) OLS regression The Ordinary Least Squares (OLS) method was employed to determine the factors that influence farmers’ ATP. The OLS regression model was estimated as follows: Y= β0+β1X1+β2X2+………+ βiXi +ei where: Y is the dependent variable, which is gross margin; β’s are the parameter estimates, X’s are the explanatory variables (Table 2). where: Y is the dependent variable, which is gross margin; β’s are the parameter estimates, X’s are the explanatory variables (Table 2). GM = TVP − ∑PiQi Data collection 2 April 2017 Published under a Creative Commons Attribution Licence 231 financially, hence motivating family members to work on the farm is a critical management factor. To calculate labour costs, a shadow price of R20/day per person was used. This value for labour costs was estimated based on discussions with farmers and extension officers during focus group discussions and key informant interviews, respectively. According to the focus group discussions and key informant interviews, labour is employed in the plots twice a week, for 4 weeks a month over a cropping season. To calculate the gross margins, the total costs were subtracted from the total revenue and then divided by the land occupied by the farmer in order to get gross margins per hectare. about market and policy environment. Therefore, the RIM is only suitable when the residual input contributes a large fraction of the output value, as is the case for water in irrigated agriculture in water-scarce regions (Speelman, 2009; Yokwe, 2006). If an input to production is omitted or underestimated, its contribution is wrongly attributed to water. To overcome this problem, all relevant inputs should be included in the model. Some important inputs like farm labour, including family labour, are often unpaid. A shadow price is then estimated, usually in terms of the opportunity cost of the workers (Young, 1996). In this study, the revenue earned by the farmers for each crop was calculated by multiplying their production by respective market prices. The portion of total production that was consumed by the household was excluded as it was taken as negligible. Inputs considered relevant in the production process, such as fertilizers, pesticides, energy, tillage, seed, land and labour, were taken into account. For fertilizers and pesticides the competitive market prices were used to determine costs. However, the value of inputs provided to farmers by extension services through the government subsidy scheme, e.g., fertilizers, pesticides and seeds, added to input costs. For inputs and the output, market and individual prices are considered to equal shadow prices. Descriptive statistics The schemes also have some distinctive features: the majority of plot holders are women (88.75%). The average age of the farmers in the schemes was 58 years. This shows that the more elderly are more involved in agriculture. This could be indicative of the fact that the males and the younger population have moved to the cities to explore better opportunities. Household size ranges from 1 to 20, with an average of 7. This could potentially indicate a sizeable amount of family labour, especially when labour availability is mainly dependent on the family size. Of the sampled population, 57% had no education and 16% had obtained secondary education. This shows very low levels of literacy in Msinga. When looking at the duration of time that the farmers have been in the scheme (also a proxy for irrigation- farming experience), the minimum is 1 year and the maximum is 40 years, with a mean duration of 19 years. Experience is expected to aid farmers in increasing both their productivity and profitability.h The production of food crops for home consumption is limited as, according to the farmers, less than 5% is consumed at home whilst the bulk of production is for sale. Farmers are commercially oriented and approach farming as a business. Almost all crops use fertilizers and crop chemicals. The use of hired labour is not common as most farmers employ household labour. An active, informal plot rental market makes it possible for many farmers to gain access to additional plots. More than 30% of farmers indicated that they were leasing some land. This level of land rental could minimise the plots that are not under cultivation in any given year. h Most farmers in Msinga employ family labour. According to Van Averbeke (2008), one of the important advantages of using family labour is flexibility. Individual family members are often faced with the choice between working on the farm and engaging in other economic activities, which may be more rewarding The crops grown by most farmers in summer are maize, which is grown by 67% of the respondents, potatoes, grown by 63%, and tomatoes, grown by 38% of the sample. In winter, beans and cabbage were the dominant crops, with beans grown by 38% of the sample and cabbage grown by 27%. http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence Descriptive statistics 2 April 2017 Published under a Creative Commons Attribution Licence 232 TABLE 2 Description of the variables Variable Variable description Expected sign Dependent variables Ability to pay (ATP) Total gross margins realized by farmers per hectare Independent variables Gender Household head gender: male= 0, female=1 - Age Household head age (years) + Land security Farmer perceptions of their land tenure security + Household labour Number of household members available to work in the plots + Household labour1 Number of household members available to work in the plots - Duration Number of years that the household has been involved in the irrigation scheme + Access to credit Access to credit in the past year: yes = 1, no = 0 + Access to extension Access to extension services in the last year in numbers (Number of visits in the last 12 months) + Training Agricultural skills training: yes = 1, no = 0 + Total land Total land holdings of household (ha) - Household assets The value of household assets (Rands) + Livestock size Livestock size in Tropical Livestock Units (TLU) + Off-farm income Off-farm income (Rands) + Pump_1 Pump used: electric pump = 1, gravity or otherwise = 0 + Pump_2 Pump used: diesel pump = 1, gravity or otherwise = 0 + Education_1 Education level of respondent: no education = 1, primary education or otherwise = 0 - Education_2 Education level of respondent: secondary education = 1, primary education or otherwise = 0 + Scheme management Farmers’ perceptions on the management of the scheme: good = 1, poor = 0 + Road conditions Farmers’ perceptions on the road conditions: good = 1, poor = 0 + 1Household labour has been included in the model to account for its marginal effect on ATP. Initially, labour is expected to have a positive effect on ATP, but as it increases, it is expected to have a negative effect on production and consequently negative affect on ATP. TABLE 2 Description of the variables Table 3 shows the crops grown by season in the schemes and the yields realized. An average of 3.56 t/ha of maize is produced by the farmers in the schemes, with the mean gross margin being 3 497.57 R/ha. According to Mnkeni et al. (2010), maize could potentially yield 8.2 t/ha. The average yield for potatoes is 21.4 t/ha with a mean gross margin of 3 501.79 R/ha. http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence Descriptive statistics Mnkeni et al. (2010) reports potato yields of 45.3 t/ha. Tomatoes had the least number of growers in summer. This may be due to the costly nature of tomato production, from fertilizers to herbicides and pesticides. On average, the yield of tomatoes was 28.94 t/ha, with an average gross margin of 17 249.41 R/ha. Mnkeni et al. (2010) reported an average potential yield of 47.1 t/ha for tomatoes. The results show that tomatoes had the highest gross margin among the summer season crops. However, gross margins are not the only determinant of farmers’ decisions to plant tomatoes, as it is a perishable commodity. For example, access to reliable markets is critical. The yields reported in this study are below those reported by Mnkeni et al. (2010). However, this was expected. Lower yields are expected under farmer-managed conditions because of the less than optimal conditions under which they produce crops. This is especially true of Msinga, which experiences unfavourable conditions for agriculture. Table 4 shows comparisons of gross margins across categorical variables, and the t-tests presented in the table indicate that there is a statistically significant difference between the gross margins by gender. Males achieve higher gross margins than females, probably because the former have better access to resources, resulting in higher productivity. hf The difference between the gross margins of those who felt land secure and their counterparts was statistically significant. On average, the land secure realised higher gross margins compared to the insecure. This is expected as those with secure land tenure invest resources and measures that lead to improved long-term productivity relative to their counterparts. There were statistically significant differences between gross margins based on the type of irrigation water conveyance method. Farmers relying on gravity had higher gross margins compared to those using motorized pumps. Gravity irrigation water is cheaper and hence leads to high gross margins. With the winter crops, the average production of beans was 0.87 t/ha with an average gross margin of R2 913.14. The number of cabbage producers was lower than for other crops. According to Mnkeni et al. (2010), the potential yield for cabbage is 64.8 t/ha for small-scale irrigators. However, in these schemes, the average yield was 19.50 t/ha and the gross margin was 1 909.19 R/ha per season, reflecting the low price for this crop. http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. Descriptive statistics 2 April 2017 Published under a Creative Commons Attribution Licence 233 TABLE 3 Crop yields and gross margins for crops dominant in summer and those dominant in winter. Source: Household survey (2013) Crop Number of growers Percentage of growers (%) Mean production (t/ha) Gross margin range (per ha) Mean gross margin (R/ha ) Min Max Summer crops Maize 108 67.08 3.56 −R5 500 R14 700 3 498 Potatoes 101 62.73 21. 35 −R2 900 R10 700 3 502 Tomatoes 61 37.89 28.94 −R5 420 R39 500 17 249 Winter crops Beans 66 41.00 0.87 −R2 593 R9 970 2 913 Cabbage 44 27.33 19.50 −R1 600 R6 489 1 909 TABLE 4 Gross margin comparisons across different categorical variables (source: household survey, 2013) Variable definition Category 0 Category 1 t-test significance Mean Std. Dev. Mean Std. Dev. Gender (0 = female, 1 = male) R11 898.01 13 743.35 R18 880.73 18 058.25 * Irrigation system (0 = gravity, 1 = motorized pump) R13 464.69 16 149.44 R11 951.22 12 455.86 ns Land security (0 = insecure, 1 = secure) R10 485.90 13 031.27 R16 693.16 15 986.95 * Road conditions (0 = poor, 1 = good) R10 561.06 14 073.41 R13 237.37 14 520.18 ns Notes: * and ** mean significant at 1%, 5% levels of significance, respectively TABLE 3 TABLE 3 TABLE 4 http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence Factors affecting ability to pay for water The parameter estimate for land security is negative and significant (p < 0.05). This implies that the ATP for farmers who do not feel secure with their land is R4 776.26 less than that of those who feel secure, ceteris paribus. An OLS model was estimated to determine the household characteristics that predict households’ ATP for improved irrigation. The model results are presented in Table 5. The results show that, collectively, all estimated coefficients are statistically significant since the F statistic is statistically significant (p < 0.01). The Adjusted R2 value of about 31% is acceptable considering this is cross-sectional data. The Akaike Information Criteria (AIC) was employed to check whether the model was correctly specified. The results showed that the OLS was a correct specification of the model. The household labour coefficient estimate is positive and significant (p < 0.01). This implies that if the household labour increases by 1 person, the ATP will increase by R3 187.98, ceteris paribus. However, the relationship between labour and gross margin is not linear. The coefficient estimate for household labour squared is negative and significant (p < 0.05). As household labour increases, the household’s ATP for improved irrigation initially increases but decreases at higher levels of household labour, ceteris paribus. hfii i Among the explanatory variables included in the analysis, 7 variables had a significant effect on respondents’ ATP. Results from the estimated model reveal that gender, land security, household labour, household labour squared, training, household assets and road conditions significantly impact on farmers’ ATP. The gender of the household head significantly affects the ATP (p < 0.1). The results indicate that if the household head is female, the ATP is R6 400.56 less than that of male-headed households, ceteris paribus. The coefficient estimate for training is positive and significant (p < 0.05). This implies that those farmers/household heads who have received some agriculture-related training have R6 590.72 more ATP for irrigation water compared to those who have not received any form of training, ceteris paribus. This was expected as training can equip farmers with the information that can help them to increase productivity, which increases their ATP. http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence 234 TABLE 5 Determinants of farmers’ ability to pay for improved irrigation (source: household survey, 2013) Variable Coefficient Std. Err. DISCUSSION TABLE 5 TABLE 5 The findings that male-headed households earn higher gross margins in this study suggests that such households generally have better access to resources which is expected to lead to increased profits and therefore higher ATP. These findings are in line with those of Quisumbing (1996) and Udry et al. (1995), who found that in Sub-Saharan Africa, women are less educated and have limited access to labour, fertilizers and other inputs. That is, women have to ‘learn by doing’ as most of them are illiterate and often learn agricultural practices that are passed from one generation to another. This differs from males who have higher literacy rates and access to resources, and therefore can achieve higher productivity and gross margins. Kyomugisha (2009) also found similar results where security of land tenure is one of the factors that influence investment to enhance land productivity and gross margins. Farmers who have secure land tenure are more likely to invest in improved production practices than their counterparts. Farmers who do have not security of land tenure may not invest much to keep the land productive, e.g., adopting practices like crop rotation, as they may be trying to maximize extraction, in the process reducing the quality of the soil and its productivity. Kyomugisha (2009) concludes that secure land tenure impacts positively on the level of investment made towards production, which increases productivity and gross margins, and the farmers ATP for water. The initial increase in ATP, which is followed by a decrease as labour increases, can be explained from a productivity perspective. Given the small farm sizes under irrigation, initial increases in labour lead to higher productivity. However, eventually further increases result in reduced productivity and, by extrapolation, a decline in ATP.h There are a number of organizations training farmers; these are from the government, non-government organizations and the private sector. The government is the major player in providing training through its various agricultural development programmes (Kinambuga, 2010). The government also collaborates with non-governmental organizations to train farmers. Nevertheless, the proportion of farmers who access training is very low and this has a bearing on their production abilities, and thus gross margins achieved and their ATP for irrigation water. Only 31% of the respondents had received training related to agriculture. Training is important as it gives farmers production information and technologies that can increase their productivity. Factors affecting ability to pay for water Gender −6 400.56* 3 324.14 Age −113.17 81.64 Land security −4 776.26 2 341.61 Household labour 3 187.98* 1 071.52 Household labour squared −208.37 108.44 Duration 7.00 49.48 Credit access −545.73 2 507.21 Extension service access −254.9 570.06 Training 6 590.72 2 129.12 Total land holdings 2 088.71 1 763.82 Household assets 0.03 0.01 Livestock size −36.29 83.88 Off-farm income 0.12 0.17 Pump_1 9 078.06 5 519.16 Pump_2 1 111.27 2 488.54 Education_1 4 368.7 2 358.14 Education_2 −285.14 3 336.54 Scheme management −3 415.6 2 244.71 Road conditions 1 598.91 752.47 Cons 16 964.1 9 410.87 F statistic *** Adjusted R-squared 0.3088 n 160 Note on statistical significance: = p < 0.10, * = p < 0.05, *** = p < 0.01 http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence CONCLUSIONS DAFF (Department of Agriculture, Forestry and Fisheries, South Africa) (2012) Revitalization of irrigation schemes. URL: http://www.nda. agric.za (Accessed 1 July 2015).f The main aim of this study was to determine the ability of farmers to pay for improved irrigation water supply in rural KZN’s Msinga Local Municipality, and to determine factors that determine their ATP for improved irrigation. Results show that farmers are realizing positive gross margins from their plots, although some made negative gross margins due to harsh weather conditions, but in general positive gross margins are realized from the irrigation schemes. The study shows that support services such as training are pivotal in improving farmers’ understanding of agricultural issues, influencing their cropping patterns, and thus their ATP for water. Furthermore, the study shows the importance of institutions, particularly security of land tenure, that also influence the ability of farmers to pay for irrigation water. The study recommends increased farmer training to expose farmers to agricultural knowledge. The study further concludes that, given the gross margins of the farmers, overall the farmers have the ATP for improved irrigation. Government policies to ensure sustainability of llh ld i i i h h ld i h i DWAF (Department of Water Affairs and Forestry, South Africa) (2004) National Water Resource Strategy. (DWAF). URL: http://www.dwaf. gov.za/Documents/Policies/NWRS/Default.htm (Accessed 26 April 2014) Department of Water Affairs and Forestry Pretoria 0 ). epa t e t o Watef a s a d o est y, eto a. FANADZO M (2012) Revitalisation of smallholder irrigation schemes for poverty alleviation and household food security in South Africa: A review. Afr. J. Agric. Res. 7 (13) 1956–1969. https://doi.org/10.5897/ ajarx11.051 FUJITA Y, FUJII A, FURUKAWA S and OGAWA T (2005) Estimation of willingness to pay (WTP) for water and sanitation services through contingent valuation method (CVM): A case study in Iquitos City, The Republic of Peru. JBICI Rev. 11 59–87. h GOMO T (2012) Water management in communally managed smallholder irrigation schemes: Technical performance of Mooi River Irrigation Scheme. MSc thesis, School of Engineering, University of KwaZulu-Natal, Pietermaritzburg, South Africa. g g KwaZulu-Natal, Pietermaritzburg, South Africa. HELLEGERS PJ and PERRY CJ (2006) Can irrigation water use be guided by market forces? Theory and Practice. Int. J. Water Resour. Dev. 22 79–86. REFERENCES ADEPOJU AA and OMONONA BT (2009) Determinants of willingness to pay for improved water supply in Osogbo Metropolis, Osun State, Nigeria. Res. J. Soc. Sci. 4 1–6. KYOMUGISHA E (2009) Land tenure and agricultural productivity in Uganda. Uganda strategy support program (USSP). Brief No. 5. IFPRI, Kampala. AKTER S (2006) Farmers’ willingness to pay for irrigation water under government managed small scale irrigation projects in Bangladesh. J. Bangladesh Stud. 9 21–31.f LANGE GM and HASSAN R (2007) Methodologies for valuation of water services. In: Lange GM and Hassan R (eds) The Economics of Water Management in Southern Africa: An Environmental Accounting Approach. Edward Elgar Publishing, Cheltenham. 203–236. ALEBEL B (2002) Analysis of affordability and willingness to pay for improved water service in urban areas, strategy for cost recovery. MSc Thesis, School of Graduate Studies of Addis Ababa University, Addis Ababa. pp g g MNKENI P, CHIDUZA C, MODI A, STEVENS J, MONDE N, VAN DER STOEP I and DLADLA R (2010) Best management practices for smallholder farming on two irrigation schemes in the Eastern Cape and KwaZulu-Natal through participatory adaptive research. WRC Report No. TT 478/10. Water Research Commission, Pretoria. ALHASSAN M (2012) Estimating farmers’ willingness to pay for improved irrigation: an economic study of the Bontanga irrigation scheme in Northern Ghana. MSc thesis, Department of Agricultural and Resource Economics, Colorado State University, Fort Collins, Colorado. g p p y p Report No. TT 478/10. Water Research Commission, Pretoria. MOFFAT B, MOTLALENG G and THUKUZA A (2011) Households’ willingness to pay for improved water quality and reliability of supply in Chobe ward, Maun. Botswana J. Econ. 8 (12) 45–61. ARIAS P, HALLAM D, KRIVONOS E and MORRISON J (2013) Smallholder integration in changing food markets. Food and Agriculture Organization of the United Nations, Rome. NDETEWIO P, MWAKANJE A, MUJWAHUZI M and NGANA J (2013) Factors influencing willingness to pay for watershed services in lower Moshi, Pangani Basin, Tanzania. Int. J. Agric. Environ. 02 57–75. BACHA D, NAMARA R, BOGALE A and TESFAYE A (2011) Impact of small-scale irrigation on household poverty: Empirical evidence from the Ambo District in Ethiopia. Irrig. Drain. 60 1–10. https://doi. org/10.1002/ird.550 g g PERRET S and GEYSER M (2007) The full financial costs of irrigation services: a discussion on existing guidelines and implications for smallholder irrigation in South Africa. CONCLUSIONS https://doi.org/10.1080/07900620500405643 KANYOKA P, FAROLFI S and MORARDET S (2008) Households’ preferences and willingness to pay for multiple use water services in rural areas of South Africa: An analysis based on choice modelling. Water SA 34 (6) 715–723.i smallholder irrigation schemes should recognize the opportunity to recoup running costs directly from farmers. smallholder irrigation schemes should recognize the opportunity to recoup running costs directly from farmers. KINAMBUGA D (2010) Evaluation of the constraints to profitable smallholder dairying: a case of Nakuru County, Kenya. MSc thesis, Egerton University, Kenya. DISCUSSION Therefore, farmers with no training risk having low production due to lack of knowledge. The average asset base of the farmers in the study was R83 916.34. The expectation is that households with a higher value of assets also do better in their level of production and make good gross margins, thus should afford higher ATP for irrigation water. This is in agreement with Kinambuga (2010). Accordingly, the parameter estimate for household assets is positive and significant (p < 0.1). This implies that an increase in the household assets will result in an increase in the household’s ATP for the improved irrigation, ceteris paribus. Household assets are a sign of wealth such that families with high-value assets are expected to have invested in more capital and even production technologies, which would allow them to produce more, realize higher gross margins and, consequently, be able to pay more for improved irrigation. Household assets in the present study include houses, cars, water tanks, motor cycles, etc., which were all valued and added together to get their total value at household level. According to Arias et al. (2013), when market integration of producers is limited, interventions to reduce barriers to market participation will often have a greater payoff. One such intervention is improving market information systems, by improving feeder roads or reducing the fees that traders pay to transport products between markets. Therefore, farmers’ perception of their roads is one way to determine their access to market (Arias et al., 2013). In the present study, farmers’ perceptions have been used as a proxy for market access, and 81% of the respondents perceived their roads to be in good condition. According to the model, farmers’ perceptions of the road conditions impact on their ATP for improved irrigation. The parameter estimate for road conditions is positive and significant (p < 0.05). That is, if farmers perceive the road conditions to be good, their ATP for irrigation water will be R1 598.91 more than that of their counterparts. Good road conditions imply easier access to markets. If farmers have better access to markets, they are able to sell their produce with ease and realize higher gross margins, which in turn leads to increased ATP for water. http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence 235 http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence REFERENCES Water SA 33 (1) 67–78.f QUISUMBING A (1996) Male-female differences in agricultural productivity: methodological issues and empirical evidence. World Dev. 24 (10) 1579–1595. https://doi. g BOGALE A and URGESSA B (2012) Households’ willingness to pay for improved rural water service provision: application of contingent valuation method in Eastern Ethiopia. J. Hum. Ecol. 38 (2) 145–154. org/10.1016/0305-750X(96)00059-9 COUSINS B (2012) Smallholder irrigation schemes, agrarian reform and ‘accumulation from below’: Evidence from Tugela Ferry, KwaZulu-Natal. Paper presented at Strategies to overcome poverty and inequality: Towards Carnegie III, 3–7 September 2012, University of Cape Town, South Africa.f SINYOLO S, MUDHARA M and WALE E (2014) The impact of smallholder irrigation on household welfare: The case of Tugela Ferry irrigation scheme in KwaZulu-Natal, South Africa. Water SA 40 (1) 145–156. https://doi.org/10.4314/wsa.v40i1.18h SINYOLO S (2013) The impact of smallholder irrigation and water security on household welfare: The case of Tugela Ferry irrigation scheme in KwaZulu-Natal, South Africa. MSc thesis, School of Agriculture, Earth and Environmental Sciences, University of KwaZulu-Natal, Pietermaritzburg, South Africa. p DAEA (Department of Agriculture and Environmental Affairs, KwaZulu- Natal Province) (2011) Mooi River Irrigation Scheme topographic survey, southern region. DAEA, Hilton. DEARLOVE PD (2007) Msinga Municipality Integrated Development Plan 2005/2006. KwaZulu-Natal Department of Local Governance and Traditional Affairs. URL: http://devplan.kzntl.gov.za/idp_ reviewed_2007_8/IDPS/KZ244 /Adopted/Msinga %20IDP.pdf (Accessed 27 March 2014). SPEELMAN S (2009) Water use efficiency and influence of management policies, analysis for the small-scale irrigation sector in South Africa. PhD thesis, Ghent University, Belgium. y g UDRY C, HODDINOTT J, ALDERMAN H and HADDAD L (1995) Gender differentials in farm productivity: implications for household efficiency and agricultural policy. Food Policy 20 (5) 407–423. https:// doi.org/10.1016/0306-9192(95)00035-D VAN AVERBEKE W (2008) Best management practices for small-scale subsistence farming on selected irrigation schemes and surrounding y g UDRY C, HODDINOTT J, ALDERMAN H and HADDAD L (1995) Gender differentials in farm productivity: implications for household efficiency and agricultural policy. Food Policy 20 (5) 407–423. https:// doi.org/10.1016/0306-9192(95)00035-D UDRY C, HODDINOTT J, ALDERMAN H and HADDAD L (1995) Gender differentials in farm productivity: implications for household efficiency and agricultural policy. Food Policy 20 (5) 407–423. https:// doi.org/10.1016/0306-9192(95)00035-D DENISON J and MANONA S (2007) Principles, approaches and guidelines for the participatory revitalization of smallholder irrigation schemes. Volume 1: A rough guide for irrigation development practitioners. WRC Report No. TT 309/07. Water Research Commission, Pretoria. areas through participatory adaptive research in Limpopo Province. Draft Final Report of WRC Project K5/1464//4. Centre for Organic and Smallholder Agriculture (COSA), Department of Crop Sciences, Tshwane University of Technology, Pretoria. http://dx.doi.org/10.4314/wsa.v43i2.07 Available on website http://www.wrc.org.za ISSN 1816-7950 (Online) = Water SA Vol. 43 No. 2 April 2017 Published under a Creative Commons Attribution Licence areas through participatory adaptive research in Limpopo Province. Draft Final Report of WRC Project K5/1464//4. Centre for Organic and Smallholder Agriculture (COSA), Department of Crop Sciences, Tshwane University of Technology, Pretoria. VAN AVERBEKE W and KHOSA TB (2011) Smallholder irrigation schemes in South Africa with a focus on Dzindzi Canal Irrigation Scheme in Limpopo: dynamic smallholders amidst contested policy priorities. In Aliber M (ed.) Strategies to support South African smallholders as a contribution to government’s second economy strategy. Vol. 2: Case studies. Institute for Poverty, Land and Agrarian Studies, School of Government, University of the Western Cape, Cape Town. 145–162. areas through participatory adaptive research in Limpopo Province. Draft Final Report of WRC Project K5/1464//4. Centre for Organic and Smallholder Agriculture (COSA), Department of Crop Sciences, Tshwane University of Technology, Pretoria. VAN AVERBEKE W and KHOSA TB (2011) Smallholder irrigation schemes in South Africa with a focus on Dzindzi Canal Irrigation Scheme in Limpopo: dynamic smallholders amidst contested policy priorities. In Aliber M (ed.) Strategies to support South African smallholders as a contribution to government’s second economy strategy. Vol. 2: Case studies. Institute for Poverty, Land and Agrarian Studies, School of Government, University of the Western Cape, Cape Town. 145–162. VAN AVERBEKE W and KHOSA TB (2011) Smallholder irrigation schemes in South Africa with a focus on Dzindzi Canal Irrigation Scheme in Limpopo: dynamic smallholders amidst contested policy priorities. In Aliber M (ed.) Strategies to support South African smallholders as a contribution to government’s second economy strategy. Vol. 2: Case studies. Institute for Poverty, Land and Agrarian Studies, School of Government, University of the Western Cape, Cape Town. 145–162. REFERENCES g VAN AVERBEKE W (2008) Best management practices for small-scale subsistence farming on selected irrigation schemes and surrounding VAN AVERBEKE W (2008) Best management practices for small-scale subsistence farming on selected irrigation schemes and surrounding 236 VAN AVERBEKE W, DENISON J and MNKENI PNS (2011) Smallholder irrigation schemes in South Africa: A review of knowledge generated by the Water Research Commission. Water SA 37 (5) 797–808. https://doi.org/10.4314/wsa.v37i5.17 g g y 37 (5) 797–808. https://doi.org/10.4314/wsa.v37i5.17 YOKWE SCB (2006) Investigation of the economics of water as used by smallholder irrigation farmers in South Africa. MSc thesis, Faculty of Natural and Agricultural Sciences, Department of Agricultural Economics, Extension and rural development, University of Pretoria, Pretoria. YOUNG RA (1996) Measuring economic benefits for water investments and policies. World Bank Technical Paper No. 338. The World Bank, Washington, D.C. 237
https://openalex.org/W4243292917	https://www.pure.ed.ac.uk/ws/files/124611277/document_1_.pdf	English	null	Randomised controlled trial of the short-term effects of OROS-methylphenidate on ADHD symptoms and behavioural outcomes in young male prisoners with attention-deficit/hyperactivity disorder (CIAO-II)	Research Square (Research Square)	2,019	cc-by	16,192	General rights g Copyright for the publications made accessible via the Edinburgh Research Explorer is retained by the author(s) and / or other copyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associated with these rights. Edinburgh Research Explorer Randomised controlled trial of the short-term effects of OROS- methylphenidate on ADHD symptoms and behavioural outcomes in young male prisoners with attention-deficit/hyperactivity disorder (CIAO-II) Citation for published version: Asherson, P, Johansson, L, Holland, R, Fahy, T, Forester, A, Howitt, S, Lawrie, S, Strang, J, Young, S, Landau, S & Thomson, L 2019, 'Randomised controlled trial of the short-term effects of OROS- methylphenidate on ADHD symptoms and behavioural outcomes in young male prisoners with attention- deficit/hyperactivity disorder (CIAO-II)', Trials, vol. 20, no. 1. https://doi.org/10.1186/s13063-019-3705-9 Digital Object Identifier (DOI): 10.1186/s13063-019-3705-9 Link: Link to publication record in Edinburgh Research Explorer Link: Link to publication record in Edinburgh Research Explorer Document Version: Publisher's PDF, also known as Version of record Take down policy Th U i i f Ed p y The University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorer content complies with UK legislation. If you believe that the public display of this file breaches copyright please contact openaccess@ed.ac.uk providing details, and we will remove access to the work immediately and investigate your claim. Download date: 24. Oct. 2024 Asherson et al. Trials (2019) 20:663 https://doi.org/10.1186/s13063-019-3705-9 Open Access Randomised controlled trial of the short- term effects of OROS-methylphenidate on ADHD symptoms and behavioural outcomes in young male prisoners with attention-deficit/hyperactivity disorder (CIAO-II) Philip Asherson1* , Lena Johansson1, Rachel Holland1, Tom Fahy1, Andrew Forester2, Sheila Howitt3, Stephen Lawrie3, John Strang1, Susan Young4, Sabine Landau1 and Lindsay Thomson3 Abstract Background: Attention-deficit/hyperactivity disorder (ADHD) is a highly prevalent disorder, seen in 20–30% of young adult prisoners. Pharmacoepidemiological studies, a small randomised controlled trial and open trial data of methylphenidate suggest clinically significant reductions in ADHD symptoms, emotional dysregulation, disruptive behaviour and increased engagement with educational activities. Yet, routine treatment of ADHD in offenders is not yet established clinical practice. There is continued uncertainty about the clinical response to methylphenidate (MPH), a first-line treatment for ADHD, in offenders, who often present with an array of complex mental health problems that may be better explained by states of inattentive, overactive, restless and impulsive behaviours. To address this problem, we will conduct an efficacy trial to establish the short-term effects of osmotic-controlled release oral delivery system (OROS)-methylphenidate (Concerta XL), an extended release formulation of MPH, on ADHD symptoms, emotional dysregulation and behaviour. Methods: This study is a parallel-arm, randomised, placebo-controlled trial of OROS-MPH on ADHD symptoms, behaviour and functional outcomes in young male prisoners aged 16–25, meeting Diagnostic and Statistical Manual of Mental Disorders, fifth edition criteria for ADHD. Participants are randomised to 8 weeks of treatment with OROS- MPH or placebo, titrated over 5 weeks to balance ADHD symptom improvement against side effects. Two hundred participants will be recruited with a 1:1 ratio of drug to placebo. The primary outcome is change in level of ADHD symptoms after 8 weeks of trial medication. (Continued on next page) * Correspondence: Philip.asherson@kcl.ac.uk 1Institute of Psychiatry, Psychology and Neuroscience, King’s College London, De Crespigny Park, London SE5 8AF, UK Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted 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. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Background studies in community ADHD samples. First, offenders present with an array of complex mental health prob- lems that may better explain the states of inattentive, overactive, restless and impulsive behaviours used to de- fine ADHD. These include problems commonly seen in offenders such as personality, anxiety, post-traumatic stress and substance abuse disorders, in addition to gen- eral and specific learning difficulties. Second, nearly all previous treatment trials of ADHD have been conducted in carefully selected samples with low levels of co- occurring psychosocial and mental health comorbidities. However, the co-occurrence of mental health disorders might modify the efficacy of drug treatments in ADHD. One example is comorbid drug abuse. Meta-analysis of treatment trials found no effect of MPH on ADHD symptoms in ADHD cases comorbid with drug abuse (standardised mean difference [SMD] = 0.08, p = 0.59), whereas there was a medium effect in non-comorbid samples (SMD = 0.51, p < 0.00001) [11]. This may be relevant to prison populations, where a history of drug abuse is common. Further, concerns have been expressed about the potential for stimulant medications such as MPH to worsen coexisting conditions. The most recent research recommendations from NICE [8] state that: ’no evidence was identified to justify different medi- cation choices in people with ADHD and a history of psychosis, mania, or personality disorder. These groups are often excluded from trials. There are reasons (for ex- ample, mechanism of action of medication options, pre- vious reports of adverse effects) to suspect that these groups may respond differently to different drugs, but a lack of trials to confirm this. Primarily there are some concerns that stimulant medication may worsen the symptoms of any of these coexisting conditions’. Attention-deficit/hyperactivity disorder (ADHD) is char- acterised by developmentally inappropriate levels of in- attentive, hyperactive and impulsive behaviours. The disorder is often accompanied by symptoms of emotional instability and leads to clinical and psychosocial impair- ments culminating in long-term negative outcomes and comorbid conditions [1]. ADHD affects around 5–7% of children [2, 3] and 3–4% of adults [4, 5]. Individuals meet- ing diagnostic criteria for ADHD are found at dispropor- tionately high rates in prison populations with an estimated prevalence rate between 20 and 30% in young offender institutes and prisons [6]. © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted 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. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Asherson et al. Trials (2019) 20:663 Page 2 of 21 (Continued from previous page) Discussion: Potential benefits include improvement in ADHD symptoms, emotional dysregulation, attitudes towards violence and critical incidents and increased engagement with educational and rehabilitation programmes. Demonstrating the efficacy and safety of MPH on ADHD symptoms and associated impairments may provide the data needed to develop effective healthcare pathways for a significant group of young offenders. Establishing efficacy of MPH in this population will provide the foundation needed to establish long-term effectiveness studies with the potential for demonstrating significant reductions in criminal behaviour and improved health-economic outcomes. Trial registration: ISRCTN registry, ISRCTN16827947, 31st May 2016; EudraCT number, 2015-004271-78, 31st May 2016. Last particpant last visit 6 June 2019. Data lock 27 August 2019. Keywords: Neurodevelopmental disorder, ADHD, OROS-methylphenidate, Prison mental health, Trial Background p Both the National Institute for Health and Care Excel- lence (NICE) and the Scottish Intercollegiate Guidelines Network (SIGN) recommend methylphenidate (MPH) for treating ADHD with significant impairment in children, adolescents and young adults [7–9]. Nevertheless, it is un- common to diagnose and treat ADHD in young adult of- fenders. The reasons for this are unclear, but concerns have been expressed that the common occurrence of mental health, neurodevelopmental and psychosocial problems might provide a better explanation for impul- sive, overactive and inattentive states in young offenders, or might interfere with the treatment response in cases of ADHD. For these reasons NICE [7] recommended that drug treatment efficacy trials are needed in offender popu- lations (page 134 of the full guideline, section 5.18.1.4) and repeated this recommendation in 2013 [10]. The guideline stated that ’there should be an assessment of ef- ficacy in these groups (i.e. forensic and drug abuse popula- tions) of the ADHD treatments already recommended for treatment in the community. Randomised controlled trial design is recommended’. Clinical trials of ADHD treat- ments have yet to be conducted in young adult offenders, and the efficacy of MPH treatment for ADHD remains unknown in this group. Previous studies Previous community studies demonstrate the efficacy of MPH in reducing ADHD symptoms in children, adoles- cents and adults with ADHD [12]. A recent comprehen- sive network meta-analysis estimated an effect size from randomised controlled trials of MPH in reduction of ADHD symptoms in adults with an SMD of 0.78 (95% confidence interval [CI] of 0.62–0.93) [12]. However, there are only limited trial data for treatment of ADHD in young offenders presenting with a more complex mix of psychosocial, mental health and behavioural problems. One small randomised controlled trial of MPH in a prison sample of 30 Swedish prisoners with ADHD showed a large effect (SMD = 2.1) on ADHD symptom reduction [13]. While this study supports the treatment of ADHD in offenders, it cannot be considered definitive for the treat- ment of young offenders more generally, because of the small sample size, older age group and selection of of- fenders with severe ADHD and with long-term sentences treated in a special prison unit in Sweden. In an earlier study of ADHD in prisoners, we found a sixfold increase in critical incidents among prison in- mates with high levels of ADHD symptoms compared to prisoners with low levels of symptoms. This increase remained significant even after controlling for antisocial personality disorder [13], thus making this an important outcome for randomised controlled trials of prisoners with ADHD. Another important outcome is symptoms of emotional dysregulation such as irritability, anger and reactive aggression, which are also commonly seen in of- fenders with ADHD. Meta-analyses of randomised con- trolled trials of ADHD medications found reductions in emotional dysregulation, including problems with tem- per control, mood lability and emotional over-reactivity [19, 20]. Hence, treatment of offenders with ADHD might also lead to significant reductions in emotional dysregulation and potentially aggressive or violent be- haviour. The symptoms of ADHD are also known to interfere with education and employment due to a com- bination of restlessness, reduced attention span, forget- fulness and problems with planning and organisation [21]. Treatment might therefore lead to greater positive engagement with prison educational and rehabilitation programmes. In our open label pilot study at HMP Isis, we also found significant effects on all the secondary outcomes proposed for this study (all p < 0.001) includ- ing measures of emotional dysregulation, attitudes to- wards violence, the number of critical incidents and positive engagement with the education and rehabilita- tion programme [14]. Previous studies p p Additional support comes from a pilot open label study that preceded this protocol; the study investigated the ef- fects of MPH in 121 young offenders in HMP Isis in Southeast London who met Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5) diag- nostic criteria for ADHD [14]. That study followed similar procedures to the protocol reported here, but it was an open trial with a single treatment arm. Potential partici- pants were screened using a DSM-IV symptoms checklist, and diagnosis was confirmed following the DSM-IV-based Diagnostic Interview for ADHD in Adults (DIVA) and medical review from a trained consultant psychiatrist. The prevalence of ADHD in the prison was estimated to be 19%, of which 78% met criteria for the combined type presentation of ADHD. Significant pre-post treatment re- ductions, unadjusted for multiple outcome measures, were seen for investigator-rated ADHD symptoms using the Conners Adult ADHD Rating Scale (CAARS). g More generally, the benefits of treating ADHD with MPH are also expected to extend to a wide range of long- term outcomes relevant to young adult offenders with ADHD. The evidence for this comes largely from pharma- coepidemiological studies using within-individual compari- sons of periods on and off medication, to control for ‘treatment by indication’ effects. For example, suicidal be- haviour was found to be higher among adults treated for ADHD with stimulants compared to those who had not been treated, suggesting that stimulants might increase the risk of suicidal behaviour [15]. However, reductions in sui- cidality were found when periods when individuals were taking medication for ADHD were compared to periods Choice of intervention and comparator To address this problem, we are conducting a randomised controlled efficacy trial of osmotic-controlled release oral delivery system methylphenidate (OROS-MPH, Concerta XL), a sustained release formulation of methylphenidate, There are two main reasons why response of ADHD symptoms to stimulant medications may be different for young offenders compared to responses for previous Page 3 of 21 Page 3 of 21 Asherson et al. Trials (2019) 20:663 Asherson et al. Trials (2019) 20:663 compared to placebo, in young adult prisoners with ADHD who are aged between 16 and 25 years. when the same individuals were not taking medication. This suggests a protective effective on suicidality during pe- riods of taking stimulants, although individuals with ADHD and suicidality are more likely to be treated, presumably be- cause their ADHD is more severe. Using a similar pharma- coepidemiological design, a study using Swedish national registry data of 25,656 male patients with ADHD found a sixfold higher rate of criminal convictions over a 4-year period in patients with ADHD compared to controls. Re- garding medication effects, they found a 32% reduction in the risk of criminal convictions, using both within and be- tween methods of analysis, to compare periods on and off medication for ADHD. Furthermore, these protective ef- fects on criminal convictions were only seen for ADHD medications (stimulants or atomoxetine) but not for com- monly prescribed antidepressants, indicating the specificity of these findings to ADHD medications [12]. Other out- comes identified using a similar methodology include vio- lent reoffending on release from prison [16], depression [17] and risk of serious transport accidents [18]. Trial objectives Th The primary objective is to establish the efficacy of OROS-MPH in reducing ADHD symptoms (inattention and hyperactivity-impulsivity) in young male offenders aged 16–25 who meet diagnostic criteria for DSM-5 ADHD. Potential risks One often-raised concern is the potential for abuse of prescribed MPH, particularly in a population of of- fenders with ADHD and high rates of substance abuse. Stimulants can be abused by crushing short-acting for- mulations such as immediate release MPH, which can then be insufflated (snorted) or injected, leading to a rapid entry of the drug into the brain and the experience of euphoria. However, when taken orally, the slow phar- macokinetic profile does not lead to euphoria [22]. This is important, because it is not possible to crush the trial medication (OROS-MPH, Concerta XL) or easily extract MPH for injection. Risk of diversion or abuse is there- fore reduced in this study using this formulation of MPH. Furthermore, in our pilot study we did not ob- serve excessive drug-seeking behaviour for stimulant medication [14]. The young adult offenders being treated for ADHD were generally cautious about increasing the dose of medication and were titrated to modest doses, comparable to community samples (18% used 18 mg daily, 37% used 36 mg, 14% used 54 mg, 26% used 72 mg and only 4% used 90 mg). There are standard operating procedures for the delivery of controlled drugs within the prisons. Other potential risks are the usual range of adverse effects observed when treating ADHD with OROS-MPH. Secondary objectives include evaluating the following: reductions in emotional dysregulation, the number of adjudications for antisocial behaviour and rule breaking in the previous 8 weeks, ratings of aggressive and/or dis- ruptive behaviour by prison officers and education staff, attitudes towards violence, self-report of well-being. Additionally, we intend to investigate the hypothesis that improvements in secondary behavioural outcomes are mediated by improvements in ADHD symptoms or emotional dysregulation. Study setting Participants are recruited from two prisons. The first is HMP & YOI Isis in London (England, UK), a prison for sentenced young adults and category C offenders, de- fined as those who cannot be trusted in open conditions but who are unlikely to try to escape. The second is HMYOI Polmont in Falkirk (Scotland, UK), a holding fa- cility for young offenders aged 16–21, with sentences ranging from 6 months to life. All participants were sen- tenced prisoners when screened for entry into the trial. Potential benefits Potential benefits of treating young adult offenders with ADHD with MPH include improvement in clinical and Asherson et al. Trials (2019) 20:663 Asherson et al. Trials (2019) 20:663 Page 4 of 21 behavioural outcomes. These include ADHD symptoms, emotional dysregulation, attitudes towards violence, crit- ical incidents and engagement with educational and re- habilitation programmes. Demonstrating efficacy and safety of MPH on ADHD symptoms and associated im- pairments may provide the data needed to develop ef- fective healthcare pathways, including the use of MPH, for a significant group of young offenders. Establishing efficacy of MPH in this population will provide the foun- dation needed to establish long-term effectiveness stud- ies with the potential for demonstrating significant reductions in criminal behaviour and improved health- economic outcomes. the trial medication. Figure 1 illustrates the prisoner’s journey through the trial as the Consolidated Standards of Reporting Trials (CONSORT) diagram. This process will be completed following database lock for the trial. Table 1 is a summary of all trial procedures and assess- ments. Additional file 2 provides the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) checklist. Eligibility criteria Inclusion criteria For inclusion into the study, one must be male, aged be- tween 16 and 25 years (at consent for screening), English speaking (defined as sufficient to complete study assess- ments), able to provide informed consent (understand the information sheet and make an informed decision taking into account the pros and cons of study participa- tion) and meeting clinical diagnostic criteria for DSM-5 ADHD. Trial design The trial design is a parallel arm, randomised placebo- controlled trial of an extended release formulation of MPH (OROS-MPH, Concerta XL) on ADHD symptoms, behaviour and functional outcomes in young male pris- oners aged 16–25 who meet DSM-5 criteria for ADHD. Participants will be randomised to 8 weeks of treatment with either OROS-MPH or placebo, titrated over 5 weeks to balance ADHD symptom improvement against side effects. Two hundred participants will be recruited with a 1:1 ratio of drug to placebo. The duration of each participant’s follow-up is 8 weeks from the start date of The diagnostic criteria are defined as five or more symp- toms of ADHD in either the inattentive or hyperactive- impulsive symptom domains, and six or more symptoms of ADHD in either the inattentive or hyperactive-impulsive symptom domains before the age of 12 years. Where it is not possible to gain enough clinical information to score childhood symptoms of ADHD, the operational criteria Asherson et al. Trials (2019) 20:663 Page 5 of 21 Fig. 1 CONSORT diagram Page 6 of 21 Asherson et al. Trial design Trials (2019) 20:663 Table 1 Table caption Visit 2 3 4 Randomisation and Initiation of treatment 5 6 7 8 9 10 Consent, randomisation and initiation of treatment Informed consent X Eligibility checks X Randomisation and initiation of treatment X Treatment with OROS-MPH or placebo Treatment with OROS-MPH or placebo X X X X X X X Baseline (pre-randomisation), titration (visit 5-7), mediator (visit 9) and outcome (visit 10) measures DIVA (taken from pre-trial records) X MINI 7.0.1 diagnostic interview X MINI 7.0.1 cross disorder symptom ratings X X Demographic data X WASI-II (general cognitive ability) X RPAQ X CTQ X Drug and alcohol use (AUDIT-C and NIDA) X ZAN-BPD X CAARS –O X X X X X X X AES X X X X X X X Blood pressure and pulse X X X X X X X WRAADS X X X ARI-S X X X MEWS X X X Demographic data X MVQ X X BSI X X CGI X X Weiss-CD X Weight X X X Number of critical Incidents (adjudications)b X X MOASP X X MOASE O O BRCP X X BRCE O O ( ) g X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X O X O 5 6 7 8 9 10 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X O X O Visit 2 3 4 Rando Consent, randomisation and initiation of treatment Informed consent X Eligibility checks X Randomisation and initiation of treatment X Treatment with OROS-MPH or placebo Treatment with OROS-MPH or placebo X Baseline (pre-randomisation), titration (visit 5-7), mediator (visit 9) and outcome (visit 10) measures DIVA (taken from pre-trial records) X MINI 7.0.1 diagnostic interview X MINI 7.0.1 cross disorder symptom ratings X Demographic data X WASI-II (general cognitive ability) X RPAQ X CTQ X Drug and alcohol use (AUDIT-C and NIDA) X ZAN-BPD X CAARS –O X AES X Blood pressure and pulse X WRAADS X ARI-S X MEWS X Demographic data X MVQ X BSI X CGI X Weiss-CD X Weight X Number of critical Incidents (adjudications)b X MOASP X MOASE O BRCP X BRCE O Page 7 of 21 Asherson et al. Trial design Trials (2019) 20:663 p ( ) Visit 2 3 4 Randomisation and Initiation of treatment 5 6 7 8 9 10 Educational engagement (proportion of education sessions attended)b O O Positive and negative IEPs (HMP ISIS only)b X X CORE-M X X Concomitant medications and compliancea X X X X X X X X Prescribed dose of trial medication and compliancea X X X X X X Withdrawal status X X X X X X Study medication guess X X indicates a measure which should always be recorded; O indicates a measure completed for a subset of participants participating in education and related activities; ataken from prescription records; btaken from Abbreviations: DIVA Diagnostic interview for adult ADHD, MINI-7.0.1 MINI international psychiatric interview for common mental health disorders, WASI-II Wechsler Abbreviated Scale of Intelligence-II, RPAQ Reactive proactive aggression questionnaire, CTQ Childhood trauma questionnaire, AUDIT-C Alcohol use screening test, NIDA NIDA quick drug screen, ZAN-BPD Zanarini borderline personality disorder scale, CAARS-O Observer rated DSM-IV ADHD, AES Adverse events scale, WRAADS Wender-Reimherr Adult ADHD Diagnostic Scale, ARI-S Affective reactivity index, MEWS Mind Excessively Wandering Scale, MVQ Maudsley violence questionnaire, BSI Brief symptom inventory, CGI Clinical global impression scale, Weiss-CD Conduct disorder scale, MOASP Modified overt aggression scale by prison officers, MOASE Modified overt aggression scale by education staff, BRCP Behaviour report cards by prison officers, BRCE Behaviour report cards by education staff, IEPs Incentive and Earned Privileges, CORE-M CORE outcome measure Page 8 of 21 Asherson et al. Trials (2019) 20:663 Page 8 of 21 Asherson et al. Trials (2019) 20:663 Marketing Authorisation for both HDPE and blister packaging. Placebo tablets are manufactured once. Trial medication over-encapsulation and packaging is under- taken in two campaigns in order to accommodate a trial duration of up to 3.5 years. Concerta 18-mg tablets typ- ically have a maximum shelf life of 3 years from the date of manufacture; however, by the time the product is repacked for the clinical trial, the remaining shelf life is likely to be under 2.5 years. applied have been adapted to include evidence of several ADHD symptoms with impairment starting before the age of 12 years, and five or more symptoms currently with moderate to severe impairment. Trial design In addition this includes persistent trait-like (non-episodic) course of symptoms, im- pairments in two or more clinical or psychosocial domains and two or more settings from symptoms of ADHD and onset of symptoms before the age of 12 years. y y Over-encapsulation uses ‘DBcaps’ capsules which are de- signed specifically for the blinding of clinical trial medica- tions. We have to over-encapsulate Concerta and placebo capsules with lactose capsules, rather than make a matching placebo capsule, because Concerta capsules have printing on them and are of a distinct shape that would be difficult to manufacture and might infringe copyright. We sought advice on this from previous investigators using OROS- MPH and from companies that provide drug and placebo supplies for studies. Studies on the use of DBcaps have shown that over-encapsulation of capsules results in a lag time of 2–3 min in disintegration compared with the unen- capsulated capsules. The pharmacokinetic properties of Concerta XL 18-mg prolonged release capsules indicate re- lease over several hours: following oral administration of Concerta XL to adults the drug overcoat dissolves, provid- ing an initial maximum drug concentration at about 1–2 h. The MPH contained in the two internal drug layers is grad- ually released over the next several hours. Peak plasma con- centrations are achieved at about 6–8 h, after which plasma levels of MPH gradually decrease (Section 5.2 of the Sum- mary of Product Charateritics: https://www.medicines.org. uk/emc/product/6872/smpv). Exclusion criteria b l A subject is excluded from the study if he meets any of the following criteria. He lacks capacity to give informed consent; he has a moderate or severe learning disability, defined as IQ < 60; he poses a serious risk of violence to the researcher; he has current major depression, psych- osis, mania or hypomania; he has a past history of bipo- lar disorder or schizophrenia (we exclude those with a clear history of episodic mania/hypomania or psychosis unrelated to acute drug intoxication, but do not exclude on the basis of chronic emotional dysregulation, i.e. irrit- ability, frustration, anger or emotional-mood instability). Subjects are also excluded if they have medical contrain- dications to the use of stimulants (e.g. glaucoma, hyper- tension, cardiovascular disease or a structural heart problem); have been taking contraindicated medications during the 4 weeks prior to randomisation; show drug- seeking behaviour or craving (defined as drug-seeking behaviour that is unusually severe and likely to affect the titration protocol due to unusual and excessive demands for drugs or where there is a current withdrawal syn- drome from an addiction disorder with drug depend- ency); receive any ADHD medication between consent for screening and randomisation. Trial medication Trial medication is delivered as prescribed daily, with participants observed to ensure they swallow the cap- sules. There is a daily record of compliance with the trial prescription. Both active medication and placebo are ti- trated in the same way. Treatment starts at an initial dose of 18 mg (1 tablet) for 1 week and is then increased weekly over the following 4 weeks, in increments of 18 mg, to a maximum of 72 mg (4 tablets). Medication is reduced by 18 mg (1 tablet) if there is a limiting adverse event, in which case there will be no further increase in medication for the duration of the trial. Medication may be provided either once or twice daily up to the max- imum daily dose. Titration upwards will be stopped if all 18 ADHD symptoms are scored as negligible (score of 0 or 1 on the CAARS) or absent. Unacceptable levels of adverse effects on the lowest dose of 18 mg might lead to a cessation of treatment in a few cases. OROS-MPH (Concerta XL) is supplied as 18-mg capsules and placebo to match. Capsules are over-encapsulated and packaged in bottles of 46. Each bottle is assigned a unique randomisation number; the randomisation system allocates the right bottle to each patient. Over-encapsulation has been successfully adopted in previous studies to generate matched placebo to OROS-MPH. Piramal Healthcare UK Ltd. supply the investigational medicinal product (IMP), placebo to match manufacture, clinical trials packaging, Qualified Person (QP) Certification and distribution for 200 patients. The Sponsor arranged the supply of Concerta 18- mg tablets from the Marketing Authorisation Holder, Janssen-Cilag Limited. Janssen-Cilag Limited provides the summary of product characteristics (SmPC), updated throughout the duration of the study. The over-encapsulated active tablets are repacked in high-density polyethylene (HDPE) bottles and exceed the remaining shelf life of the study without the need for a stability programme, as Concerta 18 mg has a A maximum dose of 72 mg was included for this trial because previous clinical trials indicated that a proportion of adults respond better at higher doses without Asherson et al. Trials (2019) 20:663 Asherson et al. Trials (2019) 20:663 Page 9 of 21 unacceptable levels of adverse events, and because current licensing for Concerta XL up to 54 mg is based on dose levels for children and adolescents, rather than adults. Trial medication NICE recommends a daily dose of MPH in adults to a maximum of 100 mg per day [7], and for Concerta XL the British National Formulary (No 62, September 2011) rec- ommends doses up to a maximum of 108 mg in adults. were always out of their cells on time to receive trial medication. 4. In the pilot study, treatment was disrupted for the Ramadan festival for several participants. We will take care to check that participants are not started on trial medication where religious customs might interfere with adherence to the trial protocol. 4. In the pilot study, treatment was disrupted for the Ramadan festival for several participants. We will take care to check that participants are not started on trial medication where religious customs might interfere with adherence to the trial protocol. 5. In the pilot study, daily adherence to the trial medication reduced when participants were not reviewed weekly. One of the findings in the pilot study was the importance that prisoners gave to the weekly follow-up meetings when they can discuss their ADHD and response to the treatment process, in addition to completing study assessments. We will therefore offer weekly meetings with offenders throughout the 8-week trial. Strategies to improve adherence We envisaged that adherence with allocated medication will present a challenge for around 20% of participants. Some offenders may not feel motivated to take the trial medication if they experience adverse effects or do not feel they are improving. They may also take medication intermittently because of the strict prison regime that al- lows for only a short time window for leaving their cells to obtain medication from the medicine hatch on the prison wings. The cases involving these persons are not expected to contribute to missing data. In our pilot study we accrued considerable experience in managing the expectations of offenders and providing the support needed to help participants adhere to the trial protocol. The following steps will be adopted to maximise adher- ence to medication: g 6. Nurse support in addition to a research assistant and medical staff will ensure that offenders are given the support they need to adhere to the protocol. Concomitant treatments Concomitant treatments are allowed with medications that are not contraindicated with MPH. All concomitant medications are recorded in the study database. Use of the following medications in the 4 weeks prior to the start of treatment with Concerta XL will lead to exclu- sion from the clinical trial, based on potential adverse drug interactions: clonidine, coumarins, monoamine oxi- dase inhibitors, moclobemide and rasagline. 1. In the pilot, minor adverse effects (13%) were the most common reason for non-adherence to medication. This was linked to the observation that this population may be more sensitive to minor adverse effects, particularly changes in appetite, than community populations, perhaps reflecting the importance of meal times to prisoners. To maximise adherence to the protocol and minimise this as a potential source of missing data, we will take care to identify the early signs of minor adverse effects such as appetite loss and adjust the medication dose accordingly. Baseline and outcome measures The schedule of baseline and outcome measures as well as the procedures for the trial are listed in Table 1. Baseline- only measures are collected on all participants prior to randomisation and include descriptors of the study popu- lation and baseline moderators for further analysis as pre- dictors of the treatment response. Primary and secondary outcome measures are collected at baseline prior to ran- domisation and 8 weeks after the initial trial prescription. The primary outcome is the investigator-rated Conners Adult ADHD Rating Scale (CAARS-O) at the 8 weeks outcome. The other 5-week and 8-week measures are sec- ondary outcomes or mediator variables. All outcome mea- sures are listed in Table 2 along with their definitions. 2. Seven percent of the pilot sample did not wish to take medication in the mornings (08:00), which was the initial protocol followed in the pilot study. We then adjusted the protocol to allow for 12:00 medication for those who got up later in the day, worked mainly in the afternoons or had a strong preference for a 12:00 dosing, which resolved the problem. This flexibility in dosing time more accurately reflects dosing decisions in the community and provides a better match to patients’ daily routines. Investigator-rated measuresThe following investigator- rated meaures are used in the study: Investigator-rated measuresThe following investigator- rated meaures are used in the study: 3. During the pilot study, prison staff did not always let patients out of their cells to receive medication or remind participants to get up on time. To resolve this problem, we initiated the use of research staff whenever possible to assist in the delivery of medication by checking that prisoners DIVA v2.0: DSM-IV-based Diagnostic Interview for ADHD in Adults [23]. DIVA 2.0. is a semi- structured interview assessment used to capture the diagnostic symptoms and criteria for DSM-IV Page 10 of 21 Asherson et al. Investigator-rated measuresThe following investigator- rated meaures are used in the study: Trials (2019) 20:663 Table 2 Primary and secondary outcome measures Concept Instrument Time points for analysis model Type of measure Summary measure ADHD symptoms Conners Adult ADHD rating scale (CAARS), investigator rated M (B, 4, 5, 8) Continuous scale Mean difference Emotional dysregulation Wender-Reimherr Adult ADHD Diagnostic Scale (WRAADS), investigator rated M (B, 5, 8) Continuous scale Mean difference Irritability Affective Reactivity Index (ARI), self-rated M (B, 5, 8) Continuous scale Mean difference Spontaneous mind wandering Mind Excessively Wandering Scale (MEWS), self-rated M (B, 5, 8) Continuous scale Mean difference Attitudes towards violence Maudsley Violence Questionnaire (MVQ), self-rated M (B, 5, 8) Continuous scale Mean difference Common psychopathological symptoms Brief Symptom Inventory (BSI), self-rated M (B, 5, 8) Continuous scale Mean difference Global impression of disease severity Clinical Global Impression (CGI) scale, clinician rated M (B, 5, 8) Categorical Mean difference Behavioural problems recorded by prison officers in prison records Number of critical incidents recorded in prison records Incident rate (B, 8) Incident rate over 8-week period Incidence rate ratio Prison office ratings of aggressive behaviour Modified Overt Aggression Scale (MOASP), Prison Officer report M (B, 8) Continuous Mean difference Educational staff ratings of aggressive behaviour Modified Overt Aggression Scale (MOASE), Education Staff report M (B, 8)a Continuous Mean difference Prison officer ratings of behaviour Behaviour Report Card (BRCP), Prison Officer report M (B, 8) Continuous Mean difference Educational ratings of behaviour Behaviour Report Card (BRCE), Educational Staff report M (B, 8)a Continuous Mean difference Engagement with the educational programme Number of educational sessions attended over 8-week period Incident rate (B, 8)a Incident rate over 8-week period Incidence rate ratio Incentive points for rewarding behaviour Number of Incentives and Earned Privileges (IEPs) Incident rate (B, 8) Incident rate over 8-week period Incidence rate ratio Current psychological distress CORE Outcome Measure (CORE-OM), self-rated M (8) Continuous Mean difference Notes on time points for analysis: Current psychological distress Notes on time points for analysis: Numbers in brackets refer to the assessment time points of measures included in the model (B = baseline, 4 = week-4 data, 5 = week-5 data, 8 = week-8 data) For incident rate variables (e.g. Investigator-rated measuresThe following investigator- rated meaures are used in the study: number of behavioural problem reports) the baseline data are from the 8 weeks prior to randomisation; for the 8-week outcome the data are from the 8-week period from the start of medication aIndicates a measure completed for the subset of participants participating in education and related activities Numbers in brackets refer to the assessment time points of measures included in the model (B = baseline, 4 = week-4 data, 5 = week-5 data, 8 = week-8 data) For incident rate variables (e.g. number of behavioural problem reports) the baseline data are from the 8 weeks prior to randomisation; for the 8-week outcome the data are from the 8-week period from the start of medication p aIndicates a measure completed for the subset of participants participating in education and related activities ADHD. The diagnostic algorithm applied to these data was modified for DSM-5 criteria. ADHD. The diagnostic algorithm applied to these data was modified for DSM-5 criteria. ADHD. The diagnostic algorithm applied to these data was modified for DSM-5 criteria. cross-disorder symptom checklist scores at base- line and at the 8 weeks assessments. cross-disorder symptom checklist scores at base- line and at the 8 weeks assessments. MINI 7.0.1: Mini International Neuropsychiatric Interview for comorbid mental health disorders [24]. MINI 7.0.1 is a semi-structured interview assessment used to capture DSM-IV diagnostic criteria for common mental health disorders. Sections completed included the following: major depressive episode, suicidality, manic episode, hypomanic episode, panic disorder, agoraphobia, social anxiety disorder, obsessive-compulsive disorder, post-traumatic stress disorder, psychotic disorder and mood disorder with psychotic features, generalised anxiety disorder, antisocial personality disorder. In addition to diagnostic categories evaluated at baseline only, we collected MINI 7.0.1: Mini International Neuropsychiatric Interview for comorbid mental health disorders [24]. MINI 7.0.1 is a semi-structured interview assessment used to capture DSM-IV diagnostic criteria for common mental health disorders. Sections completed included the following: major depressive episode, suicidality, manic episode, hypomanic episode, panic disorder, agoraphobia, social anxiety disorder, obsessive-compulsive disorder, post-traumatic stress disorder, psychotic disorder and mood disorder with psychotic features, generalised anxiety disorder, antisocial personality disorder. In addition to diagnostic categories evaluated at baseline only, we collected ZAN-BPD: Zanarini Rating Scale for Borderline Personality Disorder [25]. A validated scale for the assessment of borderline personality disorder, used as a baseline moderator variable. CAARS-O: Conners Adult ADHD Rating Scale for ADHD symptoms [26]. Investigator-rated measuresThe following investigator- rated meaures are used in the study: The 8 weeks CAARS-O assessment is the primary outcome measure for this study. CAARS-O was also used as a secondary outcome at week 5 and to assist the psychiatrist in titrating participants onto the optimal trial medication dose. CAARS-O consists of the 18 DSM-IV ADHD symptoms, rated on a 4-point Likert scale (0 not at all, never; 1 just a little, once in a while; 2 pretty often; 3 very much, frequently). This scale and other closely CAARS-O: Conners Adult ADHD Rating Scale for ADHD symptoms [26]. The 8 weeks CAARS-O assessment is the primary outcome measure for this study. CAARS-O was also used as a secondary outcome at week 5 and to assist the psychiatrist in titrating participants onto the optimal trial medication dose. CAARS-O consists of the 18 DSM-IV ADHD symptoms, rated on a 4-point Likert scale (0 not at all, never; 1 just a little, once in a while; 2 pretty often; 3 very much, frequently). This scale and other closely Page 11 of 21 Asherson et al. Trials (2019) 20:663 similar scales have been extensively validated as outcome measures in previous clinical trials of adult ADHD. similar scales have been extensively validated as outcome measures in previous clinical trials of adult ADHD. mind wandering, an aspect of psychopathology that is closely associated with ADHD and a strong predictor of ADHD-associated impairment in daily life. p WRAADS: Emotional dysregulation measured from the Wender-Reimherr Adult ADHD Diagnostic Scale [27]. We applied the emotional dysregulation items from an interview assessment of the WRAADS-ED, following previous publications on the treatment response of emotional symptoms in ADHD [20, 28]. Brief Symptom Inventory (BSI) [38]. A self-rating scale that captures comorbid symptoms. Subscales include nine symptom dimensions: somatisation, obsession‐compulsion, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, psychoticism. AES: Adverse Events Scale [29]. Scale of common adverse effects associated with stimulant medications for ADHD used with permission from the CADDRA website. Maudsley Violence Questionnaire (MVQ) [39]. This scale was designed to capture beliefs associated with violence. The Machismo subscale relates to embarrassment over backing down and justification of violence in response to threat and attack; the Acceptance subscale includes the overt enjoyment and acceptance of violence in everyday life. In previous research the Machismo subscale showed the greater relationship to actual violence [39] and the greater reduction in our pilot study for this trial [14]. Investigator-rated measuresThe following investigator- rated meaures are used in the study: CGI: Clinical Global Impression scale [30]. Scale used by the research psychiatrist to give an overall rating of clinical severity and a clinical impression of the clinical response and adverse effects of the trial medication. Participant self-rating scales Self-rating scales are given to the participants for self- completion. The scale questions are usually read out to participants who give their response accordingly. The self-rating scales are: g p y CORE Outcome Measure (CORE-OM) [40]. This scale captures subjective well-being, problems and symptoms, life functioning and risk and harm. It is designed to measure psychological distress before and after treatment. Reactive-Proactive Aggression Questionnaire (RPAQ) [31]. This scale is included as a baseline moderator capturing proactive and reactive forms of aggression. Data from prison records and prison staff Data from prison records and prison staff Data will be collected from prison records and prison nurs- ing and educational staff relating to behaviour in the 8 weeks before the collection of the baseline measures. For cases of individuals new to custody presenting with signifi- cant behavioural problems, the retrospective baseline report- ing period will be for a period of 1 month or more, to allow for initial behavioural problems that may arise when people first enter prison. The data will include: gg Weiss CD: Weiss conduct disorder scale. This scale is included to capture conduct disorder symptoms as a baseline moderator. Alcohol Use Disorders Identification Test (AUDIT-C) and National Institute on Drug Abuse (NIDA) quick screen. Alcohol and substance abuse checklist using the AUDIT-C and NIDA quick screen to capture drug and alcohol use in year prior to the current prison sentence. The NIDA quick screen was adapted from the single-question screen for drug use in primary care by Saitz and colleagues [32]. AUDIT-C is validated as a quick screen for alcohol use [33]. Number of adjudications for antisocial behaviour and rule breaking (HMP Isis and HMP/YOI Polmont) and negative Incentives and Earned Privileges (IEPs) (HMP Isis only) Childhood Trauma Questionnaire (CTQ) [34]. Included as a potential moderator of the clinical response to MPH. g y MOASP: Ratings of aggressive behaviour by Prison Officers using the Modified Overt Aggression Scale (MOAS) [41] Barkley ADHD (B-ADHD). Self-rating scale for DSM-IV ADHD symptoms [35] is as an initial screening instrument. Participants are considered to screen positive for ADHD if they have four or more symptoms scoring 2 or more in either the inattentive or hyperactive/impulsive symptom domain. BRCP: Ratings of behaviour by prison staff using Behaviour Report Cards [42] by Prison Officers MOASE: Ratings of aggressive behaviour by Education Staff using the MOAS. This item is optional, depending on whether prisoners are attending education sessions or not g BRCE: Classroom Behaviour Report Card scored by Education Staff (HMP Isis and HMP/YOI Isis). This item is optional, depending on whether prisoners are attending education sessions or not Affective Reactivity Index (ARI-S) [36]. A self-rating scale for irritability. Mind Excessively Wandering Scale (MEWS) [37]. A self-rating scale that captures excessive spontaneous Page 12 of 21 Page 12 of 21 Page 12 of 21 Asherson et al. Participant timelines A schedule of participant visits is illustrated in Fig. 2. p A schedule of participant visits is illustrated in Fig. 2. Research visits Following consent 1 for screening, and confirmation of the diagnosis of ADHD and eligibility by a psych- iatrist trained in the assessment of ADHD, informa- tion sheets and consent forms for the controlled trial (consent 2) will be provided and discussed with po- tential participants (visit 1). Information sheets will be reviewed and informed consent obtained for the clin- ical trial (visit 2). There is no limit on the time taken Primary outcome measure There are two stages of consent. Initial consent 1 (screen- ing and diagnostic step) allows for the use of screening questionnaires for ADHD, followed by a diagnostic assess- ment using the DIVA interview for adult ADHD and re- view by a trained psychiatrist. During these pre-trial steps, patients who fail eligibility criteria will not be invited to continue and will not be asked to provide consent 2, to participate further in the trial (clinical trial step). The eligi- bility criterion which they are identified as failing will be noted. Individuals who do meet the diagnostic and eligibil- ity criteria are invited to take part in the clinical trial, at which stage informed consent is requested to take part in the randomised controlled trial. Consent 2 is taken by the trial psychiatrists. The primary endpoint is the level of ADHD symptoms measured on the investigator-rated CAARS-O at 8 weeks post-treatment initiation to address the question of effi- cacy of OROS-MPH on ADHD symptoms in young of- fenders meeting DSM-5 diagnostic criteria for ADHD. Investigator-rated CAARS-O scores are a common out- come measure used in previous treatment trials of ADHD in the community; the CAARS-O score measures the same list of 18 symptoms used as the primary out- come in nearly all other studies of adult ADHD. Data from prison records and prison staff Trials (2019) 20:663 IEPs: Number of positive IEPs for positive engagement in education, occupational and rehabilitation programmes (HMP Isis only). IEPs: Number of positive IEPs for positive engagement in education, occupational and rehabilitation programmes (HMP Isis only). at baseline: CAARS-O hyperactive/impulsivity and in- attention subscores and WRAADS for emotional dysreg- ulation. These measures are hypothesised to mediate treatment response in terms of secondary behavioural outcomes (critical incidents and prison staff classroom report cards). Critical incidents are taken from the prison records at 8 weeks and are recorded over the pre- vious 4 weeks. The prison staff classroom report cards are recorded at 8 weeks and record behaviour over the preceding week. at baseline: CAARS-O hyperactive/impulsivity and in- attention subscores and WRAADS for emotional dysreg- ulation. These measures are hypothesised to mediate treatment response in terms of secondary behavioural outcomes (critical incidents and prison staff classroom report cards). Critical incidents are taken from the prison records at 8 weeks and are recorded over the pre- vious 4 weeks. The prison staff classroom report cards are recorded at 8 weeks and record behaviour over the preceding week. Secondary outcome measures Secondary outcomes address important questions about the effects on comorbid symptoms and behavioural im- pairments that are commonly seen in offenders with ADHD. These are critical incidents (adjudications) from prison records for the 8-week period (in two 4-week pe- riods) from initiation of the trial medication to the 8 weeks assessments; ratings of aggressive behaviour by prison staff using the MOASP at 8 weeks; BRCP behav- iour report cards from prison staff at 8 weeks; engage- ment with educational activities (including number of scheduled educational sessions, proportion of scheduled educational sessions attended and reports of disruptive behaviour in education session reported at 8 weeks using the BRCE and MOASE completed by education staff— only for those people involved in education); attitudes towards violence using the MVQ at 8 weeks, CORE-OM at 8 weeks; general psychopathology using the BSI at 8 weeks; excessive mind wandering measured using the MEWS at 8 weeks; symptoms of emotional dysregulation measured using the WRAADS at 8 weeks; symptoms of emotional dysregulation measured using ARI at 8 weeks; overall health measured using CGI at 8 weeks. Informed consent for the trial (consent 2) is obtained by providing the participants with written information sheets and ensuring that they understand the information pro- vided and the implications of taking part in the research, prior to obtaining signed consent (see Additional file 1 for a copy of the information and consent forms). They are offered the opportunity to be treated for ADHD as part of the trial, or alternatively they are offered treatment outside of the trial. The participants have as long as they like to decide while the trial is ongoing. Patients complete baseline measures after providing informed consent to take part in the trial (consent 2). Once the baseline assessments are complete and eligibil- ity checks completed and documented, the individuals will be randomised to one of the treatment arms. Baseline measures The following measures are recorded at baseline. CAARS-O; WRAADS (three subscores: temper, affective lability, emotional over-reactivity), weight, pulse, blood pressure, Weiss CD, IQ (Wechsler Abbreviated Scale of Intelligence [WASI]), DIVA score and ADHD diagnosis, ZAN-BPD score, drug use in lifetime and alcohol use in the past year, RPAQ (two subscales: reactive, proactive), MEWS, CTQ, MVQ, CORE-OM, CGI, ARI-S, concomi- tant medications, BSI, MINI 7.0.1 and AES. Mediator measures To address the secondary mediation hypotheses, the fol- lowing putative mediators are recorded at 5 weeks and Asherson et al. Trials (2019) 20:663 Page 13 of 21 Fig. 2 Schematic diagram of flow of participants (potential and actual) through the pre-trial assessment and trial One week after the start of trial medication, the partic- ipants are reviewed, and trial medication titrated accord- ing to their clinical response and adverse effect profile (visit 5, week-1 titration). Symptoms of ADHD are mea- sured using the CAARS-O, adverse events checked using the AES and blood pressure and pulse checked. This ti- tration procedure is repeated at weeks 2–4 (visits 6–8). Five weeks after the start of medication (visit 9, week-5 assessment) the prescription (the titrated dose) is con- firmed and maintained for the rest of the trial. At the week-5 assessment, outcome measures are completed by a research investigator for the CAARS-O, WRAADS and between visits 1 and 2 within the timeframe of the project. Potential participants will be encouraged to take as much time as they need to reach a fully in- formed decision about participation in the trial. Base- line data will be collected from participants, prison records and members of staff (visit 3). Once baseline data have been collected and eligibility confirmed fol- lowing medical review by a psychiatrist, participants will be randomised to treatment with placebo or OROS-MPH (visit 4). Trial prescriptions will be com- pleted and given to the pharmacy. Medication should start within 1 week of visit 4 (i.e. randomisation). Asherson et al. Trials (2019) 20:663 Asherson et al. Trials (2019) 20:663 Page 14 of 21 Page 14 of 21 Following clinical review, patients who meet diagnostic criteria for ADHD and the other eligibility criteria for the trial will be invited to take part in the clinical trial. MEWS and the pulse, blood pressure, weight and AES. The final visit 10 is completed 3 weeks later after 8 weeks from the first prescription of the trial medication. At this visit all outcome measures are completed. As far as possible the information on clinical response derived during the titration visits (weeks 1–4) is not shared with other members of the research team, particularly with the investigator completing the weeks 5 and 8 outcome measures. Thus, potential unblindings based on the ob- served clinical response and adverse events will be minimised. Mediator measures p Eligibility for the study will be further checked and re- corded once the consent form (consent 2) has been signed and baseline assessments have been completed, prior to randomisation. Using an algorithm that applies the DSM-5 criteria to the DIVA interview data, the po- tential participants will be checked to ensure they meet diagnostic criteria for DSM-5 ADHD. A clinical review by a psychiatrist trained in the diagnostic assessment of ADHD will review all inclusion and exclusion criteria. The exclusion criteria of IQ less than 60 will be based on the 95% CI for the IQ estimate from the WASI-II in- cluding IQ of 60, in combination with a clinical assess- ment by the psychiatrist to confirm that the participant has the ability to understand the rating scale and inter- view assessment questions; understand the information sheet and the study procedures and risks; and the ability to provide sufficiently detailed accounts of ADHD symp- toms and behaviours, consistent with an IQ score greater than 60. Since there are no validated IQ tests for the visually impaired, including WASI-II, this criterion will be based on clinical judgement alone for participants with this impairment. This will also be the procedure for anyone unable to complete the WASI-II assessment due to severity of their ADHD symptoms or other mental health problems. Withdrawal of subjects Participants have the right to withdraw from the study at any time for any reason, and healthcare staff have the right to withdraw patients from the trial if they consider the trial is having an adverse effect on the participants. However, where participants discon- tinue taking trial medication, we will invite them to remain in the study to complete trial assessments, thereby minimising loss of data. Should a participant decide to withdraw from the study, all efforts will be made to report the reason for withdrawal as thor- oughly as possible. A 25% loss is expected to be easily achievable, since in the pilot 10% left the prison due to unexpected transfers from the prison, and problems with adherence to trial medication were rarely followed by problems completing trial assessments. Due to potential concerns about the interaction of trial medication with unknown psychoactive sub- stances, if a participant discloses to any member of the research team that he has used ’spice’, i.e. syn- thetic cannabis or another unknown psychoactive sub- stance, while participating in the study, a clinical evaluation will be made. If it is current use (defined as within the last 2 days), the study medication will be stopped. If it happened earlier in the study and is considered an isolated incident, the trial medication can continue. If the trial medication is stopped, the participant will remain in the study and will be asked to complete trial assessments. A clinical assessment The total sample size to be randomised is 200. The total sample size to be randomised is 200. The primary outcome is ADHD symptoms, measured using CAARS-O. The results of a single-arm open label pilot study of young prisoners with ADHD who were given MPH showed a mean decrease of 25.0 points with a standard deviation of 9.1 [14]. This suggested a stan- dardised effect size of SMD = 2.75. It could reasonably be assumed that at least 20% of this effect might be at- tributed to the effects of MPH. On this basis, this study is powered to detect a standardised effect size of d = 0.55. Assuming a standard deviation of 9.1, this would translate into a treatment difference of 5.0 points. This effect size is consistent with the results of a recent meta- regression analysis (12), which estimated the effect of treatment to be SMD = 0.49 (95% CI 0.08, 0.64). The sample size calculation used G*Power version 3 and was based on the use of a t test to compare the means of the treatment groups. In order to have 90% power at the 5% significance level to detect a standardised effect of SMD = 0.55, this study would need to collect outcome data on 142 participants. Inflating for the expectation that loss to follow-up may be as high as 25%, a mini- mum of 190 participants should be recruited, with the target for the study set at 200. Blinding l d Blinding is maintained for all study investigators, in- cluding the on-site researchers, pharmacy, statistical and data management teams. Investigators will be un- blinded after the primary analysis is complete. The primary analysis dataset will not include any trial medication dosage data to ensure that the statistician remains blinded. We do however propose a sensitivity analysis to assess efficacy for those complying with tablets offered. This analysis will exclude those partic- ipants who took no trial medication on less than 75% of the days on which it was prescribed. Additionally, persons who withdraw from treatment or the trial or are released, transferred or deported will be excluded. We will not consider what proportion of the pre- scribed medication was taken on any given day. Recruitment procedures Participants will be recruited from HMYOI Isis (London) and HMYOI Polmont (Falkirk). Following consent to be screened for ADHD (consent 1), screening questionnaire data will be collected by the prison mental health teams using a DSM-IV ADHD symptom rating scale (25). Patients who screen positive will be invited to complete the DIVA [27]. This will be followed by a clin- ical review by a psychiatrist trained in the diagnostic as- sessment of ADHD, including collateral information obtained from an informant whenever feasible. Page 15 of 21 Asherson et al. Trials (2019) 20:663 Asherson et al. Trials (2019) 20:663 unblinded at this point, but investigators will remain blind until the primary analysis is complete. The In- vestigator must report all code breaks (with reason) as they occur on the case report form. will be made on a case-by-case basis as to the safety of restarting the trial medication after 48 h from the time of stopping the trial medication. Randomisation and allocation concealment Randomisation to OROS-MPH or placebo will be at a 1:1 ratio. Randomisation is at the participant level and is performed using the King’s Clinical Trials Unit (CTU) independent Randomisation Service, ensuring reliability and credibility in the randomisation process, with blinding of both investigators and participants. Randomisation is stratified by prison with variable block sizes to ensure that equal numbers of patients are allocated to the two arms within each prison stratum. Patient characteristics will not be considered in the randomisation process. However, we expect the drug treatment and placebo trial arms to be balanced in terms of cognitive ability, ADHD symptom severity and co-occurring psychosocial and mental health problems. Emergency unblinding g y g Emergency unblinding will follow the standard operating procedures for the King’s Health Partners Clinical Trials Office (KHPCTO). In circumstances where unblinding is deemed necessary, the starting point will always be the local investigating team. Whenever possible the decision to unblind will be made by the Chief Investigator, the Principal Investigator or clinically qualified staff working on the project. Out of hours, if clinically qualified mem- bers of the research team are not available, then the 24- h Emergency Scientific and Medical Services (ESMS) system will be used. The ESMS system consists of a call centre which is manned around the clock by information scientists who have a minimum qualification of a life sci- ence degree to include toxicology or pharmacology. These information scientists are always available and are the direct line of communication to the number on the patient card. The information scientists will be trained in the specific details of this study and have direct access to one of the ESMS consultant physicians should clinical advice be required. Our consultant physicians practice general and internal medicine and specialise in clinical pharmacology and toxicology, ensuring clinical advice is available night and day. To maintain the overall quality and legitimacy of the clinical trial, code breaks will occur only in exceptional circumstances when knowledge of the actual treatment is absolutely essential for further management of the patient. The Investigator will always maintain the blind as far as possible. Prescriptions are completed by the trial psychiatrist. Each patient is allocated a kit (labelled carton) contain- ing four labelled bottles, each containing 46 active or placebo tablets. Each kit and its bottles will be labelled according to Annex 13 guidelines and have its own ran- domisation/treatment pack number. The centralised ran- domisation system will allocate the correct treatment pack/kit to each patient during the trial. Safety checks arm, dummy variables coding assessment time point (4, 5 or 8 weeks) and trial arm x time point interaction terms as explanatory variables. Random effects that vary at the par- ticipant level will be used to model the covariance struc- ture between the repeated measures. The approach will ensure that a different trial arm effect can be estimated at 4, 5 or 8 weeks—with the estimated effect at 8 weeks pro- viding the evaluation of OROS-MPH efficacy in terms of the primary outcome. Similar models will be used to evaluate continuous secondary outcomes. Patients are monitored daily by the prison mental and healthcare teams. Safety checks will be conducted in line with NICE Guidelines (2009). Regarding the research aspect of the study (i.e. obtain- ing follow-up data), there is little risk to participant safety. Participants will be aware that should they wish to withdraw from the study they may do so. Participants who become upset or distressed by the questions in the research (this is unlikely, as the questions are similar to those asked regularly in the context of their clinical care) will be offered support by the researchers and by the prison mental health team. The secondary count outcomes at 8 weeks (e.g. number of critical incidents) will be compared between treatment arms using Poisson regressions to estimate incidence rate ratios (after conditioning on baseline counts and random- isation stratifiers). Logistic regression will be carried out for scheduled educational sessions attended. Parameters will be estimated using maximum likelihood. The healthcare team will follow national guidelines on safety, which is predominantly related to monitoring of cardiovascular function. More specifically, the clinical care will follow these procedures: Inferences will remain valid in the presence of missing data provided that the missing data generating mechan- ism is MAR. More specifically, this particular MAR as- sumption stipulates that only variables included in the analysis model drive missingness. While we model sev- eral time points simultaneously, inferences will be made only at the time point of interest (8 weeks). Using linear mixed models means that we can allow variables mea- sured and included in the model (e.g. previously ob- served values of the outcome including baseline values, trial arm, stratifier and post-treatment time point) to predict attrition, and allows us to make use of all avail- able data. Statistical analyses A d l d A detailed statistical analysis plan has been developed by the trial statisticians in collaboration with the Chief Investi- gator and approved by the Trial Steering Committee (TSC). Analyses will be carried out by the trial statistician (RH) and checked by the senior statistician co-investigator (SL). In the first instance data will be analysed under intention-to-treat assumptions; that is, participants will be analysed in the groups to which they were rando- mised irrespective of treatment received. Efficacy will be assessed by comparing primary and secondary outcomes between the OROS-MPH and placebo arms. We intend to use linear mixed modelling, which as- sumes that only variables included in the model pre- dict missingness. We will assess empirically whether this particular missing at random (MAR) assumption is reasonable, using an independent statistician to maintain blinding if necessary. If the assumption is not reasonable, multiple imputation will be used in- stead to accommodate the missing data generating process, and the statistician might need to become In order to assess the efficacy of the continuous primary CAARS-O outcome (see Table 2), a linear mixed model will be used. The model will contain CAARS-O scores from the last three (4, 5 and 8 weeks) post-randomisation time points as the dependent variable and baseline CAARS-O values, randomisation stratifier (prison), trial Asherson et al. Trials (2019) 20:663 Asherson et al. Trials (2019) 20:663 Page 16 of 21 Page 16 of 21 Safety checks We will also check empirically whether with- drawal from allocated treatments is predictive of missingness at 8 weeks. And if we find that such post-randomisation vari- ables drive missingness, we will consider using multiple im- putation to accommodate such a MAR process. Mediation analysis using structural equation modelling will be used to partition the total treatment effect into mediated and non- mediated components. 1. Before commencing treatments, checks will be made on pulse and blood pressure and review of healthcare records. 2. Potential cardiovascular abnormalities will be evaluated for risk and, if necessary, an opinion will be obtained from a cardiologist prior to commencing treatment. 3. The clinical team will check pulse and blood pressure once a week for the first 5 weeks and at the end of the 8-week trial. Other safety checks will include monitoring of adverse events during assessments. In addition, participants will be monitored daily by prison staff, and any potential ad- verse events will be reported to the prison healthcare team. Procedures for recording and reporting adverse events Safety will remain the responsibility of the prison mental healthcare team. Adverse events of any medical or non- medical intervention identified or recorded by the re- search team at each site will be verified by the clinician who is part of the research team, or by an assigned med- ical colleague at specialist registrar grade or above who is a member of the prison healthcare team or by the clinical lead for the project (Professor Asherson). The decision to stop treatment following an adverse event will remain the responsibility of the clinical team. Minor adverse events that do not come under official reporting procedures will be reported to the clinical team, e.g. sleep disturbance, minor levels of anxiety or dysthymia, small increase in pulse and blood pressures, reduced appetite and other minor physical symptoms that do not endanger patients or cause more than minor distress. All other adverse events from medication will be recorded and reported in Local trial management The project will be led by Professor Asherson (PA) in London supported by a Trial Manager. The Trial Man- ager will liaise with the trial monitors and ethical board where required and support completion of interim re- ports as well as the ongoing management of the project. The Research Psychiatrist at YOI Isis will coordinate all daily activities on site. Principal Investigator Thomson (LT) will lead the project in Edinburgh and will be sup- ported by the local RA, who will conduct similar day-to- day coordinating tasks in HMYOI Polmont. Trial Steering Committee g A TSC will be convened to provide overall supervision of the trial and ensure the trial is conducted to the rigor- ous standards set out in the Medical Research Council (MRC) guidelines for Good Clinical Practice (GCP). The TSC will monitor progress, adherence and safety. The TSC chair is Professor Jenny Shaw (Consultant Forensic Psychiatrist, University of Manchester), and members in- clude Dr. Ylva Ginsberg (Consultant Psychiatrist specia- lising in ADHD in prisoners, Stockholm, Sweden), Peter Mason (Forensic Psychiatrist and specialist in ADHD, Cheshire and Wirral), Anthony Davis, R&D Manager (Oxleas National Health Service [NHS] Foundation Trust), Dr. Ulrich Muller-Sedgwick (Barnet, Enfield and Haringey Mental Health NHS Trust) and Beverley Nolker, (POA learning and user representative). Non- independent members are the lead applicants in London and Edinburgh (Philip Asherson and Lindsay Thomson). Other members of the research management group will attend as observers and to report to the TSC. It is envis- aged that the TSC will meet before the start of the pro- ject and every 6 months, alternating between telephone conference and face-to-face meetings. The research team acting on behalf of King’s College London as Sponsors have delegated the delivery of the Sponsor’s responsibility for Pharmacovigilance, as de- fined in Regulation 5 of the Medicines for Human Use (Clinical Trials) Regulations 2004, to the KHPCTO. Reporting of serious adverse events (SAEs) will continue until the last patient last dose has been completed. For each participant, the reporting period will be from the time of first dose of the trial medication to the end of his involvement in the trial (last dose at the end of 8 weeks). All SAEs, serious adverse reactions (SARs), sus- pected unexpected serious adverse reactions (SUSARs) and Important Medical Events (IMEs) (excepting those specified in this protocol as not requiring reporting) will be reported immediately by the Chief Investigator or designated site investigators to the KHPCTO in accord- ance with the current Pharmacovigilance Policy. We will copy this information to Janssen-Cilag at the same time. The KHPCTO will report SUSARs and other SARs to the Medicines and Healthcare products Regulatory Agency (MHRA) and competent authorities of other European Economic Area (EEA) states in which the trial is taking place. Treatment stopping rules The trial may be prematurely discontinued by the Sponsor, Chief Investigator or Regulatory Authority on the basis of new safety information or for other reasons given by the Data Monitoring and Ethics Committee (DMEC)/TSC, Regulatory Authority or Ethics Commit- tee concerned. Trial discontinuation for safety reasons is not envisaged given the successful pilot study. If the study is prematurely discontinued, active participants will be informed and no further participant data will be collected. Data Monitoring and Ethics Committee (DMEC) a a o o g a d o ( ) A DMEC will be convened to monitor the safety, efficacy, ethical conduct and quality of the data. The committee will consist of three members experienced in clinical trials, including an independent statistician. The DMEC chair is Professor Seena Fazel, University of Oxford (an experi- enced Forensic Psychiatrist). Other members are Professor Chris Hollis, University of Nottingham (an expert on the Clinical Management of ADHD) and Adrian Cook (a trial statistician). DMEC meetings will be timed to occur prior to TSC meetings for timely reporting to the TSC. A sum- mary of the role and reporting structure is available from the Chief Investigator on request. The Chief Investigator will report to the relevant eth- ics committees. Reporting timelines are as follows: SUSARs which are fatal or life-threatening must be reported not later than 7 days after the Sponsor is first aware of the reaction. Any additional relevant information must be reported within a further 8 days. y SUSARs that are not fatal or life-threatening must be reported within 15 days of the Sponsor first becoming aware of the reaction. The Chief Investigator and KHPCTO (on behalf of the Co-Sponsors) will submit a Development Safety Update Report (DSUR), relating to this trial IMP, to the MHRA and Ethics Committee annually. Accountability for trial medication Accountability for trial medication All aspects of treatment and accountability for managing the medication storage and delivery are managed locally by the prison pharmacies and mental health teams, as per standard practice for this medication in the prisons. Inves- tigational medicinal product (IMP) accountability will be recorded and verified. All aspects of treatment compliance and recording of treatment administration/refusal are managed by the prison mental health teams and locally by healthcare staff as per standard practice for these sites. Pa- tients are observed when they are given medication and checked to ensure the capsules have been swallowed. This information is then recorded (signed off) by nursing staff who delivered the medication on prison pharmacy record sheets or digital records. Asherson et al. Trials (2019) 20:663 Page 17 of 21 Asherson et al. Trials (2019) 20:663 Page 17 of 21 line with The Medicines for Human Use (Clinical Trials) Regulations 2004 and Amended Regulations 2006. Data handling The Chief Investigator (PA) will act as custodian for the trial data. Data will be stored on a database to be set up by the Clinical Trials Unit (CTU) at King’s College London. Patient data will be pseudo-anonymised. All pseudo- anonymised data will be stored on a password-protected computer. All trial data will be stored, handled, processed and archived in line with the Data Protection Act and the Medicines for Human Use (Clinical Trials) Amended Reg- ulations 2006. OROS-MPH is only licensed for first-time use in young people with ADHD and severe impairment under the age of 18, although NICE also recommends MPH as the first-line treatment for ADHD in adults. The 8-week trial includes a placebo group, so we will be denying a recommended treat- ment for ADHD during this period. However, currently prisoners with ADHD are rarely treated because of uncer- tainty over validity of the ADHD diagnosis, efficacy of treat- ment and concerns about potential drug abuse and diversion in prison populations. To address the issue of equal access to treatment, we will offer treatment to all par- ticipants once the trial is completed. Care will be taken to ensure that no coercion is involved in recruiting prisoners into the study. Initial consent will be obtained by members of the prison mental health team. Following procedures in the pilot study, informed consent will be obtained at the screening and diagnostic steps as well as at the start of the trial. All participants will have the mental capacity to make informed decisions. It will be made very clear that taking part in the study will have no impact, negative or positive, on their time in the prison or the prison regime. However, some participants may benefit (and show improvements in behaviour) from the treatment that is offered as part of the clinical trial. Taking part in the study will not lead to loss of earnings. The study medication is a controlled substance. There are however standard operating procedures in place Quality assurance Monitoring of this trial will be to ensure compliance with GCP, and scientific integrity will be managed and oversight retained by the KHPCTO Quality Team. Project Management Group Project Management Group The project will be led by Chief Investigator Asherson in London. The programme manager in London will liaise with the Edinburgh study coordinator weekly Asherson et al. Trials (2019) 20:663 Asherson et al. Trials (2019) 20:663 Page 18 of 21 Page 18 of 21 Page 18 of 21 for the prescription of controlled drugs from the prison pharmacy. throughout the project, monitor progress and maintain communication about successes and barriers to progress and report back to the lead applicants. Asherson and Thomson will hold a weekly telephone conference to re- view progress with the data collection teams. A meeting of all investigators and co-applicants will review progress on a monthly basis. Participants are informed that their anonymised re- search data from this study will be stored securely and may be shared with other scientists or research groups where this helps us to understand the findings of the study. The data may also be used in combination with data from other similar studies. All personal information is stored in a secure place (locked cabinet in locked office in the prisons). Outside of the immediate healthcare team involved in this re- search, no one will be able to match personal informa- tion (name and prison number) with the information gathered for research. After completion of the study, personal information held by the local study teams will be destroyed. Personal details will only be used to con- tact participants about the study. Personal details will not be linked to clinical or prison records in the research records. The clinical data and prison records used for re- search will be identified using a study ID code. Data management Data will be stored on a trials database to be set up by the CTU at King’s College London. This allows for full audit information and checks on data entry that will be used to ensure the integrity of the data collection and monitoring of the study progress. At the end of the study, all research records will be transferred by secure courier service to the Social, Genetic and Developmental Psychiatry (SGDP) Centre, King’s College London. A detailed data manage- ment plan is available from the authors. Ethical issues specific to this project The trial will be conducted in compliance with the prin- ciples of the Declaration of Helsinki (1996), the princi- ples of GCP and in accordance with all applicable regulatory requirements including but not limited to the Research Governance Framework and the Medicines for Human Use (Clinical Trial) Regulations 2004, as amended in 2006 and any subsequent amendments. This protocol and related documents will be submitted for re- view to London South East Research Ethics Committee (REC), and to the MHRA for Clinical Trial Authorisa- tion. Annual progress and safety reports and a final re- port at conclusion of the trial will be submitted to the KHPCTO (on behalf of the Sponsor), the REC and the MHRA within the timelines defined in the regulations. We previously received ethical approval and MHRA registration for the current open label trial of OROS- MPH in HMP Isis and will follow the same recruitment and consent procedures as in the previous study. Authors’ contributions All authors contributed to the review and development of the protocol. Specials thanks go to LJ (Trial Manager) and RH (trial statistician) for their critical review of the protocol and its development. The Chief Investigator and co-applicants have sole responsibility for the study design, collection, management, analysis and interpretation of the data, writing of the report and the decision to submit reports arising from this study for publication. All authors have read and approved the final manuscript. Insurance/indemnity Clinical Trial insurance is provided by the King’s College London Clinical Trials Insurance Policy. Supplementary information f These four criteria are reproduced as follows: Supplementary information accompanies this paper at https://doi.org/10. 1186/s13063-019-3705-9. Supplementary information accompanies this paper at https://doi.org/10. 1186/s13063-019-3705-9. 1. Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; AND Additional file 1. : Information sheet and consent form for the screening stage and trial stage of the study. (DOCX 44 kb) Additional file 1. : Information sheet and consent form for the screening stage and trial stage of the study. (DOCX 44 kb) Additional file 2. : SPIRIT 2013 checklist: recommended items to address in a clinical trial protocol and related documents. (DOC 122 kb) 2. Drafting the work or revising it critically for important intellectual content; AND 3. Final approval of the version to be published; AND Abbreviations ADHD: Attention-deficit/hyperactivity disorder; AES: Adverse Events Scale; ARI-S: Affective Reactivity Index; AUDIT-C: Alcohol Use Disorders Identification Test; BRCE: Behaviour Report Card by Education Staff; BRCP: Behaviour Report Card by Prison Officers; BSI: Brief Symptom Inventory; CAARS: Conners Adult ADHD Scale; CAARS-O: Conners Adult ADHD Rating Scale; CGI: Clinical Global Impression; CORE-OM: CORE Outcome Measure; CTQ: Childhood Trauma Questionnaire; CTU: Clinical Trials Unit; DIVA: Diagnostic Interview for Adult ADHD; DMEC: Data monitoring and ethics committee; DSM: Diasgnostic and statistical manual; DSUR: Development Safety Update Report; ESMS: Emergency Scientific and Medical Services; GCP: Good clinical practice; IEPs: Incentives and Earned Privileges; IME: Important medical event; IMP: Investigational medicinal product; KHPCTO: King’s Health Partners Clinical Trials Office; MAR: Missing at random; MEWS: Mind Excessively Wandering Scale; MHRA: Medicines and Healthcare products Regulatory Agency; MINI: International Neuropsychiatric Interview for common mental health disorders; MOASE : Modified Overt Aggression Scale by Education Staff; MOASP: Modified Overt Aggression Scale by Prison Officers; MPH: Methylphenidate; MRC: Medical Research Council; MVQ: Maudsley Violence Questionnaire; NICE: National Institute for Health and Care Excellence; NIDA: National Institute on Drug Abuse; REC: Research ethics committee; RPAQ: Reactive-Proactive Aggression Questionnair; SAE: serious adverse events; SAR: serious adverse reactions; SIGN: Scottish Intercollegiate Guidelines Network; SMD: Standardised mean difference; SUSAR: suspected unexpected serious adverse reactions; TSC: Trial steering committee; WASI: Wechsler Abbreviated Scale of Intelligence; Weiss-CD: Weiss Conduct Disorder Scale; WRAADS: Wender-Reimherr Adult ADHD Diagnostic Scale; ZAB-BPD: Zanarini Rating Scale for Borderline Personality Disorder Over the duration of the study, many people will have contributed to the CIAO-II project. We will follow the principles outlined in the International Committee of Medical Journal Editors Recommendations for the Con- duct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals (ICMJE Recommendations 2013). Publication policy and access to data It is intended that the results of the study will be reported and disseminated at international conferences and in peer-reviewed scientific journals. Additionally, we will dis- seminate through internal reports to the prison services and through relevant online forums such as the UK Adult Asherson et al. Trials (2019) 20:663 Asherson et al. Trials (2019) 20:663 Page 19 of 21 Page 19 of 21 ADHD Network (UKAAN). The data collected will be shared and used in collaborative studies once the main ob- jectives of the study have been completed. We will provide efficient and rapid access to data to scientists following a formal review process by the study co-investigators. The dataset that will be shared will be anonymised, meaning that it will not include data of name, birthdate or prisoner record number. Any raw data from the trials database or randomisation system will not be shared. Investigators who wish to access the data once the study is finished should contact the Chief Investigator (PA) and request permission. A formal application will be required indicat- ing the question to be addressed and methods to be ap- plied. Data access will be allowed for all legitimate scientific enquiries following scientific review by the Chief Investigator and co-applicants. Funding h This project is funded by the Efficacy and Mechanism Evaluation (EME) programme, an MRC and National Institute for Health Research (NIHR) partnership (project ref. 14/23/17). The EME programme is funded by the MRC and NIHR, with contributions from the Chief Scientist Office (CSO) in Scotland and Health and Care Research Wales and the Health and Social Care (HSC) R&D Division, Public Health Agency in Northern Ireland. Janssen Pharmaceutical Company supplied the Concerta XL for this study. The views Acknowledgements S l h k 4. Agreement 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. Special thanks go to members of the Trial Steering Committee (Jenny Shaw, Ylva Ginsberg, Peter Mason, Ulrich Muller-Sedgwick, Anthony Davis, Beverly Nolker and Campbell Nichol), the Data Monitoring Committee (Seena Fazel, Chris Hollis and Adrian Cook) and the trial management group (Megan Bed- ding, Imogen Kretzschmar, Megan Mansfield, Grace Williamson, Lauren Wil- son, Clare McCafferty, Kathleen McCurdy, Craig Marsh, Laura Giannulli, Caroline Kelly, Sharon Jakobowitz, Prajakta Patil, Resham Abbasi, Alexandra Blackman, Donnie Macpherson and Tatsiana Luhakova) for their invaluable guidance on trial procedures and development of the protocol. We would also like to thank all prison and healthcare staff who have made this trial possible, as well as the clinical research associates working with the KHPCTO for monitoring on behalf of the Sponsor. In addition to being accountable for the parts of the work he or she has done, an author should be able to identify which co-authors are responsible for other spe- cific parts of the work. In addition, authors should have confidence in the integrity of the contributions of their co-authors. also like to thank all prison and healthcare staff who have made this trial possible, as well as the clinical research associates working with the KHPCTO for monitoring on behalf of the Sponsor. Competing interests ll d 13. Ginsberg Y, Lindefors N. Methylphenidate treatment of adult male prison inmates with attention-deficit hyperactivity disorder: randomised double- blind placebo-controlled trial with open-label extension. Br J Psychiatry. 2012;200(1):68–73. King’s College London departmental research support account received payments for work by PA: consultancy to Shire, Eli Lilly and Novartis; educational/research awards from Shire, Lilly, Novartis, Vifor Pharma, GW Pharmaceuticals and QbTech; speaker at events sponsored by Shire, Lilly, Novartis, Flynn Pharma and Medicen. PA received personal honoaria for talking at events sponsored by Shire, Flynn Pharma and Medicen. SY has received honoraria for consultation and/or educational talks in the last 5 years from Janssen, HB Pharma and/or Shire. She is author of the ADHD Child Evaluation (ACE) and ACE+ for adults and lead author of R&R2 for ADHD Youths and Adults. 14. Asherson P, Evans C, Young S: A pilot study of Concerta XL in adult offenders with ADHD. 2018. https://kclpure.kcl.ac.uk/portal/en/publications/ a-pilot-study-of-concerta-xl-in-adult-offenders-with-adhd(68d820ff-1026-4 cdc-867e-4fa0325d162c).html. 15. Chen Q, Sjolander A, Runeson B, D'Onofrio BM, Lichtenstein P, Larsson H. Drug treatment for attention-deficit/hyperactivity disorder and suicidal behaviour: register based study. BMJ. 2014;348:g3769. In the last 3 years, SL has received personal fees from Janssen, Otsuka and Sunovion and research support from Janssen, all in connection with schizophrenia meetings and trials. 16. Chang Z, Lichtenstein P, Langstrom N, Larsson H, Fazel S. Association between prescription of major psychotropic medications and violent reoffending after prison release. JAMA. 2016;316(17):1798–807. p g JS has worked with several pharmaceutical companies to seek to identify new or improved medications, but they do not have a relationship to the study reported here. This has included research grant support and consultancy payments to King’s College London and travelling and/or accommodation and/or conference expenses (including, in the past 3 years, from Martindale, Indivior, MundiPharma, Camurus/Braeburn). The other authors declare that they have no competing interests JS has worked with several pharmaceutical companies to seek to identify new or improved medications, but they do not have a relationship to the study reported here. This has included research grant support and consultancy payments to King’s College London and travelling and/or accommodation and/or conference expenses (including, in the past 3 years, from Martindale, Indivior, MundiPharma, Camurus/Braeburn). The other authors declare that they have no competing interests 17. Chang Z, D'Onofrio BM, Quinn PD, Lichtenstein P, Larsson H. Medication for attention-deficit/hyperactivity disorder and risk for depression: a nationwide longitudinal cohort study. Biol Psychiatry. 2016;80(12):916–22. 18. Trial status The protocol version is version v2.0, 30 August 2018. Recruitment started 30 May 2016. Recruitment was completed on 30 May 2019. Last particpant last visit 6 June 2019. Data lock 27 August 2019. Page 20 of 21 Asherson et al. Trials (2019) 20:663 Asherson et al. Trials (2019) 20:663 Received: 25 February 2019 Accepted: 6 September 2019 Consent for publication 12. Cortese S, Adamo N, Del Giovane C, Mohr-Jensen C, Hayes AJ, Carucci S, Atkinson LZ, Tessari L, Banaschewski T, Coghill D, et al. Comparative efficacy and tolerability of medications for attention-deficit hyperactivity disorder in children, adolescents, and adults: a systematic review and network meta- analysis. Lancet Psychiatry. 2018;5(9):727–38. Personal identifying data will not be published. Consent was obtained for publication of anonymised group-level data. Received: 25 February 2019 Accepted: 6 September 2019 expressed in this publication are those of the author(s) and not necessarily those of the MRC, NHS, NIHR or the Department of Health and Social Care. The Sponsor for the trial is King’s College London. PA and JS are NIHR Senior Investigators and are both supported by the NIHR Biomedical Research Centre for Mental Health at South London and Maudsley NHS Foundation Trust and King’s College London. Competing interests ll d Chang Z, Lichtenstein P, D'Onofrio BM, Sjolander A, Larsson H. Serious transport accidents in adults with attention-deficit/hyperactivity disorder and the effect of medication: a population-based study. JAMA Psychiatry. 2014;71(3):319–25. accommodation and/or conference expenses (including, in the past 3 years, from Martindale, Indivior, MundiPharma, Camurus/Braeburn). The other authors declare that they have no competing interests. 19. Lenzi F, Cortese S, Harris J, Masi G. Pharmacotherapy of emotional dysregulation in adults with ADHD: a systematic review and meta-analysis. Neurosci Biobehav Rev. 2018;84:359–67. Availability of data and materials London: British Psychological Society and the Royal College of Psychiatrists; 2008. 8. NICE. Diagnosis and management of ADHD in children, young people and adults. London: British Psychological Society and the Royal College of Psychiatrists; 2018. 8. NICE. Diagnosis and management of ADHD in children, young people and adults. London: British Psychological Society and the Royal College of Psychiatrists; 2018. References 1. Asherson P, Buitelaar J, Faraone SV, Rohde LA. Adult attention-deficit hyperactivity disorder: key conceptual issues. Lancet Psychiatry. 2016;3(6):568–78. 1. Asherson P, Buitelaar J, Faraone SV, Rohde LA. Adult attention-deficit hyperactivity disorder: key conceptual issues. Lancet Psychiatry. 2016;3(6):568–78. 2. Polanczyk G, de Lima MS, Horta BL, Biederman J, Rohde LA. The worldwide prevalence of ADHD: a systematic review and metaregression analysis. Am J Psychiatry. 2007;164(6):942–8. 2. Polanczyk G, de Lima MS, Horta BL, Biederman J, Rohde LA. The worldwide prevalence of ADHD: a systematic review and metaregression analysis. Am J Psychiatry. 2007;164(6):942–8. Availability of data and materials y It is intended that the results of the study will be reported and disseminated at international conferences and in peer-reviewed scientific journals. Add- itionally, we will disseminate through internal reports to the prison services and through relevant online forums such as UKAAN. The data collected will be shared and used in collaborative studies once the main objectives of the study have been completed. We will provide efficient and rapid access to data for secondary analyses. We will provide access to scientists following a formal review process by the study co-investigators, to minimise duplication of effort by different research teams. We will encourage teams with similar research questions to work collaboratively. The dataset that will be shared will be anonymised, meaning that it will not include the name, birthdate, prisoner record number or other data that could potentially identify individ- uals. In the initial phase of data sharing, investigators who wish to access the data once the study is finished should contact the Chief Investigator (PA) and request permission. A formal application will be required indicating the question to be addressed and methods to be applied. Data access will be allowed for all legitimate scientific enquiries following scientific review by the Cchief Investigator and co-applicants. 3. Thomas R, Sanders S, Doust J, Beller E, Glasziou P. Prevalence of attention- deficit/hyperactivity disorder: a systematic review and meta-analysis. Pediatrics. 2015;135(4):e994–1001. 3. Thomas R, Sanders S, Doust J, Beller E, Glasziou P. Prevalence of attention- deficit/hyperactivity disorder: a systematic review and meta-analysis. Pediatrics. 2015;135(4):e994–1001. 4. Fayyad J, De Graaf R, Kessler R, Alonso J, Angermeyer M, Demyttenaere K, De Girolamo G, Haro JM, Karam EG, Lara C, et al. Cross-national prevalence and correlates of adult attention-deficit hyperactivity disorder. Br J Psychiatry. 2007;190:402–9. 5. Simon V, Czobor P, Balint S, Meszaros A, Bitter I. Prevalence and correlates of adult attention-deficit hyperactivity disorder: meta-analysis. Br J Psychiatry. 2009;194(3):204–11. 6. Young S, Moss D, Sedgwick O, Fridman M, Hodgkins P. A meta-analysis of the prevalence of attention deficit hyperactivity disorder in incarcerated populations. Psychol Med. 2015;45(2):247–58. 7. NICE. Attention deficit hyperactivity disorder: the NICE guideline on diagnosis and managment of ADHD in children, young people and adults. London: British Psychological Society and the Royal College of Psychiatrists; 2008. 7. NICE. Attention deficit hyperactivity disorder: the NICE guideline on diagnosis and managment of ADHD in children, young people and adults. Ethics approval and consent to participate Ethical approval was obtained from the East of England Essex Research Ethics Committee (16/EE/0117). Approvals for the study were further obtained from the Health Research Authority (HRA), the National Offender Management Service (NOMS, England), the Scottish Prison Service (SPS, Scotland), Research and Development from Oxleas NHS Foundation Trust and NHS Forth Valley. All participants provided informed consent and signed approved consent forms for participation in the study. 9. SIGN. Management of attention deficit hyperactivity and hyperkinetic disorders in children and young people: a national clinical guideline. Edinburgh: Scottish Intercollegiate Guidelines Network; 2009. 9. SIGN. Management of attention deficit hyperactivity and hyperkinetic disorders in children and young people: a national clinical guideline. Edinburgh: Scottish Intercollegiate Guidelines Network; 2009. 10. Attention deficit hyperactivity disorder: evidence update 45. http://arms. evidence.nhs.uk/resources/hub/1027588/attachment. Accesssed 25 Oct 2019. 11. Koesters M, Becker T, Kilian R, Fegert JM, Weinmann S. Limits of meta- analysis: methylphenidate in the treatment of adult attention-deficit hyperactivity disorder. J Psychopharmacol. 2009;23(7):733–44. 1Institute of Psychiatry, Psychology and Neuroscience, King’s College London, De Crespigny Park, London SE5 8AF, UK. 2Healthcare Department, HM Prison Brixton, Jebb Avenue, London SW2 5XF, UK. 3Division of Psychiatry, University of Edinburgh, Morningside Park, Edinburgh EH10 5HF, UK. 4Psychology Services Limited, PO1735, Croydon CR0 7WA, UK. 19. Lenzi F, Cortese S, Harris J, Masi G. Pharmacotherapy of emotional dysregulation in adults with ADHD: a systematic review and meta-analysis. Neurosci Biobehav Rev. 2018;84:359–67. Author details 1I i f P 1Institute of Psychiatry, Psychology and Neuroscience, King’s College London, De Crespigny Park, London SE5 8AF, UK. 2Healthcare Department, HM Prison Brixton, Jebb Avenue, London SW2 5XF, UK. 3Division of Psychiatry, University of Edinburgh, Morningside Park, Edinburgh EH10 5HF, UK. 4Psychology Services Limited, PO1735, Croydon CR0 7WA, UK. 20. Moukhtarian TR, Cooper RE, Vassos E, Moran P, Asherson P. Effects of stimulants and atomoxetine on emotional lability in adults: a systematic review and meta-analysis. Eur Psychiatry. 2017;44:198–207. Page 21 of 21 Asherson et al. Trials (2019) 20:663 Asherson et al. Trials (2019) 20:663 21. Asherson P, Akehurst R, Kooij JJ, Huss M, Beusterien K, Sasane R, Gholizadeh S, Hodgkins P. Under diagnosis of adult ADHD: cultural influences and societal burden. J Atten Disord. 2012;16(5 Suppl):20S–38S. 22. Volkow ND, Swanson JM. Variables that affect the clinical use and abuse of methylphenidate in the treatment of ADHD. Am J Psychiatry. 2003;160(11):1909–18. 23. Kooij JJ: Diagnostic Interview for ADHD in Adults 2.0 (DIVA 2.0). In: Diagnostic assessment and treatment. Kooij JJ, editor. Amsterdam: Pearson Assessment and Information BV; 2010. 24. Black DW, Arndt S, Hale N, Rogerson R. Use of the Mini International Neuropsychiatric Interview (MINI) as a screening tool in prisons: results of a preliminary study. J Am Acad Psychiatry Law. 2004;32(2):158–62. 25. Zanarini MC, Vujanovic AA, Parachini EA, Boulanger JL, Frankenburg FR, Hennen J. Zanarini Rating Scale for Borderline Personality Disorder (ZAN-BPD): a continuous measure of DSM-IV borderline psychopathology. J Personal Disord. 2003;17(3):233– 42. 26. Conners CK, Erhardt D, Sparrow E. Conners’ Adult ADHD Rating Scales (CAARS). Technical manual. New York: Multi Health Systems Inc; 1999. 27. Reimherr FW, Marchant BK, Strong RE, Hedges DW, Adler L, Spencer TJ, West SA, Soni P. Emotional dysregulation in adult ADHD and response to atomoxetine. Biol Psychiatry. 2005;58(2):125–31. 28. Rosler M, Retz W, Fischer R, Ose C, Alm B, Deckert J, Philipsen A, Herpertz S, Ammer R. Twenty-four-week treatment with extended release methylphenidate improves emotional symptoms in adult ADHD. World J Biol Psychiatry. 2010;11(5):709–18. 29. CADDRA patient and medication form https://www.caddra.ca/wp-content/ uploads/CADDRA-Patient-ADHD-Medication-Form.pdf. Accessed 25 Oct 2019. 30. Guy W: Clinical Global Impressions (CGI) Scale, Modified. Task Force for the Handbook of Psychiatric Measures: Handbook of psychiatric measures. In: Rush JA, editor. Washington, DC: American Psychiatric Association; 2000. 31. Raine A, Dodge K, Loeber R, Gatzke-Kopp L, Lynam D, Reynolds C, Stouthamer- Loeber M, Liu J. Author details 1I i f P The Reactive-Proactive Aggression Questionnaire: differential correlates of reactive and proactive aggression in adolescent boys. Aggress Behav. 2006;32(2):159–71. 32. Saitz R, Palfai TP, Cheng DM, Alford DP, Bernstein JA, Lloyd-Travaglini CA, Meli SM, Chaisson CE, Samet JH. Screening and brief intervention for drug use in primary care: the ASPIRE randomized clinical trial. JAMA. 2014;312(5):502–13. 33. Bush K, Kivlahan DR, McDonell MB, Fihn SD, Bradley KA. The AUDIT alcohol consumption questions (AUDIT-C): an effective brief screening test for problem drinking. Ambulatory Care Quality Improvement Project (ACQUIP). Alcohol Use Disorders Identification Test. Arch Intern Med. 1998;158(16):1789–95. 34. Bernstein DP, Stein JA, Newcomb MD, Walker E, Pogge D, Ahluvalia T, Stokes J, Handelsman L, Medrano M, Desmond D, et al. Development and validation of a brief screening version of the Childhood Trauma Questionnaire. Child Abuse Negl. 2003;27(2):169–90. 35. Barkley R, Murphy K. A clinical workbook: attention-deficit hyperactivity disorder. 3rd ed. New York: Guilford; 1998. 36. Stringaris A, Goodman R, Ferdinando S, Razdan V, Muhrer E, Leibenluft E, Brotman MA. The Affective Reactivity Index: a concise irritability scale for clinical and research settings. J Child Psychol Psychiatry Allied Discip. 2012;53(11):1109–17. 36. Stringaris A, Goodman R, Ferdinando S, Razdan V, Muhrer E, Leibenluft E, Brotman MA. The Affective Reactivity Index: a concise irritability scale for clinical and research settings. J Child Psychol Psychiatry Allied Discip. 2012;53(11):1109–17. 37. Mowlem FD, Skirrow C, Reid P, Maltezos S, Nijjar SK, Merwood A, Barker E, Cooper R, Kuntsi J, Asherson P. Validation of the Mind Excessively Wandering Scale and the relationship of mind wandering to impairment in adult ADHD. J Atten Disord. 2019; 23(6):624–34. 38. Derogatis LR, Melisaratos N. The Brief Symptom Inventory: an introductory report. Psychol Med. 1983;13(3):595–605. 39. Walker J, Bowes N. The evaluation of violent thinking in adult offenders and non-offenders using the Maudsley Violence Questionnaire. Crim Behav Ment Health. 2013;23(2):113–23. 39. Walker J, Bowes N. The evaluation of violent thinking in adult offenders and non-offenders using the Maudsley Violence Questionnaire. Crim Behav Ment Health. 2013;23(2):113–23. 40. Evans C, Connell J, Barkham M, Margison F, McGrath G, Mellor-Clark J, Audin K. Towards a standardised brief outcome measure: psychometric properties and utility of the CORE-OM. Br J Psychiatry. 2002;180:51–60. 41. Knoedler DW. The Modified Overt Aggression Scale. Am J Psychiatry. 1989; 146(8):1081–2. 41. Knoedler DW. The Modified Overt Aggression Scale. Am J Psychiatry. 1989; 146(8):1081–2. 42. Classroom Behaviour Report Card Resource Book. https://jimwrightonline. com/pdfdocs/tbrc/tbrcmanual.pdf. Accssed 25 Oct 2019. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
https://openalex.org/W4249214663	https://epic.awi.de/id/eprint/52785/1/tc-14-2173-2020_Krumpen.pdf	English	null	The MOSAiC ice floe: sediment-laden survivor from the Siberian shelf	null	2,020	cc-by	12,115	Correspondence: Thomas Krumpen (tkrumpen@awi.de) Correspondence: Thomas Krumpen (tkrumpen@awi.de) Received: 23 February 2020 – Discussion started: 25 February 2020 Revised: 10 June 2020 – Accepted: 21 June 2020 – Published: 6 July 2020 Revised: 10 June 2020 – Accepted: 21 June 2020 – Published: 6 July 2020 ice-free summer period since reliable instrumental records began. We show, using a Lagrangian tracking tool and a ther- modynamic sea ice model, that the MOSAiC ﬂoe carrying the Central Observatory (CO) formed in a polynya event north of the New Siberian Islands at the beginning of De- cember 2018. The results further indicate that sea ice in the vicinity of the CO ( < 40 km distance) was younger and 36 % thinner than the surrounding ice with potential consequences for ice dynamics and momentum and heat transfer between ocean and atmosphere. Sea ice surveys carried out on vari- ous reference ﬂoes in autumn 2019 verify this gradient in ice thickness, and sediments discovered in ice cores (so-called dirty sea ice) around the CO conﬁrm contact with shallow Abstract. In September 2019, the research icebreaker Po- larstern started the largest multidisciplinary Arctic expedi- tion to date, the MOSAiC (Multidisciplinary drifting Obser- vatory for the Study of Arctic Climate) drift experiment. Be- ing moored to an ice ﬂoe for a whole year, thus including the winter season, the declared goal of the expedition is to better understand and quantify relevant processes within the atmosphere–ice–ocean system that impact the sea ice mass and energy budget, ultimately leading to much improved cli- mate models. Satellite observations, atmospheric reanalysis data, and readings from a nearby meteorological station in- dicate that the interplay of high ice export in late winter and exceptionally high air temperatures resulted in the longest The MOSAiC ice ﬂoe: sediment-laden survivor from the Siberian shelf Thomas Krumpen1, Florent Birrien1, Frank Kauker1, Thomas Rackow1, Luisa von Albedyll1, Michael Angelopoulos1, H. Jakob Belter1, Vladimir Bessonov2, Ellen Damm1, Klaus Dethloff1, Jari Haapala3, Christian Haas1, Carolynn Harris4, Stefan Hendricks1, Jens Hoelemann1, Mario Hoppmann1, Lars Kaleschke1, Michael Karcher1, Nikolai Kolabutin2, Ruibo Lei5, Joseﬁne Lenz1,6, Anne Morgenstern1, Marcel Nicolaus1, Uwe Nixdorf1, Tomash Petrovsky2, Benjamin Rabe1, Lasse Rabenstein7, Markus Rex1, Robert Ricker1, Jan Rohde1, Egor Shimanchuk2, Suman Singha8, Vasily Smolyanitsky2, Vladimir Sokolov2, Tim Stanton9, Anna Timofeeva2, Michel Tsamados10, and Daniel Watkins11 homas Krumpen1, Florent Birrien1, Frank Kauker1, Thomas Rackow1, Luisa von Albedyll1, 1 1 2 1 1 1Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research, Am Handelshafen 12, 27570 Bremerhaven, Germany 1Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research, Am Handelshafen 12, 27570 Bremerhaven, Germany Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research, Am Handelshafen 12, 7570 Bremerhaven, Germany 2Arctic and Antarctic Research Institute, Ulitsa Beringa, 38, Saint Petersburg, 199397, Russia 3 3Finnish Meteorological Institute, Marine Research, P.O. Box 503, 00101 Helsinki, Finland 4Dartmouth College, Department of Earth Science, 6105 Fairchild Hall, Hanover, NH 03755, US 5 5Polar Research Institute of China, MNR Key Laboratory for Polar Science, 451 Jinqiao Road, P d Sh h i 200136 Chi 5Polar Research Institute of China, MNR Key Laboratory for Polar Science, 451 Jinqiao Road, Pudong, Shanghai 200136, China g g 6Association of Polar Early Career Scientists, Alfred Wegener Institute for Polar and Marine Research g g 6Association of Polar Early Career Scientists, Alfred Wegener Institute for Polar and Marine Research, Telegrafenberg A45 14473 Potsdam Germany 7Drift & Noise Polar Services, Stavendamm 17, 28195 Bremen, Germany German Aerospace Center, Remote Sensing Technology Institute, SAR Signal Processing, 8German Aerospace Center, Remote Sensing Technology Institute, SAR Signal Processing, Am Fallturm 9 28359 Bremen Germany 9Naval Postgraduate School, Oceanography Department, 833 Dyer Road, Building 232, Monterey, CA 93943, USA 10Centre for Polar Observation and Modelling, University College London, Dept. of Earth Science, 5 Gower Place, London WC1E 6BS, UK College of Earth, Ocean, and Atmospheric Science, Oregon State University, Corvallis, OR, USA 11College of Earth, Ocean, and Atmospheric Science, Oregon State University, Corvallis, OR, US T. Krumpen et al.: The MOSAiC ice ﬂoe waters in an early phase of growth, consistent with the track- ing analysis. Since less and less ice from the Siberian shelves survives its ﬁrst summer (Krumpen et al., 2019), the MO- SAiC experiment provides the unique opportunity to study the role of sea ice as a transport medium for gases, macronu- trients, iron, organic matter, sediments and pollutants from shelf areas to the central Arctic Ocean and beyond. Com- pared to data for the past 26 years, the sea ice encountered at the end of September 2019 can already be classiﬁed as excep- tionally thin, and further predicted changes towards a season- ally ice-free ocean will likely cut off the long-range transport of ice-rafted materials by the Transpolar Drift in the future. A reduced long-range transport of sea ice would have strong implications for the redistribution of biogeochemical matter in the central Arctic Ocean, with consequences for the bal- ance of climate-relevant trace gases, primary production and biodiversity in the Arctic Ocean. vived the summer melt (hereafter called residual ice (WMO, 2017), shorthand for residual ﬁrst-year ice, which does not graduate to become second-year ice until 1 January). With the support of the Russian research vessel Akademik Fedorov, a distributed network (DN) of autonomous buoys was in- stalled in a 40 km radius around the CO on 55 additional residual ice ﬂoes of similar age (Krumpen and Sokolov, 2020). For more information about the MOSAiC expedition the reader is referred to https://www.mosaic-expedition.org (last access: 25 June 2020). vived the summer melt (hereafter called residual ice (WMO, 2017), shorthand for residual ﬁrst-year ice, which does not graduate to become second-year ice until 1 January). With the support of the Russian research vessel Akademik Fedorov, a distributed network (DN) of autonomous buoys was in- stalled in a 40 km radius around the CO on 55 additional residual ice ﬂoes of similar age (Krumpen and Sokolov, 2020). For more information about the MOSAiC expedition the reader is referred to https://www.mosaic-expedition.org (last access: 25 June 2020). The purpose of this paper is to investigate the environ- mental conditions that shaped the ice in the chosen research region prior to and at the start of the MOSAiC drift. The analyses presented here are of high importance for future work as they will provide the initial state for model-based studies and satellite-based validation planned to take place during MOSAiC. 1 Introduction In early autumn 2019 the German research icebreaker Po- larstern, operated by the Alfred Wegener Institute (AWI), Helmholtz Centre for Polar and Marine Research, was moored to an ice ﬂoe north of the Laptev Sea in order to travel with the Transpolar Drift on a 1-year-long journey to- ward the Fram Strait. The goal of the international Multidis- ciplinary drifting Observatory for the Study of Arctic Cli- mate (MOSAiC) project is to better quantify relevant pro- cesses within the atmosphere–ice–ocean system that impact the sea ice mass and energy budget. Other main goals are a better understanding of available satellite data via ground- truthing and improved process understanding that can be im- plemented in climate models. MOSAiC continues a long tra- dition of Russian north pole (NP) drifting ice stations. In the past, these stations predominantly used older multi-year ice ﬂoes as their base of operations, with small settlements set up on the surface. Using this approach, the Arctic and Antarctic Research Institute (AARI, Russia) undertook 40 NP drift sta- tions in the central Arctic between 1937 and 2013. However, as the summer melt period lasted longer every year, thick multi-year ﬂoes suitable for ice camps became more seldom, and Russia was ultimately forced to temporarily discontinue these drifting stations. We ﬁrst provide an overview of the ice conditions in the extended surroundings of the experiment and of the atmo- spheric and oceanographic processes that preconditioned the ice in the preceding winter and summer. To do so, we utilise satellite observations, NCEP atmospheric reanalysis data, and readings from a nearby meteorological station. Secondly, we evaluate the representativeness of the ice conditions in Polarstern’s immediate vicinity compared to the extended surroundings. These analyses chieﬂy employ a Lagrangian backward tracking tool (see Methods) that al- lows us to determine where the encountered ice was initially formed and to identify the dominant processes that have in- ﬂuenced the ice along its trajectory. For this work, a thermo- dynamic one-column model was coupled to the backtracking tool to simulate ice growth and melting processes along these trajectories (Methods). The coupled results are then com- pared with observational data gathered by satellites and in situ measurements made during the search for the main ﬂoe and set-up of the DN. T. Krumpen et al.: The MOSAiC ice ﬂoe In addition, it provides the foundation for the analysis and interpretation of upcoming biogeochemical and ecological studies. This study exclusively employs pre- viously described methods (Damm et al., 2018; Peeken et al., 2018; Krumpen et al., 2016, 2019) for tracking sea ice back in time and for modelling thermodynamic sea ice evolution (see Methods). These tools are used in combination with the ﬁrst ﬁeld observations made on board the accompanying re- search vessel Akademik Fedorov. A more detailed description of the CO’s physical characteristics will be the focus of fu- ture studies. Published by Copernicus Publications on behalf of the European Geosciences Union. 2174 2 Material and methods a combination of satellite-derived, low-resolution drift prod- ucts (Krumpen et al., 2019). The approach has also been ap- plied in a number of previous studies for the same purpose (Ricker et al., 2018; Damm et al., 2018; Peeken et al., 2018; Krumpen et al., 2016 and others). In summary, IceTrack uses a combination of three different publicly available ice drift products for the tracking: (i) motion estimates based on a combination of scatterometer and radiometer data provided T. Krumpen et al.: The MOSAiC ice ﬂoe T. Krumpen et al.: The MOSAiC ice ﬂoe T. Krumpen et al.: The MOSAiC ice ﬂoe 2175 Figure 1. Initial sea ice conditions in the MOSAiC study region on 25 September 2019, shortly before anchoring at the MOSAiC ﬂoe. (a) Satellite-based sea ice concentration (source: University of Bremen). (b) Ship tracks of Polarstern (white) and Akademik Fedorov (black) superimposed on a MODIS image (source: NASA) obtained on 22 September 2019. The red circle indicates the distributed network region (DNR, 40 km radius). (c) Akademik Fedorov (right) and Polarstern (left) during bunkering procedure in thin ice, (d) Sentinel-1 SAR image operated at C-band obtained on September 25 (source: ESA). The DN was mostly installed on the darker ﬂoes that correspond to older ice that had survived the summer (residual ice). The position of the Central Observatory is marked by a black rectangle. (e) Close-up of the Central Observatory based on a TerraSAR-X image (X-band) obtained on September 25 (source: DLR). The ﬂoe was initially 2.8 km × 3.8 km in size and is characterised by a strongly deformed zone in the centre, called the “fortress”. Figure 1. Initial sea ice conditions in the MOSAiC study region on 25 September 2019, shortly before anchoring at the MOSAiC ﬂoe. (a) Satellite-based sea ice concentration (source: University of Bremen). (b) Ship tracks of Polarstern (white) and Akademik Fedorov (black) superimposed on a MODIS image (source: NASA) obtained on 22 September 2019. The red circle indicates the distributed network region (DNR, 40 km radius). (c) Akademik Fedorov (right) and Polarstern (left) during bunkering procedure in thin ice, (d) Sentinel-1 SAR image operated at C-band obtained on September 25 (source: ESA). The DN was mostly installed on the darker ﬂoes that correspond to older ice that had survived the summer (residual ice). The position of the Central Observatory is marked by a black rectangle. (e) Close-up of the Central Observatory based on a TerraSAR-X image (X-band) obtained on September 25 (source: DLR). The ﬂoe was initially 2.8 km × 3.8 km in size and is characterised by a strongly deformed zone in the centre, called the “fortress”. 1 Introduction The MOSAiC project represents an attempt to adapt to the “new normal” in the Arctic (warmer and thinner Arctic sea ice) and to use the ship itself as an observational platform. Around the ship, an ice camp (Central Observatory, CO) with comprehensive instrumentation was set up to intensively ob- serve processes within the atmosphere, ice, and ocean. For this purpose, on 4 October 2019, the ship was moored to a promising ice ﬂoe measuring roughly 2.8 km × 3.8 km (see Fig. 1 at coordinates 85◦N, 136◦E). The ﬂoe was part of a loose assembly of pack ice, not yet a year old, which had sur- Thirdly, we discuss whether the ice encountered in autumn 2019 on site was unusually thin compared to previous years. For this we run the coupled thermodynamics–tracking model for the MOSAiC start region with NCEP forcing data of the past 26 years to examine interannual variability of residual ice thickness in the study region. In closing, implications for upcoming future physical, bio- geochemical and ecological MOSAiC studies due to the con- ditions encountered on site are discussed. The Cryosphere, 14, 2173–2187, 2020 https://doi.org/10.5194/tc-14-2173-2020 https://doi.org/10.5194/tc-14-2173-2020 2175 T. Krumpen et al.: The MOSAiC ice ﬂoe https://doi.org/10.5194/tc-14-2173-2020 2.1 Lagrangian sea ice trajectories To determine the origin, pathways and thickness changes of sea ice, as well as the atmospheric forcing acting on the ice cover, we use our Lagrangian drift analysis system called IceTrack that traces sea ice backward in time using 2.4 Sea ice break-up and freeze-up The timing of sea ice break-up and freeze-up (Fig. 2b) was estimated for each year based on CERSAT sea ice concen- tration data for the region between 86◦N, 100◦E and 71◦N, 160◦E. An ice-free grid point is deﬁned as the ﬁrst day in a series of at least 10 d when ice concentration exceeds and reaches zero (Janout et al., 2016). 2.3 Area ﬂux estimates To investigate the impact of winter sea ice dynamics on the summer ice cover, we calculate monthly sea ice area ﬂuxes through the northern boundary of the Laptev Sea for the win- ter season from March to April (1992–2019). The gate is lo- cated between 110 and 160◦E at 77.5◦N (black line with arrows in Fig. 2a). The ﬂux calculations follow the approach of Ricker et al. (2018), who estimated volume ﬂuxes through the Fram Strait. For ice concentration, we use the CERSAT product. For ice motion, we use merged products from CER- SAT that are based on radiometer and scatterometer data. Figure 2c shows the total ice area export from March to April of each winter, including a trend line plotted on top. 2.2.2 Satellite-based and model-based sea ice thickness estimates The satellite-based sea ice thickness observations used in this study are based on the weekly merged CryoSat-2–SMOS sea ice thickness product provided on a 25 km EASE2 grid by the AWI (Ricker et al., 2017). Weekly estimates from April were then averaged in order to obtain monthly sea ice thickness estimates for April 2019 (compare Fig. 2a). 2.2.1 Ice concentration and water depth Ice concentration along the trajectories is provided by CER- SAT and based on 85 GHz SSM/I brightness temperatures. The CERSAT product makes use of the ARTIST Sea Ice (ASI) algorithm and is available on a 12.5km×12.5 km grid (Ezraty et al., 2007). Information on water depth was ob- tained from the International Bathymetry Chart of the Arctic Ocean (IBCAO, Jakobsson et al., 2012). T. Krumpen et al.: The MOSAiC ice ﬂoe The oceanic forcing, including sea surface temperature and salinity, was derived from a climatology based on hydrographic surveys carried out in the Laptev Sea (Janout et al., 2016), where most of the ice originated. https://doi.org/10.5194/tc-14-2173-2020 The Cryosphere, 14, 2173–2187, 2020 T. Krumpen et al.: The MOSAiC ice ﬂoe 2176 T. Krumpen et al.: The MOSAiC ice ﬂoe by the Center for Satellite Exploitation and Research (CER- SAT; Girard-Ardhuin and Ezraty, 2012), (ii) the OSI-405-c motion product from the Ocean and Sea Ice Satellite Ap- plication Facility (OSI SAF; Lavergne, 2016), and (iii) Po- lar Pathﬁnder Daily Motion Vectors (v.4) from the National Snow and Ice Data Center (NSIDC; Tschudi et al., 2016). The contributions of individual products to the used motion ﬁeld are weighted based on their accuracies and availability which vary with seasons, years and study region. The Ice- Track algorithm ﬁrst checks for the availability of CERSAT motion data within a predeﬁned search range. CERSAT pro- vides the most consistent time series of motion vectors start- ing from 1991 to present and has shown good performance on the Siberian shelves (Rozman et al., 2011). During sum- mer months (June–August) when drift estimates from CER- SAT are missing, motion information is bridged with OSI SAF (2012 to present). Prior to 2012, or if no valid OSI SAF motion vector is available within the search range, NSIDC data are applied. The tracking approach works as follows: ice in user-deﬁned individual starting locations or positions on a 25 km EASE2 grid is traced backward in time on a daily basis. Tracking is discontinued if (a) the tracked ice reaches the coastline or fast ice edge or (b) the ice concentration at a speciﬁc location along the backward trajectory drops below 40 % and we assume the ice to be formed. by the Center for Satellite Exploitation and Research (CER- SAT; Girard-Ardhuin and Ezraty, 2012), (ii) the OSI-405-c motion product from the Ocean and Sea Ice Satellite Ap- plication Facility (OSI SAF; Lavergne, 2016), and (iii) Po- lar Pathﬁnder Daily Motion Vectors (v.4) from the National Snow and Ice Data Center (NSIDC; Tschudi et al., 2016). The contributions of individual products to the used motion ﬁeld are weighted based on their accuracies and availability which vary with seasons, years and study region. The Ice- Track algorithm ﬁrst checks for the availability of CERSAT motion data within a predeﬁned search range. CERSAT pro- vides the most consistent time series of motion vectors start- ing from 1991 to present and has shown good performance on the Siberian shelves (Rozman et al., 2011). T. Krumpen et al.: The MOSAiC ice ﬂoe During sum- mer months (June–August) when drift estimates from CER- SAT are missing, motion information is bridged with OSI SAF (2012 to present). Prior to 2012, or if no valid OSI SAF motion vector is available within the search range, NSIDC data are applied. The tracking approach works as follows: ice in user-deﬁned individual starting locations or positions on a 25 km EASE2 grid is traced backward in time on a daily basis. Tracking is discontinued if (a) the tracked ice reaches the coastline or fast ice edge or (b) the ice concentration at a speciﬁc location along the backward trajectory drops below 40 % and we assume the ice to be formed. counts for thermodynamic growth and melting as well as me- chanical redistributions due to ridging (e.g. Thorndike et al., 1975; Lipscomb, 2001). For the purpose of this study, the me- chanical aspect was disregarded in order to focus on thermo- dynamically grown level ice. At each time step, the growth and melt rates are derived from heat ﬂuxes based on atmo- spheric and oceanic forcing by solving conservation laws of snow and ice enthalpy (e.g. be Bitz and Lipscomp, 1999). Every simulation began with open-ocean conditions. The at- mospheric forcing was provided by NCEP reanalysis data (Kanamitsu et al., 2002) and consisted of downward short- and longwave radiation ﬂuxes, surface air temperature and speciﬁc humidity, wind ﬁeld, and precipitation. The oceanic forcing, including sea surface temperature and salinity, was derived from a climatology based on hydrographic surveys carried out in the Laptev Sea (Janout et al., 2016), where most of the ice originated. counts for thermodynamic growth and melting as well as me- chanical redistributions due to ridging (e.g. Thorndike et al., 1975; Lipscomb, 2001). For the purpose of this study, the me- chanical aspect was disregarded in order to focus on thermo- dynamically grown level ice. At each time step, the growth and melt rates are derived from heat ﬂuxes based on atmo- spheric and oceanic forcing by solving conservation laws of snow and ice enthalpy (e.g. be Bitz and Lipscomp, 1999). Every simulation began with open-ocean conditions. The at- mospheric forcing was provided by NCEP reanalysis data (Kanamitsu et al., 2002) and consisted of downward short- and longwave radiation ﬂuxes, surface air temperature and speciﬁc humidity, wind ﬁeld, and precipitation. T. Krumpen et al.: The MOSAiC ice ﬂoe Figure 2. Summary of various processes that affected ice formation in the Laptev Sea and the East Siberian Sea in winter 2018/2019: (a) CryoSat-2–SMOS sea ice thickness anomaly at the end of the winter (April 2019 minus April 2010–2018) in the eastern Eurasian Arctic. A zone of thinner ice was present prior to the onset of melting along the coastline. The ice ﬁeld in which the MOSAiC expedition was set up 5 months later is marked by a dotted line. (b) Estimate of the onset of break-up (red line) and freeze-up (blue line) with their standard deviations and trends between 86◦N, 100◦E and 71◦N, 160◦E. (c) Satellite-based late winter (March–April) ice area export through a “gate” spanning from 110 to 160◦E at 77.5◦N. A trend line is plotted on top. In (a), the gate is depicted as a solid black line. (d, e) Air temperatures (2 m) recorded at Kotelny meteorological station (yellow circle in a) between 1935 and 2019 in the summer (red line) and winter months (blue line). All trends provided in this graph are signiﬁcant at a 95 % conﬁdence level. Fi 2 S f i h ff d i f i i h S d h Sib i S i i 2018/2 Figure 2. Summary of various processes that affected ice formation in the Laptev Sea and the East Siberian Sea in winter 2018/2019: (a) CryoSat-2–SMOS sea ice thickness anomaly at the end of the winter (April 2019 minus April 2010–2018) in the eastern Eurasian Arctic. A zone of thinner ice was present prior to the onset of melting along the coastline. The ice ﬁeld in which the MOSAiC expedition was set up 5 months later is marked by a dotted line. (b) Estimate of the onset of break-up (red line) and freeze-up (blue line) with their standard deviations and trends between 86◦N, 100◦E and 71◦N, 160◦E. (c) Satellite-based late winter (March–April) ice area export through a “gate” spanning from 110 to 160◦E at 77.5◦N. A trend line is plotted on top. In (a), the gate is depicted as a solid black line. (d, e) Air temperatures (2 m) recorded at Kotelny meteorological station (yellow circle in a) between 1935 and 2019 in the summer (red line) and winter months (blue line). All trends provided in this graph are signiﬁcant at a 95 % conﬁdence level. T. Krumpen et al.: The MOSAiC ice ﬂoe While searching for a suitable ﬂoe for the CO, two addi- tional regions were visited (see Fig. 3a, R2 and R3), each consisting of a collection of smaller ﬂoes. Here, manual ice and snow thickness measurements were taken on the level ice with a drill, measuring stick, and thickness gauge. and summer ablation. According to Haas and Eicken (2001), a comparison of GEM measurements performed in the cen- tral Arctic during summer months with drill-hole data indi- cate that the accuracy of the induction measurements is better than 0.05–0.10 m and that the method is well suited for high- resolution thickness proﬁling. For further details on the data processing and handling, we refer Hunkeler et al. (2016). Table 1 summarises the mean and modal thickness of sea ice and snow for all individual sampling sites. It is important to note here that electromagnetic sounding only yields the total ice thickness (snow thickness plus sea ice thickness). Therefore the snow surface layer thickness has to be measured independently to yield ice thickness. Snow thickness measurements on L1–L3 and M8 were obtained ev- ery 2–5 m along the GEM tracks with a magnaprobe (Snow Hydro, Fairbanks, AK, USA). At R1, manual snow thickness measurements were taken at randomly selected locations. Af- ter GEM and magnaprobe measurements were converted to a drift- and rotation-corrected coordinate system using a GPS reference station, sea ice thickness was calculated by sub- tracting total ice thickness from snow thickness. 2.5.1 Snow and ice thickness measurements Ground-based electromagnetic (GEM) induction measure- ments of ice thickness were obtained on ﬁve different resid- ual ice ﬂoes between 1 and 7 October: four ﬂoes were located in the vicinity of the CO (∼15 km) and part of the DN (see Fig. 3a, L1-L3, M8). The ﬁfth ﬂoe was positioned outside the DN and will hereafter be called Reference Site R1. In addition to satellite-based mean thickness estimates, the level ice thickness was computed along the Lagrangian drift trajectories by means of the one-dimensional thermodynamic model Icepack (see CICE Consortium, 2020) that drifted with the ice. The single-column model describes the seasonal evolution of thickness distribution for a single ﬂoe from an initial ice thickness. It uses an approach combining seven ice categories and seven layers (only one layer of snow) and ac- The GEM was mounted on a plastic sledge and pulled across the snow surface. The most frequently occurring ice thickness, the mode of the distribution (compare Fig. 6), rep- resents level ice thickness and is the result of winter accretion The Cryosphere, 14, 2173–2187, 2020 https://doi.org/10.5194/tc-14-2173-2020 2177 T. Krumpen et al.: The MOSAiC ice ﬂoe 2.5.2 Ice coring Ice cores were taken at all the L sites (Fig. 3a) with a stan- dard 9 cm Kovacs ice corer. At L1, four cores were collected. At L2, three cores were taken from level ice and three cores from a ridge at different surface elevations. At L3, three cores were extracted from level ice and three cores at the lower relief area of a ridge. Within the MOSAiC central ﬂoe, ice coring took place at several sites on a weekly basis, but only the sediment-laden sea ice observed at one of the residual ice stations is discussed in this paper. The ice cores were sectioned into 10 cm samples, melted, and then ﬁltered for https://doi.org/10.5194/tc-14-2173-2020 The Cryosphere, 14, 2173–2187, 2020 2178 T. Krumpen et al.: The MOSAiC ice ﬂoe T. Krumpen et al.: The MOSAiC ice ﬂoe Figure 3. Results of Lagrangian sea ice backward tracking (see Methods). (a) Starting point of the MOSAiC expedition (black star: position of the Central Observatory), the spatial extent of the investigation areas deﬁned in this paper (DNR and EMR), and the reference sites where additional ice and snow thickness measurements were obtained. (b) Sea ice age at the start of the MOSAiC expedition on 25 September according to Lagrangian tracking. (c) Water depth at the ice formation site for each tracking position. (d) Average distance of sea ice ravelled from its formation site to its position on 25 September. (e) Sea ice concentration for each individual point, averaged over the ﬁrs 3 months (June–September) of tracking along its trajectory. (f) CryoSat-2 ice thickness estimates in late April, along the trajectory of each point. Figure 3. Results of Lagrangian sea ice backward tracking (see Methods). (a) Starting point of the MOSAiC expedition (black star: position of the Central Observatory), the spatial extent of the investigation areas deﬁned in this paper (DNR and EMR), and the reference sites where additional ice and snow thickness measurements were obtained. (b) Sea ice age at the start of the MOSAiC expedition on 25 September according to Lagrangian tracking. (c) Water depth at the ice formation site for each tracking position. (d) Average distance of sea ice travelled from its formation site to its position on 25 September. (e) Sea ice concentration for each individual point, averaged over the ﬁrst 3 months (June–September) of tracking along its trajectory. T. Krumpen et al.: The MOSAiC ice ﬂoe Table 1. Ice and snow thickness observations obtained on various residual ice ﬂoes in the immediate vicinity (grey, L1–L3, M8) and extended surroundings (R1–R3) of the Central Observatory. The positions of the sites are shown in Fig. 3a. Sample unit indicates either the distance covered by instruments like GEM/magnaprobe (in kilometres) or the number (n) of individual measurements that were performed manually. Numbers in parentheses provide the standard deviation. Table 1. Ice and snow thickness observations obtained on various residual ice ﬂoes in the immediate vicinity (grey, L1–L3, M8) and extended surroundings (R1–R3) of the Central Observatory. The positions of the sites are shown in Fig. 3a. Sample unit indicates either the distance covered by instruments like GEM/magnaprobe (in kilometres) or the number (n) of individual measurements that were performed manually. Numbers in parentheses provide the standard deviation. Ice thickness (m) Snow thickness (m) Total ice thickness (m) Site Sampling device Date Mean Mode Samples Mean Modal Samples Mean Mode L1 GEM/magnaprobe 5 Oct 0.86 (0.66) 0.43 8.7 km 0.10 (0.04) 0.07 n = 659 0.96 0.5 L2 GEM/magnaprobe 7 Oct 0.67 (0.54) 0.33 9.6 km 0.11 (0.04) 0.08 n = 519 0.78 0.41 L3 GEM/magnaprobe 9 Oct 1.0 (0.81) 0.31 7.9 km 0.11 (0.05) 0.06 n = 799 1.11 0.37 M8 GEM/magnaprobe 11 Oct 0.76 (0.75) 0.35 1.2 km 0.09 (0.04) 0.06 n = 385 0.85 0.41 R1 GEM/magnaprobe 1 Oct 0.85 (0.47) 0.62 21 km 0.11 (0.04) 0.09 n = 86 0.96 0.71 R2 Manual 2 Oct 0.55 (0.1) 0.60 n = 38 0.18 0.18 n = 38 0.73 0.78 R3 Manual 2 Oct 0.61 (0.17) 0.70 n = 20 0.06 0.06 n = 20 0.67 0.76 2.5.3 Ice observations from the bridge fect on the ice cover. During late winter months dominated by an offshore-directed drift component, newly formed ice areas are larger and remain comparatively thin and there- fore melt more rapidly once temperatures rise above freezing. This feedback mechanism is even more pronounced when temperatures at the end of winter are unusually high. Fig- ure 2 summarises the conditions and processes that shaped ice formation in the Laptev Sea and East Siberian Sea in win- ter 2018/2019. Satellite-based estimates of offshore-directed sea ice area transport between March and April are shown in Fig. 2c (1992–2019, from 110 to 160◦E at 77.5◦N). Late winter ﬂux estimates indicate that the sea ice advec- tion away from the Siberian shelves towards the central Arc- tic was approximately 70 % higher (2.32×105 km2) in 2019 than the long-term mean annual rate (∼1.36 × 105 km2). Following Krumpen et al. (2013), the strong positive trend (+0.53 × 105 km2 per decade) in late winter ice area export is associated with an increasing drift speed as a result of thin- ning ice cover and a rapid loss of thick multi-year ice. As a consequence of the intensiﬁed ice advection shortly before spring break, satellite-based sea ice thickness observations (Fig. 2a) show negative thickness anomalies throughout the entire coastal zones of the East Siberian Sea and the Laptev Sea in April 2019, except for the southern half of the area around the New Siberian Islands. On board Akademik Fedorov, visual ice observations were carried out from the bridge by a group of three specially trained ice observers. Detailed descriptions of the method- ology and protocols applied are provided in Alekseeva et al. (2019) and AARI (2011), all congruent to the WMO Sea Ice Nomenclature (2017). Continuous 24 h ice observations were available from 28 September (approaching R1) to 3 Oc- tober (approaching the DN). The observations included vi- sual descriptions of the ice cover’s main characteristics, i.e. total concentration and partial concentrations and forms of the encountered stages of ice development, hummock and ridge concentration, melting stage, and sizes and orientations of fractures and leads. In this paper, we will use the observed (within the limits of horizontal visibility) residual ice frac- tion along the ship’s track (see Fig. 5). Data were resampled to an hourly interval. 2.5.2 Ice coring (f) CryoSat-2 ice thickness estimates in late April, along the trajectory of each point. sediments using 0.45 µm ﬁlters. At all sampling sites, paral- lel cores were taken and stored at −20 ◦C for future methane concentration and isotope analysis. Since the MOSAiC ﬂoes may originate from methane supersaturated seawater near the Siberian coast, some of the residual ice may contain relict biogeochemical conditions from the initial ice formation. This further demonstrates the importance of understanding the history of the MOSAiC ﬂoe for future studies. https://doi.org/10.5194/tc-14-2173-2020 https://doi.org/10.5194/tc-14-2173-2020 The Cryosphere, 14, 2173–2187, 2020 2179 https://doi.org/10.5194/tc-14-2173-2020 T. Krumpen et al.: The MOSAiC ice ﬂoe the EMR shortly before MOSAiC’s starting date. Figure 3b shows the age of the sea ice within the EMR on 25 Septem- ber. Based on the backtracking analysis, the EMR’s resid- ual ice had an average age of 318 d and was formed on 11 November 2018 (±15 d). Second-year (SYI) or multi-year ice (MYI) was not found, either from tracking or from scat- terometer data. Most of the residual ice was originally pro- duced during or shortly after the freeze-up in polynyas (or elsewhere on the shallow Siberian shelves) (Fig. 3c), featur- ing water depths of less than 30 m. Only the ice at the far eastern and northern edges of the EMR originated from re- gions with a water depth exceeding 50 m. From the time of its formation to 25 September, the EMR ice had travelled an av- erage distance of approximately 2440 km (±205 km, Fig. 3d) and experienced low ice concentrations between June and September 2019 (Fig. 3e). Hence, the residual ice encoun- tered after our arrival on site was severely weathered, and bridge observations indicated that a large fraction was melted completely during summer months. Residual ice that sur- vived was characterised by frozen-over melt ponds with a < 10 cm deep layer of fresh snow. Based on visual observa- tion, melt pond fraction was 70 %–80 % in the undeformed ice areas, and the bottom layer experienced internal melting. According to ice coring, only the top 30 cm of ice was solid. Because both ships only reached the target region after the freeze-up had begun, large expanses of previously open wa- ter were now covered with new ice. (Fig. S1, Supplement) and observations from the Kotelny meteorological station (Fig. 2a, yellow circle), the tempera- tures during the ice growth phase (October 2018–May 2019) were elevated: reanalysis data show positive temperature anomalies of 3 ◦C in comparison to the 1981–2010 climatol- ogy, and records at Kotelny show signiﬁcantly higher temper- atures than those at the beginning of the instrumental record (Fig. 2e). In particular, temperatures at the end of the win- ter are unusually high. If this coincides, as described above, with periods of strong offshore-directed winds, the formation of new ice in coastal areas is reduced, which favours early melting of the ice cover in spring (Fig. S2, Supplement). The subsequent temperature anomalies in spring and sum- mer 2019 were even more pronounced. T. Krumpen et al.: The MOSAiC ice ﬂoe During the summer months, Kotelny meteorological monitoring station recorded the highest mean temperatures since the beginning of record- keeping (Fig. 2d), and the reanalysis data indicate a pos- itive anomaly of 2.5◦on the Siberian shelves and in ad- jacent northern regions (Fig. S1, Supplement). The rapidly rising temperatures in spring accelerated the melting of the ice cover, which was extremely thin to begin with (Fig. 2a). This resulted in the earliest ice break-up ever observed (com- pare Fig. 2b, red line) and rapid northward retreat of the ice edge, which exposed surface waters to direct solar heat- ing. Consequently, summer (August 2019) sea surface tem- peratures south of the MOSAiC starting area were approx- imately 2–4 ◦C higher than the 1982–2010 mean (Timmer- mans and Ladd, 2019), such that wind events that force ice ﬂoes back into warm waters could have caused additional ice melt (Steele and Ermold, 2015). Moreover, the intensive warming of the upper ocean (Janout et al., 2016) caused a delay in the autumnal freeze-up of sea ice (Fig. 2b, blue line) and resulted in large parts of the marginal seas remaining ice- free for up to 93 d. This means that the MOSAiC expedition started immediately after the longest recorded ice-free period in the region. Based on the backtracking analysis, the ﬂoes selected for the Central Observatory and the DN were located in a zone of comparatively young ice that formed roughly 3 weeks later than the ice within the EMR (Fig. 3b, early December 2018) and originated from a shallow (Fig. 3c) region closer to its lo- cation on 25 September (Fig. 3d, 2240 km). Figure 4a shows the trajectories obtained for the centre of the DNR (the po- sition of the CO, red line) and four adjacent positions at a distance of 25 km (grey lines). Information on water depths and ice concentration along the central trajectory is provided in Fig. 4b, c. The trajectories indicate that the ice inside the DNR was formed in a polynya event on 5 December 2018, north of the New Siberian Islands in water that was less than 10 m deep. An eastward ice drift then transported the newly formed ice along the shallow shelf, until it reached deeper water in early February 2019. T. Krumpen et al.: The MOSAiC ice ﬂoe Ice cores collected at various points in the DN and on the CO conﬁrm that the DNR ice originated in the shallow Siberian shelves, since some of the cores contained sediment inclusions of sandy silt in the up- permost 50 cm (Fig. 4c, d). Though the quantities were small in most cases, these inclusions can only be found on the shal- low Siberian Arctic shelves with average water depths of less than 30 m (Sherwood, 2000; Wegner et al., 2017). There, par- ticulate matter and organisms are incorporated into the newly formed ice by suspension freezing (Eicken et al., 2000) or, to a smaller degree, by grounded sea ice pressure ridges plough- ing through the sea ﬂoor (Darby et al., 2011). A detailed 3.1 Sea ice retreat in summer 2019: preconditioning processes Sea ice retreat during the melting period in the Laptev Sea and East Siberian Sea is the result of atmospheric and oceanic processes and regional feedback mechanisms acting on the ice cover, in both winter and summer. In the following, we will brieﬂy review the sea ice conditions on the Siberian Shelf seas prior to the start of the expedition and the main preconditioning mechanisms that contributed to the north- ward retreat of the ice edge in 2019. In this regard, our focus is on the atmospherically driven processes, since results from oceanographic surveys are not yet available. Ocean-driven preconditioning mechanisms are less well understood. However, there is indication that enhanced win- ter ventilation of the ocean can reduce sea ice formation in this area at a rate now comparable to losses from atmospheric thermodynamic forcing (Polyakov et al., 2017). Observations carried out in the eastern Eurasian Basin have shown that weakening of the halocline and shoaling of intermediate- depth Atlantic water layer result in heat ﬂux equivalent to 40– 54 cm reductions in ice growth in 2013/2014 and 2014/2015. In addition, anomalously high temperatures during the winter months can further reduce the growth of ﬁrst-year ice (FYI), resulting in thinner ice cover at the end of the win- ter (Ricker et al., 2017). According to NCEP reanalysis data Ocean-driven preconditioning mechanisms are less well understood. However, there is indication that enhanced win- ter ventilation of the ocean can reduce sea ice formation in this area at a rate now comparable to losses from atmospheric thermodynamic forcing (Polyakov et al., 2017). Observations carried out in the eastern Eurasian Basin have shown that weakening of the halocline and shoaling of intermediate- depth Atlantic water layer result in heat ﬂux equivalent to 40– 54 cm reductions in ice growth in 2013/2014 and 2014/2015. In addition, anomalously high temperatures during the winter months can further reduce the growth of ﬁrst-year ice (FYI), resulting in thinner ice cover at the end of the win- ter (Ricker et al., 2017). According to NCEP reanalysis data Ice dynamics and ice export in winter are important pre- conditioning mechanisms for the ice retreat in summer. Itkin and Krumpen (2017) observed that enhanced offshore- directed transport of sea ice in late winter has a thinning ef- https://doi.org/10.5194/tc-14-2173-2020 The Cryosphere, 14, 2173–2187, 2020 2180 T. Krumpen et al.: The MOSAiC ice ﬂoe T. Krumpen et al.: The MOSAiC ice ﬂoe T. Krumpen et al.: The MOSAiC ice ﬂoe Figure 4. (a) Lagrangian backward trajectories (see Methods) of the DNR. The multicoloured trajectory line, with colour corresponding to the month of year, indicates the centre of the DNR (Central Observatory). The dashed circle provides the conﬁdence bound of the ice origin. The grey lines provide additional trajectories for four points in the DNR at a distance of 25 km. Derived trajectories were veriﬁed by a manual tracking of the Central Observatory based on Sentinel-1, TerraSAR-X and MODIS (multicoloured circles). The bathymetry is shown in the background. Brownish zones near the coast indicate shallow-water areas of less then 30 m water depth. Panels (b) and (c) show water depth (m) and ice concentration (%) along the trajectory of the Central Observatory. (c) Sediment samples obtained from 10 cm ice core sections at L1 (left: level ice, 20–30 cm depth), L2 (middle: ridged/rafted ice, 243–253 cm depth, processed depth accounting for gaps in the core) and the central ﬂoe (right: ridged/rafted area at 49–59 cm depth). (d) Ice core taken at the central ﬂoe (c, right) with a sediment layer. Figure 4. (a) Lagrangian backward trajectories (see Methods) of the DNR. The multicoloured trajectory line, with colour corresponding to the month of year, indicates the centre of the DNR (Central Observatory). The dashed circle provides the conﬁdence bound of the ice origin. The grey lines provide additional trajectories for four points in the DNR at a distance of 25 km. Derived trajectories were veriﬁed by a manual tracking of the Central Observatory based on Sentinel-1, TerraSAR-X and MODIS (multicoloured circles). The bathymetry is shown in the background. Brownish zones near the coast indicate shallow-water areas of less then 30 m water depth. Panels (b) and (c) show water depth (m) and ice concentration (%) along the trajectory of the Central Observatory. (c) Sediment samples obtained from 10 cm ice core sections at L1 (left: level ice, 20–30 cm depth), L2 (middle: ridged/rafted ice, 243–253 cm depth, processed depth accounting for gaps in the core) and the central ﬂoe (right: ridged/rafted area at 49–59 cm depth). (d) Ice core taken at the central ﬂoe (c, right) with a sediment layer. tent in the atmosphere further reduce accuracy of low- resolution motion products (Sumata et al., 2014), IceTrack uses the OSI SAF motion product to bridge the lack of CER- SAT data. 3.2 Sea ice origin and initial conditions in September 2019 In this section we describe the predominant ice conditions at the beginning of MOSAiC, in both the ship’s immediate vicinity and its extended surroundings. The latter encompass the area within a 220 km radius of Polarstern and will here- after be referred to as the extended MOSAiC region (EMR; see Fig. 3a). A radius was selected to include various ice types, which differ in terms of their provenance (i.e. ori- gin) and/or age. The EMR includes both the ice edge to the south and thicker and more stable pack ice to the north. The ship’s immediate vicinity (distributed network region, DNR) includes the DN and has a radius of 40 km. We will ﬁrst de- scribe the ice conditions in the EMR, before turning our at- tention to the DNR. Once the MOSAiC ﬂoe had been chosen, we applied a tracking tool (see Methods) to the residual ice that was in The Cryosphere, 14, 2173–2187, 2020 https://doi.org/10.5194/tc-14-2173-2020 2181 T. Krumpen et al.: The MOSAiC ice ﬂoe The percentage of residual ice observed along the course of Akademik Fedorov (black circles) is superimposed. (b) Growth and melt of level ice in the EMR, DNR and at R1 (cf. Fig. 3a). the tracking experiment using different combinations of mo- tion products, or higher and lower ice concentration thresh- olds. ence; results not shown here). Nevertheless, the decrease in ice thickness toward the DNR is clearly recognisable in both models and is in agreement with direct ﬁeld observations: Fig. 6 shows the results of the GEM ice thickness measure- ments carried out on four ﬂoes in the distributed network (L1–L3 and M8) and compares them with measurements taken on R1. The measured difference in modal ice thick- nesses (without snow) between R1 and the DNR was 0.3 m (R1: 0.5 m vs. DNR: 0.2 m). Higher ice thicknesses were also measured at R2 and R3 located farther to the north and west, which were reached by helicopter (Table 1, Methods). Note that we originally planned to trace the provenance of the MOSAiC ﬂoe using high-resolution satellite data (Sentinel-1, TerraSAR-X and MODIS). However, only spo- radic high-resolution images of the region were available, and the combination of low summertime sea ice concen- tration and high degree of cloud cover made it extremely difﬁcult to manually track the exact position of individual ﬂoes over an extended period of time. Nevertheless, the high- resolution satellite data enabled us to track nearby large-scale patterns such as shear zones or very prominent ﬂoes. Hence, we could at least determine the approximate location of the MOSAiC ﬂoe on individual images. The resulting estimates for the different positions of the CO (brown-yellow coloured circles in Fig. 4a) correspond well to the computed trajecto- ries (red line in Fig. 4a), which lends increased conﬁdence in our results. Visual observations made from the bridge of the Akademik Fedorov as it travelled along the expedition route provided further evidence for the presence of a thickness gradient be- tween the DNR and EMR. The percentage of residual ice steadily dropped from nearly 90 % at R1 to 20 % at the DNR; conversely, the percentage of thin, newly formed ice rose from 10 % to ca. 80 %. This indicates that, given its lower initial thickness at the end of the winter, some of the ice in the DNR could have completely melted in summer. T. Krumpen et al.: The MOSAiC ice ﬂoe To quantify uncertainties of sea ice trajectories on a larger temporal and spatial scale, we reconstructed the path- ways of drifting buoys using IceTrack. For this purpose, we selected 10 buoys that had survived a full summer and winter in the Arctic. Their drift was then reproduced from October onwards in a backward direction over 12 months. Figure S3 (Supplement) shows the deviation between actual and virtual tracks, which is rather small (60 ± 24 km after 320 d) and in an acceptable range. The maximum deviation between real and virtual buoys gives a measure of the largest possible error that can occur when determining the ice origin. After 320 d it is around 105 km. The conﬁdence bound is shown in Fig. 4 as an ellipsoid (dashed line). No signiﬁcant differences in sea ice pathways and source areas were observed when repeating chemical analysis of these trapped sediments will be con- ducted at a later point in time. The validity and reliability of Lagrangian drift studies de- pend on the accuracy of the applied sea ice motion product. In this study, we primarily use the CERSAT drift dataset because it provides the most consistent time series of mo- tion vectors starting from 1991 to present (see Methods). Comparisons with buoys and high-resolution SAR images indicate that in particular during winter months, when the atmospheric moisture content is low and surface melt pro- cesses are absent, the quality of motion products from low- resolution satellites is high (Sumata et al., 2014; Krumpen et al., 2019). Restrictions may arise from the coarse resolution of the sensors in near-shore regions characterised by a com- plex coastline, extensive fast-ice areas, and polynyas (Roz- man et al., 2011). During summer months (June–August), when strong surface melt processes and high moisture con- https://doi.org/10.5194/tc-14-2173-2020 The Cryosphere, 14, 2173–2187, 2020 2182 T. Krumpen et al.: The MOSAiC ice ﬂoe T. Krumpen et al.: The MOSAiC ice ﬂoe Figure 5. (a) Level ice thickness on 25 September 2019 simulated with a thermodynamic model (see Methods). The percentage of residual ice observed along the course of Akademik Fedorov (black circles) is superimposed. (b) Growth and melt of level ice in the EMR, DNR and at R1 (cf. Fig. 3a). Figure 5. (a) Level ice thickness on 25 September 2019 simulated with a thermodynamic model (see Methods). T. Krumpen et al.: The MOSAiC ice ﬂoe The thickness gradient between the DNR and EMR is conﬁrmed by CryoSat-2–SMOS measurements from the end of win- ter 2018/2019. Already in April 2019, a negative thickness anomaly prevails at the later starting position of the drift ex- periment (Figs. 2a and 3f). To calculate the ice thickness variability in the EMR and DNR at the start of MOSAiC (Fig. 5a) and the ice thickness evolution along the drift trajectories encountered by the ice in those regions (Fig. 5b), we used the results of a thermody- namic model (see Methods). Results show that the residual ice in the DNR was not only younger and originated from a different location than the ice in the surrounding EMR, but it was also thinner: on 25 September, the averaged ice thick- ness inside the EMR was 0.58 m (±0.27 m), while the thick- ness of ice inside the DNR was 0.37 (±0.09 m), i.e. 36 % (0.21 m) less than in the EMR. To conﬁrm model results, we applied a second, simpler thermodynamic model devel- oped by Thorndike (1992) and used in Peeken et al. (2018) and Krumpen et al. (2019). The model is chieﬂy based on air temperatures, assumes a constant ocean heat ﬂux and em- ploys snow climatology, but indicates the existence of similar thickness gradients between the EMR and DNR (40 % differ- T. Krumpen et al.: The MOSAiC ice ﬂoe However, in order to verify that the model is capable to reproduce the interan- nual variability correctly, model estimates require compari- son to historical observational data from the past. Unfortu- nately, ﬁeld surveys in this exact location and that time of the year are scarce, but GEM ice thickness measurements in the surroundings of the DNR between 84 and 86.5◦N and 100 and 150◦E (compare Fig. 3) were obtained by Haas and Eicken (2001) during the ARK-12 cruise of Polarstern in August 1996. The authors obtained around 37 km of thick- ness proﬁle data at 5 m horizontal spacing. They found aver- age FYI modal thicknesses of ∼1.85 m, typical for SYI or even MYI in summer. The 1996 GEM measurements were obtained 6 weeks earlier in the melt season (10 to 22 Au- gust 1996) inside the EMR area and south of it. In com- parison, the exceptionally thick September 1996 ice is re- produced by our thermodynamic model with 1.6 m in the DNR (Fig. 7c). According to Haas and Eicken (2001), the relatively thick ice in 1996 was due to speciﬁc atmospheric circulation conditions during summer, characterised by per- sistent low sea level pressure over the central Arctic. This resulted in very weak surface melt and the absence of melt ponds north of approximately 84◦N in 1996. The model re- sults and forcing data for 1996 conﬁrm that strongly reduced net shortwave ﬂuxes led to a signiﬁcant reduction in ice melt- ing during the summer months. Even in years dominated by strong melting processes, the model seems to realistically re- produce ice thickness: in winter 2013/2014, ice formed com- Figure 7c displays the time series of September FYI thick- ness estimates in the DNR for the period between 1994 and 2019. In addition, Fig. 7d provides the annual cycle of DNR ice growth and melt. An overall decrease in residual ice thickness between 1994 and 2019 is visible (trend: −0.22 m per decade), which is subject to a high interannual variability and therefore not statistically signiﬁcant. The DNR ice en- countered in September 2019 can be classiﬁed as exception- ally thin over a longer period of time (Fig. 7c). However, for the larger region of the EMR, ice thicknesses in September 2019 agree well with the long-term average (Fig. 7d). T. Krumpen et al.: The MOSAiC ice ﬂoe Figure 6. Total (ice plus snow, a) and snow (b) thickness distribution of the ﬂoes located inside the DNR (L1-3, M8, red line) and at R1 (blue line; see Fig. 3a for positions). Ice thickness measurements were made with a ground-based electromagnetic (GEM) instrument pulled across the ice on a sledge. Snow thickness measurements were made with a magnaprobe. Figure 6. Total (ice plus snow, a) and snow (b) thickness distribution of the ﬂoes located inside the DNR (L1-3, M8, red line) and at R1 (blue line; see Fig. 3a for positions). Ice thickness measurements were made with a ground-based electromagnetic (GEM) instrument pulled across the ice on a sledge. Snow thickness measurements were made with a magnaprobe. region, supporting the notion of exceptionally thin ice in the MOSAiC starting region. In this section, we compare the conditions we encountered at the end of September 2019 with those of previous years by applying the combined tracking–thermodynamics model to the period between 1994 and 2019. Figure 7a shows the history and variation in imag- inary MOSAiC ﬂoe trajectories for the past 26 years. Track- ing was performed backwards in time starting from the DNR region on 25 September of each year. Results indicate that the climatological probability that DNR ice originates from the New Siberian Islands, like in 2019, is about 25 % (red shaded area and tracks). From a climatological perspective, it is usually more likely that the ice at the starting position has its origin in the Laptev Sea (55 %, light blue shaded area). A smaller part (∼20 %) typically comes from the East Siberian Sea (grey shaded area). The approximate age of the ice near the starting point is around either 1 or 2 years (Fig. 7b), with a tendency towards decreasing ice age. This tendency of decreasing ice age is evident from the frequency of SYI. While SYI occurred in about 64 % of all years between 1992 and 2004, it was already much less frequent during the past 15 years (20 %, 2005–2019). of April, followed by above-average melt. An in-depth anal- ysis of the applied forcing data in the thermodynamic model reveals that the intensiﬁed ice production is a consequence of reduced precipitation rates in winter 2018/2019 (Fig. S4, Supplement). Through a comparison with in situ data, we have shown above that the thermodynamic model is able to simulate regional differences in ice thickness. 3.3 MOSAiC ice conditions compared to previous years We showed that due to its younger age and different prove- nance, the DNR ice was thinner than the surrounding ice. But the thicknesses measurements summarised in Fig. 6 and Table 1 are also much smaller than what was observed by Haas and Eicken (2001) in the 1990s by similar GEM and drill-hole measurements. They found late-summer modal FYI thicknesses between 1.25 m (1995), 1.75 m (1993), and 1.85 m (1996) in regions near or south of the MOSAiC study The Cryosphere, 14, 2173–2187, 2020 https://doi.org/10.5194/tc-14-2173-2020 2183 T. Krumpen et al.: The MOSAiC ice ﬂoe Figure 7. Ice origin, age and thickness of the DNR ice on 25 September between 1994 and 2019: (a) trajectories from the past 26 years separated by the region of origin: (i) blue: Laptev Sea; (ii) red: region north of the New Siberian Islands; (iii) grey: East Siberian Sea. (b) Age of the ice in the DNR region on 25 September of each year. (c) Thickness of DNR FYI based on a thermodynamic model (Methods). (d) Annual cycle of FYI growth and melt. Figure 7. Ice origin, age and thickness of the DNR ice on 25 September between 1994 and 2019: (a) trajectories from the past 26 years separated by the region of origin: (i) blue: Laptev Sea; (ii) red: region north of the New Siberian Islands; (iii) grey: East Siberian Sea. (b) Age of the ice in the DNR region on 25 September of each year. (c) Thickness of DNR FYI based on a thermodynamic model (Methods). (d) Annual cycle of FYI growth and melt. further delayed freeze-up and led to the longest ice-free pe- riod since the beginning of satellite observations. paratively late in the season and melted completely during summer (Fig. 7d). Satellite sea ice concentration data con- ﬁrm that the DNR region and large parts of the EMR were ice-free already at the beginning of August 2014. If com- bined with reliable trajectory and realistic forcing data, the good agreement between the thermodynamic model and ob- servations for the years 1996, 2014 and 2019 shows that the model can be used to study interannual variability of FYI thickness changes and the driving mechanisms behind them. Backward trajectories of sea ice present in the large EMR around Polarstern during the initial phase of the MOSAiC drift experiment indicate that the majority of residual ice was formed shortly after freeze-up in 2018. In comparison, the ice within the smaller DNR around Polarstern was 3 weeks younger and formed on the shallow shelves north of the New Siberian Islands. Sediments discovered in ice cores conﬁrm contact of sea ice with shallow waters in an early phase of growth. While in recent years the strong ice retreat in sum- mer melts most of the shallow-water ice on its way to the cen- tral Arctic Ocean (Krumpen et al., 2019), part of the resid- ual ice encountered in the DNR has survived summer melt. T. Krumpen et al.: The MOSAiC ice ﬂoe Therefore, besides the original goals, MOSAiC will also pro- vide an excellent opportunity to better understand the role of sea ice as a transport medium for climate-relevant gases, macronutrients, iron, organic matter, sediments and pollu- tants from shelf areas to the central Arctic Ocean and be- yond. This is particularly important because with predicted changes towards a seasonally ice-free ocean under climate change a complete cut-off of the long-range transport of ice- rafted materials by the Transpolar Drift appears possible in the future. By comparing transport rates of residual ice with newly formed ice on site, one can examine the impact a re- duced long-range transport of sea ice has for the redistribu- tion of biogeochemical matter in the central Arctic Ocean. T. Krumpen et al.: The MOSAiC ice ﬂoe Both DNR and EMR ice shows above-average growth rates in win- ter 2018/2019 as well as above-average thicknesses at the end https://doi.org/10.5194/tc-14-2173-2020 The Cryosphere, 14, 2173–2187, 2020 2184 T. Krumpen et al.: The MOSAiC ice ﬂoe T. Krumpen et al.: The MOSAiC ice ﬂoe R1–R3) and the DNR are also evident in direct ﬁeld obser- vations. It is therefore to be expected that the momentum and energy transfer between the ocean and the atmosphere is subject to strong spatial variations. Future studies will show whether these regional differences can be reproduced using high-resolution models and satellite data. Whether the ob- served thickness gradients also inﬂuence ice dynamics in the immediate and extended surroundings of the Central Obser- vatory is another exciting research question, and a compari- son of the ice dynamics in the DNR and EMR derived from satellite data is work in progress. However, we assume that the encountered regional differences will balance out during the ice growth phase and thus reduce the spatial variability in ice dynamics over the course of the winter and over the course of the whole MOSAiC expedition. Competing interests. The authors declare that they have no conﬂict of interest. Competing interests. The authors declare that they have no conﬂict of interest. Acknowledgements. This work was carried out as part of the Russian-German Research Cooperation QUARCCS funded by the German Ministry for Education and Research (BMBF) under grant 03F0777A and CATS under grant 63A0028B. Data used in this pa- per were produced as part of the international Multidisciplinary drifting Observatory for the Study of the Arctic Climate (MO- SAiC) with the tag MOSAiC20192020 (AWI_PS122_1 and AF- MOSAiC-1_00). NCEP Reanalysis 2 data are made available by NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from their web- site at https://www.esrl.noaa.gov/psd/ (last access: 25 June 2020). The work on satellite remote sensing data was partly funded through the EU H2020 project SPICES (640161), the ESA Sea Ice CCI phase 1 and 2 (AO/1-6772/11/I-AM), and the Helmholtz PACES II (Polar regions And Coasts in the changing Earth System) and FRAM (FRontiers in Arctic marine Monitoring) programmes. TerraSAR-X images were provided by the German Aerospace Cen- ter (DLR) through TSX Science AO OCE3562. We thank the crew of the research vessels Akademik Fedorov and Polarstern and the helicopter company Naryan-Marsky for their great logistical sup- port during the set-up of the MOSAiC experiment and participants of the Akademik Fedorov cruise and MOSAiC School for helping hands. The ice thickness in September 2019 can be classiﬁed as exceptionally thin when compared to the last 26 years. T. Krumpen et al.: The MOSAiC ice ﬂoe In this sense, we might have already experienced the “new nor- mal” of Arctic conditions during the initial phase of MO- SAiC, which might make future follow-up campaigns of this scale increasingly difﬁcult. An only seasonally ice-covered Arctic with a reduced (or even cut-off) transport of ice-rafted material by the Transpolar Drift will have strong implications for the redistribution of biogeochemical matter in the central Arctic Ocean, with consequences for the balance of climate- relevant trace gases, primary production and biodiversity in the Arctic Ocean. Financial support. This research has been supported by the Ger- man Ministry for Education and Research (grant no. 03F0777A), the German Ministry for Education and Research (grant no. 63A0028B), the German Aerospace Center (grant no. AO OCE3562), the German Minsitry for Education and Research (MOSAiC20192020), the EU H2020 (grant no. 640161), and the European Space Agency (grant no. AO/1-6772/11/I-AM). Code and data availability. All data are archived in the MOSAiC Central Storage (MCS) and will be available on PANGAEA after ﬁ- nalisation of the respective datasets according to the MOSAiC data policy. The production of the merged CryoSat-SMOS sea ice thick- ness data was funded by the ESA project SMOS & CryoSat-2 Sea Ice Data Product Processing and Dissemination Service, and data was obtained from http://meereisportal.de (ftp://ftp.awi.de/sea_ice/ product/cryosat2_smos/v202/, Hendricks and Ricker, 2019). NCEP Reanalysis 2 data are made available by NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from their website at https://www.esrl. noaa.gov/psd/ (NOAA, 2020). The article processing charges for this open-access publication were covered by a Research The article processing charges for this open-access The article processing charges for this open-access publication were covered by a Research Centre of the Helmholtz Association. Centre of the Helmholtz Association. Review statement. This paper was edited by Yevgeny Aksenov and reviewed by two anonymous referees. Review statement. This paper was edited by Yevgeny Aksenov and reviewed by two anonymous referees. Supplement. The supplement related to this article is available on- line at: https://doi.org/10.5194/tc-14-2173-2020-supplement. 4 Conclusion and implications for future studies In this study, we investigate the initial ice conditions and pre- conditioning mechanisms at the start of the MOSAiC drift experiment. Moreover, we evaluate how representative the ice within the distributed network region (DNR) is compared to the experiment’s extended surroundings (extended MO- SAiC region, EMR), and we question whether the ice en- countered was unusually thin compared to past years. An analysis of satellite-based observations, reanalysis data and readings from the meteorological station Kotelny from 2019 indicates that sea ice retreat in the Siberian Shelf seas was strongly inﬂuenced by ice dynamics in late winter and unusually high temperatures in summer. A high offshore- directed transport of sea ice shortly before the onset of spring resulted in unusually thin ice cover throughout the entire coastal zones of the marginal seas in April. Rapidly rising temperatures with record temperatures in summer acceler- ated the melting of the thin ice cover and caused the earliest break-up since 1992. Intensive warming of the upper ocean The application of the thermodynamic model reveals that ice in the DNR is 36 % thinner than the surrounding ice due to its younger age and different provenance of origin. Dif- ferences in modal ice thickness between outer areas (sites https://doi.org/10.5194/tc-14-2173-2020 The Cryosphere, 14, 2173–2187, 2020 2185 https://doi.org/10.5194/tc-14-2173-2020 The Cryosphere, 14, 2173–2187, 2020 T. Krumpen et al.: The MOSAiC ice ﬂoe H.: Remapping the thickness distribution in sea ice models, J. Geophys. Res.-Oceans, 106, 13989–14000, 2001. NOAA: NCEP Reanalysis 2 data, available at: https://www.esrl. noaa.gov/psd/, last access: 25 June 2020 Hendricks, S. and Ricker, R.: Product User Guide & Algo- rithm Speciﬁcation: AWI CryoSat-2 Sea Ice Thickness (version 2.2), available at: https://epic.awi.de/id/eprint/50033/ (last ac- cess: 25 June 2020), 2019. Polyakov, I. V., Pnyushkov, A. V., Alkire, M. B., Ashik, I. M., Bau- mann, T. M., Carmack, E. C., Goszczko, I., Guthrie,J., Ivanov, V. V., Kanzow, T., Krishﬁeld, R., Kwok, R.,Sundfjord, A., Morison, J., Rember, R., and Yulin, A.: Greater role for Atlantic inﬂows on sea-ice loss in the Eurasian Basin of the Arctic Ocean, Science, 6335, 285–291, https://doi.org/10.1126/science.aai8204, 2017. Hunkeler, P., Hoppmann, M., Hendricks, S., Kalscheuer, T., and Gerdes, R.: A glimpse beneath Antarctic sea ice: platelet-layer from multi-frequency electromagnetic induction sounding, Geo- phys. Res. Lett., 43, 1, https://doi.org/10.1002/2015GL065074, 2016. Peeken, I., Primpke, S, Beyer, B., Guetermann, J., Katlein, C., Krumpen, T., Bergmann, M., Hehemann, L., and Gerdts, G.: Arctic sea ice is an important temporal sink and means of transport for microplastic, Nat. Commun., 9, 1509, https://doi.org/10.1038/s41467-018-03825-5, 2018. Itkin, P. and Krumpen, T.: Winter sea ice export from the Laptev Sea preconditions the local summer sea ice cover and fast ice decay, The Cryosphere, 11, 2383–2391, https://doi.org/10.5194/tc-11- 2383-2017, 2017. Ricker, R., Hendricks, S., Kaleschke, L., Tian-Kunze, X., King, J., and Haas, C.: A weekly Arctic sea-ice thickness data record from merged CryoSat-2 and SMOS satellite data, The Cryosphere, 11, 1607–1623, https://doi.org/10.5194/tc-11-1607-2017, 2017. Jakobsson, M., Mayer, L., Coakley, B, Dowdeswell, J., Forbes, S., Fridman, B., Hodnesdal, H., Noormets, R., Pedersen, R., Rebesco, M., Schenke, H. W., Zarayskaya, Y., Accettella, D., Armstrong, A., Anderson, R. M., Bienhoff, P., Camerlenghi, A., Church, I., Edwards, M., Gardner, J., Hall, J., Hell, B., Hestvik, O., Kristoffersen, Y., Marcussen, C., Mohammad, R., Mosher, D., Nghiem, S., Pedrosa, M., Travaglini, P., and Weatherall, P.: The International Bathymetric Chart of the Arctic Ocean (IBCAO) Version 3.0, Geophys. Res. Lett., 39, L12609, https://doi.org/10.1029/2012GL052219, 2012. Ricker, R., Girard-Ardhuin, F., Krumpen, T., and Lique, C.: Satellite-derived sea ice export and its impact on Arc- tic ice mass balance, The Cryosphere, 12, 3017–3032, https://doi.org/10.5194/tc-12-3017-2018, 2018. T. Krumpen et al.: The MOSAiC ice ﬂoe 2186 CICE Consortium Icepack: Icepack version 1.1.0, Zenodo, https://doi.org/10.5281/zenodo.1213462, 2020. trends in Laptev Sea ice outﬂow between 1992–2011, The Cryosphere, 7, 349–363, https://doi.org/10.5194/tc-7-349-2013, 2013. Damm, E., Bauch, D., Krumpen, T., Rabe, B., Korhonen, M., Vinogradova, E., and Uhlig, C.: The Transpolar Drift conveys methane from the Siberian Shelf to the central Arctic Ocean, Sci. Rep., 8, 4515, https://doi.org/10.1038/s41598-018-22801-z, 2018. Krumpen, T. and Sokolov, V.: The Expedition AF122/1 : Setting up the MOSAiC Distributed Network in October 2019 with Re- search Vessel AKADEMIK FEDOROV, Berichte zur Polar- und Meeresforschung = Reports on polar and marine research, Bre- merhaven, Alfred Wegener Institute for Polar and Marine Re- search, 744 , 119 pp., https://doi.org/10.2312/BzPM_0744_2020, 2020. Darby, D. A., Myers, W. B., Jakobsson, M., and Rigor, I.: Modern dirty sea ice characteristics and sources: The role of anchor ice, J. Geophys. Res.-Oceans, 116, 2156–2202, 2011. Krumpen, T., Gerdes, R., Haas, C., Hendricks, S., Herber, A., Se- lyuzhenok, V., Smedsrud, L., and Spreen, G.: Recent summer sea ice thickness surveys in Fram Strait and associated ice volume ﬂuxes, The Cryosphere, 10, 523–534, https://doi.org/10.5194/tc- 10-523-2016, 2016. Eicken, H., Koatschek, J., Lindemann, F., Dmitrenko, I., Freitag, J., and Kassens, H.: A key source area and constraints on entrain- ment for basin-scale sediment transport by Arctic sea ice, Geo- phys. Res. Lett., 27, 13, https://doi.org/10.1029/1999GL011132, 2000. Krumpen, T., Belter, J., Boetius, A., Damm, E., Haas, C., Hen- dricks, S., Nicolaus, M., Nöthig, E. M. , Paul, S., Peeken, I., Ricker, R., and Stein, R.: Arctic warming interrupts the Transpo- lar Drift and affects longrange transport of sea ice and ice-rafted matter, Sci. Rep., 9, 5459, https://doi.org/10.1038/s41598-019- 41456-y, 2019. Ezraty, R., Girard-Ardhuin, F., Piolle, J. F., Kaleschke, L., Heygster, G.: Arctic and Antarctic Sea Ice Concentration and Arctic Sea Ice Drift Estimated from Special Sensor Microwave Data, Techni- cal Report, Departement d’Oceanographie Physique et Spatiale, IFREMER, Brest, France, 2007. Girard-Ardhuin, F. and Ezraty, R.: Enhanced arctic sea ice drift estimation merging radiometer and scatterometer data, IEEE T. Geosci. Remote., 50, 2639–2648, 2012. Lavergne, T.: Validation and Monitoring of the OSI SAF Low Res- olution Sea Ice Drift Product (v5), Technical Report, The EU- METSAT Network of Satellite Application Facilities, July 2016 Haas, C. and Eicken, J.: Interannual variability of summer sea ice thickness in the Siberian and central Arctic under different atmo- spheric circulation regimes, J. Geophys. Res., 106, 4449–4462, 2001. Lipscomb, W. References AARI: Guidance to Special Shipborne Ice Observations, Technical Report, Arctic and Antarctic Research Insitute (AARI), Saint- Petersburg, Russia, 2011. Author contributions. TK conceived the study and wrote the paper. FB, FK, SH, JB, VS, LvA, CH, TR, RR, VS, ED, AT, JH and SuS undertook the data analysis, developed the methods or contributed to interpretation of results. Field observations (thickness of snow and ice, bridge observations, ice cores, etc.) were made and pro- cessed by VB, TP, AM, AT, MH, ES, NK, JR, JB, JH, MT and MA. All authors commented on the manuscript Alekseeva, T., Tikhonov, V., Frolov, S., Repina, I., Raev, M., Sokolova, J., Sharkov, E., Afanasieva, E., and Serovetnikov, S.: Comparison of Arctic Sea Ice Concentration from the NASA Team, ASI, and VASIA2 Algorithms with Sum- mer and Winter Ship Data, Remote Sensing, 11, 2481, https://doi.org/h10.3390/rs11212481, 2019. Bitz, C. M. and Lipscomb, W. H.: An energy-conserving thermody- namic sea ice model for climate study, J. Geophys. Res.-Oceans, 104, C7, https://doi.org/10.1029/1999JC900100, 1999. The Cryosphere, 14, 2173–2187, 2020 https://doi.org/10.5194/tc-14-2173-2020 The Cryosphere, 14, 2173–2187, 2020 T. Krumpen et al.: The MOSAiC ice ﬂoe Rozman, P., Hoelemann, J., Krumpen, T., Gerdes, R., Koeberle, C., Lavergne, T., and Adams, S.: Validating satellite derived and modelled sea-ice drift in the Laptev Sea with in situ mea- surements from the winter of 2007/08, Polar Res., 30, 7218, https://doi.org/10.3402/polar.v30i0.7218, 2011. Janout, M., Hölemann, Waite, A. M., Krumpen, T., Appen, W. J., and Martynov, F.: Sea-ice retreat controls timing of summer plankton blooms in the Eastern Arctic Ocean, Geophys. Res. Lett., 43, 24, https://doi.org/10.1002/2016GL071232, 2016. Sherwood, C. R.: Numerical model of frazil ice and suspended sed- iment concentrations and formation of sediment laden ice in the Kara Sea, J. Geophys. Res.-Oceans, 105, 14061–14080, 2000. Kanamitsu, M., Ebisuzaki, W., Woollen, J., Yang, S. K., Hnilo, J. J., Fiorinom M., and Potter, G., H.: NCEP-DOE AMIP-II Reanaly- sis (R-2), B. Am. Meteorol. Soc., 83, 1631–1643, 2002. Steele, M. and Ermold, W.: Loitering of the retreating sea iceedge in the Arctic Seas, J. Geophys. Res., 120, 7699–7721, 2015. Krumpen, T., Janout, M., Hodges, K. I., Gerdes, R., Girard- Ardhuin, F., Hölemann, J. A., and Willmes, S.: Variability and Sumata, H., Lavergne, T., Girard-Ardhuin, F., Kimura, N., Tschudi, M. A., Kauker, F., Karcher, M., and Gerdes, R.: An intercompar- https://doi.org/10.5194/tc-14-2173-2020 The Cryosphere, 14, 2173–2187, 2020 https://doi.org/10.5194/tc-14-2173-2020 The Cryosphere, 14, 2173–2187, 2020 T. Krumpen et al.: The MOSAiC ice ﬂoe 2187 ison of Arctic ice drift products to deduce uncertainty estimates, J. Geophys. Res.-Oceans, 119, 4887–4921, 2014. Tschudi, M., Fowler, C., Maslanik, J., and Stewart, J. S.: Polar Pathﬁnder Daily 25 km EASE-Grid Sea Ice Motion Vectors, Ver- sion 3. Technical report, NASA National Snow and Ice Data Cen- ter Distributed Active Archive Center, Boulder, Colorado USA, 2016. Thorndike, A. S.: A toy model linking atmospheric thermal radi- ation and sea ice growth, J. Geophys. Res.-Oceans, 97, 9401– 9410, 1992. Thorndike, A. S., Rothrock, D. A., Maykut, G. A., and Colony, R.: The thickness distribution of sea ice, J. Geophys. Res., 80, 4501– 4513, 1975. Wegner, C., Wittbrodt. K., Hoelemann J. A., Janout, M. A., Krumpen, T., Selyuzhenok, V, Novikhin, A., Polyakova, Y., Krykova, I., Kassens H., and Timokhov, L.: Sediment entrain- ment into sea ice and transport in the Transpolar Drift: A case study from the Laptev Sea in winter 2011/2012, Cont. Shelf Res., 141, 1–10, 2017. Timmermans, M. L. and Ladd, C.: Arctic Report Card 2019, Sea Surface Temperatures, edited by: Richter-Menge, J., Drucken- miller, M. L., and Jeffries, M., available at: http://www.arctic. noaa.gov/Report-Card (last access: 25 June 2020), 2019. WMO: Sea Ice Nomenclature, WMO/OMM/BMO – No. 259, Edi- tion 1970–2017, Terminology, Volume I, 2017. The Cryosphere, 14, 2173–2187, 2020 https://doi.org/10.5194/tc-14-2173-2020
https://openalex.org/W2740116847	https://link.springer.com/content/pdf/10.1007%2Fs40789-017-0177-y.pdf	English	null	CO hydrogenation combined with water-gas-shift reaction for synthetic natural gas production: a thermodynamic and experimental study	International Journal of Coal Science & Technology/International journal of coal science & technology	2,017	cc-by	7,541	CO hydrogenation combined with water-gas-shift reaction for synthetic natural gas production: a thermodynamic and experimental study Fanhui Meng1 • Xin Li1 • Xiaoyang Lv1 • Zhong Li1 Received: 9 March 2017 / Revised: 19 April 2017 / Accepted: 11 July 2017 / Published online: 28 July 2017 The Author(s) 2017. This article is an open access publication Abstract The hydrogenation of CO to synthetic natural gas (SNG) needs a high molar ratio of H2/CO (usually large than 3.0 in industry), which consumes a large abundant of hydrogen. The reverse dry reforming reaction (RDR, 2H2 ? 2CO $ CH4 ? CO2), combining CO methanation with water-gas-shift reaction, can signiﬁcantly decrease the H2/CO molar ratio to 1 for SNG production. A detailed thermodynamic analysis of RDR reaction was carried out based on the Gibbs free energy minimization method. The effect of temperature, pressure, H2/CO ratio and the addition of H2O, CH4, CO2, O2 and C2H4 into the feed gas on CO conversion, CH4 and CO2 selectivity, as well as CH4 and carbon yield, are discussed. Experimental results obtained on homemade impregnated Ni/Al2O3 catalyst are compared with the calculations. The results demonstrate that low temperature (200–500 C), high pressure (1–5 MPa) and high H2/CO ratio (at least 1) promote CO conversion and CH4 selectivity and decrease carbon yield. Steam and CO2 in the feed gas decrease the CH4 selectivity and carbon yield, and enhance the CO2 content. Extra CH4 elevates the CH4 content in the products, but leads to more carbon formation at high temperatures. O2 signiﬁcantly decreases the CH4 selectivity and C2H4 results in the generation of carbon. Int J Coal Sci Technol (2018) 5(4):439–451 https://doi.org/10.1007/s40789-017-0177-y Int J Coal Sci Technol (2018) 5(4):439–451 https://doi.org/10.1007/s40789-017-0177-y CO hydrogenation combined with water-gas-shift reaction for synthetic natural gas production: a thermodynamic and experimental study Keywords Synthetic natural gas Reverse dry reforming of methane Gibbs free energy minimization Experimental study CO conversion List of symbols Ak Total mass of k element in the feed fi H Standard-state fugacity of species i (Pa) fi Fugacity of species i (Pa) Gi Gibbs free energy of species i (J/mol) Gi H Standard Gibbs free energy of species i (J/mol) DGﬁ H Standard-state Gibbs free energy of formation of species i (J/mol) GC(g) Partial molar Gibbs free energy of gas carbon (J/mol) GC(s) Partial molar Gibbs free energy of solid carbon (J/mol) GfC(s) H Standard-state Gibbs function of formation of solid carbon (J/mol) 4rHm H Standard-state reaction enthalpy change (J/mol) KH Standard-state equilibrium constant ni Mole of species I (mol) nC Mole of carbon (mol) N Number of components P System pressure (Pa) pH Pressure of the standard state (Pa) R Molar gas constant (J/(mol K) T Temperature (K) yi Mole fraction of species i Greek symbols aik Number of atoms of the k element present in each molecule of species i li Chemical potential of species i (J/mol) /i Fugacity coefﬁcient of species i kk Lagrange multiplier & Zhong Li lizhong@tyut.edu.cn 1 Key Laboratory of Coal Science and Technology of Ministry of Education and Shanxi Province, Institute of Coal Chemical Engineering, Taiyuan University of Technology, No. 79 West Yingze Street, Taiyuan 030024, Shanxi, China & Zhong Li lizhong@tyut.edu.cn 1 Introduction Natural gas is a highly efﬁcient and clean fossil fuel due to its high caloriﬁc value, low sooting tendency and slag free prod- ucts, leading to its increasing consumption year by year (Gao et al. 2015; Meng et al. 2015a; Ro¨nsch et al. 2016). In 2014, the consumption of natural gas in China increased to 197.3 billion cubic meters, with a growth rate of 30.9% every year in the last decade (BP 2016). Recently, the consumption of natural gas has raised a serious concern regarding its depletion because of its limited reserves (Kopyscinski et al. 2010; Huo et al. 2013), in comparison, coal is considered as a much more abundant energy resource in many countries. The production of synthetic natural gas (SNG) from coal has been developed to be a potential route to circumvent the limited supply of natural gas, especially in China (Li et al. 2014a, b; Lu et al. 2014). In literatures, data are available for the thermodynamic analysis of methanation reactions. Miguel et al. (2015) and Sahebdelfar et al. (2015) conducted a thermodynamic cal- culation of CO2 methanation based on the method of Gibbs free energy minimization and compared with the experi- mental data. Gao et al. (2012) analyzed the thermodynamic properties of several reactions during the complete metha- nation of CO and CO2. However, these thermodynamic studies were carried out based on the complete methanation reactions. To our knowledge, there is little information on the thermodynamic analysis of the RDR reaction that occurs at low temperatures. Therefore, it needs to perform the calculations based on the Gibbs free energy minimization method and validate the data through experimental means. p y ( , ; ) Among the coal-to-SNG production processes, SNG is produced through the four major steps, i.e., coal gasiﬁca- tion, water-gas-shift (WGS) reaction (CO ? H2O $ H2 ? CO2), gas cleaning and CO methanation (3H2 ? CO $ CH4 ? H2O) (Shinde and Madras 2014; Wang et al. 2015). The CO methanation reaction is a key process for increasing SNG production (Meng et al. 2015b; Go¨tz et al. 2016; Gao et al. 2016). If one mole of CO is converted to methane, three moles of H2 are stoichiometrically required. However, the content of carbon in coal is usually more than 60 wt% (up to more than 90 wt% in bituminous coal) whereas that of hydrogen is \5 wt% (Martelli et al. Greek symbols 440 F. Meng et al. signiﬁcantly affected the activity and stability of Ni/SiO2 catalysts. Jiang et al. (2013, 2014) investigated the stepwise sulﬁdation and sulﬁdation temperature on the catalytic activity of MoO3/CeO2–Al2O3. It is because there are many advantages of this reaction. First, the feed gas has a low H2/ CO ratio of one, which needs less hydrogen; Secondly, the water-free products can diminish the damage of steam on methanation catalyst, and the by-product CO2 can be easily removed by employing low-temperature methanol puriﬁ- cation process. In other words, less H2 is needed from the gas of coal gasiﬁcation, which simpliﬁes the SNG pro- duction process and reduces the cost. 2 Methods The reverse dry reforming (RDR) reaction (2CO ? 2H2 $ CH4 ? CO2), which is the combination of CO methanation with WGS reaction, can be used to produce SNG. Recently, many studies have focused on the RDR reaction with the H2/CO ratio of one. Yan et al. (2013) found that the catalyst preparation methodologies 1 Introduction 2011; Shen et al. 2016). The high content of carbon in coal results in low H2/CO molar ratios, usually less than one, of pro- duced gas from coal gasiﬁcation (Messerle et al. 2016). For example, the produced gas of the British Gas-Lurgi (BGL) coal gasiﬁcation process is composed of 60%–70% CO, 27%–30% H2, 0%–7% CH4, 1%–4% CO2, and trace amounts of O2 and light hydrocarbons (Yu and Wang 2010). To increase the H2/CO ratio, the WGS reaction should be well controlled. It is worthwhile to mention, in order to enhance the CO conversion and CH4 yield during industry processes, an even higher H2/CO ratio is usually used. For instance, the H2/CO ratio of the Lurgi process for methanation was optimized at about 3.2, and that of the Topsøe Recycle Energy Efﬁcient Methanation (TREMP) process reached about 3.5 (Kopyscinski et al. 2010). More amount of CO needs to be converted to produce H2 by WGS reaction in order to get a high H2/CO ratio, which results in the high operating cost and energy consumption. It is well known that the produced gas from coal con- tains many impurities, such as steam, CO2, CH4, O2 and light hydrocarbons of C2H4 and C2H6. In order to increase the production of SNG and optimize the H2/CO ratio of the produced gas, effects of these substances on the catalytic performance of the RDR reaction have to be investigated. Moreover, the yield of solid carbon should be taken into account during the thermodynamic analysis. The objective of this work is to elucidate, through a ther- modynamic study supported by experimental data, the effects of temperature, pressure and the other factors affecting the RDR reaction, such as the H2/CO molar ratio and addition of H2O, CH4, CO2, O2, and C2H4 in feed gas on the catalytic activity and selectivity and the yield. For this purpose, this study does not take into account of reaction kinetics, practical heat and mass transfer processes. It is expected to produce necessary thermochemical data to describe the effectiveness of the RDR reaction and to provide useful guidance to chemical engineers for optimizing the individual processes. 2.1 Thermodynamic analysis software The HSC Chemistry software 6.0 allows simulating chemi- cal reactions and processing on the thermochemical data 123 123 123 CO hydrogenation combined with water-gas-shift reaction for synthetic natural gas production… 441 GCðgÞ ¼ GCðsÞ ﬃDGfCðsÞ ¼ 0 ð5Þ X N1 i¼1 ni DGH fi þ RT ln yiuip pH þ X k kkaik ! þ nCDGH f CðsÞ ¼ 0 ð6Þ basis. In this study, the modules of reaction equations and equilibrium compositions were utilized to calculate the effects of various substances in conversion, selectivity and yield. The calculations were performed based on an exten- sive thermochemical database, which contains enthalpy (H), entropy (S) and heat capacity (Cp) data of more than 17000 chemical compounds (Roine 2010; Kumar et al. 2016). ð5Þ ð6Þ In the HSC Chemistry software 6.0, the reaction system needs to be speciﬁed, in terms of its phases and species, and the amount of the reactants. The program calculates the amount of products at equilibrium in isothermal or isobaric condition for a heterogeneous system. At the equilibrium state, the free energy of the system is minimized. 2.2 Thermodynamic analysis method The equilibrium products at different temperatures and pressures were calculated using the Gibbs free energy minimization method, which has been widely applied for thermodynamic calculations (Adhikari et al. 2007; Nahar and Madhani 2010; Lo´pez Ortiz et al. 2015). The detailed interpretation of this theory can be referenced by Wang et al. (Wang and Cao 2012; Wang et al. 2014). It should be noted that these thermodynamic analyses do not include any reaction kinetic limitation or transport process in the real process. However, thermochemical calculations show a great importance in adjusting the fea- sibility of a reactive process under certain conditions. Here, possible reactions are summarized in Table 1 for the cal- culations which were carried out based on different types of gases including CO, H2, CO2 and CH4, and the solid product of deposition carbon (graphite). Other substances such as alcohols, acids and high hydrocarbons are not taken into account due to their trace contents in the equilibrium The total Gibbs free energy is expressed as Eq. (1). As to chemical reaction equilibrium state, fi H = pH, fi = yiuip, and Gi H = DGﬁ H are supposed. With the Lagrange multiplier method, Eqs. (2) and (3) indicate the minimum Gibbs free energy of each gas and total system without solid ones, respectively. Equation (4) is the constraining condition. The vapor–solid phase equilibrium is applied to the Gibbs free energy of carbon, as shown in Eq. (5). Thus, the mini- mization formation of Gibbs free energy Eq. (6) is obtained by substituting Eq. (1) with Eqs. (3) and (5). p p q pp energy of carbon, as shown in Eq. (5). Thus, the mini- mization formation of Gibbs free energy Eq. (6) is obtained by substituting Eq. (1) with Eqs. (3) and (5). Gt ¼ X N i¼1 niGi ¼ X N i¼1 nili ¼ X N i¼1 niGH i þ RT X N i¼1 ni ln fi f H i ð1Þ DGH fi þ RT ln yiuip pH þ X k kkaik ¼ 0 ð2Þ X N i¼1 niðDGH fi þ RT ln yiuip pH þ X k kkaikÞ ¼ 0 ð3Þ X niaik ¼ Ak ð4Þ Fig. 2.2 Thermodynamic analysis method 1 Equilibrium constants of the reactions as a function of temperature y g q q Gt ¼ X N i¼1 niGi ¼ X N i¼1 nili ¼ X N i¼1 niGH i þ RT X N i¼1 ni ln fi f H i ð1Þ DGH fi þ RT ln yiuip pH þ X k kkaik ¼ 0 ð2Þ X N i¼1 niðDGH fi þ RT ln yiuip pH þ X k kkaikÞ ¼ 0 ð3Þ X niaik ¼ Ak ð4Þ Fig. 1 Equilibrium constants of the reactions as a function of temperature Gt ¼ X N i¼1 niGi ¼ X N i¼1 nili ¼ X N i¼1 niGH i þ RT X N i¼1 ni ln fi f H i ð1Þ DGH fi þ RT ln yiuip pH þ X k kkaik ¼ 0 ð2Þ X N i¼1 niðDGH fi þ RT ln yiuip pH þ X k kkaikÞ ¼ 0 ð3Þ X niaik ¼ Ak ð4Þ X niaik ¼ Ak ð4Þ Fig. 1 Equilibrium constants of the reactions as a function of temperature X niaik ¼ Ak ð4Þ Fig. 1 Equilibrium constants of the reactions as a function of temperature ð4Þ Table 1 The relevant reactions in the reverse dry reforming reaction Table 1 The relevant reactions in the reverse dry reforming reaction Reaction no. Reaction formula 4H298K (kJ/mol) Reaction type R1 2CO ? 2H2 $ CH4 ? CO2 -247.3 Reverse dry reforming reaction R2 CO ? 3H2 $ CH4 ? H2O -206.1 CO methanation R3 CO2 ? 4H2 $ CH4 ? 2H2O -165.0 CO2 methanation R4 CO ? H2O $ H2 ? CO2 -41.2 Water-gas shift R5 2CO $ C ? CO2 -172.4 Boudouard reaction R6 CH4 $ 2H2 ? C ?74.8 Methane cracking R7 CO ? H2 $ C ? H2O -131.3 CO reduction R8 CO2 ? 2H2 $ C ? 2H2O –90.1 CO2 reduction F. Meng et al. 442 c Fig. 3 Effect of temperature and pressure on catalytic performance. a CO conversion, b CH4 selectivity, c CH4 yield, and d carbon yield c Fig. 3 Effect of temperature and pressure on catalytic performance. a CO conversion, b CH4 selectivity, c CH4 yield, and d carbon yield gas mixture. The elemental mass balance is evaluated by carbon, hydrogen, and oxygen. 2.2 Thermodynamic analysis method The conversion of CO, selectivities of CH4 and CO2, yields of CH4 and solid carbon are calculated as follows: XCOð%Þ ¼ FCO;in FCO;out FCO;in 100 ð7Þ SCH4ð%Þ ¼ FCH4;out FCH4;out þ FCO2;out þ FC;out 100 ð8Þ SCO2ð%Þ ¼ FCO2; out FCH4; out þ FCO2; out þ FC; out 100 ð9Þ YCH4ð%Þ ¼ FCH4; out P i NiFi; in 100 ð10Þ Ycarbonð%Þ ¼ FC;out P i NiFi; in 100 ð11Þ ð9Þ ð10Þ ð11Þ Here, i indicates all carbon containing species (CO, CO2, CH4 and C2H4) at inlet, and Ni indicates the number of carbon atom of i-th species. 2.3 Experimental study The alumina (191 m2/g, Shandong Aluminum Co., China) supported homemade Ni-based catalyst was prepared by the co-impregnation method, as described in Meng’s works (Meng et al. 2017). The Ni-based catalyst, with the Ni loading of 20 wt% and La loading of 4 wt%, showed the speciﬁc surface area of 128 m2/g and pore size of 5.1 nm, and the catalyst was denoted as ExCat. To validate the thermodynamic calculations, the RDR reaction was carried out in a stainless steel, high-pressure ﬁxed-bed tube reactor (10 mm 9 2 mm 9 500 mm) within the temperature p Fig. 2 Equilibrium mole fraction of related substances in RDR reaction Fig. 2 Equilibrium mole fraction of related substances in RDR reaction Fig. 2 Equilibrium mole fraction of related substances in RDR reaction 12 3 CO hydrogenation combined with water-gas-shift reaction for synthetic natural gas production… atural gas production… 44 443 c Fig. 4 Effect of H2/CO molar ratio on catalytic performance. a CO conversion, b CH4 selectivity, c CH4 yield, and d carbon yield range of 300–550 C. 300 mg of Ni/Al2O3 catalyst (20–40 mesh) was placed in the reactor. Prior to the RDR reaction, the catalyst was reduced at 550 C in a H2 (99.99%, pur- chased from Taiyuan Iron & Steel (Group) Co., Ltd., China) ﬂow diluted with 25% N2 (99.995%, purchased from Taiyuan Iron & Steel (Group) Co., Ltd., China) for 6 h. A mixed feed gases of H2/CO = 1 (the gas of CO with a purity of 99.9% was purchased from Taiyuan Iron & Steel (Group) Co., Ltd., China) were introduced and con- trolled with the mass ﬂow controller (MFC), preheat treatment was ﬁnished at 200 C in ﬁrst oven at a space velocity of 20000 mL/(g h)-1. In the second oven, two thermocouples are employed for the reaction. One is placed closely to the reactor, in the middle of the oven to control the oven temperature. The other one is placed inside of the catalyst bed for the measurement of reaction temperature of catalyst bed. The outlet gas steam was cooled by condenser (2 C) and quantitatively analyzed by an online gas chro- matography (GC, Agilent 7890A) using helium (99.999%, purchased from Taiyuan Iron & Steel (Group) Co., Ltd., China) as the carrier gas. 2.3 Experimental study The GC equipped with a ﬂame ionization detector (FID) with an HP-AL/S column was employed to analyze CH4, and a thermal conductivity detector (TCD) equipped with a Porapak-Q column, HP- PLOT/Q column, and HP-MOLESIEVE column was employed to analyze CO2, CO, and N2. 123 3.2 Equilibrium compositions Figure 2 shows the methanation products with their mole fractions at equilibrium temperatures at 0.1 MPa. The feed gas contains H2 and CO with a H2/CO stoichiometric ratio of 1. The products mainly consist of CH4 and CO2 in the temperature range of 200–400 C. The mole fractions of CH4 and CO2 decrease as the temperature increases, whereas the mole fractions of H2 and CO exhibit the opposite trend. This can be explained by that the methane- generating reactions (R1–R3) are exothermic reactions and a higher temperature inhibits them. It is also found that the mole fraction of CH4 is lower than that of CO2. At a low H2/CO ratio, the Boudouard reaction (R5) is more prefer- able, leading to a large amount of CO2 and solid carbon. On the other side, the amount of H2 is higher than that of CO, as shown in Fig. 2. when the temperature over 550 C, the CO mole fraction increases linearly with temperature, which is due to the inhibition of the RDR reaction and the CO produced by the reverse Boudouard reaction (Nahar and Madhani 2010). The solid carbon is produced between 300 and 800 C, with a maximum amount around 575 C. Boudouard reaction will not occur when the temperature reaches 700 C or higher, as shown in Fig. 1, from which the solid carbon is mainly derived from the methane cracking reaction R6. Once the solid carbon produced, it may cover the active sites and results in catalyst deacti- vation (Takenaka et al. 2008). Thus to emphasize, a proper temperature below 600 C for RDR reaction is focused, in looking for a high yield of methane. 3.1 Equilibrium analysis of the reactions The equilibrium constants K of R1–R8 at various temper- atures are shown in Fig. 1. The value of K is calculated using the Van’t Hoff equation: d ln KH dT ¼ DrHH m RT2 ð12Þ ð12Þ It can be seen in Fig. 1, as the temperature increases, all the K values decrease except that of R6, which agrees with the Le Chatelier’s principle. R1, R2, R3, R5, and R7 play important roles in the RDR reaction system. When the temperature is lower than 500 C, the equilibrium constant K reduces in the order of R1 [ R2 [ R5 [ R3 [ R7 [ R8 [ R4 [ R6. Among all these reactions, R1 and R2 show elative high K values at low temperatures, which will lead to the high conversions of CO. CO2 could be con- verted via reactions of R3 and R8; however, the CO2 cannot be fully converted, which is due to that the reactions 12 3 F. Meng et al. 444 c Fig. 5 Effect of H2O added in the feed gas on catalytic performance. a CO conversion, b CH4 selectivity, c CO2 selectivity, and d carbon yield c Fig. 5 Effect of H2O added in the feed gas on catalytic performance. a CO conversion, b CH4 selectivity, c CO2 selectivity, and d carbon yield of R1, R4, and R5 generate CO2. Moreover, the solid carbon generated from the reaction of R5 to R8, and the Boudouard reaction (R5) acts a dominant role due to its largest K value. Importantly, all these reactions may occur simultaneously in the system, resulting in a balanced composition of the products. 3.3 Effect of temperature and pressure The effects of temperature and pressure on RDR reaction are shown in Fig. 3. Figure 3a shows that the CO can be fully converted when the temperature was below 400 C, further increasing the temperature resulted in the decrease of CO conversion, from the reason of the exothermic nature of RDR reaction. At a constant temperature, a higher pressure promotes the CO conversion due to the volume reducing nature of RDR reaction. The above results 123 123 CO hydrogenation combined with water-gas-shift reaction for synthetic natural gas production… 445 c Fig. 6 Effect of CH4 imputed into the feed gas on catalytic performance. a CO conversion, b CH4 selectivity, c CO2 selectivity, and d carbon yield indicate that a lower temperature and a higher pressure are favorable for the RDR reaction, in terms of increasing CO conversion. On the other hand, the increasing range of CO conversion is not obvious when the pressure is higher than 1 MPa. In Fig. 3b, high CH4 selectivity is obtained at low temperatures and high pressures. This is because all the methane producing reactions are volume reducing and exothermic. When the temperature is higher than 550 C, the K value of R5 is comparatively larger than that of R1 (Fig. 1), and the Boudouard reaction (R5) becomes domi- nantly in the reaction system, leading to a high CH4 selectivity at relative low temperatures. In Fig. 3c, a CH4 yield close to 50% is obtained in the temperature ranges of 200–300 C from 1 to 5 MPa. Thus, to get a comparatively high CO conversion and CH4 yield, the conditions of high pressure and low temperature are recommended. Knowing that low temperature is not beneﬁt to accelerate the reac- tion rate and high pressure is harmful for the equipment, a pressure range of 2–3 MPa and a temperature range of 300–500 C are favorable for the RDR reaction. The variation of carbon yield is presented in Fig. 3d. All these carbon yield curves exhibit a volcano characteristic, with less yield of carbon at high pressures. The solid car- bon results from many reactions, including R5, R6, R7, and R8 (as shown in Table 1), from which have different K values. Since the K value of R6 is negative at 200–550 C and the value of R5 is higher than that of R7 and R8 at 200–800 C (Fig. 3.3 Effect of temperature and pressure 1), so R5 is the main reason for the deposition of carbon. At the point of 0.1 MPa and around 575 C, the carbon yield reaches the maximum (23%). Accordingly, at this condition, the occurrence of R6 triggered a higher production of carbon. However, further increase the temperature results in the decrease of carbon yield, possibly because the reverse reactions of R5, R7 and R8 consumes a comparable amount of solid carbon. 3.4 Effect of H2/CO ratio Since the produced gas derived from coal gasiﬁcation has a varying ratio of H2/CO (Zheng and Furinsky 2005), there is a need to investigate the effect of H2/CO ratio on perfor- mance of RDR reaction. Figure 4a exhibits the effect of H2/CO molar ratio on CO conversion. The CO conversion changes slightly as the H2/CO ratio increases. When the pressure is 3 MPa, the CO conversion reaches nearly 100% in the temperature range of 200–550 C. Figure 4b shows the changes of CH4 selectivity. A high H2/CO ratio improves the selectivity of CH4, as it can be found when 12 3 F. Meng et al. 446 c Fig. 7 Effect of CO2 introduced into the feed gas on catalytic performance. a CO conversion, b CH4 selectivity, c CO2 selectivity, and d carbon yield the H2/CO ratio is 0.8, the maximum CH4 selectivity of 40% can be obtained. Considering the nearly 100% CO conversion below 450 C, there should be much carbon deposition (see Fig. 4d). When the H2/CO ratio increased to 1, the corresponding initial CH4 selectivity increases to 50%. Further increase H2/CO ratio to 3, the CH4 selectivity enhanced remarkably to 100%. In addition, high pressure is favorable to improve the CH4 selectivity. So a high H2/CO ratio or pressure value is useful for SNG production. Fig- ure 4c reveals the variations of CH4 yield. Higher H2/CO ratio or pressure and lower temperature lead to a higher CH4 yield. Figure 4d demonstrates the variation of carbon yield. When the H2/CO ratio is 0.8, a signiﬁcant amount of carbon is produced, because a large amount of unreacted CO is converted to solid carbon via R5. The highest carbon yield is 28 mol% at 0.1 MPa. To avoid such carbon deposition, high H2/CO ratio and high pressure are rec- ommended for the RDR reaction. c Fig. 7 Effect of CO2 introduced into the feed gas on catalytic performance. a CO conversion, b CH4 selectivity, c CO2 selectivity, and d carbon yield 3.5 Effect of H2O content Steam controls the H2/CO ratio via WGS reaction (R4), which is mostly used in methanation and ammonia syn- thesis industrial process. Moreover, it can be also used for eliminating the carbon deposition to some extent via reverse R7 and R8. The effect of steam content in feed gas is shown in Fig. 5. As can be found in Fig. 5a, the introduction of steam slightly decreases the CO conversion at 0.1 and 3 MPa. Although the steam does not participate in the RDR reaction, it can inhibit the methanation reaction R2 and thus decrease the CO conversion. As shown in Fig. 5b, steam has a small effect on the selectivity of CH4. How- ever, the selectivity of CO2 signiﬁcantly increases as the amount of steam increases at 0.1 MPa (Fig. 5c). From Fig. 5d, the additional steam decreases the yield of carbon, especially at high pressures. When the steam ratio reaches 0.4 at 3 MPa, only trace amount of carbon is formed at the temperature higher than 650 C. This is because the added steam promotes the reverse R8 reaction. This is also veri- ﬁed by Fig. 5c, where the addition of steam results in the increase of CO2 selectivity. In industrial methanation processes, such as High Combined Shift Methanation (HICOM) (Ensell and Stroud 1983) and Ralph M. Parsons (RMP) (G. A White et al. 1975), certain amount of steam is required to eliminate the solid carbon. 123 447 CO hydrogenation combined with water-gas-shift reaction for synthetic natural gas production… ural gas production… 4 1 3 c Fig. 8 Effect of O2 imported into the feed gas on catalytic performance. a CO conversion, b CH4 selectivity, c CO2 selectivity, and d carbon yield 3.6 Effect of CH4 content The syngas, derived from gasiﬁcation of coal, frequently contains a certain amount of CH4. In addition, the con- tained CH4 in methanation products is generally cycled to dilute the feed gas, in order to avoid reaction temperature run away from the ﬁxed-bed technology (Ro¨nsch et al. 2016). Thus, to study the effect of CH4 on the RDR reac- tion is of necessity. Figure 6a shows the effect of CH4 content on CO conversion. Additional CH4 results in slight decrease of CO conversion at both 0.1 and 3 MPa. The reason is that CH4 is the product of reactions R1, R2 and R3. Introduction of CH4 can inhibit these reactions (R1, R2 and R3), and thus decrease the CO conversion. In Fig. 6b and c, additional CH4 greatly improves the CH4 content in the product and reduces the CO2 selectivity at both 0.1 and 3 MPa, especially at low temperatures. However, the introduction of CH4 into the feed gas sharply exacerbates the deposition of solid carbon, especially at high temper- atures (Fig. 6d). Increasing pressure is very effective to reduce the carbon deposition. Comparing to 0.1 MPa, the carbon yield has been greatly reduced at 3 MPa. Interest- ingly, increasing the CH4 ratio from 1 to 2 increases slightly the carbon yield, when the temperature was lower than 600 C. Therefore, to enhance the CH4 yield and avoid the deposition of carbon, a certain content of CH4 in feed gas or product gas for recycling should not be ignored and the temperature should not exceed 600 C. 3.7 Effect of CO2 content Usually, the syngas needs to be puriﬁed to remove the acidic gas, such as CO2. In this section, the effect of CO2 on the RDR reaction is elucidated. Figure 7 shows the effect of CO2 content on the performance. In Fig. 7a, the increasing of CO2 ratio results in a slightly decrease of CO conversion, because the addition of CO2 inhibits the RDR reaction and probably accelerates the reaction rate of CO2 methanation (R3), based on the Le Chatelier’s principle. Figure 7b, c reveal the effect of CO2 on the selectivities of CH4 and CO2, respectively. As the CO2 amount increases, the selectivity of CH4 decreases gradually at 0.1 and 3 MPa. However, the selectivity of CO2 exhibits the opposite trend, especially when the temperature is higher than 600 C at 0.1 MPa, the selectivity of CO2 increases signiﬁcantly under the high CO2 amount. Meanwhile, the carbon yield (see Fig. 7d) drops gradually with an increasing of CO2 ratio, due to the inhibition of Boudouard 12 3 F. Meng et al. 448 c Fig. 9 Effect of C2H4 brought into the feed gas on catalytic performance. a CO conversion, b CH4 selectivity, c CO2 selectivity, and d carbon yield c Fig. 9 Effect of C2H4 brought into the feed gas on catalytic performance. a CO conversion, b CH4 selectivity, c CO2 selectivity, and d carbon yield reaction (R5). To decrease the carbon yield, the addition of CO2 could work at both pressures. In brief, high pressure inhibits carbon formation, however, it is better to remove the CO2 in the syngas to get a high methane yield. 3.8 Effect of O2 content In addition to steam, air and oxygen are also used as the feed gases for the coal gasiﬁcation to produce the syngas. Thus, the effect of O2 amount on the RDR reaction is necessary to be studied. The results are shown in Fig. 8. Figure 8a depicts the effect of O2 amount on CO conver- sion. The CO conversion exhibits almost no change at 200–500 C at a constant pressure. However, when the temperature is higher than 500 C, as the O2 amount increases, the CO conversion decreases slightly at 3.0 MPa. It is probably due to that a small amount of O2 reacts with the CH4 (partial oxidation of methane, 1/2O2 ? CH4 $ CO ? 2H2), which changes the CO conversion. Figure 8b shows the effect of O2 amount on CH4 selectivity. In the range of 200–500 C, as the O2 content rises, the selectivity of CH4 decreases especially at a high pressure. It is because that, when O2 is introduced into the reaction system, it reacts with CO to generate CO2, resulting in the decrease of CH4 selectivity. Another reason could be the reaction of partial oxidation of methane that consumes CH4, which also decreases the CH4 selectivity. From Fig. 8c, the CO2 selectivity increases as the O2 content rises, especially when the temperature is higher than 500 C at 0.1 MPa. It is due to that the CO and the produced solid carbon are oxygenated by O2, especially at high temperatures. Both of the possibilities result in the increase of CO2 selectivity. In Fig. 8d, the addition of O2 reduces the carbon yield to a large extent at 0.1 and 3 MPa, due to the fact that solid carbon reacts with the O2. It is noticeable that no carbon is formed at 3 MPa when the O2 ratio is 0.1 or 0.2. Therefore, to obtain a high CH4 yield, from the thermodynamic point of view, the feed gas should contain none of O2. 3.9 Effect of C2H4 content In the process of coal gasiﬁcation, some trace amounts of high hydrocarbons are produced, such as C2H6 and C2H4. Here, C2H4 is taken as a typical hydrocarbon to study the effect on performance. From Fig. 9a, b, C2H4 slightly affects the CO conversion and CH4 selectivity. This is mainly due to that, C2H4 is not involved in the reactions 123 123 CO hydrogenation combined with water-gas-shift reaction for synthetic natural gas production… 449 c Fig. 10 Comparison between experimental results and calculated ones at various temperatures and pressures. a CO conversion, b CH4 selectivity, c CO2 selectivity, d Carbon yield c Fig. 10 Comparison between experimental results and calculated ones at various temperatures and pressures. a CO conversion, b CH4 selectivity, c CO2 selectivity, d Carbon yield (R1, R2, R4 and R5), and in which CO is a reactant. This observation agrees with Gao’s results (Gao et al. 2012). However, as shown in Fig. 9c, the CO2 selectivity decreases as the C2H4 ratios increases. Figure 9d shows that the carbon yield rises rapidly with increasing the C2H4 ratio, especially at 0.1 MPa. The temperature correspond- ing to the maximum carbon yield is around 600 C. It can be concluded that at low pressure (0.1 MPa) and high temperatures (about 600 C), the reaction of C2H4 cracking (C2H4 $ C ? 2H2) should not be ignored. Whereas at high pressure (3 MPa), the carbon yield is remarkably reduced. Therefore, as an impurity, the C2H4 content must be controlled to prevent the formation of solid carbon. 123 4 Conclusions Gao J, Wang Y, Ping Y, Hu D, Xu G, Gu F, Su F (2012) A thermodynamic analysis of methanation reactions of carbon oxides for the production of synthetic natural gas. RSC Adv 2:2358–2368 A detailed thermodynamic equilibrium analysis of reverse dry reforming (RDR) reaction by minimizing the Gibbs free energy method in the range of 200–800 C and 0.1–3 MPa, and an experimental results in the range of 300–550 C and 0.1–3 MPa are studied. The calculation results demonstrate that low temperature and high pres- sure are beneﬁcial for the CO conversion and CH4 yield, and high H2/CO ratio (at least 1) promotes CH4 yield and decreases carbon yield. In the range of 200–500 C and 1–5 MPa, the CO conversion and CH4 yield reach 95%– 100% and 43%–50%, respectively. Steam in the feed gas enhances the CO2 selectivity and inhibits the generation of carbon, almost no carbon formed at the H2/CO/H2O ratio of 1/1/0.4, when the temperature is below 600 C at 3 MPa. CH4 contained in the recycling product gas ele- vates the CH4 content in the products, but also leads to more solid carbon at 500–800 C, especially at 0.1 MPa. CO2 has a negative effect on CH4 selectivity, but it could result in a slightly decrease of carbon yield at the tem- perature higher than 500 C. O2 is not preferable for increasing CH4 selectivity and decreasing the CO2 selectivity although it decreases the carbon yield. C2H4 is prone to crack, creating a high carbon yield. As impuri- ties, O2 and C2H4 should be completely removed to get a high CH4 yield. The experimental data are consistent with the calculation ones, indicating that minimizing the Gibbs free energy is effective to analyze the RDR reac- tion thermodynamically. This work is expected to provide a valuable suggestion in the process optimization for SNG production by combining CO methanation with WGS reaction. Gao J, Liu Q, Gu F, Liu B, Zhong Z, Su F (2015) Recent advances in methanation catalysts for the production of synthetic natural gas. RSC Adv 5:22759–22776 Gao Y, Meng F, Li X, Wen J, Li Z (2016) Factors controlling nanosized Ni-Al2O3 catalysts synthesized by solution combus- tion for slurry-phase CO methanation: the ratio of reducing valences to oxidizing valences in redox systems. Acknowledgements This work was supported by Youth Foundation of Shanxi Province (No. 2013021007-4) and National Basic Research Program of China (No. 2012CB723105). 3.10 Comparison between thermodynamic calculations and experimental results Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://crea tivecommons.org/licenses/by/4.0/), which permits unrestricted 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. References Adhikari S, Fernando S, Haryanto A (2007) A Comparative Thermodynamic and Experimental Analysis on Hydrogen Pro- duction by Steam Reforming of Glycerin. Energy Fuels 21:2306–2310 BP (2016) Statistical Review of World Energy. June Ensell R, Stroud H (1983) The British gas HICOM methanation process for SNG production. In: Proceedings of the international Ensell R, Stroud H (1983) The British gas HICOM methanation process for SNG production. In: Proceedings of the international gas research conference, UK: British Gas Corporation pp 472–481 gas research conference, UK: British Gas Corporation pp 472–481 3.10 Comparison between thermodynamic calculations and experimental results The RDR activity test was carried out on ExCat to compare the experimental results with thermodynamic calculations. Both the results at various temperatures and pressures are shown in Fig. 10. Figure 10a shows the comparisons of CO conversion. Thermodynamically, the equilibrium conver- sion of CO is almost 100% between 200 and 400 C, fur- ther increasing the temperature decreases the CO conversion, especially at a low pressure (0.1 MPa). Experimental results show that the CO conversion ﬁrstly increases and then decreases as the temperature rises and the highest CO conversion was obtained around 400 C. The high pressure is beneﬁt to improve the CO conversion. It should be noted that the experimental conversion of CO at 400 C or below is much lower than that of the calcu- lated values, which may be due to the low reaction rate at low temperatures. When the temperature is 400 C or higher, the experimental results show good accordance with the calculated ones. Figure 10b shows the CH4 selectivity at various tem- peratures and pressures. The calculated results shows that increasing temperature decreases the CH4 selectivity, whereas the increasing pressure enhances the CH4 selec- tivity. At low temperature and pressure, the experimental results are slightly higher than the calculation ones, while at high temperature, the experimental CH4 selectivity is much higher than the calculated one. Figure 10c shows the comparison of CO2 selectivity. The calculation results exhibit that the selectivity of CO2 is constant at various temperatures and pressures, and the experimental results 12 3 450 F. Meng et al. are slightly lower than the calculation ones. The above comparison show that more amount of CO2 converted to CH4 during the reaction. Figure 10d shows the comparison of carbon yields. The calculated carbon yields rose as the temperature increased, and decreased as the pressure increased, which means a large amount of CO is converted to carbon. However, the experimental results show that the yields of carbon were nearly zero, probably due to the catalyst inhibit the formation of carbon. The above dis- cussion show that the experimental results are generally in accordance with the calculated ones at different tempera- tures and pressures. The result also indicates that the Gibbs free energy minimization method is an ideal tool for thermodynamic analysis of the RDR process. 4 Conclusions Int J Hydrogen Energy 40:8833–8843 Shinde VM, Madras G (2014) CO methanation toward the production of synthetic natural gas over highly active Ni/TiO2 catalyst. AIChE J 60:1027–1035 Takenaka S, Orita Y, Umebayashi H, Matsune H, Kishida M (2008) High resistance to carbon deposition of silica-coated Ni catalysts Meng F, Li Z, Liu J, Cui X, Zheng H (2015b) Effect of promoter Ce on the structure and catalytic performance of Ni/Al2O3 catalyst for CO methanation in slurry-bed reactor. J Nat Gas Sci Eng 23:250–258 Takenaka S, Orita Y, Umebayashi H, Matsune H, Kishida M (2008) High resistance to carbon deposition of silica-coated Ni catalysts in propane stream reforming. Appl Catal A: Gen 351:189–194 High resistance to carbon deposition of silica-coated Ni catalysts in propane stream reforming. Appl Catal A: Gen 351:189–194 in propane stream reforming. Appl Catal A: Gen 351:189–194 Wang W, Cao Y (2012) Combined Carbon Dioxide Reforming with Steam Reforming of Ethanol for Hydrogen Production: thermo- dynamic Analysis. Int J Green Energy 9:503–516 Meng F, Li X, Li M, Cui X, Li Z (2017) Catalytic performance of CO methanation over La-promoted Ni/Al2O3 catalyst in a slurry-bed reactor. Chem Eng J 313:1548–1555 Wang B, Liu S, Hu Z, Li Z, Ma X (2014) Active phase of highly active Co3O4 catalyst for synthetic natural gas production. RSC Adv 4:57185–57191 Messerle VE, Ustimenko AB, Lavrichshev OA (2016) Comparative study of coal plasma gasiﬁcation: simulation and experiment. Fuel 164:172–179 Wang Y, Su Y, Zhu M, Kang L (2015) Mechanism of CO methanation on the Ni4/c-Al2O3 and Ni3Fe/c-Al2O3 catalysts: a density functional theory study. Int J Hydrogen Energy 40:8864–8876 Wang Y, Su Y, Zhu M, Kang L (2015) Mechanism of CO methanation on the Ni4/c-Al2O3 and Ni3Fe/c-Al2O3 catalysts: a Miguel CV, Soria MA, Mendes A, Madeira LM (2015) Direct CO2 hydrogenation to methane or methanol from post-combustion exhaust streams – A thermodynamic study. J Nat Gas Sci Eng 22:1–8 density functional theory study. Int J Hydrogen Energy 40:8864–8876 White GA, Roszkowski TR, Stanbridge DW (1975) The RMProcess. Methanation of Synthesis Gas, Advances in Chemistry, Amer- ican Chemical Society 146:138–148 Nahar GA, Madhani SS (2010) Thermodynamics of hydrogen production by the steam reforming of butanol: analysis of inorganic gases and light hydrocarbons. Int J Hydrogen Energy 35:98–109 Yan X, Liu Y, Zhao B, Wang Z, Wang Y, Liu C (2013) Methanation over Ni/SiO2: effect of the catalyst preparation methodologies. 4 Conclusions Catal Sci Technol 6:7800–7811 Go¨tz M, Lefebvre J, Mo¨rs F, McDaniel Koch A, Graf F, Bajohr S, Reimert R, Kolb T (2016) Renewable Power-to-Gas: a techno- logical and economic review. Renew Energy 85:1371–1390 Huo J, Yang D, Xia F, Tang H, Zhang W (2013) Feasibility analysis and policy recommendations for the development of the coal based SNG industry in Xinjiang. Energy Policy 61:3–11 Jiang M, Wang B, Lv J, Wang H, Li Z, Ma X, Qin S, Sun Q (2013) Effect of sulﬁdation temperature on the catalytic activity of MoO3/CeO2–Al2O3 toward sulfur-resistant methanation. Appl Catal A: Gen 466:224–232 Jiang M, Wang B, Yao Y, Wang H, Li Z, Ma X, Qin S, Sun Q (2014) Effect of stepwise sulﬁdation on a MoO3/CeO2–Al2O3 catalyst for sulfur-resistant methanation. Appl Catal A: Gen 469:89–97 Kopyscinski J, Schildhauer TJ, Biollaz SMA (2010) Production of synthetic natural gas (SNG) from coal and dry biomass—A technology review from 1950 to 2009. Fuel 89:1763–1783 Kumar N, Roy A, Wang Z, L’Abbate EM, Haynes D, Shekhawat D, Spivey JJ (2016) Bi-reforming of methane on Ni-based pyrochlore catalyst. Appl Catal A: Gen 517:211–216 Li H, Yang S, Zhang J, Kraslawski A, Qian Y (2014a) Analysis of rationality of coal-based synthetic natural gas (SNG) production in China. Energy Policy 71:180–188 Li S, Ji X, Zhang X, Gao L, Jin H (2014b) Coal to SNG: technical progress, modeling and system optimization through exergy analysis. Appl Energ 136:98–109 Lo´pez Ortiz A, Pallares Sa´mano RB, Mele´ndez Zaragoza MJ, Collins- Martı´nez V (2015) Thermodynamic analysis and process sim- ulation for the H2 production by dry reforming of ethanol with CaCO3. Int J Hydrogen Energy 40:17172–17179 Lu B, Ju Y, Kawamoto K (2014) Conversion of producer gas using NiO/SBA-15 obtained with different synthesis methods. Int J Coal Sci Technol 1:315–320 12 3 CO hydrogenation combined with water-gas-shift reaction for synthetic natural gas production… 451 Shen F, Liu J, Zhang Z, Yang Y (2016) Temporal measurements and kinetics of selenium release during coal combustion and gasiﬁcation in a ﬂuidized bed. J Hazard Mater 310:40–47 Martelli E, Kreutz T, Carbo M, Consonni S, Jansen D (2011) Shell coal IGCCS with carbon capture: conventional gas quench vs. innovative conﬁgurations. Appl Energ 88:3978–3989 Meng F, Li Z, Ji F, Li M (2015a) Effect of ZrO2 on catalyst structure and catalytic methanation performance over Ni-based catalyst in slurry-bed reactor. 4 Conclusions Int J Hydrogen Energy 38:2283–2291 Roine A (2010) Chemical reaction and equilibrium software with extensive thermo-chemical database. Outokumpu HSC 6.0, Chemistry for Windows Yu Z, Wang F (2010) The technology of Coal gasiﬁcation. Chemical industry Press, Beijing Ro¨nsch S, Schneider J, Matthischke S, Schlu¨ter M, Go¨tz M, Lefebvre J, Prabhakaran P, Bajohr S (2016) Review on methanation – From fundamentals to current projects. Fuel 166:276–296 Zheng L, Furinsky E (2005) Comparison of Shell, Texaco, BGL and KRW gasiﬁers as part of IGCC plant computer simulations. Energ Conver Manag 46:1767–1779 Sahebdelfar S, Takht Ravanchi M (2015) Carbon dioxide utilization for methane production: a thermodynamic analysis. J Petrol Sci Eng 134:14–22 12 123 12
W1989342638.txt	https://zenodo.org/records/2476335/files/article.pdf	de		Beitr�ge zur Kenntnis der Hydramide	Monatshefte für Chemie	1,906	public-domain	2,960	839 Beitr ge zur Kenntnis der Hydramide von Stud. phil. Artur Fiirth. Aus dem chemischen Laboratorium der k. k. deutschen Universititt in Prag. (Vorgelegt in der Sitzung am 12. Juli 1906.) In den letzten Jahren sind am hiesigen Institute mehrfach Versuche angestellt worden, welche die Verdriingung von Atomgruppen zum Gegenstande haben, die ein mit doppelter Bindung am Kohlenstoff hiingendes Stickstoffatom enthalten. So hat F u l d a 1 die Isonitrosogruppe durch den Phenylhydrazinrest verdriingt und such die umgekehrte Reaktion ausgefiihrt. Ferner hat O fn e r ~ Benzylidenanilin mit Benzylphenylhydrazin in Benzylidenbenzylphenylhydrazin iiberfiihrt. Schliel31ich hat Ott 8 diese letztere Umwandlung bei den Schiffschen Basen mit Hydrazinen, Semicarbazid und Hydroxylamin in die entsprechenden Hydrazone, Semicarbazone und Oxime beobachtet. Herr Prof. G o l d s c h m ~ e d t stellte mir nun die Aufgabe, die Verdriingung der mit Doppelbindung am Stickstoff h~ingenden Aldehydreste in den Hydramiden zu versuchen und so gemischte Hydramide vom Typus R-CH--N R.CH__ N ) CH'R1 darzustellen. 1 Monatsheffe fiir Chemie, 23, 907 (1902). 2 Ebenda, 25, 597 (1904). a Ebenda, 26, 335 (1905). 840 A. Fiir th, Bei diesen Versuchen ist auch das bisher nicht bekannte Hydrotolylamid dargestellt und bentitzt worden. Da G a t t e r m a n n in seiner, im letzten Heffe yon L i e b i g ' s Annalen (347, p. 347) erschienenen Abhandlung ~,Synthesen aromatischer Aldehyde<< diese Verbindung und auch das Trimethylamarin und Trimethyllophin beschreibt, so sehe ich mich veranlalgt, meine BeobachtungeL~, die fibrigens in Bezug auf die Verdr/ingung yon Aldehydresten in den Hydramiden fortgesetzt werden sollen, zu verSffentlichen. Vorgreifend kann ich mitteilen, daft mir die Darstellung gemischter Hydramide nicht gelungen ist, da die Verdr/ingung der Aldehydreste durch den zweiten Aldehyd stets weiter ging, als erwartet wurde, und a l l e Aldehydreste durch den verdr/ingenden Aldehyd ersetzt wurden. Das ffir die weiteren Untersuchungen notwendige Hydrobenzamid, welches nach der yon R o c h l e d e r l angegebenen Methode (Stehenlassen des Benzaldehyds mit w/isserigem Ammoniak) dargestellt wurde, zeigte trotz mehrmaligen Umkristallisierens aus Alkohol nicht den yon L a u r e n t ~ angegebenen Schmelzpunkt 1 10~ es verfltissigte sich schon bei 102 ~ Es wurden daher mit dem Hydrobenzamid folgende Reinigungsprozeduren vorgenommen: 1. Mehrmaliges Verreiben mit 5ther, 2. Verreiben mit Methylalkohol, 3. Umkristallisieren aus Ather, 4. LSsen in Chloroform in der K/iRe und F/illen mit Methylalkoho!. Nach jeder dieser Operationen wurde die Bestimmung des Schmelzpunktes vorgenommen und diese ergab durchwegs fibereinstimmend 102 ~ Ein v o n d e r Firma E. M e r c k bezogenes Hydrobenzamid hatte nach dem Umkristallisieren aus 2kther ebenfalls den Schmelzpunkt 102 ~ Annalen 41, 89 (1842). Annalen 21, 130 (1837). Die anderenAutoren, die sich mit dem Hydrobenzamid beschiiftigten, geben einen Schmelzpunktiiberhaupt ni c ht an. Zur genntnis der Hydramide. 841 Ftir die Reinheit meines Priiparates sprechen die Analysen : I. 0 " 3 2 1 4 g gaben 0 - 9 9 6 9 g CO s und 0" 1777g H~O. II. 0 " 2 1 0 0 g gaben 0"6496 g CO 2 und 0 9 1127g H20. In 100 Teilen: Gefunden Berechnet fiir I II C2~H~sN~ C ........ 84"6 84"4 84"6 H ........ 6" 1 60 6"0 Hydrotolylamid. W/ihrend G a t t e r m a n n dasselbe dutch LSsen yon p - T o l y l a I d e h y d in aikoholischer Ammoniak15sung erhS.lt, stellte ich es dutch Schtitteln des p - T o l y l aldehyds mit tiberschiissigem w~.sserigen Ammoniak dar. Die Ausbeute war nahezu quantitativ. Das erhaltene Produkt ist 15slich in Alkohol, Methylalkohol, Chloroform, Benzol und 2~ther. Aus Ather und Alkohol umkristallisiert, ergab es den Schmelzpunkt 91 ~ (bei G a t t e r m a n n 92~ Die Kristalle sind prismatisch. 0" 2036 g ergaben 0" 6282 ~ CO 2 und 0" 1220 g H20. in 100 Teilen: Oefunden Berechnet fiir C24H~4N~ 84" 2 6"7 84-7 7"1 v C ........ H ........ Die dem Amarin homologe Base, das T r i m e t h y l a m a r i n , w u r d e a u f dem yon B e r t a g n i n i 1 angegebenen, von B a h r m a n n ~ /:lbernommenen W e g e dargestellt HydrotolyIamid wurde im Schwefels/iurebade dutch mehrere Stunden auf 130 bis 140 ~ erhitzt. Die erhaltene braungelbe, glasartige Masse wurde in heiBem Alkohol gelSst und hei13 mit Salzs~ure neutralisiert. Nach dem Erkalten schied sich das salzsaure Trimethylamarin in weil3en Kristallen aus. Dasselbe wurde ab1 Jahresb. 1853, 471 und Ann, Chem. Pharm. 88, 127 (1853). Journ. f. prakt. Chem. 27, 296 (1883), Chemie-HeiI N r . 8, 59 842 a. Fiir th, gesaugt, wieder in heitlem Alkohol gelSst und die freie Base durch Ammoniak geftillt. Dieses T r i m e t h y l a m a r i n w u r d e mehrreals mit heil3em W a s s e r aus~ezogen, dann in sehr verdtinnter Essigs/iure gelSst, das UngelSste abfiltriert und im Filtrate die Base mit A m m o n i a k gef/illt. W e n n nicht stark gektihlt wurde, fiel dieselbe als dickes, gelbliches 01 aus, welches beim Stehen zu Klumpen erstarrte. Diese w u r d e n aus Ather und Alkohol umkristallisiert. Die Kristalle w a r e n sechsseitige S~iulen. Der Schme!zpunkt 127 ~, den G a t t e r ~ a n n angibt, konnte nicht beobachtet werden. H i n g e g e n sinterte das Trimethylamarin, in der Kapillare erhitzt, bei 119 bis 120 ~ zeigte aber, trotzdem sich bei 126 bis 1 2 7 ~ ein Meniskus bildete, fortgesetzt eine Trfibung, die erst bei 136 ~ vollst/i,ndig verschwand. Bei t27 ~ konnte man aufsteigende Gasbl~ischen wahrnehmen. Da sich dieser Vorgang beim Schmelzen stets, auch bei zwei nacheinander aus Alkohot fallenden Fraktionen zeigte, d a f e r n e r G a s e n t w i c k l u n g bei einer T e m p e r a t u r fiber 100 ~ stattfand, so lie13 mich dies vermuten, daft die Substanz ebenso, wie es D 6 1 6 p i n e I beim Amarin n a c h g e w i e s e n hat, Kristallwasser enthglt. In der T a t stimmen die erhaltenen Analysen auf C2~H,4N 2 4-1/~ H~O, welch letzteren Kristallwassergehalt auch D616p i n e beobachtet hatte. I. 0"2176 g gaben 0 " 6 5 6 0 g CO~ und 0" 1433 g H20. II. 0"2209 g gaben 0"6655 g CO 2 und 0" 1461 g H20. III. 0"2007 g gaben 14"9 c m s feuchten Stickstoff bei 23 ~ und 746 m m Hg. In 100 Teilen: Gefunden C ...... H ...... N ...... Berechnet fiir I II III 82'3 7"3 -- 82"2 7"3 -- -8'2 i Compt. rend., 1 2 5 , 179 (1897). C~ H~4N~ C2~H2~N2--I-1/2H20 84"7 7"1 8"2 82"5 7"2 8"0 843 Zur Kenntnis der Hydramide. Bei l~ingerem Erhitzen der Substanz auf 125 ~ schmolz dieselbe und nahm eine schwach gelbliche Ftirbung an: I. 0"3759 g verloren 0"0101 g. II. 0. 3498 g verloren 0" 0086 g. III. 0"5829 g verloren 0"0141 g. H~O . . . . . . I Gefunden " II III 2"7 2"5 2"4 Berezhnet ftir C~H24N2- ~ 1/~ HgO 2"6 Trimethyllophin. Hydrobenzamid wurde in einer Retorte geschmolzen. Bei weiterem stgtrkeren Erhitzen trat eine stiirmische Reaktion ein, es entwickelte sich Ammoniak und Toluol (am Geruche erkennbar) destillierte fiber. Nach dem Erkalten hinterbIieb eine dunkelgrCme, undurchsichtige Masse, die in viel _Ather 15slich war. Der )~ther wurde abdestilliert und der Rtickstand mit Alkohol erw/irmt. Bis auf eine geringe Menge gelbgrtiner Niidelchen, die den Schmelzpunkt zirka 250 ~ batten, 15ste sich alles. (HtSchstwahrscheinlich ist dies derselbe KSrper, den G a t t e r m a n n ebenfalls beobachtet hat und den er als ein polymeres Tolunitril anspricht.) Der in Alkohol 1/Ssliche Anteil wurde nach dem Abdunsten des Alkohols aus 5ther umkristallisiert. Es resultierten gelbe, verfilzte, seidengltinzende Nadeln vom Schmelzpunkt 235 ~ (bei G a t t e r m a n n 234~ Dieselben wurden in Benzol geltSst und die L6sung dem Sonnenlichte ausgesetzt. Ein Tell des Benzols wurde dann abdestilliert u n d e s schied sich das Trimethyilophin farblos ab. Der Schmelzpunkt war nach einmaligem Umkristallisieren aus _~ther wieder 235 ~ 0"2013 g gaben 0"6263 g CO~ und 0" 1177 g H~O. C ........ H ........ Oefunden Berechnet fiir C24H~2N ~ 84"9 6"5 85"2 6"5 59* 844 A. Ftirth, F-Trinitrohydrobenzamid. 5 g p-Nitrobenzaldehyd wurden mit konzentriertem wg.sserigen Ammoniak 1/ingere Zeit geschtittelt. Eine LSsung fand nicht start, nur nahm der kristallisierte K6rper eine flockige Beschaffenheit an. Derselbe wurde abgesaugt. Er ist schwer 16slich in Alkohol, Methylalkohol, nahezu unlaslich in Benzol und 5.ther, hingegen leicht 15slich in Chloroform. Kristallisiert konnte derselbe nieht erhalten werden, daher wurde er gereinigt durch L6sen in Chloroform und FNlen mit ~ther unter Kfihlung. In der Kapillare erhitzt, f/irbt sich der so erhaltene gelbe K6rper bei 160 bis 170 ~ dunkel, ohne bei weiterem Erhitzen zu schmelzen. Die Ausbeute an diesem so gereinigten K6rper ist gering. 0"2121g gaben 31 c m 3 N bei 22~ und 7 4 3 ~m Hg. In 100 Teilen: N ........ Gefunden Berechnet ftir C,21H15NsO6 16"2 16"2 Verdriingungsversuche. Hydrobenzamid und m-Nitrobenzaldehyd. 3 g Hydrobenzamid und 4 g rn-Nitrobenzaldehyd wurden, in Ather gel6st, gemischt. Die L6sung wurde eine Stunde am Rtickflul3ktihler erhitzt, sodann ein grolger Teil des L6sungsmittels abdestilliert und die konzentrierte L6sung stehen gelassen. Nach zwei Tagen schied sich ein kristallinischer, gelblichweil3er Niederschlag in der Menge von 2 " 5 g ab. Derselbe wurde mit viel Ather gekocht, um die anhaftenden Spuren von Aldehyd zu entfernen, dann abfiltriert und zuerst aus Alkohol, dann aus Benzol umkristallisiert. Es schieden sich weifie, diinne, mikroskopische Nadeln vom SchmeIzpunkt 160 ~ ab, die un16slich in Wasser und Ather, schwer in Alkohol, leichter in Benzol 16slich sin& I. 0"2009 g gaben 0"4293 g CO s und 0"0685 g H~O. If. 0"2045 g gaben 0"4335 g CO s und 0"0612 g H20. III. 0 " 2 0 0 2 g gaben 0"4259 g CO~ und 0"0656 g H20. Zur Kenntnis der Hydramide, 845 IV. 0" 2196 g g a b e n 31 "7 c m ~ feuchten Stickstoff bei 23 ~ und 740" 7 m m Hg. V. 0" 2057 g g a b e n 30" 2 c m ~ feuchten Stickstoff bei 23 ~ und 729"3 m m Hg. C ........ H ........ N ........ I II III IV V Mittei 58"2 3"7 -- 57"8 3"3 -- 58"0 3"7 -- --15"9 --15"9 58"0 3"5 15"9 Aus diesen A n a l y s e n geht hervor, daft die Reaktion nicht in der erwarteten W e i s e verlaufen ist, denn das erwartete gemischte H y d r a m i d NO~. CGH~ . CH ~-- N NO~ C~H 4. CH = N ~ C H . C~H 5 hat die Z u s a m m e n s e t z u n g : In 100 Teilen: 9 C~lH16N404, v C ........ H ........ N ........ O ........ 64"9 4"1 14"4 16'5 Die gleichzeitig durchgeffihrte Reaktio n yon H y d r o t o 1y 1a m i d und r lieferte die Aufkl/irung auch ffir den v o r s t e h e n d e n Fall. 3 g H y d r o t o t y l a m i d und 4 g N i t r o b e n z a l d e h y d w u r d e n in AtherlSsung einige Zeit am RfickfluBkfihler g e k o c h t und sodann die L 6 s u n g eingeengt. N a c h einiger Zeit kristallisierte ein KSrper in der Menge von 2"5 g aus, der, aus Benzol umkristallisiert, denselben S c h m e l z p u n k t 160 ~ zeigte, wie der im v o r h e r g e h e n d e n Versuche erhaltene K6rper. Ein Mischungss c h m e l z p u n k t beider S u b s t a n z e n ergab ebenso 160 ~ D a h e r sind beide KSrper miteinander identisch. Nach diesen T a t s a c h e n w a r a n z u n e h m e n , daft der m-Nitrob e n z a l d e h y d alle drei Aldehydreste der H y d r a m i d e verdriingt hat. In diesem Falle mfiBte nach der Gleichung: 846 A. Ffirth, (C6H 5 . CH)3 N, + 3 NO~. C6H ~ . CHO - - - (NO 2 9C6H ~ 9CH)3. N 2 + 3C6H 5 . CHO das m - T r i n i t r o h y d r o b e n z a m i d a u s m - N i t r o b e n z a l d e h y d und d e m die S trukturformel entstehen, das B e r t a g ' n i n i 1 A m m o n i a k erhalten hat und m--NO2--C6H~--CH -- N )CH ~--NO2--C6H~--CH -- N C~H~. m . NO~ z u k o m m t . Derselben entspricht die Z u s a m m e n s e t z u n g C~IH~5NsO~ C ........ H ........ N ........ 58:2 3"4 16"2 die mit den yon mir erhaltenen Resultaten tibereinstimmt. Ein S c h m e l z p u n k t ist in der Arbeit B e r t a g n i n i ' s nicht angegeben. Das s y n t h e t i s c h n a c h den A n g a b e n B e r t a g n i n i ' s yon mir dargestellte m - T r i n i t r o h y d r o b e n z a m i d hatte ebenfalls den S c h m e l z p u n k t 160 ~ desgleichen der M i s c h u n g s s c h m e l z p u n k t mit den durch Verdr~ingung erhaltenen PrS.paraten. D a h e r sind die beiden KSrper miteinander identisch. In der Mutterlauge des N i t r o h y d r a m i d s mul3te sich der verdr~ingte Aldehyd auffinden lassen. Bei einem diesbeztiglichen V e r s u c h (ausgeKihrt mit der Mutterlauge der Reaktion H y d r o tolylamid 4- ~ - N i t r o b e n z a l d e h y d ) konnte ich in der T a t nach Entfernen des 5 t h e r s durch Destillation bei der F r a k t i o n i e r u n g des Rtickstandes den T o l y l a t d e h y d finden, denn es g i n g die H a u p t m e n g e z w i s c h e n 202 ~ und 208 ~ fiber (Siedepunkt des p - T o l y l a l d e h y d s 204~ aul3erdem w a r der T o l y l a l d e h y d a m intensiven Geruch kenntlich. 2g Hydrosalicylamid 2 und 3 g e n - N i t r o b e n z a l d e h y d w u r d e n in konzentriert /itherischer L b s u n g stehen gelassen. N a c h einiger Zeit bildete sich ein Niederschlag, der aus Benzol umkristallisiert w u r d e und d a n n den S c h m e l z p u n k t 160 ~ zeigte. t Ann. 79, 272 (1851). e EttIing. Ann. 35", 261 (1840). Zur Kenntnis der Hydramide. 847 Ein Gemisch dieses KSrpers mit dem synthetischen m-Trinitrohydrobenzamid schmilzt ebenfalls bei 160 ~ daher sind die beiden Substanzen identisch. D e r s e l b e V e r s u c h m i t 2 g A n i s h y d r a m i d ~ und 2"5 g N i t r o b e n z a l d e h y d ergab ebenfalls ein mit dem m-Trinitrohydrobenzamid identisches Produkt. Nach diesen vier Versuchen ist wohl kein Zweifel, dab in alien in .~ther lgslichen Hydramiden die R. CH ---Gruppe durch m - - N O ~ - - C 6 H 4 - - C H = verdr/ingt werden kann. Mannigfach in Bezug auf relative Mengen der reagierenden Stoffe sowie in Bezug auf die Reaktionsdauer variierte Versuche, die darauf abzielten, >,gemischte Hydramide<< zu erhalten, blieben erfolglos. Desgleichen ist es mir nicht gelungen, die mit m-Nitrobenzaldehyd leicht erfolgende Verdr/ingung dutch andere Aldehyde zu bewirken, wobei in Anwendung kamen: Furfurol, Piperonal, Anisaldehyd und Salicylaldehyd mit Hydrobenzamid und Hydrotolylamid. Meinem hochverehrten Lehrer Herrn Professor Dr. Golds c h m i e d t spreche ich ft'lr die wertvollen Anregungen und ftlr die mir in reichlichstem Mal3e zu teil gewordene werkt~itige Hilfe meinen ergebensten und aufrichtigsten Dank aus. t Cahours, Ann., 56, 809 (1845).
https://openalex.org/W2223895690	https://repub.eur.nl/pub/87130/REPUB_87130_OA.pdf	English	null	Hyperoxia increases arterial oxygen pressure during exercise in type 2 diabetes patients: a feasibility study	European journal of medical research	2,016	cc-by	6,956	© 2016 Rozenberg et al. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted 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. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Abstract Objective: The study investigated the feasibility and potential outcome measures during acute hyperoxia in type 2 diabetes patients (DM2). Methods: Eleven DM2 patients (7 men and 4 women) were included in the study. The patients cycled (30 min at 20 % Wmax) whilst breathing three different supplemental oxygen flows (SOF, 5, 10, 15 L min−1). During hyperoxic exercise, arterial blood gases and intra-arterial blood pressure measurements were obtained. Results: Arterial pO2 levels increased significantly (ANOVA, p < 0.05) with SOF: 13.9 ± 1.2 (0 L min−1); 18.5 ± 1.5 (5 L min−1); 21.7 ± 1.7 (10 L min−1); 24.0 ± 2.3 (15 L min−1). Heart rate (HR) and pH increased significantly after termi- nating administration of hyperoxic air. Conclusions: An SOF of 15 L min−1 appears to be more effective than 5 or 10 L min−1. Moreover, HR, blood pressure, blood lactate and pH are not recommended as primary outcome measures. Keywords: Diabetes type 2, Hyperoxia, Exercise, Oxygen, Dose inversely related to blood glucose levels as well as dura- tion and severity of DM2 [15–17]. Pathophysiological mechanisms explaining the impaired pulmonary func- tion may be micro-angiopathy, chronic inflammation and autonomic neuropathy [16, 18] resulting in a diminished alveolar micro-vascular reserve [15, 17, 19–21]. Impaired alveolar gas exchange in DM2 patients has been shown to correlate with a lower ˙V O2 and workload capacity during aerobic type of exercise [22]. European Journal of Medical Research European Journal of Medical Research Rozenberg et al. Eur J Med Res (2016) 21:1 DOI 10.1186/s40001-015-0194-5 European Journal of Medical Research Hyperoxia increases arterial oxygen pressure during exercise in type 2 diabetes patients: a feasibility study Robert Rozenberg1, Robert T. Mankowski1, Luc J. C. van Loon2, Janneke G. Langendonk3, Eric J. G. Sijbrands3, Anton H. van den Meiracker3, Henk J. Stam1 and Stephan F. E. Praet1 Abstract Objective: The study investigated the feasibility and potential outcome measures during acute hyperoxia in type 2 diabetes patients (DM2). Methods: Eleven DM2 patients (7 men and 4 women) were included in the study. The patients cycled (30 min at 20 % Wmax) whilst breathing three different supplemental oxygen flows (SOF, 5, 10, 15 L min−1). During hyperoxic exercise, arterial blood gases and intra-arterial blood pressure measurements were obtained. Results: Arterial pO2 levels increased significantly (ANOVA, p < 0.05) with SOF: 13.9 ± 1.2 (0 L min−1); 18.5 ± 1.5 (5 L min−1); 21.7 ± 1.7 (10 L min−1); 24.0 ± 2.3 (15 L min−1). Heart rate (HR) and pH increased significantly after termi- nating administration of hyperoxic air. Conclusions: An SOF of 15 L min−1 appears to be more effective than 5 or 10 L min−1. Moreover, HR, blood pressure, blood lactate and pH are not recommended as primary outcome measures. Keywords: Diabetes type 2, Hyperoxia, Exercise, Oxygen, Dose Correspondence: r.mankowski@ufl.edu 1 Subdivision MOVEFIT‑Sports Medicine, Department of Rehabilitation Medicine, Erasmus University Medical Center, Wytemaweg 80, 3000 CA Rotterdam, The Netherlands Full list of author information is available at the end of the article Background Breathing a hyperoxic gas mixture has been shown to acutely enhance power output (W) by 8–13 % [1–6], increase oxygen uptake (˙V O2) by 10–14 % [2, 3, 6–10], decrease blood lactate level [11] and lower perceived exertion [7] during aerobic type of exercise. Both healthy subjects and COPD patients show improved exer- cise performance with hyperoxia [12–14]. These find- ings suggest that certain other clinical populations with impaired cardiovascular and/or pulmonary fitness levels might benefit from exercise under hyperoxic conditions as well. Patients with type 2 diabetes (DM2) might be good candidates for hyperoxic exercise training as pre- vious research indicated that DM2 patients frequently have a reduced diffusion capacity of the lungs (8–25 %), Although beneficial effects of exercise under hyper- oxic conditions have been reported for different types of chronic disease populations [1, 4, 8, 23–26], experi- mental data on an effective dose of hyperoxic air dur- ing exercise in DM2 patients are still lacking. Despite the ongoing debate on oxygen transport and consump- tion [50], increased oxygen availability in arterial blood may improve intracellular transport and uptake of active muscle tissue, and subsequently improve exercise perfor- mance. In accordance, the aim of the present feasibility study was to establish an effective dose of supplemental Correspondence: r.mankowski@ufl.edu 1 Subdivision MOVEFIT‑Sports Medicine, Department of Rehabilitation Medicine, Erasmus University Medical Center, Wytemaweg 80, 3000 CA Rotterdam, The Netherlands Full list of author information is available at the end of the article Page 2 of 8 Rozenberg et al. Eur J Med Res (2016) 21:1 oxygen in DM2 population as a basis to guide and opti- mise future hyperoxic exercise training protocols. subjects performed a hyperoxic exercise session. The bout consisted of 25 min of submaximal cycling at 20 % of the maximal workload capacity (Wmax). The workload was chosen to ensure that the subjects reach steady state, based on the assumption that the anaerobic threshold is at least 40 % Wmax in DM2 patients. After calibration patients underwent the Allen’s test, and subsequently beat-to-beat blood pressure was obtained through a per- cutaneous intra-articular catheter in the radial artery of the non-dominant hand [51]. Data were registered in a computer and analysed using specialised software (Beat scope, Finapres Medical Systems, Amsterdam, the Neth- erlands). It was performed to minimise the risk of ischae- mia of the hand. Background The exercise protocol consisted of 6 phases: 5 min of rest without supplemental oxygen flow (SOF), 10 min without SOF, 5 min with 5 L min−1 SOF, 5 min with 10 L min−1 SOF, 5 min with 15 L min−1 SOF, 5 min without SOF (Fig. 1). A stage duration of 5 min was chosen to reach steady state during, at least, the last 2 min of each stage [27, 28]. The last stage was added to assess the effect of cessation of SOF.h Methods Subjects j Eleven patients diagnosed with DM2 for at least 2 years and not taking anti-hypertensive medication were screened and included at the outpatient clinic at Erasmus University Medical Center in Rotterdam, the Nether- lands. The characteristics of the eligible patients are pre- sented in Table 1. Out of 22 screened patients, a total of six patients were not willing to participate in the hyper- oxic exercise intervention following the maximal exer- cise test. Four patients were excluded from the hyperoxic experiment because it was not possible to introduce an intra-arterial catheter in the radial artery. One patient was excluded from our study because of abnormally high lactate levels during exercise and was diagnosed with mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes (MELAS) syndrome afterwards. Baseline characteristics of excluded patients were not dif- ferent from the experimental group. Included subjects gave their informed consent to participate in the study, approved by the medical ethical committee of the Eras- mus University Medical Center in Rotterdam (ISRCTN number: NTR2299). f The SOF was administered directly into a face mask (without a reservoir bag, Teleflex Inc. Hudson RCI adult Multi-Vent air entrainment mask), allowing the inhalation of room air to meet the subjects’ ventilatory demands. In our study, we chose to dose oxygen as a fixed flow quantity, instead of a fixed inspirational frac- tion with a maximum of 15 L min−1. This design was chosen to match the possibilities of the standard facili- ties for supplemental oxygen available in most primary and secondary healthcare settings. During the last min of each phase arterial blood gas and the rate of perceived exertion (Borg score) [29] were obtained. Procedures Prior to the hyperoxic exercise session all subjects per- formed a maximal exercise test on a cycle ergometer (Jae- ger ER800) using an incremental workload (1.85 W/6 s for men, 1.2 W/6 s for women). The oxygen uptake (˙V O2 ) (Oxycon Pro, Viasys, Houten, Netherlands) and heart rate (HR) (Polar wear-link, Finland) were measured con- tinuously. A second visit was scheduled within 1–3 weeks following a maximal exercise test. During the second visit Results Participant The characteristics of the included participants are pre- sented in Table 1. Statistical analysis An independent sample T test was used to analyse the baseline characteristics and maximal exercise test. We used a single-factor ANOVA with repeated measures to compare the means of the given variables during differ- ent hyperoxic exercise phases. Differences with a p value <0.05 were considered significant. The Bonferroni adjust- ment was applied. Data were presented as mean ± SD. Table 1 Subject characteristics No significant differences (p < 0.05) a Based on skinfold measurements (Durnin and Womersley 1969) (n = 11) Mean ± SD Sex (M:F) 7:4 Age (years) 56.3 ± 6.3 T2D duration (years) 10.5 ± 6.6 Weight (kg) 87.7 ± 16.5 Length (cm) 171.1 ± 11.0 BMI (kg/m2) 30.1 ± 6.1 Abdominal circumference (cm) 100 ± 13 Fat percentagea (%) 33.9 ± 9.1 Fasting glucose (mmol/L) 11.3 ± 3.0 HbA1c (%) 8.3 ± 1.3 Table 1 Subject characteristics No significant differences (p < 0.05) a Based on skinfold measurements (Durnin and Womersley 1969) (n = 11) Mean ± SD Sex (M:F) 7:4 Age (years) 56.3 ± 6.3 T2D duration (years) 10.5 ± 6.6 Weight (kg) 87.7 ± 16.5 Length (cm) 171.1 ± 11.0 BMI (kg/m2) 30.1 ± 6.1 Abdominal circumference (cm) 100 ± 13 Fat percentagea (%) 33.9 ± 9.1 Fasting glucose (mmol/L) 11.3 ± 3.0 HbA1c (%) 8.3 ± 1.3 Table 1 Subject characteristics Technical feasibility From a technical perspective, our results demonstrate that supplemental oxygen, applied with a standard open face- mask (5–15 L min−1), results in significant increases in arterial pO2 levels during exercise. Higher pO2 at increased SOF (i.e. 5, 10 and 15 L min−1) suggests a dose-dependent effect. The pO2 levels obtained from the radial artery dur- ing hyperoxic exercise in the present study (24.0 ± 2.3 kPa) were comparable with the arterial pO2 levels measured by Plet et al. in healthy subjects. Administration of 55 % of oxy- gen improved maximal oxygen uptake by 12 % during cycle- ergometry in comparison with normoxic exercise [9]. Other studies investigating the influence of hyperoxia during exer- cise found slightly higher pO2 levels of approximately 40 kPa obtained from the femoral artery with an inspired oxygen fraction of 60 % [32–34]. Taken together, our data show that supplemental oxygen applied during submaximal exer- cise via a standard open face mask increases arterial pO2 levels. Additional oxygen availability could compensate for the diminished diffusion capacity, endothelial function and low aerobic capacity seen in most DM2 patients [15, 17, 19–21]. The latter suggests that a hyperoxic training study in DM2 patients could be a potential solution in a medical fitness centre, since no special equipment is needed other than an open facemask and standard gas cylinders with O2. However, before investigating training effects under hyper- oxic conditions, this warrants further controlled trials on cardiovascular and pulmonary function in DM2 patients. There was a decreasing trend of the pH (Fig. 2c) at the onset of the exercise bout (T1–T2), while pH increased significantly (p value) during the last step when the sup- plemental oxygen flow administration was stopped (T5–T6). Maximal exercise testh The average maximal oxygen uptake (VO2max) of the sub- jects was 1.83 ± 0.59 L min−1. The mean VO2max was on average ~24 % (p < 0.05) below the average of a healthy population of the same age, weight, length and sex based Rozenberg et al. Eur J Med Res (2016) 21:1 Page 3 of 8 Continuous intra-arterial blood pressure & heart rate measurement T0 ABG T5 ABG T2 ABG T3 ABG T4 ABG 5 10 15 0 Phase 1 Phase 2 Phase 3 Phase 4 Phase 5 Phase 6 Cycling at 20%Wmax T1 ABG SupplementalO2 (L) Fig. 1 Hyperoxic exercise protocol. ABG arterial blood gas, RPE rate of perceived exertion (Borg score) Discussion on the regression equations by Fairbarn et al. [30]. Maximal HR was not significantly different form the predicted val- ues according to the Tanaka regression equation [31]. The results of the maximal exercise test are presented in Table 2. We tested the feasibility of hyperoxic exercise and dose– response in type 2 diabetes patients. The main find- ing of this feasibility study was that exercise under SOF 15 L min−1 increased pO2 more effectively than lower doses (5 and 10 L min−1) in DM2 patients (Table 3). Hyperoxic exercise: blood gas analysis All included subjects were able to complete 25 min of submaximal hyperoxic exercise. Arterial pO2 levels (Fig. 2a) did not change immediately after starting the exercise (T1–2). The arterial pO2 increased significantly with increased SOF (T2–T5) in a dose-dependent man- ner and returned to baseline after cessation of the SOF administration (T5–T6). The arterial pCO2 (Fig. 2b) did not change significantly in response to exercise under dif- ferent SOF. Hyperoxic exercise: HR, blood pressure and rate of perceived exertionhi [31]) Significant difference (p < 0.05) HR (bpm) 155 ± 18 169 ± 5b 92 ± 10 128 ± 17 0.83 ± 0.08* RER 1.09 ± 0.09 0.92 ± 0.07 0.85 ± 0.07* Systolic blood pressure (mmHg) 180 ± 30 Diastolic blood pressure (mmHg) 79 ± 12 RPE (Borg 15.8 ± 2.8 0 10 20 30 T1 T2 T3 T4 T5 T6 * * * a pO2 (kPa) 4.0 4.5 5.0 5.5 6.0 T1 T2 T3 T4 T5 T6 b pCO2 (kPa) 0 10 20 30 T1 T2 T3 T4 T5 T6 * * * a pO2 (kPa) 4.0 4.5 5.0 5.5 6.0 T1 T2 T3 T4 T5 T6 b pCO2 (kPa) c d 7.34 7.36 7.38 7.40 7.42 7.44 T1 T2 T3 T4 T5 T6 * pH 0 1 2 3 4 T1 T2 T3 T4 T5 T6 Lactate (mmol/L) Fig. 2 Blood gas analysis a pO2 levels during hyperoxic exercise. T3–5 are significantly different from T1, T2 and T6, and each other (p < 0.05). b pCO2 levels during hyperoxic exercise. No significant changes (p < 0.05). c pH levels during hyperoxic exercise. T6 is significantly higher than T2 and T3 (p < 0.05). d Lactate levels during hyperoxic exercise. No significant changes (p < 0.05) Fig. 2 Blood gas analysis a pO2 levels during hyperoxic exercise. T3–5 are significantly different from T1, T2 and T6, and each other (p < 0.05). b pCO2 levels during hyperoxic exercise. No significant changes (p < 0.05). c pH levels during hyperoxic exercise. T6 is significantly higher than T2 and T3 (p < 0.05). d Lactate levels during hyperoxic exercise. No significant changes (p < 0.05) The results will improve our understanding on whether additional oxygen during exercise may improve oxidative metabolism in populations such as DM2 with deficient car- diovascular and respiratory function. exclusion criterion for the present study, the major- ity (=76 %) of the eligible DM2 patients that were approached in our outpatient clinic were willing to participate in our feasibility study. Although, a train- ing study requires a more long-term commitment, the willingness to participate in our feasibility study indi- cates that it might be possible to recruit a sufficient and representative proportion of subjects for a randomized Hyperoxic exercise: HR, blood pressure and rate of perceived exertionhi The HR (Fig. 3a) increased significantly after the onset of the exercise bout (T1–T2), and remained unchanged during the SOF phases (T2–T5) and increased after stop- ping the SOF (T5–T6). After an initial increase of systolic and diastolic blood pressure after starting the exercise (T1–2), systolic, diastolic and mean arterial blood pres- sure (Fig. 3b–d) did not change during the SOF. Fur- thermore, the rate pressure product (Fig. 3e) showed a significant increase after stopping the SOF (T5–T6). The rate of perceived exertion (Fig. 3f) increased with the start of exercise (T1–2) and did not increase significantly during exercise (T2–T6). Rozenberg et al. Eur J Med Res (2016) 21:1 Page 4 of 8 Table 2 Maximal exercise test AT anaerobic threshold using V-slope method a Using the Fairbarn and Wasserman equations (Fairbarn et al. [30]) b Using the Tanaka equation (Tanaka et al. [31]) * Significant difference (p < 0.05) VO2 (ml/min) 1830 ± 593 2499 ± 773 76 ± 21 1334 ± 354 0.75 ± 0.13 ˙VO2 (ml/min kg) 21.4 ± 7.0 28.9 ± 8.8 76 ± 24* 15.5 ± 4.0 0.75 ± 0.13* HR (bpm) 155 ± 18 169 ± 5b 92 ± 10 128 ± 17 0.83 ± 0.08* RER 1.09 ± 0.09 0.92 ± 0.07 0.85 ± 0.07* Systolic blood pressure (mmHg) 180 ± 30 Diastolic blood pressure (mmHg) 79 ± 12 RPE (Borg 15.8 ± 2.8 0 10 20 30 T1 T2 T3 T4 T5 T6 * * * a pO2 (kPa) 4.0 4.5 5.0 5.5 6.0 T1 T2 T3 T4 T5 T6 b c d pCO2 (kPa) 7.34 7.36 7.38 7.40 7.42 7.44 T1 T2 T3 T4 T5 T6 * pH 0 1 2 3 4 T1 T2 T3 T4 T5 T6 Lactate (mmol/L) Fig. 2 Blood gas analysis a pO2 levels during hyperoxic exercise. T3–5 are significantly different from T1, T2 and T6, and each other (p < 0.05). b pCO2 levels during hyperoxic exercise. No significant changes (p < 0.05). c pH levels during hyperoxic exercise. T6 is significantly higher than T2 and T3 (p < 0.05). d Lactate levels during hyperoxic exercise. No significant changes (p < 0.05) AT anaerobic threshold using V-slope method a Using the Fairbarn and Wasserman equations (Fairbarn et al. [30]) b Using the Tanaka equation (Tanaka et al. Patient recruitment and study population T1 is significantly lower T2–T6 (p < 0.05) 60 80 100 120 140 T1 T2 T3 T4 T5 T6 * a Heart rate (bpm) 70 80 90 100 110 T1 T2 T3 T4 T5 T6 c Diastolic blood pressure (mmHg) 70 80 90 100 110 T1 T2 T3 T4 T5 T6 c Diastolic blood pressure (mmHg) 60 80 100 120 140 T1 T2 T3 T4 T5 T6 * * a Heart rate (bpm) 120 160 200 240 T1 T2 T3 T4 T5 T6 * b Systolic blood pressure (mmHg) 80 100 120 140 160 T1 T2 T3 T4 T5 T6 d Mean arterial pressure (mmHg) 70 80 90 100 110 T1 T2 T3 T4 T5 T6 c Diastolic blood pressure (mmHg 120 160 200 240 T1 T2 T3 T4 T5 T6 * b Systolic blood pressure (mmHg) 80 100 120 140 160 T1 T2 T3 T4 T5 T6 d Mean arterial pressure (mmHg) 0 10000 20000 30000 T1 T2 T3 T4 T5 T6 * * e Rate pressure product 0 5 10 15 20 T1 T2 T3 T4 T5 T6 * f RPE (Borg score) Fig. 3 Cardiovascular response and rate of perceived exertion a HR during hyperoxic exercise. T1 is significantly lower than T2–6. T6 is significantly higher than T1, T3, and T4 (p < 0.05). b Systolic blood pressure during hyperoxic exercise. T1 is significantly lower than T2, T3 and T6 (p < 0.05). c Diastolic blood pressure during hyperoxic exercise. No significant changes (p < 0.05). d Mean arterial blood pressure during hyperoxic exercise. No significant changes (p < 0.05). e Rate pressure product during hyperoxic exercise. T1 is significantly lower than T2–6. T6 is significantly higher than T1 and T3 (p < 0.05). f Rate of perceived exertion (Borg score) during hyperoxic exercise. T1 is significantly lower T2–T6 (p < 0.05) 0 10000 20000 30000 T1 T2 T3 T4 T5 T6 * * e Rate pressure product 0 5 10 15 20 T1 T2 T3 T4 T5 T6 * f RPE (Borg score) Fig. 3 Cardiovascular response and rate of perceived exertion a HR during hyperoxic exercise. T1 is significantly lower than T2–6. T6 is significantly higher than T1, T3, and T4 (p < 0.05). b Systolic blood pressure during hyperoxic exercise. T1 is significantly lower than T2, T3 and T6 (p < 0.05). c Diastolic blood pressure during hyperoxic exercise. Patient recruitment and study population No significant changes (p < 0.05). d Mean arterial blood pressure during hyperoxic exercise. No significant changes (p < 0.05). e Rate pressure product during hyperoxic exercise. T1 is significantly lower than T2–6. T6 is significantly higher than T1 and T3 (p < 0.05). f Rate of perceived exertion (Borg score) during hyperoxic exercise. T1 is significantly lower T2–T6 (p < 0.05) clinical trial on the medium-term effects of hyperoxic exercise training. diffusion capacity [12–14], showed the anticipated increase in exercise capacity. Less perceived exercise intensity and improvement of exercise performance will also motivate overweight and with poorly regulated DM2 patients to adhere to hyperoxic exercise training. In accordance with previous studies [35–38], the mean VO2max of the investigated patient sample is well below the average of the healthy population, even when cor- rected for a high BMI. High HbA1c and fasting glu- cose levels showed that our overweight subjects had poorly regulated DM2. As such, the present study pop- ulation may not be representative for the general well- controlled DM2 population. Long-term adherence has been reported to vary substantially (10–80 %) in con- ventional exercise programs for DM2 patients [39–44]. However, effects of hyperoxic exercise training in other patient populations, with a reduced alveolar and capillary Patient recruitment and study population Despite the invasive nature of our study and the fact that use of antihypertensive medication was an Rozenberg et al. Eur J Med Res (2016) 21:1 Page 5 of 8 60 80 100 120 140 T1 T2 T3 T4 T5 T6 * * a Heart rate (bpm) 120 160 200 240 T1 T2 T3 T4 T5 T6 * b Systolic blood pressure (mmHg) 80 100 120 140 160 T1 T2 T3 T4 T5 T6 d Mean arterial pressure (mmHg) 0 10000 20000 30000 T1 T2 T3 T4 T5 T6 * * e Rate pressure product 70 80 90 100 110 T1 T2 T3 T4 T5 T6 c Diastolic blood pressure (mmHg) 0 5 10 15 20 T1 T2 T3 T4 T5 T6 * f RPE (Borg score) Fig. 3 Cardiovascular response and rate of perceived exertion a HR during hyperoxic exercise. T1 is significantly lower than T2–6. T6 is significantly higher than T1, T3, and T4 (p < 0.05). b Systolic blood pressure during hyperoxic exercise. T1 is significantly lower than T2, T3 and T6 (p < 0.05). c Diastolic blood pressure during hyperoxic exercise. No significant changes (p < 0.05). d Mean arterial blood pressure during hyperoxic exercise. No significant changes (p < 0.05). e Rate pressure product during hyperoxic exercise. T1 is significantly lower than T2–6. T6 is significantly higher than T1 and T3 (p < 0.05). f Rate of perceived exertion (Borg score) during hyperoxic exercise. Potential outcome measures In contrast with previous hyperoxic exercise studies [8, 9, 33, 45–47], we observed no change in HR, blood pres- sure or rate of perceived exertion during exercise while increasing the supplemental oxygen flow during exer- cise. However, after stopping administration of SOF we observed a significant increase in HR and rate pres- sure product [HR systolic blood pressure (SBP)]. The Rozenberg et al. Eur J Med Res (2016) 21:1 Page 6 of 8 Table 3 Hyperoxic exercise session a T3–5 are significantly different from T1, T2 and T6, and each other (p < 0.05) b T6 is significantly higher than T2 and T3 (p < 0.05) c T1 is significantly lower than T2–6. T6 is significantly higher than T1, T3, and T4 (p < 0.05) d T1 is significantly lower than T2, T3 and T6 (p < 0.05) e T1 is significantly lower than T2–6. Potential outcome measures T6 is significantly higher than T1 and T3 (p < 0.05) f T1 is significantly lower T2–T6 (p < 0.05) T1 T2 T3 T4 T5 T6 pO2 (kPa) 13.4 ± 1.1 13.9 ± 1.2 18.5 ± 1.5a 21.7 ± 1,7a 24.0 ± 2.3a 13.5 ± 1.5 Oxygen content (mmol/L) 8.3 ± 1.0 8.3 ± 0.9 8.3 ± 1.0 8.6 ± 1.2 8.4 ± 1.0 8.2 ± 1.1 pCO2 (kPa) 4.9 ± 0.5 5.1 ± 0.3 5.1 ± 0.4 5.1 ± 0.4 5.0 ± 0.5 4.9 ± 0.5 pH 7.40 ± 0.02 7.39 ± 0.02 7.40 ± 0.01 7.40 ± 0.02 7.40 ± 0.03 7.41 ± 0.02b Lactate (mmol/L) 1.96 ± 0.80 2.36 ± 0.94 2.21 ± 1.01 2.13 ± 1.09 2.00 ± 1.09 2.03 ± 1.10 Heart rate (bpm) 86 ± 11c 102 ± 13 102 ± 13 102 ± 12 106 ± 12 110 ± 13c Systolic blood pressure (mmHg) 159.4 ± 19.5d 192.6 ± 31.7 186.1 ± 30.0 186.0 ± 29.7 183.3 ± 33.5 187.1 ± 27.3 Diastolic blood pressure (mmHg) 85.6 ± 10.2 90.5 ± 11.0 88.9 ± 11.1 89.1 ± 9.9 89.0 ± 11.7 89.7 ± 10.1 Rate pressure product 13,626 ± 2483e 19,722 ± 4569 19,106 ± 4000 19,062 ± 4115 19,471 ± 5112 20,627 ± 4137e Mean arterial pressure (mmHg) 110.2 ± 12.4 124.5 ± 17.1 121.3 ± 16.6 121.4 ± 15.6 120.5 ± 18.2 122.2 ± 15.0 Rate of perceived exertion (/20) 6.0 ± 0.0f 9.8 ± 2.0 10.8 ± 2.6 11.6 ± 3.3 12.3 ± 4.3 13.1 ± 4.9 Table 3 Hyperoxic exercise session cardiovascular response during phase 6 indicates that hyperoxia lowers the cardiovascular burden during sub- maximal steady-state exercise in patients with DM2. A number of physiological mechanisms might explain why, in comparison with previous hyperoxic exercise studies, SOF did not lower HR and systemic blood pressure dur- ing phases 3–5 in our experimental setup. First, it is pos- sible that even at an exercise intensity of 20 % ˙V O2 max max, our patients were not completely in a steady-state condition during phases 3–5. Second, in comparison with previous hyperoxic exercise studies, the absolute exer- cise intensity may have been too low to cause a signifi- cant drop in HR, blood pressure or the rate of perceived exertion (Borg score). Authors’ contributions RR d SP d d h RR and SP designed the study protocol, obtained an approval from the medi- cal ethical committee, performed the experiments and took significant part in writing this manuscript. RM assisted during the experiments and significantly contributed in writing the manuscript. JL, ES and AM advised on the contents of the protocol and contributed inclusion of type 2 diabetes patients. HS and LL advised on the contents of the protocol and chosen methods and contrib- uted significantly by revising a manuscript. All authors read and approved the final manuscript. Potential outcome measures Third, the arterial wall stiffening in combination with the diabetes-related endothelial dys- function may have impaired a normal vascular response to hyperoxia [48, 49]. perceived exertion may not be suitable primary outcome measures for a hyperoxic training study in unfit DM2 patients. Instead, direct assessment of the VO2max should be considered in a hyperoxic training study to monitor and document change in exercise performance. Conclusions B d Based on arterial pO2 measurements, a supplemental oxygen flow of 15 L min−1 appears sufficient to compen- sate for impaired alveolar and capillary oxygen transport and/or consumption in DM2 patients. Based on this fea- sibility study, we propose to first investigate acute effects of various inspiratory oxygen fractions on the cardio-res- piratory system and speed of oxygen uptake kinetics. This will improve our understanding on potential exercise performance enhancement benefits of supplementary oxygen. This would warrant future studies to investigate the medium- and long-term benefits of hyperoxic exer- cise training in patients with DM2. 1 Subdivision MOVEFIT‑Sports Medicine, Department of Rehabilitation Medi- cine, Erasmus University Medical Center, Wytemaweg 80, 3000 CA Rotterdam, The Netherlands. 2 Department of Human Movement Sciences, NUTRIM School for Nutrition, Toxicology and Metabolism, Maastricht University Limitations of the study Unfortunately, for medical ethical reasons (invasive study) it was difficult to add a healthy control group or different oxygen conditions. Because of this limitation, we can only speculate about the physiological reason for this abnormal response to hyperoxic exercise. Arte- rial blood gas collection (arterial blood withdrawal) was vastly limited because of impaired structure of arterial walls in the DM2 patients. These invasive methods may be replaced by non-invasive study measures in the future studies such as bio-impedance cardiography and near- infrared spectroscopy. Nevertheless, the present feasi- bility study suggests that HR, blood pressure and rate of References y y 6. Wilber RL, Holm PL, Morris DM, et al. Effect of F(I)O(2) on physiological responses and cycling performance at moderate altitude. Med Sci Sports Exerc. 2003;35:1153–9. 31. Tanaka H, Monahan KD, Seals DR. Age-predicted maximal HR revisited. J Am Coll Cardiol. 2001;37:153–6. 7. Ekblom B, Huot R, Stein EM, et al. Effect of changes in arterial oxygen content on circulation and physical performance. J Appl Physiol. 1975;39:71–5. 32. Mourtzakis M, González-Alonso J, Graham TE, et al. Hemodynamics and O2 uptake during maximal knee extensor exercise in untrained and trained human quadriceps muscle: effects of hyperoxia. J Appl Physiol. 2004;97:1796–802. 8. Perry CG, Talanian JL, Heigenhauser GJ, et al. The effects of training in hyperoxia vs. normoxia on skeletal muscle enzyme activities and exercise performance. J Appl Physiol. 2007;102:1022–7. 33. Pedersen PK, Kiens B, Saltin B. Hyperoxia does not increase peak muscle oxygen uptake in small muscle group exercise. Acta Physiol Scand. 1999;166:309–18. 9. Plat J, Pedersen PK, Jensen FB, et al. Increased working capacity with hyperoxia in humans. Eur J Appl Physiol Occup Physiol. 1992;65:171–7. 10. Welch HG. Hyperoxia and human performance: a brief review. Med Sci Sports Exerc. 1982;14:253–62. 34. Stellingwerff T, Leblanc PJ, Hollidge MG, et al. Hyperoxia decreases muscle glycogenolysis, lactate production, and lactate efflux during steady-state exercise. Am J Physiol Endocrinol Metab. 2006;290:1180–90. 11. Maeda T, Yasukouchi A. Blood lactate disappearance during breathing hyperoxic gas after exercise in two different physical fitness groups–on the workload fixed at 130 % AT. Appl Human Sci. 1998;17:33–40. 35. Brun JF, Bordenave S, Mercier J, et al. Cost-sparing effect of twice-weekly targeted endurance training in type 2 diabetics: a one-year controlled randomized trial. Diabetes Metab. 2008;34:258–65. i 12. Emtner M, Porszasz J, Burns M, et al. Benefits of supplemental oxygen in exercise training in nonhypoxemic chronic obstructive pulmonary disease patients. Am J Respir Crit Care Med. 2003;168:1034–42. 36. Hordern MD, Cooney LM, Beller EM, et al. Determinants of changes in blood glucose response to short-term exercise training in patients with Type 2 diabetes. Clin Sci (Lond). 2008;115:273–81. 13. Palange P, Crimi E, Pellegrino R, et al. Supplemental oxygen and heliox: ‘new’ tools for exercise training in chronic obstructive pulmonary disease Curr Opin Pulm Med. 2005;11:145–8. y 37. Leite SA, Monk AM, Upham PA, et al. Low cardiorespiratory fitness in peo- ple at risk for type 2 diabetes: early marker for insulin resistance. Diabetol Metab Syndr. 2009;1:8. 14. Author details 1 1 Subdivision MOVEFIT‑Sports Medicine, Department of Rehabilitation Medi- cine, Erasmus University Medical Center, Wytemaweg 80, 3000 CA Rotterdam, The Netherlands. 2 Department of Human Movement Sciences, NUTRIM School for Nutrition, Toxicology and Metabolism, Maastricht University Page 7 of 8 Rozenberg et al. Eur J Med Res (2016) 21:1 Medical Centre, Universiteitssingel 50, 6229 ER Maastricht, The Netherlands. 3 Section of Pharmacology, Vascular and Metabolic Diseases, Department of Internal Medicine, Erasmus University Medical Center, ‘s‑Gravendijkwal 230, 3015 CE Rotterdam, The Netherlands. 18. Sandler M. Is the lung a ‘target organ’ in diabetes mellitus? Arch Intern Med. 1990;150:1385–8. 19. Klein OL, Smith LJ, Tipping M, et al. Reduced diffusion lung capacity in patients with type 2 diabetes mellitus predicts hospitalization for pneu- monia. Diabetes Res Clin Pract. 2011;92:12–5. Received: 18 June 2015 Accepted: 29 December 2015 Received: 18 June 2015 Accepted: 29 December 2015 23. Grataloup O, Prieur F, Busso T, et al. Effect of hyperoxia on maximal O2 uptake in exercise-induced arterial hypoxaemic subjects. Eur J Appl Physiol. 2005;94:641–5. y 24. Izdebska E, Izdebski J, Trzebski A. Hemodynamic responses to brief hyperoxia in healthy and in mild hypertensive human subjects in rest and during dynamic exercise. J Physiol Pharmacol. 1996;47:243–56. Acknowledgements 20. Ozşahin K, Tuğrul A, Mert S, et al. Evaluation of pulmonary alveolo-capil- lary permeability in Type 2 diabetes mellitus: using technetium 99mTc- DTPA aerosol scintigraphy and carbon monoxide diffusion capacity. J Diabetes Complications. 2006;20:205–9. This study was supported by the Dutch Research Foundation 2008.11.007. The authors would like to thank Ashraf Fagirzada and Mery Arevalo de Weever for their assistance with patient recruitment and experimental procedures. 21. Saler T, Cakmak G, Saglam ZA, et al. The assessment of pulmonary diffus- ing capacity in diabetes mellitus with regard to microalbuminuria. Intern Med. 2009;48:1939–43. References 1. Morris DM, Kearney JT, Burke ER. The effects of breathing supplemental oxygen during altitude training on cycling performance. J Sci Med Sport. 2000;3:165–75. 25. Karlsen T, Hoff J, Støylen A, et al. Aerobic interval training improves VO2 peak in coronary artery disease patients; no additional effect from hyper- oxia. Scand Cardiovasc J. 2008;42:303–9. 2. Peltonen JE, Ritola JJ, Ahotupa, et al. Oxygen uptake response during maximal cycling in hyperoxia, normoxia and hypoxia. Aviat Space Environ Med. 2001;72:904–11. 26. Stray-Gundersen J, Chapman RF, Levine BD. “Living high-training low” altitude training improves sea level performance in male and female elite runners. J Appl Physiol. 2001;91:1113–20. 3. Peltonen JE, Tikkanen HO, Rusko HK. Cardiorespiratory responses to exercise in acute hypoxia, hyperoxia and normoxia. Eur J Appl Physiol. 2001;85:82–8. 3. Peltonen JE, Tikkanen HO, Rusko HK. Cardiorespiratory responses to exercise in acute hypoxia, hyperoxia and normoxia. Eur J Appl Physiol. 2001;85:82–8. pp y 27. Baron B, Dekerle J, Robin S, et al. Maximal lactate steady state does not correspond to a complete physiological steady state. Int J Sports Med. 2003;24:582–7. 4. Ploutz-Snyder LL, Simoneau JA, Gilders RM, et al. Cardiorespiratory and metabolic adaptations to hyperoxic training. Eur J Appl Physiol Occup Physiol. 1996;73:38–48. 4. Ploutz-Snyder LL, Simoneau JA, Gilders RM, et al. Cardiorespiratory and metabolic adaptations to hyperoxic training. Eur J Appl Physiol Occup Physiol. 1996;73:38–48. 28. Macsween A. The reliability and validity of the Astrand nomogram and linear extrapolation for deriving VO2 max from submaximal exercise data. J Sports Med Phys Fitness. 2001;41:312–7. 5. Vanhatalo A, Fulford J, DiMenna FJ, et al. Influence of hyperoxia on muscle metabolic responses and the power-duration relationship during severe- intensity exercise in humans: a 31P magnetic resonance spectroscopy study. Exp Physiol. 2010;95:528–40. 5. Vanhatalo A, Fulford J, DiMenna FJ, et al. Influence of hyperoxia on muscle metabolic responses and the power-duration relationship during severe- intensity exercise in humans: a 31P magnetic resonance spectroscopy study. Exp Physiol. 2010;95:528–40. 29. Borg G, Linderholm H. Exercise performance and perceived exertion in patients with coronary insufficiency, arterial hypertension and vasoregu- latory asthenia. Acta Med Scand. 1970;187:17–26. 30. Fairbarn MS, Blackie SP, McElvaney NG, et al. Prediction of HR and oxygen uptake during incremental and maximal exercise in healthy adults. Chest. 1994;105:1365–9. y y 6. Wilber RL, Holm PL, Morris DM, et al. Effect of F(I)O(2) on physiological responses and cycling performance at moderate altitude. Med Sci Sports Exerc. 2003;35:1153–9. Competing interests p g The authors declare that they have no competing interests. The authors declare that they have no competing interests. 22. Cooper BG, Taylor R, Alberti KG, et al. Lung function in patients with diabetes mellitus. Respir Med. 1990;84:235–9. Received: 18 June 2015 Accepted: 29 December 2015 Received: 18 June 2015 Accepted: 29 December 2015 41. Kirk A, Mutrie N, MacIntyre P, et al. Effects of a 12-month physical activity counselling intervention on glycaemic control and on the status of cardiovascular risk factors in people with Type 2 diabetes. Diabetologia. 2004;47:821–32. 42. Praet SF, van Rooij ES, Wijtvliet A, et al. Brisk walking compared with an individualised medical fitness programme for patients with type 2 diabe- tes: a randomised controlled trial. Diabetologia. 2008;51:736–46. 43. Schneider SH, Khachadurian AK, Amorosa LF, et al. Ten-year experience with an exercise-based outpatient life-style modification program in the treatment of diabetes mellitus. Diabetes Care. 1992;15:1800–10. 44. Shinji S, Shigeru M, Ryusei U, et al. Adherence to a home-based exercise program and incidence of cardiovascular disease in type 2 diabetes patients. Int J Sports Med. 2007;28:877–9. 45. Byrnes WC, Mihevic PM, Freedson PS, et al. Submaximal exercise quanti- fied as percent of normoxic and hyperoxic maximum oxygen uptakes. Med Sci Sports Exerc. 1984;16:572–7. 46. Graham TE, Wilson BA. Effects of hypercapnia and hyperoxia on metabo- lism during exercise. Med Sci Sports Exerc. 1983;15:514–9. References van Helvoort HA, Heijdra YF, Heunks LM, et al. Supplemental oxygen prevents exercise-induced oxidative stress in muscle-wasted patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2006;173:1122–9. 38. Lindegaard B, Hansen T, Hvid T, et al. The effect of strength and endur- ance training on insulin sensitivity and fat distribution in human immu- nodeficiency virus-infected patients with lipodystrophy. J Clin Endocrinol Metab. 2008;93:3860–9. 15. Chance WW, Rhee C, Yilmaz C, et al. Diminished alveolar microvascular reserves in type 2 diabetes reflect systemic microangiopathy. Diabetes Care. 2008;31:1596–601. 39. Balducci S, Leonetti F, Di Mario U, et al. Is a long-term aerobic plus resist- ance training program feasible for and effective on metabolic profiles in type 2 diabetic patients? Diabetes Care. 2004;27:841–2. 16. Hsia CC, Raskin P. Lung function changes related to diabetes mellitus. Diabetes Technol Ther. 2007;9:73–82. 17. Klein OL, Krishnan JA, Glick S, et al. Systematic review of the associa- tion between lung function and Type 2 diabetes mellitus. Diabet Med. 2010;27:977–87. 40. Dunstan DW, Vulikh E, Owen N, et al. Community center-based resistance training for the maintenance of glycemic control in adults with type 2 diabetes. Diabetes Care. 2006;29:2586–91. Rozenberg et al. Eur J Med Res (2016) 21:1 Page 8 of 8 41. Kirk A, Mutrie N, MacIntyre P, et al. Effects of a 12-month physical activity counselling intervention on glycaemic control and on the status of cardiovascular risk factors in people with Type 2 diabetes. Diabetologia. 2004;47:821–32. 47. Rousseau A, Bak Z, Janerot-Sjöberg B, et al. Acute hyperoxaemia-induced effects on regional blood flow, oxygen consumption and central circula- tion in man. Acta Physiol Scand. 2005;183:231–40. y 48. Jansson PA. Endothelial dysfunction in insulin resistance and type 2 diabetes. J Intern Med. 2007;262:173–83. 42. Praet SF, van Rooij ES, Wijtvliet A, et al. Brisk walking compared with an individualised medical fitness programme for patients with type 2 diabe- tes: a randomised controlled trial. Diabetologia. 2008;51:736–46. 49. Tabit CE, Chung WB, Hamburg NM, et al. Endothelial dysfunction in diabetes mellitus: molecular mechanisms and clinical implications. Rev Endocr Metab Disord. 2010;11:61–74. 43. Schneider SH, Khachadurian AK, Amorosa LF, et al. Ten-year experience with an exercise-based outpatient life-style modification program in the treatment of diabetes mellitus. Diabetes Care. 1992;15:1800–10. 50. Crecelius AR, Kirby BS, Dinenno FA. Intravascular ATP and the regula- tion of blood flow and oxygen delivery in humans. Exerc Sport Sci Rev. 2015;43(1):5–13. 44. References Shinji S, Shigeru M, Ryusei U, et al. Adherence to a home-based exercise program and incidence of cardiovascular disease in type 2 diabetes patients. Int J Sports Med. 2007;28:877–9. 51. Limberg JK, Kellawan JM, Harrell JW, Johansson RE, Eldridge MW, Proctor LT, Sebranek JJ, Schrage WG. Exercise-mediated vasodilation in human obesity and metabolic syndrome: effect of acute ascorbic acid infusion. Am J Physiol Heart Circ Physiol. 2014;307(6):H840–7. 45. Byrnes WC, Mihevic PM, Freedson PS, et al. Submaximal exercise quanti- fied as percent of normoxic and hyperoxic maximum oxygen uptakes. Med Sci Sports Exerc. 1984;16:572–7. 46. Graham TE, Wilson BA. Effects of hypercapnia and hyperoxia on metabo- lism during exercise. Med Sci Sports Exerc. 1983;15:514–9. • We accept pre-submission inquiries • Our selector tool helps you to ﬁnd the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and we will help you at every step: 51. Limberg JK, Kellawan JM, Harrell JW, Johansson RE, Eldridge MW, Proctor LT, Sebranek JJ, Schrage WG. Exercise-mediated vasodilation in human obesity and metabolic syndrome: effect of acute ascorbic acid infusion. Am J Physiol Heart Circ Physiol. 2014;307(6):H840–7. 41. Kirk A, Mutrie N, MacIntyre P, et al. Effects of a 12-month physical activity counselling intervention on glycaemic control and on the status of cardiovascular risk factors in people with Type 2 diabetes. Diabetologia. 2004;47:821–32. 49. Tabit CE, Chung WB, Hamburg NM, et al. Endothelial dysfunction in diabetes mellitus: molecular mechanisms and clinical implications. Rev Endocr Metab Disord. 2010;11:61–74. 50. Crecelius AR, Kirby BS, Dinenno FA. Intravascular ATP and the regula- tion of blood flow and oxygen delivery in humans. Exerc Sport Sci Rev. 2015;43(1):5–13. 45. Byrnes WC, Mihevic PM, Freedson PS, et al. Submaximal exercise quanti- fied as percent of normoxic and hyperoxic maximum oxygen uptakes. Med Sci Sports Exerc. 1984;16:572–7. 48. Jansson PA. Endothelial dysfunction in insulin resistance and type 2 diabetes. J Intern Med. 2007;262:173–83. 47. Rousseau A, Bak Z, Janerot-Sjöberg B, et al. Acute hyperoxaemia-induced effects on regional blood flow, oxygen consumption and central circula- tion in man. Acta Physiol Scand. 2005;183:231–40. 48. Jansson PA. Endothelial dysfunction in insulin resistance and type 2 diabetes. J Intern Med. 2007;262:173–83. 49. Tabit CE, Chung WB, Hamburg NM, et al. Endothelial dysfunction in diabetes mellitus: molecular mechanisms and clinical implications. Rev Endocr Metab Disord. 2010;11:61–74. 50. Crecelius AR, Kirby BS, Dinenno FA. Intravascular ATP and the regula- tion of blood flow and oxygen delivery in humans. Exerc Sport Sci Rev. 2015;43(1):5–13. 51. Limberg JK, Kellawan JM, Harrell JW, Johansson RE, Eldridge MW, Proctor LT, Sebranek JJ, Schrage WG. Exercise-mediated vasodilation in human obesity and metabolic syndrome: effect of acute ascorbic acid infusion. Am J Physiol Heart Circ Physiol. 2014;307(6):H840–7.
https://openalex.org/W2894278838	https://eprints.kingston.ac.uk/id/eprint/43694/1/Hathiramani-S-43694-VoR.pdf	English	null	Relaxation and Exercise in Lymphoma Survivors (REIL Study): A Randomised Clinical Trial Protocol	Journal of global oncology	2,018	cc-by	6,817	Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:17 https://doi.org/10.1186/s13102-019-0127-7 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:17 https://doi.org/10.1186/s13102-019-0127-7 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation https://doi.org/10.1186/s13102-019-0127-7 Open Access © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted 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. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Abstract Background: Lymphoma survivors commonly report ongoing complaints including fatigue, pain, depression and decreased quality of life (QoL) following treatment. Although evidence suggests that both relaxation and exercise can significantly improve such symptoms, there is no consensus on which intervention is more effective. This paper presents the REIL (Relaxation and Exercise In Lymphoma) Study protocol. The REIL study aims to compare the effect of two home-based interventions – relaxation and exercise – on QoL in lymphoma survivors. Methods: Eligible participants (n = 36) will be randomised to a relaxation or exercise programme to perform at least three times per week. The primary outcome measure is QoL, assessed by the European Organisation for Research and Treatment of Cancer QoL Questionnaire Core 30 (EORTC QLQ-C30). Secondary outcome measures include body composition, cardiovascular status, pulmonary function, grip strength, functional exercise capacity (six minute walk test), well-being assessed by the FACT-Lym questionnaire, and psychological status assessed by the Hospital Anxiety and Depression Scale. Total duration of the study will be twelve weeks and outcome measures will be assessed at baseline, six weeks and at the end of the study. Discussion: It is anticipated that results from this preliminary study will begin to highlight effective pathways to improve QoL following chemotherapy for this population. This will better inform healthcare professionals to optimise QoL of lymphoma patients, and enable a smooth transition from being a cancer patient to survivor. Trial registration: The REIL study has been registered on a publicly accessible database, ClinicalTrials.gov, Registration Number: NCT02272751, October 2014. Trial registration: The REIL study has been registered on a publicly accessible database, ClinicalTrials.gov, Registration Number: NCT02272751, October 2014. Keywords: Lymphoma survivors, Exercise, Relaxation, Mindfulness, Rehabilitation, Cancer survivors, Self-management, Quality of life Keywords: Lymphoma survivors, Exercise, Relaxation, Mindfulness, Rehabilitation, Cancer survivors, Self-management, Quality of life Keywords: Lymphoma survivors, Exercise, Relaxation, Mindfulness, Rehabilitation, Cancer survivors, Self-management, Quality of life Relaxation and exercise in lymphoma survivors (REIL study): a randomised clinical trial protocol Suchita Hathiramani1,4* , Ruth Pettengell2, Hannah Moir3 and Ahmed Younis1,4 Background treatment exposures. Commonly reported long-term and late effects of treatment in lymphoma survivors include both physical and psychosocial symptoms such as fa- tigue, pain, muscle weakness, neuropathies, depression, anxiety and decreased self-esteem; as well as decreased function and quality of life (QoL) [1–3]. Hence there has been a call for further research to address the ongoing needs of lymphoma survivors [1]. Significant improvements have been made in the survival rates in cancer patients due to earlier detection and ad- vances in treatment. Although initial research in the field of cancer survivorship focused on tumours such as breast and prostate, dramatic advances in haematological malignancies have also led to increased survival [1], but at a cost of increasing complications. Like survivors of other cancers, lymphoma survivors are at an increased risk of morbidity and adverse effects due to disease and When medical treatment is completed successfully and the patient is in clinical remission, the close support from medical staff is usually reduced and the patient is expected to return to ‘normal life’. This transition how- ever does not always occur smoothly and studies have highlighted that post-treatment survivors at this * Correspondence: s.hathiramani@sgul.kingston.ac.uk 1Kingston University and St. George’s University of London, London, UK 4Faculty of Health, Social Care & Education, Cranmer Terrace, London SW17 0RE, UK Full list of author information is available at the end of the article * Correspondence: s.hathiramani@sgul.kingston.ac.uk 1Kingston University and St. George’s University of London, London, UK 4Faculty of Health, Social Care & Education, Cranmer Terrace, London SW17 0RE, UK Full list of author information is available at the end of the article Page 2 of 9 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:17 Page 2 of 9 has been a call for further studies to move beyond wait- list control groups and to compare with active control or other empirically supported interventions [15, 16], for instance relaxation to exercise. As a result, there have been studies and study protocols looking at comparison of two interventions in cancer survivors, but the majority of these have focused on breast cancer [17, 18] and au- thors have called for future research to focus on other survivor groups, including haematologic cancer sites [1, 19]. transition phase often feel unprepared and uninformed [1, 3]. Survivors at this phase are sometimes in need of further support and advice on how to progress activity, manage ongoing symptoms and return to pre-morbid status and often do not receive this [2, 3]. There are cur- rently no recommended care pathways for lymphoma patients in remission following treatment, despite the fact that this transition period shortly after treatment has been highlighted to be particularly difficult [4]. This lack of care pathways has been highlighted by the government in the United Kingdom (UK). The National Cancer Survivorship Initiative (NCSI) was launched in 2008 and calls for research into the development of pathways for cancer survivors [5], recommending holis- tic assessment and personalised care-planning, emphasis on the use of patient-reported outcome measures and a shift towards self-management [6]. The REIL study aims to address some of these issues highlighted by studying a sample of lymphoma survivors post-chemotherapy, supporting patients during the tran- sition phase, comparison of two interventions, use of pa- tient-reported outcome measures and emphasis on self- management. Results from this preliminary study would provide an indication of efficacy of interventions, and with the findings from the REIL study we aim to build towards the development of evidence-based practice guidelines for lymphoma survivors. Self-management in cancer survivorship has been de- fined as ‘awareness and active participation by the per- son in their recovery, recuperation and rehabilitation, to minimise the consequences of treatment, and promote survival, health and well-being” [7]. Increasing numbers of survivors together with limited healthcare resources has led to a focus on self-management as an important approach for cancer survivors. Furthermore, self-man- agement can empower cancer patients, increase their confidence to manage problems associated with disease and treatment, and enhance QoL [8]. This paper presents the REIL study protocol using SPIRIT 2013 guidelines [20]. Study design The REIL study is a prospective, randomised, clinical intervention trial. Participants will be randomised to ex- ercise or relaxation intervention. Participants will be assessed at baseline prior to commencing the interven- tion programme, at six weeks, and at the end of the twelve week intervention. Aims The primary aim of the REIL Study is to compare the ef- fect of two interventions – relaxation and exercise – on QoL in a sample of lymphoma patients in remission post-chemotherapy. The null hypothesis is that there is no difference in QoL between the relaxation and exer- cise groups post-intervention. Literature indicates that interventions such as relax- ation and exercise both have a positive impact on QoL of cancer survivors [9–12]. It has been demonstrated that exercise interventions have a positive impact on physical and psychological factors related to QoL such as peak oxygen consumption, physical functioning, fa- tigue, self-esteem and social functioning; as well as over- all QoL of cancer patients and survivors [9, 10]. Studies on relaxation interventions including mindfulness-based stress reduction and progressive muscular relaxation have also demonstrated significant positive impact on physical and psychological symptoms such as fatigue, pain, anxiety and depression, as well as on QoL in can- cer survivors [10–12]. Hence both relaxation and exer- cise interventions are recommended by organisations providing expert advice and support to cancer patients and survivors such as Lymphoma Action [13] and Mac- millan Cancer Support [14]. Secondary aims are to investigate the effects of the two interventions on body composition, cardiovascular status, pulmonary function, muscle strength, functional exercise capacity, well-being and psychological status; and explore perceptions about participation in the post- treatment intervention programme. Ethical approval However in the majority of trials studying the effects of relaxation or exercise on cancer survivors, the benefits have only been relative to a control group; and to rule out potential placebo effects future studies need to spe- cify more rigorous comparison conditions, for instance a control intervention with similar elements that may in- fluence outcome such as attention from study personnel or time spent on the procedure [15, 16]. Hence there The REIL study has received ethical approval from Cam- den and Islington National Research Ethics Service (13/ LO/1327), who are the responsible for approval of final protocol as well as any modifications or amendments. Local site approval from St. George’s Hospital Joint Re- search and Enterprise Office (JREO) has also been ob- tained (13.0108). The JREO is an independent office Page 3 of 9 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:17 Page 3 of 9 Page 3 of 9 (2019) 11:17 providing external review and monitoring of any re- search undertaken at St. George’s NHS Trust in order to maintain clinical research governance guidelines and standards. Randomisation A random allocation list will be prepared by the depart- ment statistician, generated using GraphPad randomisa- tion software (GraphPad Prism version 6.04 for Windows, GraphPad software, CA). On entry into the study each participant will be assigned an anonymous ID number and each number will be allocated an interven- tion (exercise or relaxation) on the list. As all assessment sessions will be carried out by the single principal inves- tigator (SH), it will not be possible to blind the investiga- tor to intervention. Also due to the nature of the intervention participants cannot be blinded to group allocation. The study has been registered on a publicly accessible database, ClinicalTrials.gov, NCT02272751. Interventions Interventions for this study are reported using the tem- plate for intervention description and replication (TIDieR) checklist and guide [24]. Interventions for this study are reported using the tem- plate for intervention description and replication (TIDieR) checklist and guide [24]. For both groups, advice, instruction, demonstration and practice will be carried out at baseline. This will be delivered by the principal investigator SH, an experi- enced physiotherapist with a special interest and experi- ence in oncology. Participants will be able to contact SH by telephone in between sessions if additional support is required. All participants will be advised on potential ad- verse events and what to report including pain, fatigue, muscle soreness, lymphoedema, nausea, dyspnoea, dizzi- ness, tachycardia and cramp. A logbook will be provided to all participants to encourage adherence and document self-reported participation levels, intervention type car- ried out, frequency, intensity, duration and any adverse events. Exercise intervention While there are no guidelines for the delivery of exercise in cancer survivors, it is recommended to follow the guidelines for general UK population [25]. Hence partici- pants randomised to the exercise intervention in this study will be advised to carry out the programme for at least 50 min three times per week. Eligibility criteria l Inclusion criteria include patients with histologically confirmed lymphoma in remission post–chemotherapy, chemotherapy treatment completed within the last six weeks, age 18 years or older, able to give informed con- sent, good performance status (assessed by the Eastern Cooperative Oncology Group {ECOG} status 0–2) [21] and medically able to carry out an exercise training programme. Patients with active disease, unstable angina or unexplained electrocardiogram, poor performance status (ECOG status 3 or more), pregnancy, difficulty breathing at rest, persistent cough, fever or illness, or any cognitive impairment limiting ability to give in- formed consent or complete QoL questionnaires will be excluded. Written informed consent will be obtained from all individual participants included in the study. Recruitment Participants will be recruited from a single specialist clinical setting – the Haematology-Oncology Out-Pa- tient (HOOP) Clinic at St George’s Hospital, London. Assessment for eligibility, recruitment, medical screening and obtaining of informed consent will be carried out by the patients’ medical consultant (RP). In order to en- courage participant enrolment, potential participants will be introduced to the principal investigator (SH) who will explain about the study, answer questions and provide with the written participant information sheet to take away and read before making a decision. Flow of participants in the study is shown in Fig. 1. Sample size Sample size was calculated to determine clinically rele- vant effects on the primary outcome measure, the Euro- pean Organisation for Research and Treatment of Cancer – QoL core questionnaire (EORTC QLQ-C30, version 3.0). A minimally important difference of 5 to 10 points is generally accepted as clinically meaningful [22], and calculations were based on comparison of means be- tween two groups. Assuming a two-sided significance level level (α) of 0.05, power of 80% (ß = 0.20) and stand- ard deviation from EORTC website reference values [23], it was determined that a sample of minimum 46 participants will be required to detect a significant change in the EORTC summary score, 23 in each inter- vention group. The exercise programme includes elements of aerobic, upper and lower limb resistance training, core stability and stretches and is designed to be performed independ- ently at home. For the aerobic component, participants will be advised to walk indoors or outdoors for a mini- mum of 30 min and to aim for moderate intensity as rec- ommended by the Borg Rate of Perceived Exertion (RPE) scale [26]. Where unable to maintain moderate in- tensity due to fatigue, weakness or other complaints, participants will be advised to aim for moderate intensity aerobic exercise for any duration of time able, and to gradually increase over the next few weeks aiming to build up to 150 min over one week. They will be taught Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:17 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:17 Page 4 of 9 (2019) 11:17 Page 4 of 9 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation Fig. 1 Flow of participants in the REIL Study. CONSORT flow diagram demonstrating participant flow through each phase of the randomised controlled trial (enrolment, intervention allocation, follow-up and data analysis) Relaxation intervention resistance exercises for the major upper and lower limb muscles using body weight or TheraBand™isometric re- sistance bands to carry out following aerobic activity. The resistance training follows American College of Sports Medicine (ACSM) guidelines [27], and partici- pants will be advised to carry out three sets of each exer- cise for 8–12 repetitions. They will also be taught stretches for upper and lower limb muscles, and core abdominal exercises in supine. Participants will be also be provided with written instruction, leaflet with photographs and advice sheet. Similar programmes in- corporating aerobic, resistance and flexibility training three times a week over a twelve week period have been used with good results in improving QoL in cancer survivors [9]. The relaxation intervention consists of a bed- or chair- based program, developed based on the literature on mindfulness-based interventions in cancer survivors [10–12]. An audio CD was produced to guide partici- pants through relaxation techniques incorporating mind- fulness meditation, breathing exercises, guided visualisation and progressive muscle relaxation. Here also participants will be advised to carry out the programme for 50 min three times a week. Participants will also be provided with written information on relax- ation and tips on how to incorporate into daily life. Par- ticipants in this group will not be advised to perform any exercise outside of their normal habits, nor asked to avoid activity. Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:17 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:17 Page 5 of 9 (2019) 11:17 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation answered on a four-point scale and scores are derived according to the EORTC scoring manual [30]. Participants will be followed-up at six weeks, midway through the programme. At this stage all assessments will be repeated, any change noted and they will be ad- vised on how to modify the programme (relaxation or exercise) as appropriate. Once participants have com- pleted twelve weeks of their intervention, they will be provided with resources of the other intervention group, for their information or to carry out as desired, in ac- cordance with the Declaration of Helsinki. Body composition Due to the illness itself and treatment and side-effects (including steroid use, loss of appetite, nausea and vomiting), lymphoma patients may experience a fluctu- ation in weight and body fat percentage during the dur- ation of their treatment. Therefore body composition will be assessed using a bioelectrical impedance analyser (Tanita BC-418) to monitor whether it will stabilise to within the desired range. Participants’ standing stature will be measured using a stadiometer (SECA, Germany). Height measured will be inserted into the body compos- ition analyser to obtain weight, body mass index (BMI), and body fat percentage. Participants will be measured without footwear and in light clothing using the same equipment each time. Outcome measures All outcome measures will be assessed at baseline prior to commencing intervention, midway (six weeks) and on completion of intervention programme (twelve weeks). Outcome measures, instruments and guidelines used and time assessed are summarised in Table 1. At base- line, patient demographics including gender, age, social history and medical history will be recorded. Psychological status The Hospital Anxiety and Depression Scale (HADS) is a valid tool for assessing the severity of anxiety disorders and depression in various populations including psychi- atric, primary care patients and the general population [37]. The HADS is also recommended in cancer settings [38]. Questions are answered using a four-point scale and higher scores represent higher levels of anxiety and depression. Muscle strength Grip strength is a good indication of global muscle strength and assessment of grip is common method that is used to assess general strength characteristics [32]. Here, isometric grip strength will be measured in kilo- grams using an adjustable hand-held dynamometer (Jamar® Hydraulic Hand Dynamometer, USA) with the elbow flexed at 90 degrees and the forearm and wrist in neutral position following standardised American Col- lege of Sports Medicine (ACSM) guidelines [32]. End of study questionnaire During their final assessment at twelve weeks, partici- pants will be invited to complete the end of study ques- tionnaire. The open questions aim to explore their thoughts and feelings about the intervention programme, and to analyse their views on their particu- lar intervention including preferences, perceived advan- tages and disadvantages and reasons for participation and adherence. Quality of life p y Resting Blood Pressure (BP) and Heart Rate (HR) will be assessed using an electronic BP monitor (Omron M10- IT, Japan) and finger pulse oximeter (Solaris S10A, China) in sitting. In addition, measures of pulmonary capacity and function will be assessed using a hand-held microspirometer (Micro 1 Medical Microspirometer, England) following British Thoracic Society guidelines [31]. Forced expiratory volume in one second (FEV1), The primary outcome measure is QoL, assessed by the EORTC QLQ-C30. This self-reported questionnaire has been demonstrated to be a valid and reliable tool, takes approximately eleven minutes to complete and most subjects require no assistance [28]. This has been used extensively in cancer survivorship research, and also spe- cifically with lymphoma patients [29]. The EORTC QLQ-C30 consists of thirty items. Each question is Table 1 Primary and Secondary Outcome Measures Outcome measured Instrument Table 1 Primary and Secondary Outcome Measures Outcome measured Instrument Guidelines When assessed Baseline Midway 6 weeks End of intervention 12 weeks Primary outcome Quality of Life EORTC QLQ-C30 Self-reported Questionnaire ☑ ☑ ☑ Secondary outcomes Body Composition Height, Weight, Bioimpedance analysis – BMI, Body Fat % - TANITA BC-418 Body Composition Analyzer Same investigator and equipment, time of day and clothing kept similar where possible ☑ ☑ ☑ Cardiovascular status – Resting BP, HR, SpO2 Omron M10-IT electronic BP monitor, Solaris Finger Pulse Oximeter S10A ☑ ☑ ☑ Pulmonary Function – FEV1, FVC, FER, PEF Hand-held Micro 1 Medical Microspirometer British Thoracic Society, 2013 ☑ ☑ ☑ Muscle Strength – Grip Strength Jamar® Hydraulic Hand Dynamometer American College of Sports Medicine, 2014 ☑ ☑ ☑ Functional Exercise Capacity Six Minute Walk Test American Thoracic Society, 2002 ☑ ☑ ☑ Physical Well-being, Social/Family Well-being, Emotional Well-being, Functional Well-being FACT-Lym Self-reported Questionnaire Anxiety and Depression HADS ☑ ☑ ☑ Feelings and Perceptions about participating in programme End of Study Questionnaire Open-ended Questionnaire mailed to participants’ home ☑ Guidelines Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:17 Page 6 of 9 (2019) 11:17 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation forced vital capacity (FVC), forced expiratory ratio (FER) and expiratory peak flow (PEF) will be recorded. lymphoma subscale. The FACT-Lym has good internal consistency and validity [35]. All questions are answered on a five-point scale and scoring guidelines are recom- mended together with the questionnaire [36]. Functional exercise capacity Due to the variation in the population being studied in terms of age, co-morbidity and physical ability, a sub- maximal test was considered appropriate for this study. The Six-Minute Walk Test (6MWT) is considered a good indicator of functional ability and exercise capacity and is a common outcome measure to evaluate progress in rehabilitation programmes [33]. This test was chosen as it can be undertaken in a clinical setting with ease, takes a short duration of time to complete and appeared appropriate for this sample of patients following chemo- therapy who are generally deconditioned and fatigue quickly. The 6MWT has been validated and used exten- sively in a variety of patient populations, and has been recommended for use in cancer patients also [34]. Data analysis ll b y Data will be entered into the Microsoft Excel (2013) database by SH and analysed by all authors using the IBM SPSS version 22 (SPSS, Inc., Chicago, IL) statis- tical software package. The intent-to-treat principle will be applied and the significance level will be set at 0.05. Descriptive statistics will be used to present baseline demographic characteristics. To compare dif- ference in the primary outcome measure, QoL score between the two groups post-intervention, ANCOVA (analysis of covariance) will be used, with pre-test values as covariate to adjust for any baseline differ- ences. Data will be analysed to check fit for ANCOVA assumptions. Post-hoc paired-samples t- tests for pre-post intervention will be tested within groups with Bonferroni corrections. Missing data will be treated as recommended by scoring guidelines of questionnaires [24, 30]. These analyses will be re- peated for secondary outcome measures, providing data fit normal distribution and other assumptions for ANCOVA. The influence of potential confounding factors such as age, gender, number of co-morbidities and ECOG status will be explored using correlation analyses. Adherence and drop-out rates will also be analysed. Qualitative data from the end of study ques- tionnaire will be analysed for codes and themes using qualitative content analysis [39]. The 6MWT will be carried out following guidelines recommended by the American Thoracic Society (ATS) [33]. The test will be carried out in a 30 m marked dis- tance, and the total distance covered in 6 min will be re- corded, as well as post-test Heart Rate and oxygen saturation, RPE and Rate of Dyspnoea [26]. The test will cease prior to 6 min if the patient chooses to stop or the tester terminates testing (indications for terminating ex- ercise as per the ATS guidelines include the following: severe exhaustion or shortness of breath, wheezing, diz- ziness, chest pain or muscle cramps) [33]. At no point will the patient be encouraged to continue beyond the level at which they wish to stop. Discussion A large proportion of lymphoma survivors continue to experience unmet needs following treatment [1–3]. Long-term and late effects are often overlooked, and survivors do not routinely receive advice or interventions to target these and maximise well-being [40]. Previous studies have supported the use of both relaxation and exercise interventions to treat physical and psychosocial complaints of cancer survivors [9–12]. No trial to date has compared efficacy of these two interventions in lymphoma survivors. It will not be possible to stratify participants by factors such as age, gender, etc. during randomization due to the small sample size. However, sample size was calcu- lated using reference values (including standard devi- ation) and power calculations recommended by the EORTC reference manual [23]; similar trials on the ef- fects of exercise in cancer patients have been carried out with comparable sample sizes, and these were able to de- tect significant results [9]. The REIL study aims to study a sample of lymph- oma survivors shortly post-chemotherapy, and com- pare the effectiveness of relaxation and exercise on improving QoL. This study was developed to address some issues highlighted in survivorship literature in- cluding focus on survivors of cancers other than breast or prostate, emphasis on the transition phase immediately following treatment, addressing physical, psychological and social needs of lymphoma survivors, use of patient-reported outcome measures, self-man- agement, and moving from control groups to com- parison between two interventions. In addition to quantitative data from outcome mea- sures, the end of study questionnaire will explore partici- pants’ perceptions and feelings towards an intervention programme post-chemotherapy. Such information will highlight behaviours and patterns in this sample of lymphoma survivors including reasons for participation, adherence and non-adherence, and any preferences; it is anticipated that this data will ultimately help in the pro- motion of a healthy lifestyle in lymphoma and cancer survivors. The authors will aim to publish results from this study to add to the evidence to inform healthcare professionals on effective interventions to improve QoL of lymphoma survivors. In this study lymphoma patients in remission will be recruited within six weeks of their last chemother- apy session. Patients during this early phase of sur- vivorship who have completed initial cancer treatment have been relatively neglected [19], and this period immediately following treatment has been identified as a key time to address side effects of treatment and facilitate return to pre-morbid health [41]. Well-being h The Functional Assessment of Chronic Illness Therapy (FACIT) questionnaires have been validated in studies of cancer management and are designed to encompass a range of psychosocial factors [35]. Like the EORTC QLQ-C30, these questionnaires are also a patient re- ported measure of QoL. However these questionnaires focus on aspects of well-being including physical, social, emotional and functional well-being subscales. Here, the lymphoma-specific scale (FACT-Lym) will be used (comprising the general or FACT-G questionnaire plus a Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:17 Page 7 of 9 (2019) 11:17 (2019) 11:17 Page 7 of 9 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation flexibility, familiar surroundings, family support and reduced travel requirements [45]. Discussion Studies on cancer survivors have indicated that they would prefer to begin an exercise programme immediately or soon after treatment [42], and at this point the focus can move from the disease and treatment to ‘moving on’ to the next phase of life [43]. Conclusion It is well documented that lymphoma survivors commonly suffer from consequences of treatment such as pain, fa- tigue, decreased function, anxiety and depression which have a negative impact on their quality of life. However, there is no standardize care pathway recommended for lymphoma survivors following chemotherapy. The current proposed REIL study aims to determine the more effective intervention of two in improving QoL in lymphoma survivors. It is anticipated that re- sults from this preliminary study will help build to- wards the development of feasible and effective practice guidelines to improve QoL of lymphoma sur- vivors post-chemotherapy. While the additional psychosocial benefits of super- vised group classes for cancer survivors have been recog- nised [44], the need for a shift towards self-management in cancer survivorship has been highlighted [6, 7]. Here participants are supported to self-manage in a variety of ways, but the onus will be on them to initiate contact with the healthcare professionals for additional support when needed [7, 8]. Both the relaxation and exercise programmes in this study are designed to be performed independently and participants will be supplied with re- sources to allow convenient home-based performance, including written instructions and advice. Participants will be able to contact the investigator whenever re- quired. This study will provide results including ad- herence and drop-out rates that are more generalizable to ‘real-world’ lymphoma survivors than controlled clinic-based research. Home-exercise pro- grammes also afford patients better scheduling 6MWT: Six Minute Walk Test; ACSM: American College of Sports Medicine; ANOVA: Analysis of Variance; ATS: American Thoracic Society; BMI: Body Mass Index; BP: Blood Pressure; CD: Compact Disc; COPD: Chronic Obstructive Pulmonary Disease; ECOG: Eastern Cooperative Oncology Group; EORTC QLQ-30: European Organisation for Research and Treatment of Cancer – QoL core questionnaire; FACIT: Functional Assessment of Chronic Illness Therapy; FACT-Lym: Functional Assessment of Cancer Therapy - Lymphoma; FER: Forced Expiratory Rate; FEV1: Forced Expiratory Volume; FVC: Forced Vital Capacity; HADS: Hospital Anxiety and Depression Scale; HOOP: Haematology and Oncology Out-Patients; HR: Heart Rate; NCSI: National Cancer Survivorship Initiative; NHS: National Health Service; PEF: Peak Expiratory Flow; QoL: Quality of Life; REIL: Relaxation and Exercise In Lymphoma; RPE: Rate of Perceived Exercise; TIDieR: Template for Intervention Description and Replication; UK: United Kingdom Received: 29 November 2018 Accepted: 18 July 2019 25. Campbell A, Stevinson C, Crank H. The BASES expert statement on exercise and cancer survivorship. J Sports Sci. 2012;i:1–4. Availability of data and materials 13. Lymphoma Action. 2019. https://lymphoma-action.org.uk/ (Accessed 22 July 2019). The written material and datasets used and/or analysed during the current study will be available from the corresponding author on reasonable request. 14. Macmillan Cancer Support. 2019. https://www.macmillan.org.uk/information- and-support (Accessed 22 July 2019). 14. Macmillan Cancer Support. 2019. https://www.macmillan.org.uk/information- and-support (Accessed 22 July 2019). 15. Bower JE. Mindfulness interventions for cancer survivors: moving beyond wait-list control groups. J Clin Oncol. 2016;34:3366–8. 15. Bower JE. Mindfulness interventions for cancer survivors: moving beyond wait-list control groups. J Clin Oncol. 2016;34:3366–8. Acknowledgments Not applicable. 8. Foster C, Fenlon D. Recovery and self-management support following primary cancer treatment. Br J Cancer. 2011;105:S21–8. 8. Foster C, Fenlon D. Recovery and self-management support following primary cancer treatment. Br J Cancer. 2011;105:S21–8. Foster C, Fenlon D. Recovery and self-management support following primary cancer treatment. Br J Cancer. 2011;105:S21–8. 9. Gerritson JKW, Vincent AJPE. Exercise improves quality of life in patients with cancer: a systematic review and meta-analysis of randomised controlled trials. Br J Sports Med. 2016;50:796–803. 9. Gerritson JKW, Vincent AJPE. Exercise improves quality of life in patients with cancer: a systematic review and meta-analysis of randomised controlled trials. Br J Sports Med. 2016;50:796–803. Funding 12. Reich RR, Lengacher CA, Alinat CB, et al. Mindulness-based stress reduction in post-treatment breast cancer patients: immediate and sustained effects across multiple symptom clusters. J Pain Symptom Manag. 2017;53:85–95. There are no sources of funding for this study. Abbreviations 6MWT Si Mi 6MWT: Six Minute Walk Test; ACSM: American College of Sports Medicine; ANOVA: Analysis of Variance; ATS: American Thoracic Society; BMI: Body Mass Index; BP: Blood Pressure; CD: Compact Disc; COPD: Chronic Obstructive Pulmonary Disease; ECOG: Eastern Cooperative Oncology Group; EORTC QLQ-30: European Organisation for Research and Treatment of Cancer – QoL core questionnaire; FACIT: Functional Assessment of Chronic Illness Therapy; FACT-Lym: Functional Assessment of Cancer Therapy - Lymphoma; FER: Forced Expiratory Rate; FEV1: Forced Expiratory Volume; FVC: Forced Vital Capacity; HADS: Hospital Anxiety and Depression Scale; HOOP: Haematology and Oncology Out-Patients; HR: Heart Rate; NCSI: National Cancer Survivorship Initiative; NHS: National Health Service; PEF: Peak Expiratory Flow; QoL: Quality of Life; REIL: Relaxation and Exercise In Lymphoma; RPE: Rate of Perceived Exercise; TIDieR: Template for Intervention Description and Replication; UK: United Kingdom Page 8 of 9 Page 8 of 9 Page 8 of 9 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:17 (2019) 11:17 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation Consent for publication Not applicable. Not applicable. 20. Chan A, Tetzlaff JM, Gøtzche PC, et al. SPIRIT 2013 explanation and elaboration: guidance for protocols of clinical trials. BMJ. 2013;346:e7586. Author details 1 22. Osaba D, Rodrigues G, Myles J, et al. Interpreting the significance of changes in health-related quality-of-life scores. J Clin Oncol. 1998;16:139–44. 1Kingston University and St. George’s University of London, London, UK. 2St. George’s Healthcare NHS Trust, London, UK. 3Kingston University, Kingston Upon Thames, London, UK. 4Faculty of Health, Social Care & Education, Cranmer Terrace, London SW17 0RE, UK. 23. Scott NW, Fayers PM, Aaronson NK, et al. EORTC QLQ-C30 Reference Values. 2008. https://www.eortc.org/app/uploads/sites/2/2018/02/reference_values_ manual2008.pdf (Accessed 22 July 2019). 24. Hoffman TC, Glasziou PP, Boutron I, et al. Better reporting of interventions: template for intervention description and replication (TIDier) checklist and guide. BMJ. 2014;348:g1687. Received: 29 November 2018 Accepted: 18 July 2019 Authors’ contributions All authors contributed to the development of study protocol. SH was a major contributor in writing the manuscript. RP contributed to aims, eligibility and recruitment sections; AY to sample size, randomisation and statistical analysis; HM contributed to outcome measures, and feedback to all sections of paper. All authors read and approved the final manuscript. 10. Duncan M, Moschopoulou E, Herrington E, et al. Review of systematic reviews of non-pharmacological interventions to improve quality of life in cancer survivors. BMJ Open. 2017;7:e015860. 11. Parás-Bravo P, Salvadores-Fuentes P, Alonso-Blanco C, et al. The impact of muscle relaxation techniques on the quality of life of cancer patients, as measured by the FACT-G questionnaire. PLoS One. 2017;12:e0184147. The authors declare that they have no competing interests. The authors declare that they have no competing interests. References 26. Borg GAV. Psychophysical bases of perceived exertion. Med Sci Sports Exer. 1982;14:877–81. 1. Damlaj M, El Fakih R, Hashmi SK. Evolution of survivorship in lymphoma, myeloma and leukemia: metamorphosis of the field into long term follow- up care. Blood Rev. 2019;33:63–73. 1. Damlaj M, El Fakih R, Hashmi SK. Evolution of survivorship in lymphoma, myeloma and leukemia: metamorphosis of the field into long term follow- up care. Blood Rev. 2019;33:63–73. 27. American College of Sports Medicine. ACSM’s guidelines for exercise prescription and testing. 9th ed. Philadelphia: Lipincott Williams & Wilkins; 2014. 2. Ciavarella S, Minoia C, Quinto AM, et al. Improving provision of care for long-term survivors of lymphoma. Clin Lymphoma Myeloma Leuk. 2017;17:e1–9. 28. Aaronson NK, Ahmedzai S, Begman B, et al. The European Organisation for Research and Treatment of Cancer QLQ-C30: a quality-of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst. 1993;85: 365–76. 3. Raphael D, Frey R, Gott M. Maintaining psychosocial wellbeing for post- treatment haematological cancer survivors: strategies and potential barriers. Eur J Oncol Nurs. 2019;38:36–41. 29. Heutte N, Flechtner HH, Mourier N, et al. Quality of life after successful treatment of early-stage Hodgkin's lymphoma: 10-year follow-up of the EORTC–GELA H8 randomised controlled trial. Lancet Oncol. 2009;10:1160–70. 4. Bhuva N, Li SP, Maher J. Living with and beyond Cancer: new challenges. In: Mohan R, editor. Topics in Cancer survivorship. Croatia: InTech Europe; 2012. p. 1–12. 30. Fayers PM, Aaronson NK, Bjordal K, et al. EORTC QLQ-C30 scoring manual. 3rd ed. Brussels: EORTC; 2001. 5. Maher EJ. Managing the consequences of cancer treatment and the English National Cancer Survivorship Initiative. Acta Oncol. 2013;52:225–32. 31. The British Thoracic Society. A Guide to Performing Quality Assured Diagnostic Spirometry. 2013. https://www.brit-thoracic.org.uk/media/70454/ spirometry_e-guide_2013.pdf. (Accessed 22 July 2019). 6. Richards M, Corner J, Maher J. The national cancer survivorship initiative: new and emerging evidence on the ongoing needs of cancer survivors. Br J Cancer. 2011;105:S1–4. spirometry_e-guide_2013.pdf. (Accessed 22 July 2019). 32. American College of Sports Medicine. ACSM’s health-related physical fitness assessment manual. 4th ed. Philadelphia: Lippincott Williams & Wilkins; 2014. 7. DH, Macmillan Cancer Support, NHS Improvement. National Cancer Survivorship Initiative Vision. 2010. https://webarchive.nationalarchives.gov. uk/20100809113601/http://www.improvement.nhs.uk/cancer/LinkClick. aspx?fileticket=4apVUSvGcow%3d&tabid=214. (Accessed 22 July 2019). 33. American Thoracic Society. ATS statement: guidelines for the six-minute walk test. Am J Respir Crit Care Med. 2002;166:111–7. 33. American Thoracic Society. ATS statement: guidelines for the six-minute walk test. Ethics approval and consent to participate The REIL study has received ethical approval from Camden and Islington National Research Ethics Service (13.LO.1327), as well as local site approval from St. George’s Hospital Joint Research and Enterprise Office (13.0108). The study has been registered on a publicly accessible database, ClinicalTrials.gov (NCT02272751). 16. Buchanan DR, White JD, O’Mara AM, et al. Research-design issues in cancer- symptom-management trials using complementary and alternative medicine: lessons from the National Cancer Institute Community clinical oncology program experience. J Clin Oncol. 2005;23:6682–9. 17. Carlson LE, Doll R, Stephen J, et al. Randomized controlled trial of mindfulness-based cancer recovery versus supportive expressive group therapy for distressed survivors of breast cancer (MINDSET). J Clin Oncol. 2013;25:3119–26. Written informed consent will be obtained from all individual participants included in the study. All participants will remain anonymous and records stored securely as per St. George’s NHS Trust research guidelines. All procedures performed involving human participants will be 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. 18. Schmidt ME, Wiskemann J, Krakowski-Roosen H, et al. Progressive resistance versus relaxation training for breast cancer patients during adjuvant chemotherapy: design and rationale of a randomized clinical trial (BEATE study). Contemp Clin Trials. 2013;34:117–25. 19. Richardson A, Addington-Hall J, Amir Z, et al. Knowledge, ignorance and priorities for research in key areas of cancer survivorship: findings from a scoping review. Br J Cancer. 2011;105:S82–94. Competing interests 21. Oken MM, Creech RH, Tormey DC, et al. Toxicity and response criteria of the eastern cooperative oncology group. Am J Clin Oncol. 1982;5:649–55. The authors declare that they have no competing interests. References A qualitative study of the post- treatment experiences and support needs of survivors of lymphoma. Eur J Oncol Nurs. 2017;28:62–8. 44. Leach HJ, Covington KR, Voss C, et al. Effect of group dynamics-based exercise versus personal training in breast cancer survivors. Oncol Nurs Forum. 2019;46:185–97. 45. Olney SJ, Nymark J, Brouwer B, et al. A randomised controlled trial of supervised versus unsupervised exercise programs for ambulatory stroke survivors. Stroke. 2006;37:476–81. References Am J Respir Crit Care Med. 2002;166:111–7. Page 9 of 9 Hathiramani et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:17 34. Schmidt K, Vogt L, Thiel C, et al. Validity of the six-minute walk test in cancer patients. Int J Sports Med. 2010;34:631–6. 34. Schmidt K, Vogt L, Thiel C, et al. Validity of the six-minute walk test in cancer patients. Int J Sports Med. 2010;34:631–6. 35. Webster K, Cashy J, Cella D, et al. Measuring quality of life (QOL) in patients with non-Hodgkin’s lymphoma (NHL): the functional assessment of Cancer therapy-lymphoma (FACT-Lym). Qual Life Res. 2005;14:1650. 35. Webster K, Cashy J, Cella D, et al. Measuring quality of life (QOL) in p with non-Hodgkin’s lymphoma (NHL): the functional assessment of C therapy-lymphoma (FACT-Lym). Qual Life Res. 2005;14:1650. therapy-lymphoma (FACT-Lym). Qual Life Res. 2005;14:1650. 36. FACIT. FACT-Lym Scoring and Interpretation Materials. 2010. http://www. facit.org/FACITOrg/Questionnaires. (Accessed 22 July 2019). 36. FACIT. FACT-Lym Scoring and Interpretation Materials. 2010 36. FACIT. FACT-Lym Scoring and Interpretation Materials. 2010. http://www facit.org/FACITOrg/Questionnaires. (Accessed 22 July 2019). facit.org/FACITOrg/Questionnaires. (Accessed 22 July 2019) 37. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiat Scand. 1983;67:361–70. 37. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiat Scand. 1983;67:361–70. 38. Mitchell AJ, Meader N, Symonds P. Diagnostic validity of the hospital anxiety and depression scale (HADS) in cancer and palliative settings: a meta-analysis. J Affect Disord. 2010;126:335–48. 38. Mitchell AJ, Meader N, Symonds P. Diagnostic validity of the hospital anxiety and depression scale (HADS) in cancer and palliative settings: a meta-analysis. J Affect Disord. 2010;126:335–48. 39. Graneheim UK, Lundman B. Qualitative content analysis in nursing research: concepts, procedures and measures to achieve trustworthiness. Nurse Educ Today. 2004;24:105–12. y 40. Ganz PA. Q&A: the ‘three Ps’ of cancer survivorship careBMC. 2011;Med 9:1–3. 40. Ganz PA. Q&A: the ‘three Ps’ of cancer survivorship careBMC. 2011;Med 9:1–3. 41. Spence RR, Heesch KC, Brown WJ. Exercise and cancer rehabilitation: a systematic review. Cancer Treat Rev. 2010;36:185–94. systematic review. Cancer Treat Rev. 2010;36:185–94. 42. Courneya KS, Friedenreich CM. Physical activity and cancer control. Semin Oncol Nurs. 2007;23:242–52. 43. Monterosso L, Taylor K, Platt V, et al. A qualitative study of the post- treatment experiences and support needs of survivors of lymphoma. Eur J Oncol Nurs. 2017;28:62–8. 43. Monterosso L, Taylor K, Platt V, et al. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
https://openalex.org/W1970035873	https://escholarship.org/content/qt5cb1z014/qt5cb1z014.pdf?t=napr3t	English	null	Comparative Analysis of Genome Sequences Covering the Seven Cronobacter Species	PloS one	2,012	cc-by	12,042	UC Irvine Copyright Information This work is made available under the terms of a Creative Commons Attribution License, availalbe at https://creativecommons.org/licenses/by/4.0/ UC Irvine UC Irvine Previously Published Works Title Comparative Analysis of Genome Sequences Covering the Seven Cronobacter Species Permalink https://escholarship.org/uc/item/5cb1z014 Journal PLoS ONE, 7(11) ISSN 1932-6203 Authors Joseph, Susan Desai, Prerak Ji, Yongmei et al. Publication Date 2012-11-16 DOI 10.1371/journal.pone.0049455 Supplemental Material https://escholarship.org/uc/item/5cb1z014#supplemental Copyright Information This work is made available under the terms of a Creative Commons Attribution License, availalbe at https://creativecommons.org/licenses/by/4.0/ Peer reviewed UC Irvine UC Irvine Previously Published Works Title Comparative Analysis of Genome Sequences Covering the Seven Cronobacter Species Permalink https://escholarship.org/uc/item/5cb1z014 Journal PLoS ONE, 7(11) ISSN 1932-6203 Authors Joseph, Susan Desai, Prerak Ji, Yongmei et al. Publication Date 2012-11-16 DOI 10.1371/journal.pone.0049455 Supplemental Material https://escholarship.org/uc/item/5cb1z014#supplemental Copyright Information This work is made available under the terms of a Creative Commons Attribution License, availalbe at https://creativecommons.org/licenses/by/4.0/ Peer reviewed Abstract Read, Emory University, United States of America Received July 24, 2012; Accepted October 9, 2012; Published November 16, 2012 Received July 24, 2012; Accepted October 9, 2012; Published November 16, 2012 Received July 24, 2012; Accepted October 9, 2012; Published November 16, 2012 Copyright: 2012 Joseph et al. This is an open-access article distributed under the terms of the Creative Commons Attr unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Copyright: 2012 Joseph et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Life Technologies funded and performed the genome sequencing. The authors thank Alain Rico for the sequencing of strains using the C. sakazakii 680, C. dublinensis 1210 and C. condimenti 1330 using the Ion Torrent PGMTM. MM and PD were supported in part by United States National Institutes of Health grants AI039557 AI052237, AI073971, AI075093, AI077645, AI083646, USDA grants 2009-03579 and 2011-67017-30127, the Binational Agricultural Research and Development Fund, and CDMRP BCRP W81XWH-08-1-0720. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have the following interests. YJ was employed by Life Technologies, RS is an employee of Life Technologies, LD is an employee of Life Technologies, PB is an employee of Life Technologies, AR was employed by Life Technologies, CAC was employed by Life Technologies and owns stock in the company, MF is an employee of Life Technologies and owns stock in LIFE. This work was funded by Life Technologies who also performed the genome sequencing. In addition, one of the co-authors, Craig Cummings is now an employee of Sequenta, Inc. There are no patents, products in development or marketed products to declare. This does not alter the authors’ adherence to all the PLOS ONE policies on sharing data and materials. * E-mail: stephen.forsythe@ntu.ac.uk (SJF) ;mmcclelland@sdibr.org (MM) * E-mail: stephen.forsythe@ntu.ac.uk (SJF) ;mmcclelland@sdibr.org (MM) . These authors contributed equally to this work. ¤a Current address: The University of Chicago Booth School of Business, Chicago, Illinois, United States of America ¤b Current address: Sequenta, Inc., South San Francisco, California, United States of America the Cronobacter are most closely related to the Enterobacter and Citrobacter genera. Abstract They are ubiquitous organisms present in a wide range of environments, including water, soil, and a variety of fresh and processed foods [1]. The bacterium has been isolated from factory production lines, including powdered infant formula factories, and households [2] as well as clinical sources such as Comparative Analysis of Genome Sequences Covering the Seven Cronobacter Species Susan Joseph1., Prerak Desai2., Yongmei Ji3.¤a, Craig A. Cummings3¤b, Rita Shih3, Lovorka Degoricija3, Alain Rico4, Pius Brzoska3, Stephen E. Hamby5, Naqash Masood1, Sumyya Hariri1, Hana Sonbol1, Nadia Chuzhanova5, Michael McClelland2,6, Manohar R. Furtado3, Stephen J. Forsythe1 1 Pathogen Research Group, School of Science and Technology, Nottingham Trent University, Nottingham, United Kingdom, 2 Vaccine Research Institute San Diego, San Diego, California, United States of America, 3 Life Technologies Corporation, Foster City, California, United States of America, 4 Life Technologies Corporation, Darmstadt, Germany, 5 Bioinformatics and Biomathematics Group, School of Science and Technology, Nottingham Trent University, Nottingham, United Kingdom, 6 Department of Pathology and Laboratory Medicine, University of California Irvine, Irvine, California, United States of America Powered by the California Digital Library University of California eScholarship.org Comparative Analysis of Genome Sequences Covering the Seven Cronobacter Species Abstract Background: Species of Cronobacter are widespread in the environment and are occasional food-borne pathogens associated with serious neonatal diseases, including bacteraemia, meningitis, and necrotising enterocolitis. The genus is composed of seven species: C. sakazakii, C. malonaticus, C. turicensis, C. dublinensis, C. muytjensii, C. universalis, and C. condimenti. Clinical cases are associated with three species, C. malonaticus, C. turicensis and, in particular, with C. sakazakii multilocus sequence type 4. Thus, it is plausible that virulence determinants have evolved in certain lineages. Methodology/Principal Findings: We generated high quality sequence drafts for eleven Cronobacter genomes representing the seven Cronobacter species, including an ST4 strain of C. sakazakii. Comparative analysis of these genomes together with the two publicly available genomes revealed Cronobacter has over 6,000 genes in one or more strains and over 2,000 genes shared by all Cronobacter. Considerable variation in the presence of traits such as type six secretion systems, metal resistance (tellurite, copper and silver), and adhesins were found. C. sakazakii is unique in the Cronobacter genus in encoding genes enabling the utilization of exogenous sialic acid which may have clinical significance. The C. sakazakii ST4 strain 701 contained additional genes as compared to other C. sakazakii but none of them were known specific virulence- related genes. Conclusions/Significance: Genome comparison revealed that pair-wise DNA sequence identity varies between 89 and 97% in the seven Cronobacter species, and also suggested various degrees of divergence. Sets of universal core genes and accessory genes unique to each strain were identified. These gene sequences can be used for designing genus/species specific detection assays. Genes encoding adhesins, T6SS, and metal resistance genes as well as prophages are found in only subsets of genomes and have contributed considerably to the variation of genomic content. Differences in gene content likely contribute to differences in the clinical and environmental distribution of species and sequence types. Citation: Joseph S, Desai P, Ji Y, Cummings CA, Shih R, et al. (2012) Comparative Analysis of Genome Sequences Covering the Seven Cronobacter Species. PLoS ONE 7(11): e49455. doi:10.1371/journal.pone.0049455 Citation: Joseph S, Desai P, Ji Y, Cummings CA, Shih R, et al. (2012) Comparative Analysis of Genome Sequences Covering the Seven Cronobacter Species. PLoS ONE 7(11): e49455. doi:10.1371/journal.pone.0049455 Editor: Timothy D. November 2012 \| Volume 7 \| Issue 11 \| e49455 Comparative Analysis of Cronobacter Genomes both C. sakazakii and C. malonaticus survive and replicate in macrophages inside phagosomes, whereas C. muytjensii dies and C. dublinensis is serum sensitive. Virulence also varies within the C. sakazakii species as evident from epidemiological studies of a NICU outbreak in France where the clinical outcome of three C. sakazakii pulsetypes differed, with only one pulsetype (now recognised as ST4 strains) causing the three deaths [8,19]. To date, only strains from C. sakazakii, C. malonaticus and C. turicensis have been associated with reported neonatal infections, whereas C. malonaticus ST7 strains were associated with adult infections [19]. Therefore pathogenicity in humans may be an acquired trait in this genus. cerebrospinal fluid, blood, bone marrow, sputum, urine, and faeces [3]. The organism is an opportunistic pathogen of humans that can cause infections in all age groups [4]. However, low birth weight neonates are most at risk and in this host group it has been associated with outbreaks of necrotizing enterocolitis, meningitis, and septicaemia [3,5,6,7,8,9]. Infections with these presentations result in exceptionally high mortality rates ranging from 40 to 80 percent [10]. In recent years, several outbreaks of bacterial infection in neonatal intensive care units (NICU) have been traced to powdered formula contaminated with Cronobacter spp. [11,12]. Cronobacter was defined as ‘yellow-pigmented Enterobacter cloacae’ until 1980 when it was designated Enterobacter sakazakii by Farmer et al [13]. Later analyses of both partial 16S rDNA and hsp60 sequences showed that E. sakazakii isolates formed at least four distinct clusters, and it was proposed that these clusters could be unique species [14]. This was confirmed using DNA-DNA hybridization and phenotyping, and subsequently Enterobacter sakazakii was re-classified as the new genus Cronobacter [15]. Initially the genus was composed of only four species, but this has been revised and it currently contains the seven species: Cronobacter sakazakii, C. malonaticus, C. turicensis, C. muytjensii, C. dublinensis, C. universalis and C. condimenti [16]. Kucerova et al. [3,29] used whole genome sequence analysis of C. sakazakii strain BAA-894 and microarray based comparative genomic hybridisation (CGH) to explore the genetic basis of virulence in Cronobacter. CGH highlighted 15 clusters of genes in C. sakazakii BAA-894 that were divergent or absent in more than half of the tested strains; six of these were of probable prophage origin. Several putative virulence factors (i.e. fimbriae and multidrug efflux systems) were identified in these variable regions. Comparative Analysis of Cronobacter Genomes Here we present a comparative genomic sequence analysis of eleven recently sequenced genomes from the seven Cronobacter species fully representing the genus, and including the recently recognised species C. universalis and C. condimenti [16]. These strains were chosen as representatives of the Cronobacter genus based on seven locus MLSA [17,18]. The three C. sakazakii strains (680, 696 and 701) are well characterised clinical isolates previously reported in eight publications [3,8,17,18,19,20,29,30]. In addition, strain 701 is a C. sakazakii ST4 strain that was isolated from a fatal case of neonatal meningitis and has been used for in vitro tissue culture studies of virulence [8,26]. C. sakazakii 680 is ST8, a sequence type associated with clinical infections without being linked to infant formula consumption [17,19]. C. malonaticus 681T is ST7, a sequence type associated with adult infections. According to MLSA, C. condimenti is distantly related to the other Cronobacter species [18]. Completed genomes of C. sakazakii BAA-894 and C. turicensis z3032 were already in the public domain [29,32] and were used to accurately match reads to reference sequences. The incomplete draft genome of C. sakazakii E899 [32] became available during the completion of this study and was used as a comparator where appropriate. Thus, a total of 14 genomes were used in this comparative study. A multilocus sequence typing (MLST) scheme has been established for the entire Cronobacter genus and is available online at http://www.pubMLST.org/cronobacter [16,17,18]. The scheme is based on seven housekeeping genes (atpD, fusA, glnS, gltB, gyrB, infB, ppsA) with a concatenated length of 3036 nucleotides that can be used for multilocus sequence analysis (MLSA). Phylogenetic analysis based on this set of sequences estimated that the separate Cronobacter species evolved in the past 40 million years, with C. sakazakii and C. malonaticus emerging as definable species 11–23 million years ago [18]. The Cronobacter MLST scheme has been applied to more than 350 strains that were widely distributed geographically, temporal- ly, and by source, some of which could be traced over a 50-year period. There are currently 121 defined sequence types (ST) covering all Cronobacter species, some of which are stable clones [18,19]. Certain associations have been noted between sequence type and source. For example, C. sakazakii ST1 strains are primarily isolates from infant formula and clinical sources, whereas C. sakazakii ST8 is primarily composed of isolates from clinical sources. Of special significance is C. Comparative Analysis of Cronobacter Genomes sakazakii ST4 which appears to have a high propensity for neonatal meningitis [18,19]. This appears to be a very stable clone as clinical and non-clinical strains have been isolated from 7 countries for over 50 years. In addition, C. malonaticus ST7 is associated with adult infections though the source has not been identified. The aim of this study was to determine (i) the core Cronobacter genome, (ii) specific regions related to physiological and virulence related traits, and (iii) unique regions in C. sakazakii ST4, associated with neonatal meningitis [19,30], and in C. malonaticus ST7 associated with adult infections. Due to the severity of infant infection, a better understanding of the genomic variation between Cronobacter spp. is needed, and will be of interest to manufacturers of powdered infant formula, regulatory bodies, and those studying the evolution and diversity of bacterial pathogenicity. To date, little is known about the mechanisms of pathogenicity in Cronobacter. Candidate virulence determinants include superox- ide dismutase (SodA) for macrophage survival [20], haemolysin [21], flagella [22], a metalloprotease [23], an enterotoxin [24], and plasmid-borne virulence factors such as Cronobacter plasminogen activator (Cpa) and type six secretion systems (T6SS) [25]. The bacteria can attach to intestinal cells and survive in macrophages [20,26]. OmpA and OmpX possibly have a role in the organism penetrating the blood-brain barrier, though the mechanism leading to the destruction of the brain cells is unknown and could, in part, be a host response [27]. However, only a few studies have described the interaction of Cronobacter spp. with human cells, and the specific receptors involved remain to be determined. It is known that the disruption of tight junctions significantly enhances association of C. sakazakii with the human intestinal cell line Caco- 2 [28]. Strains from C. sakazakii and C. malonaticus show higher invasion of Caco-2 than other Cronobacter species [20]. Similarly, Introduction The Cronobacter genus (formerly Enterobacter sakazakii) is composed of Gram-negative, motile, non-sporeforming, peritrichous rods within the Enterobacteriaceae family. This family includes the well- known enteric bacterial pathogens E. coli and Salmonella, though November 2012 \| Volume 7 \| Issue 11 \| e49455 PLOS ONE \| www.plosone.org November 2012 \| Volume 7 \| Issue 11 \| e49455 1 Comparative Analysis of Cronobacter Genomes Comparative Analysis of Cronobacter Genomes revealed that pair-wise DNA sequence identity varies between 89 and 97%. A BRIG alignment for all the 14 Cronobacter genomes using the C. sakazakii BAA-894 genome as a backbone is shown in Figure S1. C. turicensis directly evolved from a common ancestor, while the ML tree of core SNPs (Figure 1A) predicted that they evolved in a step-wise fashion. This correlates well with the phylogenetic analysis of the seven loci used in the Cronobacter MLST scheme [18]. Co-incidentally, among all the speciating nodes in the consensus tree of 1117 core genes, the node where C. universalis speciates has the most incongruence. Hence, based on these two observations, it can be inferred that after speciation there might have been a large scale recombination event between ancestral C. universalis and hitherto unknown source. Clonal Frame predicted that for the observed phylogeny the ratio of substitutions introduced by recombination as compared to mutation (r/m) was 0.32, while ratio of number of recombination events to mutation events (r=h) was 0.01. C. turicensis directly evolved from a common ancestor, while the ML tree of core SNPs (Figure 1A) predicted that they evolved in a step-wise fashion. This correlates well with the phylogenetic analysis of the seven loci used in the Cronobacter MLST scheme [18]. Co-incidentally, among all the speciating nodes in the consensus tree of 1117 core genes, the node where C. universalis speciates has the most incongruence. Hence, based on these two observations, it can be inferred that after speciation there might have been a large scale recombination event between ancestral C. universalis and hitherto unknown source. Clonal Frame predicted that for the observed phylogeny the ratio of substitutions introduced by recombination as compared to mutation (r/m) was 0.32, while ratio of number of recombination events to mutation events (r=h) was 0.01. Phylogenetic and Evolutionary Analysis of the Cronobacter Genus Figure 1A represents a maximum likelihood (ML) phylogram constructed based on the 138,143 SNPs identified from a concatenated alignment of 1117 core genes (879,768 nucleotides) present in single copies in all strains. Figure 1B represents a majority rule consensus tree generated by summarizing all the individual gene trees. The topology of speciation for each Cronobacter species was similar, as predicted by the two different approaches. There were contradictions in resolving the nodes within C. sakazakii indicating that recombination events might have obscured the phylogenetic signals within individual gene trees. Within C. sakazakii there was a higher level of confidence in placing C. sakazakii 701 and C. sakazakii E899 within the same clade, indicating that these two strains might have evolved more recently from a common ancestor. MLST revealed these two strains are both ST4 (see Figure 2 for MLSA phylogenetic analysis of the fourteen sequenced). Figure 1C represents the results of Clonal Frame analysis [33] of 99 randomly chosen core loci. Clonal Frame tries to recreate the evolutionary descent of each sequence by taking into account the recombination events that may have occurred during its evolutionary history. Clonal Frame also compares the recombination and mutation rates in assessing the impact of recombination on the evolution of observed phylogeny. Consensus tree (Figure 1C) obtained from two Clonal Frame runs with 200,000 MCMC iterations had a topology similar to the ML tree obtained using core genome SNPs, with one important difference. Clonal Frame analysis predicted that C. universalis and Comparisons between the evolutionary rates of gene families did not reveal a clear variation between the species of the Cronobacter genus. However, an enrichment analysis performed to assess the statistical over-representation of relatively fast evolving genes within certain functional categories revealed that genes related to iron acquisition, copper homeostasis, phage tail proteins, and genes related to O-antigen and fimbriae evolved faster as compared to the rest of the Cronobacter pan genome. Cronobacter Genomes Cronobacter Genomes Using the SOLiDH 4 System and Ion Torrent PGMTM next generation sequencing platforms, we sequenced the genomes of eleven Cronobacter strains including three C. sakazakii, two C. malonaticus, one C. muytjensii, one C. turicensis, two C. dublinensis, one C. universalis, and one C. condimenti (Table 1). The sum of the contig lengths ranged from 4.4 to 4.9 Mb, in length, comparable to the genomes sizes of the previously sequenced C. sakazakii BAA-894 and C. turicensis z3032 genomes [29,31]. The publicly available, incomplete genome of C. sakazakii E899 is only 3.96 Mb and appeared to lack plasmid sequences [32]. Genome comparison November 2012 \| Volume 7 \| Issue 11 \| e49455 PLOS ONE \| www.plosone.org 2 Comparative Analysis of Cronobacter Genomes Figure 1. Phylogentic analysis of Cronobacter spp. based on whole genome sequencing. (A) A maximum likelihood phylogram based on 138,143 SNPs identified from a concatenated alignment of 1117 core genes (879,768 nucleotides) present in single copies in all strains. The numbers on the nodes represent the bootstrap support from 1000 replicates (in percentage). (B) A majority rule consensus tree generated by summarizing individual 1117 core gene trees. The numbers on the internal nodes indicate the fraction (in percentage) of gene trees which support the partition of the taxa into the two sets. (C) A Clonal Frame (CF) analysis based on randomly selected 99 core loci. The consensus tree was obtained from 2 CF runs with 200,000 MCMC iterations. The numbers on the nodes represent the bootstrap support in percentage. doi:10.1371/journal.pone.0049455.g001 plasmids, pESA3 and pCTU1, share a common backbone and genes encoding for potential virulence traits such as iron acquisition systems (eitCBAD and iucABCD/iutA). The differences include the presence of plasminogen activator gene (cpa) and T6SS in the pESA3 plasmid, whereas in pCTU1 there is a 27 kb region with homologies to fhaB, fhaC and five adhesins (FHA locus) [25]. The cpa gene was only found in C. sakazakii strains 696 and 701, and was absent from 680. It was also absent from all other Cronobacter species with the exception of C. universalis. The BRIG alignments with pESA3 and pCTU1 as reference backbones are shown in Figures S2 and S3, respectively. It can be seen that the plasmid backbone was conserved in all the strains with varying degrees of homology. There were also clear regions of variation between the pESA3 and pCTU1 plasmids, some of which have been previously reported. However, it should be noted that matches to the plasmid do not necessarily confirm their location. To address this issue the read coverage was used to predict the location of the plasmid-borne genes. Genes homologous with those Using the binomial mixture model approach of Snipen et al. [36], the observed pan genome was estimated to contains 6,155 genes, of which 2,271 (36.9%) genes were conserved in all fourteen Cronobacter genomes, an estimate of the core genome, and 3,884 (63.1%) were accessory genes found in at least one of the fourteen genomes studied. Because the draft genomes can have up to thousands of sequence gaps, some genes may have incomplete fragments or were unannotated. Core and Pan Genome Identification A pan genome analysis of the fourteen genomes of the Cronobacter genus was used to study the diversity of the genomes, to identify genes of phenotypic and pathogenicity interest and those unique to each species. Putative prophage regions, identified using Prophinder [34], were included in the analysis. Some of the characteristics of the genomes of each of the species are discussed below. Additional detail is given in Text S1. All genomes were annotated using RAST [35]. The number of annotated genes per genome varied between 3,700 and 4,200. Table 1. Eleven newly assembled and three publicly available Cronobacter genomes analyzed in this study. Species Strain Source Country Genome size (Mbp)d MLST Sequence Type EBI Accession Number C. sakazakii 680 Clinical USA 4.36 8 CALG01000001-CALG01000201 696 Clinical France 4.90 12 CALF01000001-CALF01000569 701 Clinical France 4.75 4 CALE01000001-CALE01000768 E899a Clinical USA 3.96 4 AFMO00000000 BAA-894a Powdered formula USA 4.53 1 NC_009778-NC_009780 C. malonaticus 507 Clinical Czech Republic 4.45 11 CALD01000001-CALD01000249 681b Clinical USA 4.50 7 CALC01000001-CALC01000171 C. turicensis 564 Clinical USA 4.50 5 CALB01000001-CALB01000114 z3032a,c Clinical Switzerland 4.60 19 NC_013282-NC_013285 C. dublinensis 582 Unknown UK 4.68 36 CALA01000001-CALA01000427 1210d Environment Ireland 4.59 106 CAKZ01000001-CAKZ01000221 C. muytjensii 530 Infant formula Denmark 4.53 49 CAKY01000001-CAKY01000365 C. universalis NCTC 9529T Environmental UK 4.45 54 CAKX01000001-CAKX01000231 C. condimenti 1330e Food Slovakia 4.46 98 CAKW01000001-CAKW01000155 aStrains for which genome sequence is publicly available; E899 appears to lack plasmid sequences. bC. malonaticus species type strain (LMG 23826T). cC. turicensis species type strain (LMG 23827T). dC. dublinensis species type strain (LMG 23823T). eC. condimenti species type strain (LMG 26250T). fSizes of newly sequenced genomes were derived from the sum of the length of all contigs from de novo assembly. doi:10.1371/journal.pone.0049455.t001 Table 1. Eleven newly assembled and three publicly available Cronobacter genomes analyzed in this study. November 2012 \| Volume 7 \| Issue 11 \| e49455 PLOS ONE \| www.plosone.org 3 Comparative Analysis of Cronobacter Genomes Comparative Analysis of Cronobacter Genomes PLOS ONE \| www.plosone.org 4 November 2012 \| Volume 7 \| Issue 11 \| e49455 November 2012 \| Volume 7 \| Issue 11 \| e49455 PLOS ONE \| www.plosone.org Comparative Analysis of Cronobacter Genomes Comparative Analysis of Cronobacter Genomes Physiological and Phenotypic Traits in the Cronobacter Core and Pan Genome [29] detailed three prophages and 3 partial prophages in C. sakazakii BAA-894 contributing to the genome diversity; none of these prophages were previously reported in C. turicensis z3032 [31]. In our study of the whole genus we found a total of 27 major prophage regions (.20 ORFs) across the 13 genomes, most of which were found to be shared between the Cronobacter species (Table S1). In addition, we found a clustered regularly interspaced short palindromic repeats (CRISPR) array, which was located in the region ESA_02830- 38. CRISPR arrays consist of a series of highly conserved repeats ,20–50 bp separated by unique DNA fragments (spacers) of a similar length. It is understood that the CRISPR system provides a mechanism of resistance to infection by phage to which the bacteria have previously been exposed. Various CRISPR-associ- ated (cas) genes and specific adjacent sequences are required to trigger these processes [37]. This specific CRISPR region was present in all Cronobacter strains except C. sakazakii 680 (ST8). Comparison of the three C. sakazakii genomes revealed 408 ORFs, which were unique within this species. A number of regions were found to be unique to the genomes of strains 696 (ST12) and 701 (ST4). The C. sakazakii 680 (ST8) genome had unique genes encoding a region comprising iron uptake (fecRABCDE), a lac operon, and arsenate resistance on the larger plasmid. It also had large regions missing compared with the reference genome C. sakazakii BAA-894, including flagella synthesis (ESA_01224- 69) and CRISPR (ESA_02831-37). The two C. malonaticus genomes revealed 92 ORFs unique to the species. Only about 15% of these ORFs were conserved between the two genomes. The C. malonaticus 681T (ST7) unique ORFs included mainly phage related proteins and two genes related to capsular polysaccharide export proteins. The C. turicensis 564 genome was analysed along with the publicly available C. turicensis z3032 genome to reveal 137 ORFs unique to the species, of which less than 10% were shared between the two genomes. Comparative analysis of C. dublinensis genomes revealed 142 ORFs unique to the species, of which less than 20% were shared between the two genomes. The C. dublinensis 582 genome also showed the presence of two additional clusters of CRISPR (938–943 and 3962–68). Physiological and Phenotypic Traits in the Cronobacter Core and Pan Genome The presence of known genes conferring physiological and phenotypic traits, categorized according to environmental stress response, cell surface composition, sugar metabolism, and metal resistance was examined. Figure 3. Graph representing the rarefaction of the Cronobacter pan genome, generated using the 14 Cronobacter genomes included in this study. doi:10.1371/journal.pone.0049455.g003 A number of key stress response genes were investigated. A gene homologous to the universal stress protein uspA (ESA_01955) was detected in all strains. Similarly, homologues were located for stringent starvation response (ESA_03615), carbon starvation sensing protein rspA (ESA_01752), and carbon starvation protein (ESA_00801) in all Cronobacter strains. However, an additional carbon starvation homologue (ESA_00339) was found in all Cronobacter strains except C. sakazakii 680 due to the long missing region mentioned above. Cronobacter is known for its ability to survive desiccation for up to two years [38,39]. Genes involved in desiccation resistance and osmotic stress adaption were found in all Cronobacter species. This included genes encoding the uptake of the osmoprotectants glycine, betaine, and trehalose (ESA_00587- 9, ESA_01944-66), which were present in all Cronobacter. Thermal tolerance in Cronobacter has been controversial due to conflicting reports [40–42]. The initiation translation factor (infB) proposed by Asakura et al. [41] was present in all Cronobacter, and is one of the alleles used for MLST. In contrast, the thermoresistance gene cluster (or most components) identified by [42] was only present in C. sakazakii strains 696, 701, and in both strains of C. malonaticus. Unfortunately, this region is poorly annotated and the association with thermal resistance is not supported by experimental data in these strains. Together with biofilm formation, the ability to adapt and persist under stressed conditions enables the bacterium to survive in dry food ingredients. All Cronobacter species had genes (crtZ; ESA_00341-4) for b-carotene production, a yellow pigmen- tation, which we speculate helps protect the organism from light produced by oxygen radicals. in pESA3 and pCTU1 were found in all sequenced strains of all species, though pESA3 was only partial in C. sakazakii 680 and C. condimenti 1330T, and pCTU1 had a partial match in C. sakazakii 680, 696, 701, and C. condimenti 1330T. in pESA3 and pCTU1 were found in all sequenced strains of all species, though pESA3 was only partial in C. sakazakii 680 and C. condimenti 1330T, and pCTU1 had a partial match in C. sakazakii 680, 696, 701, and C. condimenti 1330T. Previously, Kucerova et al. Comparative Analysis of Cronobacter Genomes Comparative Analysis of Cronobacter Genomes Figure 3. Graph representing the rarefaction of the Cronobacter pan genome, generated using the 14 Cronobacter genomes included in this study. doi:10.1371/journal.pone.0049455.g003 Cronobacter species. The C. universalis genome had 73 unique ORFs. This included a region of 36 ORFs, which appears to be a plasmid remnant encoding for conjugation (trb genes), replication initiation, and plasmid stabilization. There were 40 ORFs found to be unique to the genome of C. muytjensii strain 530. The genome of the newly defined species C. condimenti strain 1330T contained 79 unique ORFs. These included genes related to b-xylosidase and a cluster of eight genes that are indicative of a unique fucose biosynthesis pathway. Comparative Analysis of Cronobacter Genomes This could cause a slight under- estimation of the size of the core genome and a slight over- estimation of the size of the pan genome. A graph representing the rarefaction of the Cronobacter pan genome has been generated and is shown in Figure 3. The published genome of C. sakazakii BAA-894 contains two plasmids pESA2 (31 kb, 51% GC content) and pESA3 (131 kb, 56% GC), encoding for 38 and 127 genes respectively. In contrast, C. turciensis z3032 contains three plasmids pCTU2 (22.5 kb, 49%GC), pCTU3 (53.8 kb, 50% GC) and pCTU1 (138 kb, 56% GC), encoding for 32, 74 and 136 genes respectively. The larger Figure 2. Maximum Likelihood tree based on the concatenated sequences (3,036 bp) of the seven MLST loci. The tree is drawn to scale using MEGA, with 1000 bootstrap replicates. doi:10.1371/journal.pone.0049455.g002 Figure 2. Maximum Likelihood tree based on the concatenated sequences (3,036 bp) of the seven MLST loci. The tree is drawn to scale using MEGA, with 1000 bootstrap replicates. doi:10.1371/journal.pone.0049455.g002 hood tree based on the concatenated sequences (3,036 bp) of the seven MLST loci. The tree is drawn to scal tstrap replicates. 9455.g002 November 2012 \| Volume 7 \| Issue 11 \| e49455 PLOS ONE \| www.plosone.org 5 Comparative Analysis of Cronobacter Genomes Homologues of tellurite resistance genes (terACDYZ) were located in the loci ESA_01775–ESA_01804 of C. sakazakii BAA- 894. As the gene cluster was absent from all other Cronobacter strains studied, the reference strain BAA-894 or the ST1 lineage probably acquired the tellurite resistance cluster recently. Homol- ogies of these genes are also found on the IncII plasmid R478 of Serratia marcescens, pK29 of Klebsiella pneumoniae NK29, pEC-IMPQ of Enterobacter cloacae, and pAPEC-O1-R of E. coli APEC O1 [57]. In BAA-894 this cluster of genes contains an IS element (ISEhe3) just before ESA_01781 which was also identified on the pK29 plasmid as well as a fragment of an IS4 family transposase within the locus ESA_01803. antigen locus are highly divergent and were not detected previously by microarray hybridization [29]. The lipopolysaccha- ride (LPS) is one of the few structural features of Cronobacter which has been chemically investigated and is known to vary across the genus. In C. sakazakii and C. malonaticus the LPS are composed of various branched polymers, whereas they are unbranched in C. muytjensii [45–48]. Variation in the O-antigen region has been used in serotype- specific PCR-based assays [49–53]. The regions between UDP- glucose pyrophosporylase (galF) and 6-phosphogluconate (gnd) were amplified to generate distinguishable O-antigen groups. The genetic architecture of the O-antigen cluster in the reference genome C. sakazakii BAA-894 (ST1) corresponds to the serotype O:1 and both the ST4 strains 767 and NCTC 8155 are O:2. However, the current PCR-based serotyping assay has a significant limitation since some defined serotypes are found in more than one species [51–53]. Copper is essential for bacterial growth, being required for many key enzymes, but is toxic in excess. Two copper and silver resistance gene clusters are located close to each other (ESA_04238-45 and ESA_04248–55) in the C. sakazakii BAA- 894 annotated chromosome and are on the smallest plasmid (pCTU3) in C. turicensis z3032 (pCTU3_3p00600-700 and pCTU3_3p00490-590). The first region (cusRCFBA/silRECBA) was present in all C. sakazakii strains, in one C. malonaticus (681) strain, and one C. turicensis (z3032) strain. CusA is a membrane protein belonging to the resistance-nodulation-division (RND) protein superfamily. CusF is a periplasmic protein, which in E. coli interacts with the cusCBA efflux system. The second region (pcoABCDR) was present in nearly all C. sakazakii strains (except 701), in one C. malonaticus strain (681), one C. turicensis strain (z3032), and in C. universalis 581T. Virulence Associated Genes Putative virulence traits in Cronobacter include (i) attachment and invasion of host cells, (ii) iron acquisition, (iii) macrophage survival, and (iv) protein secretion systems. Since genes encoding for SodA, OmpA, OmpX, and a metalloprotease (zpx) were found in all genomes irrespective of source and species, they could not be connected to particular virulence-associated strains such as C. sakazakii ST4 and therefore to neonatal meningitis. Cronobacter are ubiquitous in the environment, and their main habitat is believed to be plant material. Cronobacter are able to ferment b-glucoside sugar substrates from plants, including cellobiose, arbutin, salicin, and esculin. The genes involved in the metabolism of b-glucosides are in the cluster ESA_02544-47. ESA_02544 is probably a transcriptional antiterminator from the BglG family, which is involved in positive control of the utilization of different sugars by transcription antitermination [56]. ESA_02545 encodes a kinase which converts b-glucosides to 6- phospho-b-glucosides and ESA_02546 encodes a 6-phospho-b- glucosidase specific to arbutin-6-phosphate and salicilin-6-phos- phate. ESA_02547 also encodes 6-phospho-b-glucosidase, which may have the same or similar function as ESA_02546. Malonate is also found in plants and C. malonaticus was initially a subspecies within C. sakazakii that was differentiated from C. sakazakii subsp. sakazakii by the malonate utilisation test [15]. However, the genes required for malonate metabolism - malonyl CoA acyl carrier protein transacylase, phosphoribosyl-dephospho-CoA transferase, and malonate decarboxylase (Ctu_34990-35070) - are found in all Cronobacter species except C. sakazakii and C. dublinensis strain 582. Therefore this test cannot be used solely for the purpose of identifying C. malonaticus. Cronobacter spp. route of infection is probably through attach- ment and invasion of the intestinal cells and therefore genes encoding surface appendages such as fimbriae (pili) are of importance. The genetic content of all fimbriae clusters was most similar to the type I chaperone/usher-assembled pilus system [58]. These clusters may encode complete and functional pili, as some degree of homology was found between the genes in the C. sakazakii fimbriae clusters and the remaining components necessary for type-I pilus assembly (the minor tip fibrillum fimG and fimbrial adhesin fimH genes). Type 1 fimbriae have been associated with E. coli K1 invasion of human brain cells [59] and are therefore of particular interest in Cronobacter pathogenicity studies. A total of ten putative fimbriae gene clusters were identified and are summarised in Table 2. Curli fimbriae are believed to play an important role in the adhesion of E. Comparative Analysis of Cronobacter Genomes PcoA encodes a homologue to multicopper oxidase, and PcoB is an outer membrane protein transfering copper into the periplasm where it is oxidised by PcoA. Both these regions are encoded on the plasmids R478 and APEC- O1-R, referred to above, which carry tellurite resistance genes [57]. The copper and silver resistance genes on the reference C. sakazkaii BAA-894 genome are separate from the tellurite resistance genes. Therefore these regions do not appear to be the result of a single plasmid integration event. It is notable that C. sakazakii 701 (ST4) also encoded an O- antigen acetylase (O-acetyl transferase; 28141.13.peg.3410) adja- cent to a prophage region which, except for fragments in C. sakazakii 696 (ST12), was not present in other C. sakazakii strains. O-antigen modification enzymes are encoded by temperate bacteriophages [54]. The possibility of this in Cronobacter was proposed by Sun et al. [51] when comparing RFLP patterns from PCR amplification of the O-antigen genes of 119 C. sakazakii strains. Seroconversion is an important virulence factor since antigenic variation enhances the survival of the bacteria because the host would have to mount a specific immune response to each different serotype. Serotype-converting bacteriophages play an important role in conferring these traits, and recognising their presence in C. sakazakii is important due to the application of the serotyping scheme. The gene cluster encoding for maltose utilisation and including a–glucosidase (ESA_02709-14) was found in all Cronobacter strains. The presence of a-glucosidase and a constitutive maltose uptake system, as well as osmotic pressure tolerance, has been used as the basis for a number of commercially available enrichment broths and chromogenic agars for the isolation of Cronobacter spp. [55]. Other sugar utilization pathways were only found in some Cronobacter species. A gene cluster for a-mannosidase utilisation (ESA_02616-18) was only found in C. sakazakii and C. turicensis, and not in the closely related C. malonaticus. Physiological and Phenotypic Traits in the Cronobacter Core and Pan Genome Only single strains were sequenced for each of the remaining Capsular polysaccharides on the bacterial cell surface can play a crucial role in the interaction of bacteria with their environment and in their pathogenicity. The genes encoding the enzymes responsible for the production and transport of these polysaccha- rides are usually clustered in large operons, such as the wca operon in E. coli K-12. This gene cluster controls the biosynthesis of the exopolysaccharide colanic acid containing fucose and glucuronic acid. Cronobacter spp. can produce copious extracellular polymers, resulting in highly mucoid colony formation and can form biofilms on inert surfaces including infant feeding equipment and nasogastric feeding tubes [8,43,44]. Colanic acid production was encoded (ESA_01155-01175; wzABCKM) in all sequenced Crono- bacter strains. The genes for capsular polysaccharide assembly and export (ESA_03350-59) were present in all Cronobacter strains except for two clinical isolates, C. sakazakii 701 (ST4) and 696 (ST12). One highly variable genomic region (ESA_01179–89) corre- sponds to the O-antigen gene locus. The locus contains two genes galF (ESA_01177) and rfbB (ESA_01178) which are conserved in all the Cronobacter strains, whereas the rest of the genes from the O- November 2012 \| Volume 7 \| Issue 11 \| e49455 November 2012 \| Volume 7 \| Issue 11 \| e49455 PLOS ONE \| www.plosone.org 6 November 2012 \| Volume 7 \| Issue 11 \| e49455 November 2012 \| Volume 7 \| Issue 11 \| e49455 Virulence Associated Genes The nanK gene (ESA_03610) encodes N-acetylmannosamine kinase; nanT (ESA_03611) encodes the sialic acid transporter N-acetylneur- aminate lyase; nanA (ESA_03612) encodes N-acetylneuraminate lyase, and nanR (ESA_03613) is a transcriptional regulator from the GntR family. The nanE locus encoding the enzyme N- acetylmannosamine-6-phosphate-2epimerase was located separate from this cluster at ESA_00529, and unlike the rest was found conserved across the genome of the Cronobacter genus. NanC (ESA_03302) encodes the N-acetylneuraminic acid outer mem- brane channel protein). It is notable that the nan cluster is adjacent to a stringent starvation gene homologue (sspA, ESA_03615) and therefore expression of the nan cluster could be responsive to environmental nutrient levels. The acquisition of genes encoding for the utilization of exogenous sialic acid may have a major role in C. sakazakii colonisation of the human intestinal tract (via mucins) and the use of sialic acid in breast milk, infant formula, and brain cells as a nutrient source [62]. Iron is an essential growth factor and bacteria have evolved mechanisms, such as siderophores, for its acquisition. Nearly all Cronobacter strains examined possess complete operons for the production of enterobactin synthesis (entABCDEFS, ESA_00791- 800), receptor, and transport (fepABCDEG; ESA_02727-31). The exception were C. sakazakii strains 701 and 696, which lacked the homologue for EntS, an enterobacterin exporter. All Cronobacter species possess a complete plasmid-borne operon for aerobactin synthesis (iucABCD) and its receptor iutA (ESA_pESA3p05547-51). Additionally, a gene cluster for hydroxamate-type siderophores (fhuABCDE, ESA_03187-90 & ESA_02242) was encoded in all Cronobacter species. As previously referred to, C. sakazakii 680 encoded a unique siderophore (fecRABCDE) which is probably plasmid-borne. According to Franco et al. [25], the only functional siderophores are those encoded on the larger plasmid. Iron uptake is often cited as a virulence mechanism, especially with respect to plausible utilisation by Cronobacter of iron in breast milk and infant formula. However, drawing further conclusions will require further work because there are a number of probable iron uptake mechanisms across the genus, including those in species not associated with neonatal infections: C. muytjensii, C. malonaticus, C. dublinensis, C. universalis, and C. condimenti. Despite detailed genomic analysis, the reason for the association of C. sakazakii ST4 with neonatal meningitis remains unclear. Initially Joseph & Forsythe [19] proposed that it could be either environmental persistence, resulting in greater neonatal exposure, or the presence of specific pathogenicity traits. Virulence Associated Genes coli to host cells by interacting with matrix proteins such as fibronectin, laminin, and plasminogen to initiate adherence and colonization. However, since C. sakazakii strains dominate clinical isolates, the absence of curli fimbriae genes infers this trait is not essential for Cronobacter pathogenicity. November 2012 \| Volume 7 \| Issue 11 \| e49455 November 2012 \| Volume 7 \| Issue 11 \| e49455 PLOS ONE \| www.plosone.org 7 Comparative Analysis of Cronobacter Genomes Table 2. Cronobacter fimbriae cluster designations. Cluster Loci C. sakazakii C. malonaticus C. muytjensii C. turicensis C. dublinensis C. universalis C. condimenti Comment BAA-894 680 696 701 507 681 530 564 582 1210 NCTC 9529T 1330 1 ESA_01970-76 + + 2 2 2 2 2 2 2 2 2 2 Also found in C. sakazakii strain2 [Kucerova et al. 2010; 35] 2 ESA_02342-45 + + + + + + + + + + + + Present in all Cronobacter 3 ESA_02538-42 + + + + + + + + + + + + ESA_02541&42 not found in C. muytjensii, C. dublinensis, C. condimenti 4 ESA_02795-99 + + + + + + 2 2 2 2 2 2 Only found in C. sakazakii and C. malonaticus 5 ESA_03512-20 + + + + 2 2 2 2 2 2 2 2 Beta-fimbriae. Only found in C. sakazakii 6 ESA_04067-73 + + + + + + + + + + + + Present in all Cronobacter 7 ESA_03812-15 + + + + + + 2 + + + + + Absent from C. muytjensii 8 Ctu_36390-450 2 + + + + + + + + + + + p–fimbriae. Absent from C. sakazakii BAA-894 9 ESA_03231-33 + + + + + + + + + + + + Type IV. Gliding motility 10 Ctu-16160-230 2 2 2 2 + + + + + + + + Curli fimbriae. Absent from C. sakazakii PLOS ONE \| www.plosone.org 8 November 2012 \| Volume 7 \| Issue 11 \| e49455 November 2012 \| Volume 7 \| Issue 11 \| e49455 PLOS ONE \| www.plosone.org 8 Comparative Analysis of Cronobacter Genomes Why the region has been lost in the C. sakazakii lineage and retained in the rest of the genus is unclear. and to pathogens [62]. In C. sakazakii ESA_03609 encodes a putative sugar isomerase (YhcH). Genes ESA_03610-13 encode the nanKTAR genes involved in the N-acetylneuraminate and N- acetylmannosamine degradation pathway. Virulence Associated Genes It was present in all Cronobacter strains except for one strain of C. malonaticus (507) and one strain of C. muytjensii (530). Six putative T6SS clusters were identified in the Cronobacter genomes (Table S2), some of which were previously described by Kucerova et al. [3,29]. These are putative T6SS clusters as it remains to be determined whether they encode functional secretion systems. Zhou et al. [61] recently reported the role of T6SS in E. coli K1 invasion of the human blood-brain barrier. Although several T6SS are found in all Cronobacter species, none were unique to C. sakazakii ST4 strain 701. All sequenced Cronobacter genomes exhibited the presence of various hemolysin and hemolysin-related genes, scattered across the genome (Table S3). These genes were present in all the genomes with the only exceptions being C. sakazakii 701 and C. malonaticus 507, which lacked the 21 kDa hemolysin precursor gene. Most of the strains had two copies of the hemolysin gene and the hemolysin activator protein precursor gene. O f h i i l i C k kii Type VI secretion system (T6SS) is a newly described system that may be involved in competing with other bacteria in adherence, cytotoxicity, host-cell invasion, growth inside macro- phages, and survival within the host. One T6SS associated gene that was located separately from the main T6SS gene cluster, was vgrG, encoding a lipoprotein (ESA_00292-4). It was present in all Cronobacter strains except for one strain of C. malonaticus (507) and one strain of C. muytjensii (530). Six putative T6SS clusters were identified in the Cronobacter genomes (Table S2), some of which were previously described by Kucerova et al. [3,29]. These are putative T6SS clusters as it remains to be determined whether they encode functional secretion systems. Zhou et al. [61] recently reported the role of T6SS in E. coli K1 invasion of the human blood-brain barrier. Although several T6SS are found in all Cronobacter species, none were unique to C. sakazakii ST4 strain 701. Strains and Culture Conditions All sequenced Cronobacter genomes exhibited the presence of various hemolysin and hemolysin-related genes, scattered across the genome (Table S3). These genes were present in all the genomes with the only exceptions being C. sakazakii 701 and C. malonaticus 507, which lacked the 21 kDa hemolysin precursor gene. Most of the strains had two copies of the hemolysin gene and the hemolysin activator protein precursor gene. Cronobacter strains were selected to represent the seven recog- nized species, including those from reported clinical cases as well as species type strains (Table 1). All Cronobacter strains were stored at 280uC in Nutrient Broth (Oxoid, UK) with 10% glycerol, subcultured on Trypticase Soy Agar (Oxoid ThermoFisher, UK) and checked for purity. Overnight Trypticase Soy Broth (Oxoid ThermoFisher, UK) cultures were used for DNA extraction. One of the most interesting gene clusters, unique to C. sakazakii is ESA_03609–13 (nanKTAR) and ESA_03302 (nanC) which encode for the uptake and utilization of exogenous sialic acid, which is functional (unpublished laboratory studies). Sialic acid catabolism is limited to commensal bacteria in the intestinal tract Virulence Associated Genes Other than one putative plasmid-borne T6SS, this study has not shown any conclusive unique regions in the ST4 strain C. sakazakii 701, originally isolated from the peritoneal fluid of a fatal neonatal meningitis case. The secretion systems ancestrally related to the bacterial conjugation machinery are referred to as the type IV secretion systems (T4SSs). The T4SSs can transfer both proteins and nucleoprotein complexes and could constitute a conjugal transfer system [60]. A T4SS, unique to C. sakazakii BAA-894 and C. turicensis z3032, was located on the smaller plasmid (pESA2 31kB and pCTU2 22.5 kB) of both strains. A second cluster of 19 genes, belonging to the Tra group of genes and annotated as IncF conjugation plasmids, was found to be present in the C. sakazakii 701 and 696 genomes and is linked to the type IV protein secretion system. This could also be a plasmid-borne region as these genes have previously been identified in the Inc group of plasmids in E. coli and Serratia. In summary, the comparison of the draft genomes representing the Cronobacter genus has revealed the core genes and accessory genes unique to each strain. It was found that C. sakazakii has acquired the ability to use exogenous sialic acid, which may be important in the colonisation of the intestinal tract. Mobile traits such as adhesins, T6SS, and metal resistance as well as prophages have contributed considerably to the variation of genomic content and probably account for much of the variation in clinical and environmental distribution of species and sequence types. The lack of clearly identifiable virulence genes unique to C. sakazakii ST4 may indicate its prevalence in neonatal meningitis cases due to environmental persistence and increasing host exposure. Further improvement of the draft genome sequences, bioinformatics searches for novel or known virulence traits present in bacteria with similar modes of infection, and in silico and in vitro ascertainment of the contribution of single nucleotide polymor- phisms, genome rearrangements, and other sequence features to pathogenicity may shed light on acquired pathogenicity of ST4 strains. p Type VI secretion system (T6SS) is a newly described system that may be involved in competing with other bacteria in adherence, cytotoxicity, host-cell invasion, growth inside macro- phages, and survival within the host. One T6SS associated gene that was located separately from the main T6SS gene cluster, was vgrG, encoding a lipoprotein (ESA_00292-4). Genome Sequencing and Assembly Genome Sequencing and Assembly Genomic DNA was isolated using the Qiagen DNeasy Blood and Tissue DNA Isolation Kit according to manufacturer’s PLOS ONE \| www.plosone.org November 2012 \| Volume 7 \| Issue 11 \| e49455 9 Comparative Analysis of Cronobacter Genomes instructions. Cronobacter genomes of all strains except 680, 1210T and 1330T were sequenced using the SOLiDTM 4 system (Life Technologies, Carlsbad, CA). Long mate-paired genomic DNA libraries with approximately 1.8 kb inserts were sequenced to generate 23–36 million of 2650 bp reads for each strain, approximating 500–800 fold coverage of the genome. The colorspace reads were error-corrected and then assembled de novo into contigs and scaffolds using the Velvet assembly engine [63]. The ultimate genome assemblies contain 1,600 to 3,100 contigs with N50 of 3.7 to 5.5 kb and 260 to 1,170 scaffolds with N50 of 230 to 600 kb (Table 3). Species Strain Sequencing platform Total cont length (bp) N50 of scaffolds (bp) Number of scaffolds N50 of contigs (bp) Number of contigs Estimated number o ORFs C. sakazakii 680 PGM 4,357,873 nda nd 51,120 194 4,178 696 SOLiD 4,897,138 297,746 920 4,336 2,659 4,661 701 SOLiD 4,752,729 346,235 1,171 3,538 3,148 4,509 C. malonaticus 507 SOLiD 4,447,701 373,979 464 3,703 2,361 4,226 681 SOLiD 4,496,745 345,762 263 5,537 1,592 4,291 C. turicensis 564 SOLiD 4,500,608 411,105 263 4,796 1,807 4,227 C. dublinensis 582 SOLiD 4,677,592 229,230 539 3,822 2,657 4,483 1210 PGM 4,594,228 nd nd 46,941 210 4,376 C. muytjensii 530 SOLiD 4,533,101 596,924 444 4,925 1,937 4,304 C. universalis NCTC 9529T SOLiD 4,450,737 331,248 389 4,506 2,085 4,316 C. condimenti 1330 PGM 4,469,5362 nd nd 83,159 137 4,307 and indicates that scaffolding was not performed due to lack of mate-paired libraries. Strains 680, 1210T and 1330T were sequenced using Ion Torrent PGM system (Life Technologies). Fragment library preparation was performed with the Ion Fragment Library Kit (Life Technologies, Darmstadt, Germany). Template preparation was carried out with the Ion XpressTM Template Kit (Life Technologies). The Ion Sequencing Kit (Life Technologies) was used with the Personal Genome MachineTM (PGMTM) sequencer. A single sequencing run (65 cycles) was performed on an Ion 316TM chip for each library. Contigs were assembled from fragment reads using the MIRA 3 assembler [http://sourceforge. net/apps/mediawiki/mira-assembler/index.php]. The assembled genome scaffolds were aligned to the most closely related publicly available genomes using MUMmer [64]. Genome Sequencing and Assembly The scaffolds of strains 680, 696, 701, 507, and 681T were aligned to the C. sakazakii BAA-894 complete genome (accession numbers NC009778 - NC009780). The scaffolds of strains 564, 582, 1210T, 530, 581T and 1330T were aligned to the C. turicensis z3032 complete genome (accession numbers NC013282-NC013285). Scaffolds were broken at points where non-contiguous regions of the reference genome were juxtaposed and then re-ordered so that they were syntenic with the reference genome. All scaffolds from a given strain were concatenated into a single pseudogenome, separated by the sequence, NNNNCACACACTTAATTAAT- TAAGTGTGTGNNNNN, which contains stop codons in all six reading frames. Scaffolds that did not match the reference genomes were concatenated in random orders at the end of the genome. The pseudogenomes were annotated using the RAST automated annotation server [35]. and indicates that scaffolding was not performed due to lack of mate-paired libraries. doi:10.1371/journal.pone.0049455.t003 The genome sequences of the eleven newly sequenced strains were deposited to Genbank (see Table 1 for accession numbers). Core and Pan Genome Identification The genomes were compared using methods similar to those previously published [65]. Orthologous and paralogous gene families were initially constructed based on RAST annotations. An ‘all-against-all’ tblatx matrix was constructed using BLAT [66]. The blat matrix was used as an input to construct orthologous/ paralogous gene families using OrthoMCL [67]. A representative nucleotide sequence from each gene family was collated with strain-specific genes not present in any gene families. This set of representative sequences was again subjected to a tblatx analysis against the whole genome nucleotide sequences to identify candidate genes that may have been missed by the annotation server RAST. A gene was considered present if it had at least 85% nucleotide sequence identity covering at least 20% of the sequence length. The phyletic table generated from the tblatx analysis was consolidated with the phyletic table generated from OrthoMCL analysis to compute a comprehensive pan genome matrix. This phyletic matrix was used as an input for the binomial mixture model software of Snipen et al. [36] to determine the Cronobacter core genome and accessory genes. PLOS ONE \| www.plosone.org 10 Comparative Analysis of Cronobacter Genomes the genus Cronobacter using the ClustalW2 program [75] and MEGA (Molecular Evolutionary Genetics Analysis) software version 5 [76]. Genetic distances and clustering were determined using Kimura’s two-parameter model [77]; evolutionary trees were reconstructed by the Maximum Likelihood method [78]. Stability of the relationships was assessed by the bootstrapping method (1000 replicates). Prophage Identification Prophage Identification ACLAME Prophinder, available at http://aclame.ulb.ac.be/ Tools/Prophinder/, was used for prophage identification [34]. Prophage Identification ACLAME Prophinder, available at http://aclame.ulb.ac.be/ Tools/Prophinder/, was used for prophage identification [34]. Phylogenetic analysis. Core genome analysis revealed that there were 1117 annotated gene families conserved across all genomes in exactly one copy. Orthologous DNA sequences within each gene family were aligned using Muscle [68]. Individual alignments were concatenated using custom scripts, and a maximum likelihood (ML) tree was constructed using RaxML (Ver. 7.2.6) [69]. As a distinct alternative, a majority rule consensus tree of all gene family trees that were conserved across all genomes and were present in exactly one copy was constructed using Phylip (Ver 3.69) [70]. A concatenated alignment of 100 randomly chosen loci from the 1117 single copy core loci was used as an input for Clonal Frame [33]. Core and Pan Genome Identification Two independent runs, each consisting of 200,000 Markov chain – Monte Carlo (MCMC) iterations were performed, where the first 100,000 iterations were discarded as burn-in. The runs were compared for convergence using the Gelman- Rubin statistic [71], which was found to be satisfactory. Author Contributions MLSA was undertaken by using existing sequences in the multilocus sequence typing (MLST) database (http://www. pubMLST.org/cronobacter) and for new ones obtained in this study. Multilocus sequence typing of strains was performed as previously described [17]. The sequences obtained were indepen- dently aligned with sequences of the type strains of all species of Conceived and designed the experiments: YJ CC SJ SJF. Performed the experiments: SJ SH HS YJ AR. Analyzed the data: YJ PD LD PB RS CC SEH NM SJ SH HS AR NC MM SJF. Contributed reagents/materials/ analysis tools: MM MF. Wrote the paper: YJ CC MM SEH NC SJ SJF. Conceived and designed the experiments: YJ CC SJ SJF. Performed the experiments: SJ SH HS YJ AR. Analyzed the data: YJ PD LD PB RS CC SEH NM SJ SH HS AR NC MM SJF. Contributed reagents/materials/ analysis tools: MM MF. Wrote the paper: YJ CC MM SEH NC SJ SJF. References 9. Hariri S, Joseph S, Forsythe SJ (2012) Predominance of Cronobacter sakazakii ST4 clonal complex strains in Cronobacter neonatal meningitis infections in US 2011. Emerg Infect Dis. In Press. 1. Friedemann M (2007) Enterobacter sakazakii in food and beverages (other than infant formula and milk powder). Int J Food Microbiol 116: 1–10. 2. Kandhai MC, Reij MW, Gorris LG, Guillaume-Gentil O, van Schothorst M (2004) Occurrence of Enterobacter sakazakii in food production environments and households. Lancet 363: 39–40. 2. Kandhai MC, Reij MW, Gorris LG, Guillaume-Gentil O, van Schothorst M (2004) Occurrence of Enterobacter sakazakii in food production environments and households. Lancet 363: 39–40. 10. Lai KK (2001) Enterobacter sakazakii infections among neonates, infants, children, and adults. Medicine 80: 113–122. 11. Himelright I, Harris E, Lorch V, Anderson M (2002) Enterobacter sakazakii infections associated with the use of powdered infant formula -Tennessee, 2001. J Am Med Assoc 287: 2204–2205. 3. Kucerova E, Joseph S, Forsythe SJ (2011) The Cronobacter genus: ubiquity and diversity. Quality Assurance and Safety of Crops & Foods 3: 104–122. 4. FAO/WHO (2008) Enterobacter sakazakii (Cronobacter spp.) in powdered follow-up formulae. Microbiological Risk Assessment Series No. 15, Rome, 90pp. Available: http://www.who.int/foodsafety/publications/micro/mra_followup/ en/. 12. van Acker J, de Smet F, Muyldermans G, Bougatef A, Naessens A, et al. (2001) Outbreak of necrotizing enterocolitis associated with Enterobacter sakazakii in powdered milk formula. J Clin Microbiol 39: 293–7. 13. Farmer JJ, Asbury MA, Hickman FW, Brenner DJ, the Enterobacteriaceae Study Group (USA) (1980) Enterobacter sakazakii: a new species of Enterobacteriaceae isolated from clinical specimens. Int J Syst Bacteriol 30: 569–84. 5. Biering G, Karlsson S, Clark NC, Jonsdottir KE, Ludvigsson P, et al. (1989) Three cases of neonatal meningitis caused by Enterobacter sakazakii in powdered milk. J Clin Microbiol 27: 2054–6. 6. Burdette JH, Santos C (2000) Enterobacter sakazakii brain abscess in the neonate: the importance of neuroradiologic imaging. Pediatr Radiol 30: 33–4. 7. Willis J, Robinson JE (1988) Enterobacter sakazakii meningitis in neonates. Pediatr Infect Dis J 7: 196–9. 8. Caubilla-Barron J, Hurrell E, Townsend S, Cheetham P, Loc-Carrillo C, et al. (2007) Genotypic and phenotypic analysis of Enterobacter sakazakii strains from an outbreak resulting in fatalities in a neonatal intensive care unit in France. J Clin Microbiol 45: 3979–85. 8. Caubilla-Barron J, Hurrell E, Townsend S, Cheetham P, Loc-Carrillo C, et al. 14. Iversen C, Waddington M, On SL, Forsythe S (2004) Identification and phylogeny of Enterobacter sakazakii relative to Enterobacter and Citrobacter species. J Clin Microbiol 42: 5368–70. 15. Iversen C, Lehner A, Mullane N, Bidlas E, Cleenwerck I, et al. (2007) The taxonomy of Enterobacter sakazakii: proposal of a new genus Cronobacter gen. nov. and descriptions of Cronobacter sakazakii comb. nov. Cronobacter sakazakii subsp. sakazakii, comb. nov., Cronobacter sakazakii subsp. malonaticus subsp. nov., Cronobacter turicensis sp. nov., Cronobacter muytjensii sp.nov., Cronobacter dublinensis sp. nov. and Cronobacter genomospecies 1. BMC Evol Biol, 7: 64. 13. Farmer JJ, Asbury MA, Hickman FW, Brenner DJ, the Enterobacteriaceae Study Group (USA) (1980) Enterobacter sakazakii: a new species of Enterobacteriaceae isolated from clinical specimens. Int J Syst Bacteriol 30: 569–84. Figure S1 BLAST Ring Image Generator (BRIG) analy- sis of the Cronobacter genomes. (DOC) Figure S1 BLAST Ring Image Generator (BRIG) analy- sis of the Cronobacter genomes. (DOC) Figure S2 BLAST Ring Image Generator (BRIG) analy- sis of Cronobacter plasmid pESA3 with matching sequence content found in other Cronobacter species. (DOC) Figure S3 BLAST Ring Image Generator (BRIG) analy- sis of Cronobacter plasmid pCTU1 to matching content in other Cronobacter species. (DOC) y Evolutionary rate of gene families. Individual gene sequences (DNA and amino acid sequences) for all genes in gene families predicted by OrthoMCL were extracted using custom Perl scripts. Codon alignments within gene families were constructed using PAL2NAL [72]. Amino acid alignments needed as input for PAL2NAL were constructed using Muscle [68]. The dS/dN ratios for all possible pairwise comparisons of the codon alignments within a gene family were calculated using SNAP [73]. Mean dS/ dN ratios were assigned for individual gene families by averaging all pairwise ratios within each family. The pan genome was ranked based on logarithmic (log10) dS/dN ratios; enrichment analysis for gene sets containing relatively fast evolving gene families was conducted using Gene Set Enrichment Analysis (GSEA) [74]. Within GSEA, controlled false discovery rates were used to adjust p values for multiple comparisons. y Evolutionary rate of gene families. Individual gene sequences (DNA and amino acid sequences) for all genes in gene families predicted by OrthoMCL were extracted using custom Perl scripts. Codon alignments within gene families were constructed using PAL2NAL [72]. Amino acid alignments needed as input for PAL2NAL were constructed using Muscle [68]. The dS/dN ratios for all possible pairwise comparisons of the codon alignments within a gene family were calculated using SNAP [73]. Mean dS/ dN ratios were assigned for individual gene families by averaging all pairwise ratios within each family. The pan genome was ranked based on logarithmic (log10) dS/dN ratios; enrichment analysis for gene sets containing relatively fast evolving gene families was conducted using Gene Set Enrichment Analysis (GSEA) [74]. Within GSEA, controlled false discovery rates were used to adjust p values for multiple comparisons. Table S1 Number of putative prophage regions in the Cronobacter genomes. (DOC) Table S2 Presence (+) or absence (2) of type six secretion systems in Cronobacter spp. (DOC) Table S3 Copy number variation of the hemolysin gene within the Cronobacter genus. (DOC) Comparative Analysis of Cronobacter Genomes 16. Joseph S, Cetinkaya E, Drahovska H, Levican A, Figueras M, et al. (2012) Cronobacter condimenti sp. nov., isolated from spiced meat and Cronobacter universalis sp. nov., a novel species designation for Cronobacter sp. genomospecies 1, recovered from a leg infection, water and food ingredients. Intl J Syst Evol Microl 62: 1277–83. 44. Hurrell E, Kucerova E, Loughlin M, Caubilla-Barron J, Forsythe SJ (2009) Biofilm formation on enteral feeding tubes by Cronobacter sakazakii, Salmonella serovars and other Enterobacteriaceae. Intl J Food Microbiol 136: 227–231. J 45. Maclean LL, Pagotto F, Farber JM, Perry MB (2009a) Structure of the antigenic repeating pentasaccharide unit of the LPS O-polysaccharide of Cronobacter sakazakii implicated in the Tennessee outbreak. Biochem Cell Biol (Biochimie et Biologie Cellulaire) 87: 459–465. 17. Baldwin A, Loughlin M, Caubilla-Barron J, Kucerova E, Manning G, et al. (2009) Multilocus sequence typing of Cronobacter sakazakii and Cronobacter malonaticus reveals stable clonal structures with clinical significance which do not correlate with biotypes. BMC Microbiol 9: 223. 46. MacLean LL, Vinogradov E, Pagotto F, Farber JM, Perry MB (2009b) Characterization of the O-antigen in the lipopolysaccharide of Cronobacter (Enterobacter) malonaticus 3267. Biochem Cell Biol 87: 927–32. 18. Joseph S, Sonbol H, Hariri S, Desai P, McClelland M, et al. (2012) Diversity of the Cronobacter genus as revealed by multi locus sequence typing. J. Clin. Microbiol. 50: 3031–3039. 47. Czerwicka M, Forsythe SJ, Bychowska A, Dziadziuszko H, Kunikowska D, et al. (2010) Chemical structure of the O-polysaccharide isolated from Cronobacter sakazakii 767. Carbohydr Res 345: 908–913. 19. Joseph S, Forsythe SJ (2011) Predominance of Cronobacter sakazakii ST4 in neonatal infections. Emerg. Infect. Dis. 17: 1713–1715. 48. Maclean LL, Pagotto F, Farber JM, Perry MB (2009c) The structure of the O- antigen in the endotoxin of the emerging food pathogen Cronobacter (Enterobacter) muytjensii strain 3270. Carbohydr Res 344: 667–671. 20. Townsend S, Hurrell E, Forsythe S (2008) Virulence studies of Enterobacter sakazakii isolates associated with a neonatal intensive care unit outbreak. BMC Microbiol 8: 64. y j y 49. Mullane N, O’Gaora P, Nally JE, Iversen C, Whyte P, et al. (2008) Molecular analysis of the Enterobacter sakazakii O-antigen gene locus. Appl Environ Microbiol 74: 3783–94. 21. Cruz A, Xicohtencatl-Cortes J, Gonza´lez-Pedrajo B, Bobadilla M, Eslava C, et al. (2011a) Virulence traits in Cronobacter species isolated from different sources. Can J Microbiol. 57: 735–44. 50. Comparative Analysis of Cronobacter Genomes Jarvis KG, Grim CJ, Franco AA, Gopinath G, Sathyamoorthy V, et al. (2011) Molecular characterization of Cronobacter lipopolysaccharide O-antigen gene clusters and development of serotype-specific PCR assays. Appl Environ Microbiol 77: 4017–26. 22. Cruz AL, Rocha-Ramirez LM, Gonzalez-Pedrajo B, Ochoa SA, Eslava C, et al. (2011b) Flagella from Cronobacter sakazakii induced an inflammatory response in human monocytes. Cytokine 56: 95. 51. Sun Y, Wang M, Liu H, Wang J, He X, et al. (2011) Development of an O- antigen serotyping scheme for Cronobacter sakazakii. Appl Environ Microbiol 77: 2209–14. 23. Kothary MH, McCardell BA, Frazar CD, Deer D, Tall BD (2007) Characterization of the zinc-containing metalloprotease encoded by zpx and development of a species-specific detection method for Enterobacter sakazakii. Appl Environ Microbiol 73: 4142–51. 52. Sun Y, Wang M, Wang Q, Cao B, He X, et al. (2012) Genetic Analysis of the Cronobacter sakazakii O4 to O7 O-antigen gene clusters and development of a PCR Assay for identification of all C. sakazakii O serotypes. Appl Environ Microbiol 78: 3966–74. 24. Pagotto FJ, Nazarowec-White M, Bidawid S, Farber JM (2003) Enterobacter sakazakii: infectivity and enterotoxin production in vitro and in vivo. J Food Prot 66: 370–5. 53. Arbatsky NP, Sun Y, Shashkov AS, Wang M, Liu B, et al. (2012) Structure and genetics of the O-antigen of Cronobacter sakazakii G2726 (serotype O3) closely related to the O-antigen of C. muytjensii 3270. Carbohydr Res. 355: 50–55. 25. Franco AA, Hu L, Grim CJ, Gopinath G, Sathyamoorthy V, et al. (2011) Characterization of putative virulence genes on the related RepFIB plasmids harbored by Cronobacter spp. Appl Environ Microbiol 77: 3255–67. 54. Allison GE, Verma NK. (2000) Serotype-converting bacteriophages and O- antigen modification in Shigella flexneri. Trends Microbiol 18: 17–23. 26. Townsend SM, Hurrell E, Gonzalez-Gomez I, Lowe J, Frye JG, et al. (2007) Enterobacter sakazakii invades brain capillary endothelial cells, persists in human macrophages influencing cytokine secretion and induces severe brain pathology in the neonatal rat. Microbiology 153: 3538–47. 55. Iversen C, Druggan P, Forsythe SJ (2004) A selective differential medium for Enterobacter sakazakii. Intl J Food Microbiol 96: 133–139. 56. Bardowski J, Ehrlich SD, Chopin A (1994) BglR protein, which belongs to the BglG family of transcriptional antiterminators, is involved in beta-glucoside utilization in Lactococcus lactis. J Bacteriol 176: 5681–5685. 27. Kim K, Kim K-P, Choi J, Lim J-A, Lee J, et al. Comparative Analysis of Cronobacter Genomes (2010) Outer membrane proteins A (OmpA) and 6 (OmpX) are essential for basolateral invasion of Cronobacter sakazakii. Appl Environ Microbiol 76: 5188–5198. pp 28. Kim KP, Loessner MJ (2008) Enterobacter sakazakii invasion in human intestinal Caco-2 cells requires the host cell cytoskeleton and is enhanced by disruption of tight junction. Infect Immun 76: 562–70. 57. Gilmour MW, Thompson NR, Sanders M, Parkhill J, Taylor D (2004) The complete nucleotide sequence of the resistance plasmid R478: defining the backbone components of incompatibility group H conjugative plasmids through comparative genomics. Plasmid 52: 182–202. g j 29. Kucerova E, Clifton SW, Xia XQ, Long F, Porwollik S, et al. (2010) Genome sequence of Cronobacter sakazakii BAA-894 and comparative genomic hybridiza- tion analysis with other Cronobacter species. PLoS ONE 5: e9556. 58. Nucci S-P, Ba¨umle AJ (2007) Evolution of the chaperone/usher assembly pathway. Fibrial classification goes Greek. Microbiol Mol Biol Rev 71: 551. 59. Teng C-H, Cai M, Shin S, Zie Y, Kim K-J, et al. (2005) Escherichia coli K1 RS218 interacts with human brain microvascular endothelial cells via type 1 fimbria bacteria in the fimbriated state. Infect Immun 73: 2923–31. 30. Hamby S, Joseph S, Forsythe SJ, Chuzhanova N (2011) In silico identification of pathogenic strains of Cronobacter from biochemical data reveals association of inositol fermentation with pathogenicity. BMC Microbiol 11: 204. 60. Mattick JS (2002) Type IV pili and twitching motility. Ann Rev Microbiol 56: 289–314. 31. Stephan R, Lehner A, Tschler P, Rattei T (2011) Complete genome sequence of Cronobacter turicensis LMG 23827, a food-borne pathogen causing deaths in neonates. J Bacteriol 193: 309–310. 61. Zhou Y, Tao J, Yu H, Ni J, Zeng L, et al. (2012) Hcp Family proteins secreted via the Type VI secretion system coordinately regulate Escherichia coli K1 interaction with human brain microvascular endothelial cells. Infect Immun. 80: 1243–51. 32. Chen Y, Strain EA, Allard M, Brown EW (2011) Genome sequence of Cronobacter sakazakii E899, a strain associated with human illness. J Bacteriol 193: 581. 33. Didelot X, Falush D (2007) Inference of bacterial microevolution using multilocus sequence data. Genetics 175: 1251–1266. 62. Almagro-Moreno S, Boyd EF (2009) Insights into the evolution of sialic acid catabolism among bacteria. BMC Evol Biol 9: 118. 63. Zerbino DR, Birney E (2008) Velvet: algorithms for de novo short read assembly using de Bruijn graphs. Genome Res 18: 821–829. 34. Comparative Analysis of Cronobacter Genomes Lima-Mendez G, Van Helden J, Toussaint A, Leplae R (2008) Prophinder: a computational tool for prophage prediction in prokaryotic genomes. Bioinfor- matics 24: 863–5. 64. Kurtz S, Phillippy A, Delcher AL, Smoot M, Shumway M, et al. (2004) Versatile and open software for comparing large genomes. Genome Biol 5: R12. 35. Aziz RK, Bartels D, Best AA, DeJongh M, Disz T, et al. (2008) The RAST Server: rapid annotations using subsystems technology. BMC Genomics 9: 75. 65. den Bakker HC, Cummings CA, Ferreira V, Vatta P, Orsi RH, et al. (2010) Comparative genomics of the bacterial genus Listeria: Genome evolution is characterized by limited gene acquisition and limited gene loss. BMC Genomics 11: 688. 36. Snipen L, Almoy T, Ussery DW (2009) Microbial comparative pan-genomics using binomial mixture models. BMC Genomics 10: 385. g 37. Karginov FV, Hannon GJ (2010) The CRISPR system: small RNA-guided defence in bacteria and archaea. Molecular Cell 37: 7–19. 66. Kent WJ (2002) BLAT-the BLAST-like alignment tool. Genome Res 12: 656– 664. 38. Caubilla-Barron J, Forsythe S (2007) Dry stress and survival time 38. Caubilla-Barron J, Forsythe S (2007) Dry stress and survival time of Enterobacter sakazakii and other Enterobacteriaceae. J Food Protect 70: 2111–7. 67. Li L, Stoeckert CJ Jr, Roos DS (2003) OrthoMCL: identification of ortholog groups for eukaryotic genomes. Genome Res 13: 2178–2189. J 39. Osaili T, Forsythe S (2009) Desiccation resistance and persistence of Cronobacter species in infant formula. Intl J Food Microbiol 136: 214–220. 39. Osaili T, Forsythe S (2009) Desiccation resistance and persistence 68. Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic acids research 32: 1792–1797. species in infant formula. Intl J Food Microbiol 136: 214–220 and high throughput. Nucleic acids research 32: 1792–1797. 40. Williams TL, Monday SR, Edelson-Mammel S, Buchanan R, Musser SM (2005) A top-down proteomics approach for differentiating thermal resistant strains of Enterobacter sakazakii. Proteomics 5: 4161–9. 69. Stamatakis A, Hoover P, Rougemont J (2008) A rapid bootstrap algorithm for the RAxML Web servers. Systematic biology 57: 758–771. 70. Felsenstein J (1989) PHYLIP - Phylogeny Inference Package (Version 3.2). Cladistics 5: 164–166. 41. Asakura H, Morita-Ishihara T, Yamamoto S, Igimi S (2007) Genetic characterization of thermal tolerance in Enterobacter sakazakii. Microbiol Immunol 51: 671–677. 71. Gelman A, Rubin DB (1992) Inference from iterative simulation using multiple sequences. Statistical science 7: 457–472. 42. References (2007) Genotypic and phenotypic analysis of Enterobacter sakazakii strains from an outbreak resulting in fatalities in a neonatal intensive care unit in France. J Clin Microbiol 45: 3979–85. PLOS ONE \| www.plosone.org November 2012 \| Volume 7 \| Issue 11 \| e49455 11 Comparative Analysis of Cronobacter Genomes 74. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, et al. (2005) Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proceedings of the National Academy of Sciences of the United States of America 102: 15545–15550. 75. Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22: 4673–80. 78. Felsenstein J (1981) Evolutionary trees from DNA sequences: a maximum likelihood approach. J Mol Evol 17: 368–376. y p y 77. Kimura M (1980) A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J Mol Evol 16: 111–120. 76. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, et al. (2011) MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolution- ary distance, and maximum parsimony methods. Mol Biol Evol 28: 2731–2739. Comparative Analysis of Cronobacter Genomes Gajdosova J, Benedikovicova K, Kamodyova N, Tothova L, Kaclikova E, et al. (2011) Analysis of the DNA region mediating increased thermotolerance at 58uC in Cronobacter sp. and other enterobacterial strains. Antonie Van Leeuwenhoek. 100: 279–89. 72. Suyama M, Torrents D, Bork P (2006) PAL2NAL: robust conversion of protein sequence alignments into the corresponding codon alignments. Nucleic acids research 34: W609–612. 43. Hurrell E, Kucerova E, Loughlin M, Caubilla-Barron J, Hilton A, et al. (2009) Neonatal enteral feeding tubes as loci for colonisation by members of the Enterobacteriaceae. BMC Infect Dis 9: 146. 73. Korber B (2000) HIV Signature and Sequence Variation Analysis. In: Allen G. Rodrigo GHL, editor. Computational Analysis of HIV Molecular Sequences: Kluwer Academic Publishers. 55–72. PLOS ONE \| www.plosone.org November 2012 \| Volume 7 \| Issue 11 \| e49455 November 2012 \| Volume 7 \| Issue 11 \| e49455 12 Comparative Analysis of Cronobacter Genomes 76. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, et al. (2011) MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolution- ary distance, and maximum parsimony methods. Mol Biol Evol 28: 2731–2739. 77. Kimura M (1980) A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J Mol Evol 16: 111–120. 78. Felsenstein J (1981) Evolutionary trees from DNA sequences: a maximum likelihood approach. J Mol Evol 17: 368–376. Comparative Analysis of Cronobacter Genomes PLOS ONE \| www.plosone.org November 2012 \| Volume 7 \| Issue 11 \| e49455 PLOS ONE \| www.plosone.org 13
https://openalex.org/W2042802872	https://revistadeindias.revistas.csic.es/index.php/revistadeindias/article/download/951/1025/1445	es		De la indiferencia a la resistencia. Los sectores populares y la Guerra de Independencia en el norte de Chile (1817-1823)	Revista de Indias	2,014	cc-by	13,233	Revista de Indias, 2014, vol. LXXIV, núm. 260 Págs. 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia. Los sectores populares y la Guerra de Independencia en el norte de Chile (1817-1823)* por Igor Goicovic Donoso Universidad de Santiago de Chile El proceso de Independencia en Chile fue liderado por la élite criolla terrateniente y mercantil radicada en Santiago y en sus inmediaciones. La posterior etapa de construcción del Estado y de la nación fue expresión de su proyecto político. No obstante, los sectores populares constituidos por peones, artesanos y labradores, esclavos e indígenas, se mantuvieron ajenos a esta propuesta y, en el contexto de la desorganización de la sociedad colonial, desplegaron estrategias de autonomización social y política. Expresión de ello fueron las montoneras, el bandolerismo y la deserción. Este artículo analiza el desarrollo de estos fenómenos en el norte de Chile entre 1817 y 1823. Palabras Clave: Independencia de Chile; sectores populares, Atacama y Coquimbo; montoneras; bandolerismo; deserción. Introducción Durante el proceso de Independencia de Chile, los sectores populares fueron interpelados compulsivamente para sostener el esfuerzo de guerra, tanto por los patriotas, como por los realistas1. No obstante su actitud general * Este artículo forma parte del proyecto de Investigación FONDECYT 1095107. Agradezco la colaboración de la estudiante Andrea Armijo Reyes, en la recopilación de las fuentes documentales. 1 La categoría pueblo o sector popular es utilizada en este artículo de acuerdo con la denominación que genéricamente le asignaban las élites del período a las clases subalternas. Ella refería al conjunto de sujetos que desempeñaba oficios «viles», como peones, labradores y artesanos, y que pertenecían a la condición de indígenas, negros y mestizos. Al respecto ver Fernández Labbé, 42/1 (Santiago de Chile, 2009): 109-139. 130 Igor Goicovic Donoso fue la de resistir el enrolamiento forzoso y, una vez recluidos en el mismo, tratar, sistemáticamente, de huir. Este tipo de situaciones permite inferir que la filiación de estos grupos con los bandos en disputa se configuró como débil e inestable. La Guerra era observada por las clases subalternas como un conflicto al interior de la élite dominante, en consecuencia, totalmente ajeno a sus intereses2. Esta indiferencia de los sectores populares frente al proceso de Independencia se explica por el carácter de «guerra civil» interoligárquica que tuvo dicho conflicto. Efectivamente, el proceso desencadenado en 1808 con la captura y cautiverio del rey de España, Fernando VII, dio origen en Hispanoamérica a múltiples y heterogéneos procesos, en los cuales las élites terratenientes y mercantiles adoptaron distintas posiciones políticas, de cara a la reorganización del poder local. No obstante, la rearticulación de las estructuras de poder no modificó sustantivamente la condición económica y social de las clases subalternas, en consecuencia éstas no se sintieron interpretadas (más allá de los consuetudinarios vínculos clientelares), con los aspectos de orden político que encarnaron en los diferentes bandos en conflicto3. La Guerra, y la depredación de los ecosistemas productivos, así como las levas de los trabajadores rurales, suponían un daño manifiesto y sustantivo a las economías campesinas. Así, cuando las clases populares se veían arrastradas al enfrentamiento intentaban medrar de los recursos disponibles para resarcirse de las pérdidas materiales que conllevaba la Guerra. No debe sorprender, entonces, que la adscripción política de estos grupos sociales tuviera un importante grado de movilidad. Efectivamente, el curso general de los aconLa compilación de Palacios (2009) entrega una mirada particularizada sobre este fenómeno, mientras que Fontana, 33/2 (Santiago de Chile, 2010): 7-21, aporta un interesante enfoque general. En el anverso de estas interpretaciones, la lectura de la Independencia de Hispanoamérica como parte del proceso revolucionario de naturaleza «liberal» del mundo atlántico, se puede revisar en Carmagnani, 33/2 (Santiago de Chile, 2010): 23-36. 3 Un análisis de conjunto al movimiento de Independencia en América del Sur en, Bushnell, 1991: 75-123. La puesta al día sobre los debates historiográficos en torno a la Independencia en Schmidt-Nowara, 45/2 (Pittsburgh, 2010): 228-235. Un primer acercamiento a la participación de los sectores populares en las guerras de Independencia lo aporta Vilar, 1972: 461-508. Para el caso de Chile, el enfoque interpretativo más clásico (la Independencia como hibridación entre tradición y modernidad), corresponde a Jocelyn-Holt, 1992. Desde otra perspectiva, Salazar Vergara (2005), instala la dicotomía entre élite mercantil y pueblo productor, como base de los procesos políticos de construcción de Estado. Mientras que Guerrero Lira (2002), devela la acentuación de las contradicciones políticas entre realistas y patriotas, en el contexto de la Restauración Borbónica. Una perspectiva más reciente, centrada en las convulsiones del proceso político, en Grez Toso, 26 (Portoviejo, 2010): 40-55. Por último, los enfoques historiográficos que han predominado en el tratamiento de la Independencia de Chile en, Moulian Emparanza, 1996. 2 Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 131 tecimientos bélicos definía las lealtades, cuestionando, incluso, los vínculos clientelares que estaban en la base de las relaciones sociales tradicionales. De esta manera se fue perfilando una conducta política marcada por la ubicuidad de los actores populares; ello significa que éstos adoptaban circunstancialmente las posturas políticas que predominaban en los diferentes y cambiantes escenarios que establecía la Guerra. En consecuencia, el surgimiento y desarrollo de la adhesión a la causa de la patria recién se puede verificar después de la batalla de Maipú, en abril de 1818. Antes de dicha fecha, a lo menos en las regiones de Atacama y Coquimbo, el reconocimiento y apoyo irrestricto a los patriotas no se puede observar con claridad. Es más, en la fase álgida de los enfrentamientos armados de la denominada «Guerra a Muerte», estas adhesiones continuaron siendo muy débiles. No obstante, las manifestaciones de descontento popular no surgieron con el proceso de Independencia. Durante la fase colonial se produjeron en América violentos levantamientos indígenas –especialmente en México y Perú– y sangrientas revueltas de esclavos –en el Caribe y Brasil–4. En Chile, por el contrario, la sociedad colonial sólo conoció del bandolerismo y de la vagancia como formas de resistencia popular frente a la dominación oligárquica. Las características específicas que definían la estructura agraria del país y las relaciones sociales y de poder que se construyeron a su interior explican, en buena medida, las peculiaridades del caso chileno5. Territorio y sociedad Hacia 1810 el territorio norte de la gobernación de Chile, se encontraba formado por cuatro partidos: Copiapó, Huasco, Coquimbo e Illapel6. Las co4 Sobre las rebeliones indígenas en la zona andina ver O’Phelan, 1988. También Cahill, 1999. Para el caso de México ver Reina Ayoma, 1980 y Escobar, LXIX/246 (Madrid, 2009): 73-100. Los alzamientos de esclavos, para el Brasil colonial, han sido tratados por, Moura, 1988, Reis, 1995 y Marquese y Parrón, LXXI/251 (Madrid, 2011): 19-52. 5 Sobre este tema ver el clásico estudio de Góngora, 1966. Más recientemente, los trabajos de Cárdenas, 11 (Santiago de Chile, 1991): 47-61 y de Araya, 1999. 6 La evolución histórica de la división político-administrativa del Reino de Chile se encuentra en, Cobos Noriega, 1989: 33-113. La reorganización republicana del espacio y de la población en Sanhueza, 41/2 (Santiago de Chile, 2008): 447-493. La macro zona denominada contemporáneamente como Norte Chico (Regiones de Atacama y Coquimbo), representó hasta la década de 1870 el límite norte de la República. Su relieve alcanza un rasgo distintivo en el perfil orográfico del país. Ello porque desde las estribaciones cordilleranas se extienden en dirección al Océano Pacífico una serie de cordones montañosos, de regular altura, que atraviesan transversalmente el territorio. Encajonados entre estos cordones montañosos discu- Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 132 Igor Goicovic Donoso marcas que forman parte de este amplio territorio desarrollaban, a comienzos del siglo XIX, una amplia y compleja gama de actividades económicas. En ellas se combinaban y superponían las actividades, mineras, agrícolas y ganaderas. La estacionalidad de las actividades agrícolas y ganaderas y los ciclos recurrentes de expansión y contracción de la minería determinaban una fuerte movilidad territorial y laboral entre la población de la zona. Estos elementos han permitido caracterizar la región, desde los albores del período colonial, como una zona de frontera7. Cabe señalar que durante el siglo XVIII se produjo en todo Chile, y particularmente en Atacama y Coquimbo, un acelerado proceso de crecimiento de la población, siendo sus fases más relevantes los períodos 1744-1766 y 1788-18138. Algunas estimaciones generales, realizadas por Jorge Pinto Rodríguez, señalan que entre 1700 y 1835 la población de la zona norte del país se incrementó a un ritmo promedio de 1,8% anual. En ello influyó, de manera importante, la carencia de dificultades para asegurar la subsistencia de una población dispersa, la disminución de las epidemias y la profunda brecha que se abrió entre el crecimiento de la tasa de natalidad y la estabilización de la tasa de mortalidad9. Efectivamente, a partir de 1780 es posible observar en el Norte Chico un período de recuperación del ritmo de crecimiento de la población, que se extendió hasta mediados del siglo XIX. Las referencias estadísticas para el período no son precisas. No obstante, cabe destacar que el Primer Censo de Población levantado en Chile, entre 1831 y 1835, fijó en 1.010.332, el número de habitantes del país. En 1843 la población llegaba a 1.083.801, mientras que a mediados del siglo XIX ya se bordeaba el millón y medio de habitantes10. rren una serie de ríos –Copiapó, Elqui, Limarí y Choapa, entre los más importantes– de escaso caudal, que permiten la formación de un manto vegetal de gran relevancia para las actividades mineras y agrícolas de la zona. Ver Cunill, 1979 y CORFO, 1950-1962. 7 El debate teórico sobre el concepto de «frontera» en Quijada, LXII/224, (Madrid, 2002): 103-142. Para el caso de Chile, y teniendo como referencia la región del Norte Chico, ver Pinto Rodríguez, 1980: 14. Otro estudio clásico sostiene que pese al aislamiento de los valles de las regiones de Atacama y Coquimbo, se crea en ellos un tipo de vida similar para toda la región, notablemente diferente del núcleo central. Ver Carmagnani, 1963: 9-12. 8 Pinto Rodríguez, 1980: 13. Salinas Meza (París, 1986): 103-112. 9 Pinto Rodríguez, 1980: 31-33. Pinto calcula en 20.000 el número de habitantes para el Norte Chico a fines del siglo XVIII. Mientras que el censo de población de 1835 estableció que la provincia de Coquimbo, compuesta por los departamentos de Copiapó, Vallenar, Freirina, Elqui, Serena, Ovalle, Combarbalá e Illapel, contaba con 99.844 habitantes el año del empadronamiento. República de Chile, 1835: 216. 10 Sobre este punto ver los censos de población de los años respectivos. República de Chile, 1835; 1844; 1855. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 133 Este proceso se explica, fundamentalmente, por una serie de factores económicos que concurren en el mismo período; a saber: El aumento de la producción agrícola, la expansión experimentada por la producción minera y el incremento del comercio interregional con Santiago, e internacional, particularmente con Inglaterra11. También habría que señalar que la configuración de la zona centro-sur del país como teatro de operaciones militares, durante la Guerra de Independencia, operó como un factor de atracción de población hacia el sector septentrional, escasamente amagado por los enfrentamientos armados12. Se configuró, entonces, un paisaje económico en el cual los centros mineros– especialmente aquéllos orientados a la explotación cuprífera– operaban fundamentalmente para satisfacer la demanda externa –primero española y luego inglesa–, mientras que las haciendas y los pequeños productores agrícolas dirigían sus productos hacia los campamentos mineros, que comenzaban a configurar un emergente mercado interno. Ello, empero, no implicó una transformación radical de la estructura económica colonial ni mucho menos conllevó un desarrollo efectivo de las fuerzas productivas. Lo anterior se explica, en buena medida, por el hecho de que los excedentes eran controlados por un corto número de hacendados y mercaderes, mientras que los pequeños productores estaban atados a una estructura arcaica, en la cual predominaba el monocultivo y la falta de semillas e instrumentos agrícolas13. Atacama y Coquimbo eran, en esta época, estructuras económicas pre-capitalistas, afectas a ciclos de expansión y contracción de sus actividades mineras, con una población reducida en volumen y dispersa a lo largo de los valles transversales. Cavieres y Cortés, sostienen que las villas coloniales –a excepción de Santiago– fueron sólo una extensión del espacio rural circundante, por ende la riqueza (o pobreza) de villas como Copiapó, La Serena o San Rafael de Rozas, quedaba condicionada a la expansión de la minería o al desarrollo de los viñedos14. El escenario social, derivado de esta estructura económica, fue heterogéneo y convulso. El pueblo, particularmente los grupos excluidos y marginados, se constituían en sus propios reductos de cotidianeidad. Los más recurrentes fueron, las fiestas religiosas y luego las fiestas cívicas; la plaza y las alamedas; y evidentemente los espacios de expendio y consumo de alcohol Pinto Rodríguez, 1980: 84 y ss. Carmagnani, 1963: 49-52. Pinto Rodríguez, 1980: 112. 13 La organización económico-social del espacio regional en Cavieres y Cortés, 1992: 81. 14 Cavieres y Cortés, 1992: 93. 11 12 Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 134 Igor Goicovic Donoso (chinganas, tabernas, pulperías, etc.)15. En ellos desplegaban costumbres y formas de vida particulares, en muchos casos asociadas a la vagancia, la trashumancia y el accionar delictual. Se trataba de sujetos atrevidos, irreverentes y violentos que preservaban celosamente su independencia. Uno de los aspectos más controversiales de estos espacios de sociabilización popular se encontraba dado por el importante nivel de violencia que se incubaba y desplegaba en las pulperías16. Efectivamente, el alcohol combinado con el juego y la disputa por los trofeos femeninos, culminaba regularmente en agresiones brutales o en desórdenes colectivos17. Al respecto el viajero francés Claudio Gay, observaba en la década de 1840: En esta especie de reuniones se sucitan con frecuencia serias disputas provocadas sobre todo por la bebida; antes el chileno se armaba inmediatamente con su cuchillo y envolviendo el brazo izquierdo con su poncho a guisa de escudo, avanzaba el uno contra el otro para herirse18. Esta imagen proyectada por Claudio Gay se encuentra ampliamente refrendada en las piezas judiciales contenidas en los archivos locales. Al realizar una revisión parcial sobre un total de 15 causas por homicidio, incoadas entre 1750 y 1835, en la jurisdicción de los juzgados del crimen de los partidos de Copiapó y Huasco, en el extremo norte de la región en estudio, encontramos los siguientes antecedentes: En las 15 causas los homicidas fueron varones, y en 14 de ellas sus víctimas fueron otros hombres. Sólo en una oportunidad la persona asesinada fue una mujer. Un dato interesante en este cúmulo de situaciones lo aporta una causa de 1818, ocasión en la cual tanto la víctima como el victimario fueron niños de aproximadamente 12 años de edad19. Por otra parte, se puede establecer que, regularmente (13 ocasiones), las muertes se producían en el contexto de una riña. Los sujetos, normalmente alterados por los efectos del alcohol, se agredían, primero, con golpes de pies y manos, para luego recurrir a armas corto-punzantes o contundentes, que eran las que, en definitiva, les provocaban la muerte. Cabe destacar que sólo en un 15 La fiesta y la violencia como diada fundante de las identidades populares en, Godoy Orellana, 2007: 5-34. 16 El surgimiento de la tendencia elitaria a criminalizar tanto las conductas como los ámbitos de sociabilización del mundo popular en León León, 41 (Bogotá, 2010): 160-183. 17 Sobre este punto ver Pinto Rodríguez, 1981: 131-137. 18 Gay, 1862: 168. 19 Causas Criminales, Archivo Histórico Nacional, Santiago de Chile (AHN), Judicial de Copiapó, legajo, 19, piezas 16, 19, 20, 26, 31 y 36; legajo 160, piezas 19, 21 y 22; legajo 20, piezas 3, 20, 34 y 48; Judicial de Vallenar, legajo 25, pieza 1; y Judicial de Elquí (Vicuña), legajo 7, pieza 1. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 135 caso (expediente de 1767), la muerte fue resultado de un ataque alevoso que buscaba neutralizar a la víctima para luego robarle. En otra circunstancia (expediente de 1835), se trató de un amotinamiento de peones, motivado por la demanda de alcohol, que provocaron desórdenes en las inmediaciones de un bodegón de la hacienda de Nantoco. Al negárseles la venta del alcohol los peones atacaron las casas de la hacienda y desde el interior de la misma se disparó contra la turba provocando la muerte de uno de los participantes. No obstante, cabe consignar que las armas más recurrentes en todos estos episodios fueron, piedras, palos y cuchillos. Sólo en un caso (el motín recién referido) se utilizó arma de fuego. La mayoría de los participantes en los homicidios eran peones de oficio. En un caso se trató de peones de minas, mientras que en los restantes los peones involucrados trabajaban en las haciendas de la zona. En la fase colonial las autoridades judiciales señalaban la condición estamental de los participantes. En los expedientes que tuvimos a la vista, en 7 de un total de 8 casos, la víctima o el victimario fueron identificados como indios y negros. Por último habría que señalar que los asesinatos se producían, mayoritariamente, en zonas rurales: Como el pueblo de indios de San Fernando (cercano a San Francisco de la Selva de Copiapó), en la casa habitación de los involucrados, a orillas de los ríos, en el Camino Real o en bodegones de expendio de licores. Ello constituía un problema adicional para las autoridades, ya que si bien éstas realizaban infructuosos esfuerzos por llevar a los victimarios ante la justicia, su tarea se veía dificultada por la huida de los sujetos de la zona o por su evasión desde las precarias cárceles locales20. Sin duda que éstos fenómenos, de gran frecuencia en la sociedad tradicional, cuestionaban la autoridad de las élites dominantes, poniendo de manifiesto el alto grado de autonomía alcanzado por la cultura popular. La violencia se encontraba en el centro de las relaciones sociales. Es más, se configuraba como un componente habitual de la existencia y era admitida como tal por el cuerpo social. Esto se encuentra relacionado con que la sociedad vive una situación de conflicto permanente, tanto en el ámbito público –conflictos de interés–, como en el privado –desacuerdos y relaciones intrafamiliares agresivas21. Los conflictos de interés se originaban por el dominio La idea de pena y penitencia en el castigo colonial en Araya, 39/2 (Santiago de Chile, 2006): 349-367. El modelo «Portaliano» de disciplinamiento social en Pinto Vallejos, 44/2 (Santiago de Chile, 2011): 401-442. 21 Sobre los escenarios y dinámicas de la violencia en Chile, durante el siglo XIX, ver: Goicovic Donoso, 21 (Valparaíso, 2004): 121-145 y Goicovic Donoso, 3/1 (Osorno, 2006): 75-80. 20 Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 136 Igor Goicovic Donoso de bienes, préstamos o pactos incumplidos y los intrafamiliares oponían, principalmente, a los esposos y, en menor medida, a padres e hijos. Pero también hay violencia fuera del hogar, en el campo, en las calles o en los lugares de diversión. Como indicamos previamente, en estos casos el grado máximo de violencia era el homicidio22. En la esfera pública, la violencia cotidiana era tan admitida como la intrafamiliar, y se recurría muy poco a la justicia para arreglar los conflictos criminales. Estamos en presencia, por lo tanto, de una sociedad en la que actúan al mismo tiempo dos fuerzas: Un cuerpo social habituado a absorber y resolver sus tensiones por si mismo, pero a la vez un ordenamiento social administrado jerárquicamente que trataba de impregnar y normalizar al cuerpo social a través del Estado y de la justicia. Estas fuerzas contrapuestas lograron coexistir, en un precario equilibrio, durante la última fase de la administración colonial pero, al desencadenarse la Guerra de Independencia, las clases populares multiplicaron sus estrategias y dinámicas autonomistas. El Bajo Pueblo y la Guerra de Independencia en Chile La estabilidad del mundo colonial se vio alterada profundamente por la Guerra de Independencia23. Las movilizaciones y enfrentamientos político-militares del período, relajaron la autoridad estatal y hacendal y permitieron la organización armada del campesinado24. En este proceso uno de los rasgos distintivos fue la indiferencia inicial que los sectores populares manifestaron frente al conflicto político. Su relación con los distintos bandos estuvo marcada por su adscripción clientelar, es decir, su movilización política o militar estaba condicionada por los vínculos sociales y laborales que los unían a la élite hacendal y mercantil25. No obstante el carácter cooptativo que adquirió el proceso, los sectores populares privilegiaron la defensa y resguardo de sus 22 Ver Salinas Meza, 4 (Santiago de Chile, 2000): 13-49 y Pinto Rodríguez, 8 (Santiago de Chile, 1988): 73-97. 23 Para un enfoque general sobre la relación entre revueltas y revoluciones ver Tenenti, 1999: 17-38. 24 Los movimientos populares premodernos han sido estudiados por: Rudé, 2009; Hobsbawm, 2001; y Thompson, 1984. Para el caso latinoamericano, en particular sobre Argentina, un trabajo muy completo en Fradkin, 2006. 25 El grueso de las tropas que formaron en Chile el ejército realista eran de origen hispano-criollo. Muchos de ellos chilotes y habitantes de la región de Concepción. Ello fue tempranamente acreditado por Campos Harriet, 1976. Un enfoque similar en Luqui-Lagleze, 6 (Santiago de Chile, 2000): 69-76. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 137 propios intereses26. Este fenómeno fue tempranamente percibido por Marcelo Segall, quien sostuvo: «La investigación objetiva en los archivos y periódicos nos entrega a las clases inferiores de Chile actuando en ambos lados, el patriota y el realista, en favor de sus intereses directos, inmediatos o lejanos, como lucha de clases»27. La retórica patriota independentista no logró seducir de manera amplia y sistemática a las clases subalternas. En algunas circunstancias, como es el caso de la población afro descendiente, la promesa de emancipación favoreció el reclutamiento de cuadros de este origen para las milicias antimonárquicas. No obstante, como sostiene Marixa Lasso, ello no supuso, necesariamente, avanzar hacia la configuración de una sociedad fundada en la igualdad racial28. Pero, de igual forma, los grupos subalternos, como en este caso la población de origen negro, vio en la Guerra de Independencia una posibilidad de movilidad social ascendente. Esta fue la situación de los artesanos mulatos que formaron en Chile el Batallón de Infantes de la Patria (una unidad militar segregada), que jugó un rol destacado en las campañas militares de 1817 y 181829. Su aspiración, no siempre concretada, fue alcanzar reconocimiento por su contribución a la lucha por la Independencia y, además, acceder a una condición social y jurídica de mayor equidad. El caso del mundo mapuche fue muy diferente. La situación de los pueblos al sur del río Bio Bío se encontraba regulada por los parlamentos que la corona española había suscrito con las comunidades aborígenes. Efectivamente, en el parlamento de Quilín, realizado en 1641, los españoles reconocieron la frontera del río Bio Bío y la independencia del territorio mapuche ubicado entre este cauce y el río Toltén. Los parlamentos posteriores, especialmente los verificados en el paraje de Negrete en 1726, 1771 y 1793, ratificaron este acuerdo30. De esta manera la condición de autonomía política y territorial de dichos pueblos en los espacios que habitaban se encontraba jurídicamente La indiferencia de los sectores populares frente al proceso de construcción de Estado y Nación desarrollado por la élite criolla, en Pinto Vallejos y Ortiz de Zarate, 2009: 120-144. 27 Segall, 125 (Santiago de Chile, 1962): 5. 28 Lasso, 27 (Bogotá, 2007): 32-45. Para el caso de la población cubana de color, enrolada en el «ejército mambi» durante la Independencia de ese país, ver el trabajo de Helg, LVIII/212, (Madrid, 1998): 47-63. Una postura similar se encuentra en el trabajo de Hébrard, LXII/225 (Madrid, 2002): 429-462, a propósito de los «soldados-ciudadanos» en Venezuela, en el contexto del proceso de Independencia y de la posterior constitución de la sociedad política. 29 Contreras Cruces, 44/1 (Santiago de Chile, 2011): 51-89 y Vivanco Cifuentes, 6 (Santiago de Chile, 2010): 7-28. 30 Al respecto ver Bengoa, 2000: 37-40. 26 Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 138 Igor Goicovic Donoso garantizada. Por ello la política patriota de subsumir al conjunto de estos grupos bajo la denominación genérica de chilenos y, a partir de ello, los esfuerzos por reducir la condición de autonomía de los indígenas, estimuló el espíritu de revuelta y, junto con ello, la incorporación masiva de los mapuche a las guerrillas y montoneras realistas31. Miles de mapuche y de hispano-criollos pasaron a engrosar las filas de las guerrillas realistas, colocando en jaque la soberanía del naciente Estado chileno en la zona sur del territorio32. Sobre este punto la historiografía nacional ha desarrollado investigaciones acuciosas, de resultados controversiales. Los primeros en tratar el tema fueron los historiadores decimonónicos, Diego Barros Arana y Benjamín Vicuña Mackenna. Al estudiar las campañas militares de uno de los montoneros más relevante del período, el criollo Vicente Benavides, Barros Arana señala que se trataba de un sujeto «feroz», motivado en sus acciones por el ánimo de venganza por antiguos agravios recibidos de los patriotas. No obstante, este autor no pone en duda el apoyo social con el cual operó este guerrillero realista, ni su capacidad para desestabilizar el nuevo orden político y militar33. Vicuña Mackenna, por su parte, no sólo dio cuenta de la brutalidad del enfrentamiento armado en la zona sur, sino que, además, estableció que dicho conflicto formaba parte de la «guerra civil» que afectó a la sociedad hispano criolla de la época y que se expresó como ruptura territorial. El reino de Chile, hallábase ciertamente dividido al acometer la empresa de su Independencia, i por las influencias combinadas de la política i de la milicia, de la sociabilidad i de la historia, en dos reinos diferentes, apartados, casi hostiles. Uno de esos reinos era Chile, el nombre tradicional de las comarcas del Maipo al Aconcagua, i se estendía del Maule al Paposo [Actualmente de Talca a Taltal]. El otro reino era el fuerte Penco [Concepción], el reino de la espada, como Santiago era de la toga i la cogulla. I tan cierto era esto que los altivos pobladores de la raya 31 En los territorios al sur del Bio Bío las guerrillas realistas, compuestas por pehuenches e hispano-criollos, fueron organizadas y lideradas por sacerdotes de la orden de San Francisco. Al respecto ver Valenzuela Márquez, 38/1 (Santiago de Chile, 2005): 139-144. Este fenómeno, por su brutalidad, fue denominado «Guerra a Muerte». Para el caso de México el trabajo de Mengus, LXIX/247 (México D.F., 2009): 144, sobre la región mixteca, ha establecido que los indígenas repudiaban tanto el bando patriota como el realista, «…pues ambos exigían contribuciones extraordinarias y provisorias para sostener la guerra». El contrapunto a este enfoque en Tutino, 78/3 (México, 1998): 367-418, quien da cuenta de la activa participación popular e indígena en los procesos revolucionarios del siglo XIX. 32 Los rasgos distintivos de la adhesión a la causa del Rey en la zona de Concepción en Cartes Montory y Puigmal, 2008: 55-77. Una buena caracterización de la zona y de la formación de élites locales durante este período en Mazzei de Grazia, 2004. 33 Barros Arana, 1850: 38. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 139 fronteriza (…) llamaron siempre Chile únicamente al primero de aquellos territorios; i así continúanlo llamando las jentes de aquellas comarcas que obedecen, sin apercibirse de ello, a una tradición inevitable34. Resulta evidente que durante la Guerra de Independencia la noción simbólico-política de Chile aún no se encontraba debidamente desarrollada. Las lealtades que operaban al interior de la sociedad referían, fundamentalmente, a los vínculos clientelares que los sujetos construían con sus patrones y con los caciques locales. Ello incidió, de manera importante, en la formación de las unidades militares y en la construcción de las lealtades políticas al interior del pueblo. Para José Bengoa la Guerra de Independencia fue un fenómeno ajeno a los mapuche y éstos se vieron involucrados en la misma cuando el teatro de operaciones militares se trasladó al sur de Chile. Los mapuche no eran parte de la sociedad hispano-criolla y, por lo tanto, percibieron la Independencia como un hecho externo y ajeno. No obstante, tras la derrota realista en Maipú (1818) y el subsecuente repliegue de las tropas del Rey al sur del país, una parte importante de las comunidades mapuche se integraron a las bandas armadas que desplegaron en el sur de Chile una estrategia de guerrillas para impedir la instauración del nuevo régimen. En esta ocasión no se trató de una guerra regular entre ejércitos convencionales; fue un enfrentamiento brutal entre un Estado en proceso de consolidación, por lo tanto débilmente asentado y precariamente constituido, y grupos armados irregulares (montoneras), compuestas mayoritariamente por indígenas y campesinos hispano-criollos, que utilizaron como método preferente de combate la emboscada y la sorpresa, y que recurrieron al pillaje, el robo y el saqueo para recabar recursos y subsistir35. Por su parte Jorge Pinto sostiene que los mapuche y en general la sociedad regional no miró con simpatía el proceso emancipador. Ello por cuanto desde el período colonial se había producido un proceso de integración regional fundado en la complementariedad de las redes indígenas con las redes capitalistas. Este proceso que no sólo reportaba autonomía política, sino que también otorgaba autonomía económico-social, se veía amenazado por un proyecto independentista planificado y dirigido por la élite santiaguina. De esta manera, la guerra de montoneras del período fue el resultado de la política de resistencia al cambio desplegado por esta comunidad regional36. Vicuña Mackenna, 1868: XX-XXI. Bengoa, 2000: 139-152. 36 Pinto Rodríguez, 2003: 64-78. 34 35 Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 140 Igor Goicovic Donoso A su vez María Isabel Contador establece una diferencia entre las fuerzas que conformaban el bando monárquico. Por una parte identifica a las tropas del ya mencionado caudillo Vicente Benavides, compuestas por soldados e indígenas, pero que fueron expresión de continuidad del ejército realista; mientras que, por otro lado, se encontraban las montoneras autonomizadas, formadas por campesinos, desertores, vagabundos, delincuentes y prisioneros fugados, que si bien se asumían en la causa del Rey, respondían a sus propios intereses y liderazgos37. La historiografía mapuche, de desarrollo reciente, ha sostenido que esta fase es expresión de la etapa más álgida en la defensa de la autonomía territorial y cultural del pueblo-nación mapuche; pero que es, a su vez, el punto de inflexión para la etapa posterior de incorporación forzada al estado chileno38. Por último Leonardo León indica que en las montoneras formadas a partir de 1818 se agruparon «los elementos más díscolos e insubordinados de la gente con los últimos remanentes del ejército de Rey». Este contingente contó con el apoyo del grueso de la población mapuche de la región. Ellos fueron los que constituyeron las partidas armadas de Vicente Benavides y de los hermanos Pincheira, inaugurando la denominada «Guerra a Muerte» que asoló los territorios del sur hasta 182239. Pero este fenómeno presenta rasgos distintivos en las diferentes regiones del país. En la zona norte de Chile la resistencia frente al reclutamiento forzoso y las compulsiones militares se expresaron tanto a través de levantamientos generales (como el alzamiento de los indígenas de Chalinga de 1818), como por la vía del las deserciones. Los sectores populares y la Guerra de Independencia en el Norte Chico Efectivamente, mientras en la zona sur del territorio la Guerra adquiría creciente virulencia, en la zona norte del país el proceso de Independencia y la subsecuente relajación de la autoridad estatal y hacendal, dio origen a dos situaciones: Por una parte, un sector de la población se estableció económicamente de manera independiente, a través de la estrategia de campesiniza37 Contador, 1998: 115-116. 38 Al respecto ver Marimán Quemenado, 2006: 53-127. Una interpretación diametralmen- te diferente entrega Villalobos Rivera, 1989: 24, que sostiene que los mapuche sólo participaron de la lucha «bajo la tentación del pillaje». 39 Leonardo León Solís ha sido el historiador chileno que ha estudiado más sistemáticamente este fenómenos; al respecto ver León Solís, 5 (Santiago de Chile, 2009): 82; 68/2 (Sevilla, 2011): 483-510; 2012: 35 y ss. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 141 ción; mientras que otro, exaltó la trashumancia por la vía del bandolerismo y de la insurgencia armada. Ambas respuestas vinieron a constituir modelos de subsistencia ampliamente legitimados entre los sectores populares40. La expresión más radical de insurgencia armada frente a la conflictividad interoligárquica en la zona norte fue el levantamiento indígena y popular de Chalinga, verificado en el Partido de Illapel el 20 de marzo de 181841. Sobre este acontecimiento el gobernador de Illapel, Tomás Echavarría, escribió al Intendente de Coquimbo: El jueves Santo a las ocho del día, hallándose este vecindario en misa en la Iglesia de Santo Domingo, nos asaltaron los indios del pueblo de Chalinga sublevados, al mando de Francisco Carbajal y de Vicente Paillante, y encontrándonos todos dentro de la dicha Iglesia, entraron a ella con lanzas, garrotes, estoques y otras armas de esta naturaleza, descargándolos por donde podían sin exceptuar a sujeto alguno. Salieron heridos muchos, pero gravemente el capitán Don Francisco Javier Monardez y Don Juan Agustín Romero y a estos y los demás sujetos visibles, y hasta los sacerdotes, nos condujeron a la cárcel con la gritería de viva el Rey y mueran todos los patriotas; y sacando de ella la recluta que tenía pronta para remitir a VS, nos aseguraron con prisiones dándonos la sentencia de degüello. La plebe de esta villa se reunió con ellos a la voz del saqueo, y se ocuparon en éste y en conducirnos reos a la cárcel nuestra42. La incursión de los indígenas causó sorpresa y desasosiego en el gobernador. La turba no sólo había asaltado la villa cabecera del partido, también había golpeado y humillado a sus vecinos más ilustres, profanado la iglesia de Santo Domingo y soliviantado al bajo pueblo de la localidad para que participara, junto con ellos, en el saqueo de las viviendas y comercios de la oligarquía. Es por ello que Tomás Echavarría intuía que este levantamiento no formaba parte de la guerra civil que dividía coyunturalmente a la élite entre patriotas y realistas. Se trataba, en este caso, de la indiada, de los vagabundos y malentretenidos, que en virtud de «oscuros designios» se habían levantado contra los tres pilares del sistema de dominación, fuera este colonial o republicano: La oligarquía, la Iglesia y la propiedad. Tras una semana de escaramuzas los indios de Chalinga fueron derrotados por las tropas organizadas por el hacendado de Illapel, Francisco José Irarrázaval. Los indios capturados tras la refriega fueron remitidos a Coquimbo, siguiendo precisas instrucciones del Director Supremo, Bernardo O’Higgins. 40 Al respecto ver Bengoa, 1988: 107-113. Sobre este episodio ver Goicovic Donoso, 4 (Santiago de Chile, 2000): 51-86. Carta del Gobernador de Illapel al Intendente de Coquimbo, 21 de marzo de 1818, AHN, Intendencia de Coquimbo, vol. 6. 41 42 Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 142 Igor Goicovic Donoso El pueblo de Chalinga quedó desolado43. De acuerdo con Ramón Guerrero, teniente de gobernador del partido de Illapel, tras la deportación de algunos y la fuga de otros, sólo quedaban en el pueblo de «diez o dose indios que no se mezclaron en tan feroz atentado». Esta pequeña comunidad fue colocada bajo la dirección de un nuevo cacique, Dionisio Pallacán, el cual, a diferencia del revoltoso Vicente Paillante, se caracterizaba, a juicio de Guerrero, por ser «un decidido patriota (…)»44. No obstante las readecuaciones políticas que se produjeron al interior de la comunidad indígena de Chalinga, el destino del pueblo se encontraba zanjado. Las autoridades locales no estaban dispuestas permitir que las fértiles tierras de la localidad permanecieran en manos de los indígenas y que éstos, rotulados de revoltosos, continuaran constituyendo una amenaza para las autoridades locales. El gobernador de Illapel, hacia 1823, Miguel Bravo de Saravia, informaba a las autoridades de Coquimbo que en las 400 cuadras de tierras de Chalinga, habitaban 150 familias de indios, a los cuales no se les podía hacer adjudicación individual de tierras, «por evitar perjuicios en las sementeras». Para completar su informe y sugerir medidas resolutorias el gobernador manifestaba: El carácter turbulento de estos Indios, de habitual holgazanería, las continuas quejas que contra ellos se presentan en este juzgado por repetidos robos que hacen de animales en las haciendas vecinas y demás individuos que los rodean, y la mancha que tienen sobre si por los alborotos que causaron en este partido el año de 1818, unido todo a la estreches del terreno que forma el pueblo y que impide que echa la adjudicación de las cuadras por familia quede un residuo de tierras de muy mala y difícil proporción para hacer la venta (…) me han hecho sentir que al Estado y al Partido sería muy conveniente destinar a estos indios a otros pueblos, y dejar el terreno que hoy forma este pueblo vacío y en disposición que el Gobierno lo pudiese enajenar en el todo, y de la manera que lo tenga por mas conveniente45. De esta manera, en 1824 el Estado, presionado por la élite minera y hacendal local, concluyó su política de desarticulación de la comunidad adjudiLa zona norte del territorio de Chile se encontraba estrechamente ligada al noroeste argentino y al territorio sur de la Audiencia de Charcas. Las relaciones económicas y sociales identificadas por Sempat Assadourian (Santiago de Chile, 1970: 65-109), operaban a través de rutas y senderos que favorecieron, durante la Guerra de Independencia, la circulación de desertores, noticias, rumores y conspiraciones. 44 Carta del Teniente de Gobernador de Illapel al Director Supremo, 20 de mayo de 1818, AHN, Ministerio del Interior, vol. 38. 45 Del Gobernador de Illapel al Intendente de Coquimbo, 17 de diciembre de 1823, AHN, Intendencia de Coquimbo, vol. 6. 43 Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 143 cando individualmente las tierras a los pocos indígenas que quedaban y facilitando el proceso de enajenación de la propiedad. Mientras que las tierras que habían sido abandonadas por los indígenas expulsados de la zona, fueron pasadas a remate. Durante este complejo e inestable período, la situación económica y social de los trabajadores rurales se vio particularmente precarizada. Los altos niveles de concentración de la propiedad agraria, la extensión de los mecanismos de expoliación que gravaban a los pequeños productores, las complejidades propias de un proceso productivo técnicamente precarizado y las dificultades para distribuir eficientemente los productos agrícolas en los mercados, facilitaron la marginalidad socioeconómica de un porcentaje importante de la población. Ello, a su vez, estimuló los procesos migratorios y las movilidades estacionales de los excedentes laborales. Una parte de este excedente estaba constituido por hombres y mujeres jóvenes, que accedían al servicio domésticos en las casas de la oligarquía, en el caso de las mujeres, y al oficio de peones o a la delincuencia rural, en el caso de los varones46. Pero Jorge Pinto Rodríguez, al analizar el problema de la violencia para el Corregimiento de Coquimbo en el siglo XVIII, concluye que no existió bandolerismo en la región. A ese efecto señala que, durante el período que analiza, no existieron importantes crisis económicas, sociales ni políticas. Por el contrario, la sociedad se encontraba férreamente controlada y la movilidad existente era sólo aquella estimulada por las élites en función de sus requerimientos de mano de obra. De esta manera la violencia existente se canalizaba como violencia interpersonal en las chinganas y bodegones47. No obstante, la crisis política, desencadenada en 1810 y profundizada a los largo de toda esa década de enfrentamientos armados, modificó de manera profunda el escenario del Norte Chico. Si bien la región no fue teatro de 46 La criminalidad rural, asociada regularmente al bandolerismo, ha sido un tópico ampliamente tratado por la historiografía. Una parte importante de las contribuciones contemporáneas se amparan en un trabajo clásico de Hobsbawm, 2000: 19-31. Un contrapunto teórico interesante en la discusión de esta perspectiva en, Daitsman, 19 (Santiago de Chile, 1990): 263-267 y Rodríguez, 26 (Santiago de Chile, 2007): 91-105. Para un análisis regional del bandolerismo ver Slatta, 1987. En Hispanoamérica los fenómenos más extendidos se produjeron en México, Brasil y Perú. Para México ver Robinson, 44/1 (Pittsburgh, 2009): 5-31. Para Brasil ver Holston, 59/6 (Washington, 2007): 21-27. Para Perú ver Aguirre y Walker, 1990. En Chile, el tema del bandolerismo para el siglo XIX ha sido tratado por Salinas Campos, 36 (Madrid, 1986): 57-75. Valenzuela Marquez, 1991. Rojas Gómez, 40/2 (Santiago de Chile, 2007): 419-444; 26 (Santiago de Chile, 2007): 33-56. Gallardo Martínez, 2007. En un contexto más amplio referido a las relaciones de producción y de organización social del agro chileno, fue abordado, también por Bengoa, 1988: 103-115. 47 Pinto Rodríguez, 8 (Santiago de Chile, 1988): 95. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 144 Igor Goicovic Donoso operaciones militares, su condición de reserva estratégica para el bando patriota, supuso sistemáticas exacciones de recursos económicos y levas permanentes de sus trabajadores para engrosar los batallones de los independentistas. De ahí que una parte importante de estos sujetos se echaran al monte y se convirtieran en bandoleros, inaugurando de esa forma una tradición delictiva que, en algunos casos, se extiende hasta nuestros días48. El bandolerismo en el Norte Chico operó con un alto grado de movilidad territorial. Los bandidos, en general buenos conocedores de su entorno, utilizaban los boquetes de la Cordillera de Los Andes para traficar el ganado robado y para buscar refugio, durante las incursiones represivas, en el territorio de Cuyo. De la misma manera, los pasos existentes en los cordones transversales eran ocupados para desplazarse de un valle a otro. Ello favorecía que los bandoleros no tuvieran una localización precisa y que las autoridades de la época recibieran noticias confusas sobre su presencia en Copiapó, Elqui, Limarí, Choapa, la Ligua o Los Andes. A ello contribuía el que una parte importante de estos sujetos se dedicaban de manera preferente a la arriería, en especial de ganado caprino, durante la temporada de verano. En consecuencia, la trashumancia este-oeste y norte-sur era parte consustancial de sus trayectorias cotidianas. También debemos destacar que la práctica del bandolerismo se apoyaba en redes sociales fundadas en una heterogénea gama de vínculos sociales: De familia y parentesco, de linaje y clan, de amistad y paisanaje, de señorío y clientela. Lo interesante del fenómeno es que en estas redes la acción solidaria era un factor fundamental para la adecuada articulación del grupo49. Tanto la comunidad en la cual se desenvolvían los bandoleros, como los integrantes del grupo, entendían que la camaradería, el apoyo mutuo y la restitución de favores, conformaban un conjunto de valores que se encontraban en la base de la relación y de la subsistencia colectiva. No obstante, se debe señalar que los bandoleros no golpearon, exclusivamente, a los ricos hacendados o mercaderes de la región. En muchas ocasio48 El año 2005 la prensa de la región de Coquimbo daba cuenta de la captura de los integrantes de una banda de cuatreros conocida como «Los Chuchetas del Río». Ésta se encontraba integrada (entre otros sujetos) por los hermanos Oscar, Narciso y Pedro Velasco Guerrero y había sido fundada, a comienzos de la década de 1970, por el campesino Pascual Velasco, padre de las personas detenidas. La prensa describía a la banda como una «tradición familiar». Ver Muñoz, Yanko, «El fin de los cuatreros del Norte», El Mercurio, Santiago de Chile, 24 de enero de 2005. 49 El concepto de «red social» fue desarrollado por Bott 1990 [1957]. La construcción comunitaria de la red social en Imizcoz Beunza, 1996: 13-50. El tema de la solidaridad como estrategia de reciprocidad en la familia rurales de Chile en el siglo XIX en Goicovic Donoso, 2006. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 145 nes sus víctimas fueron miembros de su propia clase o grupo social e incluso, más ocasionalmente, de su misma comunidad. Es más, estos bandoleros también fueron parte de los grupos armados organizados por las élites oligárquicas, a efectos de cumplir tareas de amedrentamiento y control social sobre los trabajadores rurales. No es extraño, en consecuencia, que cuando los bandoleros rompen el lazo solidario que los une a la comunidad, sea la propia comunidad la que los denuncie o entregue a las autoridades. Junto a las montoneras y al bandolerismo es posible observar, en este período, otras manifestaciones de rebeldía popular, como la resistencia al enrolamiento militar y la deserción50. Efectivamente, el reclutamiento forzoso llevado a cabo por las autoridades locales devenía, regularmente, en huídas al monte. Mientras que las deserciones se hacían especialmente frecuentes cuando los enfrentamientos militares alcanzaban mayor intensidad o cuando el abastecimiento de las unidades militares experimentaba un serio deterioro51. En ambas circunstancias los huidos escapaban con el armamento que les proporcionaba la élite. Muchos de los soldados desertores, fácilmente se convertían en peones o en delincuentes ya que, manejaban armas, no tenían tierras y su paradero era desconocido para sus familias. Por otro lado la precariedad de su condición socio-económica los obligaba a transitar expeditivamente de una actividad a otra. Esta condición itinerante, sumada al ambiente de desconfianza y temor que afectaba a las comunidades campesinas durante el enfrentamiento armado, colocaba a los huidos en un difícil trance. Los batallones de enganchadores y de cazadores de desertores apresaban a todo aquel que no portaba el pasaporte de libre tránsito respectivo o que no acreditaba oficio o residencia estable. Muchas de las capturas se nutrían, precisamente de la información que la comunidad entregaba sobre los afuerinos o personas extrañas a la localidad. Esto impelía tanto a los desertores como a aquellos que podían ser objeto de reclutamiento a convertirse rápidamente en ladrones o salteadores de caminos52. 50 Al respecto ver los trabajos de León Solís, 5 (Santiago de Chile, 2009): 75-101; 29 (Santiago de Chile, 2008): 33-66; 35 (Santiago de Chile, 2002): 251-297. También, Valdés Urrutia, 164 (Santiago de Chile, 1998): 103-126. 51 En un artículo muy interesante Martínez Martín y Otárola Cascante, 41/272 (Bogotá, 2010): 86-109, establecieron que un objetivo fundamental de las campañas militares del ciclo 1815-1819 en Venezuela y Colombia, fue la lucha por el acceso al ganado, a objeto de darle sustentabilidad a la intendencia militar. La derrota militar suponía la pérdida del recurso y ello, a su vez, daba origen a masivas deserciones en ambos bandos. 52 Cortez Ahumada, 2004: 27-32. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 146 Igor Goicovic Donoso A pesar de que la región se posicionó tempranamente en el bando patriota, la lealtad de la élite regional y de los habitantes de la zona nunca estuvo debidamente garantizada. Los vaivenes que experimentó el proceso independentista estimuló el despliegue de conductas políticas oportunistas, que situaban a los actores regionales en posiciones vacilantes y ambivalentes. La cohesión de la sociedad local en torno al Estado republicano y a la idea de nación fue el resultado de un largo proceso social y político, que recién hacia mediados de la década de 1830 comenzó a mostrar resultados. No obstante, la base de dicho proceso lo constituyó el triunfo de las armas patriotas en el enfrentamiento armado previo. Por ello, la fase 1817-1823, fue fundamental en la construcción de los soportes políticos y simbólicos en torno a los cuales, posteriormente, se forjó la identidad nacional. La recluta forzosa, como señalamos, previamente, se convirtió, en esa época, en el principal mecanismo utilizado por las autoridades patriotas para formar sus unidades militares. Una nota enviada por el gobierno central a las autoridades de la Ligua, Petorca e Illapel, en julio de 1817, es perentoria en ese sentido: Se trata de aumentar nuestras fuerzas militares a un pie que la haga respetable a los enemigos exteriores. Para el efecto aplicará V su empeño y notorio patriotismo en reclutar sin perder instante, doscientos o trescientos jóvenes, robustos, solteros y amantes de la patria en todo el territorio de su jurisdicción, excitando en ellos el honor y entusiasmo por la defensa de nuestro suelo y derechos. Concluida dicha comisión, conducirá V los que se hayan reclutado con la brevedad posible, auxiliándolos de cuenta del Estado, con advertencia que para mediados del entrante agosto deben estar en ésta y a disposición del Excelentísimo señor General en Jefe53. No obstante, los informes entregados por las autoridades locales referidos a la recluta de soldados demuestran que éstas resultaban cada vez menos exitosas. Así lo indicaba en 1817 el gobernador de Illapel: Contestando a la nota de V.S del 10 del pasado octubre debo prevenirle (…) ser moralmente imposible el verificativo de la recluta que se me ordena practicar en este partido de mi mando. Lo mismo digo a V.S pues estos reclutas no pueden hacerse sin valerme de la fuerza, atendiendo a que este es un pueblo sin opinión, de cuya noticia no carecerá V.S para hacerlas de este modo carezco de fuerza, y el emprenderla sin ella, es exponerme a que quede burlado un proyecto, tan interesante al Estado; mayormente cuando estoy penetrado de que había solo uno u otro individuo que quiera alistarse bajo nuestras banderas voluntariamente tengo pedido 53 Circular a las autoridades de La Ligua, Petorca e Illapel sobre la recluta, 19 de julio de 1817, Archivo Bernardo O’Higgins, tomo XXIII, Santiago de Chile, 1961: 308. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 147 al Supremo Gobierno con fecha 20 de septiembre dos banderas con un cabo y cuatro soltados, y al menos 100 fusiles para formar una Compañía Cívica que sirva de seguridad a los reclutas que se vengan acopiando54. Los dichos del gobernador eran categóricos. No se podía cumplir la orden del Gobierno porque la comunidad local carecía de opinión frente al conflicto y no estaba dispuesta a enrolarse de manera voluntaria. Un año más tarde el escenario no había cambiado sustancialmente. La población de la localidad enfrentada a las partidas oficiales de reclutamiento escapaba a los montes y las autoridades locales daban cuenta de la carencia de recursos para desplegar unidades de persecución y captura: (…) con su llegada a ésta [de la Comisión Reclutadora] se echaron todas las gentes a los cerros y montes (…) También me es doloroso tener que decir a V.S que agotados con la fatalidad y sequedad del año los recursos con que podía contar para el acopio de los 200 reclutas que solicita, V.S, me es de necesidad el haber de paralizar esta operación, pues no tengo ni un miliciano que tenga una cabalgadura para poder perseguir dichos reclutas que con ocasión de las remezas que tengo hechas están como dicen con la oreja parada y es muy difícil tomarlos55. Durante el período de cosecha (en el verano meridional), las huidas y deserciones se incrementaban notoriamente. Ello inducía a las autoridades locales a sugerir a sus superiores la suspensión, a lo menos transitoriamente, de las campañas de reclutamiento: Las actuales ocupaciones de las cosechas me obligan a cerrar en la recluta para dar tiempo para que los labradores cosechen sus sementeras que hasta ahora tienen en pie por falta de peones a causa de andar todos fugitivos de temor de la enunciada recluta. Puede V.S siendo servido ordenar al capitán Mora se retire dejando encargado de su comisión a algún oficial de milicias de aquellas inmediaciones para que vaya recibiendo los reclutas que vayan buenamente cayendo en adelante (…) pues el no seguir por lo pronto es por no ocasionar a los infelices labradores los grandes perjuicios que les demanda la falta de brazos56. Los continuos reveses experimentados por la política de reclutamiento forzoso y las continuas deserciones que afectaban a las unidades militares movilizadas, intentaron ser corregidos mediante la aplicación de medidas más 54 Del Gobernador de Illapel al Intendente de Coquimbo, 5 de noviembre de 1817, AHN, Intendencia de Coquimbo, vol. 6. 55 Ibidem, 16 de septiembre 1819, AHN, Intendencia de Coquimbo, vol. 6. 56 Ibidem, 28 de enero de 1820, AHN, Intendencia de Coquimbo, vol. 6. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 148 Igor Goicovic Donoso severas de castigo. Indignado, el general san Martín, informaba a las autoridades regionales en octubre de 1817: Los desertores se multiplican de un modo escandaloso al abrigo de la impunidad que les franquean sus aposentadores. Dígnese V.E expedir un severo bando contra estos criminales, o bien, haciendo responsables a los hacendados en cuyo distrito se encuentren, o a los jueces territoriales, o por los medios más coercitivos que sean del agrado supremo de V.E dirigiendo a los curas así ésta, como las demás providencias generales, para que leídas en los concursos festivos, no pueda alegarse ignorancia57. Pero el problema de la deserción no se reducía a los sujetos que abandonaban los deberes que la patria les había encomendado. También las autoridades detectaron la existencia de una amplia red de coberturas y apoyos sociales que favorecía las huídas y amparaba a los prófugos. Estos «aposentadores», podían ser hacendados locales interesados en mantener a su alcance la fuerza de trabajo necesaria para el desarrollo de las actividades agrícolas, o trabajadores y miembros de las comunidades campesinas que protegían a sus pares por vínculos de camaradería y compadrazgo. Al respecto, un bando de gobierno de 1817 establecía lo siguiente: Inútiles serían los desvelos del gobierno en la formación de cuerpos militares e infructuosos los sacrificios hechos por los virtuosos ciudadanos para este mismo objeto, si sobre las penas que aplica la ordenanza a los desertores, no se impusiese también la correspondiente a los individuos que fomentan la deserción, abrigando en sus casas y haciendas a estos delincuentes, que por entregarse al ocio y a sus vicios, resisten el servicio y la defensa de su patria. Para que no se frustre objeto tan interesante, se proviene lo siguiente: 1.° que todo propietario pudiente, en cuya casa o hacienda se encontraren desertores, pagará irremisiblemente la multa de 200 pesos por cada uno, sin que le sirva de escudo la ignorancia ni cualesquier pretexto que alegare. 2.° los mayordomos, dependientes o inquilinos, que no dieren parte al juez más inmediato de la aparición de estos criminales, serán condenados con cadenas al pie, al servicio de la Maestranza por el término de un año. 3.° los jueces que se desentendiesen de estas delaciones, y no procediesen prontamente a la aprehensión y remisión de estos reos, serán responsables con sus empleos y calificados en odio de su desobediencia y apatía, por indignos de cargos públicos. 4.° la comisión militar encargada de sustanciar las causas de los desertores antes de oírles su justificación, indagará de ellos los puntos donde han morado 57 Del General San Martín a Bernardo O’Higgins, sobre la deserción, 11 de octubre de 1817, Archivo Bernardo O’Higgins, tomo XXIII, Santiago de Chile, 1961: 330-331 Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 149 durante la deserción, los individuos con quiénes han hablado y cuanto conduzca al esclarecimiento de los aposentadores y contraventores de este decreto. Para que llegue a noticia de todos, publíquense por bando, fíjese e imprímase en Gaceta, y circúlese por el Departamento de Gobierno a todos los jueces de la comprensión del Estado, igualmente a los curas, para que en los días de concurrencia lo hagan leer a su feligresía. Sala Directorial de Santiago, 14 de octubre, 1817 Firma: Francisco Antonio Pérez – José Manuel de Astorga – Luis de la Cruz58. Las redes de protección a los desertores que las autoridades intentaban desmontar eran amplias y heterogéneas, pero, a su vez, de gran efectividad. Las noticias que informaban sobre el despliegue de los batallones de reclutamiento se conocían con antelación y ello operaba como alerta que precipitaba las huidas. Así lo indicaba el gobernador de Illapel, Ramón Guerrero, en 1819: He llegado de la capital y me he encontrado en esta con el Sargento Mayor don José Estevan Paez, e informado por el inmediatamente de la comisión que le ha traído [Reclutar nuevas tropas], he visto ser moralmente imposible que la pueda desempeñar. El mismo es sabedor de que aún antes de que llegase a ésta, ya habían cartas escritas de Coquimbo anunciando la recluta. Así es que esparcida esta noticia los hombres se han ocultado, y no hay casi ni un muchacho que entre ni leña al pueblo, todo esto tengo hecho presente al mencionado Sargento Mayor manifestándole la imposibilidad de poder evacuar su comisión. Yo le he dicho se retire a esa con su tropa a fin de dejar sosegar la gente, y que yo quedaré al cargo de la recluta, y me ha contestado no puede verificarlo sin orden de V.S en esta virtud he determinado remitir a V.S por la porta este oficio, para que impuesto de su contenido, se sirva resolver lo que estime conveniente59. Situaciones como las descritas obligaban a las nuevas autoridades a reclutar sus contingentes militares entre sujetos cuya adhesión a la causa de la patria dejaba más de alguna duda. Este era el caso de los delincuentes habituales y de aquellos que previamente habían combatido en el bando adversario. Sobre el primer caso, el gobernador de Illapel, Ramón Guerrero, informaba en 1820 a su superior en la ciudad de La Serena: Con esta fecha remito a disposición de V.S por el conducto del capitán don Isidro Mora, 25 reclutas. En este número van inclusos dos abigeos, a saber Pascual Gaona y Pascual Gallardo que en caso de no tener aplicación, para la tropa, se 58 Bando contra la deserción, 17 de octubre de 1817, Archivo Bernardo O’Higgins, tomo XXIII, Santiago de Chile, 1961: 331-332. 59 Del Gobernador de Illapel al Intendente de Coquimbo, 5 de agosto de 1819, AHN, Intendencia de Coquimbo, vol. 12. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 150 Igor Goicovic Donoso servirá V.S mandarlas dar otro destino, pues no conviene residan en esta por las muchas quejas que tengo acerca de los robos que han hecho60. El texto siguiente pone de manifiesto que las lealtades políticas de los combatientes se alteraban con facilidad; en especial cuando la suerte del ejército propio se encontraba desmedrada. He licenciado la fuerza armada que traje de Coquimbo conforme a la orden de V.E, dejando sólo tres dragones, tres artilleros y un cabo de asamblea que se han presentado a servir a la Patria, y servían al enemigo; los que me son indispensables para escolta de algunos reos que tengo resuelto llevarlos conmigo a presentarlos a V. E.61. En una nota enviada por el gobernador de Illapel a las autoridades de la región, en 1824, Miguel Bravo de Saravia indicaba que había logrado reclutar 16 personas para el ejército de la patria. De ellas, 5 eran desertores recapturados. Otro de los enganchados, José Cuevas, fue enviado al ejército por ser «perjudicial al pueblo y excelente para soldado, que aunque casado no hace vida maridal con la mujer sino es que intenta quitarle la vida». Mientras que Pablo Hurqueta, «aunque inútil para soldado, puede destinarse al trabajo de obras públicas, por no convenir en este lugar, y por prodigo en procurar consumir los cortos bienes de su padre en el todo»62. Por otro lado la composición de las autoridades políticas en el nuevo Estado debía hacerse, también, minimizando el alcance de las lealtades políticas pretéritas. Efectivamente, los miembros de las élites regionales, al igual que el bajo pueblo, habían manifestado durante la primera fase del conflicto posturas vacilantes o pusilánimes, lo cual complejizaba, habitualmente, la nominación de los cargos públicos. Ello, a su vez, desencadenaba agrias disputas entre los bandos locales. La conquista del poder local suponía reforzar la situación de dominación de una determinada fracción de la élite, por ello, entonces, la acusación de simpatías con los realistas o de obsecuencia frente al dominio de los mismos, se convertía en un poderoso anatema político. En julio de 1817 la élite de Vallenar se encontraba profundamente dividida por el nombramiento del gobernador local. Una fracción apoyaba al minero Ignacio de Urízar, proclamado como tal por el cabildo de la villa; mientras Ibidem, 17 de enero de 1820, AHN, Intendencia de Coquimbo, vol. 6. Del Intendente de Coquimbo al Director Supremo, 3 de marzo de 1817, Archivo Bernardo O’Higgins, tomo XVII, Santiago de Chile, 1957: 49. 62 Del Gobernador de Illapel al Intendente de Coquimbo, 10 de junio de 1824, AHN, Intendencia de Coquimbo, vol. 6. 60 61 Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 151 que el grupo rival apoyaba al gobernador designado por el gobierno central, Francisco Bascuñan y Aldunate. En ambas circunstancias los declarantes aseguraban que su adversario había estado en connivencia con el enemigo durante la restauración borbónica63. Más allá de la justeza de la crítica es posible inferir que, efectivamente, ambos personajes (al igual que quienes los respaldaban) tuvieron una actitud más o menos obsecuente con los realistas; y ello no es extraño si tenemos en consideración la derrota catastrófica de las armas patriotas en 1814. En otras circunstancias, quienes habían desempeñado importantes cargos de representación política en el gobierno patriota hasta 1814, sometidos a la autoridad realista tras la derrota de Rancagua, optaban por reconocer su «delito de rebelión», y solicitar a las nuevas autoridades el perdón o la reducción de las condenas que se les aplicaban. Este fue el caso del ex gobernador de Vallenar, Francisco Sainz de la Peña, que «reconoció sus delitos y pidió perdón» al gobernador de Chile Mariano Osorio64. Nuevamente la tendencia a la ubicuidad de los sujetos y la evolución del conflicto armado incidían en la adopción de posturas y actitudes. De los antecedentes reunidos se pueden inferir algunas conclusiones preliminares. Un primer aspecto a considerar es que las manifestaciones de resistencia popular a los mecanismos compulsivos desplegados por la élite, para sumar a estos grupos al esfuerzo de guerra, fueron múltiples y diversos. El más recurrente fue la oposición al reclutamiento forzoso y la deserción. No obstante, también es posible observar manifestaciones más radicales, como el levantamiento de indios de Chalinga, la guerra de montoneras de la zona sur y la extensión que alcanzó en algunas regiones el bandolerismo rural. Desde esta perspectiva, la Guerra de Independencia favoreció la constitución de un «pueblo en armas» que, circunstancialmente, alcanzó importantes niveles de autonomía social. En ese escenario las armas proporcionadas por la oligarquía reforzaron la autonomía política del bajo pueblo65. 63 Respecto de esta polémica ver Representación del Cabildo de Vallenar sobre que se reponga a Urízar como Gobernador, 21 de junio de 1817, Archivo Bernardo O’Higgins, tomo XVII, Santiago de Chile, 1957: 225-229 y Representación de vecinos de Vallenar en favor de Bascuñán, 6 de julio de 1817, Archivo Bernardo O’Higgins, XVII, Santiago de Chile, 1957: 252-253. Para zanjar el conflicto el gobierno nombró, en agosto de 1817, a Manuel Recabarren en el cargo de gobernador. Su principal función era atenuar los conflictos entre la élite local. 64 Decreto del Presidente por el que condena a Francisco Sainz de la Peña, 20 de octubre de 1815, Archivo Bernardo O’Higgins, tomo XIX, Santiago de Chile, 1959: 165-167. 65 La compleja reorganización de la sociedad y de las estructuras de poder, tras prolongados conflictos armados, para los casos de Perú y México, se puede analizar en Mallon, 2003: 461-508. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 152 Igor Goicovic Donoso Por otra parte, también es posible observar que los sectores populares también se adaptaron a los requerimientos de la élite. Pero esta adaptación fue más bien funcional. Operó de acuerdo con la trayectoria de la Guerra (los que resultaban vencedores y vencidos), y tuvo en cuenta las oportunidades de consecución de botín que la guerra suponía. No obstante, las oscilaciones o ambivalencias que presentó la conducta popular no son exclusivas de este sector social. La élite regional también manifestó dudas e intento medrar de la contingencia de guerra. En todo caso, quienes sostuvieron posturas más decididas o coherentes con el proceso emancipatorio alcanzaron una favorable ubicación política en el nuevo orden republicano. La Guerra, en consecuencia, coadyuvo no sólo a la formación del nuevo orden político, sino también a la renovación de la élite de poder. Bibliografía Aguirre, Carlos y Walker, Charles, Bandoleros, abigeos y montoneros. Criminalidad y violencia en el Perú. Siglos XVIII-XX, Lima, Instituto de Apoyo Agrario/P&P, 1990. Araya, Alejandra, Ociosos, vagabundos y malentretenidos en Chile colonial, Santiago de Chile, DIBAM/LOM, 1999. Araya, Alejandra, “El castigo físico: El cuerpo como representación de la persona, un capítulo en la historia de la occidentalización de América, siglos XVI-XVIII”, Historia, 39/2 (Santiago de Chile, 2006): 349-367. Barros Arana, Diego, Estudios históricos sobre Vicente Benavides y las campañas del sur: 1818-1822, Santiago de Chile, Imprenta de Julio Belín y Compañía, 1850. Bengoa, José, Historia social de la agricultura chilena. El poder y la subordinación, tomo I, Santiago de Chile, Ediciones Sur, 1988. Bengoa, José, Historia del pueblo mapuche. Siglos XIX y XX, Santiago de Chile, LOM Editores, 2000. Bott, Elizabeth, Familia y red social. Roles, normas y relaciones externas en las familias urbanas corrientes, Madrid, Taurus, 1990 [1.º edición, Londres, 1957]. Bushnell, David, “La independencia de la América del Sur española”, Leslie Bethell (ed.), Historia de América Latina, Barcelona, Crítica, 1991: 75-123. Cahill, David, Violencia, represión y rebelión en el sur andino: La sublevación de Tupac Amaru y sus consecuencias, Lima, IEP, 1999. Campos Harriet, Fernando, Los defensores del Rey, Santiago de Chile, Editorial Andrés Bello, 1976. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 153 Cárdenas G., Mario, “Grupos marginados en los inicios de la era republicana: Vagabundos, mendigos e indigentes”, Cuadernos de Historia, 11 (Santiago de Chile, 1991): 47-61. Carmagnani, Marcelo, El salariado minero en Chile colonial. Su desarrollo en una sociedad provincial. El Norte Chico, 1690-1800, Santiago de Chile, Universidad de Chile, Centro de Historia Colonial, Editorial Universitaria, 1963. Carmagnani, Marcello, “La Independencia: Una revolución atlántica”, Cuadernos de Historia, 33/2 (Santiago de Chile, 2010): 23-36. Cartes Montory, Armando y Puigmal, Patrick, De la Alsacia al Bio Bío. El oficial napoleónico Fredéric de Brandsen en las campañas de la Independencia de Chile (18151819), Concepción, Universidad de Los Lagos/Universidad de San Sebastián, 2008. Cavieres, Eduardo y Cortés, Hernán, “Historia regional y estructuras socioeconómicas tradicionales: la sociedad agrícola de La Serena en el siglo XVIII”, Marcela Orellana Muermann y Juan Guillermo Muñoz Correa (eds.), El agro colonial, Santiago de Chile, Universidad de Santiago de Chile, 1992: 81. Cobos Noriega, María Teresa, La división político-administrativa de Chile, 1541-1811, Valparaíso, Instituto de Historia, Universidad Católica de Valparaíso, 1989. Contador, Ana María, Los Pincheira. Un caso de bandidaje social. Chile, 1817-1832, Santiago de Chile, Bravo y Allende Editores, 1998. Contreras Cruces, Hugo, “Artesanos mulatos y soldados beneméritos. El Batallón Infantes de la Patria, en la guerra de Independencia de Chile, 1795-1820”, Historia, 44/1 (Santiago de Chile, 2011): 51-89. CORFO, Geografía económica de Chile, 4 vols., Santiago de Chile, CORFO/Editorial Universitaria, 1950-1962. Cortez Ahumada, Abel, “Delincuencia, redes sociales y espacios en la vida cotidiana rural de Chile central. Valle de Aconcagua, 1820-1850”, Santiago de Chile, Tesis de Licenciatura, Universidad de Chile, 2004. Cunill, Pedro, Geografía de Chile, Santiago de Chile, Editorial Universitaria, 1979. Daitsman, Andy, “Bandolerismo: mito y sociedad. Algunos apuntes teóricos”, Proposiciones, 19 (Santiago de Chile, 1990): 263-267. Escobar Ohmstede, Antonio, “Violencia social en la primera mitad del siglo XIX mexicano: El caso de los huastecas”, Revista de Indias, LXIX/246 (Madrid, 2009): 73-100. Fernández Labbé, Marcos, “La mano del plebeyo levantada, es la montaña que se despeña. La conceptualización del pueblo en tiempos de transformación. Chile, 1750-1850”, Historia, 42/1 (Santiago de Chile, 2009): 109-139. Fontana, Josep, “En torno al significado de la Independencia”, Cuadernos de Historia, 33/2 (Santiago de Chile, 2010): 7-21. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 154 Igor Goicovic Donoso Fradkin, Raúl, La historia de una montonera. Bandolerismo y caudillismo en Buenos Aires, 1826, Buenos Aires, Siglo XXI, 2006. Gay, Claudio, Historia física y política de Chile. Agricultura, vol. 1, París, Imprenta de E. Thunot y Cia., 1862. Gallardo Martínez, Eduardo, “Bandidaje rural en Osorno, 1917-1927: Una década de transgresión y violencia en los campos del sur en las páginas de la prensa local”, Eduardo Gallardo Martínez, Joel Lespai Silva, Jorge Muñoz Sougarret y Raul Núñez Muñoz (eds.), Amotinados, abigeos y usurpadores: Una mirada regional acerca de las formas de violencia en Osorno (1821-1931), Osorno, Universidad de Los Lagos/ PEDCH, 2007: 53-100. Godoy Orellana, Milton, “¡Cuándo el siglo se sacara la máscara! Fiesta, carnaval y disciplinamiento cultural en el Norte Chico. Copiapó, 1840-1900”, Historia, 40/1 (Santiago de Chile, enero-junio de 2007): 5-34. Goicovic Donoso, Igor, “Conflictividad social y violencia colectiva en Chile tradicional. El levantamiento indígena y popular de Chalinga (1818)”, Revista de Historia Social y de las Mentalidades, 4 (Santiago de Chile, 2000): 51-86. Goicovic Donoso, Igor, “Consideraciones teóricas sobre la violencia social en Chile (1850-1930)”, Última Década, 21 (Valparaíso, 2004): 121-145. Goicovic Donoso, Igor, “Los escenarios de la violencia popular en la transición al capitalismo”, Espacio Regional. Revista de Estudios Regionales, 3/1 (Osorno, 2006): 75-80. Goicovic Donoso, Igor, Relaciones de solidaridad y estrategia de reproducción social en la familia popular del Chile tradicional (1750-1860), Madrid, CSIC, 2006. Góngora, Mario, Vagabundaje y sociedad fronteriza en Chile (Siglos XVII a XIX), Santiago de Chile, Cuadernos del Centro de Estudios Socioeconómicos, Universidad de Chile, 1966. Grez Toso, Sergio, “Bicentenario en Chile. La celebración de una laboriosa construcción política”, Spondylus, 26 (Portoviejo, 2010): 40-55. Guerrero Lira, Cristián, La contrarrevolución de la independencia de Chile, Santiago de Chile, Editorial Universitaria/Centro de Investigaciones Diego Barros Arana, 2002. Hébrard, Veronique, “¿Patricio o soldado: Qué ‘uniforme’ para el ciudadano? El hombre en armas en la construcción de la nación (Venezuela, 1.ª mitad del siglo XIX)”, Revista de Indias, LXII/225 (Madrid, 2002): 429-462. Helg, Aline, “Sentido e impacto de la participación negra en la guerra de independencia de Cuba”, Revista de Indias, LVIII/212 (Madrid, 1998): 47-63. Hobsbawm, Eric, Bandidos, Barcelona, Editorial Crítica, 2000. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 155 Hobsbawm, Eric, Rebeldes Primitivos. Estudio de las formas arcaicas de los movimientos sociales en los siglos XIX y XX, Barcelona, Crítica, 2001. Holston, Mark, “Cangaceiros. Brazil’s bandit kings”, Americas, 59/6 (Washington, 2007): 21-27. Imizcoz Beunza, José María, “Comunidad, red social y élites. Un análisis de la vertebración social en el Antiguo Régimen”, José María Imizcoz Beunza (dir.), Élites, poder y red social. Las élites del País Vasco y Navarra en la Edad Moderna. (Estado de la cuestión y perspectivas), Bilbao, Servicio Editorial Universidad del País Vasco, 1996: 13-50. Jocelyn-Holt, Alfredo, La Independencia de Chile. Tradición, modernización y mito, Madrid, Mapfre, 1992. Lasso, Marixa, “Un mito republicano de armonía racial: Raza y patriotismo en Colombia1810-1812”, Revista de Estudios Sociales, 27 (Bogotá, 2007): 32-45. León León, Marco Antonio, “De la compulsión a la educación para el trabajo. Ocio, utilidad y productividad en el tránsito del Chile colonial al republicano, 17501850”, Historia Crítica, 41 (Bogotá, 2010): 160-183. León Solís, Leonardo, “Reclutas forzados y desertores de la patria: El bajo pueblo chileno en la Guerra de la Independencia, 1810-1814”, Historia, 35 (Santiago de Chile, 2002): 251-297. León Solís, Leonardo, “El difícil dilema de sobrevivir entre dos patrias: El bajo pueblo chileno entre Chacabuco y Maipú, 1817”, Cuadernos de Historia, 29 (Santiago de Chile, 2008): 33-66. León Solís, Leonardo, “La deserción durante la guerra de Independencia en Chile, 1818-1820”, Cuadernos de Historia Militar, 5 (Santiago de Chile, 2009): 75-101. León Solís, Leonardo, “Montoneras populares durante la gestación de la República, Chile: 1810-1820”, Anuario de Estudios Americanos, 68/2 (Sevilla, 2011), 483-510. León Solís, Leonardo, Ni patriotas ni realistas. El bajo pueblo durante la Independencia de Chile, 1810-1822, Santiago de Chile, DIBAM, 2012. Luqui-Lagleze, Julio, “El ejército real de Chile en al guerra de la Independencia 18101818”, Anuario de la Universidad Internacional SEK, 6 (Santiago de Chile, 2000): 69-76. Mallon, Florencia, Campesino y nación. La construcción de México y Perú poscoloniales, México D.F., CIESAS/Colegio de San Luis/Colegio de Michoacán, 2003. Marimán Quemenado, Pablo, “Los mapuche antes de la conquista militar chileno-argentina”, Pablo Marimán Quemenado, et al., ¡Escucha winka! Cuatro ensayos de historia nacional mapuche y un epílogo sobre el futuro, Santiago de Chile, LOM, 2004: 53-127. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 156 Igor Goicovic Donoso Marquese, Rafael y Parrón, Tamis, “Revolta escrava e política da escravidao: Brasil e Cuba, 1791-1825”, Revista de Indias, LXXI/251 (Madrid, 2011): 19-52. Martínez Martín, Abel y Otárola Cascante, Andrés, “Hambriento un pueblo lucha”. La alimentación de los ejércitos del Rey y del Libertador durante la Independencia de Colombia (1815-1819)”, Historia Crítica, 41/272 (Bogotá, 2010): 86-109. Mazzei de Grazia, Leonardo, La red familiar de los Urrejola de Concepción en el siglo XIX, Santiago de Chile, DIBAM, 2004. Mengus, Margarita, “La tradición indígena frente a los cambios liberales”, Revista de Indias, LXIX/247 (Madrid, 2009): 137-156. Moulian Emparanza, Luis, La Independencia de Chile. Balance historiográfico, Santiago de Chile, Factum Ediciones, 1996. Moura, Clovis, Rebelioes da sanzala: Quilombos, insurreçoes, guerrilhas, Porto Alegre, Mercado Aberto, 1988. O’Phelan, Scarlett, Un siglo de rebeliones anticoloniales. Perú y Bolivia, 1700-1783, Cuzco, Centro de Estudios Andinos Bartolomé de las Casa, 1988. Palacios, Marco (coord.), La Independencia hispanoamericana. Interpretaciones 200 años después, Bogotá, Norma, 2009. Pinto Rodríguez, Jorge, La población del Norte Chico en el siglo XVIII, Coquimbo, Talleres Gráficos de la Universidad del Norte, 1980. Pinto Rodríguez, Jorge, Las minas de azogue de Punitaqui. Estudio de una faena minera de fines del siglo XVIII, Coquimbo, Universidad del Norte, 1981. Pinto Rodríguez, Jorge, “La violencia en el corregimiento de Coquimbo durante el siglo XVIII”, Cuadernos de Historia, 8 (Santiago de Chile, 1988): 73-97. Pinto Rodríguez, Jorge, La formación del Estado y la nación y el pueblo mapuche. De la inclusión a la exclusión, Santiago de Chile, DIBAM, 2003. Pinto Vallejos, Julio y Ortiz de Zárate, Verónica, ¿Chilenos todos? La construcción social de la nación (1810-1840), Santiago de Chile, LOM, 2009. Pinto Vallejos, Julio, “¿La tendencia de la masa al reposo? El régimen portalino enfrenta al mundo plebeyo”, Historia, 44/2 (Santiago de Chile, 2011): 401-442. Quijada, Mónica, “Repensando la frontera sur Argentina: Concepto, contenido, continuidades y discontinuidades de una realidad espacial y étnica (Siglos XVIII-XIX”, Revista de Indias, LXII/224 (Madrid, 2002): 103-142. Reina Ayoma, Leticia, Rebeliones indígenas en México, México, Siglo XXI Editores, 1980. Reis, Joao José, Slave rebellion in Brazil. The muslim uprising of 1835 in Bahia, Baltimore, Johns Hopskins University Press, 1995. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 157 República de Chile, Repertorio chileno. Año de 1835, Santiago de Chile, Imprenta La Araucana, 1835. República de Chile, II Censo general de la República de Chile (1843), Santiago de Chile, Oficina Central de Estadísticas, 1844. República de Chile, III Censo general de la República de Chile (1854), Santiago de Chile, Oficina Central de Estadísticas, 1855. Robinson, Amy, “Mexican banditry and discourses of class. The case of Chucho El Roto”, Latin American Research Review, 44/1 (Pittsburgh, 2009): 5-31. Rodríguez, “Rebeldes campesinos. Notas sobre el estudio del bandidaje en América Latina (Siglo XIX)”, Cuadernos de Historia, 26 (Santiago de Chile, 2007): 91-105. Rojas Gómez, “Aspectos económicos relacionados con el delito de abigeato en la provincia de Concepción, 1820-1850”, Cuadernos de Historia, 26 (Santiago de Chile, 2007): 33-56. Rojas Gómez, Mauricio, “Entre la legitimidad y la criminalidad: El caso del ‘aparaguayamiento’ en Concepción, 1800-1850”, Historia, 40/2 (Santiago de Chile, 2007): 419-444. Rudé, George, Revolución industrial y revuelta agraria. El capitán Swing, Madrid, Siglo XXI Editores, 2009. Salazar Vergara, Gabriel, Construcción de Estado en Chile (1800-1837). Democracia de los pueblos. Militarismo ciudadano. Golpismo oligárquico, Santiago de Chile, Editorial Sudamericana, 2005. Salinas Campos, Maximiliano, “El bandolero chileno del siglo XIX. Su imagen en la sabiduría popular”, Araucaria de Chile, 36 (Madrid, 1986): 57-75. Salinas Meza, René, “Sur la fecondite en Amerique du Sud. Le cas du Chilí aux XVIIIe et XIXe siècles”, Annales de Démographie Historique, (París, 1986): 103-112. Salinas Meza, René, “Violencias sexuales e interpersonales en Chile tradicional”, Revista de Historia Social y de las Mentalidades, 4 (Santiago de Chile, 2000): 13-49. Sanhueza, María Carolina, “La primera división político-administrativa de Chile, 1811-1826”, Historia, 41/2 (Santiago de Chile, 2008): 447-493. Schmidt-Nowara, Cristopher, “Politics and ideas in Latin American Independence”, Latin American Research Review, 45/2 (Pittsburgh, 2010): 228-235. Segall, Marcelo, “Las luchas de clases en las primeras décadas de la República de Chile, 1810-1846”, Anales de la Universidad de Chile, 125 (Santiago de Chile, 1962): 1-47. Sempat Assdourian, Carlos, “Chile y el Tucumán en el siglo XVI. Una correspondencia de mercaderes”, Historia, 9 (Santiago de Chile, 1970): 65-109. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 158 Igor Goicovic Donoso Slatta, Richard, Bandidos. The varieties of Latin American banditry, New York, Greenwood, 1987. Tenenti, Alberto, De las revueltas a la revoluciones, Barcelona, Crítica, 1999. Thompson, Edward Palmer, Tradición, revuelta y conciencia de clase, Barcelona, Critica, 1984. Tutino, John, “The revolution in mexican Independence: Insurgency and the renegotiation of property, production and patriarchy in the Bajío, 1800-1855”, Hispanic American Historical Review, 78/3 (México, 1998): 367-418. Valdés Urrutia, Mario, “La deserción en el ejército patriota durante la Guerra de la Independencia en Chile: 1813-1818. Notas para su comprensión”, Revista Chilena de Historia y Geografía, 164 (Santiago de Chile, 1998): 103-126. Valenzuela Márquez, Jaime, “Los franciscanos de Chillán y la Independencia: Avatares de una comunidad monarquista”, Historia, 38/1 (Santiago de Chile, 2005): 113-148. Valenzuela Márquez, Jaime, Bandidaje rural en Chile central. Curicó, 1850-1900, Santiago de Chile, DIBAM, 1991. Vicuña Mackenna, Benjamín, La guerra a muerte. Memoria sobre las últimas campañas de la Independencia de Chile, 1819-1824, Santiago de Chile, Imprenta Nacional, 1868. Vilar, Pierre, “La participación de las clases populares en los movimientos de Independencia de América Latina”, VV.AA., La Independencia en el Perú, Lima, IEP, 1972: 155-174. Villalobos Rivera, Sergio, “Guerra y paz en la Araucanía: Periodificación”, VV.AA., Araucanía. Temas de historia fronteriza, Temuco, Ediciones Universidad de la Frontera, 1989: 7-30. Vivanco Cifuentes, Claudio, “La actuación del Batallón Infantes de la Patria durante la Patria Vieja, 1810-1814”, Cuadernos de Historia Militar, 6 (Santiago de Chile, 2010): 7-28. Fecha de recepción: 5 de marzo de 2012. Fecha de aceptación: 13 de diciembre de 2012. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005 De la indiferencia a la resistencia 159 From indifference to resistance. Popular sectors and the War of Independence in northern Chile (1817-1823) Chile’s independence movement was led by the Creole elite of landowners and merchants based in Santiago and its surroundings. The subsequent phase of construction of the state and the nation was an expression of their political project. Nonetheless, popular sectors consisting of laborers, craftsmen and farmers, slaves and Indians, remained aloof from this plan and, in the context of the disruption of colonial society, developed strategies of social and political empowerment. Expressions of this were the montoneras (irregular armed forces), banditry and military desertion. This article analyzes the development of these phenomena in northern Chile between 1817 and 1823. Key words: Independence of Chile; popular sectors; Atacama and Coquimbo; montoneras; banditry; military desertion. Revista de Indias, 2014, vol. LXXIV, n.º 260, 129-160, ISSN: 0034-8341 doi:10.3989/revindias.2014.005
https://openalex.org/W2416913265	http://www.zurnalai.vu.lt/kalbotyra/article/download/7672/5549	Lithuanian	null	Prancūzų kalbos veiksmažodžio devoir ‘privalėti’ semantinė funkcinė polifonija. Tarpkalbinė studija	Kalbotyra	2,015	cc-by	7,206	ISSN 1392-1517. Online ISSN 2029-8315. KALBOTYRA. 2013 • 65 ISSN 1392-1517. Online ISSN 2029-8315. KALBOTYRA. 2013 • 65 Prancūzų kalbos veiksmažodžio devoir ‘privalėti’ semantinė funkcinė polifonija. Tarpkalbinė studija Vita Valiukienė Anglų filologijos katedra Filologijos fakultetas Vilniaus universitetas Universiteto g. 5 LT-01513 Vilnius, Lietuva El. paštas: vita@francais.lt Anotacija Straipsnyje aptariama prancūzų kalbos veiksmažodžio devoir ‘privalėti’ reikšmių įvairovė iš tarpkalbinės prancūzų-lietuvių / lietuvių-prancūzų perspektyvos. Remiantis lingvistinėje literatūroje publikuotais diachroniniais tyrimais apžvelgiama šio veiksmažodžio reikšmių dinamika prancūzų kalboje. Straipsnyje taip pat pateikiama deontinio modalumo, epistemiškumo, evidencialumo kategorijų samprata prancūzų lingvistikoje. Trumpai pristatoma gramatikalizacijos proceso esmė. Apibendrinamosios bandomojo tyrimo pastabos daromos remiantis individualia straipsnio autorės introspekcija analizuojant prancūzų ir lietuvių grožinės literatūros kūrinių pavyzdžius bei jų vertimus į lietuvių ir prancūzų kalbą ir mokslinėje literatūroje vyraujančia išvada: deontiniam devoir ‘privalėti’ aspektui perteikti dažniausiai pasirenkama veiksmažodinė strategija, o epistemiškumui reikšti – adverbialinė raiška. Tyrimu atkreipiamas dėmesys į kontekstinės situacijos svarbą devoir ‘privalėti’ reikšmių identifikavimui. Raktiniai žodžiai: aletinis modalumas, deontinis modalumas, devoir ‘privalėti’, epistemiškumas, evidencialumas, gramatikalizacija 1 Įvadas Tradiciškai prancūzų kalbos mokomosiose priemonėse veiksmažodžio devoir ‘privalėti’ turiniui ir raiškai skiriamas menkas dėmesys. Dažnai apsiribojama jo priskyrimu prie modalinių veiksmažodžių kategorijos ir tik kelių pagrindinių reikšmių pristatymu (Chu 2008, 1). Dendale’is (1994, 25) kritikuoja deskriptyvines prancūzų kalbos gramatikas už tai, kad pastarosiose pristatomas neišsamus šio veiksmažodžio reikšmių potencialas, o 113 episteminiam devoir ‘privalėti’ priskiriama dažniausiai tik tikimybės (pranc. probabilité) reikšmė. Prancūzų lingvistai (Huot 1974; Sueur 1975, 1979, 1981) kalbėdami apie tikimybę, kurią gali turėti devoir ‘privalėti’, ją priešpriešina tikrumo (pranc. certitude) reikšmei, bet niuansų neanalizuoja. Kaip teigia Dendale’is (1994, 25–26), šiuo atveju prancūzų požiūris skiriasi nuo anglosaksiškos tradicijos tuo, kad pastarosiose studijose episteminiam žymikliui must ‘privalėti’ priskiriamas išties platus reikšmių spektras: tikimybė (pranc. probabilité), tikrumas (pranc. vraisemblance), beveik visiškas įsitikinimas (pranc. quasi-certitude) ir analizuojama jo evidencinė prigimtis. Tiek pačia evidencialumo kategorija, tiek evidencinėmis reikšmėmis, kurias gali turėti devoir ‘privalėti’, prancūzų kalbos tyrėjai susidomėjo palyginti neseniai (Barbet, Saussure 2012). Gal tuo iš dalies galima būtų pateisinti lietuvių vertėjus ir prancūzų kaip svetimosios kalbos mokytojus, tinkamai neišnaudojančius plataus šio žodžio reikšmių potencialo. Miliūnaitė (2005) pranešime, kuris buvo skaitytas 2005 m. lapkričio 16 d. Lietuvių literatūros vertėjų sąjungos seminare „Vaikų ir jaunimo literatūros vertimas“ pastebi, kad peržiūrėtose knygose kai kur jis [veiksm. privalėti] kartojamas tiesiog iki koktumo, kone kiekviename puslapyje, ir ta jo koncentracija verčia pasijusti, tarsi būtum atsidūręs geležinių taisyklių valdomame pasaulyje, nors knygos šiaip pasakoja apie gražius ir švelnius dalykus. Straipsnio tikslas yra parodyti, kodėl būtina ir prasminga aptarti veiksmažodžio devoir ‘privalėti’ multifunkcionalumą prancūzų kalboje ir tirti jo raišką prancūzų-lietuvių/ lietuvių-prancūzų tarpkalbinėje perspektyvoje. Šiam tikslui pasiekti bandomojoje studijoje keliami tokie uždaviniai: 1) trumpai pristatyti deontinio modalumo ir epistemiškumo bei evidencialumo kategorijų sampratą prancūzų lingvistikoje; 2) remiantis deskriptyvine-diachronine analize bei žymiausių prancūzų kalbos tyrėjų įžvalgomis aptarti veiksmažodžio devoir ‘privalėti’ gramatėjimo procesą; 3) kontrastyvinės analizės principu mėginti išsiaiškinti, kokie devoir ‘privalėti’ atitikmenys dažniausiai pasirenkami prancūzų grožinės literatūros vertimuose į lietuvių kalbą; 4) atlikti kontrastyvinį bandomąjį tyrimą iš lietuvių kalbos pozicijų: ištirti, kokie lietuvių kalbos grožinės literatūros žodžiai ar jų samplaikos dažniausiai gali tapti veiksmažodžio devoir ‘privalėti’ atitikmenimis. 1 Gramatėjimu, arba gramatikalizacija, laikomas toks procesas, kai į tam tikrą morfosintaksinį kontekstą patekę leksiniai vienetai įgyja gramatinę funkciją ir ilgainiui tampa gramatinių kategorijų (laiko, nuosakos, modalumo, veikslo ar kt.) žymikliais (Marcello-Nizia 2006, 43). 2 Teorinė tyrimo bazė ir tyrimo metodas Straipsnyje referuojami Barbet (2012), Desclés ir Guentchévos (2001), Dendale’io (1994, 1999), Kronningo (1990, 1996, 2001), Sueuro (1975, 1979, 1981), Vetters’o (2004, 2012) darbai, sudarantys teorinį ir empirinį prancūzų kalbos veiksmažodžio devoir ‘privalėti’ 114 tyrimo pagrindą. Devyniolika pavyzdžių (1, 2, 10, 11, 15–20, 24–32) surinkta rankiniu būdu iš XIX a. antros pusės – XXI a. pradžios prancūzų arba XX a. pabaigos – XXI a. pradžios lietuvių grožinės literatūros kūrinių ir jų vertimų į prancūzų arba lietuvių kalbą. Šių pavyzdžių šaltinių santrumpų sąrašas pateikiamas straipsnio pabaigoje. Kai kurie aptariamą reiškinį iliustruojantys pavyzdžiai (3–4) paimti iš žodyno, kiti (5–9, 12–14, 21–23) pateikiami iš referuojamų lingvistų darbų, jų vertimas į lietuvių kalbą – straipsnio autorės. Statistiškai pavyzdžių imtis nėra didelė (32 pavyzdžiai), bet šis bandomasis tyrimas reikšmingas tuo, kad kiekvienas ėminys atskirai analizuojamas ir aprašomas. Toks kokybinis aprašas naudingas multifunkcionalaus prancūzų kalbos veiksmažodžio devoir ‘privalėti’ bendram pristatymui, o vėlesniam tyrimui bus pasirinktas statistinis tekstynu (šiuo metu kuriamu) paremtas kalbos duomenų apdorojimas, kuris pateiks išsamų kiekybinį deontinio ar episteminio modalumo grupių pasiskirstymą originaliuose prancūzų kalbos grožinės literatūros tekstuose ir jų vertimuose į lietuvių kalbą. Analizuojamų reikšminių žodžių glosos pateikiamos laikantis Leipcigo glosavimo taisyklių (http://www.eva.mpg.de/lingua/resources/glossing-rules.php). os pateikiamos laikantis /resources/glossing-rules eorijų apžvalga prancūzų kalbos veiksmaž kinėje prancūzų kalboje ji , 32). Viena vertus, jis ga šlaikęs ir leksines reikšm 9–124): s qui lui devaient six cent as skolingas šešis šimtus parents, dit-il, la crainte d ngas tėvams, – tarė jis, – u ncūzų kalboje, būdingos dynas (LKŽ 1976, 768–76 es kaip būti reikalingam, 3 Diachroninių tyrimų ir modalumo teorijų apžvalga Žvelgiant iš diachroninės perspektyvos, prancūzų kalbos veiksmažodis devoir ‘privalėti’ yra gramatikalizacijos1 procese ir šiuolaikinėje prancūzų kalboje jis užima tarpinę padėtį tarp leksikos ir gramatikos (Vetters 2012, 32). Viena vertus, jis gali būti vartojamas kaip modalumo žymiklis, kita vertus, jis yra išlaikęs ir leksines reikšmes būti skolingam (1), jaustis dėkingam (2) ir kt. (TLF 1979, 119–124): (1) Il dit qu’il y avait à Parme des gens qui lui devaient six cents francs<...>(CP402) ‘Papasakojo, kad Parmoje jam kažkas skolingas šešis šimtus frankų<....>’ (PV356) (1) Il dit qu’il y avait à Parme des gens qui lui devaient six cents francs<...>(CP402) ‘Papasakojo, kad Parmoje jam kažkas skolingas šešis šimtus frankų<....>’ (PV356) (2) Encore une chose que je dois à mes parents, dit-il, la crainte des maladies. (HR165) ‘Dar už vieną dalyką turiu būti dėkingas tėvams, – tarė jis, – už ligų baimę.’ (RŽ109) Leksinės reikšmės, tik kitokios nei prancūzų kalboje, būdingos ir lietuviškam devoir atitikmeniui privalėti. Lietuvių kalbos žodynas (LKŽ 1976, 768–769) prie veiksmažodžio privalėti, privalyti nurodo tokias reikšmes kaip būti reikalingam, reikėti (3), mėgti (4): 115 (3) Bet mergaitės, pačios dar privalydamos auklių, jau iš lopų margas lėles madaravo. (3) (4) Ans neprivalo to valgio, gėrio. (4) Bybee et al. (1994) ir van der Auwera ir Plugianas (1998) iškėlė hipotezę, kad modaliniai veiksmažodžiai, kurių leksinė reikšmė yra reikėti (angl. need), pirmiausiai ima žymėti inherentišką veiksmo dalyvio būtinumo reikšmę, o tie veiksmažodžiai, kurių leksinė reikšmė yra būti skolingam (angl. owe) pirmiausiai tampa deontinio modalumo žymikliais. Prancūzų kalbos diachroninės studijos patvirtina šią mokslininkų hipotezę, nes jau IX– XI a. rašytiniuose šaltiniuose randami pirmieji deontinio modalumo pavyzdžiai: (5) Salvarai eo cist meon fraze Karlo< ...> si cum om per dreit son frada salvar dift. (Rossari et al. 2007, 11) ginti.INF privalėti.PRS3.SG ‘Taigi aš ginsiu savo brolį Karolį <…> kaip kad kiekvienas žmogus privalo ginti savo brolį.’ (Rossari et al. 2007, 11) Kronningo teigimu (1996, 17), prancūzų kalbos veiksmažodis devoir ‘privalėti’ leksinę skolos reikšmę perėmė iš lotynų kalbos veiksmažodžio debere ‘privalėti’. Privalėjimo ir būtinumo reikšmių atsiradimą prancūzų kalboje tyrėjas linkęs sieti ne tiek su lotynų kalbos įtaka, kiek su pačios prancūzų kalbos gramatikalizacijos procesu. Skirtingai nuo anglosaksiškos tradicijos, prancūzų kalboje skiriami du deontinio modalumo tipai: teorinis (pranc. théorique) ir praktinis (pranc. pratique) (Kronning 2001). Deontinis teorinis modalumas, iliustruojamas (5) pavydžiu, nesusijęs su situacijos dalyvio valia, o veiksmą nulemia nuo jo nepriklausomos išorinės aplinkybės – dažniausiai įvairios idėjos ar kolektyvinio elgesio normos ir valios aktai (Barbet 2012). Deontinio praktinio modalumo reikšmė irgi nesusijusi su situacijos dalyvio valia, bet jo veiksmą kontroliuoja konkretūs išoriniai materialūs ar nematerialūs veiksniai, ir tokiu būdu valios aktas gali įgauti labai ryškų teleologinį atspalvį: (6) Et de toutes icelles les premieres venues sont les plus chieres; et doivent estre mengées le jour qu’elles sont escossées, ou privalėti.PRS.3PL būti.INF autrement elles deviennent noires et aigres. (Barbet 2008, 12) ‘Ir iš visų brangiausios yra pačios ankstyviausios, ir turi būti suvalgytos būtent tą dieną, kai būna nuskintos, nes kitaip pasidaro juodos ir rūgščios.’ (6) Et de toutes icelles les premieres venues sont les plus chieres; et (6) Et de toutes icelles les premieres venues sont les plus chieres; et doivent estre mengées le jour qu’elles sont escossées, ou privalėti.PRS.3PL būti.INF autrement elles deviennent noires et aigres. (Barbet 2008, 12) ‘Ir iš visų brangiausios yra pačios ankstyviausios, ir turi būti suvalgytos būtent tą dieną, kai būna nuskintos, nes kitaip pasidaro juodos ir rūgščios.’ autrement elles deviennent noires et aigres. (4) Ans neprivalo to valgio, gėrio. (Barbet 2008, 12) ‘Ir iš visų brangiausios yra pačios ankstyviausios, ir turi būti suvalgytos būtent tą dieną, kai būna nuskintos, nes kitaip pasidaro juodos ir rūgščios.’ ‘Ir iš visų brangiausios yra pačios ankstyviausios, ir turi būti suvalgytos būtent tą dieną, kai būna nuskintos, nes kitaip pasidaro juodos ir rūgščios.’ Nagrinėdamas prancūzų kalbos veiksmažodį devoir ‘privalėti’ ir ieškodamas paralelių tarp jo ir veiksmažodžio pouvoir ‘galėti’, Vetters’as (2004) siūlo skirti dar vieną 116 reikšmę – vidinės būtinybės (pranc. auto-obligation). Tokio modalumo reikšmė jau siejama ne su išorinėmis aplinkybėmis, bet su būtinybe, sąlygojama vidinių fiziologinių subjekto savybių: reikšmę – vidinės būtinybės (pranc. auto-obligation). Tokio modalumo reikšmė jau siejama ne su išorinėmis aplinkybėmis, bet su būtinybe, sąlygojama vidinių fiziologinių subjekto savybių: (7) La femme du gardien eut pitié d’eux et leur proposa du café. Ils acceptèrent. Prévoyant une nuit blanche Aldo en avala plusieurs. Il pouvait avaler une boisson brûlante tandis que Anselme devait attendre privalėti.IPFV.3SG laukti.INF qu’elle tiédisse. (Vetters 2004, 666) ‘Sargo žmonai jų pagailo, ir ji jiems pasiūlė kavos. Jie sutiko. Žinodamas, kad jo laukia bemiegė naktis, Aldo išgėrė keletą puodelių. Jis galėjo gerti vos ne verdantį gėrimą, tuo tarpu Anzelmas turėjo palaukti, kol jis atvės.’ (7) La femme du gardien eut pitié d’eux et leur proposa du café. Ils acceptèrent. Prévoyant une nuit blanche Aldo en avala plusieurs. Il pouvait avaler une boisson brûlante tandis que Anselme devait attendre privalėti.IPFV.3SG laukti.INF ‘Sargo žmonai jų pagailo, ir ji jiems pasiūlė kavos. Jie sutiko. Žinodamas, kad jo laukia bemiegė naktis, Aldo išgėrė keletą puodelių. Jis galėjo gerti vos ne verdantį gėrimą, tuo tarpu Anzelmas turėjo palaukti, kol jis atvės.’ ‘Sargo žmonai jų pagailo, ir ji jiems pasiūlė kavos. Jie sutiko. Žinodamas, kad jo laukia bemiegė naktis, Aldo išgėrė keletą puodelių. Jis galėjo gerti vos ne verdantį gėrimą, tuo tarpu Anzelmas turėjo palaukti, kol jis atvės.’ Van der Auwera ir Plungianas (1998, 80) tokią reikšmę įvardija kaip vidinę subjekto būtinybę (angl. participant internal necessity (need)). Jie pateikia anglų kalbos pavyzdį (8) ir jo vertimą į prancūzų kalbą (9): (8) Boris needs to sleep ten hours every night for him to function properly (9) Boris doit dormir dix heures par nuit pour être en forme, être efficace. privalėti.PRS.3SG miegoti.INF (a) ‘Naktį Borisas turi išmiegoti dešimt valandų, kad jaustųsi žvalus.’ (b) ‘Kad jaustųsi žvalus, Borisas turi išmiegoti dešimt valandų.’ (9) Boris doit dormir dix heures par nuit pour être en forme, être efficace. privalėti.PRS.3SG miegoti.INF (a) ‘Naktį Borisas turi išmiegoti dešimt valandų, kad jaustųsi žvalus.’ (b) ‘Kad jaustųsi žvalus, Borisas turi išmiegoti dešimt valandų.’ (9) (a) ‘Naktį Borisas turi išmiegoti dešimt valandų, kad jaustųsi žvalus.’ (b) ‘Kad jaustųsi žvalus, Borisas turi išmiegoti dešimt valandų.’ Tokio tipo (8), (9) sakinius Wrightas (1963) vadina anankastiniais asertyvais (gr. anankè ‘būtinybė’, lot. adsertio, adserere ‘teigti’), Kronningas (1996, 117) juos įvardija kaip sakinius, reiškiančius tikslingą būtinybę, kurios aktualizacija tampa priklausoma nuo konkrečios sąlygos įvykdymo. Tam, kad būtų pasiektas tikslingas rezultatas – kad Borisas jaustųsi žvalus, turi būti įvykdyta sąlyga – jis turi išmiegoti dešimt valandų. Tikslo ar tikslingo rezultato siekimo reikšmę lietuviškame vertime žymi ne tiek jungtukas kad, kuris gali būti daugiareikšmis, o tariniais einančių veiksmažodžių santykis: šalutinio dėmens tariniu eina tariamosios nuosakos vientisinė forma jaustųsi, o veiksmažodinė dalis turi išmiegoti pagrindiniame sakinyje reiškia apgalvotą veiksmą, kuris priklauso nuo vidinės fiziologinės subjekto charakteristikos. Kadangi lietuvių kalboje dėmenų tvarka yra laisva, todėl galimi abu vertimo variantai (a), (b), nors tradiciškai šalutinis dėmuo eina po pagrindinio (DLKG, 689). Veiksmažodžio devoir ‘privalėti’ vartojimas panašiame kontekste nėra labai dažnas reiškinys prancūzų kalboje, ir greičiausiai dėl to kiti lingvistai tokią jo reikšmę šlieja prie deontinio praktinio modalumo (Barbet 2012). Abu modalumo tipai panašūs tuo, 117 kad jais apibūdinami ne teiginiai, o abstrakčios propozicijos, kurios teisingumo požiūriu nėra vertinamos, be to, abiem atvejais ryškus teleologinis bruožas. Tokiam modalumui, kuris susijęs su dalyvio pajėgumu, valiojimu, gebėjimu atlikti veiksmą, tarptautinėje mokslinėje literatūroje labiau taikomas Palmerio (1986) pasiūlytas terminas dinaminis (angl. dynamic modality). Reikėtų paminėti, kad prancūzų kalboje šalia vidinės būtinybės ar dinaminio modalumo tipų yra dar viena labai panaši būtinumo reikšmė, žyminti inherentišką veikėjo savybę, nesusijusią su išorinėmis aplinkybėmis, – tai įgimta ar įgyta subjekto vidinė nuostata paklusti visuomenės ar moralės normoms. Tokia reikšmė koduojama ne paprasta, bet sangrąžine veiksmažodžio se devoir de + bendratis konstrukcija, kuri reiškia ‘jaustis įsipareigojusiam, privalančiam’ (TLF 1979, 121): (10) <...> l‘Altesse répondit au ministre qu‘une princesse de Parme se devait à elle même d‘être toujours ainsi. (12) Li soens orgoilz le devreit bien confondre. (Rossari et al. 2007, 11) privalėti.COND.3SG sumaišyti.INF. ‘Jo puikybė turėtų tikrai jį pražudyti.’ (8) Boris needs to sleep ten hours every night for him to function properly (CP333) REFL privalėti.IPFN.3SG ‘<...> jos didenybė ministrui atsakė, kad Parmos princesė iš pagarbos sau turinti visada šitaip atrodyti.’ (PV295) (10) <...> l‘Altesse répondit au ministre qu‘une princesse de Parme se devait à elle même d‘être toujours ainsi. (CP333) REFL privalėti.IPFN.3SG ‘<...> jos didenybė ministrui atsakė, kad Parmos princesė iš pagarbos sau turinti visada šitaip atrodyti.’ (PV295) (10) Tai, kad toks atvejis nėra vienintelis iš mūsų rastų pavyzdžių prancūzų kalbos originalo tekstuose ir lietuvių kalbos vertimuose į prancūzų kalbą, leidžia daryti prielaidą, kad tokio tipo modalinė reikšmė yra gana dažna prancūzų kalboje: (11) Moteris privalėjo suteikti jam galimybę. (SŠ31) ‘La femme se devait de lui laisser une chance.’ (PD316) REFL privalėti.IPFN.3SG (11) Moteris privalėjo suteikti jam galimybę. (SŠ31) ‘La femme se devait de lui laisser une chance.’ (PD316) REFL privalėti.IPFN.3SG Dėl pačios epistemiškumo (lato sensu) sąvokos turinio prancūzų tyrėjų tarpe diskusijų nekyla – ji siejama ne su išoriniais ar vidiniais dalykais, lemiančiais situacijos dalyvio veiksmą, bet atspindi pasakymo autoriaus sprendimą, kiek propozicijos turinys atitinka tikrovę (Kronning 2001; Sueur 1979). Kitu svarbiu episteminio modalumo aspektu yra laikoma jo sąsaja su kalbėtojo prisiimama atsakomybe, nes sprendimas apie propozicijos atitikimą tikrovei neišvengiamai suponuoja ir didesnį ar mažesnį kalbėtojo prisiimamos atsakomybės laipsnį. Pirma episteminė veiksmažodžio devoir ‘privalėti’ reikšmė užfiksuota XI a. pabaigoje: Dėl pačios epistemiškumo (lato sensu) sąvokos turinio prancūzų tyrėjų tarpe diskusijų nekyla – ji siejama ne su išoriniais ar vidiniais dalykais, lemiančiais situacijos dalyvio veiksmą, bet atspindi pasakymo autoriaus sprendimą, kiek propozicijos turinys atitinka tikrovę (Kronning 2001; Sueur 1979). Kitu svarbiu episteminio modalumo aspektu yra laikoma jo sąsaja su kalbėtojo prisiimama atsakomybe, nes sprendimas apie propozicijos atitikimą tikrovei neišvengiamai suponuoja ir didesnį ar mažesnį kalbėtojo prisiimamos atsakomybės laipsnį. Pirma episteminė veiksmažodžio devoir ‘privalėti’ reikšmė užfiksuota XI a. pabaigoje: bien confondre. (Rossari et al. 2007, 11) sumaišyti.INF. bien confondre. (Rossari et al. 2007, 11) sumaišyti.INF. 118 Kaip teigia senosios prancūzų kalbos specialistai, tekstynuose tokių pavyzdžių nėra gausu, aiškiai dominuoja deontinė reikšmė. Tai leidžia tyrėjams kelti prielaidą, kad epistemine reikšme veiksmažodis devoir ‘privalėti’ buvo vartojamas labai fragmentiškai (Kronning 1990). Duomenys patvirtina, kad lotynų kalbos veiksmažodis debere ‘privalėti’ taip pat retai vartojamas epistemiškai, ir jo reikšmė akivaizdžiai krypsta link neginčijamos būtinybės raiškos (Bertocchi, Orlandini 2001, 53). Taigi prancūzų tyrėjai laikosi požiūrio, kad episteminė devoir ‘privalėti’ reikšmė galėjo išsirutulioti iš lotynų aletinės (gr. alethes ‘tikras, patikimas’), neginčijamos būtinybės reikšmės, o ne iš leksinės ar modalinės deontinės. (8) Boris needs to sleep ten hours every night for him to function properly Tekstynais paremta analizė rodo, kad aletinės reikšmės atsirado gerokai anksčiau nei episteminės, todėl tradiciškai prancūzų tyrėjai ir išskiria šiuos du epistemiškumo variantus: objektyvų aletinį (pranc. nécessité d’être aléthique objective) ir subjektyvų episteminį būtinumą (pranc. nécessité d’être épistemique subjective). Prancūzų lingvistai linkę nepritarti klasikiniam Bybee (1994) epistemiškumo kilmės modeliui: deontinis modalumas → epistemiškumas. Anglosaksiška tradicija nėra linkusi skaidyti epistemiškumo į smulkesnius vienetus. Palmer (1986) laikosi nuomonės, kad nereikėtų skirti aletinio modalumo nuo episteminio. Kitų lingvistų nuomonė panaši: since the boundaries of alethic and epistemic modal interpretations are often empirically hard to distinguish, there have been attempts at conflating the two types of modality. One line of argument has been to reduce alethic interpretations to epistemic interpretations. (Papafragou 1998, 40) Prancūzų kalbos tyrėjai (Kronning 2001) skirstymą į episteminį aletinį (objektyvų) ir episteminį (subjektyvų) argumentuoja taip: 1) aletinio modalumo, reiškiančio neginčijamą būtinybę (pranc. nécessité absolue), ir episteminės tikimybės (pranc. probabilité) reikšmės net intuityviai yra pernelyg skirtingos, kad galėtų būti talpinamos į vieną kategoriją; 2) abu tipai turi specifinių morfosintaksinių bruožų; 3) diskursiniame kontekste jie gali būti skirtingai interpretuojami; 4) diachroniniu požiūriu aletinio modalumo reikšmės išsirutuliojo gerokai anksčiau nei episteminės. Aletinio modalumo devoir ‘privalėti’ reikšmė koduoja ne privalėjimą ar reikiamybę, bet stichinę ir savaiminę būtinybę. Tai nėra skatinimas, raginimas ar kitoks kalbėtojo valios pareiškimas (Kronning 1996, 70). Remiantis deklaratyvinėmis žiniomis apie pasaulį, daroma apibendrinamoji išvada apie subjekto būseną, kurią jis neginčijamai anksčiau ar vėliau patirs. Tai biologinio proceso užkoduota būtinybė, kuriai žmonija paklūsta. Neišvengiamo būtinumo semantika prancūzų kalboje gali aktualizuotis tokiame sakinyje: 119 (13) Tous les hommes doivent privalėti.PRS.3PL privalėti.PRS.3PL mirti.INF p (a) ‘Visi žmonės turi / turės mirti.’ (b) ‘Visi žmonės mirtingi.’ (c) ‘Visi mirsim.’ Lietuvių kalbos vertime sintetinę būtinybę galima perteikti nebūtinai išlaikant modalinį veiksmažodį ‘turėti’ (a), įmanoma ir kitokia neišvengiamos būtinybės aktualizacija (b), (c). Veiksmažodžio devoir ‘privalėti’ gramatikalizacijos procesas patvirtina hipotezę (van der Auwera, Plungian 1998), kad modalumo žymikliai, reiškiantys galimybę ir būtinybę, ilgainiui gali tapti laiko žymikliais. Įdomu tai, kad viduramžių prancūzų kalbos tekstynuose neaptinkama perifrastinė konstrukcija aller ‘eiti’ + bendratis, kuria šiuolaikinėje kalboje koduojamas būsimas prospektyvus, orientuotas į ateitį, veiksmas ir kuri yra pati populiariausia būsimojo laiko reiškėja (Buridant 2000, 354). Tekstynai liudija tokį būsimojo laiko raiškos būdą atsiradus tik XV a. (ibidem). Prospektyvaus būsimojo veiksmo vektoriumi viduramžiais buvo konstrukcija devoir ‘privalėti’ + bendratis, kuri į šiuolaikinę prancūzų kalbą gali būti verčiama pasitelkiant aller ‘eiti’, être sur le point ‘rengtis, ketinti’, falloir ‘reikėti’. (8) Boris needs to sleep ten hours every night for him to function properly (14) Vois que maistre Adans fait le sage Pour che qu’il doit estre escoliers. (Barbet 2008, 13) privalėti.PRS.3SG būti.INF ‘Tik pažiūrėk, kokiu išminčiumi dedasi ponaitis Adomas, nes juk jis turi pradėti eiti mokslus!’ (14) Konstrukcija devoir ‘privalėti’ + bendratis nepakankamai sugramatiškėjo, kad šiandien galėtų būti laikoma vienu iš perifrastinių būsimųjų laikų šiuolaikinėje prancūzų kalboje (Kronning 1990, 8). Tradiciškai Vakarų lingvistikoje, kaip teigia Sueuras (1981, 165), visus kalbos reiškinius mėginama preciziškai sisteminti, kategorizuoti, klasifikuoti, o bet kokia atsiradusi dviprasmybė lingvistui atrodo tarsi didžiausia nesėkmė. Vis dėlto pastaruoju metu kalbininkai vis dažniau vartoja žodį „kontinuumas“, taip akcentuodami leksikos ir gramatikos ribų nykimą ir vis didėjančią semantinio-pragmatinio konteksto svarbą kalbos vieneto reikšmės identifikavimui. Ribų tarp leksikos ir gramatikos nykimą rodo ir tai, kad prancūzų kalboje evidencialumas laikomas ne tik gryna semantikos kategorija, kurios raiška – įvairūs leksiniai įterpiniai, dalelytės, prieveiksmiai, percepcijos ir kalbėjimo veiksmažodžiai, bet ir gramatinės konstrukcijos, kurių evidencines reikšmes 120 lemia veiksmažodžių laikai, nuosakos ir pan. (Barbet, Saussure 2012). Taigi prancūzų lingvistikoje evidencialumas laikomas semantine (konceptualiąja) gramatine kategorija, kurios paskirtis yra nurodyti kognityvinį arba (ir) komunikacinį autoriaus pasakymo pagrindą (Vetters 2012, 32). Palyginti su germanų ar neindoeuropiečių kalbomis, prancūzų kalboje evidencialumas yra žymiai mažiau tyrinėtas reiškinys, ir kol kas tvirto teorinio pagrindo nebuvimas iš dalies paaiškina tyrėjų nesutarimą dėl veiksmažodžio devoir ‘privalėti’ reikšmių ir funkcijų interpretavimo. Kai kurie lingvistai jį laiko tik modaliniu (Sueur 1979, 1983; Le Querler 1996, 2001), kiti (Kronning 1996, 2001) pusiau evidenciniu-modaliniu. Didžiausios diskusijos kyla dėl jo evidencinės ir episteminės reikšmių susipynimo (Dendale 1994; Tasmowski, Dendale 1994; Desclés, Guentchéva 2001; Dendale, Mulder 1996), o pastaruoju metu Rossari et al. (2007) skelbia, kad jis esąs grynai evidencinis. Atrodo, kad būtinos gilesnės diachroninės ir sinchroninės devoir ‘privalėti’ studijos; jų išvados galėtų patikslinti devoir ‘privalėti’ reikšmių šiuolaikinėje prancūzų kalboje aiškinimą. 4 Devoir ‘privalėti’ atitikmenys lietuvių kalboje Prancūzų grožinės literatūros pavyzdžių ir jų vertimų į lietuvių kalbą analizė rodo, kad devoir ‘privalėti’ atitikmenų raiška gali būti labai įvairi. Šio prancūzų kalbos veiksmažodžio deontinė reikšmė dažniausiai perteikiama į lietuvių kalbą veiksmažodžiais privalėti, turėti, reikėti, pavyzdžiui: (15) Mais, en réalité, il y a doute, et je dois vous isoler. (P263) privalėti.PRS.1SG izoliuoti.INF ‘Bet, tiesą sakant, šiek tiek abejoju ir privalau jus izoliuoti.’ (M227) (15) Mais, en réalité, il y a doute, et je ‘Bet, tiesą sakant, šiek tiek abejoju ir privalau jus izoliuoti.’ (M227) (16) A chaque fois je devais soulever le battant, placer privalėti.IPFV.1SG pakelti.INF la page avec minutie <...> (ST35) i k i k k l i d į k i i ld i l ( ) (16) A chaque fois je devais soulever le battant, placer privalėti.IPFV.1SG pakelti.INF la page avec minutie <...> (ST35) (16) A chaque fois je devais soulever le battant, placer privalėti.IPFV.1SG pakelti.INF la page avec minutie <...> (ST35) ’ (BD21) Kiekvieną kartą turėjau pakelti dangtį, kruopščiai paguldyti lapą< ...>’ (BD21 (17) Il s’ensuit que nous devons adopter chronologiquement privalėti.PRS.2PL taikyti.INF chronologiquement l’ordre que voici < ...> (HR123) ‘Iš to išplaukia, kad mums reikia nusistatyti tokią chronologinę tvarką <...>’ (RŽ69) Preliminarūs bandomojo tyrimo rezultatai leidžia daryti prielaidą, kad verčiant episteminę devoir ‘privalėti’ reikšmę į lietuvių kalbą vyrauja adverbialinė raiška. 121 Dažniausiai vartojami adverbialai greičiausiai, veikiausiai, turbūt ir kt. Keletas šią dominantę iliustruojančių pavyzdžių: (ST125) (19) Non, dit Lil, il doit avoir honte d’être amoureux. (HR135) privalėti.PRS.3SG gėdytis.INF ‘Ne, – nesutiko Lilė, – veikiausiai jis gėdijasi, kad įsimylėjo.’ (RŽ123) (19) Non, dit Lil, il doit i l ‘Ne, – nesutiko Lilė, – veikiausiai jis gėdijasi, kad įsimylėjo.’ (RŽ123) (20) J’avais dû m’endormir quand la sonnette retentit. (EN32) privalėti.IPFV.1SG užmigti.INF ‘Aš turbūt buvau užsnūdęs, ir tada sučirškė skambutis.’ (BN22) Kaip matyti, vertime į lietuvių kalbą informacija perteikiama tiksliai. Autoriaus abejonė, netikrumas išlieka, tik keičiasi jo kodavimo specifika: ten, kur prancūzų kalboje vartojamas episteminis devoir ‘privalėti’, lietuvių kalboje dominuoja įvairūs adverbialai (greičiausiai (18), veikiausiai (19), turbūt (20)). Įvairias kalbas tyrę van der Auwera ir Plugianas (1998) teigia, kad būtinumą ir privalėjimą reiškiantys modaliniai veiksmažodžiai linkę išrutulioti evidencinę nuogirdos (pranc. ouï-dire / information rapportée) reikšmę. Pasak jų, pavyzdžiui, olandų kalbos sakinyje (19) moeten ‘privalėti’, danų kalboje skulle ‘privalėti’ gali reikšti ne tik paties subjekto numanymą (pranc. inférence), bet ir iš kitų šaltinių išgirstą informaciją, t. y. nuogirdą. Priešingai nei germanų kalbose, prancūzų kalbos veiksmažodis negali turėti tokios reikšmės. Vetters’o (2012, 45) teigimu, atitinkamame prancūzų kalbos sakinyje doit être ‘turi būti’, negali koduoti nuogirdos reikšmės, įmanoma tik viena – inferencijos, pavyzdžiui: (21) Het moet een goede film zijn. (21) Het moet een goede film zijn. (21) (22) Cela doit être un bon film. privalėti.PRS.3SG būti.INF ‘Turbūt tai geras filmas.’ Pasakymo autorius daro tokias išvadas remdamasis asmeniniu skoniu ir numanymu, kuriam susiformuoti galėjo padėti nuogirda: daug girdėjau apie tą filmą ar pan. Prancūzų kalboje nuogirda gali būti reiškiama tariamąja nuosaka (conditionnel) (Kronning 2003). Tai itin ryškus prancūzų spaudos bruožas (pranc. conditionnel journalistique). 122 Tačiau Haillet’o (1995, 207) ir Dendal’io (1999, 16) nuomone, veiksmažodis devoir ‘privalėti’ neturi tokios savybės, t. y. net pavartotas tariamąja nuosaka jis neturi evidencinės nuogirdos reikšmės. Nuogirdos reikšmę šio veiksmažodžio tariamoji nuosaka įgyja tik tuomet, jei būna pavartota su kokiu nors įterptiniu vienetu, pvz., selon ‘pasak’, d’après ‘pagal’, kuris eksplicitiškai nurodo informacijos šaltinį (Vetters 2012, 45): (23) Selon le constructeur, ce nouveau moteur devrait Il est parti, ou il doit partir. (MB252) privalėti.PRS.3SG išvykti.INF ‘Kas man sakė? – jos šiurkštaus tono truputį nustebintas pakartojo jis. – Ogi Žiraras, kurį ką tik sutikau prie „Prancūzų kavinės“ durų. Kažkur išvykęs ar išvyksiąs.’ (PB235) Originalo kalbos sakinyje (26) kalbėjimo veiksmažodis dire ‘sakyti’ rodo, kad kalbėtojas perduoda ne savo pastebėjimus, o persako, t. y. pateikia kaip girdėtą iš kito žmogaus, informaciją (perpasakojimo reikšmė). Bet kadangi kalbama apie ateityje vyksiantį veiksmą, natūraliai susidaro netikrumo įspūdis – veiksmo dalyvis nėra tvirtai įsitikinęs pranešamo fakto validumu ar įgyvendinimu (numanymo reikšmė). Tokia dvejopa reikšmė – perpasakojimo ir numanymo – atspindi ir vertime į lietuvių kalbą. Kaip teigia Balkevičius (1998, 69), kalbėjimo ir mąstymo reikšmės bazinis veiksmažodis (verbum dicendi et sentiendi) suponuoja įspūdį, jog kalbantysis aktualizuotu dalyviu praneša tai, ką yra pasakęs ar bent įsivaizdavęs pats predikacijos denotatas (asmeninio sakinio veiksnys). Lietuviškame sakinyje abejojimo ir numanymo atspalvį dar labiau sustiprina neapibrėžtas prieveiksmis kažkur. (23) Selon le constructeur, ce nouveau moteur devrait (23) Selon le constructeur, ce nouveau moteur devrait (23) vartojimas.SG ‘Pasak inžinieriaus, šis naujas variklis šimtui kilometrų turėtų suvartoti dviem litrais mažiau.’ Vetters’o (2012, 33) ir kitų lingvistų (Dendale 1994; Desclés, Guentchéva 2001) teigimu, devoir ‘privalėti’ dažniausiai koduoja evidencinę numanymo reikšmę. Be jokios abejonės, prielaidos jai susidaryti gali būti nelygiavertės: (24) Du haut de ma tour de Babel, je regarde le parc d’Ueno et je vois des arbres enneigés: des cerisiers en fleurs –, ce doit être tai privalėti.PPR.3SG būti.INF Pâques. (ST85) ‘Iš savojo Babelio bokšto aukštybių žiūriu į Ueno parką ir matau apsnigtus medžius – taip, tai greičiausiai Velykos.’ (BD50) (24) Du haut de ma tour de Babel, je regarde le parc d’Ueno et je vois des arbres enneigés: des cerisiers en fleurs –, ce doit être tai privalėti.PPR.3SG būti.INF Pâques. (ST85) ‘Iš savojo Babelio bokšto aukštybių žiūriu į Ueno parką ir matau apsnigtus medžius – taip, tai greičiausiai Velykos.’ (BD50) (24) Du haut de ma tour de Babel, je regarde le parc d’Ueno et je vois des arbres enneigés: des cerisiers en fleurs –, ce doit être tai privalėti.PPR.3SG būti.INF Pâques. (ST85) ‘Iš savojo Babelio bokšto aukštybių žiūriu į Ueno parką ir matau apsnigtus medžius – taip, tai greičiausiai Velykos.’ (BD50) (25) Puisqu’il y a l’analphabétisme, il devrait y avoir privalėti.COND.3SG būti.INF l’anarythmétisme pour parler du drame de mon espèce. (ST88) ‘Kadangi yra analfabetų, matyt, yra ir anaritmetikų, – tai ypatingas ir dramatiškas žmonių tipas, kuriam ir priklausau.’ (BD53) (25) Pavyzdyje (24) apie galimą Velykų atėjimą spėjama remiantis vizualine percepcija ir enciklopedinėmis žiniomis, o (25) prielaida yra paremta loginiu samprotavimu (pranc. puisque nurodo priežastį, kuri žinoma pasakymo adresatui). Numanymo reikšmė prancūzų lingvistikoje dar skirstoma į tipus, o konkrečiam tipui nustatyti prancūzų kalbos tyrėjai remiasi premisų (in absentia arba in praesentia) validiškumu. Išskiriami abdukcinio ir dedukcinio numanymo tipai, tačiau jų analizė nėra šios studijos tikslas. Iš tikrųjų neretai sunku yra įvertinti, kokią funkciją sakinyje atlieka devoir ‘privalėti’ ir todėl nereikėtų apsiriboti vien tik izoliuoto sakinio lygmens analize ir bandyti įsprausti vieną ar kitą reikšmę į kurios nors vienos konkrečios kategorijos rėmus. Tai patvirtina 123 prancūziškas sakinys (26), kuriame galima įžvelgti ne vienos, bet dviejų devoir ‘privalėti’ reikšmių – perpasakojimo (pranc. reportative) ir numanymo – susipynimą. (26) – Qui me l’a dit? Répliqua-t-il un peu surpris de ce ton brusque; c’est Girard, que j’ai rencontré tout à l’heure à la Porte du café Français. 5 Lietuvių kalbos būtinumo lauko raiškos ir evidencinių konstrukcijų aktualizacija prancūziškuose atitikmenyse su devoir ‘privalėti’ Kaip parodė 4 skyrelyje nagrinėti prancūzų kalbos pavyzdžiai ir jų vertimai į lietuvių kalbą, deontinė devoir ‘privalėti’ reikšmė lietuvių kalboje dažniausiai perteikiama veiksmažodžiais privalėti, reikėti, turėti, o episteminė – adverbialais greičiausiai, matyt, turbūt ir kt. Analogiškos tendencijos išryškėjo ir analizuojant lietuvių kalbos pavyzdžius, kurie į prancūzų kalbą verčiami veiksmažodžiu devoir ‘privalėti’. Ar tokios daromos prielaidos bus validžios, bus galima vertinti tik atlikus tekstynais paremtą kokybinę ir kiekybinę analizę. Šiuo metu tokio tipo aprašomoji dvikryptė studija reikalinga tam, kad ne tik iš prancūzų kalbos perspektyvos, bet ir iš lietuvių kalbos pozicijų būtų atidžiau susipažinta su būtinumo lauko raiškos reljefiškumo ir evidencialumo problematika. Šių klausimų aptarimas, tikėtina, prisidės prie prancūzų kalbos veiksmažodžio devoir ‘privalėti’ semantikos ir funkcijų aiškinimo ir tolesnio kryptingo tyrimo. 124 Lietuvių kalbą tyrinėjusios Usonienė ir Jasionytė (2010) pastebėjo, kad gramatinami lietuvių kalbos veiksmažodžiai gauti, tekti gali koduoti tokią būtinybę, kai už susiklosčiusią situaciją veikėjas nėra atsakingas, o jam reikia atlikti ar išgyventi tai, kas priklauso nuo išorinių aplinkybių, pavyzdžiui: Lietuvių kalbą tyrinėjusios Usonienė ir Jasionytė (2010) pastebėjo, kad gramatinami lietuvių kalbos veiksmažodžiai gauti, tekti gali koduoti tokią būtinybę, kai už susiklosčiusią situaciją veikėjas nėra atsakingas, o jam reikia atlikti ar išgyventi tai, kas priklauso nuo išorinių aplinkybių, pavyzdžiui: (27) Juk teks šalty šovėti. (SŠ26) ‘Tu devras rester dans le froid.’ (PD312) privalėti.FUT.2SG (27) Juk teks šalty šovėti. (SŠ26) ‘Tu devras rester dans le froid.’ (PD312) privalėti.FUT.2SG (27) Juk teks šalty šovėti. (SŠ26) (28) Gavom įsimaišyti į juodai apsitaisiusių žmonių būrį. (PM165) (28) Gavom įsimaišyti į juodai apsitaisiusių žmonių būrį. (PM165) ‘Nous allions devoir nous mélanger à ce groupe de gens tout habillés privalėti.IPFV.1PL de noir.’ (AM163) Minimi būtinumą reiškiantys lietuvių kalbos veiksmažodžiai nėra lygiaverčiai, bet atitinkamame neigiamos konotacijos kontekste jie suponuoja situacijos dalyvių veiksmus, kurie neišvengiamai apspręsti išorinių veiksnių, todėl visiškai įmanomas tekti ir gauti vertimas į prancūzų kalbą su devoir ‘privalėti’. Prancūzų kalboje būtinumo raiška taip pat gali būti koduojama ne tik veiksmažodžiu devoir ‘privalėti’, bet ir être obligé de + bendratis ‘būti priverstam, įpareigotam’, falloir ‘reikėti’ ir kt., todėl tikėtinas dalykas, kad lietuvių kalboje jie taip pat gali būti tekti, gauti atitikmenimis. Dar vienas modalumo raiškos tipas lietuvių kalboje – bendraties būtinybės / reikiamybės konstrukcijos, kurios gali būti vartojamos ir be modalinio veiksmažodžio asmenuojamos formos (nulinis variantas). Tokiu atveju veiksmo subjektas dažnai žymimas naudininku, kuris savo ruožtu sustiprina tikslo reikšmę. 5 Lietuvių kalbos būtinumo lauko raiškos ir evidencinių konstrukcijų aktualizacija prancūziškuose atitikmenyse su devoir ‘privalėti’ Naudininkas man į prancūzų kalbą verčiamas asmeniniu įvardžiu je ‘aš’ ir dėl tokios prancūzų kalbos sintaksinės struktūros tampa būtina modalinio būtinybės žymiklio devoir ‘privalėti’ įterpimas prancūziškame atitikmenyje: (29) Kodėl man nerakinti durų? (SS10) ‘Pourquoi je ne dois pas fermer la porte ?’ (PD299) privalėti.PRS.1SG rakinti.INF (30) Tai ką man daryti? (SP274) ‘Qu’est-ce que je dois faire maintenant ?’ (PS58) privalėti.PRS.1SG daryti.INF (29) Kodėl man nerakinti durų? (SS10) ‘Pourquoi je ne dois pas fermer la porte ?’ (PD299) privalėti.PRS.1SG rakinti.INF (29) (30) Tai ką man daryti? (SP274) ‘Qu’est-ce que je dois faire maintenant ?’ (PS58) privalėti.PRS.1SG daryti.INF (30) Tarpkalbiniai tyrimai rodo, kad įvairių kalbų kitimo tendencijos yra panašios, bet jų absoliutus ekvivalentiškumas neįmanomas ne tik dėl pačių kalbų skirtingos prigimties 125 (šiuo atveju lietuvių kalba yra sintetinė, o prancūzų – analitinė), bet ir dėl to, kad kalbose pasikeitimai vyksta nevienodai. Pakitimai kiekvienoje kalboje jungiasi į savitą srautą, kuris keičia sintaksinių ryšių ir semantinių santykių kodavimo būdą, o tie santykiai jau reiškiami ne pačių žodžių formomis, o žodžių junginiais ar jų samplaikomis. Taigi žodžio, žodžio junginio ar jų samplaikos daugiareikšmiškumą neutralizuoti gali tik kontekstinė aplinka. Kontekstinės aplinkos įvertinimas labai svarbus nuostabos arba miratyvinių (pranc. mirative) reikšmių aktualizacijai. Prancūzų ir lietuvių kalbose jų statusas dar iki galo nėra išaiškintas (Vetters 2004, 46). Holvoet ir Semėnienė (2004, 111) linkę manyti, kad lietuvių kalboje miratyvinė evidencinių formų funkcija ypač ryški konstrukcijose su esamojo laiko dalyviais. Iki šiol atliktuose veiksmažodžio devoir ‘privalėti’ tyrimuose lingvistai niekur neužsimena apie jo potencialią miratyvinę reikšmę, tačiau tikrinamojo pobūdžio lietuvių kalbos klausimai, kurie, atrodytų, taip pat galėtų būti prilyginti miratyvinėms konstrukcijoms, į prancūzų kalbą dažnai verčiami pasitelkiant devoir ‘privalėti’. Tai leidžia daryti prielaidą, kad reiškiant subjekto nuostabą, devoir ‘privalėti’ galėtų būti viena iš funkcinių šios raiškos priemonių: <...> lyg ir smagus, tik kažkoks apžėlęs visas, o paakiai juodi. <...> lyg ir smagus, tik kažkoks apžėlęs visas, o paakiai juodi. <...> lyg ir smagus, tik kažkoks apžėlęs visas, o paakiai juodi. – Ar skutiesi jau? – lyg džiaugdamasi, lyg nuliūsdama sumirksėjo Morta. – Ar jau? Kaip čia taip dabar? (KR84) ‘<...> il avait l’air content, malgré sa barbe trop longue et ses yeux cernés. – Tu dois te raser, maintenant ? – dit Morta en clignant privalėti.PRS.2SG skustis.INF des yeux, comme si elle était partagée entre le contentement et l’inquiétude. Déjà?’ (GA87) (31) ‘<...> il avait l’air content, malgré sa barbe trop longue et ses yeux cernés. 5 Lietuvių kalbos būtinumo lauko raiškos ir evidencinių konstrukcijų aktualizacija prancūziškuose atitikmenyse su devoir ‘privalėti’ <...> il avait l’air content, malgré sa barbe trop longue et ses yeux cernés. des yeux, comme si elle était partagée entre le contentement et l’inquiétude. Déjà?’ (GA87) Lietuviškame pavyzdyje akivaizdus kalbėtojo noras pabrėžti, kad jam sunku patikėti tuo, ką fiksuoja akys. Šioje konstrukcijoje aiškiai atsispindi skirtumas tarp to, ką rodo informacijos šaltinis (tiesioginis matymas), ir paties veikėjo minčių. Be to, dalelytė ar, du kartus pavartotos dalelytės jau suteikia stebėjimosi ir nenoro patikėti matomu faktu atspalvį. Prancūzų kalbos vertime taip pat jaučiama labiau nuostabos nei spėjimo reikšmė, kurią sustiprina prieveiksmiai maintenant ‘dabar’ ir déjà ‘jau’. Identifikuojant devoir ‘privalėti’ reikšmę neretai remiamasi sintaksiniu-transformaciniu metodu, kuris kai kurių mokslininkų darbuose (Sueur 1983; Le Querler 1996, 2001) vis dar naudojamas kaip validus, tačiau naujausiuose tyrimuose jau pabrėžiama platesnio semantinių-pragmatinių parametrų konteksto reikšmė. Tokio konteksto svarbą puikiai iliustruoja šis pavyzdys: 126 (32) Ir nutilo Murzė, ir prie kauliuko sugrįžo, bet vis pagraužusi urzgė, rodydama dantis staklėms, ir vėl graužė, ir vėl irzo. (32) Ir nutilo Murzė, ir prie kauliuko sugrįžo, bet vis pagraužusi urzgė, rodydama dantis staklėms, ir vėl graužė, ir vėl irzo. (32) Ir Stepukas lyg paklausė, lyg pasakė: – Slapstai, ar ne? Juk slapstai ką... (KR93) ‘Tu caches quelqu’un, ou non? Tu dois privalėti.PRS.2SG quelqu’un <...>’ (GA93) Atrodytų, kad pavyzdyje (32) akivaizdi tikrinamojo pobūdžio klausimo interpretacija, bet remiantis kontekstine situacija galima įžvelgti jau ne nuostabos, o numanymo, paremto tiesioginiais percepciniais duomenimis, reikšmę, o tai yra inferencinio evidencialumo laukas. Tai tik keletas lietuvių kalbos būtinumo lauko raiškos ir evidencinių konstrukcijų aktualizacijos pavyzdžių prancūzų kalboje su veiksmažodžiu devoir ‘privalėti’. Tokia apžvalga tik dar kartą patvirtina faktą, kad veiksmažodis devoir ‘privalėti’ gali būti įvairių reikšmių kodavimo šaltinis ir vektorius, dėl to neabejotinai reikalingi tolesni išsamūs tyrimai. 6 Apibendrinamosios pastabos Teorinis diachroninis veiksmažodžio devoir ‘privalėti’ aptarimas rodo, kad gramatikalizacijos procese šis veiksmažodis turėjo patirti išties dinamišką reikšmių kaitą, ir tai, matyt, suponuoja jo semantinę polifoniją šiuolaikinėje prancūzų kalboje. Referuojamų autorių ir grožinės literatūros pavyzdžiai rodo, kad veiksmažodis yra išlaikęs pilną leksinę (skolos) reikšmę ir gali būti modalumo žymiklis. Prancūzų tyrėjai jam priskiria ištisą spektrą reikšmių, nuo deontinės (teorinės ir praktinės), vidinės būtinybės, episteminės, aletinės iki evidencinės. Remdamiesi diachroninėmis lotynų ir prancūzų kalbų studijomis lingvistai linkę nepritarti klasikinei modalumo kilmių teorijai (deontinis modalumas → epistemiškumas), o siūlo modelį leksinė reikšmė → deontinis modalumas → aletinis modalumas → epistemiškumas, kuris, jų manymu, geriau paaiškina devoir ‘privalėti’ reikšmę ir funkciją šiuolaikinėje prancūzų kalboje. Iš bandomosios tarpkalbinės studijos pavyzdžių apžvalgos ir individualia intuicija paremtų įžvalgų galima daryti prielaidą, kad iš prancūzų į lietuvių kalbą deontiniam devoir ‘privalėti’ aspektui perteikti gali būti pasirenkama veiksmažodinė strategija (gauti, privalėti, reikėti, tekti, turėti), o epistemiškumui reikšti – adverbialinė dominantė (greičiausiai, tikriausiai, turbūt, veikiausiai ir kt.). Ar tikrai daroma prielaida ir įžvalgos 127 yra pagrįstos ir teisingos tarpkalbinėje prancūzų-lietuvių kalbų erdvėje, bus galima spręsti atlikus išsamią tekstynu paremta devoir ‘privalėti’ ir jo vertimo atitikmenų kiekybinių ir kokybinių parametrų analizę. Reikėtų atkreipti dėmesį, kad kaip rodo jau atlikti kontrastyviniai anglų-lietuvių kalbų tyrimai (Usonienė, Šolienė 2010; Šolienė 2012), lietuvių kalboje episteminiam modalumui žymėti dažniausiai pasirenkama būtent adverbialinė raiška. Kadangi dėl pačių evidencinių (numanymo, nuogirdos, miratyvinės, perpasakojimo) devoir ‘privalėti’ reikšmių vis dar kyla diskusijų, ieškant šio veiksmažodžio lietuviškų atitikmenų būtina labai atidžiai įvertinti kuo platesnę semantinę-pragmatinę konteksto situaciją. Tiriant būtinumo raiškos variantus tarpkalbinėje perspektyvoje būtų naudinga ir reikalinga išsamiau panagrinėti beasmenines bendraties reikiamybės konstrukcijas bei kitus atvejus (être obligé de + bendratis ‘būti priverstam, įpareigotam’, falloir ‘reikėti’ ir kt.), palyginti šių konstrukcijų ir devoir ‘privalėti’ atitikmenis lietuvių kalboje. Ši bandomoji studija tik dar kartą patvirtina, kad daugiareikšmiam veiksmažodžio devoir ‘privalėti’ tyrimui būtina išsami kiekybinė ir kokybinė analizė. Šiuo metu kuriamas lygiagretusis dvikryptis prancūzų-lietuvių / lietuvių-prancūzų kalbų grožinės literatūros tekstynas, kurio medžiaga paremti tyrimai, be jokios abejonės, padėtų lietuvių vertėjams ir prancūzų kaip svetimosios kalbos dėstytojams geriau suprasti ir įvertinti ne tik veiksmažodžio devoir ‘privalėti’ reikšmę bei funkcijas tarpkalbinėje perspektyvoje, bet ir kitus būtinumo raiškos būdus abiejose kalbose. Gavalda, A. 2010. Guodėja. Vertė N. S. Pukinskaitė. Vilnius: Alma littera. Duomenų šaltiniai LKŽ Lietuvių kalbos žodynas. 1976. 10 tomas. Vinius: Mokslas. TLF Trésor de la langue française. 1979. CNRS. Paris: Gallimard. Santrumpų sąrašas Stiklo šalis. Vilnius: Lietuvos rašytojų sąjungos leidykla. ST Nothomb, A. 1999. Stupeur et tremblements. Paris: Albin Michelin. Santrumpų sąrašas AM Ivanauskaitė, J. 2003. L’année des muguets. Vertė M. Vitureau. Des âmes dans le brouillard. Anthologie de nouvelles lituaniennes contemporaines. Presses universitaires de Caen. AM Ivanauskaitė, J. 2003. L’année des muguets. Vertė M. Vitureau. Des âmes dans le brouillard. Anthologie de nouvelles lituaniennes contemporaines. Presses universitaires de Caen. BD Nothomb, A. 2005. Baimė ir drebėjimas. Vertė D. Gintalas. Vilnius: Alma littera. BD Nothomb, A. 2005. Baimė ir drebėjimas. Vertė D. Gintalas. Vilnius: Alma littera. BN Schmitt, E. E. 2006. Berniukas Nojus. Vertė S. Banionytė-Gervienė. Vilnius: Alma littera. BD Nothomb, A. 2005. Baimė ir drebėjimas. Vertė D. Gintalas. Vilnius: Alma littera. BN Schmitt, E. E. 2006. Berniukas Nojus. Vertė S. Banionytė-Gervienė. Vilnius: Alma littera. BN Schmitt, E. E. 2006. Berniukas Nojus. Vertė S. Banionytė-Gervienė. Vilnius: Alma littera. C Gavalda, A. 2008. La Consolante. Paris: Le Dilettante. C Gavalda, A. 2008. La Consolante. Paris: Le Dilettante. C Gavalda, A. 2008. La Consolante. Paris: Le Dilettante. CP Stendhal. 1838. La chartreuse de Parme. Editions du groupe «Ebooks libres et gratuits». CP Stendhal. 1838. La chartreuse de Parme. Editions du groupe «Ebooks libres et gratuits». EN Schmitt, E. E. 2004. L’enfant de Noé. Paris: Albin Michelin. G Gavalda, A. 2010. Guodėja. Vertė N. S. Pukinskaitė. Vilnius: Alma littera. 128 GA Meras, I. 2003. Le goût amer de l’oseille. Vertė M. Vitureau. Des âmes dans le brouillard. Anthologie de nouvelles lituaniennes contemporaines. Presses universitaires de Caen. HR Vian, B. 1950. L’herbe rouge. Paris: Ed. Jean-Jaques Pauvert. KR Meras, I. 2004. Kartus rūgštynių skonis. Stotelė vidukelėj. Vilnius: Lietuvos rašytojų sąjunga. P Camus, A. 1947. La peste. Editions du groupe «Ebooks libres et gratuits». M Kamiu, A. 1994. Maras. Vertė R. Jankevičiūtė. Vilnius: Valstybinis leidybos centras. MB Flaubert, G. 1857. Madame Bovary. Editions du groupe «Ebooks libres et gratuits». PB Flaubert, G. 1993. Ponia Bovary. Vertė S. Čiurlionienė. Vilnius: Onė. PD Juknaitė, V. 2003. Le pays de verre. Vertė A. Melkūnaitė ir L. Muhleisen. Des âmes dans le brouillard. Anthologie de nouvelles lituaniennes contemporaines. Presses universitaires de Caen. PM Ivanauskaitė, J. 1985. Pakalnučių metai. Vilnius: Tyto alba. PS Granauskas, R. 2003. Un papillon sur la bouche. Vertė M. Barakauskaitė. Des âmes dans le brouillard. Anthologie de nouvelles lituaniennes contemporaines. Presses universitaires de Caen. PV Stendalis. 1994. Parmos vienuolynas. Vertė R. Ramunienė. Kaunas: Varpas. RŽ Vian, V. 2007. Raudona žolė. Vertė S. Banionytė-Gervienė. Vilnius: Žar SP Granauskas, R. 2000. Su peteliške ant lūpų. Vilnius: Presvika. SŠ Juknaitė, V.1995. Literatūros sąrašas Balkevičius, Jonas. 1998. Lietuvių kalbos predikatinių konstrukcijų sintaksė. Vilnius: Mokslo ir enciklopedijų leidybos institutas. Barbet, Cécile. 2008. Pragmatique des verbes modaux en français: aspects synchroniques diachroniques. Interneto prieiga: www2.unine.ch/.../ProjetDeTheseCecileBarbet. pdf. Žiūrėta: 2013-05-05. 129 Barbet, Cécile. 2012. Devoir et pouvoir, des marqueurs modaux ou évidentiels? Langue française 173, 49–63. Barbet, Cécile et Louis de Saussure. 2012. Présentation: modalité et évidentialité en français. Langue française 173, 3–12. Bertocchi, Alessandra et Anna Orlandini. 2001. L’expression de la possibilité épistémique en latin. Cahiers Chronos 8, 47–65. Buridant, Claude. 2000. Grammaire nouvelle de l’ancien français. Paris: Sedes. Bybee, Joan, Revere Perkins, William Pagliuca. 1994. The evolution of Grammar: Tense, Aspect and Modality in the Languages of the Word. Chicago: University of Chicago Press. Chu, Xiaoquan. 2008. Les verbes modaux du français. Paris: Ophrys. Dabartinės lietuvių kalbos gramatika (DLKG). 2006. Vilnius: Mokslo ir enciklopedijų leidybos institutas. Dendale, Patrick. 1994. Devoir épistémique, marqueur modal ou évidentiel? Langue française 102, 24–40. Dendale, Patrick et Walter de Mulder. 1996. Déduction ou abduction: le cas de devoir inférentiel. L’énonciation médiatiseé Z. Guentchéva (sous la direction de). Paris/ Louvain: Peeters. 305–318. Dendale, Patrick. 1999. Devoir au conditionnel: valeur évidentio-modale et origine du conditionnel. Cahiers Chronos 4, 7–28. Desclés, Jean P., Zlatka Guentchéva. 2001. La notion d’abduction et le verbe devoir épistémique. Cahiers Chronos 8, 103–122. Haillet, Patrick. 1995. Le conditionnel dans le discours journalistique. Québec, Neuville: Bref. Holvoet, Axel, Loreta Semėnienė. 2004. Gramatinių kategorijų tyrimai. Lietuvių kalbos gramatikos darbai 2. Vilnius: Lietuvių kalbos institutas. Huot, Hélène. 1974. Le verbe devoir. Etude synchronique et diachronique. Paris: Klincksieck. Kronning, Hans. 1990. Modalité et diachronie: du déontique à l’épistémique. L’évolution sémantique du verbe debere/devoir. Actes du 11 Cogrès des Romanistes Scandinaves, 13-17 août 1990, Université de Trondheim. 301–312. Kronning, Hans. 1996. Modalité, cognition et polysémie: sémantique du verbe modal devoir. Uppsala/Stockholm: Almqvist & Wiksell International. Kronning, Hans. 2001. Pour une tripartition des emplois du modal devoir. Cahiers Chronos 8, 67–84. Le Querler, Nicole. 1996. Typologie des modalités. Caen: Presses Universitaires de Caen. Le Querler, Nicole. 2001. La place du verbe modal pouvoir dans une typologies des modalités.Cahiers Chronos 8, 17–32. 130 Marcello-Nizia, Christiane. 2006. Grammaticalisation et changement linguistique. Bruxelles: De Boek (Coll. Champs linguistiques). Miliūnaitė, Rita. 2005-11-16 skaitytas pranešimas: Vaikų ir jaunimo literatūros vertimas. Interneto prieiga: http://www.llvs.lt/?recensions=37&comments. Žiūrėta: 2013- 05-05. Palmer, Franc R. 1986. Mood and modality. Cambridge University Press, Cambridge. Papafragou, Anna. 1998. Inference and word meaning: The case of modal auxiliaries. Literatūros sąrašas Lingua 105, 1–47. Rossari, Corinne, Corina Cojocariu, Claudia Ricci, Adriana Spiridon. 2007. Devoir et l’évidentialité en français et en roumain. Interneto prieiga: http://discours.revues. org/116http://discours.revues.org/116. Žiūrėta: 2013-05-05. Sueur, Jean P. 1975. Etude sémantique et syntaxique des verbes devoir et pouvoir. Thèse de doctorat. Université de Paris X-Nanterre. Sueur, Jean P. 1979. Etude sémantique et syntaxique des verbes devoir et pouvoir. Le français moderne 2, 97–120. Sueur, Jean P. 1981. Les verbes modaux, sont-ils ambigus David, J. & G. Kleiber, éds. La notion sémantico-logique de modalité. Paris: Klincksieck. 165–180. Šolienė, Audronė. 2012. Epistemic necessity in a parallel corpus: Lithuanian vs. English. Usonienė, Aurelija, Nicole Nau, Ineta Dabašinskienė, eds. Multiple perspectives in linguistic research on Baltic languages. Newcastle upon Tyne: Cambridge Scholars Publishing. 10–42. Šolienė, Audronė. 2012. Epistemic necessity in a parallel corpus: Lithuanian vs. English. Tasmowski, Liliane et Patrick Dendale. 1994. Pouvoir, un marqueur d’évidentialité. Langue française 102, 41–55. Usonienė, Aurelija & Audronė Šolienė. 2010. Choice of strategies in realizations of epistemic possibility in English and Lithuanian. A corpus-based study. International Journal of Corpus Linguistics 15(2), 291–316. Usonienė, Aurelija & Erika Jasionytė. 2010. Towards grammaticalization: Lithuanian acquisitive verbs gauti (‘get’) and tekti (‘be gotten’). Acta Linguistica Hafniensia 42(2), 199–220. van der Auwera, Johan & Vladimir Plungian. 1998. Modality’s semantic map. Linguistic Typology 2, 79–124. Vetters, Carl. 2004. Les verbes modaux pouvoir et devoir en français. Revue belge de Philologie et d’Histoire 82, 657–671. Vetters, Carl. 2012. Modalité et évidentialité dans pouvoir et devoir: typologie et discussions. Langue française 173, 31–47. Wright, Georg H. 1963. Norm and action. A logical Enquiry. London: Routledge & Kegan Paul. 131 Vita Valiukienė Vita Valiukienė Résumé Le présent article vise à présenter la «plurivocité» sémantique du verbe devoir en français contemporain et sa réalisation en lituanien. La perspective interlangagière lituanien vers français est aussi prise en considération. En français contemporain le verbe devoir a un emploi lexical plein (renvoyant à la notion de dette) et tout un faisceau de valeurs modales. Un rapide regard diachronique suggère que l’emploi déontique est plus ancien que celui relevant de la modalité épistémique. La tradition linguistique française débat du modèle suivant: valeur lexicale → modalité déontique → modalité aléthique → épistémicité. L’interprétation de devoir dépend en grande partie du contexte d’où résulte un grand nombre d’effets évidentiels dont la nature fait encore débat parmi les linguistes. On a constaté que le français ne connaît pas d’emploi évidentiel d’information rapportée pour devoir et que la plupart de ses emplois évidentiels concernent l’inférence. Toutefois, dans bien des cas le contexte reste un facteur crucial pour lever certaines indéterminations et décider de quel type de construction évidentielle il s’agit. Une étude contrastive pilote révèle que la valeur déontique du verbe français devoir peut être traduite vers le lituanien à l’aide des verbes de nécessité gauti, privalėti, reikėti, tekti, turėti. La stratégie adverbiale est plutôt réservée à la traduction de son emploi épistémique: les adverbes greičiausiai, tikriausiai, turbūt, veikiausiai deviennent leurs équivalents en lituanien. Etant donné la nature différente des deux langues, l’équivalence absolue des valeurs est impossible. Toutefois ces évaluations n’éliminent point la nécessité de mener des recherches basées sur un corpus représentatif et d’analyser de manière approfondie le champ sémantique et fonctionnel de l’obligation et de la nécessité dans la perspective interlangagière. Įteikta 2013 m. balandžio mėn. Įteikta 2013 m. balandžio mėn. Įteikta 2013 m. balandžio mėn. 132
https://openalex.org/W4220747471	https://www.researchsquare.com/article/rs-1482012/latest.pdf	English	null	Prevalence of Preterm Premature Rupture of Membrane and Associated Factors Among Pregnant Women Admitted To Health Facilities in Ambotown, Ethiopia, 2021.	Research Square (Research Square)	2,022	cc-by	6,636	Prevalence of Preterm Premature Rupture of Membrane and Associated Factors Among Pregnant Women Admitted To Health Facilities in Ambotown, Ethiopia, 2021. Rebuma Muleta Gutema (  rebummul7@gmail.com ) Rebuma Muleta Gutema (  rebummul7@gmail.com Ambo University Gurmesa Daba Dina Ambo University Anisa Berhanu Ambo University Maru Mossisa Erena Ambo University Research Article License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Page 1/17 Page 1/17 Abstract Background: Preterm pre-labor rupture of the membrane is a major cause of perinatal, neonatal, and maternal morbidity and mortality both in high- and low-income countries. A woman with premature rupture of membranes is at risk of complications like intra-amniotic infection, postpartum hemorrhage, and death. Little is known about the problem in the study area, therefore, this study was designed to determine the prevalence of preterm premature rupture of membrane and its associated factors among pregnant women admitted to health facilities in ambo town. Methods: Hospitals based cross sectional study was conducted on 391 pregnant women who were admitted to the Hospitals in Ambo town from July 05/2021-August 30/2021. The data were collected through face-to-face interviews by a structured questionnaire. Descriptive analyses were performed using frequencies, percentages, binary logistic regression by adjusting for confounding factors. Results: The prevalence of preterm premature rupture of membrane was found to be 22.6%. Preeclampsia (AOR=3.2, 95% CI=0.69-0.57), Economic status (AOR = 2.64 (95% CI = 1.99–6.01), current urinary tract infection (AOR = 2.42, 95% CI = 1.32–5.19), previous history of premature rupture of membrane (AOR = 2.31, 95% CI = 1.02–6.27), andanemia (AOR = 1.85, 95% CI = 0.65–4.56) were factors associated with preterm premature rupture of membrane. Conclusions: The prevalence of preterm premature rupture of membrane in the study area was high. Based on the predisposing factors identifiedearly screening and treatment as well as health promotion isimportant to reduce the risk of preterm premature rupture of membrane. Background It also increases neonatal resuscitation, cord compression which leads to fetal distress and neonatal sepsis (24, 25). In low-income countries, more than 90% of extremely preterm newborns (less than 28 weeks of gestation) die during the first few days of life; in high-income countries, less than 10% of extremely preterm babies die within the first few days of life (26, 27). PPROM which is remote from term should only be cared for in facilities where a Neonatal intensive care unit (NICU) is available and capable of caring for the premature neonate. This is a challenge particularly in developing countries like Ethiopia where most of the health facilities are not well equipped. This studyaims to identify the related factors and helps to establish the comprehensive plan to assist the country in reducing the prevalence of PPROM.It also offers the community about the severity of the problem and informs them about how they may prevent the problem and act carefully to make it less difficult than predicted. In addition, the study may be useful to other researchers as a reference material for future studies on similar issues. Even if few studies with fewer variables have been conducted in some parts of Ethiopia, there is no study carried out in the study area.So, this study plans to assess the prevalence of preterm premature rupture of membrane and associated factors among admitted pregnant mothers in Ambo Health facilities. The result of this study is important in improving maternal health and community awareness, as well as its treatment and prevention. and maternal depression are considered as risk factors (13). The effect of PPROM ranges from maternal and neonatal morbidity and mortality (14).It complicates 3-4.5% of pregnancies worldwide (4, 15). Among women with preterm PROM, clinically evident intra-amniotic infection occurs in approximately 15–25%, and postpartum infection occurs in approximately 15–20% of the cases (16).The incidence of infection is higher at earlier gestational ages, and life-threatening maternal infections complicate expectant management of prevailed PROM (13, 17). The primary complication for the mother is infection which can leads to chorioamnionitis, placental abruption, psychological effect, lactation problem, disseminated intravascular coagulopathy, metritis after delivery, increase the need for operative deliveryand economic and resource wastage (18). PPROM raises perinatal mortality by four times and neonatal morbidity by three times (19, 20). Background The longer the time between rupture of the membrane and delivery, the higher the risk of maternal and fetal morbidity and mortality (21). Prematurityis responsible for 80–90% of perinatal mortality due to complications like Respiratory distress syndrome, Birth injury, Intraventricular hemorrhage, Hypocalcaemia, hypoglycemia, Hypothermia, Hyperbilirubinaemia and Infection (11, 22, 23). It also increases neonatal resuscitation, cord compression which leads to fetal distress and neonatal sepsis (24, 25). In low-income countries, more than 90% of extremely preterm newborns (less than 28 weeks of gestation) die during the first few days of life; in high-income countries, less than 10% of extremely preterm babies die within the first few days of life (26, 27). PPROM which is remote from term should only be cared for in facilities where a Neonatal intensive care unit (NICU) is available and capable of caring for the premature neonate. This is a challenge particularly in developing countries like Ethiopia where most of the health facilities are not well equipped. This studyaims to identify the related factors and helps to establish the comprehensive plan to assist the country in reducing the prevalence of PPROM.It also offers the community about the severity of the problem and informs them about how they may prevent the problem and act carefully to make it less difficult than predicted. In addition, the study may be useful to other researchers as a reference material for future studies on similar issues. Even if few studies with fewer variables have been conducted in some parts of Ethiopia, there is no study carried out in the study area.So, this study plans to assess the prevalence of preterm premature rupture of membrane and associated factors among admitted pregnant mothers in Ambo Health facilities. The result of this study is important in improving maternal health and community awareness, as well as its treatment and Background Preterm premature rupture of membrane (PPROM) is defined as rupture of the membrane before the beginning of labor in pregnancies that are between 28 and 37 weeks of gestation (1). It is distinguished by a painless liquid flow and other investigations like the ferning and pooling test. The lower limit of gestational age to describe Preterm in Ethiopia is a pregnancy that is > 28 weeks of gestation and < 37 weeks of gestational age(2, 3). Membrane or bag of water is the sac that maintains amniotic fluid which covers the developing baby. This fluid is important for nutrition, in defending the fetus against infection, fetal trauma, and compression of the umbilical cord(1).The magnitude of PPROM differs across the worldfrom country to country and the population.For instances evidences shows that PPROM accounts 2.2% in Manipur, India, 3.1% in Brazil, 2.3% in Canada, 19.2% in China, 3.3% in Nigeria and 7.5% in Uganda(4–9). In Ethiopia According to research conducted at Debre Tabor and Tikur Anbesa hospital, the prevalence of PPROM was 13.67% and 1.4% respectively(10–12).Even though the exact cause of PPROM is unknown; a structural defect in the membrane, maternal ethnic origin, previous preterm delivery, intrauterine infection at an early gestational age, low socioeconomic status of the women, inadequate prenatal care, smoking, UTI during pregnancy, maternal nutritional status, abnormal vaginal discharge, Page 2/17 Page 2/17 and maternal depression are considered as risk factors (13). The effect of PPROM ranges from maternal and neonatal morbidity and mortality (14).It complicates 3-4.5% of pregnancies worldwide (4, 15). Among women with preterm PROM, clinically evident intra-amniotic infection occurs in approximately 15–25%, and postpartum infection occurs in approximately 15–20% of the cases (16).The incidence of infection is higher at earlier gestational ages, and life-threatening maternal infections complicate expectant management of prevailed PROM (13, 17). The primary complication for the mother is infection which can leads to chorioamnionitis, placental abruption, psychological effect, lactation problem, disseminated intravascular coagulopathy, metritis after delivery, increase the need for operative deliveryand economic and resource wastage (18). PPROM raises perinatal mortality by four times and neonatal morbidity by three times (19, 20). The longer the time between rupture of the membrane and delivery, the higher the risk of maternal and fetal morbidity and mortality (21). Prematurityis responsible for 80–90% of perinatal mortality due to complications like Respiratory distress syndrome, Birth injury, Intraventricular hemorrhage, Hypocalcaemia, hypoglycemia, Hypothermia, Hyperbilirubinaemia and Infection (11, 22, 23). Methods Study area and design Study area and design Page 3/17 An institution-based cross-sectional study was conducted in Ambo town, Oromia Regional state, from July 05/2021-August 30/2021. The town is located 114 kilometers from Addis Ababa the capital city of Ethiopia. The total population of this town is estimated to be 83,053 of whom 41,692 are men and 41,361 women according to the 2007 census. The town has four governmental health institutions. These include Ambo University referral hospital, Ambo general hospital, Awaro health center, and Ambo health center. Ambo University referral hospital and Ambo general hospital are included in the study. Since its establishment in July 1999Ambo General Hospital is providing obstetrics and gynecology department in Page 3/17 Page 3/17 addition to other services. The department has 19 beds with 3 delivery Kochs. Ambo University referral hospital was established in 2006 and currently provides Gynecology, labor, and delivery in addition to other services with 38 beds (14 labor and delivery ward, 24 gynecology wards) and 4 deliveries Koch’s. The Selected pregnant women who were admitted in obstetrics wards of AGH and AURH at their gestation between 28 and 36 + 6weeks were the study population. Preterm Premature rupture of membrane Preterm Premature rupture of membrane Independent variable Socio demographic status, Past and current obstetric characteristics and Medical and behavioral characteristics. Sample size and Sampling technique The Sample size was calculated by using a single population proportion formula by considering P = 0.137; taken from previous Research done at Debre Tabor (11), 95% confidence level, 5% desired degree of accuracy. By adding a 10% non-response rate and using a design effect of 2, the final calculated sample size was found to be 391. Systematic random sampling was used to select study The Sample size was calculated by using a single population proportion formula by considering P = 0.137; taken from previous Research done at Debre Tabor (11), 95% confidence level, 5% desired degree of accuracy. By adding a 10% non-response rate and using a design effect of 2, the final calculated sample size was found to be 391. Systematic random sampling was used to select study participants.The average number of pregnant women who were admitted in Ambo General Hospital and Ambo University Referral Hospital during the data collection period was estimated based on the previous admission, which was obtained by referring to a six-month registration book/record before data collection. Around 755 pregnant women were admitted in labor, maternity, and high-risk wards in six months. The sampling interval (kth unit) was obtained by dividing the entire total pregnant women admitted in six months (755) by the desired sample size (391) and it was approximately 2. The first woman was randomly chosen for the survey by the lottery method, and then every second woman who was admitted in the ward was selected for the study. DATA PROCESSING AND ANALYSIS The collected data were entered into Epi data version 3.3.1 software after coding and checking the completeness and exported to statistical package for social science SPSS version 20 for analysis. Descriptive analysis using frequencies, means, percentage, and standard deviations were done and presented in text and tables. Logistic regression analyses adjusting for potential confounding factors were used to see the association between the preterm premature rupture of membrane and the explanatory variables which had a P-value less than 0.25. Finally, the strength of association was weighed using an odds ratio at 95% confidence interval and P-value < 0.05. ETHICAL CONSIDERATION Ethical clearance has been obtained from the department of midwifery Ambo university research and community service coordinator. Letter of permission was soughed from the West Shoa Zonal health department and from each health institution. Verbal consent was taken from the participants after the data collectors clarified the objectives of the study, processes, and their right to refuse not to participate at any time. Furthermore, the confidentiality of the study participants was assured. Data collection tools, procedure, and quality control The questionnaire was initially prepared in English version by reviewing different related literature (5, 10- 13). Then the questionnaire translated to Afan Oromo an area language and then translated back to English by language experts by maintaining consistency. The questionnaire includes Socio-demographic characteristics, historical and current obstetrical gynecological history, medical history, and behavioral factors. Medical and obstetric data that could not be accessed by interviews, such as gestational age, diagnosis of PPROM, urinary tract infections, STIs, and anemia were collected from patient medical Page 4/17 Page 4/17 records and charts. In addition, each woman’s MUAC was measured using nonelastic and non-stretchable MUAC tapes at the midpoint between the tips of her shoulder and elbow on her left arm. Data were collected through face-to-face interviews by using structured pretested questionnaires from women. The data collection was carried out by five trained degree holders’ health care providers and supervised by 1 Master holder. Two-days training were given for the data collectors. The collected data were reviewed and checked for consistency, clarity; completeness, and accuracy throughout the data collection process. OPERATIONAL DEFINITIONS Preterm Premature rupture of membrane (PPROM)-Preterm premature rupture of membrane (PPROM) is defined as membrane rupture before the beginning of labor in pregnancies that are less than 37 weeks of gestation. Preterm Premature rupture of membrane (PPROM)-Preterm premature rupture of membrane (PPROM) is defined as membrane rupture before the beginning of labor in pregnancies that are less than 37 weeks of gestation. Preterm-Preterm or premature birth describes neonates who are born too early. Preterm-Preterm or premature birth describes neonates who are born too PROM-is rupture of fetal membranes at least an hour before the onset of labor Preterm labor- Labor occurring after 28 weeks of gestation but before 37 completed weeks of gestation. Anemia: A pregnant woman whose Hemoglobin level is<11 gm/dl. DISSEMINATION OF THE RESULTS R l The final result of the study was submitted to Ambo University College of medicine and health science, department of midwifery and for the AGH and AURH after it was completed and presented to the department of midwifery by hard copy and soft copy. Results Page 5/17 A total of 391selected pregnant women participated in the study. The mean age of respondents was 25.23 (SD + 4.05). The majority of the participants (95.2) were married and almost all (98.1%) were Oromo and 331(84.8%) respondents were in the age range of 20–30-years. The majority (41.9%), of the participants, were attended secondary school and above. Regarding the occupational status102 (26.0%) women were governmental employers, 299(69.0%) participants’ mid-upper arm circumference measurements were greater than or equal to 23 cm (Table 1). A total of 391selected pregnant women participated in the study. The mean age of respondents was 25.23 (SD + 4.05). The majority of the participants (95.2) were married and almost all (98.1%) were Oromo and 331(84.8%) respondents were in the age range of 20–30-years. The majority (41.9%), of the participants, were attended secondary school and above. Regarding the occupational status102 (26.0%) women were governmental employers, 299(69.0%) participants’ mid-upper arm circumference measurements were greater than or equal to 23 cm (Table 1). A total of 391selected pregnant women participated in the study. The mean age of respondents was 25.23 (SD + 4.05). The majority of the participants (95.2) were married and almost all (98.1%) were Oromo and 331(84.8%) respondents were in the age range of 20–30-years. The majority (41.9%), of the participants, were attended secondary school and above. Regarding the occupational status102 (26.0%) women were governmental employers, 299(69.0%) participants’ mid-upper arm circumference measurements were greater than or equal to 23 cm (Table 1). Page 6/17 Table 1 Socio-demographic Characteristics of the respondents on the prevalence of preterm premature rupture of membrane and its associated factors among pregnant women attending health facilities in West Shoa Zone, Ethiopia, 2020. (n = 391). Socio-demographic Characteristics of the respondents on the prevalence of preterm premature rupture of membrane and its associated factors among pregnant women attending health facilities in West Shoa Zone, Ethiopia, 2020. (n = 391). pregnant women attending health facilities in West Shoa Zone, Ethiopia, 2020. (n = 391). Reason for Admission of the participants. Reason for Admission of the participants. Reason for Admission of the participants. Nearly one-fourth (108 (27.6%)) of respondents were admitted due to pre-eclampsia/eclampsia, 91 (23.2%) and 89(22.6%) of respondents were admitted for the indications of oligohydraminous and PPROM respectively (Fig. 1). Results Variables Frequency (%) Respondents Age (years) < 20 33 8.4 20–30 331 84.8 31+ 27 6.8 Marital Status Single 8 1.9 Married 372 95.2 Divorced 9 2.3 Widowed 2 0.6 Ethnicity Oromo 383 98.1 Amhara 2 0.6 Others* 6 1.3 Religion Orthodox 187 47.7 Muslim 79 20.3 Protestant 110 28.1 Others** 15 3.9 Educational Status No formal education 135 34.5 Primary education 92 23.5 Secondary and above 164 41.9 Occupational status governmental employer 102 26.0 self-employment 82 20.9 Daily labor 98 25.0 house wife 86 22 Students 23 5.8 Monthly income <= 1000.00 150 38.3 Others-Tigre, SNNP. Others- (wakefata and adventist) Page 7/17 Variables Frequency (%) 1001.00–3000.00 161 41.2 3001.00+ 80 20.5 MUAC < 23 92 31.0 ≥ 23 299 69.0 Residence Urban 246 62.9 Rural 145 37.1 Others-Tigre, SNNP. Others*- (wakefata and adventist) (%) Past and current obstetric characteristics The majority of the participants (45.8%) were Multigravida, regarding gestational age 319(81.6%) were between 34–36 weeks and nearly all of the respondents (97.4%) were had ANC follow up, 72(23.2%) respondents had the previous history of PROM and 68 (21.9%) had a past history of preterm delivery. From the selected participants 251 (64.2%) of the current presentation was cephalic, 63 (16.1%) have had urinary tract infection currently while 53(13.5%) have vaginal discharge (Table 2). Page 8/17 Page 8/17 Table 2 Past and current obstetric characteristics of respondents on the prevalence of preterm premature rupture of membrane and its associated factors among pregnant women attending health facilities in West Shoa Zone, Ethiopia, 2020. (n = 391). Past and current obstetric characteristics of respondents on the prevalence of preterm premature rupture of membrane and its associated factors among pregnant women attending health facilities in West Shoa Zone, Ethiopia, 2020. (n = 391). preterm premature rupture of membrane and its associated factors among pregnant women attending health facilities in West Shoa Zone, Ethiopia, 2020. (n = 391). Variables Frequency (%) Gravidity Primigravida 145 37.1 Multigravida 179 45.8 Grand multigravida 67 17.1 Gestational age (in weeks) 29–33 72 18.4 34–36 319 81.6 ANC follow-up Yes 381 97.4 No 10 2.6 History of previous PROM Yes 91 23.2 No 300 76.8 History of preterm birth Yes 135 21.9 No 92 78.1 Presentation Cephalic 251 64.2 Breech 129 32.9 Shoulder 11 2.9 UTI current pregnancy Yes 63 16.11 No 328 83.88 Abnormal Vaginal discharge Yes 53 13.5 No 338 86.5 GDM Yes 42 10.6 No 349 89.4 PPROM Yes 89 22.7 No 302 77.3 Lifting heavy objects Yes 8 2.6 No 383 97.4 Falling in accident Yes 5 1.0 Page 9/17 Variables Frequency (%) No 386 99.0 Anemia Yes 19 4.8 No 372 95.1 Factors of PPROM Variables Frequency (%) No 386 99.0 Anemia Yes 19 4.8 No 372 95.1 actors of PPROM Discussion The prevalence of preterm premature rupture of membrane (PPROM) in this study was found to be 22.6%. This finding was higher than the studies conducted in, tertiary care center in India (2.01%), the Rio Grande in Brazil (3.1%), in China (19.2%), Kampala International University teaching hospital in Uganda (7.5%) and Debre Tabor (13.67%) (4, 5, 7, 9, 11). The difference might be due to reduced standard quality of health care, low economic status, and social life in developing countries. Since the sample of this study was from high-risk populations/admitted pregnant woman rather than those who are in the community this may also increase the magnitude of PPROM.Women with preeclampsia were three times more likely to develop PPROM than those who did not have preeclampsia. This finding is also supported by some studies conducted in Uganda, Lithuania, and China (9, 28, 29). In preeclampsia, reactive oxygen species which are generated by oxidative stress, and some pathological conditions that develop during pregnancy and are related to hypoxic stress can affect the elevation of S100B(Anti-S100 beta antibody) concentration in the amnion and alter production and/or clearance of prolactin from the maternal compartment that can bring premature rupture of membrane (30–32). Urinary tract infection was another factor associated with the development of PPROM. Pregnant women with a history of UTI in pregnancy were two times more likely to get PPROM than those who did not have UTI. This finding was consistent with the study performed by Singh et al and dagne et al. (2, 11).This might be Elevations of inflammatory mediators such as prostaglandins, cytokines, and proteinases in the local tissue play a causative role in the disruption of fetal membrane integrity and in triggering uterine contractility. They are produced as a part of the physiologic maternal defense mechanism in response to pathogens’ invasion. The inflammatory mediators and production of matrix-degrading enzymes and TNFs are involved in mechanisms of PPROM (33).In the current study, those women whose monthly income was less than one thousand Ethiopian Birr /or less than twenty dollars were two times more at risk of developing PROM than those who earn more than three thousand Ethiopian Birr/sixty dollars.A study conducted in Northern Ethiopia supplemented this finding (11).This might be due to less income results nutritional deficiency and the mother becoming undernourished. Risk Factors of PPROM Gravidity, preeclampsia, current urinary tract infection, history of previous preterm premature rupture of membrane, economic status, and being anemic are significantly associated with preterm premature rupture of membrane. Being Grand multigravida predisposes for PPROM is nearly five times more likely than Primigravida (AOR = 5.30, 95%( CI: 2.07, 13.52). Furthermore, women who had Preeclampsia were three times and those who have a history of Previous PROM two times more likely to develop PPROM than those who do not have respectively (AOR = 3.2, 95% (CI: 1.47–7.04) (AOR = 2.31 (,95% (CI: 1.02– 6.27). A woman whose economic status was less than one thousand Ethiopian birrs was two times more likely to develop PPROM than those who earn more than three thousand Ethiopian birrs (AOR = 2.14, 95% (CI: 0.89–5.41). Women who had anemia were nearly two times more likely to develop PPROM as compared to those who are not (AOR = 1.85, 95% (CI: 0.65–4.56) (Table 3). Page 10/17 Table 3 Table 3 Bivariate and multivariable association of respondents on prevalence of preterm premature rupture of membrane and its associated factors among pregnant women attending health facilities in west Shoa Zone, Ethiopia, 2020. (n = 391). Preterm PROM Variables Yes No COR AOR Gravidity Primigravida 30(20.7%) 115(79.3%) 6.05(3.01– 14.02) 5.3(2.07– 13.52) Multigravida 41(23%) 138(77%) 1.41 (0.59– 2.87) 1.78 (1.64– 5.92) Grand multigravida 18(26.8%) 49(73.1%) 1 1 economic status <= 1000.00 28(18.7%) 122(81.3%) 1 1 1001.00– 3000.00 38(23.6%) 123(76.3%) 1.02 (0.48– 3.36) 1.27 (0.62– 3.74) 3001.00+ 24(30%) 56(70%) 1.87 (0.60– 3.90 2.14(0.89– 5.41)* Gestational DM Yes 21(50%) 21(50%) 0.26(0.10– 0.69) 0.2(0.69 − 0.57) No 77(22.5%) 272(77.5%) 1 1 Preeclampsia Yes 13(12.03%) 95(87.9%) 2.83(1.41– 5.69) 3.2(1.47– 7.04) No 59(27.3%) 157(72.7%) 1 1 Previous PROM Yes 34(37.5%) 57(62.5) 2.53 (1.25– 5.78) 2.31(1.02– 6.27)* No 54(18.0%) 246(82%) 1 1 Current UTI Yes 24(38%) 39(62%) 3.04 (1.43– 5.23) 2.62 (1.32– 5.19)∗ No 64(19.6%) 264(80.4%) 1 1 Lifting heavy objects Yes 25(58.8%) 18(41.2%) 0.15(0.73 − 0.32) 0.76(0.03– 0.18) No 63(18.1%) 285(81.9%) 1 1 Anemia Yes 10(52.6%) 9(47.4) 2.08 (0.65– 5.65) 1.85(0.65– 4.56)* No 79(25.1) 293(74.9%) 1 1 ∗p value < 0.05 Table 3 Bivariate and multivariable association of respondents on prevalence of preterm premature rupture of membrane and its associated factors among pregnant women attending health facilities in west Shoa Zone, Ethiopia, 2020. (n = 391). Page 11/17 Discussion Nutritional deficiencies, particularly micronutrient deficiencies such as vitamin C or ascorbic acid, affect collagen formation and can weaken the body’s ability to defend itself from degenerative processes caused by oxidative stress, which could lead to easy breakage of the membrane (30). In our study, having had anemia was also found to be a major risk factor. Pregnant women with anemia would be at risk of PROM nearly two times greater than non-anemic mothers. This is supported by the study done in Indonesia (34, 35). This may be justified by the fact that women with anemia are more likely to develop intra-amniotic and intrapartum infections, which may lead to PPROM. S li it ti f thi t d Th d t i ht b t i l d i bilit bi b Some limitations of this study were: The respondents might be prone to social desirability bias because some of the variables were based on self-reports. Finally, there might be a possibility of recall bias because women were asked about events that happened before the study. Consent for publication Not applicable. Conclusion And Recommendation Page 12/17 Page 12/17 The prevalence of preterm premature rupture of the membrane is high in the study area. To prevent PPROM, it is essential to check for modifiable factors during antenatal care and early screening, diagnosis, and treatments of preeclampsia and UTI are indicated to reduce PPROM. Since this study was carried out in hospital-admitted pregnant women, we recommend large low-risk population-based studies in the study area as well as in the country as a whole to interpret differences between countries. To prevent PPROM, it is essential to check for modifiable factors during antenatal care and early screening, diagnosis, and treatments of preeclampsia and UTI are indicated to reduce PPROM. Availability of data and materials The datasets produced and/or analyzed during this study are not publicly available due to some privacy reasons but are available from the corresponding author on time of request. Funding Not applicable Ethics approval and consent for participation Ethical review Committee (IRC) of Ambo University approved this study. Letter of permission was soughed from West Shoa Zonal health department and from each health institution. Verbal consent was taken from the participants after the data collectors clarified the objectives of the study, processes and their right to refuse not to participate at any time. Furthermore, confidentiality of the study participants was assured. Authors’ contributions RM; Formed and designed the study, control the data collection, analyze the data and develop the manuscript.GD, AB, and MM supervised the data collection, analyzed, edited the data and reviewed the manuscript. All authors read and agreeon the final manuscript. Competing interests The authors stated that they have no competing interests. Abbreviations PPROM Preterm Premature Rupture of Membrane PPROM Preterm Premature Rupture of Membrane UTI Urinary Tract Infection MUAC Mid Upper Arm Circumference STI Sexually transmitted Infection PROM Premature Rupture of Membrane IUFD Intra Uterine Fetal Death IUGR Intra Uterine Growth Retardation NRFHRP Non Reassuring Fetal Heart Rate Pattern Page 13/17 First and foremost, we would like to thank Ambo University for the approval of the ethical clearance for this study.We thank data collectors, supervisors and respondents for their valuable contributions and respond. Finally, our sincere thanks also go to Mr Daniel Bellema Assistant Professor) for his valuable comments and suggestion. First and foremost, we would like to thank Ambo University for the approval of the ethical clearance for this study.We thank data collectors, supervisors and respondents for their valuable contributions and respond. Finally, our sincere thanks also go to Mr Daniel Bellema Assistant Professor) for his valuable comments and suggestion. References The incidence and management outcome of preterm premature rupture of membranes (PPROM) in a tertiary hospital in Nigeria. American Journal of Clinical Medicine Research. 2014;2(1):14 – 7. 8. TC O, Enwereji J, Okoro O, Adiri C, Ezugwu E, Agu P. The incidence and management outcome of preterm premature rupture of membranes (PPROM) in a tertiary hospital in Nigeria. American Journal of Clinical Medicine Research. 2014;2(1):14 – 7. 8. TC O, Enwereji J, Okoro O, Adiri C, Ezugwu E, Agu P. The incidence and management outcome of preterm premature rupture of membranes (PPROM) in a tertiary hospital in Nigeria. American Journal of Clinical Medicine Research. 2014;2(1):14 – 7. 9. Byonanuwe S, Nzabandora E, Nyongozi B, Pius T, Ayebare DS, Atuheire C, et al. Predictors of premature rupture of membranes among pregnant women in rural Uganda: a cross-sectional study at a tertiary teaching hospital. International Journal of Reproductive Medicine. 2020;2020. 9. Byonanuwe S, Nzabandora E, Nyongozi B, Pius T, Ayebare DS, Atuheire C, et al. Predictors of premature rupture of membranes among pregnant women in rural Uganda: a cross-sectional study at a tertiary teaching hospital. International Journal of Reproductive Medicine. 2020;2020. 9. Byonanuwe S, Nzabandora E, Nyongozi B, Pius T, Ayebare DS, Atuheire C, et al. Predictors of premature rupture of membranes among pregnant women in rural Uganda: a cross-sectional study at a tertiary teaching hospital. International Journal of Reproductive Medicine. 2020;2020. 10. sewyew Addisu D, Melkie A. Prevalence of preterm premature rupture of membrane and its associated factors among pregnant women admitted in Debre tabor general hospital, North West Ethiopia: A facility based cross-sectional study. 2019. 10. sewyew Addisu D, Melkie A. Prevalence of preterm premature rupture of membrane and its associated factors among pregnant women admitted in Debre tabor general hospital, North West Ethiopia: A facility based cross-sectional study. 2019. 11. Addisu D, Melkie A, Biru S. Prevalence of preterm premature rupture of membrane and its associated factors among pregnant women admitted in Debre Tabor General Hospital, North West Ethiopia: institutional-based cross-sectional study. Obstetrics and Gynecology International. 2020;2020. 11. Addisu D, Melkie A, Biru S. Prevalence of preterm premature rupture of membrane and its associated factors among pregnant women admitted in Debre Tabor General Hospital, North West Ethiopia: institutional-based cross-sectional study. Obstetrics and Gynecology International. 2020;2020. 12. Sirak B, Mesfin E. References Maternal and perinatal outcome of pregnancies with preterm premature rupture of membranes (pprom) at tikur anbessa specialized teaching hospital, addis ababa, ethiopia. Ethiop Med J. 2014;52(4):165 – 72. 12. Sirak B, Mesfin E. Maternal and perinatal outcome of pregnancies with preterm premature rupture of membranes (pprom) at tikur anbessa specialized teaching hospital, addis ababa, ethiopia. Ethiop Med J. 2014;52(4):165 – 72. 13. Pisoh DW, Mbia CH, Takang WA, Djonsala OGB, Munje MC, Mforteh AA, et al. Prevalence, Risk Factors and Outcome of Preterm Premature Rupture of Membranes at the Bamenda Regional Hospital. Open Journal of Obstetrics and Gynecology. 2021;11(3):233 – 51. 13. Pisoh DW, Mbia CH, Takang WA, Djonsala OGB, Munje MC, Mforteh AA, et al. Prevalence, Risk Factors and Outcome of Preterm Premature Rupture of Membranes at the Bamenda Regional Hospital. Open Journal of Obstetrics and Gynecology. 2021;11(3):233 – 51. 14. Dars S, Malik S, Samreen I, Kazi RA. Maternal morbidity and perinatal outcome in preterm premature rupture of membranes before 37 weeks gestation. Pakistan journal of medical sciences. 2014;30(3):626. 14. Dars S, Malik S, Samreen I, Kazi RA. Maternal morbidity and perinatal outcome in preterm premature rupture of membranes before 37 weeks gestation. Pakistan journal of medical sciences. 2014;30(3):626. 15. Xie A, Zhang W, Chen M, Wang Y, Wang Y, Zhou Q, et al. Related factors and adverse neonatal outcomes in women with preterm premature rupture of membranes complicated by histologic chorioamnionitis. Medical science monitor: international medical journal of experimental and clinical research. 2015;21:390. 15. Xie A, Zhang W, Chen M, Wang Y, Wang Y, Zhou Q, et al. Related factors and adverse neonatal outcomes in women with preterm premature rupture of membranes complicated by histologic chorioamnionitis. Medical science monitor: international medical journal of experimental and clinical research. 2015;21:390. 16. Al-Riyami N, Al-Shezawi F, Al-Ruheili I, Al-Dughaishi T, Al-Khabori M. Perinatal outcome in pregnancies with extreme preterm premature rupture of membranes (Mid-Trimester PROM). Sultan Qaboos University medical journal. 2013;13(1):51. 16. Al-Riyami N, Al-Shezawi F, Al-Ruheili I, Al-Dughaishi T, Al-Khabori M. Perinatal outcome in pregnancies with extreme preterm premature rupture of membranes (Mid-Trimester PROM). Sultan Qaboos University medical journal. 2013;13(1):51. 17. Tchirikov M, Schlabritz-Loutsevitch N, Maher J, Buchmann J, Naberezhnev Y, Winarno AS, et al. Mid- trimester preterm premature rupture of membranes (PPROM): etiology, diagnosis, classification, international recommendations of treatment options and outcome. Journal of perinatal medicine. 2018;46(5):465 – 88. 18. References 1. C. B. Obstetrics and gynecology.Williams’s 24th. Chapter 22. 2010:213-6. 1. C. B. Obstetrics and gynecology.Williams’s 24th. Chapter 22. 2010:213-6. 2. Dagnew N, Tazebew A, Ayinalem A, Muche A. Measuring newborn foot length to estimate gestational age in a high risk Northwest Ethiopian population. PloS one. 2020;15(8):e0238169. 2. Dagnew N, Tazebew A, Ayinalem A, Muche A. Measuring newborn foot length to estimate gestational age in a high risk Northwest Ethiopian population. PloS one. 2020;15(8):e0238169. 3. Organization WH. Food, Medicine and Healthcare Administration and Control Authority of Ethiopia. Standard Treatment Guidelines for General Hospital. 2015. 3. Organization WH. Food, Medicine and Healthcare Administration and Control Authority of Ethiopia. Standard Treatment Guidelines for General Hospital. 2015. 4. Mohan SS, Thippeveeranna C, Singh NN, Singh LR. Analysis of risk factors, maternal and fetal outcome of spontaneous preterm premature rupture of membranes: a cross sectional study. Int J Reprod Contracept Obstet Gynecol. 2017;6(9):3781-7. 4. Mohan SS, Thippeveeranna C, Singh NN, Singh LR. Analysis of risk factors, maternal and fetal outcome of spontaneous preterm premature rupture of membranes: a cross sectional study. Int J Reprod Contracept Obstet Gynecol. 2017;6(9):3781-7. 5. Hackenhaar AA, Albernaz EP, Fonseca T. Preterm premature rupture of the fetal membranes: association with sociodemographic factors and maternal genitourinary infections. Jornal de pediatria. 2014;90:197–202. 5. Hackenhaar AA, Albernaz EP, Fonseca T. Preterm premature rupture of the fetal membranes: association with sociodemographic factors and maternal genitourinary infections. Jornal de pediatria. 2014;90:197–202. 6. Smith GN, Rafuse C, Anand N, Brennan B, Connors G, Crane J, et al. Prevalence, management, and outcomes of preterm prelabour rupture of the membranes of women in Canada. Journal of Obstetrics and Gynaecology Canada. 2005;27(6):547 – 53. 6. Smith GN, Rafuse C, Anand N, Brennan B, Connors G, Crane J, et al. Prevalence, management, and outcomes of preterm prelabour rupture of the membranes of women in Canada. Journal of Obstetrics and Gynaecology Canada. 2005;27(6):547 – 53. 7. Chandra I, Sun L. Third trimester preterm and term premature rupture of membranes: is there any difference in maternal characteristics and pregnancy outcomes? Journal of the Chinese Medical Association. 2017;80(10):657 – 61. 7. Chandra I, Sun L. Third trimester preterm and term premature rupture of membranes: is there any difference in maternal characteristics and pregnancy outcomes? Journal of the Chinese Medical Association. 2017;80(10):657 – 61. Page 14/17 Page 14/17 8. TC O, Enwereji J, Okoro O, Adiri C, Ezugwu E, Agu P. References Workineh Y, Birhanu S, Kerie S, Ayalew E, Yihune M. Determinants of premature rupture of membrane in Southern Ethiopia, 2017: case control study design. BMC Research Notes. 2018;11(1):1–7. 19. Anatolievna SV, Ruslanovna AS. Risk assessment for prolonged rupture to delivery interval in case of premature breaking of membranes at 34–36 weeks’ gestation. Медицинский вестник Северного Кавказа. 2015;10(2 (38)):151-5. Page 15/17 Page 15/17 20. Melamed N, Hadar E, Ben-Haroush A, Kaplan B, Yogev Y. Factors affecting the duration of the latency period in preterm premature rupture of membranes. The Journal of Maternal-Fetal & Neonatal Medicine. 2009;22(11):1051-6. 21. Morris JM, Roberts CL, Bowen JR, Patterson JA, Bond DM, Algert CS, et al. Immediate delivery compared with expectant management after preterm pre-labour rupture of the membranes close to term (PPROMT trial): a randomised controlled trial. The Lancet. 2016;387(10017):444 – 52. 21. Morris JM, Roberts CL, Bowen JR, Patterson JA, Bond DM, Algert CS, et al. Immediate delivery compared with expectant management after preterm pre-labour rupture of the membranes close to term (PPROMT trial): a randomised controlled trial. The Lancet. 2016;387(10017):444 – 52. 22. Spaeth JP, Lam JE. The extremely premature infant (micropremie) and common neonatal emergencies. A Practice of Anesthesia for Infants and Children: Elsevier; 2019. p. 841 – 67. e7. 23. Lissauer T, Fanaroff AA, Miall L, Fanaroff J. Neonatology at a Glance: John Wiley & Sons; 2020. 23. Lissauer T, Fanaroff AA, Miall L, Fanaroff J. Neonatology at a Glance: Jo 24. Gezer A, Parafit-Yalciner E, Guralp O, Yedigoz V, Altinok T, Madazli R. Neonatal morbidity mortality outcomes in pre-term premature rupture of membranes. Journal of Obstetrics and Gynaecology. 2013;33(1):38–42. 25. Abouseif HA, Mansour AF, Hassan S, Sabbour S. Prevalence and outcome of preterm premature rupture of membranes (PPROM) among pregnant women attending Ain Shams maternity hospital. Egyptian Journal of Community Medicine. 2018;36(2):99–107. 26. Liu L, Oza S, Hogan D, Chu Y, Perin J, Zhu J, et al. Global, regional, and national causes of under-5 mortality in 2000–15: an updated systematic analysis with implications for the Sustainable Development Goals. The Lancet. 2016;388(10063):3027-35. 26. Liu L, Oza S, Hogan D, Chu Y, Perin J, Zhu J, et al. Global, regional, and national causes of under-5 mortality in 2000–15: an updated systematic analysis with implications for the Sustainable Development Goals. The Lancet. 2016;388(10063):3027-35. 27. Lawn JE, Gravett MG, Nunes TM, Rubens CE, Stanton C. References Global report on preterm birth and stillbirth (1 of 7): definitions, description of the burden and opportunities to improve data. BMC pregnancy and childbirth. 2010;10(1):1–22. 28. Rasmussen S, Ebbing C, Irgens LM. Predicting preeclampsia from a history of preterm birth. PloS one. 2017;12(7):e0181016. 29. Liu L, Wang L, Yang W, Ni W, Jin L, Liu J, et al. Gestational hypertension and pre-eclampsia and risk of spontaneous premature rupture of membranes: A population‐based cohort study. International Journal of Gynecology & Obstetrics. 2019;147(2):195–201. 30. Aouache R, Biquard L, Vaiman D, Miralles F. Oxidative stress in preeclampsia and placental diseases. International journal of molecular sciences. 2018;19(5):1496. 30. Aouache R, Biquard L, Vaiman D, Miralles F. Oxidative stress in preeclampsia and placental diseases. International journal of molecular sciences. 2018;19(5):1496. 31. Boskabadi H, Zakerihamidi M. Evaluation of maternal risk factors, delivery, and neonatal outcomes of premature rupture of membrane: A systematic review study. Journal of Pediatrics Review. 2019;7(2):77–88. 31. Boskabadi H, Zakerihamidi M. Evaluation of maternal risk factors, delivery, and neonatal outcomes of premature rupture of membrane: A systematic review study. Journal of Pediatrics Review. 2019;7(2):77–88. 32. Tskitishvili E, Komoto Y, Temma-Asano K, Hayashi S, Kinugasa Y, Tsubouchi H, et al. S100B protein expression in the amnion and amniotic fluid in pregnancies complicated by pre-eclampsia. MHR: Basic science of reproductive medicine. 2006;12(12):755 – 61. 32. Tskitishvili E, Komoto Y, Temma-Asano K, Hayashi S, Kinugasa Y, Tsubouchi H, et al. S100B protein expression in the amnion and amniotic fluid in pregnancies complicated by pre-eclampsia. MHR: Basic science of reproductive medicine. 2006;12(12):755 – 61. 33. Zuo G, Dong J-X, Zhao F-F, Chen Y. Expression of matrix metalloproteinase-9 and its substrate level in patients with premature rupture of membranes. Journal of Obstetrics and Gynaecology. 2017;37(4):441-5. Page 16/17 Page 16/17 34. Arzda MI. Profil Ketuban Pecah Dini pada Ibu Bersalin di RSUD Dr. Zainoel Abidin Banda Aceh. Jurna Kedokteran Syiah Kuala. 2021;21(3). 35. Singh D, Usham R, Kamei H. Preterm prelabour rupture of membrane: 1 year study. Journal of Evolution of Medical and Dental Sciences. 2015;4(49):8495-9. 35. Singh D, Usham R, Kamei H. Preterm prelabour rupture of membrane: 1 year study. Journal of Evolution of Medical and Dental Sciences. 2015;4(49):8495-9. 34. Arzda MI. Profil Ketuban Pecah Dini pada Ibu Bersalin di RSUD Dr. Zainoel Abidin Banda Aceh. Jurnal Kedokteran Syiah Kuala. 2021;21(3). Figures Figures Figures Figure 1 Reasons (indications of admission for pregnant woman in Ambo Hospitals, 2022. (n=391) Other: (IUFD, severe anemia, severe IUGR, NRFHRP) Figure 1 Figure 1 Figure 1 Reasons (indications of admission for pregnant woman in Ambo Hospitals, 2022. (n=391) Other: (IUFD, severe anemia, severe IUGR, NRFHRP) Reasons (indications of admission for pregnant woman in Ambo Hospitals, 2022. (n=391) *Other: (IUFD, severe anemia, severe IUGR, NRFHRP) Page 17/17 Page 17/17 Page 17/17
https://openalex.org/W4317904061	https://hal.univ-lille.fr/hal-03959239/file/ChemistrySelect2023.pdf	English	null	Evidence of Surface Properties by Isopropanol Decomposition Reaction and NH<sub>3</sub>‐TPD over Ni−Fe Spinel Nanoparticles Prepared <i>via</i> Hydrothermal Route	ChemistrySelect	2,023	cc-by	6,664	Evidence of Surface Properties by Isopropanol Decomposition Reaction and NH3-TPD over Ni–Fe Spinel Nanoparticles Prepared via Hydrothermal Route Rafik Benrabaa, Annick Rubbens, Axel Löfberg, Rose-Noelle Vannier Evidence of Surface Properties by Isopropanol Decomposition Reaction and NH3-TPD over Ni–Fe Spinel Nanoparticles Prepared via Hydrothermal Route Rafik Benrabaa, Annick Rubbens, Axel Löfberg, Rose-Noelle Vannier To cite this version: Rafik Benrabaa, Annick Rubbens, Axel Löfberg, Rose-Noelle Vannier. Evidence of Surface Proper- ties by Isopropanol Decomposition Reaction and NH3-TPD over Ni–Fe Spinel Nanoparticles Pre- pared via Hydrothermal Route. ChemistrySelect, 2023, ChemistrySelect, 8 (4), pp.e202204361. 10.1002/slct.202204361. hal-03959239 Distributed under a Creative Commons Attribution 4.0 International License HAL Id: hal-03959239 https://hal.univ-lille.fr/hal-03959239v1 Submitted on 27 Jan 2023 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License 23656549, Research Article doi.org/10.1002/slct.202204361 ChemistrySelect www.chemistryselect.org Introduction CH3-CO-CH3 þ H2 : deshydrogenation, acetone (3) (1) deshydratation, diisopropylether [b] Prof. R. Benrabaa Laboratoire de Matériaux Catalytiques et Catalyse en Chimie Organique, Faculté de Chimie, USTHB, BP32, El-Alia, 16111 Bab Ezzouar, Alger, Algérie (2) deshydratation, propylene ðCH3Þ2-CH-OH ! CH3-CO-CH3 þ H2 : (3) [c] Prof. A. Rubbens, Dr. A. Löfberg, Prof. R.-N. Vannier Univ. Lille, CNRS, Centrale Lille, Univ. Artois, UMR 8181 – UCCS – Unité de Catalyse et Chimie du Solide, F-59000 Lille, France E-mail: axel.lofberg@univ-lille.fr deshydrogenation, acetone deshydrogenation, acetone Rafik Benrabaa,[a, b] Annick Rubbens,[c] Axel Löfberg,[c] and Rose-Noëlle Vannier[c] and Thermal Gravimetric Analysis- Differential Scanning Calo- rimetry (TGA-DSC). The total acidity and acid strength distribu- tion was determined by NH3-TPD. The synthesized oxide showed the presence of mixed phases containing the inverse spinel structure NiFe2O4 as the major phase and iron oxide α- Fe2O3 as an additional phase. It was shown that the surface of the powder was richer in Ni2+ species. The surface acidity (Brönsted and Lewis) determined by NH3-TPD showed that the number of acidic sites increased dramatically with the temper- ature. The catalytic conversion of isopropanol yields both acidic and redox/basic sites were available. A good correlation between textural, structural, redox and acid-base properties of catalyst is established. and Thermal Gravimetric Analysis- Differential Scanning Calo- rimetry (TGA-DSC). The total acidity and acid strength distribu- tion was determined by NH3-TPD. The synthesized oxide showed the presence of mixed phases containing the inverse spinel structure NiFe2O4 as the major phase and iron oxide α- Fe2O3 as an additional phase. It was shown that the surface of the powder was richer in Ni2+ species. The surface acidity (Brönsted and Lewis) determined by NH3-TPD showed that the number of acidic sites increased dramatically with the temper- ature. The catalytic conversion of isopropanol yields both acidic and redox/basic sites were available. A good correlation between textural, structural, redox and acid-base properties of catalyst is established. In this work, surface properties evaluated by catalytic con- version of isopropanol and by ammonia Temperature Pro- grammed Reduction (NH3-TPD) over ferrite spinel nanoparticles was investigated. Bulk NiFe2O4 with specific surface area 114 m2/g and crystallites size 6 nm was prepared by hydro- thermal synthesis using nitrates as precursors. X-ray diffraction (XRD), Raman spectroscopy, specific surface area by B.E.T, Scanning Electron Microscopy (SEM) and X-ray energy disper- sive microanalysis (EDS), Transmission Electron Microscopy (TEM) and X-ray Photoelectron Spectroscopy (XPS) techniques were used for their structural and textural characterizations. The reducibility by hydrogen at variable temperatures was investigated by Temperature Programmed reduction (H2-TPR) infrared (IR), nuclear magnetic resonance (NMR) or X-ray photoelectron spectroscopy (XPS) have also been extensively used to study the nature of acid-base properties.[5,6] Introduction The study of the acidity of bulk oxides is of great importance for applications of these materials, particularly in the petro- chemical industry and environmental sciences. In order to design materials with high activity and selectivity, a detailed characterization of the active sites is required. A question of basic interest is to determine the correlation between these properties and the promotion of catalytic activity of oxides. The adsorption of alcohols over oxide catalysts has been used as a chemical probe reaction for many years. Isopropanol decomposition reaction is one of the main methods for determining surface acid-base properties.[7,8] The decomposi- tion leads to the formation of olefins, ether, ketones, carbon monoxide or dioxide. If the solid exhibits acidic properties, only the dehydration products (ether+olefins) are obtained, where- as if the catalyst displays basic or redox properties, dehydro- genation of the alcohol is observed (ketone formation).[9,10] Depending on the nature of the catalyst and the operating conditions chosen, one or the other of the reaction paths can be favoured. Several methods have been successfully used to study the active sites. These methods are based on the adsorption/ desorption of probe molecules. Among the methods com- monly applied, temperature-programmed desorption (TPD)[1] or adsorption microcalorimetry[2–4] of probe molecules give in- formation regarding the distribution and strength of active sites located on the surface. Spectroscopic methods, such as The various reactions usually accepted to describe the decomposition of isopropanol can be summarized as follows: Without oxygen: [a] Prof. R. Benrabaa Laboratoire de Physico-Chimie des Matériaux, Faculté des Sciences et de la Technologie, Université Chadli BENDJEDID-El Tarf, B.P 73 El Tarf 36000, Algérie E-mail: r.benrabaa@univ-eltarf.dz [b] Prof. R. Benrabaa Laboratoire de Matériaux Catalytiques et Catalyse en Chimie Organique, Faculté de Chimie, USTHB, BP32, El-Alia, 16111 Bab Ezzouar, Alger, Algérie [c] Prof. A. Rubbens, Dr. A. Löfberg, Prof. R.-N. Vannier Univ. Lille, CNRS, Centrale Lille, Univ. Artois, UMR 8181 – UCCS – Unité de Catalyse et Chimie du Solide, F-59000 Lille, France E-mail: axel.lofberg@univ-lille.fr © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH. This is an open access article under the terms of the Creative Commons At- tribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. 2ðCH3Þ2-CH-OH ! ðCH3Þ2CHO-CHðCH3Þ2 þ H2O : deshydratation, diisopropylether (1) ðCH3Þ2-CH-OH ! CH3CH¼CH2 þ H2O : deshydratation, propylene (2) ðCH3Þ2-CH-OH ! [a] Prof. R. Benrabaa Laboratoire de Physico-Chimie des Matériaux, Faculté des Sciences et de la Technologie, Université Chadli BENDJEDID-El Tarf, B.P 73 El Tarf 36000, Algérie E-mail: r.benrabaa@univ-eltarf.dz ChemistrySelect 2023, 8, e202204361 (1 of 9) Evidence of Surface Properties by Isopropanol Decomposition Reaction and NH3-TPD over NiFe Spinel Nanoparticles Prepared via Hydrothermal Route k Benrabaa,[a, b] Annick Rubbens,[c] Axel Löfberg,[c] and Rose-Noëlle Vannier[c] [b] Prof. R. Benrabaa Laboratoire de Matériaux Catalytiques et Catalyse en Chimie Organique, Faculté de Chimie, USTHB, BP32, El-Alia, 16111 Bab Ezzouar, Alger, Algérie Structural (XRD and Raman), surface (BET, XPS, SEM and TEM) and bulk composition (EDS) It appears that the activities vary considerably from one solid to another. Selectivity is also very variable but in general, a solid having a basic character like MgO and CaO, exhibits a very low activity but a selectivity almost of 100% in acetone product. In contrast, an acidic solid such as V2O5 has high activity and selectivity for propylene of up to 94%.[11] The isopropanol decomposition is a method that does not allow to distinguish between the Brønsted and Lewis acidity. However, Aramendia et al.[12] have reported that the dehydration activity during isopropanol decomposition could be correlated with Brønsted acidity. The phase composition, purity and structure of the products were ascertained by XRD and Raman analysis at room temper- ature. Figure 1 displays the XRD patterns of NiFe2O4 spinel oxide. The registered diffractogram shows only the lines characteristic of pure NiFe2O4 (PDF 00-054-0954) structure with the main peaks at 2θ°18.4 w (111), 30.4 m (220), 35.7 vs (311), 37.3 w (222), 43.4 m (400), 53.9 w (422), 57.3 m (511) and 62.9 m (440). (intensity of lines, w: weak, m: medium, s: strong; vs: very strong; hkl in brackets) (Figure 1, left). No lines corresponding to nickel or iron free oxides, could be observed. Spinels family are excellent materials in electronics for their ferromagnetic properties, in catalysis such as the reduction of NOx, the decomposition of alcohols and the reforming of hydrocarbons.[8,12–15] The catalytic activity of spinels depends essentially on (i) the degree of substitution of the cations and the degree of inversion of the spinel structure, (ii) the preparation method and departure precursors, (iii) the presence of a support and (iv) the distribution of species on the catalytic surface. These parameters considerably influence the acid-base behavior of the catalysts and consequently its catalytic activity. The XRD pattern of NiFe2O4 spinel oxide was refined using Rietveld method. The structural model corresponding to NiFe2O4 spinel (Fd-3 m)[17] was introduced in the refinement. The refinement led to a=8.377(5) Å and crystallite size Cs= 6 nm. The bulk ferrite powder was further investigated using laser Raman spectroscopy in order to ascertain the structure of the catalyst. The Raman spectrum is given in Figure 1 (right). The Raman spectra of NiFe2O4 spinel oxide show the characteristic vibrational modes of the NiFe2O4 inverse spinel[18–20] There are located at wavenumbers: 702 cm1 (s), 665 cm1 (s), 579 cm1 Figure 1. With oxygen: © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH. This is an open access article under the terms of the Creative Commons At- tribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH ChemistrySelect 2023, 8, e202204361 (1 of 9) Research Article doi.org/10.1002/slct.202204361 ChemistrySelect ðCH3Þ2-CH-OH þ 1=2O2 ! CH3-CO-CH3 þ H2O : Oxy-deshydrogenation (4) ðCH3Þ2-CH-OH þ 9=2O2 ! 3CO2 þ 4H2O : total Combustion (5) ðCH3Þ2-CH-OH þ 3O2 ! 3CO þ 4H2O : incomplete combustion (6) Indeed, the study of the acidity of the catalysts is a key point for understanding the mechanism of the catalytic act. A previous work in dry reforming of methane,[16] we showed that the activity of nickel ferrite could be correlated to the surface acidity that would limit the relative contribution of Reverse Water Gas Shift reaction. For a better insight of the acidity of catalyst prepared via hydrothermal synthesis, it was reinvestigated. In this work, NiFe2O4 spinel oxide issued from hydrothermal method was prepared under other conditions compared to our previous. The structural, textural and surface composition as well as redox properties was studied in details and then correlated to the catalytic conversion of isopropanol. (6) incomplete combustion The decomposition reaction of isopropanol can be carried out in the absence (reactions 1, 2 and 3) or in the presence of gaseous oxygen (reactions 4, 5 and 6). In the presence of oxygen, isopropanol can undergo an oxidation-hydrogenation reaction to form acetone (reaction 4) or a combustion reaction which leads to the formation of carbon dioxide or carbon monoxide (reactions 5 and 6) depending on whether the reaction mixture is poor or rich in oxygen. Results and discussion Structural (XRD and Raman), surface (BET, XPS, SEM and TEM) and bulk composition (EDS) © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH Structural (XRD and Raman), surface (BET, XPS, SEM and TEM) and bulk composition (EDS) Structural properties by XRD (left) and Raman (right) of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. Figure 1. Structural properties by XRD (left) and Raman (right) of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH ChemistrySelect 2023, 8, e202204361 (2 of 9) Research Article doi.org/10.1002/slct.202204361 ChemistrySelect (m, broad), 487 cm1 (m), 456 cm1 (s), 336 cm1 (w). In contrast to XRD results which showed no peak corresponding to free oxides, two bands with low intensity at ∼1140 cm1 and 1340 cm1 were observed and attributed to α-Fe2O3 hematite.[21,22] the pores are cylindrical and open at both ends (Figure 2-b). the obtained results confirmed that NiFe2O4 prepared by HT technique at 180°C/1 h is in mesoporous category (25–500 Å). SEM micrograph (Figure 3-left) confirmed monodisperse and highly nanoparticles distribution with small grain size. However, TEM image (Figure 3-right) showed ununiform nanosized particles with diffraction fringes were observed confirming that the high crystallinity of the sample as evidenced by the XRD pattern. The heterogeneity of NiFe2O4 powder suggested that the presence of a mixed phase in accordance with Raman spectroscopy showing the presence of NiFe2O4 and α-Fe2O3 hematite as supplementary phase. The crystal size measured in TEM image of the sample is smaller-around 10 nm. The elemental composition of NiFeO4 sample was performed by EDS analysis. The measured Fe:Ni atomic ratio (∼1.9) of analyzed sample was very similar to the nominal atomic ratio (Fe:Ni=2) used in the hydrothermal preparation. The surface composition of catalyst was determined by XPS at room temperature. The Figure 4 represent the photoemission spectra of 2p levels of nickel (Ni2p3/2 line) and iron (Fe2p3/2 line). The presence of Ni2+ and Fe3+ was detected by Ni2p3/2 (854.9 eV) and Fe2p3/2 (710.7 eV) photopeaks, respectively.[23,24] The bind- ing energies values obtained and the recorded shape (Figure 4) The specific surface area of the catalyst was determined from the nitrogen adsorption-desorption isotherms at 77 K. NiFe2O4 sample treated at 180°C for 1 hours under hydro- thermal conditions showed a value of 114 m2/g (Table 1). The N2 adsorption-desorption isotherm of the catalyst (Figure 2-a), is type IV according to IUPAC classification with H3-type hysteresis loop, indicating mesoporous materials. The porosity of the catalyst was studied by the B.J.H method (Barrer, Joyner and Halenda) from the N2 desorption isotherm assuming that Table 1. Structural (XRD and Raman), surface (BET, XPS, SEM and TEM) and bulk composition (EDS) Crystallographic parameters, specific surface area and Ni/Fe ratio of NiFe2O4 catalyst. Cs[a] (nm) a[b] (Å) SB.E.T [c] (m2/g) Ni/Fe[d] Ni/Fe[e] NiFe2O4 6 8.377(5) 114 1.9 1.9 [a] Crystallite size, [b] Lattice parameter, [c] Specific Surface Area, [d] Obtained by EDX, [e] Obtained by XPS Figure 2. N2 adsorption-desorption isotherm (a) and porous distribution of NiFe2O4 solid prepared via hydrothermal synthesis at 180°C/1 h. Figure 2. N2 adsorption-desorption isotherm (a) and porous distribution of NiFe2O4 solid prepared via hydrothermal synthesis at 180°C/1 h. Figure 3. SEM (left) and TEM (right) micrographs of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. Figure 3. SEM (left) and TEM (right) micrographs of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. ChemistrySelect 2023, 8, e202204361 (3 of 9) © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH Figure 4. XPS spectra of Ni 2p3/2 (a) and Fe 2p3/2 (b) for NiFe2O4 solid prepared via hydrothermal synthesis at 180°C/1 h. Figure 5. H2-TPR profiles of bulk NiFe2O4 spinel oxide prepared by hydrothermal synthesis at 180°C/1 h. ChemistrySelect Research Article doi.org/10.1002/slct.202204361 23656549, 2023, 4, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/slct.202204361 by Cochrane France, Wiley Online Library on [24/01/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.co Research Article doi.org/10.1002/slct.202204361 ChemistrySelect Figure 4. XPS spectra of Ni 2p3/2 (a) and Fe 2p3/2 (b) for NiFe2O4 solid prepared via hydrothermal synthesis at 180°C/1 h. Figure 4. XPS spectra of Ni 2p3/2 (a) and Fe 2p3/2 (b) for NiFe2O4 solid prepared via hydrothermal synthesis at 180°C/1 h. Figure 5. H2-TPR profiles of bulk NiFe2O4 spinel oxide prepared by hydrothermal synthesis at 180°C/1 h. Figure 5. H2-TPR profiles of bulk NiFe2O4 spinel oxide prepared by hydrothermal synthesis at 180°C/1 h. variable temperature under H2 atmosphere. The amount of H2- consumption is ∼16 mmol/g. As shown in Figure 5, the H2-TPR profile is composed of a (i) shoulder at lower temperature (ca. 200– 260°C) that could correspond to the reduction of some loosely bound Ni2+ based particles not engaged in the spinel in accordance with XPS experiments showing a slight excess of Ni (Fe:Ni=1.9) and (ii) two neat reduction peaks with a maximum H2- consumption at 341 and 543°C which likely correspond to the reduction of most Ni2+ species to Ni° and Fe3+ to Fe°, respectively. Structural (XRD and Raman), surface (BET, XPS, SEM and TEM) and bulk composition (EDS) These three steps were confirmed by TGA-DSC under H2 which showed two main mass losses between 280°C and 400°C and between 400°C and 600°C, with a much larger loss for the second in good agreement with the reduction of nickel first followed by the reduction of iron at higher temperature (Figure 6). These two mass losses are preceded by two small mass losses, one before 100°C likely due to traces of physisorbed water and the second between 200 and 300°C, likely related to the reduction of nickel species at the catalyst surface. Assuming that all species of NiFe2O4 spinel structure are reduced, the theoretical mass loss of oxygen of the spectra for Ni and Fe cations confirm that both species are in NiFe2O4 spinel structure and in (II+) et (III+) oxidation state, respectively. The Fe:Ni surface ratio (∼1.9) was very close to the expected value (Fe:Ni=2) and in good agreement with Ni:Fe obtained by EDX experiments (Table 1). EDS and XPS results are not in the same trend of Raman spectroscopy which confirming a mixed phase in the NiFe2O4 structure. We were waiting an atomic ratio Fe:Ni>2 (bulk and surface richer in Fe species) for justify the presence of α-Fe2O3 observed by Raman analysis. These results suggest that, in our hydrothermal synthesis conditions, a very low amount of α-Fe2O3 hematite was formed beside NiFe2O4 structure. © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH Reducibility study (H2-TPR and H2-TGA-DSC) The reduction behavior of bulk NiFe2O4 spinel oxide prepared by hydrothermal method was followed by H2-TPR and TGA-DSC at ChemistrySelect 2023, 8, e202204361 (4 of 9) Research Article doi.org/10.1002/slct.202204361 ChemistrySelect related to NiFe2O4 is 27.3%. This value is close to value recorded from TGA curve under H2 atmosphere (∼30.6%) with a difference of ∼3.3% which confirms a total reduction of Ni2+ and Fe3+ to metallic species in NiFe2O4 inverse spinel structure 350 C 600 C : Fe2O3 þ 3H2 ! 2Fe þ 3H2O (2) during reduction by hydrogen (5% H2 in N2) up to 1000°C Figure 6. TGA-DSC thermograms of NiFe2O4 prepared via hydrothermal route at 180°C/1 h Figure 7. Ammonia TPD profile of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. ChemistrySelect Research Article doi.org/10.1002/slct.202204361 56549, 2023, 4, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/slct.202204361 by Cochrane France, Wiley Online Library on [24/01/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules o Figure 6. TGA-DSC thermograms of NiFe2O4 prepared via hydrothermal route at 180°C/1 h 3, 4, Downloaded from https://chemistry europe.onlinelibrary.wiley.com/doi/10.1002/slct.202204361 by Cochrane France, Wiley On Figure 6. TGA-DSC thermograms of NiFe2O4 prepared via hydrothermal route at 180°C/1 h , Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/slct.202204361 by Cochrane France, Wiley Figure 6. TGA-DSC thermograms of NiFe2O4 prepared via hydrothermal route at 180°C/1 h Figure 7. Ammonia TPD profile of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. Figure 7. Ammonia TPD profile of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. 350 C 600 C : Fe2O3 þ 3H2 ! 2Fe þ 3H2O (2) 350 C 600 C : Fe2O3 þ 3H2 ! 2Fe þ 3H2O (2) related to NiFe2O4 is 27.3%. This value is close to value recorded from TGA curve under H2 atmosphere (∼30.6%) with a difference of ∼3.3% which confirms a total reduction of Ni2+ and Fe3+ to metallic species in NiFe2O4 inverse spinel structure. during reduction by hydrogen (5% H2 in N2) up to 1000°C. The reduction process of NiFe2O4 sample can be summar- ized as follows: Acid properties (NH3-TPD) Figure 7 shows NH3-TPD profile of NiFe2O4 sample. The desorption temperature indicates the acid strength of the sites, with weaker sites desorbing at lower temperature. After deconvolution of the NH3-TPD curve, three regions could be distinguished for the Amb- 350 C : NiFe2O4 þ H2 ! Ni þ Fe2O3 þ H2O (1) ChemistrySelect 2023, 8, e202204361 (5 of 9) ChemistrySelect 2023, 8, e202204361 (5 of 9) © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH Research Article doi.org/10.1002/slct.202204361 ChemistrySelect temperatures higher than 400°C,[28,29] which is not the case in our sample. analyzed sample thus allowing to distinguish three types of acid sites: weak acid sites (∼261°C), moderate acid sites (316°C) and strong acid sites (395°C). The number of acid sites for each stage was calculated from the ammonia thermodesorption curves and the values are reported on Figure 7. The total number of acid sites corresponds to 873 μmol(NH3)/g. It can be seen that the number of acid sites increased with the increase of temperature and the highest value (720 μmol(NH3)/g) was observed at 395°C. The low acid character observed at 261°C can be linked to the presence of Ni2+ species on the surface which may act as basic character in accordance with XPS data at room temperature. The increase in the temperature induces certainly a change in surface composition and it seems that the acid character of the sample is favored at high temperature by the migration of Fe3+ species of the bulk, in particular those at octahedral sites, on the surface and increasing the number of acid sites because Fe3+ has an optical basicity greater than Ni2+ species (ΛFe3+ (Λtet=0.659)<Fe3+ (Λoct= 0.756)<Ni2+ (Λoct=0.953).[25–27] Moreover, from these results, we can speculate about the absence of Brönsted sites, since it has been reported that Brönsted sites correspond to desorption analyzed sample thus allowing to distinguish three types of acid sites: weak acid sites (∼261°C), moderate acid sites (316°C) and strong acid sites (395°C). The number of acid sites for each stage was calculated from the ammonia thermodesorption curves and the values are reported on Figure 7. The total number of acid sites corresponds to 873 μmol(NH3)/g. It can be seen that the number of acid sites increased with the increase of temperature and the highest value (720 μmol(NH3)/g) was observed at 395°C. Acid properties (NH3-TPD) The low acid character observed at 261°C can be linked to the presence of Ni2+ species on the surface which may act as basic character in accordance with XPS data at room temperature. The increase in the temperature induces certainly a change in surface composition and it seems that the acid character of the sample is favored at high temperature by the migration of Fe3+ species of the bulk, in particular those at octahedral sites, on the surface and increasing the number of acid sites because Fe3+ has an optical basicity greater than Ni2+ species (ΛFe3+ (Λtet=0.659)<Fe3+ (Λoct= 0.756)<Ni2+ (Λoct=0.953).[25–27] Moreover, from these results, we can speculate about the absence of Brönsted sites, since it has been reported that Brönsted sites correspond to desorption Catalytic data in isopropanol decomposition As already mentioned in the introduction, the decomposition occurs via two parallel elimination path- ways: decomposition either leads to the formation of propylene and di-isopropyl either through dehydration on strong acidic sites (Brönsted and Lewis acid sites) or the formation of dehydrogenated product, acetone through the involvement of acid-base or redox sites. The catalytic activity in isopropanol decomposition reaction of NiFe2O4 catalyst show a dual character in the temperature range 200–400°C suggesting that the role of the surface composition is responsible for the catalytic activity. The acidity of oxides can be compared using the optical basicity Λ, which allows comparison of the acidity of cations by taking into account their coordination and symmetry in the oxide.[25–27] The optical basicity Λ of inverse spinel NiFe2O4 is Λ=0.771. For cations, it increases along the Fe3+ (Λtet=0.659)<Fe3+ (Λoct=0.756)<Ni2+ (Λoct=0.953) series (tet and oct for tetrahedral and octahedral coordination, respectively). Both acidic and basic/redox sites were available during the temper- ature range studied was recorded in our spinel synthetized. In the temperature range 200–350°C, a dehydrogenation pathway indicating the acetone formation was recorded and this behavior cover up when the reaction temperature increased up to 400°C by propylene product formation. The binary character of NiFe2O4 sample is well correlated to their structural, textural and their redox properties. The formation of acetone in the temperature range 200–350°C can be attributed to the local- ization of Ni2+ species on the surface of NiFe2O4 sample as showed by XPS and EDX experiments (Fe:Ni=1.9). As mentioned previously the optical basicity Λ of Ni2+ cations are higher than those of Fe3+ species at octahedral or tetrahedral sites. In addition, the basic character of NiFe2O4 sample in the temperature range 200–350°C can be well linked to the weak and moderate acid sites observed on this sample by NH3-TPD experiments (Figure 7) at 261°C (36 μmol(NH3)/g) and 316°C (117 μmol(NH3)/g), respectively. The temperature range 350– 400°C is a turning point for NiFe2O4 spinel oxide prepared by hydrothermal synthesis, when the acid character appeared; this change is correlated certainly to the change in the surface composition and especially to the change in redox properties as observed by H2-TPR and TGA-DSC analyses as well as to NH3- TPD measurements. Catalytic data in isopropanol decomposition The activity of NiFe2O4 catalyst was tested for isopropanol decomposition in the temperature range 200–400°C without oxygen at atmospheric pressure. The catalyst shows no catalytic activity in isopropanol decomposition below 200°C. The conversion of isopropanol and the selectivity to propylene and acetone, obtained on nickel ferrite oxide, are presented in Figures 8 and 9, respectively. Propylene and acetone are the main products; propylene is due to the presence of acid sites, and acetone to that of basic/ redox sites. The formation of isopropyl ether (condensation product) was not observed over our catalyst. The isopropanol conversion (Figure 8) increased with the reaction temperature. The evolution of acetone and propylene selectivity of the Figure 8. Isopropanol conversions of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. Figure 8. Isopropanol conversions of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. Figure 9. Acetone (a) and propylene (b) selectivity of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. ChemistrySelect 2023, 8, e202204361 (6 of 9) © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH Figure 8. Isopropanol conversions of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. Figure 9. Acetone (a) and propylene (b) selectivity of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. Figure 9. Acetone (a) and propylene (b) selectivity of NiFe2O4 prepared via hydrothermal route at 180°C/1 h. © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH ChemistrySelect 2023, 8, e202204361 (6 of 9) Research Article doi.org/10.1002/slct.202204361 ChemistrySelect catalyst are depicted in Figure 9. The behavior of NiFe2O4 sample varies depending on the reaction temperature; acetone and propylene were observed throughout the temperatures range 200–400°C, which means that both basic and acidic sites were available. At lower temperatures (200–350°C), the for- mation of acetone is faster compared to that of propylene. However, when the reaction temperature increases from 350 to 400°C, an opposite character was observed, the surface of NiFe2O4 catalyst becomes more acidic, indeed, the propylene selectivity increases at the expense of acetone by 27 to 74% between 350 and 400°C (Figure 9). The proportion of acetone, with a maximum of 73% observed around 350°C, undergoes a rapid drop to 400°C (up to 26%). heterogeneous catalysts. In isopropanol decomposition reac- tion, little work can be found in the literature on the study of this model reaction. Catalytic data in isopropanol decomposition Indeed, the temperature range of character change to acidity through the formation of propylene of NiFe2O4 sample well coincides with the reduction temperature of Ni2+ and Fe3+ as affirmed by H2-TPR and TGA-DSC analyses and also corroborate to the temperature of strong acid sites observed by NH3-TDP- (Figure 7). The reducibility of a catalyst is a factor which greatly influences its catalytic activity towards redox reactions.[30,31] As affirmed previously by XRD analysis at higher temperatures under H2 atmosphere,[32,33] NiFe2O4 is stable up to 300°C and above this temperature Ni and Fe metallic were observed suggesting a change in redox proper- ties from 300°C leading to a change in acid-base behavior. The formation of propylene at 350–400°C can be well explained to the high number of strong acid sites (720 μmol(NH3)/g) at The reaction rate R (μmol/h) calculated in the temperature range 200–350°C is plotted in Figure 10. It is observed that the rate increases with increasing the reaction temperature up to 300°C. Above this temperature, the rate mitigates dramatically and dropped up to 350°C in accordance with acetone and propylene products distribution as observed above. © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH Correlation of the structural, textural and redox properties with catalytic decomposition The observed behavior in catalytic activity is linked in our opinion to the textural properties and structural properties of the solid. The surface composition observed by XPS (Fe/Ni) is one of the main parameters responsible for surface properties such as acidity and basicity. Isopropanol dehydrogenation and dehydration reactions to acetone and propylene, respectively, related to the contribu- tion of both acidic and basic sites, is favored on our catalyst. In the light of properties determined by physicochemical charac- terizations, some considerations follow to tentatively account for these results. It is worth noting that the structural and textural as well as redox properties affect significantly the catalytic activity of Figure 10. Variation of the reaction rate R (μmol/h) as function of the reaction temperature of NiFe2O4 prepared via hydrothermal route at 180°C/ 1 h. Figure 10. Variation of the reaction rate R (μmol/h) as function of the reaction temperature of NiFe2O4 prepared via hydrothermal route at 180°C/ 1 h. ChemistrySelect 2023, 8, e202204361 (7 of 9) Research Article doi.org/10.1002/slct.202204361 ChemistrySelect power (3 mW on the sample). Four accumulations were used in each spectral range. No damage of the material by the laser was observed. The scattered light was analysed with an XY Raman Dilor spectrometer equipped with an optical multichannel charge coupled device liquid nitrogen-cooled detector. The spectral resolution was 0.5 cm1 in the 130–1500 cm1 range. Acquisition and data processing were performed with the LABSPEC software. The specific surface area (SBET) of the catalysts was determined by nitrogen adsorption at 196°C with Micromeritics ASAP2010 apparatus. Scanning electron microscopy (SEM) and X-ray energy dispersive microanalysis (EDS) were carried out on HITACHI 4100S apparatus at 15 kV. Transmission electron microscopy (TEM) study was performed with a Philips CM200 and with a JEOL 2000EX. A suspension of the sample in ethanol was placed on a holey-carbon grid. X-ray photoelectron spectroscopy (XPS) at room temperature was carried out on Escalab 220 XL spectrometer (Vacuum Generators). A monochromatic Al Kα X-ray source was used and electron energies were measured in the constant analyzer energy mode. The pass energy was 100 eV for the survey of spectra and 40 eV for the single element spectra. All XPS binding energies were referred to C1s core level at 285 eV. The angle between the incident X-rays and the analyzer was 58°, photoelectrons being collected perpendicularly to the sample surface. Spectra were analyzed with the Casa-XPS software. Characterization techniques X-ray powder diffraction (XRD) was performed on a Bruker AXS D8 Advance diffractometer working in Bragg-Brentano geometry using Cu Kα radiation (λ=1.54 Å), equipped with an energy dispersive detector (Sol-X). Patterns were collected at room temperature, in the 2θ=10–90° range, with a 0.02° step and 10 s counting time per step. The EVA software was used for phase identification. Laser Raman Spectroscopy (LRS) was performed using a Spectra Physics krypton ion laser at room temperature using the 647.1 nm excitation line. The beam was focused onto the samples using the macroscopic configuration of the apparatus. To avoid damage due to laser heating, all compounds were studied at a very low laser Catalysts Preparation Nickel ferrite spinel was synthesized by hydrothermal method, using nitrates salts as precursors. NaOH (2 M) was added to an aqueous solution of Ni(NO3)2⋅6H2O (0.05 M) and Fe(NO3)3×9H2O (0.07 M) up to pH=10 before loading in the hydrothermal cell. The cell was heated up to 180°C for 1 hours. After reaction, the product was washed with distilled water and dried at 80°C. Correlation of the structural, textural and redox properties with catalytic decomposition The reducibility of the catalyst was also studied by temperature programmed reduction by hydrogen (H2-TPR) and by thermal analysis TGA-DSC under H2. H2-TPR was carried out on Micromeritics-Autochem II 2920 with a TCD detector to monitor the H2 consumption. After calibration of H2 on the TCD, the sample displayed in a U-shaped quartz reactor was pre-treated in argon, and heated at 5°C/min from 25 to 1000°C in 5% H2 in Ar gas mixture. In-situ TGA-DSC under H2 atmosphere was performed using a HidenIsochema gravimetric analyzer (model IGA-003) under 5% H2 in N2 flow from 25 to 500°C. Ca. 20 mg of sample was used. Measurements at 400°C for 5 h under H2 atmosphere were also carried out. The total acidity (Lewis + Brönsted) and acid strength distribution of catalysts were determined by temperature programmed desorption of ammonia (NH3-TPD). The NH3-TPD was performed using 150 mg of catalyst saturated with NH3 (40 cm3/min). After saturation, NH3 weakly adsorbed was desorbed in a He flow, at the adsorption temper- ature, until no NH3 was detected in the outlet gas. The NH3-TPD was performed by raising the temperature up 550°C at a heating rate of 5°C/min. The NH3 was measured by gas chromatography using a TCD detector. 395°C as obtained by NH3-TPD (Figure 7). Another suggestion to explain the dehydration pathway of our NiFe2O4 sample in 350–400°C could be the migration, towards the surface, of Fe3+ species which is most acidic compared to Ni2+[25–27,34] as confirmed by Raman technique, the presence of supplementary α-Fe2O3 oxidized phase in addition to the main NiFe2O4 spinel phase. The decrease of the reaction rate in the temperature range 300–350°C is probably due to two reasons: (i) first, to the drop of the specific surface area when the reaction temperature increased and (ii) the decrease of the surface basicity per unit mass (g) of catalyst as ascertained by NH3-TPD. Catalytic tests The catalytic tests corresponding to isopropanol decomposition were carried out between 200 and 500°C in a fixed bed continuous flow reactor under atmospheric pressure and in the absence of oxygen, by passing the carrier gas (Nitrogen) through a saturator containing the alcohol and maintained at 273 K. The weight of catalyst was ca. 0.05 g and the flow rate was 40 mL STP min1. Product analysis was performed by gas chromatography using a 10% Carbowax 20 M on Chromosorb 200 column and a Flame Ionization Detector. Conclusions Bulk ferrite spinel prepared using hydrothermal route with 114 m2/g was investigated in isopropanol catalytic decomposi- tion in the temperature range 200–400°C. Structural properties evaluated by Raman spectroscopy shows, beside spinel structure, the presence of α-Fe2O3 as supplementary phase. The decomposition of isopropanol leads, on our catalyst, to the formation of propylene and acetone suggesting respectively the presence both acid sites and basic/redox sites on the surface of NiFe2O4 solid. The distribution of these products depends strongly to the distribution of Ni2+ and Fe3+ species on the surface; it changes depending on the reaction temper- ature. This evolution is probably linked to the evolution of the surface composition. Fe3+ would promote the formation of propylene, while nickel species would catalyze the formation of acetone. Our bulk NiFe2O4 issued from hydrothermal conditions having shown basic properties at low temperatures and acidic properties at high temperatures have probably undergone an evolution of the surface by migration of Fe3+ species from bulk to the surface. © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH ChemistrySelect 2023, 8, e202204361 (8 of 9) Data Availability Statement [22] M. Giarola, G. Mariotto, D. Ajo, J. Raman Spectrosc. 2012, 43(4), 556–558. [22] M. Giarola, G. Mariotto, D. Ajo, J. Raman Spectrosc. 2012, 43(4), 556–558. [23] http://www.xpsfitting.com. The data that support the findings of this study are available from the corresponding author upon reasonable request. [22] M. Giarola, G. Mariotto, D. Ajo, J. Raman Spectrosc. 2012, 43(4), 556 558. [23] http://www.xpsfitting.com. [23] http://www.xpsfitting.com. [23] http://www.xpsfitting.com. [24] V. K. Mittal, S. Bera, R. Nithya, M. P. Srinivasan, S. Velmurugan, S. V. Narasimhan, J. Nucl. Mater. 2004, 335(3), 302–310. [25] P. Moriceau, A. Lebouteiller, E. Bordes, P. Courtine, Phys. Chem. Chem. Phys. 1999, 1(24), 5735–5744. Keywords: acid-base properties · characterization · isopropanol · reducibility · spinel. Keywords: acid-base properties · characterization · isopropanol · reducibility · spinel. Phys. 1999, 1(24), 5735–5744. [26] E. Bordes-Richard, Top. Catal. 2008, 50(1-4), 82–89. [27] E. Bordes-Richard, P. Courtine, Optical Basicity: a Scale of Acidity/Basicity of Solids and its Application to Oxidation Catalysis, in: J. L. J. Fierro (Ed.), Metal Oxides: Chemistry and Applications, Marcel Dekker, 2005. [1] H. G. Karge, V. Dondur, J. Weitkamp, J. Phys. Chem. 1991, 95, 283–288. [1] H. G. Karge, V. Dondur, J. Weitkamp, J. Phys. Chem. 1991, 95, 283–288. Metal Oxides: Chemistry and Applications, Marcel Dekker, 2005 [2] A. Auroux, Top. Catal. 2002, 19(3–4), 205–213. [28] D. Das, H. K. Mishra, A. K. Dalai, K. M. Parida, Appl. Catal. A. 2003, 243(2), 271–284. [3] A. Auroux, Top. Catal. 1997, 4(1–2), 71–89. [4] S. Bennici, A. Auroux, J. Jackson D (eds), Thermal Analysis and Calorimetric Methods, Wiley-VCH Verlag GmbH & Co. KGaA, 2009, pp. 391–441. [29] H. Zou, Y. S. Lin, Appl. Catal. A Gener. 2004, 265(1), 35–42. [30] W. Shan, W. Shen, C. Li, Chem. Mater. 2003, 15(25), 4761–4767. [31] B. M. Mogudi, P. Ncube, R. Meijboom, Applied Catalysis B: Environm. 2016, 198, 74–82. [5] L. M. Kustov, Top. Catal. 1997, 4(1–2), 131–144. [6] M. Stocker, Microporous Mater. 1996, 6(5–6), 235–237. [32] R. Benrabaa, H. Boukhlouf, A. Löfberg, A. Rubbens, R. N. Vannier, Bordes-Richard, A. Barama, J. Nat. Gas Chem. 2012, 21(5), 595–604. 32] R. Benrabaa, H. Boukhlouf, A. Löfberg, A. Rubbens, R. N. Va [7] S. A. Bocanegra, A. D. Ballarini, O. A. Scelza, S. R. De Miguel, Mater. Chem. Phys. 2008, 111(2–3), 534–544. Bordes-Richard, A. Barama, J. Nat. Gas Chem. 2012, 21(5), 595–604. [33] R. Benrabaa, A. Löfberg, A. Rubbens, E. Bordes-Richard, R. N. Vannier, A. Barama, Catal. Acknowledgements The Fonds Européen de Développement Régional (FEDER), CNRS, Région Hauts-de-France, Chevreul Institute (FR 2638) and Minis- tère de l’Education Nationale de l’Enseignement Supérieur et de la ChemistrySelect 2023, 8, e202204361 (8 of 9) Research Article doi.org/10.1002/slct.202204361 Research Article doi.org/10.1002/slct.202204361 ChemistrySelect Recherche are acknowledged for funding XPS/LEIS/ToF-SIMS spectrometers XRD and SEM instruments. The authors are also grateful to Olivier Gardol, Martine Trentesaux, Laurence Burylo and Nora Djelal for the technical assistance. [14] A. Gervasini, A. Auroux, J. Phys. Chem. 1990, 94(16), 6371–1379. [15] P. N. Trikalitis, P. J. Pomonis, Appl. Catal. A 1995, 131(2), 309–322. [16] R. Benrabaa, A. Löfberg, J. Guerrero Caballero, E. Bordes-Richard, A. Rubbens, R. N. Vannier, H. Boukhlouf, A. Barama, Catal. Commun. 2015, 58, 127–131. [17] K. N. Subramanyam, J. Phys. C. 1971, 4, 2266. [18] M. N. Iliev, D. Mazumdar, J. X. Ma, A. Gupta, F. Rigoto, J. Fonctcuberta, Phys. Rev. 2011, B83, 014108. Conflict of Interest [19] A. Ahlawat, V. G. Sathe, V. R. Reddy, A. Gupta, J. Magn. Magn. Mater. 2011, 323(15), 2049–2054. All co-authors declare no conflict of interest. [20] D. Varshney, K. Verma, Mater. Chem. Phys. 2013, 140, 412–418. [21] L. C. Prinsloo, P. Colomban, J. D. Brink, I. Meiklejohn, J. Raman Spectrosc. 2011, 42(4), 626–632. © 2023 The Authors. ChemistrySelect published by Wiley-VCH GmbH ChemistrySelect 2023, 8, e202204361 (9 of 9) [1] H. G. Karge, V. Dondur, J. Weitkamp, J. Phys. Chem. 1991, 95, 283–288. Submitted: November 10, 2022 Accepted: January 12, 2023 [2] A. Auroux, Top. Catal. 2002, 19(3–4), 205–213. Data Availability Statement Today 2013, 203, 188–195. [8] J. A. Wang, X. Bokhimi, O. Novaro, T. Lopez, R. Gomez, J. Mol. Catal. A 1999, 145(1–2), 291–300. [34] C. G. Ramankutty, S. Sugunan, B. Thomas, J. Mol. Catal. A 2002, 187(1), 105–117. [9] C. Ancion, G. Poncelet, Appl. Catal. A 1994, 108, 31–40. [10] J. C. Luy, J. M. Parera, Appl. Catal. 1984, 13(1), 39–48. [11] R. Grabowski, B. Grzybowska, K. Samson, J. Słoczyński, J. Stoch, K. Wcisło, Applied Catalysis A. Gen. 1995, 125(1), 129–144. [12] M. A. Aramendia, V. Borau, C. Jimenez, J. M. Marinas, A. Porras, F. G. Urbano, J. Catal. 1996, 161, 829–838. Submitted: November 10, 2022 A t d J 12 2023 Submitted: November 10, 2022 Accepted: January 12, 2023 [13] A. Gervasini, A. Auroux, J. Catal. 1991, 131(1), 190–198. ChemistrySelect 2023, 8, e202204361 (9 of 9)
https://openalex.org/W1578247965	https://revistas.pucp.edu.pe/index.php/economia/article/download/2159/2089, https://revistas.pucp.edu.pe/index.php/economia/article/download/2159/2089/	es		Impacto del libre flujo de capitales en la intermediación bancaria en el Perú: 1988-1997	Economía/Economía	1,998	cc-by	15,053	ECONOMÍA. Vol. XX1 N” 42 Diciembre 1998 IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA EN EL PERÚ: 1988-1997 Lourdes Coll Calderón RESUMEN La afluencia de los capitales hacia América Latina en la década del noventa, atraídos por los indicadores más favorables de las variables fundamentales de crecimiento del producto y control de la inflación, ha permitido la expansión del volumen de fondos a ser intermediados, con lo coral se ha puesto en riesgo la calidad de los activos bancarios. El artículo examina las medidas que se han tomado para controlar el efecto de los flujos de capital en el crédito interno del Perú, y se contrasta estos resultados con los de otros países de la región; asimismo, se hace una revisión de las nuevas formas de supervisión implementadas a raíz del cambio de régimen. Se analiza también los riesgos que todavía no están cubiertos con la aplicación de las normas internacionales de Basilea. ABSTRACT The inflow of capitals into Latin America during the 1990s attracted by more favorab1.e economic indicators, such as a growing product and inflation under control, has allowed for the expansion in the volume of funds handled Lourdes Coll Calderón by the banking system, risking a deterioration of the quality of bank assets. This article examines the measures taken to control this problem in Peru, as well as in other countries in the region, in order to compare their results. We also make a review of the new bank supervision norms, and an analysis of the risk that has not yet been covered with the application of the Base1 norms. INTRODUCCIÓN Desde principios de los años noventa el sistema financiero peruano ha pasado por un proceso de crecimiento; sin embargo, anteriormente se han vivido etapas de crisis financieras y políticas, especialmente durante 1987, cuando se intentó estatizar el sistema financiero peruano. Esa crisis fue muy severa y se agravó por el manejo económico que provocó la hiperinflación. A finales de 1990, el gobierno peruano dictó las medidas necesarias para entrar a un proceso de ajuste y estabilización a fin de combatir estas crisis. Dentro de estas medidas se incluyeron dispositivos relativos a la regulación bancaria y la promoción del mercado de capitales. El objetivo era restaurar los mecanismos básicos de financiamiento del sector productivo, profundizar el desarrollo del sistema financiero y contribuir de esta manera al crecimiento económico, siguiendo la tesis de McKinnon 1. Por otra parte, la nueva legislación ha permitido al sector bancario, además de intermediar depósitos y realizar colocaciones, acceder a mercados diferentes promocionando inversiones reales, como es el caso de la banca de inversión, aparte de la posibilidad de ofrecer nuevos productos. Además, aunque ya existían mecanismos informales de transferencia de dinero, con las medidas de liberalización financiera y cambiaria se oficializó el libre flujo de capitales, lo cual ha permitido incrementar en términos reales, entre 1990 y 1997, los niveles de liquidez de 697 a 3,203 millones de nuevos soles de agosto de 1990, es decir en un 360%. Mientras que el grado de intermediación, 1. 60 Según la cual existe una fuerte relación de causualidad entre el desarrollo financiero y el desarrollo económico. Es más, existe evidencia que el nivel de desarrollo financiero, alcanzado por un país, predice un crecimiento futuro, además de una mayor acumulación de capital y mayor cambio tecnológico. Ver Levine (1997). IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA subió de 5.1% a 21.0% del PB1 entre 1990 y 1998 (ver Gráfico 1)2. Estos indicadores sugieren que la etapa de “represión financiera”-? que vivió la economía peruana antes de estas medidas estaba prácticamente acabando. En el contexto internacional también se ha visto un proceso de liberalización generalizada de los movimientos de capital. Esta creciente dimensión de los mercados tinancieros ha contribuido en los últimos años a aliviar el sector externo de América Latina. Los flujos financieros han crecido espectacularmente en los años recientes. Esta recuperación de los flujos financieros desde inicios de los noventa fue muy positiva para América Latina, por cuanto se removió la restricción externa dominante que deprimía los niveles de inversión. Esta afluencia de capitales ha resurgido por los atractivos prospectos en los indicadores de las variables fundamentales del crecimiento y el control de la inflación que se vienen dando en países de América Latina; también contribuyó a este redireccionamiento la disminución de las tasas de interés internacionales. Sin embargo, existen repercusiones en otros aspectos macroeconómicos, como la apreciación del tipo de cambio, la expansión de los bienes no transables a expensas de los transables y el incremento de los déficits en cuenta corriente. Respecto del destino de los capitales, esta nueva ola de flujos financieros tiene características diferentes sobre los anteriores flujos de capital, como los de fines de los años sesenta y principios de los setenta que estuvieron dirigidos al sector público, y es que esta vez los flujos se están dirigiendo predominantemente al sector privado en forma de patrimonio (equity capital) en lugar de llegar en forma de deuda. Una de las consecuencias de estos grandes flujos de capital que han llegado al país es la expansión del volumen de los fondos a ser intermediados a través del sistema financiero doméstico. Esta expansión podría implicar 2. Debe grado aclararse que los niveles de liquidez de intermediación tiene como fuente 3. La represión con controles tienen como al BCRP fuente al INEI, mientras que el financiera se identifica con controles de la tasa de interés y particularmente que generan tasas de inteh real negativas sobre los depósitos. 61 Lourdes Coll Calderón CUADRO I INDICADOR DEL GRADO DE Ih’TERMEDIACIóN (liquidez. como % del PBI) EN EL PERÚ 25 20 15 10 5 0 T Fuente: I I 90 91 92 Memorias del BCRP (varios I 93 I 94 95 I 96 I 97 98 números) problemas en la calidad de activos del sistema financiero, debido a la posibilidad que el crédito de alto riesgo se incrementeq tal vez hasta el punto de atentar contra la estabilidad del sistema y se genere una crisis de endeudamiento. Dada las profundas medidas de liberalización, es posible preguntarse si las fuerzas de mercado y de política económica dictadas ejercen un control suficiente sobre los mercados financieros como para controlar un entorno latinoamericano que históricamente se ha comportado demasiado inestable. El control de esta volatilidad en los mercados financieros presenta dos enfoques alternativos. De un lado, los partidarios de la reforma liberal aducen que los onerosos requisitos reglamentarios sobre los bancos pueden reducirse mediante el fortalecimiento de la supervisión bancaria. Estos controles, basados en normas que actualmente se aplican en las economías de mercado 4. 62 Entendiéndose por “cddito de alto riesgo” aquél sujeto a que el deudor o la contraparte de un contrato financiero no cumpla con las condiciones del contrato. La Ley del Sistema Finaaciero y del Sistema de Seguros y Orgánica de la Superintendencia de Banca y Seguros clasifica a los deudores en las siguientes categorías: normal, con problemas potenciales, deficiente, dudoso y pérdida. Dentro de la definición de “ctiito de alto riesgo” entran todas las categorías a excepción de la categoría “normal”. IMPACTO DEL LIBRE FXUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA... industrializadas de la Organización de Cooperación y Desarrollo Económicos (OCDE), están diseñados para prevenir el crecimiento excesivo del crédito otorgado a prestatarios riesgosos. Los escépticos, de otro lado, sin embargo, sostienen que las fuerzas de mercado y la supervisión aunque apropiadas para los países de la OCDE no constituyen restricciones suficientemente fuertes en los mercados mas volátiles de América Latina. El presente documento presenta un análisis de las medidas que se han tomado para controlar el efecto de los flujos de capital del exterior en el crédito interno tanto en el Perú como en países seleccionados de la región latinoamericana. Posteriormente, se hace una revisión de las nuevas normas de supervisión implementadas en el Perú, a raíz del cambio de régimen. El análisis se concentra en las debilidades del ratio del capital adecuado frente a la necesidad de mejorar la supervisión bancaria como respuesta a la volatilidad causada por los flujos de capital. 1. FLUJOS DE CAPITAL EN EL CONTEXTO LATINOAMERICANO 1.1. Un Alcance Tebrico acerca de la Dependencia de los Flujos de Capital en el Desarrollo Económico Uno de los temas claves que. enfrenta la economía internacional, y que se discute tanto académicamente como en la práctica, esta relacionada con los niveles y mecanismos de intermediación financiera internacional. Más aun, los problemas se han vuelto en cierto modo más complejos a medida que la intermediación financiera internacional, además de mantener aún su clásico propósito de canalizar ahorros hacia inversión productiva, ha adquirido nuevas y más complicadas facetas. A medida que la superestructura de los mercados de inversiones y de divisas se ha vuelto enorme y compleja, y que las necesidades mundiales se han vuelto más diferenciadas, el resultado ha sido que muchos más de los flujos brutos a través de los mercados de capitales del mundo de hoy reflejan decisiones financieras que no están directamente atadas a (ni necesariamente conducen a) la formación física de capital como ocurría en el pasado. Sin embargo, la mayoría de autores están de acuerdo con que el proceso de desarrollo económico normalmente incluye un requerimiento tanto de ca63 Lourdes Coil Calderón pital externo a largo plazo como de crédito a corto plazo en la cuenta corriente de los países en desarrollo; más aun, la mayoría de países que hoy están industrializados (y particularmente los EE.UU., Canadá, Australia y Nueva Zelandia) descansaron fuertemente en el capital externo y los préstamos externos para su desarrollos. Debe ser reconocido que hay muy poco consenso entre teóricos del desarrollo económico sobre la exacta relación entre capital externo y desarrollo, aunque la mayoría de pensadores acepta que el capital externo es necesario para acelerar el desarrollo. En América Latina, los críticos del rol de capital externo debido a sus potenciales efectos negativos en la autonomía nacional han sido los analistas ligados a la escuela de la “dependencia”. Al otro lado del espectro, no solamente se afirma categóricamente la necesidad del capital externo para el desarrollo, sino también se ha tratado de cuantificar, a través de un modelo matemático relativamente simple, cuánto capital externo debería ser transferido a un país en desarrollo y a qué tasa podría ser absorbido. Un influyente estudio en esta línea es el famoso modelo de “dos brecha@, en el que las brechas aludidas son aquélla! que resultan de comparar las necesidades de inversión con la proyección de ahorro nacional, y las necesidades de importación con la proyección de exportaciones. A pesar de que esta aproximación fue útil para resaltar el rol del capital externo en el desarrollo, ha sido más tarde criticada por su excesivo y casi mecánico énfasis en la falta de divisas como la única limitación al desarrollo; así como su concentración a nivel agregado, descuidando consideraciones sectoriales. Otra explicación7 de la ligazón entre capital externo y desarrollo señala que la aceleración de la tasa de crecimienjo de los países en desarrollo implicaba un aumento más que proporcional en la demanda de importaciones. Se enfatizaba, además, que la tendencia al desequilibrio externo es inherente al proceso de desarrollo, y que la transferencia de recursos financieros puede jugar, junto con otras medidas a nivel nacional como internacional, un importante rol en ayudar a eliminar o controlar estos desequilibrios externos. Finalmente, concluye que para que los países del Tercer Mundo crezcan y se .5 Seiber 6. Ver 7. Prebisch 64 (1977). Chenely y Strout(1966). (1959). IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA.. desarrollen rápidamente necesitan obtener flujos netos externos sostenidos de recursos financieros bajo términos que no sean excesivamente onerosos. Un estudio más reciente revela que en América Latina existe una fuerte correlación entre el crecimiento y las entradas de capitals. El documento muestra que durante el periodo de afluencia de 1976-1982, la economía de América Latina creció cerca de un 4% al año, mientras recibía una afluencia neta de capitales de cerca del 5% del PBI. Asimismo, en el período de escasez de capitales que se inició en 1982 y finalizó en 1990, el crecimiento se redujo a menos del 1.5% y la afluencia neta de capitales a la región decreció a menos de 1%. Mientras que en el período 1991-1994 el crecimiento de alrededor de 3.5% anual estuvo acompañado de una afluencia de capitales que crecía a un 6% anual. Esta relación no es totalmente causal, debido a que el crecimiento y la afluencia de capitales dependen en gran medida de aspectos’ similares. Y la causalidad se manifiesta claramente en más de una’dirección, ya que los inversionistas internacionales tienden a buscar regiones en las ¿lue puede esperarse un rápido crecimiento económico. Sin embargo, las cifras ilustran lo que la teoría y la experiencia latinoamericana llevan a pensar: la afluencia de capitales puede proporcionar un impulso expansivo a la economía interna y, a la inversa, una marcada reducción de la tasa de afluencia puede resultar fuertemente contraccionaria. 1.2. Los Flujos Financieros hacia América Latina 1.2.1 Los Años Cincuenta A medida que la Segunda Guerra Mundial llegaba a su término, dos desarrollos tuvieron una influencia importante en la naturaleza futura de los flujos financieros internacionales: la creación de las instituciones de Bretton Woods y el gran peso de los EE.UU. en la economía mundial. Las dos Instituciones de Bretton Woods -el Fondo Monetario Intemacional y el Banco Mundial- fueron creadas como parte de un intento por 8. Ver Gavin, Hausmann y Leidermau (1996). 65 LourdesColl Calderón evitar la inestabilidad de los tipos de cambio y el caos financiero que había acompañado la depresión mundial de los años treinta, así como los controles proteccionistas y devaluaciones competitivas del período de entreguerras. Bretton Woods y las instituciones que creó representaron el primer intento coherente de imponer orden y estabilidad en el sistema monetario internacional a fin de preservar la unidad del sistema monetario. El intento de controlar el sistema monetario internacional a través de las instituciones oficiales, sin embargo, no fue acompañado en la práctica por importantes tlujos provenientes de ellas, particularmente, hacia los países en desarrollo. La principal característica de los movimientos internacionales de capital durante los años de posguerra fue el gran flujo de fondos desde EE.UU. Durante los últimos años de la década de 1940 y los años cincuenta, los Estados Unidos eran claramente más ricos y poderosos que nunca, y fue el único país capaz de ofrecer tanto ayuda económica como protección militar. A pesar de que controlaban las instituciones de Bretton Woods, los EE.UU. preferían canales bilaterales para sus flujos financieros. Por ello, tanto las instituciones oficiales bilaterales como las multilaterales jugaron un rol insignificante durante los años cincuenta. Los flujos financieros hacia América Latina tomaron la forma de inversión directa a través de corporacior.es multinacionales. Los países latinoamericanos empezaron a implementar políticas de sustitución de importaciones a fin de combatir los déficits en cuenta corriente. Dicha inversión estaba protegida por altas barreras arancelarias y favorecida por crédito interno barato. Las corporaciones multinacionales saltaron las altas barreras arancelarias impuestas por estos países, a través del establecimiento de nuevas empresas o la compra de empresas nacionales ya existentes. 1.2.2 Los Años Sesenta En los años sesenta, la actitud estadounidense hacia la ayuda oficial comenzó a cambiar rápidamente. Como señala Prout (1976), la motivación del aumento de la ayuda fue “desvergonzadamente política”. Los americanos creían que al fomentar el desarrollo económico y social de las naciones no comprometidas contribuían al desarrollo de las sociedades que posteriormente apoyarían su posición y forma de vida. 66 IMPACTO DEL LIBRE F%UJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA... Con la Revolución Cubana, se puso en boga plantear que la totalidad de América Latina enfrentaba sólo dos alternativas: reforma o revolución. Los EE.UU. necesitaban hacer un gran esfuerzo para procurar que la primera alternativa prevaleciera. Es por ello que con la administración Kennedy se creó el programa de la Alianza para el Progreso, cuya retórica inicial estuvo fuertemente influida por un sector muy importante del pensamiento latinoamericano. Particularmente influyente fue el documento preparado por Prebisch y otros destacados economistas latinoamericanos, donde se hacía un llamado especial a los EE.UU. para cooperar con aquellos países latinoamericanos deseosos de hacer cambios estructurales. La retórica inicial de la alianza para el Progreso fue bastante progresista, y sus tres objetivos oficialmente establecidos fueron: crecimiento económico, cambio estructural y democraciatización política. Después de la muerte de Kennedy, la política de los EE.UU. hacia Latinoamérica se volvió más conservadora. Las reformas estructurales y la democratización de los países latinoamericanos dejaron de ser objetivos oficiales de la Alianza. Hacia el final de 1966, la política de los EE.UU. hacia América Latina estaba claramente en su fase más pragmática El nivel total de la asistencia financiera bruta de los EE.UU. hacia América Latina durante los años sesenta fue muy impresionante; sin embargo los desembolsos netos fueron sustancialmente menores. Más de la mitad de la asistencia económica fue dedicada a repagos, amortización de préstamos oficiales previos, e intereses. Los créditos oficiales y la ayuda estaban supuestamente destinados a financiar proyectos de inversión gubernamental. Sin embargo, el flujo de salida neto de los fondos de inversión privada estadounidense desde Latinoamérica hacia los EE.UU. excedió el flujo de entrada neta de los fondos oficiales estadounidenses. Esta tendencia explica, quizás, el respaldo de las multinacionales estadounidenses a la Alianza para el Progreso; sin embargo no se puede inferir que la ayuda oficial facilitó la salida del capital privado extranjero. 67 Lourdes Coll Calderón 1.2.3 Los Años Setenta Durante los años setenta existieron cuatro claras tendencias en los flujos externos hacia América Latina. En primer lugar, comparando este periodo con los años sesenta, la proporción de financiamiento externo proveniente de fuentes privadas aumentó notablemente, de 39.8% en 1961-65 a 92.7% en 1978 (ver Cuadro 1). El segundo cambio está relacionado con la composición de tales flujos privados. Como puede verse en el mismo cuadro, durante los años sesenta los fondos eran canalizados predominante a través de inversiones directas. Desde los años setenta, sin embargo, eran los bancos multinacionales los que proporcionaban la principal fuente de fondos, superando tanto a la inversión directa como a los fondos públicos. Una tercera tendencia, que puede observarse en el mismo cuadro, es que durante los años setenta -y particularmente en 1973- hubo un agudo crecimiento en el nivel nominal y real del total de flujos financieros netos hacia America Latina. Estos préstamos fueron principalmente usados para cubrir los crecientes déficits de la cuenta corriente de la balanza de pagos. Como resultado del aumento de precio del petróleo y de otros bienes importados, combinando coli una reducida demanda para sus exportaciones a raíz de la recesión de los países industrializados, los déficits en cuenta corriente de los países latinoamericanos -y particularmente los importadores de petróleo crecieron sustancialmente. La cuarta característica de los flujos financieros después de 1973 fue que los miembros de la OPEP reemplazaron en algunos años a los países industrializados como proveedores de capital, tal como puede apreciarse en el Cuadro 2. De esta manera, el deficit de ahorro nacional respecto a la inversión doméstica en el mundo importador de petróleo fue financiado por el excedente en ahorros de los exportadores de petróleo. 68 IMPACTO DEL ESTRUCTURA LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA.. CUADRO 1 Y NIVEL DE LOS FLUJOS NETOS DE RECURSOS EXTERNOS HACIA AMÉRICA LATINA, 1961-1978 (A) (promedios anuales en porcentajes) J. Flujos públicos netos A. Multilaterales 1. Desarrollo 2. Compensatorios B. Bilaterales 1. EE.UU. 2. Otros países (b) 60.2 19.5 16.6 2.9 40.7 36.9 3.8 40. I 15.1 17.1 -1.4 24.4 223.6 0.8 25.2 13.4 ll.6 1.8 ll.8 6.8 5.0 19.6 14.4 6.6 7.8 5.2 2.6 2.6 12.0 7.4 8.4 -1.0 4.6 1.7 2.9 7.3 3.1 7.2 -4.1 4.2 0.8 3.4 JJ. Flujos privados netos (c) A. Bancos (d) B. Proveedores C. Bonos D. Inversión Directa 39.8 2.1 7.7 5.0 25.2 59.9 9.3 13.8 2.5 33.3 74.8 43.8 2.3 2.5 26.2 80.4 61.0 3.7 3.3 12.4 88.0 48.3 5.8 14.8 20. I 92.7 56.6 9.8 10.3 16.0 100.0 100.0 1575.8 2641.3 100.0 100.0 7561.9 15301.5 III. Total (%) Total (millones de dólares) 100.0 100.0 15637.0 21807.2 (a) Incluye los paises miembros del Banco Interamericano cias subregionales. (b) Incluye los países socialistas y los miembros de la OCDE excepto los EE.UU. (c) Incluye créditos para nacionalizaciones. (d) Jncluye instituciones financieras de Desarrollo y Agen- no bancarias. Fuente: GrJfJJth-Jones (1984). 69 Lourdes Coll Calderón ESTRUCTURA CUADRO 2 DE LOS FLUJOS DE CAPITAL (miles de millones de dólares) Grupo de los Diez países industrializadosy Suiza Paísesdesarrolladospequeños Paísesen desarrollo no exportadoresde Petróleo Paísesen desarrollo exportadoresde petróleo -8.25 GLOBALES (A) -0.75 -19.5 23.5 51.5 I .5 17.75 10.5 12.0 22.0 6.0 21.0 22.5 36.0 51.0 -1.25 -38.0 -4.5 -68.0 -116.0 (a) El total de flujos de capital incluye movimientos monetarios (esto es, posición en cuenta corriente con el signo cambiado); signo negativo indica exportación del capital. (b) Promedios anuales. Fuente: Griftith-Jones (1984). El principal medio para “reciclar” este importante volumen de fondos de los países superavitarios hacia los deficitarios fue el sistema bancario multinacional privado de los países industrializados. En consecuencia, se puede decir que el principal factor del crecimiento de los préstamos a los países en desarrollo fue el rápido aumento de los depósitos provenientes tanto de países exportadores de petróleo como de otras fuentes, acompañado de una disminución de la demanda de crédito por parte de los clientes tradicionales a causa de la recesión en los países industrializados. Ello motivó a que los bancos prestaran a clientes hasta entonces catalogados como marginales. Al mismo tiempo hubo factores que introdujeron una mayor inclinación a endeudarse, tanto en el sector público como privado de los países en desa70 IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA.. rrollo. Los gobiernos se embarcaron en programas que implicaban una fuerte expansión de la inversión pública, la frecuentemente tenía una alta proporción de bienes de capital importados que, por lo general, no estaban íntegramente financiados con recursos del país. Ello estuvo acompañado de un relativo estancamiento en el flujo neto de la asistencia oficial para el desarrollo. Además, los gobiernos de los países en desarrollo desarrollaron actitudes tendientes a favorecer una expansión del financiamiento externo proveniente de fuentes bancarias privadas. Los gobiernos comenzaron a preferir los préstamos privados en la medida en que éstos implicaban un mínimo tipo de condicionalidad en los desembolsos. cual’ Los préstamos podían hacerse efectivos en forma muy rápida y formalmente casi no venían acompañados de atadura o condiciones en cuanto a la forma en que los respectivos fondos deberían ser empleados o sobre la política económica del país receptor. El desarrollo de nuevas técnicas operativas (innovación financiera) en los mercados de euromonedas, especialmente hacia fines de los años sesenta, redujeron el riesgo en los eurobancos por los préstamos efectuados a los países en desarrollo. Las innovaciones que se introdujeron disminuían los riesgos derivados de grandes préstamos sujetos a largos periodos de amortización, que son el tipo de créditos requeridos por los países en desarrollo. Entre las innovaciones que se introdujeron está el crédito rotatorio (roll-over). Esta modalidad se basa en una tasa de interés flotante que varía aproximadamente según el costo de la moneda para el prestamista, el cual obtiene sus fondos en el mercado interno de corto plazo. -La base para fijar la tasa de interés en estos préstamos a los países en desarrollo es la tasa LIBOR (London Inter-Bank Offered Rate) o bien el “Prime Rate” de los EE.UU., que constituyen una medida del costo de los fondos para los bancos. En adición a esta variable básica, se carga un margen (spread) sobre el LIBOR o Prime; éste refleja tanto la liquidez del mercado como la capacidad de crédito (credit worthiness) del prestatario en un momento determinado. El costo del crédito incluye además una “comisión de servicio” y una “comisión de compromiso”, que se aplican sobre la porción no utilizada de los préstamos. Aunque el préstamo otorgado a un país en desarrollo pueda tener un largo periodo de amortización, la tasa de interés es modificada cada vez que el préstamo es extendido (normalmente cada 3 ó 6 meses). Esta “tasa de interés flotante” es decisiva porque transfiere al prestatario uno de los riesgos más importantes 71 Lourdes Col1 Calderón del mercado (es el prestatario el que absorbe tanto los cambios cíclicos como de largo plazo que experimentan las tasas de interés). Otra innovación importante en esta época fueron los “préstamos sindicados”, pues una proporción importante de las transacciones realizadas por los países en desarrollo en los mercados de euromonedas había sido canahzada a través de préstamos sindicados. Estos préstamos, que se iniciaron a finales de los sesenta, son compartidos por un gran número de eurobancos. Este mecanismo ha permitido que el riesgo por incumplimiento que se presenta en los grandes préstamos se distribuya entre un gran número de bancos. 1.2.4 Los Años Ochenta : La década perdida Desde 1982 en adelante, el tema del financiamiento en los países en desarrollo se centró en cómo enfrentar las reiteradas dificultades de pago; lo que se puso en juego no sólo fueron los resultados de las finanzas particulares de cada país que no pudo afrontar sus pagos, sino el propio sistema financiero internacional. Las preocupaciones principales de los países en desarrollo en cuanto a las finanzas de largo plazo pasaron a ser el financiamiento de.sus deudas vencidas. La razón más importante para la continua demanda de financiamiento externo se relacionó con el retinanciamiento del servicio de las obligaciones vencidas. La característica principal de los años ochenta fue la aguda escasez de divisas, lo cual llevó a un retroceso de la tasa de crecimiento del producto de muchos países. Particularmente, en el Perú la tasa de crecimiento promedio del producto entre 1980 y 1989 fue -0.9 %. Lo cual ha llevado a caracterizar la década de 1980 como “la década perdida”. Los años ochenta fueron “la década perdida” también para Latinoamérica, en general, debido no sólo al problema de la deuda externa, sino también por la declinación de las economías, y por la reducción de los niveles de vida; ademas, las aspiraciones respecto a un futuro mejor se vieron tremendamente disminuidas. Aun cuando ningún país latinoamericano repudió su deuda formalmente, hubo una serie de atrasos en el servicio de la deuda, lo cual dio lugar a una suspicacia extrema por parte de los acreedores que dudaron en incrementar su exposición en la región. Muy por el contrario, los bancos comerciales, que en realidad constituían los principales acreedores privados, 72 IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA.. redujeron su exposición liquidando sus préstamos, rechazando renovar líneas de crédito rotatorio, y /o denegando crédito fresco para amortizar antiguas obligaciones. La severidad de la crisis, originada por el problema de la deuda externa, al final de la década, dejó pesimistas a muchos observadores sobre las posibilidades de renovación de los flujos externos de capital hacia la región. En cambio, la expectativa prevaleciente era’ que las regiones fuertemente endeudadas debían esperar, en el futuro cercano, un alivio financiero proveniente, principalmente, de las fuentes oficiales de crédito. Sin embargo, al final de la década resurgieron los flujos externos de capital privado hacia América Latina, lo cual sorprendió incluso a los observadores de los países receptores, tanto por el volumen como por la composición de los flujos. 1.3 Los Flujos Financiyos en Iòs Años Recientes Tras un paréntesis de casi una década, la afluencia de capitales intemacionales a América Latina se reinició a principios de los noventa. Después de haberse ajustado a la escasez, la región pronto debió enfrentar el desafío relativamente poco familiar de tener que manejar una abundancia de capitales. El renovado flujo de capitales fue ampliamente percibido como un voto de confianza internacional a las medidas de estabilización y, liberalización adoptadas en los años anteriores por la mayor parte de las economías latinoamericanas, y como una’ valiosa oportunidad para canalizar el ahorro intemacional hacia la promoción del desarrollo de la economía interna. Estas fluctuaciones en el volumen de los flujos de capital hacia América Latina revelan, en primer lugar, su volatilidad; y en segundo lugar.que los flujos de capital hacia América Latina están determinados en gran medida por’ los acontecimientos que ocurren en la economía mundial, y que son esencialmente exógenos a América Latina. Según Gavin, Haussmann y Feiderman (1996), los flujos netos a la región están altamente correlacionados con la tasa de interés mundial; sobre la que los acontecimientos económicos ocurridos en América Latina tienen escasa influencia. La correlación positiva entre los flujos de capital a América Latina y el superávit en cuenta Corriehte de las principales economías industrializadas, según el mismo estudio, también pone de relieve la vinculación que existe entre las fluctuaciones en la disponibilidad de capitales internacionales y los flujos de capital a la región. 73 Lourdes Coll Calderón Ello no significa que los flujos de capital no se vean afectados por el entorno interno de políticas. Durante los años noventa, los países latinoamericanos que llevaron a cabo agresivas políticas de liberalización, redujeron la intlación y mantuvieron o crearon un sistema comercial y financiero abierto, recibieron capitales que aquellos que no lograron esos resultados. En resumen, independientemente del entorno interno de políticas, es probable que los flujos internacionales de capital sigan constituyendo un importante mecanismo a través del cual se trasmiten a América Latina los shocks que afectan a las economías industrializadas. 1.4 Los jlujos de capital en el caso peruano Los instrumentos financieros utilizados para efectuar la transferencia de capitales tienen un rol, importante en el comportamiento del grupo de residentes del país receptor y de los inversionistas internacionales. De esta manera, los flujos de capital también pueden tener un impacto significativo sobre el grado de volatilidad a los cuales se ve sujeta la economía. Es probable que los flujos sean mayores si los flujos son a corto plazo y de naturaleza puramente especulativa que si en su mayoría reflejan inversiones extranjeras directas, las cuales presumiblemente están guiadas por variables fundamentales a mediano o largo plazo. Esto es particularmente cierto en la deuda a corto plazo que es líquida y puede ser vendida por los inversionistas internacionales en cualquier momento; en el caso de capital a largo plazo los propietarios asumen gran parte del riesgo del precio, además del riesgo asociado a una fuga generalizada de capitales de la economía. La composición de los flujos también está determinada por el grado de sensibilidad respecto a las diferencias en las tasas de interés. Los mecanismos a través de los cuales el capital se canalizó al Perú difieren notablemente en los años recientes de los que tipificaron la afluencia de fines de los setenta y de los ochenta. Como puede verse en el Gráfico 2, los préstamos a largo plazo al Sector Público, principalmente provenientes de fos bancos comerciales, constituyeron la forma dominante de intermediación internacional a fines de los setenta, representando más del 70% de la afluencia de capitales hacia Perú. 14 IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA... En marcado contraste, los préstamos de bancos comerciales constituyeron cerca del 5% en 1996, mientras que la inversión extranjera directa y la inversión por privatización constituyeron el 46% y 41%, respectivamente9. Estos cambios tienen importantes implicaciones con respecto al manejo de los shocks que afectan la cuenta de capital. La participación mucho mayor de la inversión extranjera de capital significa que los inversionistas extranjeros están expuestos en forma más directa y comparten más plenamente el riesgo país. Las políticas pueden afectar la composición de la afluencia de capitales en diferentes formas. Las políticas económicas sustentables y verosímiles probablemente se traduzcan en flujos de capital menos volátiles, ya que los inversionistas internacionales estarán más dispuestos a asumir compromisos relativamente irreversibles con respecto a la economía. 2. ENFOQUE MONETARIO DE LA BALANZA DE PAGOS Tal como se ha mostrado anteriormente, históricamente ha existido una gran afluencia de capitales hacia América Latina, especialmente en la segunda mitad de este siglo. El objeto de estudio de este documento es analizar las implicancias que ha tenido este movimiento de capitales en el manejo de la política monetaria de 10s países de la región. Para realizar este análisis se utiliza el Enfoque Monetario de la Balanza de Pagos, según el cual todo desequilibrio en la balanza de pagos refleja un desequilibrio monetario, ya que el superávit o déficit afecta la liquidez del sistema, y por tanto a la intermediación bancaria. Aun cuando la Aproximación Monetaria a la Balanza de Pagos data de la época de David Humeto, su estudio cobra nuevamente vigencia después de 9. No se considera dentro del cáIcuIo de los porcentajes el correspondiente a “Capitales de Corto Plazo” debido a que el año 1996 dicho rubro alcanzó la cifra negativa de 57 millones de US dólares. 10. David Hume (1711-1776) argumentó que un superávit permanente no era posible, por lo que no tenía sentido centrar los objetivos en subsanar la balanza, en igual forma el déficit 7.5 GRhICO 2 FLUJOS DE CAPITAL AL PERÚ POR TIPO DE INVERSIÓN 1992 1979 Inversión Extraniera Directa Capital a CP 1979-1997 Inversión Extranjera Directa 8% Prkstamos de LP S. Publ. 72% héstatnos de LP S. Publ. 12% Capital a CP 86% 1996 1996 Inversión Extranjera Directa Prlvatización 41% Inversión en Cartera 8% Préstamos de LP S. Publ. 5% Inversión Extraniera Caoital a CP ~~~~~ Publ. 14% Privatización 2% Inversión en pubI, Cartera 6% 6% IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA. _. la Segunda Guerra Mundial con la introducción de dos importantes aportes. El primero, que precedió al interés académico en políticas monetarias, estuvo desarrollado por funcionarios del FMI. Su exponente más importante fue J.J. Polack’ l. Robert Mundell y Harry Johnson, de la Universidad de Chicago, realizaron similares aportes. Los funcionarios del Fondo desarrollaron este enfoque a fin de contar con módulos que sirvieran para monitorear el manejo macroeconómico, pues sólo se disponía de estadísticas rudimentarias como fuente de información. Cuando el Fondo desarrolló los primeros estudios en las economías de los países en desarrollo encontró que muy pocos de los problemas subyacentes se podían resolver con la ayuda de la estructura analítica Keynesiana; principalmente, porque a principios de los cincuenta muchos de los países en vías de desarrollo no contaban con Cuentas Nacionales en ingreso y producción lo suficientemente detalladas para hacer un análisis de ingreso nacional y de determinación de la balanza de pagos en concordancia a la línea Keynesiana y a la Teoría de la Absorci6nl2. Sin embargo, sí se podía contar con las estadísticas monetarias de los Bancos Centrales, obtenidas en el trabajo de supervisión del sistema bancario. También existía información de comercio exterior. En vista de la disponibilidad de este tipo de datos en muchos países en vías de desarrollo, donde otra clase de información estadística era escasa, se pensó en utilizar esta estructura de análisis para aprovechar totalmente los datos disponibles. En opinión de algunos académicos, como Chacholiades (1992), las limitaciones del enfoque monetario son bastantes fuertes, pues considera a la balanza de pagos como una válvula de seguridad que se abre automáticamente para eliminar un exceso de oferta de dinero en forma de déficit en la balanza de pagos o para admitir dentro del país una cantidad adicional de dinero en forma de superávit en la balanza de pagos, con el fin de satisfacer un exceso existente de demanda de dinero. se curaría por sí mismo. Esto era posible porque el patrón oro (gold standard) tenfa un mecanismo de ajuste automático. Visto de otra manera, cada país tiene una cantidad “natural” de oferta monetaria que corresponde con su capacidad productiva. II. Cuyo artkulo and Payments más impoltante en este sentido, Problem”, apareció en 1957. “Monetary !T. Según la Teoría de la Absorción la Balanza Comercial entre la producción de una economía y su capacidad Analysis of Income Formation es igual a la diferencia de absorción. que existe 77 Lourdes Coll Calderón Al considerar los desequilibrios como un fenómeno temporal y autocorrector, el enfoque monetario conduce prácticamente a la conclusión de que, por lo general, no es necesaria ninguna medida de política para corregir los desequilibrios en la balanza de pagos. Si las autoridades son lo suhcientemente pacientes y permanecen pasivas, un desequilibrio de pagos se corregirá de manera automática, tarde o temprano. El problema, claro está, es que la corrección puede darse muy tarde y mientras tanto la economía tiene que sufrir, más adelante, un innecesario costo de ajuste. Sin embargo, en opinión de Mario Blejer (1993), es verdad que, en su más estilizada caracterización, para el enfoque monetario los cambios en las reservas internacionales están directamente relacionados al mercado monetario de modo que, bajo tipo de cambio fijo, el cambio en reservas debe igualar la diferencia entre el cambio real de la demanda monetaria y el cambio en el componente doméstico de la oferta monetaria. Pero esta caracterización se basa en supuestos muy restrictivos, como la total integración entre bienes internos y externos y de los mercados de capital. Esta también es una caracterización de equilibrio de largo plazo, pero es consistente con distintas alternativas de ajuste de corto plazo. En la actualidad, muchos de los supuestos restrictivos se han levantado, y en su lugar se han tenido resultados más realistas. Pero la estructura más general preserva la relación entre los sectores externos y monetarios como lo postula el enfoque monetario (esto es, que los desequilibrios monetarios recaen en la balanza de pagos), y esta relación es lo suficientemente fuerte como para no perder su ~validez ante diferentes conjuntos de supuestos. Se puede llegar a la conclusión que esperar utilizar el enfoque monetario exclusivamente para formular políticas de largo plazo basados en que se cumplan los supuestos ideales es caracterizar este enfoque de una manera estrecha y distorsionada. El Enfoque Monetario de la Balanza de Pagos nos dice que cuando el tipo de cambio es controlado por las autoridades monetarias, la balanza de pagos viene a ser uno de los medios por el cual la oferta monetaria nominal es ajustada. De tal forma que la oferta monetaria nominal no está mas bajo el control directo de las autoridades monetarias y viene a ser una variable endógena dentro del sistema. Lo que las autoridades monetarias controlan es el volumen del crédito interno. La distinción entre la creación de dinero y el crédito interno es crucial, ya que en este caso, el banco central lo que controla es el crédito interno. Por otro lado, cuando el tipo de cambio es flexible, 78 IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA hipotéticamente, no deberían existir cambios en el nivel de reservas. Sin embargo, por rigideces del mercado esta afirmación no es del todo cierta. En este caso, las autoridades monetarias están en disposición de controlar el nivel de reservas, con lo cual ejercen un control sobre la oferta monetaria real. En el contexto latinoamericano y peruano se ha visto que ha existido un incremento de los flujos externos de capital, lo cual ha repercutido en la elevación de las Reservas Internacionales, y en consecuencia en la Liquidez del Sistema. Los incrementos en los movimientos crediticios internos también ha sido característica de estos últimos años. Ya que el Enfoque Monetario de la Balanza de Pagos relaciona estas tres variables, siendo su ecuación fundamental: Reservas + Crédito Interno = Liquidez del Sistema. Obtenemos de esta manera el sustento teórico que relaciona las tres variables que nos interesan para el análisis del control monetario en la entrada de los flujos de capital. La siguiente sección tratará de evaluar la efectividad de las políticas monetarias en controlar el crédito interno, cuya fuente son los flujos de capital extranjero. Un indicador de dichos esfuerzos, en el contexto latinoamericano, es la relación que se encuentra entre el Crédito Interno y la Liquidez. 3. IMPLICANCIAS DEL LIBRE FLUJO DE CAPITALES EN EL CRÉDITO INTERNO EN LATINOAMERICA 3.1 Introduccidn Los mercados internacionales han crecido espectacularmente desde mediados del decenio de 1960. Los movimientos de capital reflejan la expansión de la economía mundial, el aumento del comercio internacional, la globalización de la producción; todos estos factores se vinculan a los netamente financieros, los cuales han aumentado a un ritmo notablemente más rápido. En los decenios de 1970 y 1980, muchos países comenzaron a liberalizar sus sectores financieros nacionales y a abandonar la regulación de las “transacciones de divisas”. Todo lo cual, combinado con los revolucionarios avances tecnológicos en la gestión de la información y de las telecomunicaciones, ademas del surgimiento de técnicas financieras cada vez más perfeccionadas, contribuyó a un auge de los flujos financieros nacionales e internacionales. 79 Lourdes Coll Calderón Hablar de mercados financieros integrados resulta prematuro, puesto que la movilidad del capital internacional está lejos de ser perfecta. No obstante, no hay duda que los flujos de capital y la integración financiera internacional aumentan velozmente. Esta circunstancia ha causado una gran controversia. En un extremo están aquellos que ven la creciente integración como un signo de mayor eficiencia; según esta interpretación, los mercados están superando la represión financiera que caracterizaba la ineficiente regulación gubernamental. En el otro extremo están aquéllos que consideran que el auge de los flujos de capital constituye una especulación, con un alto riesgo implícito. Para contrarrestar estas dudas, se ha hecho una evaluación siguiendo el enfoque monetario de la Balanza de Pagos para Perú y otros países de la región. Se ha escogido cuatro países de mayor desarrollo -Argentina, Brasil, Chile y México-, y otros tres países de desarrollo similar -Bolivia, Ecuador y Colombia. Las razones para haber escogido estos países son, en primer lugar, porque son socios comerciales y su desarrollo’ influye notablemente en el desenvolvimiento de la economía peruana; en segundo lugar, porque han seguido políticas económicas similares, abocándose a las reformas y apertura de sus sistemas financieros, procurando una adecuada adaptación de las nuevas políticas financieras al ritmo de la liberalización y de la apertura. No obstante que las reformas financieras se han efectuado atendiendo a las características particulares de cada país, se advierten elementos comunes tales como: mayor participación de los mecanismos de mercado en la determinación de las tasas de interés, mayor captación de recursos financieros, disminución de los requerimientos de encajet3, y cambio en el diseño de los enfoques en la supervisión bancaria. Además, se ha tenido presente que el requisito básico para la implementación exitosa de las políticas de apertura y reforma del sistema financiero es contar con un entorno macroeconómico estable. En este aspecto, la mayorfa de países de la región ha logrado avances importantes en el terreno de la estabilización económica y el control de la inflación. Al mismo tiempo, se ha decretado la liberalización de las tasas de interés y el tipo de cambio, como parte de la implementación de políticas monetarias restrictivas. 13. 80 En el Perú, los requerimientos para incentivar los préstamos invariable. de encaje en MN, por lo general, en MN, sin embargo, el encaje han tendido a disminuir en ME ha permanecido IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA... En el Cuadro 3, se puede observar que casi todos los países en estudio han tenido un crecimiento económico similar, a excepción de Chile, Perú y Argentina, que han estado por encima del promedio, y de México que ha estado bastante rezagado, en el período 1991-1995. Por otro lado, se observa que casi todos los países presentan una apreciación en su moneda, producto de la masiva afluencia de capitales hacia esta región. Ademas, a excepción de Chile, la mayoría de los países muestran un bajo ratio de ahorro interno/PBI. La información de las principales variables monetarias del Sistema Monetario Peruano y Latinoamericano se encuentra en los Cuadros 4 al ll, donde se consignan las cuentas del Sistema Monetario de Perú, Brasil, Chile, México, Argentina, Bolivia, Colombia y Ecuador. Estos cuadros han servido para poder graficar la efectividad de las políticas monetarias para controlar las variables monetarias frente al flujo de capitales externos. Un indicador de dicha efectividad es la relación Crédito Interno/Liquidez del Sistema, los cuales se presentan en los Gráficos 3 al 10, conservándose el orden de paísest4. De los gráficos mencionados, se observa que Chile ha mantenido durante todo el período analizado un índice cercano al 0.5 y Perú se ha ido acercando paulatinamente a ese ratio, mientras que el resto de países ha seguido trayectorias similares, aunque con tendencia a mejorar el ratio. Por ejemplo, Argentina, que al inicio del periodo mantenía un ratio de 1.25, lo ha ido disminuyendo paulatinamente hasta llegar a 0.87 en 1997. El caso de Brasil es fluctuante, pues si bien al inicio del periodo sustentaba un ratio de 0.94, en 1993 alcanzó el 1.03, para luego disminuir a 0.81 en 1997. México, si bien 14. Esta relación es el complemento de la relación Reservas InternacionalesiLiquidez del Sistema. Su importancia está muy ligada con el nuevo concepto para el manejo efectivo de las Reservas Internacionales. La literatura tradicional destaca el papel de los flujos de comercio en la determinación del nivel óptimo de reservas de un país. Los analistas típicamente relacionan el nivel óptimo de las reservas con la vulnerabilidad de los ingresos por exportación y el grado de apertura de las economías. Mientras tanto, en un mundo con una mayor movilidad del capital, una mayor integración y sofisticación financiera, la pregunta acerca del nivel óptimo de las reservas tiene que ver menos con las corrientes de comercio que con la necesidad de evitar las crisis financieras. Sebastián Edwards plantea que la manera de analizar el nivel de reservas esté relacionado mis con los pasivos del gobierno o del sector bancario, y que todavía se está planteando la manera cómo efectuar este análisis. Ver Edwards (1997). 81 Lourdes Coll Calderón comenzó Con un ratio bajo de 0.76, en 1994 y 1995 alcanzó 1.02; este ratio se justifica por la crisis suscitada en ese país, to cual llevó al retiro de los capitales externos. Colombia comienza el periodo con un nivel razonable de 0.52, subiendo paulatinamente hasta alcanzar el 0.72. Ecuador, si bien comienza con un nivel alto de 1.11, disminuye solamente hasta el nivel de 0.76 en 1997. Bolivia mantiene niveles estables entre 0.72 y 0.80. De los países analizados se observan dos grupos claramente diferenciados, los que han alcanzado el ratio de 0.5 y el resto. Estos dos grupos de países han enfrentado diferentes criterios de política. Se puede ver que aun cuando el crecimiento de la .liquidez ha sido notable en Chile y Perú, 554% y 260% entre 1988 y 1997 respectivamente, este primer grupo ha controlado el crecimiento del crédito interno y por consiguiente aumentado sus Reservas Internacionalesl5. 3.2 Políticas de intervención En general, dentro de un ámbito de liberalización financiera, los instrumentos adoptados se han orientado a dos niveles de intervención: intervención esterilizada e .intervención no esterilizada. Estos mecanismos de intervención obedecen a la inesperada abundancia de financiamiento externo a los que se enfrenta un país. Dado que dicho flujo de capital puede ser en parte transitorio o, en su defecto, fluir con demasiada rapidez como para que las economías lo absorban con eficacia es necesario, entonces, intervenir. El objetivo es estimular flujos cuyo volumen esté en concordancia con la capacidad interna de absorción de la economía, y encauzarlos hacia proyectos de inversión’ productivos. Las posibles combinaciones de políticas monetarias y cambiarias llevan a distinguir dos alternativas de intervención principales: a) Intervención no esterilizada. Este primer nivel de intervención ha sido adoptado con frecuencia por países que se plantean como principal objetivo la estabilidad de los precios, usando como ancla un tipo de cambio nominal fijo, y que están dispuestos a aplicar una política 15. En Perú, cerca del 40% de las Reservas corresponden a encajes de los bancos que son mantenidos en el BCR. Esto denota el esfuerzo del BCR por esterilizar los flujos externos. 82 PRINCIPALES CUADRO 3 VARIABLES MACROECONóh4lCAS Per6 23.6 7.1 5,572 -3,447 85.6 22.5 27.8 16.2 20.1 18.5 32.5 15.5 18.9 4.0 -0.3 0.2 1.2 46,506 16,039 79,542 20,969 -18,136 160 -1,576 -2,399 70.6 92.4 121.1 87.0 16.0 25.2 20.2 ll.3 26.1 16.0 4.7 8.9 4.7 1,042 10,396 4,358 -302 -4,054 -826 126.6 74.9 78.8 Países más desarrollados Brasil Chile México Argentina 2.6 7.2 0.8 5.2 Paísesde Desarrollo Similar Bolivia Colombia Ecuador 4.0 4.5 3.8 1.5 2.8 1.5 Fuente: CEPAL, Estudio Económico de America Latina y el Caribe, 1995-1996 Lourdes Coll Calderón monetaria pasiva. Se puede singularizar a Argentina en el decenio de 1990 entre los países que han estado más cerca de la forma más pura de esta alternativa. La inflación ha disminuido marcadamente y la economía ha experimentado una fuerte recuperación de la actividad económica. Sin embargo, la apreciación del tipo de cambio, junto con la liberalización de las importaciones ha contribuido a un acentuado deterioro de la balanza comercial, y el déficit en cuenta corriente se ha acrecentado haciendo al país vulnerable a caídas de ingresos de capital, en tanto que el ahorro interno ha mostrado una tendencia a estancarse o bajar. Puesto que el ahorro externo estimula desproporcionalmente la actividad interna, la trayectoria de la economía depende en gran medida de la sostenibilidad de los flujos de capital16. b) Intervención esterilizada. Este enfoque supone una esterilización de los efectos monetarios al acumular reservas durante períodos de, intensificación de las entradas de capital. El propósito es aislar la oferta monetaria de las fuertes fluctuaciones derivadas de la movilidad del capital. Este tipo de esterilización, si es eficaz, impide que bajen las tasas internas reales de interés y limita la expansión de la demanda agregada. Este tipo de intervención ha sido preferido por los países que ya han salido de situaciones recesivas y que mantienen una política monetaria activa y, al mismo tiempo, una postura más cautelosa en lo que respecta a las entradas de capital. Refleja una preocupación por el desarrollo sostenido del sector rubros comerciales y la canalización del capital extranjero hacia el ahorro y la inversión. En las economías que hacen uso pleno de su capacidad productiva, la esterilización tiene la ventaja de controlar el gasto e impedir una mayor apreciación real cambiaría. Sin embargo, existe una opción intermedia para los gobiernos que quieren utilizar un tipo de cambio flexible y la intervención del Banco Central como instrumentos. Esta opción es más realista. Siendo el monto de acumulación de reservas una opción de política. Cuanto más agresiva es la acumulación de reservas más se afslan las presiones sobre el. tipo de cambio nominal. Si las reservas son bajas al comienzo,>del período’de afluencia de Rujos 16. 84 Devlin, Ffmch-Davk y Giifith-Jones (1995). ‘- IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA... de capital, puede esperarse una acumulación de reservas; después puede existir una disminución de la tendencia a acumular reservas. 3.3 Situación en el Perú En julio de 1990, si bien existía un control de cambios con un sistema de cambios múltiples, en la práctica existía ya una apertura de capitales con el exterior. Una gran parte de los agentes económicos tenían atesorada una magnitud considerable de divisas, existiendo además mecanismos informales mediante los cuales se podía transferir moneda extranjera del país al exterior y viceversa. De esta manera existía de facto un libre flujo de capitales, por lo que una liberalización oficial de la cuenta de capitales representaba básicamente una disminución de los costos de transacciónt7. En el caso peruano, la liberalización de la cuenta de capitales involucraba decidir sobre aspectos importantes del sistema cambiario. Existía un sistema de tipo de cambio múltiple en el cual, junto a un mercado oficial con tipos de cambio diferenciados para distintos grupos de importaciones y exportaciones, coexistían dos mercados cambiarios libres: uno referido a los certificados en moneda extranjera que se otorgaba a los exportadores y el otro correspondía al mercado informal. De otro lado, una buena parte de los depósitos bancarios estaban constituidos en moneda extranjera, estando éstos sujetos a topes en las tasas de interés que eran ligeramente superiores a las tasas vigentes del mercado internacional. En consecuencia, la apertura de capitales con el exterior no solamente incluía la oficialización del libre flujo de capitales, sino también la liberalización de las tasas domésticas de interés en moneda extranjera. Frente a las expectativas de mayor rentabilidad a raíz de la liberalización de la economía se produce un flujo rápido de capitales del exterior que no puede ser completamente absorbido por la economía, lo cual contribuye a la apreciación del tipo de cambio. 17. Los mecanismos informales de transferencia de moneda extranjera están referidos al uso que tenían ciertas instituciones financiens no formalmente establecidas de transferir dinero al exterior a un costo relativamente alto, por actuar fuera de las reglas establecidas por entidades reguladoras en ese momento. Por lo cual, liberalizar asfixiantes, automáticamente bajaba los costos de transacción. el mercado de estas normas Lourdes Coll Calderón Posteriormente, en setiembre de 1990, con la administración Fujimori se implementó un exigente programa de estabilización y reforma económica. Este programa de estabilización adoptó una política de control monetario basado en metas decrecientes de la base monetaria. Sin embargo, la presión revaluatoria sobre el tipo de cambio real que ejerce la entrada de capitales, y el reconocimiento de una potencial vulnerabilidad del sistema financiero ante una salida repentina de capitales, llevó al BCR, casi desde un inicio -1992-, a emplear instrumentos monetarios para mediatizar estos efectos macroeconómicos: por un lado, el mantenimiento de un encaje marginal de 50% a los depósitos en moneda extranjera, que posteriormente se ha reducido a 45%, y de otro lado, la compra de divisas. Sin embargo, el hecho más importante es que el BCR ha sido consciente que la entrada de capitales externos al Perú requerí; la aplicación de un alto tipo de encaje en moneda extranjera como medida precautorial. Con esta medida se pretendía restringir, en alguna medida, la significativa cantidad de recursos financieros por efecto de la entrada de capitales, por los problemas potenciales de solvencia que pudieran surgir en el sistema financiero, ya que las entidades del sistema financiero -al no poder absorber adecuadamente estos fondos- los colocarían en alternativas de alto riesgo, pudiendo desembocar en situaciones de deterioro en la calidad de los activost9. 3.4 Situación en países de desarrollo similar al peruano Al igual que en el resto de la región, en Bolivia, Ecuador y Colombia se han llevado a cabo procesos de liberalización financiera en medio de fuertes desequilibrios macroeconómicos e insuficiencia de ahorro interno. 18. Este efecto se ve potenciado por el hecho de que la liquidez en moneda extranjera representa alrededor de un 70% de la liquidez total del sistema. Sin embargo, la banca nacional puede endeudarse con el exterior en condiciones bastantes favorables ya que no necesitan guardar encaje con los créditos que tienen del exterior. 19. Ver Memoria 86 del BCR-l993’, p. 69. IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA.. . La interacción entre las condiciones macroeconómicas y la situación del sistema: bancario en América Latina ha demostrado que los riesgos de crisis bancarias son grandes si los equilibrios macroeconómicos no están plenamente logrados. La puesta en marcha de un programa de fuerte liberalización o desregularización, sin que se hayan consolidado las mejoras institucionales correspondientes, incluida una supervisión más eficaz, tiende a aumentar la vulnerabilidad del sistema financiero. Por otra parte, las incertidumbres y los desórdenes de índole político también propician en gran medida la aparición de crisis bancarias en la región. Por ejemplo, en Ecuador, aunque se produjeron crisis políticas, problemas de confianza y de desajustes macroeconómicos, las dificultades bancarias obedecieron en especial a los bajos estándares de información y a un marco regulatorio y de información deficiente. Por otro lado, aun cuando la economía de Bolivia ha gozado de estabilidad en los últimos años, se ha debido hacer frente a las crisis suscitadas en su sistema bancario. Lo cual puede atribuirse, sin perjuicio de los avances registrados desde fines de la década pasada, a las situaciones microeconómicas e institucionales, tales como la ausencia de marcos regulatorios acordes con los riesgos de un mercado poco intervenido y en rápida expansión. 3.5 Situacih en patses de mayor desarrollo al peruano Para fines comparativos se han tomado en estudio las experiencias en Brasil, Chile, México y Argentina, como países de mayor desarrollo a los anteriormente mencionados. A excepción de Chile, cuyo mercado bancario ha dado muestras de extraordinaria bonanza, pues ya lleva 11 años de crecimiento consecutivo con depósitos del orden del 60% del PBI, retornos reales sobre el capital cercanos al 20% anual, y una morosidad de cartera no mayor a un 0.7%;:10s otros dos países han enfrentado situaciones de alta inestabilidad. .i. Es bien conocida ya la perdida de confianza que suscitó la crisis mexicana en 1994, la cual llevó no sólo al retiro del capital externo del país. Ademas, desencadenó en la región una ola de réplicas que lesionaron las instituciones financieras de algunos países. Esta ola afectó también a Argentina y Brasil cuya situación se vio agravada por la necesidad de realizar fuertes ajustes macroeconómicos. , Lourdes Coll Calderón Es importante notar que el alto índice Crédito Interno/Liquidez que detentan Brasil México, y Argentina (ver Gráficos 3, 5 y 6) implica un alto grado de riesgo para la región por el fuerte peso económico y financiero que tienen estos tres países. 3.6 Composición reservas-crédito interno en la liquidez del sistema Se puede decir que esta composición dependerá de la evaluación que las autoridades económicas hagan al respecto y de lo que consideren sus necesidades. Si las reservas son bajas al comienzo de un período de afluencia, puede esperarse que los Bancos Centrales acumulen reservas en las etapas iniciales, aun en el caso en que no deban hacerlo en virtud de un compromiso cambiario formal. A lo largo del tiempo, puede decrecer la tendencia a acumular reservas. En consecuencia, el nivel de Reservas Internacionales que se quiere tener depende de las posibilidades de reversión de los flujos entrantes de capital. Si existe una alta probabilidad de que el capital que está fluyendo decida retirarse en un futuro próximo, entonces serfa prudente acumular un mayor nivel de Reservas. En el caso concreto de los países en estudio, ha existido un incremento de Reservas Internacionales particularmente fuerte en Chile, Bolivia, Perú y Colombia con aumento de 554%, 323%, 260% y 258% respectivamente, en el periodo 1988-1997. Por otro lado, los países que detentan un menor ratio Reservas / Liquidez, o alternativamente un mayor ratio Crédito Internokquidez son, coincidentemente, los que menos Reservas Internacionales han acumulado en este último periodo. En orden decreciente, tenemos que Argentina, Ecuador, México y Brasil han variado sus Reservas en 159%, 125%, 88% y -36%, respectivamente. Es importante distinguir los objetivos de la política de un gobierno cuando se trata de determinar su nivel de reservas internacionales. Por un lado, el objetivo puede ser obtener un ancla cambiaria que evite el incremento en el nivel de precios; y por otro, el objetivo puede ser aislar la oferta monetaria de fluctuaciones, a fin de evitar una expansión de la demanda agregada. En ambos casos, la determinación de la política que se decida seguir determina al, mismo tiempo el nivel del crédito interno. Y en este sentido, el objetivo es ‘también evitar que éste genere situaciones de riesgo en caso de una reversión de los flujòs. 88 IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA... 3.7,~ ComideracioneS de enlace En resumen, el tema tratado ha sido determinar los instrumentos que permiten aminorar los efectos de los flujos de capital que ingresan al país, como los incrementos los niveles de Reservas Internacionales que han ayudado a controlar el excesivo crecimiento de la liquidez. Sin embargo, por otro lado, existe otro tipo de tratamiento de los efectos en los flujos de capital sobre el sistema bancario, el cual está referido a las técnicas de supervisión bancaria Los defensores de la liberalización han argumentado que la adopción de mejores técnicas podría reemplazar a los controles directos sobre el sistema bancario. A diferencia de los instrumentos de control de la liquidez que sólo controlan el volumen, la supervisión bancaria idealmente afecta la cantidad y la calidad de la ,expansión del crédito. Existe una importante herramienta de supervisión que se ha aplicado durante la última década, y es una norma internacional para la relación entre el capital de un banco y sus activos ponderados en función del riesgo. Esta norma requiere que los bancos mantengan más capital en relación con los activos de mayor riesgo, lo cual debería estimular a los bancos a mantener carteras con menor riesgo. Sin embargo, resulta difícil determinar si las normas de capital están logrando restringir el incremento de .los activos con riesgo, o si algunos bancos están declarando en forma inapropiada sus coeficientes capital/préstamos. Si los activos son valorados en forma inapropiada, el capital bancario refleja una situación falsa de-calidad del balance de un banco. Por lo que el desafío fundamental que enfrentan los responsables de la formulación de políticas es diseñar herramientas de supervisión para controlar el riesgo financiero aplicable a instituciones individuales. En el siguiente punto se abordará el tema, en forma precisa, de los alcances que ha tenido la reforma de la supervisión bancaria en nuestro medio y, en general, las limitaciones que tiene el ratio aludido (Ratio de Basilea) en cuanto a su capacidad de cubrir otros riesgos. 89 CUADRO g 4 CUENKASDEL SISTEMA MONETARIO PERUANO 1988 (0918) 1989 1990 1992 2.035,OO 4.372,OO 2.544.00 5.029,oo 1993 1994 8.311,OO 14.802,OO 8.947,OO 9.588,OO 1991 8,86 0,68 0.000129 2,63 17,04 333,00 988,00 19,67 1.321,OO 1995 17.089,OO 14.025,OO 1996 1997 24.784,OO 25.570,OO 17.181,OO 24.885,OO (1.395,35) 6.658,91 5.263,57 0,002666 988,00 6.391,60 1.771,28 2.632,60 5.255,32 3.291,07 7.379,59 7.026,60 4.579,oo 0,188000 0,773ooo 9.401,OO 1,246OOO 3.508,83 4.036,12 7.544,94 17.258,OO 24.390,oo 1.988000 2,195oOO 4.180,58 6.743,5 1 4.500,50 4.368,11 8.681,09 11.111,62 31.114,oo 41,965.OO 2,253OOO 7.585,OO 10.103,55 6.225,03 13.810,03 7.004,08 9.355,26 17.107,62 50.455,oo * Corresponde hasta el mes de agosto Fuente: Estadísticas del Fondo Monetario Internacional 2,453OOO 2,660OOO 9.612,78 5.923,67 18.968,05 IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA., 8 91 CUADRO 5 CUENTASDEL SISTEMA MONETARIO BRASILE&fO 1988 1989 1990 1991 1992 1993 1994 1995 19% 1997* 0,002578 0,048779 0,710OOo (L376000) 65,951000 0,038905 0.041483 o,ooooooo95 0,684056 0,732835 0,000001031 9,674OOO 10,384OoO 0,000024836 51,964OOo 50,588OOO 0,000148000 755,285OOO 821,236OOO 0,001641000 2,384,OOOOOO20.885,OOOOOO23.268,OOOooO 0,0320OOOOO 30.724,OOOOOO 179.763,OOOOOO 210.487,OOOOOO0,639OOOOOO 44.817,OOOOOO 229.570,000000 274.387,OOOOOO0,918000000 53.327,OOOOOO 254.684,OOOOOO 308.011,000000 1,005000000 56.923,OOOOOO 246.392,OOOOOOO 303.315,000000 1,088000000 * Corresponde hasta el mes de agosto Fuente: Estadísticas del Fondo Monetario Internacional 27.137 47.312 28.588 (9.297) 40.190 74.500 48.081 48.820 53.062 52.319 409.526 663.488 389.515 351.108 460.259 652.656 281.319 250.076 253.417 226.463 436.663 710.800 418.103 341.811 500.449 727.156 329.401 298.897 306.479 278.782 IMPACTO DEL LIBRE FLUJO DE CAPITALES c. 7 9 r w o- EN LA w o- d o- INTERMEDIACIÓN (v 0” BANCARIA. ._ 0 o- 93 CUADRO6 $ CUENTAS DEL SISTEMA MONETARIO PERUANO 1988 -1283,0 5003,o 3.720,O 245,05 3.965,1 4.210,1 8.175,2 1989 -720,O 5737,7 5.017,7 267,16 5.284,9 5.552,O 10.836,9 1990 263,2 6707,9 6.971,1 304,90 7.276,O 7.580,9 1991 1071,9 7652,0 8.723,9 349,22 9.073,1 9.422,3 18.495,5 ’ 14.856,9 1992 1684,7 9172,8 10.857,5 362,58 11.220,1 11.582,7 22.802,7 1993 2229,8 11397,4 13.627,2 404,17 14.031,4 14.435,5 28,466,9 1994 3341,8 12406,6 15.748,4 420,18 16.168,6 16,588,8 32.757,3 1995 4376,0 15182,0 19.558,O 396.77 19.954,8 20.351,5 40.306,3 1996 5939,7 17676,8 23.616,m5 412,27 24.028,8 24,441,O 48,469,8 1997* 7532,9 18582,l 26.115,O 414,85 26.529,9 26.944,7 53.474,6 * Corresponde hasta el mes de agosto Fuente: Estadísticas del Fondo Monetario Internacional IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN 8 8 .. Y 3 BANCARIA.. CUADRO 7 CUENTASDEL SISTEMAMONETARIO MEXICANO 1988 6,493 115.328 121.821 2,273 1 2.856 50.736 1989 7.388 167.930 175.318 2,4615 68.223 1990 1991 9.868 23.958 244.341 333.747 254.209 357.705 2,8126 3,0184 3.001 3,508 1992 31.250 406,436 437.686 3,0949 1993 1994 48.848 (10.953) 457.940 602.391 506.788 591.438 3,1156 3,3751 (3,243 1995 1996 1997* (13.134) 21.914 91.133 739.848 758.932 702.287 726.714 780.846 793.420 6,4194 7,5994 7.8857 (2.046) 2,884 ll.557 L * Corresponde hasta el mes de agosto Fuente: Estadísticas del Fondo Monetario Internacional 7.937 10.097 15.679 53.592 71.224 90,382 86.874 110.571 118.508 131.324 141,422 146.983 178.481 162.661 175.236 115.252 99,867 113.206 102.751 89.058 100.615 IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA... 97 CUADRO 8 CUENTASDEL SISTEMAMONETARIO ARGENTINO 1988 1989 - 1990 1991 (4.413,OO) (6.145,OO) 22.272,OO 40.987,OO 17.859.00 34.842,OO 0,558500 0,998500 (7.901,52) 39.878,25 31.976,72 (6.154,23) 41.048,57 34.894,34 1992 1993 2.537,OO 7.877,OO 6.402,OO 50.479,OO 60.811,OO 53.016,OO 68.688,OO 0,990500 0,998500 2,561,33 7.888,83 50.963,15 60.902,35 53.524,48 68.791,19 1994 - 0,133700 - 0,179500 67.373,OO 73.775,OO 0,999500 6.405,20 67.406,70 73.811,91 1995 1996 2.574,OO 70.927,OO 73.501,oo 1,000000 2.574,OO 70.927,OO 73.501,oo 7.698,OO 1997* 12.409,OO 75.239,C-O 84.520,OO 82.937,OO 96.929,OO 0,999500 0,999500 7.701,85 12.415,21 75.276,64 84.562,28 82.978,49 96.977,49 * Corresponde hasta el mes de agosto Fuente: Estadísticas del Fondo Monetario Internacional IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA.. 8 x 99 CUADRO9 CUENTASDEL SISTEMAMONETARIO BOLIVIANO 739 1.937 2.676 2,3505 3 14,40 824,08 í.138,48 756 957 3.202 4.749 3.958 5.706 2,6917 3,1727 280,86 301,64 1.189,58 1.496,83 1.470,45 1.798,47 1.335 6.242 7.577 3,5806 372,84 1.743,28 2.116,13 1.480 1.607 9.056 13.821 10.536 15.428 3,9005 4,265 1 379,44 376,78 2.321,75 3.240,49 2.701,19 3.617,27 1.730 16.917 18.647 4,6205 374,42 3.661,29 4.035,71 2.341 3.970 5.044 16.541 18.853 20.182 18.882 22.823 4,8003 5,0746 486,68 782,33 3,445,83 3.715,17 3,933,50 4.497,50 25.226 5,2366 963,22 3.854,03 4.817,25 * Corresponde hasta el mes de agosto Fuente: Estadísticas del Fondo Monetario Internacional IMPACTO DEL LIBRE FWJJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA.. CUADRO 10 CUENTAS DEL SISTEMA MONETARIO 1988 970,l 3.665,5 1989 1990 1991 2.695,4 - 2.210,7 6.348,4 3.947,8 4.137,7 4.424,O 8.371,8 1992 1993 5.374,l 5.592,4 5.877,8 8.819,7 1994 5.664,6 1995 1996 1997* 3.965 4.264 8.229 502,26 633,05 4.402 8.238 12.640 6.236 6:988 13.225 11.251,9 14.412,l 759,28 863,06 7.078 7.741 14.819 12.669,1 18.333,7 844,84 6.480 6.705 10.219 14.996 16.699 21.701 6.772.9 17.256,2 23.979,l 912,83 7.365 18.904 26,269 7.981,7 9.069,4 21.834,9 23,480,4 29.916,6 32.549,8 1.036,69 1.102,40 7.699 8.227 21.062 21.299 28.761 29.526 * Corresponde hasta el mes de agosto Fuente: Estadísticas del Fondo Monetario Internacional 299,17 COLOMBIANO 382,57 IMPACTO DELLIBREFLUJO DECAPITALESEN ou3omomomomo ^ ^ ^ <D <o LT?e t 00000000000 LA INTERMEDIACIÓN BANCARIA... -?. c? @? $ 103 CUADRO ll CUENTAS DEL SISTEMA MONETARIO ECUATORIANO (98,7) CV) 995,8 1.099,7 897,l 1.089,8 301,60 526,30 (327) 3.302 2.974 (19) 2.089 2.071 276,9 374,s 455,l 1.228,2 2.217,O 3.262,6 1.505,1 2.591,5 3.717,7 767,80 1.046,20 1.534,oo 361 358 296 1.600 2.119 2.127 1.960 2.477 2.424 1993 1994 2.67,7 3.638,1 11.667,5 15.929,2 14.340,2 19.567,3 1.919,10 2.196,70 1.393 1.656 6.080 7.251 7.472 8.908 1995 3.705,5 12.41&6 16.124,1 2.564,50 1.445 4.843 6.287 1996 6.096,O 16.573,3 22.669,3 3.189,50 1.911 5.196 7.107 1997* 6.495,7 20.861,8 27.357,5 4.087,60 1.589 5.104 6.693 1988 1989 1990 1991 1992 * Corresponde hasta el mes de agosto Fuente: E!stadIsticasdel Fondo Monetario Internacional IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA... 105 Lourdes 4. Coll Calderón LA REFORMA DE LA SUPERVISIÓN BANCARIA En América Latina, al igual que en otros países emergentes, se ha seguido una orientación hacia una economía de mercado en la propiedad pública, gasto público, comercio internacional, seguridad social, así como en el sistema financiero; para esto último ha sido necesaria una reforma de la supervisión bancaria. La reforma de la supervisión bancaria está planteada, entonces, como t’na de las medidas medulares del proceso de cambio del sistema económico. Si se toma en cuenta la historia de alta volatilidad de las economías latinoamericanas en las últimas décadas, es posible llegar a considerar que el entorno económico no es lo suficientemente estable como para depender exclusivamente de las fuerzas de mercado, por lo menos no en las etapas iniciales del cambio estructural. Es posible, entonces, asumir que la reforma de la supervisión bancaria podría ser el factor atenuante de la volatilidad que genera el comportamiento de los agentes económicos internos, en el contexto de libre flujo de fondos en una economía de mercado. Más aun, si encontramos que los agentes económicos externos no son los que han jugado un papel fundamental en generar las crisis recientes en América Latina. Muy por el contrario, aun en contra del punto de vista popular que considera a los inversionistas extranjeros como los responsables de las crisis financieras, un examen detallado revela una realidad diferente: muchas de las crisis financieras han sido anticipadas y algunas veces iniciadas por la acciónde los residentes al cambiar sus activos en divisas y sacarlas del país. 4.1 Crisis bancarias, inversiones extranjeras y supervisión Un caso extremo del rol que los residentes jugaron en originar y desarrollar la crisis de la deuda latinoamericana de los años ochenta, es el caso de Argentina. Se ha estimado que todo el incremento de la deuda externa de ese país entre 1974 y 1982 fue usado, directamente o indirectamente, en financiar las salidas de capital privado. Algo similar sucedió en Venezuela y en menor medida en México. Sin embargo, éste no fue el caso de Brasil que entró en crisis sin estar plagado de los elementos característicos de una fuga de capitales. Más recientemente, el rol que han jugado los mexicanos en la crisis del peso en 1994-1995 ha sido destacado por el Fondo Monetario Internacional como un ejemplo de crisis que proviene de dentro. “La presión por el tipo de cambio mexicano no vino de la fuga de capitales de los 106 IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA.. inversionistas extranjeros o de posiciones especulativas hechas por éstos, sino por los residentes mexicanos” (IMF 1995, Capital Market Report). Un nuevo reporte del Fondo2n, sobre el rol de los fondos mutuos y de pensiones en la crisis asiática, sugiere que los bancos y compañías locales estuvieron en la vanguardia de aquéllos que vendieron la moneda local tailandesa, anticipándose a su devaluación en 1997. Los argumentos anteriormente presentados ponen de relieve la reciente discusión sobre la posible respuesta a la actual crisis en los controles de capital. Como se ha visto, no se ejerce el control monetario al poner trabas a la salida de los-capitales de los inversionistas internacionales (hedge funds) ya que son los propios agentes locales los que generan la falta de confianza en la moneda local. En el otro sentido, también se ha sobrevendido la receta de poner controles a la entrada de capitales. Y sin embargo, ésta parece ser la respuesta que ahora lidera el FMI, aun cuando va en contra de su posición tradicional de ser bastión del libre mercado. La poca efectividad de esta receta se pone de manifiesto considerando el caso de Chile, pues no ha podido disminuir su inestabilidad macroeconómica. Si las instituciones multilaterales consideraran a los controles de capital como la solución, pondrían en peligro el no centrarse en lo que es verdaderamente importante: “alcanzar la transparencia, fortificar la regulación bancaria, abolir la corrupción, reforzar las reformas económicas, y en muchos casos, perseguir una mayor reestructuración corporativa”2 t . 4.2 Reforma de la supervisión bancaria en el contexto peruano Conforme se ha señalado, el Perú, desde inicios de la década de los noventa, ha iniciado una serie de reformas estructurales con miras a modernizar su economía. En lo que va de este periodo, se han realizado reformas en distintos sectores importantes de la economía, reformas tales como: la reforma del sistema tributario, la privatización de las empresas públicas, como forma de reducir el aparato estatal, y la reforma laboral. Todas estas medidas sumadas a la reinserción de la economía peruana dentro del contexto inter20. Eichengreen 21. Edwards (1998). (1998). 107 Lourdes Coll Calderón nacional, tras casi siete años de moratoria en el pago de su deuda externa, se han orientado a lograr que el Perú se convierta en un país atractivo para el capital extranjero. Durante muchos años el sistema bancario peruano desempeñó un rol excesivamente tradicional, de manera tal que la operatividad de este sistema se reducía a captar recursos del público en forma de depósitos para su posterior colocación en forma de créditos. Las operaciones de alcance intemacional se encontraban limitadas para los bancos nacionales, en tanto que el resto del sistema financiero (financieras, cajas municipales, etc.) tenía una participación más reducida. A la difícil coyuntura de la banca peruana en los años ochenta, debe agregarse el fenómeno de la banca informal, el cual se explica por la hiperinflación que sufrió el país en la segunda parte de la década pasada, el creciente desempleo y la recesión. El fenómeno de la banca informal restringió aun más el rol de la banca como agente de intermediación financiera. A fin de lograr la modernización del sistema financiero nacional, conforme a los estándares internacionales, se han dictado hasta tres leyes generales, cuyos objetivos centrales pueden ser resumidos en: Modernización del sistema financiero. A fin de permitir que las empresas supervisadas puedan ser competitivas con las empresas extranjeras. Con tal finalidad, el concepto de banca tradicional se vio modificado con uno mas dinámico y complejo: la banca múltiple. Fortalecimiento del ente supervisor. A fin de que la Superintendencia de Banca y Seguros (SBS) se encuentre en condiciones de poder supervisar un sistema dinámico y moderno se le otorgó tanto un rol preventivo (visitas de inspección, verificación de la observancia de los límites operativos) como reparador (enfrentar las situaciones de crisis que se pudieran presentar). Adecuación de la legislación nacional a las recomendaciones intemacionales sobre supervisión bancaria, que se encuentran contenidas en el Concordato de Basilea*. I. 22. 108 El Concordato de Basilea es el resultado de una reunión realizado por los supervisores bancarios de los principales países del mundo. Sirve de modelo para el desarrollo ,de las legislaciones internas. IMPACTO - DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA.. Implementación de un sistema financiero altamente competitivo. En este sistema se establece, por ejemplo, la libertad de fijar las tasa de interés para las operaciones que realicen las empresas supervisadas con el público. En el sistema financiero es el propio mercado el que va a establecer la tasa de compra-venta. Las normas dictadas en los últimos años que recogen lo indicado en los párrafos anteriores son: - 4.3 Ley General de Instituciones, Bancarias, Financieras y de Seguros. Decreto Legislativo N” 637. Dictada en abril de 1991 Ley General de Instituciones Bancarias, Financieras y de Seguros. Decreto Legislativo No 770. Dictada en octubre de 1993. Ley General del Sistema Financiero y del Sistema de Seguros y Orgánica de las Superintendencia de Banca y Seguros. Ley N” 26702. Dictada en diciembre de 1996. Alcance del Acuerdo de Basilea Parte de las recomendaciones adoptadas han sido ya formuladas por los supervisores de los países industrializados. Estos entes internacionales formulan muchas recomendaciones, las cuales no son de obligatorio cumplimiento para las partes, ya que éstas deben adecuarse a sus realidades particulares. El objetivo principal para Bqilea es que las instituciones mantengan el capital adecuado, y éste se determina de tal forma que la institución financiera tenga el respaldo patrimonial que asegure que las pérdidas que pudiesen resultar de la actividad bancaria sean absorbidas por los accionistas y no por el público ahorrador. Para tener un capital adecuado, la relación Activos Ponderados por Riesgo VS. Capital debe por lo menos sobrepasar cierto nivel (8%). Al hacer estas consideraciones se piensa que el banco cubre sus riesgos dentro y fuera del balance23. Sin embargo, este nivel es adecuado siempre que el banco no esté expuesto a otros tipos de riesgo, tales como: 23. Como riesgos los derivados, fuera de balaace etc. se consideraa los cn5ditos contingentes, las cartas fianzas, Lourdes - Coll Calderón Riesgo de tasa de interés Riesgo operacional Riesgo de liquidez Riesgo de concentración de activos Riesgo de tipo de cambio Estos riesgos, que no se perciben en la lectura de los balances, deben incorporarse al Ratio de Basilea para mejorar el nivel de apalancamiento. Asimismo, existen otras debilidades en términos conceptuales y de aplicabilidad práctica que tienen que ser superadas24: - - La ponderación de riesgos. No aparece claro en el ratio cómo las diferentes clases de créditos son resumidas y la forma como éstas se hacen comparables. Por ejemplo, un crédito hipotecario es tratado en igual forma que uno comercial o de consumo; sólo se tiene en cuenta su capacidad de cumplimiento en el pago. La diversificación de carteras. Tal como se encuentra definido el cálculo de capital adecuado, éste ignora las correlaciones o covarianzas entre los activos o grupos de activos que permiten determinar si la disminución del riesgo, mediante una diversificacion eficiente, pues sólo considera la calidad del cliente. Todos los créditos tienen igual ponderación25. Dentro de cada clase de crédito no se toma en cuenta las diferencias en la calidad del riesgo crediticio entre los clientes, y por lo tanto, incrementa el riesgo del banco. Por ejemplo, los créditos dirigidos al sector comercial tienen igual peso si están dirigido a una empresa clase A (e.g: Saga Falabella) o, a una pyme (e.g.: de las que se encuentran en Gamarra). En el contexto peruano no se han tenido en cuenta estos “otros riesgos” ya mencionados, en parte debido a que las recomendaciones formuladas por los distintos entes supervisores de los principales países industrializados que 24. Saunders 25. Un posible argumento a favor de tener el mismo riesgo para todos los créditos de una categoría es que si el banco mantiene un portafolio de ckditos muy diversificado, el riesgo no sistemático de cada préstamo individual puede ser eliminado dejando sólo el riesgo sistemático. 110 (1992) IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA.. conforman el Concordado de Basilea emiten sus recomendaciones de acuerdo al contexto propio de cada país que lo conforma. Por otro lado, desde Basilea, se siguen emitiendo recomendaciones pues se está viviendo un proceso de cambio continuo, por lo que se espera que los riesgos mencionados se incorporen en un futuro próximo. 4.4 Reflexiones finales Las limitaciones de las normas internacionales de Basilea y la actual crisis llevan a plantear la necesidad de reexaminar la forma como está organizado el sistema financiero internacional. Plantearse dudas sobre el funcionamiento del sistema internacional, un año antes de la crisis, era como expresar dudas sobre algo que es bueno y sagrado: la santidad del mercado y su infalible habilidad para ajustarse, la libertad de comercio y de movilidad del capital. Se aducía que las dificultades subyacían no en el sistema mismo, sino en la forma como las políticas de mercado se aplicaban en las economías pequeñas. Sin embargo, el problema no es de ahora, ni de las economías pequeñas, el problema es sistémico y se cuestiona al capitalismo en sí mismo. El primer recurso utilizado en el pasado fue establecer mecanismos para lidiar con las crisis financieras, y para ello se constituyó el Fondo Monetario Internacional, el Tesoro y se han agrupado los G-7 y los G-20. Es prometedor ver que, por primera vez en muchas dtcadas, los líderes políticos de Europa y EE.UU. han comenzado a reconocer la necesidad de una reforma básica de la arquitectura financiera. La reforma requerida necesita ir más allá de contemplar al FMI como prestamista de último recurso, pues su rol va más allá. Llevando a mayores profundidades los planteamientos de reforma, es necesario preguntarse por la-reforma del sistema que rige los tipos de cambio de las monedas. Y como decía Keynes, para velar por la estabilidad del sistema, se debe regular las Balanzas de Pagos no sólo de los países deficitarios sino de los superavitarios también26. Estas consideraciones sobre el entorno económico internacional permiten adicionar que la supervisión bancaria, al defender fuertemente el ratio del capital adecuado, no sólo permitirá resolver las crisis del sistema financiero 26. Dell (1981). 111 Lourdes Coll Calderón mundial, sino también reducirá el riesgo (volatilidad) de aquellos países que tengan una supervisión exigente. De esta manera se concluye sobre la necesidad de su perfeccionamiento, a la luz de las experiencias negativas de otros países al querer controlar las crisis con barreras a la entrada y salida de capitales. 5. CONCLUSIONES La revisión de los hechos revela que ha habido un crecimiento de la intermediación que ha tendido a aliviar la restricción externa que contribufa a los bajos niveles de inversión en Perú y América Latina y favorecido el crecimiento económico. Las relaciones entre el crecimiento económico y la afluencia de capitales a América Latina en los años noventa están documentados en el informe presentado por Gavin, Hausmann y Leidermann (1996), donde revelan que existe una fuerte correlación entre el crecimiento y las entradas de capital, pero que esta relación no necesariamente es causal, ya que los inversionistas internacionales (hedge funds) tienden a buscar regiones en las que puede esperarse un rápido crecimiento económico. El estudio de los flujos financieros a través de los años en América Latina revela que estos flujos dependen tanto de las políticas como del entomo económico no relacionado con las políticas. El fin de la segunda Guerra Mundial, los acuerdos de Bretton Woods, la Revolución Cubana, la formación de la Alianza para el Progreso, el precio del petróleo, la formación de la OPEP, son todos acontecimientos políticos que, en igual o mayor forma, han contribuido notablemente a la formación de capitales y su direccionamiento. Entre los factores no políticos se encuentran el nivel de la tasa de interés en el mercado internacional, notándose que existe una fuerte correlación entre los flujos netos hacia la región y la tasa de interés mundial, por lo que los acontecimientos ocurridos al interior de América Latina tenían escasainfluencia. Asimismo, dentro de este contexto se debe tomar en cuenta los superávits en cuenta corriente de las principales economías industrializadas y la disponibilidad de capitales internacionales para evaluar sus decisiones retorno/ riesgo de sus inversiones en el área internacional. 112 IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA.. Sin embargo, el objeto de este documento ha sido destacar las implicancias que ha tenido el libre flujo de capitales en la intermediación bancaria en países latinoamericanos, especialmente el Perú. Para ello, se ha estudiado esta relación en el contexto peruano y en los países de su entorno. En primer lugar, se encuentra que existe una correlación directa entre el incremento del flujo internacional de capitales y el nivel de liquidez de cada país. Y de acuerdo al Enfoque Monetario de la Balanza de Pagos, este incremento de Liquidez puede incrementar tanto el nivel de Reservas Internacionales como el Crédito Interno, o ambos. En consecuencia, en esta cadena de relaciones causales, la relación entre Flujo de Capitales e Intermediación Bancaria es directa. Esta relación presenta, adicionalmente, una necesidad de comprensión de los efectos de los instrumentos de la política monetaria en las variables externas y el manejo monetario, crediticio y cambiario. Por tanto, este análisis permite observar los movimientos de los instrumentos de la política monetaria de cada, país para determinar el monto de estos flujos de capital neto que se dirige a la intermediación. Por ejemplo, del análisis empírico realizado en este documento, se puede concluir que fue Chile el país que llevó una política más conservadora en el periodo 19881997. Debido a que el índice de Crédito Internoniquidez se acercó a 0.5 durante todo el periodo, es decir, del total de la liquidez de su sistema sólo el 50% fue canalizado a la intermediación bancaria. Mientras que Perú, a inicios del mismo periodo de estudio, detentó un nivel de 1.2 en el índice, lo cual significa que toda la liquidez de su sistema más un 20% era canalizado hacia la intermediación, con lo cual la economía en su conjunto realizaba un desahorro; sin embargo, posteriormente, el Perú ha tendido hacia el nivel conservador de 0.5 en ese índice a partir de 1991. Mientras que el resto de países ha seguido trayectorias similares, aunque con tendencia a mejorar el ratio. Por ejemplo, Argentina que al inicio del periodo mantenía un ratio de 1.25, lo ha ido disminuyendo paulatinamente hasta llegar a 0.87 en 1997. El caso de Brasil es fluctuante, pues si bien al inicio del periodo sustentaba un ratio de 0.94, en 1993 alcanzó el 1.03, para luego disminuir a 0.81 en 1997. México si bien comenzó con un bajo ratio de 0.76. en 1994 y 1995 alcanzó 1.02, este ratio se justifica por la crisis suscitadaén ese país que llevó al retiro de los capitales externos. Colombia comienza el periodo con un nivel razonable de 0.52, subiendo paulatinamente hasta alcanzar el 0.72. Ecuador, si bien comienza con un nivel alto de 1.11, disminuye sólo hasta el nivel de 0.76 en 1997. Bolivia mantiene niveles casi estables entre 0.72 y 0.80 Estos países 113 Lourdes Coll Calderón han seguido políticas que han tendido más a la volatilidad. La preocupación es mayor en los casos de Brasil, México y Argentina por el tamaño de su economía y las repercusiones que ello ocasionaría al resto de la región. Junto a los instrumentos monetarios para controlar el crecimiento de la liquidez existe otro nivel de tratamiento. Este nivel está referido a las técnicas de supervisión bancaria, cuya reforma ha sido concebida como medida medular de todo proceso de reforma económica orientado hacia la economía de mercado, proceso que ha sido característico de los países emergentes durante los años noventa. Sin embargo, todavía existe el desafío fundamental que enfrentan los responsables de la formulación de políticas de mejorar el diseño de herramientas de supervisión para controlar el riesgo financiero. Fundamentalmente, el avance está basado en los acuerdos llegados en el Concordato de Basilea, en donde se propone una norma para la relación entre el capital de un banco y los activos ponderados en función del riesgo. Sin embargo, resulta difícil determinar si las normas de capital están logrando restringir el incremento de los activos con riesgos, debido a que el ratio de Basilea, tiene sus limitaciones. Dicho ratio no alcanza a incorporar otros riesgos. En el contexto peruano se encuentran íimitaciones similares, pues se han seguido las recomendaciones de Basilea, aun cuando los avances han sido considerables respecto a la década pasada. Las limitaciones del ratio de Basilea y la actual crisis llevan a plantear la necesidad de reexaminar la forma como está organizado el sistema financiero internacional. Estas consideraciones sobre el entorno económico internacional permite concluir que la supervisión bancaria, al defender fuertemente el ratio del capital adecuado, no sólo permitirá resolver las crisis del sistema financiero, sino también reducirá el riesgo (volatilidad) de aquellos países que tengan una supervisión exigente. REFERENCIAS BLEJER, 1983 114 Mario “Recent Economic Policies of the Southern Cone Countries and the Monetary Approach to the Balance of Payments”, IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA... en Economic Liberalization and Stabilization Policies in Argentina, Chile, and Uruguay: Applications of the Monetaty Approach to the Balance of Payments, pp. 3-20. Editado por Nicolas Ardetto Barletta, Mario Blejer y Luis Landau. Washington, D.C.: The World Bank. CEPAL, 1996 Estudio Económico de América Latina y el Caribe, 19951996. Santiago de Chile, Naciones Unidas. CIGUEÑAS, Benedicto “El Sistema Financiero y las Nuevas Disposiciones Legales”. En Moneda, pp. 56-60, N” 103. CORBO, Vittorio y HERNANDEZ, Leonardo “Macroeconomic Adjustment to Capital Inflows: Lessons from Recent Latin Ameritan and East Asian Experience”. The World Bank Economic Review, vol 11, N” 1, february 1996. CHACHOLIADES, Miltiades 1992 Economía Internacional, Santa’Fe de Bogotá: McGraw-Hill Interamerican. CHENERY, HOLLIS y STROUT Alan “Foreign Assistance and Economic Development”. American Economic Review, N” 56, setiembre 1966. DELL, Sidney 1981 En “On Being Grandmotherly: the Evolution of IMF Conditionality”. En Essays in International Finance. Princeton University, N” 144. DEVLIN, Robert, FFRENCH-DAVIS, Ricardo y GRIFFITH-JONES, Stephany “Flujos de Capital y el Desarrollo. en los Noventa: 1995 Implicaciones para las Políticas Económicas”. En Pensa- miento Iberoamericano, Revista de Economía Política: El Sistema Financiero, Globalización e Inestabilidad, pp. 4 l76, volumen 27, Enero-Junio. 115 Lourdes Coll Calderón EDWARDS, 1997 EDWARDS, 1998 Sebastían Comentarios a “Las raíces de las crisis bancarias: el contecto .maroeconómico” de Michael Gavin y Ricardo Hausmann. En Las crisis bancarias en Américu Latina. Editado por Ricardo Hausmann y Liliana Rojas-Suárez. Santiago de Chile: Fondo de Cultura económica. Sebastian “Barking Up the Wrong Tree”. En Financia1 Times, 8 de Octubre. Barry, MATHIENSON, Donald y otros “Hedge Funds and Financia1 Market Occasional Papers N” 166, Mayo. EICHENGREEN, 1998 ESTEiFANíA, 1998 GAVIN, 1996 Michael, HAUSMANN, Ricardo y LEIDERMAN, Ricardo “Los Aspectos Macroeconómicos del Flujo de Capitales en América Latina: Experiencia y Cuestiones Políticas”. En La Volatilidad de los Flujos de Capitales. Editado por Ricardo Hausmann y Liliana Rojas-Suárez. Washington, D.C.: Bank for International Development. 1995 GRIFFITH-JONES, 116 IMF Joaquín “Miedo al Pánico”. En El País, 26 de Octubre. GRIFFITH-JONES, 1984 Dynamics”. Stephany “Globalización de los Mercados Financieros y el Impacto de los Flujos hacia la Regulación”. En Pensamiento Iberoamericano, Revista de Economía Política: El Sistema Financiero, Globalización e Inestabilidad, pp, 41-76, volumen 27, Enero-Junio. Stephany Finanzas Internacionales y Latinoamérica: Pasado y Futuro. Montevideo: Asociación Latinoamericana de Instituciones Financieras de Desarrollo-ALIDE. IMPACTO DEL LIBRE FLUJO DE CAPITALES EN LA INTERMEDIACIÓN BANCARIA.. Alejandro “No hay Mercado para Tantos”. En Revista “Si”, pp. 4-5. Lima, 30 de Marzo. INDACOCHEA, 1998 INTERNATIONAL MONETARY 1957 The Monetary Approach to the Balance of Payments. Washington, D.C.: International Monetary Fund. KLINGERBERGER, Juan “Adecuación de Capital”. En Los Bancos, Junio. 1997 Ross LEVINE, 1977 “Financia1 Development and Economic Gr0wth:~View.s and Agenda”. Journal of Economic Literatut-e, vol XXXV, No 2, June. MARTHANS, 1998 Juan José “Si hay un Banco en Dificultades”. En Revista “Si”, pp. 2027. Lima, 30 Marzo. Felipe “Un Análisis Crítico de la Ley Genera1 del Sistema Financiero y del Sistema de Seguros”. En Moneda, pp. 52-55, N” 103. MORRIS GUERINONI, OLCESE, Jorge 1998 POLACK, “Hay dos Bancos que pueden quebrar”. En Revista “Si”, pp. 24-25. Lima, 13 de Abril. J. J. 1957 PREBISCH, 1959 FUND “Monetary Analysis of Income Formation”. Internutional Monetary Fund Staff Papers, Nov. R. “Commercial Policy in the Underdeveloped Countries”. En American Economic Review, vol. 49, N” 2. 117 Lourdes Coll Calderón PROUT, Christopher 1976 “Finance for Developing Countries: An Essay”. En fnternational Economic Relations of de Western World . Editado por A. Schonfield. Oxford University Press. ROJAS-SUAREZ, Liliana y WEISBROD, Steven “El Logro de la Estabilidad en los Mercados Financieros Latinoamericanos frente a la Volatilidad de los Flujos de Capital”. En Lu Volatilidad de los Flujos de Capital. Editado por Ricardo Hausmann y Liliana Rojas-Suárez. Washington, D.C.: Bank for International Development. 1996 ROZENWURCEL, G. 1997 “El Sistema Bancario Argentino en los Noventa: De la Profundización Financiera a La Crisis Sistémica”. En Desarrollo Económico, Revista de Ciencias Sociales, vol 37, No 146, Julio-Setiembre. SAUNDERS, 1992 SEIBER, 1977 SINKEY 1992 118 Anthony Financia1 Institutions: Irwing Editor. A Modern Perspectives, Sydney: M.J. Developing Countries in the Euro-currency Market: A debt Overhang Escalator: Departamento de Comercio Exterior de los EE.UU., Washington D.C. F., Joseph Commercial Bank, Financial, Management in the Financia1 Service Industry. New York: Macmillan.
https://openalex.org/W2473401852	https://ccsenet.org/journal/index.php/ibr/article/download/61280/33120	English	null	Inhibiting Factors of Inter-organizational Cost Management Complementary Study	International business research	2,016	cc-by	12,079	International Business Research; Vol. 9, No. 8; 2016 ISSN 1913-9004 E-ISSN 1913-9012 Published by Canadian Center of Science and Education International Business Research; Vol. 9, No. 8; 2016 ISSN 1913-9004 E-ISSN 1913-9012 Published by Canadian Center of Science and Education Inhibiting Factors of Interorganizational Cost Management Complementary Study Rafael Araújo Sousa Farias¹,², Valdirene Gasparetto² ¹Department of Administrative Sciences, University of Brasília, Brasília, Brazil ²Department of Accounting Sciences, Federal University of Santa Catarina, Florianópolis, Brazil Correspondence: Rafael Araújo Sousa Farias, Universidade de Brasília, Campus Universitário Darcy Ribeiro, Prédio da FACE, Programa de Pós-Graduação em Administração. CEP: 70910-900, Brasília-DF, Brazil. E-mail: farias-rafael@hotmail.com Accepted: June 8, 2016 Online Published: July 6, 2016 URL: http://dx.doi.org/10.5539/ibr.v9n8p91 Received: May 23, 2016 doi:10.5539/ibr.v9n8p91 Received: May 23, 2016 doi:10.5539/ibr.v9n8p91 Accepted: June 8, 2016 Abstract The research problem of this study is based on the discussion of inhibitors related Interorganizational Cost Management (IOCM). Taking on an inductive logic, the study´s objective is exploratory by way of a qualitative approach, with the overall goal of analyzing which factors inhibit the applicability of Interorganizational Cost Management. This study is a complement and completion of the debate initiated by Farias (2016). Fifty-four surveys retrieved from the literature were analyzed, which demonstrate the difficulties faced by companies in managing costs in a cooperative manner. Analysis on these studies could illustrate the perceptions held by different businesses, and list the difficulties faced by them, leading to the identification of 30 inhibiting factors. The diversity of the same highlights the interdisciplinary nature, as well as complexity, of the phenomenon in question. The study chose to divide the inhibiting factors into three groups, which relate to the developmental stages of Interorganizational relationships (planning, start of operations and maturation), with the inhibitors present in the three stages. Inhibitory factors related to People were found to be most predominant; the implementation of Interorganizational approaches requires not only changes in processes, but also in the adaptation of organizational behavior on part of those involved. Thus, the application of IOCM cannot be seen as a technical approach, guided by technology and management programs alone, and companies need to overcome internal barriers. Keywords: inhibitors, Interorganizational cost management, strategic cost management, supply chain 1 Introduction Keywords: inhibitors, Interorganizational cost management, strategic cost management, supply chain 1. Introduction The evolution of markets and the increasing complexity in supply chains led to the rise of new management techniques and new systems of information exchange among companies, growing beyond the internal environment and reaching Interorganizational relationships (Kulmala, Paranko, & Uusi-Rauva, 2002). Interorganizational networks have emerged as a widespread alternative for business needs to be met; moreover, Interorganizational networks can be a viable option for companies that cannot compete individually or have difficulties in operating in the face of the current economic complexity (Wincent, 2008). Businesses should take a stand and coordinate relationships that will help them address these challenges, such as globalization, as an opportunity (Drucker, 1997). Businesses can obtain a competitive advantage through Interorganizational Cost Management, the goal of which is to find solutions that have lower costs when compared to the sum of the costs of businesses acting individually (Kulmala et al., 2002). Cooper and Slagmulder (1999) explain that Interorganizational Cost Management is a structured approach to coordinating the activities of companies in a supply chain, aiming at reducing total network costs. Studies concerning Interorganizational relationships say that this approach is a tool for businesses to grow in the market in which they operate, generating benefits for all those involved (Ellram, 1994; Ellram & Siferd, 1998; Cooper & Slagmulder, 1999; Ferrin & Plank, 2002; LaLonde, 2003). Nevertheless, from an empirical point of view, there is a significant number of businesses that fail to participate in cooperative processes, and many networks are unable to consolidate their structures and management models (Sadowski & Duysters, 2008; Pereira, Alves, & Silva, 2010). Ongoing research on IOCM has generated approaches and concepts which explain the advantages of acting collaboratively with members of the supply chain to manage costs (Munday, 1992; LaLonde & Pohlen, 1996; Kulmala et al., 2002; Cooper & Slagmulder, 1999, 2003a, 2003b) and explain the implementation steps of the approach (Ellram, 2002a; Cooper & Slagmulder, 2003b). However, few studies discuss the difficulties and potential obstacles which may 91 International Business Research Vol. 9, No. 8; 2016 www.ccsenet.org/ibr occur during the implementation process (Hakansson & Snehota, 1995; Park & Ungson, 2001; Bastl, Grubic, Templar, Harrison, & Fan, 2010; Pereira et al., 2010). occur during the implementation process (Hakansson & Snehota, 1995; Park & Ungson, 2001; Bastl, Grubic, Templar, Harrison, & Fan, 2010; Pereira et al., 2010). 1. Introduction In light of this, the inter-relationship is possible and IOCM is presented as an opportunity, but its application can be a difficult task. It is observed that, given the advantages it affords, it would be natural that most companies should seek the interrelationship; paradoxically, the opposite occurs (Kulmala et al., 2002). This situation is viewed as a research opportunity, thus, the question posed here is: what are the factors that inhibit the applicability of the Interorganizational Cost Management approach? To answer the research question, this study determined the overall goal to analyze which factors inhibit the applicability of the Interorganizational Cost Management. Managers are facing new challenges in the pursuit of competitive advantages outside their organizations. However, they have little guidance on the potential challenges related to the implementation of approaches to Interorganizational Cost Management (Dekker, 2004; Vosselman & Van Der Meer-Kooistra, 2006; Bastl et al., 2010). Studies that help businesses successfully implement the Interorganizational program costs can generate benefits while increasing their knowledge of business processes (Stapleton, Pati, Beach, & Julmanichoti, 2004). Along these lines, a diagnosis of the reasons why the Interorganizational relationships fail could contribute during all stages of building relationships, highlighting the importance of the initial phase. This analysis can help managers create statutes, formalize contracts and select chain members (especially suppliers). As such, finding out the reasons why businesses leave Interorganizational relationships can mean an improvement in their management process (Pereira et al., 2010). 2. Theoretical Framework 2.1 Interorganizational Relationships 2.1 Interorganizational Relationships Cooperative Interorganizational relationships are characterized by the interaction between two or more organizations, both of similar and different natures (Oliver, 1990). Businesses do not operate in isolation, but rather in contact with suppliers, customers, business partners, lenders and government agencies (Coad & Scapens, 2006), thus, cooperation between companies is an alternative so they can increase competitiveness before the market (Ebers, 1997). When expanding the traditional and isolated concepts of businesses, Interorganizational relationships can be configured in different ways (Zylbersztajn & Neves, 2000). Cândido (2002) emphasizes the importance of organizations to act together, in order to adapt to the current competitive environment. Companies seek competitive advantages including the possibility of combination of competences, use of know-how from other companies, shared burden by carrying out technological research, sharing of risks and costs of exploiting new opportunities, resource-sharing and strengthening buying power, among others (Amato Neto, 2000). 2.2 Interorganizational Cost Management 2.2 Interorganizational Cost Management The Interorganizational context requires that the cost management of a business not be restricted by its internal environment, as businesses are links within a chain, thus, it is influenced by external factors and, at the same time, it, too, bears an influence on chain costs. The production of goods is the result of several interconnected steps that affect one other, for the supply chain is not a sequence of independent activities, but a system of interdependent activities (Porter, 1989). Traditional practices and management techniques do not often play the role of managing Interorganizational relationships (Bastl et al., 2010). The generation of administrative and financial information should assist managers to make decisions in that scenario (Seal, Berry, & Cullen, 2004; Tomkins, 2001). This requires companies to manage not only their internal operations but, in unison, the operations of other chain companies (Cooper & Slagmulder, 1999; Kulmala et al., 2002; Kajüter & Kulmala, 2005; Coad & Cullen, 2006; Souza & Rocha, 2009). From this context, the Interorganizational Cost Management emerges in the area of strategic management of costs. Interorganizational Cost Management is a structured approach to the coordination of activities of companies in a supply chain, so that its total costs are reduced (Cooper & Slagmulder, 1999). Souza and Rocha (2009, p. 25) state that "Interorganizational Cost Management is a cooperative process of cost management comprising other organizations from a value chain besides the company itself". IOCM aims to reduce costs related to the distribution, logistics and manufacturing of products (Cooper & Slagmulder, 1999). The different methods for cost reduction share a focus on the coordination of businesses in the supply chain and in cost management policies across organizational boundaries, aiming at an "intelligent" cost management, so that the cooperative relations of the chain are more efficient before competitor chains (Christopher, 2007). 2.3 Inhibiting Factors to Interorganizational Cost Management 3. Methodological Aspects The methodological procedures described by Farias (2016) were performed First, the field of exploration was delimited, and the databases selected after these steps were followed: (i) the databases which appeared in the Portal de Periódicos Capes (501 databases) were selected; (ii) in the field “Knowledge Areas”, the “Social Sciences” option was specified, and the sub-area “Business Administration-Public Administration-Accounting” was chosen, which resulted in 73 databases; (iii) the databases were selected to provide “Full Text” (24 databases); and (iv) a selection was made of the databases that allowed searches on “All text fields” and the use of at least two axes by using Boolean expressions. The process resulted in 13 databases: ANNUAL REVIEWS; CAMBRIDGE JOURNALS ONLINE; EMERALD INSIGHT; JSTOR; OECD LIBRARY; SAGE JOURNALS ONLINE; APA PSYCARTICLES; WILEY ONLINE LIBRARY; PROQUEST; WEB OF SCIENCE; SCIENCEDIRECT; SCOPUS; and EBSCO. the use of at least two axes by using Boolean expressions. The process resulted in 13 databases: ANNUAL REVIEWS; CAMBRIDGE JOURNALS ONLINE; EMERALD INSIGHT; JSTOR; OECD LIBRARY; SAGE JOURNALS ONLINE; APA PSYCARTICLES; WILEY ONLINE LIBRARY; PROQUEST; WEB OF SCIENCE; SCIENCEDIRECT; SCOPUS; and EBSCO. In the selected databases, searches were conducted with terms in English, using two axes. Axis 1 contained words whose meaning is similar to “inhibitor”, i.e., those words that refer to the idea of “barrier”, something that gets in the way or hinders. Axis 2 contained terms related to “Interorganizational cost management”. Searches were formulated with the following keywords combined with the axes: Axis 1 – “inhibit” OR “barrier” OR “problem” OR “disturb” OR “difficult”. Axis 2 – “Interorganizational cost” OR “Interorganizational cost” OR “interorganisational cost” OR “inter-organisational cost”. Searches were formulated with the following keywords combined with the axes: Axis 1 – “inhibit” OR “barrier” OR “problem” OR “disturb” OR “difficult”. Axis 2 – “Interorganizational cost” OR “Interorganizational cost” OR “interorganisational cost” OR “inter-organisational cost”. The searches were conducted in January 2015, selecting the option “all text fields” using the Boolean operator “AND” between the axes, the Boolean operator “OR” between terms, without delimitation of time period. The searches were conducted in January 2015, selecting the option “all text fields” using the Boolean operator “AND” between the axes, the Boolean operator “OR” between terms, without delimitation of time period. The search was conducted in the 13 selected databases; however, there were search results in only seven databases (at least one article). 2.3 Inhibiting Factors to Interorganizational Cost Management Companies stand to receive many benefits by coordinating efforts and acting collaboratively with members of the Companies stand to receive many benefits by coordinating efforts and acting collaborative 92 International Business Research www.ccsenet.org/ibr Vol. 9, No. 8; 2016 supply chain and there are many potential rewards offered justifying the creation of Interorganizational networks (Cooper & Slagmulder 1999). However, many problems can occur during this process, which can make it difficult or impossible to achieve the goals aspired by companies. supply chain and there are many potential rewards offered justifying the creation of Interorganizational networks (Cooper & Slagmulder 1999). However, many problems can occur during this process, which can make it difficult or impossible to achieve the goals aspired by companies. Failures in Interorganizational relationships are recurrent and many companies are far too optimistic about the prospect of benefits when they begin in a coordination of efforts with others (Hakansson & Snehota, 1995). This coordination can cause conflicts – common in relationships of this type –, which pose difficulties to obtaining positive results. Bastl et al. (2010), aimed to point out the factors that inhibit the implementation of Interorganizational cost management approaches, among other goals. According to the authors, their study was the first to systematically address the inhibiting factors in the implementation of cost management approaches within the Interorganizational scope systematically. The study revealed the complexity of inhibiting factors involved in the implementation of approaches, in addition to demonstrating the interdisciplinarity of the topic. Farias (2016) aimed to identify, in the literature, the factors and situations that inhibit the applicability of the Interorganizational Cost Management approach. In a general way, the study analyzed 35 academic studies that show the difficulties faced by companies to manage costs in a cooperative manner. The author identified 25 inhibiting factors and classified them into seven distinct groups for the purposes of analysis. 4.1 Group 1 – Formation of Collaborative Relationships and Strategies of Companies Group 1 is formed by inhibiting factors concerning the formation of collaborative relationships and strategies of companies: it refers to the moment when companies identify opportunities and advantages that can get when they act collaboratively with suppliers and customers. At this stage, companies analyze their profiles and strategies and map opportunities and processes that can be put into practice (Miles & Snow, 1992; Rogers, 1995; Stephenson, 2003; Shipilov, Rowley, & Aharonson, 2006). Table 1 shows the inhibitory factors for the Group 1. y p Table 1. Inhibitory factors related the formation of collaborative relationships and strategies of companies Inhibitors Authors 1 Lack of resources to execute the project Glad and Becker (1995); LaLonde and Pohlen (1996); Nicolini, Tomkins, Holti, Oldman, and Smalley (2000); Stapleton et al. (2004); Kajüter and Kulmala (2005) 2 Differences between companies’ strategic plans Hitt, Dancin, Levitas, Arregle, and Borza (2000); Edelman, Bresnan, Newell, Scarborough, and Swan (2004); Ojala and Hallikas (2007) 3 Different expectations of behavior between companies Hitt et al. (2000); Kajüter and Kulmala (2005) 4 Lack of long-term management goals Messner and Meyer-Stamer (2000); Ferrin and Plank (2002); Edelman et al. (2004); Möller, Windolph, and Isbruch (2011) 5 Dependence of mutual learning Ellram (2002b); Gareth (2005); Thomson and Gurowka (2005) 6 Lack of recognition that costing systems are necessary Kulmala et al. (2002); Gupta and Gunasekaran (2004) 7 Lack of interest on the part of the partner companies in sharing information Munday (1992); LaLonde and Pohlen (1996); Norek and Pohlen (2001); Kulmala (2002); Kulmala et al. (2002); Ellram (2002a); Cooper and Slagmulder (2003a); Kulmala (2004); Agndal and Nilsson (2008); Windolph and Moeller (2012) 8 Lack of understanding of costs along the chain LaLonde and Pohlen (1996); Nicolini et al. (2000) 9 Lack of perception of the value resulting from the relationship Cooper and Slagmulder (1999); Cannon and Homburg (2001) 10 Difficulty in defining mutual benefits Dekker (2003); Kajüter and Kulmala (2005) Source: Survey data (2016). Analysis of the cost behavior in the chain, project execution and construction of relationships with the members of Table 1. Inhibitory factors related the formation of collaborative relationships and strategies of companies Inhibitors Authors 1 Lack of resources to execute the project Glad and Becker (1995); LaLonde and Pohlen (1996); Nicolini, Tomkins, Holti, Oldman, and Smalley (2000); Stapleton et al. 3. Methodological Aspects The databases that provided the articles were EMERALD INSIGHT; WILEY ONLINE LIBRARY; PROQUEST; WEB OF SCIENCE; SCIENCEDIRECT; SCOPUS; and EBSCO. The search yielded a total of 418 results. The search was conducted in the 13 selected databases; however, there were search results in only seven databases (at least one article). The databases that provided the articles were EMERALD INSIGHT; WILEY ONLINE LIBRARY; PROQUEST; WEB OF SCIENCE; SCIENCEDIRECT; SCOPUS; and EBSCO. The search yielded a total of 418 results. In order to ensure the relevance of the search results, the content was filtered through the following steps: (i) the 418 articles arising from the search were downloaded; however, only 225 articles were fully available and free; (ii) the 225 studies were imported into a bibliographic management software, Mendeley® , which was used to delete 134 repeated or misaligned studies, thus yielding 91 works; (iii) the titles and abstracts were read, and the articles that were not aligned with the search criteria were excluded. In order to be included, a study should: a) address the issue of cost management in the Interorganizational context; b) contribute to the debate about inhibitors and barriers to Interorganizational Cost Management. After filtering, 35 studies were selected to compose the bibliographic portfolio to be analyzed. Then, the articles were fully read. The reading of each article identified the factors that hampered/hindered Interorganizational Cost Management in each environment, enabling the creation of a report with all the identified factors. Twenty-five factors inhibiting IOMC practices were identified. Following the phases carried out by Farias (2016), a new search started from the analysis of the references of the works comprising the bibliographic portfolio, with the aim of filling the gaps that might exist after the previous phases. Nineteen more studies were then selected. In total, 54 studies were read and 30 inhibiting factors were identified. 93 International Business Research Vol. 9, No. 3. Methodological Aspects 8; 2016 www.ccsenet.org/ibr In regards to the limitations of the research methods, specifically to the process carried out to identify inhibitors, the following items merit attention: a) the articles were read by a single researcher, which may have caused bias; b) the analyzed literature contains academic papers from a determined time frame (1992 - 2012); over the years, new circumstances may unfold and new inhibiting factors of the IOCM practice may be identified; and c) the use of works from different sources and in different contexts may have allowed the identification of excessively generic and comprehensive factors, without being specific to a certain branch or economic sector. The possibilities of inhibiting factors found on the IOCM are not limited to those identified in this study. Therefore, the identified inhibiting factors and conclusions and debates on the subject may vary depending on the approach used, the methodology applied, the sector analyzed, the type of Interorganizational relationships, among others. 4. Demonstration and Analysis of Results Bastl et al. (2010) suggest that the inhibitory factors that affect the cost management approaches among companies should be analyzed in three dimensions, namely: People, Process and Technology. Farias (2016) elected to group the inhibiting factors into seven groups: Corporate Strategy; Integration Between Companies; People; Intra- and Interorganizational Processes; Corporate Training and Education; Conflicts Between Companies; and Lack of Trust Between Companies. Despite the noteworthy, pioneering research emerging in regards to the subject, this study opted to classify the inhibiting factors into three groups distinct from those used by Bastl et al. (2010) and Farias (2016). For the purposes of the present study, the 30 inhibiting factors identified were divided into three groups which refer to the developmental stages that Interorganizational relationships go through (planning, start of operations and maturation) (Ring & Van de Ven, 1994; Bryson, Crosby, & Stone, 2006; Chen, 2010). Group 1 is formed of inhibiting factors concerning the formation of collaborative relationships and strategies of companies. Group 2 is formed by inhibiting factors concerning the development of processes and construction of relationships. Group 3 is formed by inhibiting factors related to the results, reviews and adjustments that occur in relationships. 4.1 Group 1 – Formation of Collaborative Relationships and Strategies of Companies 4.1 Group 1 – Formation of Collaborative Relationships and Strategies of Companies Besides the initial costs, upgrading accounting systems and maintaining relationships can require significant effort and resources (Kajüter & Kulmala, 2005). The importance of analyzing not only the financial condition of the company, but also the other members of the chain is emphasized (Kajüter & Kulmala, 2005). One of the barriers to the success of coordinated actions are the differences between companies’ strategic plans (Hitt et al., 2000). An understanding of the goals of partners and a compatibility with the same can improve the performance of the chain, otherwise, differences may result in conflicting relationships (Hitt et al., 2000). Before companies formalize any agreement, they may use standards and criteria for the selection of partners, analyzing the compatibility of goals and expectations (Hitt et al., 2000; Cooper & Slagmulder, 1999). This situation can be a barrier to Interorganizational Cost Management, so that the exchange of information and coordinated actions may be damaged (Edelman et al., 2004). Kajüter and Kulmala (2005) emphasize the importance of questioning how each party wishes to behave prior to entering into a partnership. Different expectations of behavior between companies, that is, the difference between the behavior that managers believe they would have in a possible joint action, and the expectation that they have in relation to their potential partners, could be a significant limiting factor in the implementation of IOCM (Kajüter & Kulmala, 2005). Companies need to signal, in a credible manner, their intentions to continue the long-term relationship when seeking to implement the IOCM (Messner & Meyer-Stamer, 2000; Edelman et al., 2004; Möller et al., 2011). It is understood that a lack of long-term management goals could be a barrier to the approach. Being aware of the effort that exists to build the relationships and the existing delay in building trust between companies, co-coordinated management will have better results if there are long-term prospects (Ferrin & Plank, 2002), since a commitment to long-term cost savings develops confidence between the members of the supply chain (Kajüter & Kulmala, 2005). Companies use the strategy of cooperation networks in order to increase their benefits however, sometimes they ignore that relations are not always synonymous with good results (Gareth, 2005). One factor that inhibits Interorganizational relationship in the long run is the fact that the partners, as they learn what others know, tend to break relations (Gareth, 2005). 4.1 Group 1 – Formation of Collaborative Relationships and Strategies of Companies The partner's know-how learning can make that relationship unnecessary, leading companies leave the cooperative relationship. Thus, the dependence of mutual learning is a barrier to IOCM. The company and its partners need to be aware of the importance of costing systems and the information generated by them. Lack of recognition that costing systems are necessary is a barrier to IOCM (Kulmala et al., 2002; Gupta & Gunasekaran, 2004). If the company has no information on the costs available for internal use, it is impossible to share with other companies. The willingness to share information is not enough, thus, the company should assess its capacity to produce such information, since it is required a lot of work to develop costing systems (Kulmala et al., 2002). Cooper and Slagmulder (2003a) and Kulmala (2004) explain that the relationship between buyers and suppliers must be based on confidence and great information sharing, however, companies might not be willing to share information in light of uncertainty. According to the authors, there are two measures that help reduce uncertainty, namely: the increased amount of information shared – in both directions – and the reduction of transaction processing time. Lack of interest on the part of the partner companies in sharing information can be a barrier to Interorganizational relationship (LaLonde & Pohlen, 1996; Norek & Pohlen, 2001; Kulmala, 2002; Ellram, 2002a). Suppliers may reject the idea of sharing internal information due to the fear that buyers will use the data to increase the pressure on their profit margins, in an opportunistic fashion (Munday, 1992; Windolph & Moeller, 2012). Even if the initiative to share information occurs by the buyer, the transparency of costs can bring benefits to suppliers (Agndal & Nilsson, 2008). In this way, suppliers should be aware that information sharing, in order to conduct a strategic cost management, can bring benefits to the whole chain. LaLonde and Pohlen (1996) and Nicolini et al. (2000) state that one of the barriers to IOCM is lack of understanding of costs along the chain by companies. Generally, companies initially develop projects and then contact the prices of suppliers who were not involved in the development of the project design. Companies analyze the continuation, or cancellation, of relationships through the capacity of these to generate demonstrable value to the participants (Cannon & Homburg, 2001). 1 In the text, the inhibiting factors are highlighted in bold. 4.1 Group 1 – Formation of Collaborative Relationships and Strategies of Companies (2004); Kajüter and Kulmala (2005) 2 Differences between companies’ strategic plans Hitt, Dancin, Levitas, Arregle, and Borza (2000); Edelman, Bresnan, Newell, Scarborough, and Swan (2004); Ojala and Hallikas (2007) 3 Different expectations of behavior between companies Hitt et al. (2000); Kajüter and Kulmala (2005) 4 Lack of long-term management goals Messner and Meyer-Stamer (2000); Ferrin and Plank (2002); Edelman et al. (2004); Möller, Windolph, and Isbruch (2011) 5 Dependence of mutual learning Ellram (2002b); Gareth (2005); Thomson and Gurowka (2005) 6 Lack of recognition that costing systems are necessary Kulmala et al. (2002); Gupta and Gunasekaran (2004) 7 Lack of interest on the part of the partner companies in sharing information Munday (1992); LaLonde and Pohlen (1996); Norek and Pohlen (2001); Kulmala (2002); Kulmala et al. (2002); Ellram (2002a); Cooper and Slagmulder (2003a); Kulmala (2004); Agndal and Nilsson (2008); Windolph and Moeller (2012) 8 Lack of understanding of costs along the chain LaLonde and Pohlen (1996); Nicolini et al. (2000) 9 Lack of perception of the value resulting from the relationship Cooper and Slagmulder (1999); Cannon and Homburg (2001) 10 Difficulty in defining mutual benefits Dekker (2003); Kajüter and Kulmala (2005) Source: Survey data (2016). Analysis of the cost behavior in the chain, project execution and construction of relationships with the members o e 1. Inhibitory factors related the formation of collaborative relationships and strategies of companies Analysis of the cost behavior in the chain, project execution and construction of relationships with the members of the 94 International Business Research Vol. 9, No. 8; 2016 www.ccsenet.org/ibr chain results in the consumption of financial resources that include, for example, labor, facilities, utilities, materials, among others, so, the lack of resources to execute the project1 could be a barrier to IOCM (LaLonde & Pohlen, 1996; Stapleton et al., 2004). Besides the initial costs, upgrading accounting systems and maintaining relationships can require significant effort and resources (Kajüter & Kulmala, 2005). The importance of analyzing not only the financial condition of the company, but also the other members of the chain is emphasized (Kajüter & Kulmala, 2005). chain results in the consumption of financial resources that include, for example, labor, facilities, utilities, materials, among others, so, the lack of resources to execute the project1 could be a barrier to IOCM (LaLonde & Pohlen, 1996; Stapleton et al., 2004). 4.1 Group 1 – Formation of Collaborative Relationships and Strategies of Companies In fact, the reduction of the total value of the product offered to the final consumer is an indicator of the efficiency of Interorganizational relationships, but not the only one. The lack of perception of the value resulting from the relationship between companies is mentioned by Cannon and Homburg (2001) as a factor that may jeopardize the continuity of long-term relationships. The 95 International Business Research www.ccsenet.org/ibr Vol. 9, No. 8; 2016 measurement of the value creation, or generated intangibles, by the relationships between the companies is a difficult task, and it is their duty to appeal to a variety of perspectives to assess the value creation of relationships (Cannon & Homburg, 2001). measurement of the value creation, or generated intangibles, by the relationships between the companies is a difficult task, and it is their duty to appeal to a variety of perspectives to assess the value creation of relationships (Cannon & Homburg, 2001). Dekker (2003), Kulmala et al. (2002) and Kajüter and Kulmala (2005) identified issues related to the definition of mutual benefits as inhibitors of the practice of IOCM. They report that relationships and partnerships tend to break when they find difficulty in defining mutual benefits to the parties involved. 4.2 Group 2 – Development of Processes and Construction of Relationships 4.2 Group 2 – Development of Processes and Construction of Relationships Group 2 is formed by inhibiting factors related to the development of processes and construction of relationships: the second stage concerns the relationships and processes carried out in order to achieve better results. At this stage, companies increase information sharing, resources and coordinate activities (Oliver, 1990). That is, the moment when planning performed in the early stage is put into practice (Nassimbeni, 1998; Olson, 1999; Prahalad & Ramaswamy, 2004). Table 2 shows the inhibiting factors for the Group 2. Table 2 shows the inhibiting factors for the Group 2. g p Table 2. Inhibiting factors related to the development of processes and construction of relationships Inhibitors Authors 11 Lack of trust among the partners in IOCM Barney and Hesterly (1996); Dekker (2004) 12 Opportunism Cooper and Yoshikawa (1994); Ellram and Siferd (1998); Messner and Meyer-Stamer (2000); Nicolini et al. (2000); Axelsson, Laage-Hellman, and Nilsson (2002); Ellram (2002b); Seal et al. (2004); Edelman et al. 4.1 Group 1 – Formation of Collaborative Relationships and Strategies of Companies (2004); Kajüter and Kulmala (2005) 13 Lack of incentives to innovate and prosper Gareth (2005); Pereira et al., 2010 14 Resistance to change Ellram (1994); Ellram and Siferd (1998); Cokins (1998); Axelsson et al. (2002); Lin, Collins, and Su (2001) 15 Lack of management support from members of organizations Ellram (2002a); Seal et al. (2004); Ellram (2006); Himme (2012) 16 Difficulties in integrating information between companies Cokins (1998); Cullen, Berry, Seal, Dunlop, Ahmend, and Marson (1999); Ellram (2002a, 2002b); LaLonde (2003); Ramos (2004); Mouritsen and Thrane (2006) 17 Unreliability of internal cost data Ellram and Siferd (1998); Milligan (1999); Ellram (2002b) 18 Inadequate level of data sharing Munday (1992); Cooper and Slagmulder (1999) 19 Poor internal availability of data Lin et al. (2001); Kulmala et al. (2002); Cokins (2003); LaLonde (2003); Kajüter and Kulmala (2005) 20 Lack of expertise and education programs geared towards employees LaLonde and Pohlen (1996); Cokins (1998); Cullen et al. (1999); Nicolini et al. (2000); Kulmala et al. (2002); Ellram (2002b); Ramos (2004); Seal et al. (2004); Thomson and Gurowka (2005); Himme (2012) Source: Survey data (2016). e 2. Inhibiting factors related to the development of processes and construction of relationships Source: Survey data (2016). Lack of trust among the partners in IOCM is seen as an inhibiting factor to the practice (Barney & Hesterly, 1996; Cooper & Slagmulder, 1999; Dekker, 2004). Cooper and Slagmulder (1999) explain that trust is the foundation of IOCM and that, once the activities begin, it should receive the necessary adjustments to its long-term maintenance. Dekker (2004) explains that even for companies that have a willingness to share information, confidence tends to be weakened when there are risks in transactions. The larger the uncertainty in agreements and partnerships, the more control mechanisms will be needed, such as the use of contracts, which increases the complexity of the relationships (Barney & Hesterly, 1996). Messner and Meyer-Stamer (2000) explain that opportunism is a recurrent problem in relationships. The authors consider that cooperative relationships must be based on trust among members, but even loyal members can sometimes be deceived by the partners during negotiations. The opportunistic characteristic in an Interorganizational relationship appears when one of the companies operates exclusively in pursuit of their own interests and violates rules established between the parties (Williamson, 1985). 4.1 Group 1 – Formation of Collaborative Relationships and Strategies of Companies Opportunism is the breaking of agreements, standards and principles governing an Interorganizational relationship (Edelman et al., 2004), and the pursuit of exclusively individual interests when the company has cooperation agreements with others, it is unacceptable (Cooper & Yoshikawa, 1994; Ellram & Siferd, 1998; Dekker, 2003, Kajüter & Kulmala, 2005). Another barrier to the continuity of Interorganizational relationships is the lack of incentives to innovate and prosper (Gareth, 2005). A partnership between companies can insulate them from external pressure and can make them less efficient, that is, the lack of external pressures could cause the chain links to weaken. For example, according to Pereira et al., 2010, companies operating in cartel are protected from the pressures of competition, thus the excessive protection 96 International Business Research Vol. 9, No. 8; 2016 www.ccsenet.org/ibr can cause them to fail to innovate and prosper, rendering them less efficient. can cause them to fail to innovate and prosper, rendering them less efficient. Implemented models of Interorganizational Cost Management create new scenarios, requiring new behaviors by the members of the organizations, so whenever an organization tries to change its operations, the nature and extent need to be discussed with employees (Ellram & Siferd, 1998). The change in the organizational behavior refers not only to the matter of designing management systems, but rather be a systematic effort (Cokins, 1998; Axelsson et al., 2002). The flexibility to change is an important factor for the successful implementation of management programs, thus, resistance to change by members of organizations may be a barrier to the success of IOCM (Ellram, 1994; Ellram & Siferd, 1998; Cokins, 1998; Axelsson et al., 2002; Lin et al., 2001). Employees need to be encouraged to participate in all levels of the cost reduction process (Himme, 2012), since the lack of management support from members of organizations impairs the cost management (Seal et al., 2004). Companies should evaluate if managers are supported by the company's members in seeking ways to build interest in reducing costs. Another factor that inhibits the success of IOCM are the difficulties in integrating information between companies (Cokins, 1998; LaLonde, 2003; Ramos, 2004). Ramos (2004) explains that the process of cooperation between businesses, new characteristics arise, such as increased complexity and the need to work in and outside of organizational boundaries; so, to work collaboratively requires the use of external data to the company, this requires the integration of such information. 4.1 Group 1 – Formation of Collaborative Relationships and Strategies of Companies It is important that the company’s buying team be able to communicate with the internal team of suppliers, because without the support of specialists from the suppliers, the analysis may be too complex and time-consuming (Cullen et al., 1999; Ellram, 2002a, 2002b). It is expected that companies have the ability to adapt their technologies and integrate the external environment to themselves to enable the exchange of information and the application of Interorganizational Cost Management. Nevertheless, taking decisions based on information transferred by other companies can be risky, as organizations can share incorrect data "without knowing" even believing that their costing systems are efficient and provide reliable information (Milligan, 1999; Ellram, 2002b; Cokins 2003). Milligan (1999) and Ellram (2002b) emphasize the importance of credibility of the numbers reported to the success of strategic cost management. The authors explain that the credibility of the numbers happens when all key members of the chain agree on how the figures are calculated, that is, the criteria applied. The numbers must be determined and calculated by a reliable source in order to avoid mistakes and manipulation of data (Ellram & Siferd, 1998; Ellram, 2002b). Unreliability of internal cost data can be seen as a barrier to the success of IOCM, since management decisions are based on these (Ellram & Siferd, 1998; Milligan 1999; Ellram, 2002b). Munday (1992) indicates that when the shared data is used in a constructive manner, not just to pressure the margins of suppliers, the occurrence of efficient cooperation is possible. In fact, cost data, previously used only for internal report purposes, can now provide information to external partners. However, there is often inadequate level of data sharing between the companies, which can be a barrier to IOCM (Munday, 1992). In addition, poor internal availability of data is also pointed out as an obstacle to the approach (Kulmala et al., 2002; Cokins 2003; LaLonde, 2003; Kajüter & Kulmala, 2005). Lin et al. (2001) explain that a perfect costing system would trace all costs and activities of companies. Hundreds, even thousands, of transactions take place in a company every day, and there are activities that can only be identified and measured with significant effort of the company, which can lead to a benefit not worth its own costs. Group 3 – Results, Reviews and Adjustments that Occur in Relationships 4.3 Group 3 – Results, Reviews and Adjustments that Occur in Relationships Group 3 is formed by inhibiting factors related to the results, reviews and adjustments that occur in relationship. The third stage refers to the moment when companies receive the initial results and then have the capacity to evaluate if the relationships brought benefits, and if the planning and initial goals have been achieved, so they can make the necessary adjustments to keep the relationships, if necessary (Miles & Snow, 1992; Stephenson, 2003). Table 3 shows the inhibiting factors for Group 3. Table 3. Inhibitory factors related to the results, reviews and adjustments that occur in relationships Inhibitors Authors 21 Imbalance between short-term and long-term gains Park and Ungson (2001); Edelman et al. (2004); Pereira et al. (2010) 22 Uneven distribution of benefits Cooper and Yoshikawa (1994); Cooper and Slagmulder (1999); Kulmala (2004); Kajüter and Kulmala (2005); Christopher (2007) 23 Poorly prepared IOCM model design Waeytens and Bruggeman (1994); Cokins (1998, 2003); Gareth (2005) 24 Highly complex IOCM systems Cokins (1998); Messner, Meyer-Stamer (2000); Kaplan and Anderson (2004) 25 Inflexibility in IOCM Park and Ungson (2001); Dubois (2003); Gareth (2005) 26 Inability to determine market prices Cooper and Slagmulder (1999); Nicolini et al. (2000) 27 Loss of customer focus Cokins (1998); Lin et al. (2001); Himme (2012) 28 Unfulfilled initial goals Ebers (1997); Wegner and Padula (2010); Pereira, Venturini, Wegner, and Braga (2010) 29 Constant conflicts Ebers (1997); Ariño and De La Torre (1998); Messner, Meyer-Stamer (2000); Kajüter and Kulmala (2005); Lui (2009) 30 Lack of regular performance reviews Cooper and Slagmulder (1999); Gareth (2005) Source: Survey data (2016). e 3. Inhibitory factors related to the results, reviews and adjustments that occur in relationships Table 3. Inhibitory factors related to the results, reviews and adjustments that occur in relationships Pereira et al., 2010 explain that the imbalance between short-term and long-term gains can be an inhibiting factor of the Interorganizational relationship. According to the authors, the partner companies are generally geared for the short term and the involved parties feel motivated to explore and achieve benefits quickly. However, if the partners are far too oriented to the long term, their motivation to maintain the relationship can weaken. It is thus important that the parties determine the balance between the benefits in the short and long term, in order to avoid demotivation and the erosion of the relationship. 4.1 Group 1 – Formation of Collaborative Relationships and Strategies of Companies The author emphasizes that managers have to understand that the availability of internal data is a reality and that companies need to evaluate if this issue does not jeopardize the interrelationship. Ramos (2004) points out that when the company seeks to strengthen the relationship with its suppliers, a new context arises. The environment becomes more complex; the necessity to coordinate the activities of accounting with those of logistics emerges; a need to work across organizational boundaries grows, and a drive to create new management systems, among others, comes to the forefront. In this situation, companies must evaluate if the department responsible for cost management has the competence necessary to fulfill the duties of the new context (Cokins, 1998; Cullen et al., 1999; Ramos, 2004). Kulmala et al. (2002) explain that companies need to evaluate whether the relationships with suppliers are beneficial for their business or not. This way, they can calculate the amount of cost reduction that relationships will offer them. Toward this end, they need to have a clear understanding of cost behavior so that their analysis can illustrate whether the relationship is beneficial or not. In order to prepare employees for the cost reduction programs, companies must provide training and education to describe the changes that may occur (Himme, 2012). Thus, the lack of expertise and education programs geared towards 97 International Business Research www.ccsenet.org/ibr Vol. 9, No. 8; 2016 employees regarding the IOCM are regarded as barriers to the success of the approach (LaLonde & Pohlen, 1996; Cokins, 1998; Cullen et al., 1999; Nicolini et al., 2000; Kulmala et al., 2002; Ellram, 2002b; Ramos, 2004; Seal et al., 2004; Thomson & Gurowka, 2005; Himme, 2012). employees regarding the IOCM are regarded as barriers to the success of the approach (LaLonde & Pohlen, 1996; Cokins, 1998; Cullen et al., 1999; Nicolini et al., 2000; Kulmala et al., 2002; Ellram, 2002b; Ramos, 2004; Seal et al., 2004; Thomson & Gurowka, 2005; Himme, 2012). Group 3 – Results, Reviews and Adjustments that Occur in Relationships 8; 2016 Internationall Business Reseaarch www.ccsenet.orrg/ibr While joint m adjustments in IOCM can le Relationships Partners shou necessary rigi management tak n the relations ead to a prob among compa uld develop a dity to keep th kes place, learn ship between c blematic situat anies that are t relationship in he structure as ning about the companies may tion due to th too rigid or too n which they otherwise agre e behavior of c y occur (Dubo he dynamics a o flexible may can mold to eed among the costs succeeds, ois, 2003; Gar and developm y find it difficu the changes t e parties (Garet , then, continu reth, 2005). Th ment of the ne ult to achieve s that might occ th, 2005). ous improvem hus, the inflex eeds of both success (Gareth cur without lo ments and ibility in partners. h, 2005). osing the Nicolini et a interrelationsh determination difficulties to determining p share, compan subject to vary al. (2000) em hips could be n of prices to o determine th prices, for exam ny strategies a ying influence mphasizes that a barrier to IO consumers is he prices of o mple, the pric and the prices o s from compan t the inability OCM. Accordi s easier, once other products e of competito of substitute g nies. y to determin ing to the auth prices are se s. There are m ors, customer goods, among o ne market pri hors, for some t by the mark many factors t characteristics others. In othe ices of produ products, suc ket, however, that managers s, the threat of er words, price ucts arising f ch as commod companies m s must consid f new entrants e determination from the ities, the may have der when s, market n will be Lin et al. (20 companies mu become exces to meet their n customer, for 001) state that ust build their ssively alert to needs. Group 3 – Results, Reviews and Adjustments that Occur in Relationships Another obstacle to Interorganizational relationships is the uneven distribution of benefits (Cooper & Yoshikawa, 1994; Cooper & Slagmulder, 1999; Kulmala, 2004). Companies tend not to want to cooperate and share cost information when the benefits are not shared fairly. Christopher (2007) points out that this does not mean that the benefits should be shared equally, but that those involved are in agreement and pleased with the benefits concerning each one. Kajüter and Kulmala (2005) explain that there is no general rule on how the benefits generated by the inter-relationship should be shared, which is justified by the fact that situations vary from case to case. Consequently, a variety of contracts has arisen, for example, the equitable sharing of benefits and the granting of benefits only for the supplier – which in turn commits not to increase the price of their products –, among others (Kajüter & Kulmala, 2005). Waeytens and Bruggeman (1994) point out that problems in the design of models related to cost management can hamper companies from being able to determine their costs correctly and, as a result, deter collaboration with other members of their chain. A poorly prepared IOCM model design, with poorly formulated activities, which does not illustrate the real cost of the company's activities, is seen as a barrier to IOCM (Cokins, 1998, 2003; Kaplan & Anderson, 2004; Gareth, 2005). A project poorly developed from the onset is a factor that can lead to an unsuccessful partnership (Waeytens & Bruggeman, 1994; Cokins, 1998). Kaplan and Anderson (2004) warn that highly complex IOCM systems can inhibit their applicability. The models implemented by companies tend to evolve while they learn more about the variety and complexity of their processes, applications, suppliers and customers. In order to increase the accuracy and details of the information, the models may become too complex, including the sharing of unnecessary information (Cokins, 1998; Messner & Meyer-Stamer, 2000; Kaplan & Anderson, 2004). The simpler the system and the more a company understands about the processes from their partners, the greater the potential for implementing the approach (Park & Ungson, 2001; Kajüter & Kulmala, 2005). 98 Vol. 9, No. Group 3 – Results, Reviews and Adjustments that Occur in Relationships A strate the work is us t companies s relationships b issues related egic cost mana eless when yo should be aler based on these d to the costs in agement plan r u lose the focu rt so that ther e, as well as im n a way that th requires thorou us on them (Co e is no loss o mprove profit p hey will lose fo ugh planning i okins, 1998; H of customer fo potential. How ocus on their c in order to ma Himme, 2012). ocus. For the wever, managem customers and aintain the focu authors, ment can thus fail us on the Ebers (1997) develop skills maintaining re conditioned to Interorganizat companies aba the desired ini explains that c s. According elationships. W o its ability to tional relations andon partners itial goals can companies join to the author, Wegner and Pa o achieve the ship can lead to ships (Ebers, 1 cause the conf n efforts in ord , unfulfilled in adula (2010) s goals set by o dissatisfactio 1997; Pereira, flicts, among t der to achieve nitial goals ca state that the c the companie on and disagre Venturini, Weg the companies, common goal an lead to dis continuation of s. Failure to a eement betwee gner, and Brag , to happen mo ls, among othe ssatisfaction a f Interorganiza achieve the go n the parties, o ga 2010). Furth ore often (Eber ers, to reduce c and lack of in ational relation oals establishe one of the reas hermore, not a rs, 1997). costs and nterest in nships is ed by an sons why achieving Problems and development Ariño & De L relationships parties, financ d difficulties ar of relationship La Torre, 199 can end after cial resources a re expected in ps. Neverthele 98; Messner & the wear caus and commitme n partnerships ess, constant c & Meyer-Stam sed by conflic ent. and conflicts conflicts can l mer, 2000; Kajü cts, even if the among the pa lead to the en üter & Kulma ey have begun arties are a nat nd of the relat ala, 2005). Group 3 – Results, Reviews and Adjustments that Occur in Relationships Lu n based on tru tural compone tionship (Eber ui (2009) expla ust, goodwill f nt to the rs, 1997; ains that from the Gareth (2005) necessary adju is seen as a b reviews, are a ) explains tha ustments so th barrier to a su also the most vu at relationships hat companies uccessful IOC ulnerable to m s among comp can overcome M. Excessive meet a definitiv panies must g e their difficult ly stable relat ve end. go through reg ties, once the l tionships, whi gular reviews, lack of regular ch, in turn, d in order to m r performance o not undergo make the reviews o regular Lastly, it bear environment Interorganizat contribute to t occur in their s mentioning t in which the tional Cost Ma the analysis of companies. that this study companies op anagement app f the factors, a does not inten perate. It is b proach, even w and in practice nd to exhaust th elieved that d with the limita , it means help he topic, as IO diagnosing the ations that eve ping managers OCM inhibitors e factors that ery study has, s to minimize s vary accordin inhibit the su it means, in th the problems t ng to the uccess of heory, to that may 4.4 Discussionn Bastl et al. (20 analyzed in th these criteria. Tables 1, 2 an 010) suggest th hree dimension The numbers nd 3. hat the inhibit ns, namely: Pe in the figure r tors that affect eople, Process refer to the inh the cost mana s and Technolo hibitors presen agement appro ogy. Figure 1 c nted in the theo oaches between classifies the i oretical framew n companies sh inhibitors acco work correspo hould be ording to onding to Tables 1, 2 an Figure 1. Inhib nd 3. bitors classified according too their relation nship with the ddimensions: Peeople, Processs and Technology Figure 1. Inhibbitors classified according too their relationnship with the ddimensions: Peeople, Processs and Technology Figure 1. Inhibbitors classified according too their relationnship with the ddimensions: Peeople, Processs and Technology 99 Vol. 9, No. 8; 2016 International Business Research www.ccsenet.org/ibr Source: Adapted from Bastl et al. (2010) Upon employing the perspective in Bastl et al. (2010), it can be observed that the predominant inhibitors in the companies surveyed are mainly related to People. This is to say that if Interorganizational efforts are to be successfully undertaken, the first aspect to be worked on are the people, who need to know about the importance of vision processes, information sharing, and the establishment of collaborative long-term relationships, in order to improve the performance of all companies and chains they belong to. The use of technology and automation processes can make organizations more efficient; however, they remain predominantly dependent on actions taken by people. Companies can make specific investments in different sectors, such as technology and operational processes, so logically investments in the apprenticeship of employees can improve the organization as a whole (Barney & Hesterly, 1996). Some behaviors, such as opportunism, resistance to change processes, lack of interest in supporting new projects, disloyalty to other companies, among others, may reflect and trigger other inhibiting factors that interfere with the success of IOCM. Applying the classification used in this study, Table 4 shows a model which represents all identified Applying the classification used in this study, Table 4 shows a model which represents all identified inhibiting factors. T bl 4 I hibit id tifi d i thi t d pp y g y, p g Table 4. 5. Conclusions and Recommendations The advantages in establishing partnerships with suppliers and customers are many, however, there are countless difficulties and factors which inhibit the relationships that aim to improve performance. The implementation of IOCM may be hampered if companies do not have the internal capabilities to achieve it. In light of this, companies must overcome internal barriers and inefficiencies first, and then undertake Interorganizational efforts. It is concluded that Interorganizational Cost Management is an interdisciplinary phenomenon and collaborative relationships between suppliers and customers are based on many different factors. The diversity of identified inhibitors (30 factors) confirms the complexity and interdisciplinarity of the phenomenon studied. The application of the IOCM cannot be regarded as a purely technical approach, ruled by technology and management programs alone. Most identified inhibitors relate to people, their behavior and skills. That is, before the company has well-structured processes and technologies that help in the relationship between suppliers and customers, it needs to develop its human capital. Based on this, it is concluded that, in order for companies to implement IOCM successfully, it is necessary that the people involved in the processes cooperate with the approach. Inhibitors were found in all stages of building relationships. This suggests that the difficulties and barriers faced by managers are not restricted to a particular time, but rather occur constantly. This study stresses that managers make proactive us of mechanisms to help regulate relationships and resolve potential problems with partners. It is expected that the identification of inhibitors contributes to reduce the risks related to the implementation of the approach, as well as risks related to the building of relationships. As a general goal, this study sought to contribute by negating the view that there are no risks and problems in the inter-relationship, from the both academy’s point of view and that of managers working in the field. It is suggested that future studies provide continuity for studies related to barriers that hinder coordination of efforts between suppliers and customers. The lack of studies on the subject is seen as a gap in the literature. Another research opportunity is the analysis of the relationship between the inhibitors of IOCM. This calls into question the existence of correlations among factors and how a given inhibiting factor affects another. Additionally, it is also suggested that studies analyze how the strategic focus and idiosyncrasies of companies interplays with the decision to form partnerships. Source: Adapted from Bastl et al. (2010) Inhibitors identified in this study Stage 1 Stage 2 Stage 3 People Differences between companies’ strategic plans Different expectations of behavior between companies Lack of recognition that costing systems are necessary Lack of interest on the part of the partner companies in sharing information Lack of perception of the value resulting from the relationship Lack of trust among the partners in IOCM Opportunism Lack of incentives to innovate and prosper Resistance to change Lack of management support from members of organizations Lack of expertise and education programs geared towards employees Imbalance between short-term and long-term gains Uneven distribution of benefits Constant conflicts People – Process Lack of resources to execute the project Lack of long-term management goals Dependence of mutual learning Difficulty in defining mutual benefits Poorly prepared IOCM model design Highly complex IOCM systems Inflexibility in IOCM Inability to determine market prices Loss of customer focus Unfulfilled initial goals Lack of regular performance reviews People – Process – Technology Lack of understanding of costs along the chain Unreliability of internal cost data Inadequate level of data sharing Technology – Process Difficulties in integrating information between companies Poor internal availability of data Table 4. Inhibitors identified in this study Table 4. Inhibitors identified in this study Imbalance between short-term and long-term gains Uneven distribution of benefits Constant conflicts Poorly prepared IOCM model design Highly complex IOCM systems Inflexibility in IOCM Inability to determine market prices Loss of customer focus Unfulfilled initial goals Lack of regular performance reviews Unreliability of internal cost data Inadequate level of data sharing Source: Survey data (2016). Source: Survey data (2016). Coincidentally, it is observed that the 30 factors are equally divided into three groups: Group 1 is formed of inhibiting factors related to the formation of collaborative relationships and strategies of companies; Group 2, of inhibiting factors related to the development of processes and construction of relationships; and Group 3, of inhibiting factors related to the results, reviews and adjustments that produced by relationships. The groups indicate that inhibitors may be present at all stages of construction of relationships between companies. As indicated previously, a strong relationship between companies based in stability, cooperation, trust and 100 International Business Research Vol. 9, No. 8; 2016 www.ccsenet.org/ibr interdependence is a necessary component of the implementation of IOCM (Cooper & Slagmulder, 1999; Dekker, 2003, 2004; Kajüter & Kulmala, 2005; Souza & Rocha, 2009; Windolph & Moeller, 2012). Source: Adapted from Bastl et al. (2010) Thus, managers cannot ignore the above-mentioned factors at any stage of the construction of relationships with chain members. interdependence is a necessary component of the implementation of IOCM (Cooper & Slagmulder, 1999; Dekker, 2003, 2004; Kajüter & Kulmala, 2005; Souza & Rocha, 2009; Windolph & Moeller, 2012). Thus, managers cannot ignore the above-mentioned factors at any stage of the construction of relationships with chain members. An interest in forming strategic partnerships must be accompanied by concerns about the protection of the companies in order to avoid or solve possible problems. The process of development of Interorganizational relationships requires, in most cases, the application of mechanisms to help regulate relationships (Cooper & Slagmulder, 1999; Pereira et al., 2010). These are tools for planning and control used to aid the implementation of IOCM, seeking to hone, control, guide and incentivize the process (Cooper & Slagmulder, 1999, 2003a, 2003b, 2004; Kajüter & Kulmala, 2005; Souza & Rocha, 2009). Being aware that the inhibitors are present at all stages of relationships, companies must establish rules, specify rights and duties, codes of ethics and a committee that will oversee the completion of stated rules, and use mechanisms to, among other purposes, punish opportunistic behavior (Abbade, 2005; Janowicz-Panjaitan & Noorderhaven, 2009). As pointed out, there are many factors that influence the applicability of the IOCM, however, the factors may affect the interactions among companies in different ways. The characteristics and nuances of each organization leave them susceptible to factors in different ways. There are many characteristics of the companies that influence the applicability of management approaches, be they intra or Interorganizational in nature (Mintzberg, 1995). After all, the traits of the larger environmnet bear a great deal of influence on the applicability of IOCM (Cooper & Slagmulder, 1999; Souza & Rocha, 2009). Mintzberg (1995) highlights the following characteristics in particular: a) age and size of the organization; b) technical system (formalization, technical sophistication, operational automation) used for the production of their goods or services; c) environmental characteristics in which the organization operates (dynamics, complexity and sector diversity); and d) power over the organization (considering competitors and suppliers). As such, the study emphasizes the understanding that the environment in which the company operates and its characteristics interfere with the use of management approaches. References Abbade, E. B. (2005). Cooperação interorganizacional na associação londrinense de empresários supermercadistas e no APL têxtil de Goioerê: Uma análise sob a ótica da Teoria dos Custos de Transação. (Master’s thesis, University of Paraná, Curitiba, Brazil). Retrieved from http://zip.net/bksHpw Agndal, H., & Nilsson, U. (2008). Supply chain decision-making supported by an open books policy. International Journal of Production Economics, 116(1), 154-167. http://dx.doi.org/10.1016/j.ijpe.2008.08.038 Amato, N. J. (2000). Redes de cooperação produtiva e clusters regionais: Oportunidades para as pequenas e médias empresas (3rd ed.). São Paulo, SP: Atlas. Arino, A., & De La, T. J. (1998). Learning from failure: Towards an evolutionary model of collaborative ventures. Organization Science, 9(3), 306-325. http://dx.doi.org/10.1287/orsc.9.3.306 Axelsson, B., Laage-Hellman, J., & Nilsson, U. (2002). Modern management accounting for modern purchasing. European Journal of Purchasing & Supply Chain Management, 8(1), 53-62. http://dx.doi.org/10.1016/S0969-7012(01)00017-X Barney, J. B., & Hesterly, W. (1996). Organizational economics: Understanding the relationship between organizations and economic analysis (1st ed.). London, UK: Sage. Bastl, M., Grubic, T., Templar, S., Harrison, A., & Fan, I. (2010). Inter-organisational costing approaches: The inhibiting factors. The International Journal of Logistics Management, 21(1), 65-88. http://dx.doi.org/10.1108/09574091011042188 Bryson, J., Crosby, B., & Stone, M. M. (2006). The design and implementation of cross-sector collaborations: Prepositions from the literature. Public Administration Review, 66(1), 44-55. http://dx.doi.org/10.1111/j.1540-6210.2006.00665.x Cândido, G. A. (2002). A formação de redes interorganizacionais como mecanismo para geração de vantagem competitiva e para promoção do desenvolvimento regional: O papel do estado e das políticas públicas neste cenário. Revista Eletrônica de Administração, 8(4), 32-47. Cannon, J. P., & Homburg, C. (2001). Buyer-supplier relationships and customer firm costs. Journal of Marketing, 65(1), 29-43. http://dx.doi.org/10.1509/jmkg.65.1.29.18136 Chen, B. (2010). Antecedents or processes? Determinants of perceived effectiveness of Interorganizational collaborations for public service delivery. International Public Management Journal, 13(4), 381-407. http://dx.doi.org/10.1080/10967494.2010.524836 Christopher, M. (2007). Logística e gerenciamento da cadeia de suprimentos: Criando redes que agregam valor (2nd ed.). São Paulo, SP: Thomson Learning. Coad, A., & Cullen, J. (2006). Inter-organisational cost management: Towards an evolutionary perspective. Management Accounting Research, 17(4), 342-369. http://dx.doi.org/10.1016/j.mar.2006.02.003 Coad, A., & Scapens, R. (2006). Inter-organisational cost management. Financial Management, 35(2), 38-39. Cokins, G. (1998). Why is traditional accounting failing managers? Hospital Material Management Quarterly, 20(2), 72-80. Cokins, G. (2003). Measuring profits and costs across the supply chain for collaboration. Cost Management, 17(5), 22-29. Cooper, R., & Slagmulder, R. (1999). Supply chain development for the lean enterprise: Interorganizational cost management (1st ed.). 5. Conclusions and Recommendations This analysis allows to discover if there are profiles of companies that are likely to act collaboratively with suppliers and customers. The recommendation is also made that research with an interdisciplinary approach be carried out to study mechanisms to minimize and resolve the difficulties brought by inhibitors. This can assist 101 International Business Research Vol. 9, No. 8; 2016 www.ccsenet.org/ibr in overcoming the difficulties brought about by the larger context in which the company operates. Finally, it is suggested that studies evaluate if IOCM inhibitors vary according to the location where the companies operate within the supply chain. Thus, it will be possible to identify which factors are common to companies that are closer to raw materials and the companies that are closer to the consumer of the chain’s finished product. Along these lines, the study of Interorganizational Cost Management and the challenges encountered in their approach present opportunities for further investigation aimed at developing knowledge about the task of managing costs in an Interorganizational manner. Finally, it is suggested that studies evaluate if IOCM inhibitors vary according to the location where the companies operate within the supply chain. Thus, it will be possible to identify which factors are common to companies that are closer to raw materials and the companies that are closer to the consumer of the chain’s finished product. Along these lines, the study of Interorganizational Cost Management and the challenges encountered in their approach present opportunities for further investigation aimed at developing knowledge about the task of managing costs in an Interorganizational manner. References Trenton, NJ: The IMA Foundation for Applied Research. Cooper, R., & Slagmulder, R. (2003a). Interorganizational costing: Part 1. Cost Management, 17(5), 14-21. Retrieved from http://maaw.info/ArticleSummaries/ArtSumCooperSlagmulder03%281%29.htm Cooper, R., & Slagmulder, R. (2003b). Interorganizational costing: Part 2. Cost Management, 17(6), 12-24. Retrieved 102 Vol. 9, No. 8; 2016 International Business Research www.ccsenet.org/ibr from http://maaw.info/ArticleSummaries/ArtSumCooperSlagmulder03%282%29.htm Cooper, R., & Slagmulder, R. (2004). Interorganizational cost management and relational context. Accounting, Organizations and Society, 29(1), 1-26. http://dx.doi.org/10.1016/S0361-3682(03)00020-5 Cooper, R., & Yoshikawa, T. (1994). Interorganizational cost management systems: The case of the Tokyo-Yokohama-Kamakura supplier chain. International Journal of Production Economics, 37(1), 51-62. http://dx.doi.org/10.1016/0925-5273(94)90007-8 Cullen, J., Berry, A., Seal, W., Dunlop, A., Ahmend, M., & Marson, J. (1999). Interfirm supply chains: The contribution of management accounting. Management Accounting, 77(6), 30-32. Dekker, H. C. (2003). Value chain analysis in interfirm relationships: A field study. Management Accounting Research, 14(1), 1-23. http://dx.doi.org/10.1016/S1044-5005(02)00067-7 Dekker, H. C. (2004). Control of Interorganizational relationships: Evidence on appropriation concerns and coordination requirements. Accounting, Organizations and Society, 29(1), 27-49. http://dx.doi.org/10.1016/S0361-3682(02)00056-9 Drucker, P. F. (1997). Toward the new organization. Executive Excellence, 14(2), 17-28. http://dx.doi.org/10.1002/ltl.40619970304 Dubois, A. (2003). Strategic cost management across boundaries of firms. Industrial Marketing Management, 32(5), 365-374. http://dx.doi.org/10.1016/S0019-8501(03)00010-5 Ebers, M. (1997). The formation of Interorganizational networks (1st ed.). London, UK: Oxford University Press. Edelman, L., Bresnan, M., Newell, S., Scarborough, H., & Swan, J. (2004). The benefits and pitfalls of social capital: Empirical evidence of two organizations in the United Kingdom. British Journal of Management, 15(1), 59-69. http://dx.doi.org/10.1111/j.1467-8551.2004.00406.x Ellram, L. (2002a). Supply management’s involvement in the target costing process. European Journal of Purchasing & Supply Management, 8(4), 235-244. http://dx.doi.org/10.1016/S0969-7012(02)00019-9 Ellram, L. (2006). The implementation of target costing in the United States: Theory versus practice. Journal of Supply Chain Management, 42(1), 13-26. http://dx.doi.org/10.1111/j.1745-493X.2006.04201003.x Ellram, L. M. (1994). A taxonomy of Total Cost of Ownership models. Journal of Business Logistics, 1 Ellram, L. M. (2002b). Strategic cost management in the supply chain: A purchasing and supply management perspective (1st ed.). Phoenix, AZ: CAPS Research. Ellram, L. M., & Siferd, S. P. (1998). Total Cost of Ownership: A key concept in strategic cost management decisions. Journal of Business Logistics, 19(1), 55-84. Farias, R. A. S. (2016). Interorganizational Cost Management study on inhibitor. International Business Research, 9(3), 68-78. http://dx.doi.org/10.5539/ibr.v9n3p68 Ferrin, B. G., & Plank, R. E. (2002). Total Cost of Ownership Models: An exploratory study. Journal of Supply Chain Management, 38(3), 18-29. http://dx.doi.org/10.1111/j.1745-493X.2002.tb00132.x Gareth, B. (2005). Perfecting partnerships. References Professional Engineering, 17(21), 34-44. d, E., & Becker, H. (1995). Activity-Based Costing and management (1st ed.). New York, NY: Wiley. Gupta, K. M., & Gunasekaran, A. (2004). Costing in new enterprise environment: A challenge for managerial accounting researchers and practitioners. Managerial Accounting Journal, 20(4), 337-357. http://dx.doi.org/10.1108/02686900510592034 Hakansson, H., & Snehota, I. (1995). Developing relationships in business networks (1st ed.). London, UK: Routledge. Himme, A. (2012). Critical success factors of strategic cost reduction. Metrika, 23(3), 183-210. http://dx.doi.org/10.1007/s00187-012-0157-8 Hitt, M. A., Dancin, M. T., Levitas, E., Arregle, J., & Borza, A. (2000). Partner selection in emerging and developed market contexts: Resource-based and organizational learning perspectives. Academy of Management Journal, 43(3), 449-467. http://dx.doi.org/10.2307/1556404 Janowicz-Panjaitan, M., & Noorderhaven, N. G. (2009). Trust, calculation, and interorganizacional learning of tacit knowledge: An organizational roles perspective. Organization Studies, 30(10), 1021-1044. http://dx.doi.org/10.1177/0170840609337933 103 Vol. 9, No. 8; 2016 International Business Research www.ccsenet.org/ibr Kajüter, P., & Kulmala, H. (2005). Open book accounting in networks: Potencial achievements and reasons for failures. Management Accounting Research, 16(2), 179-204. http://dx.doi.org/10.1016/j.mar.2005.01.003 Kaplan, R. S., & Anderson, S. R. (2004). Time-Driven Activity-Based Costing. Harvard Business Review, 82(11), 131-138. http://dx.doi.org/10.2139/ssrn.485443 Kulmala, H. (2002). Open-Book Accounting in networks. Finnish Journal of Business Economics, 52( Kulmala, H. (2004). Developing cost management in cutomer-supplier relationships: Three case studies. Journal of Purchasing & Supply Management, 10(2), 65-77. http://dx.doi.org/10.1016/j.pursup.2004.02.003 Kulmala, H., Paranko, J., & Uusi-Rauva, E. (2002). The role of cost management in network relationships. International Journal of Production Economics, 79(1), 33-43. http://dx.doi.org/10.1016/S0925-5273(00)00061-X LaLonde, B. (2003). Three problems that linger. Supply Chain Management Review, 7(2), 17-28. LaLonde, B., & Pohlen, T. (1996). Issues in Supply Chain Costing. The International Journal of Logistics Management, 7(1), 1-11. Lin, B., Collins, J., & Su, R. K. (2001). Supply chain activity: Na activity-based perspective. International Journal of Physical Distribution & Logistics Management, 31(9), 702-713. http://dx.doi.org/10.1108/EUM0000000006286 Lui, S. S. (2009). The roles of competence trust, formal contract, and time horizon in Interorganizational learning. Organization Studies, 30(4), 333-353. http://dx.doi.org/10.1177/0170840608101139 Messner, D., & Meyer-Stamer, J. (2000). Governance and networks: Tools to study the dynamics of clusters and global value chains (1st ed.). Disburg, RU: INEF. Miles, R., & Snow, C. (1992). Causes of failure in network organizations. California Management Review, 34(4), 53-72. http://dx.doi.org/10.2307/41166703 Milligan, B. (1999). Tracking total cost of ownership proves elusive. Purchasing, 127(3), 22-23. Mintzberg, H. (1995). Criando organizações eficazes (1st ed.). São Paulo, SP: Atlas. Möller, K., Windolph, M., & Isbruch, F. (2011). Inovação, 7(1), 62-83. http://dx.doi.org/10.5585/rai.v7i1.341 New York, NY: Free Press Wincent, J. (2008). An exchange approach on firm cooperative orientation and outcomes of strategic multilateral network participants. Group Organization Management, 33(3), 303-329. http://dx.doi.org/10.1177/1059601108318184 Windolph, M., & Moeller, K. (2012). Open-book accounting: Reason for failure of inter-firm cooperation? Management Accounting Research, 23(1), 47-60. http://dx.doi.org/10.1016/j.mar.2011.07.001 Windolph, M., & Moeller, K. (2012). Open-book accounting: Reason for failure of inter-firm cooperation? Management Accounting Research, 23(1), 47-60. http://dx.doi.org/10.1016/j.mar.2011.07.001 Zylbersztajn, D., & Neves, M. F. (2000). Economia e gestão dos negócios agroalimentares (1st ed.). São Paulo, SP: Pioneira. Zylbersztajn, D., & Neves, M. F. (2000). Economia e gestão dos negócios agroalimentares (1st ed.). São Paulo, SP: Pioneira. yright for this article is retained by the author(s), with first publication rights granted to the journal. References The effect of rational factors on open-book accounting and inter-organisational cost management in buyer-supplier partnerships. Journal of Purchasing & Supply Management, 17(2), 121-131. http://dx.doi.org/10.1016/j.pursup.2011.01.002 Mouritsen, J., & Thrane, S. (2006). Accounting, network complementarities and the development of inter-organisational relations. Accounting, Organizations and Society, 31(3), 241-275. http://dx.doi.org/10.1016/j.aos.2005.04.002 Munday, M. (1992). Accounting cost data disclosure and buyer-supplier partnerships: A research note. Management Accounting Research, 3(3), 245-250. http://dx.doi.org/10.1016/S1044-5005(92)70014-6 Nassimbeni, G. (1998). Network structures and coordination mechanisms: A taxonomy. International Journal of Operations & Production Management, 18(6), 538-554. Nicolini, D., Tomkins, C., Holti, R., Oldman, A., & Smalley, M. (2000). Can Target Costing be applied in the construction industry? Evidence from two case studies. British Journal of Management, 11(4), 303-324. http://dx.doi.org/10.1111/1467-8551.00175 Norek, C. D., & Pohlen, L. (2001). Cost knowledge: A foundation for improving supply chain relationships. International Journal of Logistics Management, 12(1), 37-51. http://dx.doi.org/10.1108/09574090110806217 Ojala, M., & Hallikas, J. (2007). Financial decision risks in supplier networks. Managing Supply Chain Risks, 8(2), 80-91. http://dx.doi.org/10.1080/16258312.2007.11517184 Oliver, C. (1990). Determinants of Interorganizational relationships: Integration and future directions. Academy of Management Review, 15(2), 241-265. http://dx.doi.org/10.5465/AMR.1990.4308156 Olson, M. (1999). A lógica da ação coletiva: Os benefícios públicos e uma teoria dos grupos sociais (1st ed.). São Paulo, SP: EDUSP. Park, S. H., & Ungson, G. (2001). Interfirm rivalry and managerial complexity: A conceptual framework of alliance failure. Organization Science, 12(1), 37-53. http://dx.doi.org/10.1287/orsc.12.1.37.10118 Pereira, B. A. D., Alves, J. N., & Silva, P. E. (2010). O lado obscuro dos relacionamentos interorganizacionais: Por que as empresas saem dos processos cooperativos? Espacios, 31(4), 25-27. Pereira, B. A. D., Venturini, J. C., Wegner, D., & Braga, A. L. (2010). Desistência da cooperação e encerramento de redes interorganizacionais: Em que momento essas abordagens se encontram? Revista de Administração e 104 Vol. 9, No. 8; 2016 International Business Research www.ccsenet.org/ibr This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). Inovação, 7(1), 62-83. http://dx.doi.org/10.5585/rai.v7i1.341 Inovação, 7(1), 62-83. http://dx.doi.org/10.5585/rai.v7i1.341 Inovação, 7(1), 62-83. http://dx.doi.org/10.5585/rai.v7i1.341 Porter, M. (1989). Vantagem competitiva: Criando e sustentando um desempenho superior (12th ed.). Rio de Janeiro, RJ: Elsevier. Prahalad, C., & Ramaswamy, V. (2004). O Futuro da competição: Como desenvolver diferenciais inovadores em parceria com os clientes (1st ed.). Rio de Janeiro, RJ: Elsevier. Ramos, M. M. (2004). Interaction between management accounting and supply chain management. Supply Chain Management: An International Journal, 9(2), 134-138. http://dx.doi.org/10.1108/13598540410527033 Ring, P. S., & Van de Ven, A. H. (1994). Developmental processes of cooperative Interorganizational relationships. The Academy of Management Review, 19(1), 90-118. http://dx.doi.org/10.5465/AMR.1994.9410122009 Rogers, E. (1995). Diffusion of innovations (4th ed.). New York, NY: The Free Press. Rogers, E. (1995). Diffusion of innovations (4th ed.). New York, NY: The Free Press. Sadowski, B., & Duysters, G. (2008). Strategic technology alliance termination: An empirical investigation. Journal of Engineering and Technology Management, 25(1), 305-320. http://dx.doi.org/10.1016/j.jengtecman.2008.10.002 Seal, W., Berry, A., & Cullen, J. (2004). Disembedding the supply chain: Institutionalized reflexivity and inter-firm accounting. Accounting, Organizations and Society, 29(1), 73-92. http://dx.doi.org/10.1016/S0361-3682(02)00055-7 Shipilov, A., Rowley, T., & Aharonson, B. (2006). When do network matter? A study of tie formation and decay. Advances in Strategic Management, 23(2), 481-515. http://dx.doi.org/10.1016/S0742-3322(06)23015-8 Souza, B. C., & Rocha, W. (2009). Gestão de Custos Interorganizacionais: Ações coordenadas entre clientes e fornecedores para otimizar resultados (1st ed.). São Paulo, SP: Atlas. Stapleton, D., Pati, S., Beach, E., & Julmanichoti, P. (2004). Activity-Based Costing for logistics and marketing. Business Process Management Journal, 10(5), 584-597. http://dx.doi.org/10.1108/14637150410559243 Stephenson, G. (2003). The somewhat flawed theoretical foundation of the extension service. Journal of Extension, 41(4), 16-30. Thomson, J., & Gurowka, J. (2005). Sorting out the clutter. Strategic Finance, 87(2), 27-33. Tomkins, C. (2001). Interdependencies, trust and information in relationships, alliances and networks. Accounting, Organizations and Society, 26(2), 161-191. http://dx.doi.org/10.1016/S0361-3682(00)00018-0 Vosselman, E., & Van Der Meer-Kooistra, J. (2006). Efficiency seeking behavior in changing management control in interfirm transactional relationships: An extended transaction cost economics perspective. Journal of Accounting & Organizational Change, 2(2), 123-143. http://dx.doi.org/10.1108/18325910610675970 Waeytens, D., & Bruggeman, W. (1994). Barriers to successful implementation of ABC for continuous improvement: A case study. International Journal of Production Economics, 36(2), 39-52. http://dx.doi.org/10.1016/0925-5273(94)90147-3 Wegner, D., & Padula, A. D. (2010). Governance and management of horizontal business networks: An analysis of retail networks in Germany. International Journal of Business & Management, 5(12), 74-88. http://dx.doi.org/10.5539/ijbm.v5n12p74 Williamson, O. E. (1985). The economic institutions of capitalism (5th ed.). Copyrights Copyright for this article is retained by the author(s), with first publication rights granted to the journal. This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license Copyright for this article is retained by the author(s), with first publication rights granted to the journal. This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons org/licenses/by/3 0/) Copy g t o t s a t c e s eta ed by t e aut o (s), w t st pub cat o g ts g a ted to t e jou a . This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). 105
https://openalex.org/W2108533573	https://bg.copernicus.org/articles/11/4679/2014/bg-11-4679-2014.pdf	English	null	Biophysical controls on net ecosystem CO&lt;sub&gt;2&lt;/sub&gt; exchange over a semiarid shrubland in northwest China	Biogeosciences	2,014	cc-by	12,500	Biophysical controls on net ecosystem CO2 exchange over a semiarid shrubland in northwest China X. Jia1,2, T. S. Zha1,2, B. Wu1,2, Y. Q. Zhang1,2, J. N. Gong3, S. G. Qin1,2, G. P. Chen4, D. Qian1,2, S. Kellomäki3, and H Peltola3 1Yanchi Research Station, College of Soil and Water Conservation, Beijing Forestry University, Beijing 100083, China 2Key Laboratory of soil and Water Conservation and Desertiﬁcation Combating, Ministry of Education, Beijing Forestry University, Beijing 100083, China y j g 3School of Forest Sciences, University of Eastern Finland, P.O. Box 111, 80101 Joensuu, Finland 4Institute of Forest Sciences, Bailongjiang Forestry Management Bureau, Wudu, Gansu 746010, China Correspondence to: T. S. Zha (tianshanzha@bjfu.edu.cn) Correspondence to: T. S. Zha (tianshanzha@bjfu.edu.cn) Received: 13 March 2014 – Published in Biogeosciences Discuss.: 31 March 2014 Revised: 10 July 2014 – Accepted: 23 July 2014 – Published: 8 September 2014 Abstract. The carbon (C) cycling in semiarid and arid areas remains largely unexplored, despite the wide distribution of drylands globally. Rehabilitation practices have been carried out in many desertiﬁed areas, but information on the C se- questration capacity of recovering vegetation is still largely lacking. Using the eddy-covariance technique, we measured the net ecosystem CO2 exchange (NEE) over a recovering shrub ecosystem in northwest China throughout 2012 in or- der to (1) quantify NEE and its components and to (2) exam- ine the dependence of C ﬂuxes on biophysical factors at mul- tiple timescales. The annual budget showed a gross ecosys- tem productivity (GEP) of 456 g C m−2 yr−1 (with a 90 % prediction interval of 449–463 g C m−2 yr−1) and an ecosys- tem respiration (Re) of 379 g C m−2 yr−1 (with a 90 % pre- diction interval of 370–389 g C m−2 yr−1), resulting in a net C sink of 77 g C m−2 yr−1 (with a 90 % prediction interval of 68–87 g C m−2 yr−1). The maximum daily NEE, GEP and Re were −4.7, 6.8 and 3.3 g C m−2 day−1, respectively. Both the maximum C assimilation rate (i.e., at the optimum light intensity) and the quantum yield varied over the growing sea- son, being higher in summer and lower in spring and autumn. At the half-hourly scale, water deﬁcit exerted a major control over daytime NEE, and interacted with other stresses (e.g., heat and photoinhibition) in constraining C ﬁxation by the vegetation. Low soil moisture also reduced the temperature sensitivity of Re (Q10). Biophysical controls on net ecosystem CO2 exchange over a semiarid shrubland in northwest China At the synoptic scale, rain events trig- gered immediate pulses of C release from the ecosystem, fol- lowed by peaks of CO2 uptake 1–2 days later. Over the entire growing season, leaf area index accounted for 45 and 65 % of the seasonal variation in NEE and GEP, respectively. There was a linear dependence of daily Re on GEP, with a slope of 0.34. These results highlight the role of abiotic stresses and their alleviation in regulating C cycling in the face of an in- creasing frequency and intensity of extreme climatic events. growing season, leaf area index accounted for 45 and 65 % of the seasonal variation in NEE and GEP, respectively. There was a linear dependence of daily Re on GEP, with a slope of 0.34. These results highlight the role of abiotic stresses and their alleviation in regulating C cycling in the face of an in- creasing frequency and intensity of extreme climatic events. Biogeosciences, 11, 4679–4693, 2014 www.biogeosciences.net/11/4679/2014/ doi:10.5194/bg-11-4679-2014 © Author(s) 2014. CC Attribution 3.0 License. Biogeosciences, 11, 4679–4693, 2014 www.biogeosciences.net/11/4679/2014/ doi:10.5194/bg-11-4679-2014 © Author(s) 2014. CC Attribution 3.0 License. 1 Introduction Drylands (semiarid and arid areas) cover over 40 % of the earth’s land surface and have been rapidly expanding as a result of climate change and human activities (Asner et al., 2003). For example, the total desertiﬁed area in China in- creased by 2460 km2 yr−1 from the 1980s to the mid-1990s (Yang et al., 2005). Although dryland ecosystems are char- acterized by low precipitation, soil fertility and productivity, they are important to the global carbon (C) budget as they ac- count for approximately 20 % of total terrestrial net primary productivity (Whittaker, 1975) and 15 % of total soil organic carbon (Lal, 2004). A recent study showed a dramatic contri- bution of semiarid ecosystems to the interannual variability of the global carbon cycle (Poulter et al., 2014). The C cy- cling in desert ecosystems is particularly sensitive to climate and land-use changes, and may feed back to the climate sys- tem (Li et al., 2005). In order to accurately predict global C cycling under changing climate, it is necessary to under- stand how CO2 exchange in dry areas responds to variations X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4680 X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland in climatic conditions (Gao et al., 2012). Currently, the C dynamics of desert shrub ecosystems and their responses to environmental factors are less well known compared to those of forests and grasslands (Gao et al., 2012). in climatic conditions (Gao et al., 2012). Currently, the C dynamics of desert shrub ecosystems and their responses to environmental factors are less well known compared to those of forests and grasslands (Gao et al., 2012). Shrubland ecosystems at the south edge of the Mu Us desert (also referred to as the Mu Us sandland) lie in a critical ge- ographic transition zone between arid and semiarid climates, and between agricultural and pastoral land uses. Overgrazing on the natural shrublands and steppes caused severe desertiﬁ- cation in this region (Chen and Duan, 2009). Grazing of natu- ral vegetation has been prohibited since the late 1990s. Thus, the vegetation has been given the opportunity to recover for over a decade. The rehabilitation of desertiﬁed lands has been evidenced by the increasingly ﬁne soil texture, increased nu- trient contents and biodiversity, and reduced wind erosion (Chen and Duan, 2009). The ability of the recovering ecosys- tems to sequester CO2 has not yet been assessed. This infor- mation, however, is essential to adaptive management under changing climate. in climatic conditions (Gao et al., 2012). Currently, the C dynamics of desert shrub ecosystems and their responses to environmental factors are less well known compared to those of forests and grasslands (Gao et al., 2012). g ( , ) Whether a dryland ecosystem is a net sink or source of CO2 is affected by the way it responds to climatic vari- ability (Liu et al., 2012). In semiarid and arid ecosystems, moisture-related factors such as precipitation, soil water con- tent (SWC) and vapor pressure deﬁcit (VPD) usually exert strong inﬂuences on diurnal, seasonal and interannual varia- tions in the net ecosystem CO2 exchange (NEE) (Fu et al., 2006; Gao et al., 2012). Water deﬁcit may depress gross ecosystem productivity (GEP) by limiting plant physiolog- ical processes (e.g., stomatal closure) and altering plant phe- nology (e.g., delayed leaf emergence) and canopy structure (e.g., reduced leaf area index, LAI) (Zhou et al., 2013). X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland Low water availability may also limit ecosystem respiration (Re) by reducing root activity, suppressing microbial decomposi- tion of organic matter and restricting the diffusion of extra- cellular enzymes and C substrates in the soil (Wang et al., 2014). Moreover, the effects of water availability on GEP and Re depend not only on the sensitivity of related biotic processes and the magnitude of water stress but also on the temporal pattern of water supply. For example, NEE in dry- land ecosystems showed complex and inconsistent responses to rainfall events (Liu et al., 2011; Gao et al., 2012), indi- cating our lack of understanding on how dryland ecosystems respond to water stress and its relief. Using the eddy-covariance technique, we measured NEE over a shrub ecosystem at the south edge of the Mu Us desert throughout 2012. Our objectives were (1) to quantify NEE and its partitioning into GEP and Re at diurnal, seasonal and annual scales and (2) to examine the dependence of NEE and its components on abiotic and biotic factors at multi- ple timescales. We hypothesized that soil water shortage is dominant over other stresses in controlling NEE of dryland ecosystems, and could modify the responses of NEE to other environmental factors. We also proposed that the seasonal dynamics of LAI is an important determinant of productivity over the growing season, whereas at shorter timescales (e.g., hourly) abiotic stresses could impose critical constraints on CO2 ﬂuxes. Besides water availability, NEE in arid and semiarid ecosystems is also affected by other abiotic and biotic fac- tors. Drought stress often accompanies thermal and irradia- tion stresses, as the cloudiness is usually low and the soil is readily heated up by solar radiation during dry periods. High leaf temperature can deactivate photosystem II, enhance the evaporative demand for plants and stimulate respiration (Fu et al., 2006). Strong irradiation is common in arid and semi- arid areas, and is likely to induce midday photosynthetic de- pression (Fu et al., 2006). In many ecosystems, canopy de- velopment (e.g., changes in LAI) is critical to the seasonal evolution of CO2 ﬂuxes (Xu and Baldocchi, 2004; Li et al., 2005). However, the large stochasticity of precipitation and variability of soil moisture in arid and semiarid ecosystems can obscure the effects of LAI (Wang et al., 2008). X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland Consid- ering the inconsistent effects of these environmental stresses and biotic factors on CO2 ﬂuxes (e.g., Fu et al., 2006; Aires et al., 2008; Wang et al., 2008), it is needed to examine the relative importance of these biophysical controls and their interactions in desert shrub ecosystems. Published by Copernicus Publications on behalf of the European Geosciences Union. Published by Copernicus Publications on behalf of the European Geosciences Union. 2.3 Meteorological measurements Incident photosynthetically active radiation (PAR) was mea- sured using a quantum sensor (PAR-LITE, Kipp and Zo- nen, the Netherlands). Net radiation (Rn) was measured us- ing a four-component radiometer (CNR-4, Kipp & Zonen, the Netherlands). Air temperature (Ta) and relative humid- ity were measured with a thermohygrometer (HMP155A, Vaisala, Finland). All these meteorological sensors were mounted on the tower at 6 m above the ground. Soil temper- ature (Ts) and water content (SWC) proﬁles were monitored adjacent to the tower using ECH2O-5TE sensors (Decagon Devices, USA) at four depths (0.1, 0.3, 0.7 and 1.2 m). Five soil heat plates (HFP01, Hukseﬂux Thermal Sensors, the Netherlands) were placed at 10 cm below the soil surface, within about 5 m of the tower base. Rainfall was measured from 15 May 2012. The measurements were done with a manual rain bucket before 22 July, and thereafter with a tip- ping bucket rain gauge (TE525WS, Campbell Scientiﬁc Inc., USA) at a distance of about 50 m from the tower. All mi- crometeorological variables were measured every 10 s and then averaged or summed to the 30 min resolution before be- ing stored on data loggers (CR200X for rainfall, CR3000 for all others, Campbell Scientiﬁc Inc., USA). 2.5.1 Flux calculation Raw data were processed using the EddyPro 4.0.0 software (LI-COR Inc., USA). Processing steps included spike re- moval, tilt correction (double axis rotation), correction for sensor separation, spectral correction, detrending (Reynolds averaging) and ﬂux computation (Burba and Anderson, 2010). Correction for density ﬂuctuations (WPL terms) was not used, however, as LI-7200 is capable of outputting CO2 mixing ratios; that is, thermal expansion and water dilution of the sampled air have already been accounted for (Burba and Anderson, 2010). Half-hourly ﬂuxes were rejected if missing records, removed spikes and absolute limit viola- tions together exceeded 10 % of the total records of any of the three components of wind velocity and/or CO2 concen- tration. CO2 ﬂuxes were also excluded from analyses when turbulent mixing was low during calm nights (friction ve- locity u∗< 0.18 m s−1). The u∗threshold was estimated fol- lowing the ChinaFLUX standard method (Zhu et al., 2006). Half-hourly CO2 ﬂuxes were despiked following Papale et al. (2006). Instrument malfunction, power failure and sensor calibration together led to 3 % missing of half-hourly ﬂux data in 2012, while the data quality control procedure re- jected 26 % of the annual data set. Eighty-seven percent of all the missing and rejected NEE values occurred during night- time. For estimating annual sums of C ﬂuxes, only 7 % of all daytime data needed to be gap-ﬁlled, compared to a pro- portion of 52 % at nighttime. Downward ﬂuxes are counted as negative and upward ﬂuxes as positive. The overall per- formance of the eddy ﬂux measurement system was evalu- ated by the degree of energy balance closure (Li et al., 2005), which was 82 % in 2012 when taking into account the heat stored in the soil above soil heat plates. 2.4 LAI measurements The eddy-covariance instruments were mounted at a height of 6.2 m on a scaffold tower and oriented in the prevailing wind direction (northwest). A 3-D ultrasonic anemometer (CSAT3, Campbell Scientiﬁc Inc., USA) was used to mea- sure ﬂuctuations in wind speed, direction and sonic temper- ature. A closed-path fast response infrared gas analyzer (LI- 7200, LI-COR Inc., USA) was used to measure ﬂuctuations in CO2 and water vapor concentrations. The tube between the air inlet and the Li-7200 was 100 cm; the tube ﬂow rate was 15 L min−1. The tube inlet was situated about 14 cm south of, 16 cm east of, and 8 cm below the anemometer sampling volume. We calibrated the LI-7200 every 3 months, using 99.99 % nitrogen gas to calibrate zeros for both CO2 and water vapor, and a 650 ppm CO2 standard and a dew point generator (LI-610, LI-COR Inc., USA) to calibrate the span for CO2 and water vapor, respectively. A data logger (LI- 7550, LI-COR Inc., USA) was used to store 10 Hz real-time data. The underlying surface of the shrubland was ﬂat and ex- tended over 250 m in all directions. Footprint analysis using the ﬂux source area model (FASM) (Schmid, 1997) showed that > 90 % of the ﬂuxes originated from within 200 m of the tower. The CO2 storage term was not added in estimating NEE because of the short canopy (1.4 m) that usually makes the term negligible (Zhang et al., 2007). In addition, CO2 storage term tends to be close to zero when summed to daily and annual timescales (Baldocchi, 2003). For measuring LAI, we deployed a 4 × 4 grid of 16 quadrats (10 × 10 m each) within a 100 × 100 m plot centered on the ﬂux tower in late March 2012. LAI was measured at roughly weekly intervals. The starting and ending dates for LAI mea- surements were speciﬁed for each species based on pheno- logical observations. The plot-level LAI was calculated as the sum over all component species. The methods for LAI measurements were detailed in the Supplementary Material. 2.1 Study site NEE measurements were made at the Yanchi Research Station (37◦42.51′ N, 107◦13.62′ E; 1530 m a.s.l.), Ningxia, northwest China. The area lies on the southern edge of the Mu Us desert and is characterized by a mid-temperate semi- arid continental monsoon climate. The mean annual temper- ature (1954–2004) is 8.1 ◦C and the frost-free season lasts for 165 days on average (Chen and Duan, 2009; Wang et al., 2014). The mean annual precipitation is 287 mm, 62 % of which falls from July to September (Feng et al., 2013). The mean annual potential evapotranspiration is 2024 mm. The soil is sandy and has a bulk density of 1.54 ± 0.08 g cm−3 (mean ± standard deviation (SD), n = 16) in the upper 10 cm of the soil proﬁle. The area is dominated by a mixture of deciduous shrub species including Artemisia ordosica, Hedysarum mongolicum and Hedysarum scoparium, and also has sparsely distributed patches of Salix psammophila and Agropyron cristatum. The canopy height is about 1.4 m. Water deﬁcit is a limiting factor for plant photosynthesis and soil respiration (Rs) at the study site (Feng et al., 2013; Wang et al., 2014). China is one of the most threatened countries by deser- tiﬁcation, with the total desertiﬁed area (2.62 million km2) accounting for 27 % of the national land area (Yang et al., 2005). Extensive revegetation and conservation practices have been carried out in northern China (Li et al., 2004). However, little has been done to quantify the C sequestra- tion potential of the recovering vegetation (Gao et al., 2012). www.biogeosciences.net/11/4679/2014/ Biogeosciences, 11, 4679–4693, 2014 4681 www.biogeosciences.net/11/4679/2014/ X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4682 X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland Figure 1. Seasonal variations of air temperature (Ta) at 6 m above ground and soil temperature (Ts) at 10 cm depth (a), incident photosyn- thetically active radiation (PAR) (b), vapor pressure deﬁcit (VPD) (C), soil water content (SWC) and rainfall (d), leaf area index (LAI) (e) and CO2 ﬂuxes (f) in 2012. NEE, net ecosystem CO2 exchange; GEP, gross ecosystem productivity; Re, ecosystem respiration. Daily means are shown for Ta, Ts, VPD and SWC; Daily sums are shown for PAR, rainfall and CO2 ﬂuxes; interpolated daily values are shown for LAI. The vertical dashed lines separate each month, and the horizontal dashed lines in (a) and (f) represent y = 0. Figure 1. Seasonal variations of air temperature (Ta) at 6 m above ground and soil temperature (Ts) at 10 cm depth (a), incident photosyn- thetically active radiation (PAR) (b), vapor pressure deﬁcit (VPD) (C), soil water content (SWC) and rainfall (d), leaf area index (LAI) (e) and CO2 ﬂuxes (f) in 2012. NEE, net ecosystem CO2 exchange; GEP, gross ecosystem productivity; Re, ecosystem respiration. Daily means are shown for Ta, Ts, VPD and SWC; Daily sums are shown for PAR, rainfall and CO2 ﬂuxes; interpolated daily values are shown for LAI. The vertical dashed lines separate each month, and the horizontal dashed lines in (a) and (f) represent y = 0. used to estimate missing daytime data because net CO2 uptake declined at high PAR, especially in summer (Fig. 2). In Eqs. 2–6, Qc is the PAR intensity at the maximum rate of net CO2 uptake (NEEmax), ϕ0 and ϕc are the quantum yield at Q = 0 and Q = Qc, respectively, and Rd is the model- derived bulk ecosystem respiration. For the estimation of missing NEEday, Eq. (1) was ﬁt to consecutive windows of 500 non-missing daytime data points to obtain seasonally varying parameter values. NEEday = α 1 −βQ 1 + γ Q(Q −Qc), (1) (1) where NEEday is daytime NEE (µmol CO2 m−2 s−1), Q is incident PAR in units of µmol m−2 s−1, Qc is the light com- pensation point, and α, β, and γ are ﬁt values for the follow- ing calculations (Ye, 2007; see also Appendix A for a list of parameters and abbreviations). 2.5.2 Gap-ﬁlling and partitioning NEE into GEP and Re Linear interpolation was used to ﬁll small gaps (≤2 h). For larger gaps during daytime (i.e., PAR ≥5 µmol photons m−2 s−1), NEE–PAR relationships were used for gap-ﬁlling. A light response model (Eq. 1) which was modiﬁed from the rectangular hyperbola to incorporate photoinhibition at high radiation (Ye, 2007) was www.biogeosciences.net/11/4679/2014/ Biogeosciences, 11, 4679–4693, 2014 X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland where NEEday is daytime NEE (µmol CO2 m−2 s−1), Q is incident PAR in units of µmol m−2 s−1, Qc is the light com- pensation point, and α, β, and γ are ﬁt values for the follow- ing calculations (Ye, 2007; see also Appendix A for a list of parameters and abbreviations). The Q10 model was used for ﬁlling nighttime gaps (Zha et al., 2004): NEEnight = Re10Q(Ts−10)/10 10 , (7) (7) Qm = √(β + γ )(1 + γ Qc)/β −1 γ (2) NEEmax = α 1 −βQm 1 + γ Qm (Qm −Qc) (3) φ0 = NEE′ day (Q = 0) = α[1 + (γ + β)Qc] (4) φc = NEE′ day (Qc) = α 1 + (γ −β)Qc −βγ Q2 c (1 + γ Qc)2 (5) Rd = NEEday (Q = 0) = −αQc (6) (2) where NEEnight is nighttime NEE, Ts the soil temperature at 10 cm depth, Re10 the Re at Ts = 10◦C, Q10 the temperature sensitivity of ecosystem respiration. Equation (7) was only ﬁt to the annual data set because short-term data points were too scattered to establish any valid NEEnight−Ts relationships. Ts at 10 cm depth was selected because it produced a higher co- efﬁcient of determination (R2) than Ts at other depths and Ts. In order to estimate annual CO2 ﬂuxes, missing Ts val- ues were gap-ﬁlled with the mean diurnal variation (MDV) (6) Biogeosciences, 11, 4679–4693, 2014 www.biogeosciences.net/11/4679/2014/ Biogeosciences, 11, 4679–4693, 2014 X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4683 Figure 2. Daytime net ecosystem CO2 exchange (NEEday) as a function of incident photosynthetically active radiation (PAR) for each month from May to October in 2012. Half-hourly NEEday was bin-averaged into 50 µmol photons m−2 s−1 PAR increments. Bars indicate standard errors. Light response curves were ﬁt with Eq. (1). The dashed lines represent y = 0. Figure 2. Daytime net ecosystem CO2 exchange (NEEday) as a function of incident photosynthetically active radiation (PAR) for each month from May to October in 2012. Half-hourly NEEday was bin-averaged into 50 µmol photons m−2 s−1 PAR increments. Bars indicate standard errors. Light response curves were ﬁt with Eq. (1). The dashed lines represent y = 0. method (Moffat et al., 2007), while missing PAR values were gap-ﬁlled using an empirical relationship to half-hourly PAR data from a meteorological tower about 3 km east. and a stepwise multiple linear regression against VPD, Ta and SWC (Z-transformed data were used in the step- wise regression). In order to test NEEday–Ta and NEEday– VPD relationships, as well as their dependence on SWC, NEEday was compiled with respect to SWC at 30 cm depth (SWC ≤0.1m3 m−3, SWC > 0.1m3 m−3), and then bin- averaged into 1 ◦C Ta and 0.2 kPa VPD intervals, respec- tively. NEEday–Ta and NEEday–VPD relationships were ﬁt with the quadratic model. Daytime Re during the growing season was extrapolated from the temperature response function for NEEnight (Eq. 7). Off-season Re was considered as 24 h NEE ﬂuxes. GEP was estimated as GEP = Re −NEE. (8) (8) GEP = Re −NEE. q For examining the effects of SWC on the Re–Ts (10 cm depth) relationship, we classiﬁed NEEnight when Ts > 0◦C into two groups with respect to SWC at 30 m depth (SWC ≤ 0.1m3 m−3, SWC > 0.1m3 m−3), and then bin-averaged NEEnight into 1 ◦C Ts intervals. Re10 and Q10 in Eq. (7) were estimated separately for each SWC group. A minimum of 10 data points were required for a valid bin for all abovemen- tioned bin averages. The following surface ﬁtting (Ts–REW model) was then used to further examine the interaction be- tween temperature and water availability in regulating half- hourly NEEnight: Biogeosciences, 11, 4679–4693, 2014 www.biogeosciences.net/11/4679/2014/ www.biogeosciences.net/11/4679/2014/ www.biogeosciences.net/11/4679/2014/ NEEnight = (a + bREW)(c + dREW)(Ts−10)/10, (9) (9) where a, b, c and d are ﬁt parameters (Re10 = a + b REW; Q10 = c + d REW). REW means relative extractable water content (Zhou et al., 2013), which was calculated as REW = SWC −SWCmin SWCmax −SWCmin , (10) REW = SWC −SWCmin SWCmax −SWCmin , (10) (10) where SWCmax and SWCmin are the minimum and maximum SWC during the period of Ts > 0◦C, respectively. Linear re- gressions were used to compare measured vs. predicted half- hourly NEEnight values and to examine the dependence of 2.5.3 Statistical analysis Equation (1) was ﬁt monthly from May to October to evaluate the seasonal variation in light response of NEE. The regressions were conducted on bin-averaged data us- ing 50 µmol m−2 s−1 PAR intervals. In order to test the de- pendency of the NEEday–PAR relationship on abiotic factors and exclude the confounding effects of plant phenology, we compiled NEEday during the peak growing season (June– August) into multiple groups according to VPD (VPD ≤ 1 kPa, 1kPa < VPD ≤2kPa, VPD > 2kPa), Ta (Ta ≤20◦C, 20◦C < Ta ≤25◦C, Ta > 25◦C) and SWC at 30 cm depth (SWC ≤0.1m3 m−3, SWC > 0.1m3 m−3). The NEEday val- ues were then bin-averaged before parameters were ﬁt for each group. These threshold values were chosen to most clearly show the differences between levels, and to avoid having too few data points in a certain group. In addi- tion, the values were equal or close to those used by pre- vious studies in dryland areas. To evaluate the relative im- portance of different abiotic factors, Eq. (1) was ﬁt to all half-hourly NEEday values during June–August, and the residuals were then subjected to least-square regressions NEEnight = (a + bREW)(c + dREW)(Ts−10)/10, (9) X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4684 X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland Figure 3. Daytime net ecosystem CO2 exchange (NEEday) as a function of incident photosynthetically active radiation (PAR) under different environmental conditions (a–c) and the relationships between residuals of the light response curve and environmental factors (d–f). Only data from the peak growing season (June–August) were used to minimize the confounding effects of phenology. Half-hourly NEEday was bin-averaged into 50 µmol photons m−2 s−1 PAR increments in (a)–(c). Bars indicate standard errors. Light response curves were ﬁt with Eq. (1). The horizontal dashed lines represent y = 0. Figure 3. Daytime net ecosystem CO2 exchange (NEEday) as a function of incident photosynthetically active radiation (PAR) under different environmental conditions (a–c) and the relationships between residuals of the light response curve and environmental factors (d–f). Only data from the peak growing season (June–August) were used to minimize the confounding effects of phenology. Half-hourly NEEday was bin-averaged into 50 µmol photons m−2 s−1 PAR increments in (a)–(c). Bars indicate standard errors. Light response curves were ﬁt with Eq. (1). The horizontal dashed lines represent y = 0. NEEnight residuals on REW for both the Q10 model (Eq. 7) and the Ts–REW model (Eqs. 9–10). SWC at 30 cm depth was chosen for both daytime and nighttime analyses because its effects were most pronounced among SWC at different layers. gap-ﬁlling and ﬂux partitioning. Again, the 90 % prediction limits of all (N = 2000) simulated annual GEP and Re val- ues were calculated. The resulting GEP and Re uncertainties encompass sources from both measurement error and model parameterization (Hagen et al., 2006). Linear regressions were used to examine the seasonal re- lationships between GEP and Re, and the responses of NEE and GEP to seasonal changes in LAI. For this purpose, daily- integrated values were calculated for C ﬂuxes, and daily LAI was derived by linear interpolation between measurements. www.biogeosciences.net/11/4679/2014/ www.biogeosciences.net/11/4679/2014/ Biogeosciences, 11, 4679–4693, 2014 3.2 Seasonal variation in NEE and its biophysical controls Daily NEE ranged from −4.71 g C m−2 day−1 (largest net CO2 uptake) on 30 June to 1.63 g C m−2 day−1 on 30 July (Fig. 1f). GEP reached a maximum of 6.78 g C m−2 day−1 on 30 June. Maximum Re was 3.26 g C m−2 day−1 on 25 July. Annual net ecosystem productivity (NEP = −NEE) was 77 g C m−2 yr−1 (with a 90 % prediction interval of 68– 87 g C m−2 yr−1). Re contributed 379 g C m−2 yr−1 (with a 90 % prediction interval of 370–389 g C m−2 yr−1) to NEP, leading to an annual GEP of 456 g C m−2 yr−1 (with a 90 % prediction interval of 449–463 g C m−2 yr−1). p g y PAR had strong inﬂuences on NEEday during the grow- ing season (Fig. 2), accounting for > 80 % of the variabil- ity in NEEday in most months (except for October) (Table 1). A third-order polynomial pattern was observed for the NEEmax–PAR relationship in summer months (Figs. 2b, c and 3). The absolute values of NEEmax, ϕ0, ϕc and Rd were all highest in July, while lower in spring and autumn (Table 1). The effect of PAR was modiﬁed by other environmen- tal factors (Table 2; Fig. 3). The magnitude of NEEmax and ϕc decreased, whereas Qc increased, with increasing VPD and Ta. In addition, Rd increased with Ta. The magnitude of NEEmax, Qc and Rd were all lower under dry soil condi- tions. The NEEday residuals were positively correlated with VPD and Ta, and negatively correlated with SWC (Fig. 3d– f). The stepwise regression produced the following relation- ship: residual = −0.30 SWC + 0.17 Ta + 0.11 VPD (R2 = 0.16, P < 0.01). The responses of GEP to PAR resembled those of NEEday. In addition, the effects of VPD, Ta and SWC on the GEP–PAR relationship were similar to those on the NEEday–PAR relationship (data not shown). Figure 4. Responses of daytime net ecosystem CO2 exchange (NEEday) and gross ecosystem productivity (GEP) to air temper- ature (Ta) and vapor pressure deﬁcit (VPD) under different soil water contents (SWC) measured at 30 cm depth. Only data from the peak growing season (June–August) were used to minimize the confounding effects of phenology. Half-hourly NEEday was bin- averaged into 1 ◦C Ts (a and c) and 0.2 kPa VPD (b and d) inter- vals, respectively. X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4685 deeper layers (30 and 70 cm) only responded to large rainfall events (Fig. 1d). Figure 4. Responses of daytime net ecosystem CO2 exchange (NEEday) and gross ecosystem productivity (GEP) to air temper- ature (Ta) and vapor pressure deﬁcit (VPD) under different soil water contents (SWC) measured at 30 cm depth. Only data from the peak growing season (June–August) were used to minimize the confounding effects of phenology. Half-hourly NEEday was bin- averaged into 1 ◦C Ts (a and c) and 0.2 kPa VPD (b and d) inter- vals, respectively. Bars indicate standard errors. Response curves were ﬁt using the quadratic model. The horizontal dashed lines in (a) and (b) represent y = 0. 3.2 Seasonal variation in NEE and its biophysical controls Bars indicate standard errors. Response curves were ﬁt using the quadratic model. The horizontal dashed lines in (a) and (b) represent y = 0. correlation with REW (Fig. 7b). However, the pattern disap- peared after incorporating REW into the model (Fig. 7d). There was a linear dependence of daily Re on GEP (R2 = 0.65), with a slope of 0.34 (Fig. 8a). Both daily GEP and NEE responded linearly to the seasonal variation of LAI (R2 = 0.65 and 0.45, respectively), with a slope of 4.12 for GEP, and −2.03 for NEE (Fig. 8b and c). y NEEday ﬁrst decreased (towards higher CO2 uptake), and then increased, with increasing Ta and VPD (Fig. 4a and b). Moreover, NEEday was more responsive to Ta and VPD, and showed higher maximum CO2 uptake rates under wetter soil conditions (SWC ≥0.1m3 m−3). For most of the VPD range, NEEday was more negative under higher soil water avail- ability (Fig. 4b). Responses of GEP to Ta and VPD were also consistent with those of NEEday (Fig. 4c and d), indi- cating that environmental controls on NEEday were largely attributable to photosynthetic responses. 3.1 Seasonal variation in environmental conditions We evaluated the cumulative effect of random measure- ment uncertainty on annual NEE with the so-called succes- sive days approach (Hollinger and Richardson, 2005; Drag- oni et al., 2007). This method infers the statistical properties of the random error from the difference between half-hourly NEE measurements made exactly 24 h apart. The effects of imperfect environmental similarity between the successive days were controlled for following Dragoni et al. (2007). A Monte Carlo approach was then used to generate a ran- dom error for each measured half-hourly NEE. The simu- lation was repeated 2000 times, and the uncertainty of the measured annual NEE was estimated by calculating the 90 % prediction limits of all simulated annual NEE values. Simi- larly, we evaluated the random uncertainty for annual GEP and Re following a Monte Carlo algorithm detailed by Ha- gen et al. (2006). The algorithm infers the statistical proper- ties of the random error from the residuals of the model for Environmental variables showed clear seasonal patterns (Fig. 1). The daily mean Ta ranged from −8.5 on 23 January to 23.4 ◦C on 11 July (Fig. 1a). Ts had a min- imum of −12.1 ◦C on 8 February and a maximum of 25.8 ◦C on 22 June (Fig. 1a). The daily PAR reached a maximum of 61.5 mol m−2 day−1 on 15 June, and was < 30 mol m−2 day−1 during winter (Fig. 1b). Daily mean VPD reached a maximum of 2.2 kPa on 9 June, and was lower than < 0.5 kPa during winter (Fig. 1c). Rainfall summed to 304.9 mm from mid-May to December, > 60 % of which fell between June and August. There were three rain events larger than 20 mm day−1, among which the largest occurred on 27 June (49.8 mm day−1) (Fig. 1d). Snowmelt and soil thaw in early spring resulted in a relatively wet soil (Fig. 1d). During the growing season, SWC (except for that at 120 cm depth) followed the pattern of rainfall, and SWC in www.biogeosciences.net/11/4679/2014/ Biogeosciences, 11, 4679–4693, 2014 3.3 Synoptic variation in NEE as related to rain pulses Pulses of NEE were observed during the growing season when rainfall occurred (Fig. 9a). In order to examine the effects of rain events and related environmental factors on NEE, half-hourly measurements around the largest rainfall event (DOY 179–180, 61 mm) were scrutinized (Fig. 9b and c). NEE during daytime was markedly depressed on DOY 179–180 compared to the day before rain (DOY 178), and NEE during nighttime was slightly higher on DOY 179– 180 than on days without rain (DOY 181–184). As a result, a positive pulse in daily NEE was observed on DOY 179– 180 (Fig. 9a). The positive NEE pulse was accompanied by a sharp increase in SWC but decreases in Ta, Ts and PAR (Fig. 9b and c). Daytime NEE was dramatically stimulated by high PAR, temperatures and SWC on days immediately NEEnight related positively with Ts at 10 cm depth for both SWC groups (Fig. 5). However, Q10 was much larger, with Re10 slightly smaller, for the higher SWC group. The sur- face ﬁtting showed that Q10 increased from 1.9 to 3.2, and Re10 increased with from 0.73 to 0.83 µmol CO2 m−2 s−1, as REW increased from 0 to 1 (Fig. 6). The Ts–REW model ﬁt the measured half-hourly values better than the Q10 model (Fig. 7). Half-hourly NEEnight residuals showed a positive www.biogeosciences.net/11/4679/2014/ NEE (µmol CO2 m−2 s−1), net ecosystem CO2 exchange; PAR (µmol photons m−2 s−1), photosynthetically active radiation; Qc (µmol photons m−2 s−1), light compensation point; NEEmax (µmol CO2 m−2 s−1), the maximum magnitude of daytime NEE; Qc (µmol photons m−2 s−1), the PAR intensity corresponding to NEEmax; ϕ0 and ϕc (µmol CO2 µmol photons−1), the quantum yield when PAR is equal to zero and Qc, respectively; Rd (µmol CO2 m−2 s−1), model-derived bulk ecosystem respiration; Adj. R2, adjusted coefﬁcient of determination. Table 2. Parameter values describing the response of daytime NEE to incident PAR during the peak growing season (June–August) under different environmental conditions. Table 2. Parameter values describing the response of daytime NEE to incident PAR during the peak gro different environmental conditions. escribing the response of daytime NEE to incident PAR during the peak growing season (June–August) und ditions. Table 2. Parameter values describing the response of daytime NEE to incident PAR during the peak growing season (June–August) under different environmental conditions. Treatment α(× 103) β(× 104) γ (× 104) Qc Qm NEEmax ϕ0 ϕc Rd Adj. R2 VPD ≤1 −14.46 ± 1.58 5.18 ± 0.47 −2.58 ± 1.46 152.80 ± 17.76 1184.57 −8.30 −0.0150 −0.0139 2.21 0.96 1 < VPD ≤2 −15.33 ± 2.84 2.61 ± 0.55 6.28 ± 4.07 150.00 ± 20.32 1482.84 −6.49 −0.0174 −0.0135 2.30 0.90 VPD > 2 −13.86 ± 6.81 0.89 ± 1.49 17.85 ± 17.80 150.20 ± 33.65 2332.48 −4.64 −0.0178 −0.0108 2.08 0.80 Ta ≤20 −16.10 ± 2.47 3.91 ± 0.63 1.48 ± 2.84 146.88 ± 19.01 1260.76 −7.65 −0.0174 −0.0149 2.36 0.94 20 < Ta ≤25 −17.35 ± 3.27 2.99 ± 0.49 6.42 ± 4.11 150.83 ± 20.22 1334.66 −6.64 −0.0198 −0.0151 2.62 0.89 Ta > 25 −13.96 ± 8.36 1.77 ± 1.17 12.00 ± 15.1 208.20 ± 50.47 1763.11 −4.79 −0.0180 −0.0108 2.91 0.79 SWC ≤0.1 −17.63 ± 3.13 3.06 ± 0.36 9.62 ± 4.39 129.53 ± 17.06 1204.58 −5.54 −0.0205 −0.0151 2.28 0.88 SWC > 0.1 −16.61 ± 2.38 2.84 ± 0.48 4.05 ± 2.84 177.30 ± 17.22 1512.46 −7.85 −0.0186 −0.0147 2.94 0.94 Fit parameters in Eq. (1) (α,β,γ and Qc) are presented as mean ± SE. www.biogeosciences.net/11/4679/2014/ NEE (µmol CO2 m−2 s−1), net ecosystem CO2 exchange; PAR (µmol photons m−2 s−1), photosynthetically active radiation; Qc (µmol photons m−2 s−1), light compensation point; NEEmax (µmol CO2 m−2 s−1), the maximum magnitude of daytime NEE; Qc (µmol photons m−2 s−1), the PAR intensity corresponding to NEEmax; ϕ0 and ϕc (µmol CO2 µmol photons−1), the quantum yield when PAR is equal to zero and Qc, respectively; Rd (µmol CO2 m−2 s−1), model-derived bulk ecosystem respiration; Adj. R2, adjusted coefﬁcient of determination; VPD (kPa), vapor pressure deﬁcit; Ta (°C), air temperature at 6 m above ground, SWC (m3 m−3), soil water content at 30 cm depth. photosynthetically active radiation; Qc (µmol photons m−2 s−1), light compensation point; NEEmax (µmol CO2 m−2 s−1), the maximum magnitude of daytime NEE; Qc (µmol photons m−2 s−1), the PAR intensity corresponding to NEEmax; ϕ0 and ϕc (µmol CO2 µmol photons−1), the quantum yield when PAR is equal to zero and Qc, respectively; Rd (µmol CO2 m−2 s−1), model-derived bulk ecosystem respiration; Adj. R2, adjusted coefﬁcient of determination; VPD (kPa), vapor pressure deﬁcit; Ta (°C), air temperature at 6 m above ground, SWC (m3 m−3), soil water content at 30 cm depth. (µmol photons m 2 s 1), the PAR intensity corresponding to NEEmax; ϕ0 and ϕc (µmol CO2 µmol photons 1), the quantum yield when PAR is equal to zero and Qc, respectively; Rd (µmol CO2 m−2 s−1), model-derived bulk ecosystem respiration; Adj. R2, adjusted coefﬁcient of determination; VPD (kPa), vapor pressure deﬁcit; Ta (°C), air temperature at 6 m above ground, SWC (m3 m−3), soil water content at 30 cm depth. following the rain event (DOY 181–182), leading to a clear post-rain peak in CO2 uptake. www.biogeosciences.net/11/4679/2014/ www.biogeosciences.net/11/4679/2014/ Biogeosciences, 11, 4679–4693, 2014 X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4686 Table 1. Parameter values describing the response of daytime NEE to incident PAR for each month and the entire growing season (GS). Month α(± 103) β (± 104) γ (± 104) Qc Qm NEEmax ϕ0 ϕc Rd Adj. R2 May −6.59 ± 1.65 4.01 ± 0.61 −0.04 ± 3.14 186.94 ± 36.34 1344.19 −3.54 −0.0071 −0.0061 1.23 0.82 Jun −14.87 ± 2.18 3.04 ± 0.37 5.61 ± 2.97 141.22 ± 16.28 1340.05 −6.03 −0.0167 −0.0132 2.10 0.93 Jul −17.67 ± 2.56 3.12 ± 0.48 3.96 ± 2.86 154.99 ± 17.23 1395.40 −7.98 −0.0196 −0.0158 2.74 0.94 Aug −15.71 ± 3.01 3.69 ± 0.39 7.83 ± 4.47 146.00 ± 18.69 1105.94 −4.79 −0.0184 −0.0133 2.29 0.86 Sep −7.33 ± 1.92 2.31 ± 1.79 6.23 ± 8.04 159.79 ± 23.65 1629.61 −3.33 −0.0083 −0.0064 1.17 0.90 Oct −1.59 ± 0.93 6.59 ± 2.46 3.92 ± 18.90 324.98 ± 86.41 869.97 −0.28 −0.0021 −0.0011 0.52 0.36 GS −11.80 ± 1.41 1.10 ± 0.58 12.20 ± 4.08 150.06 ± 9.91 2281.00 −4.98 −0.0142 −0.0098 1.77 0.96 Fit parameters in Eq. (1) (α,β,γ and Qc) are presented as mean ± SE. NEE (µmol CO2 m−2 s−1), net ecosystem CO2 exchange; PAR (µmol photons m−2 s−1), photosynthetically active radiation; Qc (µmol photons m−2 s−1), light compensation point; NEEmax (µmol CO2 m−2 s−1), the maximum magnitude of daytime NEE; Qc (µmol photons m−2 s−1), the PAR intensity corresponding to NEEmax; ϕ0 and ϕc (µmol CO2 µmol photons−1), the quantum yield when PAR is equal to zero and Qc, respectively; Rd (µmol CO2 m−2 s−1), model-derived bulk ecosystem respiration; Adj. R2, adjusted coefﬁcient of determination. Table 1. Parameter values describing the response of daytime NEE to incident PAR for each month and the entire growing season (GS). Table 1. Parameter values describing the response of daytime NEE to incident PAR for each month and the e Fit parameters in Eq. (1) (α,β,γ and Qc) are presented as mean ± SE. 3.4 Diurnal variations in NEE and meteorological factors Dryland ecosystems can vary from a net sink to source of CO2 on an annual basis, as mainly determined by the vari- ations in water availability and the amount of primary pro- ducers (Liu et al., 2012). The total amount of C sequestered by the studied shrubland in 2012 (NEP = 77 g C m−2 yr−1), with an annual rainfall of at least 305 mm and a peak LAI of 1.2, was generally lower than that sequestered by forests and grasslands in humid and subhumid areas (e.g., Suyker and Verma, 2001; Zha et al., 2004; Zhou et al., 2013). How- ever, the NEP was higher than many reported values from semiarid and arid nonforest ecosystems (Wang et al., 2008; Gao et al., 2012). For example, a revegetated shrub ecosys- tem ∼200 km west of our site dominated by Caragana kor- shinskii and A. ordosica had an NEP of 14–23 g C m−2 yr−1, with an annual precipitation of < 150 mm (Gao et al., 2012). A semiarid steppe in central Mongolia showed an NEP of 41 g C m−2 yr−1, an annual precipitation of 260 mm and a peak LAI of 0.57 (Li et al., 2005). Liu et al. (2012) reported that a salt desert shrubland (with LAI < 0.4) in northwest The monthly mean diurnal variations of non-rainy days showed that the diel amplitude of NEE varied dramatically over the growing season, being largest in July and smallest in October (Fig. 10a). NEE during nighttime was relatively low in magnitude in comparison to that during daytime in most months (except for October). The ecosystem was a C sink on non-rainy days from May to September. Net CO2 uptake peaked before noon (at 09:30–10:00 LST, LST = GMT + 8) on summer (June–August) days, leading to an asymmetric distribution of NEE around noon. PAR did not show such an asymmetry, being highest between 12:30 and 13:30 LST (GMT + 8) (Fig. 10b). Both Ta and VPD were lowest in early morning, and peaked in late afternoon (Fig. 10c and d). Biogeosciences, 11, 4679–4693, 2014 www.biogeosciences.net/11/4679/2014/ X. Jia et al.: Biophysical controls on NEE over a semiarid Figure 5. Nighttime net ecosystem CO2 exchange (NEEnight) as a function of soil temperature (Ts) at 10 cm depth under different soil water contents (SWC) measured at 30 cm depth. Only data when Ts >0 ◦C were used. 3.4 Diurnal variations in NEE and meteorological factors At our site, precipitation in 2012 was at least 18 mm higher than the long-term mean, which calls for future studies that assess how semiarid shrub ecosystems re- spond to interannual variability in water availability. interacted to reduce CO2 uptake under water-stressed condi- tions (Fig. 4b and d). Li et al. (2005) also suggested that the sensitivity of stomata to VPD becomes stronger once leaf wa- ter potential starts to drop because of the deﬁciency of water supply from the soil. Low soil water availability may aggra- vate VPD-induced stomatal closure. Water limitation of Re or Rs has been found in various types of ecosystems (Gao et al., 2012). Our results showed a marked decrease in both the magnitude and temperature sen- sitivity (Q10) of Re under low SWC (Figs. 5 and 6). Wang et al. (2014) measured Rs in a nearby community dominated by A. ordosica (800 m north of the eddy-ﬂux tower). They found that Rs was closely correlated with Ts when SWC at 10 cm depth was higher than 0.08 m3 m−3, whereas Rs was decoupled from temperature during dry periods. The re- duction in Q10 of Re under drought conditions was most likely associated with decreased C transportation to roots due to suppressed photosynthesis, deactivated rhizosphere and switched C pool being respired (e.g., from labile to recal- citrant) (Zhang et al., 2007; Wang et al., 2008; Gao et al., 2012). Our result that the Ts–REW model over-performed the Ts-only model (Fig. 7) indicated the need to take water avail- ability into account when modeling short-term (e.g., hourly) changes of respiration in dryland ecosystems. 3.4 Diurnal variations in NEE and meteorological factors Half-hourly NEEnight was bin-averaged into 1 ◦C Ts intervals. Error bars indicate standard errors. X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4687 Figure 5. Nighttime net ecosystem CO2 exchange (NEEnight) as a function of soil temperature (Ts) at 10 cm depth under different soil water contents (SWC) measured at 30 cm depth. Only data when Ts >0 ◦C were used. Half-hourly NEEnight was bin-averaged into 1 ◦C Ts intervals. Error bars indicate standard errors. Figure 6. Nighttime net ecosystem CO2 exchange (NEEnight) as a function of soil temperature (Ts) at 10 cm depth and relative ex- tractable water content (REW, see Eq. 10). Only data when Ts >0 ◦C were used. X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4687 J p y Figure 6. Nighttime net ecosystem CO2 exchange (NEEnight) as a function of soil temperature (Ts) at 10 cm depth and relative ex- tractable water content (REW, see Eq. 10). Only data when Ts >0 ◦C were used. Figure 5. Nighttime net ecosystem CO2 exchange (NEEnight) as a function of soil temperature (Ts) at 10 cm depth under different soil water contents (SWC) measured at 30 cm depth. Only data when Ts >0 ◦C were used. Half-hourly NEEnight was bin-averaged into 1 ◦C Ts intervals. Error bars indicate standard errors. Figure 6. Nighttime net ecosystem CO2 exchange (NEEnight) as a function of soil temperature (Ts) at 10 cm depth and relative ex- tractable water content (REW, see Eq. 10). Only data when Ts >0 ◦C were used. China was near C neutral in a dry year (17 mm annual precip- itation) but a C sink of 40 g C m−2 yr−1 in a wet year (48 mm annual precipitation). The shrub ecosystem we studied has a wide regional distribution, and it thus represents a huge C ﬁxation potential. At our site, precipitation in 2012 was at least 18 mm higher than the long-term mean, which calls for future studies that assess how semiarid shrub ecosystems re- spond to interannual variability in water availability. China was near C neutral in a dry year (17 mm annual precip- itation) but a C sink of 40 g C m−2 yr−1 in a wet year (48 mm annual precipitation). The shrub ecosystem we studied has a wide regional distribution, and it thus represents a huge C ﬁxation potential. 4.2 Effects of abiotic stresses on NEE The nature and relative importance of different biophysical factors in controlling NEE differ among ecosystems (Fu et al., 2006). Water stress, which varies signiﬁcantly at the sea- sonal and interannual scales, is the most common limitation to vegetation growth in dryland ecosystems (Fu et al., 2006). Our study revealed the dominant role of low soil water con- tent in limiting photosynthesis, which was also observed by Fu et al. (2006) in a semiarid steppe. However, they found that high temperature and strong solar radiation together con- tributed to a decrease of NEEday in an alpine shrub. Our ﬁndings that low SWC and high VPD depressed the maxi- mum rate of CO2 uptake (NEEmax) (Table 2; Fig. 3) are in agreement with previous studies in dryland ecosystems (Li et al., 2005; Wang et al., 2008; Yang et al., 2011). Both SWC and VPD affect plant hydraulic status; however, they reduce C assimilation through different mechanisms. Dry soil leads to reduced water supply for metabolism and cell expansion, while VPD affects CO2 supply for photosynthesis by regulat- ing stomatal conductance and evaporative demand (Zhou et al., 2013). These two mechanisms did not act in isolation but Air temperature is another factor affecting VPD besides humidity, which may explain the reduction in NEEmax at high temperatures (Fig. 3b). Alternatively, high Ta might have suppressed photosystem II, resulting in smaller NEEmax. The quadratic relationship between NEEday or GEP and Ta (Fig. 4a and c) was likely because C assimilation was lim- ited by low PAR and temperature on cloudy and rainy days, www.biogeosciences.net/11/4679/2014/ Biogeosciences, 11, 4679–4693, 2014 4688 X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland Figure 7. Comparisons of measured and predicted half-hourly nighttime net ecosystem CO2 exchange (NEEnight) using the Q10 model (a) and the Ts–REW model (c); relationships between NEEnight residuals and REW for the Q10 model (b) and the Ts–REW model (d). REW means the relative extractable water content (see Eq. 10). Only data when Ts >0 ◦C were used. The dashed lines in (a) and (c) represent y = x, and those in (b) and (d) represent y = 0. The ﬁt line in (a): y = 1.08x−0.14, R2 = 0.33, P < 0.01; the ﬁt line in (c): y = 1.06x−0.11, R2 = 0.39, P < 0.01. X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4688 X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland Figure 7. Comparisons of measured and predicted half-hourly nighttime net ecosystem CO2 exchange (NEEnight) using the Q10 model (a) and the Ts–REW model (c); relationships between NEEnight residuals and REW for the Q10 model (b) and the Ts–REW model (d). REW means the relative extractable water content (see Eq. 10). Only data when Ts >0 ◦C were used. The dashed lines in (a) and (c) represent y = x, and those in (b) and (d) represent y = 0. The ﬁt line in (a): y = 1.08x−0.14, R2 = 0.33, P < 0.01; the ﬁt line in (c): y = 1.06x−0.11, R2 = 0.39, P < 0.01. The slope and intercept as well as their 95 % conﬁdence intervals (CI) are 1.27 (1.04, 1.50) and −0.40 (−0.48, −0.32) for the relationship between REW and the NEEnight residuals from the Q10 model (b), and are −0.21 (−0.43, 0.02) and 0.03 (−0.05, 0.11) for the relationship between REW and the NEEnight residuals from the Ts–REW model (d). Figure 8. Relationship between gross ecosystem productivity (GEP) and ecosystem respiration (Re) (a), and between leaf area index (LAI) and GEP (b) and net ecosystem CO2 exchange (NEE) (c). Daily values are shown for the growing period when LAI > 0. The horizontal dashed line in (c) represents y = 0. Figure 8. Relationship between gross ecosystem productivity (GEP) and ecosystem respiration (Re) (a), and between leaf area index (LAI) and GEP (b) and net ecosystem CO2 exchange (NEE) (c). Daily values are shown for the growing period when LAI > 0. The horizontal dashed line in (c) represents y = 0. whereas limited by heat and water stresses on clear days (Fu et al., 2006). Similar to Li et al. (2005), we found that NEEday became less responsive to Ta under drought con- ditions (Fig. 4a), reﬂecting drought limitations to plant ac- tivities. The asymmetric distribution of net photosynthesis around noon, which we observed (Fig. 10a), is common in arid areas (Zhang et al., 2007). It can be ascribed to both VPD-induced stomatal closure and temperature-induced in- creases in Re in the afternoon (Fig. 10c and d). similation, i.e., insufﬁcient thermal dissipation of leaves and consequent damage to photosynthetic apparatus under exces- sive light (Fu et al., 2006). 4.2 Effects of abiotic stresses on NEE The slope and intercept as well as their 95 % conﬁdence intervals (CI) are 1.27 (1.04, 1.50) and −0.40 (−0.48, −0.32) for the relationship between REW and the NEEnight residuals from the Q10 model (b), and are −0.21 (−0.43, 0.02) and 0.03 (−0.05, 0.11) for the relationship between REW and the NEEnight residuals from the Ts–REW model (d). X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland Alternatively, high VPD and tem- perature may have depressed CO2 uptake under high PAR. The third-order polynomial pattern in the light response of NEEday (Figs. 2b, c and 3) was unexpected yet interesting. We propose that it may be related to confounding factors such as VPD and temperature. Although VPD and temper- ature covaried with PAR at the diurnal scale, they lagged PAR by 3–4 h (Fig. 10b–d). Therefore, their depression ef- fects on CO2 assimilation could be strongest when PAR is below its daily maximum. A detailed understanding of how In our study, the CO2 uptake decreased under strong so- lar radiation in summer months (Figs. 2 and 3). One pos- sible explanation to this is the photoinhibition of CO2 as- www.biogeosciences.net/11/4679/2014/ Biogeosciences, 11, 4679–4693, 2014 X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4689 Figure 9. Pulses in daily net ecosystem CO2 exchange (NEE) as related to rainfall events during the growing season (a), and synoptic variations of half-hourly NEE, soil water content (SWC) at 30 cm depth, air temperature (Ta), soil temperature at 10 cm depth (Ts) and incident photosynthetically active radiation (PAR) before and after a selected rainfall event (b–c). The horizontal dashed lines in (a) and (b) represent y = 0, and the vertical dashed lines in (b) and (c) separate each day. The shadow areas in (b) and (c) indicate rainy days. Rainfall was 50 mm on DOY 179 and 11 mm on DOY 180. decomposition of soil organic matter due to the “Birch ef- fect”, which has been widely reported in regions with dry soils when a rainfall event occurs after a period of drought (Jarvis et al., 2007). Alternatively, soil water from rainfall may have degassed the CO2 stored in soil pores, consider- ing the sandy soils with high porosity in desert areas (Lee et al., 2004). The lagged responses of photosynthesis may be associated with physiological acclimation and recovery of plants from the preceding dry period (Liu et al., 2011). It may also be related to the post-rainfall environmental conditions that are optimum for C assimilation (Fig. 9c). These results showed that fast and slow responses contribute together to the pulsed ecosystem behavior. X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland The asynchronous and dif- ferential responses of biotic processes to rainfall pulses may affect C sequestration capacity of arid and semiarid areas un- der the projected climate change scenarios associated with increased rainfall variability (Chen et al., 2009). It is worthy of note that not all rain events caused an equal response of NEE (Fig. 9a). For example, NEE seemed relatively insensitive to a smaller rain event on DOY 202 (31 mm). This may be due to other biophysical factors that confound the NEE responses to sudden increases in water availability (Chen et al., 2009). Both temperature and radi- ation were much less affected over the DOY 202 rain event (data not shown) than over the DOY 179–180 event (61 mm, Fig. 9b and c), which could partially explain the result that the DOY 202 rain event did not cause a large ﬂuctuation in NEE. The behavior of NEE over a rain event also depends on the size and timing of water pulse, the environmental condi- tions prior to the rain, plant phenology, functional type and rooting depth, all of which affect the rainfall response of NEE (Aires et al., 2008; Liu et al., 2011; Gao et al., 2012). Figure 9. Pulses in daily net ecosystem CO2 exchange (NEE) as related to rainfall events during the growing season (a), and synoptic variations of half-hourly NEE, soil water content (SWC) at 30 cm depth, air temperature (Ta), soil temperature at 10 cm depth (Ts) and incident photosynthetically active radiation (PAR) before and after a selected rainfall event (b–c). The horizontal dashed lines in (a) and (b) represent y = 0, and the vertical dashed lines in (b) and (c) separate each day. The shadow areas in (b) and (c) indicate rainy days. Rainfall was 50 mm on DOY 179 and 11 mm on DOY 180. these interacting environmental factors regulate ecophysio- logical processes is needed to develop mechanistic models suitable for arid and semiarid ecosystems, which is a focus of our ongoing research. 4.4 Biotic controls on CO2 ﬂuxes Leaf area relates to both the amount of photosynthetic tis- sues and the amount of intercepted light by the vegetation (Yang et al., 2011). Our results that photosynthetic param- eters varied seasonally (Table 1; Fig. 2) with canopy de- velopment (e.g., changes in LAI) have been reported pre- viously for different vegetation (Zha et al., 2004; Yang et al., 2011). The small magnitudes of NEEday and its weak response to PAR in October (Table 1; Figs. 2f and 10) re- sulted partially from senescent leaves and reduced LAI at the end of the growing season. Temperature and radiation also decreased at the late season, contributing to reduced CO2 uptake by the vegetation. Furthermore, we found that LAI accounted for 45 % and 65 % of the seasonal variation in NEE and GEP, respectively (Fig. 8b and c), indicating the importance of canopy development in controlling C balance. Similar GEP–LAI and NEE–LAI relationships have been re- ported for steppe, grassland and pasture ecosystems (e.g., Tappeiner and Cernusca, 1998, Flanagan et al., 2002; Yang et al., 2011). The slope of the GEP–LAI relationship reported here (4.1 g C m−2 leaf area day−1) was comparable to that www.biogeosciences.net/11/4679/2014/ X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4690 et al.: Biophysical controls on NEE over a semiarid shrubland p y Figure 10. Monthly mean diurnal cycles of net ecosystem CO2 exchange (NEE) (a), incident photosynthetically active radiation (PAR) (b), air temperature (Ta) at 6 m above ground (c) and vapor pressure deﬁcit (VPD) (d). The vertical dashed lines indicate noon, and the horizontal dashed line in (a) represents y = 0. Figure 10. Monthly mean diurnal cycles of net ecosystem CO2 exchange (NEE) (a), incident photosynthetically active radiation (PAR) (b), air temperature (Ta) at 6 m above ground (c) and vapor pressure deﬁcit (VPD) (d). The vertical dashed lines indicate noon, and the horizontal dashed line in (a) represents y = 0. in a semiarid steppe (3.1 g C m−2 day−1) (Li et al., 2005) and two Mediterranean grasslands (3.9–4.1 g C m−2 day−1; Xu and Baldocchi, 2004; Aires et al., 2008). However, it was much smaller than that found in a Canadian temperate grass- land (7.5–8.7 g C m−2 day−1; Flanagan et al., 2002). A small GEP–LAI slope may be indicative of water and nutrient lim- itations (Li et al., 2005). 4.3 Effects of rain pulses on NEE Rain events triggered pulsed dynamics of NEE in the shrub ecosystem during the growing period (Fig. 9a). Large posi- tive daily NEE occurred immediately after the heavy rain on DOY 179–180 (61 mm), but the peak of CO2 uptake lagged the pulse of CO2 release by 1–2 days (Fig. 9a and b). Simi- larly, Gao et al. (2012) found that a large precipitation event resulted in a rapid burst of positive C ﬂux (CO2 release) be- fore negative values set in 1–2 days later in a revegetated shrubland in northwest China. Wang et al. (2014) also re- ported immediate pulses of Rs following rain events in an A. ordosica-dominated community at our site. In a semiarid steppe in central Mongolia, the respiration enhancement ef- fect was even higher during the rain period itself (Li et al., 2005). In contrast, Liu et al. (2011) found in a saline desert that NEE took 4–5 days to reach its peak of CO2 uptake after rain. The rapid stimulation of CO2 release by precipitation may arise from the rapid activation of microorganisms and www.biogeosciences.net/11/4679/2014/ Biogeosciences, 11, 4679–4693, 2014 Biogeosciences, 11, 4679–4693, 2014 Appendix A t of abbreviations. Variables/parameters Description Carbon ﬂuxes GEP Gross ecosystem productivity NEP Net ecosystem productivity NEE Net ecosystem CO2 exchange NEEday Daytime net ecosystem CO2 exchange (µmol m−2 s−1) NEEnight Nighttime net ecosystem CO2 exchange (µmol m−2 s−1) Re Ecosystem respiration Rs Soil respiration Biophysical variables LAI Leaf area index (m2 m−2) PAR Photosynthetically active radiation Q PAR intensity (µmol photons m−2 s−1) REW Relative extractable water content Rn Net radiation (Wm−2) Ta Air temperature (°C) Ts Soil temperature (°C) SWC Soil water content (m3 m−3) SWCmax Maximum SWC during the period when Ts at 10 cm depth > 0 ◦C SWCmin Minimum SWC during the period when Ts at 10 cm depth > 0 ◦C Model parameters Qc Light compensation point (µmol photons m−2 s−1) Qc PAR intensity at the maximum rate of net CO2 uptake (µmol photons m−2 s−1) NEEmax Maximum rate of net CO2 uptake at Q = Qm (µmol m−2 s−1) Q10 Sensitivity of ecosystem respiration to changes in temperature Rd Bulk ecosystem respiration derived from Eq. (1) (µmol CO2 m−2 s−1) Re10 Re at Ts =10 ◦C (µmol CO2 m−2 s−1) ϕ0 Quantum yield at Q = 0 (µmol CO2 µmol photons−1) ϕc Quantum yield at Q = Qc (µmol CO2 µmol photons−1) α, β, γ Fit parameters in Eq. (1) a, b, c, d Fit parameters in Eq. (9) Table A1. List of abbreviations. Biogeosciences, 11, 4679–4693, 2014 5 Conclusions In line with our hypotheses, we found in the semiarid shrub ecosystem that (1) water stress exerted a strong control over half-hourly changes in NEE during the peak growing sea- son, and interacted with heat stress and photoinhibition in constraining C ﬁxation; (2) rain pulses regulated NEE at the synoptic scale, highlighting the role of water supply in the al- leviation of abiotic stresses; (3) canopy development largely determined NEE and GEP over the entire growing season. Climate modeling suggests a warmer and drier future cli- mate in the semiarid and arid regions of Asia (McCarthy et al., 2001). Hence, more stressful environmental conditions in the future may lead to substantially lower carbon sequestra- tion capacity in temperate semiarid areas. Also, the predicted higher variability in precipitation (Easterling et al., 2000), i.e., more extreme but less frequent rainfall events intervened by longer dry periods, accentuates the role of the temporal pattern of water availability in controlling NEE in the future. Canopy productivity was shown to have a shadowing ef- fect on Re as photosynthesis provides substrates to both au- totrophic and heterotrophic respiration (Wan and Luo, 2003; Flanagan and Johnson, 2005). In our study, GEP accounted for 65 % of the seasonal variation in Re, with a slope of 0.34 (Fig. 8a). Li et al. (2005) reported a similar slope (0.33) for the Re–GEP relationship in a semiarid steppe. Xu and Bal- docchi (2004) and Aires et al. (2008) showed slopes of 0.47 and 0.53 in Mediterranean grasslands, respectively. However, Liu et al. (2011) reported a larger slope (0.8) in a saline desert shrub ecosystem, which may reﬂect greater C alloca- tion to respiratory tissues (stems, branches and coarse roots) or functions (e.g., maintenance respiration). Biogeosciences, 11, 4679–4693, 2014 www.biogeosciences.net/11/4679/2014/ X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland Appendix A Table A1. List of abbreviations. 5 Conclusions Variables/parameters Description Carbon ﬂuxes GEP Gross ecosystem productivity NEP Net ecosystem productivity NEE Net ecosystem CO2 exchange NEEday Daytime net ecosystem CO2 exchange (µmol m−2 s−1) NEEnight Nighttime net ecosystem CO2 exchange (µmol m−2 s−1) Re Ecosystem respiration Rs Soil respiration Biophysical variables LAI Leaf area index (m2 m−2) PAR Photosynthetically active radiation Q PAR intensity (µmol photons m−2 s−1) REW Relative extractable water content Rn Net radiation (Wm−2) Ta Air temperature (°C) Ts Soil temperature (°C) SWC Soil water content (m3 m−3) SWCmax Maximum SWC during the period when Ts at 10 cm depth > 0 ◦C SWCmin Minimum SWC during the period when Ts at 10 cm depth > 0 ◦C Model parameters Qc Light compensation point (µmol photons m−2 s−1) Qc PAR intensity at the maximum rate of net CO2 uptake (µmol photons m−2 NEEmax Maximum rate of net CO2 uptake at Q = Qm (µmol m−2 s−1) Q10 Sensitivity of ecosystem respiration to changes in temperature Rd Bulk ecosystem respiration derived from Eq. (1) (µmol CO2 m−2 s−1) Re10 Re at Ts =10 ◦C (µmol CO2 m−2 s−1) ϕ0 Quantum yield at Q = 0 (µmol CO2 µmol photons−1) ϕc Quantum yield at Q = Qc (µmol CO2 µmol photons−1) α, β, γ Fit parameters in Eq. (1) a, b, c, d Fit parameters in Eq. (9) 4691 X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4692 The Supplement related to this article is available online at doi:10.5194/bg-11-4679-2014-supplement. piration in biologically crusted soil during the dry season, Sci. World J., 2013, 408560, doi:10.1155/2013/408560, 2013. piration in biologically crusted soil during the dry season, Sci. World J., 2013, 408560, doi:10.1155/2013/408560, 2013. Flanagan, L. B. and Johnson, B. G.: Interacting effects of temper- ature, soil moisture and plant biomass production on ecosystem respiration in a northern temperate grassland, Agr. Forest Mete- orol., 130, 237–253, 2005. Acknowledgements. This research work has been supported especially by the National Natural Science Foundation of China (NSFC, proj. no. 31361130340, 31270755 and 31200537). It is also related to the ongoing Finnish–Chinese research collaboration project EXTREME, between Beijing Forestry University, School of Soil and Water Conservation (team led by prof. Tianshan Zha) and University of Eastern Finland (UEF), School of Forest Sciences (team led by Prof. Heli Peltola), funded jointly by NSFC, the Academy of Finland and University of Eastern Finland (proj. no. 14921) for years 2013–2016. We are grateful to the two anonymous reviewers and the Handling Editor (Dr. Paul Stoy) who provided very helpful comments and suggestions. We thank H. S. Shi, Y. M. Zhang, X. W. Yang, S. J. Liu, B. Wang and S. L. Tang for their assistance with ﬁeld measurements and instrument maintenance. Flanagan, L. B., Wever, L. A., and Carlson, P. J.: Seasonal and inter- annual variation in carbon dioxide exchange and carbon balance in a northern temperate grassland, Glob. Change Biol., 8, 599– 615, 2002. Fu, Y. L., Yu, G. R., Sun, X. M., Li, Y. N., Wen, X. F., Zhang, L. M., Li, Z. Q., Zhao, L., and Hao, Y. B.: Depression of net ecosystem CO2 exchange in semi-arid Leymus chinensis steppe and alpine shrub, Agr. Forest Meteorol., 137, 234–244, 2006. Gao, Y., Li, X., Liu, L., Jia, R., Yang, H., Li, G., and Wei, Y.: Sea- sonal variation of carbon exchange from a revegetation area in a Chinese desert, Agr. Forest Meteorol., 156, 134–142, 2012. Hagen, S. C., Braswell, B. H., Linder, E., Frolking, S., Richard- son, A. D., and Hollinger, D. Y.: Statistical uncertainty of eddy ﬂux-based estimates of gross ecosystem carbon exchange at Howland Forest, Maine, J. Geophys. Res., 111, D08S03, doi:10.1029/2005JD006154, 2006. Edited by: P. Stoy Edited by: P. Stoy Hollinger, D. Y. and Richardson, A. References Aires, L. M. I., Pio, C. A., and Pereira, J. S.: Carbon dioxide ex- change above a Mediterranean C3/C4 grassland during two cli- matologically contrasting years, Glob. Change Biol., 14, 539– 555, 2008. Lal, R.: Carbon sequestration in dryland ecosystems, Environ. Man- age., 33, 528–544, 2004. Asner, G. P., Archer, S., Hughes, R. F., Ansley, R. J., and Wess- man, C. A.: Net changes in regional woody vegetation cover and carbon storage in Texas drylands, 1937–1999, Glob. Change Biol., 9, 316–335, 2003. Lee, X., Wu, H. J., Sigler, J., Oishi, C., and Siccama, T.: Rapid and transient response of soil respiration to rain, Glob. Change Biol., 10, 1017–1026, 2004. Li, S. G., Asanuma, J., Eugster, W., Kotani, A., Liu, J. J., Urano, T., Oikawa, T., Davaa, G., Oyunbaatar, D., and Sugita, M.: Net ecosystem carbon dioxide exchange over grazed steppe in cen- tral Mongolia, Glob. Change Biol., 11, 1941–1955, 2005. Baldocchi, D. D.: Assessing the eddy covariance technique for evaluating carbon dioxide exchange rates of ecosystems: past, present and future, Glob. Change Biol., 9, 479–492, 2003. Burba, G. and Anderson, D.: A Brief Practical Guide to Eddy Covariance Flux Measurements: Principles and Workﬂow Ex- amples for Scientiﬁc and Industrial Applications, LI-COR Bio- sciences, 2010. Li, X. R., Ma, F. Y., Xiao, H. L., Wang, X. P., and Kim, K. C.: Long- term effects of revegetation on soil water content of san dunes in arid regions of Northern China, J. Arid Environ., 57, 1–16, 2004. Liu, R., Li, Y., and Wang, Q. X.: variations in water and CO2 ﬂuxes over a saline desert in western China, Hydrol. Process., 26, 513– 522, 2011. Chen, X. and Duan, Z.: Changes in soil physical and chemical properties during reversal of desertiﬁcation in Yanchi County of Ningxia Hui autonomous regions, China, Environ. Geol., 57, 975–985, 2009. Liu, R., Pan, L. P., Jenerette, G. D., Wang, Q. X., Cieraad, E., and Li, Y.: High efﬁciency in water use and carbon gain in a wet year for a desert halophyte community, Agr. Forest Meteorol., 162– 163, 127–135, 2012. Chen, S., Lin, G., Huang, J., and Jenerette, D.: Dependence of car- bon sequestration on the differential responses of ecosystem pho- tosynthesis and respiration to rain pulses in a semiarid steppe, Glob. Change Biol., 15, 2450-2461, 2009. McCarthy, J., Canziani, O., Leary, N., Dokken, D. J., and White, K. X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland D.: Uncertainty in eddy covari- ance measurements and its application to physiological models, Tree Physiol., 25, 873–885, 2005. Jarvis, P., Rey, A., Petsikos, C., Wingate, L., Reyment, M., Pereira, J., Banza, J., David, J., Miglietta, F., Borghetti, M., Manca, G., and Valentini, R.: Drying and wetting of Mediter- ranean soils stimulates decomposition and carbon dioxide emis- sion: the “Birch effect”, Tree Physiol., 27, 929–940, 2007. X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland 4693 Whittaker, R. H.: Communities and Ecosystems, MacMillan pub- lishing Co., New York, 1975. Whittaker, R. H.: Communities and Ecosystems, MacMillan pub- lishing Co., New York, 1975. Papale, D., Reichstein, M., Aubinet, M., Canfora, E., Bernhofer, C., Kutsch, W., Longdoz, B., Rambal, S., Valentini, R., Vesala, T., and Yakir, D.: Towards a standardized processing of Net Ecosys- tem Exchange measured with eddy covariance technique: algo- rithms and uncertainty estimation, Biogeosciences, 3, 571–583, doi:10.5194/bg-3-571-2006, 2006. Xu, L. and Baldocchi, D. D.: Seasonal variation in carbon dioxide exchange over a Mediterranean annual grassland in California, Agr. Forest Meteorol., 123, 79–96, 2004. Yang, X., Zhang, K., Jia, B., and Ci, L.: Desertiﬁcation assessment in China: An overview, J. Arid Environ., 63, 517–531, 2005. Poulter, B., Frank, D., Ciais, P., Myneni, R. B., Andela, N., Bi, J., Broquet, G., Canadell, J. G., Chevallier, F., Liu, Y. Y., Running, S. W., Sitch, S., and van der Werf, G. R.: Contribution of semi- arid ecosystems to interannual variability of the global carbon cycle, Nature, 509, 600–603, 2014. Yang, F., Zhou, G., Hunt, J. E., and Zhang, F.: Biophysical regula- tion of net ecosystem carbon dioxide exchange over a temperate desert steppe in Inner Mongolia, China, Agr. Forest Meteorol., 142, 318–328, 2011. Schmid, H. P.: Experimental design for ﬂux measurements: match- ing scales of observations and ﬂuxes, Agr. Forest Meteorol., 87, 179–200, 1997. Ye, Z. P.: A new model for relationship between irradiance and the rate of photosynthesis in Oryza sativa, Photosynthetica, 45, 637– 640, 2007. Suyker, A. E. and Verma, S. B.: Year-round observations of the net ecosystem exchange of carbon dioxide in a native tallgrass prairie, Glob. Change Biol., 7, 279–289, 2001. Zha, T., Kellomäki, S., Wang, K., and Rouvinen, I.: Carbon seques- tration and ecosystem respiration for 4 years in a scots pine for- est, Glob. Change Biol., 10, 1492–1503, 2004. Zhang, W. L., Chen, S. P., Chen, J., Wei, L., Han, X. G., and Lin, G. H.: Biophysical regulations of carbon ﬂuxes of a steppe and a cultivated cropland in semiarid Inner Mongolia, Agr. Forest Meteorol., 146, 216–229, 2007. Tappeiner, U. and Cernusca, A.: Model simulation of spatial distri- bution of photosynthesis in structural differing plant communi- ties in the Central Caucasus, Ecol. Model., 113, 201–223, 1998. References S.: Climate Change 2001: Impacts, Adaptation, and Vulnerability, Cambridge University Press, New York, 2001. Dragoni, D., Schmid, H. P., Grimmond, C. S. B., and Loescher, H. W.: Uncertainty of annual net ecosystem productivity estimated using eddy covariance ﬂux measurements, J. Geophys. Res., 112, D17102, doi:10.1029/2006JD008149, 2007. Vulnerability, Cambridge University Press, New York, 2 Moffat, A. M., Papale, D., Reichstein, M., Hollinger, D. Y., Richard- son, A. D., Barr, A. G., Beckstein, C., Braswell, B. H., Churk- ina, G., Desai, A. R., Falge, E., Gove, J. H., Heimann, M., Hui, D., Jarvis, A. J., Kattge, J., Noormets, A., and Stauch, V. J.: Comprehensive comparison of gap-ﬁlling techniques for eddy covariance net carbon ﬂuxes, Agr. Forest Meteorol., 147, 209– 232, 2007. Easterling, D. R., Meehl, G. A., Parmesan, C., Changnon, S. A., Karl, T. R., and Mearns, L. O.: Climate extremes: observations, modeling, and impacts, Science, 289, 2068–2074, 2000. Feng, W., Zhang, Y., Wu, B., Zha, T., Jia, X., Qin, S., Shao, C., Liu, J., Lai, Z., and Fa, K.: Inﬂuence of disturbance on soil res- www.biogeosciences.net/11/4679/2014/ Biogeosciences, 11, 4679–4693, 2014 X. Jia et al.: Biophysical controls on NEE over a semiarid shrubland Wang, Y., Zhou, G., and Wang, Y.: Environmental effects on net ecosystem CO2 exchange at half-hour and month scales over Stipa krylovii steppe in northern China, Agr. Forest Meteorol., 148, 714-722, 2008. Zhu, Z., Sun, X., Wen, X., Zhou, Y., Tian, J., and Yuan, G.: Study on the processing method of nighttime CO2 eddy covariance ﬂux data in ChinaFLUX, Sci. China Ser. D, 49, supp. II, 36–46, 2006. Wan, S. and Luo, Y.: Substrate regulation of soil res- piration in tallgrass prairie: results of a clipping and shading experiment, Global Biogeochem. Cy., 17, 1054, doi:10.1029/2002GB001971, 2003. Zhou, J., Zhang, Z., Sun, G., Fang, X., Zha, T., McNulty, S., Chen, J., Jin, Y., and Noormets, A.: Response of ecosystem car- bon ﬂuxes to drought events in a poplar plantation in Northern China, Forest Ecol. Manag., 300, 33–42, 2013. Wang, B., Zha, T. S., Jia, X., Wu, B., Zhang, Y. Q., and Qin, S. G.: Soil moisture modiﬁes the response of soil respiration to temper- ature in a desert shrub ecosystem, Biogeosciences, 11, 259–268, doi:10.5194/bg-11-259-2014, 2014. Biogeosciences, 11, 4679–4693, 2014 www.biogeosciences.net/11/4679/2014/
https://openalex.org/W2740452420	https://aip.scitation.org/doi/pdf/10.1063/1.4994989	English	null	Compartmentalized embryoid body culture for induction of spatially patterned differentiation	Biomicrofluidics	2,017	cc-by	2,766	RESEARCH ARTICLE \| AUGUST 02 2017 Compartmentalized embryoid body culture for induction of spatially patterned differentiation RESEARCH ARTICLE \| AUGUST 02 2017 RESEARCH ARTICLE \| AUGUST 02 2017 Compartmentalized embryoid body culture for induction of spatially patterned differentiation Shohei Kaneda; Jiro Kawada; Hidenori Akutsu; Justin Ichida; Yoshiho Ikeuchi; Teruo Fujii Shohei Kaneda; Jiro Kawada; Hidenori Akutsu; Justin Ichida; Yoshiho Ikeuchi; Teruo Fujii Shohei Kaneda; Jiro Kawada; Hidenori Akutsu; Justin Ichida; Yoshiho Ikeuchi; Teruo Fujii Articles You May Be Interested In Articles You May Be Interested In Flip channel: A microfluidic device for uniform-sized embryoid body formation and differentiation 24 October 2024 04:01:55 BIOMICROFLUIDICS 11, 041101 (2017) a)Authors to whom correspondence should be addressed: shk@iis.u-tokyo.ac.jp and tfujii@iis.u-tokyo.ac.jp. b)S. Kaneda and J. Kawada contributed equally to this work. Compartmentalized embryoid body culture for induction of spatially patterned differentiation Shohei Kaneda,1,2,a),b) Jiro Kawada,1,b) Hidenori Akutsu,3 Justin Ichida,4,5 Yoshiho Ikeuchi,6,7 and Teruo Fujii1,a) 1Center for International Research on Integrative Biomedical Systems, Institute of Industrial Science, The University of Tokyo, Tokyo, Japan 2LIMMS-CNRS/IIS, Institute of Industrial Science, The University of Tokyo, Tokyo, Japan 3Department of Reproductive Medicine, Centre for Regenerative Medicine, National Research Institute for Child Health and Development, Tokyo, Japan 4Department of Stem Cell Biology and Regenerative Medicine, Keck School of Medicine, University of Southern California, Los Angeles, California 90033, USA 5Eli and Edythe Broad CIRM Center for Regenerative Medicine and Stem Cell Research, University of Southern California, Los Angeles, California 90033, USA 6Institute of Industrial Science, The University of Tokyo, Tokyo, Japan 7Department of Chemistry and Biotechnology, Graduate School of Engineering, The University of Tokyo, Tokyo, Japan (Received 9 July 2017; accepted 24 July 2017; published online 2 August 2017) We developed a compartmentalized culture system of single embryoid bodies (EBs) utilizing a through-hole on a membrane to induce spatially patterned differ- entiation. An EB derived from mouse pluripotent stem cells was immobilized on the through-hole. By introducing a stem cell maintenance medium and a differenti- ation medium into upper and lower culture compartments, respectively, a localized differentiated state was achieved only in the lower part of EB, which is exposed to the medium in the lower compartment. This system may enable us to reconstruct complex tissues and to recapitulate developmental processes using EBs. V C 2017 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://creativecommons.org/ licenses/by/4.0/). [http://dx.doi.org/10.1063/1.4994989] 24 October 2024 04:01:55 I. INTRODUCTION Spherical aggregates of pluripotent stem cells (PSCs), i.e., embryoid bodies (EBs), are use- ful for various applications, e.g., developmental1 and disease models,2 regenerative medicine,3 and in vitro drug testing.4 Since spatially patterned distributions of differentiation factors are generated during developmental processes,5–7 the reconstruction of more complex tissues may be possible by applying patterned distributions of factors to the EB culture. However, in con- ventional EB culture methods, EBs are incubated in a medium containing these factors; there- fore, the whole EB surface is exposed to the factors homogenously, and it is consequently difﬁ- cult to apply a patterned distribution of factors. The use of microﬂuidic technologies has been proposed to generate a patterned distribution of factors by controlled diffusion8 or laminar ﬂow.9–14 Using these methods, a patterned distribution of factors has been applied to cell popu- lations and cell responses have been observed, e.g., the direction of migration11,12 and the spa- tial distribution of gene expression13,14 in populations have been successfully controlled. However, these methods sometimes require a special culture apparatus (e.g., syringe pump or pressure controller) to maintain the ﬂow as well as a skilled operator. To overcome these limi- tations, we propose a simple compartmentalized culture system to expose an EB to two differ- ent agents, such as differentiation factors [Fig. 1(a)]. V C Author(s) 2017. 11, 041101-1 1932-1058/2017/11(4)/041101/4 041101-2 Kaneda et al. Biomicrofluidics 11, 041101 (2017) II. MATERIALS AND METHODS A microﬂuidic device made of polydimethylsiloxane (PDMS) was fabricated [Figs. 1(b) and S1, supplementary material]. The device consists of upper and lower compartments and a thin membrane with a through-hole sandwiched between the compartments. The membrane was fabricated by a spin-coating method,15 and other parts of the device were fabricated by punch- ing holes in PDMS slabs using biopsy punches (Kai Industries, Gifu, Japan). Mouse induced pluripotent stem cells (miPSCs) (iPS-MEF-Ng-20D-17 cell line16) and mouse embryonic stem cells (mESCs) were maintained in a stem cell maintenance medium (ESGRO-2i Medium; Merck Millipore, Darmstadt Germany) on a gelatin-coated dish (Iwaki, Tokyo, Japan) to keep the undifferentiated state. To form undifferentiated EBs, miPSCs or mESCs were seeded in KnockOut DMEM (Life Technologies, Carlsbad, CA, USA) containing 15% KnockOut Serum Replacement (HyClone, Logan, UT, USA), 1% GlutaMAX (Life Technologies), 1% MEM Non-Essential Amino Acids (Life Technologies), 0.09% 2-mercaptoethanol (Life Technologies), and 0.1% Leukemia Inhibitory Factor (Wako, Osaka, Japan) at 20 000 cells per 200 ll in a V-shaped well of a 96-well plate (Sumitomo Bakelite, Tokyo, Japan) with an ultra-low-cell adhesion surface, and cultured at 37 C under 5% CO2. On day 2, the formed EB was collected from the plate using a pipetter with a wide-oriﬁce tip and transferred to the device. To validate that the EBs can be exposed to two different agents separately in the device, three ﬂuorescent dyes, i.e., Calcein AM (Dojindo, Kumamoto, Japan), Hoechst 33342 (Dojindo), and MitoTracker Orange (Life Technologies), were used. A neural differentiation medium (RHB-A; StemCells, Inc., Newark, CA, USA) was used to induce the differentiation of EBs. An inverted ﬂuorescence microscope (IX71; Olympus Corp., Tokyo, Japan) and a CCD camera (DP71; Olympus Corp.) were used to obtain both phase contrast and ﬂuorescence images. An incubator integrated with a motorized inverted microscope system (CCM-1.3XYZ/CO2; Astec, Fukuoka, Japan) was used to obtain time-lapse images. By tilting the device 90 from the regular position for observation, images of the side view of the EB in the device were obtained. To reduce the roughness of side surfaces of the device, an extra PDMS component was placed next to the device in the observation (Fig. S2, supplementary material). A raster graphics editor (Photoshop CS5; Adobe Systems, San Jose, CA, USA) was used to measure the ﬂuorescence. 24 October 2024 04:01:55 III. RESULTS AND DISCUSSION To immobilize EBs on the through-hole in the device, a ﬂow induced by hydrostatic pres- sure between the upper and lower compartments was used. First, both the upper and lower com- partments were preﬁlled with culture medium. Then, an aliquot of medium with EB was added to generate the ﬂow, causing the EB to spontaneously move towards and become immobilized on the through-hole [Fig. 2(a) and Movie S1, supplementary material]. We conﬁrmed that through-holes of various sizes, ranging from 100 to 400 lm, can be used to immobilize EBs (Fig. S3, supplementary material). A side view of the iPSC-derived EB just after immobiliza- tion shows that the whole EB expresses Nanog-GFP, which indicates the undifferentiated state FIG. 1. Compartmentalized embryoid body culture for induction of spatially patterned differentiation. (a) Conceptual dia- gram of a compartmentalized embryoid body culture system. (b) Schematic drawing of the design of the compartmentalized culture device, consisting of PDMS (blue indicates the upper compartment, and light blue indicates the lower compartment) and PDMS/Glass support (shown in white). FIG. 1. Compartmentalized embryoid body culture for induction of spatially patterned differentiation. (a) Conceptual dia- gram of a compartmentalized embryoid body culture system. (b) Schematic drawing of the design of the compartmentalized culture device, consisting of PDMS (blue indicates the upper compartment, and light blue indicates the lower compartment) and PDMS/Glass support (shown in white). 041101-3 Kaneda et al. Biomicrofluidics 11, 041101 (2017) FIG. 2. Induction of spatially patterned differentiation. (a) Immobilized EB on a through-hole. The through-hole diameter is 200 lm. (b) Side view of an immobilized EB on a through-hole. The middle part of the EB in contact with the membrane was blurry due to the imaging setup (Fig. S1, supplementary material). (c) EBs partially labeled with various ﬂuorescent dyes. The dotted white lines and black lines show the upper parts and lower parts of EBs, respectively. (d) Induction of spa- tially patterned differentiation of the EB. The upper and lower parts of the EB were simultaneously incubated with the stem cell maintenance medium and the neural differentiation medium for 4 days. (e) EBs separately incubated with the stem cell maintenance medium (upper EB) and neural differentiation medium (lower EB) as a control, respectively. (f) Relative ﬂuo- rescence intensity of Nanog-GFP in control and compartmentalized EB culture. III. RESULTS AND DISCUSSION Average intensities of areas indicated by dotted lines in (d), (e) were measured and normalized by intensities in the area framed by a white-dotted line, i.e., the upper EB (e) or upper part (d) incubated with the stem cell maintenance medium, respectively. Scale bars: 200 lm (a), (b), 500 lm (c), (d), (e). EBs derived from miPSCs (b), (d), (e) and mESCs (a), (c) were used. 24 October 2024 04:01:55 24 October 2024 04:01:55 FIG. 2. Induction of spatially patterned differentiation. (a) Immobilized EB on a through-hole. The through-hole diameter is 200 lm. (b) Side view of an immobilized EB on a through-hole. The middle part of the EB in contact with the membrane was blurry due to the imaging setup (Fig. S1, supplementary material). (c) EBs partially labeled with various ﬂuorescent dyes. The dotted white lines and black lines show the upper parts and lower parts of EBs, respectively. (d) Induction of spa- tially patterned differentiation of the EB. The upper and lower parts of the EB were simultaneously incubated with the stem cell maintenance medium and the neural differentiation medium for 4 days. (e) EBs separately incubated with the stem cell maintenance medium (upper EB) and neural differentiation medium (lower EB) as a control, respectively. (f) Relative ﬂuo- rescence intensity of Nanog-GFP in control and compartmentalized EB culture. Average intensities of areas indicated by dotted lines in (d), (e) were measured and normalized by intensities in the area framed by a white-dotted line, i.e., the upper EB (e) or upper part (d) incubated with the stem cell maintenance medium, respectively. Scale bars: 200 lm (a), (b), 500 lm (c), (d), (e). EBs derived from miPSCs (b), (d), (e) and mESCs (a), (c) were used. [Fig. 2(b)]. By incubating the EB with the stem cell maintenance medium, cell proliferation at both the upper and lower parts of the EB was observed and leading to an articulated shape (cul- ture for 44.5 h is shown in Movie S2, supplementary material). In the present system, it is possible to change the culture medium in the device after the immobilization of the EB using conventional tools, a pipetter, and an aspirator. Figure 2(c) shows EBs collected from devices after staining with various ﬂuorescent dyes. These images 041101-4 Kaneda et al. Biomicrofluidics 11, 041101 (2017) clearly show that EBs could be separately exposed to two different agents above and below the membrane. SUPPLEMENTARY MATERIAL See supplementary material for photographs of fabricated devices (Fig. S1), schematic illustration of the observation system for the side view of the EB in the device (Fig. S2), micro- photographs of EBs immobilized on through-holes of various size (Fig. S3), video showing an immobilization process of EB on the through-hole (Movie S1), and video showing an EB cul- ture for 44.5 h on the device (Movie S2). 24 October 2024 04:01:55 IV. CONCLUSIONS In this study, we developed a novel cell culture device using a through-hole membrane that enables compartmentalized EB culture for induction of spatially patterned differentiation. Our system is easy to use because it requires only a conventional pipetter and aspirator for induc- tion. It enables the induction of two different differentiated states in a single EB, which is not possible with conventional EB culture systems. This spatially patterned differentiation method has potential applications in formation of organoids that have functionalities when the two sub- domains are located in a coordinated positional relationship.17 III. RESULTS AND DISCUSSION The upper and lower parts of the EB were incubated with the stem cell maintenance medium and the neural differentiation medium, respectively. After 4 days, the EB was collected to observe Nanog-GFP ﬂuorescence [Fig. 2(d), referred to as the compartmentalized EB cul- ture]. The ﬂuorescence in the upper part was higher than that in the lower part. This distinct pattern of ﬂuorescence in the two parts corresponded to controls incubated individually with the stem cell maintenance medium and neural differentiation medium [Fig. 2(e), referred to as the control culture]. Relative ﬂuorescent intensities of Nanog-GFP were similar in the control and the compartmentalized EB culture [Fig. 2(f)]. These results suggest that the compartmentalized culture device can be used to induce spatially patterned differentiation on a single EB. 17F. Birey, J. Andersen, C. D. Makinson, S. Islam, W. Wei, N. Huber, H. C. Fan, K. R. C. Metzler, G. Panagiotakos, N. Thom, N. A. O’Rourke, L. M. Steinmetz, J. A. Bernstein, J. Hallmayer, J. R. Huguenard, and S. P. Pas¸ca, Nature 545(7652), 54–59 (2017). 16K. Okita, T. Ichisaka, and S. Yamanaka, Nature 448(7151), 313–317 (2007). 17 This work was partially supported by JSPS KAKENHI Grant No. JP 17K14985. This work was partially supported by JSPS KAKENHI Grant No. JP 17K14985. 1M. Eiraku, N. Takata, H. Ishibashi, M. Kawada, E. Sakakura, S. Okuda, K. Sekiguchi, T. Adachi, and Y. Sasai, Nature 472(7341), 51-U73 (2011). 2 A. Lancaster, M. Renner, C. A. Martin, D. Wenzel, L. S. Bicknell, M. E. Hurles, T. Homfray, J. M. Penninger, A. P. kson, and J. A. Knoblich, Nature 501(7467), 373 (2013). 2M. A. Lancaster, M. Renner, C. A. Martin, D. Wenzel, L. S. Bicknell, M. E. Hurles, T. Homfray, J. M. Penninger, A. P Jackson, and J. A. Knoblich, Nature 501(7467), 373 (2013). , , ( ), ( ) 3C. E. Murry and G. Keller, Cell 132(4), 661–680 (2008). y , ( ), ( ) 4A. A. Kindberg, R. M. Bendriem, C. E. Spivak, J. Chen, A. Handreck, C. R. Lupica, J. Liu, W. J. Freed, and C. T. Lee Dis. Models Mech. 7(12), 1397–1405 (2014). y , ( ), ( ) 4A. A. Kindberg, R. M. Bendriem, C. E. Spivak, J. Chen, A. Handreck, C. R. Lupica, J. Liu, W. J. Freed, and C. T. Lee, Dis. Models Mech. 7(12), 1397–1405 (2014). 5B. L. Allen, T. Tenzen, and A. P. McMahon, Genes Dev. 21(10), 1244–1257 (2007). 6S. Shimozono, T. Iimura, T. Kitaguchi, S. Higashijima, and A. Miyawaki, Nature 496( 7 B. L. Allen, T. Tenzen, and A. P. McMahon, Genes Dev. 21(10), 1244 1257 (2007). 6S. Shimozono, T. Iimura, T. Kitaguchi, S. Higashijima, and A. Miyawaki, Nature 496(7445), 363 (2013). 7 g g j y 7K. S. Zaret, Nat. Rev. Genet. 3(7), 499–512 (2002). 8 7K. S. Zaret, Nat. Rev. Genet. 3(7), 499–512 (2002). 8 8V. V. Abhyankar, M. A. Lokuta, A. Huttenlocher, and D. J. Beebe, Lab Chip 6(3), 389–393 (2006). 9B G Ch L A Fl S W Rh P H S h A P L E S M ki d N L J 9B. G. Chung, L. A. Flanagan, S. W. Rhee, P. H. Schwartz, A. P. Lee, E. S. Monuki, and N. L. Jeon, Lab Chip 5(4) 401–406 (2005). , , , g ( ), ( ) 11D. Irimia, S. Y. Liu, W. G. Tharp, A. Samadani, M. Toner, and M. C. Poznansky, Lab Chip 6(2), 191–198 (2006). 12 p y p 12H. B. Mao, P. S. Cremer, and M. D. Manson, Proc. Natl. Acad. Sci. U.S.A. 100(9), 5449–5454 (2003). 13 J. Kawada, H. Kimura, H. Akutsu, Y. 16K. Okita, T. Ichisaka, and S. Yamanaka, Nature 448(7151), 313–317 (2007). 17F Bi J A d C D M ki S I l W W i N H b H C F K R C M l G P i k N 13J. Kawada, H. Kimura, H. Akutsu, Y. Sakai, and T. Fujii, Lab Chip 12(21), 4508–4515 (2012). 14W. T. Fung, A. Beyzavi, P. Abgrall, N. T. Nguyen, and H. Y. Li, Lab Chip 9(17), 2591–2595 (2009 15 This work was partially supported by JSPS KAKENHI Grant No. JP 17K14985. Sakai, and T. Fujii, Lab Chip 12(21), 4508 4515 (2012). 14W. T. Fung, A. Beyzavi, P. Abgrall, N. T. Nguyen, and H. Y. Li, Lab Chip 9(17), 2591–2595 (2009). 15 5D. Huh, B. D. Matthews, A. Mammoto, M. Montoya-Zavala, H. Y. Hsin, and D. E. Ingber, Science 328(5986), 1662–1668 (2010). 6 6K. Okita, T. Ichisaka, and S. Yamanaka, Nature 448(7151), 313–317 (2007). 7 17F. Birey, J. Andersen, C. D. Makinson, S. Islam, W. Wei, N. Huber, H. C. Fan, K. R. C. Metzler, G. Panagiotakos, N. Thom, N. A. O’Rourke, L. M. Steinmetz, J. A. Bernstein, J. Hallmayer, J. R. Huguenard, and S. P. Pas¸ca, Nature 545(7652), 54–59 (2017).
https://openalex.org/W4389306096	https://www.ijfmr.com/papers/2023/6/9907.pdf	English	null	Unravelling the Premises of Power Sharing, Democratic Ideals and Nation-Building and its Institutional Challenges	International Journal For Multidisciplinary Research	2,023	cc-by-sa	3,756	Abstract This paper explores nation-building as a complex term, highlighting its ambiguity and historical growth. Nation-building is described as a societal process that happens gradually and is impacted by political, technological, and economic shifts. For a new state to be legitimate, citizens must actively shape a shared political will. This is the central idea. Power-sharing is a contentious and potentially pacifying practise that is being studied in nations where identity groups are divided. The study emphasises the significance of institutional design as political violence around the world increases. For long-term stability, nation-building is important because it fosters a unified identity, frequently via education. Problems, however, lead to critical questions about its suitability, especially in light of the democratic peace paradigm, and they end with a variety of factors to take into account for successful nation-building initiatives. Keywords: nation-building, identity, legitimacy, power-sharing, democratic Unravelling the Premises of Power Sharing, Democratic Ideals and Nation-Building and its Institutional Challenges Dr. E. Benrithung Patton essor, Department of Political Science, Fazl Ali College, Mokokchung: Nagaland, India. Assistant Professor, Department of Political Science, Fazl Ali College, Mokokchung: Na International Journal for Multidisciplinary Research (IJFMR) E-ISSN: 2582-2160 ● Website: www.ijfmr.com ● Email: editor@ijfmr.com International Journal for Multidisciplinary Research (IJFMR) E-ISSN: 2582-2160 ● Website: www.ijfmr.com ● Email: editor@ijfmr.com rnational Journal for Multidisciplinary Research (IJFMR) E-ISSN: 2582-2160 ● Website: www.ijfmr.com ● Email: editor@ijfmr.com of representative citizens if ‘nation-building’ is to be successful. The citizens must grant the new state the essential legitimacy. Effective collective decision-making frequently necessitates forcing individual parties to make sacrifices for the common good, therefore commitment to the common good and to a shared community are crucial (eg. to protect the environment for future generations). The foundational components of community, shared values and ties, must support these sacrifices in order for them to be considered legitimate and, as a result, either forced onto the community or ineffectively accomplished. The modern nation-state is supported in large part by citizenship. People must embrace and comprehend their role as citizens, together with the rights and obligations that this includes, in order for important institutions like the political and judiciary systems to function effectively. Citizenship serves as the cornerstone of the social contract that binds individuals to a particular nation-state and vice versa. The sense of belonging to a larger political community can be weakened or undermined when individual identities and allegiances are based primarily on membership in nationality and other highly localised groups. The appeal of sub-national identities is likely to endure or even grow stronger in places where states have either never operated effectively or have stopped to do so, hurting efforts to forge a sense of national community. of representative citizens if ‘nation-building’ is to be successful. The citizens must grant the new state the essential legitimacy. Effective collective decision-making frequently necessitates forcing individual parties to make sacrifices for the common good, therefore commitment to the common good and to a shared community are crucial (eg. to protect the environment for future generations). The foundational components of community, shared values and ties, must support these sacrifices in order for them to be considered legitimate and, as a result, either forced onto the community or ineffectively accomplished. Th d ti t t i t d i l t b iti hi P l t b d h d ‘Nation-building’: its foundational framework The phrasing ‘nation-building’ is frequently used, yet it is still ambiguous and contentious. Its meaning is implied rather than defined in a large portion of the policy documentation. The more abstract process of creating a feeling of community or shared identity among the diverse groups who make up the population of a single state is attributed to as ‘nation-building.’ “Nation-building involves the creation of a national identity that supersedes local identities and loyalties that might compete with and preclude broader identification with the state” (Encyclopedia of Violence, Peace, and Conflict, 2022). It is crucial to consider how theories of nation building have changed over time, as well as the other ideas they have both replaced and embraced. Many people think that nation building is an evolutionary process rather than a revolutionary one, that it takes time, and that it is a social process that cannot be sped up from the outside. It took a very long time for the city-states of Greek to develop into a nation, the city-states of Germany to form the union and then a nation, the diverse languages and cultural groups of France to form the nation of France, and the development of China from the warring kingdoms. These developments were the result not only of changes in political leadership but also of changes in technology and economic processes (the agricultural and then the industrial revolutions), as well as of other factors. The community of citizens who wish to create a nation and take charge of this process is the subject of ‘nation-building’ (Dinnen, 2006). A shared political will must be shaped with the active involvement Volume 5, Issue 6, November-December 2023 IJFMR23069907 1 rnational Journal for Multidisciplinary Research (IJFMR) Besley & Persson, 2011; Burgess et al. 2015). Surprisingly little conclusive statistical evidence linking power sharing to peace exists despite the increased interest in it (Francois et al. 2015). Despite a few significant exceptions, such as those provided by Cederman et al. (2013), Gates et al. (2015), and Strom et al. (2017), identification concerns are still crucial since the introduction of power sharing may be endogenous and because national limits are insufficient as a safety net. Since most research rely on larger institutional elements, the lack of quantitative data on power sharing is another barrier. Political violence has reached distressing heights throughout human history, and efforts to prevent it have been made consistently. These extremes were notably evident in the 20th century. On the one hand, it was characterised by two World Wars, numerous instances of mass murder of civilians, horrifying suppression carried out by a number of totalitarian regimes, as well as numerous ongoing ethnic civil wars. On the other hand, there are significant growth in democratic initiatives and peacekeeping activities. We are currently experiencing a turnaround, with formerly peaceful regions of the world destabilising into extreme levels of violence. This is happening after an age of optimism about humanity’s ability to reduce armed conflict (Pinker, 2012). The problem has gained prominence among major international organisations (the United Nations, the World Bank, the IMF, and the OECD), and policy evaluations from the academic community can be a significant aid in this endeavour. Power sharing to prevent violence has been one permeating notion. Power sharing connote the precise divisions and combinations of power that guarantee that different groups have some degree of equal ‘participation’ in the institutions of a state and/or shared ‘ownership’ of its assets. This can be accomplished by creating a federation, a central government of national unity, or autonomous, non- governmental entities that serve as a check on the executive branch of government. Countries like Switzerland, Belgium, and Lebanon have long-standing systems of power sharing across various religious or linguistic groups. Since the 1998 Agreement, a similar system has also been in effect in Northern Ireland. Other nations, including as Afghanistan, Iraq, and Mali, have attempted it with less success. Power sharing, as a paradigm, builds stable pillars in the political architecture of the country. Like load-bearing buildings, these pillars sustain the weight of governance and keep it from collapsing into anarchy or chaos. rnational Journal for Multidisciplinary Research (IJFMR) Power sharing establishes institutions that act as sentinels defending against corruption and the misuse of power by upholding the values of openness, responsibility, and the rule of law. This innovative strategy guarantees that the nation’s governance is firmly established and resilient to both internal and foreign shocks. As a consequence, the nation is resilient and the rule of law is upheld, giving its people the confidence that their rights are upheld and their opinions are taken seriously. Power sharing becomes more prevalent on the international scene and becomes a diplomatic language that is in line with the values of inclusivity and collaboration. Nations that engage in domestic power sharing tend to be better at cooperative diplomacy. International relationships become anchored by the capacity to recognise and value other points of view and work together to find answers. This skill at diplomacy not only raises the nation’s profile internationally, but also puts it in a position to spur multilateral endeavours. A nation can become a global partner and contribute to a world where nations work together on the basis of mutual respect and shared values by introducing the concepts of power sharing to the international arena. Power sharing as a paradigm for nation-building A stable and inclusive nation is built on an architectural foundation, which is what power sharing looks like as a nation-building paradigm. Power sharing is fundamental to the governance plan because it prevents any one party from controlling the political landscape. Power sharing orchestrates a political composition that harmonises varied interests, much like a well-composed symphony where several instruments contribute to the main tune. This method avoids discordant notes that could cause social splits by acknowledging the inherent significance of inclusive decision-making. Power is distributed across different various sectors and communities, creating a resilient national structure that can withstand the challenges posed by its heterogeneous structure. Power sharing has been the subject of a contentious public discussion, with proponents emphasising its possible pacifying qualities while opponents highlight its spectacular failures, the risk of a lack of accountability, and the possibility of political rivalry in large coalition administrations. Power sharing, at its core, creates a democratic fabric that encompasses the principles of participation and representation. Citizens participate actively in shaping the fabric of their country rather than acting as passive observers in the loom of governance. By ensuring that political threads are woven throughout the populace’s collective will, power sharing helps to weave a storey of inclusivity and shared ownership. By promoting a political environment where various views are not only heard but also included in the decision-making process, it is consistent with democratic ideals. When this democratic tapestry is unfolded, it portrays a picture of a dynamic and involved society in which power is not centralised in the hands of a select few but rather distributed across the populace. There is a strong trend for some nationality, ethnically or religiously divided countries to implement some form of power sharing. And there is a wealth of abstract evidence and journalistic stories that point to power sharing having the ability to play a significant role in reducing violence. The consensus model of democracy, which is characterised by power sharing and the decentralisation of power at all levels, was chosen by many prosperous and peaceful countries that are ethnically and religiously divided (Lijphart, 1999). These illustrations support recent research showing that democratic institutions have a substantial impact on how political power can be dispersed and regulated (Acemoglu & Robinson, 2001; Volume 5, Issue 6, November-December 2023 IJFMR23069907 2 Volume 5, Issue 6, November-December 2023 rnational Journal for Multidisciplinary Research (IJFMR) E-ISSN: 2582-2160 ● Website: www.ijfmr.com ● Email: editor@ijfmr.com utilisation of power are the functions of institutions, which include both formal structures and informal norms. Comprehending these dynamics is essential for grasping decision-making procedures, distribution of resources, and the general stability of social structures. In order to stack the playing field in favour of democracy and arbitrate disputes in calm, if not productive, scholars and practitioners agree that institutional design is the most important policy tool at our disposal (Barnes, 2001; Belmont et al. 2002; Horowitz, 2002; Norris, 2002). A sighting of wide comparative politics discussion on democracy and power sharing helps in evaluating the externally sponsored endeavour to create democratic institutions and the subsequent consolidation of democracy in post-conflict developing countries. The majority of the power sharing discourse is focused on the requirement to give institutional assurances and protections to groups within divided community, even multinational states. utilisation of power are the functions of institutions, which include both formal structures and informal norms. Comprehending these dynamics is essential for grasping decision-making procedures, distribution of resources, and the general stability of social structures. In order to stack the playing field in favour of democracy and arbitrate disputes in calm, if not productive, scholars and practitioners agree that institutional design is the most important policy tool at our disposal (Barnes, 2001; Belmont et al. 2002; Horowitz, 2002; Norris, 2002). A sighting of wide comparative politics discussion on democracy and power sharing helps in evaluating the externally sponsored endeavour to create democratic institutions and the subsequent consolidation of democracy in post-conflict developing countries. The majority of the power sharing discourse is focused on the requirement to give institutional assurances and protections to groups within divided community, even multinational states. What insight the institutional prescriptions of power sharing can provide to post-conflict rebuilding efforts that are not primarily hampered by the problem of nationality in this construct in an effort to generalise from a body of literature that emphasises nationality as the fundamental political cleavage in a nation. For instance, the post-conflict democracy-building processes in Cambodia, East Timor, and Afghanistan are similar in that they place more emphasis on leaders’ efforts to hold onto power and their relationship to popular involvement than on the prominence of group identity in politics. Institutional patterns and power dynamics The way governments, organisations, and societies work is shaped by the interplay between power dynamics and institutional patterns. Establishing the ground rules and determining the allocation and Volume 5, Issue 6, November-December 2023 IJFMR23069907 3 ational Journal for Multidisciplinary Research (IJFMR) power relations change too. When power dynamics are analysed, it is necessary to look at how various actors or groups within institutions negotiate their interests, exercise influence, and deal with difficulties. In decision-making processes, institutional structures and power dynamics are vital. Which interests are given priority, how decisions are made, and who gets to speak up are all influenced by the official and unofficial conventions that exist inside organisations. Predicting and interpreting decision outcomes requires an understanding of power dynamics. It also highlights possible areas of disagreement or opposition within the framework. Addressing injustices and improving the efficiency of organisations and communities require an understanding of institutional patterns and power relations as well as a critical evaluation of them. Targeted interventions are made possible by identifying locations where power is concentrated or where particular people are marginalised. Restructuring institutional patterns and establishing more equal power relations can be aided by initiatives that support inclusivity, accountability, and transparency. As a result, any organisation or society must consider the interaction between institutional patterns and power dynamics. Comprehending these components in detail offers valuable perspectives on decision-making processes, the individuals with sway, and the distribution and contestation of power. It provides the framework for formulating policies that promote equity, diversity, and efficient government. rnational Journal for Multidisciplinary Research (IJFMR) Instead of competition among identity groups concerned about their security and political power after a conflict, the transition to democracy and its subsequent consolidation have focused on fights among elites in each case. The framework for organisational and societal functioning is provided by formal institutions, such as governmental entities, legal frameworks, and bureaucratic structures. The formal distribution of power is determined by the defined roles and hierarchies found in these institutions. Formal institutions can exhibit hierarchical, decentralised, or a hybrid pattern that influences the decision-making and implementation process. Under a federal system, for instance, authority may be distributed throughout several levels of government, or in a centralised government structure, power may be concentrated at the top. On the other hand, unwritten social conventions, rules, and practises that govern conduct inside a society or organisation are referred to as informal institutions. The dynamics of power can be greatly impacted by these unofficial patterns. Decision-making processes are influenced by social hierarchies, cultural norms, and networks of influence that frequently function outside the boundaries of official systems. Gaining a comprehensive understanding of power dynamics in a particular setting requires the recognition and comprehension of these informal institutions. The readings on power sharing does, however, provide insight into how legitimate governance systems might be developed in political systems that do not accept all-or-nothing decisions. Both theoretical arguments and actual information can be found in abundance in the description on power sharing and democracy. Understandably, power sharing is construed as a systems inclusively as the behaviours and frameworks that encourage broad-based governing coalitions that are typically inclusive of all significant mobilised groups in society (Sisk, 1996). The notion that institutions and practises can be put together in many ways to enhance democratic conflict management is highlighted by understanding power sharing in this way. The tactics used to exercise, negotiate, and challenge power inside a system are referred to as power dynamics. Politics, the economy, society, and culture are just a few of the ways that power can appear. Decision-making authority, policy beneficiaries, and potential marginalisation are all influenced by the allocation of power. As conditions change, society changes, and outside factors come into play, Volume 5, Issue 6, November-December 2023 4 References References 1. Acemoglu, D. and J. A. Robinson (2001). ‘A theory of political transitions’. American Economic Review. 91 (4): 938-963. 1. Acemoglu, D. and J. A. Robinson (2001). ‘A theory of political transitions’. American Economic Review. 91 (4): 938-963. 2. Anonymous. Nation building: Why some countries come together. Princeton University. Retrieved from https://assets.press.princeton.edu/chapters/i11197.pdf 2. Anonymous. Nation building: Why some countries come together. Princeton University. Retrieved from https://assets.press.princeton.edu/chapters/i11197.pdf 3. Barnes, S. H. (2001). ‘The contribution of democracy to rebuilding post conflict societies’. American Journal of International Law. 95: 86–101. 3. Barnes, S. H. (2001). ‘The contribution of democracy to rebuilding post conflict societies’. American Journal of International Law. 95: 86–101. 4. Belmont, K., Mainwaring, S. and Reynolds, A. (2002). ‘Introduction: institutional design, conflict management, and democracy’. In Reynolds, Bendix, Reinhard, eds., Nation building and Citizenship Berkeley: University of California Press. 4. Belmont, K., Mainwaring, S. and Reynolds, A. (2002). ‘Introduction: institutional design, conflict management, and democracy’. In Reynolds, Bendix, Reinhard, eds., Nation building and Citizenship Berkeley: University of California Press. 5. Besley, T. and T. Persson (2011). ‘The logic of political violence’. Quarterly Journal of Economics. 126: 1411–1445. 5. Besley, T. and T. Persson (2011). ‘The logic of political violence’. Quarterly Journal of Economics. 126: 1411–1445. 6. Burgess, R, R. Jedwab, E. Miguel, A. Morjaria and G. Padró i Miquel (2015). ‘The value of democracy: Evidence from road building in Kenya’. American Economic Review. 105: 1817–1851. 7. Cederman, L-E, K. Skrede Gleditsch and H. Buhaug (2013). Inequality, grievances and civil war. Cambridge University Press. 8. Encyclopedia of Violence, Peace, and Conflict. Third Edition. (2022). 9. Francois, P, I. Rainer and F. Trebbi (2015). ‘How is power shared in Africa?’. Econometrica. 83: 465– 503. 10. Gates, S, B. A. Graham, Y. Lupu, H. Strand and K. W. Strøm (2016). ‘Power sharing, protection, and peace’. The Journal of Politics. 78 (2): 512–526. 11. Horowitz, D. L. (2002). ‘Constitutional design: proposals versus processes’. In Bendix, Reinhard. Nation building and Citizenship. Berkeley: University of California Press. 15–36. 11. Horowitz, D. L. (2002). ‘Constitutional design: proposals versus processes’. In Bendix, Reinhard. Nation building and Citizenship. Berkeley: University of California Press. 15–36. 12. Kant, Immanuel (1983). Perpetual Peace, and Other Essays on Politics, History, and Morals. Hacket Publishing Company. 12. Kant, Immanuel (1983). Perpetual Peace, and Other Essays on Politics, History, and Morals. Hacket Publishing Company. 13. Lijphart (1999). International Journal for Multidisciplinary Research (IJFMR) E-ISSN: 2582-2160 ● Website: www.ijfmr.com ● Email: editor@ijfmr.com Decoding Nation-Building: Challenges, Cooperation, Prosperity A thorough grasp of the fundamental components that determine the fate of nations emerges from disentangling the complex web of power sharing, democratic ideals, and nation-building challenges. A robust and peaceful nation is created by skilfully weaving together the diverse strands of a nation, as demonstrated by the investigation of power sharing as a paradigm for nation-building. Democratic values are like a participatory governance symphony; they reverberate through the power-sharing corridors, highlighting the role that inclusion and representation play in the orchestration of a thriving society. But there are difficulties along the way. Nations have complicated environments that need to be carefully navigated due to institutional patterns and power relations. Power dynamics impact the fair allocation of opportunities and resources, and formal institutions and informal norms influence decision- making. As we delve into the heart of nation-building, it becomes evident that addressing these challenges is imperative for fostering stability, justice, and progress. Recognizing that nation-building is a continuous process requiring flexibility and a dedication to democratic values is crucial while taking on these problems. Formal or informal, the institutions we establish become the foundation of a nation. It becomes crucial to foster inclusive, accountable, and transparent institutions as a result. Furthermore, it is critical that we recognise past wrongs, promote social cohesiveness, and accept the complexity of variety as we defend power sharing and democratic principles. Conclusively, the dissection of power dynamics, democratic principles, and institutional challenges in the context of nation-building is a sophisticated exploration into the complex forces that shape the course of national history. It is a call to action, imploring societies to navigate these complexities with discernment and wisdom, understanding that real nation-building is a cooperative effort in which diversity’s threads are not only preserved but also celebrated and power is distributed in a way that secures the prosperity and well-being of all citizens. Volume 5, Issue 6, November-December 2023 IJFMR23069907 5 5 References Patterns of Democracy: Government Forms and Performance in Thirty- Six Countries. New Haven/London: Yale University Press. 14. Mueller, Hannes, Rauh, Christopher and Rohner, Dominic (2003). ‘Power sharing, conflict and state building Hannes’. In Nation Building Big Lessons from Successes and Failures, Dominic Rohner and Ekaterina Zhuravskaya (Ed). Centre for Economic Policy Research Press. London: UK. 15. Norris, P. (2002). ‘Ballots not bullets: testing consociational theories of ethnic conflict, electoral systems, and democratization’. In Reynolds, ed., Bendix, Reinhard. Nation building and Citizenship. Berkeley: University of California Press. 206–47. 16. Pinker, S. (2012). The better angels of our nature: Why violence has declined. Penguin Books. 17. Sisk, T. (1996). Power Sharing and International Mediation in Ethnic Conflicts. Washington DC: Institute of Peace Press. 18. Sisk, T. (2003). ‘Power-sharing after civil wars: matching problems to solutions’. In Darby and Ginty Mac, Contemporary Peacemaking. London: Macmillan. 139–50. 19. Stephenson, Carolyn (2005). Nation Building- Beyond Intractability. Retrieved from https://www.beyondintractability.org/essay/nation_building/%3B 20. Tolz, Vera (1998). ‘Forging the Nation: National Identity and Nation Building in Post-Communist Russia’. Europe-Asia Studies. 50 (6): 993–1022. Volume 5, Issue 6, November-December 2023 IJFMR23069907 6
https://openalex.org/W2766437424	https://pubs.rsc.org/en/content/articlepdf/2017/re/c7re00164a	English	null	Continuous direct anodic flow oxidation of aromatic hydrocarbons to benzyl amides	Reaction chemistry & engineering	2,017	cc-by	3,576	Open Access Article. Published on 02 November 2017. Downloaded on 10/24/2024 This article is licensed under a Creative Commons Attribution 3.0 Un DOI: 10.1039/c7re00164a rsc.li/reaction-engineering subsequent chemical reactions.9 It has previously been shown that replacing batch electrolysis with the continuous methods can reduce the formation of the overoxidation prod- ucts since the reactor is constantly replenished with the starting materials and the desired products are removed to a benign environment.7d The continuous production of benzyl amides by anodic oxidation in flow was developed. The stability and productivity of the equipment was examined over time and monitored by means of in-line UV analysis. The applicability of the method to twelve sub- strates was demonstrated. Methods for the site-selective C–H functionalisation of hydro- carbons, without the need of installing an additional func- tional group, are becoming more widely used.1 Their exploita- tion in multistep synthesis allows reducing the number of steps, increases atom efficiency and reduces waste.2 Neverthe- less, most of the modern C–H activation methods, although being usually chemo- and regioselective, require expensive and moisture- or air sensitive transition metal catalysts.3 The discovery reported here relates to the development of an electrochemical method to convert aromatic hydrocarbons into the benzyl amides. It is well known that alcohols can be converted into amides by the Ritter reaction – when reacting with nitriles and water under strong acidic conditions.10 This reaction proceeds through the formation of a carbocation which is sequentially trapped first by nitrile followed by a molecule of water. If instead the carbocation is generated by two-electron oxidation from the hydrocarbon, following the same trapping steps, the amide can be obtained analogously. This reaction has been observed when toluene and a few electron-rich toluene derivatives were oxidized electro- chemically at the anode in batch mode using acetonitrile as a solvent and an undivided cell.11 Unfortunately, these amides could only be isolated in low yields mainly due to overoxidation.11a Considering that flow electrochemistry methods are often advantageous over the batch alternatives in terms of synthetic applicability, efficiency and throughput, we focused on the development of a continuous direct anodic oxidation of aromatic hydrocarbons to benzyl amides. Department of Chemistry, University of Cambridge, Lensfield Road, CB2 1EW, UK. E-mail: svl1000@cam.ac.uk † Electronic supplementary information (ESI) available: The experimental methods, general procedure for the anodic oxidation, characterization of the syn- thesized compounds and copies of NMR spectra. See DOI: 10.1039/c7re00164a Reaction Chemistry & Engineering Reaction Chemistry & Engineering This journal is © The Royal Society of Chemistry 2017 Open Access Article. Published on 02 November 2017. Downloaded on 10/24/2024 6:19:25 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Cite this: React. Chem. Eng., 2017, 2, 822 Cite this: React. Chem. Eng., 2017, 2, 822 Mikhail A. Kabeshov, Biagia Musio and Steven V. Ley * Received 3rd October 2017, Accepted 2nd November 2017 Received 3rd October 2017, Accepted 2nd November 2017 Open Access Article. Published on 02 November 2017. Downloaded on 10/24/2024 6:19:25 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. a Black triangle represents input solution, white triangle represents output solution. b Determined by 1H NMR using the calibrated external standard (neat TMSCl contained in the inner tube of a coaxial NMR tube). electrolyte, electric current and the concentration of water (Table 1).13 desired product 3a (48% conversion; entry 9, Table 1). Apply- ing higher current did not improve the conversion due to the overoxidation side reactions (entries 12 and 13, Table 1). Un- der these new conditions, the system remained stable and fully homogeneous over time. First, the efficiency of LiBF4 as an electrolyte was exam- ined. It was shown that the presence of water was crucial for the reaction, since without water added only small amount of the product was detected (entry 1, Table 1). Stoichiometric amounts of water were found to be optimal as an excess led to a decrease in conversion to 3a (entries 2 and 4, Table 1). Similarly, the use of stoichiometric amount of LiBF4 also pro- vided the highest conversion to 3a (compare entries 5 and 6 with entry 3, Table 1). The lower conversion to 3a was ob- served at a higher current due to over-oxidation with the p- tert-butylbenzaldehyde formation (determined by 1H NMR in the crude reaction mixture) while some of the starting mate- rial 1a was still present (entry 7, Table 1).14 One possible reason of the higher efficiency of the system based on the LiBF4 electrolyte during the initial period, when compared to Bu4NPF6, could be the difference in the pH of the two reaction mixtures (Scheme 1). In the case of LiBF4, the lithium cation Li+ is initially read- ily reduced at the cathode not affecting the pH of the reac- tion mixture (red rectangle, Scheme 1). On the other hand, the anodic oxidation produces 1 eq. of H+ per 1 F of electric current passed, resulting in an acidic pH of the reaction mix- ture as confirmed experimentally (pH ∼0–1). Conversely, when Bu4NPF6 is used as an electrolyte, as 1 eq. of H+ per 1 F of electric current passed is produced in the same way at the anode, 1 eq. This journal is © The Royal Society of Chemistry 2017 Open Access Article. Published on 02 November 2017. Downloaded on 10/24/2024 This article is licensed under a Creative Commons Attribution 3.0 Un Electrochemical methods for direct oxidation and reduc- tion are increasingly important for modern synthetic chemis- try mainly because they employ electric current instead of more obnoxious agents.4 Electrochemistry is therefore be- coming more frequently used in the design of new atom-effi- cient, safe and more sustainable chemical processes.5 Also a range of products can sometimes be obtained from a com- mon starting material by controlling either the electric cur- rent or potential.6 Carrying out electrochemical experiments in a continuous manner increases greatly their throughput and general appli- cability.7 A higher surface-to-volume ratio, characteristic for micro- and mesofluidic conditions, causes better conductivity of the reaction media, resulting in a lower concentration of the ancillary electrolyte required.8 Furthermore, a shorter dis- tance between the anode and cathode enables the processes which involve highly unstable, reactive intermediates to pro- ceed smoothly and with high level of selectivity due to fast All the experimental studies were performed using an Am- monite® 8 reactor (V = 1 mL)12 equipped with a platinum disk anode, circular stainless steel cathode, and perfluoroelastomer FFKM gasket. This unit, with the carbon polymer anode, was previously applied to the electrochemical methoxylation of N-formylpyrrolidine and p-methoxybenzyl al- cohol deprotection allowing high conversion of the substrate in a single run.13 The reaction conditions for the continuous electrolysis at the constant current of p-tert-butyltoluene 1a in acetonitrile (0.1 M) as a model substrate were explored varying 822 \| React. Chem. Eng., 2017, 2, 822–825 This journal is © The Royal Society of Chemistry 2017 Reaction Chemistry & Engineering 1 Anodic oxidation of p-tert-butyltoluene 1a in acetonitrile (0.1 M; flow rate 500 μL min−1)a Table 1 Anodic oxidation of p-tert-butyltoluene 1a in acetonitrile (0.1 M; flow rate 500 μL min−1)a Table 1 Anodic oxidation of p-tert-butyltoluene 1a in acetonitrile (0.1 M; flow rate 500 μL min−1)a Table 1 Anodic oxidation of p-tert-butyltoluene 1a in acetonitrile (0.1 M; flow rate 500 μL min−1)a Entry [2], M [H2O], M Current, A (charge, F) Conversion 3a,b % 1 LiBF4, 0.1 None 0.20 (2.5) 8 2 LiBF4, 0.1 0.05 0.20 (2.5) 60 3 LiBF4, 0.1 0.1 0.20 (2.5) 66 4 LiBF4, 0.1 0.2 0.20 (2.5) 56 5 LiBF4, 0.05 0.1 0.20 (2.5) 54 6 LiBF4, 0.15 0.1 0.20 (2.5) 59 7 LiBF4, 0.1 0.1 0.23 (2.9) 57 8 Bu4NPF6, 0.01 None 0.20 (2.5) 7 9 Bu4NPF6, 0.01 0.1 0.20 (2.5) 48 10 Bu4NPF6, 0.01 0.12 0.20 (2.5) 46 11 Bu4NPF6, 0.01 0.2 0.20 (2.5) 43 12 Bu4NPF6, 0.01 0.1 0.23 (2.5) 46 13 Bu4NPF6, 0.01 0.1 0.26 (3.3) 45 a Black triangle represents input solution, white triangle represents output solution. b Determined by 1H NMR using the calibrated external standard (neat TMSCl contained in the inner tube of a coaxial NMR tube). Open Access Article. Published on 02 November 2017. Downloaded on 10/24/2024 6:19:25 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. The continuous flow electrolysis could not be performed in pure acetonitrile as a solvent and toluenesulfonic acid (pKa = −2.8) as an additive due to a solid formation (entry 1, Table 2). The addition of dichloromethane as a co-solvent (25% vol) was beneficial to keep the system homogeneous, but the amide 3a was still obtained with low conversion (48% conversion; entry 2, Table 2). When using trifluoroacetic acid (TFA; pKa = 0.23), a higher conversion to 3a was observed (68% conversion; entry 3, Table 2). This result could be fur- ther improved by using methanesulfonic acid (MSA; pKa = −1.9; 72% conversion; entry 4, Table 2). Changing the concen- tration of MSA did not lead to any further improvement in the formation of 3a (entries 5 and 6, Table 2). Under the optimised conditions (entry 4, Table 2) the system remained fully homogeneous over the time. Next, the stability of the electrolytic system was investi- gated (Fig. 1). Thus a mixture of the hydrocarbon 1a (0.1 M), H2O (0.1 M), MSA (0.1 M) and Bu4NPF6 (0.01 M) in acetoni- trile was continuously passed (500 μL min−1) through the re- actor applying constant current of 0.20 A (2.5 F). Fig. 1 Anodic oxidation of p-tert-butyltoluene 1a monitored by Flow- UV™. ucts 3 under strong acidic conditions. The evaporation of sol- vent in vacuo was sufficient to obtain the crude products 3 (Scheme 2). An online steady state monitoring was best achieved by means of a UV/vis detector (Flow-UV™)15 which was installed after the reactor (Fig. 1). The reaction pleasingly proceeded with the constant productivity of the amide 3a equal to 0.38 g h−1 over 9 hours (Fig. 1). The amide derivatives of various para- (3a–3g; 3k–3l), meta- (3i) and ortho-substituted (3j) benzene derivatives were obtained in good yields. The anodic oxidation was successful for the preparation of either moderately electron-rich (3a, 3c), electron-neutral (3b, 3f, 3k) or electron-poor (3d, 3f–3j, 3l) Once the stability and the applicability of the developed procedure to a large scale production were demonstrated, the generality of the methodology over a range of aromatic hydro- carbons was studied (Scheme 2). To simplify the downstreaming process, the efflux of the reactor was mixed with a second stream (500 μL min−1) containing an ammonia solution in methanol (0.7 M). Open Access Article. Published on 02 November 2017. Downloaded on 10/24/2024 6:19:25 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. of H+ per 1 F is consumed at the cathode, maintaining the overall neutral pH of the reaction mixture Under the optimum conditions when using LiBF4 as an electrolyte (entry 3, Table 1), it was found that the system was not stable over time and that the conversion towards the am- ide 3a deteriorated significantly (from 66% to 38% in 3 hours). The formation of an inorganic solid inside the cell channel was observed over the time that caused clogging and voltage instability. This solid, which reacted violently with water with the formation of hydrogen gas, is likely to contain lithium metal in agreement with the previous reports.11c Sub- sequent loss of the current efficiency was therefore the most probable reason for the decreased conversion over time. Scheme 1 Electrochemical processes for the systems using LiBF4 and Bu4NPF6 as electrolytes. In order to find sustainable alternative conditions, the an- odic oxidation of p-tert-butyltoluene 1a was optimised using an electroneutral electrolyte Bu4NPF6 (entries 8–13, Table 1). As in the case of LiBF4, a stoichiometric amount of water was found to be optimal for the experiments with Bu4NPF6 (en- tries 8–11, Table 1) providing the highest conversion to the Scheme 1 Electrochemical processes for the systems using LiBF4 and Bu4NPF6 as electrolytes. React. Chem. Eng., 2017, 2, 822–825 \| 823 823 View Article Online Fig. 1 Anodic oxidation of p-tert-butyltoluene 1a monitored by Flow- UV™. Reaction Chemistry & Engineering Communication Reaction Chemistry & Engineering (green rectangle, Scheme 1). Knowing that the reduction potential EĲ2H+/H2) at the cathode is proportional to lnĲ[H+]) according to the Nernst equation, the reduction of H+ to H2 and the overall electrolysis should proceed easier and at a lower voltage in the presence of a Brønsted acid. With this information available, the influence of various Brønsted acid additives on the anodic oxidation of p-tert- butyltoluene 1a was explored (Table 2). Open Access Article. Published on 02 November 2017. Downloaded on 10/24/2024 6:19:25 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Full neutralisation of the acid was beneficial to reduce decomposition of the amide prod- Scheme 2 Scope of the anodic oxidation of aromatic hydrocarbons (isolated yields are listed). Table 2 Anodic oxidation of p-tert-butyltoluene 1a in acetonitrile (0.1 M) using Bu4NPF6 as an electrolyte and Brønsted acid additivesa Entry Acid additive [Acid], M Conversion 3a,b % 1 TsOH·2H2O 0.1 N/A 2c TsOH·2H2O 0.1 48 3 TFA 0.1 68 4 MSA 0.1 72 (63) 5 MSA 0.05 60 6 MSA 0.13 70 a cĲ1a) = cĲH2O) = 0.1 M, cĲBu4NPF6) = 0.01 M, flow rate 500 mL min−1. b Determined by 1H NMR using the calibrated external standard (neat TMSCl contained in the inner tube of a coaxial NMR tube); isolated yield in brackets. c Solvent system: acetonitrile : dichloromethane = 3 : 1. Table 2 Anodic oxidation of p-tert-butyltoluene 1a in acetonitrile (0.1 M) using Bu4NPF6 as an electrolyte and Brønsted acid additivesa Table 2 Anodic oxidation of p-tert-butyltoluene 1a in acetonitrile (0.1 M) using Bu4NPF6 as an electrolyte and Brønsted acid additivesa a cĲ1a) = cĲH2O) = 0.1 M, cĲBu4NPF6) = 0.01 M, flow rate 500 mL min−1. b Determined by 1H NMR using the calibrated external standard (neat TMSCl contained in the inner tube of a coaxial NMR tube); isolated yield in brackets. c Solvent system: acetonitrile : dichloromethane = 3 : 1. a cĲ1a) = cĲH2O) = 0.1 M, cĲBu4NPF6) = 0.01 M, flow rate 500 mL min−1. b Determined by 1H NMR using the calibrated external standard (neat TMSCl contained in the inner tube of a coaxial NMR tube); isolated yield in brackets. c Solvent system: acetonitrile : dichloromethane = 3 : 1. Scheme 2 Scope of the anodic oxidation of aromatic hydrocarbons (isolated yields are listed). Scheme 2 Scope of the anodic oxidation of aromatic hydrocarbons (isolated yields are listed). This journal is © The Royal Society of Chemistry 2017 This journal is © The Royal Society of Chemistry 2017 824 \| React. Chem. Eng., 2017, 2, 822–825 View Article Online Reaction Chemistry & Engineering Communication derivatives. A number of functional groups were tolerated by the method: such as halogen (3d–3f, 3k–3l), nitrile (3g, 3j), ketone (3i), ester (3h). The synthesis of both primary (3a–3j) and secondary (3k–3l) benzyl amides was also successfully achieved. In all the cases, the amides 3 were isolated as the major products. Notes and references 1 (a) H. Yi, G. Zhang, H. Wang, Z. Huang, J. Wang, A. K. Singh and A. Lei, Chem. Rev., 2017, 117, 9016; (b) T. Gensch, M. N. Hopkinson, F. Glorius and J. Wencel-Delord, Chem. Soc. Rev., 2016, 45, 2900; (c) S. M. Paradine, J. R. Griffin, J. Zhao, A. L. Petronico, S. M. Miller and M. C. White, Nat. Chem., 2015, 7, 987. 1 (a) H. Yi, G. Zhang, H. Wang, Z. Huang, J. Wang, A. K. Singh and A. Lei, Chem. Rev., 2017, 117, 9016; (b) T. Gensch, M. N. Hopkinson, F. Glorius and J. Wencel-Delord, Chem. Soc. Rev., 2016, 45, 2900; (c) S. M. Paradine, J. R. Griffin, J. Zhao, A. L. Petronico, S. M. Miller and M. C. White, Nat. Chem., 2015, 7, 987. 11 (a) I. N. Rozhkov and I. Y. Alyev, Tetrahedron, 1975, 31, 977; (b) L. Eberson and B. Olofsson, Acta Chem. Scand., 1969, 23, 2355; (c) E. A. Mayeda and L. L. Miller, Tetrahedron, 1972, 28, 3375; (d) T. Tajima, H. Ishii and T. Fuchigami, Electrochem. Commun., 2002, 4, 589. 12 Ammonite by Cambridge reactor Design (CRD): http://www. cambridgereactordesign.com/ammonite/ammonite.html (last visited on 19/09/17). 2 (a) J. Wencel-Delord and F. Glorius, Nat. Chem., 2013, 5, 369; (b) D. Y.-K. Chen and S. W. Youn, Chem. – Eur. J., 2012, 18, 9452; (c) C. Leitner and T. Gaich, Chem. Commun., 2017, 53, 7451. 13 (a) R. A. Green, R. C. D. Brown, D. Pletcher and B. Harji, Electrochem. Commun., 2016, 73, 63; (b) R. A. Green, R. C. D. Brown and D. Pletcher, Org. Process Res. Dev., 2015, 19, 1424; (c) R. A. Green, K. E. Jolley, A. A. M. Al-Hadedi, D. Pletcher, D. C. Harrowven, O. De Frutos, C. Mateos, D. J. Klauber, J. A. Rincón and R. C. D. Brown, Org. Lett., 2017, 19, 2050. 3 (a) K. S. Egorova and V. P. Ananikov, Angew. Chem., Int. Ed., 2016, 55, 12150; (b) F. Rodesly, J. Oble and G. Poli, J. Mol. Catal. A: Chem., 2017, 426, 275; (c) N. V. Tzouras, I. K. Stamatopoulos, A. T. Papastavrou, A. A. Liori and G. C. Vougioukalakis, Coord. Chem. Rev., 2017, 343, 25. 4 (a) M. Yan, Y. Kawamata and P. S. Baran, Angew. Chem., Int. Ed., 2017, DOI: 10.1002/anie.201707584, ASAP. Conflicts of interest There are no conflicts or financial interest to declare. Open Access Article. Published on 02 November 2017. Downloaded on 10/24/2024 This article is licensed under a Creative Commons Attribution 3.0 Un 8 8 (a) R. A. Green, R. C. D. Brown and D. Pletcher, J. Flow Chem., 2016, 6, 191; (b) M. Atobe, H. Tateno and Y. Matsumura, Chem. Rev., 2017, DOI: 10.1021/acs. chemrev.7b0353, ASAP; (c) Y. Liu and X. Jiang, Lab Chip, 2017, DOI: 10.1039/c7lc00627f, ASAP. Acknowledgements 9 M. A. Kabeshov, B. Musio, P. R. D. Murray, D. L. Browne and S. V. Ley, Org. Lett., 2014, 16, 4618. The authors are grateful to EPSRC (grants EP/K009494/1 and EP/K039520/1) for financial support. The authors are also grateful to Dr Bashir Harji (Cambridge Reactor Design Ltd) for technical support and assistance. 10 (a) D. Jiang, T. He, L. Ma and Z. Wang, RSC Adv., 2014, 4, 64936; (b) A. Guérinot, S. Reymond and J. Cossy, Eur. J. Org. Chem., 2012, 19; (c) J. J. Ritter and P. P. Minieri, J. Am. Chem. Soc., 1948, 70, 4045; (d) R. Sanz, A. Martinez, V. Guilarte, J. M. Alvarez-Gutierrez and F. Rodriguez, Eur. J. Org. Chem., 2007, 4642. Article. Published on 02 November 2017. Downloaded on 10/24/2024 6:19:25 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Open Access Article. Published on 02 November 2017. Downloaded on 10/24/2024 6:19:25 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. 7 (a) D. S. P. Cardoso, B. Šljukić, D. M. F. Santos and C. A. C. Sequeira, Org. Process Res. Dev., 2017, 21, 1213–1226; (b) K. Watts, W. Gattrell and T. Wirth, Beilstein J. Org. Chem., 2011, 7, 1108; (c) S. Suga, M. Okajima, K. Fujiwara and J.-I. Yoshida, J. Am. Chem. Soc., 2001, 123, 7941; (d) J. Kuleshova, J. T. Hill-Cousins, P. R. Birkin, R. C. D. Brown, D. Pletcher and T. Underwood, Electrochim. Acta, 2012, 69, 197; (e) R. Hayashi, A. Shimizu, Y. Song, Y. Ashikari, T. Nokami and J.-I. Yoshida, Chem. – Eur. J., 2017, 23, 61; ( f ) D. Pletcher, R. A. Green and R. C. D. Brown, Chem. Rev., 2017, DOI: 10.1021/acs.chemrev.7b00360, ASAP. Open Access Article. Published on 02 November 2017. Downloaded on 10/24/2024 6:19:25 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Residual starting materials were the second major components of all the reaction mixtures while only traces of other by-products, such as substituted benzalde- hydes or benzyl alcohols, were minor components (<2%). E. J. Horn, B. R. Rosen and P. S. Baran, ACS Cent. Sci., 2016, 2, 302; (e) A. G. O'Brien, A. Maruyama, Y. Inokuma, M. Fujita and P. S. Baran, Angew. Chem., Int. Ed., 2014, 53, 11868; ( f ) H.-B. Zhao, Z.-J. Liu, J. Song and H.-C. Xu, Angew. Chem., Int. Ed., 2017, 56, 12732; (g) X.-Y. Qian, S.-Q. Li, J. Song and H.-C. Xu, ACS Catal., 2017, 7, 2730. 6 G. Laudadio, N. J. W. Straathof, M. D. Lanting, B. Knoops, V. Hessel and T. Noël, Green Chem., 2017, 19, 4061. Conclusions In conclusion, the preparation of benzyl amides from a num- ber of aromatic hydrocarbons was achieved by a stable and continuous flow anodic oxidation. The addition of a Brønsted acid was crucial to maintain the process stability over 9 hours with the productivity of 9.12 g per day. The results described here open up to the application of electrochemical methodol- ogies to large-scale production. Notes and references 14 Lower flow rates with the current adjusted did not affect conversion to 3a whereas the concentration of the substrate 1a is reported to be optimal for 2e−oxidation processes using the Ammonite 8 set-up. 5 (a) K. D. Moeller, Synlett, 2009, 8, 1208; (b) B. R. Rosen, E. W. Werner, A. G. O'Brien and P. S. Baran, J. Am. Chem. Soc., 2014, 136, 5571; (c) J. Yoshida, K. Kataoka, R. Horcajada and A. Nagaki, Chem. Rev., 2009, 108, 2265; (d) 15 Flow-UV by Uniqsis Ltd: http://www.uniqsis.com/ paProductsDetail.aspx?ID=Flow-UV (last visited on 19/09/17). React. Chem. Eng., 2017, 2, 822–825 \| 825 This journal is © The Royal Society of Chemistry 2017
https://openalex.org/W4285029016	https://www.researchsquare.com/article/rs-1309115/latest.pdf	English	null	Enhanced Thermoelectric Properties of Bi<sub>2</sub>Te<sub>3</sub>‐Based Micro–Nano Fibers via Thermal Drawing and Interfacial Engineering	Advanced materials	2,022	cc-by	7,651	Enhanced thermoelectric properties of Bi2Te3- based micro-nano ¦bers via thermal drawing and interfacial engineering Min Sun South China University of Technology https://orcid.org/0000-0002-7494-3584 Guowu Tang South China University of Technology Hanfu Wang National Center for Nanoscience and Technology Ting Zhang Nanyang Technological University https://orcid.org/0000-0001-5967-0525 Pengyu Zhang South China University of Technology Bin Han South China University of Technology Ming Yang Institute of Engineering Thermophysics, Chinese Academy of Science, Beijing 100190, P. R. China Hang Zhang Chinese Academy of Sciences Yicong Chen Sun Yat-sen University Jun Chen School of Electronics and Information Technology, Sun Yat-sen University https://orcid.org/0000-00 7397-2714 Dongdan Chen South China University of Technology Jiulin Gan State Key Laboratory of Luminescent Materials and Devices, Institute of Optical Communication Materials, South China University of Technology https://orcid.org/0000-0001-9842-0510 Qi Qian South China University of Technology Zhongmin Yang (  yangzm@scut.edu.cn ) State Key Laboratory of Luminescent Materials and Devices, South China University of Technology, h h Min Sun Bin Han South China University of Technology Hang Zhang Chinese Academy of Sciences Yicong Chen Sun Yat-sen University Dongdan Chen South China University of Technology Dongdan Chen South China University of Technology Zhongmin Yang (  yangzm@scut.edu.cn ) State Key Laboratory of Luminescent Materials and Devices, South China University of Technology, Guangzhou 510640, China Zhongmin Yang (  yangzm@scut.edu.cn ) State Key Laboratory of Luminescent Materials and Devices, South China University of Technology, Guangzhou 510640, China Enhanced thermoelectric properties of Bi2Te3-based micro-nano fibers via thermal drawing and interfacial engineering Min Sun1,2,#, Guowu Tang1,2,3,#, Hanfu Wang4,#, Ting Zhang5,6, Pengyu Zhang2, Bin Han2, Ming Yang5, Hang Zhang5, Yicong Chen7, Jun Chen7, Dongdan Chen2, Jiulin Gan2, Qi Qian2,, and Zhongmin Yang1,2, Min Sun1,2,#, Guowu Tang1,2,3,#, Hanfu Wang4,#, Ting Zhang5,6, Pengyu Zhang2, Bin Han2, Ming Yang5, Hang Zhang5, Yicong Chen7, Jun Chen7, Dongdan Chen2, Jiulin Gan2, Qi Qian2,, and Zhongmin Yang1,2, 1 School of Physics and Optoelectronics, South China University of Technology, Guangzhou 510640, China 2 State Key Laboratory of Luminescent Materials and Devices, Institute of Optical Communication Materials, Guangdong Provincial Key Laboratory of Fiber Laser Materials and Applied Techniques, and Guangdong Engineering Technology Research and Development Center of Special Optical Fiber Materials and Devices, School of Materials Science and Engineering, South China University of Technology, Guangzhou 510640, China. 3 School of Physics and Optoelectronic Engineering, Guangdong University of Technology, Guangzhou 510006, China. 4CAS Key Laboratory of Nanosystem and Hierarchical Fabrication, National Center for Nanoscience and Technology of China, Beijing 100190, China. 5 Institute of Engineering Thermophysics, Innovation Academy for Light-duty Gas Turbine, Chinese Academy of Sciences, Beijing 100190, China. 5 Institute of Engineering Thermophysics, Innovation Academy for Light-duty Gas Turbine, Chinese Academy of Sciences, Beijing 100190, China. 6 Nanjing Institute of Future Energy System, Nanjing 211135, China 7 State Key Laboratory of Optoelectronic Materials and Technologies, Guangdong Province Key Laboratory of Display Material and Technology, School of Electronics and Information Technology, Sun Yat-sen University, Guangzhou, 510275 China. # The authors contributed equally: Min Sun, Guowu Tang, Hanfu Wang. * E-mail: qianqi@scut.edu.cn; yangzm@scut.edu.cn. 6 Nanjing Institute of Future Energy System, Nanjing 211135, China State Key Laboratory of Optoelectronic Materials and Technologies, Guangdong Province Key Laboratory of Display Material and Technology, School of Electronics and Information Technology, Sun Yat-sen University, Guangzhou, 510275 China. # The authors contributed equally: Min Sun, Guowu Tang, Hanfu Wang. * E-mail: qianqi@scut.edu.cn; yangzm@scut.edu.cn. DOI: https://doi.org/10.21203/rs.3.rs-1309115/v1 DOI: https://doi.org/10.21203/rs.3.rs-1309115/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Abstract: High-performance thermoelectric (TE) materials with great flexibility and stability are urgently needed to efficiently convert heat energy into electrical power. Recently, intrinsically crystalline, mechanically stable, and flexible inorganic TE fibers that show TE properties comparable to their bulk counterparts are of interest to researchers. Despite remarkable progress moving TE fibers towards room-temperature TE conversion, the figure-of-merit value (ZT) and bending stability still need enhancement. Herein, we report interfacial engineering enhanced thermoelectric properties of micro-nano polycrystalline TE fibers fabricated by thermally drawing Bi2Te3-based bulks in a glass-fiber template. The interfacial engineering effect comes from generating stress-induced oriented nanocrystals to increase electrical conductivity and producing strain-distorted interfaces to decrease thermal conductivity. The resulting fibers achieve a 40% higher ZT (~1.4 at 300 K) than their 1 / 20 bulk counterparts and show a reversible bending radius of 50 µm, approaching the theoretical elastic limit. This fabrication strategy works for a wide range of inorganic TE materials and benefits the development of fiber-based micro-TE devices. bulk counterparts and show a reversible bending radius of 50 µm, approaching the theoretical elastic limit. This fabrication strategy works for a wide range of inorganic TE materials and benefits the development of fiber-based micro-TE devices. Keywords: thermoelectric fibers; Bi2Te3; thermal drawing; interfacial engineering Keywords: thermoelectric fibers; Bi2Te3; thermal drawing; interfacial engineering Keywords: thermoelectric fibers; Bi2Te3; thermal drawing; interfacial engineering 1. Introduction 2 / 20 As achieving carbon-dioxide-emission peak and carbon neutrality are significant goals for alleviating global warming and energy problems, an effective solution is to recycle the distributed waste heat energy at medium-or-low temperatures (300–600 K). The capacities of direct conversion between heat energy and electricity have enabled TE devices to achieve power generation and solid-state cooling, in a zero-carbon and sustainable way.1-3 The efficiency of TE materials is a key factor for the waste heat energy conversion, which is gauged by its dimensionless figure-of-merit value, ZT=S2σT/κ, where S is the Seebeck coefficient, σ is the electrical conductivity, T is the absolute temperature, and κ is the thermal conductivity. However, the ZT value of commercial Bi2Te3-based materials is restricted to ~1 at room temperature owing to the adverse interdependence of the Seebeck coefficient, electrical and thermal conductivity.4,5 Since Bi2Te3-based bulk materials with nanostructure have been demonstrated to be one of the best room-temperature TE materials with the highest ZT~1.4 at 300 K,6-8 they possess large-scale application prospects for TE generators and refrigerators. Moreover, the bulk TE devices are usually made of mechanically cut Bi2Te3-based pellets, whose elastic strain maxima are experimentally reported to be half lower than the theoretical limit of 5.5%.9 The TE devices can only offer limited flexibility, which makes them difficult to be used on small heat sources with curved shapes.10-12 Current solutions mainly focus on embedding/coating the Bi2Te3-based materials into/on flexible substrates such as yarns, textiles, films, paintings, and papers.13-17 However, low TE performance, weak mechanical stability, and complex fabrication process still impede their applications. Hence, simultaneously improving ZT and flexibility of Bi2Te3-based TE materials is a great 2 / 20 challenge, thus it is necessary to explore new strategies to meet the demand. Over the past 50 years, optical fibers have been an extremely successful platform for optical materials and devices research. 1. Introduction The internet and most telecommunication are delivered across the globe today through billions of kilometers of silica fibers, which are typically produced by thermally drawing a preform at the softening temperature of silica glass.18 With the same facile technique, multiple materials with disparate properties, including semiconductors, metals, and insulators, can be co-drawn into micro-/nanofibers, which paves the way to a new generation of fibers endowed with unique functionalities at fiber length scales and costs.19,20 Based on these studies, ultralong and flexible TE fibers (Bi2Te3, SnSe, etc.) have been fabricated by thermal drawing crystalline TE materials in a glass-fiber preform.21-26 In addition, the TE properties of the Bi2Te3 TE fibers can be enhanced due to the preferential orientation of the Bi2Te3 polycrystalline nanosheets in the core, which were induced by the thermal stress during the fiber drawing process.27 However, some disadvantages of the reported TE fibers need to be improved, such as large bending radius, limited ZT value at bulk counterpart level, and low power density for thick glass cladding. On the other hand, the characterization of the comprehensive TE properties (electrical conductivity, Seebeck coefficient, and thermal conductivity) of a single micro-nano fiber is still challenging due to factors such as difficulties in forming good electrical/thermal contact and reducing heat radiation loss.28 In this work, we fabricated micro-nano Bi2Te3-based fibers with diameters ranging from 0.1 to 5 µm based on a two-step thermal drawing method. Moreover, a measuring chip was made to characterize the TE properties of fibers and the relative measuring errors were evaluated. Interestingly, we found an interface engineering effect during the fiber drawing process by characterizing the microstructure of microfiber interfaces. The fiber cores were then tuned into oriented nanosheet crystals with rough interfaces, showing increased electrical transport and significantly decreased thermal transport. The ZT value of 4-µm-diameter Bi2Te3-based fibers was measured to be 1.4 at 300 K, which is 40% higher than that of their bulk counterpart. Furthermore, the fibers exhibit mechanical stability at a 3 / 20 curvature radius of 50 µm, which is three orders of magnitude smaller than the reported inorganic TE fibers,21-24 approaching the theoretical limit of ~5.5% elastic strain for Bi2Te3 along the c axis at 300 K.9 This work benefits the further development of high-performance and highly flexible TE fibers towards widespread applications in energy harvesting and thermal management. 2.1 Micro-nano thermoelectric fibers fabricated via thermal drawing 2.1 Micro-nano thermoelectric fibers fabricated via thermal drawing 2.1 Micro nano thermoelectric fibers fabricated via thermal drawing Figure 1. (a) Schematic of micro-nano TE fibers prepared by the glass-fiber-template thermal drawing method. (b) Schematic of the defect structure in the fiber core. (c) Thermal conductivities and (d) ZT values of the p-type Bi2Te3/BST bulks and fibers at 300 K. Figure 1. (a) Schematic of micro-nano TE fibers prepared by the glass-fiber-template thermal drawing method. (b) Schematic of the defect structure in the fiber core. (c) Thermal conductivities and (d) ZT values of the p-type Bi2Te3/BST bulks and fibers at 300 K. Figure 1a shows a schematic of thermal drawing Bi2Te3-based bulks into fibers. Micro-nano TE fibers with Bi2Te3 or Bi0.5Sb1.5Te3 (BST) core and borosilicate cladding were fabricated by thermal drawing preforms. Especially, the preforms have been endured 700 ℃ for a two-hour preheating time before drawing, to form a wetting interface between the core and the cladding by elemental diffusion (see Methods and 4 / 20 Supplementary Figure S1). The scanning electron microscope (SEM) and energy disperse spectroscope (EDS) were carried out on the micro-nano fiber cross-section with cladding and the hydrogen fluoride (HF) etched fiber core without cladding (Supplementary Figure S2). It shows that the average nanosheet thickness in the fiber core is ~20 nm. Furthermore, the fibers undergo rapid cooling with a rate of ~150℃/s right after thermal drawing, which should cause an interfacial defect as our reported Bi2Te3 fibers after preheating and fast drawing.23 The defect illustration of the fiber cores is exhibited in Figure 1b, which shows that the carrier charges can quickly transport in the oriented nanosheets and multiscale defects can hinder phonon transport. The TE properties parallel to the fiber-axis direction of the fibers were studied. Figure 1c-1d shows the thermal conductivities (κ) and ZT values of the Bi2Te3 fibers (BTF), Bi0.5Sb1.5Te3 fibers (BSTF), and their corresponding bulks. The κ of the BTF and BSTF are 50% lower than that of the corresponding bulks. Importantly, the BSTF with a diameter of 4 µm shows an ultra-low κ of 0.57 W/m·K and a high ZT of ~1.4 at 300 K. 2.1 Micro-nano thermoelectric fibers fabricated via thermal drawing The micro-nano fiber without glass cladding should possess a higher output power density than the reported mW/cm2 level of TE fibers with glass cladding.22 To highlight the advantages of the micro-nano TE fibers, the thermal conductivities and the ZT values of other state-of-the-art Bi2Te3-based fibers are also compared in Figure 1c-1d.22,23,29 The enhanced TE properties and the corresponding mechanism of the interfacial engineering effect will be discussed as follows. 2.2 Interfacial microstructure 2.2 Interfacial microstructure 5 / 20 Figure 2. Linear elemental analysis on the cross-section of (a) BTF and (b) BSTF. Three dimensional AFM images of the surfaces of the BSTF cores with different diameters: (c) 100 µm; (d) 20 µm; (e) 5 µm; (f) 1 µm. Figure 2. Linear elemental analysis on the cross-section of (a) BTF and (b) BSTF. Three dimensional AFM images of the surfaces of the BSTF cores with different diameters: (c) 100 µm; (d) 20 µm; (e) 5 µm; (f) 1 µm. Fi 2 Li l l l i h Figure 2. Linear elemental analysis on the cross-section of (a) BTF and (b) BSTF. Three dimensional AFM images of the surfaces of the BSTF cores with different diameters: (c) 100 µm; (d) 20 µm; (e) 5 µm; (f) 1 µm. To understand the mechanism of the ultra-low thermal conductivities and high electrical conductivities in the micro-nano fibers, further analysis was performed on the microfiber by electro-probe microanalyzer (EPMA), atomic force microscopy (AFM), and transmission electron microscopy (TEM). The detailed information can be found in Methods. Figure 2a-2b exhibits the linear elemental (Bi, Sb, Te, Si, and O) profiles across the polished interface of the BTF and BSTF by EPMA measurement. There exists little diffusion of Bi, Sb, and Te into the cladding, while there is some diffusion of Si and O into the core region. The diffused elements should form some nanoscale oxides in the core, such as SiO2 and TeO2, which will have strong phonons scattering, resulting in lower thermal conductivity as in our previous studies.23 The elemental diffusion introduces an oxidic surface of the fiber core, which could protect the fiber core from atmospheric oxidation. 6 / 20 Figure 2c-2f shows the surface roughness of the BSTF cores with diameters of 100, 20, 5, and 1 µm, respectively. When the diameter is 100 µm, the surface of the as-prepared fiber core is relatively flat, however, it becomes rougher as the core 6 / 20 diameter decreases. When the diameter is 1 µm, the height of the humps dramatically increases. The surface roughness was measured to be 9, 33, 64, and 78 nm, corresponding to the fiber diameters of 100, 20, 5, and 1 µm, respectively. The enhancement in the surface roughness with the decreasing core diameter could be attributed to the increased depth of elemental diffusion as mentioned above. 2.2 Interfacial microstructure Notably, besides the well-recognized classical size effect due to the phonon boundary scattering, the surface roughness can also significantly alter the thermal conductivity.30 Figure 3. (a) Oriented crystal growth model of the Bi2Te3 fiber core via thermal drawing. (b) HAADF image of a BSTF chip and (c)-(e) enlarged view of the circled region in (b). The insets of (c)-(e) are corresponding SAED patterns. Figure 3. (a) Oriented crystal growth model of the Bi2Te3 fiber core via thermal drawing. (b) HAADF image of a BSTF chip and (c)-(e) enlarged view of the circled region in (b). The insets of (c)-(e) are corresponding SAED patterns. To reveal the growth process of the oriented nanosheet crystal in the fiber confined field, which is shown in Figure 1 and Supplementary Figure S2, the morphology evolution during the fiber-drawing process is schematically shown in Figure 3a. During the cooling process of the first thermal drawing (part i), the Bi2Te3 melt gradually nucleates and grows into hexagonal nanosheets around the interface to minimize potential energy. The nanosheet crystals will continuously stack around the interface and then the crystals grow along the (0 0 l) plane and compete in several 7 / 20 7 / 20 directions into a flower shape (Supplementary Figure S2a). More interestingly, during the second drawing process (part ii), the nanosheet crystals stack around the interface at the beginning. Accompanied by smaller fiber-core space and larger interface stress than the first drawing process, then the crystals preferentially grow along the (0 0 l) plane in one direction (Supplementary Figure S2b). Meanwhile, elemental diffusion takes place between the cladding and the core, which causes an increase in the surface roughness of the fiber core as found above. Figure 3b presents a HAADF image of a BSTF sample processed by using a focused ion beam (FIB) method, including the upside core and the underside cladding of the fiber. Figures 3c-3e show the enlarged views of the circled region in Figure 3b, from the interface to the core of 0, 0.5, and 1 µm depth. The insects of Figures 3c-3e present the corresponding images of selected area electron diffraction (SAED). These measurements confirm the crystalline state of Bi0.5Sb1.5Te3 in the fiber core. Figure 3c shows that the core nearby the interface is polycrystalline with a lattice-plane spacing of 0.317 nm, corresponding to the (0 1 5) plane. 2.2 Interfacial microstructure It is noted that the core in 1 µm depth away from the interface is an ordered polycrystalline state, as shown in Fig. 3d. In addition, it includes diverse dislocations and strain distortions (Supplementary Figure S3), which is similar to the reported mosaic crystal.31 And the insert of Figure 3d marks the corresponding lattice planes, including strong (0 0 15) and weak (3 0 0). Moreover, Figure 3e shows a single-crystal state with clear (3 0 0) planes, suggesting the direction of the Bi0.5Sb1.5Te3 (3 0 0) planes are parallel to the fiber axis. These results verify the effect of interfacial engineering on the ordered and defective microstructure of the fiber core (Figure 1b), which will enhance the electrical transport rather than the thermal transport.32 As a result, the fiber cores could exhibit high µ/κl (µ is the carrier mobility and κl is the lattice thermal conductivity), which is proportional to the ZT value for high TE performance evaluation.33 2.3 Measured room-temperature thermoelectric properties 8 / 20 Figure 4. (a) SEM image of the BSTF bonded on the Si test chip. (b) FEM model for simulating the data-reduction process of the self-heating 3ω method with COMSOL software. (c) Frequency-dependent V3ω values measured from two fibers and (d) peak amplitude of the temperature oscillation calculated from COMSOL for a 4-µm-diameter fiber. (e) Temperature-dependent electrical resistance and (f) temperature-difference-dependent thermoelectric voltage. (g) Schematic of the fiber bending model. (h) Bending-radius-dependent relative electrical resistance; R and R0 are the resistances in the bending state and original flat state, respectively; Inset: an optical microscopic image of a bent BSTF. Figure 4. (a) SEM image of the BSTF bonded on the Si test chip. (b) FEM model for Figure 4. (a) SEM image of the BSTF bonded on the Si test chip. (b) FEM model for simulating the data-reduction process of the self-heating 3ω method with COMSOL software. (c) Temperature-dependent electrical resistance and (f) temperature-difference-dependent thermoelectric voltage. (g) Schematic of the fiber bending model. (h) Bending-radius-dependent relative electrical resistance; R and R0 are the resistances in the bending state and original flat state, respectively; Inset: an optical microscopic image of a bent BSTF. The electrical conductivity (σ), Seebeck coefficient (S), and thermal conductivity (κ) of the fiber were measured by using a pre-patterned silicon (Si) chip. Therefore the TE properties can be in-situ measured from one fiber. As shown in Figure 4a, a Bi2Te3 fiber is placed over four golden electrodes for the thermal conductivity and electrical conductivity measurements on the test chip, in which two outer ones (I+ and I-) and two inner ones (V+ and V-) are used for sourcing current and extracting voltage, respectively. One end of the fiber is located close to a golden heater which is used to create a temperature gradient for the S measurement. Once the heater was turned on, the two inner electrodes were also used as thermometers to probe the temperature 9 / 20 difference across the central portion of the fiber. The bonding between the fiber and the electrodes was created by depositing platinum (Pt) by applying FIB. The chip was subsequently annealed at 40°C for 10 minutes to cure the welding spot of Pt. The axial thermal conductivity κ\|\| of the suspended fiber was measured by the self-heating 3ω method.28 In this method, the fiber is periodically heated by an alternating current (AC) at an angular frequency ω. 2.3 Measured room-temperature thermoelectric properties The magnitude of the average temperature oscillation of the central portion of the fiber is proportional to a 3ω voltage: (1) 3 2 3 4 2 4 1 (2 ) rms s s rms s t I LR V A w b p k wg = + P (1) where and Irms are the root-mean-square amplitude of the 3ω voltage and the exciting current, respectively. L is the length of the central portion of the fiber. Rs, βs, and As are the resistance, temperature coefficient of the resistance (TCR), and the cross-sectional area, respectively. γt represents a characteristic thermal time constant associated with the heat transfer along the fiber axial direction. The measured data can be used to fit Equation (1) to extract the axial thermal conductivity κ\|\|. Notice that a 30 µm wide trench is created between the two inner electrodes (V+ and V-) on the silicon substrate to make the central portion of the fiber (sensing region) suspended. This will eliminate the heat flow between the sensing region and the substrate, which is necessary for properly implementing the self-heating 3ω method. Figure 4c shows the experimentally measured values of the two fibers as a function of exciting current frequency f=ω/2π. The obtained κ\|\| of 3.5-µm-diameter BTF and 4-µm-diameter BSTF are 0.69±0.06 and 0.57±0.05 W/m·K, respectively. 3 rms V w 3 rms V w 3 rms V w where and Irms are the root-mean-square amplitude of the 3ω voltage and the exciting current, respectively. L is the length of the central portion of the fiber. Rs, βs, and As are the resistance, temperature coefficient of the resistance (TCR), and the cross-sectional area, respectively. γt represents a characteristic thermal time constant associated with the heat transfer along the fiber axial direction. The measured data can be used to fit Equation (1) to extract the axial thermal conductivity κ\|\|. Notice that a 30 µm wide trench is created between the two inner electrodes (V+ and V-) on the silicon substrate to make the central portion of the fiber (sensing region) suspended. This will eliminate the heat flow between the sensing region and the substrate, which is necessary for properly implementing the self-heating 3ω method. Figure 4c shows the experimentally measured values of the two fibers as a function of exciting current frequency f=ω/2π. The obtained κ\|\| of 3.5-µm-diameter BTF and 4-µm-diameter BSTF are 0.69±0.06 and 0.57±0.05 W/m·K, respectively. 2.3 Measured room-temperature thermoelectric properties 3 rms V w 3 rms V w 3 rms V w where and Irms are the root-mean-square amplitude of the 3ω voltage and the exciting current, respectively. L is the length of the central portion of the fiber. Rs, βs, and As are the resistance, temperature coefficient of the resistance (TCR), and the cross-sectional area, respectively. γt represents a characteristic thermal time constant associated with the heat transfer along the fiber axial direction. The measured 3 rms V w 3 rms V w and Irms are the root-mean-square amplitude of the 3ω voltage and the 3 rms V w data can be used to fit Equation (1) to extract the axial thermal conductivity κ\|\|. Notice that a 30 µm wide trench is created between the two inner electrodes (V+ and V-) on the silicon substrate to make the central portion of the fiber (sensing region) suspended. This will eliminate the heat flow between the sensing region and the substrate, which is necessary for properly implementing the self-heating 3ω method. Figure 4c shows the experimentally measured values of the two fibers as a function of exciting current frequency f=ω/2π. The obtained κ\|\| of 3.5-µm-diameter BTF and 4-µm-diameter BSTF are 0.69±0.06 and 0.57±0.05 W/m·K, respectively. 3 rms V w The reliability of the current experimental configuration for measuring κ\|\| could be affected by the surface radiation of the fiber and the thermal contact of Pt bonding. To address these issues, we have applied the COMSOLTM software to perform a finite element modeling (FEM) calculation on a 4-µm-diameter fiber with a pre-set κ\|\| value of 0.5 W/m·K (Figure 4b). In the simulation, we assigned on purpose a unity emissivity to the surfaces of the fiber and Pt bonding. The thermophysical properties 10 / 20 and experimental dimension of the Pt bonding were also carefully considered. Figure 4d shows the theoretically predicted peak amplitude of temperature oscillation for the 4-µm-diameter fiber. By performing a similar nonlinear fitting on the data shown in Figure 4d (see Methods), the κ\|\| derived from the FEM simulation is found to be 0.455 W/m·K, which is about 9% smaller than its pre-set value. Thus, we believe that the 3ω measurements based on the current configuration are reliable. The temperature-dependent electrical resistances of two fibers were measured by the four-probe method, from which the TCRs in Equation (1) and the σ can be calculated (Figure 4e). 2.3 Measured room-temperature thermoelectric properties The S of the two fibers were measured by using a steady-state method.34 In this method, the heater was powered by a DC current source, which can create temperature gradients of different magnitude across the fiber, and the corresponding TE voltages (dV) were measured by a Keithley 2400 soucemeter (Figure 4f). The S can be calculated from the slope of the dV-dT linear fitting. Detailed descriptions of the experimental implementation and simulation of the self-heating 3ω method can be found in Methods. The obtained TE properties of the fiber samples are listed in Table 1 with measuring uncertainties (Supplementary Note 1) in comparison with those reported in the literature. Except for S, the values of σ, κ\|\|, and ZT of the prepared fibers are all improved with respect to those of bulk materials at different levels. Remarkably, the fiber core exhibits an ultralow κ and enhanced σ, which are mainly contributed to the interfacial defect structure and inner-oriented nanosheet crystals. The nanosheet thickness of ~20 nm is close to the wavelength of phonons and larger than that of majority carriers, so the nanosheet boundaries could selectively scatter phonons and make carriers pass through. 35 The phonon modulation is then investigated by Vienna Ab initio Simulation Package (VASP) software based on the Density Functional Theory (DFT). The calculation details can be found in Supplementary Note 2. The study confirms the effect of different mean-free-path phonons on the lattice thermal conductivity of Bi2Te3 at 300 K (Supplementary Figure S5). What’s more, the nanostructure of Bi2Te3 under 20 nm could efficiently scatter phonons and depress the lattice thermal conductivity at 300 K. 11 / 20 The fiber core without cladding can keep 90% of its original electrical conductivity value after being preserved in the atmosphere for more than one year. (Supplementary Figure S6). And these fibers do not require using ion milling/etching as reported for nanowires to cleave surface oxides and measure TE properties.29,36 Beyond these, we carry out a bending procedure and consecutively measure I-V characteristics of the representative BTF and BSTF (fiber diameter D~4 µm) by applying two gold-plated tungsten probes (Figure 4g and Supplementary Figure S7). Using the original resistance R0 as a reference, the bending-state resistance R is normalized to exhibit a change of the electrical resistance as a function of 1–5 mm bending radius r (Figure 4h). 2.3 Measured room-temperature thermoelectric properties After exerting 100-time bending cycles at r~5 mm, the resistance increases less than 5%, indicating its tremendous mechanical flexibility (Supplementary Figure S7 and Supplementary Movie 1). Meanwhile, the bending radius minima (rmin) of all referred samples are also listed in Table 1. Their flexibility can also be estimated by the maximum bending strain (ε=D/2r).37 It exhibits that these microfibers possess rmin~50 µm an excellent ε of 4%, approaching that of the reported nanowires and the ~5.5% theoretical limit of Bi2Te3 elastic strain.9,38 These results demonstrate that the thermally drawn micro-nano Bi2Te3 fibers with interfacial engineering effects possess both high TE performance and great bending stability. Table1. Thermoelectric and bending properties of Bi2Te3 bulk and fiber samples Table1. Thermoelectric and bending properties of Bi2Te3 bulk and fiber samples Table1. Thermoelectric and bending properties of Bi2Te3 bulk and fiber samples Samples at 300 K S (µV/K) σ (104 S/m) κ\|\| (W/m·K) ZT rmin Bulk Bi2Te3 114 5.0 1.50 0.21 / 3.5 µm BTF 107±12 5.9±0.3 0.69±0.06 0.3±0.07 45±5 µm Bulk BST 212 10.5 1.4 1.01 / 4 µm BSTF 145±16 12.5±0.7 0.57±0.05 1.38±0.34 50±5 µm 30 µm BST 22 150 15.6 0.84 1.25 1 cm 50 µm Bi2Te3 23 131 7.4 0.52 0.73 5 cm 65 nm BST 29 125 4.5 0.3 0.7 500 nm 12 / 20 3. Conclusions In conclusion, we have demonstrated the high-performance micro-nano polycrystalline Bi2Te3 TE fibers with great flexibility and stability through thermal drawing Bi2Te3-based bulks in a glass-fiber template. A special interface between the glass cladding and the Bi2Te3 core was observed to induce oriented nanosheet crystals in the fiber cores, enabling the fibers to display ultra-low thermal conductivity and high electrical conductivity. Furthermore, this interfacial engineering efficiently regulates the electrical and thermal transport, while keeping the electrical properties stably after 100-time 5-mm-diameter bending and long-time preservation. This strategy could be universal to other inorganic TE materials, particularly those made from nontoxic and abundant elements. The high TE performance and stability in the micro-nano TE fibers will stimulate further research on the TE fiber materials (e.g. fibrous crystal anisotropy and quantum size effect) and the development of integrated TE fiber devices (e.g. self-powered wearable electronics and micro-area temperature management). 4. Methods 4.1 Fabrication of micro-nano TE fibers. The glass-fiber-template method and two-step thermal drawing were applied for micro-nano Bi2Te3 fibers fabrication. First, a 10-mm-diameter Bi2Te3 or Bi0.5Sb1.5Te3 rod (99.99% purity, Santech Materials, Changsha) was sealed in a Pyrex3.3-borosilicate-glass cladding to assemble a 30-mm-diameter macroscopic cylindrical preform. The preforms were preheated at 700 ℃ for two-hour before drawing to form a wetting interface of the core and the cladding by elemental diffusion. The TE fibers were then drawn at ~950 °C using an optical fiber drawing tower. The fibers endure flash cooling with a rate of ~150 ℃/s right after thermal drawing, but all fibers maintain a complete core-cladding structure and the fiber cores show a relative density of >99.2%. The 400-µm-diameter Bi2Te3-based fiber prepared by the first thermal drawing was then inserted into a glass tube of Pyrex3.3 borosilicate to form a new fiber preform. Cascading two thermal 13 / 20 drawing steps, continuous meters of micro-nano Bi2Te3-based TE fibers were produced. 4.2 Composition and microstructure characterizations. Utilizing an argon ion beam cross-section polisher (CP, JEOL IB-09020CP, Japan), the as-drawn fibers were polished to obtain a cross-sectional sample and a longitudinal section sample. And the two fiber samples were observed under a field electron-scanning electron microscope (ZEISS Merlin, FE-SEM, Germany) operated at 20 kV with a working distance of 8.4 mm. The Bi2Te3 core was obtained by etching the fibers in HF acid solution to strip the glass cladding and then identified by an X-ray diffractometer (XRD, X’Pert PROX, Cu Kα, Netherland). Using an electron probe micro-analyzer (Shimadzu EPMA-1600, Japan) equipped with a wavelength-dispersive spectrometer, elemental and microstructural analyses were performed on the fibers at several locations traversing the core in approximately 1 µm increments to examine the distribution of elements. A chip sample, containing the fiber core/cladding interface region of the fiber, was cut out from the fiber by a focused gallium ion beam instrument (FIB, FEI Helios 450S, America). High-resolution transmission electron microscopy (HR-TEM) observations on this chip sample were carried out at 200 kV using a transmission electron microscope (FEI Titan Themis 200, X-FEG TEM). 14 / 20 4.3 Self-heating 3ω measurement. The thermal conductivity κ\|\| and volumetric heat capacity CV of the suspended Bi2Te3 fiber were measured by a self-heating 3ω method proposed originally by Lv and co-workers. 4. Methods In this method, the electrically conductive sample acts as not only a heater but also a resistive temperature sensor, since it owns a substantially large temperature coefficient of resistance (TCR). Figure S4 shows a schematic of the electrical circuit for carrying out the 3ω measurement on the fiber integrated test chip using a four-probe configuration.39,40 (To reduce the convective heat loss, the test chip was installed inside a vacuum chamber evacuated to below 5×10-4 Pa during the measurement). A sinusoidal current at frequency ω supplied by a Keithley 6221 current source was injected into the fiber from two outer contacts (I+ and I-) on the test chip. The periodic heating induced a resistance fluctuation of the fiber at frequency 2ω. Multiplication of the 2ω resistance variation by the 1ω driving 14 / 20 current gives rise to a small 3ω voltage which can be used to extract the thermo-physical properties of the sample by using Equation (1). The voltage extracted from two inner contacts (V+ and V-) contain both 1ω component (V1ω) and 3ω component (V3ω). Since the amplitude of V1ω is usually 3–4 orders of magnitude larger than that of V3ω, V1ω should be filtered out to achieve faithfully recording of the 3ω voltage signal. Hence the fiber was electrically connected with a potentiometer in series. The resistance of the potentiometer was adjusted to be slightly larger than that of the fiber. After passing through a differential amplifier, the voltage signal extracted from the fiber was fed into the input A of a Stanford SR850 lock-in amplifier, while the 1ω voltage signal V1ω,P from the potentiometer was passed to a 12-bit DAC converter in which the amplitude of V1ω,P was adjusted to be equal to that of V1ω by controlling the gain of the DAC with a computer. The attenuated V1ω,P signal was then sent to the input B of the lock-in amplifier to deduct out V1ω by using a differential input mode. After the signal subtraction step, V3ω can be measured by the same lock-in amplifier. 4.4 Finite element simulation of 3ω measurement. A 3D FEM simulation has been carried out to evaluate the reliability of measuring the thermal conductivity of the fiber on the test chip with the geometrical configuration shown in Figure 4a. 4. Methods A simulation model was constructed to imitate the chip setup, in which the diameter and length of the suspended portion of the fiber were 4 µm and 30 µm, respectively. (Figure 4b) The fiber was assigned a thermal conductivity of 0.5 W/m·K and a volumetric heat capacity of 1.22×106 J/(m3K).41 Lee and co-workers found that the platinum film created by the FIB deposition mainly consists of carbon (45%–55%) and platinum (40%–50%).42 The thermal conductivity of the Pt/C film was measured to be 40.97 W/m·K. In the simulation, the thermal conductivity of the Pt bonding was set to be 40 W/m·K. We further assumed that the Pt/C bonding is composed of 50% amorphous carbon and 50% platinum in volume,42 therefore the volumetric heat capacity of the Pt/C bonding was approximated to be 2.98×106 J/(m3K). From Lee’s work,43 we also estimated that the electrical conductivity of the Pt/C material is about 72.04 S/m. To initiate the simulation, an AC electrical current with a peak amplitude 15 / 20 I0 of 0.4 mA and an angular frequency of ω was injected into the fiber from two outer electrodes to create periodical heating at frequency 2ω in vacuum at room temperature. The temperature distribution of the test chip was calculated by coupling the heat transfer module with the AC/DC module implemented in the COMSOLTM software. The surface heat radiation loss was modeled by employing the Stefan-Boltzmann law. The material properties of the fiber, Pt/C bonding, gold strips,44,45 and silicon substrate used in the FEM modeling are summarized in Table 2, which are all assumed to be temperature independent. Table 2. Materials properties used in the FEM simulation Bi2Te3 41 Pt/C 43 Au 44 Si 45 （W m-1 K-1） CV （J m-3 K-1）（S m-1） 0.5 1.22×106 1 1.23×105 40 2.98×106 1 72.04 240.00 1.60×106 0.03 5.23×106 142.2 1.63×106 0.5 / k g s Table 2. Materials properties used in the FEM simulation Supplementary Figure S7 illustrates a typical time-dependent spatially averaged temperature oscillation curve extracted from the central portion of the fiber heated by the electrical current at f=15 Hz (f=ω/2π). The peak amplitude of the temperature oscillation can be displayed as a function of exciting current frequency, as shown in Figure 4d. Acknowledgments The authors acknowledge funding from the Natural Science Foundation of China (52002131, U1830203, 62005080, 52172249, 51888103, 91833303), 2021 Talent Revitalization Plan Project for New High Performance Material Industry in Qingyuan City (2021YFJH02001), Local Innovative and Research Teams Project of Guangdong Pearl River Talents Program (2017BT01X137), Key R&D Program of Guangzhou (202007020003), and Innovation Academy for Light-duty Gas Turbine, Chinese Academy of Sciences (CXYJJ21-ZD-02). The authors acknowledge Dr. Ben Huang and Prof. Pengcheng Zhai at Wuhan University of Technology for providing mechanical discussions of thermoelectric fibers. 4. Methods Notice that in the real experiment can be derived from 2 p T w D 2 p T w D the root-mean-square amplitude of the 3ω voltage by: (2) 3 rms V w 2 3 = w b D rms s s p I R T V the root-mean-square amplitude of the 3ω voltage by: 3 rms V w (2) 2 3 = 2 w w b D rms s s p rms I R T V (2) where Irms is the root-mean-square amplitude of the exciting current. Rs is the resistance of the suspended portion of the fiber. βs is the TCR of the fiber. Substituting Equation (2) into Equation (1) yields temperature oscillation: 16 / 20 (3) ( ) 2 2 2 4 t 8 1 2 w p k wg D = + rms s p s I R L T A P (3) where As is the cross-sectional area of the fiber. By fitting the vs. f curve in Figure 4d to Equation (3), the thermal conductivity was found to be 0.455 W/(mK). 2 p T w D Author contributions M.S., G.W.T., P.Y.Z., and B.H. prepared the micro-nano TE fibers and conducted EPMA, AFM, and TEM characterization. M.S., G.W.T., and H.F.W. performed TE properties measurements and wrote the original manuscript. M.Y., H.Z., and T.Z. performed theoretical simulation. M.S., Y.C.C., and J.C. conducted stability tests of electrical performance. M.S., G.W.T., H.F.W., D.D.C., and J.L.G. analyzed results. Q.Q. and Z.M.Y. directed this project. 5 Yao, Z. et al. Solute manipulation enabled band and defect engineering for thermoelectri References 1 Snyder, G. J. & Toberer, E. S. Complex thermoelectric materials. Nature Materials 7, 105-114, doi:10.1038/nmat2090 (2008). 1 Snyder, G. J. & Toberer, E. S. Complex thermoelectric materials. Nature Materials 7, 105-114, doi:10.1038/nmat2090 (2008). 2 Tarancon, A. Powering the IoT revolution with heat. Nat Electron 2, 270-271, doi:10.1038/s41928-019-0276-4 (2019). 2 Tarancon, A. Powering the IoT revolution with heat. Nat Electron 2, 270-271, doi:10.1038/s41928-019-0276-4 (2019). 3 Geffroy, C., Lilley, D., Parez, P. S. & Prasher, R. Techno-economic analysis of waste-heat conversion. Joule 5, 3080-3096, doi:10.1016/j.joule.2021.10.014 (2021). 3 Geffroy, C., Lilley, D., Parez, P. S. & Prasher, R. Techno-economic analysis of waste-heat conversion. Joule 5, 3080-3096, doi:10.1016/j.joule.2021.10.014 (2021). 4 Pei, J., Cai, B. W., Zhuang, H. L. & Li, J. F. Bi2Te3-based applied thermoelectric materials: research advances and new challenges. National Science Review 7 (2020). 4 Pei, J., Cai, B. W., Zhuang, H. L. & Li, J. F. Bi2Te3-based applied thermoelectric materials: research advances and new challenges. National Science Review 7 (2020). Yao, Z. et al. Solute manipulation enabled band and defect engineering for thermoelectric 17 / 20 enhancements of SnTe. InfoMat 1, 571-581, doi:10.1002/inf2.12044 (2019). enhancements of SnTe. InfoMat 1, 571-581, doi:10.1002/inf2.12044 (2019). 6 Yang, G. S. et al. Ultra-High Thermoelectric Performance in Bulk BiSbTe/Amorphous Boron Composites with Nano-Defect Architectures. Advanced Energy Materials 10, doi: 10.1002/aenm.202000757 (2020). 7 Zhu, B. et al. Realizing record high performance in n-type Bi2Te3-based thermoelectric materials. Energy & Environmental Science 13, 2106-2114, doi:10.1039/d0ee01349h (2020). 8 Wu, Z. H., Zhang, S., Liu, Z. K., Mu, E. R. & Hu, Z. Y. Thermoelectric converter: Strategies from materials to device application. Nano Energy 91, doi: 10.1016/j.nanoen.2021.106692 (2022). 9 Huang, B., Li, G., Yang, X. & Zhai, P. Capturing anharmonic and anisotropic natures in the thermotics and mechanics of Bi2Te3 thermoelectric material through an accurate and efficient potential. Journal of Physics D: Applied Physics 52, 425303, doi:10.1088/1361-6463/ab2f3a (2019). 10 Zhao, W. et al. Multi-localization transport behaviour in bulk thermoelectric materials. Nature Communications 6, 6197, doi:10.1038/ncomms7197 (2015). 11 Sarma, D. D. Essential Considerations for Reporting Thermoelectric Properties. ACS Energy Lett 6, 3715-3718, doi:10.1021/acsenergylett.1c02043 (2021). 12 Yang, C. et al. Transparent flexible thermoelectric material based on non-toxic earth-abundant p-type copper iodide thin film. Nature Communications 8, 16076, doi:10.1038/ncomms16076 (2017). 13 Jin, Q. et al. Flexible layer-structured Bi2Te3 thermoelectric on a carbon nanotube scaffold. Nature Materials 18, 62, doi:10.1038/s41563-018-0217-z (2019). 14 Zhang, T. et al. References Mechanically Durable and Flexible Thermoelectric Films from PEDOT:PSS/PVA/Bi0.5Sb1.5Te3 Nanocomposites. Advanced Electronic Materials 3, doi: 10.1002/aelm.201600554 (2017). 15 Rojas, J. P. et al. Paper-based origami flexible and foldable thermoelectric nanogenerator. Nano Energy 31, 296-301, doi:10.1016/j.nanoen.2016.11.012 (2017). 16 Park, S. H. et al. High-performance shape-engineerable thermoelectric painting. Nature Communications 7, doi:10.1038/ncomms13403 (2016). 17 Lee, J. A. et al. Woven-Yarn Thermoelectric Textiles. Advanced Materials 28, 5038-5044, doi:10.1002/adma.201600709 (2016). 18 Yan, W. et al. Thermally drawn advanced functional fibers: New frontier of flexible electronics. Materials Today, doi:10.1016/j.mattod.2019.11.006 (2020). 19 Loke, G., Yan, W., Khudiyev, T., Noel, G. & Fink, Y. Recent Progress and Perspectives of Thermally Drawn Multimaterial Fiber Electronics. Advanced Materials 32, doi:10.1002/adma.201904911 (2020). 20 Tao, G., Stolyarov, A. M. & Abouraddy, A. F. Multimaterial Fibers. International Journal of Applied Glass Science 3, 349-368, doi:10.1111/ijag.12007 (2012). 21 Sun, M. et al. Sn-Se alloy core fibers. Journal of Alloys and Compounds 725, 242-247, doi:10.1016/j.jallcom.2017.07.150 (2017). 22 Zhang, T. et al. High-performance, flexible, and ultralong crystalline thermoelectric fibers. Nano Energy 41, 35-42, doi:10.1016/j.nanoen.2017.09.019 (2017). 23 Sun, M. et al. Enhanced thermoelectric properties of polycrystalline Bi2Te3 core fibers with preferentially oriented nanosheets. APL Materials 6, doi:10.1063/1.5018621 (2018). 24 Zhang, J. et al. Single-Crystal SnSe Thermoelectric Fibers via Laser-Induced Directional 18 / 20 Crystallization: From 1D Fibers to Multidimensional Fabrics. Advanced Materials 32, doi: 10.1002/adma.202002702 (2020). Crystallization: From 1D Fibers to Multidimensional Fabrics. Advanced Materials 32, doi: 10.1002/adma.202002702 (2020). 25 Yan, W. et al. Advanced Multimaterial Electronic and Optoelectronic Fibers and Textiles. Advanced Materials 31, doi: 10.1002/adma.201802348 (2019). 26 Shi, X.-L., Chen, W.-Y., Zhang, T., Zou, J. & Chen, Z.-G. Fiber-based thermoelectrics for solid, portable, and wearable electronics. Energy & Environmental Science 14, 729-764, doi:10.1039/d0ee03520c (2021). 27 Sun, M. et al. Tailoring microstructure and electrical transportation through tensile stress in Bi2Te3 thermoelectric fibers. Journal of Materiomics 6, 467-475, doi:10.1016/j.jmat.2020.02.004 (2020). 28 Lu, L., Yi, W. & Zhang, D. L. 3ω method for specific heat and thermal conductivity measurements. Review of Scientific Instruments 72, 2996-3003, doi:10.1063/1.1378340 (2001). 29 Li, L., Xu, S. & Li, G. Enhancement of Thermoelectric Properties in Bi-Sb-Te Alloy Nanowires by Pulsed Electrodeposition. Energy Technology 3, 825-829, doi:10.1002/ente.201500071 (2015). 30 Chen, R., Lee, J., Lee, W. & Li, D. Thermoelectrics of Nanowires. Chemical Reviews 119, 9260-9302, doi:10.1021/acs.chemrev.8b00627 (2019). 31 He, Y. et al. Ultrahigh Thermoelectric Performance in Mosaic Crystals. Advanced Materials 27, 3639-3644, doi:10.1002/adma.201501030 (2015). 32 Wu, Y. et al. Lattice Strain Advances Thermoelectrics. 45 Bodenschatz, N., Liemert, A., Schnurr, S., Wiedwald, U. & Ziemann, P. Extending the 3 omega method: Thermal conductivity characterization of thin films. Review of Scientific Instruments 84, doi: 10.1063/1.4817582 (2013). cross-plane thermal conductivity of thin films. Phys Status Solidi A 210, 106-118, doi:10.1002/pssa.201228478 (2013). References Joule 3, 1276-1288, doi:10.1016/j.joule.2019.02.008 (2019). 33 Snyder, G. J. et al. Weighted Mobility. Advanced Materials 32, doi: 10.1002/adma.202001537 (2020). 34 Wang, H. et al. Determination of the thermopower of microscale samples with an AC method. Measurement 131, 204-210, doi:10.1016/j.measurement.2018.08.027 (2019). 35 Xie, W., Tang, X., Yan, Y., Zhang, Q. & Tritt, T. M. Unique nanostructures and enhanced thermoelectric performance of melt-spun BiSbTe alloys. Applied Physics Letters 94, doi:10.1063/1.3097026 (2009). 36 Yang, L. et al. High thermoelectric figure of merit of porous Si nanowires from 300 to 700K. Nature Communications 12, doi: 10.1038/s41467-021-24208-3 (2021). 37 Peng, J. & Snyder, G. J. A figure of merit for flexibility. Science 366, 690-+, doi:10.1126/science.aaz5704 (2019). 38 Li, X. et al. In situ physical examination of Bi2S3 nanowires with a microscope. Journal of Alloys and Compounds 798, 628-634, doi:10.1016/j.jallcom.2019.05.319 (2019). 39 Huxtable, S. T. Heat transport in superlattices and nanowire arrays. Ph.D thesis, University of California, Berkeley (2002). 40 Cahill, D. G. Thermal conductivity measurement from 30 to 750 K: The 3 omega method (vol 61, pg 802, 1990). Review of Scientific Instruments 73, 3701-3701, doi:10.1063/1.1505652 (2002). 41 Lee, H. F. A modified 3ω method for thermal conductivity measurement of one-dimensional nanostructures. Master thesis, The Pennsylvania State University (2009). 42 Madelung, O. R., U.; Schulz, M. Non-tetrahedrally Bonded Elements and Binary Compounds. Vol. Subvolume 41C (1998). 43 Morath, C. J. M., H. J.; Cuomo, J. J.; Pappas, D. L.; Grill, A.; Patel, V. V.; Doyle, J. P.; Saenger, K. L. Picosecond optical studies of amorphous diamond and diamondlike carbon:Thermal conductivity and longitudinal sound velocity. Journal of Applied Physics, 2636 (1994). 44 Volklein, F., Reith, H. & Meier, A. Measuring methods for the investigation of in-plane and 19 / 20 cross-plane thermal conductivity of thin films. Phys Status Solidi A 210, 106-118, doi:10.1002/pssa.201228478 (2013). 45 Bodenschatz, N., Liemert, A., Schnurr, S., Wiedwald, U. & Ziemann, P. Extending the 3 omega method: Thermal conductivity characterization of thin films. Review of Scientific Instruments 84, doi: 10.1063/1.4817582 (2013). cross-plane thermal conductivity of thin films. Phys Status Solidi A 210, 106-118, doi:10.1002/pssa.201228478 (2013). 45 Bodenschatz, N., Liemert, A., Schnurr, S., Wiedwald, U. & Ziemann, P. Extending the 3 omega method: Thermal conductivity characterization of thin films. Review of Scientific Instruments 84, doi: 10.1063/1.4817582 (2013). 45 Bodenschatz, N., Liemert, A., Schnurr, S., Wiedwald, U. & Ziemann, P. Extending the 3 omega method: Thermal conductivity characterization of thin films. Review of Scientific Instruments 84, doi: 10.1063/1.4817582 (2013). 45 Bodenschatz, N., Liemert, A., Schnurr, S., Wiedwald, U. & Ziemann, P. Extending the 3 omega method: Thermal conductivity characterization of thin films. Review of Scientific Instruments 84, doi: 10.1063/1.4817582 (2013). 20 / 20 Supplementary Files This is a list of supplementary ¦les associated with this preprint. Click to download. SupplementaryInfomation12.docx SupplementaryMovie1.mov
https://openalex.org/W3015106251	https://www.mdpi.com/1996-1073/13/7/1654/pdf?version=1585832581	English	null	Production of Itaconic Acid from Cellulose Pulp: Feedstock Feasibility and Process Strategies for an Efficient Microbial Performance	Energies	2,020	cc-by	8,264	Received: 27 February 2020; Accepted: 31 March 2020; Published: 2 April 2020 Received: 27 February 2020; Accepted: 31 March 2020; Published: 2 April 2020 Abstract: This study assessed the feasibility of using bleached cellulose pulp from Eucalyptus wood as a feedstock for the production of itaconic acid by fermentation. Additionally, diﬀerent process strategies were tested with the aim of selecting suitable conditions for an eﬃcient production of itaconic acid by the fungus Aspergillus terreus. The feasibility of using cellulose pulp was demonstrated through assays that revealed the preference of the strain in using glucose as carbon source instead of xylose, mannose, sucrose or glycerol. Additionally, the cellulose pulp was easily digested by enzymes without requiring a previous step of pretreatment, producing a glucose-rich hydrolysate with a very low level of inhibitor compounds, suitable for use as a fermentation medium. Fermentation assays revealed that the technique used for sterilization of the hydrolysate (membrane ﬁltration or autoclaving) had an important eﬀect in its composition, especially on the nitrogen content, consequently aﬀecting the fermentation performance. The carbon-to-nitrogen ratio (C:N ratio), initial glucose concentration and oxygen availability, were also important variables aﬀecting the performance of the strain to produce itaconic acid from cellulose pulp hydrolysate. By selecting appropriate process conditions (sterilization by membrane ﬁltration, medium supplementation with 3 g/L (NH4)2SO4, 60 g/L of initial glucose concentration, and oxygen availability of 7.33 (volume of air/volume of medium)), the production of itaconic acid was maximized resulting in a yield of 0.62 g/g glucose consumed, and productivity of 0.52 g/L·h. Keywords: lignocellulosic biomass; cellulose pulp; hydrolysis; oxygen availability; C:N ratio; fermentation; bioreﬁnery; itaconic acid; Aspergillus terreus Keywords: lignocellulosic biomass; cellulose pulp; hydrolysis; oxygen availability; C:N ratio; fermentation; bioreﬁnery; itaconic acid; Aspergillus terreus energies energies Energies 2020, 13, 1654; doi:10.3390/en13071654 www.mdpi.com/journal/energies Abraham A. J. Kerssemakers , Pablo Doménech , Marco Cassano , Celina K. Yamakawa Giuliano Dragone and Solange I. Mussatto * Novo Nordisk Foundation Center for Biosustainability, Technical University of Denmark, Kemitorvet, Building 220, 2800 Kongens Lyngby, Denmark; aajker@biosustain.dtu.dk (A.A.J.K.); pablo.domenech@ciemat.es (P.D.); marcocassano08@gmail.com (M.C.); celinayamakawa@biosustain.dtu.dk (C.K.Y.); giudra@biosustain.dtu.dk (G.D.) * Correspondence: smussatto@biosustain.dtu.dk or solangemussatto@hotmail.com; Tel.: +45-93-511-891 Novo Nordisk Foundation Center for Biosustainability, Technical University of Denmark, Kemitorvet, Building 220, 2800 Kongens Lyngby, Denmark; aajker@biosustain.dtu.dk (A.A.J.K.); pablo.domenech@ciemat.es (P.D.); marcocassano08@gmail.com (M.C.); celinayamakawa@biosustain.dtu.dk (C.K.Y.); giudra@biosustain.dtu.dk (G.D.) * Correspondence: smussatto@biosustain.dtu.dk or solangemussatto@hotmail.com; Tel.: +45-93-511-891 Article Production of Itaconic Acid from Cellulose Pulp: Feedstock Feasibility and Process Strategies for an Eﬃcient Microbial Performance Abraham A. J. Kerssemakers , Pablo Doménech , Marco Cassano , Celina K. Yamakawa Giuliano Dragone and Solange I. Mussatto * 1. Introduction The development of new process technologies using lignocellulosic feedstock as a carbon source for the production of fuels and chemicals is, currently, one of the main drivers of society to move towards a more sustainable future [1]. Second-generation biofuel plants are already a reality at a commercial scale and, to become truly sustainable and circular, industry is also increasingly viewing the production of chemicals from renewable resources as an attractive area for investment. Biofuels and biochemicals can be produced in single product processes; however, their production in an integrated bioreﬁnery is seen as a more eﬃcient and interesting approach to solve economic challenges related to biomass conversion processes since, currently, the cost of single biobased production processes, in many cases, still exceeds the cost of petrochemical production [2]. One of the main reasons for these high costs is the recalcitrant nature of biomass that, therefore requires a two-step Energies 2020, 13, 1654; doi:10.3390/en13071654 www.mdpi.com/journal/energies Energies 2020, 13, 1654 2 of 12 processing to obtain sugars for fermentation as follows: a pretreatment step to fractionate the material and solubilize especially hemicellulose sugars, and a subsequent hydrolysis step to recover glucose from cellulose. Pretreatment, in particular, is an energy intensive step and signiﬁcantly contributes to the ﬁnal cost of the process [3]. Itaconic acid stands out as one of the most relevant among the variety of chemicals that can be produced from lignocellulosic biomass, since it is a platform chemical with extensive applications in diﬀerent ﬁelds. Some of the main interests around itaconic acid arise from its potential to substitute petrochemically produced acrylic acid. However, it can also be used to produce biodegradable polymers, paints, varnishes, and diﬀerent organic compounds. Moreover, itaconic acid and its derivatives support the synthesis of a wide range of innovative polymers through crosslinking, with applications in special hydrogels for water decontamination, drug delivery, nanohydrogels for food applications, coatings, and elastomers [4]. Currently, itaconic acid is produced industrially from aerobic fungal fermentation using pure glucose as a carbon source, which is not the cheapest or the most sustainable substrate option. Moreover, although the production is done by fermentation, at present, the cost to produce itaconic acid is high and has been a bottleneck preventing its application in diﬀerent sectors [5]. 1. Introduction With a market in expansion due to the increased number of potential applications (its market was worth USD 126.4 million in 2014 and with an expected growth rate of 60% it is predicted to reach around USD 204.6 million by 2023 [6]), the establishment of a more sustainable and cost-competitive process for the production of itaconic acid from renewable feedstock has been strongly encouraged. The present study aims to contribute with new knowledge to advance this area by using industrially produced bleached cellulose pulp as a feedstock for itaconic acid production. Bleached cellulose pulp, which is the material used for paper manufacturing, is one of the most abundant raw materials worldwide. With a huge volume of production, and a weak demand from the paper industry in the last years, the stocks of bleached cellulose pulp have been extremely high and are posing a major problem for the entire pulp market, according to industry experts [7]. To overcome this problem, diﬀerent alternative uses for the pulp have been explored with the aim of promoting innovation and new business opportunities [8], including the production of biofuels, nanocellulose, and biocomposites. Recently, attempts have been done to produce itaconic acid from diﬀerent types of biomass including beech wood [9], corn stover [10], wheat chaﬀ[11], rice husks [12], and corn cobs [13]. A comprehensive examination of the itaconic acid production from these diﬀerent feedstocks clearly demonstrates an important impact of biomass pretreatment steps, presence of inhibitor compounds, and fermentation conditions on itaconic acid yield and on the feasibility of the process in general. This paper is the ﬁrst study on the use of bleached cellulose pulp for the production of itaconic acid. In this study, the composition of cellulose pulp, as well as its degradation by enzymes and fermentability were some of the points explored to evaluate its feasibility for application in the production of itaconic acid. Then, eﬀorts were done to select process conditions able to result in an improved bioconversion eﬃciency. Sterilization of the cellulose pulp hydrolysate through diﬀerent techniques, medium composition in terms of carbon-to-nitrogen ratio (C:N ratio) and initial glucose concentration, and oxygen availability were evaluated and discussed in detail. At the end, the process conditions that maximize the production of itaconic acid were selected and the results were compared with literature data from other feedstocks to conclude on the potential of this new bioprocess. 2.1. Microorganism and Inoculum Preparation The ﬁlamentous fungus Aspergillus terreus NRRL 1960 was used in the experiments. The strain was obtained from the ARS Culture Collection (Peoria, IL, USA) and preserved in the form of spores in 20% (v/v) glycerol stock solution at −80 ◦C. Energies 2020, 13, 1654 3 of 12 For inoculum preparation, the stock culture was activated on 2.4% (w/v) potato dextrose broth (PDB) medium at 35 ◦C for 3 days, and subsequently on 3.9% (w/v) potato dextrose agar (PDA) plates at 35 ◦C for 7 days. Then, spores were collected from plates by using a sterilized solution of 4% (w/v) Tween 80. The spore suspension was diluted with sterile MilliQ water in order to obtain a concentration of 106 spores/mL at the beginning of the fermentation. 2.3. Hydrolysate Sterilization Three diﬀerent methods (membrane ﬁltration, autoclave at 112 ◦C for 15 min, and autoclave at 121 ◦C for 20 min) were tested for sterilization of the hydrolysate prior its use as fermentation medium. The autoclave assays were carried out in an autoclave MultiControl 2 (CertoClav, Austria); while for the membrane method, Nalgene RapidFlow™PES-membrane ﬁlters with a pore size of 0.2 µm (Thermo Fisher Scientiﬁc, USA) were used. 2.2. Cellulose Pulp Characterization and Hydrolysis Bleached cellulose pulp from Suzano S/A (Brazil) was used as raw material for the production of itaconic acid. The cellulosic material, which was produced from Eucalyptus wood and had a moisture content of approximately 5% (w/w), was ground to particle size ≤1.0 mm by means of a mill Polymix PX-MFC 90D (Kinematica, Switzerland) and its composition was determined by following standard methods [14–16]. Enzymatic hydrolysis of the cellulose pulp was carried out using the enzyme concentrate Cellic® CTec2, kindly supplied by Novozymes (Bagsværd, Denmark). The cellulase activity of the concentrate, which was measured according to standard protocol [17] and expressed in ﬁlter paper units (FPU), was 217.5 FPU/mL. One unit of FPU was deﬁned as the amount of enzyme required to liberate 1 µmol of glucose from Whatman no.1 ﬁlter paper per minute at 50 ◦C. For the experiments, an enzyme load of 10 FPU/g cellulose was added to 0.1 M sodium citrate buﬀer (pH 4.7), and then mixed with the cellulose pulp in a concentration of 12% (w/v). The reactions were carried out in 2-L Duran laboratory bottles with vertical baﬄes containing 0.6 L of working volume. The bottles were accommodated horizontally in a Bottle/Tube Roller system (Thermo Scientiﬁc, USA) placed inside an incubator, and kept at 50 ◦C and 20 rpm for 96 h. Afterwards, the hydrolysate was separated by centrifugation (10,000 rpm, 5 ◦C, 20 min). 2.6. Fermentation Parameters To evaluate the performance of itaconic acid production by A. terreus the following fermentation parameters were considered: 1. Itaconic acid yield per sugar consumed as YP/S (g/g) = −∆P ∆S; 2. Biomass yield per sugar consumed as YX/S (g/g) = −∆X ∆S ; 3. Itaconic acid productivity as QP (g/L·h) = ∆P ∆t . where P is the concentration of itaconic acid (g/L), S is the concentration of sugar (g/L), and t is the fermentation time (h). where P is the concentration of itaconic acid (g/L), S is the concentration of sugar (g/L), and t is the fermentation time (h). where P is the concentration of itaconic acid (g/L), S is the concentration of sugar (g/L), and t is the fermentation time (h). 2.5. Analytical Methods and Statistical Analysis Nitrogen content in the hydrolysates was determined by using an elemental analyzer Vario MACRO cube (Elementar Analysensysteme GmbH, Germany), following the Dumas method. Cell mass concentration during the fermentations was estimated by dry weight measurement. The fermentation broth was centrifuged at 10,000 rpm for 10 min and the biomass pellet was rinsed two times with deionized water and dried at 60 ◦C for 48 h. The supernatant of centrifuged samples was used for pH measurement and determination of sugars, itaconic acid, and potential by-products. The concentrations of glucose, cellobiose, xylose, sucrose, mannose, glycerol, organic acids (itaconic, acetic and formic), 5-hydroxymethylfurfural (5-HMF) and furfural were determined by high-performance liquid chromatography (HPLC) using a Dionex Ultimate 3000 HPLC equipment (Thermo Scientiﬁc, USA) coupled with a Biorad Aminex® HPX-87H column (300 × 7.8 mm). For analysis, the column was maintained at 65 ◦C and a 5 mM H2SO4 solution was used as mobile phase at a ﬂow rate of 0.5 mL/min. Sugars, glycerol, and organic acids were detected using a Shodex RI-101 refractive index detector, whereas 5-HMF and furfural were detected using an ultraviolet detector at 254 nm. Statistical analysis including graphs and quantitative information such as mean and standard deviation was performed using the software OriginPro 9.1.0 (OriginLab Corporation, USA). 2.6. Fermentation Parameters 2.4. Fermentation Media and Conditions Initially, diﬀerent synthetic media were tested for the production of itaconic acid by A. terreus, which contained only one type of carbon source (glucose, xylose, sucrose, mannose, or glycerol) at a concentration of 50 g/L. Later, the cellulose pulp hydrolysate was used as fermentation medium, which contained around 53 g/L of glucose as carbon source. For all the experiments, the initial pH of the media was adjusted to 3.0. All the fermentation media, synthetic and hydrolysate, were supplemented with the following nutrients (in g/L): KH2PO4 (0.2), (NH4)2SO4 (3.0), MgSO4·7H2O (3.0), CaCl2·1H2O (0.2), ZnSO4·7H2O (0.15), FeSO4·7H2O (0.16), and CuSO4·5H2O (0.015). To assess the eﬀect of nitrogen concentration on itaconic acid production from cellulosic hydrolysate, the following three diﬀerent concentrations of (NH4)2SO4 were evaluated: 1, 3, and 5 g/L. For comparison, hydrolysate without any nutrient supplementation was also used as fermentation medium. Fermentation experiments were carried out in 250-mL Erlenmeyer ﬂasks at 35 ◦C and 200 rpm for 3 to 5 days (72 h to 120 h). A working volume of 50 mL was used in the experiments with pure carbon sources. Experiments performed to assess the eﬀect of aeration on itaconic acid production from cellulosic hydrolysate were carried out with diﬀerent working volumes varying from 20 to 50 mL in order to result in diﬀerent air-to-liquid ratios (Vair/Vm) as shown in Table 1. All other fermentations from cellulosic hydrolysate medium were performed using a working volume of 30 mL. All experiments were carried out in duplicate. 4 of 12 Energies 2020, 13, 1654 Table 1. Diﬀerent air-to-liquid ratios used for the fermentation experiments. Volume of Medium (mL) Air Column (mL) Vf/Vm Vair/Vm 20 230 12.5 11.5 30 220 8.33 7.33 50 200 5.00 4.00 Vf, volume of the ﬂask; Vm, volume of medium; Vair, volume of air. Table 1. Diﬀerent air-to-liquid ratios used for the fermentation experiments. 2.5. Analytical Methods and Statistical Analysis 3.2. Cellulose Pulp as Feedstock for Itaconic Acid Production Since glucose was the best carbon source for itaconic acid production by A. terreus, a cellulose-rich material can be considered to be the ideal candidate for use in this bioprocess as an alternative to replace pure glucose as the carbon source. During recent years, studies have reported the use of cellulose pulp as a substrate for the production of diﬀerent compounds by fermentation, including ethanol and hydrogen [19,20]. However, there are no studies reporting the use of bleached cellulose pulp for the production of itaconic acid. This study conﬁrms that bleached cellulose pulp can be a feasible feedstock for use in the production of itaconic acid since this material is produced in high amounts in the pulp and paper industries and has attracted great interest for use in the production of valuable compounds (rather than for use in the production of paper only). In addition to its great availability, other important advantages of using bleached cellulose pulp for itaconic acid production include its high content of cellulose (which can be converted into glucose) and the possibility of applying an enzymatic hydrolysis directly, without previous pretreatment, saving time and energy, which can positively impact on the ﬁnal costs of the itaconic acid production process, making it more economical. The cellulose pulp used in this study was produced from Eucalyptus wood. The chemical composition of this material was cellulose 89.7% and hemicellulose 10.3% (dry weight). Lignin was present in trace amount. Enzymatic hydrolysis of this material under the conditions used in this study yielded a hydrolysate containing 72.3 g/L of glucose and 14.8 g/L of xylose, representing a rich carbon source for use as fermentation medium by A. terreus. Fermentation results from this medium are discussed in the next sections. 3.1. Carbon Source Assessment Cellulose Pulp as Feedstock for Itaconic Acid Production 3.1. Carbon Source Assessment With the aim of identifying potential raw materials for the production of itaconic acid, initial experiments were carried out to evaluate the performance of A. terreus when cultivated in diﬀerent carbon sources. As can be seen in Table 2, hexoses, especially glucose, were the preferred carbon sources used by the strain to produce itaconic acid. Glucose has also been reported as being the preferred carbon source for other ﬁlamentous fungi, and this could be explained by the fact that this hexose enables the most direct conversion pathway, eliminating the need for additional biochemical steps [18]. In the present study, the production of itaconic acid from other carbon sources including xylose, sucrose, mannose, and glycerol were clearly lower that that observed from glucose (Table 2). The relevance of glucose as compared with the other carbon sources is also highlighted by the values of itaconic acid yield per substrate consumed, which was of 0.61 g/g of glucose consumed, more than three times higher than that observed for mannose, which was the second best carbon source evaluated in our study. 5 of 12 Energies 2020, 13, 1654 Table 2. Itaconic acid concentration and yield and biomass yield for the fermentations with A. terreus using diﬀerent carbon sources. Table 2. Itaconic acid concentration and yield and biomass yield for the fermentations with A. terreus using diﬀerent carbon sources. Carbon Source Itaconic Acid (g/L) YP/S (g/g) YX/S (g/g) Glucose 24.85 0.61 0.22 Xylose 1.85 0.09 0.61 Sucrose 2.15 0.05 0.28 Mannose 9.16 0.18 0.40 Glycerol 2.20 0.05 0.43 YP/S, itaconic acid yield per sugar consumed; YX/S, biomass yield per sugar consumed. It is also interesting to note that the strain presented a very good ability to consume all the carbon sources, including C3 (glycerol), C5 (xylose), and C12 (sucrose), being a potential candidate for use in bioprocesses. However, unlike glucose, the other carbon sources were mainly used for biomass growth instead of itaconic acid formation, which is evidenced by the higher values of biomass yield (YX/S) obtained (Table 2). The biomass formation in glucose was the lowest as compared with the other carbon sources, which was due to the main use of this carbon source for product formation. These results play a crucial role in selecting novel lignocellulosic biomass sources for use on the production of itaconic acid, opening up good possibilities for integrating the production of this acid in a bioreﬁnery. 3.2. 3.3. Hydrolysate Sterilization Concentration of 5-HMF and furfural in the cellulose pulp hydrolysate after sterilization by three diﬀerent methods. Figure 1. Concentration of 5-HMF and furfural in the cellulose pulp hydrolysate after sterilization by h diff h d Figure 1. Concentration of 5-HMF and furfural in the cellulose pulp hydrolysate after sterilization by three diﬀerent methods. The carbon-to-nitrogen ratio (C:N ratio) is another important characteristic of the hydrolysate that can strongly affect the microbial performance during fermentation, being of high importance to define a suitable C:N ratio to obtain high product yield during fermentation [23]. Analyses of the carbon and nitrogen contents in the cellulose pulp hydrolysate revealed that the carbon composition was not affected by any of the sterilization methods evaluated in this study. However, the nitrogen content was changed, leading to hydrolysates with different C:N ratios (Table 3). Sterilization by filtration clearly resulted in a medium with lower content of nitrogen, which was also visually cleaner and more translucent than the hydrolysates sterilized by autoclaving (figure not shown). According to the literature, nitrogen limitation can be beneficial for the production of organic acid [18]. Since nitrogen is required for biomass production, lack of nitrogen can slow down cell growth, to which some fungi respond by increasing the organic acid production [18]. In addition, high C:N ratios would direct more carbon into the tricarboxylic acid (TCA) cycle, allowing for higher productivities [24]. Ratios that are too high, however, could lead to reduced productivity due to substrate inhibition The carbon-to-nitrogen ratio (C:N ratio) is another important characteristic of the hydrolysate that can strongly aﬀect the microbial performance during fermentation, being of high importance to deﬁne a suitable C:N ratio to obtain high product yield during fermentation [23]. Analyses of the carbon and nitrogen contents in the cellulose pulp hydrolysate revealed that the carbon composition was not aﬀected by any of the sterilization methods evaluated in this study. However, the nitrogen content was changed, leading to hydrolysates with diﬀerent C:N ratios (Table 3). Sterilization by ﬁltration clearly resulted in a medium with lower content of nitrogen, which was also visually cleaner and more translucent than the hydrolysates sterilized by autoclaving (ﬁgure not shown). According to the literature, nitrogen limitation can be beneﬁcial for the production of organic acid [18]. 3.3. Hydrolysate Sterilization Sterilization is an important step to be performed when a pure culture has to be used in a bioprocess. However, since sterilization conditions can aﬀect the composition of the hydrolysate, three diﬀerent methods were evaluated in this study with the aim of selecting the option that most favors the production of itaconic acid by fermentation. Sugar degradation with consequent formation of toxic compounds (5-HMF and furfural) and carbon-to-nitrogen ratio (C:N ratio) were the responses considered to select the best sterilization method. Although the temperatures used for autoclaving could promote some degradation of glucose and xylose into 5-HMF and furfural, respectively, the formation of these compounds was low for all three sterilization methods evaluated (Figure 1). For the membrane sterilization, the 5-HMF obtained 6 of 12 12 Energies 2020, 13, 1654 Energies 2019 12 x FOR was mostly likely residual and was already present after the enzymatic hydrolysis. It is well known that 5-HMF and furfural are toxic compounds that can aﬀect the microbial performance during fermentation [21]. In the case of itaconic acid production, 5-HMF and furfural concentrations of 0.4 g/L have been reported as being toxic for A. terreus, inhibiting the growth, sugar utilization, and product formation [22]. These values are much higher than the concentrations found in the cellulose pulp hydrolysate, independent of the sterilization method applied (Figure 1). Therefore, it can be concluded that none of the sterilization methods was able to promote signiﬁcant degradation of sugars and formation of toxic compounds at a level unsuitable for fermentation. was mostly likely residual and was already present after the enzymatic hydrolysis. It is well known that 5-HMF and furfural are toxic compounds that can affect the microbial performance during fermentation [21]. In the case of itaconic acid production, 5-HMF and furfural concentrations of 0.4 g/L have been reported as being toxic for A. terreus, inhibiting the growth, sugar utilization, and product formation [22]. These values are much higher than the concentrations found in the cellulose pulp hydrolysate, independent of the sterilization method applied (Figure 1). Therefore, it can be concluded that none of the sterilization methods was able to promote significant degradation of sugars and formation of toxic compounds at a level unsuitable for fermentation. Figure 1. Concentration of 5-HMF and furfural in the cellulose pulp hydrolysate after sterilization by three different methods Figure 1. 3.3. Hydrolysate Sterilization Since nitrogen is required for biomass production, lack of nitrogen can slow down cell growth, to which some fungi respond by increasing the organic acid production [18]. In addition, high C:N ratios would direct more carbon into the tricarboxylic acid (TCA) cycle, allowing for higher productivities [24]. Ratios that are too high, however, could lead to reduced productivity due to substrate inhibition [18]. T bl 3 C b d it t t i th h d l t ft t ili ti b diff t th d Table 3. Carbon and nitrogen contents in the hydrolysate after sterilization by diﬀerent methods. Table 3 Carbon and nitrogen contents in the hydrolysate after sterilization by different methods Table 3. Carbon and nitrogen contents in the hydrolysate after sterilization by diﬀerent methods. ] Table 3. Carbon and nitrogen contents in the hydrolysate after sterilization by different methods. Table 3. Carbon and nitrogen contents in the hydrolysate after sterilization by diﬀerent methods. St ili ti M th d C b (% / ) Nit (% / ) Table 3. Carbon and nitrogen contents in the hydrolysate after sterilization by different methods. Sterilization Method Carbon (% w/w) Nitrogen (% w/w) Membrane filter 0.2 μm 36.16 ± 0.04 0.10 ± 0.01 Autoclave 112 °C 36.35 ± 0.14 0.17 ± 0.00 Autoclave 121 °C 36.30 ± 0.07 0.25 ± 0.06 Table 3. Carbon and nitrogen contents in the hydrolysate after sterilization by diﬀerent methods. Sterilization Method Carbon (% w/w) Nitrogen (% w/w) Membrane ﬁlter 0.2 µm 36.16 ± 0.04 0.10 ± 0.01 Autoclave 112 ◦C 36.35 ± 0.14 0.17 ± 0.00 Autoclave 121 ◦C 36.30 ± 0.07 0.25 ± 0.06 In order to select the sterilization method more suitable for the production of itaconic acid by A. terreus, fermentation assays were performed using the sterilized hydrolysates, with and without nutrient supplementation. No biomass growth or glucose consumption were observed from media without nutrient addition, thus, confirming the necessity of adding nutrients to all the hydrolysates. In order to select the sterilization method more suitable for the production of itaconic acid by A. terreus, fermentation assays were performed using the sterilized hydrolysates, with and without nutrient supplementation. No biomass growth or glucose consumption were observed from media without nutrient addition, thus, conﬁrming the necessity of adding nutrients to all the hydrolysates. 3.3. Hydrolysate Sterilization The results obtained from fermentation of sterilized hydrolysates with nutrient supplementation showed a clear advantage for the method of sterilization by membrane ﬁltration (Figure 2), which provided 7 of 12 on ch Energies 2020, 13, 1654 The results obtain showed a clear ad the highest itaconic acid yield (0.52 g/g) and productivity (0.40 g/L·h) after 72 h (Table 4). It is also interesting to note that glucose consumption and itaconic acid production were maximum at 72 h of fermentation but decreased afterwards, indicating possible consumption of the product when glucose, the main carbon source, was exhausted. is also interesting to note that glucose consumption and itaconic acid production were maximum at 72 h of fermentation but decreased afterwards, indicating possible consumption of the product when glucose, the main carbon source, was exhausted. 0 20 40 60 80 100 120 0 10 20 30 40 50 60 70 Glucose (g/L) Time (h) 0 5 10 15 20 25 30 Itaconic acid (g/L) Figure 2. Glucose (solid lines) and itaconic acid (dashed lines) concentrations during the fermentation of the cellulose pulp hydrolysate sterilized by different methods. Membrane filter 0.2 μm (); autoclave at 112 °C (); and autoclave at 121 °C (). All the hydrolysates were supplemented with b d f d Figure 2. Glucose (solid lines) and itaconic acid (dashed lines) concentrations during the fermentation of the cellulose pulp hydrolysate sterilized by diﬀerent methods. Membrane ﬁlter 0.2 µm (■); autoclave at 112 ◦C (•); and autoclave at 121 ◦C (▲). All the hydrolysates were supplemented with nutrients to be used as fermentation medium. 0 20 40 60 80 100 120 0 10 20 30 40 50 60 70 Glucose (g/L) Time (h) 0 5 10 15 20 25 30 Itaconic acid (g/L) Figure 2. Glucose (solid lines) and itaconic acid (dashed lines) concentrations during the fermentation of the cellulose pulp hydrolysate sterilized by different methods. Membrane filter 0.2 μm (); autoclave at 112 °C (); and autoclave at 121 °C (). All the hydrolysates were supplemented with Figure 2. Glucose (solid lines) and itaconic acid (dashed lines) concentrations during the fermentation of the cellulose pulp hydrolysate sterilized by diﬀerent methods. Membrane ﬁlter 0.2 µm (■); autoclave at 112 ◦C (•); and autoclave at 121 ◦C (▲). All the hydrolysates were supplemented with nutrients to be used as fermentation medium. 3.3. Hydrolysate Sterilization Analysis of the fermentation parameters (Table 4) clearly indicated that the hydrolysate sterilized by the filtration method promoted the best fermentation performance, resulting in the highest values of itaconic acid yield (YP/S) and productivity (QP). Such a result could be attributed to the C:N ratio present in the hydrolysate. The filtration method resulted in lower nitrogen content in the hydrolysate, and, as a consequence, in a higher C:N ratio that could have changed the metabolism towards acid production as opposed to fungal growth, thus, explaining the highest production of itaconic acid and the lowest production of biomass obtained from this medium. It is also worth noting that the production of itaconic acid from cellulose pulp hydrolysate sterilized by membrane filtration obtained in the present study compares very well to other studies on the production of this acid from different biomass hydrolysates [25,26]. When corn starch and wheat bran hydrolysates were used as fermentation medium for A. terreus, itaconic acid yields of 0.41-0.42 (g/g) were obtained [26]. Analysis of the fermentation parameters (Table 4) clearly indicated that the hydrolysate sterilized by the ﬁltration method promoted the best fermentation performance, resulting in the highest values of itaconic acid yield (YP/S) and productivity (QP). Such a result could be attributed to the C:N ratio present in the hydrolysate. The ﬁltration method resulted in lower nitrogen content in the hydrolysate, and, as a consequence, in a higher C:N ratio that could have changed the metabolism towards acid production as opposed to fungal growth, thus, explaining the highest production of itaconic acid and the lowest production of biomass obtained from this medium. It is also worth noting that the production of itaconic acid from cellulose pulp hydrolysate sterilized by membrane ﬁltration obtained in the present study compares very well to other studies on the production of this acid from diﬀerent biomass hydrolysates [25,26]. When corn starch and wheat bran hydrolysates were used as fermentation medium for A. terreus, itaconic acid yields of 0.41–0.42 (g/g) were obtained [26]. Table 4. Fermentation parameters obtained for the production of itaconic acid by A. terreus from cellulose pulp hydrolysate sterilized by different methods. Table 4. Fermentation parameters obtained for the production of itaconic acid by A. terreus from cellulose pulp hydrolysate sterilized by diﬀerent methods. 3.3. Hydrolysate Sterilization Sterilization Method YP/S (g/g) YX/S (g/g) QP (g/L·h) Membrane filter 0.2 μm 0.52 ± 0.01 0.13 ± 0.02 0.40 ± 0.01 Autoclave 112 °C 0.23 ± 0.04 0.23 ± 0.17 0.10 ± 0.03 Autoclave 121 °C 0.42 ± 0.02 0.28 ± 0.03 0.26 ± 0.01 YP/S, itaconic acid yield per sugar consumed; YX/S, biomass yield per sugar consumed; and QP, itaconic acid productivity. Sterilization Method YP/S (g/g) YX/S (g/g) QP (g/L·h) Membrane ﬁlter 0.2 µm 0.52 ± 0.01 0.13 ± 0.02 0.40 ± 0.01 Autoclave 112 ◦C 0.23 ± 0.04 0.23 ± 0.17 0.10 ± 0.03 Autoclave 121 ◦C 0.42 ± 0.02 0.28 ± 0.03 0.26 ± 0.01 YP/S, itaconic acid yield per sugar consumed; YX/S, biomass yield per sugar consumed; and QP, itaconic acid productivity. On the basis of the above, sterilization by membrane filtration was selected as the most suitable sterilization technique as it provided the best results of itaconic acid titer, yield and productivity, and therefore was the sterilization method used in all the subsequent experiments. On the basis of the above, sterilization by membrane ﬁltration was selected as the most suitable sterilization technique as it provided the best results of itaconic acid titer, yield and productivity, and therefore was the sterilization method used in all the subsequent experiments. 3.4. Eﬀect of Aeration on the Production of Itaconic Acid 3.4. Eﬀect of Aeration on the Production of Itaconic Acid A suﬃcient oxygen supply is a fundamental requirement for a successful performance of the microbial strain during fermentation processes. To better understand its eﬀect on the fermentation of cellulose pulp hydrolysate by A. terreus, three diﬀerent aeration conditions were tested in a shake ﬂask 8 of 12 Energies 2020, 13, 1654 setup (Table 1), which were promoted by varying the working volume used in the ﬂasks. A similar setup has been used and discussed in other studies to understand the eﬀect of aeration during fermentation in ﬂasks [27,28]. Results of these experiments are summarized in Table 5. setup (Table 1), which were promoted by varying the working volume used in the ﬂasks. A similar setup has been used and discussed in other studies to understand the eﬀect of aeration during fermentation in ﬂasks [27,28]. Results of these experiments are summarized in Table 5. Table 5. Fermentation parameters obtained during the production of itaconic acid by A. terreus from cellulose pulp hydrolysate under diﬀerent aeration conditions. Table 5. Fermentation parameters obtained during the production of itaconic acid by A. terreus from cellulose pulp hydrolysate under diﬀerent aeration conditions. Vair/Vm YP/S (g/g) YX/S (g/g) QP (g/L·h) 11.5 0.20 ± 0.02 0.28 ± 0.02 0.17 ± 0.03 7.33 0.52 ± 0.01 0.13 ± 0.02 0.40 ± 0.06 4.00 0.21 ± 0.06 0.29 ± 0.04 0.13 ± 0.06 YP/S, itaconic acid yield per sugar consumed; YX/S, biomass yield per sugar consumed; and QP, itaconic acid productivity. Interestingly, the two boundary conditions, Vair/Vm of 11.5 and 4, showed a decreased fermentation performance as compared with that observed for the intermediate condition, Vair/Vm of 7.33 (Table 5). In addition, the biomass yield was lower for a Vair/Vm of 7.33, revealing that an increased ﬂux of carbon was deviated to the product formation under this oxygen condition. These results indicate that oxygen plays an important role in the production of itaconic acid by A. terreus from cellulose pulp hydrolysate. Therefore, selecting the ideal condition is highly important to maximize the product formation since conditions of excess or limitation of oxygen did not provide the best results. According to some authors, interrupting aeration can completely stop the production of itaconic acid by A. terreus [29]. Moreover, experiments using diﬀerent shaking speeds in ﬂasks showed that lowering the RPMs had a negative eﬀect on the production of itaconic acid [22]. 3.4. Eﬀect of Aeration on the Production of Itaconic Acid On the other hand, research with Aspergillus niger revealed that a reduced level of dissolved oxygen has a positive eﬀect on the production of itaconic acid since high levels of dissolved oxygen increase the production of other organic acids such as citric and oxalic acid, which redirects carbon away from itaconic acid production [30]. Therefore, it is important to manage the aeration of the system carefully according to the strain and medium conditions used. Low and high concentrations of dissolved oxygen could both have an adverse eﬀect on the production of itaconic acid. Research is, therefore, required to establish the best oxygen level to be used during fermentation. This is also of great importance for upscaling experiments in bioreactors. Since a V i /V of 7 33 was the oxygen condition that provided the best results of itaconic acid g g p p g p Since a Vair/Vm of 7.33 was the oxygen condition that provided the best results of itaconic acid production, this condition was selected and used in the subsequent experiments. 3.5. Eﬀect of C:N Ratio on the Production of Itaconic Acid Considering that the previous experiments on the sterilization method suggested a signiﬁcant inﬂuence of the C:N ratio on the production of itaconic acid, additional experiments were performed at this step to explore such eﬀect with the aim of selecting conditions able to improve the production of itaconic acid from cellulose pulp hydrolysate. As a ﬁrst approach, experiments consisted in changing the nitrogen availability in the medium by varying the concentration of (NH4)2SO4 added to it. As can be seen in Table 6, the addition of 1 g/L (NH4)2SO4 did not provide suﬃcient nitrogen for the microorganism to properly metabolize the carbon source and convert it into itaconic acid. Better results were obtained for the other two nitrogen concentrations tested, 3 and 5 g/L. From these, supplementation of the medium with 3 g/L (NH4)2SO4 gave the best results of itaconic acid production, with yield and productivity of 0.52 g/g and 0.40 g/L·h, respectively. These results conﬁrm that the production of itaconic acid can be improved by using an appropriate C:N ratio. Nitrogen limitation or excess are both non ideal conditions for the metabolism of A. terreus go through the itaconic acid formation. The inﬂuence of diﬀerent nitrogen sources and concentrations on the production of itaconic acid has also been reported in other studies using diﬀerent microbial strains and fermentation media. For example, the production of itaconic acid by the fungus Ustilago maydis in medium containing 200 g/L of glucose was improved when the concentration of NH4Cl added as nitrogen source was 9 of 12 Energies 2020, 13, 1654 increased from 15 to 75 mM [31]; while the production of itaconic acid by A. terreus ATCC 10020 from rice husk hydrolysate containing 15 g/L of glucose was improved when the medium was supplemented with 1.3 g/L NaNO3 and 1.1 g/L (NH4)2SO4 [12]. This makes it possible to conclude that diﬀerent strains have diﬀerent nitrogen requirements for their metabolism and, according to the medium used for fermentation, diﬀerent concentrations of nitrogen should be added to promote the best performance of the strain towards product formation. Table 6. Eﬀect of the nitrogen concentration on the fermentation parameters obtained during the production of itaconic acid by A. terreus from cellulose pulp hydrolysate. 3.5. Eﬀect of C:N Ratio on the Production of Itaconic Acid Nitrogen (g/L) YP/S (g/g) YX/S (g/g) QP (g/L·h) 1 0.01 ± 0.08 0.30 ± 0.25 0.002 ± 0.08 3 0.52 ± 0.01 0.13 ± 0.02 0.40 ± 0.06 5 0.49 ± 0.01 0.21 ± 0.01 0.33 ± 0.00 YP/S, itaconic acid yield per sugar consumed; YX/S, biomass yield per sugar consumed; and QP, itaconic acid productivity. Table 6. Eﬀect of the nitrogen concentration on the fermentation parameters obtained during the production of itaconic acid by A. terreus from cellulose pulp hydrolysate. As a second approach to explore the eﬀect of the C:N ratio on the performance of A. terreus to produce itaconic acid from cellulose pulp hydrolysate, small changes in the carbon composition were made for a ﬁxed medium supplementation of 3 g/L (NH4)2SO4 (which gave the best results of itaconic acid production in the previous experiments). According to the results, when the initial concentration of glucose in the medium was increased from 45 to 60 g/L, a signiﬁcant increase in the production of itaconic acid could be observed, which resulted in 2.3 times and 3 times higher values of yield and productivity, respectively (Table 7). These results reinforce that increasing the initial concentration of carbon source is an important strategy to result in a higher production of itaconic acid by A. terreus. However, within the scope of this study, higher concentrations of initial glucose were not evaluated since, as the nitrogen supplementation was ﬁxed, increased carbon sources would lead to much higher C:N ratios, which could negatively impact on the production of itaconic acid. For future experiments, higher concentrations of glucose should be tested using an appropriate nitrogen supplementation to oﬀer the ideal C:N balance required by the strain to maximize the formation of itaconic acid from cellulose pulp hydrolysate. Table 7. Eﬀect of the substrate concentration on the fermentation parameters obtained during the production of itaconic acid by A. terreus from cellulose pulp hydrolysate. Glucose (g/L) YP/S (g/g) YX/S (g/g) QP (g/L·h) 45 0.27 ± 0.02 0.24 ± 0.02 0.17 ± 0.00 53 0.52 ± 0.01 0.13 ± 0.18 0.40 ± 0.01 60 0.62 ± 0.02 0.18 ± 0.04 0.52 ± 0.02 YP/S, itaconic acid yield per sugar consumed; YX/S, biomass yield per sugar consumed; and QP, itaconic acid productivity. Table 7. Eﬀect of the substrate concentration on the fermentation parameters obtained during the production of itaconic acid by A. terreus from cellulose pulp hydrolysate. 4. Conclusions To accelerate the use of lignocellulosic feedstocks in fermentative processes it is crucial to select the right biomass for the desired process. This study demonstrated that bleached cellulose pulp is a potential candidate for use on the production of itaconic acid by fermentation since it is highly rich in cellulose that can easily be converted into glucose by enzymatic hydrolysis without requiring a previous step of pretreatment. This is in fact an important aspect contributing to the economic feasibility of the fermentation process for itaconic acid production, since pretreatment is usually a very energy-intensive step and impacts signiﬁcantly on the ﬁnal costs of the process and the product. Other important ﬁndings of this study were related to the fermentation of the glucose-rich hydrolysate produced from cellulose pulp. Due to the presence of glucose as the main component, no lignin or sugar degradation products in the medium that could negatively aﬀect the strain performance, the only concern is to establish conditions that can direct the metabolism of the strain towards the product formation with minimum use of carbon source for biomass growth. Within this study, it was demonstrated that the C:N ratio and the oxygen availability play important roles in the production of itaconic acid by A. terreus from cellulose pulp hydrolysate and should be carefully considered in subsequent studies in a bioreactor setup. Increasing the initial carbon source was also a strategy able to result in better production of itaconic acid and should be further explored taking into account the use of an appropriate C:N ratio during the experiments. Finally, sterilization of the hydrolysate before fermentation is a required step that can also aﬀect the medium composition leading to an unbalance in the C:N ratio, being the sterilization by membrane ﬁltration the most recommended method to result in a better fermentation performance. Author Contributions: Conceptualization, S.I.M.; Methodology, S.I.M., G.D., and C.K.Y.; Investigation, M.C. and P.D.; Resources: S.I.M.; Data curation, A.A.J.K., M.C., P.D., G.D., C.K.Y., and S.I.M.; Writing—original draft preparation, A.A.J.K., G.D., and S.I.M.; Writing—review and editing, S.I.M.; Supervision, S.I.M., G.D. and C.K.Y.; Project administration: S.I.M.; Funding Acquisition, S.I.M. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by the Novo Nordisk Foundation, Denmark (grant number NNF10CC1016517). 4. Conclusions Acknowledgments: Special thanks to Suzano S/A (Brazil) for supplying the cellulose pulp used in this research, and to Novozymes (Denmark) for providing the enzyme concentrate Cellic® CTec2. Conﬂicts of Interest: The authors declare no conﬂict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. 3.5. Eﬀect of C:N Ratio on the Production of Itaconic Acid Glucose (g/L) YP/S (g/g) YX/S (g/g) QP (g/L·h) Table 7. Eﬀect of the substrate concentration on the fermentation parameters obtained during the production of itaconic acid by A. terreus from cellulose pulp hydrolysate. Finally, considering the diﬀerent strategies evaluated in the present study, sterilization of the cellulose pulp hydrolysate by membrane ﬁltration, medium supplementation with 3 g/L (NH4)2SO4, 60 g/L of initial glucose concentration, and oxygen availability of 7.33 (volume of air/volume of medium) were the most suitable to maximize the production of itaconic acid by A. terreus, resulting in a production of 37.5 g/L, corresponding to a yield of 0.62 g/g glucose consumed, and productivity of 0.52 g/L·h. These values compare very well to other recent studies on the production of itaconic acid by A. terreus from diﬀerent lignocellulosic feedstocks (Table 8) and conﬁrm the feasibility of using bleached cellulose pulp for this application. These results can still be improved by optimization of the fermentation conditions using a bioreactor setup, which will be investigated in a next study. 10 of 12 10 of 12 Energies 2020, 13, 1654 Table 8. Production of itaconic acid by A. terreus from diﬀerent lignocellulosic feedstocks. All the values correspond to experiments in ﬂasks. Feedstock, Glucose Concentration (g/L) Biomass Processing Aspergillus terreus Strain Itaconic Acid (g/L) Reference Corn stover, 54 Steam explosion pretreatment + Enzymatic hydrolysis CICC 2452, mutant AT-90 19.3 [10] Wheat chaﬀ, ≈50 Alkaline pretreatment + Enzymatic hydrolysis DSM 23081 27.7 [11] Rice husks, 35 Acid pretreatment + detoxiﬁcation ATCC 10020 1.9 [12] Corn cobs, 7.7 Enzymatic hydrolysis DSM 826 0.9 [13] Bleached cellulose pulp, 60 Enzymatic hydrolysis NRRL 1960 37.5 Present study 2. De Jong, E.; Stichnothe, H.; Bell, G.; Jørgensen, H. Bio-Based Chemicals. A 2020 Update. IEA Bioenergy Task 42. Available online: https://www.academia.edu/42073867/Bio-Based_Chemicals_A_2020_Update (accessed on 26 February 2020). 1. Dragone, G.; Kerssemakers, A.A.J.; Driessen, J.L.S.P.; Yamakawa, C.K.; Brumano, L.P.; Mussatto, S.I. Innovation and strategic orientations for the development of advanced bioreﬁneries. Bioresour. Technol. 2020, 302, 122847. [CrossRef] [PubMed] References 1. Dragone, G.; Kerssemakers, A.A.J.; Driessen, J.L.S.P.; Yamakawa, C.K.; Brumano, L.P.; Mussatto, S.I. Innovation and strategic orientations for the development of advanced bioreﬁneries. Bioresour. Technol. 2020, 302, 122847. [CrossRef] [PubMed] 2. De Jong, E.; Stichnothe, H.; Bell, G.; Jørgensen, H. Bio-Based Chemicals. A 2020 Update. IEA Bioenergy Task 42. Available online: https://www.academia.edu/42073867/Bio-Based_Chemicals_A_2020_Update (accessed on 26 February 2020). 2. De Jong, E.; Stichnothe, H.; Bell, G.; Jørgensen, H. Bio-Based Chemicals. A 2020 Update. IEA Bioenergy Task 42. Available online: https://www.academia.edu/42073867/Bio-Based_Chemicals_A_2020_Update (accessed on 26 February 2020). Energies 2020, 13, 1654 11 of 12 3. Mussatto, S.I.; Dragone, G.M. Biomass pretreatment, bioreﬁneries, and potential products for a bioeconomy development. In Biomass Fractionation Technologies for a Lignocellulosic Feedstock Based Bioreﬁnery; Elsevier: Amsterdam, The Netherlands, 2016; pp. 1–22. ISBN 978-0-12-802323-5. . Teleky, B.-E.; Vodnar, D.C. Biomass-derived production of itaconic acid as a building block in speci polymers. Polymers 2019, 11, 1035. [CrossRef] [PubMed] 5. Analysts Global Industry Inc. Itaconic Acid (IA)—A Global Strategic Business Report. Available online: https://www.strategyr.com/pressMCP-6465.asp (accessed on 26 February 2020). 6. Transparency Market Research. Rising Demand in Manufacturing of Superabsorbent Polymers to Inundate Itaconic Acid Market. Available online: https://www.transparencymarketresearch.com/pressrelease/itaconic- acid-market.htm (accessed on 26 February 2020). 7. Matthis, S. Pulp Price Erosion Persist in October. Available online: https://www.pulpapernews.com/20191028/ 10851/pulp-price-erosion-persists-october (accessed on 26 February 2020). 8. Suzano. NDR Boston and Itaú BBA 14th Annual LatAm CEO Conference. Available online: http://ri.suzano. com.br/ptb/7597/ApresentaoNDR_Conferncia_USA_EN.pdf (accessed on 26 February 2020). 9. Regestein, L.; Klement, T.; Grande, P.; Kreyenschulte, D.; Heyman, B.; Maßmann, T.; Eggert, A.; Sengpiel, R.; Wang, Y.; Wierckx, N.; et al. From beech wood to itaconic acid: Case study on bioreﬁnery process integration. Biotechnol. Biofuels 2018, 11, 279. [CrossRef] [PubMed] 10. Li, X.; Zheng, K.; Lai, C.; Ouyang, J.; Yong, Q. Improved itaconic acid production from undetoxiﬁed enzymatic hydrolysate of steam-exploded corn stover using an Aspergillus terreus mutant generated by atmospheric and room temperature plasma. BioResources 2016, 11, 9047–9058. [CrossRef] 11. Krull, S.; Eidt, L.; Hevekerl, A.; Kuenz, A.; Prüße, U. Itaconic acid production from wheat chaﬀby Aspergillus terreus. Process Biochem. 2017, 63, 169–176. [CrossRef] 12. Pedroso, G.B.; Montipó, S.; Mario, D.A.N.; Alves, S.H.; Martins, A.F. Building block itaconic acid from left-over biomass. Biomass Convers. Bioreﬁnery 2017, 7, 23–35. [CrossRef] 13. Jimenez-Quero, A.; Pollet, E.; Zhao, M.; Marchioni, E.; Averous, L.; Phalip, V. Itaconic and fumaric acid production from biomass hydrolysates by Aspergillus strains. J. Microbiol. Biotechnol. 2016, 26, 1557–1565. [CrossRef] 4. References Sluiter, A.; Hames, B.; Ruiz, R.; Scarlata, C.; Sluiter, J.; Templeton, D. Determination of Ash in Biomass; Techn Report NREL/TP-510-42622; National Renewable Energy Laboratory: Golden, CO, USA, 2008. 15. Sluiter, A.; Hames, B.; Ruiz, R.; Scarlata, C.; Sluiter, J.; Templeton, D.; Crocker, D. Determination of Structural Carbohydrates and Lignin in Biomass; Technical Report NREL/TP-510-42618; National Renewable Energy Laboratory: Golden, CO, USA, 2012. 16. Sluiter, A.; Ruiz, R.; Scarlata, C.; Sluiter, J.; Templeton, D. Determination of Extractives in Biomass; Technical Report NREL/TP-510-42619; National Renewable Energy Laboratory: Golden, CO, USA, 2008. 17. Adney, B.; Baker, J. Measurement of Cellulase Activities; Technical Report NREL/TP-510-42628; National Renewable Energy Laboratory: Golden, CO, USA, 2008. 18. Mondala, A.H. Direct fungal fermentation of lignocellulosic biomass into itaconic, fumaric, and malic acids: Current and future prospects. J. Ind. Microbiol. Biotechnol. 2015, 42, 487–506. [CrossRef] 19. Aierkentai, G.; Liang, X.; Uryu, T.; Yoshida, T. Eﬀective sacchariﬁcatio produce bioethanol. J. Fiber Sci. Technol. 2017, 73, 261–269. [CrossRef] 19. Aierkentai, G.; Liang, X.; Uryu, T.; Yoshida, T. Eﬀective sacchariﬁcation and fermentation of kraft pulp to produce bioethanol. J. Fiber Sci. Technol. 2017, 73, 261–269. [CrossRef] 20. Moreau, A.; Montplaisir, D.; Sparling, R.; Barnabé, S. Hydrogen, ethanol and cellulase production from pulp and paper primary sludge by fermentation with Clostridium thermocellum. Biomass Bioenergy 2015, 72, 256–262. [CrossRef] 21. Mussatto, S.I.; Roberto, I.C. Alternatives for detoxiﬁcation of diluted-acid lignocellulosic hydrolyzates for use in fermentative processes: A review. Bioresour. Technol. 2004, 93, 1–10. [CrossRef] [PubMed] 22. Saha, B.C.; Kennedy, G.J.; Bowman, M.J.; Qureshi, N.; Dunn, R.O. Factors aﬀecting production of itaconic acid from mixed sugars by Aspergillus terreus. Appl. Biochem. Biotechnol. 2019, 187, 449–460. [CrossRef] [PubMed] 23. Liu, Z.; Feist, A.M.; Dragone, G.; Mussatto, S.I. Lipid and carotenoid production from wheat straw hydrolysates by diﬀerent oleaginous yeasts. J. Clean. Prod. 2020, 249, 119308. [CrossRef] 24. Casas López, J.L.; Sánchez Pérez, J.A.; Fernández Sevilla, J.M.; Acién Fernández, F.G.; Molina Grima, E.; Chisti, Y. Production of lovastatin by Aspergillus terreus: Eﬀects of the C:N ratio and the principal nutrients on growth and metabolite production. Enzyme Microb. Technol. 2003, 33, 270–277. [CrossRef] 12 of 12 12 of 12 Energies 2020, 13, 1654 25. Reddy, C.S.K.; Singh, R. Enhanced production of itaconic acid from corn starch and market refuse fruits by genetically manipulated Aspergillus terreus SKR10. Bioresour. Technol. 2002, 85, 69–71. [CrossRef] 26. Kuenz, A.; Krull, S. Biotechnological production of itaconic acid—Things you have to know. Appl. Microbiol. References Biotechnol. 2018, 102, 3901–3914. [CrossRef] 7. Klöckner, W.; Büchs, J. Advances in shaking technologies. Trends Biotechnol. 2012, 30, 307–314. [CrossR 28. Silva, J.P.A.; Mussatto, S.I.; Roberto, I.C. The inﬂuence of initial xylose concentration, agitation, and aeration on ethanol production by Pichia stipitis from rice straw hemicellulosic hydrolysate. Appl. Biochem. Biotechnol. 2010, 162, 1306–1315. [CrossRef] 29. Gyamerah, M.H. Oxygen requirement and energy relations of itaconic acid fermentation by Aspergillus terreus NRRL 1960. Appl. Microbiol. Biotechnol. 1995, 44, 20–26. [CrossRef] 30. Li, A.; Pfelzer, N.; Zuijderwijk, R.; Brickwedde, A.; van Zeijl, C.; Punt, P. Reduced by-product formation and modiﬁed oxygen availability improve itaconic acid production in Aspergillus niger. Appl. Microbiol. Biotechnol. 2013, 97, 3901–3911. [CrossRef] 31. Maassen, N.; Panakova, M.; Wierckx, N.; Geiser, E.; Zimmermann, M.; Bölker, M.; Klinner, U.; Blank, L.M. Inﬂuence of carbon and nitrogen concentration on itaconic acid production by the smut fungus Ustilago maydis. Eng. Life Sci. 2014, 14, 129–134. [CrossRef] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W3092856724	https://www.nature.com/articles/s41467-020-19085-1.pdf	English	null	Non-dispersive infrared multi-gas sensing via nanoantenna integrated narrowband detectors	Nature communications	2,020	cc-by	9,469	1 School of Optical and Electronic Information and Wuhan National Research Center for Optoelectronics (WNLO), Huazhong University of Science and Technology, Wuhan 430074, China. 2 Department of Electrical Engineering, Yale University, New Haven, CT 06511, USA. 3 School of Physics, Shandong University, Jinan 250100, China. ✉email: feiyi@hust.edu.cn NATURE COMMUNICATIONS \| (2020) 11:5245 \| https://doi.org/10.1038/s41467-020-19085-1 \| www.nature.com/naturecommunications ARTICLE ARTICLE Non-dispersive infrared multi-gas sensing via nanoantenna integrated narrowband detectors Xiaochao Tan1, Heng Zhang1, Junyu Li1, Haowei Wan1, Qiushi Guo2, Houbin Zhu3, Huan Liu1 & Fei Yi 1✉ Non-dispersive infrared (NDIR) spectroscopy analyzes the concentration of target gases based on their characteristic infrared absorption. In conventional NDIR gas sensors, an infrared detector has to pair with a bandpass ﬁlter to select the target gas. However, mul- tiplexed NDIR gas sensing requires multiple pairs of bandpass ﬁlters and detectors, which makes the sensor bulky and expensive. Here, we propose a multiplexed NDIR gas sensing platform consisting of a narrowband infrared detector array as read-out. By integrating plasmonic metamaterial absorbers with pyroelectric detectors at the pixel level, the detectors exhibit spectrally tunable and narrowband photoresponses, circumventing the need for separate bandpass ﬁlter arrays. We demonstrate the sensing of H2S, CH4, CO2, CO, NO, CH2O, NO2, SO2. The detection limits of common gases such as CH4, CO2, and CO are 63 ppm, 2 ppm, and 11 ppm, respectively. We also demonstrate the deduction of the con- centrations of two target gases in a mixture. 1 School of Optical and Electronic Information and Wuhan National Research Center for Optoelectronics (WNLO), Huazhong University of Science and Technology, Wuhan 430074, China. 2 Department of Electrical Engineering, Yale University, New Haven, CT 06511, USA. 3 School of Physics, Shandong University, Jinan 250100, China. ✉email: feiyi@hust.edu.cn 1 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 T he mid-infrared (mid-IR) spectral range (wavelength λ ~ 2 to 20 µm) is known as the “molecular ﬁngerprint” region, where a wide variety of gas molecules exhibit highly characteristic rotational or vibrational transition bands1–5. Notably, in the mid-IR region, the absorption strengths of molecular transitions are typically 10–1000 times greater than those in the visible or near-IR region6. As such, mid-IR spec- troscopic gas sensors can be employed to uniquely identify and quantify the presence of substances with high sensitivity and selectivity. Non-dispersive infrared (NDIR) spectroscopy is one of mid-IR spectroscopic gas sensors that analyzes gases based on their characteristic absorption wavelengths in the mid-IR caused by their molecular vibrations1,7, which can ﬁnd profound appli- cations in traced gas sensing8,9, breadth analysis10,11, environ- mental monitoring12,13, to name a few. T is to introduce pixel-level spectral selectivity to mid-IR detectors by integrating plasmonic metamaterial absorbers (PMAs) onto the detector pixel. PMA is composed of arrays of metallic plas- monic resonators that can selectively absorb a certain spectral band of light and, therefore, can be regarded as nanoscale absorption ﬁlters19–23. Non-dispersive infrared multi-gas sensing via nanoantenna integrated narrowband detectors Xiaochao Tan1, Heng Zhang1, Junyu Li1, Haowei Wan1, Qiushi Guo2, Houbin Zhu3, Huan Liu1 & Fei Yi 1✉ p Following this idea, we propose a new NDIR architecture in which an array of narrowband PMA integrated pyroelectric ele- ments are used to spectrally resolve the absorption of multiple gases at the same time24. By tuning the geometry of metallic plasmonic resonators, the central detection wavelength of each element can be independently adjusted to match the characteristic absorption bands of different target gases. The multiplexed sen- sing platform can thus be used to analyze multiple target gases in a mixture with signiﬁcantly reduced device footprint and oper- ating time. Leveraging the proposed gas sensing platform, we demonstrate the sensing of 8 different gases: H2S, CH4, CO2, CO, NO, CH2O, NO2, SO2, with the detection limits of 489, 63, 2, 11, 17, 27, 54, 104 ppm, respectively. We also demonstrate that the concentrations of two target gases in a mixture can be deduced from the voltage responses of two narrowband detectors. Although the sensing platform in its current form is still bulkier than the commercial NDIR sensors, we believe that by reducing the thickness of the pyroelectric elements and improving the quality factors of the narrowband PMA, integrated multiplexed gas sensors with centimeter long sizes can be achieved. In conventional NDIR gas sensor, the light source is broadband and not pre-ﬁltered. When the light beam containing a wide range of wavelengths passes through and interacts with the sample gases in a chamber, only a portion of the optical energy is absorbed by the gases at their characteristic absorption wave- lengths14. To analyze the concentration of a target gas, a bandpass optical ﬁlter is typically added before the detectors to eliminate all unwanted wavelengths in the light beam and only allow the characteristic absorption wavelengths of the gas to reach the detector. In other words, the spectral selectivity in conventional NDIR architecture is enabled by the added ﬁlters, rather than the detectors (see Supplementary Note 1). In order to analyze several target gases in a mixture at the same time, one can simply implement multiple pairs of “bandpass ﬁlter + optical detector” in the NDIR gas sensor15,16. However, this scheme greatly increases the cost, the system complexity as well as the operating time, especially when the number of target gases is large17,18. Such challenge is fundamentally rooted in the lack of spectral selectivity of most commercially available mid-IR detectors. One way to avoid the need for separate optical ﬁlters Results Principle of operation. The schematic diagram of the proposed NDIR multiplexed gas sensing platform, which is composed of three parts: the broadband light source, the gas cell and the multiplexed sensor with necessary focusing optics is plotted in Fig. 1a. An example of packaged multiplexed pyroelectric sensor with different detection wavelengths for spectral sensing is shown b Au LiTaO3 SiO2 a c Mid-IR light source Gas cell Multiplexed sensor x y z P R tp Fig. 1 An overview of the proposed non-dispersive infrared (NDIR) architecture using metamaterial enabled narrowband pyroelectric detectors. a A schematic diagram of the gas sensing system based on the proposed NDIR architecture with an array of narrowband PMA integrated pyroelectric elements used as the spectral sensor. The system is composed of three parts: the broadband light source, the gas cell, and the multi-element sensor together with necessary focusing optics. b The joint package of the multiple pyroelectric elements with different detection wavelengths. c The device geometry of the narrowband detection element. From the top to the bottom are: the Au nanodisk antenna, the silicon dioxide spacer, the gold backplate that is also used as the top electrode of the pyroelectric element, the lithium tantalate (LT) substrate, and the gold bottom electrode. The length of the smallest repeatable unit is the periodicity P and the radius of the nanodisk is R. The area size of each absorber is 1 × 1 mm and the thickness tp of the LT substrate is 75 μm. To provide electrical access to the gold backplate buried underneath the silicon dioxide spacer, a window area beside each absorber is opened by removing the spacer. b Au LiTaO3 SiO2 a c Mid-IR light source Gas cell Multiplexed sensor x y z P R tp a Multiplexed sensor Multiplexed sensor Gas cell b c Fig. 1 An overview of the proposed non-dispersive infrared (NDIR) architecture using metamaterial enabled narrowband pyroelectric detectors. a A schematic diagram of the gas sensing system based on the proposed NDIR architecture with an array of narrowband PMA integrated pyroelectric elements used as the spectral sensor. The system is composed of three parts: the broadband light source, the gas cell, and the multi-element sensor together with necessary focusing optics. b The joint package of the multiple pyroelectric elements with different detection wavelengths. c The device geometry of the narrowband detection element. Results From the top to the bottom are: the Au nanodisk antenna, the silicon dioxide spacer, the gold backplate that is also used as the top electrode of the pyroelectric element, the lithium tantalate (LT) substrate, and the gold bottom electrode. The length of the smallest repeatable unit is the periodicity P and the radius of the nanodisk is R. The area size of each absorber is 1 × 1 mm and the thickness tp of the LT substrate is 75 μm. To provide electrical access to the gold backplate buried underneath the silicon dioxide spacer, a window area beside each absorber is opened by removing the spacer. NATURE COMMUNICATIONS \| (2020) 11:5245 \| https://doi.org/10.1038/s41467-020-19085-1 \| www.nature.com/naturecommunications 2 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 a 5μm c J (A/m2) Y (μm) Z (μm) b 1.0 0.3 0.2 0.1 0 –0.1 –0.2 –0.3 –1.5 –1 –0.5 0 0.5 1 1.5 ×106 2 1.5 1 0.5 0 0.8 0.6 0.4 0.2 Absorbance (a.u.) Absorption 0.0 0.8 0.6 0.4 0.2 0.0 2 3 4 Wavelength (μm) 5 6 7 8 The spectral and near-ﬁeld properties of the plasmonic metamaterial absorbers. a The scanning electron microscope (SEM) image of the gold isk antenna array. b The distribution of light-induced current density magnitude \|J \| and current density vector J in the YZ cut-plane of the MIM ber at λpeak = 5.73 μm, obtained by COMSOL, a ﬁnite element method-based solver. c The measured absorption spectra of 8 fabricated MIM bers compared to the infrared absorption bands of eight target gases: H2S, CH4, CO2, CO, NO, CH2O, NO2, SO2. See Supplementary Note 7 for the ption spectra in the full wavelength range. Source data are provided as a Source Data ﬁle. a 5μm J (A/m2) Y (μm) Z (μm) b 0.3 0.2 0.1 0 –0.1 –0.2 –0.3 –1.5 –1 –0.5 0 0.5 1 1.5 ×106 2 1.5 1 0.5 0 J (A/m2) Y (μm) Z (μm) b 0.3 0.2 0.1 0 –0.1 –0.2 –0.3 –1.5 –1 –0.5 0 0.5 1 1.5 ×106 2 1.5 1 0.5 0 a 5μm b a Z (μm) c c 1.0 0.8 0.6 0.4 0.2 Absorbance (a.u.) Absorption 0.0 0.8 0.6 0.4 0.2 0.0 2 3 4 Wavelength (μm) 5 6 7 8 Fig. 2 The spectral and near-ﬁeld properties of the plasmonic metamaterial absorbers. a The scanning electron microscope (SEM) image of the gold nanodisk antenna array. Results b The distribution of light-induced current density magnitude \|J \| and current density vector J in the YZ cut-plane of the MIM absorber at λpeak = 5.73 μm, obtained by COMSOL, a ﬁnite element method-based solver. c The measured absorption spectra of 8 fabricated MIM absorbers compared to the infrared absorption bands of eight target gases: H2S, CH4, CO2, CO, NO, CH2O, NO2, SO2. See Supplementary Note 7 for the absorption spectra in the full wavelength range. Source data are provided as a Source Data ﬁle. should be designed to match the characteristic absorption bands of different target gases in the mid-IR. Figure 2a shows the scanning electron microscope (SEM) image of the fabricated gold nanodisk antenna array (see Supplementary Note 2 for the details about the design of the MIM absorbers for 8 target gases, and Supplementary Note 3 about the fabrication and package of the narrowband detectors). To reveal the microscopic picture of resonant light trapping and dissipation of the optical energy, we need to look at the distribution of the optically induced currents in the absorber. Figure 2b plots the local distribution of the light- induced current density magnitude \|J\|and current density vector J in an MIM absorber at its resonant wavelength λpeak = 5.73 μm. Under the excitation of the y-polarized (Cartesian coordinate system in Fig. 1c) plane wave, the induced local currents oscil- lating along the y-direction are maximized in the center region of the nanodisk antenna and decrease in magnitude towards the edges of the antenna29, resulting in net electric charges and enhanced local electric ﬁelds near the antenna edges. The ohmic loss caused by the oscillating currents is mainly distributed at the lower surface of the antenna and the upper surface of the gold backplate, which serves to heat up the LT underneath. Figure 2c shows the measured absorption spectra of the 8 fabricated MIM absorbers and the characteristic IR absorption spectra of their target gases, whose absorption bands are far away from each other. It ought to be noted that by optimizing the antenna in Fig. 1b. The absorption ﬁlter is essentially a metal-insulator- metal (MIM)-based metamaterial absorber that consists a top layer of gold nanodisk antennas, a SiO2 spacer, and a gold backplate. The absorber is directly fabricated on top of a com- mercially available thin lithium tantalate (LT) substrate with pre- deposited gold electrodes on the top and bottom surfaces. Results The top gold electrode of the LT substrate is also used as the gold back- plate of the MIM absorbers for simplicity. LT as the sensing material offers a very broadband infrared response, enough to cover the characteristic absorption bands of typical gases, and a high pyroelectric coefﬁcient25. For simplicity, here we cut the LT substrate with integrated MIM absorbers into separate elements with different spectral responsivity. Apparently, one can increase the number of elements in the package to monitor more gases with different spectral response. Figure 1c illustrates the operating principle of one narrowband detection element. The gold nano- disk antennas serve to resonantly absorb the mid-IR radiation in a narrowband fashion and convert absorbed optical energy into heat26, which elevates the temperature of the LT substrate. The resulting temperature increase ΔT in turn causes the LT layer to generate the pyroelectric readout current ΔIout27,28, which is then converted to readout voltage ΔVout by the readout electronics. Design of plasmonic metamaterial absorbers. To sense multiple target gases, the spectral absorption of the narrowband detectors structure and the pattern of the array30, or replacing the metallic antennas with dielectric antennas31, the line-widths of the nar- rowband detectors can be further reduced, which in turn, improves the sensor selectivity. Photoresponse of the narrowband detectors. In order to assess the spectral response of the narrowband detectors, we measured the IR absorption spectra (Fig. 4a black curve) and the wavelength-dependent voltage responses (Fig. 4a red curve) by using a frequency-tunable quantum cascade laser (QCL). When the radius of Au nanodisk is 0.94 μm and the periodicity is 3 μm, the detector has a narrowband absorption spectrum peaked at 5.52 μm with a full width at half maximum (FWHM) of 670 nm. Importantly, the wavelength-dependent voltage responses of the detectors reproduce the IR absorption spectra of the PMAs very well. Figure 4b shows the voltage response of the detector as a function of the modulation frequency. It is found that when the modulation frequency is 7 Hz, the output voltage of the detector drops to 70.7% (3 dB) relative to the output voltage at the fre- quency of 4 Hz. Therefore, the modulation frequency should be set below 7 Hz. Moreover, the inset is the dynamic response of the narrowband detector (5.52 μm) in the time-domain at the mod- ulation frequency of 5 Hz recorded by an oscilloscope (Tektronix, DPO2024B). The average measured voltage response at the modulation frequency of 5 Hz is 90 V W−1. Factors that cause the measured responsivity to be lower than the calculated respon- sivity include: (1) The heat conduction between the LT element and the printed circuit board containing the impedance matching circuit is more signiﬁcant than the calculated case. (2) The actual spot size of the optical beam arriving at the LT element can be smaller than the area size of the LT element (see Supplementary Note 5 for the theoretical calculation of the voltage response). Heat generation in the detector. Since the sensitivity of the NDIR sensor to the gases is strongly correlated to the mid-IR responsivity of each narrowband detector element, it is necessary to accurately model the photothermal and the temporal response of the detector element. We used the heat transfer module in COMSOL Multiphysics, a ﬁnite element method based solver, to calculate the steady-state distribution of the temperature increase ΔT in the PMA integrated LT detector caused by the dissipated electromagnetic energy. Design of plasmonic metamaterial absorbers. To sense multiple target gases, the spectral absorption of the narrowband detectors Design of plasmonic metamaterial absorbers. To sense multiple target gases, the spectral absorption of the narrowband detectors TURE COMMUNICATIONS \| (2020) 11:5245 \| https://doi.org/10.1038/s41467-020-19085-1 \| www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 A K 297 297 296 295 294 293 A’ b a Heat sink 297.5 296 295 294 293 297.0 296.5 Temperature (K) Temperature (K) 296.0 295.5 295.0 –1.2 0 5 10 15 Time (s) 20 0.1Hz T =294.75K T =0.66 s 25 30 –0.6 X (mm) ΔT = 4.17K 0.0 0.6 1.2 IR Source ~ F ΔS 1.5 mm 2.4 mm Heat insulation tp=75 μm B ∫∫ T =295.68K 1 Ad c d Fig. 3 The thermal analysis of the pyroelectric detector. a The schematic diagram of the suspended LT substrate used for calculating the steady-state temperature distribution and the dynamic temperature change in the time-domain. b The steady-state temperature distribution at the surface of the LT layer. c The steady-state temperature distribution along the cut line A-A’. The red dotted line indicates that the average temperature is 295.68 K. d The time-domain average temperature at the upper surface of the LT substrate assuming that the power is modulated by a square wave with the frequency of 0.1 Hz. Source data are provided as a Source Data ﬁle. A K 297 297 296 295 294 293 A’ b a Heat sink IR Source ~ F ΔS 1.5 mm 2.4 mm Heat insulation tp=75 μm B b IR Source 296 295 294 293 Temperature (K) 0 5 10 15 Time (s) 20 0.1Hz T =294.75K T =0.66 s 25 30 d 297.5 297.0 296.5 Temperature (K) 296.0 295.5 295.0 –1.2 –0.6 X (mm) ΔT = 4.17K 0.0 0.6 1.2 ∫∫ T =295.68K 1 Ad c d d c Fig. 3 The thermal analysis of the pyroelectric detector. a The schematic diagram of the suspended LT substrate used for calculating the steady-state temperature distribution and the dynamic temperature change in the time-domain. b The steady-state temperature distribution at the surface of the LT layer. c The steady-state temperature distribution along the cut line A-A’. The red dotted line indicates that the average temperature is 295.68 K. d The time-domain average temperature at the upper surface of the LT substrate assuming that the power is modulated by a square wave with the frequency of 0.1 Hz. Source data are provided as a Source Data ﬁle. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 a b d c 1.0 0.8 lR absorption spectrum Measurement noise Johnson noise Temperature noise Narrowband detector voltage response Absorption Voltage noise (VHz-1/2) NEP (W Hz-1/2) Vout(V) Vout(V) Vout(V) 0.6 0.4 0.4 5.52 μm 0.45 0.40 0.35 0.30 0.25 0.20 0.0 0.2 0.4 Time (s) 0.6 0.8 0.40 0.32 0.24 0.16 0.08 0.00 0.3 0.2 0.1 0.0 0.2 0.0 10-5 8.0x10-8 6.0x10-8 4.0x10-8 2.0x10-8 10-6 10-7 10-8 10-9 3 4 5 5 10 Chopper frequency (Hz) 15 20 5 10 15 Frequency (Hz) 20 5 10 15 Frequency (Hz) 20 Wavelength (μm) 6 7 8 Fig. 4 Photoresponse of detector. a The measured voltage responses of one narrowband detector as a function of the wavelength of the input beam, compared with the absorption spectra of the integrated PMAs. The input beam is from a tunable quantum cascade laser and is modulated by an optical chopper at the frequency of 5 Hz. b The voltage response of the detector as a function of the chopping frequency. The inset is the time-domain voltage response of the narrowband detector with a peak wavelength of 5.52 µm. c The measured ﬂuctuation (total noise) in the output voltage of the detector without input optical power, and the calculated values of two major sources of noise: temperature noise and Johnson noise as a function of the frequency. d The noise equivalent power calculated from the measured total noise and the responsivity of the narrowband detector (5.52 µm). Source data are provided as a Source Data ﬁle. b Vout(V) Vout(V) 0.4 5.52 μm 0.45 0.40 0.35 0.30 0.25 0.20 0.0 0.2 0.4 Time (s) 0.6 0.8 0.3 0.2 0.1 0.0 5 10 Chopper frequency (Hz) 15 20 a 1.0 0.8 lR absorption spectrum Narrowband detector voltage response Absorption Vout(V) 0.6 0.4 0.40 0.32 0.24 0.16 0.08 0.00 0.2 0.0 3 4 5 Wavelength (μm) 6 7 8 b c Measurement noise Johnson noise Temperature noise Voltage noise (VHz-1/2) 10-5 10-6 10-7 10-8 10-9 5 10 15 Frequency (Hz) 20 d NEP (W Hz-1/2) 8.0x10-8 6.0x10-8 4.0x10-8 2.0x10-8 5 10 15 Frequency (Hz) 20 d c Fig. 4 Photoresponse of detector. a The measured voltage responses of one narrowband detector as a function of the wavelength of the input beam, compared with the absorption spectra of the integrated PMAs. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 The input beam is from a tunable quantum cascade laser and is modulated by an optical chopper at the frequency of 5 Hz. b The voltage response of the detector as a function of the chopping frequency. The inset is the time-domain voltage response of the narrowband detector with a peak wavelength of 5.52 µm. c The measured ﬂuctuation (total noise) in the output voltage of the detector without input optical power, and the calculated values of two major sources of noise: temperature noise and Johnson noise as a function of the frequency. d The noise equivalent power calculated from the measured total noise and the responsivity of the narrowband detector (5.52 µm). Source data are provided as a Source Data ﬁle. power. The two major sources of noise here are: (1) thermal ﬂuctuation noise ~uNT accounting for the random ﬂuctuations in temperature due to the statistical nature of the heat exchange between the suspended LT detector and the supporting pins on the circuit board; (2) Johnson noise ~uNR originating from the thermal agitation of the electrons inside the electrical conductors at equilibrium. The LT element is a capacitive structure with loss resistance, and its Johnson noise exhibits a frequency dependence due to the product of the loss resistance and the capacitance. Figure 4c plots the calculated thermal ﬂuctuation noise ~uNT and Johnson noise ~uNR (see Supplementary Note 5 for the theoretical calculation of the noises). The total voltage noise of the detector at different modulation frequency is also plotted in Fig. 4c. The measured noise level agrees well with the calculated value when the modulation frequency is below 20 Hz. As presented in Fig. 4d, At the modulation frequency of 5 Hz, the NEP is found to be 1.90 × 10−8 W Hz−1/2. From the comparison of the narrowband detectors with commercial LT detectors in Supplementary Note 8, it is seen that the performance of the narrowband detectors enabled by plasmonic metamaterial absorber developed in this work is comparable to the performance of the commercial LT detectors that use metal black coating as the IR absorber. Supplementary Note 9 for the arrangement of the components in the NDIR system. A collimated light beam from a SiC broadband infrared source is modulated by an optical chopper before it entered the gas cell. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 The light beam passing through the gas cell is then focused onto the sensing area of the packaged pyroelectric detector. The generated pyroelectric current is converted into the output voltage by the integrated circuit (IC) and the output voltage is then measured using the lock-in ampliﬁer. Figure 5b presents the results of the single-target gas sensing experiment for eight gases: H2S, CH4, CO2, CO, NO, CH2O, NO2, and SO2. Since the absolute value of the detector output voltage may vary among each experiment, we use the relative change in the output voltage ΔV/V0 to represent the voltage response of the detector. V0 is the initial voltage output of the detector when the chamber is ﬁlled with pure nitrogen gas. V is the voltage output of the detector when the target gas is mixed into the chamber and ΔV ≡V −V0 is the change in the output voltage caused by the target gas. To ﬁt the measured detector response ΔV/V0 as a function of the target gas concentration, we use a modiﬁed Beer–Lambert equation32–36: power. The two major sources of noise here are: (1) thermal ﬂuctuation noise ~uNT accounting for the random ﬂuctuations in temperature due to the statistical nature of the heat exchange between the suspended LT detector and the supporting pins on the circuit board; (2) Johnson noise ~uNR originating from the thermal agitation of the electrons inside the electrical conductors at equilibrium. The LT element is a capacitive structure with loss resistance, and its Johnson noise exhibits a frequency dependence due to the product of the loss resistance and the capacitance. Figure 4c plots the calculated thermal ﬂuctuation noise ~uNT and Johnson noise ~uNR (see Supplementary Note 5 for the theoretical calculation of the noises). The total voltage noise of the detector at different modulation frequency is also plotted in Fig. 4c. The measured noise level agrees well with the calculated value when the modulation frequency is below 20 Hz. As presented in Fig. 4d, At the modulation frequency of 5 Hz, the NEP is found to be 1.90 × 10−8 W Hz−1/2. From the comparison of the narrowband detectors with commercial LT detectors in Supplementary Note 8, it is seen that the performance of the narrowband detectors enabled by plasmonic metamaterial absorber developed in this work is comparable to the performance of the commercial LT detectors that use metal black coating as the IR absorber. Design of plasmonic metamaterial absorbers. To sense multiple target gases, the spectral absorption of the narrowband detectors Figure 3a shows the heat transfer model of the suspended LT substrate. The four corners of the LT sub- strate are suspended by silicon posts. The simulated steady-state distribution of the temperature across the upper surface of the LT layer is shown in Fig. 3b. We also plot the steady-state tem- perature distribution along the cut line A-A’ in Fig. 3c. The maximum steady-state temperature at the center point B is found to be Tcenter = 297.32 K. The average value of the steady-state temperature Taverage, deﬁned as the steady-state temperature distribution integrated over the upper surface of the LT substrate and divided by the area size of the upper surface of the LT substrate, is then calculated be 295.68 K, as shown by the red dotted line in Fig. 3c. We also plot the increase of the average temperature Taverage(t) in the time-domain in Fig. 3d. The ther- mal time constant τT is found to be 0.66 s. (The temperature change as a function of the modulation frequency of the optical chopper is shown in Supplementary Note 4). g p To determine the noise equivalent power (NEP), which is a direct measure of the smallest optical power that can be measured by the detector, we need to evaluate the noise in the detector, or the ﬂuctuation in the output voltage, without any input optical NATURE COMMUNICATIONS \| (2020) 11:5245 \| https://doi.org/10.1038/s41467-020-19085-1 \| www.nature.com/naturecommunications 4 ARTICLE NATURE COMMUNICATIONS \| (2020) 11:5245 \| https://doi.org/10.1038/s41467-020-19085-1 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 5 The NDIR experiment on single-target gas using the narrowband pyroelectric detectors. a The gas sensing system. b The voltage responses of the narrowband detectors as a function of the concentrations of the target gases. The purple squares stand for the measured data while the red dashed lines are the ﬁtting curves based on the Beer–Lambert law. Source data are provided as a Source Data ﬁle. the details about the calculation of k from the gas absorption lines obtained from the HITRAN database. l = 5 m represents the optical path length of the White type multipass cell; x is the gas concentration. The parameter c is added into the power term as a linearization coefﬁcient to account for the variations in the optical path length and light scattering for accurately ﬁtting the equation to the actual absorption data. In practice, the parameters span and c are ﬁtting parameters that are adjusted to match the ﬁtting curves to the measured data as close as possible. See Supple- mentary Table 7 in Supplementary Note 12 for the detection limit of eight target gases in single-target gas measurement. the details about the calculation of k from the gas absorption lines obtained from the HITRAN database. l = 5 m represents the optical path length of the White type multipass cell; x is the gas concentration. The parameter c is added into the power term as a linearization coefﬁcient to account for the variations in the optical path length and light scattering for accurately ﬁtting the equation to the actual absorption data. In practice, the parameters span and c are ﬁtting parameters that are adjusted to match the ﬁtting curves to the measured data as close as possible. See Supple- mentary Table 7 in Supplementary Note 12 for the detection limit of eight target gases in single-target gas measurement. D2 ¼ span21 ek21lxc21 1 1 þ span22 ek22lxc22 2 1 : ð4Þ ð4Þ We chose CO (gas 1) and SO2 (gas 2) as the two target gases for the mixed-gas experiment. The characteristic absorption wave- lengths of the two gases are: λ1 = 4.67 µm and λ2 = 7.35 µm, respectively. Correspondingly, the detection wavelengths of the two detectors are tuned to be 4.67 µm (detector I) and 7.35 µm (detector II), respectively. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 To determine the values of spanij, and cij, we ﬁrst performed four single-target gas measurements: (1) CO measured by detector I; (2) SO2 measured by detector I; (3) CO measured by detector II; (4) SO2 measured by detector II, as shown in Fig. 6a. By ﬁtting the measured detector responses, we can obtain the values of spanij, and cij. When the values of spanij, cij and kij are determined, the mathematical model of the “two- gas-two-detector” problem is established. We then performed four mixed-gas experiments to verify the mathematical model: (1) CO with varying concentration and SO2 with ﬁxed concentration measured by detector I; (2) SO2 with varying concentration and CO with ﬁxed concentration measured by detector I; (3) CO with varying concentration and SO2 with ﬁxed concentration mea- sured by detector II; (4) SO2 with varying concentration and CO with ﬁxed concentration measured by detector II, respectively. In each mixed-gas experiment, the ﬁxed concentration is chosen to be 7500 ppm. The measured responses of detector I and detector II are plotted in Fig. 6b using purple squares. We also plot Fig. 6b the calculated detector responses based on Eq. (3) and Eq. (4) using red dashed lines (see Supplementary Note 11 for the model parameters of the “two-gas-two-detector” problem). It is seen that the detector responses predicted by Eq. (3) and Eq. (4) agree well with the measurements, which conﬁrms the effectiveness of the mathematical model. See Supplementary Note 11 about using the mathematical model to work out the gas concentrations x1 and x2 from the detector responses D1 and D2. The NDIR experiment on mixed target gases. When there are multiple target gases in the gas cell, one can use multiple nar- rowband detectors to measure the gas mixture in the cell and back-calculate the concentration of each target gas based on the measured responses of the detectors. For example, assuming there are M target gases in the gas cell, and N narrowband detectors are used for measurement. The voltage response of the ith detector Di ≡ΔVi/V0i is related to the concentration of each target gas by: Di ¼ X M j¼1 spanij ekijlx cij j 1 : ð2Þ ð2Þ Here, i is the number of the detector and j is the number of the target gas. The parameter kij are calculated using Supplementary Equation (8) in Supplementary Note 10. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 Δv=v0 ¼ span eκlxc 1 : ð1Þ ð1Þ The modiﬁed version of the Beer–Lambert Law is required by the practical considerations in the NDIR implementation. The coef- ﬁcient span accounts for the fact that not all the IR radiation that impinges upon the detector is absorbed by the gas, even at high concentrations. The value of span ranges from 0 to 1 because of the optical ﬁlter bandwidth and the ﬁne structure of the absorption spectra. The coefﬁcient κ represents the effective absorption coefﬁcient of the gas. See Supplementary Note 10 for The NDIR experiment on single-target gas. We then con- structed an NDIR system to examine the performances of the fabricated narrowband detectors in gas sensing. As shown in Fig. 5a, the target gas is mixed with pure nitrogen gas and injected into the gas cell. The gas cell used in this work is a White type multipass cell with an effective optical length of 5 m. See 5 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 a In Out Target Gas Lock-in Amplifier Computer MFC2 MFC1 Pure N2 Gas Lens Detector Globar 0.0 -0.1 -0.2 -0.3 ΔV/V0 -0.4 Measurement data Fitting data -0.5 H2S CH4 CO2 CH2O NO2 SO2 CO NO -0.6 0 5 10 15 20 0 5 10 15 20 0 5 10 15 20 0 5 10 Gas concentration (103 ppm) 15 20 0 0.0 0.1 0.2 0.3 5 10 15 20 0 5 10 15 20 0 5 10 15 20 Chopper Power supply Gas cell b Fig. 5 The NDIR experiment on single-target gas using the narrowband pyroelectric detectors. a The gas sensing system. b The voltage responses of the narrowband detectors as a function of the concentrations of the target gases. The purple squares stand for the measured data while the red dashed lines are the ﬁtting curves based on the Beer–Lambert law. Source data are provided as a Source Data ﬁle. a In Out Target Gas Lock-in MFC2 MFC1 Pure N2 Gas Lens Detector Globar Chopper Power supply Gas cell a Detector 0.0 -0.1 -0.2 -0.3 ΔV/V0 -0.4 Measurement data Fitting data -0.5 H2S -0.6 0 5 10 15 20 b b CH4 0 5 10 15 20 CO2 0 5 10 15 20 NO ation (103 ppm) 0 5 10 15 20 CO 0 5 10 G t 15 20 Gas concentration (103 ppm) Fig. NATURE COMMUNICATIONS \| (2020) 11:5245 \| https://doi.org/10.1038/s41467-020-19085-1 \| www.nature.com/naturecommunications Discussion detectors based on LT have been massively produced using standard tools for making IC chips37. The nanoantenna array was also fabricated using regular fabrication processes in IC industry such as electron beam lithography, electron beam evaporation, and metal lift-off. In the next step, electron beam lithography can be replaced by more standard tools such as project litho- graphy (stepper) for wafer-scale mass production38. Although gold is not CMOS compatible, we can use materials such as aluminum39, or TiN40 that are CMOS compatible to make nanoantennas. Thus, our design is compatible with standard fabrication processes in IC industry and meet demands for low- cost and mass production. In summary, we have presented the design, fabrication and measurement of narrowband pyroelectric detectors that are enabled by directly integrating plasmonic metamaterial absorbers onto the sensing area of lithium tantalate elements as on-chip absorption ﬁlters. The detection wavelength, or the peak absorption wavelength of the integrated absorber can be tuned to cover the full mid-infrared spectrum by varying the design of the metamaterials. We fabricated eight narrowband detectors whose detection wavelengths are aligned with the characteristic absorption wavelength of eight target gases: H2S, CH4, CO2, CO, NO, CH2O, NO2, SO2, and implement the detectors in a home- built NDIR system to measure the gas concentrations. The mathematical model for a simple case in which two target gases (CO and SO2) in a mixture are measured by two narrowband detectors is also provided. The model can successfully predict the measured responses of the two detectors based on the con- centrations of the two target gases. We also build a computer program based on the mathematical model and demonstrate the deduction of the gas concentrations x1 and x2 from the detector responses D1 and D2. The presented work thus goes beyond the conventional NDIR gas sensor in multiplexed gas sensing by circumventing the need for multiple pairs of “bandpass ﬁlter + optical detector”. As the prospects for improving the limit of detection and make a ~cm long NDIR sensor while maintaining the same integration, the two key aspects we are working on are: (1) Reduce the thickness of the LT element from 75 µm to 700 nm (two orders of magnitude)20. This can be done by replacing the self-supported LT plate with thin ﬁlm LT on silicon. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 a 0.0 –0.1 –0.2 ΔV/V0 –0.3 –0.4 0 Detector l Detector l Detector l Fixed SO2,changed CO Detector ll Detector ll Fixed SO2,changed CO Fixed CO,changed SO2 Fixed CO,changed SO2 Measurement data Fitting data Measurement data Fitting data Changed CO Detector ll Changed CO Detector l Changed SO2 Detector ll Changed SO2 5 10 15 20 0 5 Gas concentration (103 ppm) 10 15 20 0 5 10 15 20 0 5 10 15 20 0 5 10 15 0 5 Gas concentration (103 ppm) 10 15 0 5 10 15 0 5 10 15 0.0 –0.1 –0.2 ΔV/V0 –0.3 –0.5 –0.4 b Fig. 6 The NDIR experiment on two target gases measured by two narrowband detectors. a The detector responses of four single-target gas measurements: CO measured by detector I; SO2 measured by detector I; CO measured by detector II; SO2 measured by detector II, respectively. b The detector responses of four mixed-gas experiments: ﬁxed SO2 concentration and varying CO concentration measured by detector I; ﬁxed CO concentration and varying SO2 concentration measured by detector I; ﬁxed SO2 concentration and varying CO concentration measured by detector II; ﬁxed CO concentration and varying SO2 concentration measured by detector II, respectively. In each mixed-gas experiment, the ﬁxed concentration is chosen to be 7500 ppm. The purple squares stand for the measured detector responses while the red dashed lines are the calculated detector responses based on the mathematical models provide by Eq. (3) and Eq. (4). Source data are provided as a Source Data ﬁle. a 0.0 –0.1 –0.2 ΔV/V0 –0.3 –0.4 0 Detector l Changed CO 5 10 15 20 Detector ll Changed SO2 0 5 10 15 20 a Measurement data Fitting data Detector ll Changed CO Detector l Changed SO2 0 5 Gas concentration (103 ppm) 10 15 20 0 5 10 15 20 Gas concentration (103 ppm) Detector l Fixed SO2,changed CO Measurement data Fitting data 0 5 10 15 0.0 –0.1 –0.2 ΔV/V0 –0.3 –0.5 –0.4 b Detector ll Fixed CO,changed SO2 0 5 10 15 b Detector l Fixed CO,changed SO2 0 5 Gas concen 10 15 Detector ll Fixed SO2,changed CO ation (103 ppm) 0 5 10 15 Gas concentration (103 ppm) Fig. 6 The NDIR experiment on two target gases measured by two narrowband detectors. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 a The detector responses of four single-target gas measurements: CO measured by detector I; SO2 measured by detector I; CO measured by detector II; SO2 measured by detector II, respectively. b The detector responses of four mixed-gas experiments: ﬁxed SO2 concentration and varying CO concentration measured by detector I; ﬁxed CO concentration and varying SO2 concentration measured by detector I; ﬁxed SO2 concentration and varying CO concentration measured by detector II; ﬁxed CO concentration and varying SO2 concentration measured by detector II, respectively. In each mixed-gas experiment, the ﬁxed concentration is chosen to be 7500 ppm. The purple squares stand for the measured detector responses while the red dashed lines are the calculated detector responses based on the mathematical models provide by Eq. (3) and Eq. (4). Source data are provided as a Source Data ﬁle. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 The parameters spanij and cij are ﬁtting parameters that account for the contribution of the jth target gas to the response Di of the ith detector. j g g p Taking a simple case where two target gases are measured by two narrowband detectors (“two-gas-two-detector”) as an example. The voltage responses of the two detectors are related to the concentrations of the two target gases by: D1 ¼ span11 ek11lxc11 1 1 þ span12 ek12lxc12 2 1 ; ð3Þ 6 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 NATURE COMMUNICATIONS \| (2020) 11:5245 \| https://doi.org/10.1038/s41467-020-19085-1 \| www.nature.com/naturecommunications References 1. Popa, D. & Udrea, F. Towards integrated mid-infrared gas sensors. Sensors 19, 2076 (2019). To evaluate the impact of the heat transfer between the sensing area and the environment via convection, we also use the heat transfer module in COMSOL to simulate the temperature change in the detector. The bottom sides of the four silicon posts are all set to be constant temperature T0 = 293.15 K while all other sides of the structure are set as heat insulation with the dimensions of the LT substrate are: length = 2.4 mm, width = 1.5 mm and thickness = 75 μm. The height and radius of the silicon posts are 250 and 50 μm, respectively. As for the MIM absorber, the top layer of nanodisk antennas is ignored for simplicity while the silicon dioxide spacer and gold backplate are included and set as the heat source with a Gaussian proﬁle: g(x, y) = g0 exp(−x2/r02) exp(−y2/r02). Here g0 = P0/(π r02 t) is the power density of the heat source; t = (tSiO2+ tbackplate) = 0.2 μm is the total thickness of the spacer and the backplate, the corresponding material parameters are in Supplementary Table 2. 2. Gordon, I. E. et al. The HITRAN2016 molecular spectroscopic database. J. Quant. Spectrosc. Radiat. Transf. 203, 3–69 (2017). 3. Mantsch, H. H. & Chapman, D. Infrared Spectroscopy of Biomolecules (Wiley- Liss, New York, 1996). 4. Barth, A. Infrared spectroscopy of proteins. Biochim. Biophys. Acta Bioenerg. 1767, 1073–1101 (2007). 5. Schliesser, A. & Picqué, N. Hänsch TWJNP. Mid-infrared frequency combs. Nat. Photon. 6, 440 (2012). 6. Heard, D. Analytical Techniques for Atmospheric Measurement (John Wiley & Sons, 2008). 7. Dinh, T.-V., Choi, I.-Y., Son, Y.-S. & Kim, J.-C. A review on non-dispersive infrared gas sensors: Improvement of sensor detection limit and interference correction. Sensor. Actuat. B Chem. 231, 529–538 (2016). Device fabrication. Mid-IR detectors were fabricated at the Center of Micro- Fabrication and Characterization (CMFC) of Wuhan National Research Center for Optoelectronics. A schematic of the fabrication and package of the narrowband detector is shown in Supplementary Fig. 3. The fabrication of the narrowband detectors begins with the deposition of the 80 nm silicon dioxide spacer on top of the gold electrode (100 nm Au/20 nm Cr) pre-deposited on the 75 µm LT substrate (Yamaju Ceramics Co., LTD.) using plasma-enhanced chemical vapor deposition (PECVD). ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 from the human operator. Finally, the measurement lab is well ventilated to minimize the remaining water vapor in the room. Thus in the future, to make sure that the ~cm long NDIR sensors can accurately measure gases in the 5–7 µm region in standard atmospheric conditions, water vapor need to be removed to a very high degree. optical chopper (Thorlabs, MC2000B, fmod ≥4 Hz) was used to mechanically modulate the output beam from the QCL before it reaches the narrowband detectors. The beam modulated by the chopper was then focused by a reﬂective objective (Thorlabs, LMM-15X-P01) to the area of antenna array. A power supply (GWINSTEK, GPS-3303C) that provides +5 V bias voltage for the impedance matching circuit in the TO-package and a lock-in ampliﬁer (Stanford Research System, SRS-830) connected to a computer controlled by LabVIEW is used to measure the output electrical signal of the detector. The input reference signal of the lock-in ampliﬁer is same as the chopper input signal. optical chopper (Thorlabs, MC2000B, fmod ≥4 Hz) was used to mechanically modulate the output beam from the QCL before it reaches the narrowband detectors. The beam modulated by the chopper was then focused by a reﬂective objective (Thorlabs, LMM-15X-P01) to the area of antenna array. A power supply (GWINSTEK, GPS-3303C) that provides +5 V bias voltage for the impedance matching circuit in the TO-package and a lock-in ampliﬁer (Stanford Research System, SRS-830) connected to a computer controlled by LabVIEW is used to measure the output electrical signal of the detector. The input reference signal of the lock-in ampliﬁer is same as the chopper input signal. Owing to the high insertion loss caused by the gas cell, the current NDIR system is not implemented with a reference channel. In future designs of ~cm long NDIR sensors, one of the detection elements in the 5–7 μm range can be made as a refer- ence detector. Measuring errors caused by dust or diminishing radiation intensity are removed by the use of the reference channel. We also build a gas sensing system. A schematic of the gas sensing system is shown in Fig. 5a. Data availability The data that support the ﬁndings of this study are available from the authors on reasonable request; Source data are provided with this paper. ARTICLE The gas sensing system is composed of three subsystems: the optical subsystem contains a collimated SiC broadband IR source (Thorlabs, SLS203L/M, see Supplementary Note 6 for details), an optical chopper, a gas chamber (GAINWAY, GW-1020IR-5M), a reﬂective objective (Thorlabs, LMM- 15X-P01) and the prepared narrowband pyroelectric detector; the electrical subsystem contains a power supply (GWINSTEK, GPS-3303C) that provides +5 V bias voltage for the detector and a lock-in ampliﬁer (SRS-830); the gas supply subsystem contains mass ﬂow controllers (Sevenstar, CS200C) that control real- time ﬂow of each gas. The target gas gets mixed with pure nitrogen gas and sent into the gas chamber. In the future, the demonstrated NDIR sensor architecture can be expanded by increasing the number of narrowband detection elements to analyze more target gases (see Supplementary Note 13 for methods of minimizing the thermal cross-talk between the neighboring elements). It can also be extended to other thermal detector platforms such as thermopile detectors and vanadium oxide microbolometers. When combined with the large size focal plane array technology, the presented device architecture will evolve as on-chip infrared spectrometers, which can serve as tools for spectroscopic analysis of gases, chemicals, explosives, and other types of substances. References Electron beam resist (AR-P 6200.09) was spin-coated and nanodisk arrays were deﬁned by electron beam lithography (Vistec EBPG 5000plus ES) followed by electron beam evaporation(EB-500S) to deposit 50 nm Au/10 nm Ni (adhesive layer). Another round of electron beam lithography was then performed to deﬁne the 600 µm side-length window areas for wire-bonding. The silicon dioxide spacer in the wire-bonding areas was then removed by dry etching (Plasmalab system 100 ICP 180) to expose the gold electrode underneath the spacer. The LT substrate was then cut into separate single-element detectors using a laser cutter, each of which has an antenna array and a wire-bonding area. The detector was then mounted onto a TO-5 packaged impedance matching circuit with a customized special JFET (low drift, low noise) in voltage mode. The bottom electrode of the LT detector is in direct contact with the one input pin of the circuit. The top electrode of the LT detector was then wire-bonded to the other input pin to ﬁnish the electrical connection. The optical window is an infrared high trans- missive CaF2 glass. 8. Hodgkinson, J. & Ralph, P. T. Optical gas sensing: a review. Meas. Sci. Technol. 24, 012004 (2013). 9. Hodgkinson, J., Smith, R., Ho, W. O., Saffell, J. R. & Tatam, R. P. Non-dispersive infra-red (NDIR) measurement of carbon dioxide at 4.2 μm in a compact and optically efﬁcient sensor. Sensor. Actuat. B Chem. 186, 580–588 (2013). optically efﬁcient sensor. Sensor. Actuat. B Chem. 186, 580–588 (20 10. Vincent, T. A. & Gardner, J. W. A low cost MEMS based NDIR system for the monitoring of carbon dioxide in breath analysis at ppm levels. Sensor. Actuat. B Chem. 236, 954–964 (2016). 11. Vincent, T. A., Urasinska-Wojcik, B. & Gardner, J. W. Development of a low- cost NDIR system for ppm detection of carbon dioxide in exhaled breath analysis. Procedia Eng. 120, 388–391 (2015). 12. Mendes, L. et al. NDIR gas sensor for spatial monitoring of carbon dioxide concentrations in naturally ventilated livestock buildings. Sensors 15, 11239–11257 (2015). 13. Kwon, J., Ahn, G., Kim, G., Kim, J.-C., Kim, H. A study on NDIR-based CO2 sensor to apply remote air quality monitoring system. ICCAS-SICE 2009- ICROS-SICE International Joint Conference 2009, Proceedings (IEEE, 2009). 14. Hodgkinson, J. & Tatam, R. P. Optical gas sensing: a review. Meas. Sci. Technol. 24, 012004 (2013). 15. Tan, Q. L. et al. Code availability The code that support the ﬁndings of this study is available from the authors on reasonable request; Received: 28 September 2019; Accepted: 24 September 2020; Methods Fi it l Finite element simulations. We used COMSOL, a ﬁnite element method-based solver to numerically study the optical properties of the absorbers employing periodic boundary conditions. With plane wave excitation polarized along the y- axis in the electromagnetic waves module. The domain boundaries parallel to the x−z and y−z planes and port1 is the top face of simulation area, port2 is the bottom face of simulation area. Using the corresponding parameter in Supplementary Table 1, we can obtain the optimized absorption curve. Discussion (2) Increase the quality factor of the MIM absorber that determines the overlap between the spectral response of the detector and the gas absorption lines. This can be done by optimizing the design of the nanoantennas in the MIM absorbers30. Since water vapor (humidity) is a strong source of cross- response in the 5–7 μm, we took several measures to minimize the intereference from the water vapor. First, the NDIR system is installed with a pump that can take out the air and water vapor inside the gas pipelines. Second, we also put desiccants at the outlets of the pipelines to remove water vapor from the ambient. Third, the NDIR system is located on an optical table that is covered by a hood. The hood can help isolate the water vapor Since pyroelectric materials have long been used to construct low-cost and uncooled detectors throughout the infrared spec- trum, the demonstrated narrowband detectors ﬁnd immediate commercial applications such as the detection of ﬁre ﬂame and human bodies based on their infrared absorption features and non-contact temperature measurement. In particular, pyroelectric 7 Author contributions X.C.T. and F.Y conceived the idea. X.C.T. and J.Y.L. designed and simulated the prototype structures. X.C.T. fabricated the detector and built up the test system. X.C.T. and H.Z. performed the photoresponse and gas sensing measurement. X.C.T., H.Z., and H.W.W. performed the electrical calculation, data analysis. Q.S.G. helped with data analysis. H.B.Z. provided the Lithium Tantalate materials. F.Y. organized the project, analyzed the results, and provided the support. All authors contributed to the preparation of the manuscript. 24. Tan, X., Li, J., Yang, A., Liu, H., Yi, F. Narrowband plasmonic metamaterial absorber integrated pyroelectric detectors towards infrared gas sensing. 2018 Conference on Lasers and Electro-Optics (CLEO)) (San Jose, California United States, 2018). 25. Aggarwal, M. et al. Pyroelectric Materials for Uncooled Infrared Detectors: Processing, Properties, and Applications (National Aeronautics and Space Administration, 2010). ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-19085-1 Acknowledgements 18. Dong, M. et al Development and measurements of a mid-infrared multi-gas sensor system for CO, CO2 and CH4 detection. Sensors 17, 2221 (2017). F.Y. acknowledges funding support from National Natural Science Foundation of China (NSFC) (11774112, 11604110); National key research and development program of China (2016YFC0201300, 2019YFB2005700); The Fundamental Research Initiative Funds for Huazhong University of Science and Technology (2017KFYXJJ031, 2019kfyRCPY122); Graduates’ Innovation Fund, Huazhong University of Science and Technology (5003182041). We thank Li Pan engineer in the Center of Micro-Fabrication and Characterization (CMFC) of WNLO for the support in PECVD fabrication. We thank Zeng Tiantian engineer in the Huazhong University of Science & Technology Analytical & Testing Center for the support in FTIR test. We thank the technical support from Experiment Center for Advanced Manufacturing and Technology in School of Mechanical Science & Engineering of HUST. 19. Yi, F., Zhu, H., Reed, J. C. & Cubukcu, E. Plasmonically enhanced thermomechanical detection of infrared radiation. Nano Lett. 13, 1638–1643 (2013). 20. Suen, J. Y. et al. Multifunctional metamaterial pyroelectric infrared detectors. Optica 4, 276–279 (2017). 21. Montoya, J. A., Tian, Z.-B., Krishna, S. & Padilla, W. J. Ultra-thin infrared metamaterial detector for multicolor imaging applications. Opt. Exp. 25, 23343–23355 (2017). 22. Ogawa, S., Komoda, J., Masuda, K. & Kimata, M. Wavelength selective wideband uncooled infrared sensor using a two-dimensional plasmonic absorber. Optice 52, 127104–127104 (2013). 23. Yi, F., Zhu, H., Reed, J. C., Zhu, A. Y. & Cubukcu, E. Thermoplasmonic membrane-based infrared detector. IEEE Photon. Technol. Lett. 26, 202–205 (2014). The authors declare no competing interests. 27. Batra, A. K. & Aggarwal, M. D. Pyroelectric Materials: Infrared Detectors, Particle Accelerators and Energy Harvesters. (SPIE Press, Bellingham, Washington USA, 2013). Additional information Supplementary information is available for this paper at https://doi.org/10.1038/s41467- 020-19085-1. Supplementary information is available for this paper at https://doi.org/10.1038/s41467- 020-19085-1. 28. Chatard, J.-P., Norkus, V. & Dennis, P. N. J. Pyroelectric infrared detectors based on lithium tantalate: state of art and prospects. Proc. SPIE 5251, Detectors and Associated Signal Processing, 121–128 (2004). Correspondence and requests for materials should be addressed to F.Y. 29. Baffou, G., Girard, C. & Quidant, R. Mapping heat origin in plasmonic structures. Phys. Rev. Lett. 104, 136805 (2010). Peer review information Nature Communications thanks Daniel Popa and the other, anonymous, reviewers for their contribution to the peer review of this work. Peer reviewer reports are available. Peer review information Nature Communications thanks Daniel Popa and the other, anonymous, reviewers for their contribution to the peer review of this work. Peer reviewer reports are available. 30. Kang, S. et al. Ultra-narrowband metamaterial absorbers for high spectral resolution infrared spectroscopy. Adv. Opt. Mater. 7, 1801236 (2019). 31. Ren, Z., Sun, Y., Lin, Z. & Wang, C. Ultra-narrow band perfect metamaterial absorber based on dielectric-metal periodic conﬁguration. Opt. Mater. 89, 308–315 (2019). Competing interests 26. Qian, Z. et al. Zero-power infrared digitizers based on plasmonically enhanced micromechanical photoswitches. Nat. Nanotechnol. 12, 969–973 (2017). References Three-gas detection system with IR optical sensor based on NDIR technology. Opt. Laser Eng. 74, 103–108 (2015). 16. Rubio, R. et al. Non-selective NDIR array for gas detection. Sensor. Actuat. B- Chem. 127, 69–73 (2007). Measurement of photoresponse. We use a wavelength tunable quantum cascade laser (QCL) to examine the characteristic of the fabricated detectors. The working wavelength of the QCL (Block Engineering, LaserTune) can be continuously tuned from 5.4 to 6.0 μm. See Supplementary Note 6 for the details about the QCL. An 17. Fonollosa, J. et al. Limits to the integration of ﬁlters and lenses on thermoelectric IR detectors by ﬂip-chip techniques. Sensor. Actuat. A-Ph 149, 65–73 (2009). NATURE COMMUNICATIONS \| (202 NATURE COMMUNICATIONS \| (2020) 11:5245 \| https://doi.org/10.1038/s41467-020-19085-1 \| www.nature.com/naturecommunications 8 © The Author(s) 2020 Reprints and permission information is available at http://www.nature.com/reprints Reprints and permission information is available at http://www.nature.com/reprints Reprints and permission information is available at http://www.nature.com/reprints Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. 32. Lee, R. & Kester, W. Complete gas sensor circuit using nondispersive infrared (NDIR). Anal. Dialog. 50, 10–18 (2016). 33. A Background to Gas Sensing by Nondispersive Infrared. A1A-Infrared_AN1, 6, 1–9 (SGX Sensortech (IS) Ltd, 2007). 34. NDIR: Gas Concentration Calculation Overview, Alphasense Application Note, 201–06. 1–5 (Alphasense Limited, 2014). Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, 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. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. 35. NDIR: Determination of Linearisation and Temperature Correction Coefﬁcients, Alphasense Application Note, 203–04. 1–4 (Alphasense Limited, 2009). 36. NDIR: Origin of Nonlinearity and SPAN, Alphasense Application Note, 204–02. 1–2 (Alphasense Limited, 2009). 37. Norkus, V. Pyroelectric Infrared Detectors Based on Lithium Tantalate: State of Art and Prospects (SPIE, 2004). p 38. Park, J.-S. et al. All-glass, large metalens at visible wavelength using deep- ultraviolet projection lithography. Nano Lett. 19, 8673–8682 (2019). 39. Knight, M. W. et al. Aluminum for plasmonics. ACS Nano 8, 834–840 (2014). 40. Li, W. et al. Refractory plasmonics with titanium nitride: broadband metamaterial absorber. Adv. Mater. 26, 7959–7965 (2014). NATURE COMMUNICATIONS \| (2020) 11:5245 \| https://doi.org/10.1038/s41467-020-19085-1 \| www.nature.com/naturecommunications 9
https://openalex.org/W2163191136	https://europepmc.org/articles/pmc3847847?pdf=render	English	null	Comparison of the effects of ibuprofen and acetaminophen on PGE2 levels in the GCF during orthodontic tooth movement: a human study	Progress in orthodontics	2,013	cc-by	4,144	RESEARCH Open Access © 2013 Shetty et al.; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Background: Pain is among the most cited negative effects of orthodontic treatment. Non-steroidal anti-inflammatory drugs seem to be an effective option for minimizing this but can have adverse effects on tooth movement owing to their ability to block prostaglandin synthesis. Acetaminophen has been suggested as the analgesic of choice during orthodontic treatment as it showed no effect on orthodontic tooth movement in previous animal studies. The purpose of this study was to compare the effects of ibuprofen and acetaminophen on the prostaglandin E2 (PGE2) levels of the gingival crevicular fluid (GCF) during orthodontic tooth movement in human subjects. Methods: A total of 42 patients (mean age 18 ± 4.5 years) were randomly divided into three equal groups: ibuprofen, acetaminophen, and control groups. Maxillary canines were distalized with 150 g of force delivered by NiTi coil springs. GCF samples were obtained before (baseline) and after spring activation at 24, 48, and 168 h. The PGE2 content of the GCF was determined using enzyme-linked immunosorbent assay. Results: PGE2 levels in all groups increased significantly by 24 and 48 h of force application and decreased to baseline levels by 168 h. No significant difference was found between the acetaminophen and control groups at any time point. There was a significant decrease in PGE2 levels in the ibuprofen group at 24 and 48 h when compared to the other two groups. Conclusions: Acetaminophen showed no significant effect on prostaglandin synthesis and may be the safe choice compared to ibuprofen for relieving pain associated with orthodontic tooth movement. Comparison of the effects of ibuprofen and acetaminophen on PGE2 levels in the GCF during orthodontic tooth movement: a human study Niveditha Shetty1, Anand K Patil2, Sanjay V Ganeshkar2 and Srinidhi Hegde3 Shetty et al. Progress in Orthodontics 2013, 14:6 http://www.progressinorthodontics.com/content/14/1/6 Shetty et al. Progress in Orthodontics 2013, 14:6 http://www.progressinorthodontics.com/content/14/1/6 Correspondence: dr.nivedithashetty@gmail.com 1Department of Orthodontics, A.J. Institute of Dental Sciences, Mangalore, Karnataka 575004, India Full list of author information is available at the end of the article Background tooth movement in both rats and humans [3-8]. A reflection of the inflammatory process during orthodontic tooth movement can be observed as elevated concentra- tions of chemical mediators in the gingival crevicular fluid (GCF) of the moving teeth. Nowadays, this is commonly used as a biomarker assay to assess the level of various chemical mediators such as prostaglandins and indicators of root resorption such as dentin degradation products [9]. Orthodontic tooth movement is known to cause inflam- matory reactions in the periodontium and dental pulp, which stimulate release of various biochemical media- tors. This is often associated with painful reactions, which have been rated as the greatest dislike during and fourth among major fears prior to orthodontic treatment [1]. Prostaglandins, particularly prostaglandin E2 (PGE2), have been implicated as one of the main mediators of this inflammatory reaction which increase the vascular dilatation and permeability and induce bone resorption through osteoclastic cell activation [2]. Evidence indi- cates that the local application of prostaglandins in the form of subperiosteal injections resulted in increased Non-steroidal anti-inflammatory drugs (NSAIDs), which inhibit cyclooxygenase activity thereby affecting the synthesis of prostaglandins, remain the most preferred method for pain control during orthodontics [10,11]. Acetaminophen, a non-steroidal anti-inflammatory drug in the family of para-aminophenols, said to have a central analgesic effect, showed no effect on orthodontic tooth movement in previous animal studies [12-14]. However, human studies in this regard, which evaluate the direct ef- fect of acetaminophen on the production of prostaglandins * Correspondence: dr.nivedithashetty@gmail.com 1Department of Orthodontics, A.J. Institute of Dental Sciences, Mangalore, Karnataka 575004, India Full list of author information is available at the end of the article * Correspondence: dr.nivedithashetty@gmail.com 1Department of Orthodontics, A.J. Institute of Dental Sciences, Mangalore, Karnataka 575004, India Full list of author information is available at the end of the article Shetty et al. Progress in Orthodontics 2013, 14:6 http://www.progressinorthodontics.com/content/14/1/6 Page 2 of 5 micropipettes by capillary action (Figure 1). The area was isolated using cotton rolls to prevent saliva contamin- ation, and GCF was collected by placing the microcapillary pipettes at the entrance of the gingival sulcus, gently touching the marginal gingiva. From each test site, a standardized volume of 2 μl was col- lected using the calibration on white color-coded 1 to 5 μl calibrated volumetric microcapillary pipettes (Sigma- Aldrich Chemical Company, St. Louis, MO, USA) using an extracrevicular approach (‘unstimulated’). Methods The sample and standard solutions were first added to the antibody precoated microplate. Next, the diluted enzyme conjugate was added, and the mixture was shaken and incubated at room temperature for 1 h. The plate was then washed removing all the unbound material. The bound enzyme conjugate was detected by the addition of substrate which generated an optimal color after 30 min. Hydrochloric acid (1 N) was added to each well to stop the enzyme reaction. The plate was read using a 450-nm microplate reader. Quantification of PGE2 in the sam- ples was achieved by comparison with a standard curve generated from known quantities of PGE2 (standards) that had gone through the assay. Study subjects consisted of 42 patients (mean age 18 ± 4.5 years) seeking orthodontic treatment, in whom bilateral maxillary first premolar extraction was planned. They were randomly divided into three equal groups of 14 subjects each. The first group was prescribed with ibupro- fen, the second with acetaminophen, and the third group was taken as the control group, without administration of any drug. All patients were checked for periodontal status, and those with a history of systemic diseases, gastric disor- ders, or history of intake of any medication within the past 6 months were excluded from the study. The research was carried out in compliance with the Declaration of Helsinki after obtaining ethical approval from the Institutional Review Board, SDM College of Dental Sciences and Hospital, Dharwad, India. Background Each sample collection was allotted a maximum of 30 min, and some test sites that did not express any volume of GCF within the allotted time were excluded from the study. The sam- ples were diluted in phosphate buffer solution and stored at −20°C. Once all the samples were obtained, immuno- assay for PGE2 was performed. The quantitative PGE2 content of the crevicular fluid was determined using the Neogen prostaglandin E2 enzyme-linked immunosorbent assay kit (Neogen Corporation, Lexington, KY, USA), according to the manufacturer's instructions. and thereby on orthodontic tooth movement, are lacking. It is debatable whether findings from animal experiments can be extrapolated to the human situation as morphological and physiological differences between animal and human alveolar bone and periodontal ligament have to be consid- ered [15]. Moreover, acetaminophen has been shown to either inhibit or stimulate prostaglandin synthesis, de- pending on the tissue, preparation of the tissue, and constituents of the incubation milieu [12,16,17]. It is also known to reduce the levels of prostacyclins after systemic administration in humans [18]. The aim of the present study was to compare the effects of ibuprofen and acetaminophen on the PGE2 levels in the GCF during orthodontic tooth movement in human subjects. By studying the alterations in the levels of these mediators, the possible effects of these drugs on the biologic processes mediating orthodontic tooth movement in humans may be evaluated. Statistical analysis Friedman test was performed to determine if statistically significant differences were present between the four Fixed orthodontic therapy was started on all patients after obtaining informed consent. All the patients were on strict mechanical oral hygiene regimen. After initial leveling and aligning, the maxillary canines were retracted on a 0.018-in. stainless steel wire with 150 g of force delivered by nickel titanium tension springs (Orthoforce G4-Nickel Titanium, G&H Wire Company, Hanover, Germany) placed between the maxillary molars and ca- nines. The subjects received a Nance button for anchorage control. At the appliance activation, the subjects in the first group were given ibuprofen, 400 mg three times daily for 2 days. The second group received acetaminophen, 500 mg three times a day for 2 days. The third group did not receive any analgesics. Figure 1 GCF collection using micropipettes at baseline (T0). GCF sampling was done before the placement of the closed coil springs (baseline T0) and after the activation of the springs at 24 (T1), 48 (T2), and 168 (T3) h. Each sample of GCF was collected from the gingival crevicular sulcus of the maxillary canine using calibrated Figure 1 GCF collection using micropipettes at baseline (T0). Shetty et al. Progress in Orthodontics 2013, 14:6 http://www.progressinorthodontics.com/content/14/1/6 Shetty et al. Progress in Orthodontics 2013, 14:6 http://www.progressinorthodontics.com/content/14/1/6 Page 3 of 5 tissues with osteoblastic and osteoclastic remodeling. Reported patient discomfort as well as pain is generally at its highest during the first 24 h after the application of an orthodontic force. The periodicity of these complaints peaks at 24 h and decreases to baseline levels by 7 days [19,20]. NSAID usage is the most routinely used pain management method during orthodontic therapy. different time points in the three groups. Bonferroni correc- tion was applied after which within-group differences of the PGE2 levels in the gingival crevicular fluid between 0 and 168 h were evaluated by the Wilcoxon matched-pairs test. Differences of the PGE2 levels at different time periods between the three groups were determined by the Kruskal- Wallis test. Pairwise comparisons of the groups were done using the Mann-Whitney U test. The present study analyzed with the use of GCF the effects of two popularly used analgesics, ibuprofen and acetaminophen, on PGE2 levels. Statistical analysis While high doses of NSAIDs have been reported to disrupt tooth movement in previous animal studies [12,21,22], the possible effects of commonly used analgesics in over-the-counter doses on the biologic processes underlying tooth movement have not been evaluated. We have tried to evaluate the effects of ibuprofen and acetaminophen in over-the-counter doses in our study. Intergroup differences The PGE2 levels of the ibuprofen group at 24 h were significantly different when compared to the acetaminophen and control groups (p = 0.0006 and p = 0.002, respectively, in Table 1). At 48 h, the PGE2 levels in the ibuprofen group showed statistically significant differences when compared to the acetaminophen and control groups (p = 0.011, Table 1). No significant differences in PGE2 levels were found between the acetaminophen and control groups at any time measured (Table 1). No significant difference was found between the three groups at baseline and 168 h (Table 1). Intragroup differences The output of the Friedman test showed the presence of statistically significant differences (p < 0.05) between the four time points in all the three groups. Prostaglandin E2 levels in the GCF significantly increased in all groups by 24 h when compared to baseline values (p = 0.001). The PGE2 levels at 48 h in the three groups were also signifi- cantly high compared with baseline values (p = 0.001). Prostaglandin E2 levels at 168 h were not significantly different from baseline values (Table 1). Prostaglandin E2 levels of GCF decreased significantly in all the groups between 24 and 48 h, between 24 and 168 h, and between 48 and 168 h (p = 0.001, Table 1). Quantitative evaluation of PGE2 from GCF samples of human subjects in our study showed that the PGE2 levels in all the experimental groups increased signifi- cantly by 24 h and maintained the elevated levels until 48 h of orthodontic force application when compared with baseline measurements (p = 0.001). We observed a statistically significant decrease in the level of inflammatory mediators as time progressed which is evident in the peri- odical evaluation at 48 and 168 h. The gradual suppression of the inflammatory reaction and the decay in orthodontic force most likely account for this decrease in levels of PGE2. Our findings are in concordance with the studies of Grieve et al. [23] and Sari et al. [2] who also quantified GCF PGE levels after 24 and 48 h of appliance activation and found significant elevations when compared with the baseline values. Lee et al. [24] also found an increase in PGE2 levels in the GCF after 24 h of orthodontic force application. The clinically undetectable gingival inflamma- tion usually caused during fixed orthodontic appliance therapy [25,26] might have contributed to the increase in PGE2 levels at 24 and 48 h of appliance activation. T0 = baseline, T1 = 24 h, T2 = 48 h, T3 = 168 h. I, ibuprofen group; A, acetaminophen group; C, control group. Discussion In our study, the control group received no pharmaco- logical agent; therefore, it is assumed that the mean The early phase of orthodontic tooth movement is char- acterized by inflammatory responses of the periodontal Table 1 Intragroup and intergroup comparisons of PGE2 levels (ng/mL) T0 T1 T2 T3 PGE2 levels (p) T0 vs. T1 T1 vs. T2 T2 vs. T3 T0 vs. T2 T0 vs. T3 T1 vs. T3 Intragroup Group I 31.90 ± 8.88 47.94 ± 10.28 36.99 ± 8.80 32.35 ± 8.96 0.001 0.001 0.001 0.001 0.463 0.001 Group A 32.85 ± 9.09 63.00 ± 9.08 47.70 ± 9.24 33.01 ± 10.42 0.001 0.001 0.001 0.001 0.875 0.001 Group C 34.15 ± 13.59 67.15 ± 16.79 51.60 ± 16.65 35.04 ± 13.83 0.001 0.001 0.001 0.001 0.753 0.001 Intergroup (p) I vs.A 0.963 0.0006 0.011 0.490 I vs.C 0.963 0.002 0.011 0.679 A vs.C 0.747 0.421 0.520 0.963 T0 = baseline, T1 = 24 h, T2 = 48 h, T3 = 168 h. I, ibuprofen group; A, acetaminophen group; C, control group. able 1 Intragroup and intergroup comparisons of PGE2 levels (ng/mL) Shetty et al. Progress in Orthodontics 2013, 14:6 http://www.progressinorthodontics.com/content/14/1/6 Page 4 of 5 Page 4 of 5 concentrations of PGE2 released in the GCF were unaltered and purely the result of orthodontic force application. The samples of the acetaminophen group showed some decrease in the PGE2 levels when compared to the control group, but the difference was not statisti- cally significant. No statistically significant difference was observed between the acetaminophen and control groups at any time point, indicating that PGE2 levels were not affected significantly by acetaminophen administration. This finding supports the theory that acetaminophen fails to exhibit peripheral anti-inflammatory activity, i.e., it does not block prostaglandins peripherally because it does not concentrate in areas of inflammation where the peroxide level is high [18,27,28]. The explanation for this could also be that NSAIDs block COX-1 and/or COX-2, whereas paracetamol blocks a third isoform, COX-3, which is expressed only in the brain and spinal cord and therefore has minimal effects on prostaglandin synthesis [29-32]. similar to those of naproxen sodium and aspirin [20]. Therefore, acetaminophen maybe deemed to be as effective as ibuprofen, naproxen sodium, or aspirin in alleviating pain following appliance placement. The following conclusions were drawn from this study: The following conclusions were drawn from this study: 1. The PGE2 levels of the three groups peaked at 24 h and decreased nearly to baseline levels by 168 h. Therefore, the pain suppressant drugs prescribed in this period could adversely affect the PGE2 synthesis and hence the rate of orthodontic tooth movement. 2. Ibuprofen inhibited PGE2 synthesis significantly more than acetaminophen and when compared to the control group. 3. Acetaminophen was not found to affect PGE2 levels significantly during the experimental period. We found a statistically significant decrease in PGE2 levels in the ibuprofen group at 24 h (p = 0.002) and 48 h (p = 0.011) when compared to the control group. There was also a highly significant difference when comparing the mean concentrations of PGE2 between the two drug groups at 24 h (p = 0.006) and 48 h (p = 0.011). This indicated that ibuprofen inhibits PG synthesis more than acetaminophen during the first and second days of orthodontic tooth movement. This result was similar to that of Kehoe et al. [13], who found a significant decrease in the mean concentration of PGE2 levels in the PDL exudates of guinea pigs in whom ibuprofen had been administered. This implies that by inhibiting prosta- glandins, ibuprofen may have an effect on the rate of orthodontic tooth movement [16,21,22,33]. The highly significant decrease in PGE2 levels in the ibuprofen group when compared to the other two groups might be attributed to its anti-inflammatory action on peripheral inflamed tissues. The results of this study suggest that NSAIDs like ibuprofen have an inhibitory effect on the release of prostaglandins during initial tooth movement and thereby may cause an impediment in the rate of tooth movement. Acetaminophen may be suggested as the drug of choice for safe and effective management of orthodontic pain. Authors’ contributions h k d h The work presented here was carried out in collaboration among all authors. NS and AKP defined the research theme and designed the methods and experiments. NS carried out the GCF collection and laboratory assay, analyzed the data, interpreted the results, and drafted the manuscript. SH worked on the data interpretation. SH and SVG discussed the analyses, interpretation, and presentation of data. All authors read and approved the final manuscript. Author details 1 In our study, we have evaluated the effects of two analgesics, ibuprofen and acetaminophen, on the PGE2 levels in the GCF during orthodontic tooth movement. Although it is implied that NSAIDs like ibuprofen may impede tooth movement by inhibiting PG synthesis, the rate of tooth movement in the groups has not been measured. Long-term comparison of the rate of tooth movement in patients with administration of these drugs would yield definitive results. No evaluation of the efficacy of the two drugs prescribed in relieving pain associated with appliance activation was done in our study. However, comparison of the efficacy of ibuprofen and acetaminophen in controlling pain after orthodontic tooth movement in a previous study [34] has shown that both are equally effective in reducing discomfort after initial orthodontic appliance placement. In another study, acetaminophen group showed visual analogue scale results 1Department of Orthodontics, A.J. Institute of Dental Sciences, Mangalore, Karnataka 575004, India. 2Department of Orthodontics and Dentofacial Orthopedics, S.D.M College of Dental Sciences and Hospital, Dharwad, Karnataka 580009, India. 3Department of Conservative Dentistry and Endodontics, A.J. Institute of Dental Sciences, Mangalore, Karnataka 575004, India. Received: 17 April 2013 Accepted: 17 April 2013 Published: 17 May 2013 Competing interests Th h d l h Competing interests The authors declare that they have no competing interests. Shetty et al. Progress in Orthodontics 2013, 14:6 http://www.progressinorthodontics.com/content/14/1/6 Aspirin, acetaminophen, and ibuprofen: their effects on orthodontic tooth movement. Am J Orthod Dentofacial Orthop. 2006; 130:364–70. 15. Ren Y, Maltha JC, Kuijpers-Jagtman AM. The rat as a model for experimental tooth movement- a critical review and a proposed solution. Eur J Orthod. 2006; 26:483–90. 16. Botting RM. Mechanism of action of acetaminophen: is there a cyclooxygenase 3? Clin Infect Dis. 2000; 31(Suppl 5):S202–10. 17. Kis B, Snipes JA, Simandle SA, Busija DW. Acetaminophen-sensitive prostaglandin production in rat cerebral endothelial cells. Am J Physiol Regul Integr Comp Physiol. 2005; 288:R897–902. 18. Green K, Drvota V, Vesterqvist O. Pronounced reduction of in vivo prostacyclin synthesis in humans by acetaminophen (paracetamol). Prostaglandins. 1989; 37(3):311–15. 19. Ngan P, Wilson S, Shanfeld J, Amini H. The effect of ibuprofen on the level of discomfort in patients undergoing orthodontic treatment. Am J Orthod Dentofacial Orthop. 1994; 106:88–95. 20. Polat O, Karaman AI. Pain control during fixed orthodontic appliance therapy. Angle Orthod. 2005; 75:214–19. 20. Polat O, Karaman AI. Pain control during fixed orthodontic appliance therapy. Angle Orthod. 2005; 75:214–19. 21. Sandy JR, Harris M. Prostaglandins and tooth movement. Eur J Orthod. 1984; 6:175–82. 22. Chumbley AB, Tuncay OC. The effect of indomethacin (an aspirin-like drug) on the rate of orthodontic tooth movement. Am J Orthod. 1986; 89:312–14. 23. Grieve WG, Johnson GK, Moore RN, Reinhardt RA, DuBois LM. Prostaglandin E (PGE) and interleukin- lβ (IL- lβ) levels in gingival crevicular fluid during human orthodontic tooth movement. Am J Orthod Dentofacial Orthop. 1994; 105:369–74. 24. Lee KJ, Park YC, Yu HS, Choi SH, Yoo YJ. Effects of continuous and interrupted orthodontic force on interleukin-1β and prostaglandin E2 production in gingival crevicular fluid. Am J Orthod Dentofacial Orthop. 2004; 125:168–77. 25. Perinetti G, Paolantonio M, Emanuela Serra E, D'Archivio D, D'Ercole S, FeliceFesta F, Spoto G. Longitudinal monitoring of subgingival colonization by Actinobacillus actinomycetemcomitans, and crevicular alkaline phosphatase and aspartate aminotransferase activities around orthodontically treated teeth. J Clin Periodontol. 2004; 31(1):60–7. Submit your manuscript to a journal and beneﬁ t from: 7 Convenient online submission 7 Rigorous peer review 7 Immediate publication on acceptance 7 Open access: articles freely available online 7 High visibility within the ﬁ eld 7 Retaining the copyright to your article Submit your next manuscript at 7 springeropen.com 26. Pender N, Samuels RH, Last KS. The monitoring of orthodontic tooth movement over a 2-year period by analysis of gingival crevicular fluid. Shetty et al. Progress in Orthodontics 2013, 14:6 http://www.progressinorthodontics.com/content/14/1/6 5. Yamasaki K, Shibata Y, Imai S, Tani Y, Shibasaki Y, Fukuhara T. Clinical application of prostaglandin E1 (PGE1) upon orthodontic tooth movement. Am J Orthod. 1984; 85:508–18. 31. Bonnefont J, Courade JP, Alloui A, Eschalier A. Mechanism of the antinociceptive effect of paracetamol. Drugs. 2003; 63:1–4. 32. Botting R, Ayoub SS. COX-3 and the mechanism of action of paracetamol/acetaminophen. Prostaglandins Leukot Essent Fatty Acids. 2005; 72(2):85–7. 6. Leiker BJ, Nanda RS DDS, Currier GF, Howes RI, Sinha PK. The effects of exogenous prostaglandins on orthodontic tooth movement in rats. Am J Orthod Dentofacial Orthop. 1995; 108:380–88. 33. Giunta D, Keller J, Nielsen FF, Melsen B. Influence of indomethacin on bone turnover related to orthodontic tooth movement in miniature pigs. Am J Orthod Dentofacial Orthop. 1995; 108:361–66. 7. Seifi M, Eslami B, Saffar AS. The effect of prostaglandin E2 and calcium gluconate on orthodontic tooth movement and root resorption in rats Eur J Orthod. 2003; 2:199–204. 34. Salmassian R, Oesterle LJ, Shellhart WC, Newman SM. Comparison of the efficacy of ibuprofen and acetaminophen in controlling pain after orthodontic tooth movement. Am J Orthod Dentofacial Orthop. 2009; 135:516–21. 8. Kale S, Kocadereli I, Atilla P, Asan E. Comparison of the effects of 1,25 dihydroxycholecalciferol and prostaglandin E2 on orthodontic tooth movement. Am J Orthod Dentofacial Orthop. 2004; 125:607–14. 9. Krishnan V, Davidovitch Z. Cellular, molecular, and tissue-level reactions to orthodontic force. Am J Orthod Dentofacial Orthop. 2006; 129:469e.1–460e.32 doi:10.1186/2196-1042-14-6 Cite this article as: Shetty et al.: Comparison of the effects of ibuprofen and acetaminophen on PGE2 levels in the GCF during orthodontic tooth movement: a human study. Progress in Orthodontics 2013 14:6. doi:10.1186/2196-1042-14-6 Cite this article as: Shetty et al.: Comparison of the effects of ibuprofen and acetaminophen on PGE2 levels in the GCF during orthodontic tooth movement: a human study. Progress in Orthodontics 2013 14:6. 10. Krishnan V. Orthodontic pain: from causes to management—a review. Eur J Orthod. 2007; 29:170–79. 11. Xiaoting L, Yin T, Yangxi C. Interventions for pain during fixed orthodontic appliance therapy. Angle Orthod. 2010; 80(5):925–32. 12. Kehoe MJ, Cohen SM, Zarrinnia K, Cowan A. The effect of acetaminophen, ibuprofen and misoprostol on prostaglandin E2 synthesis and the degree and rate of orthodontic tooth movement. Angle Orthod. 1996; 66(5):339–50. 13. Roche JJ, Cisneros GJ, Acs G. The effect of acetaminophen on tooth movement in rabbits. Angle Orthod. 1997; 67(3):231–36. 14. Arias OR, Marquez-Orozco MC. References 1. O'Connor PJ. Patients' perceptions before, during, and after orthodontic treatment. J Clin Orthod. 2000; 34(10):591–92. 2. Sari E, Olmez H, Gurton AU. Comparison of some effects of acetylsalicylic acid and rofecoxib during orthodontic tooth movement. Am J Orthod Dentofacial Orthop. 2004; 125:310–15. 3. Yamasaki K, Miura F, Suda T. Prostaglandin as a mediator of one resorption induced by experimental tooth movement in rats. J Dent Res. 1980; 59:1635–42. 4. Yamasaki K, Shibata Y, Fukuhara T. The effect of prostaglandins on experimental tooth movement in monkeys. J Dent Res. 1982; 61:1444–46. Yamasaki K, Shibata Y, Fukuhara T. The effect of prostaglandins on experimental tooth movement in monkeys. J Dent Res. 1982; 61:1444–46. Page 5 of 5 Page 5 of 5 Page 5 of 5 Shetty et al. Progress in Orthodontics 2013, 14:6 http://www.progressinorthodontics.com/content/14/1/6 Shetty et al. Progress in Orthodontics 2013, 14:6 http://www.progressinorthodontics.com/content/14/1/6 Eur J Orthod. 1994; 16(6):511–20. 27. Graf P, Glatt M, Brune K. Acidic nonsteroid anti-inflammatory drugs accumulating in inflamed tissue. Experientia. 1975; 31:951–53. 28. Bianchi M, Panerai AE. The dose-related effects of paracetamol on hyperalgesia and nociception in the rat. Br J Pharmacol. 1996; 117:130–32. 29. Bartzela T, Türp JC, Motschall E, Malthad JC. Medication effects on the rate of orthodontic tooth movement: a systematic literature review. Am J Orthod Dentofacial Orthop. 2009; 135:16–26. 30. Graham GG, Scott KF. Mechanism of action of paracetamol. Am J Ther. 2005; 12:46–55. 30. Graham GG, Scott KF. Mechanism of action of paracetamol. Am J Ther. 2005; 12:46–55.
https://openalex.org/W4318997011	https://pubs.rsc.org/en/content/articlepdf/2023/na/d2na00668e	English	null	A versatile platform for graphene nanoribbon synthesis, electronic decoupling, and spin polarized measurements	Nanoscale advances	2,023	cc-by	6,724	Open Access Article. Published on 02 February 2023. Downloaded on 10/24/2024 This article is licensed under a Creative Commons Attribution 3.0 U The on-surface synthesis of nano-graphenes has led the charge in prototyping structures with perspectives beyond silicon-based technology. Following reports of open-shell systems in graphene-nanoribbons (GNRs), a ﬂurry of research activity was directed at investigating their magnetic properties with a keen eye for spintronic applications. Although the synthesis of nano-graphenes is usually carried out on Au(111), the substrate is diﬃcult to use for electronic decoupling and spin-polarized measurements. Using a binary alloy Cu3Au(111), we show possibilities for gold-like on-surface synthesis compatible with spin polarization and electronic decoupling known from copper. We prepare copper oxide layers, demonstrate the synthesis of GNRs, and grow thermally stable magnetic Co islands. We functionalize the tip of a scanning tunneling microscope with carbon-monoxide, nickelocene, or attach Co clusters for high-resolution imaging, magnetic sensing, or spin-polarized measurements. This versatile platform will be a valuable tool in the advanced study of magnetic nano-graphenes. Received 29th September 2022 Accepted 2nd February 2023 DOI: 10.1039/d2na00668e rsc.li/nanoscale-advances signatures in carbon-based systems,5–7,11,12,15,21 further investi- gation may depend on our ability to provide orbital-imaging, spin-polarization and decoupling from itinerant electrons that are hard to simultaneously satisfy for either substrate. When looking for model-systems of spin-polarization, we notice in cobalt nanoislands on copper22,23 a similar monoculture with a heavy focus on this platform despite it being plagued by the rapid intermixing of Co and Cu at room temperature,24 which precludes its use for thermally induced nano-graphene forma- tion. Perhaps more liberty and exibility are found in investi- gations of decoupling layers such as Cu2N/Cu,25,26 MgO/Ag,27,28 or NaCl3,29 because they appear motivated by top-down ques- tions about the pristine properties of single atoms or molecules. A natural question is therefore to ask whether tailored on- surface synthesis, spin-polarization, and decoupling are exclu- sive and if not whether we can identify systems that simulta- neously host the properties that made one specic platform so popular. Nanoscale Advances Open Access Article. Published on 02 February 2023. Downloaded on 10/24/2024 6:18:00 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. View Article Online View Journal \| View Issue aDepartment of Physics, University of Zurich, Winterthurerstrasse 190, CH-8057 Zurich, Switzerland. E-mail: ales.cahlik@physik.uzh.ch; fabian.natterer@physik.uzh. ch bInstitute of Physics, EPFL, Station 3, CH-1015 Lausanne, Switzerland † Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2na00668e aDepartment of Physics, University of Zurich, Winterthurerstrasse 190, CH-8057 Zurich, Switzerland. E-mail: ales.cahlik@physik.uzh.ch; fabian.natterer@physik.uzh. ch © 2023 The Author(s). Published by the Royal Society of Chemistry bInstitute of Physics, EPFL, Station 3, CH-1015 Lausanne, Switzerland † Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2na00668e A versatile platform for graphene nanoribbon synthesis, electronic decoupling, and spin polarized measurements† Aleˇs Cahl´ık, a Danyang Liu, a Berk Zengin, a Mert Taskin, a Johannes Schwenk b and Fabian Donat Natterer a Introduction Through experience, we have come to accept that one cannot always have one's cake and eat it, too. Unfortunately, this is also true in scientic endeavors where it may even be self-induced. This partially originates in the way investigations grow from the bottom-up, expanding on previous progress and following a curiosity driven path. This style of research works well until a top-down challenge arises that may be incompatible within the so far used framework. We see parallels to this conundrum in research following the pioneering studies on on-surface synthesis of graphene nanoribbons (GNRs)1 and chemically tailored nano-graphenes.2 For their synthesis, an overwhelming number of contributions have converged on the coinage metals gold3–15 or copper.2,16–19 Although both provide simple prepara- tion, gold is deemed favorable for tailored synthesis because it reliably facilitates polymerization directly at the halogenated carbon site. For example, the on-surface reaction of 10,10′- dibromo-9,9′-bianthracene (DBBA) results in the growth of straight N = 7 armchair graphene nanoribbons (7-AGNRs) on Au(111),1 whilst on Cu(111) partially chiral nanoribbons are formed.17,20 However, with the emergence of magnetic Here we establish with Cu3Au such a system that combines the key properties of the coinage metals Au and Cu. We show the preparation of clean Cu3Au(111) surfaces and the growth of copper oxide decoupling layers. Using the prototypical halo- carbon precursor DBBA, we demonstrate its polymerization and cyclodehydrogenation into 7-AGNR. We show that the latter step can be temperature induced and possibly also be achieved via deliberate tip-manipulation using a scanning tunneling microscope (STM). We use a high-resolution STM with carbon- monoxide (CO) functionalized tips to verify the successful synthesis into 7-AGNRs and observe a spontaneous formation © 2023 The Author(s). Published by the Royal Society of Chemistry 1722 \| Nanoscale Adv., 2023, 5, 1722–1728 Nanoscale Advances View Article Online Nanoscale Advances View Article Online Paper Paper of GNRs exhibiting zero-bias peaks, suggestive of a Kondo resonance. In order to demonstrate spin-polarization on Cu3Au, we prepare cobalt nanoislands that are thermally and magnet- ically stable and serve as an easy means to produce spin- polarized tips. Together, these salient features demonstrated in our work equip the community with a versatile platform to tackle pressing top-down questions in surface science and in the advanced investigation of carbon-based systems. resolved photoelectron emission spectroscopy.33 The darker line-features in Fig. Cu3Au substrate and its oxidation We prepare clean surfaces of Cu3Au following previous work,30 using standard surface cleaning procedures described in the Experimental section. Fig. 1A shows an STM overview topog- raphy of a Cu3Au(111) surface termination with large terraces routinely exceeding 150 nm width and separated by monoa- tomic steps. High-resolution imaging using a carbon-monoxide functionalized tip shows the atomic lattice and the (2 × 2) supercell of the L12 ordered phase (Fig. 1B and C).31 Similar to the pure coinage metals, Cu3Au(111) exhibits a nearly-free electron like surface state with an eﬀective mass m/me = 0.31 ± 0.02 and a band onset of E0 = 0.42 eV below the Fermi level when measured using quasiparticle interference imaging,32 comparable to the observed value in previous work using angle- Fig. 1 Properties of Cu3Au(111) and copper oxide overlayers. (A) Overview topography of clean Cu3Au(111) terraces (V = 1 V, I = 140 pA) with the ideal L12 surface termination indicated in the inset, partially visible by the atomically resolved image in (B) and emphasized in its Fourier transform in (C), showing the atomic lattice (small circles) and the (2 × 2) supercell (large circles). (D) The surface hosts a nearly-free electron like surface state with an eﬀective mass of m/me = 0.31 ± 0.02 as determined by the parabolic ﬁt (red-line) to the dispersion. The dispersion plot was produced from the measurement of the energy dependent local density of states on a ﬁeld-of-view of 105 nm (drive frequency 1600 Hz, modulation amplitude 0.5 V, oﬀset −0.25 V). (E) Topographic image of the Cu3Au surface with oxide patches after exposing it to molecular oxygen. Inset shows the apparent height of the oxide patch relative to the Cu3Au surface (marked by the red line) (−0.5 V, 50 pA). (F) and (G) Threefold symmetric contrast in zoomed images of the oxide patch for two diﬀerent bias values (V = −0.3 V and V = 0.5 V respectively). (H) Point tunneling spectroscopy of the oxide phase showing a gap-size of about 1.3 V. (I) LEED image of the oxidized substrate at 80 V beam energy showing spots in addition to the Cu3Au lattice. Fig. 1 Properties of Cu3Au(111) and copper oxide overlayers. ss Article. Published on 02 February 2023. Downloaded on 10/24/2024 6:18:00 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Open Access Article. Published on 02 February 2023. Downloaded on 10/24/2024 6:18:00 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Having established the preparation of clean surfaces of Cu3Au and veried their properties, we proceed to demonstrate the growth of copper oxide overlayers. We deliberately dose small amounts of molecular oxygen while annealing the crystal (see the Experimental section) to form sub-monolayer patches of an oxidized substrate (Fig. 1E). The threefold symmetric contrast (Fig. 1F and G) is reminiscent of cuprous oxide in previous reports for Cu2O/Pt(111).34 Furthermore, the spectro- scopic signatures (Fig. 1H) of the oxide patch show a charac- teristic bandgap of about 1.3 eV in good agreement with observations for Cu2O/Au(111).35 The oxide formation is further conrmed by a distinct change in the LEED pattern aer oxygen exposure (Fig. 1I). We also observe the formation of oxide layers from interstitial oxygen that naturally segregates towards the surface during repeated and prolonged annealing at tempera- tures above 600 °C. In this case, two coexisting phases are formed (Fig. S2†). Besides the threefold symmetric phase Introduction 1A and E occasionally form complete hexagonal networks with a long-range periodicity of about (36 ± 3) nm (Fig. S1†). We propose that these networks are associated with the previously reported 29 nm surface reconstruction, which had been noticed as faint spots in low-energy electron diﬀraction (LEED) experiments and described as a “herring- bone-like” reconstruction30 (for details see the ESI†). Open Access Article. Published on 02 February 2023. Downloaded on 10/24/2024 6:18:00 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Remarkably, when scanning at elevated bias, we notice instabilities in the protrusions of polyanthrylene chains, indi- cating a mechanism for their deliberate dehydrogenation, similar to previous observations.38 We accordingly use the tip of the STM to selectively strip-oﬀhydrogen from selected parts of the polymerized chains (Fig. S3†). Although the thus formed ribbons show a similar apparent height than the thermally produced GNRs, they exhibit more irregularities and variation in their topographic and electronic structures. Whether tip mediated GNR formation is a matter of tuning the manipula- tion parameters or intrinsically limited, could become the subject of further study. Magnetic properties of cobalt nanostructures and spin- polarization We nally demonstrate the sub-monolayer (ML) growth of cobalt nanoislands on Cu3Au(111) and characterize their magnetic signatures. In contrast to the fast intermixing that is typical of Co/Cu(111) already at room-temperature, the Co/ Cu3Au interface is thermally robust. We do not observe signs of intermixing for substrate temperatures up to about 300 °C. Aer deposition of 0.3 ML of Co at about 200 °C, we observe the Fig. 2 On-surface synthesis of nano-graphenes on Cu3Au(111). (A) Two step 7-AGNR annealing scheme from DBBA precursor molecules. (B) Overview topography after dosing DBBA onto the substrate kept at 200 °C and post annealing at 300 °C (V = 50 mV, I = 30 pA). While DBBA readily polymerizes according to the scheme in (A), the cyclodehydrogenation was interrupted after 15 minutes annealing at the higher temperature to leave some chains polymerized but still hydrogenated (marked by the red rectangle). The black arrow marks a spontaneous tip functionalization with a CO molecule. (C) Upper: High-resolution STM image of a fully formed N = 7 armchair GNR using a carbon-monoxide functionalized tip. Lower: Laplace-ﬁltered version highlighting the GNR structure. (D) High-resolution STM image of a defective GNR measured using a carbon-monoxide functionalized tip. (E) A zero-bias peak in tunneling spectroscopy, suggestive of a Kondo resonance, obtained above a spontaneously formed defective GNR [positions marked by dots in panel (D)]. Fig. 2 On-surface synthesis of nano-graphenes on Cu3Au(111). (A) Two step 7-AGNR annealing scheme from DBBA precursor molecules. (B) Overview topography after dosing DBBA onto the substrate kept at 200 °C and post annealing at 300 °C (V = 50 mV, I = 30 pA). While DBBA readily polymerizes according to the scheme in (A), the cyclodehydrogenation was interrupted after 15 minutes annealing at the higher temperature to leave some chains polymerized but still hydrogenated (marked by the red rectangle). The black arrow marks a spontaneous tip functionalization with a CO molecule. (C) Upper: High-resolution STM image of a fully formed N = 7 armchair GNR using a carbon-monoxide functionalized tip. Lower: Laplace-ﬁltered version highlighting the GNR structure. (D) High-resolution STM image of a defective GNR measured using a carbon-monoxide functionalized tip. (E) A zero-bias peak in tunneling spectroscopy, suggestive of a Kondo resonance, obtained above a spontaneously formed defective GNR [positions marked by dots in panel (D)]. Fig. 2 On-surface synthesis of nano-graphenes on Cu3Au(111). Cu3Au substrate and its oxidation (A) Overview topography of clean Cu3Au(111) terraces (V = 1 V, I = 140 pA) with the ideal L12 surface termination indicated in the inset, partially visible by the atomically resolved image in (B) and emphasized in its Fourier transform in (C), showing the atomic lattice (small circles) and the (2 × 2) supercell (large circles). (D) The surface hosts a nearly-free electron like surface state with an eﬀective mass of m*/me = 0.31 ± 0.02 as determined by the parabolic ﬁt (red-line) to the dispersion. The dispersion plot was produced from the measurement of the energy dependent local density of states on a ﬁeld-of-view of 105 nm (drive frequency 1600 Hz, modulation amplitude 0.5 V, oﬀset −0.25 V). (E) Topographic image of the Cu3Au surface with oxide patches after exposing it to molecular oxygen. Inset shows the apparent height of the oxide patch relative to the Cu3Au surface (marked by the red line) (−0.5 V, 50 pA). (F) and (G) Threefold symmetric contrast in zoomed images of the oxide patch for two diﬀerent bias values (V = −0.3 V and V = 0.5 V respectively). (H) Point tunneling spectroscopy of the oxide phase showing a gap-size of about 1.3 V. (I) LEED image of the oxidized substrate at 80 V beam energy showing spots in addition to the Cu3Au lattice. Nanoscale Adv., 2023, 5, 1722–1728 \| 1723 © 2023 The Author(s). Published by the Royal Society of Chemistry View Article Online Nanoscale Advances Paper enrichment of the surface layer by Au atoms compared to the ideal Cu3Au(111) surface.30 observed aer the deliberate oxygenation, a new stripe-like phase emerges exhibiting a similar bandgap in tunneling spectroscopy. We have found no direct precedent for this phase, but Cu2O is known to constitute a wealth of structures including phases of reduced symmetry, such as in the Pt(111) system.36 In addition to pristine 7-AGNRs, we also observe the spon- taneous growth of defective ribbons (Fig. 2D). The interest in GNR defects lies in their connection to single electron spins in open-shell structures exhibiting magnetic signatures.15 Simi- larly, we observe the characteristic zero bias peaks in scanning tunneling spectroscopy for several defective structures, suggestive of Kondo resonances (Fig. 2E) – an established feature of the stabilized radical. Synthesis of nano-graphenes We proceed to verify the utility of Cu3Au(111) for the on-surface synthesis of nano-graphenes. The most prominent example is the growth of GNRs from DBBA precursor molecules on diﬀerent coinage metal substrates.1,17,20 We establish the growth of straight 7-AGNRs on Cu3Au(111) by pursuing the common two-step formation procedure (Fig. 2A). To that end, we deposit the DBBA precursor molecules on the substrate kept at 200 °C, leading to the dehalogenation and subsequent polymerization of the bi-anthracene moieties into one-dimensional chains. The second annealing step at a higher temperature of 300 °C yields through cyclodehydrogenation the fully formed GNRs (Fig. 2B). To obtain a peek into the two-step synthesis, we interrupt the high temperature annealing aer 15 minutes to ensure the coexistence of fully formed GNRs and the polymerized but still hydrogenated chains. The polyanthrylene chains can be distinguished from the GNRs by their bead-like protrusions and larger apparent-height as reported in previous work.1 We verify the structure of the 7-AGNRs using high-resolution STM imaging with a CO functionalized tip (Fig. 2C). The image clearly reveals the straight 7-AGNR structure corresponding to the nanoribbon formation on Au(111)1 in contrast to the growth of partially chiral nanoribbons from DBBA on Cu(111).17 We tentatively attribute this behavior to: (1) a vertical relaxation of Au atoms of the surface layer above the Cu atoms,37 exposing them more for Ullmann coupling reactions, and (2) a signicant Open Access Article. Published on 02 February 2023. Downloaded on 10/24/2024 6:18:00 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Magnetic properties of cobalt nanostructures and spin- polarization For the magnetic characterization, we focus on triangular islands due to their resemblance with the model system Co/ Cu(111).22 To that end, we use nickelocene (NiCp2) molecules, whose spin-excitation from the S = 0 ground state to the S = ±1 Fig. 3 Growth of cobalt islands on Cu3Au and facile preparation of spin-polarized tips. (A) Overview topography of two-monolayer high Co islands that were grown on Cu3Au at 200 °C with the line proﬁle across the island shown in the inset (V = 50 mV, I = 5 pA). The circular and triangular islands have diﬀerent stacking with respect to the substrate. (B) Tunneling conductance spectrum and (C) its derivative taken with a nickelocene functionalized tip for two diﬀerent heights [Dz = 4.2 Å, inset of (A)] above a Co atom in the middle of an island [marked in the inset of (B)]. The spectrum taken closer to the Co atom shows a splitting of ∼3 meV, corresponding to the exchange ﬁeld in the order of 13 T (39). (D) Topography image containing two similarly sized islands (dotted lines) with identical stacking used for the magnetic characterization of their magnetic bistability and spin-polarization. (E) Histogram of the d-level peak position with a mean of −310 mV as determined from 66 randomly picked triangular islands. (F) Closed loop conductance image taken at −290 mV, measured with a spin-averaging tip and the corresponding point spectra (G) of the orange and red islands showing no contrast diﬀerence (10 mV modulation). (H and I) Conductance map and tunneling spectra measured over the same region with a spin-polarized tip (10 mV modulation). The intensity of the d-level shows a clear diﬀerence that is attributed to two distinct magnetic orientations between the two islands with respect to the tip magnetization and that is also visible as contrast variation in the conductance map. (J and K) Conductance map and tunneling spectra with a spin-polarized tip whose magnetization sponta- neously switched, leading to a reversed contrast between the two magnetically diﬀerently oriented islands. (L) Spin-polarized tips can be produced by transferring an island from the substrate to the tip as shown by the before and after topographic images (V = 400 mV, I = 50 pA). Fig. 3 Growth of cobalt islands on Cu3Au and facile preparation of spin-polarized tips. Magnetic properties of cobalt nanostructures and spin- polarization (A) Two step 7-AGNR annealing scheme from DBBA precursor molecules. (B) Overview topography after dosing DBBA onto the substrate kept at 200 °C and post annealing at 300 °C (V = 50 mV, I = 30 pA). While DBBA readily polymerizes according to the scheme in (A), the cyclodehydrogenation was interrupted after 15 minutes annealing at the higher temperature to leave some chains polymerized but still hydrogenated (marked by the red rectangle). The black arrow marks a spontaneous tip functionalization with a CO molecule. (C) Upper: High-resolution STM image of a fully formed N = 7 armchair GNR using a carbon-monoxide functionalized tip. Lower: Laplace-ﬁltered version highlighting the GNR structure. (D) High-resolution STM image of a defective GNR measured using a carbon-monoxide functionalized tip. (E) A zero-bias peak in tunneling spectroscopy, suggestive of a Kondo resonance, obtained above a spontaneously formed defective GNR [positions marked by dots in panel (D)]. © 2023 The Author(s). Published by the Royal Society of Chemistry 1724 \| Nanoscale Adv., 2023, 5, 1722–1728 Nanoscale Advances View Article Online Paper excited state is sensitive to the magnetic or exchange eld applied along the molecular axis.40,41 At zero-eld, both ground state to S = ±1 transitions are degenerate but in an applied eld, the Zeeman splitting separates the single step in the tunneling conductance into two. Exposing a NiCp2 functional- ized tip to the exchange eld of a Co island on Cu3Au results in a clearly observable splitting of the step feature in STS when we move the NiCp2 tip closer to the Co island (Fig. 3B and C). The proximity leads to an enhanced exchange eld and concomi- tantly to a larger Zeeman splitting. This exchange interaction related shiin the spin-excitation step is equivalent to previous reports for NiCp2 measured against Co islands on Cu(111).40 growth of round and triangular Co islands of ∼3.4 Å apparent height (Fig. 3A). Compared to the cobalt islands on Cu(111), the islands appear smaller with the edge length spanning from 15 nm to a few nm, attributed to the larger lattice mismatch (5% for Co–Cu3Au and 2% for Co–Cu).39 At the same time, the islands are larger than cobalt clusters on Au(111) (13% lattice mismatch for Co–Au), where edge eﬀects lead to noncollinear magnetization.39 Open Access Article. Published on 02 February 2023. Downloaded on 10/24/2024 6:18:00 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. © 2023 The Author(s). Published by the Royal Society of Chemistry Magnetic properties of cobalt nanostructures and spin- polarization (A) Overview topography of two-monolayer high Co islands that were grown on Cu3Au at 200 °C with the line proﬁle across the island shown in the inset (V = 50 mV, I = 5 pA). The circular and triangular islands have diﬀerent stacking with respect to the substrate. (B) Tunneling conductance spectrum and (C) its derivative taken with a nickelocene functionalized tip for two diﬀerent heights [Dz = 4.2 Å, inset of (A)] above a Co atom in the middle of an island [marked in the inset of (B)]. The spectrum taken closer to the Co atom shows a splitting of ∼3 meV, corresponding to the exchange ﬁeld in the order of 13 T (39). (D) Topography image containing two similarly sized islands (dotted lines) with identical stacking used for the magnetic characterization of their magnetic bistability and spin-polarization. (E) Histogram of the d-level peak position with a mean of −310 mV as determined from 66 randomly picked triangular islands. (F) Closed loop conductance image taken at −290 mV, measured with a spin-averaging tip and the corresponding point spectra (G) of the orange and red islands showing no contrast diﬀerence (10 mV modulation). (H and I) Conductance map and tunneling spectra measured over the same region with a spin-polarized tip (10 mV modulation). The intensity of the d-level shows a clear diﬀerence that is attributed to two distinct magnetic orientations between the two islands with respect to the tip magnetization and that is also visible as contrast variation in the conductance map. (J and K) Conductance map and tunneling spectra with a spin-polarized tip whose magnetization sponta- neously switched, leading to a reversed contrast between the two magnetically diﬀerently oriented islands. (L) Spin-polarized tips can be produced by transferring an island from the substrate to the tip as shown by the before and after topographic images (V = 400 mV, I = 50 pA). © 2023 The Author(s). Published by the Royal Society of Chemistry Nanoscale Adv., 2023, 5, 1722–1728 \| 1725 View Article Online Paper Sample preparation The Cu3Au(111) crystal was purchased from Surface Preparation Laboratory. We mount it on a ferromagnetic sample plate to allow for the attachment of permanent magnets for eld- dependent studies46 and prepare atomically clean surfaces analogous to the literature30 by repeated cycles of Ar+ ion bombardment and annealing (3 keV, 0.8 mA cm−2, 550 °C). To verify the impact of the order/disorder transition, we perform extended annealing procedures (16 hours at 340 °C) to promote ordering of the L12 structure.31 We nd no evidence that the ordering level inuences the GNR formation or the cobalt island growth. The oxidized surface was prepared by annealing a clean Cu3Au substrate at 400 °C while exposing to an 8 × 10−7 mbar O2 atmosphere for 20 minutes. We grow cobalt islands from thoroughly degassed Co rods using a commercial e-beam evaporator (Focus EFM3) at sample temperatures from 60 °C to 300 °C. Increasing the substrate temperature beyond 300 °C results in visible Co intermixing at the step edges, emergence of vacancies and sinking of the Co islands into the substrate. For the GNR growth, we dose commercially available 10,10′- dibromo-9,9′-bianthracene precursors (Chemie Brunschwig AG, CAS: 121848-75-7) using an eﬀusion cell (Kentax, TCEorg-3BSC) or a home-made evaporator at a source temperature of 160 °C. For completeness, we nally demonstrate how to readily produce spin-polarized tips by simply picking up a Co island from the Cu3Au substrate which transfers Co to the tip apex (see the Experimental section for details). This procedure leaves the neighboring structures intact (Fig. 3L), causing minimal disruption to local details. Discussion and outlook When we look at the overall properties of Cu3Au, we identify interesting opportunities and perspectives. The most notable behavior is the thermal stability of the interface to increased temperature, which may be crucial for temperature driven on- surface synthesis. This holds, even in the presence of the copper oxide overlayer and the magnetic Co islands; the latter being thermally stable up to 300 °C. The value of high temperature stability is immediately clear when we compare it to the requirements for dehalogenation and cyclodehydrogenation as elementary steps in Ullmann coupled systems (Fig. 2A). Our work thus combines the controlled on-surface synthesis of GNRs known from Au(111) with the magnetism typical of Co/ Cu(111). Preliminary experiments prove that Co islands and GNRs can coexist (Fig. S4†). The absence of the short-range herringbone reconstruction on Cu3Au(111) is helpful in recip- rocal space analyses of synthesis steps and the possibility for magnetic Co islands oﬀers facile spin-polarization and magnetic references for the study of open-shell GNRs. In addi- tion, the copper oxide may serve as an alternative decoupling layer that isolates molecular structures synthesized via Ullmann coupling reactions from itinerant electrons. Additional surface chemistry can become active at elevated temperatures (see the discussion in the ESI and Fig. S5†) and may need to be scruti- nized against the requirements of one's experiment. The Cu3Au system can also stabilize a lattice matched copper-nitride for Nanoscale Advances which electronic decoupling was successfully demonstrated with single atoms.43,44 To establish the magnetic bistability of our Co islands, we select two triangular islands of identical stacking and similar size to avoid structural ambiguities (Fig. 3D). To further ensure their electronic equivalency without strain induced eﬀects,39,42 we measure the tunneling spectra with a spin-averaging tip. The two selected islands are spectroscopically equivalent and share the same d-level position at −290 mV of comparable intensity (Fig. 3F and G). We note that the average d-level position at −310 mV in tunneling spectroscopy (see the histogram for tunneling spectra taken above 66 random triangular islands in Fig. 3E) is comparable with the reported peak position for Co on Cu(111),39 although the shitowards lower energies could be expected due to larger lattice mismatch between cobalt and Cu3Au. We tentatively attribute this to edge induced eﬀects on the peak position due to the smaller island size.39 Using a spin- polarized tip on the same two islands, the intensity of the d-level peak becomes distinct between the orange and red islands (Fig. 3H and I), showing an unequivocal signature of tunnel magnetoresistance stemming from the diﬀerent magnetic orientations of the two islands with respect to the magnetiza- tion of the tip. This is further supported by a contrast reversal between the two islands aer a spontaneous switch of the tip magnetization (Fig. 3J and K). The tip functionalization on Cu3Au is straightforward which is valuable for the verication of synthesis steps or to enhance certain contrast modes. We demonstrate this with the routine use of CO, NiCp2, and Co cluster functionalized tips for high- resolution imaging, magnetic imaging, and spin-polarization, respectively. The facile tip-functionalization promotes further work on magnetism relying on direct tunneling45 or STM enabled electron-spin resonance experiments.28 In addition to functionalization, we also show how the tip can be used to selectively dehydrogenate polymerized molecules. This invites further work utilizing defects or barriers in circular or one- dimensional polymerized carbon structures. Tip functionalization For high-resolution imaging, we functionalize the tip with a carbon monoxide molecule that we pick oﬀthe substrate spontaneously during scanning (Fig. 2B) or by stabilizing the tip above the molecule (100 mV, 10 pA), switching oﬀthe feedback loop and increasing the bias to 3 V. We dose nickelocene (NiCp2) from a commercially available powder (Chemie Brunschwig AG, CAS: 1271-28-9) directly onto the cooled sample in the microscope head. The nickelocene tip functionalization is performed with feedback loop oﬀ, by manually approaching the tip to the molecule at 20 mV bias until a sharp jump in the current channel is observed. For spin-polarized measurements, we deliberately transfer a Co island from the Cu3Au substrate to the tip. This is done by approaching the island from 20 pA, 50 mV setpoint by 400–500 pm while applying 2–2.5 V bias. © 2023 The Author(s). Published by the Royal Society of Chemistry 1726 \| Nanoscale Adv., 2023, 5, 1722–1728 1726 Nanoscale Advances View Article Online Nanoscale Advances View Article Online Nanoscale Advances View Article Online View Article Online Paper Data availability 9 S. Mishra, D. Beyer, R. Berger, J. Liu, O. Gröning, J. I. Urgel, K. Müllen, P. Ruﬃeux, X. Feng and R. Fasel, J. Am. Chem. 9 S. Mishra, D. Beyer, R. Berger, J. Liu, O. Gröning, J. I. Urgel, K. Müllen, P. Ruﬃeux, X. Feng and R. Fasel, J. Am. Chem. Soc., 2020, 142, 1147–1152. 9 S. Mishra, D. Beyer, R. Berger, J. Liu, O. Gröning, J. I. Urgel, K. Müllen, P. Ruﬃeux, X. Feng and R. Fasel, J. Am. Chem. Soc., 2020, 142, 1147–1152. 10 S. Mishra, D. Beyer, K. Eimre, R. Ortiz, J. Fern´andez-Rossier, R. Berger, O. Gröning, C. A. Pignedoli, R. Fasel, X. Feng and P. Ruﬃeux, Angew. Chem., Int. Ed., 2020, 59, 12041–12047. 11 A. S´anchez-Grande, J. I. Urgel, A. Cahl´ık, J. Santos, S. Edalatmanesh, E. Rodr´ıguez-S´anchez, K. Lauwaet, P. Mutombo, D. Nachtigallov´a, R. Nieman, H. Lischka, B. de la Torre, R. Miranda, O. Gröning, N. Mart´ın, P. Jel´ınek and D. ´Ecija, Angew. Chem., Int. Ed., 2020, 59, 17594–17599. K. Müllen, P. Ruﬃeux, X. Feng and R. Fasel, J. Am. Chem. Soc., 2020, 142, 1147–1152. All data are available in the main text, or the ESI† can be found in the Zenodo repository, https://doi.org/10.5281/ zenodo.7034378. 10 S. Mishra, D. Beyer, K. Eimre, R. Ortiz, J. Fern´andez-Rossier, y R. Berger, O. Gröning, C. A. Pignedoli, R. Fasel, X. Feng and P. Ruﬃeux, Angew. Chem., Int. Ed., 2020, 59, 12041–12047. There are no conicts to declare. There are no conicts to declare. Open Access Article. Published on 02 February 2023. Downloaded on 10/24/2024 This article is licensed under a Creative Commons Attribution 3.0 U In summary, we demonstrate the unique utility of the binary alloy Cu3Au for nanographene synthesis, electronic decoupling and magnetic characterization. Our work enables advanced investigation of open-shell carbon systems using known on- surface chemistry steps required for Ullmann coupling while preserving magnetic structures and decoupling layers. We also show the substrate's versatility for diverse SPM tip-functionali- zation used for magnetic sensing, high-resolution imaging, and spin-polarized tunneling, and demonstrate a deliberate and bond-specic dehydrogenation of precursor molecules. Together, our work oﬀers a dependable foundation for advanced characterization of nanographenes. 6 S. Mishra, D. Beyer, K. Eimre, S. Kezilebieke, R. Berger, O. Gröning, C. A. Pignedoli, K. Müllen, P. Liljeroth, P. Ruﬃeux, X. Feng and R. Fasel, Nat. Nanotechnol., 2020, 15, 22–28. 7 S. Mishra, X. Yao, Q. Chen, K. Eimre, O. Gröning, R. Ortiz, M. Di Giovannantonio, J. C. Sancho-Garc´ıa, J. Fern´andez- Rossier, C. A. Pignedoli, K. Müllen, P. Ruﬃeux, A. Narita and R. Fasel, Nat. Chem., 2021, 13, 581–586. 8 J. Su, W. Fan, P. Mutombo, X. Peng, S. Song, M. Ondr´aˇcek, P. Golub, J. Brabec, L. Veis, M. Telychko, P. Jel´ınek, J. Wu and J. Lu, Nano Lett., 2021, 21, 861–867. ss Article. Published on 02 February 2023. Downloaded on 10/24/2024 6:18:00 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Open Access Article. Published on 02 February 2023. Downloaded on 10/24/2024 6:18:00 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. 4 J. Su, M. Telychko, S. Song and J. Lu, Angew. Chem., Int. Ed., 2020, 59, 7658–7668. Author contributions Conceptualization: AC and FDN, methodology: AC, FDN, and JS, formal analysis: AC, DL, and FDN, investigation: AC, BZ, DL, FDN, and MT, visualization: AC and DL, data curation: AC, writing—original dra: FDN, writing—review & editing: AC and FDN, supervision: FDN and AC, project administration: FDN, and funding acquisition: FDN, AC, and DL. 12 K. Biswas, L. Yang, J. Ma, A. S´anchez-Grande, Q. Chen, ll i d ´ ij l´ k 12 K. Biswas, L. Yang, J. Ma, A. S´anchez-Grande, Q. Chen, K. Lauwaet, J. M. Gallego, R. Miranda, D. ´Ecija, P. Jel´ınek, X. Feng and J. I. Urgel, Nanomaterials, 2022, 12, 224. eng and J. I. Urgel, Nanomaterials, 2022, 12, 224. Conﬂicts of interest g g 13 B. Cirera, A. S´anchez-Grande, B. de la Torre, J. Santos, S. Edalatmanesh, E. Rodr´ıguez-S´anchez, K. Lauwaet, B. Mallada, R. Zboˇril, R. Miranda, O. Gröning, P. Jel´ınek, N. Mart´ın and D. Ecija, Nat. Nanotechnol., 2020, 15, 437–443. 14 Y.-C. Chen, D. G. de Oteyza, Z. Pedramrazi, C. Chen, F. R. Fischer and M. F. Crommie, ACS Nano, 2013, 7, 6123– 6128. STM measurements and tunneling spectroscopy STM measurements and tunneling spectroscopy All experiments are performed in ultra-high vacuum using a commercial STM (CreaTec Fischer & Co. GmbH) operating at about 4 K. Our tip is made from a mechanically cut PtIr wire and sharpened by gently plunging it into the Cu3Au sample until we notice sharp step edges in topographic scans. For point-spec- troscopy, we use a conventional lock-in technique at a frequency of 932 Hz. For the surface state mapping, we use a multifre- quency lock-in amplier (Intermodulation Products SA, MLA-3) as described in ref. 32. 1 J. Cai, P. Ruﬃeux, R. Jaafar, M. Bieri, T. Braun, S. Blankenburg, M. Muoth, A. P. Seitsonen, M. Saleh, X. Feng, K. Müllen and R. Fasel, Nature, 2010, 466, 470–473. 2 M. Treier, C. A. Pignedoli, T. Laino, R. Rieger, K. Müllen, D. Passerone and R. Fasel, Nat. Chem., 2011, 3, 61–67. 3 P. Ruﬃeux, S. Wang, B. Yang, C. S´anchez-S´anchez, J. Liu, T. Dienel, L. Talirz, P. Shinde, C. A. Pignedoli, D. Passerone, T. Dumslaﬀ, X. Feng, K. Müllen and R. Fasel, Nature, 2016, 531, 489–492. 1 J. Cai, P. Ruﬃeux, R. Jaafar, M. Bieri, T. Braun, S. Blankenburg, M. Muoth, A. P. Seitsonen, M. Saleh, X. Feng, K. Müllen and R. Fasel, Nature, 2010, 466, 470–473. 2 M. Treier, C. A. Pignedoli, T. Laino, R. Rieger, K. Müllen, D. Passerone and R. Fasel, Nat. Chem., 2011, 3, 61–67. 3 P. Ruﬃeux, S. Wang, B. Yang, C. S´anchez-S´anchez, J. Liu, T. Dienel, L. Talirz, P. Shinde, C. A. Pignedoli, D. Passerone, T. Dumslaﬀ, X. Feng, K. Müllen and R. Fasel, Nature, 2016, 531, 489–492. Conclusions 5 J. Li, S. Sanz, J. Castro-Esteban, M. Vilas-Varela, N. Friedrich, T. Frederiksen, D. Peña and J. I. Pascual, Phys. Rev. Lett., 2020, 124, 177201. Open Access Article. Published on 02 February 2023. Downloaded on 10/24/2024 6:18:00 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. B. Wang, X. Zhou, D. Deng, B. Yang, Y. Yang, Z. Liu, X. Bao and F. Yang, Nat. Commun., 2020, 11, 2312. 20 P. Han, K. Akagi, F. Federici Canova, H. Mutoh, S. Shiraki, K. Iwaya, P. S. Weiss, N. Asao and T. Hitosugi, ACS Nano, 2014, 8, 9181–9187. 35 H. Str¨ater, H. Fedderwitz, B. Groß and N. Nilius, J. Phys. Chem. C, 2015, 119, 5975–5981. 36 A. Gloystein and N. Nilius, J. Phys. Chem. C, 2019, 123, 26939–26946. 21 S. Song, J. Su, M. Telychko, J. Li, G. Li, Y. Li, C. Su, J. Wu and J. Lu, Chem. Soc. Rev., 2021, 50, 3238–3262. 37 M. Dion´ızio Moreira, G. N. Fontes, H. Niehus, C. A. Achete and R. B. Capaz, J. Vac. Sci. Technol., B: Nanotechnol. Microelectron.: Mater., Process., Meas., Phenom., 2012, 30, 051802. Open Access Article. Published on 02 February 2023. Downloaded on 10/24/2024 6 This article is licensed under a Creative Commons Attribution 3.0 Un 22 O. Pietzsch, A. Kubetzka, M. Bode and R. Wiesendanger, Phys. Rev. Lett., 2004, 92, 057202. 23 O. Pietzsch, S. Okatov, A. Kubetzka, M. Bode, S. Heinze, A. Lichtenstein and R. Wiesendanger, Phys. Rev. Lett., 2006, 96, 237203. 38 S. Blankenburg, J. Cai, P. Ruﬃeux, R. Jaafar, D. Passerone, X. Feng, K. Müllen, R. Fasel and C. A. Pignedoli, ACS Nano, 2012, 6, 2020–2025. 24 A. Rabe, N. Memmel, A. Steltenpohl and Th. Fauster, Phys. Rev. Lett., 1994, 73, 2728–2731. 39 B. W. Heinrich, C. Iacovita, M. V. Rastei, L. Limot, P. A. Ignatiev, V. S. Stepanyuk and J. P. Bucher, Eur. Phys. J. B, 2010, 75, 49–56. 25 C. D. Ruggiero, T. Choi and J. A. Gupta, Appl. Phys. Lett., 2007, 91, 253106. 40 B. Verlhac, N. Bachellier, L. Garnier, M. Ormaza, P. Abufager, 26 S. Loth, K. von Bergmann, M. Ternes, A. F. Otte, C. P. Lutz and A. J. Heinrich, Nat. Phys., 2010, 6, 340–344. R. Robles, M.-L. Bocquet, M. Ternes, N. Lorente and L. Limot, Science, 2019, 366, 623–627. 27 H.-J. Shin, J. Jung, K. Motobayashi, S. Yanagisawa, Y. Morikawa, Y. Kim and M. Kawai, Nat. Mater., 2010, 9, 442–447. 41 G. Czap, P. J. Wagner, F. Xue, L. Gu, J. Li, J. Yao, R. Wu and W. Ho, Science, 2019, 364, 670–673. 28 S. Baumann, W. Paul, T. Choi, C. P. Lutz, A. Ardavan and A. J. Heinrich, Science, 2015, 350, 417–420. 42 M. V. Rastei, B. © 2023 The Author(s). Published by the Royal Society of Chemistry Acknowledgements Swiss National Science Foundation PP00P2_176866 (FDN, DL, BZ, and AC), Swiss National Science Foundation 200021_200639 (FDN and AC), Swiss National Science Foundation PP00P2_211014 (FDN, DL, BZ), Swiss Government Excellence Fellowship (AC), Oﬃce of Naval Research N00014-20-1-2352 (FDN and BZ), and UZH Forschungskredit FK-20-093 (DL). We thank T. Diulus, A. Kinikar, M. Hengsberger, N. Bachellier, J. Osterwalder, R. Wiesendanger, and R. Fasel for fruitful discussion. 15 J. Li, S. Sanz, M. Corso, D. J. Choi, D. Peña, T. Frederiksen and J. I. Pascual, Nat. Commun., 2019, 10, 200. 16 J. D. Teeter, P. S. Costa, M. M. Pour, D. P. Miller, E. Zurek, A. Enders and A. Sinitskii, Chem. Commun., 2017, 53, 8463– 8466. © 2023 The Author(s). Published by the Royal Society of Chemistry Nanoscale Adv., 2023, 5, 1722–1728 \| 1727 View Article Online Nanoscale Advances Paper Paper 32 B. Zengin, J. Oppliger, D. Liu, L. Niggli, T. Kurosawa and F. D. Natterer, Phys. Rev. Res., 2021, 3, L042025. 17 C. S´anchez-S´anchez, T. Dienel, O. Deniz, P. Ruﬃeux, R. Berger, X. Feng, K. Müllen and R. Fasel, ACS Nano, 2016, 10, 8006–8011. 33 R. Courths, M. Lau, T. Scheunemann, H. Gollisch and R. Feder, Phys. Rev. B: Condens. Matter Mater. Phys., 2001, 63, 195110. 18 K. Sun, X. Li, L. Chen, H. Zhang and L. Chi, J. Phys. Chem. C, 2020, 124, 11422–11427. 34 W. Huang, Q. Liu, Z. Zhou, Y. Li, Y. Ling, Y. Wang, Y. Tu, 19 M. Telychko, G. Li, P. Mutombo, D. Soler-Polo, X. Peng, J. Su, 19 M. Telychko, G. Li, P. Mutombo, D. Soler-Polo, X. Peng, J. Su, S. Song, M. J. Koh, M. Edmonds, P. Jel´ınek, J. Wu and J. Lu, Sci. Adv., 2021, 7, eabf0269. Open Access Article. Published on 02 February 2023. Downloaded on 10/24/2024 6:18:00 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Heinrich, L. Limot, P. A. Ignatiev, V. S. Stepanyuk, P. Bruno and J. P. Bucher, Phys. Rev. Lett., 2007, 99, 246102. 29 J. Repp, G. Meyer, S. M. Stojkovi´c, A. Gourdon and C. Joachim, Phys. Rev. Lett., 2005, 94, 026803. 43 J. Gobeil, D. Coﬀey, S.-J. Wang and A. F. Otte, Surf. Sci., 2019, 679, 202–206. 30 O. Bauer, C. H. Schmitz, J. Ikonomov, M. Willenbockel, S. Soubatch, F. S. Tautz and M. Sokolowski, Phys. Rev. B, 2016, 93, 235429. 44 R. Rejali, D. Coﬀey, J. Gobeil, J. W. Gonz´alez, F. Delgado and A. F. Otte, npj Quantum Mater., 2020, 5, 1–7. 31 H. Okamoto, D. J. Chakrabarti, D. E. Laughlin and T. B. Massalski, J. Phase Equilib., 1987, 8, 454. 45 R. Wiesendanger, Rev. Mod. Phys., 2009, 81, 1495. 46 D. Liu, J. Oppliger, A. Cahl´ık, C. Witteveen, F. O. von Rohr and F. D. Natterer, MethodsX, 2023, 10, 101964. 1728 \| Nanoscale Adv., 2023, 5, 1722–1728 © 2023 The Author(s). Published by the Royal Society of Chemistry
W4292749725.txt	http://journal-app.uzhnu.edu.ua/article/download/258837/255600	en		Counteraction to customs offenses in the international and Ukrainian regulatory framework	Analìtično-porìvnâlʹne pravoznavstvo	2,022	cc-by	3,215	РОЗДІЛ VII. АДМІНІСТРАТИВНЕ ПРАВО І ПРОЦЕС; ФІНАНСОВЕ ПРАВО; ІНФОРМАЦІЙНЕ ПРАВО 141 УДК 342.9:339.543 DOI https://doi.org/10.24144/2788-6018.2022.01.26 ПРОТИДІЯ МИТНИМ ПРАВОПОРУШЕННЯМ У МІЖНАРОДНОМУ ТА УКРАЇНСЬКОМУ НОРМАТИВНО-ПРАВОВОМУ ПОЛІ Герчаківська О.Я., аспірант кафедри конституційного, адміністративного та фінансового права Західноукраїнського національного університету, https://orcid.org/0000-0001-6648-662X Герчаківська О.Я. Протидія митним правопорушенням у міжнародному та українському нормативно-правовому полі. В статті проаналізовано міжнародне законодавство у сфері протидії митним правопорушенням. Приділено увагу розгляду діяльності міжнародних інституцій, які сформували пласт нормативно-правового забезпечення митної сфери, зокрема Світової організації торгівлі, Всесвітньої митної організації. Визначено особливості митних девіацій в контексті Міжнародної конвенції про взаємну адміністративну допомогу у відверненні, розслідуванні та припиненні порушень митного законодавства. З’ясовано, що відповідно до норм Йоганнесбурзької конвенції використовуються спеціальні заходи взаємодії та протидії митним правопорушенням: нагляд, контрольована поставка, залучення експертів і свідків, спільні групи контролю та розслідування та інші. Проаналізовано нормативно-правові аспекти стимулювання суб’єктів зовнішньоекономічної діяльності не порушувати митні правила через надання статусу авторизованих економічних операторів. Досліджено правову основу механізму протидії митним правопорушенням в ЄС, зокрема в контексті первинного й вторинного законодавства. Встановлено, що гармонізація міжнародного митного законодавства та України здійснюється у двох основних векторах: відповідно до базових норм міжнародного права та відповідно до регламентів і стандартів регулювання ЄС. Ключові слова: митні правопорушення, Всесвітня митна організація, заходи протидії, корупція, митні органи, гармонізація. Gerchakivska O.Y. Counteraction to customs offenses in the international and Ukrainian regulatory framework. The article analyzes international legislation in the field of counteraction to customs offenses. Attention is paid to the consideration of the activities of international institutions that have formed a layer of regulatory and legal support for the customs sphere, in particular, the World Trade Organization, the World Customs Organization. The peculiarities of customs deviances in the context of the International Convention on Mutual Administrative Assistance in the Treatment, Investigation and Termination of Violations of Customs Legislation have been determined. It was clarified that in accordance with the norms of the Johannesburg Convention, special measures of interaction and counteraction to customs offenses are used: supervision, controlled delivery, involvement of experts and witnesses, joint control and investigation teams and others. The normative and legal aspects of stimulation of subjects of foreign economic activity not to violate customs rules through granting the status of authorized economic operators are analyzed. The legal basis for the work of the mechanism for combating customs offenses in the EU, in particular in the context of primary and secondary EU legislation, is investigated. It is established that the harmonization of international customs legislation and Ukraine is carried out in two main vectors: in accordance with the basic norms of international law and in accordance with the regulations and standards of regulation of the EU. Key words: customs offenses, World Customs Organization, countermeasures, corruption, customs authorities, harmonization. Митні правопорушення є одними із найпоширеніших та найнебезпечніших девіацій у сфері суспільного життя. Незаконне переміщення через митний кордон держави товарів, транспортних засобів, наркотичних засобів, зброї, боєприпасів, вибухових речовин, культурних та історичних цінностей традиційно перебуває в центрі уваги як кримінальних угруповань, так і пересічних громадян як один із способів швидкого збагачення. Актуалізація протидії митним злочинам пов’язана із загальним глобальним трендом відкриття кордонів та повної лібералізації торговельних відносин між державами. Без сумніву, що на динаміку митної злочинності та індексів світової економіки в світі негативно 142 Електронне наукове видання «Аналітично-порівняльне правознавство» вплине російсько-українська війна. Так, «використовуючи глобальну модель економічного моделювання, Секретаріат Світової організації торгівлі прогнозує, що криза може знизити зростання світового ВВП на 0,7%-1,3%, в результаті чого зростання складе десь між 3,1% і 3,7% на 2022 рік. Модель також передбачає, що зростання світової торгівлі в цьому році може бути скорочено майже вдвічі з 4,7%, прогнозованих СОТ в жовтні минулого року, до діапазону 2,4% і 3%» [1]. Для України проблематика протидії митним правопорушенням перебуває в площині напрацювання ефективних механізмів митного контролю, використання новітніх технічних та спеціальних засобів пошуку та ідентифікації предметів контрабанди, застосування мотиваційних механізмів для митних служб в частині мінімізації корупційних ризиків. Такі завдання можна втілити через імплементацію в національне правове поле стандартів та регламентів, які впродовж десятиліть відпрацьовувались та впроваджувались на рівні міжнародних конвенцій, декларацій, директив, регламентів, стандартів тощо. Теоретико-правові контури дослідження митних девіацій, запобігання та протидії митним правопорушенням містяться у працях В. Авер’янова, М. Аракелян, Ю. Бисаги, К. Борисова, В. Бурлакова, Т. Гавриша, Г. Дейвис, О. Джужі, А. Долгової, Л. Ентін, А. Зелінського, В. Емінова, А. Капустін, В. Кудрявцева, Н. Кузнєцової, А. Крисоватого, О. Литвинова, В. Лунєєва, А. Наку, М. Палінчак, П. Пашка, Д. Приймаченка та ін. Водночас динамічність розвитку міжнародного та українського нормативно-правового поля в сфері митних правовідносин потребують осучаснення наукового пізнання, дослідження гармонізації європейських та вітчизняних підходів протидії митній злочинності. Метою статті є вивчення правових домінант щодо протидії митним правопорушенням в світовій та українській практиці. Для досягнення поставленої мети виникла необхідність у дослідженні наступних питань: 1) розглянути генезис ухвалення міжнародних правових документів у сфері регулювання митних правовідносин та участь у них України; 2) проаналізувати наявність в міжнародних конвенціях та директивах складової, що стосується протидії митним правопорушенням; 3) окреслити вектори покращення вітчизняного нормативно-правового поля щодо боротьби з митним порушеннями. Важливою компонентою міжнародного митного законодавства виступають митні угоди міждержавного характеру, які регламентують усталені стандарти та регламенти митної справи. До інституцій, які сформували пласт нормативно-правового забезпечення митної сфери в глобальному вимірі відносять Світову організацію торгівлі (World Trade Organization) і Всесвітню митну організацію (World Customs Organization). Україна бере участь у таких міжнародних угодах, укладених на рівні зазначених організацій, серед яких варто проаналізувати наступні. Конвенція про створення Ради Митного Співробітництва 1950 р. [2] (Україна підписала Конвенцію в 1992 р., ставши членом цієї організації), що визначила організаційні засади функціонування Ради Митного Співробітництва, а згодом Всесвітньої митної організації. Міжнародна конвенція про спрощення і гармонізацію митних процедур [3] (Кіотська конвенція 1973 р.) (приєднання України відбулось у 2006 р.). Кіотська конвенція регламентувала, що національне законодавство визначає митні правопорушення та встановлює умови, за яких вони можуть розслідуватися, встановлюватися та, у разі необхідності, бути об’єктом адміністративного врегулювання. Національне законодавство визначає порядок дій митної служби після виявлення нею митного правопорушення та заходи, яких вона може вжити. Обставини вчинених митних правопорушень і вжитих заходів митною службою потрібно зазначати в адміністративних протоколах. Багато підходів з протидії митним девіаціям Кіотської конвенції лягло в основу митного законодавства України, зокрема, у випадку виявлення під час митного оформлення правопорушення, що вважається незначним, декларанту чи уповноваженій особі повинна бути надана можливість адміністративного врегулювання цього правопорушення. Така особа має право на подачу скарги в орган, незалежний від митної служби. Міжнародна конвенція про взаємну адміністративну допомогу у відверненні, розслідуванні та припиненні порушень митного законодавства [4] (Найробська конвенція 1980 р.) (Україна бере участь із 2000 р. відповідно до Указу Президента України «Про приєднання України до Міжнародної конвенції про взаємну адміністративну допомогу у відверненні, розслідуванні та припиненні порушень митного законодавства та додатків до неї)»[5]. Ухвалення цього міжнародного документу відбулось в контексті визнання того, що порушення митного законодавства становлять загрозу економічним, соціальним, фіскальним інтересам держав та законним інтересам торгівлі. Разом з тим, протидія порушенням митного законодавства може бути ефективнішою за умови співробітництва між митними інституціями. Найробська конвенція визначила ряд дефініцій у сфері протидії митним правопорушенням (рис.1). РОЗДІЛ VII. АДМІНІСТРАТИВНЕ ПРАВО І ПРОЦЕС; ФІНАНСОВЕ ПРАВО; ІНФОРМАЦІЙНЕ ПРАВО 143 Митні правопорушення відповідно до Найробської конвенції 1980 р. порушення митного законодавства митне шахрайство контрабанда будь-яке порушення або порушення митного законодавства, за допомогою якого особа вводить митні органи в оману і, таким чином, уникає, повністю або частково, сплати імпортних або експортних мит та податків, чи виконання заборон і обмежень, встановлених митним законодавством, або отримує будь-яку вигоду від порушення митного законодавства. митне шахрайство, що полягає у переміщенні товарів через митний кордон у будь-який прихований спосіб. спроба порушення митного законодавства. Рис. 1. Митні правопорушення та їх дефініції відповідно до Найробської конвенції [4] Найбільш вдалим, на нашу думку, терміном, що характеризує девіантну поведінку суб’єктів зовнішньоекономічної діяльності та громадян при перетині митного кордону держави є митне шахрайство. Така дефініція вказує на спосіб вчинення митного правопорушення та наслідки його заподіяння. Крім цього, Найробська конвенція 1980 р. в Розділі ІІ розшифровує «митне шахрайство, що не є контрабандою» та інші види шахрайства, включаючи підлог, фальсифікацію та підробку. Щодо протидії митним правопорушенням учасники Конвенції мають повідомляти один одного про кожне застосування відомих, а також нових, незвичних та можливих методів вчинення контрабанди та інших видів шахрайства з метою виявлення сучасних тенденцій у цій галузі та відпрацювання заходів протидії. В додатках до Конвенції визначено особливості надання допомоги у боротьбі з контрабандою наркотиків та психотропних речовин та протидії контрабанді творів мистецтва, антикваріату та інших культурних цінностей. Таким чином, Міжнародна конвенція про взаємну адміністративну допомогу у відверненні, розслідуванні та припиненні порушень митного законодавства стала одним із перших профільних нормативно-правових документів міжнародного характеру у сфері визначення митних правопорушень та їх протидії. Міжнародна конвенція про Гармонізовану систему опису та кодування товарів 1983 р. [6] (приєднання України відбулось у 2002 р.), яка регламентувала підходи до кодування товарів в міжнародній торгівлі, номенклатури товарів та застосування в митному адмініструванні ставок мита. Конвенція про тимчасове ввезення [7] (Стамбульська конвенція 1990 р.) (Україна - учасник з 2004 р.) спрямована на спрощення митних формальностей, які пов’язані з тимчасовим ввезенням. В статті 20 цієї Конвенції визначено особливості порушень та встановлення відповідальності за такі дії. Якщо неможливо визначити територію, на якій трапилося порушення, воно вважається скоєним на території договірної сторони, де воно було виявлено. Ще одним документом, що регламентує співробітництво в сфері протидії митним правопорушенням є Міжнародна Конвенція про адміністративну допомогу у сфері митних відносин від 27.06.2003 р. (Йоганнесбурзька Конвенція). Аналізуючи зміст Конвенції, відзначимо, що основною метою її учасників є забезпечення безпеки міжнародної системи постачання товарів, шляхом профілактики, розслідування та припинення митних правопорушень. Тобто протидія порушень у митній сфері є пріоритетним завданням країн, що уклали цю Конвенцію або ж приєдналися до неї пізніше. Так зокрема, стаття 2 Йоганнесбурзької Конвенції встановлює, що співпраця, здійснювана в рамках цієї Конвенції договірною стороною, повинна відповідати її законодавчим й адміністративним нормам, виконуватись у межах компетенції митної адміністрації договірної сторони та її наявних ресурсів, передбачає можливість здійс- 144 Електронне наукове видання «Аналітично-порівняльне правознавство» нення оперативно-розшукових заходів. У цьому контексті поділяємо застереження деяких вітчизняних науковців та практиків, які стверджують: «Недоліком вітчизняної системи правоохоронних органів є відсутність у митній службі України права на ведення оперативно-розшукової діяльності, що не дозволяє отримувати оперативно важливу інформацію від відповідних органів сусідніх держав» [8, c.18]. Стаття 5 Йоганнесбурзької Конвенції встановлює, що митні органи на прохання чи за власною ініціативою обмінюються інформацією з метою недопущення або розслідування, припинення митних правопорушень, а також для убезпечення міжнародної системи постачання товарів. Ця інформація може містити відомості про наступне: новітні підходи імплементації митного законодавства, що виявилися ефективними; сучасні тенденції, інструментарій та способи вчинення митних правопорушень; товари, що є предметом митного правопорушення, а також транспортні засоби та способи зберігання (приховування) цих товарів; осіб, які вчинили митне правопорушення (здійснили спробу) або підозрюються в його вчиненні тощо. Розділ V Йоганнесбурзької конвенції передбачає спеціальні заходи взаємодії та протидії митним правопорушенням: нагляд, контрольована поставка, залучення експертів і свідків, спільні групи контролю та розслідування та інші. Рамкові стандарти безпеки і полегшення всесвітньої торгівлі [9], (прийняті за участю представників України та 166 держав-членів ВМО). Цей міжнародний документ, що ухвалений 23.06.2005 р. сформував нову парадигму безпеки міжнародної торгівлі та створив новітню архітектоніку взаємовідносин між митними інституціями в світі. Рамкові стандарти ВМО складаються з 4 основних елементів, які скеровані на запобігання та протидію митним правопорушенням: 1) гармонізація вимог щодо попереднього електронного декларування вантажних відправлень; 2) управління митними ризиками в контексті реалізації митної безпеки держав; 3) використання технічних та спеціальних засобів митного контролю (великоформатних рентгенапаратів та детекторів радіації) при догляді вантажів з високим рівнем ризику; 4) застосування системи пільг та преференцій для ефективних суб’єктів зовнішньоекономічної діяльності, які не притягувались до відповідальності за порушення митного законодавства. В Україні така система митних пільг та преференцій знайшла відображення в запровадженні інституту автоматизованого економічного оператора (АЕО), що засвідчило наближення вітчизняного митного законодавства до європейських стандартів. Запровадження АЕО спрощує та полегшує міжнародну торгівлю, забезпечує участь у формуванні безпечних ланцюгів постачання товарів, підвищує їх конкурентоспроможність як на внутрішньому, так і на зовнішньому ринках. АЕО – це особливий статус, який засвідчує високий ступінь довіри до суб’єкта господарювання та надає йому суттєві спрощення при здійсненні митних формальностей [10]. Упродовж останніх років Мінфіном, Держмитслужбою України за сприяння Програми ЄС з підтримки управління державними фінансами в Україні (EU4PFM) забезпечено підготовку та прийняття нормативно-правової бази в цій сфері, зокрема: Закону України від 02.10.2019 р. № 141-IX «Про внесення змін до Митного кодексу України щодо деяких питань функціонування авторизованих економічних операторів» (набрав чинності 7 листопада 2019 р.) та Постанови Кабінету Міністрів України від 29.07.2020 р. №665 «Про деякі питання функціонування авторизованих економічних операторів». Однак через процедурні складнощі та жорсткі вимоги щодо отримання статусу, інститут АЕО не знайшов в Україні значного поширення. Важливим аспектом протидії митним девіаціям, зокрема корупційним правопорушенням на митниці став Стандарт 10 Рамкових стандартів ВМО «Морально-етичні якості персоналу». У ньому визнається Арушська декларація як джерело формування антикорупційних систем митних органів світу. Саме базові принципи професійної етики та боротьби з корупцією у митній сфері регламентовано Декларацією про сумлінність митних службовців, відомої як Арушська декларація 1993 р. Арушська декларація складається з 10 основних принципів, що підлягають застосуванню національними митними органами під час розроблення національного законодавства щодо забезпечення боротьби з корупцією на митницях [11, c.70]. 2003 р. була прийнята нова редакція цієї Декларації, що включає в себе ряд ключових елементів, реалізація яких хоч і носить для національних митних служб рекомендаційний характер, але підкреслює важливість їх послідовного впровадження в практику митної служби. Положення Арушської декларації знайшли своє відображення в Дисциплінарному статуті митних органів України [12], однак після багаторічних інституційних трансформацій (ДМСУ – Міндоходів – ДФСУ – знову ДМСУ) повної імплементації вимог Декларації у внутрішньовідомчих документах митної служби не відбулось. Вважаємо за необхідне проаналізувати і європейські підходи щодо протидії митним правопорушенням. Правовою основою роботи механізму співробітництва у сфері боротьби з митними правопорушеннями в ЄС, є, перш за все, первинне законодавство ЄС, джерелами якого виступають статутні договори, що закріплюють поняття і принципи митного союзу країн ЄС. До РОЗДІЛ VII. АДМІНІСТРАТИВНЕ ПРАВО І ПРОЦЕС; ФІНАНСОВЕ ПРАВО; ІНФОРМАЦІЙНЕ ПРАВО базових актів первинного законодавства ЄС, що містять норми митного права, відносять: Договір про заснування Європейського об’єднання вугілля і сталі 1951 р.; Договір про заснування Європейської комісії з атомної енергії 1957 р.; Договір про заснування Європейського Економічного Співтовариства 1957 р.; Договір про Європейський Союз 1992 р.; Лісабонська угода про внесення змін у Договір про Європейський союз та Договір заснування Європейського співтовариства 2007 р.; Угода між Україною і Європейським Союзом про участь України у програмі Європейського Союзу Горизонт 2020 (ратифіковано Законом № 604-VIII від 15.07.2015 р.) та інші [8, c.22]. До актів вторинного законодавства ЄС відносять акти, що видаються інститутами Союзу, а також всі інші акти, які приймаються на основі статутних договорів. Акти вторинного законодавства ЄС у сфері митного співробітництва мають такі основні правові форми: регламенти, директиви, рішення. Основним документом ЄС, який регламентує митно-тарифне регулювання, є Модернізований митний кодекс Європейського Союзу (Modernised Customs Code, MСС), який затверджено Регламентом від 23 квітня 2008 року № 450/2008, з 1 червня 2016 року у повному обсязі набув чинності Митний кодекс Європейського Союзу (Union Customs Code, UСС), затверджений Регламентом Європейського парламенту і Ради ЄС № 952/2013 [13]. Митний кодекс Союзу (UCC) є основним міжнародним документом, що забезпечує митні правила й процедури на митній території ЄС, що адаптовані до сучасних торгових реалій та новітніх комунікаційних інструментів. На сьогодні існує перехідний період для деяких його норм. Це пов’язано, в першу чергу, з тим, що існує необхідність розробки нових ІТ-систем або модернізації існуючих з метою повного виконання законодавчих вимог. Цей перехідний період в даний час триває і планується продовжити до 2025 року для певної кількості митних формальностей, що адмініструються електронними системами. Вагоме місце в європейському митному праві посідають окремі, регламенти, директиви та стандарти. Так, зокрема 29 квітня 2021 року було прийнято Регламент (ЄС) 2021/785 Європейського парламенту та Ради «Про створення Програми боротьби з шахрайством Союзу» [14], який вимагає від ЄС та держав-членів протидіяти шахрайству та будь-якій іншій незаконній діяльності. Союз повинен підтримувати діяльність у цих сферах. Основними цілями програми є: запобігання шахрайству, корупції та будь-якій іншій незаконній діяльності, що зачіпає фінансові інтереси Союзу; підтримувати повідомлення про порушення, в тому числі шахрайство, щодо спільних управлінських фондів та фондів допо- 145 моги перед вступом до Союзу; забезпечити інструменти для обміну інформацією та підтримки операційної діяльності у сфері взаємної адміністративної допомоги в митних та сільськогосподарських питаннях. Протидія шахрайству, в тому числі митним правопорушенням, буде здійснюватися через декілька оперативних заходів. До них відносять створення платформи інформаційних технологій. Бюджет на реалізацію цієї програми на період 2021-2027 рр. становить 181 207 млн євро [14]. Платформа включає в себе кілька модулів, в тому числі митну інформаційну систему, яка спрямована на допомогу державам-членам у запобіганні, розслідуванні та переслідуванні осіб, які порушують митне законодавство. Це відбуватиметься шляхом більш швидкого поширення інформації, ефективності процедур митного співробітництва та контролю митних служб, повноваження яких охоплюють митні операції. Загалoм міжнародний правoвий механiзм прoтидiї пoрушенням митним правопорушенням мiстить наступні елементи: нoрми адмiнiстративнoгo права; адмiнiстративнo-правoвi вiднoсини; адмiнiстративнo-правoвi метoди; прoцедури реалiзацiї нoрм адмiнiстративнoгo та митного права. Підтримуємо позицію, що адмiнiстративнo-правoвий механiзм прoтидiї митним правопорушенням – це динамiчна сукупнiсть взаємoпoв’язаних і взаємoдiючих у певнiй пoслiдoвнoстi адмiнiстративнo-правoвих захoдiв та засoбiв, з дoпoмoгoю яких девіантна пoведiнка суб’єктiв зовнішньоекономічної діяльності та громадян привoдиться у вiдпoвiднiсть з вимoгами законодавства. Це цiлий прoцес і сукупнiсть oперацiй, спрямoваних на oрганiзацiю дії та впливу нoрм адмiнiстративнoгo права на волю і пoведiнку суб’єктiв вiднoсин з привoду запoбiгання правoпoрушенням, за дoпoмoгoю якoгo виникають вiдпoвiднi правoвiднoсини [15, c.81]. Висновки. Підсумовуючи зазначимо, що проблематика протидії митним правопорушенням є складною, багаторівневою та актуальною незалежно від рівня економічного розвитку країни. Гармонізація міжнародного митного законодавства та України здійснюється у двох основних векторах: відповідно до базових норм міжнародного права та відповідно до регламентів і стандартів регулювання Європейського Союзу. Наразі перед юридичною спільнотою України постає пласт невідкладних завдань: методологічних, пов’язаних з переосмисленням сучасної специфіки розвитку митно-правових відносин у світі, науково-практичних, зумовлених перебудовою нормативно-правового поля України на європейські стандарти та появою нових схем здійснення митних девіацій, форс-мажорних, що викликані новими викликами та ризиками воєнного стану в нашій державі. 146 1. 2. 3. 4. 5. 6. 7. Електронне наукове видання «Аналітично-порівняльне правознавство» СПИСОК ВИКОРИСТАНИХ ДЖЕРЕЛ: Нота Секретаріату СОТ розглядає вплив кризи в Україні на світову торгівлю та розвиток. URL: https://www.wto.org/english/news_e/ news22_e/devel_08apr22_ e.htm (дата звернення: 04.04.2022). Конвенція про створення Ради Митного Співробітництва. URL: https://zakon.rada.gov.ua/ laws/show/995_184#Text (дата звернення: 06.04.2022). Міжнародна конвенція про спрощення і гармонізацію митних процедур. URL: https:// zakon.rada.gov.ua/laws/show/995_643#Text (дата звернення: 03.04.2022). Міжнародна конвенція про взаємну адміністративну допомогу у відверненні, розслідуванні та припиненні порушень митного законодавства. URL: https://zakon.rada.gov.ua/ laws/show/995_926#Text (дата звернення: 03.04.2022). Про приєднання України до Міжнародної конвенції про взаємну адміністративну допомогу у відверненні, розслідуванні та припиненні порушень митного законодавства та додатків I, II, III, V, VI, VII, VIII до неї). URL : https:// zakon.rada.gov.ua/laws/show/699/2000#Text (дата звернення: 02.04.2022). Міжнародна конвенція про Гармонізовану систему опису та кодування товарів. URL : https:// zakon.rada.gov.ua/laws/show/995_079#Text (дата звернення: 02.04.2022). Конвенція про тимчасове ввезення. URL: h t t p s : / / z a k o n . r a d a . g o v. u a / l a w s / s h o w / 8. 9. 10. 11. 12. 13. 14. 15. 995_472#Text (дата звернення: 02.04.2022). Хома В.О., Царенко В.І. Актуальні проблеми удосконалення системи протидії митним правопорушенням: монографія. Хмельницький: ФОП Мельник А.А., 2018. 194 с. Рамкові стандарти безпеки і полегшення всесвітньої торгівлі ВМО. URL : https://zakon.rada. gov.ua/laws/show/976_003#Text (дата звернення: 07.04.2022). Авторизовані економічні оператори. URL : https://mof.gov.ua/uk/authorized_ economic_ operators-470 (дата звернення: 07.04.2022). Ільченко О.В., Денисенко С.І. Міжнародно-правові аспекти боротьби з корупцією в митних органах: досвід Сінгапуру. Юридичний науковий електронний журнал. 2021. №5. С. 68–72. Бережнюк І.Г. та ін. Митна енциклопедія. Хмельницький : ПП Мельник А.А., 2013. 472 с. Union Customs Code. URL : https:// ec.europa.eu/taxation_customs/customs-4/ union-customs-code_en (дата звернення: 08.04.2022). Регламент (ЄС) 2021/785 Європейського парламенту та Ради від 29 квітня 2021 року «Про створення Програми боротьби з шахрайством Союзу». URL : https://eur-lex.europa.eu/legal-content/ EN/TXT/?uri=CELEX%3A32021R0785&qid = 1522853648735 (дата звернення: 10.04.2022). Переверзєва Г. Г. Удосконалення митно-тарифного регулювання в Україні. Вісн. КрНУ ім. М. Остроградського. 2016. Вип. 4 (75). С. 79–82.
https://openalex.org/W2802942913	https://www.nature.com/articles/s41419-018-0560-7.pdf	English	null	SMPD3 deficiency perturbs neuronal proteostasis and causes progressive cognitive impairment	Cell death and disease	2,018	cc-by	10,048	A R T I C L E O p e n A c c e s s © The Author(s) 2018 OpenAccessThisarticleislicensedunderaCreativeCommonsAttribution4.0InternationalLicense,whichpermitsuse,sharing,adaptation,distributionandreproduction 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. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Abstract Neutral sphingomyelinase smpd3 is most abundantly expressed in neurons of brain. The function of SMPD3 has remained elusive. Here, we report a pathogenetic nexus between absence of SMPD3 in the Golgi compartment (GC) of neurons of the smpd3-/- mouse brain, inhibition of Golgi vesicular protein transport and progressive cognitive impairment. Absence of SMPD3 activity in the Golgi sphingomyelin cycle impedes remodeling of the lipid bilayer, essential for budding and multivesicular body formation. Importantly, we show that inhibition of the Golgi vesicular protein transport causes accumulation of neurotoxic proteins APP, Aβ and phosphorylated Tau, dysproteostasis, unfolded protein response, and apoptosis, which ultimately manifests in progressive cognitive decline, similar to the pathognomonic signatures of familial and sporadic forms of Alzheimer´s disease. This discovery might contribute to the search for other primary pathogenic mechanisms, which link perturbed lipid bilayer structures and protein processing and transport in the neuronal Golgi compartment and neurodegeneration and cognitive deﬁcits. reticulum (ER) and of SMPD3 in the GC in all tissues except brain7. Ligand cell surface receptor mediated acti- vation of nSMases is believed to trigger the breakdown of SM to ceramide, acting as lipid-signaling molecule in multiple cellular signaling pathways ranging from cell growth, differentiation and apoptosis, anti-apoptosis, tumor-suppressive, and anti-proliferative cellular processes to cell senescence8–11. However, new aspects on the role of nSMases arose from the discovery and characterization of two Mg2+-dependent nSMases, SMPD2, and SMPD312–16. SMPD3 is most abundantly expressed in brain. The smpd3- null allellic mouse model has been instrumental in deﬁning the enigmatic systemic and cell-speciﬁc role in extra- neuronal tissues and in brain. Systemic deﬁciency of SMPD3 results in a complex phenotype characterized by impaired secretion of proteohormones from hypothalamic neurosecretory neurons during the postnatal growth phase thereby perturbing the hypothalamus–pituitary growth axis, and resulting in combined pituitary hormone deﬁciency17. SMPD3 deﬁciency perturbs neuronal proteostasis and causes progressive cognitive impairment Wilhelm Stoffel1,2,3, Britta Jenke2, Inga Schmidt-Soltau2, Erika Binczek1, Susanne Brodesser3 and Ina Hammels1 Correspondence: Wilhelm Stoffel (wilhelm.stoffel@uni-koeln.de) 1Laboratory of Molecular Neuroscience, Institute of Biochemistry, University of Cologne, 50931 Cologne, Germany 2CMMC (Centre for Molecular Medicine), University of Cologne, 50931 Cologne, Germany Full list of author information is available at the end of the article Edited by G. Raschellà Stoffel et al. Cell Death and Disease (2018) 9:507 DOI 10.1038/s41419-018-0560-7 Stoffel et al. Cell Death and Disease (2018) 9:507 DOI 10.1038/s41419-018-0560-7 Cell Death & Disease Cell Death & Disease Ofﬁcial journal of the Cell Death Differentiation Association Introduction Sphingomyelin (SM) is a major component of the lipid bilayer of subcellular membranes particularly of mamma- lian central nervous system (CNS). Acid (SMPD1) and neutral sphingomyelinases (sphingomyelin phosphor-dies- terases, SMases) (SMPD2 and SMPD3), hydrolyze SM and release ceramide and phosphoryl-choline. SMases differ in their enzymatic properties, regulation, tissue distribution, and subcellular localization. SM is degraded constitutively by acid SMase (SMPD1) in lysosomes1–4. Genetic defects in smpd1 lead to the fatal human neurovisceral form of Niemann–Pick disease, type A, characterized by lysosomal SM storage in cells of the reticuloendothelial system and neurons. The SMPD1-deﬁcient mouse mutant is a mimicry of NPD type A5,6. SMPD1 activity exceeds non-lysosomal neutral SMPD2 activity (nSMase1) in the endoplasmic Studies in primary chondrocyte culture provided initial mechanistic insight into SMPD3 function18. Chondrocytes of the epiphyseal growth plate highly express SMPD3 Ofﬁcial journal of the Cell Death Differentiation Association Ofﬁcial journal of the Cell Death Differentiation Association Page 2 of 14 Page 2 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 during the postnatal growth phase18–20. Primary chon- drocytes of juvenile smpd3-/- mice revealed retarded transport and secretion of extracellular matrix proteins in the Golgi secretory pathway, impeded proteostasis, and induced ER stress, manifested in growth retardation, and chondrodysplasia. The available data of these studies sug- gested a novel mechanism underlying a basic function of SMPD3 in the canonical Golgi secretory pathway. This notion and the abundant expression of smpd3 in brain motivated further studies of SMPD3 function in the CNS. subcellular localization revealed SMPD3 as key enzyme in the SM cycle of the Golgi compartment (GC) of extra- neuronal tissues18. We ﬁrst assessed the cell type speciﬁc expression of smpd3 in brain, and the subcellular localization of SMPD3. Neurons, oligodendrocytes and astrocytes of p4- mice were isolated by magnetic cell separation (MACS) from brain homogenates and grown in culture. Immunohistochemistry (IHC) detected SMPD3 most abundantly in neurons, isolated by depletion of non- neuronal cells from total brain cell homogenate as described under Materials and Methods, characterized by NF68 (Fig. 1a). Minor expression in oligodendrocytes, separated by antiO4-beads and characterized by immuno- labeling with anti- 2′,3′-cyclic-nucleotide 3′-phospho- diesterase (CNPase) antibody (Fig. 1c), and astrocytes, separated by anti-ACSA-2-tagged beads and identiﬁed by glial-ﬁbrillary acidic protein (GFAP) (Fig. 1e). Progressive depositions of neuronal proteins characterize several age-related neurodegenerative diseases, the most common disorder of which is Alzheimer’s disease (AD). Introduction Major pathognomonic signatures of rare familial AD (FAD) and the majority of late-onset sporadic forms of AD (SAD) are the accumulation of neurotoxic proteins amyloid β (Aβ) and hyperphosphorylated Tau and cognitive decline with the loss of memory21–23. Mutations in FAD have been identiﬁed in the amyloid precursor protein (APP), pre- senilin 1 (PSEN1), and presenilin 2 (PSEN2) genes. Per- turbation of the non-amyloidogenic processing pathway of APP causes the accumulation of mainly neurotoxic Aβ in FAD and the majority of SAD. Although comprehensive insight into the processing has been elaborated from several genetically engineered mouse models24, upstream mole- cular pathogenic mechanisms, leading to neurotoxic end products, are still enigmatic, and the validity of the pathognomonic “amyloid cascade hypothesis”21,25 has been challenged by several caveats26,27. As a ﬁrst step in under- standing the pathogenesis of AD risk, genome-wide asso- ciated studies and exome sequencing have identiﬁed polymorphisms and coding variants of several genes, which are involved in inﬂammatory response, endocytosis, and lipid metabolism. As well as the apoEε4 allele28,29, the neuronal apoE-receptor, apolipoprotein apoJ30,31 and recently phospholipase D332 have been recognized as strong risk factors associated with late-onset SAD33. The role of sphingolipids in neurodegeneration has received increasing attention recently, for review see34. Western blots underscore the prevalent expression of SMPD3 in neurons as well as the purity of the isolated cell types (Fig. 1b, d, f, h). High-resolution microscopy of immune-stained neu- rons colocalizes SMPD3 with Golgi-marker K58 and GALNT2 in the Golgi complex (Fig. 1g, i) and with SMS1 in Golgi of HEK293 cells (Fig. 1j) in addition to coronal sections of cortex (Fig. 2f). We mapped SMPD3 expression in sagittal (Fig. 2a) and coronal sections (Fig. 2b) of adult Cntr brains (6 mo) immuno-histochemically, using anti-SMPD3 and horse- radish peroxidase-tagged secondary antibody. Neurons of cortical layers III and V, of subiculum and CA1–CA3 regions of hippocampus, and of hypothalamus showed highest SMPD3 expression. Cortical external pyramidal cell layer III and internal ganglionic cell layer V, hippocampus, dentate gyrus and nuclei of the hypothalamus of adult brain (6 mo) showed high SMPD3 expression (Fig. 2c, d). Neurons, oligoden- drocytes, and astrocytes were also identiﬁed in adult control (Cntr) mouse brain by double labeling with anti- SMPD3 and NF68, anti-CNPase and anti-GFAP anti- bodies, respectively (Fig. 2g–n). These studies in vivo and in vitro allocated SMPD3 expression predominantly to neurons. Introduction High-resolution confocal microscopy coloca- lized SMPD3 with Golgi-marker K58 and with SMS1 in the neuronal Golgi complex (Fig. 2e, f and Supplementary Fig. 1a). In this study, we describe a molecular pathogenic link between SMPD3 deﬁciency in the neuronal Golgi complex, perturbed secretory pathway and dysproteostasis, thereby eliciting cytotoxicity and AD-like cognitive deﬁciencies. These data provide a direct neuron-based pathogenetic pathway connecting the perturbation of the Golgi lipid bilayer structure and neurodegeneration in the SMPD3- deﬁcient mouse brain. This phenotype suggests the impli- cation of smpd3 as susceptibility gene of AD. Results SMPD3 deﬁciency in hypothalamic neurosecretory neurons perturbs the GHRH/GH/Igf1-axis17. A sig- niﬁcantly expanded lifespan (up to 20%) (Fig. 3a) is caused by postnatal and juvenile Igf1- serum levels, which remain constantly low during adulthood (Fig. 3b)35,36. Neutral sphingomyelinase (SMPD3) expression in CNS is restricted to the Golgi complex of neurons Multi-tissue expression pattern disclosed the most abundant expression of smpd3 in brain14. Studies on the Ofﬁcial journal of the Cell Death Differentiation Association Page 3 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 SMPD3 expression in CNS is most abundant to neurons. Immuno-magnetic cell separation and characterization by IHC of a neurons of ol C57BL/6 mouse brain (p4), labeled with anti-SMPD3 and NF68 antibodies, c oligodendrocytes, labeled with anti-SMPD3 and anti-CNPase odies, e astrocytes, double labeled with anti-SMPD3 and anti-GFAP antibodies. Western blot of isolated cell types, using cell-speciﬁc antibodies: 8, d CNPase, f GFAP, and h SMPD3 in astrocytes (A), neurons (N), and oligodendrocytes (O). Colocalization of g SMPD3 and K58 and of i SMPD3 ALNT2 in Golgi complex of neurons. j Colocalization of SMPD3 and SMS1 in Golgi of HEK293 cells Fig. 1 SMPD3 expression in CNS is most abundant to neurons. Immuno-magnetic cell separation and characterization by IHC of a neurons of control C57BL/6 mouse brain (p4), labeled with anti-SMPD3 and NF68 antibodies, c oligodendrocytes, labeled with anti-SMPD3 and anti-CNPase antibodies, e astrocytes, double labeled with anti-SMPD3 and anti-GFAP antibodies. Western blot of isolated cell types, using cell-speciﬁc antibodies: b NF68, d CNPase, f GFAP, and h SMPD3 in astrocytes (A), neurons (N), and oligodendrocytes (O). Colocalization of g SMPD3 and K58 and of i SMPD3 and GALNT2 in Golgi complex of neurons. j Colocalization of SMPD3 and SMS1 in Golgi of HEK293 cells ression in CNS is most abundant to neurons. Im Ofﬁcial journal of the Cell Death Differentiation Association Page 4 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 Fig. 2 SMPD3 expression in CNS is restricted to neurons predominantly of cortex, hippocampus and hypothalamus. a Sagittal and b coronal sections of adult brains (6 mo) IHC-labeled with anti-SMPD3 and HRP-tagged secondary antibody. Encased areas of cortex (cor), cerebellum (cereb), hippocampus (hip), inner cortex (ico), and N. arcuatus (arc) are magniﬁed. c, d SMPD3 expression in neurons of cortical layers III and V. Colocalization of e SMPD3 and K58 and f SMPD3 and SMS1 in Golgi complex of neurons. Merged images of g, h hippocampus and i, j cortical sections, labeled with anti-SMPD3 and neuron-speciﬁc NF68 antibodies. k, l Minor expression of SMPD3 in CNPase-labeled oligodendrocytes (arrows), and m, n GFAP- labeled astrocytes Fig. 2 SMPD3 expression in CNS is restricted to neurons predominantly of cortex, hippocampus and hypothalamus. Inhibition of APP, Aβ, and pTau- protein transport, perturbation Golgi secretory pathway and apoptosis in SMPD3-deﬁcient neurons Inhibition of APP, Aβ, and pTau- protein transport, perturbation Golgi secretory pathway and apoptosis in SMPD3-deﬁcient neurons The marked increase of APP, Aβ, and pTau in western blot hybridization analysis of 12 mo of age smpd3-/- brain lysates (Fig. 4g, h, j) became also apparent in IHC (Fig. 5a–j). Anti-Aβ-antibodies displayed Aβ-deposition in neurons of cortex of 6-mo-old smpd3-/- mice (Fig. 5a), heavy Aβ-loaded Golgi of cortical neurons (Fig. 5d) and neurons of dentate gyrus (Fig. 5e) of smpd3-/- mice 24 mo of age. APP and intra- and extracellular pTau deposits were observed in neurons of the cortex already in smpd3-/- mice 12 mo of age (Fig. 5g–j). We addressed the impaired proteostasis by quantitative western blot analysis of unfolded protein response using marker pIRE. Elevated pIRE expression strongly indicated the activation of endoplasmatic reticulum stress signaling pathways (Fig. 5k). TUNEL staining was strongly increased in cor- onal and sagittal brain sections of 12-mo-old smpd3-/- compared with Cntr mice (Fig. 5l, m). IHC of coronal brain sections of smpd3-/- mice revealed an age- dependent increased number of AnnexinV-positive neu- rons, (Fig. 5n, o). Neutral sphingomyelinase (SMPD3) expression in CNS is restricted to the Golgi complex of neurons p values of ≤0.05 , ≤0.01 , ≤0.001 ** were considered signiﬁcant Combined HPTLC- MS/MS of the Golgi DAG pool of Cntr mice revealed a a cluster of DAG species, reﬂecting the backbone of Golgi-PC, and another with the major species 18:0/20:4-DAG, released by speciﬁc phospholipase C from phosphatidyl inositol-phosphates in the plasma membrane and ER (Supplementary Fig. 2a–c). (presenilin) in 6- and 12-mo-old Cntr and smpd3-/- mice. Smpd3-/- mice, at age 12 mo, showed signiﬁcantly elevated steady-state mRNA levels of the neuron-speciﬁc marker genes mapt and ttbk, and likewise of plp, whereas app expression remained unchanged (Fig. 4a, b). Cntr and smpd3-/- Golgi contained similar pattern and equal concentrations of very long chain (C20-C26)-Cer, serving as substrates of UDP-glycosyl-transferase for mono- and poly-hexosyl-ceramide synthesis (Fig. 4u–w). The phospholipidomes of the Golgi fraction of Cntr and smpd3-/- brain were similar (Fig. 4r–t, Supplementary Fig. 1d–i). We isolated the Golgi fraction of brain of 12-mo-old Cntr and smpd3-/- mice and their detergent-insoluble membrane (DIM) domains for western blot hybridization and lipidomic analyses. The purity of the Golgi fraction was conﬁrmed by western blot analyses probing the Golgi complex by anti- K58, and plasma membrane contaminants by the anti-Na +/K+-ATPase antibody (Supplementary Fig. 1b, c). Western blot hybridization analysis of protein lysates of mutant Golgi proved the absence of SMPD3 (Fig. 2k). WB of total brain, the Golgi-, and its DIM-fraction of Cntr and smpd3-/- mice revealed the association of APP, Aβ, and of Pen2, the cata- lytic subunit of γ-secretase with the Golgi-DIM-fraction, (Fig. 4c–o). Concentrations of APP and its amyloidogenic Aβ considerably increased in brain of smpd3-/- mice between age 6 mo (Fig. 4c, d) and 12 mo (Fig. 4g, h). Aβ of Golgi fraction in smpd3-/- mice of 12 mo is increased in com- parison with Cntr (Fig. 4n). Similar to the elevated steady- state, RNA concentration of neuron-speciﬁc markers mapt and ttbk (Fig. 4a, b), pTau protein concentration in the lysate of total brain of smpd3-/- mice signiﬁcantly increased between 6 and 12 mo of age (Fig. 4f, j) and is elevated in Golgi fraction of 12-mo-old smpd3-/- mice in comparison with Cntr (Fig. 4o). APP and Pen2 concentrations remained unchanged in Golgi and DIMs of Golgi of 12-mo-old Cntr and smpd3-/- mice (Fig. 4l, m, p, q). Neutral sphingomyelinase (SMPD3) expression in CNS is restricted to the Golgi complex of neurons a Sagittal and b coronal sections of adult brains (6 mo) IHC-labeled with anti-SMPD3 and HRP-tagged secondary antibody. Encased areas of cortex (cor), cerebellum (cereb), hippocampus (hip), inner cortex (ico), and N. arcuatus (arc) are magniﬁed. c, d SMPD3 expression in neurons of cortical layers III and V. Colocalization of e SMPD3 and K58 and f SMPD3 and SMS1 in Golgi complex of neurons. Merged images of g, h hippocampus and i, j cortical sections, labeled with anti-SMPD3 and neuron-speciﬁc NF68 antibodies. k, l Minor expression of SMPD3 in CNPase-labeled oligodendrocytes (arrows), and m, n GFAP- labeled astrocytes speciﬁc glast1 (glutamate-aspartate transporter 1) and eaac1 (excitatory amino acid carrier 1) and AD-related app (amyloid precursor protein), mapt (microtubule associated tau potein), ttbk1 (tau tubulin kinase1), and psen 1 and 2 speciﬁc glast1 (glutamate-aspartate transporter 1) and eaac1 (excitatory amino acid carrier 1) and AD-related app (amyloid precursor protein), mapt (microtubule associated tau potein), ttbk1 (tau tubulin kinase1), and psen 1 and 2 We investigated age-dependent gene expression in brain by real-time PCR, using primer pairs (Table S1), of neuron-speciﬁc syntaxin, synapsin and synaptophysin, oligodendrocyte-speciﬁc plp (proteolipidprotein), astrocyte- Ofﬁcial journal of the Cell Death Differentiation Association Page 5 of 14 Page 5 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 Fig. 3 Longevity and reduced IGF-1 serum levels of SMPD3- deﬁcient mice. (a Lifespan of Cntr and smpd3-/- mice (n = 20). b IGF- 1 serum levels of Cntr and smpd3-/- mice (n = 10). p values of ≤0.05 , ≤0.01 , ≤0.001 ** were considered signiﬁcant neurons are the scaffold of SMPD3 and SMS118. High- performance thin-layer chromatography (HPTLC)- MS/ MS species analysis of the SM- and Cer-pools of Golgi membranes of Cntr and smpd3-/- mice revealed that >80% of SM-pool consisted of the C18-SM species (Fig. 4w and Supplementary Fig. 2d, e). Only the Golgi Cer-pool of Cntr neurons contained the distinct C18-Cer fraction (Fig. 4r), which is absent in SMPD3-deﬁcient Golgi membranes (Fig. 4r–t). Absence of SMPD3 in the Golgi–SM cycle leads to restricted production of C18- Cer, which is the acceptor substrate of the SMS1- catalyzed group transfer of phosphoryl-choline from phosphatidyl choline, and of DAG (Fig. 4r–w). Fig. 3 Longevity and reduced IGF-1 serum levels of SMPD3- deﬁcient mice. (a Lifespan of Cntr and smpd3-/- mice (n = 20). b IGF- 1 serum levels of Cntr and smpd3-/- mice (n = 10). Impaired Golgi secretory pathway in smpd3-/- neurons triggers progressive AD-like cognitive decline mice, age 3, 6 and 12 mo in a battery of tasks, motor activity, and coordination in the Rota-rod- and beam walk and locomotor activity I the open ﬁeld test. Despite ske- letal deformations of the smpd3-/- mutant17, their per- formance in the Rota-rod test matched that of Cntr cohorts up to age 6 mo, but signiﬁcantly declined at age 12 mo (Fig. 6a). Performance in the beam walk task remained unchanged (Fig. 6b). Open ﬁeld ambulatory movement (distance run/5 min) was reduced in 3-mo-old smpd3-/- mice, and further decreased by 25% at age 6 mo and 25% at age 12 mo, similar to Cntr mice (Fig. 6c). mice, age 3, 6 and 12 mo in a battery of tasks, motor activity, and coordination in the Rota-rod- and beam walk and locomotor activity I the open ﬁeld test. Despite ske- letal deformations of the smpd3-/- mutant17, their per- formance in the Rota-rod test matched that of Cntr cohorts up to age 6 mo, but signiﬁcantly declined at age 12 mo (Fig. 6a). Performance in the beam walk task remained unchanged (Fig. 6b). Open ﬁeld ambulatory movement (distance run/5 min) was reduced in 3-mo-old smpd3-/- mice, and further decreased by 25% at age 6 mo and 25% at age 12 mo, similar to Cntr mice (Fig. 6c). Cognitive defects were explored by running a behavioral test battery38,39 with two groups of 12-mo-old Cntr and smpd3-/- mice. The performance of smpd3-/- mice in the T-maze at age 12 mo showed severe deﬁcits in the working memory (Fig. 6d), and in the radial arm water maze (RAWM) the impaired reference memory task (Fig. 6e). Cell fractionation, biochemical, and immuno- histochemical studies identiﬁed DIMs of the neuronal GC as the subcellular platform for the concerted action of integral polytopic SMPD3 and SMS1 and their substrates and products SM, Cer, PC, and DAG in the SM cycle. SM and cholesterol are major constituents segregated into DIM domains, lo—(liquid ordered) DIMS of the Golgi lipid bilayer and embedded in ld (liquid disordered) domains (van Meer et al. 2008). Sub-fractionation of the Golgi membrane stacks of peripheral tissue allocated SMPD3 and SMS1 to DIM domains18. The association of SMS1 with the trans Golgi network has been reported earlier40. Explorative motivation was measured in the open ﬁeld task with the video-tracking technique, (Fig. 6f). Impaired Golgi secretory pathway in smpd3-/- neurons triggers progressive AD-like cognitive decline Unlike the Cntr cohorts, 6- and 12-mo-old smpd3-/- mice barely overcame the initial thigmotaxis phase to enter the second phase of exploring the environment for cues of spatial representation. Furthermore, we assessed the anxiety- related behavior of Cntr and smpd3-/- mice, age 3, 6, and 12 mo, in the elevated plus-maze. smpd3-/- mice devel- oped a strongly reduced anxiogenic phenotype with age (Fig. 6g, h), indicated by loss of proclivity toward dark and loss of avoidance of open space and height. smpd3-/- mice, age 6 mo spent twice and age 12 mo and four times longer time-intervals in the open arm and entered the closed, safe arm four times more frequently than Cntr mice. Lipidomic analysis of the GC-membrane lipid bilayer revealed that the Golgi–SM fraction consists >80% of the stearoyl-SM species (C18-SM), asymmetrically distributed in the luminal leaﬂet of the Golgi lipid bilayer, and two pools of Cer species in Cntr Golgi, a distinct C18-Cer fraction and a long chain Cer fraction. C18-Cer reﬂects the backbone of Golgi-C18-SM. This C18-Cer fraction is absent in SMPD3-deﬁcient Golgi membranes. A second pool consists of very long chain (>C20) de novo synthe- sized ceramides (Cer) in Cntr and smpd3-/- Golgi com- plexes only being utilized for glycosphingolipid synthesis. Impaired Golgi secretory pathway in smpd3-/- neurons triggers progressive AD-like cognitive decline SM/cholesterol-enriched DIM domains are embedded in the phospholipid-rich domains of the luminal leaﬂet of Golgi membranes37. DIMs of the Golgi complex of Cntr We next tested the age-dependent performance of cohorts of gender and weight matched Cntr and smpd3-/- Ofﬁcial journal of the Cell Death Differentiation Association Page 6 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 Fig. 4 (See legend on next page.) ( ) g Fig. 4 (See legend on next page.) Ofﬁcial journal of the Cell Death Differentiation Association Page 7 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 (see ﬁgure on previous page) Fig. 4 Quantitative analysis of gene- and protein-expression and lipidomes of Cntr and smpd3-/- Golgi complexes. Real-time-PCR of total brain RNA of cohorts (n = 5) of a 6- and b 12-mo-old Cntr and smpd3-/- mice, using neuron-speciﬁc markers syntaxin, synapsin, and synaptophysin, of AD-related genes mapt, ttbk1, app, and psen1 and 2; oligodendrocyte-speciﬁc markers plp and mbp; and astrocytes-speciﬁc markers eaac1 and glast1, mean ± SD. Representative WB of lysates of c–f 6 and g–j 12-mo-old total brain using anti -APP, -Aβ, -Tau1, and -pTau, of k–o. Golgi complex using anti –SMPD3, -APP, -Pen2, -Aβ, and -pTau of p, q Golgi DIMs, using anti-APP and -Pen2 antibodies of Cntr and SMPD3-/- mice (n = 3). r HPTLC of Cer and DAG pools of the Golgi lipid extract. Markers: m1 18:0-Cer, m2 1,2-18:1/18:1-DAG s Quantiﬁcation of Cer species, t “C18-Cer” species (arrow in r), u identical ceramide-species in >20-ceramides. v HPTLC of Golgi lipids. w MS/MS-analysis of SM species. C cholesterol, CMH ceramide monohexoside, PE phosphatidyl ethanolamine, PS phosphatidyl serine, PC phosphatidyl choline, SM sphingomyelin Our mechanistic studies demanded the unambiguous assignment of the cellular and subcellular localization of SMPD3 in CNS. IHC of isolated neurons, oligoden- drocytes, and astrocytes of p4-mouse brain, using anti- SMPD3 antibodies and cell-speciﬁc markers convincingly demonstrated that anti-NF68 positive neurons are the main scaffold of brain SMPD3, minor expression is found in anti-CNPase-positive oligodendrocytes and anti-GFAP- positive astrocytes. IHC-mapping of neuronal SMPD3 expression in sagittal and coronal sections of brain uncovered strongest expression of SMPD3 in neurons in cortical layers III and V, of CA1 and III of hippocampus and dentate gyrus, cerebellum, and hypothalamus. Ofﬁcial journal of the Cell Death Differentiation Association Discussion Here, we describe a novel pathogenetic process involving neuronal SMPD3 deﬁciency in CNS of the smpd3-/- mouse mutant. We unveiled a molecular link between SMPD3 deﬁciency in the SM cycle of the neuronal Golgi complex, impeded remodeling of the lipid bilayer of the Golgi membrane, essential for budding, vesicle formation and protein transport of APP, Aβ and pTau, dysproteostasis, causing neurodegeneration and AD—like cognitive decline. These observations suggest a central role of C18-SM and C18-Cer in the SM cycle conﬁned to DIMS of Golgi membranes. The domain structure compartmentalizes cellular processes. SMPD3 and SMS1 are embedded in the C18-SM-enriched lipid bilayer of DIMS, the platform of the Golgi–SM cycle. Ofﬁcial journal of the Cell Death Differentiation Association Page 8 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 g. 5 Age-dependent neuronal dysproteostasis in smpd3-/- mice. Merged epiﬂuorescence images of brain sections of 6- and 24-mo-old Cntr d smpd3-/- mice revealed age-dependent heavy accumulation of Aβ and pTau in neurons of smpd3-/- mice with age, Aβ in a, d cortex and b, e ntate gyrus and c, f pTau in dentate gyrus and hippocampus, using anti-Aβ/anti-PLP and anti-pTau/anti-Synapsin-antibodies. Confocal images of j cortex and dentate gyrus of 12- and 24-mo-old Cntr and smpd3-/- brain sections double labeled with anti-APP and anti-pTau (AT8) antibody, i, j larged images of encased areas. k Representative WB of pIRE in 12-mo-old smpd3-/- mice (n = 4). l, m TUNEL assay in section of cortex of 12- and -mo-old Cntr and smpd3-/- mouse brains, counter-stained with haemalum Fig. 5 Age-dependent neuronal dysproteostasis in smpd3-/- mice. Merged epiﬂuorescence images of brain sections of 6- and 24-mo-old Cntr and smpd3-/- mice revealed age-dependent heavy accumulation of Aβ and pTau in neurons of smpd3-/- mice with age, Aβ in a, d cortex and b, e Fig. 5 Age-dependent neuronal dysproteostasis in smpd3-/- mice. Merged epiﬂuorescence images of brain sections of 6- and 24-mo-old Cntr and smpd3-/- mice revealed age-dependent heavy accumulation of Aβ and pTau in neurons of smpd3-/- mice with age, Aβ in a, d cortex and b, e dentate gyrus and c, f pTau in dentate gyrus and hippocampus, using anti-Aβ/anti-PLP and anti-pTau/anti-Synapsin-antibodies. Confocal images of g–j cortex and dentate gyrus of 12- and 24-mo-old Cntr and smpd3-/- brain sections double labeled with anti-APP and anti-pTau (AT8) antibody, i, j enlarged images of encased areas. Discussion k Representative WB of pIRE in 12-mo-old smpd3-/- mice (n = 4). l, m TUNEL assay in section of cortex of 12- and 20-mo-old Cntr and smpd3-/- mouse brains, counter-stained with haemalum Fig. 5 Age-dependent neuronal dysproteostasis in smpd3-/- mice. Merged epiﬂuorescence images of brain sections of 6- and 24-mo-old Cntr and smpd3-/- mice revealed age-dependent heavy accumulation of Aβ and pTau in neurons of smpd3-/- mice with age, Aβ in a, d cortex and b, e dentate gyrus and c, f pTau in dentate gyrus and hippocampus, using anti-Aβ/anti-PLP and anti-pTau/anti-Synapsin-antibodies. Confocal images of g–j cortex and dentate gyrus of 12- and 24-mo-old Cntr and smpd3-/- brain sections double labeled with anti-APP and anti-pTau (AT8) antibody, i, j enlarged images of encased areas. k Representative WB of pIRE in 12-mo-old smpd3-/- mice (n = 4). l, m TUNEL assay in section of cortex of 12- and 20-mo-old Cntr and smpd3-/- mouse brains, counter-stained with haemalum SMPD3-hydrolysis of C18-SM in the Golgi luminal leaﬂet triggers a reaction-cascade: the polar head group of SM is detached by SMPD3, thereby removing the “umbrella” effect of SM in the SM/C complex with the dissociation of C18-ceramide, which aggregates owing to its physical properties to a lamellar Cer-barrier boarding the lo-DIM-domain in the luminal leaﬂet, Fig. 7. This concept is in agreement with kinetic studies in model systems consisting of phase-separated ternary SM/ to its physical properties to a lamellar Cer-barrier boarding the lo-DIM-domain in the luminal leaﬂet, Fig. 7. This concept is in agreement with kinetic studies in model systems consisting of phase-separated ternary SM/ Ofﬁcial journal of the Cell Death Differentiation Association Page 9 of 14 Page 9 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 PC/C domains embedded in a liquid disordered PC lipid bilayer, when treated with bacterial sphingomyelinase41,42. C18 SM h b b d t bi d t iﬁ i Fig. 6 Age-dependent decline of motor activity, coordination, and cognitive tasks of SMPD3-deﬁcient mice. Assessment of motor activity and coordination by a Rota-rod task, b beam walk task, and c open ﬁeld test, working memory by d T-maze, reference memory by e RAWM of Cntr and smpd3-/- mice, age 12 mo, f thigmotaxis in open ﬁeld task and anxiety response by g, h elevated plus-maze. N = 8, mean ± SD. Discussion p values of ≤0.05 , ≤0.01 , ≤0.001 ** were considered signiﬁcant the anterograde Golgi-ER-carriers, which are recruited in the Ld-domain43. Remarkably, this C18-SM signature sequence is also present in the Psen2-subunit of γ- secretase, and in Aβ-protein44. We located Pen2 and APP in DIMs of the Golgi membranes of Cntr neurons. Palmitoylated beta- secretase1 (beta-site amyloid precursor protein cleaving enzyme 1), which initiates the amyloidogenic pathway and the γ-secretase tetrameric complex of PSN1, nicastrin, APH1, and Pen2 are also segregated to lo- domains for APP processing (Kitazume et al. 2001)45. Absence of SMPD3 in smpd3-/- Golgi DIMS leads to partitioning of SMS1 between lo- and ld-domains18. SMS1-catalyzed group transfer of phosphoryl-choline from phosphatidyl choline to Cer releases DAG. Unlike Cer, DAG has membrane fusogenic properties lipid, freely equilibrates between the bilayer, disintegrating locally the bilayer structure. In addition the molecular shape of DAG imposes a large negative curvature upon the membrane, which facilitates constriction of the membrane into a neck-structure46 and triggers budding initiating vesicle formation and cargo transport in the Golgi secretory pathway. Whether this lipid-driven bud formation requires ESCRT (endosome sorting complex required for transport) remains to be answered. p ) SMPD3 deﬁciency prohibits the “shuttle” function of Cer in the dynamic modulations of the Golgi lipid bilayer for budding, a fundamental process in neuronal protein transport, documented in the age-dependent accumula- tion of APP, Aβ and pTau in neurons of cortex, hippo- campus, dentate gyrus and hypothalamus of 6- and 12- mo-old SMPD3-/- mice. Impeded proteostasis and accu- mulation of neurotoxic end products in neurons activated ER- stress signaling. Elevated expression of pIRE in wes- tern blot based quantitative analysis and TUNEL assay of coronal brain sections of smpd3-/- mice (12 mo) assessed increased age-dependent apoptosis of smpd3-/- neurons. Studies on the sorting of membrane into vesicles of multivesicular bodies in PLP-egfp-overexpressing oligo- dendrocytes transfected with nSMase2 and in giant vesicles I experiments in vitro when treated with bacterial SMase released the ceramide47. Ceramides were held responsible for the transfer of exosome-associated domains into the lumen of the endosome in an ESCRT independent manner. Motor activity and coordination of SMPD3-deﬁcient and Cntr mice were evaluated in the Rota-rod and hor- izontal beam walk crossing tasks. Discussion Performance in the Rota-rod task indicated impaired motor activity and coordination in male and female smpd3-/- mice compared with Cntr mice at age 12 mo, but similar performance in the beam walk task. SMPD3 deﬁciency prohibits the “shuttle” function of Cer in the dynamic modulations of the Golgi lipid bilayer for budding, a fundamental process in neuronal protein transport, documented in the age-dependent accumula- tion of APP, Aβ and pTau in neurons of cortex, hippo- campus, dentate gyrus and hypothalamus of 6- and 12- mo-old SMPD3-/- mice. Impeded proteostasis and accu- mulation of neurotoxic end products in neurons activated ER- stress signaling. Elevated expression of pIRE in wes- tern blot based quantitative analysis and TUNEL assay of coronal brain sections of smpd3-/- mice (12 mo) assessed increased age-dependent apoptosis of smpd3-/- neurons. Studies on the sorting of membrane into vesicles of multivesicular bodies in PLP-egfp-overexpressing oligo- dendrocytes transfected with nSMase2 and in giant vesicles I experiments in vitro when treated with bacterial SMase released the ceramide47. Ceramides were held responsible for the transfer of exosome-associated domains into the lumen of the endosome in an ESCRT independent manner. Fig. 6 Age-dependent decline of motor activity, coordination, and cognitive tasks of SMPD3-deﬁcient mice. Assessment of motor activity and coordination by a Rota-rod task, b beam walk task, and c open ﬁeld test, working memory by d T-maze, reference memory by e RAWM of Cntr and smpd3-/- mice, age 12 mo, f thigmotaxis in open ﬁeld task and anxiety response by g, h elevated plus-maze. N = 8, mean ± SD. p values of ≤0.05 , ≤0.01 , ≤0.001 ** were considered signiﬁcant Motor activity and coordination of SMPD3-deﬁcient and Cntr mice were evaluated in the Rota-rod and hor- izontal beam walk crossing tasks. Performance in the Rota-rod task indicated impaired motor activity and coordination in male and female smpd3-/- mice compared with Cntr mice at age 12 mo, but similar performance in the beam walk task. PC/C domains embedded in a liquid disordered PC lipid bilayer, when treated with bacterial sphingomyelinase41,42. C18-SM has been observed to bind to a speciﬁc amino acid sequence (signature) in the transmembrane domain of p24, a subunit of the COPI complex of COPI vesicles, PC/C domains embedded in a liquid disordered PC lipid bilayer, when treated with bacterial sphingomyelinase41,42. PC/C domains embedded in a liquid disordered PC lipid bilayer, when treated with bacterial sphingomyelinase41,42. Discussion C18-SM has been observed to bind to a speciﬁc amino acid sequence (signature) in the transmembrane domain of p24, a subunit of the COPI complex of COPI vesicles, Cognitive defects were explored by running a behavioral test battery38,39 with two groups of gender and weight matched 3, 6- and 12-mo-old Cntr and smpd3-/- mice. Ofﬁcial journal of the Cell Death Differentiation Association Page 10 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 Fig. 7 Suggested mechanistic model of lipid-driven bilayer remodeling in bud formation. C cholesterol, SM sphingomyelin, PI phosphatidyl inositol, PS phosphatidyl serine, DAG diacyl glycerol, PC phosphatidyl choline, PE phosphatidyl ethanolamine, GSL glycosphingolipid, Cer ceramide, ESCRT endosomal sorting complexes required for transport nistic model of lipid-driven bilayer remodeling in bud formation. C cholesterol, SM sphingomyelin, PI phosphatidyl erine, DAG diacyl glycerol, PC phosphatidyl choline, PE phosphatidyl ethanolamine, GSL glycosphingolipid, Cer ceramide, complexes required for transport Fig. 7 Suggested mechanistic model of lipid-driven bilayer remodeling in bud formation. C cholesterol, SM sphingomyelin, PI phosphatidyl inositol, PS phosphatidyl serine, DAG diacyl glycerol, PC phosphatidyl choline, PE phosphatidyl ethanolamine, GSL glycosphingolipid, Cer ceramide, ESCRT endosomal sorting complexes required for transport Working and reference memory were probed in the T- maze and the RAWM. smpd3-/- mice developed severe deﬁcits in the working memory- and reference memory tasks at age 12 mo. T- maze and RAWM have proven to be sensitive and reliable tasks in uncovering the moderately impaired cognitive functions of AD mouse models48 asso- ciated with dysfunction of hippocampus, and in detecting memory deﬁcits in APP transgenic mouse mutants49,50. lipid bilayer modiﬁcation essential for lipid-driven bud- ding and vesicle formation, and triggers dysproteostasis, apoptosis, and neurodegeneration. Further studies are needed to support the suggested function of smpd3 as Alzheimer´s disease susceptibility gene. We anticipate this discovery might contribute to the search for other primary pathogenic mechanisms in remodeling the lipid bilayer of neuronal Golgi membranes essential for protein processing and transport triggering age-dependent neurodegeneration and cognitive deﬁcits. Explorative motivation was measured in the open ﬁeld task. Unlike the 6- and 12-mo-old Cntr cohorts, smpd3-/- mice barely overcame the initial thigmotaxis phase to enter the second phase, exploring the environment for cues of spatial representation. Materials and Methods Mouse lines The smpd3-/- mouse line was developed in this laboratory, and after 10 back-crossings maintained on a C57Bl/6 background17. Control (Cntr) mice were obtained from heterozygous SMPD3-/-×C57Bl/6 cross- ings. Mice were genotyped by PCR analysis of tail DNA. Cohorts of gender, age and weight matched Cntr and smpd3-/- mice were used in this study. Animals were kept under pathogen-free conditions. The ARRIVE Guide- lines55 have been followed in the animal studies reported in this manuscript. Smpd3-/- mice developed a strongly reduced anxiogenic phenotype with age, as indicated in the elevated plus- maze by loss of proclivity toward dark and loss of avoid- ance of open space and height. A recent study51 to assess the function of SMPD3 in AD-related pathology, using a 5xfad;fro-/- double mutant mouse52, is incompatible with the study reported here, using the unbiased smpd3-/- mutant, as substantiated by the high postnatal lethality of the fro-/- mutant compared with the extended lifespan (>20%) of the smpd3-/- mouse mutant and by the contrasting skeletal phenotypes of the fro-/- and smpd3-/- mutants19. Ofﬁcial journal of the Cell Death Differentiation Association Motor activity, behavioral, and cognitive performance Mice (n = 8) were accommodated in home cages under constant lightning, noise levels and temperature 2 h prior to testing. Equipment was cleaned with 70% ethanol between test sessions. Dissociated neuronal cultures The low-serum insulin-like growth factor 1 concentra- tion in juvenile smpd3 mice17 persisted in adult smpd3 mice and resulted in an extended lifespan (>20%), which is owing to the hypothylamic pituitary growth axis53,54. Brains of p5 pups of Cntr mice were dissociated to single-cell suspensions using the Neural Tissue Dissocia- tion Kit, Miltenyi Biotec, following the manufacturer’s protocol. Mouse neurons were isolated by depletion of non-neuronal cells, following the manufacturer’s protocol. Non-neuronal cells were magnetically labeled with biotin- conjugated monoclonal antibodies speciﬁc for non- Our study provides a novel concept of the pathogenic basis of the nexus between smpd3-gene ablation, which perturbs the SM metabolism in the neuronal GC and the Ofﬁcial journal of the Cell Death Differentiation Association Page 11 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 Triton X-100 and stained with hematoxylin–eosin for transmission microscopy or immuno-stained with afﬁnity puriﬁed rabbit polyclonal or monoclonal antibodies. A Zeiss microscope Axio ImagerM1 and the AxioVision Imaging Software were used for light- and ﬂuorescence microscopy The TCS SP8X confocal microscope (Leica Microsystems), equipped with a PL Apo 63 × /1.40 Oil CS2 objective, white light laser (NKT Photonics) and HyD detectors was used for confocal microscopy. Triton X-100 and stained with hematoxylin–eosin for transmission microscopy or immuno-stained with afﬁnity puriﬁed rabbit polyclonal or monoclonal antibodies. A Zeiss microscope Axio ImagerM1 and the AxioVision Imaging Software were used for light- and ﬂuorescence microscopy The TCS SP8X confocal microscope (Leica Microsystems), equipped with a PL Apo 63 × /1.40 Oil CS2 objective, white light laser (NKT Photonics) and HyD detectors was used for confocal microscopy. neuronal cells to be retained by anti-biotin monoclonal antibodies conjugated MicroBeads packed into MACS Column. Unlabeled neuronal cells were collected in the ﬂow-through and cultured in six well Costar-plates. Astrocytes were isolated using the anti-ACSA-2 (astrocyte cell surface antigen-2) MicroBead Kit, Milte- nyi Biotec. Fc receptors were blocked with FcR blocking reagent and then the ACSA-2+ cells were magnetically labeled with anti-ACSA-2 and separated according to the manufacturer’s instructions. neuronal cells to be retained by anti-biotin monoclonal antibodies conjugated MicroBeads packed into MACS Column. Unlabeled neuronal cells were collected in the ﬂow-through and cultured in six well Costar-plates. Astrocytes were isolated using the anti-ACSA-2 (astrocyte cell surface antigen-2) MicroBead Kit, Milte- nyi Biotec. Dissociated neuronal cultures Fc receptors were blocked with FcR blocking reagent and then the ACSA-2+ cells were magnetically labeled with anti-ACSA-2 and separated according to the manufacturer’s instructions. Motor activity and coordination tests Rotor rod test Mice were placed on a rotating rod (3.5 cm diameter) 50 cm from the ﬂoor in lanes 12 cm wide in the same direction, whereas the bar is rotating (Brooks and Dun- nett, 2009. The speed of rotation was set at constant 16 rpm during the 60 s trial. Latencies (seconds) of main- taining balance on the bar were recorded. Cell fractionation Cell fractionation, isolation of Golgi fractions56, and Triton X-100 insoluble DIMs57 from Golgi fractions were performed following established procedures. Real-time PCR Neuronal apoptosis was visualized using the TUNEL- Apoptosis Detection Kit #17-141, EMD Millipore, Merck. RNA from Cntr and smpd3-/- brains of littermates (n = 5) was isolated using Trizol, Invitrogen. In total, 10 µg of total RNA was reverse-transcribed using a transcriptase kit, life technologies18. Primer pairs used in quantitative PCR reactions are listed in Table S1. Hgprt was used as internal standard. Quantitative PCR reactions were per- formed with the ABI Prism 7900HT employing a 96well format and the Fast SYBR Green Master Mix, Applied Biosystems, following the manufacturer´s protocol. Data analysis was performed using the 2-ΔΔCt method. Lipidome analysis Total lipids were extracted from brain or subcellular fractions for phospholipidomic analysis as outlined before59. MS/MS of complex lipids in lipid extracts are described in detail under Supplemental Information. Acid and neutral sphingomyelinase assays Radioactive assay for acid and neutral sphingomyelinase activity was performed as described previously5,13. Protein analysis by western blotting Mouse brain proteins were isolated and solubilized for western blot hybridization. Protein aliquots (50–100 µg) were analyzed using the NuPAGE western Blot system (Invitrogen). The following antibodies were used: anti- Amyloid Precursor Protein, 1:20000, abcam, #ab32136; anti-α tubulin, 1:12000, Santa Cruz, #sc5546; anti- Caveolin, 1:5000, BD Transduction Laboratories #610059; anti-pTau, 1:1000, abcam, #ab926767; anti-pIRE, 1:1000, abcam, #ab48187; anti-Pen2, 1:500, abcam, #ab18189, GalNAc-T2, ThermoFisher PA 5-2141. Horseradish per- oxidase conjugated secondary antibodies were used for detection with the ECL system. Signals were quantiﬁed by densitometry using the IMAGE J2X program. Histology and IHC Cntr and smpd3-/- mice, 6 and 24 mo of age, were perfused from the left ventricle with 25 ml phosphate- buffered saline (PBS) and with 50 ml PBS-buffered 4% paraformaldehyde for cryo- and parafﬁn embedding and processing for light‐ and immunoﬂuorescence micro- scopy. Coronal and sagittal sections (5 µm) were per- meabilized with 0.5% Triton X-100/PBS at 4 °C, blocked with 3% bovine serum albumin /PBS and treated with respective antibody dilutions in tris-buffered saline, sup- plemented with 5% non-fat dry milk at 4 °C over-night. The following antibodies and dilutions were used: afﬁnity puriﬁed anti-PLP, 1:10058; anti-SMPD3, 1:200,14; anti- Amyloid 1:500 clone WO2, Merk Millipore, MABN10; anti-CNP, 1:250, abcam #ab6319; anti-Synapsin I, 1:300, abcam, #ab64581; anti-Tau1 1:100. Anti-pTau, 1:500, ThermoFisher Scientiﬁc, #MN1020; anti-GFAP, 1:400, Sigma-Aldrich, #G3893, anti-NF68, 1:300, Sigma-Aldrich, #N5139. Ofﬁcial journal of the Cell Death Differentiation Association Beam walk maze Sections were washed with PBS/0,5% Triton X-100, incu- bated with Cy3-conjugated second IgG antibody (Jackson Immuno Research) for 1 h at 37 °C, washed with PBS/0.5% The beam walk maze used in this study followed established protocols60. In brief, mice were trained to walk along an 80 cm long and 1.2 cm wide wooden beam, Ofﬁcial journal of the Cell Death Differentiation Association Page 12 of 14 Page 12 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 bridging two platforms (50 cm above the bench). Mice were placed on the beam and allowed to. The time to full- length traverse the beam and to reach the goal box reach, dropping off the bridge, and the time until the drop off were recorded. cm50. The water level of the pool was adjusted to ~ 2 cm above an invisible 10-cm circular platform placed in the back of one arm ~ 7 cm inside the back wall. The pool was located in the center of a dimmed room. Black poster walls with geometric visual cues bordered the pool horse-shoe like. The RAWM protocol consisted of a 2-day staggered training schedule for cohorts of six mice, alternating through the trials over day 1 and day 2 of the test. During block 1 (six trials) and block 2 (six trials), mice were trained to identify the platform location by alternating between a visible and a hidden platform in the goal arm, with three hidden and three visible platform trials. Block 3 consisted of three trials with a hidden platform. For day 2, mice were tested in three blocks of ﬁve trials each (15 total trials), with only the hidden escape platform employed, forcing the mice to use a spatial strategy to identify the goal arm location. Data are presented as average errors per block. Only errors during the hidden platform trials represent the spatial memory component and are included in the analysis of the RAWM task. Statistical analysis h The homemade T-maze apparatus was adapted in size to mice, as described before62. Mice were food deprived until their body weight attained 85%. They were placed at the base of the T and allowed to choose one of the goal arms. Only one of the arms contained a food reward at the end of the arm. Mice were trained 7–10 days to enter the arm with the food reward, eating the reward and starting again at the base. The food reward was always placed on the side of the maze not entered by the mouse on the previous trial. Spatial working memory was measured by recording the number of correct entries into baited arms and number of re-entries into an unbaited arm. The cutoff time of each trial was 60 s. In the cross-sectional study design, experimental groups consisted of two genotypes, the Cntr and the smpd3-/- genotype at ages 3, 6, 12, 20, and 24 mo. Sizes of animal cohorts were six genders and weight matched mice, or otherwise listed under respective ﬁgures and Material and Methods. Results are expressed as mean ± SD. Statistical signiﬁcance of differences between individual experi- mental groups was compared by the unpaired t-test using Graph Pad Quick Calcs: t-test calculator. p values of ≤ 0.05 , ≤0.01 , ≤0.001 ** were considered signiﬁcant Conﬂict of Interest The authors declare no conﬂict of interest. Open ﬁeld test The Open ﬁeld test was used to assess spontaneous ambulatory movement and spontaneous motor activity as well as the degree of exploring the novel environment resembling anxiety-like behaviors61. Distance and speed in ambulatory movement were measured. Mice were placed in an open ﬁeld maze chamber 50 cm long × 50 cm wide. Activity was recorded using video-tracking software (Stassen, Department of Physiology, University of Cologne). Data were collected continually for ﬁve minutes and the distance traveled (cm) and velocity (cm/second), recorded, and scored automatically. Distance traveled and movement speeds assessed ambulatory movement. Thig- motaxis was documented by video-tracking. Acknowledgements We gratefully acknowledge the support of this work by the Center of Molecular Medicine, University of Cologne, CECAD (Cluster of Excellence, Cellular Stress Response in Aging-Related Diseases), University of Cologne and Deutsche Forschungsgemeinschaft (Sto32/38-2). Author details 1 1Laboratory of Molecular Neuroscience, Institute of Biochemistry, University of Cologne, 50931 Cologne, Germany. 2CMMC (Centre for Molecular Medicine), University of Cologne, 50931 Cologne, Germany. 3CECAD (Cluster of Excellence: Cellular Stress Responses in Aging-Associated Diseases), University of Cologne, 50931 Cologne, Germany Elevated Plus-Maze Anxiety-related behavior was assessed by the passive avoidance test. A homemade elevated plus-maze con- sisting of two enclosed arms and two open arms, elevated 50 cm above the bench was used63. To assess anxiety- related behavior, deﬁned as the degree to which the subject avoided the open arms (perceived unsafe arms) of the maze and preferring the closed arms (perceived safe arm) of the maze. Each mouse was placed in the center of the maze The latency and number of entries spent in the preferred closed arms (“safe arm”) and open arms (“unsafe arm”) were recorded over 5 min to deﬁne the degree of anxiety-related behavior. One session of 5 min per mouse was carried out. Author contributions d d h W.S. designed research, B.J., E.B., I.H., I.S.S. performed experiments, S.B. analyzed data, W.S. and I.H. wrote the manuscript. Conﬂict of Interest The authors declare no conﬂict of interest. Conﬂict of Interest The authors declare no conﬂict of interest. References Effects of ceramide on liquid- ordered domains investigated by simultaneous AFM and FCS. Biophys. J. 90, 4500–4508 (2006). 11. Hannun, Y. A. & Obeid, L. M. Many ceramides. J. Biol. Chem. 286, 27855–27862 (2011). 12. Hofmann, K. & Dixit, V. M. Ceramide in apoptosis--does it really matter? Trends Biochem. Sci. 23, 374–377 (1998). 42. Ira & Johnston, L. J. Sphingomyelinase generation of ceramide promotes clustering of nanoscale domains in supported bilayer membranes. Biochim Biophys. Acta 1778, 185–197 (2008). 13. Tomiuk, S., Hofmann, K., Nix, M., Zumbansen, M. & Stoffel, W. Cloned mam- malian neutral sphingomyelinase: functions in sphingolipid signaling? Proc. Natl. Acad. Sci. USA 95, 3638–3643 (1998). 43. Brügger, B. et al. Evidence for Segregation of Sphingomyelin and Cholesterol during Formation of Copi-Coated Vesicles. J. Cell Biol. 151, 507–518 (2000). 14. Hofmann, K., Tomiuk, S., Wolff, G. & Stoffel, W. Cloning and characterization of the mammalian brain-speciﬁc, Mg2+-dependent neutral sphingomyelinase. Proc. Natl. Acad. Sci. USA 97, 5895–5900 (2000). 44. Contreras, F. X. et al. Molecular recognition of a single sphingolipid species by a protein’s transmembrane domain. Nature 481, 525–529 (2012). . Co t e as, . . et a . o ecu a ecog t o o a s g e sp go p d spec es by a protein’s transmembrane domain. Nature 481, 525–529 (2012). 45. Vetrivel, K. S. et al. Association of gamma-secretase with lipid rafts in post-Golgi and endosome membranes. J. Biol. Chem. 279, 44945–44954 (2004). 45. Vetrivel, K. S. et al. Association of gamma-secretase with lipid rafts in post-Golgi and endosome membranes. J. Biol. Chem. 279, 44945–44954 (2004). 15. Tomiuk, S., Zumbansen, M. & Stoffel, W. Characterization and Subcellular Localization of Murine and Human Magnesium-dependent Neutral Sphin- gomyelinase. J. Biol. Chem. 275, 5710–5717 (2000). 46. Shemesh, T., Luini, A., Malhotra, V., Burger, K. N. & Kozlov, M. M. Preﬁssion constriction of Golgi tubular carriers driven by local lipid metabolism: a the- oretical model. Biophys. J. 85, 3813–3827 (2003). 16. Zumbansen, M. & Stoffel, W. Neutral sphingomyelinase 1 deﬁciency in the mouse causes no lipid storage disease. Mol. Cell Biol. 22, 3633–3638 (2002). 47. Trajkovic, K. et al. Ceramide triggers budding of exosome vesicles into mul- tivesicular endosomes. Science 319, 1244–1247 (2008). 17. Stoffel, W., Jenke, B., Block, B., Zumbansen, M. & Koebke, J. Neutral sphingo- myelinase 2 (smpd3) in the control of postnatal growth and development. Proc. Natl. Acad. Sci. USA 102, 4554–4559 (2005). 48. References Acid sphingomyelinase-deﬁcient mice mimic the neurovisceral form of human lysosomal storage disease (Niemann-Pick dis- ease). Cell 81, 1053–1061 (1995). 5. Otterbach, B. & Stoffel, W. Acid sphingomyelinase-deﬁcient mice mimic the neurovisceral form of human lysosomal storage disease (Niemann-Pick dis- ease). Cell 81, 1053–1061 (1995). 35. van Heemst, D. et al. Reduced insulin/IGF-1 signalling and human longevity. Aging Cell 4, 79–85 (2005). 36. Selman, C. et al. Evidence for lifespan extension and delayed age-related bio- markers in insulin receptor substrate 1 null mice. FASEB J. 22, 807–818 (2008). 6. Horinouchi, K. et al. Acid sphingomyelinase deﬁcient mice: a model of types A and B Niemann- Pick disease. Nat. Genet. 10, 288–293 (1995). 6. Horinouchi, K. et al. Acid sphingomyelinase deﬁcient mice: a model of types A and B Niemann- Pick disease. Nat. Genet. 10, 288–293 (1995). 37. van Meer, G., Voelker, D. R. & Feigenson, G. W. Membrane lipids: where they are and how they behave. Nat. Rev. Mol. Cell Biol. 9, 112–124 (2008). 7. Spence, M. W. & Burgess, J. K. Acid and neutral sphingomyelinases of rat brain. Activity in developing brain and regional distribution in adult brain. J. Neu- rochem. 30, 917–919 (1978). 7. Spence, M. W. & Burgess, J. K. Acid and neutral sphingomyelinases of rat brain. Activity in developing brain and regional distribution in adult brain. J. Neu- rochem. 30, 917–919 (1978). 38. Brooks, S. P. & Dunnett, S. B. Tests to assess motor phenotype in mice: a user’s guide. Nat. Rev. Neurosci. 10, 519–529 (2009). 8. Kolesnick, R. N. & Kronke, M. Regulation of ceramide production and apop- tosis. Annu. Rev. Physiol. 60, 643–665 (1998). 39. Bryan, K. J., Lee, H., Perry, G., Smith, M. A. & Casadesus, G. Transgenic Mouse Models of Alzheimer’s Disease: Behavioral Testing and Considerations Methods of Behavior Analysis in Neuroscience (Taylor & Francis; Francis Group, LLC., Boca Raton FL, 2009). 9. Hannun, Y. A. Functions of ceramide in coordinating cellular responses to stress. Science 274, 1855–1859 (1996). 10. Hannun, Y. A. & Obeid, L. M. The Ceramide-centric universe of lipid-mediated cell regulation: stress encounters of the lipid kind. J. Biol. Chem. 277, 25847–25850 (2002). 40. Huitema, K., van den Dikkenberg, J., Brouwers, J. F. & Holthuis, J. C. Identiﬁcation of a family of animal sphingomyelin synthases. EMBO J. 23, 33–44 (2004). 41. Chiantia, S., Kahya, N., Ries, J. & Schwille, P. Publisher's note In brief, the maze consisted of six arms 160 cm in dia- meter with arm length 30 cm and circular swim area of 40 Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. Ofﬁcial journal of the Cell Death Differentiation Association Page 13 of 14 Page 13 of 14 Page 13 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 Supplementary Information accompanies this paper at https://doi.org/ 10.1038/s41419-018-0560-7. References Webster, S. J., Bachstetter, A. D., Nelson, P. T., Schmitt, F. A. & Van Eldik, L. J. Using mice to model Alzheimer’s dementia: an overview of the clinical disease and the preclinical behavioral changes in 10 mouse models. Front. Genet. 5, 88 (2014). 18. Stoffel, W. et al. Neutral sphingomyelinase (SMPD3) deﬁciency disrupts the Golgi secretory pathway and causes growth inhibition. Cell Death Dis. 7, e2488 (2016). 49. Morgan, D. et al. A[beta] peptide vaccination prevents memory loss in an animal model of Alzheimer’s disease. Nature 408, 982–985 (2000). 19. Stoffel, W. et al. Neutral sphingomyelinase (SMPD3) deﬁciency causes a novel form of chondrodysplasia and dwarﬁsm that is rescued by Col2A1-driven smpd3 transgene expression. Am. J. Pathol. 171, 153–161 (2007). 50. Alamed, J., Wilcock, D. M., Diamond, D. M., Gordon, M. N. & Morgan, D. Two- day radial-arm water maze learning and memory task; robust resolution of amyloid-related memory deﬁcits in transgenic mice. Nat. Protoc. 1, 1671–1679 (2006). p g p 20. Li, J. et al. Smpd3 expression in both chondrocytes and osteoblasts is required 20. Li, J. et al. Smpd3 expression in both chondrocytes and osteoblasts is required for normal endochondral bone development. Mol. Cell Biol. 36, 2282–2299 (2016). for normal endochondral bone development. Mol. Cell Biol. 36, 2282–2299 (2016). 51. Dinkins, M. B. et al. Neutral sphingomyelinase-2 deﬁciency ameliorates Alz- heimer’s disease pathology and improves cognition in the 5XFAD mouse. J. Neurosci. 36, 8653–8667 (2016). 21. Hardy, J. & Selkoe, D. J. The amyloid hypothesis of Alzheimer’s disease: pro- gress and problems on the road to therapeutics. Science 297, 353–356 (2002). 22. Piaceri, I., Nacmias, B. & Sorbi, S. Genetics of familial and sporadic Alzheimer’s disease. Front Biosci. 5, 167–177 (2013). 52. Oakley, H. et al. Intraneuronal beta-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with ﬁve familial Alzheimer’s disease mutations: potential factors in amyloid plaque formation. J. Neurosci. 26, 10129–10140 (2006). 23. Ashe, K. H. & Aguzzi, A. Prions, prionoids and pathogenic proteins in Alzheimer disease. Prion 7, 55–59 (2013). 24. Goate, A. et al. Segregation of a missense mutation in the amyloid precursor protein gene with familial Alzheimer’s disease. Nature 349, 704–706 (1991). 53. Butler, A. A. & Le Roith, D. Control of growth by the somatropic axis: growth hormone and the insulin-like growth factors have related and independent roles. Annu. Rev. Physiol. 63, 141–164 (2001). 25. Small, S. A. & Duff, K. Supplementary Information accompanies this paper at https://doi.org/ 10.1038/s41419-018-0560-7. 26. Perry, G., Nunomura, A., Raina, A. K. & Smith, M. A. Amyloid-beta junkies. Lancet (Lond., Engl.) 355, 757 (2000). 27. Castellani, R. J. et al. Neuropathology of Alzheimer disease: pa 27. Castellani, R. J. et al. Neuropathology of Alzheimer disease: pathognomonic but not pathogenic. Acta Neuropathol. 111, 503–509 (2006). Received: 22 March 2018 Accepted: 5 April 2018 but not pathogenic. Acta Neuropathol. 111, 503–509 (2006). Received: 22 March 2018 Accepted: 5 April 2018 p g p 28. Corder, E. H. et al. Gene dose of apolipoprotein E type 4 allele and the risk of 28. Corder, E. H. et al. Gene dose of apolipoprotein E type 4 allele 28. Corder, E. H. et al. Gene dose of apolipoprotein E type 4 allele and the risk o Alzheimer’s disease in late onset families. Science 261, 921–923 (1993). 28. Corder, E. H. et al. Gene dose of apolipoprotein E type 4 allele and the Alzheimer’s disease in late onset families. Science 261, 921–923 (1993). Alzheimer’s disease in late onset families. Science 261, 921–923 (1993). 29. Strittmatter, W. J. et al. Apolipoprotein E: high-avidity binding to beta-amyloid and increased frequency of type 4 allele in late-onset familial Alzheimer dis- ease. Proc. Natl. Acad. Sci. USA 90, 1977–1981 (1993). References 1. Kanfer, J. N., Young, O. M., Shapiro, D. & Brady, R. O. The metabolism of sphingomyelin. I. Puriﬁcation and properties of a sphingomyelin-cleaving enzyme from rat liver tissue. J. Biol. Chem. 241, 1081–1084 (1966). 1. Kanfer, J. N., Young, O. M., Shapiro, D. & Brady, R. O. The metabolism of sphingomyelin. I. Puriﬁcation and properties of a sphingomyelin-cleaving enzyme from rat liver tissue. J. Biol. Chem. 241, 1081–1084 (1966). 30. Harold, D. et al. Genome-wide association study identiﬁes variants at CLU and PICALM d h Al h ’ d N G ( ) 30. Harold, D. et al. Genome-wide association study identiﬁes variants at CLU and PICALM associated with Alzheimer’s disease. Nat. Genet. 41, 1088–1093 (2009). PICALM associated with Alzheimer’s disease. Nat. Genet. 41, 1088–1093 31. Holtzman, D. M., Herz, J. & Bu, G. Apolipoprotein E and apolipoprotein E receptors: normal biology and roles in Alzheimer disease. Cold Spring Harb. Perspect. Med. 2, a006312 (2012). y 2. Quintern, L. E. et al. Acid sphingomyelinase from human urine: puriﬁcation and characterization. Biochim. Biophys. Acta 922, 323–336 (1987). 2. Quintern, L. E. et al. Acid sphingomyelinase from human urine: puriﬁcation and characterization. Biochim. Biophys. Acta 922, 323–336 (1987). 3. Quintern, L. E. et al. Isolation of cDNA clones encoding human acid sphin- gomyelinase: occurrence of alternatively processed transcripts. Embo J. 8, 2469–2473 (1989). 3. Quintern, L. E. et al. Isolation of cDNA clones encoding human acid sphin- gomyelinase: occurrence of alternatively processed transcripts. Embo J. 8, 2469–2473 (1989). 32. Cruchaga, C. et al. Rare coding variants in the phospholipase D3 gene confer risk for Alzheimer’s disease. Nature 505, 550–554 (2014). 33. Karch, C. M. & Goate, A. M. Alzheimer’s disease risk genes and mec disease pathogenesis. Biol. Psychiatry 77, 43–51 (2015). 33. Karch, C. M. & Goate, A. M. Alzheimer’s disease risk genes 4. Newrzella, D. & Stoffel, W. Molecular cloning of the acid sphingomyelinase of the mouse and the organization and complete nucleotide sequence of the gene. Biol. Chem. Hoppe Seyler 373, 1233–1238 (1992). 4. Newrzella, D. & Stoffel, W. Molecular cloning of the acid sphingomyelinase of the mouse and the organization and complete nucleotide sequence of the gene. Biol. Chem. Hoppe Seyler 373, 1233–1238 (1992). disease pathogenesis. Biol. Psychiatry 77, 43–51 (2015). 34. Haughey, N. J. Sphingolipids in neurodegeneration. Neuromolecular Med. 12, 301–305 (2010). 5. Otterbach, B. & Stoffel, W. Ofﬁcial journal of the Cell Death Differentiation Association Stoffel et al. Cell Death and Disease (2018) 9:507 References Linking Abeta and tau in late-onset Alzheimer’s disease: a dual pathway hypothesis. Neuron 60, 534–542 (2008). Ofﬁcial journal of the Cell Death Differentiation Association Page 14 of 14 Ofﬁcial journal of the Cell Death Differentiation Association 57. Fiedler, K., Kobayashi, T., Kurzchalia, T. V. & Simons, K. Glycosphingolipid-enri- ched, detergent-insoluble complexes in protein sorting in epithelial cells. Biochemistry 32, 6365–6373 (1993). Page 14 of 14 Stoffel et al. Cell Death and Disease (2018) 9:507 59. Stoffel, W. et al. Obesity resistance and deregulation of lipogenesis in Delta6-fatty acid desaturase (FADS2) deﬁciency. EMBO Rep. 15, 110–120 (2014). 54. Bartke, A. Growth hormone and aging. Endocrine 8, 103–108 (1998). 55. Kilkenny, C., Browne, W., Cuthill, I. C., Emerson, M. & Altman, D. G. Animal research: reporting in vivo experiments: the ARRIVE guidelines. Br. J. Pharmacol. 160, 1577–1579 (2010). 60. Carter, R. J., Morton, J. & Dunnett, S. B. Motor coordination and balance in rodents. Curr. Protoc. Neurosci. Chapter 8, Unit 8.12 (2001). 56. Graham, J. M. Isolation of Golgi membranes from tissues and cells by differ- ential and density gradient centrifugation. Curr. Protoc. Cell Biol. Chapter 3, Unit 3.9 (2001). 61. Webster, S. J., Bachstetter, A. D. & Van Eldik, L. J. Comprehensive behavioral characterization of an APP/PS-1 double knock-in mouse model of Alzheimer’s disease. Alzheimers Res. Ther. 5, 28 (2013). 57. Fiedler, K., Kobayashi, T., Kurzchalia, T. V. & Simons, K. Glycosphingolipid-enri- ched, detergent-insoluble complexes in protein sorting in epithelial cells. Biochemistry 32, 6365–6373 (1993). 62. Wenk, G. L. Assessment of spatial memory using the T maze. Curr. Protoc. Neurosci. Chapter 8, Unit8.5B (2001). 63. Walf, A. A. & Frye, C. A. The use of the elevated plus-maze as an assay of anxiety-related behavior in rodents. Nat. Protoc. 2, 322–328 (2007). 58. Uschkureit, T., Sporkel, O., Stracke, J., Bussow, H. & Stoffel, W. Early onset of axonal degeneration in double (plp-/-mag-/-) and hypomyelinosis in triple (plp-/-mbp-/-mag-/-) mutant mice. J. Neurosci. 20, 5225–5233 (2000). 63. Walf, A. A. & Frye, C. A. The use of the elevated plus-maze as an assay of anxiety-related behavior in rodents. Nat. Protoc. 2, 322–328 (2007). Ofﬁcial journal of the Cell Death Differentiation Association
https://openalex.org/W2794747700	https://europepmc.org/articles/pmc5898764?pdf=render	English	null	Prospective evaluation of a rapid diagnostic test for Trypanosoma brucei gambiense infection developed using recombinant antigens	PLoS neglected tropical diseases	2,018	cc-by	11,992	RESEARCH ARTICLE Prospective evaluation of a rapid diagnostic test for Trypanosoma brucei gambiense infection developed using recombinant antigens Crispin Lumbala1,2☯, Sylvain Bie´ler3☯, Simon Kayembe3, Jacquies Makabuza1, Stefano Ongarello3, Joseph Mathu Ndung’u3 Crispin Lumbala1,2☯, Sylvain Bie´ler3☯, Simon Kayembe3, Jacquies Makabuza1, Stefano Ongarello3, Joseph Mathu Ndung’u3 1 Programme National de Lutte contre la Trypanosomiase Humaine Africaine, Kinshasa, Democratic Republic of the Congo, 2 Global Health Institute, University of Antwerp, Antwerp, Belgium, 3 Foundation for Innovative New Diagnostics (FIND), Geneva, Switzerland a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 ☯These authors contributed equally to this work. * sylvain.bieler@finddx.org ☯These authors contributed equally to this work. * sylvain bieler@finddx org ☯These authors contributed equally to this work. * sylvain.bieler@finddx.org Abstract OPEN ACCESS Citation: Lumbala C, Bie´ler S, Kayembe S, Makabuza J, Ongarello S, Ndung’u JM (2018) Prospective evaluation of a rapid diagnostic test for Trypanosoma brucei gambiense infection developed using recombinant antigens. PLoS Negl Trop Dis 12(3): e0006386. https://doi.org/10.1371/ journal.pntd.0006386 Methodology/Principal findings The objective of this study was to evaluate the accuracy of a new RDT developed using recombinant antigens (SD BIOLINE HAT 2.0), in comparison with an RDT produced using native antigens (SD BIOLINE HAT) and CATT. A total of 57,632 individuals were screened in the Democratic Republic of the Congo, either passively at 10 health centres, or actively by 5 mobile teams, and 260 HAT cases were confirmed by parasitology. The highest sensitivity was achieved with the SD BIOLINE HAT 2.0 (71.2%), followed by CATT (62.5%) and the SD BIOLINE HAT (59.0%). The most specific test was CATT (99.2%), while the specificity of the SD BIOLINE HAT and SD BIOLINE HAT 2.0 were 98.9% and 98.1%, respectively. Sensitivity of the tests was lower than previously reported, as they identified cases from par- tially overlapping sub-populations. All three tests were significantly more sensitive in passive than in active screening. Combining two or three tests resulted in a markedly increased sen- sitivity: When the SD BIOLINE HAT was combined with the SD BIOLINE HAT 2.0, sensitivity reached 98.4% in passive and 83.0% in active screening. Copyright: © 2018 Lumbala et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Background Diagnosis and treatment are central elements of strategies to control Trypanosoma brucei gambiense human African trypanosomiasis (HAT). Serological screening is a key entry point in diagnostic algorithms. The Card Agglutination Test for Trypanosomiasis (CATT) has been the most widely used screening test for decades, despite a number of practical limitations that were partially addressed by the introduction of rapid diagnostic tests (RDTs). However, current RDTs are manufactured using native antigens, which are challenging to produce. Editor: Guy Caljon, Universiteit Antwerpen, BELGIUM Received: September 25, 2017 Accepted: March 13, 2018 Published: March 28, 2018 Copyright: © 2018 Lumbala et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Editor: Guy Caljon, Universiteit Antwerpen, BELGIUM Received: September 25, 2017 Accepted: March 13, 2018 Published: March 28, 2018 Copyright: © 2018 Lumbala et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Editor: Guy Caljon, Universiteit Antwerpen, BELGIUM Prospective evaluation of HAT RDT developed using recombinant antigens Author summary Sleeping sickness, or human African trypanosomiasis (HAT), is a neglected tropical dis- ease that represents a risk to more than seventy million people in Sub-Saharan Africa. Most cases are caused by infection with Trypanosoma brucei gambiense. Diagnosis of HAT relies on the identification of suspected cases by serological methods, which include recently developed rapid diagnostic tests (RDTs). Current RDTs are produced using native antigens that are purified from live parasites in a laborious and dangerous process. The objective of this study was to evaluate the performance of a new RDT made using recombinant antigens, by screening people in fifteen endemic sites in the Demo- cratic Republic of the Congo. The new RDT was found to be more sensitive than, and as specific as, the reference RDT made using native antigens. It was also more sensitive than CATT, a serological test that has been widely used for decades. While one third of HAT cases were correctly diagnosed by all tests, the other cases were only identified by one or two of the tests. In order to enhance case detection and accelerate elimination of HAT, there may be a need to explore diagnostic strategies that combine two or more screening tests. Conclusions/Significance Competing interests: The authors have declared that no competing interests exist. The recombinant antigen-based RDT was more sensitive than, and as specific as, the SD BIOLINE HAT. It was as sensitive as, but slightly less specific than CATT. While the practi- cality and cost-effectiveness of algorithms including several screening tests would need to be investigated, using two or more tests appears to enhance sensitivity of diagnostic algo- rithms, although some decrease in specificity is observed as well. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Funding: This work was funded by the Bill and Melinda Gates Foundation (http://www. gatesfoundation.org/), grant OPP1033712 (JMN). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. gatesfoundation.org/), grant OPP1033712 (JMN). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 1 / 20 Prospective evaluation of HAT RDT developed using recombinant antigens settings, the sensitivity and specificity of CATT have also been reported as being problematic [6]. In an effort to address the shortcomings of CATT, two rapid diagnostic tests (RDTs) that detect host antibodies have recently been developed, the HAT Sero-K-SeT manufactured by Coris BioConcept (Belgium), and the SD BIOLINE HAT, hereinafter referred to as “RDT1”, produced by Alere/Standard Diagnostics (SD, South Korea), which include the same two anti- gens, VSG LiTat 1.3 and VSG LiTat 1.5. These RDTs were evaluated in retrospective studies, with very promising performance results [7,8]. Evaluation of a prototype of the RDT1 in a pro- spective study in three endemic countries, Angola, the Democratic Republic of the Congo (DRC) and the Central African Republic, showed that the sensitivity of the RDT was not differ- ent from the sensitivity of CATT, while its specificity was 1.3% lower [9]. A prospective study using the HAT Sero-K-SeT also reported excellent performance [10]. A comparison of both RDTs in an independent study using stored plasma samples collected in Guinea and Coˆte d’Ivoire concluded that there was no difference in diagnostic accuracy between the two tests [11]. The RDTs have now been introduced in multiple HAT endemic countries, where they are being used in HAT elimination programmes. However, production of the native antigens used in the manufacture of the RDTs remains a challenge, as it relies on a labor-intensive, costly and risky process that involves inoculating rats with human-infective trypanosomes. To address this challenge, and to improve standardization and quality of manufacturing, a new RDT that is produced exclusively using recombinant antigens, the SD HAT BIOLINE 2.0 (“RDT2”), has been developed in a partnership facilitated by the Foundation for Innovative New Diagnostics (FIND). The primary objective of this study was to evaluate the diagnostic accuracy of RDT2 in a multi-centric, prospective study in the DRC, and to demonstrate its non-inferiority to RDT1. As a secondary objective, the accuracy of RDT2 was compared to that of CATT. Introduction Human African trypanosomiasis (HAT) is a vector-borne, neglected tropical disease, which puts 70 million people living in sub-Saharan African countries at risk [1]. The most common form of the disease is caused by infection with the protozoan parasite Trypanosoma brucei gambiense (g-HAT), which in 2015, accounted for more than 97% of all reported HAT cases [2]. Patients progress from an early disease stage that is characterized by the presence of try- panosomes in the blood and lymphatic system, to a late stage that is associated with the inva- sion of the central nervous system by parasites [3]. If left undiagnosed and untreated, the disease is generally fatal, although asymptomatic cases and others that progress spontaneously to apparently pathogen-free status have been reported [4]. Identification of serological suspects is the main entry point into diagnostic algorithms for g-HAT. The card agglutination test for trypanosomiasis (CATT/T.b. gambiense) has been the most commonly used screening test for g-HAT. It detects antibodies using a sus- pension of purified, fixed and stained bloodstream-form trypanosomes expressing LiTat 1.3 variant surface glycoprotein (VSG), a predominant variant antigen of T.b. gambiense [5]. While CATT has played a central role in the control of HAT, its large-scale implementation for passive screening in health facilities in remote locations has been limited due to opera- tional challenges such as the need for an agitator, electricity and refrigeration. In some 2 / 20 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 Prospective evaluation of HAT RDT developed using recombinant antigens Table 1. Study sites and the corresponding numbers of HAT cases and controls that were enrolled. Site name Site type HAT cases Controls Nkara Fixed facility (HS) 41 1,698 Masi-Manimba Fixed facility (HGR) 31 1,521 Masamuna Fixed facility (CS) 18 2,041 Kwamouth Fixed facility (HGR) 8 784 Kitoy Fixed facility (HS) 8 951 Bagata Fixed facility (HGR) 5 420 Yasa Bonga Fixed facility (HGR) 5 761 Bangumi Fixed facility (CS) 3 321 Bandundu Fixed facility (CDTC) 3 1,184 Bandundu Fixed facility (HGR) 0 934 Kwamouth Mobile team 45 7,326 Mushie Mobile team 39 7,279 Bandundu Mobile team 20 19,457 Idiofa Mobile team 21 7,737 Mokala Mobile team 13 3,855 All sites 260 56,269 HS: "Hoˆpital Secondaire"; HGR: “Hoˆpital Ge´ne´ral de Re´fe´rence”; CS: “Centre de Sante´”; CDTC: “Centre de De´pistage, Traitement et Controˆle”. https://doi org/10 1371/journal pntd 0006386 t001 Table 1. Study sites and the corresponding numbers of HAT cases and controls that were enrolled. HS: "Hoˆpital Secondaire"; HGR: “Hoˆpital Ge´ne´ral de Re´fe´rence”; CS: “Centre de Sante´”; CDTC: “Centre de De´pistage, Traitement et Controˆle”. Enrolment of participants Study participants were enrolled from 6 June 2015 to 5 January 2016 in the Bandundu Prov- ince of the DRC by passive screening in ten health facilities, and by active screening using five mobile teams of the Programme National de Lutte contre la Trypanosomiase Humaine Afri- caine (PNLTHA) of the DRC (Table 1). In the health facilities, participants were enrolled among patients presenting themselves or referred from other health facilities after suspicion of HAT, and among relatives who accompanied patients. During active screening, anybody who presented to the mobile team was eligible for enrolment in the study. Study sites were visited by an external monitor prior to commencement of the study to verify that they were ade- quately prepared and personnel properly trained, and during the study to verify that the proto- col was being adhered to. HAT cases were defined as subjects in whom trypanosomes were demonstrated by microscopy in either lymph node aspirate, blood or cerebrospinal fluid (CSF). All positive parasitology results were verified by the site supervisor. Cases were classi- fied as early stage when no trypanosomes were observed in their CSF, and the CSF white cell count was lower than or equal to 5 cells/μL, while those with trypanosomes in the CSF and/or a cell count above 5 cells/μL were classified as late stage [12]. Controls were subjects living in the same areas as cases, with no known history of HAT infection, and who were either negative with all three screening tests, or who were positive with one or several screening tests, but in whom no parasites were detected in any body fluid. Clinical signs and symptoms were not considered exclusion criteria for controls. 3 / 20 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 Prospective evaluation of HAT RDT developed using recombinant antigens teams and to keep the study design as simple as possible, CATT was only performed on whole blood, and not on diluted plasma. Similarly, the trypanolysis test was not performed during this study, as this would have required additional resources to collect and transport samples for analysis, which at that time could not be performed in DRC. y In both active and passive enrolment, any subject who was positive with at least one of the screening tests, or who showed symptoms highly suggestive of HAT, was eligible for immedi- ate enrolment in the subsequent parasitological work-up. Written informed consent was sought from these subjects prior to enrolment. Any individual who declined to participate in the study was managed according to the standard procedures of the PNLTHA. Individuals who were negative to all three screening tests and who had no symptoms highly suggestive of HAT were not investigated further. Persons with palpable cervical lymph nodes had a lymph node aspirate taken and examined for motile parasites by bright field microscopy. A sample of 5 ml of venous blood was collected from each participant in a heparinized tube. Three hundred μL of blood was used to perform the capillary tube centrifugation (CTC) test (4 capil- lary tubes of approximately 65–70 μL) [15]. If the result of CTC was negative, 500 μL of whole blood was used to perform the mini anion exchange centrifugation technique (mAECT-wb) [16] and the remaining volume of blood (4.2 ml) was centrifuged to perform mAECT on buffy coat (mAECT-bc) as described by Camara et al. [17]. Since the mAECT-bc procedure had only been evaluated in one study in DRC, we took advantage of this study to collect some addi- tional performance data to compare it with mAECT-wb, even though mAECT was only per- formed on a subset of cases. A lumbar puncture was performed on all HAT cases confirmed by any of the parasitological methods, as well as on other participants with clinical signs that were strongly suggestive of HAT, according to routine procedures. Parasitological examination of CSF was done using the modified single centrifugation technique [16]. The technicians who performed the tests were employees of the PNLTHA, with experience in performing routine parasitological tests for detection of trypanosomes. Training of personnel of mobile teams and fixed health facilities included how to perform, read and interpret results of the RDTs, the study protocol and related SOPs, completion of CRFs and data management. Any positive or doubtful parasitology result was verified and confirmed by the site supervisor. Participants with any missing screening test or parasitology results were excluded from the study. All the HAT cases that were identified during the study were treated according to national guidelines. Two levels of blinding were adopted. During the initial screening of participants using blood from a finger prick, three health workers were each responsible for performing one of the three screening tests. The health workers operated independently (but used blood from the same finger prick), without exchanging results (first level of blinding), and did not have access to any clinical information. A supervisor was responsible for collecting results of the tests and deciding whether or not to collect venous blood for parasitological tests. Samples of venous blood were labelled with blinding codes by the supervisor (second level of blinding). The same codes were used to identify all samples collected from the participants (i.e. blood, lymph node aspirate, CSF) and constituted the anonymisation process that was maintained throughout the entire study. Tests performed The RDT2 (SD, South Korea) is an immuno-chromatographic test for qualitative detection of antibodies of all isotypes (IgG, IgA and IgM). It includes a nitrocellulose membrane strip with two test regions (T1 and T2) that are pre-coated with two recombinant antigens. T1 is coated with Invariant Surface Glycoprotein 65–1 (ISG65) expressed in Escherichia coli [13] and T2 with the N-terminal domain of Variant Surface Glycoprotein LiTat 1.5 (VSG LiTat 1.5) pro- duced using a Baculovirus expression system. A procedural control line (C) is also included. The test is stable for at least 24 months at 40˚C, or 5 weeks at 55˚C. The test is performed in the same way as RDT1, as described by Lumbala et al. [14]. In summary, a sample of 20 μl of whole blood is taken from a finger prick and transferred into a sample well using a disposable plastic capillary tube, and 4 drops (approximately 120 μl) of test diluent are then added. The sample flows along the membrane by capillarity, passing through the test regions T1 and T2. Results are read after 15 to 20 minutes by comparing the intensity of the test lines against a col- our chart provided by the manufacturer. A result is considered positive when the control line C and either one or both T1 and T2 test lines are visible (regardless of their intensity), negative when only the C line is observed, and invalid if the C line is not observed. In active screening, all participants found positive with a HAT screening test were also tested for malaria using an RDT (SD BIOLINE Ag P.f), while in passive screening, all participants were tested with a malaria RDT (S3 Table). However, results of malaria RDTs were only recorded for subjects who were eligible for enrolment (see below). Those who tested positive for malaria were exam- ined, and if necessary, treated in line with national guidelines. Three screening tests (CATT, RDT1 and RDT2) were performed on finger-prick blood from each subject who presented to mobile teams, any subject who presented to a health facil- ity with symptoms indicative of HAT, and accompanying individuals who consented to partic- ipate in the study. The results of screening tests were read by two independent laboratory technicians or nurses, and the results recorded separately. To avoid overburdening study 4 / 20 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 Prospective evaluation of HAT RDT developed using recombinant antigens generated (one reading per patient and per test) and used to calculate the performance metrics. This process was repeated (2,000 iterations per metric) to generate an empirical distribution of values for each metric, from which it was possible to derive values for the sample mean and 95% confidence intervals as bootstrapped percentiles. Estimates of sensitivity and specificity were calculated for each screening test, on the overall data, and stratified by disease stage and by screening method (i.e. active and passive screening). The diagnostic performance of each antigen in the RDTs was also calculated. Sensitivity and specificity were defined as the percentage of HAT cases that were found positive and the per- centage of controls that were found negative, respectively. Accuracy was assessed by calculat- ing Youden’s index [19]. To evaluate the agreement between readers, Cohen’s Kappa factor was calculated. The statistical analysis was performed in the R statistical environment (version 3.2.3). Sample size calculation The sample size was calculated to demonstrate non-inferiority of the sensitivity and specificity of RDT2 in comparison to RDT1. Based on the sensitivity of RDT1 of 92.0% that was reported by Lumbala et al. [14], using a non-inferiority margin of 8%, a confidence level of 5% and a power of 80%, the required number of HAT cases was calculated to be at least 143. Based on the same report, the expected specificity of RDT1 was 97.1%. Using a non-inferiority margin of 1%, a confidence level of 5% and a power of 90%, it was calculated that a minimum of 4,775 controls would be needed [20]. Based on the expected prevalence of HAT in the study area, the minimum number of subjects estimated to be screened in order to enrol 143 cases was 44,700. Ethical considerations The study received ethical clearance from the School of Public Health of the University of Kin- shasa (authorization number ESP/CE/012/2015). Participants provided written informed con- sent before being enrolled in the study. For children below 18 years, consent was provided by a parent or guardian. All individuals who presented at study sites during the period of enrolment and consented to being screened were eligible. Those who presented for screening but did not wish to participate in the study were screened according to the procedures of the PNLTHA. All participants’ samples were blinded and further analysed anonymously. Data management and statistical analysis Participant information and test results were recorded at study sites on paper case report forms, which were transferred to PNLTHA in Kinshasa for double data entry using a web- based clinical data management platform (VisionForm). Since two independent readings were available for each test and each sample, an approach based on bootstrapping resampling [18] was adopted: At each iteration, a random sequence of readings from the available data was 5 / 20 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 Results A total of 260 HAT cases and 56,269 controls were enrolled after screening 56,942 people. 413 individuals could not be included in the study because they did not provide informed consent (Fig 1). A total of 138 (53%) cases and 45,654 controls were enrolled by active screening, while 122 (47%) cases and 10,615 controls were enrolled by passive screening. Among cases, the early stage to late stage ratio was 4.3 in active screening and 0.53 in passive screening. The HAT prevalence was 0.30% in active and 1.13% in passive screening. On average, 255 persons were tested per day by each mobile team. The estimates of sensitivity, specificity and accuracy of the RDT2, RDT1 and CATT tests in active screening, passive screening, and active and passive screening combined are shown in Fig 2. When the results of active and passive screening were combined, the sensitivity of the three screening tests was unexpectedly low. While RDT2 detected 71.2% [CI: 65.7%; 76.6%] of the HAT cases, CATT detected only 62.5% [CI: 56.2%; 68.4%] and RDT1 only 59.0% [CI: 53.0%; 64.6%] of the cases. Sensitivity was particularly low in active screening, with only 54.8% [CI: 46.8%; 63.5%], 51.8% [CI: 43.1%; 59.9%] and 49.2% [CI: 40.9%; 57.6%] of cases being detected by the RDT2, CATT and RDT1, respectively. In passive screening, the three tests PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 6 / 20 Prospective evaluation of HAT RDT developed using recombinant antigens Fig 1. STARD diagram describing the flow of participants through the study. https://doi.org/10.1371/journal.pntd.0006386.g001 Fig 1. STARD diagram describing the flow of participants through the study. https://doi.org/10.1371/journal.pntd.0006386.g001 were more sensitive, with RDT2 achieving the highest sensitivity (90.1% [CI: 84.7%; 95.3%]), followed by CATT (74.6% [CI: 66.7%; 82.3%]) and RDT1 (70.0% [CI: 61.5%; 77.9%]). CATT had the best specificity (99.2% [CI: 99.1%; 99.2%]), followed closely by RDT1 (98.9% [CI: 98.8%; 99.0%]) and RDT2 (98.1% [CI: 98.0%; 98.2%]). With all the screening tests, speci- ficity was significantly higher in active than in passive screening. In active screening, specificity was highest with CATT (99.5% [CI: 99.5%; 99.6%]), which was followed by RDT1 (99.4% [CI: 99.3%; 99.5%]) and RDT2 (99.1% [99.0%; 99.2%]). Similarly, in passive screening, specificity was highest with CATT (97.6% [CI: 97.3%; 97.9%]), while lower results were obtained with RDT1 (96.7% [CI: 96.3%; 97.0%]) and RDT2 (93.7% [CI: 93.2%; 94.2%]). Results RDT2 had the highest accuracy (69.3% [CI: 63.5%; 74.5%]), followed by CATT (61.7% [CI: 55.7%; 67.4%]) and RDT1 (57.9% [CI: 51.8%-63.7%]). All tests had a higher accuracy in passive than in active screening. The agreement between the two technicians who read the screening tests was excellent. Cohen’s Kappa factor was above 99.8% with all the tests, both in active and passive screening. The differences in sensitivity and specificity between two screening tests are shown in Table 2 for each possible pair of tests. The RDT2 was 12.3% [CI: 3.8%; 20.7%] more sensitive than the RDT1 when the results of active and passive screening were considered together. The difference was particularly pronounced in passive screening, where the sensitivity of RDT2 was 20.1% [CI: 9.4%; 29.8%] higher than that of RDT1. By contrast, there was no evidence of a difference in sensitivity between RDT1 and RDT2 in active screening (+5.6% [CI: -7.4%; 18.7%]). The objective of non-inferiority using a margin of 8% was met in both active and pas- sive screening. The RDT2 was also more sensitive than CATT (+8.7% [CI: 1.0%; 16.6%]) when the results of active and passive screening were combined, and again, this effect was stronger in passive than in active screening (+15.5% [CI: 7.5%; 24.3%]). There was no evidence of a dif- ference in sensitivity between RDT1 and CATT in both active (-2.6% [CI: -15.2%; 9.5%]) and passive screening (-4.6% [CI: -13.3%; 5.0%]). The RDT2 was 0.83% less specific [CI: -0.96%; -0.70%] than RDT1 when results of active and passive screening were combined, which was within the non-inferiority margin of 1%. While the difference in specificity was minimal in active screening (-0.33% [CI: -0.44%; -0.23%]), it was more pronounced in passive screening (-2.98% [CI: -3.50%; -2.48%]). The RDT2 was also less specific than CATT (-1.10% [CI: -1.22%; -0.98%]), and this difference was PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 7 / 20 Prospective evaluation of HAT RDT developed using recombinant antigens more pronounced in passive (-3.86% [CI: -4.37%; -3.39%]) than in active screening (-0.46% [CI: -0.56%; -0.37%]). The RDT1 was slightly less specific than CATT (-0.27% [CI: -0.37%; -0.17%]), and this difference was also more pronounced in passive (-0.88% [CI: -1.28%; -0.50%]) than in active screening (-0.13% [CI: -0.21%; -0.05%]). Fig 2. Sensitivity (A), specificity (B) and accuracy (C) of the RDT2, RDT1 and CATT tests, by screening method. more pronounced in passive (-3.86% [CI: -4.37%; -3.39%]) than in active screening (-0.46% [CI: -0.56%; -0.37%]). The RDT1 was slightly less specific than CATT (-0.27% [CI: -0.37%; -0.17%]), and this difference was also more pronounced in passive (-0.88% [CI: -1.28%; -0.50%]) than in active screening (-0.13% [CI: -0.21%; -0.05%]). Prospective evaluation of HAT RDT developed using recombinant antigens Table 2. Differences in sensitivity and specificity between screening tests, and by method of screening. Tests compared Screening method Difference (%) Sensitivity (95% CI) Specificity (95% CI) RDT2−RDT1 Both active and passive 12.3 (3.8;20.7)# -0.83 (-0.96;-0.70) # Active 5.6 (-7.4;18.7) -0.33 (-0.44;-0.23) # Passive 20.1 (9.4;29.8) # -2.98 (-3.50;-2.48) # RDT2−CATT Both active and passive 8.7 (1.0;16.6) # -1.10 (-1.22;-0.98) # Active 3.1 (-9.8;15.5) -0.46 (-0.56;-0.37) # Passive 15.5 (7.5;24.3) # -3.86 (-4.37;-3.39) # RDT1−CATT Both active and passive -3.6 (-11.3;4.4) -0.27 (-0.37;-0.17) # Active -2.6 (-15.2;9.5) -0.13 (-0.21;-0.05) # Passive -4.6 (-13.3;5.0) -0.88 (-1.28;-0.50) # RDT1: SD BIOLINE HAT rapid diagnostic test; RDT2: SD BIOLINE HAT 2.0 rapid diagnostic test; CATT: card agglutination test for trypanosomiasis. # Difference that is significant at the 5% level. “RDT2−RDT1” corresponds to the performance of RDT2 minus the performance of RDT1, and likewise for the other pairs of tests. htt //d i /10 1371/j l td 0006386 t002 Table 2. Differences in sensitivity and specificity between screening tests, and by method of screening. RDT1: SD BIOLINE HAT rapid diagnostic test; RDT2: SD BIOLINE HAT 2.0 rapid diagnostic test; CATT: card agglutination test for trypanosomiasis. # Difference that is significant at the 5% level. “RDT2−RDT1” corresponds to the performance of RDT2 minus the performance of RDT1, and likewise for the other pairs of tests. https://doi org/10 1371/journal pntd 0006386 t002 RDT1: SD BIOLINE HAT rapid diagnostic test; RDT2: SD BIOLINE HAT 2.0 rapid diagnostic test; CATT: card agglutination test for trypanosomiasis. # Difference that is significant at the 5% level. “RDT2−RDT1” corresponds to the performance of RDT2 minus the performance of RDT1 and likewise for the other pairs of tests Considering that RDT1 and RDT2 are each made using two different antigens, we calcu- lated the sensitivity and specificity of individual antigens. Table 3 shows that for each RDT, individual antigens detected partially overlapping groups of HAT cases, since the sensitivity obtained with single antigens was lower than the result of the RDT. Therefore, having two anti- gens in these tests resulted in higher sensitivity than if only one antigen had been used. Results RDT1: SD BIOLINE HAT rapid diagnostic test; RDT2: SD BIOLINE HAT 2.0 rapid diagnostic test; CATT: card agglutination test for trypanosomiasis. https://doi.org/10.1371/journal.pntd.0006386.g002 Fig 2. Sensitivity (A), specificity (B) and accuracy (C) of the RDT2, RDT1 and CATT tests, by screening method. RDT1: SD BIOLINE HAT rapid diagnostic test; RDT2: SD BIOLINE HAT 2.0 rapid diagnostic test; CATT: card agglutination test for trypanosomiasis. Fig 2. Sensitivity (A), specificity (B) and accuracy (C) of the RDT2, RDT1 and CATT tests, by screening method. RDT1: SD BIOLINE HAT rapid diagnostic test; RDT2: SD BIOLINE HAT 2.0 rapid diagnostic test; CATT: card agglutination test for trypanosomiasis. https://doi.org/10.1371/journal.pntd.0006386.g002 more pronounced in passive (-3.86% [CI: -4.37%; -3.39%]) than in active screening (-0.46% [CI: -0.56%; -0.37%]). The RDT1 was slightly less specific than CATT (-0.27% [CI: -0.37%; -0.17%]), and this difference was also more pronounced in passive (-0.88% [CI: -1.28%; -0.50%]) than in active screening (-0.13% [CI: -0.21%; -0.05%]). PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 8 / 20 In other words, combining these two RDTs would mean that only 1.6% of cases would have been missed in passive screening, while 9.9% of them would have remained undiagnosed Table 4. Sensitivity of the three screening tests and individual antigens, by disease stage. Test Antigen (s) Disease stage Sensitivity (95% CI) RDT2 Recombinant ISG65 and recombinant VSG LiTat 1.5 Early 59.8 (52.2;67.3) Late 87.9 (81.3;93.7) Recombinant ISG65 Early 40.5 (32.6;47.9) Late 76.6 (67.9;84.5) Recombinant VSG LiTat 1.5 Early 38.2 (30.7;45.9) Late 80.4 (72.7;87.8) RDT1 Native VSG LiTat 1.3 and native VSG LiTat 1.5 Early 49.7 (41.9;57.8) Late 73.2 (64.4;81.7) Native VSG LiTat 1.3 Early 41.9 (34.0;49.4) Late 68.9 (60.2;77.5) Native VSG LiTat 1.5 Early 34.6 (27.2;42.4) Late 65.7 (56.5;74.5) CATT Trypanosomes expressing VSG LiTat 1.3 Early 50.6 (42.7;58.8) Late 80.2 (72.3;87.8) RDT1: SD BIOLINE HAT rapid diagnostic test; RDT2: SD BIOLINE HAT 2.0 rapid diagnostic test; CATT: card agglutination test for trypanosomiasis; VSG: variant surface glycoprotein. https://doi.org/10.1371/journal.pntd.0006386.t004 Prospective evaluation of HAT RDT developed using recombinant antigens Table 4. Sensitivity of the three screening tests and individual antigens, by disease stage. Test Antigen (s) Disease stage Sensitivity (95% CI) RDT2 Recombinant ISG65 and recombinant VSG LiTat 1.5 Early 59.8 (52.2;67.3) Late 87.9 (81.3;93.7) Recombinant ISG65 Early 40.5 (32.6;47.9) Late 76.6 (67.9;84.5) Recombinant VSG LiTat 1.5 Early 38.2 (30.7;45.9) Late 80.4 (72.7;87.8) RDT1 Native VSG LiTat 1.3 and native VSG LiTat 1.5 Early 49.7 (41.9;57.8) Late 73.2 (64.4;81.7) Native VSG LiTat 1.3 Early 41.9 (34.0;49.4) Late 68.9 (60.2;77.5) Native VSG LiTat 1.5 Early 34.6 (27.2;42.4) Late 65.7 (56.5;74.5) CATT Trypanosomes expressing VSG LiTat 1.3 Early 50.6 (42.7;58.8) Late 80.2 (72.3;87.8) RDT1: SD BIOLINE HAT rapid diagnostic test; RDT2: SD BIOLINE HAT 2.0 rapid diagnostic test; CATT: card agglutination test for trypanosomiasis; VSG: variant surface glycoprotein. RDT1: SD BIOLINE HAT rapid diagnostic test; RDT2: SD BIOLINE HAT 2.0 rapid diagnostic test; CATT: card agglutination test for trypanosomiasis; VSG: variant surface glycoprotein. https://doi.org/10.1371/journal.pntd.0006386.t004 The three screening tests were significantly more sensitive in late stage than in early stage patients, as shown in Table 4. The strongest difference in sensitivity between stages was observed with CATT, whose sensitivity went up from 50.6% [CI: 42.7%; 58.8%] in early stage patients to 80.2% [CI: 72.3%; 87.8] in late stage patients. The RDT2 was the most sensitive in both early stage (59.8% [CI: 52.2%; 67.3%]) and late stage patients (87.9% [81.3%; 93.7%]). While each of the antigens in RDT2 detected almost the same number of cases and contributed almost equally to the sensitivity of this test, one of the antigens of RDT1 (native VSG LiTat 1.3) detected a larger number of cases than the other antigen (native VSG LiTat 1.5). Similarly, Table 3 shows that the individual antigens of RDT2 contributed almost equally to specificity, while in the case of RDT1, native VSG LiTat 1.3 gave a slightly greater number of false positive results than the other antigen. All antigens were significantly more sensitive in passive than in active screening. The strongest difference was observed with recombinant ISG65 and recombi- nant VSG LiTat 1.5, whose sensitivity was two times higher in passive than in active screening. Table 3. Sensitivity and specificity of individual antigens in screening tests, by screening method. Antigen Screening method Sensitivity (95% CI) Specificity (95% CI) Recombinant ISG65 Both active and passive 55.0 (49.0;60.9) 98.9 (98.8;99.0) Active 37.4 (29.2;45.7) 99.5 (99.4;99.5) Passive 74.7 (65.0;82.3) 96.6 (96.3;97.0) Recombinant VSG LiTat 1.5 Both active and passive 55.3 (49.2;61.3) 98.8 (98.7;98.9) Active 37.4 (29.3;45.5) 99.4 (99.4;99.5) Passive 75.4 (67.3;82.7) 96.1 (95.7;96.4) Native VSG LiTat 1.3 Both active and passive 52.7 (46.7;58.6) 99.1 (99.0;99.2) Active 40.2 (31.7;48.3) 99.5 (99.5;99.6) Passive 66.8 (59.0;74.0) 97.3 (97.1;97.6) Native VSG LiTat 1.5 Both active and passive 47.0 (40.8;52.8) 99.3 (99.3;99.4) Active 38.8 (30.6;47.2) 99.6 (99.5;99.7) Passive 56.4 (47.2;65.8) 98.2 (98.0;98.5) CATT antigen (trypanosomes expressing VSG LiTat 1.3) Both active and passive 62.5 (56.2;68.4) 99.2 (99.1;99.2) Active 51.8 (43.1;59.9) 99.5 (99.5;99.6) Passive 74.6 (66.7;82.3) 97.6 (97.3;97.9) https://doi org/10 1371/journal pntd 0006386 t003 Sensitivity and specificity of individual antigens in screening tests, by screening method. PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 9 / 2 9 / 20 Prospective evaluation of HAT RDT developed using recombinant antigens The three screening tests were significantly more sensitive in late stage than in early stage patients, as shown in Table 4. The strongest difference in sensitivity between stages was observed with CATT, whose sensitivity went up from 50.6% [CI: 42.7%; 58.8%] in early stage patients to 80.2% [CI: 72.3%; 87.8] in late stage patients. The RDT2 was the most sensitive in both early stage (59.8% [CI: 52.2%; 67.3%]) and late stage patients (87.9% [81.3%; 93.7%]). Similarly, all the individual RDT antigens were significantly more sensitive in late than in early stage patients. The largest difference between stages was observed with recombinant VSG LiTat 1.5 (42.1%), while the smallest difference was with native VSG LiTat 1.3 (27.0%). The most sensitive antigen in early stage patients was native VSG LiTat 1.3 (41.9% [34.0%; 49.4%]), while the most sensitive antigen in late stage patients was recombinant VSG LiTat 1.5 (80.4% [72.7%; 87.8%]). We also calculated the diagnostic performance that would be achieved by combining two or three screening tests, with the goal of improving the overall sensitivity of screening, which is important in enhancing control of HAT, as humans are considered the main reservoirs of the disease [21]. The sensitivity and specificity of all possible combinations of two or three screen- ing tests is shown in Fig 3. As expected, the highest sensitivity was achieved by combining all three tests (99.6% [CI: 98.7; 100.0]). This did not reach 100% because there were some differ- ences between the two readers who interpreted test results. The most sensitive combination of two tests was RDT1 and RDT2, which detected 90.1% of cases [CI: 86.2%; 93.6%] and was markedly more sensitive than the individual tests. Lower sensitivity values were obtained by combining CATT and RDT2 (87.8% [CI: 83.7%; 91.6%]) and even more so by combining CATT and RDT1 (81.4% [CI: 76.4; 85.9]). Combining screening tests provided a greater increase in sensitivity in active than in passive screening. In active screening, sensitivity increased from 54.8% [CI: 46.8%; 63.5] with RDT2 to 83.0% [CI: 76.2%; 89.3%] when combin- ing RDT1 and RDT2. In passive screening, this same combination achieved a remarkable sen- sitivity of 98.4% [CI: 95.6%; 100.0%], compared to 90.1% [CI: 84.7%; 95.3%] with RDT2 alone. Similarly, all the individual RDT antigens were significantly more sensitive in late than in early stage patients. The largest difference between stages was observed with recombinant VSG LiTat 1.5 (42.1%), while the smallest difference was with native VSG LiTat 1.3 (27.0%). The most sensitive antigen in early stage patients was native VSG LiTat 1.3 (41.9% [34.0%; 49.4%]), while the most sensitive antigen in late stage patients was recombinant VSG LiTat 1.5 (80.4% [72.7%; 87.8%]). We also calculated the diagnostic performance that would be achieved by combining two or three screening tests, with the goal of improving the overall sensitivity of screening, which is important in enhancing control of HAT, as humans are considered the main reservoirs of the disease [21]. The sensitivity and specificity of all possible combinations of two or three screen- ing tests is shown in Fig 3. As expected, the highest sensitivity was achieved by combining all three tests (99.6% [CI: 98.7; 100.0]). This did not reach 100% because there were some differ- ences between the two readers who interpreted test results. The most sensitive combination of two tests was RDT1 and RDT2, which detected 90.1% of cases [CI: 86.2%; 93.6%] and was markedly more sensitive than the individual tests. Lower sensitivity values were obtained by combining CATT and RDT2 (87.8% [CI: 83.7%; 91.6%]) and even more so by combining CATT and RDT1 (81.4% [CI: 76.4; 85.9]). Combining screening tests provided a greater increase in sensitivity in active than in passive screening. In active screening, sensitivity increased from 54.8% [CI: 46.8%; 63.5] with RDT2 to 83.0% [CI: 76.2%; 89.3%] when combin- ing RDT1 and RDT2. In passive screening, this same combination achieved a remarkable sen- sitivity of 98.4% [CI: 95.6%; 100.0%], compared to 90.1% [CI: 84.7%; 95.3%] with RDT2 alone. In other words, combining these two RDTs would mean that only 1.6% of cases would have been missed in passive screening, while 9.9% of them would have remained undiagnosed PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 10 / 20 Prospective evaluation of HAT RDT developed using recombinant antigens Fig 3. Sensitivity (A) and specificity (B) of all possible combinations of two or three screening tests, by screenin method and by disease stage. Test combinations are shown in descending order of sensitivity. RDT1: SD BIOLINE HAT rapid diagnostic test; RDT2: SD BIOLINE HAT 2.0 rapid diagnostic test; CATT: card agglutination test for trypanosomiasis. For the sake of simplicity, only results obtained by the first reader are shown. The total number of true positives does not equal the total number of cases enrolled in the study (N = 260), as the first reader missed two cases in active screening. https://doi.org/10.1371/journal.pntd.0006386.g004 using RDT2 only. However, using such combinations resulted in some trade-off in specificity, which went down to 96.9% [CI: 96.8%; 97.1%] when the three tests were taken together, or to 97.3% [CI: 97.2%; 97.4%] when combining RDT1 and RDT2. The contribution of each screening test to the detection of cases and to false positive results is demonstrated using Venn diagrams in Figs 4 and 5. Fig 4 shows that for true positive results, the degree of overlap between the tests was much higher in passive than in active screening. Fig 5 shows that for false positive results, the degree of overlap between the tests was also higher in passive than in active screening, but this difference was much less pronounced than for true positive results. Both mAECT tests were performed on 124 HAT cases. Ninety cases (72.6%) were positive by mAECT-bc, while only 64 cases (51.6%) were positive by mAECT-wb. There was a high degree of overlap between the two tests, with 58 cases detected using both methods. The number of HAT cases identified using the different parasitological tests used in the study, as well as the corresponding positivity rates of each of the screening tests, are included as supporting information (S2 Table). This data indicates that among the three parasitological tests performed on blood samples, the positivity rates of screening tests were highest in cases identified by mAECT-bc and lowest in cases diagnosed by CTC. The positivity rates of Fig 5. Venn diagrams showing the number of false positive results obtained with the RDT2, RDT1 and CATT tests. (A) Results from active and passive screening combined (N = 1,768 false positives); (B) results from active screening (N = 769 false positives); (C) results from passive screening (N = 999 false positives). For the sake of simplicity, only results obtained by the first reader are shown. https://doi org/10 1371/journal pntd 0006386 g005 Fig 5. Venn diagrams showing the number of false positive results obtained with the RDT2, RDT1 and CATT tests. The result of the combination of tests is positive if at least one of the tests is positive, while the resu is negative if all the tests of the combination are negative. https://doi.org/10.1371/journal.pntd.0006386.g003 Fig 3. Sensitivity (A) and specificity (B) of all possible combinations of two or three screening tests, by screening method and by disease stage. Test combinations are shown in descending order of sensitivity. RDT1: SD BIOLINE HAT rapid diagnostic test; RDT2: SD BIOLINE HAT 2.0 rapid diagnostic test; CATT: card agglutination test for trypanosomiasis. The result of the combination of tests is positive if at least one of the tests is positive, while the result is negative if all the tests of the combination are negative. https://doi.org/10.1371/journal.pntd.0006386.g003 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 11 / 20 Prospective evaluation of HAT RDT developed using recombinant antigens Fig 4. Venn diagrams showing the number of true positive results obtained with the RDT2, RDT1 and CATT tests. (A) Results from active and passive screening combined (N = 258 true positives); (B) results from active screening (N = 136 true positives); (C) results from passive screening (N = 122 true positives). For the sake of simplicity, only results obtained by the first reader are shown. The total number of true positives does not equal the total number of cases enrolled in the study (N = 260), as the first reader missed two cases in active screening. https://doi.org/10.1371/journal.pntd.0006386.g004 Fig 4. Venn diagrams showing the number of true positive results obtained with the RDT2, RDT1 and CATT tests. Fig 4. Venn diagrams showing the number of true positive results obtained with the RDT2, RDT1 and CATT tests. (A) Results from active and passive screening combined (N = 258 true positives); (B) results from active screening (N = 136 true positives); (C) results from passive screening (N = 122 true positives). For the sake of simplicity, only results obtained by the first reader are shown. The total number of true positives does not equal the total number of cases enrolled in the study (N = 260), as the first reader missed two cases in active screening. htt //d i /10 1371/j l td 0006386 004 (A) Results from active and passive screening combined (N = 258 true positives); (B) results from active screening (N = 136 true positives); (C) results from passive screening (N = 122 true positives). https://doi.org/10.1371/journal.pntd.0006386.g005 Discussion The main objective of this study, to demonstrate the non-inferiority of the sensitivity and spec- ificity of the RDT2 in comparison to the RDT1, was successfully achieved. However, all three screening tests that were evaluated were unexpectedly insensitive, particularly in active screen- ing, which is in contrast with earlier reports. While CATT has been extensively evaluated and used in clinical settings, and its sensitivity has been reported to range between 68.8% and 100% [22], in this study, the test missed almost half of the cases in active screening. Previous retrospective studies also reported the sensitivity of the RDT1 to be between 82% and 99.6% [8,11]. A sensitivity of 89% was reported in a prospective study of a prototype version of the RDT1 [9], while in another trial, the sensitivity of the commercialized RDT1 was 92% [14]. This apparent discrepancy could be explained by assuming that each of the three screening tests detected cases with different serological profiles, which were only partially overlapping, as evidenced by the results shown in Fig 4. The design of the study, which included three screen- ing tests to identify suspects during enrolment, would be responsible for the low sensitivity of an individual screening test. By contrast, earlier studies only included one, or sometimes two screening tests during enrolment, and as a result, the sensitivity of screening tests could have been significantly overestimated, since cases with serological profiles that were different from the ones identified by the particular test could have been missed. Therefore, there is the need to explore the possibility of including two or more screening tests in diagnostic algorithms, in order to increase sensitivity and accelerate interruption of disease transmission, particularly by enhancing detection of patients in early stage disease. Based on the results presented here, strategies combining RDT2 with either RDT1 or CATT in active screening, and combining RDT2 with RDT1 in passive screening, could be considered to enhance case detection. In active screening, each test detected a particularly large number of cases that were missed by the other tests, and combining several screening tests would therefore result in a stronger gain in sensitivity than in passive screening. However, operational aspects would also need to be con- sidered, and cost-effectiveness analyses may provide helpful information to select the most appropriate strategies that would ensure optimal detection of cases. Prospective evaluation of HAT RDT developed using recombinant antigens screening tests in cases identified by examining lymph node aspirates were not significantly different from the positivity rates obtained in cases that were positive with parasitological tests performed on blood samples. The highest screening test positivity rates were obtained in cases with trypanosomes detected in the cerebrospinal fluid. (A) Results from active and passive screening combined (N = 1,768 false positives); (B) results from active screening (N = 769 false positives); (C) results from passive screening (N = 999 false positives). For the sake of simplicity, only results obtained by the first reader are shown. PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 12 / 20 Prospective evaluation of HAT RDT developed using recombinant antigens number of suspects would need to undergo confirmatory testing, which often requires travel- ling long distances. Such limitations will become increasingly relevant as progress is being made towards elimination of the disease, since the positive predictive value of screening tests will decline along with the disease prevalence. Alternatively, investing in the development of a new screening test that would be more sen- sitive than the tests that were evaluated here, and which would include multiple antigens, could be considered. Such a test might be developed by combining three or more antigens, which could include those in the RDT2, as well as other promising candidates identified in previous studies [23–28]. Other RDTs being developed using recombinant antigens will also need to be considered once they are available and their performance has been evaluated [25]. With the increasing prospects of new, safer treatments for g-HAT that would be effective for both stages of the disease [29,30], a test with high sensitivity and specificity could make a “test and treat” approach possible, without requiring any parasitological confirmation. pp p q g y p g A number of hypotheses could be formulated to try and explain the low sensitivity of indi- vidual screening tests observed in this study, which would require further investigations. Afri- can trypanosomes are notorious for having evolved a mechanism of escaping the host immune system by regularly changing the variant surface glycoprotein (VSG) that composes their cell coat, using a large repertoire of dedicated genes [31,32]. It is therefore likely that HAT patients who have been infected recently could have raised an immune response to only a limited num- ber of VSG antigens, while patients who are in a more advanced disease stage could harbour antibodies against a larger panel of VSGs. This could explain why screening tests that include some specific VSGs, such as the three tests evaluated here, would detect different HAT cases, and why screening tests were more sensitive in late than in early disease stage patients. Similarly, this would provide an explanation for the lower sensitivity that was observed in active screening, since the disease is generally less advanced in most cases among the people screened. Discussion In particular, there is the need to determine whether the extra complexity of the diagnostic algorithm and workload that would result from performing two or more screening tests and having more serological sus- pects to test by microscopy would cause a significant reduction in the number of people screened by a mobile team in a day, and balance it against the gain in detection of a larger pro- portion of cases among the people screened. Performing several screening tests would also be a logistical challenge in terms of transportation and storage of tests. Some patients could also refuse to have two or more tests performed on them, an unlikely possibility since blood is taken from the same finger prick. If only one screening test had to be used, the results pre- sented here support using RDT2 in order to enhance case detection, as it was more sensitive than RDT1 and CATT, in both active and passive screening settings. RDT2 was also the most sensitive test in both early and late stage patients, which indicates that the test is able to detect patients with various clinical profiles. Maximizing sensitivity would be a sensible strategy in a disease elimination context, but the marginally lower specificity of RDT2 would also need to be considered, as it would result in an increase in workload to confirm suspects, decrease in confidence in test results and would also have a negative impact on patients, since a larger PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 13 / 20 Prospective evaluation of HAT RDT developed using recombinant antigens test. This could be due to the nature of the CATT reagents, which in addition to VSG LiTat 1.3, would include other trypanosome antigens that could react with corresponding antibodies in the blood of HAT patients. Another explanation might be the difference in test formats, which may be associated with different binding or exposure characteristics of antigens and epi- topes. While in the RDT, antigens are printed on a nitrocellulose membrane, CATT is per- formed by mixing a suspension containing fixed parasites with the test sample on a plasticised card. In addition, although the exact composition of the RDT buffer is unknown, it is likely to be different from the CATT buffer (phosphate buffered saline, pH 7.2 with 0.1% sodium azide), which could have an impact on antigenic binding. In an earlier prospective study that was conducted in the DRC to evaluate the performance of RDT1, the VSG LiTat 1.5 antigen was more sensitive than the VSG LiTat 1.3 antigen (83.6% [CI: 76.3%; 89.0%] and 76.0% [CI: 68.0%; 82.5%], respectively) [14], which is in contrast to what was found in the present study. In another multi-country study that evaluated the perfor- mance of the prototype RDT1, identical sensitivity values were reported for each antigen (85.9% [CI: 79.4%; 90.6%]) [9]. While these differences may be due to slightly different study designs, they do not appear to be statistically significant, and would therefore tend to support the view that both antigens contribute equally to the sensitivity of RDT1. While the three screening tests were highly specific in active screening, they were signifi- cantly less specific in passive screening. This difference might be due to serological differences between the two populations, with the population presenting to fixed health facilities being more likely to be infected with other pathogens that could trigger immune responses cross- reacting with the tests. Alternatively, this difference might be explained by the relatively low sensitivity of routine parasitological methods [34]. Indeed, since the HAT prevalence was higher in passive than in active screening, this population was also more likely to have included HAT patients who could have been found positive by screening tests but missed by parasitol- ogy, which would have resulted in an underestimate of the specificity of screening tests. Alternatively, the difference in sensitivity between early and late-stage patients could be due to higher antibody titres in the latter because of a longer period of exposure to parasite antigens, and hence stronger immune response. This explanation would better support the finding that the sensitivity of an invariant antigen like ISG65, which is expressed throughout the infection, was higher in late-stage patients. These hypotheses could be tested using animal models infected with T.b. brucei [32]. Some patients could have also been infected with try- panosome strains lacking the genes encoding the VSG antigens present in these screening tests. In particular, deletions of the gene encoding VSG LiTat 1.3 have been reported in some T.b. gambiense isolates from Cameroon [33], and such deletions could be among the factors responsible for the low sensitivity of CATT and RDT1. Although there is currently no evidence to directly support this hypothesis, it is also conceivable that these deletions could have become increasingly frequent due to the selection pressure applied by the extensive use of CATT in HAT-endemic populations. This phenomenon could have remained unnoticed, since most studies conducted until recently only included CATT during enrolment. Finally, it cannot be excluded that some HAT cases could have corresponded to false positive parasitological test results, which would have been negative with screening tests. It is likely that several of the hypotheses described here could partially explain the observed low sensitivity of screening tests that was found in this study. Other studies comparing the performance of different screening tests in various settings will hopefully help clarify this point. The fact that RDT1 detected 49 HAT cases that were missed by CATT (Fig 4) could be explained by the presence in RDT1 of the VSG LiTat 1.5 antigen, which is not included in CATT, and also possibly by differences in test formats. On the other hand, it is noteworthy that CATT also detected 58 HAT cases that were missed by RDT1, yet RDT1 contains VSG LiTat 1.3, the antigen that is predominantly expressed by the fixed trypanosomes in the CATT PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 14 / 20 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 Prospective evaluation of HAT RDT developed using recombinant antigens batches. No failure to follow storage procedures was observed during the study, screening tests were used according to manufacturers’ instructions and staff performing the tests ensured that positive and negative controls (for CATT) as well as procedural controls (for RDTs) reacted according to instructions. Yet it is possible that some tests could have deteriorated within the limits of the controls, thereby affecting performance. The mAECT-bc method [17] may be considered as a replacement of mAECT-wb, which is routinely used in the DRC and other endemic countries. Although based on a subset of par- ticipants, the data presented here are in agreement with earlier results showing a significant increase in sensitivity using mAECT-bc. In a first study that was conducted in Guinea, the sensitivity of mAECT-bc was 96.5%, while the sensitivity of mAECT-wb was 78.9% [17]. Another study that was conducted in DRC reported a somehow smaller difference in sensitiv- ity between these two methods (90.9% and 80.4%, respectively) [34]. The lower sensitivity val- ues that were found here (72.6% and 51.6%, respectively) could be explained by the fact that the mAECT methods were only performed on a subset of participants who had been negative with other parasitological methods, and who were therefore likely to include cases with a lower parasitaemia than the other cases that were enrolled in the study. This selection bias could also explain why the difference in sensitivity was higher than in previous reports, since patients with a low parasitaemia could have provided a better dynamic range to evaluate subtle differ- ences in sensitivity. While the difference in sensitivity could be an overestimate of the true dif- ference that would be observed in an unbiased population, implementing the mAECT-bc protocol could be considered to enhance case finding, for a minimal additional workload. Since there was a high degree of overlap between the mAECT-wb and mAECT-bc results, per- forming both methods may not be justified, as it would increase costs without resulting in any significant increase in sensitivity. Although introducing mAECT-bc would require specific training to prepare buffy coat samples, it did not present a particular challenge during this study, and therefore, implementing it at other sites that are already equipped to perform mAECT-wb should be relatively straightforward. While mAECT-bc has been shown to be more sensitive than mAECT-wb, mAECT-wb is known to be more sensitive than CTC [6,34]. The difference in specificity between active and passive screening could thus be an artefact related to the imperfect parasitological reference standard, rather than reflect a real difference in test specificity. RDT1 and CATT were previously evaluated in another prospective study that was con- ducted in the DRC [14], which reported that the sensitivity of RDT1 (92.0% [CI: 86.1%- 95.5%]) was significantly higher than that of CATT (69.1% [CI: 60.7%-76.4%]). Surprisingly, there was no evidence of any difference in sensitivity between RDT1 and CATT in the present study. The reasons for this discrepancy are unclear, and several hypotheses could be drawn. First, although the studies shared some of the sites, it is possible that the two study populations may have had significantly distinct serological profiles, resulting in different degrees of overlap between the tests. According to this hypothesis, the degree of overlap between screening tests should not be viewed as constant and specific to the tests, but instead, considered as a dynamic phenomenon that may exhibit significant variability in time and in space, depending on the population that is sampled and the underlying immune responses of individual patients. Although this hypothesis seems rather unlikely since the studies were conducted in similar populations, it would be useful to conduct additional studies to establish the reproducibility of such differences. In spite of efforts to ensure compliance with the study protocol and proce- dures through training, supervision and monitoring, it is still possible that some of the sites could have performed less well, which could have had an impact on study results. Alterna- tively, differences between these studies could be due to operational or logistical factors caus- ing some of the tests to have a lower performance than expected. While this seems unlikely, subtle changes during the production of antigens or other components of one of the tests could have occurred and gone undetected, resulting in the lower sensitivity of some test PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 15 / 20 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 Prospective evaluation of HAT RDT developed using recombinant antigens high degree of overlap of true positive results, and to a lesser extent of false positive results, between the tests that was found in passive screening. The results presented here have confirmed that the RDT2 would be a useful test for both active and passive screening, either as a single test or in combination with other screening tests. Since it is produced using recombinant antigens exclusively, it will also be easier and safer to manufacture than screening tests that are made with native antigens. The RDT2 is thus a welcome addition to the set of tools that are currently available to control and eventually eliminate HAT. S1 Table. STARD checklist. (DOCX) S2 Table. Results of parasitological tests performed on HAT cases and positivity of screen- ing tests in cases that were positive with specific parasitological tests. CTC: capillary tube centrifugation; mAECT-wb: mini anion exchange centrifugation technique on whole blood; mAECT-bc: mini anion exchange centrifugation technique on buffy coat. (DOCX) S3 Table. Malaria RDT results obtained in HAT cases and serological suspects. Malaria prevalence values correspond to the percentage of positive malaria RDT results obtained among participants that were tested with a malaria RDT. For the sake of simplicity, only results obtained by the first reader are shown. (DOCX) S3 Table. Malaria RDT results obtained in HAT cases and serological suspects. Malaria prevalence values correspond to the percentage of positive malaria RDT results obtained among participants that were tested with a malaria RDT. For the sake of simplicity, only results obtained by the first reader are shown. (DOCX) S3 Table. Malaria RDT results obtained in HAT cases and serological suspects. Malaria prevalence values correspond to the percentage of positive malaria RDT results obtained among participants that were tested with a malaria RDT. For the sake of simplicity, only results obtained by the first reader are shown. (DOCX) S4 Table. Sensitivity of HAT screening tests in malaria RDT positive and negative partici- pants in passive screening. For the sake of simplicity, only results obtained by the first reader are shown. (DOCX) S1 Data. Study database. (CSV) Thus, the observation that the positivity rates of the screening tests were highest in cases found positive by mAECT-bc and lowest in cases that were positive by CTC could suggest that screening tests would be more sensitive in low-parasi- taemia than in high-parasitaemia cases. Although this would need to be further investigated, it would be in agreement with the assumption that patients with a low parasitaemia would have a stronger immune response, which would facilitate their identification using antibody-detec- tion screening tests. Conversely, patients unable to mount a strong immune response against trypanosomes and therefore more likely to have a high parasitaemia could be more difficult to identify using these screening tests. This study confronted a number of challenges, which could have somehow impacted the quality of the results. Although study sites were carefully selected based on the available epide- miological data, the HAT prevalence was generally low, making it necessary to enroll patients at 15 different sites. This presented a significant challenge to the study team in terms of coordi- nation, in particular when considering that most of the sites are located in remote, rural areas that were difficult to access. In addition, there was significant turn-over of personnel at some of the sites, requiring additional training. Enrolment was also interrupted at some sites due to stock-outs of supplies, such as mAECT kits. While the study was blinded, it is possible that technicians performing the tests could have been aware of the clinical status of some partici- pants. This is probably more likely to be true in passive screening, since the number of patients presenting daily to health facilities was sometimes very low, making blinding more difficult. Although it is hypothetical, this imperfect blinding could be one of the factors leading to the PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 16 / 20 Supporting information S1 Fig. Venn diagrams showing the number of true positive results obtained with the RDT2, RDT1 and CATT tests among early-stage (A) and late-stage (B) cases. For the sake of simplicity, only results obtained by the first reader are shown. The total number of true posi- tives does not equal the total number of cases enrolled in the study (N = 260), as the first reader missed two cases in active screening. (TIF) Acknowledgments We acknowledge the personnel of PNLTHA in the DRC for their commitment and efforts in conducting this study under difficult field conditions. We would also like to thank Patrick Mitashi and Edmond Mulamba for monitoring study sites, as well as De´borah Mujinga Tshishiku and Nhora Lubanda for data entry. We acknowledge Mike Ferguson (University of Dundee, United Kingdom) for providing the plasmid and necessary information to express recombinant ISG65-1, and Mark Carrington (University of Cambridge, United Kingdom) and Cambridge Enterprise Limited (United Kingdom) for supplying the amino acid sequence to express the N-terminal domain of VSG LiTat 1.5. We thank Sanjeev Krishna (St. George’s Uni- versity of London, United Kingdom), Olaf Valverde (Drugs for Neglected Diseases initiative, PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 17 / 20 Prospective evaluation of HAT RDT developed using recombinant antigens Switzerland), Jose Ramon Franco (World Health Organization, Switzerland), Mark Carring- ton, Mike Fergusson, Epco Hasker (Institute of Tropical Medicine, Belgium), Christian Burri (Swiss Tropical and Public Health Institute, Switzerland), Franc¸ois Chappuis (Geneva Univer- sity Hospital / Me´decins Sans Frontières, Switzerland), Mike Barrett (University of Glasgow, United Kingdom), Mark Perkins (FIND, Switzerland) and Albert Picado (FIND, Switzerland) for participating in an experts’ meeting to review interim results of the study. Finally, we thank Beatrice Gordis and Lauren Jacobson (FIND, Switzerland) for proofreading the manuscript. Author Contributions Conceptualization: Crispin Lumbala, Sylvain Bie´ler, Joseph Mathu Ndung’u. Conceptualization: Crispin Lumbala, Sylvain Bie´ler, Joseph Mathu Ndung’u. Data curation: Sylvain Bie´ler, Simon Kayembe, Jacquies Makabuza. Formal analysis: Sylvain Bie´ler, Stefano Ongarello. Formal analysis: Sylvain Bie´ler, Stefano Ongarello. Funding acquisition: Sylvain Bie´ler, Joseph Mathu Ndung’u. Investigation: Crispin Lumbala, Simon Kayembe, Jacquies Makabuza. Methodology: Crispin Lumbala, Sylvain Bie´ler, Joseph Mathu Ndung’u. Project administration: Crispin Lumbala, Sylvain Bie´ler, Jacquies Makabuza, Joseph Mathu Ndung’u. Resources: Crispin Lumbala. Supervision: Crispin Lumbala, Sylvain Bie´ler, Joseph Mathu Ndung’u. Visualization: Sylvain Bie´ler. Writing – original draft: Sylvain Bie´ler. Writing – original draft: Sylvain Bie´ler. Writing – review & editing: Crispin Lumbala, Stefano Ongarello, Joseph Mathu Ndung’u. Prospective evaluation of HAT RDT developed using recombinant antigens centric Prospective Study. PLoS Negl Trop Dis 10: e0004608. https://doi.org/10.1371/journal.pntd. 0004608 PMID: 27058033 centric Prospective Study. PLoS Negl Trop Dis 10: e0004608. https://doi.org/10.1371/journal.pntd. 0004608 PMID: 27058033 10. Buscher P, Mertens P, Leclipteux T, Gilleman Q, Jacquet D, et al. (2014) Sensitivity and specificity of HAT Sero-K-SeT, a rapid diagnostic test for serodiagnosis of sleeping sickness caused by Trypano- soma brucei gambiense: a case-control study. Lancet Glob Health 2: e359–363. https://doi.org/10. 1016/S2214-109X(14)70203-7 PMID: 25103304 11. Jamonneau V, Camara O, Ilboudo H, Peylhard M, Koffi M, et al. (2015) Accuracy of individual rapid tests for serodiagnosis of gambiense sleeping sickness in West Africa. PLoS Negl Trop Dis 9: e0003480. https://doi.org/10.1371/journal.pntd.0003480 PMID: 25642701 12. Lejon V, Buscher P (2005) Review Article: cerebrospinal fluid in human African trypanosomiasis: a key to diagnosis, therapeutic decision and post-treatment follow-up. Trop Med Int Health 10: 395–403. https://doi.org/10.1111/j.1365-3156.2005.01403.x PMID: 15860085 13. Sullivan L, Wall SJ, Carrington M, Ferguson MA (2013) Proteomic selection of immunodiagnostic anti- gens for human African trypanosomiasis and generation of a prototype lateral flow immunodiagnostic device. PLoS Negl Trop Dis 7: e2087. https://doi.org/10.1371/journal.pntd.0002087 PMID: 23469310 14. Lumbala C, Bessell PR, Lutumba P, Baloji S, Bieler S, et al. (2017) Performance of the SD BIOLINE(R) HAT rapid test in various diagnostic algorithms for gambiense human African trypanosomiasis in the Democratic Republic of the Congo. PLoS One 12: e0180555. https://doi.org/10.1371/journal.pone. 0180555 PMID: 28672036 15. Woo PT (1970) The haematocrit centrifuge technique for the diagnosis of African trypanosomiasis. Acta Trop 27: 384–386. PMID: 4396363 16. Buscher P, Mumba Ngoyi D, Kabore J, Lejon V, Robays J, et al. (2009) Improved Models of Mini Anion Exchange Centrifugation Technique (mAECT) and Modified Single Centrifugation (MSC) for sleeping sickness diagnosis and staging. PLoS Negl Trop Dis 3: e471. https://doi.org/10.1371/journal.pntd. 0000471 PMID: 19936296 17. Camara M, Camara O, Ilboudo H, Sakande H, Kabore J, et al. (2010) Sleeping sickness diagnosis: use of buffy coats improves the sensitivity of the mini anion exchange centrifugation test. Trop Med Int Health 15: 796–799. https://doi.org/10.1111/j.1365-3156.2010.02546.x PMID: 20497407 18. Efron B, Tibshirani R (1993) An Introduction to the Bootstrap. 18. Efron B, Tibshirani R (1993) An Introduction to the Bootstrap. 19. Eisenhart C, Rosenblatt JR (1972) W. J. Youden, 1900–1971. The Annals of Mathematical Statistics 43: 1035–1040. 20. Blackwelder WC (1982) "Proving the null hypothesis" in clinical trials. Control Clin Trials 3: 345–353. PMID: 7160191 21. References 1. Simarro PP, Cecchi G, Franco JR, Paone M, Diarra A, et al. (2012) Estimating and mapping the popula- tion at risk of sleeping sickness. PLoS Negl Trop Dis 6: e1859. https://doi.org/10.1371/journal.pntd. 0001859 PMID: 23145192 2. WHO (2016) Global Health Observatory data repository, Human African Trypanosomiasis. Geneva, Switerzland: World Health Organization. 3. Franco JR, Simarro PP, Diarra A, Ruiz-Postigo JA, Jannin JG (2014) The journey towards elimination of gambiense human African trypanosomiasis: not far, nor easy. Parasitology 141: 748–760. https:// doi.org/10.1017/S0031182013002102 PMID: 24709291 4. Jamonneau V, Ilboudo H, Kabore J, Kaba D, Koffi M, et al. (2012) Untreated Human Infections by Try- panosoma brucei gambiense Are Not 100% Fatal. PLoS Negl Trop Dis 6: e1691. https://doi.org/10. 1371/journal.pntd.0001691 PMID: 22720107 5. Magnus E, Van Meirvenne N, Vervoort T, Le Ray D, Wery M (1978) Use of freeze-dried trypanosomes in the indirect fluorescent antibody test for the serodiagnosis of sleeping sickness. Ann Soc Belg Med Trop 58: 103–109. PMID: 361001 6. Chappuis F, Loutan L, Simarro P, Lejon V, Buscher P (2005) Options for field diagnosis of human afri- can trypanosomiasis. Clin Microbiol Rev 18: 133–146. https://doi.org/10.1128/CMR.18.1.133-146. 2005 PMID: 15653823 7. Buscher P, Gilleman Q, Lejon V (2013) Rapid diagnostic test for sleeping sickness. N Engl J Med 368: 1069–1070. https://doi.org/10.1056/NEJMc1210373 PMID: 23484849 8. Sternberg JM, Gierlinski M, Bieler S, Ferguson MA, Ndung’u JM (2014) Evaluation of the diagnostic accuracy of prototype rapid tests for human African trypanosomiasis. PLoS Negl Trop Dis 8: e3373. https://doi.org/10.1371/journal.pntd.0003373 PMID: 25521120 9. Bisser S, Lumbala C, Nguertoum E, Kande V, Flevaud L, et al. (2016) Sensitivity and Specificity of a Prototype Rapid Diagnostic Test for the Detection of Trypanosoma brucei gambiense Infection: A Multi- 18 / 20 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 Franco JR, Simarro PP, Diarra A, Jannin JG (2014) Epidemiology of human African trypanosomiasis. Clin Epidemiol 6: 257–275. https://doi.org/10.2147/CLEP.S39728 PMID: 25125985 22. Mitashi P, Hasker E, Lejon V, Kande V, Muyembe JJ, et al. (2012) Human african trypanosomiasis diag- nosis in first-line health services of endemic countries, a systematic review. PLoS Negl Trop Dis 6: e1919. https://doi.org/10.1371/journal.pntd.0001919 PMID: 23209860 23. Bieler S, Waltenberger H, Barrett MP, McCulloch R, Mottram JC, et al. (2016) Evaluation of Antigens for Development of a Serological Test for Human African Trypanosomiasis. PLoS One 11: e0168074. https://doi.org/10.1371/journal.pone.0168074 PMID: 27936225 24. Roge S, Van Nieuwenhove L, Meul M, Heykers A, Brouwer de Koning A, et al. (2014) Recombinant anti- gens expressed in Pichia pastoris for the diagnosis of sleeping sickness caused by Trypanosoma brucei gambiense. PLoS Negl Trop Dis 8: e3006. https://doi.org/10.1371/journal.pntd.0003006 PMID: 25032684 25. Rooney B, Piening T, Buscher P, Roge S, Smales CM (2015) Expression of Trypanosoma brucei gam- biense Antigens in Leishmania tarentolae. Potential for Use in Rapid Serodiagnostic Tests (RDTs). PLoS Negl Trop Dis 9: e0004271. https://doi.org/10.1371/journal.pntd.0004271 PMID: 26650813 26. Van Nieuwenhove L, Buscher P, Balharbi F, Humbert M, Dieltjens T, et al. (2012) Identification of Mimo- topes with Diagnostic Potential for Trypanosoma brucei gambiense Variant Surface Glycoproteins Using Human Antibody Fractions. PLoS Negl Trop Dis 6: e1682. https://doi.org/10.1371/journal.pntd. 0001682 PMID: 22720103 27. Van Nieuwenhove LC, Roge S, Balharbi F, Dieltjens T, Laurent T, et al. (2011) Identification of peptide mimotopes of Trypanosoma brucei gambiense variant surface glycoproteins. PLoS Negl Trop Dis 5: e1189. https://doi.org/10.1371/journal.pntd.0001189 PMID: 21695105 28. Van Nieuwenhove L, Buscher P, Balharbi F, Humbert M, Guisez Y, et al. (2013) A LiTat 1.5 variant sur- face glycoprotein-derived peptide with diagnostic potential for Trypanosoma brucei gambiense. Trop Med Int Health 18: 461–465. https://doi.org/10.1111/tmi.12058 PMID: 23289521 19 / 20 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006386 March 28, 2018 Prospective evaluation of HAT RDT developed using recombinant antigens 29. Steinmann P, Stone CM, Sutherland CS, Tanner M, Tediosi F (2015) Contemporary and emerging strategies for eliminating human African trypanosomiasis due to Trypanosoma brucei gambiense: review. Trop Med Int Health 20: 707–718. https://doi.org/10.1111/tmi.12483 PMID: 25694261 30. Tarral A, Blesson S, Mordt OV, Torreele E, Sassella D, et al. (2014) Determination of an optimal dosing regimen for fexinidazole, a novel oral drug for the treatment of human African trypanosomiasis: first-in- human studies. Clin Pharmacokinet 53: 565–580. https://doi.org/10.1007/s40262-014-0136-3 PMID: 24535888 31. Schwede A, Macleod OJ, MacGregor P, Carrington M (2015) How Does the VSG Coat of Bloodstream Form African Trypanosomes Interact with External Proteins? PLoS Pathog 11: e1005259. https://doi. org/10.1371/journal.ppat.1005259 PMID: 26719972 32. Mugnier MR, Cross GA, Papavasiliou FN (2015) The in vivo dynamics of antigenic variation in Trypano- soma brucei. Science 347: 1470–1473. https://doi.org/10.1126/science.aaa4502 PMID: 25814582 33. Dukes P, Gibson WC, Gashumba JK, Hudson KM, Bromidge TJ, et al. (1992) Absence of the LiTat 1.3 (CATT antigen) gene in Trypanosoma brucei gambiense stocks from Cameroon. Acta Trop 51: 123– 134. PMID: 1354930 34. Mumba Ngoyi D, Ali Ekangu R, Mumvemba Kodi MF, Pyana PP, Balharbi F, et al. (2014) Performance of parasitological and molecular techniques for the diagnosis and surveillance of gambiense sleeping sickness. PLoS Negl Trop Dis 8: e2954. https://doi.org/10.1371/journal.pntd.0002954 PMID: 24921941 20 / 20
https://openalex.org/W4226193026	https://ora.ox.ac.uk/objects/uuid:ab816b39-076c-4125-9514-6b8cf238cef5/files/rh128nf30f	English	null	The Digital Brain Bank, an open access platform for post-mortem imaging datasets	eLife	2,022	cc-by	21,642	The Digital Brain Bank, an open access platform for post-mortem imaging datasets Benjamin C Tendler1, Taylor Hanayik1, Olaf Ansorge2, Benjamin C Tendler1, Taylor Hanayik1, Olaf Ansorge2, Sarah Bangerter-Christensen2, Gregory S Berns3, Mads F Bertelsen4, Katherine L Bryant1, Sean Foxley1,5, Martijn P van den Heuvel6,7, Amy FD Howard1, Istvan N Huszar1, Alexandre A Khrapitchev8, Anna Leonte2, Paul R Manger9, Ricarda AL Menke1, Jeroen Mollink1, Duncan Mortimer1, Menuka Pallebage-Gamarallage2, Lea Roumazeilles10, Jerome Sallet10,11, Lianne H Scholtens6, Connor Scott2, Adele Smart1,2, Martin R Turner1,2, Chaoyue Wang1, Saad Jbabdi1†, Rogier B Mars1,12†, Karla L Miller1† 1Wellcome Centre for Integrative Neuroimaging, FMRIB, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, United Kingdom; 2Division of Clinical Neurology, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, United Kingdom; 3Psychology Department, Emory University, Atlanta, United States; 4Centre for Zoo and Wild Animal Health, Copenhagen Zoo, Frederiksberg, Denmark; 5Department of Radiology, University of Chicago, Chicago, United States; 6Department of Complex Trait Genetics, Centre for Neurogenomics and Cognitive Research, Amsterdam Neuroscience, Vrije Universiteit Amsterdam, Amsterdam, Netherlands; 7Department of Child Psychiatry, Amsterdam Neuroscience, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, Netherlands; 8Medical Research Council Oxford Institute for Radiation Oncology, University of Oxford, Oxford, United Kingdom; 9School of Anatomical Sciences, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa; 10Wellcome Centre for Integrative Neuroimaging, Department of Experimental Psychology, University of Oxford, Oxford, United Kingdom; 11Stem Cell and Brain Research Institute, Université Lyon 1, INSERM, Bron, France; 12Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen, Nijmegen, Netherlands For correspondence: benjamin.tendler@ndcn.ox.ac. uk (BCT); karla.miller@ndcn.ox.ac.uk (KLM) †These authors contributed equally to this work Competing interest: See page 18 Funding: See page 19 Preprinted: 22 June 2021 Received: 19 August 2021 Accepted: 17 March 2022 Published: 17 March 2022 For correspondence: benjamin.tendler@ndcn.ox.ac. uk (BCT); karla.miller@ndcn.ox.ac.uk (KLM) †These authors contributed equally to this work Competing interest: See page 18 Funding: See page 19 Preprinted: 22 June 2021 Received: 19 August 2021 Accepted: 17 March 2022 Published: 17 March 2022 For correspondence: benjamin.tendler@ndcn.ox.ac. uk (BCT); karla.miller@ndcn.ox.ac.uk (KLM) †These authors contributed equally to this work Competing interest: See page 18 Abstract Post-mortem magnetic resonance imaging (MRI) provides the opportunity to acquire high-resolution datasets to investigate neuroanatomy and validate the origins of image contrast through microscopy comparisons. We introduce the Digital Brain Bank (open.win.ox.ac.uk/Digi- talBrainBank), a data release platform providing open access to curated, multimodal post-mortem neuroimaging datasets. Tools and Resources The Digital Brain Bank, an open access platform for post-mortem imaging datasets Datasets span three themes—Digital Neuroanatomist: datasets for detailed neuroanatomical investigations; Digital Brain Zoo: datasets for comparative neuroanatomy; and Digital Pathologist: datasets for neuropathology investigations. The first Digital Brain Bank data release includes 21 distinctive whole-brain diffusion MRI datasets for structural connectivity inves- tigations, alongside microscopy and complementary MRI modalities. This includes one of the highest-resolution whole-brain human diffusion MRI datasets ever acquired, whole-brain diffusion MRI in fourteen nonhuman primate species, and one of the largest post-mortem whole-brain cohort imaging studies in neurodegeneration. The Digital Brain Bank is the culmination of our lab’s invest- ment into post-mortem MRI methodology and MRI-microscopy analysis techniques. This manuscript Introduction Magnetic resonance imaging (MRI) occupies a unique position in the neuroscience toolkit. In humans, MRI is used at the single-subject level diagnostically and is increasingly deployed at the population level in epide- miology (Marcus et al., 2007; Miller et al., 2016; Snoek et al., 2021; Van Essen et al., 2013). MRI is well- established in the context of imaging causal manipulations in experimental organisms ranging from mice (Denic et al., 2011; Thiessen et al., 2013) to nonhuman primates (Absinta et al., 2017; Klink et al., 2021) and provides precise measurements in cellular and tissue preparations (Wilhelm et al., 2012). This extensive landscape of overlap with the broader neuroscience toolkit creates the potential for MRI to facilitate inte- gration between technologies and investigations. Although MRI hardware and acquisition protocols often need to be tailored to a specific domain, the underlying technology associated with all MRI measurements gives rise to a common set of signal forming mechanisms, facilitating cross-domain comparisons. There are few methods available to neuroscientists that span this breadth of domains. One challenge to the use of MRI as a bridging technology is the need for common measurements — for example, the same MRI measurements made across multiple species, or MRI and microscopy measurements in the same brain tissue (Mars et al., 2021). Post-mortem MRI provides unique oppor- tunities for such common measurements. MRI in post-mortem tissue can be used to identify the origins of image contrast through integration with microscopy (Keren et al., 2015; Langkammer et al., 2012; Mollink et al., 2017), directly addressing concerns over the nonspecificity of MRI signals. In this context, post-mortem MRI data are important because they share common signal forming mechanisms with in vivo MRI and a common tissue state with microscopy, providing a framework for investigation across multiple spatial scales. Post-mortem MRI facilitates comparative anatomy inves- tigations in species that are not traditionally accessible for in vivo imaging (Berns et al., 2015; Bhag- wandin et al., 2017; Grewal et al., 2020; Heuer et al., 2019), including extinct species (Berns and Ashwell, 2017). Long post-mortem scans provide the opportunity to push the boundaries of spatial resolution, providing whole human brain coverage reaching voxel sizes of 100–500 μm (Edlow et al., 2019; Foxley et al., 2016; Fritz et al., 2019; Weigel et al., 2021), edging closer to microscopy tech- niques but benefitting from compatibility with in vivo imaging. Editor's evaluation This paper describes a new open-access digital brain bank of post-mortem brains that have been scanned with high-resolution, multimodal magnetic resonance imaging and with select datasets accompanied by histological data. This valuable resource can be used to study healthy human brains, pathological human brains, and the brains of other species, opening new opportunities for comparative neuroanatomy and the biological validation of non-invasive neuroimaging signals. Reviewing Editor: Alex Fornito, M h U i it A t li Reviewing Editor: Alex Fornito, Monash University, Australia Reviewing Editor: Alex Fornito, Monash University, Australia ‍ ‍ Copyright Tendler et al. This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited. Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 1 of 35 Tools and resources Tools and resources Neuroscience provides a detailed overview of our work with post-mortem imaging to date, including the devel- opment of diffusion MRI methods to image large post-mortem samples, including whole, human brains. Taken together, the Digital Brain Bank provides cross-scale, cross-species datasets facilitating the incorporation of post-mortem data into neuroimaging studies. provides a detailed overview of our work with post-mortem imaging to date, including the devel- opment of diffusion MRI methods to image large post-mortem samples, including whole, human brains. Taken together, the Digital Brain Bank provides cross-scale, cross-species datasets facilitating the incorporation of post-mortem data into neuroimaging studies. Introduction As a nondestructive technique, post- mortem MRI enables the examination of tissue microstructure whilst preserving tissue, facilitating repeat MRI measurements with novel contrasts and technologies; and more generally, its integration with tools for post-mortem investigations (e.g., histopathology or proteomics). In this work, we introduce the Digital Brain Bank (open.win.ox.ac.uk/DigitalBrainBank), a data release platform resulting from a decade of post-mortem MRI research at the University of Oxford. The Digital Brain Bank provides open access to several post-mortem neuroimaging datasets spanning investigations into human neuroanatomy, cross-species neuroanatomy, and neuropathology. All data- sets provide post-mortem MRI, including diffusion MRI, with complementary microscopy data (e.g., immunohistochemistry or PLI) included with some datasets. Our post-mortem imaging research has been specifically aimed at achieving whole-brain post- mortem MRI to support the investigation of multiple brain systems/regions and long-range connec- tions (Foxley et al., 2014; Miller et al., 2011; Miller et al., 2012), and the first release to the Digital Brain Bank contains 21 distinct whole-brain post-mortem MRI datasets, including from whole human brains. All datasets are available to access, and prospective users of the Digital Brain Bank can explore Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 2 of 35 Tools and resources Tools and resources Neuroscience a subset of data directly on the Digital Brain Bank website using Tview: a bespoke, open-source, and web-based image viewer. Tview has been developed for efficient browsing of imaging data spanning drastically different spatial scales, from submicron resolution microscopy to millimeter MRI acquisi- tions. It enables real-time visualization and interaction (zooming/panning) of both MRI and microscopy images, and with flexible overlays of different modalities. All datasets uploaded to the Digital Brain Bank are associated with researchers at the University of Oxford, or from close collaborators. Limited Derived Outputs from users of Digital Brain Bank datasets will also be considered for data upload. The first release to the Digital Brain Bank includes data from multiple published projects covering a breadth of neuroimaging research, including whole-brain diffusion MRI in 14 nonhuman primate species (Bryant et al., 2021; Roumazeilles et al., 2020; Roumazeilles et al., 2021), and one of the largest post-mortem whole-brain cohort imaging studies combining whole-brain MRI and microscopy in human neurodegeneration (Pallebage-Gamarallage et al., 2018). In addition, we present a previously unpublished project providing one of the highest-resolution whole-brain human diffusion MRI datasets ever acquired (500 μm isotropic resolution). Results The Digital Brain Bank is accessible at open.win.ox.ac.uk/DigitalBrainBank. Datasets have been orga- nized into categories reflecting three predominant themes of post-mortem neuroimaging research: • Digital Anatomist: datasets for detailed neuroanatomical investigations. • Digital Anatomist: datasets for detailed neuroanatomical investigations. • Digital Brain Zoo: datasets for comparative neuroanatomy. l h l d f h l • Digital Brain Zoo: datasets for comparative neuroanatomy. • Digital Pathologist: datasets for neuropathology investigations. • Digital Pathologist: datasets for neuropathology investigations. Here, we provide an overview of each theme, with examples from available datasets in the first release to the Digital Brain Bank. A brief description of all the datasets provided with the first release, alongside relevant publications, is provided in Table 1. Introduction The Digital Brain Bank will continue to grow over the coming years, with a number of further datasets already at the early stages of curation (Howard et al., 2019a; Martins- Bach et al., 2021; Martins-Bach et al., 2020; Wu et al., 2021). Digital Anatomist the first release were acquired using a balanced SSFP (bSSFP) or T2-weighted sequence, which yields strong gray-white matter contrast m tissue. Diffusion MRI datasets were acquired using a combination of diffusion-weighted steady-state free precession (DW-SSFP) echo (DW-SE) sequences. Full details of the motivation behind the choice of sequences and available contrasts are described in the c susceptibility maps are currently available in 9 out of 12 ALS brains and all control brains. The remaining datasets were either lost of insufficient data quality for public release. Contents: MRI Resolution (MRI) Contents: Microscopy Relevant publications h-Resolution RI-PLI Whole-brain diffusion MRI, structural MRI, quantitative T1 and T2 maps: – Control human brain: 1× Diffusion MRI: (500 μm, 1 and 2 mm iso.) Structural MRI: 312.5×312.5×500 μm3 T1 map: (0.75×0.75×1.6 mm3) T2 map: (0.75×0.75×1.6 mm3) Polarised light imaging (4 μm in-plane) in the anterior commissure, corpus callosum, pons, thalamus, and visual cortex (same brain) Dataset described in this publication (Methodology in Appendix 1), Diffusion MRI processing described in Tendler et al., 2020b, T2 mapping described in Tendler et al., 2021 osum MRI- Corpus callosum diffusion MRI: – Excised control human corpus callosum samples: Polarised light imaging (4 μm in-plane), bright-field microscopy images of immunohistochemistry stains (0.25 μm in- plane) for PLP (myelin) and GFAP (astrocyte) et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 Table 1. Description of all datasets provided in the first release to the Digital Brain Bank. All Structural MRI datasets in the first release were acquired using a balanced SSFP (bSSFP) or T2-weighted sequence, which yields strong gray-white matter contrast in formalin-fixed post-mortem tissue. Diffusion MRI datasets were acquired using a combination of diffusion-weighted steady-state free precession (DW-SSFP) and diffusion-weighted spin-echo (DW-SE) sequences. Full details of the motivation behind the choice of sequences and available contrasts are described in the Discussion. †T2* and magnetic susceptibility maps are currently available in 9 out of 12 ALS brains and all control brains. The remaining datasets were either lost during scanner export, or are of insufficient data quality for public release. Digital Anatomist Category Name Contents: MRI Resolution (MRI) Contents: Microscopy Relevant publications Digital Anatomist Human High-Resolution Diffusion MRI-PLI Whole-brain diffusion MRI, structural MRI, quantitative T1 and T2 maps: – Control human brain: 1× Diffusion MRI: (500 μm, 1 and 2 mm iso.) Structural MRI: 312.5×312.5×500 μm3 T1 map: (0.75×0.75×1.6 mm3) T2 map: (0.75×0.75×1.6 mm3) Polarised light imaging (4 μm in-plane) in the anterior commissure, corpus callosum, pons, thalamus, and visual cortex (same brain) Dataset described in this publication (Methodology in Appendix 1), Diffusion MRI processing described in Tendler et al., 2020b, T2 mapping described in Tendler et al., 2021 Digital Anatomist Human Callosum MRI- PLI-Histology Corpus callosum diffusion MRI: – Excised control human corpus callosum samples: 3× Diffusion MRI: (400 μm iso.) Polarised light imaging (4 μm in-plane), bright-field microscopy images of immunohistochemistry stains (0.25 μm in- plane) for PLP (myelin) and GFAP (astrocyte) (same human corpus callosum samples) Mollink et al., 2017 Whole-brain diffusion MRI and structural MRI (available in brains marked with a ): • Bushbaby (Galago senegalensis): 1× • Capuchin monkey (Sepajus apella): 1× • Chimpanzee (Pan troglodytes): 2× • Colobus monkey (Colobus guereza): 1× • Cotton-Top tamarin (Saguinus oedipus): 1× • Golden Lion tamarin (Leontopithecus rosalia): 1× • Hamadryas baboon* (Papio hamadryas): 1× • Macaque monkey (Macaca mulatta): 3× • Mangabey (Lophocebus albigena): 1× • Night monkey, (Aotus lemurinus): 1× • Ring-tailed lemur (Lemur catta): 3× • Saki monkey (Pithecia pithecia): 1× • Western Lowland gorilla* (Gorilla gorilla): 1× • Woolly monkey (Lagothrix lagotricha): 1× Diffusion MRI: 300 μm iso.: Bushbaby, Cotton-Top tamarin & Golden Lion Tamarin 400 μm iso: Night monkey 500 μm iso: Ring-tailed lemur and Saki monkey 600 μm iso: Capuchin monkey, Chimpanzee, Colobus monkey, Hamadryas baboon, Macaque monkey, Mangabey, Western Lowland Gorilla and Woolly Monkey Structural MRI 200 μm iso: Western Lowland Gorilla 220 μm iso: Hamadryas Baboon 0.22×0.22×0.19 mm3: 1× Chimpanzee 0.375×0.375×0.40 mm3: 1× Chimpanzee None 1× Western Lowland gorilla and 1× Chimpanzee described in Roumazeilles et al., 2020, 3× Macaque monkey and 3× Ring-Tailed Lemur described in Roumazeilles et al., 2021. Hamadryas baboon, Cotton-Top tamarin and Golden Lion tamarin datasets described in this publication (Methodology in Appendix 1). Digital Anatomist Datasets within the Digital Anatomist provide a new direction for answering fundamental questions in neuroanatomy, through ultra-high resolution MRI data and complementary microscopy within the same sample in humans and model nonhuman species. The long scan times available in post-mortem MRI affords imaging at ultra-high spatial resolutions, facil- itating the delineation of small tissue structures within the human brain, one of the key aims of the Digital Anatomist. Often, post-mortem investigations are limited to small sections of excised brain tissue that repre- sent a limited anatomical region. However, our developments in whole-brain post-mortem diffusion imaging (Foxley et al., 2014; McNab et al., 2009; Miller et al., 2011; Miller et al., 2012; Tendler et al., 2020b) provide the opportunity to investigate structural connectivity and gross neuroanatomy, at scales that are unobtainable in vivo. These developments have culminated in the Human High-Resolution Diffusion MRI-PLI dataset, providing one of the highest-resolution whole-brain human diffusion MRI datasets ever acquired (500 μm isotropic resolution), as shown in Figure 1. Companion datasets acquired at 1 mm and 2 mm (isotropic) provide a comparison at cutting-edge and conventional in vivo resolutions (Figure 1a). In addition to providing a new insight into human neuroanatomy, these data can be used to inform experimental design and the interpretation of results. Here, the Human High-Resolution Diffusion MRI-PLI dataset enables users to identify the resolution required to visualize certain brain structures (Figure 1a), and how spatial resolution impacts tractography performance (e.g., overcoming ‘gyral bias’—Figure 1b; Cottaar et al., 2021; Schilling et al., 2018). Polarised light imaging (PLI) provides estimates of myelinated fiber orientation (Axer et al., 2011), and complementary PLI data acquired in a subset of brain regions (4 μm in-plane) facilitates cross-scale comparisons (Figure 1b and c). A further aim of the Digital Anatomist is to perform quantitative validations across modalities, relating MRI to microscopic measures. These kinds of analyses can only be achieved with accurately coregistered data, enabling pixel-wise comparisons across modalities acquired at drastically different spatial resolutions. This potential is most clearly seen in the Human Callosum MRI-PLI-Histology dataset, which provides diffusion MRI (400 μm isotropic), alongside complementary PLI (4 μm in-plane), and histology (myelin and astrocytes) (0.25 μm in-plane) in three excised human corpus callosum samples Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 3 of 35 Tools and resources tasets provided in the first release to the Digital Brain Bank. Digital Anatomist All other datasets described in Bryant et al., 2021 Digital Brain Zoo Marsupials Whole-brain diffusion MRI and structural MRI: • Tasmanian devil (Sarcophilus harrisii): 2× • Thylacine (Thylacinus cynocephalus): 2× Diffusion MRI: 1 mm iso: 1× Tasmanian devil 1.5 mm iso: 1× Tasmanian devil 1.1 mm iso: 1× Thylacine 1.0×1.1×0.8 mm3: 1× Thylacine Structural MRI 330 μm iso: 1× Tasmanian devil and 1× Thylacine 330×330×300 μm3: 1× Tasmanian devil 500 μm iso: 1× Thylacine None Berns and Ashwell, 2017 Digital Brain Zoo Cetaceans Whole-brain diffusion MRI and structural MRI • Common dolphin (Delphinus delphis): 1× • Pantropical dolphin (Stenella attenuata): 1× Diffusion MRI: (1.3 mm iso.) Structural MRI: (640×640×500 μm3) None Berns et al., 2015 Digital Brain Zoo Carnivora Whole-brain diffusion MRI and structural MRI: – European wolf (Canis lupus): 1× Diffusion MRI: (600 μm iso.) Structural MRI: (220 μm iso.) None Dataset described in this publication (Methodology in Appendix 1) Table 1 continued on next page Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 4 of 35 Tools and resources Tools and resources Neuroscience org/10.7554/eLife.73153 Category Name Contents: MRI Resolution (MRI) Contents: Microscopy Relevant publications Digital Pathologist Human ALS MRI- Histology Whole-brain diffusion MRI, structural MRI, quantitative T1, T2, and T2* maps, magnetic susceptibility maps (selected brains†): • Amyotrophic lateral sclerosis (ALS) human brains: 12× • Control human brains: 3× Diffusion MRI: (850 μm iso.) Structural MRI: (230–250 μm in-plane; 270–500 μm slice) T1 map: (0.65–1 mm in-plane; 0.90– 1.6 mm slice) T2 map: (0.65–1 mm in-plane; 0.90– 1.6 mm slice) T2/magnetic susceptibility maps: (0.5 mm in-plane; 1.1–1.3 mm slice) Bright-field microscopy immunohistochemistry stains (0.50 μm in-plane, exception pTDP43 – 0.25 μm in-plane): pTDP-43, IBA1 (pan microglia), CD68 (activated microglia/ macrophages), PLP (myelin), SMI-312 (axonal phosphorylated neurofilaments), and ferritin (iron storage, subset of regions) Regions: Anterior cingulate cortex, corpus callosum, hippocampus, primary motor cortex, and visual cortex (same brains). Selected multimodal histology available in two brains (1× ALS and 1× Control), and multiregional PLP (available in 10 out of 12 ALS brains and all control brains, 5–8 regions per brain) in first data release – remaining histology being actively curated. Digital Anatomist Pallebage-Gamarallage et al., 2018, Magnetic susceptibility and T2 mapping protocol described in Wang et al., 2020, Diffusion MRI processing described in Tendler et al., 2020b, T2 mapping described in Tendler et al., 2021 Table 1 continued /10 7554/ Lif 73153 Category Name Contents: MRI Resolution (MRI) Contents: Microscopy Relevant publications Digital Pathologist Human ALS MRI- Histology Whole-brain diffusion MRI, structural MRI, quantitative T1, T2, and T2* maps, magnetic susceptibility maps (selected brains†): • Amyotrophic lateral sclerosis (ALS) human brains: 12× • Control human brains: 3× Diffusion MRI: (850 μm iso.) Structural MRI: (230–250 μm in-plane; 270–500 μm slice) T1 map: (0.65–1 mm in-plane; 0.90– 1.6 mm slice) T2 map: (0.65–1 mm in-plane; 0.90– 1.6 mm slice) T2/magnetic susceptibility maps: (0.5 mm in-plane; 1.1–1.3 mm slice) Bright-field microscopy immunohistochemistry stains (0.50 μm in-plane, exception pTDP43 – 0.25 μm in-plane): pTDP-43, IBA1 (pan microglia), CD68 (activated microglia/ macrophages), PLP (myelin), SMI-312 (axonal phosphorylated neurofilaments), and ferritin (iron storage, subset of regions) Regions: Anterior cingulate cortex, corpus callosum, hippocampus, primary motor cortex, and visual cortex (same brains). Selected multimodal histology available in two brains (1× ALS and 1× Control), and multiregional PLP (available in 10 out of 12 ALS brains and all control brains, 5–8 regions per brain) in first data release – remaining histology being actively curated. Pallebage-Gamarallage et al., 2018, Magnetic susceptibility and T2 mapping protocol described in Wang et al., 2020, Diffusion MRI processing described in Tendler et al., 2020b, T2 mapping described in Tendler et al., 2021 Table 1 continued Relevant publications Table 1 continued 5 of 35 5 of 35 Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 Tools and resources Neuroscience Figure 1. The Digital Anatomist. (a) Whole-brain diffusion MRI data available in the Human High-Resolution Diffusion MRI-PLI dataset reveals the wealth of information provided at increased spatial scales, one of the key aims of the Digital Anatomist. Here, the 500 μm dataset uncovers the information hidden at lower spatial resolutions, for example, visualizing the interdigitating transverse pontine fibers with the corticospinal tract or striations through the internal capsule. (b) Similarly, datasets across multiple spatial scales can inform us of the limitations at reduced imaging resolutions. Here, gyral tractography (occipital lobe) reveals an overall pattern of fibers turning into the gyral bank at 0.5 mm. Digital Anatomist At 1 mm, an underestimation of connectivity at the gyral banks is observed, known as the ‘gyral bias’ (Cottaar et al., 2021; Schilling et al., 2018). At 2 mm, tractography bears little resemblance to the expected architecture. Multimodal comparisons enable us to validate our findings, with complementary polarised light imaging (PLI) data at over 2 orders of magnitude increase in resolution (125×) revealing a similar pattern of gyral connectivity, and (c) excellent visual agreement with tractography across the pons. (a) displays diffusion tensor principal diffusion direction maps (modulated by fractional anisotropy). Figure 1. The Digital Anatomist. (a) Whole-brain diffusion MRI data available in the Human High-Resolution Diffusion MRI-PLI dataset reveals the wealth Figure 1. The Digital Anatomist. (a) Whole-brain diffusion MRI data available in the Human High-Resolution Diffusion MRI-PLI dataset reveals the wealth of information provided at increased spatial scales, one of the key aims of the Digital Anatomist. Here, the 500 μm dataset uncovers the information hidden at lower spatial resolutions, for example, visualizing the interdigitating transverse pontine fibers with the corticospinal tract or striations through the internal capsule. (b) Similarly, datasets across multiple spatial scales can inform us of the limitations at reduced imaging resolutions. Here, gyral tractography (occipital lobe) reveals an overall pattern of fibers turning into the gyral bank at 0.5 mm. At 1 mm, an underestimation of connectivity at the gyral banks is observed, known as the ‘gyral bias’ (Cottaar et al., 2021; Schilling et al., 2018). At 2 mm, tractography bears little resemblance to the expected architecture. Multimodal comparisons enable us to validate our findings, with complementary polarised light imaging (PLI) data at over 2 orders of magnitude increase in resolution (125×) revealing a similar pattern of gyral connectivity, and (c) excellent visual agreement with tractography across the pons. (a) displays diffusion tensor principal diffusion direction maps (modulated by fractional anisotropy). (Mollink et al., 2017). These data offer multiple pathways of investigation, including the identification of the origins of image contrast; validation of microstructural models of tissue (Mollink et al., 2017); and developing unique models explicitly linking MRI with microscopy (Howard et al., 2019b). Digital Pathologist Datasets within the Digital Pathologist provide a new direction for examining neuropathology and MRI-pathology correlates in humans and established laboratory models. One of the biggest challenges in the use of MRI clinically is the lack of specificity to disease mech- anisms. Many neurological diseases are characterized by changes at the cellular and subcellular level, which cannot be directly visualized with the limited resolution of MRI. Nevertheless, MRI contrast can be made sensitive to cellular-level phenomena that are relevant to disease. Acquisition of MRI and histology in the same tissue enables us to relate microscopic changes in the neural microenvironment to MRI image contrast. The primary aim of the Digital Pathologist is to facilitate these cross-scale comparisons, imaging brain tissue associated with a neurological disease. Such data are provided in the Human ALS MRI-Histology dataset (Figure 3a), which aims to iden- tify how neuropathological changes in amyotrophic lateral sclerosis (ALS) give rise to altered MRI contrast, and answer specific questions related to ALS pathology. The Human ALS MRI-Histology dataset provides whole-brain multimodal MRI and selective histology in a cohort of 12 ALS (diag- nosis during lifetime, confirmed ALS neuropathology) and 3 control (no known neuropathology) brains (Pallebage-Gamarallage et al., 2018) provided by the Oxford Brain Bank. MRI data includes diffu- sion, structural, quantitative susceptibility maps (via quantitative susceptibility mapping, QSM), and quantitative T1, T2, and T2* maps. Histology includes markers for proteinopathy (pTDP-43), microglia (CD68 and IBA1), myelin (PLP), neurofilaments (SMI-312), and iron (ferritin) in order to detect changes in a range of microstructures within cortical and subcortical regions (anterior cingulate cortex, corpus callosum, hippocampus, primary motor cortex, and visual cortex) associated with different proposed stages of ALS disease progression (Jucker and Walker, 2013). Different MRI modalities have known sensitivities to different components of the cellular environ- ment. Combined with multimodal histology, these data provide the opportunity to relate neuropatho- logically induced changes in tissue microstructure to MRI image contrast. While these aims could be partially achieved by dissecting and scanning subregions of the brain, our approach of scanning whole post-mortem brains enables us to investigate neuropathological spread across the entire brain (Jucker and Walker, 2013). This facilitates investigations across long-range fiber-tracts associated with pathology (Figure 3b), or microstructural changes in multiple brain regions (Figure 3c). Digital Brain Zoo The Digital Brain Zoo provides curated datasets to investigate neuroanatomy in nonhuman species and compare anatomy across species. Post-mortem MRI has enormous potential to inform comparative neuroanatomy for three reasons. First, it enables the scanning of species that would be extremely difficult or impossible to study in vivo. Second, samples can be imaged with minimal handling and without invasive procedures, enabling the study of rare specimens that would not be appropriate to dissect. Third, MRI investigations can be performed in whole-brain samples, rather than excised tissue sections. This makes post-mortem MRI ideally placed to Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 6 of 35 Tools and resources Tools and resources Neuroscience characterize macroscopic brain structure, long-range structural connectivity, and tissue microstructure in species that are not traditional experimental models, and in particular rare species where very few brain samples may be available (Berns and Ashwell, 2017; Bhagwandin et al., 2017; Grewal et al., 2020; Mars et al., 2014). MRI data from multiple species allows one to formally compare brain organization, important for large-scale comparative neuroscience which has traditionally relied on very limited measures (e.g., whole or regional brain size measures of brain organization) (Mars et al., 2014). The ability to acquire data from whole brains opens up the possibility of elucidating principles of neural diversity across larger orders of mammalian species (Friedrich et al., 2021), and create between-species mappings to formally identify homologies and quantify unique aspects of any given brain (Mars et al., 2018). This also allows one to improve translational neuroscience by better understanding the relationship between the human brain and that of model species (e.g., macaque, marmoset, rat, and mouse). The Digital Brain Zoo provides access to post-mortem imaging datasets in nonhuman species covering multiple taxonomic ranks (Figure 2a), including nonhuman primate species (Bryant et al., 2021; Rouma- zeilles et al., 2020; Roumazeilles et al., 2021), Carnivora (Grewal et al., 2020), Marsupials (Berns and Ashwell, 2017), and Cetaceans (Berns et al., 2015). As with other collections in the Digital Brain Bank, the Digital Brain Zoo currently focuses primarily on whole-brain diffusion MRI. These datasets offer multiple pathways of investigation in comparative neuroanatomy, for example, through the examination of structural connections across brains (Figure 2b; Bryant et al., 2021). Digital Brain Zoo Furthermore, our developments in imaging large post-mortem samples have enabled us to acquire several high-quality post-mortem imaging datasets in species with brains that are too large to fit into specialized preclinical MRI systems, conventionally used to improve image quality in post-mortem MRI (see Discussion). Digital Pathologist Notably, these analyses are being facilitated by accurate cross-modality image coregistrations (Huszar et al., 2019), enabling us to perform pixel-wise evaluations and integrate structural analyses to identify how pathology influences MR image contrast (Figure 3d) in a subset of brain regions associated with Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 7 of 35 Tools and resources Neuroscience Figure 2. The Digital Brain Zoo. (a) The first release of the Digital Brain Zoo provides whole-brain MRI datasets spanning multiple species and axonomic ranks. Notably, we provide whole-brain diffusion MRI datasets from 14 nonhuman primate species, with samples selected for their high quality and to ensure sampling of all major branches of the primate evolutionary tree (Prosimian, New World monkey, Old World monkey, and Great Ape). (b) compares the relative volume of four tracts derived from nine nonhuman primate post-mortem datasets provided in the Digital Brain Zoo Bryant et al., 2021), where increased distance from the centre corresponds to an increased volume. Figure 2. The Digital Brain Zoo. (a) The first release of the Digital Brain Zoo provides whole-brain MRI datasets spanning multiple species and taxonomic ranks. Notably, we provide whole-brain diffusion MRI datasets from 14 nonhuman primate species, with samples selected for their high quality and to ensure sampling of all major branches of the primate evolutionary tree (Prosimian, New World monkey, Old World monkey, and Great Ape). (b) compares the relative volume of four tracts derived from nine nonhuman primate post-mortem datasets provided in the Digital Brain Zoo (Bryant et al., 2021), where increased distance from the centre corresponds to an increased volume. Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 8 of 35 Tools and resources Neuroscience Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 Figure 3. The Digital Pathologist. One of the key aims of the Digital Pathologist is the examination of neuropathological spread The Human ALS MRI-Histology dataset (a) facilitates these investigations, combining whole-brain multimodal MRI and histology n a cohort of 12 ALS and 3 control brains. (b) Displays the reconstruction of five white matter pathways associated with different post-mortem brain (Kassubek et al., 2014). Comparisons between ALS and control brains over the corpus callosum of the coh Figure 3 continued on next page Figure 3. The Digital Pathologist. One of the key aims of the Digital Pathologist is the examination of neuropathological spread in neurological disease. Figure 3 continued Figure 3 continued in fractional anisotropy (FA, normalized to Par/Temp/Occ lobe), with biggest changes associated with motor and prefrontal regions (Hofer and Frahm, 2006) (=p<0.05; =p<0.05 following multiple comparison correction) (full details of the corpus callosum analysis provided in Appendix 2). This reflects the anticipated changes in ALS with brain regions associated with motor function, in good agreement with a previous study (Chapman et al., 2014), which identified the greatest FA difference between ALS and controls in these regions. Accurate MRI-histology coregistrations facilitate cross-modality comparisons, and (d) displays an example of MRI-histology coregistration over the visual cortex of a single ALS brain achieved using the Tensor Image Registration Library (TIRL) (Huszar et al., 2019). V1=principal diffusion direction, FA=fractional anisotropy, MD=mean diffusivity, D⊥=radial diffusivity, MO=mode from diffusion tensor output, Dyad1=principal dyad orientation, f1=principal fiber fraction and D=diffusivity from Ball and Two-Stick output, swMRI=susceptibility-weighted MRI, χ=magnetic susceptibility. Details of stain contrasts in (b) and (d) are provided in Table 1. ALS, amyotrophic lateral sclerosis; MRI, magnetic resonance imaging. different proposed stages of ALS disease progression (Jucker and Walker, 2013). MRI data for the Human ALS MRI-Histology dataset for all 12 ALS and 3 control brains are immediately available to access, alongside a subset of histology data. Remaining histology data and MRI-histology coregistra- tions are being actively curated for future release to the Digital Brain Bank. different proposed stages of ALS disease progression (Jucker and Walker, 2013). MRI data for the Human ALS MRI-Histology dataset for all 12 ALS and 3 control brains are immediately available to access, alongside a subset of histology data. Remaining histology data and MRI-histology coregistra- tions are being actively curated for future release to the Digital Brain Bank. Digital Pathologist The Human ALS MRI-Histology dataset (a) facilitates these investigations, combining whole-brain multimodal MRI and histology (selected brain regions) in a cohort of 12 ALS and 3 control brains. (b) Displays the reconstruction of five white matter pathways associated with different ALS stages in a single post-mortem brain (Kassubek et al., 2014). Comparisons between ALS and control brains over the corpus callosum of the cohort (c) reveals changes Figure 3 continued on next page Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 9 of 35 Tools and resources Tview The Digital Brain Bank website enables users to browse a subset of data easily. A key feature of many datasets is that they contain both MRI and microscopy data. Few available viewers, whether downloadable or online, support both MRI and microscopy file formats, creating a barrier of entry for potential users. Moreover, 2D microscopy datasets are extremely high resolution: single images can exceed 10,000,000,000 pixels, running to gigabytes in size. We aim to provide online viewing of microscopy and MRI data using standard internet browsers, with a viewer that can handle data at very different spatial scales, provide flexible image overlays, and support color visualization for diffusion-derived measures, PLI, and multiple histological counterstains. Unable to identify existing software with these features, we developed a web-based image viewer, Tview. Tview is based on software originally used to display satellite imagery at multiple elevations, and enables real-time visualization, interaction (zooming/panning), and flexible overlays of different modalities in a single 2D plane of MRI and microscopy data. Visualization of multimodal (i.e., MRI and microscopy) datasets on the Digital Brain Bank website is achieved with Tview. An example Tview implementation is available at open.win.ox. ac.uk/DigitalBrainBank/#/tileviewer, where cross-modality coregistrations were performed using the Tensor Image Registration Library (TIRL) (Huszar et al., 2019) and FNIRT (Andersson et al., 2007; Jenkinson et al., 2012), both available as part of FSL. Code for Tview, the website, and server implementation are available at https://git.fmrib.ox.ac.uk/thanayik/dbb. The benefits of Tview are most readily realized with datasets incorporating MRI and microscopy images, enabling visualization of distinct contrasts over multiple spatial scales. However, many datasets provided in the first release to the Digital Brain Bank do not contain any microscopy data. For these datasets, a detailed static image is currently used for visualization on the Digital Brain Bank website. Tools and resources Tools and resources Neuroscience Discussion The Digital Brain Bank represents one of the most substantial resources of its kind, providing data from 45 brains across all three themes in the first release. It has been specifically designed to cover the breadth of spatial scales and modalities encountered in post-mortem imaging. Features on the website facilitate data discovery, with users able to interact with a subset of available datasets prior to access. The Digital Brain Bank is envisioned as a growing resource reflecting a range of post-mortem neuroimaging projects. Alongside the first release, we aim to first bring together datasets that have been accumulated over the past decade at the University of Oxford. Beyond this, the Digital Brain Bank will be the primary resource to release new post-mortem imaging datasets associated with both departmental and collaborative projects. The Digital Brain Bank is a comprehensive resource focusing on post-mortem MRI spanning multiple investigative themes in neuroanatomy, neuropathology, and comparative neuroanatomy. Given this broad coverage, we anticipate that datasets provided by the Digital Brain Bank will complement existing open- science initiatives in both human and nonhuman neuroimaging. There are several existing resources providing outputs derived from post-mortem data focusing on other domains, including the Allen Brain Map (tran- scriptomics) (https://portal.brain-map.org/), the BigBrain Project (histology) (https://bigbrainproject.org/), and databases compiling datasets from multiple sources such as EBRAINS (https://ebrains.eu/). Primarily, we foresee the greatest overlap and integration between the Digital Brain Bank and existing databases for in vivo and post-mortem MRI. For example, the high-resolution diffusion MRI datasets provided by the Digital Anatomist complement the aims of existing studies such as the Human Connectome Project (Van Essen et al., 2013), providing the opportunity to validate in vivo findings with higher spatial resolution. The Digital Brain Zoo’s current focus on nonhuman primates complements several existing in vivo and post- mortem MRI databases, including PRIME-DE (Milham et al., 2018) and the JMC Primates Brain Imaging Repository (Sakai et al., 2018). The multiple taxonomic ranks covered by the Digital Brain Zoo draws direct parallels with resources such as the Brain Catalogue (Toro et al., 2014), which provides nonhuman post- mortem MRI datasets for structural investigations; and the mammalian MRI database (Assaf et al., 2020), containing diffusion and T2-/T1-weighted scans of 123 different species (datasets available on request as described in Assaf et al., 2020). Requirements for data access and referencing datasets Requirements for data access and referencing datasets The Digital Brain Bank has been designed to minimize the burden on the user to access datasets, within ethical constraints. For many datasets, we have developed conditions of use terms via a material transfer agreement (MTA), which users agree to prior to access. The MTAs are primarily designed to ensure that datasets are used for research/educational purposes, to prevent misuse, and to satisfy funding requirements. For datasets restricted by MTAs, when possible, a subset of example data (e.g., data from a single subject) is available to download directly on the website. Upon signing the MTA, users will be granted access to the full dataset. MTAs are currently approved by the University of Oxford. This is currently achieved via the email address provided with each dataset on the Digital Brain Bank website. We are actively exploring alternatives to streamline this process. Upon downloading Digital Brain Bank datasets, users agree to acknowledge the source of the data in any outputs. Users are asked to cite the original study for any given dataset and the Digital Brain Bank. Details of associated publications and citation instructions are available on an information page associated with each dataset on the Digital Brain Bank website. Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 10 of 35 Tools and resources Tools and resources Neuroscience Discussion For the Digital Pathologist, we anticipate the strongest integration of our datasets with existing in vivo cohort studies in human or animal models of neuropathology. For the Human ALS MRI-Histology dataset, this includes multimodal MRI and biofluid biomarker sampling platforms such as the Oxford Study for Biomarkers in Motor Neurone Disease (Menke et al., 2014; Menke et al., 2015; Menke et al., 2016; Menke et al., 2018) and the Canadian ALS Neuroimaging Consortium (Kalra et al., 2020). All of these comparisons are supported further by the microscopy data available in select Digital Brain Bank datasets, providing the opportunity to link MRI contrast to microscopy-derived features across multiple domains. Tools and resources Post-mortem MRI Post-mortem MRI facilitates the noninvasive investigation of brain anatomy, tissue composition, and structural connectivity through the acquisition of high-resolution datasets and subsequent microscopy comparisons. Despite this potential, post-mortem MRI remains a relatively niche approach, in part due to technical challenges and the need for multidisciplinary expertise. In order to provide post-mortem MRI as an experimental technique to neuroscientists in Oxford, we have had to develop a broad range of underpinning technologies, including: (i) pulse sequences that provide high-quality data under the harsh imaging conditions of post-mortem tissue (McNab et al., 2009; Miller et al., 2011); (ii) analyses that account for the signal formation mechanisms of these sequences (Tendler et al., 2020a; Tendler et al., 2020b) or properties unique to post-mortem tissue (Tendler et al., 2021); (iii) experimental approaches that enable the use of ultra-high field MRI to increase SNR for high-resolution imaging (Foxley et al., 2014; Tendler et al., 2020b); (iv) development of custom sample holders to maximize SNR and minimize imaging artifacts (Appendix 3—figure 1 and Appendix 3—figure 2); (v) tools for aligning small 2D microscopy images into 3D whole-brain MRI (Huszar et al., 2019); (vi) strategies for co-analyzing MRI and microscopy data (Howard et al., 2019b; Mollink et al., 2017); and (vii) tech- niques for between-species comparisons (Eichert et al., 2020; Mars et al., 2018). The investment of multidisciplinary expertise and effort required to create these datasets will inevi- tably be a barrier to similar studies elsewhere. The Digital Brain Bank makes our data openly available to researchers worldwide to enable a much broader range of investigations. Considerable work has been performed to process images in a manner that users can immediately incorporate into their own Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 11 of 35 Neuroscience Table 2. Acquisition site and MRI scanner associated with all projects in the first release to the Digital Brain Bank. Post-mortem MRI Category Dataset(s) Acquisition location MRI scanner Digital Anatomist Human High-Resolution Diffusion MRI-PLI University of Oxford Siemens 7T Magnetom 32-channel receive/1-channel transmit head coil (Nova Medical) Digital Anatomist Human Callosum MRI-PLI- Histology University of Oxford 9.4T 160 mm horizontal bore VNMRS preclinical MRI system 100 mm bore gradient insert (Varian Inc) 26 mm ID quadrature birdcage coil (Rapid Biomedical GmbH) Digital Brain Zoo NonHuman Primates University of Oxford Baboon, Chimpanzee, Gorilla Siemens 7T Magnetom 28-channel receive/1 channel transmit knee coil (QED) All other brains 7T magnet with Agilent Direct-Drive console 72 mm ID quadrature birdcage RF coil (Rapid Biomedical GmbH) Digital Brain Zoo Marsupials Emory University Siemens 3T Trio 32-channel receive/1-channel transmit head coil Digital Brain Zoo Cetaceans Emory University 2× Tasmanian devil and 1× Thylacine Siemens 3T Trio 32-channel receive/1-channel transmit head coil 1× Thylacine Bruker 9.4T BioSpec preclinical MR system Digital Brain Zoo Carnivora University of Oxford Siemens 7T Magnetom 28-channel receive/1 channel transmit knee coil (QED) Digital Pathologist Human ALS MRI-Histology University of Oxford Siemens 7T Magnetom 32-channel receive/1-channel transmit head coil (Nova Medical) Tools and resources Tools and resources Neuroscience analyses (e.g., diffusion tensor and ball and sticks signal models) reducing user burden to develop their own data processing methods. Further details of these outputs and the types of data available are provided in the Materials and methods. Here, we provide an overview of how we overcame challenges associated with imaging these samples, notably those associated with imaging large post-mortem brains. Datasets were acquired over many years from multiple imaging sites, resulting in evolving experimental setup, acquisition, and processing methods between datasets. To avoid an exhaustive list of different imaging approaches, below we describe the methodology undertaken for acquisitions performed at the University of Oxford, where the majority of datasets in the first release were acquired. Details of the acquisition location and scanner used for all datasets are provided in Table 2. Sample preparation All brains and tissue samples in the first release of the Digital Brain Bank were chemically fixed using aldehyde solutions (e.g., formalin) to prevent decomposition (D’Arceuil and de Crespigny, 2007) and minimize deformation during the course of scanning. All fixed nonhuman brains and excised tissue blocks scanned at the University of Oxford were prepared by soaking the samples in phosphate- buffered saline (PBS) prior to scanning, which increases image SNR by raising T2-values closer to those found in vivo (Shepherd et al., 2009). This was not performed in whole human brains, as brain size necessitates a soaking time of multiple weeks for the buffer fluid to penetrate throughout tissue (Dawe et al., 2009; Tendler et al., 2021; Yong-Hing et al., 2005) which was incompatible with our experimental design. We note that soaking tissue for an insufficient time can lead to artificial ‘bound- aries’ in resulting images, where PBS has not penetrated into deep tissue (Miller et al., 2011). Scanning medium and sample holder Susceptibility artifacts (arising due to air-tissue or air-medium interfaces) can be exacerbated in post- mortem imaging without an appropriate scanning medium. All samples scanned at the University of Oxford were imaged in a proton-free susceptibility-matched fluid (Fomblin LC08, Solvay Solexis; or Fluorinert FC-3283, 3M). The choice of a proton-free fluid means that there is no signal outside of the brain, bringing the additional advantage of minimizing the required field-of-view for any acquisitions, and addressing scaling issues arising from a bright background signal. For whole human brain imaging, we built a two-stage custom holder (Appendix 3—figure 1), which has become the standard for all of our whole-brain human imaging experiments. The holder was designed to fit into a 32-channel receive/1-channel transmit head coil (Nova Medical), securing brains throughout the acqui- sition, and contains a spherical cavity to minimize field-inhomogeneities across the brain. The holder enables brains to be placed in a consistent position (equivalent to an in vivo supine scan), minimizing variability of B0-orientation dependent effects (e.g., susceptibility anisotropy; Liu, 2010), as well as avoiding any potential motion. While motion is clearly less problematic than in vivo, samples must be well-secured, as even small motions can give rise to coregistration challenges and artifacts across the acquisition period (often >24 hr). All human brains were scanned in this holder, with the exception of the Human High-Resolution Diffusion MRI-PLI dataset (data acquired prior to holder construction). Full information on this experimental setup is provided in Wang et al., 2020. Large nonhuman brains scanned at the University of Oxford (Gorilla, Chimpanzee, Wolf, and Baboon) were placed inside a 28-channel receive/1-channel transmit knee coil (QED) to boost SNR (smaller distance between sample and the imaging coil). These brains were placed inside a cylindrical brain holder (Appendix 3—figure 2), with a cylindrical cavity that is compatible with the shape of the knee coil, and is a shape that minimizes B0 field inhomogeneities. Small nonhuman brains/excised tissue blocks scanned on preclinical systems were placed in simpler containers, for example, syringes filled with fluorinert. Choice of MRI scanner Specialized RF coils and imaging gradients facilitate the acquisition of high-resolution, high-SNR post- mortem MRI datasets. Preclinical systems often deliver in this space, notably with powerful gradient sets, and where possible should be adopted for post-mortem imaging. Specifically, post-mortem tissue that has undergone chemical preservation with aldehyde solutions (e.g., formalin) is character- ized by short relaxation time constants (T1, T2, and T2) (Birkl et al., 2016; Birkl et al., 2018; Dawe et al., 2009; Kamman et al., 1985; Nagara et al., 1987; Pfefferbaum et al., 2004; Shepherd et al., 2009; Thelwall et al., 2006), and low diffusivity (D’Arceuil et al., 2007; Shepherd et al., 2009; Sun et al., 2003; Sun et al., 2005; Thelwall et al., 2006) when compared to in vivo tissue. Powerful Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 12 of 35 Tools and resources Tools and resources Neuroscience gradient sets provide rapid signal sampling and strong diffusion weighing, which boosts SNR versus conventional gradients (Dyrby et al., 2011; Roebroeck et al., 2019) in this environment. Broadly speaking, post-mortem MRI data provided in the first release to the Digital Brain Bank can be categorised into two experimental designs. Small post-mortem samples (e.g., small NHP brains and excised tissue blocks) were scanned using specialized preclinical systems with powerful gradient sets. At the University of Oxford, these scans were performed with either a 7T preclinical system with Agilent DirectDrive console (Agilent Technologies, CA) or a 9.4T 160 mm horizontal bore VNMRS preclinical MRI system equipped with a 100-mm bore gradient insert (Varian Inc, CA). Whole brains of larger species do not physically fit into these preclinical systems (maximum sample diameter 7–8 cm) and can only be accommodated in human scanners. These systems often have comparatively low gradient strengths, reducing the available SNR. At the University of Oxford, these brains were scanned on a Siemens 7T Magnetom human scanner. Here, we addressed the imaging environment of fixed, post-mortem tissue and comparatively low gradient-strengths by investing in alternative MR sequences to increase SNR. Further details of this are provided below. Diffusion MRI Post-mortem diffusion MRI is particularly challenging due to the MR-relevant properties of fixed tissue, with reductions in measured relaxation time constants T1, T2, and T2* (Birkl et al., 2016; Birkl et al., 2018; Dawe et al., 2009; Kamman et al., 1985; Nagara et al., 1987; Pfefferbaum et al., 2004; Shepherd et al., 2009; Thelwall et al., 2006), and diffusivity (D’Arceuil et al., 2007; Shepherd et al., 2009; Sun et al., 2003; Sun et al., 2005; Thelwall et al., 2006) routinely reported in literature. To achieve high SNR in these conditions, specialized preclinical systems and tissue preparation methods are often required (Roebroeck et al., 2019), with many groups focusing on tissue sections that can be scanned on rodent scanners with specialized hardware (Beaujoin et al., 2018; Calabrese et al., 2015). Unfortunately, size constraints restrict the use of preclinical systems to small post-mortem tissue samples (e.g., small NHP brains or excised tissue blocks). As described above, large whole brains do not physically fit into preclinical systems and can only be accommodated in human scanners. These systems often have comparatively low gradient strengths; combined with conventional methods (e.g., diffusion-weighted spin-echo, DW-SE), this can lead to low-SNR diffusion imaging volumes. Post-mortem diffusion MRI is particularly challenging due to the MR-relevant properties of fixed tissue, with reductions in measured relaxation time constants T1, T2, and T2* (Birkl et al., 2016; Birkl et al., 2018; Dawe et al., 2009; Kamman et al., 1985; Nagara et al., 1987; Pfefferbaum et al., 2004; Shepherd et al., 2009; Thelwall et al., 2006), and diffusivity (D’Arceuil et al., 2007; Shepherd et al., 2009; Sun et al., 2003; Sun et al., 2005; Thelwall et al., 2006) routinely reported in literature. To achieve high SNR in these conditions, specialized preclinical systems and tissue preparation methods are often required (Roebroeck et al., 2019), with many groups focusing on tissue sections that can be scanned on rodent scanners with specialized hardware (Beaujoin et al., 2018; Calabrese et al., 2015). Unfortunately, size constraints restrict the use of preclinical systems to small post-mortem tissue samples (e.g., small NHP brains or excised tissue blocks). As described above, large whole brains do not physically fit into preclinical systems and can only be accommodated in human scanners. These systems often have comparatively low gradient strengths; combined with conventional methods (e.g., diffusion-weighted spin-echo, DW-SE), this can lead to low-SNR diffusion imaging volumes. Structural MRI Structural MRI enables the delineation of fine tissue structures and cortical surface reconstruction through high contrast, high-resolution imaging datasets. However, the convergence of T1 relaxation Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 13 of 35 13 of 35 Tools and resources Tools and resources Neuroscience Neuroscience times for gray and white matter in formalin-fixed post-mortem tissue leads poor contrast with conven- tional T1-weighted structural protocols (Miller et al., 2011). All structural MRIs available in the first data release were acquired using either a balanced SSFP (bSSFP) or T2-weighted sequence, which demonstrate excellent gray/white matter contrast in fixed post-mortem tissue. Notably, bSSFP signal forming mechanisms lead to an extremely high SNR-efficiency (even when considering the reduced T1 and T2 of post-mortem tissue), affording the acquisition of ultra-high resolution (<500 μm) imaging volumes to delineate fine tissue structures in large post-mortem samples. Contrast in bSSFP and T2-weighted structural MRI datasets is reversed in comparison to conven- tional in vivo T1-weighted acquisitions (i.e., gray matter appears bright, and white matter appears dark). For these datasets, image contrast is predominantly driven by gray and white matter, facilitating the delineation of fine tissue structures and surface reconstructions (Roumazeilles et al., 2020). An example bSSFP dataset is displayed in Appendix 4—figure 1. Integration with conventional structural MRI processing pipelines often needs to account for the reversal of image contrast. Diffusion MRI Over the past decade, our lab has invested considerably into the use of an alternative diffusion imaging technique, diffusion-weighted steady-state free precession (DW-SSFP) (Kaiser et al., 1974; Le Bihan, 1988; Merboldt et al., 1989a; Merboldt et al., 1989b), to achieve high-SNR datasets in large post-mortem samples. DW-SSFP is well suited to the environment of fixed post-mortem tissue, achieving strong diffu- sion weighting and rapid signal sampling, even when hardware achieves limited gradient amplitudes and when T2 values are low (McNab et al., 2009; Vasung et al., 2019; Wilkinson et al., 2016). The DW-SSFP sequence has demonstrated improved SNR-efficiency compared to conventional DW-SE when imaging post-mortem tissue (Miller et al., 2012), further enhanced at ultra-high field (7T) (Foxley et al., 2014). For more details regarding DW-SSFP, please see McNab and Miller, 2010. Broadly, two separate diffusion imaging approaches were used for the first release of data to the Digital Brain Bank. Small brains and excised tissue blocks imaged on preclinical systems were scanned using conventional DW-SE sequences, where tissue preparation and powerful diffusion gradients provide imaging volumes with high SNR. Diffusion imaging for larger post-mortem samples scanned on a human scanner (Siemens 7T Magnetom) was performed using DW-SSFP. To facilitate cross-dataset comparisons, the majority of diffusion datasets from the Digital Brain Bank provide derived diffusivity estimates in the form of diffusion tensor and/or ball and sticks model parameters (Behrens et al., 2007). Whilst there are a number of standard software packages available for DW-SE data, this was achieved for DW-SSFP datasets using a custom imaging pipelines incorpo- rating the full DW-SSFP model (including T1, T2, and B1 dependencies) (Tendler et al., 2020b). There are some differences between derived diffusivity estimates from DW-SSFP and DW-SE data. Importantly, the DW-SSFP signal does not have a well-defined b-value (McNab and Miller, 2010; Tendler et al., 2020a). For all DW-SSFP datasets acquired in whole human brains (e.g., the Human ALS MRI-Histology dataset and the Human High-Resolution Diffusion MRI-PLI dataset), we utilized a recently proposed approach to transform DW-SSFP datasets acquired at two flip angles into equiv- alent measurements at a single, well-defined b-value (Tendler et al., 2020a; Tendler et al., 2020b). This facilitates within dataset comparisons, alongside comparisons with datasets acquired with the DW-SE sequence. However, this approach was not possible for the nonhuman DW-SSFP datasets Tendler et al. eLife 2022;11:e73153. Cross-scale comparisons Post-mortem imaging experiments combining MRI and microscopy are routinely used to validate the origins of image contrast. However, these comparisons are often restricted to simple summary statistics (e.g., ROI averages), rather than utilizing all the available data through pixel-wise comparisons and structural analyses (Mollink et al., 2017). These more detailed approaches are facilitated by accurate cross-modality coregis- trations, a considerable challenge given differences in image contrast and tissue deformations arising from microscopy processing (Huszar et al., 2019; Iglesias et al., 2018; Ohnishi et al., 2016). These challenges are further exacerbated when considering small tissue sections excised from large post-mortem samples, where the corresponding microscopy sampling region must be identified in a 3D imaging volume. To address this, our group has developed TIRL, a novel MR-microscopy coregistration toolbox (Huszar et al., 2019). Further details are provided in the Materials and methods. Future directions: dataset visualization To improve visualization of MRI-only datasets on the Digital Brain Bank website, we are currently inte- grating NiiVue (Rorden et al., 2021), a web-based 3D volumetric viewer for navigating MRI datasets. NiiVue additionally supports binary overlays, which will be used to visualize the location of tissue sampling in the brain. Further details are available at https://github.com/niivue/niivue, (copy archived at swh:1:rev:e67273337430a378a41d6753d91364e9e89b4d33, Hanayik, 2022). Other sequences Quantitative T1 and T2 maps are provided with the post-mortem whole-brain human datasets, acquired using conventional turbo inversion-recovery (TIR) and turbo spin-echo (TSE) sequences. Notably, T1-convergence of gray and white matter in fixed post-mortem tissue leads to low contrast on T1 maps, as described in the Structural MRI section above. T1 maps were fitted assuming mono- exponential signal recovery. T2 maps were processed using an extended phase graph (EPG) fitting scheme (Weigel, 2015), which accounts for B1-inhomogeneity at 7T (details of acquisition and processing are described in Tendler et al., 2021). Whole human brain quantitative T2* and quantitative susceptibility maps are available in a subset of brains provided with the Human ALS MRI-Histology dataset. These data were acquired using a multiecho gradient-echo sequence and processed following the procedure in Wang et al., 2020. Diffusion MRI DOI: https://doi.org/10.7554/eLife.73153 14 of 35 Tools and resources Tools and resources Neuroscience Neuroscience due to differences in the acquisition protocol. Although the diffusivity estimates for the nonhuman DW-SSFP datasets directly relate to the underlying diffusivity of tissue, the DW-SSFP signal forming mechanisms lead to varying effective b-values within and between these datasets (Tendler et al., 2020a). Conservatively, we recommend the nonhuman DW-SSFP datasets to be primarily used for structural connectivity (e.g., tractography) investigations. More generally, differences in the number of diffusion directions, choice of b-value (for DW-SE and DW-SSFP transformed datasets), imaging resolution, and SNR exist across datasets, a result of available scanning hardware, scanning time, experimental design, and sample properties (e.g., type of fixative used and size of the brain). These limitations can lead to differences in resulting diffusivity estimates and should be considered when performing comparisons across different datasets in the Digital Brain Bank. Full details of the acquisitions are provided in the original publications, alongside information on the Digital Brain Bank website and dataset files. Future directions: available datasets Several datasets are under active preparation for future release to the Digital Brain Bank, notably extending the Digital Anatomist and Digital Pathologist categories beyond human tissue for neuroanatomical and neuropathological investigations. These datasets include — Digital Anatomist: (1) Forget-Me-Not devel- oping Human Connectome Project (dHCP) study (Wu et al., 2021), providing diffusion MRI datasets acquired in unfixed, post-mortem neonatal brains; (2) BigMac dataset (Howard et al., 2019a) providing in vivo MRI, post-mortem MRI, PLI, and immunohistochemistry in a single, whole macaque brain. Digital Brain Zoo: further primate species are currently in preparation, as are extensions into orders Carnivora and Rodentia. Digital Pathologist: a cohort study combining multimodal MRI and histology to investigate mouse models of ALS (Martins-Bach et al., 2020; Martins-Bach et al., 2021). Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 15 of 35 Tools and resources Tools and resources Neuroscience Datasets The Digital Brain Bank is not designed as a stand-alone resource—when possible, datasets are asso- ciated with available publications which extensively describe the methodology used. This approach facilitates the referencing of available datasets, and similarly ensures that sufficient detail is provided on how data were acquired and processed. A list of the associated publications with the first release of datasets is provided in Table 1. However, as part of the first release of the Digital Brain Bank, we provide a human dataset that has not yet been described in literature, the Human High-Resolution Diffusion MRI-PLI dataset. We additionally provide four new species datasets for the Digital Brain Zoo, the Hamadryas baboon, Golden Lion Tamarin, Cotton-Top tamarin, and European wolf. A full descrip- tion of the acquisition and data processing for these data are provided in Appendix 1. Tensor Image Registration Library The Digital Brain Bank makes use of the TIRL to perform cross-modality MRI-microscopy coregistra- tions (Huszar et al., 2019). TIRL can be automated for coregistering 3D MR volumes to 2D microscopy images, typically given a set of sequential block-face photographs taken during the tissue dissection process. These coregistrations are available for all released histology in the Human ALS MRI-Histology (Digital Pathologist) dataset (Figure 3d). Remaining coregistrations are being actively curated and will be provided in a future release to the Digital Brain Bank. The decision to present MRI-microscopy coregistrations in the 2D histology space (Figure 3d) was chosen to facilitate visualization. During manual histology sampling, the cutting process introduces non-linear deformities. Furthermore, the cutting angle is not constrained to be parallel to the MRI voxel plane. In general, the excised tissue used for histology will pass through multiple MRI voxel planes nonlinearly, limiting visualization in the 3D MRI space. Importantly, TIRL has been specifically designed to deal with the deformations induced during cutting. Our TIRL pipeline has a specific stage for estimating these deformations from photographs and MRI, and a later stage that refines those deformations for the specific histology slice. Further information is provided in Huszar et al., 2019. Materials and methods Web development and Tview The Digital Brain Bank is a web application made up of individual service components, created using a combination of open-source software. Services include the dataset downloader, the website, and Tview. The web application is hosted on our own server hardware, and the various services of the application are orchestrated using container management system Docker (docker.com). The core website and user interface were created using Vue (vuejs.org), a front-end web frame- work for composing reusable application components. The Vue website communicates with multiple back-end services (via HTTP requests) to retrieve information (e.g., which Tview tiles to display, which datasets are available for download). The reactive web application ensures that changes to website are seen in real time, or on the next possible page reload.l Tview provides real-time zooming/panning of high-resolution microscopy and MRI overlays using leafletjs, a software library originally used to display satellite imagery at multiple elevations. We have adapted our microscopy datasets to be compatible with this software library in order to take advantage of the features it offers. This is achieved by converting images into tiles at multiple zoom levels (via libvips), and uploading these tiles to the local web server. Only the relevant subset of tiles is downloaded when users interact with images using Tview. Individual tiles comprise a small file size, facilitating real-time interaction. Continued use of the Digital Brain Bank requires a simple process for collaborators to upload post-mortem datasets, and an application administration layer is under active development to facilitate this process. Code for the Digital Brain Bank web application is available at https://git.fmrib.ox.ac.uk/thanayik/dbb. Types of data provided The Digital Brain Bank uses the following definitions: • Raw Data: MRI and microscopy images obtained directly from the MRI or slide scanner. These data require further processing be useful (i.e., the value in individual pixels is not directly inform- ative about biological processes or properties).i • Raw Data: MRI and microscopy images obtained directly from the MRI or slide scanner. These data require further processing be useful (i.e., the value in individual pixels is not directly inform- ative about biological processes or properties).i • Pre-Processed Outputs: imaging outputs from individual brains produced by a first stage of processing suitable for a broad range of subsequent analysis. For example, diffusion parameter estimates (e.g., diffusion tensor outputs), quantitative relaxometry maps, or PLI fiber orientation maps. These outputs can be immediately fed into an analysis to answer a hypothesis- or data- driven neuroscientific question. i • Derived Outputs: results produced from subsequent analysis that use Pre-Processed Outputs as an input specific to a domain of neuroscientific investigation. For example, tractography- derived pathway segmentations or group-averaged atlases. • Derived Outputs: results produced from subsequent analysis that use Pre-Processed Outputs as an input specific to a domain of neuroscientific investigation. For example, tractography- derived pathway segmentations or group-averaged atlases. The Digital Brain Bank aims to capture a broad range of datasets under the umbrella of post- mortem neuroimaging, and as such we have aimed to keep the resource flexible for uploaded data- sets. There are no strict criteria regarding the types and structure of post-mortem data released to the Digital Brain Bank. In addition, the Digital Brain Bank aims to facilitate the investigation of research hypotheses designed by the user. To reflect this, we primarily provide Pre-Processed Outputs. Limited Derived Outputs associated with specific projects will be also made available at the discretion of the Digital Brain Bank. Raw data for a given dataset is available on request. These pre-processed outputs reduce the burden on the user to develop their own processing pipelines, of particular importance when considering datasets acquired with alternative sequences not addressed with commonly used imaging software (e.g., DW-SSFP) (Tendler et al., 2020b), or datasets which required specialized fitting approaches (e.g., EPG fitting required for T2 mapping at 7T, used in the Human ALS MRI-Histology and Human High-Resolution Diffusion MRI-PLI datasets) (Tendler et al., 2021; Weigel, 2015). Conditions for data uploading All datasets hosted on the Digital Brain Bank are associated with projects performed at the University of Oxford, in collaboration with members of the University of Oxford, or from close collaborators. In addition, limited Derived Outputs from users of Digital Brain Bank datasets will also be considered for data upload, subject to quality control on an individual basis. Information regarding the primary contributors to the dataset are explicitly stated on the Digital Brain Bank website. All projects must have been granted ethical approval from the relevant brain banks and departmental ethics boards. Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 16 of 35 Tools and resources Tools and resources Neuroscience Datasets will be shared on the Digital Brain Bank website on the condition that data providers do not require co-authorship for any subsequent outputs based on the use of the datasets alone. Types of data provided For diffusion MRI datasets acquired with DW-SE, we also provide the Pre-Processed Outputs of individual diffusion volumes, which users can feed into a broad range of available software. For DW-SSFP, we do not by default provide the individual diffusion volumes, as no widely available diffusion MRI software packages incorporate the DW-SSFP signal model. We share our custom software for analyzing these data through the Digital Brain Bank. DW-SSFP diffusion volumes are available upon request, with the caveat that care needs to be taken in analyzing these data in light of the unusual dependences (e.g., T1 and T2) and signal model of DW-SSFP (Buxton, 1993; Tendler et al., 2020a; Tendler et al., 2020b). First and foremost, the Digital Brain Bank is a data sharing resource. Details of the acquisition and processing methodology associated with each dataset is provided with the accompanying manu- scripts, on the Digital Brain Bank website (Information page associated with each project), and with the downloaded dataset. However, data hosts are encouraged to provide pre-processing code when available. This code will be linked with each dataset on the Digital Brain Bank website (on the Infor- mation page), or packaged with the dataset download. For multimodal (MRI and microscopy) datasets in the first release (Human High-Resolution Diffusion MRI-PLI, Human Callosum MRI-PLI-Histology, and Human ALS MRI-Histology dataset), raw high-resolution microscopy images are provided. A full set of coregistered data to enable MRI-microscopy voxelwise comparisons via TIRL (Huszar et al., 2019) are being actively curated for future release (Figure 3d). These are currently available for the histology released with the Human ALS MRI-Histology dataset. Metadata specific to the analysis of post-mortem tissue (e.g., fixative type, post-mortem interval, etc.) or relevant to distinguishing individual datasets in a cohort study (e.g., control brain or brain with a neurological disease) is provided when available. Acknowledgements The Digital Brain Bank is supported by the Wellcome Trust (202788/Z/16/Z) and Medical Research Council (MRC, MR/K02213X/1). The Wellcome Centre for Integrative Neuroimaging is supported by core funding from the Wellcome Trust (203139/Z/16/Z). BCT is supported by funding from the Well- come Trust (202788/Z/16/Z and 222829/Z/21/Z). KLM, AS, and JM are supported by funding from the Wellcome Trust (202788/Z/16/Z), RBM is supported by funding from the Biotechnology and Biological Sciences Research Council (BBSRC) UK (BB/N019814/1) and the Netherlands Organiza- tion for Scientific Research NWO (452-13-015), SJ is supported by funding from the Wellcome Trust (221933/Z/20/Z and 215573/Z/19/Z) and the MRC (MR/L009013/1), TH and DM are supported by funding from the Wellcome Centre for Integrative Neuroimaging, OA is supported by funding from the Medical Research Council, Alzheimer’s UK, and NIHR Oxford Biomedical Research Centre, MFB is supported by funding from the Alfred Benzon’s Foundation, KLB is supported by funding from the Biotechnology and Biological Sciences Research Council (BBSRC) UK (BB/N019814/1), SF and MPG are supported by funding from the MRC (MR/K02213X/1), MPvdH is supported by the Neth- erlands Organization for Scientific Research NWO (VIDI-452-16-015 and ALW-179) and the Euro- pean Research Council (ERC-COG 101001062) AFDH and IH are supported by funding from the Engineering and Physical Sciences Research Council (EP/L016052/1) and Medical Research Council (MR/L009013/1), AAK was funded by Cancer Research UK (C5255/A15935), PRM is supported by funding from the National Research Foundation of South Africa, RALM is supported by funding from the Medical Research Council (MR/K01014X/1) and the Wellcome Trust (202788/Z/16/Z), LR is supported by funding from the Biotechnology and Biological Sciences Research Council (BBSRC) UK (BB/M011224/1), JS is supported by funding from the IDEXLYON IMPULSION 2020 (IDEX/ IMP/2020/14) and Labex CORTEX (ANR-11-LABX-0042) grant (Université de Lyon), CS is supported by funding from the NIHR Oxford Biomedical Research Centre (BRC), MRT is supported by funding from the Motor Neurone Disease Association, CW is supported by funding from the China Schol- arship Council (CSC). Human post-mortem brain datasets for the Digital Anatomist and Digital Pathologist used tissue provided by the Oxford Brain Bank, a research ethics committee (REC) approved, HTA regulated research tissue bank (REC reference 15/SC/0639). The Oxford Brain Bank is supported by the MRC, Brains for Dementia Research (BDR) (Alzheimer Society and Alzheimer Research UK), and the NIHR Oxford Biomedical Research Centre. Acknowledgements The views expressed are those of the authors and not necessarily those of the NHS, the NIHR, or the Department of Health. Datasets for the Digital Brain Zoo used tissue provided by the Australian Museum, Copenhagen Zoo, Primate Brain Bank, Save the Tasmanian Devil, Smithsonian, University of Oxford, and Zoological Society of London Data storage database As is the nature of a data resource associated with both completed and ongoing projects, some datasets will be updated over time. These future releases will typically be associated with new images being made available or improvements to existing processing pipelines. Until more 17 of 35 Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 Tools and resources Tools and resources Neuroscience streamlined data access methods are in place, we will contact users directly to inform them of any updates made to a given dataset. This approach is aligned with the current framework for data access, with users required to contact the Request Data Contact on the Digital Brain Bank website to request access. A future ambition of the Digital Brain Bank is to streamline data access procedures by integrating user sign-up, authentication, and approval combined with access to specific dataset versions within our database over time. To achieve this, we are continuously developing the platform and incorpo- rating feedback and feature requests including enabling a programmatic interface to datasets for approved users, and detailed dataset versioning. We will investigate associating dataset versions with DOIs directly on the Digital Brain Bank website (or through known providers), to facilitate the tracking and reproducibility of individual datasets and analysis pipelines. Author contributions Benjamin C Tendler, Led the Digital Brain Bank project, developed processing pipelines for Digital Anatomist (Human High-Resolution Diffusion MRI-PLI), Digital Brain Zoo (Hamadryas baboon, Gorilla, Chimpanzee and European Wolf), and Digital Pathologist (Human ALS MRI-Histology) MRI datasets, contributed to the establishment of MRI acquisition protocols for Digital Brain Zoo (Baboon and Euro- pean Wolf) datasets, established the Digital Brain Bank website text and figures, contributed to Digital Brain Bank website design, and wrote the manuscript; Taylor Hanayik, Implemented the Digital Brain Bank website, designed and implemented Tview, designed and implemented the website download/ upload framework, contributed to the Digital Brain Bank website design and drafted sections of the manuscript; Olaf Ansorge, Contributed to the development the of tissue sampling strategy, neuro- pathological analysis and recruitment of post-mortem brains tissue for Digital Anatomist (Human High- Resolution Diffusion MRI-PLI, Human Callosum MRI-PLI-Histology) and Digital Pathologist (Human ALS MRI-Histology) datasets; Sarah Bangerter-Christensen, Carried out immunohistochemistry and histology image acquisition that generated the Human ALS MRI-Histology (Digital Pathologist) hippo- campus histology dataset; Gregory S Berns, Led the projects which acquired the Marsupial (Tasma- nian Devil and Thylacine) and Cetacean (Common and Pantropical Dolphin) datasets (Digital Brain Zoo); Mads F Bertelsen, Acquired Carnivora (European Wolf) and Non-Human Primate (Hamadryas baboon, Cotton-topped tamarin, Golden lion tamarin and Ring-tailed lemur) samples (Digital Brain Zoo); Katherine L Bryant, Led the project which acquired Primate Brain Bank (Bushbaby, Capuchin monkey, Chimpanzee, Colobus monkey, Mangabey, Night monkey, Saki monkey, Woolly monkey) datasets; Sean Foxley, Conceived study design of the Human High-Resolution Diffusion MRI-PLI (Digital Anatomist) dataset, contributed to the establishment of MRI protocols and carried out MRI acquisition for the Human High-Resolution Diffusion MRI-PLI (Digital Anatomist), Non-Human Primate (Gorilla and Chimpanzee) (Digital Brain Zoo), and Human ALS MRI-Histology (Digital Pathologist) datasets, performed the tractography-PLI analysis for the Human High-Resolution Diffusion MRI-PLI (Digital Anatomist) dataset (Figs. Funder Rogier B Mars Netherlands Organization for Scientific Research NWO 452-13-015 The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication. Additional information Competing interests Saad Jbabdi, Karla L Miller: Reviewing editor, eLife. The other authors declare that no competing interests exist. 18 of 35 18 of 35 Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 Tools and resources Tools and resources Neuroscience Funding Funder Grant reference number Author Wellcome Trust 202788/Z/16/Z Benjamin C Tendler Ricarda AL Menke Jeroen Mollink Adele Smart Karla L Miller Wellcome Trust 222829/Z/21/Z Benjamin C Tendler Wellcome Trust Taylor Hanayik Duncan Mortimer Medical Research Council, Alzheimer's UK and NIHR Oxford Biomedical Research Centre Olaf Ansorge Alfred Benzon's Foundation Mads F Bertelsen Biotechnology and Biological Sciences Research Council BB/N019814/1 Katherine L Bryant Rogier B Mars Medical Research Council MR/K02213X/1 Sean Foxley Menuka Pallebage- Gamarallage Netherlands Organization for Scientific Research NWO VIDI-452-16-015 Martijn P van den Heuvel Netherlands Organization for Scientific Research NWO ALW-179 Martijn P van den Heuvel European Research Council ERC-COG 101001062 Martijn P van den Heuvel Engineering and Physical Sciences Research Council EP/L016052/1 Amy FD Howard Istvan N Huszar Medical Research Council MR/L009013/1 Istvan N Huszar Amy FD Howard Saad Jbabdi Cancer Research UK C5255/A15935 Alexandre A Khrapitchev National Research Foundation of South Africa Paul R Manger Medical Research Council MR/K01014X/1 Ricarda AL Menke Biotechnology and Biological Sciences Research Council BB/M011224/1 Lea Roumazeilles IDEXLYON IMPULSION 2020 IDEX/IMP/2020/14 Jerome Sallet Labex CORTEX (Université de Lyon) ANR-11-LABX-0042 Jerome Sallet NIHR Oxford Biomedical Research Centre Connor Scott Motor Neurone Disease Association Martin R Turner China Scholarship Council Chaoyue Wang Wellcome Trust 221933/Z/20/Z Saad Jbabdi Wellcome Trust 215573/Z/19/Z Saad Jbabdi Grant reference number Author Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 19 of 35 Tools and resources Tools and resources Neuroscience Grant reference number Author Author contributions 1b and c), and drafted sections of the manuscript; Martijn P van den Heuvel, Contributed to acquisition of resources for the Digital Brain Zoo through the Primate Brain Bank (Bushbaby, Capuchin monkey, Chimpanzee, Colobus monkey, Mangabey, Night monkey, Saki monkey, Woolly monkey); Amy FD Howard, Produced Tview images for the Digital Brain Bank website implementation, provided critical assessment of the Digital Brain Bank website, and contributed to Digital Brain Bank website design; Istvan N Huszar, Established the MRI-microscopy coregistration software (TIRL) and performed the MRI-histology coregistrations for the Human ALS MRI-Histology dataset (Digital Pathologist); Alexandre A Khrapitchev, Established the MRI acquisition protocols and carried out MRI acquisition for the small Non-Human Primate (Bushbaby, Capuchin monkey, Colobus monkey, Cotton-topped tamarin, Golden lion tamarin, Macaque monkey, Mangabey, Night monkey, Ring-tailed lemur, Saki monkey, Woolly monkey) datasets (Digital Brain Zoo); Anna Leonte, Carried out immunohistochemistry and image acquisition for the Human ALS MRI-Histology (Digital Pathologist) anterior cingulate histology dataset; Paul R Manger, Acquired Carnivora (European Wolf) and Non- Human Primate (Hamadryas baboon, Cotton-topped tamarin, Golden lion tamarin, Ring-tailed lemur) samples (Digital Brain Zoo); Ricarda AL Menke, Provided critical assessment of the processing pipe- lines for Human ALS MRI-Histology dataset (Digital Pathologist); Jeroen Mollink, Led the project which acquired the Human Callosum MRI-PLI-Histology (Digital Anatomist) dataset, and performed the PLI acquisition and analysis for the Human High-Resolution Diffusion MRI-PLI (Digital Anatomist) dataset; Duncan Mortimer, Contributed to the Digital Brain Bank website design and established the data server implementation; Menuka Pallebage-Gamarallage, Led the project which obtained the Human ALS MRI-Histology (Digital Pathologist) histology dataset, establishing protocols and carrying out systematic human brain sampling, immunohistochemistry, image acquisition and analyses of histology datasets. Supported Digital Anatomist projects (Human High-Resolution Diffusion MRI-PLI, Human Tendler et al. eLife 2022;11:e73153. Author contributions DOI: https://doi.org/10.7554/eLife.73153 20 of 35 Tools and resources Tools and resources Tools and resources Neuroscience Callosum MRI-PLI-Histology) with tissue sampling and histology data generation; Lea Roumazeilles, Led projects which acquired Non-Human Primate (Chimpanzee, Gorilla, Macaque monkey, Ring-tailed lemur) datasets and facilitated the scanning preparation for Digital Brain Zoo samples scanned at the University of Oxford; Jerome Sallet, Facilitated the acquisition and scanning preparation for all Digital Brain Zoo samples scanned at the University of Oxford; Lianne H Scholtens, Selected and prepared samples from the Primate Brain Bank (Bushbaby, Capuchin monkey, Chimpanzee, Colobus monkey, Mangabey, Night monkey, Saki monkey, Woolly monkey) (Digital Brain Zoo); Connor Scott, Supported the recruitment and preparation of human whole brain samples for the Human High-Resolution Diffu- sion MRI-PLI (Digital Anatomist) and Human ALS MRI-Histology (Digital Pathologist) datasets; Adele Smart, Carried out immunohistochemistry and image acquisition for the Human ALS MRI-Histology (Digital Pathologist) histology dataset; Martin R Turner, Supported the characterisation of the ALS cohort in the Human ALS MRI-Histology (Digital Pathologist) dataset; Chaoyue Wang, Established the analysis pipeline and performed data processing of the T2* and magnetic susceptibility maps for the Human ALS MRI-Histology (Digital Pathologist) dataset; Saad Jbabdi, Established analysis tools for processing diffusion MRI data in the Digital Anatomist (Human High-Resolution Diffusion MRI-PLI), Digital Brain Zoo (Baboon, Chimpanzee, Gorilla and European Wolf), and Digital Pathologist (Human ALS MRI-Histology) datasets, contributed to the Digital Brain Bank website design, and critically appraised the manuscript; Rogier B Mars, Leads the Digital Brain Zoo, led projects associated with the first data release in the Digital Brain Zoo performed at the University of Oxford, contributed to the Digital Brain Bank website design, and critically appraised the manuscript; Karla L Miller, Conceived the Digital Brain Bank, led projects associated with the first data release in the Digital Anatomist and Digital Pathologist, contributed to the establishment of MRI acquisition protocols for Digital Anato- mist (Human High-Resolution Diffusion MRI-PLI), Digital Brain Zoo (Hamadryas baboon, Chimpanzee, Gorilla and European Wolf), and Digital Pathologist (Human ALS MRI-Histology) datasets, contributed to the Digital Brain Bank website design, and critically appraised the manuscript Author ORCIDs Author ORCIDs Benjamin C Tendler ‍ ‍ http://orcid.org/0000-0003-2095-8665 Katherine L Bryant ‍ ‍ http://orcid.org/0000-0003-1045-4543 Amy FD Howard ‍ ‍ http://orcid.org/0000-0003-1154-1913 Alexandre A Khrapitchev ‍ ‍ http://orcid.org/0000-0002-7616-6635 Duncan Mortimer ‍ ‍ http://orcid.org/0000-0002-7483-2024 Jerome Sallet ‍ ‍ http://orcid.org/0000-0002-7878-0209 Connor Scott ‍ ‍ http://orcid.org/0000-0003-2316-1707 Adele Smart ‍ ‍ http://orcid.org/0000-0002-4293-5942 Chaoyue Wang ‍ ‍ http://orcid.org/0000-0001-9402-1563 Karla L Miller ‍ ‍ http://orcid.org/0000-0002-2511-3189 Tools and resources References Absinta M, Ha SK, Nair G, Sati P, Luciano NJ, Palisoc M, Louveau A, Zaghloul KA, Pittaluga S, Kipnis J. 2017. Human and nonhuman primate meninges harbor lymphatic vessels that can be visualized noninvasively by MRI. eLife 6:e29738. DOI: https://doi.org/10.7554/eLife.29738 Andersson JLR, Jenkinson M, Smith S. 2007. Non-Linear Registration, Aka Spatial Norm Technical Report TR07JA2. FMRIB Analysis Group of the University of Oxford Press. Assaf Y, Bouznach A, Zomet O, Marom A, Yovel Y. 2020. Conservation of brain connectivity and wiring across the mammalian class. Nature Neuroscience 23:805–808. DOI: https://doi.org/10.1038/s41593-020-0641-7, PMID: 32514137 Avants BB, Tustison NJ, Song G, Cook PA, Klein A, Gee JC. 2011. A reproducible evaluation of ANTs similarity metric performance in brain image registration. NeuroImage 54:2033–2044. DOI: https://doi.org/10.1016/j. neuroimage.2010.09.025, PMID: 20851191 g Axer M, Amunts K, Grässel D, Palm C, Dammers J, Axer H, Pietrzyk U, Zilles K. 2011. A novel approach to the human connectome: Ultra-high resolution mapping of fiber tracts in the brain. NeuroImage 54:1091–1101. DOI: https://doi.org/10.1016/j.neuroimage.2010.08.075, PMID: 20832489 p g j g Bangerter NK, Hargreaves BA, Vasanawala SS, Pauly JM, Gold GE, Nishimura DG. 2004. Analysis of multiple- acquisition SSFP. Magnetic Resonance in Medicine 51:1038–1047. DOI: https://doi.org/10.1002/mrm.20052, PMID: 15122688 Beaujoin J, Palomero-Gallagher N, Boumezbeur F, Axer M, Bernard J, Poupon F, Schmitz D, Mangin JF, Poupon C. 2018. Post-mortem inference of the human hippocampal connectivity and microstructure using ultra-high field diffusion MRI at 11.7 T. Brain Structure & Function 223:2157–2179. DOI: https://doi.org/10. 1007/s00429-018-1617-1, PMID: 29387938 Behrens TEJ, Berg HJ, Jbabdi S, Rushworth MFS, Woolrich MW. 2007. Probabilistic diffusion tractography with multiple fibre orientations: What can we gain? NeuroImage NeuroImage 34:144–155. DOI: https://doi.org/10. 1016/j.neuroimage.2006.09.018, PMID: 17070705 j g Berns GS, Cook PF, Foxley S, Jbabdi S, Miller KL, Marino L. 2015. Diffusion tensor imaging of dolphin brains reveals direct auditory pathway to temporal lobe. Proceedings. Biological Sciences 282:20151203. DOI: https://doi.org/10.1098/rspb.2015.1203, PMID: 26156774 p g p Berns GS, Ashwell KWS. 2017. Reconstruction of the cortical maps of the Tasmanian tiger and comparison to the Tasmanian devil. PLOS ONE 12:e0168993. DOI: https://doi.org/10.1371/journal.pone.0168993, PMID: 28099446 Bhagwandin A, Haagensen M, Manger PR. 2017. The brain of the black (Diceros bicornis) and white (Ceratotherium simum) African rhinoceroses: morphology and volumetrics from magnetic resonance imaging. Frontiers in Neuroanatomy 11:74. DOI: https://doi.org/10.3389/fnana.2017.00074, PMID: 28912691 y p g , Birkl C, Langkammer C, Golob-Schwarzl N, Leoni M, Haybaeck J, Goessler W, Fazekas F, Ropele S. 2016. Ethics Human subjects: All human post-mortem datasets described in the first release to the Digital Brain Bank used tissue provided by the Oxford Brain Bank, a research ethics committee (REC) approved, HTA regulated research tissue bank. The studies were conducted under the Oxford Brain Bank's generic Research Ethics Committee approval (15/SC/0639). There are four new species datasets (Hamadryas baboon, Golden Lion tamarin, Cotton-Top tamarin, and European wolf) provided in the first release to the Digital Brain Bank which have not been previ- ously described in literature. These datasets all used post-mortem tissue from animals which died of causes unrelated to research, and therefore do not require a Home Office license under the Animals (Scientific Procedures) Act 1986. Ethics statements associated with all remaining Digital Brain Bank datasets are described in the original manuscript associated with each dataset, as provided in Table 1. Decision letter and Author response Decision letter https://doi.org/10.7554/eLife.73153.sa1 Author response https://doi.org/10.7554/eLife.73153.sa2 21 of 35 Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 Tools and resources Tools and resources Neuroscience Data availability Data availability The Digital Brain Bank (https://open.win.ox.ac.uk/DigitalBrainBank) is a data release platform providing open access to curated, multimodal post-mortem neuroimaging datasets. All datasets described in this manuscript are available through the Digital Brain Bank, with details of access provided within the manuscript and on the website. Code for the Digital Brain Bank resource is available at https://git. fmrib.ox.ac.uk/thanayik/dbb. When available, details of associated processing code for each dataset is linked to the dataset's Information page on the Digital Brain Bank website. Source data for the corpus callosum analysis in Fig 3c is provided in a Supplementary File. Additional files Supplementary files • Transparent reporting form • Supplementary file 1. Corpus Callosum Analysis. • Supplementary file 1. Corpus Callosum Analysis. References Magnetic Resonance in Medicine 61:810–818. DOI: https://doi.org/10.1002/mrm.21909, PMID: 19189294 p g Denic A, Macura SI, Mishra P, Gamez JD, Rodriguez M, Pirko I. 2011. MRI in rodent models of brain disorders. Neurotherapeutics 8:3–18. DOI: https://doi.org/10.1007/s13311-010-0002-4, PMID: 21274681 Dyrby TB, Baaré WFC, Alexander DC, Jelsing J, Garde E, Søgaard LV. 2011. An ex vivo imaging pipeline for producing high-quality and high-resolution diffusion-weighted imaging datasets. Human Brain Mapping 32:544–563. DOI: https://doi.org/10.1002/hbm.21043, PMID: 20945352 g D’Arceuil H, de Crespigny A. 2007. The effects of brain tissue decomposition on diffusion tensor imaging and tractography. NeuroImage 36:64–68. DOI: https://doi.org/10.1016/j.neuroimage.2007.02.039, PMID: 17433879 D’Arceuil HE, Westmoreland S, de Crespigny AJ. 2007. An approach to high resolution diffusion tensor imaging in fixed primate brain. NeuroImage 35:553–565. DOI: https://doi.org/10.1016/j.neuroimage.2006.12.028, PMID: 17292630 Edlow BL, Mareyam A, Horn A, Polimeni JR, Witzel T, Tisdall MD, Augustinack JC, Stockmann JP, Diamond BR, Stevens A, Tirrell LS, Folkerth RD, Wald LL, Fischl B, van der Kouwe A. 2019. 7 Tesla MRI of the ex vivo human brain at 100 micron resolution. Scientific Data 6:244. DOI: https://doi.org/10.1038/s41597-019-0254-8, PMID: 31666530 Eichert N, Robinson EC, Bryant KL, Jbabdi S, Jenkinson M, Li L, Krug K, Watkins KE, Mars RB. 2020. Cross- species cortical alignment identifies different types of anatomical reorganization in the primate temporal lobe. eLife 9:e53232. DOI: https://doi.org/10.7554/eLife.53232, PMID: 32202497 p g Eickhoff SB, Stephan KE, Mohlberg H, Grefkes C, Fink GR, Amunts K, Zilles K. 2005. A new SPM toolbox for combining probabilistic cytoarchitectonic maps and functional imaging data. NeuroImage 25:1325–1335. DOI: https://doi.org/10.1016/j.neuroimage.2004.12.034, PMID: 15850749 p g j g Foxley S, Jbabdi S, Clare S, Lam W, Ansorge O, Douaud G, Miller K. 2014. Improving diffusion-weighted imaging of post-mortem human brains: SSFP at 7T. NeuroImage 102 Pt 2:579–589. DOI: https://doi.org/10. 1016/j.neuroimage.2014.08.014, PMID: 25128709 j g Foxley S, Mollink J, Jbabdi S, Clare S, Hernandez Fernandez M, Scott C, Ansorge O, Miller KL. 2016. Validating tractography of high resolution post-mortem human brain at 7T with polarized light imaging. ISMRM 24th Annual Meeting Singapore. . Friedrich P, Forkel SJ, Amiez C, Balsters JH, Coulon O, Fan L, Goulas A, Hadj-Bouziane F, Hecht EE, Heuer K, Jiang T, Latzman RD, Liu X, Loh KK, Patil KR, Lopez-Persem A, Procyk E, Sallet J, Toro R, Vickery S, et al. 2021. Imaging evolution of the primate brain: the next frontier. NeuroImage 228:117685. References Effects of formalin fixation and temperature on MR relaxation times in the human brain. NMR in Biomedicine 29:458– 465. DOI: https://doi.org/10.1002/nbm.3477, PMID: 26835664 465. DOI: https://doi.org/10.1002/nbm.3477, PMID: 26835664 p g , Birkl C, Soellradl M, Toeglhofer AM, Krassnig S, Leoni M, Pirpamer L, Vorauer T, Krenn H, Haybaeck J, Fazekas F, Ropele S, Langkammer C. 2018. Effects of concentration and vendor specific composition of formalin on postmortem MRI of the human brain. Magnetic Resonance in Medicine 79:1111–1115. DOI: https://doi.org/10. 1002/mrm.26699, PMID: 28382642 Bryant KL, Ardesch DJ, Roumazeilles L, Scholtens LH, Khrapitchev AA, Tendler BC, Wu W, Miller KL, Sallet J, van den Heuvel MP, Mars RB. 2021. Diffusion MRI data, sulcal anatomy, and tractography for eight species from the Primate Brain Bank. Brain Structure & Function 226:2497–2509. DOI: https://doi.org/10.1007/s00429-021- 02268-x, PMID: 34264391 Buxton RB. 1993. The diffusion sensitivity of fast steady‐state free precession imaging. Magnetic Resonance in Medicine 29:235–243. DOI: https://doi.org/10.1002/mrm.1910290212, PMID: 8429788 Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 22 of 35 Tools and resources Tools and resources Neuroscience Calabrese E, Hickey P, Hulette C, Zhang J, Parente B, Lad SP, Johnson GA. 2015. Postmortem diffusion MRI of the human brainstem and thalamus for deep brain stimulator electrode localization. Human Brain Mapping 36:3167–3178. DOI: https://doi.org/10.1002/hbm.22836, PMID: 26043869 67–3178. DOI: https://doi.org/10.1002/hbm.22836, PM Cardenas AM, Sarlls JE, Kwan JY, Bageac D, Gala ZS, Danielian LE, Ray-Chaudhury A, Wang HW, Miller KL, Foxley S, Jbabdi S, Welsh RC, Floeter MK. 2017. Pathology of callosal damage in ALS: An ex-vivo, 7 T diffusion tensor MRI study. NeuroImage. Clinical 15:200–208. DOI: https://doi.org/10.1016/j.nicl.2017.04.024, PMID: 28529876 Chapman MC, Jelsone-Swain L, Johnson TD, Gruis KL, Welsh RC. 2014. Diffusion tensor MRI of the corpus callosum in amyotrophic lateral sclerosis. Journal of Magnetic Resonance Imaging 39:641–647. DOI: https:// doi.org/10.1002/jmri.24218, PMID: 23843179 g j , Cottaar M, Bastiani M, Boddu N, Glasser MF, Haber S, van Essen DC, Sotiropoulos SN, Jbabdi S. 2021. Modelling white matter in gyral blades as a continuous vector field. NeuroImage 227:117693. DOI: https://doi. org/10.1016/j.neuroimage.2020.117693, PMID: 33385545 g j g Dammers J, Axer M, Grässel D, Palm C, Zilles K, Amunts K, Pietrzyk U. 2010. Signal enhancement in polarized light imaging by means of independent component analysis. NeuroImage 49:1241–1248. DOI: https://doi.org/ 10.1016/j.neuroimage.2009.08.059, PMID: 19733674 10.1016/j.neuroimage.2009.08.059, PMID: 19733674 Dawe RJ, Bennett DA, Schneider JA, Vasireddi SK, Arfanakis K. 2009. Postmortem MRI of human brain hemispheres: T 2 relaxation times during formaldehyde fixation. References DOI: https://doi.org/10. 1016/j.neuroimage.2020.117685, PMID: 33359344 j g Fritz FJ, Sengupta S, Harms RL, Tse DH, Poser BA, Roebroeck A. 2019. Ultra-high resolution and multi-shell diffusion MRI of intact ex vivo human brains using kT-dSTEAM at 9.4T. NeuroImage 202:116087. DOI: https:// doi.org/10.1016/j.neuroimage.2019.116087, PMID: 31408716 Grewal JS, Gloe T, Hegedus J, Bitterman K, Billings BK, Chengetanai S, Bentil S, Wang VX, Ng JC, Tang CY, Geletta S, Wicinski B, Bertelson M, Tendler BC, Mars RB, Aguirre GK, Rusbridge C, Hof PR, Sherwood CC, Manger PR, et al. 2020. Brain gyrification in wild and domestic canids: Has domestication changed the gyrification index in domestic dogs? The Journal of Comparative Neurology 528:3209–3228. DOI: https://doi. org/10.1002/cne.24972, PMID: 32592407 Hanayik T. 2022. NiiVue. swh:1:rev:e67273337430a378a41d6753d91364e9e89b4d33. Software Heritage. https://archive.softwareheritage.org/swh:1:dir:051571e97b09ee5db431df92725dc5514f67b5ef;origin=https:// github.com/niivue/niivue;visit=swh:1:snp:fa2ed9d7abde8e653a834b21b2afd51ee640877b;anchor=swh:1:rev: e67273337430a378a41d6753d91364e9e89b4d33 Heuer K, Gulban OF, Bazin PL, Osoianu A, Valabregue R, Santin M, Herbin M, Toro R. 2019. Evolution of neocortical folding: A phylogenetic comparative analysis of MRI from 34 primate species. Cortex; a Journal Devoted to the Study of the Nervous System and Behavior 118:275–291. DOI: https://doi.org/10.1016/j. cortex.2019.04.011, PMID: 31235272 Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 23 of 35 Tools and resources Tools and resources Neuroscience Hofer S, Frahm J. 2006. Topography of the human corpus callosum revisited-Comprehensive fiber tractography using diffusion tensor magnetic resonance imaging. NeuroImage 32:989–994. DOI: https://doi.org/10.1016/j. neuroimage.2006.05.044, PMID: 16854598 g Howard A, Jbabdi S, Khrapitchev A, Sallet J, Daubney G, Mollink J, Scott C, Sibson N, Miller K. 2019a. The BigMac dataset: ultra-high angular resolution diffusion imaging and multi-contrast microscopy of a whole macaque brain. ISMRM 27th Annual Meeting. . q g Howard AF, Mollink J, Kleinnijenhuis M, Pallebage-Gamarallage M, Bastiani M, Cottaar M, Miller KL, Jbabdi S. 2019b. Joint modelling of diffusion MRI and microscopy. NeuroImage 201:116014. DOI: https://doi.org/10. 1016/j.neuroimage.2019.116014, PMID: 31315062 Huszar IN, Pallebage-Gamarallage M, Foxley S, Tendler BC, Leonte A, Hiemstra M, Mollink J, Smart A, Bangerter-Christensen S, Brooks H, Turner MR, Ansorge O, Miller KL, Jenkinson M. 2019. Tensor Image Registration Library: Automated Non-Linear Registration of Sparsely Sampled Histological Specimens to Post-Mortem MRI of the Whole Human Brain. Neuroscience 1:e849570. DOI: https://doi.org/10.1101/849570 Iglesias JE, Modat M, Peter L, Stevens A, Annunziata R, Vercauteren T, Lein E, Fischl B, Ourselin S, Initiative ADN. 2018. Joint registration and synthesis using a probabilistic model for alignment of MRI and histological sections. Medical Image Analysis 50:127–144. DOI: https://doi.org/10.1016/j.media.2018.09.002, PMID: 30282061 Jenkinson M, Smith S. 2001. References A global optimisation method for robust affine registration of brain images. Medical Image Analysis 5:143–156. DOI: https://doi.org/10.1016/s1361-8415(01)00036-6, PMID: 11516708 g y p g Jenkinson M, Beckmann CF, Behrens TEJ, Woolrich MW, Smith SM. 2012. FSL. NeuroImage 62:782–790. DOI: https://doi.org/10.1016/j.neuroimage.2011.09.015, PMID: 21979382 Jones RC. 1941. A New Calculus for the Treatment of Optical Systems. Journal of the Optical Society of America 31:488. DOI: https://doi.org/10.1364/JOSA.31.000488 Jones RC. 1941. A New Calculus for the Treatment of Optical Systems. Journal of the Optical Society of America 31:488. DOI: https://doi.org/10.1364/JOSA.31.000488 31:488. DOI: https://doi.org/10.1364/JOSA.31.000488 p g Jucker M, Walker LC. 2013. Self-propagation of pathogenic protein aggregates in neurodegenerative diseases Nature 501:45–51. DOI: https://doi.org/10.1038/nature12481, PMID: 24005412 p g Kaiser R, Bartholdi E, Ernst RR. 1974. Diffusion and field-gradient effects in NMR Fourier spectroscopy. The Journal of Chemical Physics 60:2966–2979. DOI: https://doi.org/10.1063/1.1681477 y p g Kalra S, Müller H-P, Ishaque A, Zinman L, Korngut L, Genge A, Beaulieu C, Frayne R, Graham SJ, Kassubek J. 2020. A prospective harmonized multicenter DTI study of cerebral white matter degeneration in ALS. Neurology 95:e943–e952. DOI: https://doi.org/10.1212/WNL.0000000000010235, PMID: 32646955 gy p g Kamman RL, Go KG, Stomp GP, Hulstaert CE, Berendsen HJC. 1985. Changes of relaxation times T1 and T2 in rat tissues after biopsy and fixation. Magnetic Resonance Imaging 3:245–250. DOI: https://doi.org/10.1016/ 0730-725X(85)90353-4, PMID: 3908869 Kassubek J, Müller H-P, Del Tredici K, Brettschneider J, Pinkhardt EH, Lulé D, Böhm S, Braak H, Ludolph AC. 2014. Diffusion tensor imaging analysis of sequential spreading of disease in amyotrophic lateral sclerosis confirms patterns of TDP-43 pathology. Brain : A Journal of Neurology 137:1733–1740. DOI: https://doi.org/ 10.1093/brain/awu090, PMID: 24736303 , Kellner E, Dhital B, Kiselev VG, Reisert M. 2016. Gibbs-ringing artifact removal based on local subvoxel-shifts. Magnetic Resonance in Medicine 76:1574–1581. DOI: https://doi.org/10.1002/mrm.26054, PMID: 26745823 Keren NI, Taheri S, Vazey EM, Morgan PS, Granholm ACE, Aston-Jones GS, Eckert MA. 2015. Histologic validation of locus coeruleus MRI contrast in post-mortem tissue. NeuroImage 113:235–245. DOI: https://doi. org/10.1016/j.neuroimage.2015.03.020, PMID: 25791783 g j g Klink PC, Aubry JF, Ferrera VP, Fox AS, Froudist-Walsh S, Jarraya B, Konofagou EE, Krauzlis RJ, Messinger A, Mitchell AS, Ortiz-Rios M, Oya H, Roberts AC, Roe AW, Rushworth MFS, Sallet J, Schmid MC, Schroeder CE, Tasserie J, Tsao DY, et al. 2021. Combining Brain Perturbation and Neuroimaging in Non-human Primates. NeuroImage 235:118017. DOI: https://doi.org/10.1016/j.neuroimage.2021.118017, PMID: 33794355 Tasserie J, Tsao DY, et al. 2021. Combining Brain Perturbation and Neuroimaging in Non human Primates. NeuroImage 235:118017. References . Martins-Bach AB, Tachrount M, Tisca C, Qiu L, Spring S, Ellegood J, Nieman BJ, Sled JG, Raghavan-Nair Martins-Bach AB, Tachrount M, Tisca C, Qiu L, Spring S, Ellegood J, Nieman BJ, Sled JG, Raghavan-Nair R, Fisher E, Thomas C, Lerch J, M. K. L. 2021. Anatomical and microstructural 925 brain alterations in the TDP-M323K mouse model of amyotrophic lateral sclerosis. International Society of Magnetic Resonance in Medicine. . McNab JA, Jbabdi S, Deoni SCL, Douaud G, Behrens TEJ, Miller KL. 2009. High resolution diffusion-weighted imaging in fixed human brain using diffusion-weighted steady state free precession. NeuroImage 46:775–785. DOI: https://doi.org/10.1016/j.neuroimage.2009.01.008, PMID: 19344686 DOI: https://doi.org/10.1016/j.neuroimage.2009.01.008, PMID: 19344686 p g j g McNab JA, Miller KL. 2010. Steady-state diffusion-weighted imaging: Theory, acquisition and analysis. NMR in Biomedicine 23:781–793. DOI: https://doi.org/10.1002/nbm.1509, PMID: 20886565 Menke RAL, Körner S, Filippini N, Douaud G, Knight S, Talbot K, Turner MR. 2014. Widespread grey matter pathology dominates the longitudinal cerebral MRI and clinical landscape of amyotrophic lateral sclerosis. Brain : A Journal of Neurology 137:2546–2555. DOI: https://doi.org/10.1093/brain/awu162, PMID: 24951638 Menke RAL, Gray E, Lu CH, Kuhle J, Talbot K, Malaspina A, Turner MR. 2015. CSF neurofilament light chain reflects corticospinal tract degeneration in ALS. Annals of Clinical and Translational Neurology 2:748–755. DOI: https://doi.org/10.1002/acn3.212, PMID: 26273687 g Menke RAL, Proudfoot M, Wuu J, Andersen PM, Talbot K, Benatar M, Turner MR. 2016. Increased functional connectivity common to symptomatic amyotrophic lateral sclerosis and those at genetic risk. Journal of Neurology, Neurosurgery, and Psychiatry 87:580–588. DOI: https://doi.org/10.1136/jnnp-2015-311945, PMID: 26733601 Menke RAL, Proudfoot M, Talbot K, Turner MR. 2018. The two-year progression of structural and functional cerebral MRI in amyotrophic lateral sclerosis. NeuroImage. Clinical 17:953–961. DOI: https://doi.org/10.1016/j. nicl.2017.12.025, PMID: 29321969 Merboldt KD, Bruhn H, Frahm J, Gyngell ML, Hänicke W, Deimling M. 1989a. MRI of “diffusion” in the human brain: New results using a modified CE‐FAST sequence. Magnetic Resonance in Medicine 9:423–429. DOI: https://doi.org/10.1002/mrm.1910090316, PMID: 2710008 p g , Merboldt KD, Hxnicke W, Gyngell ML, Frahm J, Bruhn H. 1989b. Rapid NMR imaging of molecular self-diffusion using a modified CE-FAST sequence. Journal of Magnetic Resonance (1969) 82:115–121. DOI: https://doi.org/ 10.1016/0022-2364(89)90170-4 ( ) Milham MP, Ai L, Koo B, Xu T, Amiez C, Balezeau F, Baxter MG, Blezer ELA, Brochier T, Chen A, Croxson PL, Damatac CG, Dehaene S, Everling S, Fair DA, Fleysher L, Freiwald W, Froudist-Walsh S, Griffiths TD, Guedj C, et al. 2018. References DOI: https://doi.org/10.1016/j.neuroimage.2021.118017, PMID: 33794355 Langkammer C, Schweser F, Krebs N, Deistung A, Goessler W, Scheurer E, Sommer K, Reishofer G, Yen K, Fazekas F, Ropele S, Reichenbach JR. 2012. Quantitative susceptibility mapping (QSM) as a means to measure brain iron? A post mortem validation study. NeuroImage 62:1593–1599. DOI: https://doi.org/10.1016/j. neuroimage.2012.05.049, PMID: 22634862 g Le Bihan D. 1988. Intravoxel incoherent motion imaging using steady‐state free precession. Magnetic Resonance in Medicine 7:346–351. DOI: https://doi.org/10.1002/mrm.1910070312, PMID: 3205150 p g Liu C. 2010. Susceptibility tensor imaging. Magnetic Resonance in Medicine 63:1471–1477. DOI: https://doi.o 10.1002/mrm.22482, PMID: 20512849 Marcus DS, Wang TH, Parker J, Csernansky JG, Morris JC, Buckner RL. 2007. Open Access Series of Imaging Studies (OASIS): cross-sectional MRI data in young, middle aged, nondemented, and demented older adults. Journal of Cognitive Neuroscience 19:1498–1507. DOI: https://doi.org/10.1162/jocn.2007.19.9.1498, PMID: 17714011 Mars RB, Neubert FX, Verhagen L, Sallet J, Miller KL, Dunbar RIM, Barton RA. 2014. Primate comparative neuroscience using magnetic resonance imaging: promises and challenges. Frontiers in Neuroscience 8:298. DOI: https://doi.org/10.3389/fnins.2014.00298, PMID: 25339857 p g Mars RB, Sotiropoulos SN, Passingham RE, Sallet J, Verhagen L, Khrapitchev AA, Sibson N, Jbabdi S. 2018. Whole brain comparative anatomy using connectivity blueprints. eLife 7:e35237. DOI: https://doi.org/10.7554/ eLife.35237, PMID: 29749930 Mars RB, Jbabdi S, Rushworth MFS. 2021. A common space approach to comparative neuroscience. Annual Review of Neuroscience 44:69–86. DOI: https://doi.org/10.1146/annurev-neuro-100220-025942, PMID: 33534614 Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 24 of 35 Tools and resources Tools and resources Martins-Bach AB, Qiu L, Ellegood J, Wang N, Nieman BJ, Sled JG, Raghavan-Nair R, Fisher EMC, Cunningham TJ Lerch J Miller KL 2020 Brain structure in the homozygous FUSDelta14 mouse re Martins-Bach AB, Qiu L, Ellegood J, Wang N, Nieman BJ, Sled JG, Raghavan-Nair R, Fisher EMC, Cunningham TJ, Lerch J, Miller KL. 2020. Brain structure in the homozygous FUSDelta14 mouse recapitulate amyotrophic lateral sclerosis phenotypes. International Society of Magnetic Resonance in Medicine. . Martins Bach AB, Qiu L, Ellegood J, Wang N, Nieman BJ, Sled JG, Raghavan Nair R, Fisher EMC, Cunningham TJ, Lerch J, Miller KL. 2020. Brain structure in the homozygous FUSDelta14 mouse recapitulates amyotrophic lateral sclerosis phenotypes. International Society of Magnetic Resonance in Medicine. . Cunningham TJ, Lerch J, Miller KL. 2020. Brain structure in the homozygous FUSDelta14 mouse recapi amyotrophic lateral sclerosis phenotypes. International Society of Magnetic Resonance in Medicine. . g , , yg p amyotrophic lateral sclerosis phenotypes. International Society of Magnetic Resonance in Medicine. References Ex vivo diffusion MRI of the human brain: Technical challenges and recent advances. NMR in Biomedicine 32:e3941. DOI: https://doi.org/10.1002/nbm.3941, PMID: 29863793 g Rorden C, Hanayik T, Drake C, Hardcastle N, Newman-Norlund R. 2021. niivue/niivue. 0. doi.org/10.5281/ZENODO.5786270DOI: https://doi.org/10.5281/ZENODO.5786270 g p g Roumazeilles L, Eichert N, Bryant KL, Folloni D, Sallet J, Vijayakumar S, Foxley S, Tendler BC, Jbabdi S, Roumazeilles L, Eichert N, Bryant KL, Folloni D, Sallet J, Vijayakumar S, Foxley S, Tendler BC, Jbabdi S, Reveley C, Verhagen L, Dershowitz LB, Guthrie M, Flach E, Miller KL, Mars RB. 2020. Longitudinal connections and the organization of the temporal cortex in macaques, great apes, and humans. PLOS Biology 18:e3000810. DOI: https://doi.org/10.1371/journal.pbio.3000810, PMID: 32735557 Roumazeilles L, Eichert N, Bryant KL, Folloni D, Sallet J, Vijayakumar S, Foxley S, Tendler BC, Jbabdi S, Reveley C, Verhagen L, Dershowitz LB, Guthrie M, Flach E, Miller KL, Mars RB. 2020. Longitudinal connections and the organization of the temporal cortex in macaques, great apes, and humans. PLOS Biology 18:e3000810. DOI: https://doi.org/10.1371/journal.pbio.3000810, PMID: 32735557 p g j p Roumazeilles L, Lange FJ, Benn RA, Andersson JLR, Bertelsen MF, Manger PR, Flach E, Khrapitchev AA, Bryant KL, Sallet J, Mars RB. 2021. Cortical Morphology and White Matter Tractography of Three Phylogenetically Distant Primates: Evidence for a Simian Elaboration. Cerebral Cortex (New York, N.Y) 1:bhab285. DOI: https://doi.org/10.1093/cercor/bhab285, PMID: 34518890 Phylogenetically Distant Primates: Evidence for a Simian Elaboration. Cerebral Cortex (N 1:bhab285. DOI: https://doi.org/10.1093/cercor/bhab285, PMID: 34518890 1:bhab285. DOI: https://doi.org/10.1093/cercor/bhab285, PMID: 34518890 Sakai T, Hata J, Ohta H, Shintaku Y, Kimura N, Ogawa Y, Sogabe K, Mori S, Okano HJ, Hamada Y, Shibata S, Okano H, Oishi K. 2018. The Japan Monkey Centre Primates Brain Imaging Repository for comparative neuroscience: an archive of digital records including records for endangered species. Primates; Journal of Primatology 59:553–570. DOI: https://doi.org/10.1007/s10329-018-0694-3, PMID: 30357587i Schilling K, Gao Y, Janve V, Stepniewska I, Landman BA, Anderson AW. 2018. Confirmation of a gyral bias in diffusion MRI fiber tractography. Human Brain Mapping 39:1449–1466. DOI: https://doi.org/10.1002/hbm. 23936, PMID: 29266522 Shepherd TM, Thelwall PE, Stanisz GJ, Blackband SJ. 2009. Aldehyde fixative solutions alter the water relaxation and diffusion properties of nervous tissue. Magnetic Resonance in Medicine 62:26–34. DOI: https://doi.org/10. 1002/mrm.21977, PMID: 19353660 Snoek L, van der Miesen MM, Beemsterboer T, van der Leij A, Eigenhuis A, Steven Scholte H. 2021. The Amsterdam Open MRI Collection, a set of multimodal MRI datasets for individual difference analyses. Scientific Data 8:85. References DOI: https://doi.org/10.1038/s41597-021-00870-6, PMID: 33741990 Sun SW, Neil JJ, Song SK. 2003. Relative indices of water diffusion anisotropy are equivalent in live and formalin-fixed mouse brains. Magnetic Resonance in Medicine 50:743–748. DOI: https://doi.org/10.1002/mrm. 10605, PMID: 14523960 Sun SW, Neil JJ, Liang HF, He YY, Schmidt RE, Hsu CY, Song SK. 2005. Formalin fixation alters water diffusion coefficient magnitude but not anisotropy in infarcted brain. Magnetic Resonance in Medicine 53:1447–1451. DOI: https://doi.org/10.1002/mrm.20488, PMID: 15906292 p g Tendler BC, Foxley S, Cottaar M, Jbabdi S, Miller KL. 2020a. Modeling an equivalent b-value in diffusion- weighted steady-state free precession. Magnetic Resonance in Medicine 84:873–884. DOI: https://doi.org/10. 1002/mrm.28169, PMID: 31922283 Tendler BC, Foxley S, Hernandez-Fernandez M, Cottaar M, Scott C, Ansorge O, Miller KL, Jbabdi S. 2020b. Use of multi-flip angle measurements to account for transmit inhomogeneity and non-Gaussian diffusion in DW-SSFP. NeuroImage 220:117113. DOI: https://doi.org/10.1016/j.neuroimage.2020.117113, PMID: 32621975 g g j g Tendler BC, Qi F, Foxley S, Pallebage-Gamarallage M, Menke RAL, Ansorge O, Hurley SA, Miller KL. 2021. A method to remove the influence of fixative concentration on postmortem T2 maps using a kinetic tensor model. Human Brain Mapping 42:5956–5972. DOI: https://doi.org/10.1002/hbm.25661, PMID: 34541735 g g Thelwall PE, Shepherd TM, Stanisz GJ, Blackband SJ. 2006. Effects of temperature and aldehyde fixation on tissue water diffusion properties, studied in an erythrocyte ghost tissue model. Magnetic Resonance in Medicine 56:282–289. DOI: https://doi.org/10.1002/mrm.20962, PMID: 16841346 p g Thiessen JD, Zhang Y, Zhang H, Wang L, Buist R, Del Bigio MR, Kong J, Li XM, Martin M. 2013. Quantitative MRI and ultrastructural examination of the cuprizone mouse model of demyelination. NMR in Biomedicine 26:1562– 1581. DOI: https://doi.org/10.1002/nbm.2992, PMID: 23943390 1. DOI: https://doi.org/10.1002/nbm.2992, PMID: 23943 Toro R, Grisanti F, Herbin M, Santin M. 2014. The brain catalogue: an open portal for comparative neuroanatomy research. 20th Annual Meeting of the Organization for Human Brain Mapping (OHBM. . DOI: https://doi.org/ 10.13140/2.1.2141.0243 Van Essen DC, Smith SM, Barch DM, Behrens TEJ, Yacoub E, Ugurbil K, WU-Minn HCP Consortium. 2013. The WU-Minn Human Connectome Project: an overview. NeuroImage 80:62–79. DOI: https://doi.org/10.1016/j. neuroimage.2013.05.041, PMID: 23684880 g Vasung L, Rezayev A, Yun HJ, Song JW, van der Kouwe A, Stewart N, Palani A, Shiohama T, Chouinard-Decorte F, Levman J, Takahashi E. 2019. Structural and Diffusion MRI Analyses With Histological Observations in Patients With Lissencephaly. Frontiers in Cell and Developmental Biology 7:124. References An open resource for non-human primate imaging. Neuron 100:61–74. DOI: https://doi.org/10. 1016/j.neuron.2018.08.039, PMID: 30269990 Miller KL, Stagg CJ, Douaud G, Jbabdi S, Smith SM, Behrens TEJ, Jenkinson M, Chance SA, Esiri MM, Voets NL, Jenkinson N, Aziz TZ, Turner MR, Johansen-Berg H, McNab JA. 2011. Diffusion imaging of whole, post-mortem human brains on a clinical MRI scanner. NeuroImage 57:167–181. DOI: https://doi.org/10.1016/j.neuroimage. 2011.03.070, PMID: 21473920 Miller KL, McNab JA, Jbabdi S, Douaud G. 2012. Diffusion tractography of post-mortem human brains: Optimization and comparison of spin echo and steady-state free precession techniques. NeuroImage 59:2284– 2297. DOI: https://doi.org/10.1016/j.neuroimage.2011.09.054, PMID: 22008372 Miller KL, Alfaro-Almagro F, Bangerter NK, Thomas DL, Yacoub E, Xu J, Bartsch AJ, Jbabdi S, Sotiropoulos SN, Andersson JLR, Griffanti L, Douaud G, Okell TW, Weale P, Dragonu I, Garratt S, Hudson S, Collins R, Jenkinson M, Matthews PM, et al. 2016. Multimodal population brain imaging in the UK Biobank prospective epidemiological study. Nature Neuroscience 19:1523–1536. DOI: https://doi.org/10.1038/nn.4393, PMID: 27643430 Mollink J, Kleinnijenhuis M, Cappellen van Walsum AM, Sotiropoulos SN, Cottaar M, Mirfin C, Heinrich MP, Jenkinson M, Pallebage-Gamarallage M, Ansorge O, Jbabdi S, Miller KL. 2017. Evaluating fibre orientation dispersion in white matter: Comparison of diffusion MRI, histology and polarized light imaging. NeuroImage 157:561–574. DOI: https://doi.org/10.1016/j.neuroimage.2017.06.001, PMID: 28602815 p g j g Nagara H, Inoue T, Koga T, Kitaguchi T, Tateishi J, Goto I. 1987. Formalin fixed brains are useful for magnetic resonance imaging (MRI) study. Journal of the Neurological Sciences 81:67–77. DOI: https://doi.org/10.1016/ 0022-510x(87)90184-5, PMID: 3681342 Ohnishi T, Nakamura Y, Tanaka T, Tanaka T, Hashimoto N, Haneishi H, Batchelor TT, Gerstner ER, Taylor JW, Snuderl M, Yagi Y. 2016. Deformable image registration between pathological images and MR image via an optical macro image. Pathology, Research and Practice 212:927–936. DOI: https://doi.org/10.1016/j.prp.2016. 07.018, PMID: 27613662 Pallebage-Gamarallage M, Foxley S, Menke RAL, Huszar IN, Jenkinson M, Tendler BC, Wang C, Jbabdi S, Turner MR, Miller KL, Ansorge O. 2018. Dissecting the pathobiology of altered MRI signal in amyotrophic lateral sclerosis: A post mortem whole brain sampling strategy for the integration of ultra-high-field MRI and quantitative neuropathology. BMC Neuroscience 19:11. DOI: https://doi.org/10.1186/s12868-018-0416-1, PMID: 29529995 Pfefferbaum A, Sullivan EV, Adalsteinsson E, Garrick T, Harper C. 2004. Postmortem MR imaging of formalin- fixed human brain. NeuroImage 21:1585–1595. DOI: https://doi.org/10.1016/j.neuroimage.2003.11.024, PMID: 15050582 Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 25 of 35 Tools and resources Tools and resources Roebroeck A, Miller KL, Aggarwal M. 2019. References DOI: https:// doi.org/10.3389/fcell.2019.00124, PMID: 31355197 Wang C, Foxley S, Ansorge O, Bangerter-Christensen S, Chiew M, Leonte A, Menke RA, Mollink J, Pallebage-Gamarallage M, Turner MR, Miller KL, Tendler BC. 2020. Methods for quantitative susceptibility and R2* mapping in whole post-mortem brains at 7T applied to amyotrophic lateral sclerosis. NeuroImage 222:117216. DOI: https://doi.org/10.1016/j.neuroimage.2020.117216, PMID: 32745677 Weigel M. 2015. Extended phase graphs: Dephasing, RF pulses, and echoes - Pure and simple. Journal of Magnetic Resonance Imaging 41:266–295. DOI: https://doi.org/10.1002/jmri.24619, PMID: 24737382 Weigel M, Dechent P, Galbusera R, Bahn E, Nair G, Lu PJ, Kappos L, Brück W, Stadelmann C, Granziera C. 2021. Imaging multiple sclerosis pathology at 160 μm isotropic resolution by human whole-brain ex vivo magnetic resonance imaging at 3 T. Scientific Reports 11:1–13. DOI: https://doi.org/10.1038/s41598-021-94891-1, PMID: 34326420 Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 26 of 35 26 of 35 Tools and resources Tools and resources Wilhelm MJ, Ong HH, Wehrli SL, Li C, Tsai PH, Hackney DB, Wehrli FW. 2012. Direct magnetic resonance detection of myelin and prospects for quantitative imaging of myelin density. PNAS 109:9605–9610. DOI: https://doi.org/10.1073/pnas.1115107109, PMID: 22628562 https://doi.org/10.1073/pnas.1115107109, PMID: p g p Wilkinson M, Wang R, van der Kouwe A, Takahashi E. 2016. White and gray matter fiber pathways in autism spectrum disorder revealed by ex vivo diffusion MR tractography. Brain and Behavior 6:e00483. DOI: https:// doi.org/10.1002/brb3.483, PMID: 27247853 g Winkler AM, Ridgway GR, Webster MA, Smith SM, Nichols TE. 2014. Permutation inference for the general linear model. NeuroImage 92:381–397. DOI: https://doi.org/10.1016/j.neuroimage.2014.01.060, PMID: 24530839 Wu W, Baxter L, Rieger SW, Adams E, Andersson JL, Andrade MC, Andritsou F, Bastiani M, Fry RE, Frost R, Fitzgibbon S, Foxley S, Fowler D, Gallagher C, Howard AF, Hajnal JV, Moultrie F, Monk V, Porter DA, Papp D, et al. 2021. The Forget-Me-Not DHCP Study: 7 Tesla High Resolution Diffusion Imaging in the Unfixed Post-Mortem Neonatal Brain. Bioengineering 1:449666. DOI: https://doi.org/10.1101/2021.06.24.449666 Yarnykh VL. 2007. Actual flip-angle imaging in the pulsed steady state: A method for rapid three-dimensional mapping of the transmitted radiofrequency field Magnetic Resonance in Medicine 57:192–200 DOI: https:// Wu W, Baxter L, Rieger SW, Adams E, Andersson JL, Andrade MC, Andritsou F, Bastiani M, Fry RE, Frost R, Fitzgibbon S, Foxley S, Fowler D, Gallagher C, Howard AF, Hajnal JV, Moultrie F, Monk V, Porter DA, Papp D, et al. 2021. The Forget-Me-Not DHCP Study: 7 Tesla High Resolution Diffusion Imaging in the Unfixed Post-Mortem Neonatal Brain. Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 References Bioengineering 1:449666. DOI: https://doi.org/10.1101/2021.06.24.449666 et al. 2021. The Forget-Me-Not DHCP Study: 7 Tesla High Resolution Diffusion Imaging in the Unfixed Post-Mortem Neonatal Brain. Bioengineering 1:449666. DOI: https://doi.org/10.1101/2021.06.24.449666 Yarnykh VL. 2007. Actual flip-angle imaging in the pulsed steady state: A method for rapid three-dimensional mapping of the transmitted radiofrequency field. Magnetic Resonance in Medicine 57:192–200. DOI: https:// doi.org/10.1002/mrm.21120, PMID: 17191242 Yarnykh VL. 2007. Actual flip-angle imaging in the pulsed steady state: A method for rapid three-dimensional mapping of the transmitted radiofrequency field. Magnetic Resonance in Medicine 57:192–200. DOI: https:// doi.org/10.1002/mrm.21120, PMID: 17191242 g Yong-Hing CJ, Obenaus A, Stryker R, Tong K, Sarty GE. 2005. Magnetic resonance imaging and mathematical modeling of progressive formalin fixation of the human brain. Magnetic Resonance in Medicine 54:324–332. DOI: https://doi.org/10.1002/mrm.20578, PMID: 16032673 p g Zhang Y, Brady M, Smith S. 2001. Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm. IEEE Transactions on Medical Imaging 20:45–57. DOI: https://doi.org/10.1109/42.906424, PMID: 11293691 27 of 35 Tools and resources Tools and resources Neuroscience Neuroscience Human high-resolution diffusion MRI-PLI dataset MRI preparation and scanning Data were acquired from a post-mortem human brain (n=1) with no known neuropathology. The brain was extracted from the skull within 72 hr after death and fixed in 10% PBS buffered formalin (4% formaldehyde) for 6 weeks prior to scanning. The brain was removed from formalin and placed in plastic bags filled with Fomblin LC08 (Solvay Solexis), a susceptibility-matched perfluoropolyether liquid that contributes no signal to the imaging experiment. The brain was imaged with a Siemens 7T whole-body scanner (1 Tx/32 Rx head coil). Diffusion- weighted volumes were acquired using DW-SSFP sequence. As highlighted in the main text, the choice of DW-SSFP was motivated by the sequences potential to simultaneously address the short T2 and low diffusivity of fixed, post-mortem tissue, when limited to human scanners. Whole-brain diffusion MRI datasets were acquired at 500 μm, 1 mm, and 2 mm isotropic resolution. Details of the acquisition parameters are provided in Appendix 1—table 1, where we note that the 500 μm dataset took approximately 6 days of continuous scanning to acquire. DW-SSFP datasets were obtained at two flip angles to address B1-inhomogeneity at 7T, as previously described in Tendler et al., 2020b. The DW-SSFP signal is dependent on tissue relaxation time-constants (T1 and T2) and the acquisition flip angle, which must be estimated for accurate modeling. These parameters were estimated using a TIR, TSE, and actual flip angle imaging (AFI) (Yarnykh, 2007) sequence. A structural scan was additionally acquired using a true fast imaging with steady-state precession (TRUFI) sequence (bSSFP), which produces high gray/white matter contrast in post-mortem tissue. Details of the acquisition parameters are provided in Appendix 1—table 2. MRI processing DOI: https://doi.org/10.7554/eLife.73153 28 of 35 Tools and resources Tools and resources Neuroscience DW-SSFP (0.5 mm) DW-SSFP (1.0 mm) Resolution (μm3) 500·500·500 Resolution (mm3) 1.0·1.0·1.0 TE (ms) 21 TE (ms) 21 TR (ms) 30 TR (ms) 30 EPI factor 1 EPI factor 1 Bandwidth (Hz per pixel) 198 Bandwidth (Hz per pixel) 130 Acquisition time (per direction/non-DW) 45 min 03 s No. of averages 1 Acquisition time (total) 6 days 0 hr No. of averages 1 DW-SSFP (2.0 mm) q-value (cm–1) 300 Diffusion gradient duration (ms) 14.10 Diffusion gradient strength (mTm–1) 50 Flip angles (°) 33 and 98 No. of directions (per flip angle) 221 No. of non-DW (per flip angle) 6 (q=20 cm–1) Resolution (mm3) 2.0·2.0·2.0 TE (ms) 21 TR (ms) 30 EPI factor 1 Bandwidth (Hz per pixel) 130 No. of averages 1 Appendix 1—table 2. Acquisition parameters for the TIR, TSE, structural (TRUFI), and B1-mapping (AFI) sequences. Turbo inversion-recovery (TIR) True-Fast Imaging with SSFP (TRUFI) Resolution (mm3) 0.75·0.75·1.60 Resolution (μm3) 312.5·312.5·500 Number of inversions 6 TE (ms) 5.95 TE ( ) 12 TR ( ) 11 9 Appendix 1—table 1 Continued 3. DOI: https://doi.org/10.7554/eLife.73153 29 of 35 DW-SSFP (0.5 mm) DW-SSFP (1.0 mm) Resolution (μm3) 500·500·500 Resolution (mm3) 1.0·1.0·1.0 TE (ms) 21 TE (ms) 21 TR (ms) 30 TR (ms) 30 EPI factor 1 EPI factor 1 Bandwidth (Hz per pixel) 198 Bandwidth (Hz per pixel) 130 Acquisition time (per direction/non-DW) 45 min 03 s No. of averages 1 Acquisition time (total) 6 days 0 hr No. of averages 1 DW-SSFP (2.0 mm) q-value (cm–1) 300 Diffusion gradient duration (ms) 14.10 Diffusion gradient strength (mTm–1) 50 Flip angles (°) 33 and 98 No. of directions (per flip angle) 221 No. of non-DW (per flip angle) 6 (q=20 cm–1) Resolution (mm3) 2.0·2.0·2.0 TE (ms) 21 TR (ms) 30 EPI factor 1 Bandwidth (Hz per pixel) 130 No. of averages 1 Appendix 1—table 2. Acquisition parameters for the TIR, TSE, structural (TRUFI), and B1-mapping (AFI) sequences. MRI processing A Gibbs ringing correction was applied to the DW-SSFP, TIR, and TSE datasets (Kellner et al., 2016). All coregistrations within-and-between modalities were performed using a six degrees-of-freedom coregistration using FSL FLIRT (Jenkinson and Smith, 2001). T1 maps were estimated from the TIR volumes assuming mono-exponential signal evolution. T2 maps were estimated from the TSE volumes using an EPG framework (Weigel, 2015), as described in Tendler et al., 2021. B1 maps were estimated from the AFI volumes as described in the original AFI publication (Yarnykh, 2007). Structural scans were estimated from the TRUFI volumes, with banding artifacts minimized by taking the maximum intensity across volumes (Bangerter et al., 2004). For the diffusion outputs, Tensor and Ball and Two-Stick models were fit to the DW-SSFP data as described in Tendler et al., 2020b. In brief, fitting was performed using the full DW-SSFP Buxton model (Buxton, 1993), estimating a shared set of diffusion orientations (e.g., tensor eigenvectors), and a unique set of diffusivity estimates (e.g., tensor eigenvalues) per DW-SSFP flip angle. The fitting process incorporated the estimated T1, T2, and B1 maps, in addition to a noise-floor correction. The DW-SSFP sequence does not have a well-defined b-value. To address this, the diffusivity estimates at each flip angle were combined to generate diffusivity estimates at an effective b-value of 4000 s/mm2. Details of this procedure, in addition to the motivation behind the choice of 4000 s/mm2 are detailed in Tendler et al., 2020b. Note that a small modification was made to the original minimization procedure, as described below. Appendix 1—table 1. DW-SSFP Acquisition parameters at 0.5, 1.0, and 2.0 mm. DW-SSFP (0.5 mm) DW-SSFP (1.0 mm) q-value (cm–1) 300 q-value (cm–1) 300 Diffusion gradient duration (ms) 14.10 Diffusion gradient duration (ms) 14.10 Diffusion gradient strength (mTm–1) 50 Diffusion gradient strength (mTm–1) 50 Flip angles (°) 33 and 98 Flip angles (°) 33 and 98 No. of directions (per flip angle) 90 No. of directions (per flip angle) 60 No. of non-DW (per flip angle) 6 (q=20 cm–1) No. of non-DW (per flip angle) 5 (q=20 cm–1) Appendix 1—table 1 Continued on next page Appendix 1—table 1. DW-SSFP Acquisition parameters at 0.5, 1.0, and 2.0 mm. Tendler et al. eLife 2022;11:e73153. Modification to minimization procedure T dl l 2020b d ib d h i Modification to minimization procedure T dl l 2020b d ib d h i Tendler et al., 2020b described an approach to estimate DW-SSFP diffusivity estimates at a single effective b-value, achieved by incorporating a non-Gaussian diffusion model into the DW-SSFP signal equations. In Tendler et al., 2020b, non-Gaussianity was modeled using a Gamma distribution of diffusivities, estimating a mean (‍Dm‍) and standard deviation (‍Ds‍) of the Gamma distribution per voxel. Here, the Gamma fitting procedure (Equation 4 in Tendler et al., 2020b was replaced with: min Dmi,Dsi ( Lisim : αlow ( Dmi,Dsi ) −Liexp : αlow ) SD ( Liexp : αlow ) 2 2 + ( Lisim : αhigh ( Dmi,Dsi ) −Liexp : αhigh ) SD ( Liexp : αhigh ) 2 2‍, (1) (1) where ‍αlow‍ and ‍αhigh‍ are the voxelwise DW-SSFP flip angles, ‍Liexp : αlow/αhigh‍ are the voxelwise experimental diffusivity estimates (Tensor eigenvalues or Ball and Two-Stick diffusivity estimates) at each flip angle, ‍Lisim : αlow/αhigh‍ are the simulated diffusivity estimates for a given ‍Dmi‍ and ‍Dsi‍ , and ‍SD ( Liexp : αlow/αhigh ) ‍ are the estimated experimental standard deviation of the diffusivity estimates. This approach was found to reduce spurious diffusivity estimates in regions of low SNR. For further details of the modeling approach, see Tendler et al., 2020b. are the voxelwise PLI preparation, scanning, and processing p p g p g Tissue samples from the anterior commissure, corpus callosum, occipital lobe gurus, pons, thalamus, and external capsule were extracted from the post-mortem brain. Samples were stored in a 30% sucrose solution with PBS and 0.025% azide for 3 weeks. Tissue blocks were subsequently embedded in optimal cutting temperature compound (Sakura, Finetek Inc, USA) and frozen to –80°C. 60-μm sections were cut from the tissue blocks with a cryostat microtome (Leica, Germany). No tissue staining was performed, as birefringence is naturally expressed by the myelin sheath. PLI was performed using a Leica DM4000B microscope, equipped with a polarizing filter, a quarter wave plate (QWP), and a rotatable analyzer with orientation ‍ρ‍. Samples were illuminated with a white LED (pE-100wht Cooled). The fast axis of the QWP was oriented 45° with respect to the transmission axis of the polarising filter to create circular polarization. The rotating analyzer captured the phase shift induced by the myelin sheath. A total of 18 images were acquired for each field of view at equidistant analyzer orientation angles, ‍ρ = { 0◦, 10◦, . . . , 170 ◦} ‍. Images were magnified 1.25× (0.04 NA, Leica) and captured with a Leica DFC420 CCD camera (4 μm/pixel). The green color image channel was used for further analysis. The entire sample was imaged via raster scanning, with each row composed of multiple contiguous field-of-views (FOVs). These FOVs were automatically stitched together using in-house software (MATLAB 2015b, MathWorks, Natick, MA). For each specimen, a series of background images were acquired to correct for illumination inhomogeneities (Dammers et al., 2010). Microscopic fiber orientations were derived using Jones calculus (Jones, 1941), as described below. PLI fiber orientations MRI processing Turbo inversion-recovery (TIR) True-Fast Imaging with SSFP (TRUFI) Resolution (mm3) 0.75·0.75·1.60 Resolution (μm3) 312.5·312.5·500 Number of inversions 6 TE (ms) 5.95 TE (ms) 12 TR (ms) 11.9 TR (ms) 1000 Flip angle (°) 35 TIs (ms) 31, 62, 125, 250, 500, and 850 Bandwidth (Hz per pixel) 130 Flip angle (°) 180 Phase increments (o) 0 and 180 Bandwidth (Hz per pixel) 199 Number of averages (per set of increments) 16 Number of averages 1 Turbo spin-echo (TSE) – T2 Actual flip-angle imaging (AFI) – B1 Resolution (mm3) 0.75·0.75·1.60 Resolution (mm3) 1.50·1.50·1.50 Number of echoes 6 TE (ms) 1.5 Appendix 1—table 2 Continued on next page Turbo inversion-recovery (TIR) True-Fast Imaging with SSFP (TRUFI) Resolution (mm3) 0.75·0.75·1.60 Resolution (μm3) 312.5·312.5·500 Number of inversions 6 TE (ms) 5.95 TE (ms) 12 TR (ms) 11.9 TR (ms) 1000 Flip angle (°) 35 TIs (ms) 31, 62, 125, 250, 500, and 850 Bandwidth (Hz per pixel) 130 Flip angle (°) 180 Phase increments (o) 0 and 180 Bandwidth (Hz per pixel) 199 Number of averages (per set of increments) 16 Number of averages 1 Turbo spin-echo (TSE) – T2 Actual flip-angle imaging (AFI) – B1 Resolution (mm3) 0.75·0.75·1.60 Resolution (mm3) 1.50·1.50·1.50 Number of echoes 6 TE (ms) 1.5 Appendix 1—table 2 Continued on next page Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 Tools and resources Tools and resources Neuroscience Appendix 1—table 2 Continued Turbo inversion-recovery (TIR) True-Fast Imaging with SSFP (TRUFI) TEs (ms) 14, 28, 42, 56, 70, and 84 TR1/TR2 (ms) 6/30 TR (ms) 1,000 Flip angle (°) 60 Flip angle (°) 180 Bandwidth (Hz per pixel) 630 Bandwidth (Hz per pixel) 130 Number of averages 1 Number of averages 1 Digital Brain Zoo datasets Here, we provide the acquisition and processing protocol for four previously unreleased datasets t the Digital Brain Zoo. PLI fiber orientations i The light intensity (‍I‍) for a birefringent specimen inside a PLI-setup is described using Jones calculus (Jones, 1941), defining: I ( ρ ) = I0 2 [ 1 + sin ( 2ρ −2φ ) · sin δ ] ,‍ (2) I ( ρ ) = I0 2 [ 1 + sin ( 2ρ −2φ ) · sin δ ] ,‍ (2) 30 of 35 Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 Tools and resources Tools and resources Neuroscience where ‍I0‍ is the average light intensity, ‍ρ‍ is the polarizer orientation, ‍φ‍ is the in-plane orientation of the myelin sheath, and ‍δ‍ is the phase shift, defined as: ere ‍I0‍ is the average light intensity, ‍ρ‍ is the polarizer orientation, ‍φ‍ is the in-plane orientation of e myelin sheath, and ‍δ‍ is the phase shift, defined as: δ ≈2π d·∆n λ · cos2 α,‍ (3) (3) where ‍d‍ is the sample thickness, ‍∆n‍ is the sample birefringence, ‍λ‍ is the light wavelength, and ‍α‍ is the inclination angle of the myelin sheath (α). Microscopic fiber orientations were derived using as above, fitting to each pixel in the raw PLI images as previously reported in Axer et al., 2011. Appendix 1—table 3 Continued on next page Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 European wolf and Hamadryas baboon p y Formalin-fixed European wolf and Hamadryas baboon brains were provided by Copenhagen zoo. Prior to scanning, the brains were rehydrated using a PBS solution. The size of the European wolf and Hamadryas baboon brain necessitated scanning on the Siemens 7T whole-body scanner (1 Tx/28 Rx knee coil, QED), using the brain holder displayed in Appendix 3—figure 2. The brain holder was filled with fluorinert (FC-3283, 3M) during the scanning procedure, a susceptibility matched fluid that gives off no signal. Diffusion-weighted volumes were acquired using DW-SSFP sequence. As highlighted in the main text, the choice of DW-SSFP was motivated by the sequences potential to simultaneously address the short T2 and low diffusivity of fixed, post-mortem tissue, when limited to human scanners. A structural scan was additionally acquired using a TRUFI sequence (bSSFP), which produces high gray/white matter contrast in post-mortem tissue. Acquisition parameters are provided in Appendix 1—table 3. Structural scans were formed by averaging over all 16 TRUFI datasets (root-mean sum of squares). Diffusion datasets were processed using a similar approach to the great ape datasets in Bryant et al., 2021 and (Roumazeilles et al., 2020). In brief, a Gibbs ringing correction (Kellner et al., 2016) was applied to the diffusion and non-diffusion weighted datasets, with all coregistrations performed using FSL FLIRT (Jenkinson and Smith, 2001). Fitting was performed using the full DW- SSFP Buxton model (Buxton, 1993) adapted to incorporate Tensor and Ball and Two-Stick estimates. The fitting process incorporated estimated T1, T2, and B1 maps derived from a TIR, TSE, and AFI (Yarnykh, 2007) sequence acquired in the same session. Appendix 1—table 3. Acquisition parameters for the DW-SSFP and structural (TRUFI) sequences. The only difference between the European wolf and Hamadryas baboon acquisition was the number of non-diffusion weighted directions acquired (13 for wolf and 11 for baboon). DW-SSFP True-Fast Imaging with SSFP (TRUFI) q-value (cm–1) 300 Resolution (μm3) 217·217·220 Diffusion gradient duration (ms) 13.56 TE (ms) 7.33 Diffusion gradient strength (mTm–1) 52 TR (ms) 14.65 Flip angle (°) 39 Flip angle (°) 30 No. of directions 160 Bandwidth (Hz per pixel) 100 No. of non-DW 13/11 (q=20 cm–1) Phase increments (o) 0, 45, 90, 135, 180, 225, 270, and 315 Resolution (μm3) 600·600·600 No. of averages (per set of increments) 2 TE (ms) 21 TR (ms) 29 EPI factor 1 Appendix 1—table 3 Continued on next page Tendler et al. eLife 2022;11:e73153. European wolf and Hamadryas baboon DOI: https://doi.org/10.7554/eLife.73153 31 of 35 Tools and resources Tools and resources Neuroscience DW-SSFP True-Fast Imaging with SSFP (TRUFI) Bandwidth (Hz per pixel) 100 Acquisition time (per direction/non-DW) 16 min 25 s Acquisition time (total) 1 day 20 hr No. of averages 1 Appendix 1—table 3 Continued Appendix 1—table 3 Continued Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 Human ALS MRI-histology callosum analysis gy y A comparison of diffusivity properties between the ALS and control cohort (12 ALS and 3 control brains) was performed in the corpus callosum, as displayed in Figure 3c (Main Text). To achieve this, a standard-space mask of the corpus callosum was first generated using the Jülich atlas (Eickhoff et al., 2005). The callosum mask was subsequently split into five distinct regions of interest (ROIs) associated with specific fiber projections as proposed by Hofer and Frahm (Hofer and Frahm, 2006), and transformed into the space of each post-mortem brain. Briefly, a standard space FA template (FMRIB58_FA, available as part of FSL) was modified to display similar contrast to the post-mortem FA maps. Coregistration matrices were subsequently estimated between the FA map of each post- mortem brain and the modified standard space FA template using a non-linear coregistration (ANTS) (Avants et al., 2011). The callosum masks were subsequently coregistered into the space of each post-mortem brain using the estimated coregistration matrices, and multiplied by a white matter mask (generated using FAST; Zhang et al., 2001) to remove any remaining gray matter regions. Diffusion estimates were obtained by taking the mean over each ROI. Values were normalized to the splenium (Par/Temp/Occ) estimate, which has been proposed as a control region with little pathological burden in ALS (Cardenas et al., 2017). Differences in the normalized FA, MD, axial, and radial diffusivity between the ALS and control cohort were assessed with a two-tailed, family- wise error rate (FWER) corrected t-test using PALM (Winkler et al., 2014). Full results are provided in Appendix 2—table 1. Although our statistical analysis does account for sample size, it does not consider other confounds that may contribute to differences between the two groups (e.g., age and sex). Source data for the corpus callosum analysis are provided in a Supplementary file 1. Appendix 2—table 1. p-values associated with differences between the ALS and control cohort for the diffusivity estimates. Here ‘p’ defines the p-value, and ‘PFWER’ defines the FWER-corrected p-value (‍∗‍=p<0.05; ‍∗∗‍=pFWER<0.05). The largest differences between the ALS and control cohort were found in the Body (Pre/Supp Motor) category, followed by the Genu (PreFrontal) and Body (Motor) category. No differences were found in the Body (Sensory) category. Cotton-Top and Golden Lion tamarinsi Formalin-fixed Cotton-Top and Golden Lion tamarin brains were provided by Copenhagen zoo. Prior to scanning, the brains were rehydrated using a phosphate-buffered saline solution. Scanning was performed using a 7T magnet with Agilent Direct-Drive console and 72 mm ID quadrature birdcage RF coil (Rapid Biomedical GmbH). The brain holder was filled with fluorinert during the scanning procedure, a susceptibility matched fluid that gives off no signal. Diffusion-weighted volumes were acquired using diffusion-weighted spin-echo protocol with single line readout (DW-SEMS) sequence. Acquisition parameters are provided in Appendix 1—table 4. The diffusion datasets were processed using a similar approach to the prosimian and monkey data in Bryant et al., 2021. Datasets were preprocessed using FSL tools implemented in the Phoenix module of the MR Comparative Anatomy Toolbox (Mr Cat, https://www.neuroecologylab. org/). Tensor and ball and Two/Three-stick estimates were derived using FSL’s dtifit and bedpostX (Behrens et al., 2007). Appendix 1—table 4. Acquisition parameters for the DW-SEMS sequence. Appendix 1—table 4. Acquisition parameters for the DW-SEMS sequence. DW SEMS Appendix 1—table 4. Acquisition parameters for the DW-SEMS sequence. DW-SEMS b-value (s/mm2) 4000 δ (ms) 7 Δ (ms) 13 Diffusion gradient strength (mTm–1) 320 No. of directions 128 No. of non-DW 16 Resolution (μm3) 300·300·300 TE (ms) 25 TR (s) 10 EPI factor 1 Bandwidth (kHz) 100 Acquisition time (per direction/non-DW) 21 min 20 s Acquisition time (total) 2 days 4 hr No. of averages 1 Appendix 1—table 4. Acquisition parameters for t DW-SEMS b-value (s/mm2) 4000 δ (ms) 7 Δ (ms) 13 Diffusion gradient strength (mTm–1) 320 No. of directions 128 No. of non-DW 16 Resolution (μm3) 300·300·300 TE (ms) 25 TR (s) 10 EPI factor 1 Bandwidth (kHz) 100 Acquisition time (per direction/non-DW) 21 min 20 s Acquisition time (total) 2 days 4 hr No. of averages 1 32 of 35 Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 Tools and resources Tools and resources Neuroscience Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 Human ALS MRI-histology callosum analysis Body (Sensory) Body (Motor) Body (Pre/Supp Motor) Genu (PreFrontal) Fractional anisotropy (FA) p=0.34 pFWER=0.70 p=0.042* pFWER=0.11 p=0.0044* pFWER=0.013** p=0.013* pFWER=0.037** Mean diffusivity (MD) p=0.99 pFWER=1.00 p=0.18 pFWER=0.48 p=0.015* pFWER=0.053 p=0.037* pFWER=0.12 Axial diffusivity (AD) p=0.53 pFWER=0.92 p=0.58 pFWER=0.95 p=0.084 pFWER=0.27 p=0.11 pFWER=0.32 Radial diffusivity (RD) p=0.66 pFWER=0.97 p=0.073 pFWER=0.23 p=0.0022* pFWER=0.015** p=0.022* pFWER=0.062 33 of 35 Neuroscience Appendix 3 Post-mortem brain holders Appendix 3—figure 1. Post-mortem human brain holder. The brain holder ensures consistent placement during scanning. Here, the custom holder tightly seals the brain in place, whilst the 3D printed shell (provided by Dr Alard Roebroeck, Maastricht University) prevents pressure on the brain. The holder is designed with a spherical cavity to maximize field homogeneity. For further information on our scanning procedure for human post-mortem brains, see Wang et al., 2020. Tools and resources Tools and resources Neuroscience Appendix 3 Post-mortem brain holders Post-mortem brain holders Post-mortem brain holders Appendix 3—figure 1. Post-mortem human brain holder. The brain holder ensures consistent placement during scanning. Here, the custom holder tightly seals the brain in place, whilst the 3D printed shell (provided by Dr Alard Roebroeck, Maastricht University) prevents pressure on the brain. The holder is designed with a spherical cavity to maximize field homogeneity. For further information on our scanning procedure for human post-mortem brains, see Wang et al., 2020. Appendix 3—figure 1. Post-mortem human brain holder. The brain holder ensures consistent placement during scanning. Here, the custom holder tightly seals the brain in place, whilst the 3D printed shell (provided by Dr Alard Roebroeck, Maastricht University) prevents pressure on the brain. The holder is designed with a spherical cavity to maximize field homogeneity. For further information on our scanning procedure for human post-mortem brains, see Wang et al., 2020. Appendix 3—figure 2. Post-mortem cylindrical brain holder. The brain holder used for scanning large nonhuman brains which fit inside the 28 channel QED knee coil. This consisted of a cylindrical container, with plastic gauze (black) used to secure samples during the acquisition. Appendix 3—figure 2. Post-mortem cylindrical brain holder. The brain holder used for scanning large nonhuman brains which fit inside the 28 channel QED knee coil. This consisted of a cylindrical container, with plastic gauze (black) used to secure samples during the acquisition. 34 of 35 Tendler et al. eLife 2022;11:e73153. Tendler et al. eLife 2022;11:e73153. DOI: https://doi.org/10.7554/eLife.73153 Human ALS MRI-histology callosum analysis DOI: https://doi.org/10.7554/eLife.73153 Tools and resources Tools and resources Neuroscience Appendix 4 Example structural MRI dataset Example structural MRI dataset Example structural MRI dataset Example structural MRI dataset Appendix 4—figure 1. Structural MRI. Example structural MRI dataset acquired using a bSSFP sequence in the European wolf (Canis lupus) at a resolution of 220 μm (isotropic). bSSFP Structural MRI datasets display excellent gray-white matter contrast, facilitating the delineation of fine tissue structures and integration with processing pipelines for surface reconstruction. Contrast in bSSFP datasets is reversed compared to conventional T1-weighted structural MRI scans (gray matter appears bright, and white matter appears dark), which must be accounted for in any analysis pipeline. Appendix 4—figure 1. Structural MRI. Example structural MRI dataset acquired using a bSSFP sequence in the European wolf (Canis lupus) at a resolution of 220 μm (isotropic). bSSFP Structural MRI datasets display excellent gray-white matter contrast, facilitating the delineation of fine tissue structures and integration with processing pipelines for surface reconstruction. Contrast in bSSFP datasets is reversed compared to conventional T1-weighted structural MRI scans (gray matter appears bright, and white matter appears dark), which must be accounted for in any analysis pipeline. 35 of 35
https://openalex.org/W2905807945	https://europepmc.org/articles/pmc6309238?pdf=render	English	null	Vascular Targeting to Increase the Efficiency of Immune Checkpoint Blockade in Cancer	Frontiers in immunology	2,018	cc-by	7,660	Vascular Targeting to Increase the Efﬁciency of Immune Checkpoint Blockade in Cancer Maria Georganaki †, Luuk van Hooren † and Anna Dimberg* Department of Immunology, Genetics and Pathology, Science for Life Laboratory, Uppsala University, The Rudbeck Laboratory, Uppsala, Sweden Boosting natural immunity against malignant cells has had a major breakthrough in clinical cancer therapy. This is mainly due to the successful development of immune checkpoint blocking antibodies, which release a break on cytolytic anti-tumor-directed T-lymphocytes. However, immune checkpoint blockade is only effective for a proportion of cancer patients, and a major challenge in the ﬁeld is to understand and overcome treatment resistance. Immune checkpoint blockade relies on successful trafﬁcking of tumor-targeted T-lymphocytes from the secondary lymphoid organs, through the blood stream and into the tumor tissue. Resistance to therapy is often associated with a low density of T-lymphocytes residing within the tumor tissue prior to treatment. The recruitment of leukocytes to the tumor tissue relies on up-regulation of adhesion molecules and chemokines by the tumor vasculature, which is denoted as endothelial activation. Tumor vessels are often poorly activated due to constitutive pro-angiogenic signaling in the tumor microenvironment, and therefore constitute barriers to efﬁcient leukocyte recruitment. An emerging possibility to enhance the efﬁciency of cancer immunotherapy is to combine pro-inﬂammatory drugs with anti-angiogenic therapy, which can enable tumor-targeted T-lymphocytes to access the tumor tissue by relieving endothelial anergy and increasing adhesion molecule expression. This would pave the way for efﬁcient immune checkpoint blockade. Here, we review the current understanding of the biological basis of endothelial anergy within the tumor microenvironment, and discuss the challenges and opportunities of combining vascular targeting with immunotherapeutic drugs as suggested by data from key pre-clinical and clinical studies. Edited by: Patrik Andersson, Harvard Medical School, United States Patrik Andersson, Harvard Medical School, United States Reviewed by: Lasse Dahl Ejby Jensen, Linköping University, Sweden Krithika Kodumudi, Mofﬁtt Cancer Center, United States Correspondence: Anna Dimberg anna.dimberg@igp.uu.se †These authors have contributed equally to this work Reviewed by: Lasse Dahl Ejby Jensen, Linköping University, Sweden Krithika Kodumudi, Mofﬁtt Cancer Center, United States Reviewed by: Lasse Dahl Ejby Jensen, Linköping University, Sweden Krithika Kodumudi, Mofﬁtt Cancer Center, United States Correspondence: Anna Dimberg anna.dimberg@igp.uu.se †These authors have contributed equally to this work †These authors have contributed equally to this work Specialty section: This article was submitted to Cancer Immunity and Immunotherapy, a section of the journal Frontiers in Immunology Received: 28 September 2018 Accepted: 12 December 2018 Published: 21 December 2018 INTRODUCTION The ﬁeld of cancer immunotherapy has made signiﬁcant improvements during the last decade due to the development of new eﬀective means to boost tumor immune responses and achieve long-term remission or even cures in patients that were previously deemed to be untreatable. A major breakthrough was the development of antibodies targeting negative regulators of T-cell activation, termed immune checkpoints. Ipilimumab, an antagonistic antibody targeting cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) improved overall survival in metastatic melanoma patients in 2010 (1). Following the success of anti-CTLA-4 therapy, antibodies targeting programmed cell death protein 1 (PD-1), or its ligand PD-L1, MINI REVIEW MINI REVIEW published: 21 December 2018 doi: 10.3389/ﬁmmu.2018.03081 published: 21 December 2018 doi: 10.3389/ﬁmmu.2018.03081 Citation: Georganaki M, van Hooren L and Dimberg A (2018) Vascular Targeting to Increase the Efﬁciency of Immune Checkpoint Blockade in Cancer. Front. Immunol. 9:3081. doi: 10.3389/ﬁmmu.2018.03081 December 2018 \| Volume 9 \| Article 3081 Frontiers in Immunology \| www.frontiersin.org Vascular Targeting to Improve Immunotherapy Georganaki et al. is still under active debate (16), blocking CTLA-4 in cancer enhances T-cell activation, but can also lead to autoimmune responses. proved to be eﬀective at improving overall survival in a wide variety of cancers (2–7). Importantly, a proportion of patients achieve long-term remission, highlighting the potential of immune checkpoint blockade to induce durable responses (8). The encouraging results of these studies has sparked an interest from the cancer research ﬁeld and inspired further investigations into targeting of alternative immune checkpoint molecules. p After activation, which generally occurs in secondary lymphoid organs, T-cells circulate and extravasate through the vasculature at sites of inﬂammation to locate and kill target cells displaying the cognate peptide antigen on their MHC molecules. At the tumor site, T-cell activity can be hampered by several types of immunosuppression, including engagement of PD-1 expressed on T-cells by its ligand PD-L1 expressed on stromal cells and/or malignant cells (18, 19). Thus, anti-cancer immunity can be enhanced by antibodies that block the PD-1/PD-L1 interaction. Although manipulating T-cell activation status by blocking inhibitory receptors or enhancing co-stimulatory molecules has proven to be eﬃcacious in boosting anti-tumor immune responses, these treatments strictly rely on eﬃcient transport of lymphocytes from the site of T-cell activation to the tumor tissue. It is therefore not surprising that tumors that are not inﬁltrated by T-cells, and tumors where T-cell inﬁltration is only observed at the tumor border but not in the core, do not respond well to immune checkpoint blockade (3, 9). Several mechanisms contribute to regulating the inﬂammatory state, including the mutational landscape of the tumor, expression of chemokines, and checkpoint molecules and recruitment of immunosuppressive cells (20, 21). In cases where an immune response is correctly mounted but where lymphocyte recruitment to the tumor tissue is lacking, pharmacologically altering vascular phenotype to allow eﬃcient leukocyte traﬃcking may sensitize resistant tumors for immunotherapy. While checkpoint blockade represents a breakthrough in cancer therapy, a majority of cancer patients do not respond and some tumor types appear to be intrinsically resistant. Citation: The treatment is designed to boost an ongoing immune response and is ineﬃcient in cases where initial immune activation is lacking, including tumors that are devoid of inﬁltrating T-cells (3, 9). Development of therapeutic strategies to enhance immune cell recruitment may therefore increase the proportion of patients responding to immune checkpoint blockade. Circulating T- cells are recruited through expression of adhesion molecules and chemokines on the endothelial cells, collectively mediating capture, rolling, and transmigration of leukocytes from the blood stream into the inﬂamed tissue (10). In many types of cancer, constitutive stimulation by pro-angiogenic factors secreted in the tumor microenvironment renders the vasculature morphologically and functionally abnormal, constituting a barrier to eﬃcient leukocyte recruitment. In this mini-review we summarize phenotypical diﬀerences between normal vessels and tumor vessels in mediating leukocyte recruitment, the molecular mechanisms that underlie these functional changes and current eﬀorts to improve immune checkpoint blockade through vascular targeting. Frontiers in Immunology \| www.frontiersin.org IMMUNE CHECKPOINT BLOCKADE THERAPY RELIES ON EFFICIENT T-LYMPHOCYTE RECRUITMENT FIGURE 1 \| Anti-angiogenic therapy can relieve endothelial anergy, improve vessel function and enhance T-cell inﬁltration. (A) Aberrant pro-angiogenic signaling in the tumor microenvironment gives rise to an anergic endothelium with reduced pericyte coverage, disrupted endothelial cell junctions, and suboptimal activation status. Anti-angiogenic therapy reverts those defects and permits for enhanced leukocyte recruitment, through the leukocyte adhesion cascade. Chemokines and adhesion molecules on the activated endothelial surface allow for leukocyte capture, rolling, arrest, and transendothelial migration into the tumor tissue. (B) Aberrant pro-angiogenic signaling in tumors is associated with dysfunctional and anergic tumor vessels, which are not capable of recruiting tumor-targeted leukocytes (left panel). Vascular targeting can relieve endothelial anergy, improve perfusion and increase the recruitment of leukocytes into the tumor microenvironment (right panel). FIGURE 1 \| Anti-angiogenic therapy can relieve endothelial anergy, improve vessel function and enhance T-cell inﬁltration. (A) Aberrant pro-angiogenic signaling in the tumor microenvironment gives rise to an anergic endothelium with reduced pericyte coverage, disrupted endothelial cell junctions, and suboptimal activation status. FIGURE 1 \| Anti-angiogenic therapy can relieve endothelial anergy, improve vessel function and enhance T-cell inﬁltration. (A) Aberrant pro-angiogenic signaling in the tumor microenvironment gives rise to an anergic endothelium with reduced pericyte coverage, disrupted endothelial cell junctions, and suboptimal activation status. Anti-angiogenic therapy reverts those defects and permits for enhanced leukocyte recruitment, through the leukocyte adhesion cascade. Chemokines and adhesion molecules on the activated endothelial surface allow for leukocyte capture, rolling, arrest, and transendothelial migration into the tumor tissue. (B) Aberrant pro-angiogenic signaling in tumors is associated with dysfunctional and anergic tumor vessels, which are not capable of recruiting tumor-targeted leukocytes (left panel). Vascular targeting can relieve endothelial anergy, improve perfusion and increase the recruitment of leukocytes into the tumor microenvironment (right panel). one of the hallmarks of cancer (24). Angiogenesis is induced as a result of enhanced growth factor secretion in the tumor microenvironment, shifting the balance from predominantly angiostatic to pro-angiogenic signaling (25). This “angiogenic switch,” observed as a shift from avascular to vascular tumors, can occur in dormant, and slow growing tumors and be associated with tumor progression to higher malignancy grades. to eﬃcient leukocyte capture and integrin activation through mechanical forces (22). Transendothelial migration can occur through either through paracellular or transcellular pathways (10, 23). Finally, leukocytes migrate through the basement membrane and pericyte layer to reach the inﬂamed tissue (10). IMMUNE CHECKPOINT BLOCKADE THERAPY RELIES ON EFFICIENT T-LYMPHOCYTE RECRUITMENT Recruitment of lymphocytes to the tumor tissue strictly depends on eﬃcient regulation of molecules required for cell-cell interactions during capture, rolling, adhesion, and transendothelial migration. Several mechanisms can trigger neovascularization in tumors, including hypoxia, genetic alterations in tumor cells, expression of cytokines, and growth factorsm and recruitment of bone marrow-derived circulating cells (26–28). When proliferation of malignant cells results in a tumor mass that cannot be suﬃciently oxygenated by pre-existing vasculature this leads to hypoxia. Hypoxia-induced stabilization of hypoxia- inducible factor (HIF)-1α triggers up-regulation of its target genes, including several pro-angiogenic genes such as vascular endothelial growth factor (VEGF) (29). VEGF secreted by IMMUNE CHECKPOINT BLOCKADE THERAPY RELIES ON EFFICIENT T-LYMPHOCYTE RECRUITMENT Leukocyte recruitment by activated endothelial cells and subsequent migration through the vessel wall is mediated by direct molecular interactions between proteins expressed by leukocytes and endothelial cells (Figure 1A). This ﬁnely tuned process, known as the leukocyte adhesion cascade, involves leukocyte capture, rolling, adhesion, arrest, and transendothelial migration (10). This is enabled by up-regulation of adhesion molecules and chemokines on the surface of endothelial cells, denoted “endothelial activation.” Leukocyte capture and rolling are mainly mediated by interaction between selectins expressed on endothelial cells (P-selectin and E- selectin) and leukocytes (L-selectin) to carbohydrate ligands including P-selectin glycosylated ligand 1. Firm adhesion of leukocytes is mediated through interaction of leukocyte integrins with endothelial adhesion molecules. For T-lymphocytes, ﬁrm adhesion is mainly induced by lymphocyte function-associated antigen (LFA)-1 and very late activation antigen (VLA)-4 binding to endothelial intercellular adhesion molecule (ICAM)- 1 and vascular cell adhesion molecule (VCAM)-1, respectively. Activation of integrins via inside-out signaling associated with chemokine stimulation triggers leukocyte arrest to the endothelium (10). Blood ﬂow-derived shear stress contributes Immune checkpoint blockade works through inhibiting negative feedback loops that downregulate T-cell activation following an initial immune response. T-cell activation and T-cell receptor signaling has recently been reviewed in detail (11, 12). T-cells remain naïve until they encounter licensed antigen-presenting cells (APC)s that present the correct peptide antigen on major histocompatibility complex (MHC) molecules together with the appropriate co-stimulatory molecules. T-cell activation requires recognition of the MHC-antigen complex displayed on an APC, engagement of co-stimulatory molecules such as CD28 on the T cell with B7 family members on the APC and stimulation by inﬂammatory cytokines. In response to T-cell activation, other co-stimulatory molecules such as ICOS and OX40 are expressed, but also molecules that instigate negative feedback loops to prevent over-activation of T-cells. One of those negative feedback molecules is CTLA-4, which competes with CD28 for binding to B7 family members expressed on the surface of APCs (13– 16). CTLA-4 is also highly expressed on regulatory T cells, and antibodies targeting CTLA-4 have been suggested to deplete them from the tumor microenvironment through Fc eﬀector functions (17). Although the relative importance of the immune checkpoint and regulatory depletion mechanisms for therapeutic eﬃcacy December 2018 \| Volume 9 \| Article 3081 Frontiers in Immunology \| www.frontiersin.org Frontiers in Immunology \| www.frontiersin.org 2 Vascular Targeting to Improve Immunotherapy Georganaki et al. TUMOR ANGIOGENESIS RESULTS IN MORPHOLOGICALLY AND FUNCTIONALLY DISTINCT VESSELS Tumors need access to capillary network to proliferate, and the ability of tumors to stimulate angiogenesis is recognized as December 2018 \| Volume 9 \| Article 3081 Frontiers in Immunology \| www.frontiersin.org 3 Vascular Targeting to Improve Immunotherapy Georganaki et al. adhesion molecules and chemokines (51). Pro-inﬂammatory cytokines are abundantly expressed in many cancers, but pro- angiogenic factors present in the tumor microenvironment can suppress expression of adhesion molecules and chemokines that attract cytolytic T-cells and NK cells such as CXCL10 and CXCL11 (41, 52, 53). VEGF-induced signaling pathways can directly interfere with TNF-α-induced NF-κB activation, globally repressing TNF-α-induced gene expression in endothelial cells (53). Consistent with this, antagonizing VEGFR2 signaling sensitizes endothelial cells to TNF-α (54). However, the interplay between angiogenesis and inﬂammation is context dependent. TNF-α stimulation synergistically primes endothelial cells for VEGF-induced angiogenesis (55). Notably, VEGF stimulation can induce leukocyte inﬁltration in some systems, and pathways downstream of VEGF signaling can both induce and repress adhesion molecule expression (56–58). Nitric oxide and molecules such as epidermal growth factor-like domain 7 can also regulate adhesion molecule expression and clustering in tumors (59, 60). Another less studied feature of endothelial regulation of tumor immunity is the selective recruitment of immunosuppressive leukocytes through expression of speciﬁc adhesion molecules such as the common lymphatic endothelial and vascular endothelial receptor-1 (CLEVER-1) (41). tumor cells diﬀuses through the tissue and activates its receptor VEGFR2 expressed on endothelial cells (30). Downstream of VEGFR2 activation, multiple intracellular pathways are induced that regulate cell division, survival, sprouting, and migration of endothelial cells (30). Several other pro-angiogenic growth factors contribute to tumor angiogenesis, including the family of angiopoietins and their cognate receptor TIE-2 and the ﬁbroblast growth factor family (31). Some tumors harbor mutations of the gene coding for the von Hippel-Lindau protein, a crucial member of the ubiquitin ligase complex that degrades HIF-1α (32). These mutations stabilize HIF-1α, allowing expression of pro-angiogenic factors under normoxic conditions. Myeloid cells, including macrophages, neutrophils, and myeloid derived suppressor cells (MDSCs), can also stimulate vessel formation through expression of pro-angiogenic factors and/or matrix metalloproteases that release VEGF from extracellular matrix (33). Physiological angiogenesis is a well-controlled process that is attenuated when the need for new vessels have been met, but tumor angiogenesis is deregulated and continuous due to excessive expression of pro-angiogenic factors (34). TUMOR ANGIOGENESIS RESULTS IN MORPHOLOGICALLY AND FUNCTIONALLY DISTINCT VESSELS Tumor angiogenesis can give rise to disorganized vessels that are tortuous, dilated and poorly covered by pericytes (35). The tumor vasculature is often leaky due to endothelial junctional defects, blood ﬂow is generally slow and perfusion is irregular (25). Gene expression analyses have shown that tumor vessels diﬀer molecularly from their normal counterparts and have revealed a high level of vessel heterogeneity depending on the resident tumor tissue (36–40). Importantly, tumor vessels can have multiple phenotypes ranging from normal to dysfunctional and the morphology and functionality signiﬁcantly diﬀer depending on tumor type and anatomical site. In addition to regulating leukocyte entry, tumor endothelial cells can alter the anti-tumor immune response by modulating immune cell activity or viability. This can occur as a response of endothelial cells to tumor-derived growth factors (61). The concept of a “tumor endothelial barrier” refers to molecules expressed on endothelial cells that inhibit promote T-cell arrest. An example of this is tumor endothelial upregulation of FasL in response to tumor-derived VEGF, IL-10 and prostaglandin E2, which has been shown to selectively kill eﬀector CD8 T- cells but not Treg cells (44, 62). Endothelial cells can express several inhibitory molecules including immune checkpoint molecules [PD-L1, T-cell immunoglobulin domain and mucin domain (TIM3), B7-H3 and B7-H4], death receptor-ligands (TNF-related apoptosis-inducing ligand (TRAIL) and secreted immunomodulatory factors (IL-6, prostaglandin E (PGE) 2, IL- 10, and TGF-β) (44, 62).The relative importance of endothelial expression of these molecules in immunosuppression and their regulation in tumor vessels need further investigation. Antigen presentation by endothelial cells suggests that they can function as potential antigen presenting cells (63). Whether tumor endothelial cells present antigen and if this is suﬃcient for activation of T-cells, or alternatively induces T-cell anergy, is still unknown. As discussed below, anti-angiogenic therapies can alleviate endothelial anergy and enhance T-cell recruitment in tumors (53, 64–67). Immunosuppressive molecules expressed on tumor endothelial cells represent new potential targets for novel combination treatments with immunotherapy. TUMOR BLOOD VESSELS ARE BARRIERS TO EFFICIENT LEUKOCYTE RECRUITMENT Immune cells in the circulation are dependent on the vascular network to reach the tumor and kill malignant cells. However, functional abnormalities of tumor blood vessels represent diﬃcult hurdles for leukocyte recruitment. The architectural defects of tumor vessels limit perfusion and alter sheer stress, and diﬀerential protein expression in tumor endothelial cells can dampen the immune response (34, 41–44). Tumor endothelial cell respond ineﬃciently to pro-inﬂammatory signaling, and fail to express suﬃcient levels of molecules involved in the leukocyte capture, adhesion and extravasation process (Figure 1). Downregulation or ineﬀective clustering of adhesion molecules on tumor endothelial cells limits T-cell inﬁltration and inhibit anti-tumor immunity (45–47). Reduced expression of adhesion molecules in tumor vessels has been observed in several types of human cancer (48–50). Endothelial activation is generally induced by binding of pro-inﬂammatory cytokines such as tumor necrosis factor (TNF)α and interleukin (IL)-1 to their endothelial receptors, leading to activation of the transcription factor nuclear factor-κB (NF-κB) and up-regulation of selectins, Frontiers in Immunology \| www.frontiersin.org ANTI-ANGIOGENIC THERAPY CAN IMPROVE THE EFFECT OF IMMUNE CHECKPOINT BLOCKADE nutrients (68) led to intense research eﬀorts and sparked numerous clinical trials. A number of anti-angiogenic drugs have been approved to date, several of which are antibodies or small tyrosine kinase inhibitors that target VEGF/VEGFR signaling (69). The ﬁrst clinically approved drug was a humanized antibody targeting VEGF named Bevacizumab. Treatment with Bevacizumab slows tumor growth in patients with non-small cell lung and colorectal cancer, though with only a marginal improvement of long- term survival (70, 71). It has also been approved for patients with cervical cancer, glioblastoma, ovarian cancer and renal cell carcinoma (72). In breast, melanoma, pancreatic, and prostate cancer no improvement of overall survival has been observed (73). The importance of a functional vasculature for immune cell recruitment justiﬁes eﬀorts of combining immunotherapy with vascular targeting to improve vessel function and enhance up-regulation of adhesion molecules and chemokines. Inhibition of angiogenic signaling using sub-maximal doses of anti- angiogenic drugs may result in a normalization of vascular function and improve the eﬃcacy of other anti-cancer drugs, as proposed by Jain (78). Anti-angiogenic therapy provides relief of continuous angiogenic signaling, which at sub-maximal doses can result in vessel pruning, maturation, and improved perfusion (69). For cancer immunotherapy, there is an added beneﬁt that anti-angiogenic drugs enhance expression of adhesion molecules and chemokines involved in T-cell recruitment (53, 64–67). Therefore, combining immunotherapy with anti- angiogenic drugs may relieve endothelial anergy and induce lymphocyte inﬁltration into tumors that prior to treatment were of an immune-excluded phenotype (Figure 1B). Indeed, by combining adoptive T-cell transfer with anti-VEGF therapy in murine melanoma, tumor T-cell inﬁltration was increased and survival was prolonged (79). An important challenge in this concept is that the dosing of anti-angiogenic drugs is crucial for normalizing vessels and improving T-cell recruitment, and that the optimal dose may diﬀer between patients (80). Nevertheless, the combination of immunotherapy and anti-angiogenic therapy has shown beneﬁt in various therapeutic settings (Table 1). Treatment of colorectal cancer patients with Bevacizumab results in an initial response with decreased tumor growth or even regression. However, relapse is common, associated with rapid rebound angiogenesis, and tumor regrowth is often more aggressive than before anti-angiogenic treatment (74). Several mechanisms have been proposed for the resistance to anti-angiogenic treatment, including co-option of normal vessels in the surrounding tissue, recruitment of pro-angiogenic myeloid cells and upregulation of alternative pro-angiogenic factors (25). SUCCESSES AND FAILURES OF ANTI-ANGIOGENIC THERAPY The idea that anti-angiogenic therapy could block tumor progression by depriving the tumor cells of oxygen and December 2018 \| Volume 9 \| Article 3081 Frontiers in Immunology \| www.frontiersin.org Frontiers in Immunology \| www.frontiersin.org 4 Vascular Targeting to Improve Immunotherapy Georganaki et al. TABLE 1 \| Selected studies combining anti-angiogenic therapy with immune checkpoint blockade in preclinical models and clinical trials. Anti-angiogenic target Immune checkpoint target Cancer model Survival References PRECLINICAL MODELS VEGF (B20-4.1.1) PD-L1 (6E11) SCLC + (81) VEGFR2 (DC101) PD-1 (RMPI-14) Colon-26 adenocarcinoma + (82) VEGF and ANG2 (Vanucizumab) PD-1 (RMPI-14) MMTV-PyMT, RIP1-Tag2, Melanoma, Neuroendocrine + * (83) VEGFR-1,-2 and−3 (Axitinib) CTLA-4 (9H10) Melanoma + (84) VEGFR2* (Sunitinib) PD-1 (RMPI-14) Colon cancer + (85) VEGFR2 (DC101) PD-L1 (10F.9G2) Pancreatic cancer, breast cancer and glioblastoma + (86) VEGF + ANG2 (10F.9G2 + CVX-241) PD-L1 (10F.9G2) Breast cancer +/- (87) Anti-angiogenic target Immune checkpoint target Cancer type Trial References CLINICAL TRIALS VEGFR-1,-2 and−3 (Axitinib) PD-1 (Pembrolizumab) Renal cell cancer Phase 3 (88) VEGF (Bevacizumab) CTLA-4 (Ipilimumab) Metastatic melanoma Phase 1 (89) (90) (91) VEGFR-1,-2 and−3 (Axitinib) PD-L1 (Avelumab) Advanced clear-cell renal cell carcinoma Phase 1b (92) VEGFR-1,-2 and−3 (Lenvatinib) PD-1 (Pembrolizumab) Renal cell cancer Retrospective (93) VEGF (Bevacizumab) PD-L1 (Atezolizumab) Metastatic renal cell carcinoma Phase 1b (94) Antibody clone or brand name in brackets. SCLC = small-cell lung cancer, * broad tyrosine kinase inhibitor, increased T-cell exhaustion, * increased T-cell numbers and endothelial activation. Ongoing clinical trials are available at www.clinicaltrials.gov and were recently reviewed by Fukumura et al. (95). TABLE 1 \| Selected studies combining anti-angiogenic therapy with immune checkpoint blockade in preclinical models and clinical trials. Survival References body clone or brand name in brackets. SCLC = small-cell lung cancer, * broad tyrosine kinase inhibitor, increased T-cell exhaustion, * incre ation. Ongoing clinical trials are available at www.clinicaltrials.gov and were recently reviewed by Fukumura et al. (95). Frontiers in Immunology \| www.frontiersin.org ANTI-ANGIOGENIC THERAPY CAN IMPROVE THE EFFECT OF IMMUNE CHECKPOINT BLOCKADE Notably, anti-angiogenic treatment can increase invasiveness and promote metastasis formation in experimental models of cancer (75, 76). Although metastasis-promoting eﬀects of anti-angiogenic therapy have not been observed in clinical studies, the pre-clinical work has cautioned the ﬁeld and questioned how anti-angiogenic therapy should best be administrated (77). December 2018 \| Volume 9 \| Article 3081 Frontiers in Immunology \| www.frontiersin.org 5 Vascular Targeting to Improve Immunotherapy Georganaki et al. Drugs targeting VEGF/VEGFR2 signaling have been observed to enhance the response to immune checkpoint antibodies in pre-clinical tumor models. The combination of anti-VEGF and anti-VEGFR2 antibodies prolonged survival in a murine model of adenocarcinoma in combination with PD-1 blockade (82). Similarly, an antibody targeting both VEGF and Angiopoeitin- 2 improved responses to PD-1 inhibition in preclinical cancer models (83). The VEGFR inhibitor axitinib combined with anti-CTLA-4, but neither monotherapy, prolonged survival of mice bearing murine melanoma (84). This was associated with increased numbers of CD4+ and CD8+ T cells in the tumor after the combination treatment. In addition to their eﬀect on vessel phenotype, therapies targeting pro-angiogenic factors can alleviate immunosuppression by directly aﬀecting the immune cells. For example, the tyrosine kinase inhibitor Sunitinib can decrease MDSCs and Tregs (67, 96, 97). in enhanced T-cell inﬁltration and reduced tumor growth (86). Vessel normalization in combination with a vascular targeting peptide coupled to LIGHT, a ligand for the lymphotoxin β receptor, induced HEVs and tertiary lymphoid structures in Rip1-Tag5 pancreatic neuroendocrine tumors. Importantly, this therapeutic approach sensitized these tumors to anti-PD-1 and anti-CTLA-4 antibody therapy (101). These studies indicate that beyond normalizing vessels, transforming tumor vessels to HEVs can be of additional beneﬁt in enhancing the response to cancer immunotherapy. Furthermore, HEVs may promote formation of tertiary lymphoid structures which have been associated with a beneﬁcial response to cancer immunotherapy in several types of cancer (100, 102). Current eﬀorts in vascular targeting aim to improve the eﬃcacy of cancer immunotherapy through inhibition of pro- angiogenic signaling. However, several immunosuppressive molecules that contribute to the tumor endothelial barrier are regulated through alternative pathways, and may be induced secondary to immune activation. This aspect has not yet been suﬃciently explored. An increased understanding of the cross- talk between tumor cells, endothelial cells, and immune cells during immune checkpoint blockade therapy may lead to new combinatorial treatment regimens that enhance the abundance of activated T-cells in tumor tissue. ANTI-ANGIOGENIC THERAPY CAN IMPROVE THE EFFECT OF IMMUNE CHECKPOINT BLOCKADE This can ultimately increase the proportion of patients that respond to immune checkpoint blockade. The ﬁrst phase I clinical trial combining anti-angiogenic therapy with immune checkpoint blockade was a study using Bevacizumab and ipilimumab (anti-CTLA-4). The combination therapy modulated tumor vessel morphology and induced endothelial activation, associated with increased inﬁltration of dendritic cells and cytotoxic T-cells in melanoma tumors (89, 98). Similarly, combining atezolizumab (anti-PD-L1) with Bevacizumab in patients with metastatic renal cell carcinoma resulted in enhanced traﬃcking of lymphocytes, and increased cytotoxic T cells (94). Following these promising results, several clinical trials with the same therapeutic rationale have been initiated (95, 98, 99). ACKNOWLEDGMENTS We apologize to the authors of original work that was not cited in this mini-review due to space constraint. FUNDING An emerging concept is that vascular targeting in combination with immune checkpoint blockade may promote tumor immunity by inducing formation of high-endothelial venules (HEV)s. HEVs are specialized vessels found in secondary lymphoid organs that are adapted for lymphocyte traﬃcking (100). The combination of anti-VEGFR2 antibodies with PD-L1 antibodies induced formation of HEVs and improved T-cell inﬁltration in the polyoma middle T oncoprotein (PyMT) breast cancer model and the Rip1-Tag2 pancreatic neuroendocrine tumor model (RT2-PNET) (86). Formation of HEVs in glioblastoma models required further stimulation using a lymphotoxin β receptor agonistic antibody, resulting Swedish Cancer Society (CAN 2017/502), the Swedish Childhood Cancer Society (PR2015-0133, NCP2015-0075), the Swedish Research Council (Dnr 2016-02495), the Emil and Wera Cornells Stiftelse foundation, Senior Investigator Award from the Swedish Cancer Society (CAN 2015/1216) to AD. FUTURE DIRECTIONS BEYOND NORMALIZATION AND ENDOTHELIAL ACTIVATION All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication. REFERENCES Cold Spring Harb Perspect Med. (2012) 2:a006429. doi: 10.1101/cshperspect.a006429 17. Arce Vargas F, Furness AJS, Litchﬁeld K, Joshi K, Rosenthal R, Ghorani E, et al. Fc eﬀector function contributes to the activity of human Anti-CTLA-4 antibodies. Cancer Cell. (2018) 33:649–63 e4. doi: 10.1016/j.ccell.2018.02.010 40. Zhang L, He L, Lugano R, Roodakker K, Bergquist M, Smits A, et al. IDH-mutation status is associated with distinct vascular gene expression signatures in lower grade gliomas. Neuro Oncol. (2018) 20:1505–16. doi: 10.1093/neuonc/noy088 18. Dong H, Zhu G, Tamada K, Chen L. B7-H1, a third member of the B7 family, co-stimulates T-cell proliferation and interleukin-10 secretion. Nat Med. (1999) 5:1365–9. doi: 10.1038/70932 41. Mauge L, Terme M, Tartour E, Helley D. Control of the adaptive immune response by tumor vasculature. Front Oncol. (2014) 4:61. doi: 10.3389/fonc.2014.00061 19. Freeman GJ, Long AJ, Iwai Y, Bourque K, Chernova T, Nishimura H, et al. Engagement of the PD-1 immunoinhibitory receptor by a novel B7 family member leads to negative regulation of lymphocyte activation. J Exp Med. (2000) 192:1027–34. doi: 10.1084/jem.192.7.1027 42. Slaney CY, Kershaw MH, Darcy PK. Traﬃcking of T cells into tumors. Cancer Res. (2014) 74:7168–74. doi: 10.1158/0008-5472.CAN-14-2458 43. Peske JD, Woods AB, Engelhard VH. Control of CD8 T-cell inﬁltration into tumors by vasculature and microenvironment. Adv Cancer Res. (2015) 128:263–307. doi: 10.1016/bs.acr.2015.05.001 20. Popovic A, Jaﬀee EM, Zaidi N. Emerging strategies for combination checkpoint modulators in cancer immunotherapy. J Clin Invest. (2018) 128:3209–18. doi: 10.1172/JCI120775 44. Lanitis E, Irving M, Coukos G. Targeting the tumor vasculature to enhance T cell activity. Curr Opin Immunol. (2015) 33:55–63. doi: 10.1016/j.coi.2015.01.011 21. Binnewies M, Roberts EW, Kersten K, Chan V, Fearon DF, Merad M, et al. Understanding the tumor immune microenvironment (TIME) for eﬀective therapy. Nat Med. (2018) 24:541–50. doi: 10.1038/s41591-018-0014-x 45. Bouzin C, Brouet A, De Vriese J, Dewever J, Feron O. Eﬀects of vascular endothelial growth factor on the lymphocyte-endothelium interactions: identiﬁcation of caveolin-1 and nitric oxide as control points of endothelial cell anergy. J Immunol. (2007) 178:1505–11. doi: 10.4049/jimmunol.178.3.1505 22. Alon R, Dustin ML. Force as a facilitator of integrin conformational changes during leukocyte arrest on blood vessels and antigen-presenting cells. Immunity (2007) 26:17–27. doi: 10.1016/j.immuni.2007.01.002 23. Filippi MD. Mechanism of diapedesis: importance of the transcellular route. Adv Immunol. (2016) 129:25–53. doi: 10.1016/bs.ai.2015.09.001 46. Buckanovich RJ, Facciabene A, Kim S, Benencia F, Sasaroli D, Balint K, et al. REFERENCES in cancer patients. Nature (2014) 515:563–7. doi: 10.1038/nature 14011 1. Hodi FS, O’Day SJ, McDermott DF, Weber RW, Sosman JA, Haanen JB, et al. Improved survival with ipilimumab in patients with metastatic melanoma. N Engl J Med. (2010) 363:711–23. doi: 10.1056/NEJMoa1003466 1. Hodi FS, O’Day SJ, McDermott DF, Weber RW, Sosman JA, Haanen JB, et al. Improved survival with ipilimumab in patients with metastatic melanoma. N Engl J Med. (2010) 363:711–23. doi: 10.1056/NEJMoa1003466 4. Powles T, Eder JP, Fine GD, Braiteh FS, Loriot Y, Cruz C, et al. MPDL3280A (anti-PD-L1) treatment leads to clinical activity in metastatic bladder cancer. Nature (2014) 515:558–62. doi: 10.1038/nature13904 2. Hamid O, Robert C, Daud A, Hodi FS, Hwu WJ, Keﬀord R, et al. Safety and tumor responses with lambrolizumab (anti-PD-1) in melanoma. N Engl J Med. (2013) 369:134–44. doi: 10.1056/NEJMoa1305133 5. Topalian SL, Sznol M, McDermott DF, Kluger HM, Carvajal RD, Sharfman WH, et al. Survival, durable tumor remission, and long-term safety in patients with advanced melanoma receiving nivolumab. J Clin Oncol. (2014) 32:1020–30. doi: 10.1200/JCO.2013. 53.0105 3. Herbst RS, Soria JC, Kowanetz M, Fine GD, Hamid O, Gordon MS, et al. Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A December 2018 \| Volume 9 \| Article 3081 Frontiers in Immunology \| www.frontiersin.org 6 Vascular Targeting to Improve Immunotherapy Georganaki et al. 30. Olsson AK, Dimberg A, Kreuger J, Claesson-Welsh L. VEGF receptor signalling - in control of vascular function. Nat Rev Mol Cell Biol. (2006) 7:359–71. doi: 10.1038/nrm1911 6. Topalian SL, Hodi FS, Brahmer JR, Gettinger SN, Smith DC, McDermott DF, et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N Engl J Med. (2012) 366:2443–54. doi: 10.1056/NEJMoa1200690 7. Ribas A, Wolchok JD. Cancer immunotherapy using checkpoint blockade. Science (2018) 359:1350–5. doi: 10.1126/science.aar4060 31. Welti J, Loges S, Dimmeler S, Carmeliet P. Recent molecular discoveries in angiogenesis and antiangiogenic therapies in cancer. J Clin Invest. (2013) 123:3190–200. doi: 10.1172/JCI70212 8. Schadendorf D, Hodi FS, Robert C, Weber JS, Margolin K, Hamid O, et al. Pooled analysis of long-term survival data from phase II and phase III trials of ipilimumab in unresectable or metastatic melanoma. J Clin Oncol. (2015) 33:1889–94. doi: 10.1200/JCO.2014.56.2736 32. Miller F, Kentsis A, Osman R, Pan ZQ. Inactivation of VHL by tumorigenic mutations that disrupt dynamic coupling of the pVHL.hypoxia-inducible transcription factor-1alpha complex. J Biol Chem. (2005) 280:7985–96. doi: 10.1074/jbc.M413160200 9. REFERENCES Tumeh PC, Harview CL, Yearley JH, Shintaku IP, Taylor EJ, Robert L, et al. PD-1 blockade induces responses by inhibiting adaptive immune resistance. Nature (2014) 515:568–71. doi: 10.1038/nature13954 33. Rivera LB, Bergers G. Intertwined regulation of angiogenesis and immunity by myeloid cells. Trends Immunol. (2015) 36:240–9. doi: 10.1016/j.it.2015.02.005 10. Ley K, Laudanna C, Cybulsky MI, Nourshargh S. Getting to the site of inﬂammation: the leukocyte adhesion cascade updated. Nat Rev Immunol. (2007) 7:678–89. doi: 10.1038/nri2156 34. De Bock K, Cauwenberghs S, Carmeliet P. Vessel abnormalization: another hallmark of cancer? Molecular mechanisms and therapeutic implications. Curr Opin Genet Dev. (2011) 21:73–9. doi: 10.1016/j.gde.2010.10.008 11. Gaud G, Lesourne R, Love PE. Regulatory mechanisms in T cell receptor signalling. Nat Rev Immunol. (2018) 18:485–97. doi: 10.1038/s41577-018-0020-8 35. Baluk P, Hashizume H, McDonald DM. Cellular abnormalities of blood vessels as targets in cancer. Curr Opin Genet Dev. (2005) 15:102–11. doi: 10.1016/j.gde.2004.12.005 12. Golstein P, Griﬃths GM. An early history of T cell-mediated cytotoxicity. Nat Rev Immunol. (2018) 18:527–35. doi: 10.1038/s41577-018-0009-3 36. Langenkamp E, Molema G. Microvascular endothelial cell heterogeneity: general concepts and pharmacological consequences for anti- angiogenic therapy of cancer. Cell Tissue Res. (2009) 335:205–22. doi: 10.1007/s00441-008-0642-4 13. Walunas TL, Lenschow DJ, Bakker CY, Linsley PS, Freeman GJ, Green JM, et al. CTLA-4 can function as a negative regulator of T cell activation. Immunity (1994) 1:405–13. doi: 10.1016/1074-7613(94)90071-X 37. Dieterich LC, Mellberg S, Langenkamp E, Zhang L, Zieba A, Salomäki H, et al. Transcriptional proﬁling of human glioblastoma vessels indicates a key role of VEGF-A and TGFβ2 in vascular abnormalization. J Pathol. (2012) 228:378–90. doi: 10.1002/path.4072 14. Linsley PS, Greene JL, Brady W, Bajorath J, Ledbetter JA, Peach R. Human B7–1 (CD80) and B7–2 (CD86) bind with similar avidities but distinct kinetics to CD28 and CTLA-4 receptors. Immunity (1994) 1:793–801. doi: 10.1016/S1074-7613(94)80021-9 15. van der Merwe PA, Bodian DL, Daenke S, Linsley P, Davis SJ. CD80 (B7–1) binds both CD28 and CTLA-4 with a low aﬃnity and very fast kinetics. J Exp Med. (1997) 185:393–403. doi: 10.1084/jem.185.3.393 38. Masiero M, Simões FC, Han HD, Snell C, Peterkin T, Bridges E, et al. A core human primary tumor angiogenesis signature identiﬁes the endothelial orphan receptor ELTD1 as a key regulator of angiogenesis. Cancer Cell (2013) 24:229–41. doi: 10.1016/j.ccr.2013.06.004 16. Wei SC, Duﬀy CR, Allison JP. Fundamental mechanisms of immune checkpoint blockade therapy. Cancer Discov. (2018) 8:1069–86. doi: 10.1158/2159-8290.CD-18-0367 39. Aird WC. Endothelial cell heterogeneity. REFERENCES Tumor angiogenesis is accompanied by a decreased inﬂammatory response of tumor-associated endothelium. Blood (1996) 88:667–73. j 70. Manegold C. Bevacizumab for the treatment of advanced non- small-cell lung cancer. Expert Rev Anticancer Ther. (2008) 8:689–99. doi: 10.1586/14737140.8.5.689 53. Huang H, Langenkamp E, Georganaki M, Loskog A, Fuchs PF, Dieterich LC, et al. VEGF suppresses T-lymphocyte inﬁltration in the tumor microenvironment through inhibition of NF-kappaB-induced endothelial activation. FASEB J. (2015) 29:227–38. doi: 10.1096/fj.14-2 50985 71. Arkenau HT, Brunetto AT, Barriuso J, Olmos D, Eaton D, de Bono J, et al. Clinical beneﬁt of new targeted agents in phase I trials in patients with advanced colorectal cancer. Oncology (2009) 76:151–6. doi: 10.1159/000195884 54. Cindrova-Davies T, Sanders DA, Burton GJ, Charnock-Jones DS. Soluble FLT1 sensitizes endothelial cells to inﬂammatory cytokines by antagonizing VEGF receptor-mediated signalling. Cardiovasc Res. (2011) 89:671–9. doi: 10.1093/cvr/cvq346 72. Calmes-Miller J. FDA approves ﬁrst biosimilar to treat cancer. Cancer Discov. (2017) 7:1206. doi: 10.1158/2159-8290.CD-NB2017-183 73. Vasudev NS, Reynolds AR. Anti-angiogenic therapy for cancer: current progress, unresolved questions and future directions. Angiogenesis (2014) 17:471–94. doi: 10.1007/s10456-014-9420-y q 55. Sainson RC, Johnston DA, Chu HC, Holderﬁeld MT, Nakatsu MN, Crampton SP, et al. TNF primes endothelial cells for angiogenic sprouting by inducing a tip cell phenotype. Blood (2008) 111:4997–5007. doi: 10.1182/blood-2007-08-108597 74. Ruegg C, Mutter N. Anti-angiogenic therapies in cancer: achievements and open questions. Bull Cancer. (2007) 94:753–62. doi: 10.1684/bdc.2007.0451 75. Paez-Ribes M, Allen E, Hudock J, Takeda T, Okuyama H, Vinals F, et al. Antiangiogenic therapy elicits malignant progression of tumors to increased local invasion and distant metastasis. Cancer Cell (2009) 15:220– 31. doi: 10.1016/j.ccr.2009.01.027 56. Melder RJ, Koenig GC, Witwer BP, Safabakhsh N, Munn LL, Jain RK. During angiogenesis, vascular endothelial growth factor and basic ﬁbroblast growth factor regulate natural killer cell adhesion to tumor endothelium. Nat Med. (1996) 2:992–7. doi: 10.1038/nm0996-992 76. Ebos JM, Lee CR, Cruz-Munoz W, Bjarnason GA, Christensen JG, Kerbel RS. Accelerated metastasis after short-term treatment with a potent inhibitor of tumor angiogenesis. Cancer Cell (2009) 15:232–9. doi: 10.1016/j.ccr.2009.01.021 57. Goebel S, Huang M, Davis WC, Jennings M, Siahaan TJ, Alexander JS, et al. VEGF-A stimulation of leukocyte adhesion to colonic microvascular endothelium: implications for inﬂammatory bowel disease. Am J Physiol Gastrointest Liver Physiol. (2006) 290:G648–54. doi: 10.1152/ajpgi.00466.2005 j 77. De Bock K, Mazzone M, Carmeliet P. Antiangiogenic therapy, hypoxia, and metastasis: risky liaisons, or not? Nat Rev Clin Oncol. (2011) 8:393–404. doi: 10.1038/nrclinonc.2011.83 58. REFERENCES Kim I, Moon SO, Kim SH, Kim HJ, Koh YS, Koh GY. Vascular endothelial growth factor expression of intercellular adhesion molecule 1 (ICAM-1), vascular cell adhesion molecule 1 (VCAM-1), and E-selectin through nuclear factor-kappa B activation in endothelial cells. J Biol Chem. (2001) 276:7614– 20. doi: 10.1074/jbc.M009705200 78. Jain RK. Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy. Science (2005) 307:58–62. doi: 10.1126/science.1104819 79. Shrimali RK, Yu Z, Theoret MR, Chinnasamy D, Restifo NP, Rosenberg SA. Antiangiogenic agents can increase lymphocyte inﬁltration into tumor and enhance the eﬀectiveness of adoptive immunotherapy of cancer. Cancer Res. (2010) 70:6171–80. doi: 10.1158/0008-5472.CAN-10-0153 59. Gehad AE, Lichtman MK, Schmults CD, Teague JE, Calarese AW, Jiang Y, et al. Nitric oxide-producing myeloid-derived suppressor cells inhibit vascular E-selectin expression in human squamous cell carcinomas. J Invest Dermatol. (2012) 132:2642–51. doi: 10.1038/jid.2012.190 80. Huang Y, Yuan J, Righi E, Kamoun WS, Ancukiewicz M, Nezivar J, et al. Vascular normalizing doses of antiangiogenic treatment reprogram the immunosuppressive tumor microenvironment and enhance immunotherapy. Proc Natl Acad Sci USA. (2012) 109:17561–6. doi: 10.1073/pnas.1215397109 60. Delfortrie S, Pinte S, Mattot V, Samson C, Villain G, Caetano B, et al. Egﬂ7 promotes tumor escape from immunity by repressing endothelial cell activation. Cancer Res. (2011) 71:7176–86. doi: 10.1158/0008-5472.CAN-11-1301 p 81. Meder L, Schuldt P, Thelen M, Schmitt A, Dietlein F, Klein S, et al. Combined VEGF and PD-L1 blockade displays synergistic treatment eﬀects in an autochthonous mouse model of small cell lung cancer. Cancer Res. (2018) 78:4270–81. doi: 10.1158/0008-5472.CAN-17-2176 61. Mulligan JK, Lathers DM, Young MR. Tumors skew endothelial cells to disrupt NK cell, T-cell and macrophage functions. Cancer Immunol Immunother. (2008) 57:951–61. doi: 10.1007/s00262-007-0425-x 62. Motz GT, Santoro SP, Wang LP, Garrabrant T, Lastra RR, Hagemann IS, et al. Tumor endothelium FasL establishes a selective immune barrier promoting tolerance in tumors. Nat Med. (2014) 20:607–15. doi: 10.1038/nm.3541 82. Yasuda S, Sho M, Yamato I, Yoshiji H, Wakatsuki K, Nishiwada S, et al. Simultaneous blockade of programmed death 1 and vascular endothelial growth factor receptor 2 (VEGFR2) induces synergistic anti-tumour eﬀect in vivo. Clin Exp Immunol. (2013) 172:500–6. doi: 10.1111/cei.12069 63. Mai J, Virtue A, Shen J, Wang H, Yang X-F. An evolving new paradigm: endothelial cells–conditional innate immune cells. J Hematol Oncol. (2013) 6:61. doi: 10.1186/1756-8722-6-61 83. Schmittnaegel M, Rigamonti N, Kadioglu E, Cassara A, Wyser Rmili C, Kiialainen A, et al. REFERENCES Endothelin B receptor mediates the endothelial barrier to T cell homing to tumors and disables immune therapy. Nat Med. (2008) 14:28–36. doi: 10.1038/nm1699 24. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell (2000) 100:57–70. doi: 10.1016/S0092-8674(00)81683-9 25. Bergers G, Benjamin LE. Tumorigenesis and the angiogenic switch. Nat Rev Cancer. (2003) 3:401–10. doi: 10.1038/nrc1093 47. Clark RA, Huang SJ, Murphy GF, Mollet IG, Hijnen D, Muthukuru M, et al. Human squamous cell carcinomas evade the immune response by down- regulation of vascular E-selectin and recruitment of regulatory T cells. J Exp Med. (2008) 205:2221–34. doi: 10.1084/jem.20071190 26. Ferrara N, Gerber H-P, LeCouter J. The biology of VEGF and its receptors. Nat Med. (2003) 9:669–76. doi: 10.1038/nm0603-669 27. Yadav L, Puri N, Rastogi V, Satpute P, Sharma V. Tumour angiogenesis and angiogenic inhibitors: a review. J Clin Diagn Res. (2015) 9:XE01-XE5. doi: 10.7860/JCDR/2015/12016.6135 48. Griﬃoen AW, Damen CA, Martinotti S, Blijham GH, Groenewegen G. Endothelial intercellular adhesion molecule-1 expression is suppressed in human malignancies: the role of angiogenic factors. Cancer Res. (1996) 56:1111–17. 28. Fang S, Salven P. Stem cells in tumor angiogenesis. J Mol Cell Cardiol. (2011) 50:290–5. doi: 10.1016/j.yjmcc.2010. 10.024 49. Piali L, Fichtel A, Terpe HJ, Imhof BA, Gisler RH. Endothelial vascular cell adhesion molecule 1 expression is suppressed by melanoma and carcinoma. J Exp Med. (1995) 181:811–6. doi: 10.1084/jem.181.2.811 29. Liao D, Johnson RS. Hypoxia: a key regulator of angiogenesis in cancer. Cancer Metastasis Rev. (2007) 26:281–90. doi: 10.1007/s10555-007-9066-y December 2018 \| Volume 9 \| Article 3081 7 Frontiers in Immunology \| www.frontiersin.org Vascular Targeting to Improve Immunotherapy Georganaki et al. MDSCs and synergistically improving endothelial activation and T-cell recruitment. Oncotarget (2016) 7:50277–89. doi: 10.18632/oncotarget.10364 MDSCs and synergistically improving endothelial activation and T-cell recruitment. Oncotarget (2016) 7:50277–89. doi: 10.18632/oncotarget.10364 50. Madhavan M, Srinivas P, Abraham E, Ahmed I, Vijayalekshmi NR, Balaram P. Down regulation of endothelial adhesion molecules in node positive breast cancer: possible failure of host defence mechanism. Pathol Oncol Res. (2002) 8:125–8. doi: 10.1007/BF03033721 68. Folkman J. Tumor angiogenesis: therapeutic implications. New Eng J Med. (1971) 285:1182–6. doi: 10.1056/NEJM197111182852108 51. Pober JS, Sessa WC. Evolving functions of endothelial cells in inﬂammation. Nat Rev Immunol. (2007) 7:803–15. doi: 10.1038/nri2171 69. Jain RK. Antiangiogenesis strategies revisited: from starving tumors to alleviating hypoxia. Cancer Cell. (2014) 26:605–22. doi: 10.1016/j.ccell.2014.10.006 52. Griﬃoen AW, Damen CA, Blijham GH, Groenewegen G. REFERENCES Dual angiopoietin-2 and VEGFA inhibition elicits antitumor immunity that is enhanced by PD-1 checkpoint blockade. Sci Transl Med. (2017) 9:eaak9670. doi: 10.1126/scitranslmed.aak9670 64. Griﬃoen AW, Damen CA, Mayo KH, Barendsz-Janson AF, Martinotti S, Blijham GH, et al. Angiogenesis inhibitors overcome tumor induced endothelial cell anergy. Int J Cancer (1999) 80:315–9. 65. Dirkx AEM, oude Egbrink MGA, Castermans K, van der Schaft DWJ, Thijssen VLJL, Dings RPM, et al. Anti-angiogenesis therapy can overcome endothelial cell anergy and promote leukocyte-endothelium interactions and inﬁltration in tumors. FASEB J. (2006) 20:621–30. doi: 10.1096/fj.05-4493com 84. Du Four S, Maenhout SK, Niclou SP, Thielemans K, Neyns B, Aerts JL. Combined VEGFR and CTLA-4 blockade increases the antigen-presenting function of intratumoral DCs and reduces the suppressive capacity of intratumoral MDSCs. Am J Cancer Res. (2016) 6:2514–31. 85. Voron T, Colussi O, Marcheteau E, Pernot S, Nizard M, Pointet AL, et al. VEGF-A modulates expression of inhibitory checkpoints on CD8+ T cells in tumors. J Exp Med. (2015) 212:139–48. doi: 10.1084/jem.201 40559 66. Dings RPM, Vang KB, Castermans K, Popescu F, Zhang Y, Oude Egbrink MGA, et al. Enhancement of T-cell-mediated antitumor response: angiostatic adjuvant to immunotherapy against cancer. Clin Cancer Res. (2011) 17:3134–45. doi: 10.1158/1078-0432.CCR-1 0-2443 86. Allen E, Jabouille A, Rivera LB, Lodewijckx I, Missiaen R, Steri V, et al. Combined antiangiogenic and anti-PD-L1 therapy stimulates tumor immunity through HEV formation. Sci Transl Med. (2017) 9:aak9679. doi: 10.1126/scitranslmed.aak9679 67. van Hooren L, Georganaki M, Huang H, Mangsbo SM, Dimberg A. Sunitinib enhances the antitumor responses of agonistic CD40-antibody by reducing December 2018 \| Volume 9 \| Article 3081 Frontiers in Immunology \| www.frontiersin.org 8 Vascular Targeting to Improve Immunotherapy Georganaki et al. 87. Wu FT, Man S, Xu P, Chow A, Paez-Ribes M, Lee CR, et al. Eﬃcacy of Cotargeting Angiopoietin-2 and the VEGF pathway in the adjuvant postsurgical setting for early breast, colorectal, and renal cancers. Cancer Res. (2016) 76:6988–7000. doi: 10.1158/0008-5472.CAN-16-0888 96. Finke JH, Rini B, Ireland J, Rayman P, Richmond A, Golshayan A, et al. Sunitinib reverses type-1 immune suppression and decreases T-regulatory cells in renal cell carcinoma patients. Clin Cancer Res. (2008) 14:6674–82. doi: 10.1158/1078-0432.CCR-07-5212 97. Ko JS, Zea AH, Rini BI, Ireland JL, Elson P, Cohen P, et al. Sunitinib mediates reversal of myeloid-derived suppressor cell accumulation in renal cell carcinoma patients. Clin Cancer Res. (2009) 15:2148–57. doi: 10.1158/1078-0432.CCR-08-1332 88. Rexer H, Bedke J. REFERENCES [First-line therapy in advanced renal cell carcinoma : A randomized, open-label phase III study evaluating the eﬃcacy and safety of pembrolizumab (MK-3475) in combination with axitinib compared to sunitinib monotherapy as ﬁrst-line treatment for locally advanced or metastatic renal cell carcinoma (mRCC) (Keynote-426) - AN 39/16 of the AUO]. Urologe A. (2017) 56:385–6. doi: 10.1007/s00120-017-0335-2 98. Ott PA, Hodi FS, Buchbinder EI. Inhibition of immune checkpoints and vascular endothelial growth factor as combination therapy for metastatic melanoma: an overview of rationale, preclinical evidence, and initial clinical data. Front Oncol. (2015) 5:202. doi: 10.3389/fonc.2015.00202 89. Hodi FS, Lawrence D, Lezcano C, Wu X, Zhou J, Sasada T, et al. Bevacizumab plus ipilimumab in patients with metastatic melanoma. Cancer Immunol Res. (2014) 2:632–42. doi: 10.1158/2326-6066.CIR-14-0053 data. Front Oncol. (2015) 5:202. doi: 10.3389/fonc.2015.00202 99. Comunanza V, Bussolino F. Therapy for cancer: strategy of combining anti-angiogenic and target therapies. Front Cell Dev Biol. (2017) 5:101. doi: 10.3389/fcell.2017.00101 90. Wu X, Giobbie-Hurder A, Liao X, Lawrence D, McDermott D, Zhou J, et al. VEGF neutralization plus CTLA-4 blockade alters soluble and cellular factors associated with enhancing lymphocyte inﬁltration and humoral recognition in melanoma. Cancer Immunol Res. (2016) 4:858–68. doi: 10.1158/2326-6066.CIR-16-0084 100. Ager A. High endothelial venules and other blood vessels: critical regulators of lymphoid organ development and function. Front Immunol. (2017) 8:45. doi: 10.3389/ﬁmmu.2017.00045 91. Wu X, Li J, Connolly EM, Liao X, Ouyang J, Giobbie-Hurder A, et al. Combined Anti-VEGF and Anti-CTLA-4 therapy elicits humoral immunity to galectin-1 which is associated with favorable clinical outcomes. Cancer Immunol Res. (2017) 5:446–54. doi: 10.1158/2326-6066.CIR-16-0385 101. Johansson-Percival A, He B, Li ZJ, Kjellen A, Russell K, Li J, et al. De novo induction of intratumoral lymphoid structures and vessel normalization enhances immunotherapy in resistant tumors. Nat Immunol. (2017) 18:1207–17. doi: 10.1038/ni.3836 102. Sautes-Fridman C, Lawand M, Giraldo NA, Kaplon H, Germain C, Fridman WH, et al. Tertiary lymphoid structures in cancers: prognostic value, regulation, and manipulation for therapeutic intervention. Front Immunol. (2016) 7:407. doi: 10.3389/ﬁmmu.2016.00407 92. Choueiri TK, Larkin J, Oya M, Thistlethwaite F, Martignoni M, Nathan P, et al. Preliminary results for avelumab plus axitinib as ﬁrst-line therapy in patients with advanced clear-cell renal-cell carcinoma (JAVELIN Renal 100): an open-label, dose-ﬁnding and dose-expansion, phase 1b trial. Lancet Oncol. (2018) 19:451–60. Frontiers in Immunology \| www.frontiersin.org December 2018 \| Volume 9 \| Article 3081 REFERENCES doi: 10.1016/S1470-2045(18)30107-4 Conﬂict of Interest Statement: 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. 93. Iyer PC, Dadu R, Gule-Monroe M, Busaidy NL, Ferrarotto R, Habra MA, et al. Salvage pembrolizumab added to kinase inhibitor therapy for the treatment of anaplastic thyroid carcinoma. J Immunother Cancer (2018) 6:68. doi: 10.1186/s40425-018-0378-y Copyright © 2018 Georganaki, van Hooren and Dimberg. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 94. Wallin JJ, Bendell JC, Funke R, Sznol M, Korski K, Jones S, et al. Atezolizumab in combination with bevacizumab enhances antigen-speciﬁc T-cell migration in metastatic renal cell carcinoma. Nat Commun. (2016) 7:12624. doi: 10.1038/ncomms12624 95. Fukumura D, Kloepper J, Amoozgar Z, Duda DG, Jain RK. Enhancing cancer immunotherapy using antiangiogenics: opportunities and challenges. Nat Rev Clin Oncol. (2018) 15:325–40. doi: 10.1038/nrclinonc.2018.29 December 2018 \| Volume 9 \| Article 3081 Frontiers in Immunology \| www.frontiersin.org 9
https://openalex.org/W2802582384	https://www.frontiersin.org/articles/10.3389/fnbeh.2018.00109/pdf	English	null	Unilateral Application of Cathodal tDCS Reduces Transcallosal Inhibition and Improves Visual Acuity in Amblyopic Patients	Frontiers in behavioral neuroscience	2,018	cc-by	7,483	ORIGINAL RESEARCH published: 29 May 2018 doi: 10.3389/fnbeh.2018.00109 Tommaso Bocci 1,2, Francesco Nasini 3, Matteo Caleo 4, Laura Restani 4, Davide Barloscio 1, Gianluca Ardolino 2, Alberto Priori 2,5, Lamberto Maffei 4, Marco Nardi 3 and Ferdinando Sartucci 1,4* 1Section of Neurophysiopathology, Department of Clinical and Experimental Medicine, University of Pisa, Pisa, Italy, 2Clinical Center for Neurotechnologies, Neuromodulation, and Movement Disorders, Fondazione IRCCS Ca’Granda Ospedale Maggiore Policlinico, Milan, Italy, 3Department of Surgical, Medical, and Molecular Pathology and Critical Care, University of Pisa, Pisa, Italy, 4CNR Institute of Neuroscience, University of Pisa, Pisa, Italy, 5Department of Health Sciences University of Milan and Ospedale San Paolo, Milan, Italy Objective: Amblyopia is a neurodevelopmental disorder characterized by visual acuity and contrast sensitivity loss, refractory to pharmacological and optical treatments in adulthood. In animals, the corpus callosum (CC) contributes to suppression of visual responses of the amblyopic eye. To investigate the role of interhemispheric pathways in amblyopic patients, we studied the response of the visual cortex to transcranial Direct Current Stimulation (tDCS) applied over the primary visual area (V1) contralateral to the “lazy eye.” Keywords: amblyopia, tDCS, amblyopia treatment in adults, corpus callosum, ocular dominance, visual system plasticity Edited by: Nicholas J. Kelley, Northwestern University, United States Methods: Visual acuity (logMAR) was assessed before (T0), immediately after (T1) and 60’ following the application of cathodal tDCS (2.0 mA, 20’) in 12 amblyopic patients. At each time point, Visual Evoked Potentials (VEPs) triggered by grating stimuli of different contrasts (K90%, K20%) were recorded in both hemispheres and compared to those obtained in healthy volunteers. Reviewed by: Adriana Salatino, Università degli Studi di Torino, Italy Winston D. Byblow, University of Auckland, New Zealand University of Auckland, New Zealand Results: Cathodal tDCS improved visual acuity respect to baseline (p < 0.0001), whereas sham polarization had no signiﬁcant effect. At T1, tDCS induced an inhibitory effect on VEPs amplitudes at all contrasts in the targeted side and a facilitation of responses in the hemisphere ipsilateral to the amblyopic eye; compared with controls, the facilitation persisted at T2 for high contrasts (K90%; Holm–Sidak post hoc method, p < 0.001), while the stimulated hemisphere recovered more quickly from inhibition (Holm–Sidak post hoc method, p < 0.001). Correspondence: Ferdinando Sartucci ferdinando.sartucci@med.unipi.it Received: 28 February 2018 Accepted: 09 May 2018 Published: 29 May 2018 Unilateral Application of Cathodal tDCS Reduces Transcallosal Inhibition and Improves Visual Acuity in Amblyopic Patients Tommaso Bocci 1,2, Francesco Nasini 3, Matteo Caleo 4, Laura Restani 4, Davide Barloscio 1, Gianluca Ardolino 2, Alberto Priori 2,5, Lamberto Maffei 4, Marco Nardi 3 and Ferdinando Sartucci 1,4 Tommaso Bocci 1,2, Francesco Nasini 3, Matteo Caleo 4, Laura Restani 4, Davide Barloscio 1, Gianluca Ardolino 2, Alberto Priori 2,5, Lamberto Maffei 4, Marco Nardi 3 and Ferdinando Sartucci 1,4* Participants and Experimental Protocol Participants and Experimental Protocol Twelve adult patients with unilateral amblyopia (Table 1) and 12 sex and age-matched healthy volunteers were enrolled in the study (mean age 26.1 ± 6.0 years; range 24–44, five females). Patients had an intraocular acuity difference of at least 0.2 LogMAR and were classified as strabismic, anisometropic or mixed amblyopia (both strabismus and anisometropia; see Table 1). Anisometropia was defined as a spherical equivalent difference of 1 dioptre or more between the eyes. Best refractive correction was worn during testing. Healthy volunteers did not have history of neurological or psychiatric disorders and they were all drug-free. Controls with normal vision had 0.1 LogMAR acuity or better in each eye and no history of visual disorders. Permanent monocular visual impairment is a risk for blindness, if the dominant eye is injured or becomes affected later in life (Williams et al., 2003). For this reason, the early treatment is critical. Eye-patching has been used for centuries, whereas the use of atropine has only recently emerged (Repka et al., 2005). In the past few years, new approaches are being developed, such as dichoptic visual training aimed at stimulating the amblyopic eye, reducing the interocular suppression by balancing stimulus contrast between visual hemifields (Stewart et al., 2007; Vedamurthy et al., 2015b; Žiak et al., 2017). Nonetheless, all these treatments appear to be effective for up to 7 years of age (Holmes et al., 2011), showing transient and inconclusive results in older patients (Gao et al., 2018). Moreover, current treatments are often associated with residual monocular and binocular deficits (Pediatric Eye Disease Investigator Group et al., 2008), with a high rate of recurrence (Bhola et al., 2006). In patients, visual acuity was assessed at baseline (T0), immediately after (T1) and 60’ following the completion of tDCS applied over the primary visual area (V1) contralateral to the amblyopic eye. At same time points, VEPs were recorded both in amblyopic patients and controls, at two different luminance contrasts (K90% and K20%). Patients were enrolled by a clinician (FN), whereas electrophysiological recordings were performed by a different neurologist (DB), both blinded to the tDCS condition. INTRODUCTION reasoned that changes in transcallosal inhibition may explain the unbalanced mechanisms of contrast gain control and ocular dominance in amblyopia. To this aim, we enrolled 12 patients and compared changes in visual acuity and Visual Evoked Potential (VEP) amplitudes induced by inhibitory cathodal transcranial Direct Current Stimulation (tDCS) applied to the occipital lobe contralateral to the amblyopic eye. Amblyopia is a neurodevelopmental disorder clinically characterized by visual acuity and contrast sensitivity loss, refractory to pharmacological and mechanical treatments in adulthood (Holmes and Clarke, 2006): given the lack of any organic cause, it has been also defined as a disorder ‘‘in which the patient sees nothing and the doctor sees nothing’’ (Holmes and Clarke, 2006). Amblyopia results in an abnormal binocular experience due to a mismatch between the images perceived with each eye. Although the retina is generally spared, microscopic anatomical and structural abnormalities in lateral geniculate bodies and visual cortex can occur (von Noorden and Crawford, 1992; Davis et al., 2003); fMRI studies are consistent with the hypothesis of a selective involvement of the parvocellular stream at a precortical or early cortical site, thus leading to detection and processing deficit for high-contrast stimuli (Li et al., 2007; Hess et al., 2010). Citation: Bocci T, Nasini F, Caleo M, Restani L, Barloscio D, Ardolino G, Priori A, Maffei L, Nardi M and Sartucci F (2018) Unilateral Application of Cathodal tDCS Reduces Transcallosal Inhibition and Improves Visual Acuity in Amblyopic Patients. Front. Behav. Neurosci. 12:109. doi: 10.3389/fnbeh.2018.00109 Conclusions: tDCS is a promising treatment for amblyopia in adults. The rapid recovery of excitability and the concurrent transcallosal disinhibition following perturbation of cortical activity may support a critical role of interhemispheric balance in the pathophysiology of amblyopia. May 2018 \| Volume 12 \| Article 109 Frontiers in Behavioral Neuroscience \| www.frontiersin.org 1 Cathodal tDCS in the Treatment of Amblyopia Bocci et al. Participants and Experimental Protocol In animal models, the corpus callosum (CC) plays a critical role in the suppression of deprived eye responses after a period of monocular occlusion (Restani et al., 2009; Cerri et al., 2010); in humans, callosal connections appear to inhibit the responsiveness of the neurons located in the opposite hemisphere (Bocci et al., 2014). Moreover, reduced visual cortex excitability has been observed in patients with amblyopia, possibly reflecting abnormally high levels of cortico-cortical inhibition (Thompson et al., 2008; Hess and Thompson, 2015). Thus, the reduced responses of the amblyopic eye may be due to active inhibition (suppression) within the primary visual cortex. Here we tested the contribution of interhemispheric pathways to such inhibition. Written informed consent was signed by all subjects prior to participation in the study, approved by the local ethical Committee in accordance with the tenets of Helsinki. The study was approved by the local ethical committee (registration number 3135), at the University of Pisa (formally named ‘‘Comitato Etico di Area Vasta Nord Ovest della Toscana’’). Frontiers in Behavioral Neuroscience \| www.frontiersin.org Transcranial Direct Current Stimulation (tDCS) Each patient underwent both cathodal and sham tDCS, elapsed by at least 1 week. Visual acuity, expressed as LogMAR, refers to the ﬁrst clinical evaluation, immediately after the enrollment, but it was checked again before the second treatment (either sham or cathodal; in this case, checked values of the “lazy eye” are expressed within brackets). LE, left eye; RE, right eye; Aniso, anisomoetropic amblyopia; Strab, strabismic amblyopia. LogMAR was calculated according to the formula: LogMAR = −log (decimal acuity) Each patient underwent both cathodal and sham tDCS, elapsed by at least 1 week. Visual acuity, expressed as LogMAR, refers to the ﬁrst clinical evaluation, immediately after the enrollment, but it was checked again before the second treatment (either sham or cathodal; in this case, checked values of the “lazy eye” are expressed within brackets). LE, left eye; RE, right eye; Aniso, anisomoetropic amblyopia; Strab, strabismic amblyopia. LogMAR was calculated according to the formula: LogMAR = −log (decimal acuity). (SofTaxic optically-tracked by EMS, Italy). Tridimensional space positions of the head and electrode were reproduced on the computer screen in relation to an average brain anatomy based on a 3D realistic MR-constructed brain model: in accordance with previous articles, the lower horizontal border of the electrode was marked by a scalp point superficial to the tentorium cerebelli, while the medial vertical one corresponded to a scalp point superficial to the brain location 1 cm lateral to the interhemispheric falx cerebri (Olma et al., 2013; Behrens et al., 2017). The BCVA was tested for both eyes by means of standard ‘‘Early Treatment Diabetic Retinopathy Study (ETDRS) Revised’’ translucent visual acuity charts (with the following features: same number of letters per line, equal spacing between lines on a log scale, equal spacing of letters on a log scale and balanced letter difficulty in the individual lines). Retro-illuminated ETDRS viewing cabinet was used. Both eyes were separately tested at a distance of 4 m (about 13 feet). Chart 1 was used for visual acuity testing of the right eye and chart 2 for testing the left eye. The patient was asked to read slowly, beginning from the top line of the chart, from left to right. The patient was told that one chance is given to read each letter. If the patient changed a response (e.g., ‘that was a ‘‘C’’ not an ‘‘O’’’) before he/she has read the next letter, then the change was accepted. Transcranial Direct Current Stimulation (tDCS) Direct currents were applied for 20 min with an intensity of 1.5 mA (current density 0.06 mA/cm2). The intensity and duration of stimulation were comparable to those used in previous studies (Antal et al., 2004, 2006; Lang et al., 2007), below the threshold for tissue damage (Nitsche et al., 2003). tDCS strength remained below the sensory threshold throughout the experimental session. At the offset of tDCS, the current was decreased in a ramp-like manner, a method shown to achieve a good level of blinding among sessions (Gandiga et al., 2006; Galea et al., 2009). In the sham condition, the current was turned on for 5 s and then turned off in a ramp-shaped fashion. • If a patient was able to read at least 20 letters on the chart, the visual acuity score of the tested eye was recorded as the number of letters read correctly at 4 m (sum = A) plus 30 (credit of 30 score points = B). • If a patient could not read at least 20 letters on the chart at 4 m, the test was repeated at a distance of 1 m. In this case, the visual acuity score for the tested eye was recorded as the number of letters read correctly at 1 m (sum = C) plus the number of letters read correctly at 4 m (sum = A). Transcranial Direct Current Stimulation (tDCS) ( ) We applied tDCS over the V1, using a battery-driven constant current stimulator (HDCStim, Newronika, Italy) and a pair of electrodes in two saline-soaked synthetic sponges with a surface area of 25 cm2 (5 × 5 cm). Amblyopic patients underwent both cathodal (real) and sham stimulation, while in healthy controls only the cathodal polarization was applied. For cathodal stimulation, the cathode was centered either on O1 or O2 (according to the 10–20 international EEG system) and the anode on the right shoulder. It has been recently proved that inter-hemispheric connections regulate cortical gain by dampening neural responses to high-contrast stimuli in the target hemisphere (Bocci et al., 2011). Concurrently, we have suggested that the rapid recovery of excitability and the persistent transcallosal disinhibition following perturbation of cortical activity may exert a key role in the pathophysiology of photosensitive epilepsy (Bocci et al., 2016). Altogether, we Anatomical correspondence between the target region and V1 was confirmed by a navigated stimulation system Frontiers in Behavioral Neuroscience \| www.frontiersin.org May 2018 \| Volume 12 \| Article 109 2 Cathodal tDCS in the Treatment of Amblyopia Bocci et al. TABLE 1 \| Demographic and clinical assessment. Patient’s number Previous treatment Type of Amblyopia Visual acuity (logMAR) Timeline of intervention 1 None LE 0.0 cathodal/sham RE Aniso 0.55 (0.55) 2 None LE 0.0 cathodal/sham RE Aniso 1.03 (1.00) 3 None RE 0.0 sham/cathodal LE Aniso 0.40 (0.38) 4 None RE 0.0 cathodal/sham LE Strab 1.03 (0.94) 5 None RE 0.0 sham/cathodal LE Aniso 0.22 (0.22) 6 None LE 0.0 cathodal/sham RE Strab 0.55 (0.58) 7 Patching LE 0.0 sham/cathodal RE Aniso 0.38 (0.42) 8 None RE 0.0 sham/cathodal LE Strab 0.40 (0.40) 9 None LE 0.0 sham/cathodal RE Aniso 1.00 (1.05) 10 None LE 0.0 cathodal/sham RE Mixed 0.83 (0.75) 11 None LE 0.0 cathodal/sham RE Aniso 0.92 (0.84) 12 Patching RE 0.0 sham/cathodal LE Strab 0.26 (0.26) Each patient underwent both cathodal and sham tDCS, elapsed by at least 1 week. Visual acuity, expressed as LogMAR, refers to the ﬁrst clinical evaluation, immediately after the enrollment, but it was checked again before the second treatment (either sham or cathodal; in this case, checked values of the “lazy eye” are expressed within brackets). LE, left eye; RE, right eye; Aniso, anisomoetropic amblyopia; Strab, strabismic amblyopia. LogMAR was calculated according to the formula: LogMAR = −log (decimal acuity). TABLE 1 \| Demographic and clinical assessment. Electrophysiological Measures (VEPs) A three-way repeated measures ANOVA assessed the effects of ‘‘time’’ × ‘‘stimulation’’ × ‘‘contrast’’ interaction in amblyopic patients. At each time interval, a two-way RM ANOVA compared peak-to-peak amplitudes between cathodal and sham polarization at different contrasts; significant effects were checked by post hoc Holm-Sidak test. VEP amplitudes were defined as the difference between the N1 negative peak and the P1 positive peak amplitudes in microvolts (Ding et al., 2016). The N1 was defined as a negative peak 60–110 ms after the pattern reversal and the peak of the first positive wave after N1 was named as P1. Comparison Between Amblyopic Patients and Controls A three-way RM ANOVA analyzed the effects of ‘‘group’’ × ‘‘time’’ × ‘‘contrast’’ interaction between patients and healthy controls. At each time interval, a two-way ANOVA on ranks compared peak-to-peak amplitudes between amblyopic participants and healthy controls; significant effects were followed by post hoc Holm-Sidak test to compare VEP changes over time. VEPs were recorded before (T0), at the end (T1) and 45’ (T2) after tDCS. Grating stimuli were centered on the fixation point and tDCS was applied to V1. We analyzed 18 blocks of 100 averaged VEP responses (6 blocks at T0, 6 at T1 and 6 at T2), in terms of both mean amplitude (expressed as µV) and latency (ms) for two contrast levels (K90% and 20%). Visual stimuli at different contrasts were presented randomly and the obtained electrophysiological responses for each contrast were then averaged. Statistical significance was set at p < 0.05. Data were analyzed using SPSS v. 21.0 for Windows (SPSS Inc., Chicago, IL, USA) or SigmaPlot v. 12.0. RESULTS Parametric analyses were used, as all data sets successfully passed the Shapiro-Wilk test for normality (p > 0.05). A one-way analyses of variance (ANOVA) was used to compare baseline values for each subject between sham and cathodal condition. As VEP amplitudes are higher in healthy subjects and in the fellow eyes compared with the amblyopic ones, all values were normalized at baseline (T0, i.e., before tDCS: (T1/T0) × 100%). Clinical Assessment: Visual Acuity Baseline (T0) logMAR values did not change between real and sham sessions (p = 0.83). A remarkable improvement occurred at T1 when cathodal polarization was delivered within the hemisphere contralateral to the amblyopic eye, with changes lasting up to 1 h after tDCS completion (F(2,22) = 8.14, p = 0.0023, two-way ANOVA, with ‘‘time’’ and ‘‘treatment’’ as factors). This reduction ranged from 0.11 to 0.88 logMAR, with a mean of about 0.27 logMAR (see Table 2 and Figure 1), and it was significant both at T1 (p = 0.0029, Holm-Sidak post hoc comparison) and T2 (p = 0.0019) when compared to the sham group. Visual Evoked Potentials (VEPs) A detailed description of the protocol has been reported elsewhere (Bocci et al., 2011, 2016). VEPs were recorded in response to abrupt reversal (3 Hz) of a horizontal square wave grating (spatial frequency 2 c/deg), generated by computer on a display (Sony; refresh rate 60 Hz; subtending 20 × 15◦of visual angle) by a VSG card (Cambridge Research Systems). The display was centered on the vertical meridian. VEPs were recorded simultaneously in both hemispheres, with Ag/AgCl electrodes positioned 2 cm above the inion (active) and at the right mastoid (reference). Visual Acuity Assessment Each patient underwent a complete ophthalmologic examination to exclude other causes of poor vision, thus confirming that the patient’s refractive correction (where applicable) was accurate in order to perform the The Best-Corrected Visual Acuity (BCVA) testing. For each eye, the visual acuity score was the sum of A, B and C. If no letters were read correctly at either 4 m or 1 m, the visual acuity score was recorded as ‘‘0’’. All procedures were done by an expert and certified ophthalmologist (FN). May 2018 \| Volume 12 \| Article 109 Frontiers in Behavioral Neuroscience \| www.frontiersin.org 3 Cathodal tDCS in the Treatment of Amblyopia Bocci et al. Holm-Sidak post hoc method. The Pearson’s correlation was used to compare the average changes in visual acuity respect to the baseline values. mpared to baseline values are shown for the amblyopic eyes; notably, cathodal tDCS improved visual function both at T1 and T2. Changes in logMAR score compared to baseline values are shown for the amblyopic eyes; notably, cathodal tDCS improved visua Visual Acuity In each patient, changes in visual acuity (logMAR) were assessed by using a two-way repeated-measures (RM) ANOVA, with ‘‘stimulation’’ (two levels: cathodal and sham) and ‘‘time’’ (three levels: T0, T1 and T2) as experimental factors, followed by TABLE 2 \| Visual acuity following transcranial direct current stimulation (tDCS). Cathodal tDCS Sham tDCS Patient’s number T1 T2 T1 T2 1 −0.11 −0.33 −0.05 −0.11 2 −0.33 −0.47 0.03 −0.09 3 −0.14 −0.12 −0.02 −0.10 4 −0.61 −0.27 −0.19 −0.10 5 −0.07 −0.07 0.05 0.08 6 −0.22 −0.27 0.17 0.08 7 −0.22 −0.16 0.04 0.14 8 −0.25 −0.20 0.02 0.00 9 −0.35 −0.37 0.05 −0.06 10 −0.41 −0.08 −0.05 −0.07 11 −0.10 −0.88 0.07 0.08 12 −0.02 −0.08 0.02 −0.04 Changes in logMAR score compared to baseline values are shown for the amblyopic eyes; notably, cathodal tDCS improved visual function both at T1 and T2. Frontiers in Behavioral Neuroscience \| www.frontiersin.org May 2018 \| Volume 12 \| Article 109 4 Cathodal tDCS in the Treatment of Amblyopia Bocci et al. FIGURE 1 \| Changes in visual acuity (LogMAR). Amblyopic patients showed a signiﬁcant improvement of visual acuity following cathodal transcranial direct current stimulation (tDCS) compared to sham polarization, with effects lasting for up to 1 h. Data are given as mean values ± standard error (SE); ∗∗p < 0.001. compared to low-contrast (F(2,66) = 14.9, p < 0.0001, two-way ANOVA on ranks) and sham stimulation (F(2,66) = 35.9, p < 0.0001, two-way RM ANOVA), remaining persistently elevated at T2 (p < 0.0001, Holm-Sidak test). A significant correlation between the enhancement of visual acuity and the relative increase of VEP amplitudes in the amblyopic side was found (Pearson’s correlation: p = 0.002). On the opposite side, as expected due to the inhibitory effect of cathodal polarization, we observed a reduction of VEP amplitudes at T1, both at high and low contrasts. At T2 all values returned to baseline, both for high (Holm-Sidak test, T1 vs. T2: p < 0.0001) and low contrasts (p = 0.001). FIGURE 1 \| Changes in visual acuity (LogMAR). Amblyopic patients showed a signiﬁcant improvement of visual acuity following cathodal transcranial direct current stimulation (tDCS) compared to sham polarization, with effects lasting for up to 1 h. Data are given as mean values ± standard error (SE); ∗∗p < 0.001. Visual Evoked Potentials (VEPs): Comparison Between Patients and Controls When VEPs recorded from the side contralateral to tDCS were analyzed, a three-way ANOVA showed significant effects of time (F(2,132) = 20.4, p < 0.0001), contrast (F(1,132) = 64.9, p < 0.0001), time × contrast (F(2,132) = 30.7, p < 0.0001), group × contrast (F(1,132) = 11.4, p = 0.001) and contrast × group × time interaction (F(2,132) = 8.0, p = 0.0005). When analyzed separately, at high contrasts (K90%), we found a persistent enhancement of VEP amplitude in amblyopic subjects but not controls at T2 (Holm-Sidak test, p < 0.0001): thus, transcallosal disinhibition persisted in amblyopic patients, while it vanished in controls (compare Figures 3, 4, contralateral side). Changes in logMAR score linearly correlated with baseline values. Indeed, patients with greater impaired at baseline showed a more robust improvement in visual acuity (p = 0.0004, Pearson’s correlation; Figure 2). Visual Evoked Potentials (VEPs): Amblyopic Patients At baseline, in agreement with previous data reported elsewhere (Ding et al., 2016), mean VEP amplitudes for amblyopic eyes were significantly lower than those recorded by stimulating the fellow eyes (6.54 ± 0.91 vs. 9.67 ± 1.99 µV at K90%: t = 4.61, p < 0.0001, unpaired t-test). On the hemisphere in which inhibitory cathodal polarization was applied, patients and controls showed a similar reduction in VEP amplitudes at T1; a three-way ANOVA showed significant effects of time (F(2,132) = 104.3, p < 0.0001), contrast (F(2,132) = 3.9, p = 0.049) and group × time interaction (F(2,132) = 49.8, p < 0.0001). At high contrasts, values returned to baseline in patients (Holm-Sidak test, T1 vs. T2 in patients, p < 0.0001), remaining significantly reduced in controls (Holm- Sidak test, T1 vs. T2 in controls, p = 0.42; compare Figures 3, 4). p Representative VEPs from one patient are shown in Figure 3A. A three-way repeated measures ANOVA revealed significant effects of stimulation (F(1,132) = 24.2, p < 0.0001), contrast (F(1,132) = 96.8, p < 0.0001), stimulation × contrast (F(1,132) = 23.3, p < 0.0001), time × stimulation (F(2,132) = 8.7, p = 0.0003) and time × contrast interaction (F(2,132) = 5.7, p = 0.0041). In particular, at high-contrast, VEP amplitudes recorded ipsilaterally to amblyopic eyes dramatically improved DISCUSSION FIGURE 2 \| Correlation with baseline values. Changes in logMAR score linearly correlated with baseline values: patients with greater impairment at baseline showed a more robust enhancement of the visual acuity (p = 0.0004, Pearson’s correlation). Our results suggest that cathodal tDCS applied over the V1 contralateral to the ‘‘lazy eye’’ improves visual acuity, supporting the use of non-invasive brain stimulation techniques (NIBS) for the treatment of adult patients with amblyopia. Inhibitory cathodal tDCS dampened VEP amplitudes in both healthy and amblyopic subjects; concurrently, facilitation of visual responses in the contralateral side occurred, possibly due to the removal of interhemispheric inhibitory influences (Restani et al., 2009; Bocci et al., 2011). Significant differences were found at T2, with a faster normalization of VEP amplitudes in the stimulated side and a persistent disinhibition in the opposite hemisphere in amblyopic patients. This disinhibition may be at the basis of the behavioral improvement of visual acuity, which was detected at T1 and persisted at T2 (Figure 1). This interpretation is supported by the significant correlation between VEP changes and the enhancement of clinical outcome (i.e., reduction of logMAR). FIGURE 2 \| Correlation with baseline values. Changes in logMAR score linearly correlated with baseline values: patients with greater impairment at baseline showed a more robust enhancement of the visual acuity (p = 0.0004, Pearson’s correlation). May 2018 \| Volume 12 \| Article 109 Frontiers in Behavioral Neuroscience \| www.frontiersin.org 5 Cathodal tDCS in the Treatment of Amblyopia Bocci et al. FIGURE 3 \| Changes in Visual Evoked Potential (VEP) amplitudes in amblyopic subjects. (A) Representative VEP responses to central stimulation (contrast, 90%) of the amblyopic eye, in the hemisphere contralateral (top traces) and ipsilateral (bottom traces) to tDCS intervention, respectively. (B) VEP amplitudes signiﬁcantly increased ipsilaterally to the amblyopic eye, at high contrasts (K90%), and remained persistently elevated at T2 (p < 0.0001, Holm-Sidak post hoc test). On the opposite side, we observed a reduction of VEP amplitudes at T1, but at T2 all values returned to baseline, both for high (Holm-Sidak test, T1 vs. T2: p < 0.0001) and low contrasts (p = 0.001). ∗∗∗p < 0.001. FIGURE 3 \| Changes in Visual Evoked Potential (VEP) amplitudes in amblyopic subjects. (A) Representative VEP responses to central stimulation (contrast, 90%) of the amblyopic eye, in the hemisphere contralateral (top traces) and ipsilateral (bottom traces) to tDCS intervention, respectively. ACKNOWLEDGMENTS We gratefully acknowledge the participation of all subjects and Mr. C. Orsini for his excellent technical assistance. We thank Dr. Francesca Bovis (Institute Gaslini, Genova) for her valuable statistical revision. The article was supported in part by the Italian operating and development MIUR PRIN grant year 2006, n.2006062332_002. Second, the reduction of cortical excitability mediated by cathodal tDCS in the stimulated hemisphere could potentiate weak responses from the lazy eye via homeostatic mechanisms (Turrigiano, 2012). For example, it has been previously shown that brief period of monocular deprivation in adult subjects strongly alters ocular balance, producing a DISCUSSION Frontiers in Behavioral Neuroscience \| www.frontiersin.org May 2018 \| Volume 12 \| Article 109 6 Cathodal tDCS in the Treatment of Amblyopia Bocci et al. Our hypothesis is consistent with data in animals, showing that transcallosal connections are primarily involved in the weakening of deprived eye responses during monocular deprivation (Restani et al., 2009; Pietrasanta et al., 2014). Since transcallosal neurons are excitatory, interhemispheric inhibition depends upon the activation of GABAergic neurons in the target side, which contact local cortical pyramids via GABA-B receptors (Irlbacher et al., 2007; Palmer et al., 2012). Along this line, in primate models of amblyopia, the magnitude of side-to-side suppression seems to be closely related to the behavioral loss of contrast sensitivity in the amblyopic eye (Bi et al., 2011; Li et al., 2011, 2013; Tao et al., 2014). perceptual boost of the deprived eye (Lunghi et al., 2011). Along this line, the binocular imbalance that characterizes amblyopia can be reduced by occluding the amblyopic eye with a translucent patch for a few hours (Zhou et al., 2013). Another possibility is that tDCS affects brainstem nuclei or thalamic structures, such as the lateral geniculate nucleus. In this case, the effects of the manipulation on acuity and VEP responses could be at least partly due to an action at subcortical level. Although this hypothesis cannot be definitely ruled out, VEP changes following hemifield visual stimulation seem to be consistent with a selective modulation of the interhemispheric route, as described in more detail elsewhere (Bocci et al., 2011, 2016). Moreover, direct geniculocortical connections are mildly affected by monocular deprivation in animals, with effects requiring at least 20 days of ocular deprivation (Antonini et al., 1999). Further support for a key role of interhemispheric pathways in amblyopia comes from recent data showing a higher vulnerability of the parvocellular pathway to the effects of visual deprivation, thus affecting the chromatic vision in humans (Hess et al., 2010; Lunghi et al., 2013). Notably, the callosum preferentially processes high-contrast stimuli and robustly transfers chromatic information related to the activation of the parvocellular stream (Berardi and Fiorentini, 1987; Berardi et al., 1987; Corballis, 1996; Roser and Corballis, 2003). Also in our sample, the persistent facilitation of visual responses has been observed for high contrasts only (see Figure 3). Limitations and Alternative Explanations p The main limitation of our study is the small number of patients, due to the difficulty in recruiting a homogeneous group of subjects. High-powered studies, with largest samples, are needed in the future to confirm our data and assess the efficacy of unilateral cathodal tDCS as a valuable option for the long-term treatment of amblyopia. Despite the low number of cases, cathodal tDCS displayed a consistent effect on visual acuity (see Figure 2). AUTHOR CONTRIBUTIONS Although our results appear to fit an explanation based on imbalance of V1 cortical excitability between hemispheres, additional possibilities need to be considered. First, the rapid changes triggered by tDCS (in terms of both visual acuity and VEP amplitudes) strongly support alterations in the excitatory/inhibitory balance within the visual system rather than structural rearrangements of inputs from the lazy eye. In this context, there is evidence that responses of the weak eye are actively suppressed by GABAergic inhibition (Duffy et al., 1976), and tDCS may alter GABA concentrations in the cerebral cortex (Stagg et al., 2009). TB: conception and design of research, data analysis, interpretation of the results, writing article; MC, LR, AP, MN and FS: conception and design of research, interpretation of the results, writing article; FN: data collection and analysis, patients’ enrollment; GA and LM: conception and design of research, interpretation of the results; DB performed the experiments, writing the article. DISCUSSION (B) VEP amplitudes signiﬁcantly increased ipsilaterally to the amblyopic eye, at high contrasts (K90%), and remained persistently elevated at T2 (p < 0.0001, Holm-Sidak post hoc test). On the opposite side, we observed a reduction of VEP amplitudes at T1, but at T2 all values returned to baseline, both for high (Holm-Sidak test, T1 vs. T2: p < 0.0001) and low contrasts (p = 0.001). ∗∗∗p < 0.001. A growing bulk of literature suggests that the adult visual system retains a high degree of plasticity (Lunghi et al., 2010, 2011; Lo Verde et al., 2017), indicating that the excitatory/inhibitory balance that modulates gain control mechanisms could be particularly susceptible to NIBS interventions, even at short timescales (Reinhart et al., 2016). In humans, previous articles have reported a significant effect of tDCS for the recovery of contrast sensitivity and stereopsis in amblyopia, providing a novel and safe approach to improve outcome in adults (Spiegel et al., 2013a; Ding et al., 2016). Authors demonstrated an enhancement of both monocular (visual acuity) and binocular (stereopsis) measures of visual function, especially when the polarization of the visual cortex was associated with dichoptic videogame-based treatment (Spiegel et al., 2013b). Nonetheless, these studies have used excitatory, anodal tDCS bilaterally applied over the V1. Here, we reasoned to dampen the excitability of the visual area contralateral to the ‘‘lazy eye’’, with the aim to restore the balance of transcallosal inhibitory influences between hemispheres. FIGURE 4 \| VEP amplitudes in subjects with normal visual acuity. VEP amplitudes increased on the side contralateral to the application of cathodal tDCS, while they were dampened ipsilaterally. Nonetheless, different from amblyopic participants, at T2 there was a loss of the contralateral facilitation, paralleled by a persistent inhibition of the responses recorded from the polarized hemisphere. Data are given as mean values ± standard error (SE); ∗∗∗p < 0.0001. FIGURE 4 \| VEP amplitudes in subjects with normal visual acuity. VEP amplitudes increased on the side contralateral to the application of cathodal tDCS, while they were dampened ipsilaterally. Nonetheless, different from amblyopic participants, at T2 there was a loss of the contralateral facilitation, paralleled by a persistent inhibition of the responses recorded from the polarized hemisphere. Data are given as mean values ± standard error (SE); ∗∗∗p < 0.0001. Antal, A., Nitsche, M. A., and Paulus, W. (2006). Transcranial direct current stimulation and the visual cortex. Brain Res. Bull. 68, 459–463. doi: 10.1016/j. brainresbull.2005.10.006 Antonini, A., Fagiolini, M., and Stryker, M. P. (1999). Anatomical correlates of functional plasticity in mouse visual cortex. J. Neurosci. 19, 4388–4406. doi: 10.1523/JNEUROSCI.19-11-04388.1999 CONCLUSION Overall, our data support the use of unilateral cathodal tDCS for the treatment of amblyopia in adults, when pharmacological and mechanical therapies are completely ineffective; in order to improve and prolong the clinical outcome, both in adults and children, tDCS may be also combined with novel behavioral methods, comprising dichoptic training, perceptual learning and video gaming (Tsirlin et al., 2015; Vedamurthy et al., 2015a,b). Although promising, these therapies are currently influenced by visual attention, possibly narrowing their application in clinical practice. Overall, our data support the use of unilateral cathodal tDCS for the treatment of amblyopia in adults, when pharmacological and mechanical therapies are completely ineffective; in order to improve and prolong the clinical outcome, both in adults and children, tDCS may be also combined with novel behavioral methods, comprising dichoptic training, perceptual learning and video gaming (Tsirlin et al., 2015; Vedamurthy et al., 2015a,b). Although promising, these therapies are currently influenced by visual attention, possibly narrowing their application in clinical practice. Antonini, A., Fagiolini, M., and Stryker, M. P. (1999). Anatomical correlates of functional plasticity in mouse visual cortex. J. Neurosci. 19, 4388–4406. doi: 10.1523/JNEUROSCI.19-11-04388.1999 Antal, A., Nitsche, M. A., and Paulus, W. (2006). Transcranial direct current stimulation and the visual cortex. Brain Res. Bull. 68, 459–463. doi: 10.1016/j. brainresbull.2005.10.006 REFERENCES Transcallosal inhibition dampens neural responses to high contrast stimuli in human visual cortex. Neuroscience 187, 43–51. doi: 10.1016/j.neuroscience. 2011.04.050 Lunghi, C., Binda, P., and Morrone, M. C. (2010). Touch disambiguates rivalrous perception at early stages of visual analysis. Curr. Biol. 20, R143–R144. doi: 10.1016/j.cub.2009.12.015 Bocci, T., Caleo, M., Restani, L., Barloscio, D., Rossi, S., and Sartucci, F. (2016). Altered recovery from inhibitory repetitive transcranial magnetic stimulation (rTMS) in subjects with photosensitive epilepsy. Clin. Neurophysiol. 127, 3353–3361. doi: 10.1016/j.clinph.2016.06.013 Lunghi, C., Burr, D. C., and Morrone, C. (2011). Brief periods of monocular deprivation disrupt ocular balance in human adult visual cortex. Curr. Biol. 21, R538–R539. doi: 10.1016/j.cub.2011.06.004 Lunghi, C., Burr, D. C., and Morrone, M. C. (2013). Long-term effects of monocular deprivation revealed with binocular rivalry gratings modulated in luminance and in color. J. Vis. 13:1. doi: 10.1167/13.6.1 Bocci, T., Pietrasanta, M., Cerri, C., Restani, L., Caleo, M., and Sartucci, F. (2014). Visual callosal connections: role in visual processing in health and disease. Rev. Neurosci. 25, 113–127. doi: 10.1515/revneuro-2013-0025 Nitsche, M. A., Liebetanz, D., Lang, N., Antal, A., Tergau, F., and Paulus, W. (2003). Safety criteria for transcranial direct current stimulation (tDCS) in humans. Clin. Neurophysiol. 114, 2220–2222; author reply 2222–2223. doi: 10.1016/s1388-2457(03)00235-9 Cerri, C., Restani, L., and Caleo, M. (2010). Callosal contribution to ocular dominance in rat primary visual cortex. Eur. J. Neurosci. 32, 1163–1169. doi: 10.1111/j.1460-9568.2010.07363.x Corballis, M. C. (1996). A dissociation in naming digits and colors following commissurotomy. Cortex 32, 515–525. doi: 10.1016/s0010-9452(96)80008-6 Olma, M. C., Dargie, R. A., Behrens, J. R., Kraft, A., Irlbacher, K., Fahle, M., et al. (2013). Long-term effects of serial anodal tDCS on motion perception in subjects with occipital stroke measured in the unaffected visual hemifield. Front. Hum. Neurosci. 7:314. doi: 10.3389/fnhum.2013.00314 Davis, A. R., Sloper, J. J., Neveu, M. M., Hogg, C. R., Morgan, M. J., and Holder, G. E. (2003). Electrophysiological and psychophysical differences between early- and late-onset strabismic amblyopia. Invest. Ophthalmol. Vis. Sci. 44, 610–617. doi: 10.1167/iovs.02-0240 Palmer, L. M., Schulz, J. M., Murphy, S. C., Ledergerber, D., Murayama, M., and Larkum, M. E. (2012). The cellular basis of GABAB-mediated interhemispheric inhibition. Science 335, 989–993. doi: 10.1126/science.1217276 Ding, Z., Li, J., Spiegel, D. P., Chen, Z., Chan, L., Luo, G., et al. (2016). The effect of transcranial direct current stimulation on contrast sensitivity and visual evoked potential amplitude in adults with amblyopia. Sci. Rep. 6:19280. REFERENCES Antal, A., Nitsche, M. A., Kincses, T. Z., Kruse, W., Hoffmann, K. P., and Paulus, W. (2004). Facilitation of visuo-motor learning by transcranial direct current stimulation of the motor and extrastriate visual areas in humans. Eur. J. Neurosci. 19, 2888–2892. doi: 10.1111/j.1460-9568.2004.03367.x May 2018 \| Volume 12 \| Article 109 Frontiers in Behavioral Neuroscience \| www.frontiersin.org 7 Cathodal tDCS in the Treatment of Amblyopia Bocci et al. Behrens, J. R., Kraft, A., Irlbacher, K., Gerhardt, H., Olma, M. C., and Brandt, S. A. (2017). Long-lasting enhancement of visual perception with repetitive noninvasive transcranial direct current stimulation. Front. Cell. Neurosci. 11:238. doi: 10.3389/fncel.2017.00238 Lang, N., Siebner, H. R., Chadaide, Z., Boros, K., Nitsche, M. A., Rothwell, J. C., et al. (2007). Bidirectional modulation of primary visual cortex excitability: a combined tDCS and rTMS study. Invest. Ophthalmol. Vis. Sci. 48, 5782–5787. doi: 10.1167/iovs.07-0706 Berardi, N., Bisti, S., and Maffei, L. (1987). The transfer of visual information across the corpus callosum: spatial and temporal properties in the cat. J. Physiol. 384, 619–632. doi: 10.1113/jphysiol.1987.sp016473 Li, X., Dumoulin, S. O., Mansouri, B., and Hess, R. F. (2007). Cortical deficits in human amblyopia: their regional distribution and their relationship to the contrast detection deficit. Invest. Ophthalmol. Vis. Sci. 48, 1575–1591. doi: 10.1167/iovs.06-1021 j y Berardi, N., and Fiorentini, A. (1987). Interhemispheric transfer of visual information in humans: spatial characteristics. J. Physiol. 384, 633–647. doi: 10.1113/jphysiol.1987.sp016474 Li, J., Hess, R. F., Chan, L. Y., Deng, D., Yang, X., Chen, X., et al. (2013). Quantitative measurement of interocular suppression in anisometropic amblyopia: a case-control study. Ophthalmology 120, 1672–1680. doi: 10.1016/j. ophtha.2013.01.048 Bhola, R., Keech, R. V., Kutschke, P., Pfeifer, W., and Scott, W. E. (2006). Recurrence of amblyopia after occlusion therapy. Ophthalmology 113, 2097–2100. doi: 10.1016/j.ophtha.2006.04.034 Li, J., Thompson, B., Lam, C. S., Deng, D., Chan, L. Y., Maehara, G., et al. (2011). The role of suppression in amblyopia. Invest. Ophthalmol. Vis. Sci. 52, 4169–4176. doi: 10.1167/iovs.11-7233 Bi, H., Zhang, B., Tao, X., Harwerth, R. S., Smith, E. L. III., and Chino, Y. M. (2011). Neuronal responses in visual area V2 (V2) of macaque monkeys with strabismic amblyopia. Cereb. Cortex 21, 2033–2045. doi: 10.1093/cercor/bhq272 Lo Verde, L., Morrone, M. C., and Lunghi, C. (2017). Early cross-modal plasticity in adults. J. Cogn. Neurosci. 29, 520–529. doi: 10.1162/jocn_a_01067 Bocci, T., Caleo, M., Giorli, E., Barloscio, D., Maffei, L., Rossi, S., et al. (2011). REFERENCES doi: 10.1038/srep19280 Pediatric Eye Disease Investigator Group, Repka, M. X., Kraker, R. T., Beck, R. W., Holmes, J. M., Cotter, S. A., et al. (2008). A randomized trial of atropine vs. patching for treatment of moderate amblyopia: follow-up at age 10 years. Arch. Ophthalmol. 126, 1039–1044. doi: 10.1001/archopht.126. 8.1039 Duffy, F. H., Burchfiel, J. L., and Conway, J. L. (1976). Bicuculline reversal of deprivation amblyopia in the cat. Nature 260, 256–257. doi: 10.1038/260256a0 Pietrasanta, M., Restani, L., Cerri, C., Olcese, U., Medini, P., and Caleo, M. (2014). A switch from inter-ocular to inter-hemispheric suppression following monocular deprivation in the rat visual cortex. Eur. J. Neurosci. 40, 2283–2292. doi: 10.1111/ejn.12573 Galea, J. M., Jayaram, G., Ajagbe, L., and Celnik, P. (2009). Modulation of cerebellar excitability by polarity-specific noninvasive direct current stimulation. J. Neurosci. 29, 9115–9122. doi: 10.1523/JNEUROSCI.2184 -09.2009 Gandiga, P. C., Hummel, F. C., and Cohen, L. G. (2006). Transcranial DC stimulation (tDCS): a tool for double-blind sham-controlled clinical studies in brain stimulation. Clin. Neurophysiol. 117, 845–850. doi: 10.1016/j.clinph.2005. 12.003 Reinhart, R. M., Xiao, W., McClenahan, L. J., and Woodman, G. F. (2016). Electrical stimulation of visual cortex can immediately improve spatial vision. Curr. Biol. 26, 1867–1872. doi: 10.1016/j.cub.2016.05.019 Repka, M. X., Wallace, D. K., Beck, R. W., Kraker, R. T., Birch, E. E., Cotter, S. A., et al. (2005). Two-year follow-up of a 6-month randomized trial of atropine vs. patching for treatment of moderate amblyopia in children. Arch. Ophthalmol. 123, 149–157. doi: 10.1001/archopht.123.2.149 Gao, T. Y., Guo, C. X., Babu, R. J., Black, J. M., Bobier, W. R., Chakraborty, A., et al. (2018). Effectiveness of a binocular video game vs. placebo video game for improving visual functions in older children, teenagers, and adults with amblyopia: a randomized clinical trial. JAMA Ophthalmol. 136, 172–181. doi: 10.1001/jamaophthalmol.2017.6090 Restani, L., Cerri, C., Pietrasanta, M., Gianfranceschi, L., Maffei, L., and Caleo, M. (2009). Functional masking of deprived eye responses by callosal input during ocular dominance plasticity. Neuron 64, 707–718. doi: 10.1016/j.neuron.2009. 10.019 j Hess, R. F., Li, X., Lu, G., Thompson, B., and Hansen, B. C. (2010). The contrast dependence of the cortical fMRI deficit in amblyopia; a selective loss at higher contrasts. Hum. Brain Mapp. 31, 1233–1248. doi: 10.1002/hbm.20931 Roser, M., and Corballis, M. C. (2003). Interhemispheric neural summation in the split brain: effects of stimulus colour and task. Neuropsychologia 41, 830–846. doi: 10.1016/s0028-3932(02)00290-7 Hess, R. F., and Thompson, B. (2015). REFERENCES Amblyopia and the binocular approach to its therapy. Vision Res. 114, 4–16. doi: 10.1016/j.visres.2015.02.009 Holmes, J. M., and Clarke, M. P. (2006). Amblyopia. Lancet 367, 1343–1351. doi: 10.1016/S0140-6736(06)68581-4 Spiegel, D. P., Byblow, W. D., Hess, R. F., and Thompson, B. (2013a). Anodal transcranial direct current stimulation transiently improves contrast sensitivity and normalizes visual cortex activation in individuals with amblyopia. Neurorehabil. Neural Repair 27, 760–769. doi: 10.1177/1545968313 491006 Holmes, J. M., Lazar, E. L., Melia, B. M., Astle, W. F., Dagi, L. R., Donahue, S. P., et al. (2011). Effect of age on response to amblyopia treatment in children. Arch. Ophthalmol. 129, 1451–1457. doi: 10.1001/archophthalmol.2011.179 Spiegel, D. P., Li, J., Hess, R. F., Byblow, W. D., Deng, D., Yu, M., et al. (2013b). Transcranial direct current stimulation enhances recovery of stereopsis in Irlbacher, K., Brocke, J., Mechow, J. V., and Brandt, S. A. (2007). Effects of GABAA and GABAB agonists on interhemispheric inhibition in man. Clin. Neurophysiol. 118, 308–316. doi: 10.1016/j.clinph.2006.09.023 May 2018 \| Volume 12 \| Article 109 Frontiers in Behavioral Neuroscience \| www.frontiersin.org 8 Cathodal tDCS in the Treatment of Amblyopia Bocci et al. adults with amblyopia. Neurotherapeutics 10, 831–839. doi: 10.1007/s13311- 013-0200-y Vedamurthy, I., Nahum, M., Huang, S. J., Zheng, F., Bayliss, J., Bavelier, D., et al. (2015b). A dichoptic custom-made action video game as a treatment for adult amblyopia. Vision Res. 114, 173–187. doi: 10.1016/j.visres.2015.04.008 adults with amblyopia. Neurotherapeutics 10, 831–839. doi: 10.1007/s13311- 013-0200-y Stagg, C. J., Best, J. G., Stephenson, M. C., O’Shea, J., Wylezinska, M., Kincses, Z. T., Stagg, C. J., Best, J. G., Stephenson, M. C., O’Shea, J., Wylezinska, M., Kincses, Z. T., et al. (2009). Polarity-sensitive modulation of cortical neurotransmitters by transcranial stimulation. J. Neurosci. 29, 5202–5206. doi: 10.1523/JNEUROSCI. 4432-08.2009 von Noorden, G. K., and Crawford, M. L. (1992). The lateral geniculate nucleus in human strabismic amblyopia. Invest. Ophthalmol. Vis. Sci. 33, 2729–2732. Williams, C., Northstone, K., Harrad, R. A., Sparrow, J. M., Harvey, I., and ALSPAC Study Team. (2003). Amblyopia treatment outcomes after preschool screening v school entry screening: observational data from a prospective cohort study. Br. J. Ophthalmol. 87, 988–993. doi: 10.1136/bjo.87.8.988 Stewart, C. E., Stephens, D. A., Fielder, A. R., Moseley, M. J., and MOTAS Cooperative (2007). Modeling dose-response in amblyopia: toward a child- specific treatment plan. Invest. Ophthalmol. Vis. Sci. 48, 2589–2594. doi: 10.1167/iovs.05-1243 y p j Zhou, J., Thompson, B., and Hess, R. F. (2013). Frontiers in Behavioral Neuroscience \| www.frontiersin.org REFERENCES A new form of rapid binocular plasticity in adult with amblyopia. Sci. Rep. 3:2638. doi: 10.1038/srep02638 ˇ Tao, X., Zhang, B., Shen, G., Wensveen, J., Smith, E. L. III., Nishimoto, S., et al. (2014). Early monocular defocus disrupts the normal development of receptive-field structure in V2 neurons of macaque monkeys. J. Neurosci. 34, 13840–13854. doi: 10.1523/JNEUROSCI.1992-14.2014 ˇZiak, P., Holm, A., Halicka, J., Mojzis, P., and Piñero, D. P. (2017). Amblyopia treatment of adults with dichoptic training using the virtual reality oculus rift head mounted display: preliminary results. BMC Ophthalmol. 17:105. doi: 10.1186/s12886-017-0501-8 Thompson, B., Mansouri, B., Koski, L., and Hess, R. F. (2008). Brain plasticity in the adult: modulation of function in amblyopia with rTMS. Curr. Biol. 18, 1067–1071. doi: 10.1016/j.cub.2008.06.052 Conﬂict of Interest Statement: 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. Tsirlin, I., Colpa, L., Goltz, H. C., and Wong, A. M. (2015). Behavioral training as new treatment for adult amblyopia: a meta-analysis and systematic review. Invest. Ophthalmol. Vis. Sci. 56, 4061–4075. doi: 10.1167/iovs.15 -16583 Copyright © 2018 Bocci, Nasini, Caleo, Restani, Barloscio, Ardolino, Priori, Maffei, Nardi and Sartucci. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Turrigiano, G. (2012). Homeostatic synaptic plasticity: local and global mechanisms for stabilizing neuronal function. Cold Spring Harb. Perspect. Biol. 4:a005736. doi: 10.1101/cshperspect.a005736 Vedamurthy, I., Nahum, M., Bavelier, D., and Levi, D. M. (2015a). Mechanisms of recovery of visual function in adult amblyopia through a tailored action video game. Sci. Rep. 5:8482. doi: 10.1038/srep08482 May 2018 \| Volume 12 \| Article 109 Frontiers in Behavioral Neuroscience \| www.frontiersin.org 9
https://openalex.org/W4288562810	https://zenodo.org/record/2582899/files/2.pdf	Latin	null	EFL STUDENTS' PERCEPTIONS OF WHATSAPP AND ITS POTENTIAL BENEFITS IN ELT PRACTICUM	Zenodo (CERN European Organization for Nuclear Research)	2,019	cc-by	10,357	* Received: 9/1/2018, Accepted: 23/2/2018. Associate Professor/Al-Quds Open University/Palestine. Associate Professor/Al-Quds Open University/Palestine. * Received: 9/1/2018, Accepted: 23/2/2018. Dr. Khaled Abdel Jaleel Dweikat ** 14 Palestinian Journal of Technology & Applied Sciences - No. 2 - January 2019 تصورات متعلمي اللغة االجنليزية كلغة �أجنبية حول تطبيق الواتساب وفوائده املحتملة يف الرتبية العملية لتخصص اللغة االجنليزية تصورات متعلمي اللغة االجنليزية كلغة �أجنبية حول تطبيق الواتساب وفوائده املحتملة يف الرتبية العملية لتخصص اللغة االجنليزية Abstract Technology has become an indispensable part of our daily life including the educational system. The current spread of the Internet and its enormous applications have provided various benefits, specifically, the use of web-based technologies in education. Since the introduction of m-learning into education, mobile phones, smartphones and tablets have become a great part of foreign language learning and teaching. Therefore, the widespread availability of smartphone technology seems to have a promising impact on education. In particular, WhatsApp, as a smartphone application, is considered one of the most popular communication platforms in the 21st century. Hence, the current study aimed at investigating the perceptions of WhatsApp among EFL Students, and its potential benefits in ELT practicum at al-Quds Open University in Palestine. The participants were 30 senior students who were enrolled in the English major (28 females and 2 males). These students were enrolled in EFL practicum during the academic year 2017-2018. To achieve the objectives of the study, a mixed research method was employed using multiple tools for data collection, including a questionnaire, a focus group and phone interviews via WhatsApp. Results revealed that the total degree of the students’ perceptions of WhatsApp was high. The participating students used WhatsApp application to send and receive different types of materials and activities for communicating with the supervisor, the total number of which was 752. In light of these results, a number of recommendations were provided including, encouraging faculty members to integrate social media into the teaching environments by providing them with training programs, workshops, and online training. Moreover, the need for providing faculty members with curricula plans and activities that require implementation of WhatsApp in the learning environments. Introduction and theoretical background contextual learning environment that can exist anywhere. As a result, users can control the time, pace and speed of their own learning. M-learning is also more personalized than other methods of computerized instruction, as mobile devices can be more easily customized, resulting in the creation of an emotional bond between the user and the machine. Sweeney and Moore (2012) discuss four aspects which need to be considered for the design of mobile language learning apps: the mobile app design should allow for interactivity, the learning resources should include multimedia contents, utility and functionality of the learning resources should be fostered by contextual relevancy, and the app design should increase autonomous and personalized learning. The 21st century has witnessed unprecedentedly great technological innovations as a result of the emergence of Information & Communication Technology (ICT), impacting all aspects of life. The Internet transformed our universe into a small global village that is easily interconnected. The use of technology became an inevitable part of almost every aspect of our daily life and the educational systems are no exception, as technology has a positive impact on the teaching/ learning process. The great spread of the Internet and its enormous applications encouraged the use of web-based technologies in education and exploitation of their numerous benefits. This paved the way for mobile learning to be an innovative mean to deliver content and to embed technology in the university education. In this respect, the last two decades have witnessed a remarkable shift in Mobile Assisted Language Learning (MALL) to become of a high interest among instructors and researchers. Obari & Lambacher (2015) argued that mobile technologies can be used in language learning with diverse advantages than the other methods, especially with the spread of English news programs, language-learning apps, podcasting and video-casting, that are easily accessible mostly for free. Mobiles, smartphones and tablets have been greatly introduced to foreign language learning as a result of introducing m-learning into education in the developed and developing countries alike. Especially when the features of computers are merged with mobile, increasing the access to technology at any time and in any place. Martin and Ertzberger (2013) studied the difference between using computers and mobile phones within a classroom setting and found that students showed more enthusiasm towards mobile devices. ملخص: غدت التكنولوجيا جزءاً ال غنى عنه يف حياتنا اليومية مبا فيها النظم التعليمية، ال 6سيما يف ظل االنتشار الكبري لشبكة الإنرتنت وتطبيقاتها الهائلة يف وقتنا احلا9رض، وهناك فوائد وفرية ال6ستخدام التكنولوجيا القائمة على 7شبكة الإنرتنت يف التعليم. ومع ولوج التعلم النقال حلقل التعليم، حلقت الهواتف النقالة والأجهزة اللوحية هذا الركب حني دخلت هي الأخرى جماالت تعليم اللغة الأجنبية، وبناء عليه، يبدو وا9ضحا مدى ت�أثري وفرة تكنولوجيا الهواتف الذكية. ويعد تطبيق الواتساب بشكل خا8ص، كونه تطبيقا من تطبيقات الهواتف الذكية، واحداً من منصات االتصاالت الأكرث 7شعبية يف القرن احلادي والعرشين. من هنا هدفت الدرا6سة احلالية �إىل التعرف على تصورات طالب اللغة الإجنليزية كلغة �أجنبية جتاه تطبيق الواتساب وفوائده املحتملة يف مقرر الرتبية العملية لطلبة اللغة الإجنليزية يف جامعة القد6س املفتوحة يف فلسطني. وتكوَّن جمتمع الدرا6سة من ثالثني طالبا وطالبة من طلبة من الإناث 28( السنة الرابعة يف تخصص اللغة الإجنليزية من الذكور) الذين 6سجلوا يف مقرر الرتبية العملية 2 و . 2018-2017 يف الفصل الأول من العام الدرا6سي ولتحقيق �أهداف الدرا6سة، جرى ا6ستخدام �أ6سلوب بحث خمتلط 7شمل مصادر متعددة جلمع البيانات مبا يف ذلك الإ6ستبيان، وجمموعة ب�ؤرية، ومقابالت هاتفية عرب تطبيق الواتساب. و�أظهرت النتائج �أن الدرجة الكلية لتصورات الطلبة حول تطبيق الواتساب كانت عالية، وقد ا6ستخدم الطالب املشاركون تطبيق الواتساب لإر6سال وا6ستقبال �أنواع خمتلفة من املواد والأنشطة للتوا8صل مع املرشف مادة ونشاطا. ويف 752 بلغ العدد الإجمايل منها نشاطا 9ضوء هذه النتائج، قدمت جملة من التو8صيات. الكلمات املفتاحية: تعلم اللغات مبساعدة التعلم النقال، التعلم النقال، الواتساب، تصورات الطلبة، الرتبية العلمية لتخصص اللغة الإجنليزية، جامعة القد6س املفتوحة. Keywords: MALL, M-Learning, WhatsApp, Students Perceptions, ELT Practicum, QOU. Keywords: MALL, M-Learning, WhatsApp, Students Perceptions, ELT Practicum, QOU. 15 EFL STUDENTS’ PERCEPTIONS OF WHATSAPP AND ITS POTENTIAL BENEFITS IN ELT PRACTICUM EFL STUDENTS’ PERCEPTIONS OF WHATSAPP Dr. Khaled Abdel Jaleel Dweikat Introduction and theoretical background Such results indicate that mobile devices in general and smartphones in particular seem to be promising tools for deepening learning and making it further personalized and learner-centered. Kukulska-Hulme (2012) identifies it as the use of mobile technologies in language learning, especially in situations where device portability offers specific advantages. Traxler (2013) similarly believes that mobile learning is learning through mobile technologies such as mobile phones, smartphones, e-readers and tablets. These devices offer unparalleled access to communication and information. Mobile learning (m-learning) thus, is considered a learner-centric approach which focuses on the mobility of the learner as well as the mobility of the learning process itself (Böhm & Constantine, 2016) In mobile language learning, the language learning process is assisted and enhanced by the use of a mobile device. Often, the acronym MALL is used to describe this approach of language learning. By the same token, Obari and Lambacher (2014), argue that m-learning can highly motivate learners since it provides them with a rich, informal and Meishar-Tal and Ronen (2016) found that teachers’ attitudes towards using smartphones in education were positive in all aspects .Such findings go in line with some previous studies which indicated that employing smartphones in education has the potential to decrease teachers’ resistance to technology and increase their willingness to use these devices in teaching (Mac Callum & Jeffrey, 2014). Billions of people these days, as a result of the affordability of smartphones all over the world, can access different forms of social media platforms such as Facebook, WhatsApp, Messenger, WeChat, Instagram Twitter, and Snapchat to keep in touch with their family and friends, and to exchange information. Claire 16 Palestinian Journal of Technology & Applied Sciences - No. 2 - January 2019 Moreover, the number of smartphone users is rapidly increasing among high school and college students to the extent that there seems to be an emotional bond between the smartphone user and this device. One reason for this widespread use is the availability and affordability of mobile devices in the Palestinian market. Moreover, smartphones allow users to make personal profiles, create content and share messages by connecting with other users in the system anytime anywhere. The small screen size and touch interface of smartphones are also factors behind the widespread use of these devices. Introduction and theoretical background Thus, the current study aimed at investigating the perceptions of EFL Students towards WhatsApp and its potential benefits in ELT Practicum at al-Quds Open University in Palestine. More specifically, the research questions of this study are as follows: O’Malley et al., (2005) stated that the availability of the technical infrastructure for m-learning has paved the way for social networking and media applications such as WhatsApp and Blogs, that enable learners to interact with each other outside the classroom. Patient (2011) argued that contents of conversations among students using smartphone application revolved around exchanging academic information. Mistar and Amin Embi (2016) stated the benefit of integrating online discussion through smartphone applications in the learning process is the provision of the students with more opportunities to interact with their teachers and friends, which contributes to their learning and help in overcoming learning difficulties. Smartphone is simply defined as a mobile phone with advanced computing capabilities and connectivity. Due to the accelerated development of smartphones, people started to move beyond mobile brands and features to look for mobile applications. Therefore, the widespread availability of smartphone technology seems to have a promising impact on education when it enables learners to communicate with their instructors and their classmates in real time, especially with the availability of wireless internet (Wifi) and the latest 3G technology. The third generation of mobile phone network technology is capable of fast rates of data transmission that supports e-mail communications, high-speed Internet access, video streaming, and so forth (Collins English Dictionary). All these innovative technological advancements make it much cheaper to communicate via social networking sites. Consequently, teachers everywhere including Palestine can benefit from the unique features of the latest release of smartphones to create a more interactive learning experience that helps students become engaged inside and outside the classroom. As a result, students’ motivation will be increased especially when smartphones can be used 24/7, nearly for free. According to Sharples and Vavoula (2007), smartphones provide students with the ability to learn outside the classroom environment where the students become the center of learning. 1. What are the perceptions of WhatsApp among EFL students in practicum? 2. What are the benefits obtained by students of ELT practicum when using WhatsApp? What is up with WhatsApp? Furthermore, Mistar and Amin Embi (2016) indicated that 97% of Malaysians were found to be using WhatsApp because it has a variety of features and functions that enable the user to exchange easily text messages, audio files, video , images and website links. Regarding cost saving, Riyanto (2013) argued that users take advantage of WhatsApp to text their friends in other countries without paying the exorbitant international texting costs that come with traditional communications. Roffe (2013) agrees with this point of view when he argued that a big reason for the popularity of such apps is that they allow their users to message one another without paying high fees for text messages. This means that WhatsApp users need not worry about expensive call charges when it’s made through the phone’s internet connection, instead of cell plan’s voice minutes, to have face-to-face conversations and voice and video calls in addition to text messages. All these features imply that WhatsApp is easy, cheap, convenient, reliable, entertaining, which might explain why WhatsApp has currently around 1.5 billion users over the globe. The good news is that this application is available to all brands of smart phones and the software is available for Apple iOS, Google Android, Blackberry OS, Microsoft windows phone and others. Once the user has an efficient smartphone with a connection to the internet, whether through Wi-Fi or the user’s cell phone network, he /she can install the application from Google Play Store and start communicating with friends, relatives, classmates, and teachers as well. Regarding the history of WhatsApp, the service was founded by Jan Koum and Brian Acton who had previously worked 20 years combined at Yahoo. Later, in 2014, WhatsApp joined Facebook but continues to operate as a separate app with laser focus on building a messaging service that works fast and reliably anywhere in the world (https://www.whatsapp.com/about/). According to its official site, WhatsApp Messenger is a cross platform mobile messaging app for smart phones such as iPhone, Android phones, Windows Mobile or Blackberry. WhatsApp allows users to send and receive messages, photos and other information, and it is considered an alternative to text messages or SMS. Furthermore, WhatsApp can be used for text and voice messaging, chatting, audio and video calling, sending-receiving documents and files, and creating groups. What is up with WhatsApp? Church and de Oliveira (2013) carried out a study titled “What’s up with WhatsApp?” to investigate the difference in the perceptions and motives of the use of WhatsApp and traditional SMS, in order to provide a deeper understanding of the reasons and mechanism behind adopting and appropriating WhatsApp in the smartphone users’ daily lives. The study also aimed at exploring which factors influence the acceptance, usage and popularity of such application. The results revealed that the nature and intent of WhatsApp messages tend to be more social, informal and conversational, while SMS is seen as more privacy preserving, more formal and generally more reliable. Mistar and Amin Embi (2016) reported that WhatsApp is the most popular communication platform in the 21st century depending on the fact that it uses real-time messages and fast knowledge resources. Ramakrishnan and Johnsi Priya (2016) found that WhatsApp ranks at the top unlike the other applications. The results came after revealing that 100% of arts and science college students chat In the Palestinian context, people, especially teenagers and university students, are increasingly conversing online using a variety of social sites like Facebook, Twitter, Instagram and others. 17 EFL STUDENTS’ PERCEPTIONS OF WHATSAPP AND ITS POTENTIAL BENEFITS IN ELT PRACTICUM Dr. Khaled Abdel Jaleel Dweikat gray arrow and the two blue arrows respectively. The user in WhatsApp group can name the group, mute or customize notifications, use voice calls, share messages, photos and videos with up to 256 people at once. with friends. In their study, 96.33% of students chat with classmates and less than 1% chat with neighbors, parents and teachers. This popularity was also confirmed by Constine (2018) who stated that WhatsApp hit 1.5 billion monthly users in January 2018. Furthermore, Mistar and Amin Embi (2016) indicated that 97% of Malaysians were found to be using WhatsApp because it has a variety of features and functions that enable the user to exchange easily text messages, audio files, video , images and website links. with friends. In their study, 96.33% of students chat with classmates and less than 1% chat with neighbors, parents and teachers. This popularity was also confirmed by Constine (2018) who stated that WhatsApp hit 1.5 billion monthly users in January 2018. Literature review Other supporting evidence was given by Ida Sanjaya (2012) whose study revealed that users of WhatsApp tend to use it because it has attractive facilities and functions that enable the user to share information in the form of text files, audio files , videos, messages, images and others. These features and others explain why this networking tool is one of the most popular sites according to Aifan (2015). She found that the most frequently used tool by her sample of 510 students was WhatsApp, which has become a mobile learning style for faster and easier communication between students and teachers, in particular when teachers use it to attract their students interest through fun- based learning experiences. Students’ perceptions to the use of WhatsApp in their educational processes have been examined in other studies. For instance, Malecela (2016) examined students’ perceptions of WhatsApp as a learning tool in Malaysia. A qualitative research design was used in which interviews were conducted with a number of students. The results showed that the students believed WhatsApp is helpful in their learning by facilitating communication with other students and with the instructor, by enhancing collaborative learning, and by sharing educational information. Concerning perceptions and attitudes towards WhatsApp and its benefits, Gasaymeh (2017) aimed at investigating 154 university students’ use of WhatsApp for personal and educational purposes in addition to examining their perceptions of the formal integration of WhatsApp into education. The results revealed that WhatsApp was common among participants for personal and social purposes on a daily basis while the use of WhatsApp for educational purposes was limited. Nevertheless, the participants perceived the integration of WhatsApp into education to be easy, fun, and useful and they had positive feelings and intentions about using WhatsApp in their formal learning .To some extent, Tang and Hew (2017) had a similar view when they reported that WhatsApp has been used in different academic disciplines to support students’ learning. So (2016) found that university students who had experience with the formal use of WhatsApp to support their learning, and those who had no experience with the use of WhatsApp for learning had positive perceptions of its use to support teaching and learning. Literature review Due to the widespread adoption of WhatsApp by approx. 1.5 billion users around the world, several researchers have examined the effects of using this innovation on students’ learning and on their attitudes and perceptions as well. Several studies have pointed that WhatsApp can be easily accessed and used by all people for communicative purposes. For example, Devi and Tevera (2014) found that WhatsApp was one of the most popular social sites among university students, while Ahad and Lim (2014) found that WhatsApp is popular among undergraduate students, and that they use it on a daily basis. They attributed their extensive use of WhatsApp to its ease of use, speed, real-time messaging, and low cost. This finding is in line with Amshah and Thabian (2017) study, which found that the top preferred social media among 450 students at the Jordanian universities was Facebook 88 %, WhatsApp 78.7%, YouTube 51.6%, Instagram 32.4%, Viber 26.9% and Twitter 13.6%. The students used WhatsApp for sending text messages, pictures, music and videos (10.7 %) while 16.7 % used them to share ideas with friends With regard to the importance of WhatsApp groups, Zayed (2016) highlighted the possible communication pattern between the teacher and the students when creating a group that can be used to answer students’ questions and inquiries about the course material and requirements, and to deliver announcements to them. The teacher can also check if the students have received and read the messages sent to them through the single 18 Palestinian Journal of Technology & Applied Sciences - No. 2 - January 2019 and experts and to exchange experience. WhatsApp facilitates mobile learning when the students use it as a communication tool outside the classroom to exchange information about meetings and projects. In addition, the authors found that students had positive attitudes towards using WhatsApp in education. These results go in line with Yeboah and Ewur (2014) who found that most higher education students were extensive users of WhatsApp; more than 96% used it for more than three hours a day. Most participants used it for chatting, while only 7% reported using it for academic work. Purpose of the study When teachers effectively integrate mobile learning technologies in their teaching, they can create engaging learning environments, especially as today students have already adopted mobile phones in their lives and use it increasingly for social interaction. Mobile phones as a result, allow students to communicate anytime anywhere they go. What makes this possible is the abundance of mobile apps specifically designed for language learning. Instructors everywhere can use a variety of mobile applications to facilitate students’ learning and initiate a real purpose for communication. Hence, the purpose of this study was twofold: (a) To identify the use of WhatsApp as a learning tool by a group of students in EFL practicum, (b) To identify students’ perceptions of the use of WhatsApp. Mistar and Amin Embi (2016) examined students’ perceptions to using WhatsApp as a learning tool to enhance language learning of 20 students from Kuala Pilah Pre-University. The results indicated that WhatsApp helped the students learn the language better and enhanced their English language proficiency. The researchers also indicated that WhatsApp should be used to encourage institutions to provide internet facilities as a top priority in today’s education, to invest in this application as a means to improve students’ confidence in learning and using the language. Aicha (2014) sought to explore the impact of using WhatsApp mobile learning activities on the achievement and attitudes of online students at the university. Specifically, this study compares an independent sample of students in an experimental group,15 students, with a control group, 15 students, at a university class. The e-learning process of the experimental group is based on Literature review carried out the tasks through WhatsApp in EFL speaking courses for four weeks. Participants’ views on the mobile application activities were also examined through face-to-face interviews. Results showed that WhatsApp experiences significantly affected students’ language learning by lowering EFL speaking anxiety. Similarly, the findings of Mar and Christine (2013) revealed that most students were highly satisfied with using WhatsApp in learning and revealed that WhatsApp increased their willingness to read in English. It also left positive impact on their reading habits which in turn enhanced their confidence. Minalla (2018) sought to explore the possibility of utilizing WhatsApp Group to enhance EFL learners’ verbal interaction. Minalla divided the students into two groups who were taught the content using the traditional approach combined with WhatsApp Chat groups via text message as communicative platforms for practicing outside classroom contexts. The participants of experimental group were restrictively interacted via voice messages while the participants of control group were only interacted via text messages. Results revealed that the participants who received the voice messages on WhatsApp treatment significantly outperformed those who received text messages on WhatsApp. Aifan (2015) found that the most frequently used social media tool among students at King Abdul-Aziz University in Jeddah, Saudi Arabia, was WhatsApp, whereas YouTube came second and Twitter came third. Furthermore, students reported facing two major obstacles when utilizing social media. First, some of the social media content was against the students’ religion. The second obstacle was related to concerns about privacy and security issues when using social media. Literature review Examples of reported possible advantages for the use of WhatsApp included: providing immediate messaging support, bringing new opportunities of learning, facilitating communication between students and teachers, enabling fast feedback in learning, providing flexible learning, supporting multimedia learning, and supporting collaborative learning. However, the participants said that the use of WhatsApp in their learning might interfere with their private lives. Selva Kumar (2016) aimed at understanding the respondents’ perceptions towards using WhatsApp as a learning tool for Mandarin language learning at Universiti Kuala Lumpur. The data was collected via a questionnaire on students’ attitudes towards using WhatsApp and their perceptions. Results indicated that the utilization of WhatsApp improved students’ learning performance inside and outside the classroom. The main advantage of using WhatsApp is that it can be used anywhere- anytime, and it can be adopted to enhance students’ interaction and learning experiences in Mandarin Language lessons. With regard to WhatsApp activities and its educational benefits to students’ learning, Turgay and Keskin (2016) examined the effect of using WhatsApp activities in undergraduate EFL speaking classes on students’ speaking anxiety, and their feelings about the activities conducted. Thirty-nine undergraduate level participants Ngaleka & Uys (2013) reported that 19 EFL STUDENTS’ PERCEPTIONS OF WHATSAPP AND ITS POTENTIAL BENEFITS IN ELT PRACTICUM Dr. Khaled Abdel Jaleel Dweikat WhatsApp mobile learning activities while the e-learning process of the control group is without WhatsApp mobile learning activities and receives only face-to-face learning in the classroom. The results of the experiment showed that the experimental group, which uses mobile learning through WhatsApp mobile instant messaging, performed better than the control group on the achievement test following the experimental period. The attitudes of the students suggested that WhatsApp instant messaging made learning easy, favored problem solving and resolved learning difficulties, related to the learning process or to learning content distributed through WhatsApp, knowledge sharing. carried out the tasks through WhatsApp in EFL speaking courses for four weeks. Participants’ views on the mobile application activities were also examined through face-to-face interviews. Results showed that WhatsApp experiences significantly affected students’ language learning by lowering EFL speaking anxiety. Similarly, the findings of Mar and Christine (2013) revealed that most students were highly satisfied with using WhatsApp in learning and revealed that WhatsApp increased their willingness to read in English. It also left positive impact on their reading habits which in turn enhanced their confidence. Statement of the problem According to the aforementioned literature review, WhatApps has several educational benefits, 20 Palestinian Journal of Technology & Applied Sciences - No. 2 - January 2019 including creating strong social ties between students and teachers, and overall, students are more positive towards the learning environment. Added to this, wherever we go inside today universities, one can eyewitness students sit with their smartphones ignoring the presence of others. The popularity of the best –seller smartphones among university students might suggest that they would be suitable and valuable tools which could be used for educational purposes. Furthermore, the mobile chat applications have no barriers with regard to age group, nativity, social status and economic status as these can be used for free. have greater perspectives of the Net Generation technology expertise and how student learning is connected with technology as this is an essential component for higher education institutions. Based on the knowledge of the researcher, there were no studies conducted on students’ perceptions towards the use of WhatsApp application in the Palestinian context, in particular at QOU. This research gap led the researcher to conduct the present study on the EFL students’ perception to the use of WhatsApp as a learning tool in the practicum course. Exploring such perceptions will provide Palestinian education system, particularly at QOU, with resources that help the instructors deal with the Net Generation and help them utilize these innovative tools to improve students’ learning. Regardless of these promising benefits, mobile phones are seen by many teachers only as means for recreation and entertainment and not for learning. For example, Yeboah and Ewur (2014) found that most participants in their study used WhatsApp for chatting while only 7% reported using it for academic work. Ramakrishnan and Johnsi Priya (2016) as well, found that although WhatsApp ranks top among all the other applications, 96.33% of students have chat with classmates and less than 1 % of students have chat with teachers. Findings of this study may also help the instructors change their attitudes towards the adoption of WhatsApp in order to improve students’ learning in practicum. Additionally, the findings may help the instructors to expand their 21st century information and technical knowledge and skills in order to keep up with knowledge pace of their students. Statement of the problem Furthermore, some teachers even oppose the use of smartphones at school claiming that these smartphones are distractive and therefore reduce attention and concentration in class. So (2016), for example, found that university students who had experience with the formal use of WhatsApp said that the use of WhatsApp in their learning might interfere with their private lives. By the same token, several experimental studies have examined the effect of using WhatsApp in several subjects, but little research has explored the impact of using these applications in ELT practicum. Therefore, the current study aims to investigate ELT students’ use of WhatsApp and their perceptions regarding its possible integration into their practicum course. ELT Practicum at QOU Practicum is considered an essential strategy for preparing students to be competent teachers. It is a pre-service teacher educational program. The practicum course at QOU seeks to provide an opportunity for the student teachers to practice the most important concepts, principles and educational theories in real teaching contexts at schools. This practice aims at providing the student teachers with good opportunities to acquire the required educational competencies to be competent teachers in the future. To achieve such objectives, student teachers are required to spend six weeks (150) hours at a nearby school where they practice a variety of roles under the supervision of a cooperative teacher and a university practicum supervisor. The course consists of three main stages: Instruments and Data Collection At this stage, which lasts for three weeks (75 hours), the student teacher is given the opportunity to practice teaching in real contexts, where he/she plans, develops behavioral objectives, teaches, uses different methods of teaching, uses different types of audiovisual aids, evaluates, and gives feedback. During this stage, the faculty member visits the student teacher for the second time in order to evaluate and assess his/her performance in class. 3. 3. Actual practice: This is the most important stage as it consists of the actual practices of teaching, because at this final stage student teachers are expected to be familiar with the multi tasks of a classroom teacher. At this stage, which lasts for three weeks (75 hours), the student teacher is given the opportunity to practice teaching in real contexts, where he/she plans, develops behavioral objectives, teaches, uses different methods of teaching, uses different types of audiovisual aids, evaluates, and gives feedback. During this stage, the faculty member visits the student teacher for the second time in order to evaluate and assess his/her performance in class. At the end of the course and after all students finished their practicum training at the cooperating schools, the researcher, who was the instructor of the course, sent a message that requested students to attend a wrap-up meeting at the university. During this meeting, the questionnaire was completed by 27 students who attended the meeting. The questionnaire took between 15-20 minutes to complete. The other three students completed the questionnaire and sent it to the researcher (the instructor) via the academic portal of QOU. The meeting lasted for 90 minutes and covered the students experience in the course and their positive and negative comments. At the end of the meeting the researcher asked 7 volunteers to participate in a focus group session which lasted for 50 minutes. Significance of the Study Higher educational institutions cannot ignore the role of technology in fulfilling their strategic mission and in responding to the expectations of a diverse student body in many aspects. Lohnes and Kinzer (2007) argued that faculty need to 1. Observation Stage: at this stage, the student teacher is expected to spend one week (25 hours) at the cooperative school where he/she is given the chance and the required support 21 EFL STUDENTS’ PERCEPTIONS OF WHATSAPP AND ITS POTENTIAL BENEFITS IN ELT PRACTICUM Dr. Khaled Abdel Jaleel Dweikat 4 years at QOU. The students were selected purposefully and based on voluntary willingness. The researcher was the academic supervisor of this EFL practicum course. to observe different in-class and outside practices and activities such as: observing disciplinary procedures, school management, staff meetings, ordering and ranking students, attending meetings and so forth. to observe different in-class and outside practices and activities such as: observing disciplinary procedures, school management, staff meetings, ordering and ranking students, attending meetings and so forth. Instruments and Data Collection 2. Participation stage: the student teacher spends two weeks (50 hours) practicing educational tasks and activities inside and outside the classroom. For example, he/she is advised to participate in school committees along with their activities, write and conduct lesson plans, prepare worksheets, take class attendance, etc. Such participation can be performed independently or with the help of the cooperative teacher. At the beginning of this stage, the university practicum supervisor visits the student to give guidelines and orientation that can help the student improve the educational practices and engage in the school’s activities. 2. Participation stage: the student teacher spends two weeks (50 hours) practicing educational tasks and activities inside and outside the classroom. For example, he/she is advised to participate in school committees along with their activities, write and conduct lesson plans, prepare worksheets, take class attendance, etc. Such participation can be performed independently or with the help of the cooperative teacher. At the beginning of this stage, the university practicum supervisor visits the student to give guidelines and orientation that can help the student improve the educational practices and engage in the school’s activities. To achieve the objectives of the study, a mixed research method was employed using multiple sources of data collection including a questionnaire, semi-structured focus-group and phone interviews via WhatsApp. The study was conducted during the first semester of the academic year 2017 /2018. A 28- item questionnaire was used that was adapted and developed based on several previous studies. The reliability of the questionnaire was calculated through Cornbach Alpha formula and it was (0.88.8), which is acceptable for the purpose of the study. In addition to the 28 items, the questionnaire consisted of two open-ended questions that requested students to express additional comments or ideas related to WhatsApp use. The first question was “Do you think that WhatsApp can be used to help you succeed in practicum? How?” While the second was, “Do you have other comments / thoughts regarding using WhatsApp as an educational tool to support learning of practicum? 3. Actual practice: This is the most important stage as it consists of the actual practices of teaching, because at this final stage student teachers are expected to be familiar with the multi tasks of a classroom teacher. Results and Discussion This study aimed at investigating the perceptions of EFL students towards WhatsApp and its potential benefits in ELT Practicum at Al-Quds Open University. To achieve these objectives, the quantitative results will be presented first, followed by the qualitative results obtained from the focus group session and personal interviews. Firstly, the quantitative data were analyzed in accordance with the study questions as follows: ♦♦ Results related to the First Question. What are the perceptions of EFL students towards WhatsApp in practicum? To answer this question, means and standard deviations were used and the results are shown in Table (1). ♦♦ Results related to the First Question. What are the perceptions of EFL students towards WhatsApp in practicum? To answer this question, means and standard deviations were used and the results are shown in Table (1). Context and Participants The context of the present study was Al- Quds Open University (QOU), which has 21 branches in Palestine, including Nablus Branch. The participants were 30 senior English Major Students (28 females and 2 males) who were enrolled in EFL practicum, in Nablus Branch, during the first semester of the academic year 2017-2018. They are all Arabic native students who have studied English as a foreign language for 6 years at governmental schools and nearly 22 Palestinian Journal of Technology & Applied Sciences - No. 2 - January 2019 Two volunteer students helped the researcher register the answers. deviations to answer the main question of the study. For analytical purposes, the following scale was used to represent the estimation level of students’ responses. Two days later, seven personal interviews were conducted using WhatsApp voice calls. The interviews were with a randomly selected sub- sample of students chosen from the participants, in order to enrich the quantitative and qualitative data. The questions were predetermined and directed to them after they took the same scale. They were interviewed about the feelings they experienced during the practicum procedures. It is worth mentioning here that both the focus group session and the interviews were conducted in the students’ native language, Arabic, to give the students more freedom and more comfort to express their feelings and opinions. 4.5 and above: Very High, 4 – 4.49: High, 3.50 – 3.99: Medium, 3 - 3.49: Low less than 3 : Very Low Very Low Data Analysis The data collected were analyzed using (SPSS) in the form of means and standard Table (1): 24 WhatsApp enabled me to express my opinions and thoughts more freely than in face- to –face discussions with my supervisor. 4.33 0.92 High 21. 25 WhatsApp helped me to learn collaboratively with those who have similar courses. 4.30 0.79 High 22. 13 WhatsApp enabled me to communicate with my classmates about practicum requirements. 4.30 0.65 High 23. 15 WhatsApp enabled me to discuss ideas about the course with my classmates. 4.20 0.76 High 24. 21 WhatsApp helped me to increase my confidence level. 4.16 0.74 High 25. 16 WhatsApp enabled me to seek help from students who are taking practicum. 4.10 0.80 High 26. 9 WhatsApp helped me to communicate with the cooperative teacher after school day. 4.06 0.98 High 27. 3 WhatsApp enabled me to share texts and messages with my classmates. 3.96 1.03 Medium 28. 18 I think that using WhatsApp in practicum caused some social problems. 1.90 0.99 Very Low Total Degree 4.34 0.34 High Rank No. Item Means standard deviations Estimation level 12. 2 WhatsApp enabled me to send videos about my teaching practices to the supervisor. 4.53 0.77 Very High 13. 23 WhatsApp enabled me to debate ideas and exchange opinions with my supervisors and classmates. 4.50 0.68 Very High 14. 20 WhatsApp allowed me to academically engage with my supervisor and classmates anytime any place. 4.46 0.68 High 15. 8 WhatsApp enabled me to get immediate feedback from my supervisor. 4.43 0.67 High 16. 19 WhatsApp enabled me to use live phone and instant message communications with my classmates and the supervisors. 4.40 0.49 High 17. 26 I think that using WhatsApp for learning purposes improved my academic performance. 4.40 0.85 High 18. 14 WhatsApp helped me to receive announcements from the supervisor. 4.36 0.66 High 19. 6 Using WhatsApp reduced my shyness when communicating with my supervisor. 4.33 0.80 High 20. 24 WhatsApp enabled me to express my opinions and thoughts more freely than in face- to –face discussions with my supervisor. 4.33 0.92 High 21. 25 WhatsApp helped me to learn collaboratively with those who have similar courses. 4.30 0.79 High 22. 13 WhatsApp enabled me to communicate with my classmates about practicum requirements. 4.30 0.65 High 23. 15 WhatsApp enabled me to discuss ideas about the course with my classmates. 4.20 0.76 High 24. 21 WhatsApp helped me to increase my confidence level. 4.16 0.74 High 25. Table (1): Means and Standard Deviations of perceptions of EFL students towards WhatsApp Use in practicum in a descending order, according to the means. Rank No. Item Means standard deviations Estimation level 1. 22 WhatsApp allowed me to send and receive images, video, audio and messages to my supervisor at no cost. 4.70 0.53 Very High 2. 1 WhatsApp enabled me to communicate with my supervisor about the practicum requirements. 4.70 0.46 Very High 3. 28 WhatsApp allowed for all time availability of the supervisor to facilitate immediate discussion. 4.66 0.47 Very High 4. 17 I think using WhatsApp in practicum was helpful. 4.66 0.47 Very High 5. 4 WhatsApp helped me to lower my anxiety during interaction with my supervisor. 4.63 0.49 Very High 6. 10 WhatsApp helped me to arrange for the two visits conducted by the supervisor. 4.63 0.49 Very High 7. 11 WhatsApp enabled me to send samples of lessons plans and visual aids done by me. 4.60 0.67 Very High 8. 7 Generally, I would like to use WhatsApp in the educational process. 4.56 0.50 Very High 9. 27 I suggest a WhatsApp group for the practicum course next semester. 4.53 0.81 Very High 10. 12 WhatsApp enabled me to feel comfortable when chatting with my supervisor. 4.53 0.50 Very High 11. 5 WhatsApp helped me to inquire about meetings and assignments. 4.53 0.50 Very High 23 EFL STUDENTS’ PERCEPTIONS OF WHATSAPP AND ITS POTENTIAL BENEFITS IN ELT PRACTICUM Dr. Khaled Abdel Jaleel Dweikat Rank No. Item Means standard deviations Estimation level 12. 2 WhatsApp enabled me to send videos about my teaching practices to the supervisor. 4.53 0.77 Very High 13. 23 WhatsApp enabled me to debate ideas and exchange opinions with my supervisors and classmates. 4.50 0.68 Very High 14. 20 WhatsApp allowed me to academically engage with my supervisor and classmates anytime any place. 4.46 0.68 High 15. 8 WhatsApp enabled me to get immediate feedback from my supervisor. 4.43 0.67 High 16. 19 WhatsApp enabled me to use live phone and instant message communications with my classmates and the supervisors. 4.40 0.49 High 17. 26 I think that using WhatsApp for learning purposes improved my academic performance. 4.40 0.85 High 18. 14 WhatsApp helped me to receive announcements from the supervisor. 4.36 0.66 High 19. 6 Using WhatsApp reduced my shyness when communicating with my supervisor. 4.33 0.80 High 20. Table (1): 16 WhatsApp enabled me to seek help from students who are taking practicum. 4.10 0.80 High 26. 9 WhatsApp helped me to communicate with the cooperative teacher after school day. 4.06 0.98 High 27. 3 WhatsApp enabled me to share texts and messages with my classmates. 3.96 1.03 Medium 28. 18 I think that using WhatsApp in practicum caused some social problems. 1.90 0.99 Very Low Total Degree 4.34 0.34 High Total Degree Table (1) shows that the total degree of students’ perceptions of WhatsApp in practicum was high (4.34). The highest mean was given to the item” WhatsApp allowed me to send and receive images, video, audio and messages from my supervisor at no cost” which scored(4.70). On the other hand, the lowest mean was given to the item “I think that using WhatsApp in practicum caused some social problems” which scored (1.99). The quantitative data analysis above suggests a high level of positive perceptions of WhatsApp after experiencing its potentials in practicum. Furthermore, the overall responses and the high total degree of students’ perceptions reveal that WhatsApp experience significantly impacted the students’ awareness of the benefits of this smartphone application in ELT practicum. This might be due to the fact that students were able to communicate with the supervisor anytime anywhere for educational purposes to ask for support or help. In this respect, the supervisor used to send some encouraging words and expressions at certain times to build a good rapport with the participating students and pave the way for extra evaluative feedback. This feedback greatly helped the students according to their comments and responses in the focus group and the personal interviews. The result regarding students’ positive perceptions of WhatsApp practicum aligns with the findings of other studies (e.g. Gasaymeh, 2017; Malecela, 2016; Ngaleka and Uys, 2013). 24 Palestinian Journal of Technology & Applied Sciences - No. 2 - January 2019 As shown in Table 1, the highest means were given to the items 22, 1, 28, 17, 4, 10, 11, 7, 27, 12, 5, 2 and 23 respectively. These high means indicate that the participants believe that WhatsApp was helpful since it enabled them to send and receive images, video, audio and messages to the supervisor at no cost. They used WhatsApp to communicate with the supervisor regarding the practicum requirements, send samples of lessons plans, visual aids and videos on teaching practices. Table (1): This result is consistent with Mistar & Amin Embi (2016) who indicated that WhatsApp provides various functions, for instance text messages, audio files, attached images, and link to any websites and video files, which can be shared. The result is also consistent with Amshah and Thabian (2017) who found that WhatsApp was the second preferred social media tool among Jordanian students for sending text messages, pictures, music and videos. The result is also in line with Ida Sanjaya (2012) who indicated that the majority of users are interested in using WhatsApp application because it provides facilities such as sharing information in the form of audios, videos, images. , announcements, learning resources, links, worksheets and exams and so on . All of these visual and verbal interactions were practiced beyond the formal educational setting as mentioned previously. For example as shown in their high responses to item number 24 “WhatsApp enabled me to express my opinions and thoughts more freely than in face- to –face discussions with my supervisor”, “WhatsApp encouraged them to participate more freely and more actively to express their ideas and exchange opinions with the supervisor”. However, students’ interaction and communication with the supervisor and the peers as well was made in Arabic as it was somehow difficult for them to use English, as their English proficiency was quite limited outside the university walls. The result, however, is not in line with So (2016) who found that university students who had experience with the formal use of WhatsApp to support their learning, said that the use of WhatsApp might interfere with their private lives. ♦♦ Results related to the second question: What are the benefits gained by students of ELT practicum by using WhatsApp? Students answered this question via the focus group session and the personal interviews through WhatsApp voice call. Furthermore, the researcher examined the types of materials sent by the students to the supervisor (the researcher himself) via WhatsApp messenger during the application of the study to conclude the benefits. ♦♦ Results related to the second question: What are the benefits gained by students of ELT practicum by using WhatsApp? Students answered this question via the focus group session and the personal interviews through WhatsApp voice call. Qualitative results Table (2): In addition to the 28- item questionnaire, a semi-structured focus group and phone interviews via WhatsApp were used to collect some qualitative data that complement the quantitative data. Overall, students’ responses to the interviews and the focus group questions support the results of the questionnaire analysis. Their responses showed that the use of WhatsApp was received positively by seven female students who were interviewed via the WhatsApp call service. One student, for example mentioned that WhatsApp greatly helped her to improve her self- confidence especially when communicating with the supervisor .This student added that it was possible for her to call the supervisor nearly for free to inquire about some information, mainly the dates of the two visits to the cooperative school. Another student who lives far from the university said that using WhatsApp saved time, effort and money when she was able to communicate with the supervisor and with her classmates. She asked for information and inquired about the best way to teach a particular lesson from the textbook “English for Palestine.” Table (2): Types of materials sent via WhatsApp and their percentages. Type of material Number Percentage Visual aids 224 29.78 Text messages 181 24.07 Pictures 79 10.50 Videos 35 4.65 Lesson plans 42 5.59 Sample extracts of teaching practices 43 5.72 Worksheets/ exams. 28 3.72 Voice calls 63 8.38 Miscellaneous 57 7.58 Total 752 100 ypes of materials sent via WhatsApp and their percentages Type of material Number Percentage Table (2) reveals that the students used the WhatsApp application to send different types of materials and activities when communicated with the supervisor. The total number of these activities was 752. The highest number and percentage was the visual aids either made or used by the students in their practicum. Students in this regard sent 224 samples of visual aids that they used within the 30 days of their practicum. Such high percentage indicates that these students were active users of WhatsApp and they were interested in using visual aids to enhance pupils’ learning. The lowest number and percentage was that of worksheets and exams .This low percentage might be due to the fact that most students taught young students from 1st to 4th grade where exams are rare and too many worksheets are not approved by the Ministry of Education nowadays. Table (1): The researcher archived and stored in files all the text messages, verbal and visual materials for later classification and analysis. For the preliminary analysis of the data, all the data were read through to obtain a general sense of the information and to understand what kind of benefits emerged through using WhatsApp. During this process, students’ messages were analyzed by using quantitative and qualitative techniques. Consequently, categories were generated as shown in Table (2) below: school. The researcher archived and stored in files all the text messages, verbal and visual materials for later classification and analysis. For the preliminary analysis of the data, all the data were read through to obtain a general sense of the information and to understand what kind of benefits emerged through using WhatsApp. During this process, students’ messages were analyzed by using quantitative and qualitative techniques. Consequently, categories were generated as shown in Table (2) below: Table (1): Furthermore, the researcher examined the types of materials sent by the students to the supervisor (the researcher himself) via WhatsApp messenger during the application of the study to conclude the benefits. One of the major benefits revealed by the students was their willingness and motivation to collaborate and communicate with the instructor /supervisor in formal & informal way unlike the traditional pedagogic limits and constraints of a typical classroom. This benefit created an intrinsic motivation within students who felt that they were guided and supported by the supervisor, most of the time, through WhatsApp, besides face-to– face meetings and the two visits of the practicum. Furthermore, WhatsApp empowered students and allowed them to interact with their peers in the process of practicum learning and to provide each other assistance, guidance, support and feedback when necessary. This complementary role seems to be in line with Sharples, & Vavoula,( 2007) who claimed that Smartphones provide students with the ability to learn outside a classroom environment where the students is the center of learning. All students reported that WhatsApp enabled them to send videos, pictures and photos , messages, inquiries, samples of teaching practices, links, lesson plans , worksheets, etc. The quantitative analysis of data revealed that the highest mean of students’ responses was the item number 22 “ WhatsApp allowed me to send and receive images, video, audio and messages to my supervisor at no cost “ which scored (4.70). The material and the activities sent to the researcher’s WhatsApp were examined in terms of quality and quantity. Data were collected from students’ messages in these accounts over a period of three months which covered the 30 required days of practicum application at the collaborating Through WhatsApp application, students were able to exchange with the supervisor and with their classmates a great number of images, instant picture, videos, comments, feedback 25 EFL STUDENTS’ PERCEPTIONS OF WHATSAPP AND ITS POTENTIAL BENEFITS IN ELT PRACTICUM Dr. Khaled Abdel Jaleel Dweikat results achieved through the focus group session and the personal interviews conducted via the WhatsApp voice call. All in all , these results are in line with those of Amshah and Thabian (2017, Mistar & Amin Embi (2016), Ida Sanjaya (2012) who stated that WhatsApp provides various functions, for instance text messages, audio files, attached images, link to any websites and video files which can be shared. school. Qualitative results Such results are in line with the quantitative results achieved through the questionnaire and the qualitative The focus group data showed that all students stated that WhatsApp increased the level of their interaction with the supervisor which helped to break the fear barrier and helped the student to be more comfortable even in the presence of the supervisor at the school. Their overall responses and comments indicated that the students felt that WhatsApp helped them to be more engaged and active in practicum. Second, the students reported that this experience gave them a chance to exchange ideas, lesson plans, videos, English songs related to the EFP textbook and visual aids, thus enabling them to teach better. The overall responses in the focus group indicated that the students were 26 Palestinian Journal of Technology & Applied Sciences - No. 2 - January 2019 mostly enthusiastic towards using WhatsApp in their practicum course and this enthusiasm contributed to their successful collaboration and communication with the supervisor and their classmates as well. ideas via various features of WhatsApp platform such as pictures, videos, web-links, recorded videos and many more. It also helped the students to be actively engaged in e-learning activities via the various features of this application. In light of the study results, the following recommendations are proposed: The interviews conducted via WhatsApp calling service revealed that the students approved using a WhatsApp closed group next time so as to use it for posting materials and information about the practicum course. This group should involve students in previous practicum courses and recent students to exchange ideas and experiences. The supervisor himself should administer this group, according to most students, while two students stated that one or two students can help the supervisor to be joint-admins. In this respect, three students reported that WhatsApp groups are more convenient and less troubling than Facebook groups and helps students to benefit more. 1. Faculty members and university supervisors should be encouraged to utilize WhatsApp group in practicum to enable students to use it for educational purposes via their smartphones. 2. Al-Quds Open University should organize training programs, workshops, and online training to train faculty members on how to utilize WhatsApp in blended learning. 3. Al-Quds Open University should organize training programs and workshops for students to train them on how to utilize WhatsApp in their learning. 3. References 1. Ahad, A. D., & Lim, S. M. A. (2014). Convenience or nuisance?: The ‘WhatsApp’ dilemma. Procedia- Social and Behavioral Sciences, 155, 189-196 2. Aicha, B. A. (2014). The impact of WhatsApp mobile social learning on the achievement and attitudes of female students compared with face to face learning in the classroom. European Scientific Journal. Vol.10, No.22. (online) http://www.eujournal.org/index. php/esj/article/viewfile/3909/3700 Qualitative results Some students expressed favorable attitudes towards the use of peer-centered interaction and collaboration with their classmates as well as with other students who took the practicum course in previous courses. Such comments go in line with the items number 3, 13, 15, 16, 18, 23 , 25. These items received high means of estimated levels of perceptions among students. One student said in the focus group, “WhatsApp enabled us to support each other when we exchanged information, opinions and practicum related resources including lesson plans, worksheets and visual aids related to “English for Palestine” textbooks.” A second student responded enthusiastically by saying “I benefited a lot from the features of WhatsApp especially the use of free voice calls, which enabled me to chat with my classmates for hours during the application stages of practicum and after”. 4. Further studies may add a few aspects such as skills, knowledge and problems faced by the students in using WhatsApp application. 5. Further research should include cooperative teachers & supervisors’ perceptions of WhatsApp and its benefits in practicum of different specializations and courses at Al- Quds Open University. Conclusion and recommendations The quantitative and qualitative results of this study revealed that WhatsApp was helpful in ELT practicum and enabled the students to accomplish educational outcomes and the requirements of the course in several ways. WhatsApp provided faster and easier communication among students and with the supervisor, and promoted sharing ideas. It allowed the students to express thoughts and 3. Aifan , Hanan Ahmad . (2015). Saudi Students ’ Atittudes towards Using Social Media to Support Learning..P hd. Dissertation. King Abdu-Aziz University, Jeddah. 27 EFL STUDENTS’ PERCEPTIONS OF WHATSAPP AND ITS POTENTIAL BENEFITS IN ELT PRACTICUM Dr. Khaled Abdel Jaleel Dweikat ISSN 1692-5777. No.12. (January - June) 2016. pp. 29-50. 4. Amshah and Thabian (2017). The use of Jordanian University Students for Richness of Innovation of Interactive Communication Networks and Gratification Achieved. .Journal of Al-Quds Open University For Research and Studies. Vol 2, No 42 (2017) 13. Ida Sanjaya. (2012). Pemanfaatan “WhatsApp Messenger” Sebagai Media Komunikasi Pada Remaja Akhir. http://publication.gunadarma. ac.id/bitstreamI. 5. Church, K., & de Oliveira, R. (2013, August). What’s up with whatsapp?: comparing mobile instant messaging behaviors with traditional SMS. In Proceedings of the 15th international conference on Human-computer interaction with mobile devices and services (pp. 352- 361). 14. Kukulska-Hulme, A. (2012), “Mobile- Assisted Language Learning”, in Chapelle, C.A. (Ed.), The Encyclopedia of Applied Linguistics, Blackwell Publishing Ltd, Oxford, UK. 15. Lohnes, S. & Kinzer, C. (2007). Questioning assumptions about students’ expectations for technology in college classrooms. Journal of Online Education, 3(5) Retrieved from http://innovateonline.info/index. php?view=article&id=431 6. Claire O’Malley, Giasemi Vavoula, Jp Glew, Josie Taylor, Mike Sharples, et al..(2005). Guidelines for learning/ teaching /tutoring in a mobile environment. Public deliverable from the MOBILearn project (D.4.1) 16. Mac Callum, K., & Jeffrey, L., 2014. Comparing the role of ICT literacy and anxiety in the adoption of mobile learning. Computers in Human Behavior, Vol. 39, pp. 8-19. 7. Constine , Josh .(2018). WhatsApp hits 1.5 billion monthly users. $19B? Not so bad. TechCrunch . Retrieved from : https:// techcrunch.com/2018/01/31 8. Definition of 3G in Collins English Dictionary .com . Retrieved February 23 , 2018, from https://www.collinsdictionary. com/dictionary/english/3g 17. Malecela, I. O. (2016). Usage of Whatsapp among Postgraduate Students of Kulliyyah of Education, International Islamic University Malaysia. International Journal of Advanced Engineering Research and Science. 9. Devi, T. S., & Tevera, S. (2014). Use of social networking site in the University of Swaziland by the health science students: A Case study. Conclusion and recommendations Jounal of Information Management, 1(1), 19-26. 18. Martin, F., & Ertzberger, J. (2013). Here and now mobile learning: An experimental study on the use of mobile technology. Computers & Education, 68, 76-85. http://dx.doi. org/10.1016/j. compedu.2013.04.021 10. Gachago, D., Strydom, S., Hanekom, P., Simons, S., & Walters, S. (2015). Lectures’ perspectives on the use of WhatsApp to support teaching and learning in higher education. Progressio, 37(1), 172-187. 19. Meishar-Tal , Hagit and Ronen , Miky .(2016) . Experiencing A Mobile Game and its Impact on Teachers’ Attitudes Towards Mobile Learning. 12th International Conference Mobile Learning. 11. Gasaymeh , Al-Mothana M.(2017). University Students’ use of Whatsapp and their Perceptions Regarding its Possible Integration into their Education. Global Journal of Computer Science and Technology: G Interdisciplinary ,Volume 17 Issue 1 Version 1. 20. Minalla , Amir Abdalla. (2018). The Effect of WhatsApp Chat Group in Enhancing EFL Learners’ Verbal Interaction outside Classroom Contexts. English Language Teaching; Vol. 11, No. 3 21. Mistar , Izyani binti and Amin Embi , Mohamed. (2016). Students ’ Perception on the Use of WhatsAPP as a Learning Tool in ESL Classroom. Journal of Education and Social Sciences, Vol. 4, (June)ISSN 2289- 12. Han , Turgay and Keskin, Fırat .(2016). Using a Mobile Application (WhatsApp) to Reduce EFL Speaking Anxiety. Gist Education and Learning Research Journal. 28 Palestinian Journal of Technology & Applied Sciences - No. 2 - January 2019 9855. the perception of undergraduates on the use of WhatsApp as a tool for Mandarin Language Teaching and Learning : A case in UNiKL RCMP. Conference: Language Education and Civilisation International Conference (LECIC, , At Ipoh, Malaysia. 22. Ngaleka, A., & Uys, W. (2013, June). M-learning with whatsapp: A conversation analysis. In International Conference on e-Learning (p. 282). Academic Conferences International Limited. 31. Sharples, M., Taylor, J., & Vavoula, G. (2007). A Theory of Learning for the Mobile Age. In C. Andrews & R. Haythornthwaite (Eds.), The Sage Handbook of Elearning Research 23. Norfaezah Mohd Hamidin. (2015). Penggunaan Aplikasi ‘WhatsApp’ Dalam Pembelajaran & Pengajaran (P&P) di KUIS. 1st Global Conference on Technology in Language Learning 20015 32. So, S. (2016). Mobile instant messaging support for teaching and learning in higher education. The Internet and Higher Education, 31, 32-42. 24. Obari, H., & Lambacher, S. (2015). Successful EFL teaching using mobile technologies in a flipped classroom. In F. Helm, L. Bradley, M. Guarda, & S. Conclusion and recommendations Thouësny (Eds), Critical CALL – Proceedings of the 2015 EUROCALL Conference, Padova, Italy (pp. 433-438). Dublin: Research-publishing.net. 33. Stephan Böhm, Georges Philip Constantine. (2016) “Impact of contextuality on mobile learning acceptance: An empirical study based on a language learning app”, Interactive Technology and Smart Education, Vol. 13 Issue: 2, pp.107-122, https://doi.org/10.1108/ ITSE-02-2016-0003. 25. Obari, H., & Lambacher, S. (2014). Impact of a blended environment with m-learning on EFL skills. In S. Jager, L. Bradley, E. J. Meima, & S. Thouësny (Eds.), CALL Design: Principles and Practice - Proceedings of the 2014 EUROCALL Conference, Groningen, The Netherlands (pp. 267-272). Dublin Ireland: Research-publishing.net. doi:10.14705/rpnet.2014.000229. 34. Sweeney, P. and Moore, C. (2012), “Mobile Apps for Learning Vocabulary”, International Journal of Computer-Assisted Language Learning and Teaching, Vol. 2 No. 4, pp. 1–16. 26. Patient, R., Crispen, C. (2011). Using Mobile Devices to Leverage Student Access tp Collaboratively-generated resources: A Case of WhatsApp Instant Messaging, South Africa University. (online) http://versys. uitm.edu.my/prisma/view/viewPDF. 35. Tang, Y., & Hew, K. F. (2017). Is mobile instant messaging (MIM) useful in education? Examining its technological, pedagogical, and social affordances. Educational Research Review. Retrieved June 8, 2017, from http://www.direct.com/science/article/pii/ S0747563216305039 27. Ramakrishnan , N. and Johnsi Priya .J. (2016). Mobile Chatting Behaviour of Arts and Science College Students. International Journal of Research – Granthaalayah, Vol. 4, No. 7: SE (2016): 32-39. 36. Traxler, J. (2013). Mobile learning: Shaping the frontiers of learning technologies in global context. In Reshaping Learning (pp. 237-251). Springer Berlin Heidelberg. 28. Riyanto, A. (2013). English Language Learning Using ‘WhatsApp Application’. Retrieved June 28, 2014, from http:// a k h m a d r i y a n t o b l o g . w o r d p r e s s . com/2013/07/21/ 37. Yeboah, J., & Ewur, G. D. (2014). The impact of whatsApp messenger usage on students performance in Tertiary Institutions in Ghana. Journal of Education and practice, 5(6), 157-164. 29. Roffe, I. (2002). E-learning: Engagement, enhancement and execution. Quality Assurance in Education, 10(1), 40-50. 38. Zayed, N. (2016). Special Designed Activities for Learning English Language through the Application of WhatsApp. https://doi. org/10.5539/elt.v9n2p199 30. Selva Kumar, Vishalini. (2016). A study on 29
https://openalex.org/W2910163675	https://peerj.com/articles/6211.pdf	English	null	Outcomes of hospitalizations with atrial fibrillation-flutter on a weekday versus weekend: an analysis from a 2014 nationwide inpatient sample	PeerJ	2,019	cc-by	5,757	Outcomes of hospitalizations with atrial fibrillation-flutter on a weekday versus weekend: an analysis from a 2014 nationwide inpatient sample Dinesh C. Voruganti1, Ghanshyam Palamaner Subash Shantha2, Abhishek Deshmukh3 and Michael C. Giudici4 1 Division of Internal Medicine, Roy and Lucille J. Carver College of Medicine, University of Iowa Hospitals and Clinics, Iowa City, IA, United States of America 2 Division of Cardiovascular Medicine, University of Michigan - Ann Arbor, Ann Arbor, MI, United States of America vision of Cardiovascular Medicine, Mayo Clinic, Rochester, MN, United States of America 4 Division of Cardiovascular Medicine, Roy and Lucille J. Carver College of Medicine, University of Iowa Hospitals and Clinics, Iowa City, IA, United States of America 4 Division of Cardiovascular Medicine, Roy and Lucille J. Carver College of Medicine, University of Iowa Hospitals and Clinics, Iowa City, IA, United States of America 4 Division of Cardiovascular Medicine, Roy and Lucille J. Carver College of Medicine, University of Iowa Hospitals and Clinics, Iowa City, IA, United States of America How to cite this article Voruganti DC, Subash Shantha GP, Deshmukh A, Giudici MC. 2019. Outcomes of hospitalizations with atrial fibrillation-flutter on a weekday versus weekend: an analysis from a 2014 nationwide inpatient sample. PeerJ 7:e6211 http://doi.org/10.7717/peerj.6211 ABSTRACT Background. Patients with atrial fibrillation-flutter (AF) admitted on the weekends were initially reported to have poor outcomes. The primary purpose of this study is to re-evaluate the outcomes for weekend versus weekday AF hospitalization using the 2014 Nationwide Inpatient Sample (NIS). Methods. Included hospitalizations were aged above 18 years. The hospitalizations with AF were identified using the international classification of diseases 9 (ICD-9) codes (427.31, 427.32). In-hospital mortality, length of stay (LOS), other co-morbidities, cardioversion procedures, and time to cardioversion were recorded. All analysis was performed using SAS 9.4 statistical software (Cary, North Carolina). Methods. Included hospitalizations were aged above 18 years. The hospitalizations with AF were identified using the international classification of diseases 9 (ICD-9) codes (427.31, 427.32). In-hospital mortality, length of stay (LOS), other co-morbidities, cardioversion procedures, and time to cardioversion were recorded. All analysis was performed using SAS 9.4 statistical software (Cary, North Carolina). p g y Results. A total of 453,505 hospitalizations with atrial fibrillation and flutter as primary discharge diagnosis were identified. Among the total hospitalizations with a primary diagnosis of AF, 20.3% were admitted on the weekend. Among the weekend hospitalizations, 0.19% died in hospital compared to 0.74% among those admitted during the week. After adjusting for patient characteristics, hospital characteristics and disease severity, the adjusted odds for in-hospital mortality were not significantly different for weekend vs. weekday hospitalizations (OR = 0.91, 95% CI [0.77–1.11]; p = 0.33). The weekend admissions were associated with significantly lower odds of cardioversion procedures (OR = 0.72, 95% CI [0.69–0.76], P < 0.0001), lower cost of hospitalization (USD 8265.8 on weekends vs. USD 8966.5 on the weekdays, P < 0.001), slightly lower rate of anticoagulation (17.09% on the weekends vs. 18.73% on the weekdays. P < 0.0001), and slightly increased time to cardioversion (1.94 days on the weekend vs. 1.73 days on weekdays, P < 0.0005). The mean length of hospital stay (LOS) was statistically not different in both groups: (3.49 days ± 3.70 (SD) in the weekend group vs. 3.47 days ± 3.50 (SD) in the weekday group, P = 0.42) Submitted 14 August 2018 Accepted 4 December 2018 Published 17 January 2019 Corresponding author Dinesh C. Voruganti, dinesh-voruganti@uiowa.edu Academic editor Dennis Lau Additional Information and Declarations can be found on page 9 DOI 10.7717/peerj.6211 Copyright 2019 Voruganti et al. ABSTRACT Distributed under Creative Commons CC-BY 4.0 OPEN ACCESS Submitted 14 August 2018 Accepted 4 December 2018 Published 17 January 2019 Corresponding author Dinesh C. Voruganti, dinesh-voruganti@uiowa.edu Academic editor Dennis Lau Additional Information and Declarations can be found on page 9 DOI 10.7717/peerj.6211 Copyright 2019 Voruganti et al Discussion. The weekend AF hospitalizations did not have a clinically significant difference in mortality and LOS compared to those admitted on a weekday. However, the use of cardioversion procedures and cost of hospitalization was significantly lower on the weekends. Distributed under Creative Commons CC-BY 4.0 OPEN ACCESS Subjects Cardiology, Global Health, Health Policy Keywords Atrial fibrillation, Weekend, Cardioversion, Anticoagulation, In-hospital mortality, Time to cardioversion INTRODUCTION Atrial fibrillation is the most common persistent cardiac arrhythmia in clinical practice which had an estimated worldwide prevalence of 33.5 million in 2010 (Markides & Schilling, 2003). Patients with atrial fibrillation-flutter (AF) hospitalized on weekends were previously reported to have higher mortality and increased the length of hospital stay (Deshmukh et al., 2012; Shawn Lee et al., 0000). The factors thought to be contributing to the poor outcomes are: limited availability of the staff and access to procedures such as cardioversion. Similar studies on acute myocardial infarction have demonstrated higher inpatient mortality for weekend hospitalizations (De Cordova et al., 2017). A similar analysis on AF weekend hospitalizations reported improved mortality, (Weeda et al., 2016) but the improvement in these outcomes were consistently not replicated by studies performed on a large database on inpatient hospitalizations. We sought to investigate the outcomes in the year 2014 through publically available nationwide inpatient sample database (NIS) to assess the outcomes (in-hospital mortality, rates of cardioversion and time to cardioversion) and to determine whether such differences resulted from the disparities in the utilization and timing of cardioversion. Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 METHODS The NIS is a part of the Healthcare Cost and Utilization Project (HCUP) which is sponsored by the Agency for Healthcare Research and Quality (AHRQ) (HCUP, 2012). Each year of the NIS records over 7 million inpatient hospitalizations. The NIS is one of the largest all-payer databases of hospital inpatient stays available in the United States of America (USA). The 2014 (1st January to 31st December) NIS sampling frame is comprised of 44 States and the District of Columbia, covering more than 96 percent of the U.S.A population and including more than 94 percent of discharges from the USA community hospitals. Our main interest group was the hospitalizations who had a primary diagnosis (dx1) of atrial fibrillation or flutter. All hospitalizations with international classification of diseases, 9th revision, code 427.31 or 427.32 as the principal diagnosis were included. Per AHRQ-HCUP, the weekend admissions were defined as admissions on Saturday–Sunday (Ananthakrishnan, McGinley & Saeian, 2009). The use of cardioversion is denoted by the presence of ≥1 of the following ICD 9 procedural codes in any position: 99.61, 99.62 and 99.69. Our primary outcome was in-hospital mortality, and the secondary outcomes included utilization of cardioversion procedures, length of hospital stay (LOS), time to cardioversion, anticoagulation and total hospitalization charges. The study protocol was reviewed by the University of Iowa, Iowa City, Institutional Review Board (IRB) (review: 201806023), and the study was exempt from human subject research as it includes only de-identified, publically available data. All analyses Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 2/11 were performed using SAS, version 9.4 (SAS Institute, Cary, North Carolina). Survey procedures available within the SAS were applied in the analysis to account for design features of the complex sample survey. Descriptive statistics were generated for the individual and hospital characteristics for both weekend and weekday admissions. Univariate tests were applied to compare the equality of the mean or proportions for the motioned outcomes between the weekday and weekend admissions, which consisted of the Rao-Scott chi-square test for categorical outcomes and t-tests (along with standard deviation (SD)) for continuous outcomes. The cost of hospitalization was calculated from cost to charge ratio files and the total charges provided by the AHRQ (https://www.hcup-us.ahrq.gov/db/state/costtocharge.jsp). The number of chronic conditions was obtained from ‘NCHRONIC’ variable listed in the NIS database. The data element ‘NCHRONIC’ contains the count of unique chronic diagnoses reported on the discharge. RESULTS We identified a national estimate of 453,505 hospitalizations with AF as the primary diagnosis. Of these, 92,220 were characterized as weekend hospitalizations and 361,285 as weekday hospitalizations. The mean age among weekend and weekday was statistically not different (weekday 70.1 years ± 13.5 (SD) and weekend 70.2 years ± 14.2 (SD)) with a P-value of 0.53 (T-test). The proportion of males was somewhat lower in the weekend group (48.13% on weekends vs. 50.53% on the weekdays, P value < 0.0001). A slightly lower proportion of the white population was hospitalized over the weekends (80.31% on the weekends vs. 82.60% on the weekday), and a higher proportion of the Hispanic population was admitted on the weekends (6.25 on weekends vs. 5.34 on the weekdays, P < 0.0001). Hospitalizations with Medicare constituted the majority of overall hospitalizations (67.13%) for AF. AF hospitalizations were relatively higher in the Urban teaching hospitals (59.61%), and the weekday hospitalizations were higher in the urban teaching hospitals vs. the weekend (60.12 on the weekday vs. 57.63% on the weekends, P < 0.0001). The same pattern was observed in the large hospitals which constituted about 50.64% of total AF hospitalizations, and a slightly higher rate of AF hospitalizations in the large hospitals was on the weekdays (51.08% on the weekday vs. 48.93% on the weekends, P < 0.0001). Table 1 summarizes the demographics and baseline characteristics for the weekday and weekend hospitalizations along with the P values (T-test for continuous variables and Chi-square test for the categorical variables). Comparing the in-hospital mortality (primary outcome) in two groups, we have identified that the mortality for weekend hospitalizations did not significantly vary from the weekday hospitalizations (0.19% on the weekends vs. 0.74% on the weekdays, P = 0.90). Secondary outcomes were the number of inpatient cardioversion procedures, interval to the procedure (time to cardioversion), length of stay, anticoagulation and the cost of Comparing the in-hospital mortality (primary outcome) in two groups, we have identified that the mortality for weekend hospitalizations did not significantly vary from the weekday hospitalizations (0.19% on the weekends vs. 0.74% on the weekdays, P = 0.90). y p g y y the weekday hospitalizations (0.19% on the weekends vs. 0.74% on the weekdays, P = 0.90). Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 METHODS The long-term (current) use of anticoagulants was determined using the ICD-9 CM code ‘V58.61’. The national estimates for hospitalization were calculated by applying the weights provided by the HCUP-AHRQ in the NIS file. Finally, multivariate logistic regression models were applied to test the adjusted associations between the outcomes of weekend versus weekday admissions. The level of significance (α) was chosen as 5%. RESULTS Secondary outcomes were the number of inpatient cardioversion procedures, interval to the procedure (time to cardioversion), length of stay, anticoagulation and the cost of Secondary outcomes were the number of inpatient cardioversion procedures, interval to the procedure (time to cardioversion), length of stay, anticoagulation and the cost of Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 3/11 Table 1 Baseline characteristics of atrial fibrillation (AF) hospitalizations. The baseline characteristics indicate the percentage of AF hospitaliza- tions admitted on the weekends and the weekdays. The total hospitalizations include both groups. The P value indicates the chi-square test for the baseline characteristic differences among the weekday and the weekend groups. Table 1 Baseline characteristics of atrial fibrillation (AF) hospitalizations. The baseline characteristics indicate the percentage of AF hospitaliza- tions admitted on the weekends and the weekdays. The total hospitalizations include both groups. The P value indicates the chi-square test for the baseline characteristic differences among the weekday and the weekend groups. Table 1 Baseline characteristics of atrial fibrillation (AF) hospitalizations. The baseline characteristics indicate the percentage of AF hospitaliza tions admitted on the weekends and the weekdays. The total hospitalizations include both groups. The P value indicates the chi-square test for the baseline characteristic differences among the weekday and the weekend groups. Characteristic Weekday hospitalization (n = 361,285) Weekend hospitalizations (n = 92,220) Total hospitalizations (n = 453,505) P Value Mean age (years) ± (Standard deviation) 70.1 ± 13.5 70.2 ± 14.2 70.1 ± 13.6 0.5338 Gender <0.0001 Male 50.53% 48.13% 50.0% Female 49.46% 51.86% 49.95% Race <0.0001 White 82.6% 80.31% 82.1% Black 8.18% 9.03% 8.36% Hispanic 5.34% 6.25% 5.53% Asian 1.36% 1.66% 1.42% Native American 0.38% 0.43% 0.39% Other 2.10% 2.30% 2.14% Primary Payer <0.0001 Medicare 66.95% 67.83% 67.13% Medicaid 5.84% 6.41% 5.96% Private 22.61% 20.63% 22.21% Self-pay 2.43% 2.94% 2.54% No charge 0.27% 0.38% 0.29% Other 1.86% 1.78% 1.84% Hospital region <0.0001 Northeast 21.10% 20.07% 20.89% Midwest 24.55% 23.65% 24.37% South 39.71% 40.34% 39.84% West 14.62% 15.92% 14.89% Type of admission <0.0001 Elective 14.34% 4.14% 12.26% Non-Elective 85.65% 95.85% 87.73% Hospital location/teaching status 11.32% <0.0001 Rural 11.22% 11.72% 29.05% Urban non-teaching 28.65% 30.63% 59.61% Urban teaching 60.12% 57.63% Hospital bed size <0.0001 Small 18.88% 19.60% 19.03% Medium 30.02% 31.45% 30.31% Large 51.08% 48.93% 50.64% CHA2DS2VASc score (mean ± standard deviation) 2.73 ± 1.44 2.79 ± 1.47 2.74 ± 1.45 <0.0001 hospitalization. These characteristics are listed in Table 2. Variable Variable Weekend admission (n = 92,220) Weekday admission (n = 361,285) P-value Cardioversion 2.90% 14.83% P < 0.0001 Mean length of stay (days) ± Standard deviation 3.49 ± 3.70 3.47 ± 3.50 P = 0.4233 In-hospital mortality 0.19% 0.74% P = 0.9058 Mean cost of hospitalization (USD) 8265.8 8966.5 P < 0.001 Time to cardioversion (days) ± (standard deviation) 1.94 ± 2.40 1.73 ± 3.96 P = 0.0005 Anticoagulation 17.09% 18.73% <0.0001 Table 3 Unadjusted estimates for in-hospital mortality for AF hospitalizations. The unadjusted odds ratios indicate the univariate association between the comorbidity listed in the first column to the in- hospital mortality. This indicates the strength of association without adjusting for other variables. Table 3 Unadjusted estimates for in-hospital mortality for AF hospitalizations. The unadjusted odds ratios indicate the univariate association between the comorbidity listed in the first column to the in- hospital mortality. This indicates the strength of association without adjusting for other variables. Unadjusted odds ratio’s to in-hospital mortality OR 95% Confidence interva Unadjusted odds ratio’s to in-hospital mortality OR 95% Confidence interval P-value Unadjusted odds ratio’s to in-hospital mortality OR 95% Confidence interval P-value Stroke 2.67 2.07 3.45 <0.0001 Hypertension 0.73 0.63 0.85 <0.0001 Anticoagulation 0.52 0.42 0.66 <0.0001 Obesity 0.59 0.48 0.73 <0.0001 Congestive Heart Failure 8.27 5.14 13.32 <0.0001 ≥5 Chronic Conditions 3.12 2.44 3.98 <0.0001 Weekend admission 1.01 0.85 1.19 0.0140 Female 1.24 1.09 1.42 0.0011 73 days on a weekday, P = 0.0005). The weekend AF admission was associated with a 1.73 days on a weekday, P = 0.0005). The weekend AF admission was associated with a y y lower cost of hospitalization (USD 8265.8 on weekends vs. USD 8966.5 on the weekdays, P < 0.001). The weekend hospitalizations had a slightly lower rate of anticoagulation (17.09% on the weekends vs. 18.73% on the weekdays. P < 0.0001). The univariate (Table 3) and the multivariate logistic regression analysis for in-hospital mortality were performed, after adjusting for significant covariates such as age, sex, hypertension, obesity, diabetes mellitus, congestive heart failure, stroke, anticoagulation, length of stay (LOS), primary expected payer, race, hospital location and the teaching status. We observed that the weekend hospitalizations did not have significant difference for in-hospital mortality OR = 0.917 (95% CI [0.77–1.092]; P = 0.3299) (described in Table 4, Fig. 1). Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 RESULTS The univariate (Table 3) and the multivariate logistic regression analysis for in-hospital mortality were performed, after adjusting for significant covariates such as age, sex, hypertension, obesity, diabetes mellitus, congestive heart failure, stroke, anticoagulation, length of stay (LOS), primary expected payer, race, hospital location and the teaching status. We observed that the weekend hospitalizations did not have significant difference for in-hospital mortality OR = 0.917 (95% CI [0.77–1.092]; P = 0.3299) (described in Table 4, Fig. 1). Also, we observed that the hospitalizations with a diagnosis of stroke and presence of 5 or more chronic conditions had the most significant association with in hospital mortality OR= 1 609 (95% CI [1 214 2 132] P = 0 0009) and OR = 1 423 Table 2 Differences between outcomes for the weekday and weekend hospitalizations for Atrial Fibril- lation (AF). The differences in the the weekday and the weekend hospitalizations indicates the percentage of hospitalizations for AF. The P value indicates the differences in these groups after performing the chi- square and t-test. Variable Weekend admission (n = 92,220) Weekday admission (n = 361,285) P-value Cardioversion 2.90% 14.83% P < 0.0001 Mean length of stay (days) ± Standard deviation 3.49 ± 3.70 3.47 ± 3.50 P = 0.4233 In-hospital mortality 0.19% 0.74% P = 0.9058 Mean cost of hospitalization (USD) 8265.8 8966.5 P < 0.001 Time to cardioversion (days) ± (standard deviation) 1.94 ± 2.40 1.73 ± 3.96 P = 0.0005 Anticoagulation 17.09% 18.73% <0.0001 Table 2 Differences between outcomes for the weekday and weekend hospitalizations for Atrial Fibril- lation (AF). The differences in the the weekday and the weekend hospitalizations indicates the percentage of hospitalizations for AF. The P value indicates the differences in these groups after performing the chi- square and t-test. RESULTS We noted that the weekend AF hospitalizations underwent fewer cardioversion procedures than those hospitalized on a weekday (2.90% vs. 14.83%, p < 0.0001). The average time to cardioversion was not very different among both groups, though statistically significant (1.94 days on the weekend vs. Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 4/11 hospitalization. These characteristics are listed in Table 2. We noted that the weekend AF hospitalization. These characteristics are listed in Table 2. We noted that the weekend AF hospitalizations underwent fewer cardioversion procedures than those hospitalized on a weekday (2.90% vs. 14.83%, p < 0.0001). The average time to cardioversion was not very different among both groups, though statistically significant (1.94 days on the weekend vs. Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 4/11 Table 2 Differences between outcomes for the weekday and weekend hospitalizations for Atrial Fibril- lation (AF). The differences in the the weekday and the weekend hospitalizations indicates the percentage of hospitalizations for AF. The P value indicates the differences in these groups after performing the chi- square and t-test. Variable Weekend admission (n = 92,220) Weekday admission (n = 361,285) P-value Cardioversion 2.90% 14.83% P < 0.0001 Mean length of stay (days) ± Standard deviation 3.49 ± 3.70 3.47 ± 3.50 P = 0.4233 In-hospital mortality 0.19% 0.74% P = 0.9058 Mean cost of hospitalization (USD) 8265.8 8966.5 P < 0.001 Time to cardioversion (days) ± (standard deviation) 1.94 ± 2.40 1.73 ± 3.96 P = 0.0005 Anticoagulation 17.09% 18.73% <0.0001 Table 3 Unadjusted estimates for in-hospital mortality for AF hospitalizations. The unadjusted odds ratios indicate the univariate association between the comorbidity listed in the first column to the in- hospital mortality. This indicates the strength of association without adjusting for other variables. Unadjusted odds ratio’s to in-hospital mortality OR 95% Confidence interval P-value Stroke 2.67 2.07 3.45 <0.0001 Hypertension 0.73 0.63 0.85 <0.0001 Anticoagulation 0.52 0.42 0.66 <0.0001 Obesity 0.59 0.48 0.73 <0.0001 Congestive Heart Failure 8.27 5.14 13.32 <0.0001 ≥5 Chronic Conditions 3.12 2.44 3.98 <0.0001 Weekend admission 1.01 0.85 1.19 0.0140 Female 1.24 1.09 1.42 0.0011 1.73 days on a weekday, P = 0.0005). The weekend AF admission was associated with a lower cost of hospitalization (USD 8265.8 on weekends vs. USD 8966.5 on the weekdays, P < 0.001). The weekend hospitalizations had a slightly lower rate of anticoagulation (17.09% on the weekends vs. 18.73% on the weekdays. P < 0.0001). Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 Variable Also, we observed that the hospitalizations with a diagnosis of stroke and presence of 5 or more chronic conditions had the most significant association with in-hospital mortality OR= 1.609 (95% CI [1.214–2.132], P = 0.0009) and OR = 1.423 (95% CI [1.074–1.886], P = 0.014) respectively. Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 5/11 Table 4 Multivariate logistic regression analysis showing the adjusted odds ratio’s predicting the in-hospital mortality for Atrial Fibrillation (AF) hospitalizations. The adjusted odds ratio’s, 95% confidence intervals and their P-values represent the odds of in-hospital mortality after ad- justing for the covariates listed in the table. Table 4 Multivariate logistic regression analysis showing the adjusted odds ratio’s predicting the in-hospital mortality for Atrial Fibrillation (AF) hospitalizations. The adjusted odds ratio’s, 95% confidence intervals and their P-values represent the odds of in-hospital mortality after ad- justing for the covariates listed in the table. justing for the covariates listed in the table. Odds ratio estimates Effect Adjusted odds ratio 95% confidence limits P-value Weekend hospitalization 0.917 0.77 1.092 0.3299 Stroke 1.609 1.214 2.132 0.0009 Hypertension 0.376 0.32 0.441 <.0001 Anticoagulation 0.538 0.427 0.679 <.0001 Obesity 0.529 0.42 0.667 <.0001 Congestive Heart Failure 1.383 0.76 2.516 0.2886 ≥5 Chronic Conditions 1.423 1.074 1.886 0.014 AGE 1.048 1.039 1.057 <.0001 Length of stay 1.059 1.043 1.076 <.0001 EXPECTED PRIMARY PAYER Medicare (Reference group) Medicaid 1.385 0.941 2.039 0.0984 Private insurance 1.124 0.873 1.448 0.3643 Self-pay 1.868 1.065 3.275 0.0292 No charge 1.14 0.172 7.552 0.8918 Other pay 1.978 1.175 3.329 0.0103 Female gender 0.936 0.807 1.086 0.3814 RACE White (Reference group) Black 1.099 0.837 1.441 0.4971 Hispanic 1.273 0.956 1.695 0.0986 Asian or pacific islander 0.869 0.43 1.757 0.6962 Native American 0.702 0.169 2.911 0.6259 Other 0.569 0.282 1.146 0.1144 Elixhauser comorbidity index 1.474 1.412 1.538 <.0001 HOSPITAL LOCATION AND TEACHING STATUS Rural hospital (Reference group) Urban non-teaching hospital 0.832 0.651 1.064 0.143 Urban teaching hospital 0.99 0.788 1.244 0.933 DISCUSSION DISCUSSION The main inferences of our analysis on the NIS 2014 data are: (1) AF weekend hospitalizations showed no clinically significant differences in mortality, length of stay, time to cardioversion; (2) we also found that weekend hospitalizations were less likely to undergo cardioversion, and (3) they had overall lower mean cost of hospitalization. Prior study on AF weekend hospitalization by Deshmukh et al. (2012) reported the adjusted in-hospital mortality to be higher for weekend admissions (OR 1.23, CI [1.03–1.51]), longer weekend length of hospitalization and lower rates of utilization of cardioversion (7.92% weekend vs. 16.2% weekday). Similarly, another study by Weeda et al. (2016) has reported that there were no differences in the adjusted in-hospital mortality The main inferences of our analysis on the NIS 2014 data are: (1) AF weekend hospitalizations showed no clinically significant differences in mortality, length of stay, time to cardioversion; (2) we also found that weekend hospitalizations were less likely to undergo cardioversion, and (3) they had overall lower mean cost of hospitalization. Prior study on AF weekend hospitalization by Deshmukh et al. (2012) reported the adjusted in-hospital mortality to be higher for weekend admissions (OR 1.23, CI [1.03–1.51]), longer weekend length of hospitalization and lower rates of utilization of cardioversion (7.92% weekend vs. 16.2% weekday). Similarly, another study by Weeda et al. (2016) has reported that there were no differences in the adjusted in-hospital mortality Prior study on AF weekend hospitalization by Deshmukh et al. (2012) reported the adjusted in-hospital mortality to be higher for weekend admissions (OR 1.23, CI [1.03–1.51]), longer weekend length of hospitalization and lower rates of utilization of cardioversion (7.92% weekend vs. 16.2% weekday). Similarly, another study by Weeda et al. (2016) has reported that there were no differences in the adjusted in-hospital mortality Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 6/11 Figure 1 Multivariate logistic regression analysis with adjusted odds ratio’s for in-hospital mortality. The adjusted odds ratios, 95% confidence intervals and their P-values represent the odds of in-hospital mortality after adjusting for the covariates listed in the table. The blue dots indicate the adjusted odds ra- tio for the listed variable, and the red lines indicate 95% confidence intervals. OR to the right of midline (where OR = 1) indicate higher odds of in-hospital mortality while OR to the left of the midline indicate lower odds of in-hospital mortality. Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 DISCUSSION Full-size DOI: 10.7717/peerj.6211/fig-1 Figure 1 Multivariate logistic regression analysis with adjusted odds ratio’s for in-hospital mortality. The adjusted odds ratios, 95% confidence intervals and their P-values represent the odds of in-hospital mortality after adjusting for the covariates listed in the table. The blue dots indicate the adjusted odds ra- tio for the listed variable, and the red lines indicate 95% confidence intervals. OR to the right of midline (where OR = 1) indicate higher odds of in-hospital mortality while OR to the left of the midline indicate lower odds of in-hospital mortality. Full-size DOI: 10.7717/peerj.6211/fig-1 rate (OR 1.02; 95% CI [0.94–1.11]) and the length of hospitalization. However, they were found to have longer time-to-procedure and lesser treatment costs with weekend admissions. In comparison to the prior studies, our results match Deshmukh et al. (2012) where the utilization of cardioversion and cost of hospitalization was lower in weekend AF hospitalizations. On the other hand, our analysis also matches Weeda et al. (2016) in demonstrating a lack of significant difference in mortality and the length of hospitalization stay among both groups comparing weekend and weekday admissions. The outcomes on the weekend are informally referred to as the ‘weekend effect’. It is a phenomenon which is often highlighted to associate poor outcomes in weekend hospitalizations. It has been proposed that it might be a result of lack of healthcare management organizations to improve practices of care, which includes ensuring round the clock accessibility to life- saving procedures (Mathew et al., 2018). Also, published meta-analysis has associated poor outcomes in patients admitted with myocardial infarction and other medical conditions (Sorita et al., 2014). In patients with AF, the higher mortality, length of stay and lower utilization of cardioversion procedures were thought to be secondary to limited availability of services on the weekends. It was proposed that the subtle primary signs of acute problems Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 7/11 go unnoticed until later on the weekends. Studies focused on the weekend effects, in general, have emphasized for a better-organized model of care which could help in bridging the gap of the weekend effect. We notice an improvement of outcomes in hospitalizations with AF. The difference of in-hospital mortality, length of hospitalization, and time to cardioversion has been gradually decreasing since Deshmukh et al.’s (2012) publication. DISCUSSION These changes might occur due to the implementation of robust patient care across the hospitals in the United States to provide 24/7 accessibility to procedures such as cardioversion and prompt recognition of subtle clinical parameters such as atrial fibrillation with rapid ventricular response with improvised computerized telemetry, resulting in early transfer to the intensive care unit for cardioversion on the weekends (Albright et al., 2009; Conway et al., 2018). However, utilization rates of cardioversion continue to be low among the weekend AF hospitalizations. One of the reasons might be due to limited staff availability and sometimes, delay in identifying subtle signs of acute problems, such as hypotension, may go unnoticed until later. On the other hand, the lack of mortality difference on the weekends with a lower cost of hospitalization and lower rates of utilization of cardioversion procedures also raise a concern about the higher costs of hospitalization on the weekdays. We notice a higher rate of cardioversion procedures on the weekdays which might also sometimes imply overutilization of the procedures, resulting in a higher cost of hospitalization. While opportunities to improve care on the weekends are constantly being explored, a cost- effective management strategy may also be pursued to reduce the costs of hospitalizations on the weekdays. In our analysis, we noted low anticoagulation rates in both the weekend and the weekday groups, which is in concordance with the report from the ‘Get with the Guidelines’, registry that showed similarly low rates (15% to 17%) of anticoagulation in their AF patients (Piccini et al., 2016). The low anticoagulation rates are probably because of high bleeding risk, higher prevalence of contraindications to anticoagulation in AF patients, or due to the coding inconsistencies. Future directions should be focused on improving the utilization rates of cardioversion procedures and assessing the reasons for the disparity between the hospitalization costs associated with weekend hospitalizations. Our study findings provide valuable data demonstrating the improved mortality outcomes and length of hospitalization. Understanding the reasons behind the decreased cost of admission on weekends and reduced utilization of cardioversion procedures might help to bridge the gap difference. Though our study had essential strengths of including a large sample, our study is subject to some limitations. First, the NIS relies on claims data which can incur inaccurate billing and underestimation of covariates of interest, thus leading to coding bias (Yoshihara & Yoneoka, 2014). Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 Author Contributions • Dinesh C. Voruganti conceived and designed the experiments, analyzed the data, contributed reagents/materials/analysis tools, prepared figures and/or tables, approved the final draft. • Ghanshyam Palamaner Subash Shantha conceived and designed the experiments, analyzed the data, contributed reagents/materials/analysis tools, approved the final draft. • Abhishek Deshmukh and Michael C. Giudici contributed reagents/materials/analysis tools, authored or reviewed drafts of the paper, approved the final draft. Ethics The following information was supplied relating to ethical approvals (i.e., approving body and any reference numbers): The following information was supplied relating to ethical approvals (i.e., approving body and any reference numbers): The University of Iowa Human Subjects Office determined that the project described in the application (project number: 201806023) does not meet the regulatory definition of human subjects research and does not require review by the IRB because no personally identifiable information was used. DISCUSSION Missing values in our data prevented us from including specific variables in the multivariate analysis. Furthermore, we did not evaluate the causes that could have accounted for this difference that are not patient related but related to the hospital (e.g., staffing differences on weekends). Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 8/11 ADDITIONAL INFORMATION AND DECLARATIONS The authors received no funding for this work. The authors received no funding for this work. Competing Interests The authors declare there are no competing interests. CONCLUSION In the nationwide US practice, the weekend AF hospitalizations appear to have lower rates of cardioversion utilization and lower hospitalization cost. Further studies are required to identify the differences and explore the opportunities to improve AF weekend care. Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 Supplemental Information Supplemental Information Supplemental information for this article can be found online at http://dx.doi.org/10.7717/ peerj.6211#supplemental-information. Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 Data Availability The following information was supplied regarding data availability: The NIS database raw files were purchased through online HCUP (health care cost and utilization project) distributor. All purchasers and users of HCUP data must complete the online Data Use Agreement (DUA) training so that they are familiar with the rules and restrictions for the use of HCUP data. If this raw data is made public, then the likelihood of maintaining standards to complete the DUA training might be violated. Moreover, anyone who uses the database files we purchased must complete an HCUP Data Use Agreement form (a form which the owner of the data has to generate and acknowledge). Therefore, given these limitations, we have provided raw data for review purposes and not for publishing (i.e., to be shared with the public). Additionally, as per AHRQ-HCUP, before the reviewers/editors may access any HCUP data (the raw files), they are required to complete the 15-minute Data Use Agreement Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 9/11 (DUA) training and read and sign the HCUP DUA. Both of these may be completed online at this link: https://www.hcup-us.ahrq.gov/tech_assist/dua.jsp. (DUA) training and read and sign the HCUP DUA. Both of these may be completed online at this link: https://www.hcup-us.ahrq.gov/tech_assist/dua.jsp. Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 REFERENCES Albright KC, Raman R, Ernstrom K, Hallevi H, Martin-Schild S, Meyer BC, Meyer DM, Morales MM, Grotta JC, Lyden PD, Savitz SI. 2009. Can comprehensive stroke centers erase the ‘weekend effect’? Cerebrovascular Diseases 27:107–113 DOI 10.1159/000177916. Ananthakrishnan AN, McGinley EL, Saeian K. 2009. Outcomes of weekend admissions for upper gastrointestinal hemorrhage: a nationwide analysis. Clinical Gastroenterol- ogy and Hepatology 7:296–302 DOI 10.1016/j.cgh.2008.08.013. Conway R, Cournane S, Byrne D, O’Riordan D, Silke B. 2018. Improved mortality outcomes over time for weekend emergency medical admissions. Irish Journal of Medical Science 187:5–11 DOI 10.1007/s11845-017-1627-7. De Cordova PB, Johansen ML, Martinez ME, Cimiotti JP. 2017. Emergency department weekend presentation and mortality in patients with acute myocardial infarction. Nursing Research 66:20–27 DOI 10.1097/nnr.0000000000000196. Deshmukh A, Pant S, Kumar G, Bursac Z, Paydak H, Mehta JL. 2012. Comparison of outcomes of weekend versus weekday admissions for atrial fibrillation. American Journal of Cardiology 110:208–211 DOI 10.1016/j.amjcard.2012.03.011. HCUP. 2012. HCUP National Inpatient Sample (NIS) Healthcare Cost and Utilization Project (HCUP). 2012. Rockville: Agency for Healthcare Research and Quality. Markides V, Schilling RJ. 2003. Atrial fibrillation: classification, pathophysiology, mechanisms and drug treatment. Heart 89:939–943 DOI 10.1136/heart.89.8.939. Mathew A, Fyyaz SA, Carter PR, Potluri R. 2018. The enigma of the weekend effect. Journal of Thoracic Disease 10:102–105 DOI 10.21037/jtd.2017.12.115. Piccini JP, Simon DN, Steinberg BA, Thomas L, Allen LA, Fonarow GC, Gersh B, Hylek E, Kowey PR, Reiffel JA, Naccarelli GV, Chan PS, Spertus JA, Peterson ED. 2016. Differences in clinical and functional outcomes of atrial fibrillation in women and men: two-year results from the ORBIT-AF registry. JAMA Cardiology 1:282–291 DOI 10.1001/jamacardio.2016.0529. Shawn Lee AL, Gholitabar F, Gongora CA, Dominguez AC, Kroner PT, Hurtado C, Castaneda D, Mehta D. 0000. Comparison of weekday vs weekend admission on inpatient mortality and total hospital charge on atrial flutter and fibrillation: a nationwide analysis. Available at https://www.ahajournals.org/doi/abs/10.1161/circ. 134.suppl_1.14567. Sorita A, Ahmed A, Starr SR, Thompson KM, Reed DA, Prokop L, Shah ND, Murad MH, Ting HH. 2014. Off-hour presentation and outcomes in patients with acute Voruganti et al. (2019), PeerJ, DOI 10.7717/peerj.6211 10/11 myocardial infarction: systematic review and meta-analysis. BMJ 348:f7393 DOI 10.1136/bmj.f7393. Weeda ER, Hodgdon N, Do T, Unachukwu K, Cui A, Lundbye JB, Coleman CI. 2016. Association between weekend admission for atrial fibrillation or flutter and in- hospital mortality, procedure utilization, length-of-stay and treatment costs. International Journal of Cardiology 202:427–429 DOI 10.1016/j.ijcard.2015.09.053. Yoshihara H, Yoneoka D. 2014. REFERENCES Understanding the statistics and limitations of large database analyses. Spine 39:1311–1312 DOI 10.1097/BRS.0000000000000352. Weeda ER, Hodgdon N, Do T, Unachukwu K, Cui A, Lundbye JB, Coleman CI. 2016. Association between weekend admission for atrial fibrillation or flutter and in- hospital mortality, procedure utilization, length-of-stay and treatment costs. International Journal of Cardiology 202:427–429 DOI 10.1016/j.ijcard.2015.09.053. Yoshihara H, Yoneoka D. 2014. Understanding the statistics and limitations of large database analyses. Spine 39:1311–1312 DOI 10.1097/BRS.0000000000000352. Weeda ER, Hodgdon N, Do T, Unachukwu K, Cui A, Lundbye JB, Coleman CI. 2016. Association between weekend admission for atrial fibrillation or flutter and in- hospital mortality, procedure utilization, length-of-stay and treatment costs. International Journal of Cardiology 202:427–429 DOI 10.1016/j.ijcard.2015.09.053. Yoshihara H, Yoneoka D. 2014. Understanding the statistics and limitations of large database analyses. Spine 39:1311–1312 DOI 10.1097/BRS.0000000000000352. 11/11
https://openalex.org/W2016904622	https://open.library.ubc.ca/media/download/pdf/52383/1.0216000/2	English	null	Evaluation of supply-side initiatives to improve access to coronary bypass surgery	BMC health services research	2,012	cc-by	10,920	Correspondence: boris.sobolev@ubc.ca 1School of Population and Public Health, The University of British Columbia, Vancouver, BC, Canada 3Centre for Clinical Epidemiology and Evaluation, Vancouver Coastal Health Research Institute, Vancouver, BC, Canada Full list of author information is available at the end of the article RESEARCH ARTICLE Open Access Abstract Forty percent of patients in the 1998, 1999, 2004 and 2005 cohorts (years when supplementary funding was provided) underwent surgery within 16 to 20 weeks following the median waiting time, while it took between 27 and 37 weeks for the cohorts registered in the years when supplementary funding was not available. Times between decision and surgery were shorter for direct admissions than for wait-listed patients. Among patients who were directly admitted to hospital, time between decision and surgery was longest in 1992–1993 and then has been steadily decreasing through the late nineties. The rate of surgery among these patients was the highest in 1998–1999, and has not changed afterwards, even for years when supplementary funding was provided. Conclusions: Waiting times for non-emergency coronary bypass surgery shortened after supplementary funding was granted to increase volume of cardiac surgical care in a health system with publicly-funded universal coverage for the procedure. The eﬀect of the supplementary funding was not uniform for patients that access the services through wait lists and through direct admission. Keywords: Access to care, CABG, Surgical wait lists, Provincial registry, Health policy care, regional Health Authorities budget a ﬁxed number of open heart surgeries on an annual basis using population- based rates of the disease [2]. As argued elsewhere, those who make allocation decisions have no tools to predict the volume of demand at each hospital and at each point dur- ing the calendar period [3]. When demand exceeds funded capacity, cardiac centers across Canada use wait lists to manage access to the procedure. As a result, operations for patients with less severe coronary artery disease may be delayed when a surgical service experiences an extended demand for more urgent procedures [4-10]. Guided by the Abstract Background: Guided by the evidence that delaying coronary revascularization may lead to symptom worsening and poorer clinical outcomes, expansion in cardiac surgery capacity has been recommended in Canada. Provincial governments started providing one-time and recurring increases in budgets for additional open heart surgeries to reduce waiting times. We sought to determine whether the year of decision to proceed with non-emergency coronary bypass surgery had an eﬀect on time to surgery. Methods: Using records from a population-based registry, we studied times between decision to operate and the procedure itself. We estimated changes in the length of time that patients had to wait for non-emergency operation over 14 calendar periods that included several years when supplementary funding was available. We studied waiting times separately for patients who access surgery through a wait list and through direct admission. Results: During two periods when supplementary funding was available, 1998–1999 and 2004–2005, the weekly rate of undergoing surgery from a wait list was, respectively, 50% and 90% higher than in 1996–1997, the period with the longest waiting times. We also observed a reduction in the diﬀerence between 90th and 50th percentiles of the waiting-time distributions. Forty percent of patients in the 1998, 1999, 2004 and 2005 cohorts (years when supplementary funding was provided) underwent surgery within 16 to 20 weeks following the median waiting time, while it took between 27 and 37 weeks for the cohorts registered in the years when supplementary funding was not available. Times between decision and surgery were shorter for direct admissions than for wait-listed patients. Among patients who were directly admitted to hospital, time between decision and surgery was longest in 1992–1993 and then has been steadily decreasing through the late nineties. The rate of surgery among these patients was the highest in 1998–1999, and has not changed afterwards, even for years when supplementary funding was provided. Results: During two periods when supplementary funding was available, 1998–1999 and 2004–2005, the weekly rate of undergoing surgery from a wait list was, respectively, 50% and 90% higher than in 1996–1997, the period with the longest waiting times. We also observed a reduction in the diﬀerence between 90th and 50th percentiles of the waiting-time distributions. Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 © 2012 Sobolev et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Sobolev et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Correspondence: boris.sobolev@ubc.ca 1 Background In Canada, provincial health insurance plans provide uni- versal, single-payer coverage for surgical coronary revas- cularization, a procedure indicated for the treatment of coronary artery disease [1]. Responsible for the delivery of Page 2 of 15 Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Page 2 of 15 Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 for deciding on policy to reduce wait lists by adding extra funding. evidence that delaying the operation may lead to symp- tom worsening and poorer clinical outcomes, expansion in cardiac surgery capacity has been recommended in Canada. Federal and provincial governments started pro- viding one-time and recurring increases in budgets for additional open heart surgeries to reduce the number of patients waiting for coronary artery bypass grafting (CABG) and their waiting times. To evaluate the eﬀect of providing the additional fund- ing to tertiary care hospitals on access to non-emergency surgical revascularization within a single publicly funded health system, we estimated temporal changes in the length of time between decision to proceed with surgery and performed CABG. We used all relevant records from the provincial population-based registry of patients with angiographically-proven coronary artery disease iden- tiﬁed as needing bypass surgery on a non-emergency basis. To adjust for changes in time between cardiac catheterization and decision to operate, we used the most recent catheterization date from hospital discharge reports assuming that the results of this procedure (coro- nary angiography or intervention) were most likely linked to the decision to operate. Primary comparisons have been done across synthetic cohorts of patients deﬁned by the calendar period of the decision to proceed with surgery. The temporal changes in treatment delays have been estimated separately for patients who access surgery by registration on a wait list and among patients who access surgery by direct admission. g In a previous study of access to non-emergency CABG in British Columbia, Canada, between 1991 and 2000, we found that waiting times for the procedure shortened after 1998 when annual supplementary funding was granted to tertiary care hospitals that had been providing cardiac sur- gical care to adult residents of the province [8]. Between 1995–1996 and 1999–2000, there was a 12% increase in the total number of CABG operations and a decline in median waiting time from 15 to 10 weeks, although the change in waiting times was diﬀerent across urgency groups. Data sources Data were obtained from the British Columbia Cardiac Registries (BCCR) to identify the study participants and their characteristics. This population-based patient reg- istry contains demographic, clinical and treatment data, along with the dates of booking request for operating room time and procedures for all adult patients under- going CABG in any of the four cardiac centers in the province [8]. To identify cardiac catheterization dates, hospital admission and discharge dates, and coexisting medical conditions, we used each patient’s provincial health number to deterministically link BCCR records to the Canadian Institute for Health Information (CIHI) Discharge Abstract Database (DAD) [11]. Since our previous analysis, another $2 million of addi- tional funding from the provincial Ministry of Health was directed in each of 2003 and 2004 toward open heart surgery to increase the volume. However, the eﬀect of this increase in funding on wait-list sizes and waiting times for CABG remains unknown. The longer time frame after the original increase in annual budgets for CABG since 1998 and the additional funding in 2003 and 2004 makes it feasible to generate more precise assessment of this supply-side initiative to improve access to care. The eﬀect of the additional funding has not been contrasted for two diﬀerent pathways of accessing surgical coronary revascu- larization for non-emergency patients as well. Access to non-emergency CABG could be provided either through a wait list or through direct admission to hospital, and it remains unclear whether the eﬀect of the supplementary funding was uniform for the wait-listed and directly- admitted patients. However this information is important Background In addition, between 1995–1996 and 1999–2000 there was a decrease from 54% to 41% in the proportion of patients undergoing the procedure through wait lists. Considering that cardiac surgeons in British Columbia have discretion for direct admission of their patients on the basis of the estimated urgency of treatment, place of residence and other factors, this might indicate that sup- plementary funding had been used to provide more treat- ments without delay. One plausible explanation for these results was that the hospitals had capacity to increase the number of operations and thereby reduce wait times [10]. However, a limitation of the previous study was the short period for analysis of the eﬀects of supplementary fund- ing (two years). In addition, data were not available for the total amount of time between cardiac catheterization and surgical revascularization. As such, in the previous study we did not estimate the eﬀect of the supplementary funding on waiting times in full. Potential confounders h l The existing literature suggests that elderly patients are more likely to undergo revascularization as an urgent pro- cedure [14], that smaller diameter of the coronary vessels may account for the higher risk of adverse cardiovascu- lar events among women [15], that co-existing conditions may delay open heart surgery [16], that post-operative survival depends on institutional constraints and indi- vidual care providers [17], and that changes in practice may reduce the waiting time until surgery [8]. To iden- tify comorbidities at the time of decision to operate, we used diagnoses reported in the DAD within one year prior to decision. The reference category was deﬁned as no coexisting conditions. The ﬁrst comparison category was deﬁned as patients with any of the following conditions at presentation: congestive heart failure, diabetes mel- litus, chronic obstructive pulmonary disease, cancer, or rheumatoid arthritis [4]. The second comparison category was deﬁned as patients presenting with other coexisting chronic conditions, as deﬁned elsewhere [18]. The ﬁnal study cohort had a total of 12,030 wait-listed patients who had a decision to undergo ﬁrst-time isolated CABG surgery and 12,818 direct admissions who under- went ﬁrst-time isolated CABG surgery from January 1, 1992 through December 31, 2005. Among the wait-listed patients 10,339 (85.9%) underwent surgery within 1 year of registration and the remaining were removed from the list without surgery: 104 (0.9%) died, 257 (2.1%) continued to receive medical treatment, 231 (1.9%) declined surgery, 86 (0.7%) were transferred to another surgeon or hospi- tal, 321 (2.7%) were removed for other reasons, and 692 (5.8%) remained on the list after 52 weeks or at the end of the study period. In addition, we used time between cardiac catheteri- zation and decision to operate as a covariate that might reﬂect changes in practice over time. The time between catheterization and decision measured in calendar weeks. The catheterization dates were obtained from the DAD and deﬁned as the most recent diagnostic (Canadian Clas- siﬁcation of Procedure (CCP) codes 4892–4898, 4996, 4997) or therapeutic (CCP codes 4802, 4803, 4809) catheterization performed within one year preceding and including the date of booking for wait-listed patients or within one year preceding and including the date of surgery for direct admissions. We used the date of most recent catheterization procedures (diagnostic or thera- peutic) because the results of this procedure are most likely linked to decision to operate [19]. Outcome groups were classiﬁed as urgent, semiurgent, or nonurgent based on the patient’s need for treatment, as deﬁned elsewhere [8]. The outcome was time between decision to operate and surgical revascularization. This time was measured in cal- endar weeks for wait-listed patients and in days for direct admissions. We used the date of the registration on a wait list as a proxy for the decision to operate for wait- listed patients and the date of catheterization or the date of admission to hospital, whichever was most recent, as a proxy for the date of decision to operate for directly- admitted patients. This latter rule reﬂects variation in care paths of the patients: following angiography the patient may be admitted for in-hospital consultation with a car- diac surgeon, or patients who live far away are admitted for angiography, stay in hospital to undergo tests, and are booked for surgery or discharged with planned re- admission. The inception cohort of wait-listed patients had a total of 14,049 records of registration for CABG from January 1, 1991 through December 31, 2005. We excluded 567 records of patients for various reasons: procedure at registration was not isolated CABG (312), procedure at registration or at surgery was not ﬁrst-time CABG (62), emergency cases at the time of registration (34), missing operating room reports (4), removed on the registration date (101), registration was on a weekend and admission was the day after (14), or the patient had multiple episodes (40). We also excluded 1,452 records of patients who were registered in 1991 (797) or did not have a catheterization date (655). The inception cohort of direct admissions had a total of 16,014 records of CABG surgery from January 1, 1991 through December 31, 2005. We excluded 1,282 records of patients for various reasons: procedure at surgery was not isolated CABG (211), procedure at surgery was not ﬁrst- time CABG (54), emergency case at the time of surgery (861), or the patient had multiple episodes (156). We also excluded 1,914 records of patients who had surgery in 1991 (1,031), did not have a catheterization date (838), did not have an admission date (38), or had a decision to operate in 1991 (7). Patients Using records from the registry, we studied two groups of patients: (1) those who were registered on a wait list for ﬁrst-time isolated CABG surgery; and (2) those who underwent the procedure by direct admission to hospital on a non-emergency basis. Patients who accessed surgery through a wait list were registered by the surgeon’s oﬃce on the wait list after an outpatient consultation with a car- diac surgeon. In contrast, patients who accessed surgery through direct admission were admitted to a hospital’s cardiac ward directly from the catheterization labora- tory or after an outpatient consultation with a cardiac surgeon if the patient had disabling symptoms or high- risk anatomy of the coronary lesion(s). Patients in both Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Page 3 of 15 Page 3 of 15 Page 3 of 15 Potential confounders h l We used calendar weeks as the unit of time because scheduling of surgical procedures is done on a weekly basis [13]. Statistical analysis For wait-listed patients, the distribution of patient char- acteristics varied across periods (Table 1). The majority of the wait-listed patients who had a decision for CABG were men (83%). Later periods tended to have older patients (p < 0.001), fewer urgent cases at decision (p < 0.001), more patients with major comorbidities at the time of decision (p < 0.001), and more limited coronary anatomy aﬀected (p < 0.001). The distribution of cases by insti- tution at decision seemed to increase over periods for hospital 1, but decreased for hospital 2 (p < 0.001). The majority of wait-listed patients had a decision to oper- ate within a week of catheterization. This majority ranged from about 42% in 1992–1993 and increased to 65% in 2002–2003 (p < 0.001). We used chi-square testing to compare the distributions of patient characteristics across calendar period of deci- sion for wait-listed patients and direct admissions. We estimated percentiles and conditional median times to surgery to characterize the variation in times to surgi- cal revascularization, by calendar period for each type of access. Percentiles of time to surgery were estimated using the product-limit method [13,20]. The conditional median waiting time at a given moment after decision is deﬁned as the period during which one half of the patients who wait for surgery are expected to have it [13]. For wait-listed patients, average weekly surgery rates were calculated as the number of procedures divided by the sum of observed times from decision to surgery or removal from the wait list. For direct admissions, the average daily surgery rates were calculated as the number of procedures divided by the sum of observed times from decision to surgery. For direct admissions, the trends in changes in the distributions of patient characteristics over periods were similar to wait-listed patients (Table 2). Later periods tended to have older patients (p < 0.001), fewer urgent cases at decision (p < 0.001), and more patients with major comorbidities at the time of decision (p < 0.001). In contrast to wait-listed patients, later periods tended to have more direct admissions who had less limited coro- nary anatomy aﬀected (p < 0.001) and more decisions to operate within a week of catheterization (p < 0.001). About two-thirds of direct admissions had a decision to operate within a week of catheterization across all calen- dar periods. Primary study variable Th i d The primary study variable was calendar period of decision to operate classiﬁed into 7 two-year periods: 1992–93, 1994–95, 1996–97, 1998–99, 2000–01, 2002– 03, 2004–05. Calendar periods are often used as a proxy for changes in the availability of hospital resources, such as, surgical staﬃng, operating room time, special equip- ment, and beds, in studies that attempted to explain vari- ations in the patient’s waiting time for surgery [12,13]. Because the budget for open heart surgeries is deter- mined annually, calendar period provides an appropriate indicator for changes in the funded procedures. Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Page 4 of 15 Access to surgery through wait-list registration h h l d b f g y g g Figure 1 shows the cumulative distribution functions of waiting time for each calendar period, which could be used to derive the number of weeks required for a spec- iﬁed proportion of patients to undergo the operation. The diﬀerences in the proportion of patients undergo- ing CABG within a certain time of decision were sig- niﬁcant across periods (Log-rank statistic = 545.6, df=6, p < 0.001), with longer waiting times when the deci- sion was made in 1996–1997 and 2002–2003. The waiting times for these years were such that half of the wait-listed patients underwent surgery within 16 to 17 weeks, and 90% underwent surgery within 46 to 51 weeks. In contrast, during the other years about half of patients underwent surgery within 8 to 14 weeks of decision. Comparing the 1998, 1999, 2004 and 2005 cohorts (the periods when sup- plement funding was provided) with the rest, we observed a compression in access to surgery, i.e., reduction in the The Behavioural Research Ethics Board of the Univer- sity of British Columbia approved the study protocol, Certiﬁcate of Approval H06-80651. Statistical analysis Discrete-time survival regression methods were used to model the relation between the time to surgical revascu- larization and calendar period of decision for each type of access [13,21]. We restricted the regression analysis to the ﬁrst 52 weeks after decision for wait-listed patients and to the ﬁrst 7 days after decision for direct admissions. The calendar period was entered into the regression model as a set of 6 binary indicators. The 1996–1997 group (i.e., the reference group) took a value of 0 for all indicator vari- ables. The exponential of the regression coeﬃcient for an indicator variable for a period gave the odds of surgery in that period relative to the odds of surgery in 1996–1997. In a multivariable model, we adjusted for sex, age at decision, urgency at decision, institution at decision, comorbidi- ties at decision, coronary anatomy at decision, and time between catheterization and decision. When compared over calendar periods, wait-listed patients were more prevalent in the low risk group and directly-admitted patients were more prevalent in high risk group (see Table 3). The percentage of low risk patients accessing surgery through direct admis- sion declined considerably in years when supplementary funding was provided. To explore the eﬀect of the supplementary funding on direct admissions and wait-listed patients, we classiﬁed all patients using the algorithm developed by Northern New England Cardiovascular Disease Study Group [22]. For each patient, we calculated the prognostic risk of in-hospital death that summarized the eﬀect of clinical and patient characteristics. We then compared the dis- tribution of these risks between direct admissions and wait-listed patients for each calendar period of surgery using chi-square testing. Patients characteristics Overall 24,848 patients had a decision for ﬁrst-time, iso- lated CABG between 1992 and 2005: 12,030 (48%) were registered on a wait list and 12,818 (52%) were directly admitted to hospital. Patients characteristics Table 1 Characteristics of 12,030 wait-list registered patients, who had decision for coronary artery bypass grafting in British Columbia 1992–2005, by calendar period of decision All periods 1992–1993 1994–1995 1996–1997 1998–1999 2000–2001 2002–2003 2004–2005 Characteristic (n = 12030) (n = 1726) (n = 1793) (n = 1862) (n = 1610) (n = 1791) (n = 1889) (n = 1359) Age group at decision (years) <50 761 (6.3) 140 (8.1) 147 (8.2) 118 (6.3) 99 (6.1) 113 (6.3) 85 (4.5) 59 (4.3) 50–59 2559 (21.3) 362 (21.0) 390 (21.8) 367 (19.7) 337 (20.9) 419 (23.4) 426 (22.6) 258 (19.0) 60–69 4468 (37.1) 703 (40.7) 655 (36.5) 717 (38.5) 551 (34.2) 619 (34.6) 717 (38.0) 506 (37.2) 70–79 3786 (31.5) 500 (29.0) 542 (30.2) 602 (32.3) 567 (35.2) 573 (32.0) 551 (29.2) 451 (33.2) ≥80 456 (3.8) 21 (1.2) 59 (3.3) 58 (3.1) 56 (3.5) 67 (3.7) 110 (5.8) 85 (6.3) Sex Men 9981 (83.0) 1425 (82.6) 1487 (82.9) 1500 (80.6) 1334 (82.9) 1502 (83.9) 1595 (84.4) 1138 (83.7) Women 2049 (17.0) 301 (17.4) 306 (17.1) 362 (19.4) 276 (17.1) 289 (16.1) 294 (15.6) 221 (16.3) Urgency at decision∗ Urgent 739 (6.1) 113 (6.5) 162 (9.0) 177 (9.5) 66 (4.1) 72 (4.0) 87 (4.6) 62 (4.6) Semiurgent 8769 (72.9) 1331 (77.1) 1240 (69.2) 1295 (69.5) 1107 (68.8) 1315 (73.4) 1401 (74.2) 1080 (79.5) Nonurgent 2304 (19.2) 269 (15.6) 375 (20.9) 369 (19.8) 432 (26.8) 390 (21.8) 354 (18.7) 115 (8.5) Institution at decision 1 2668 (22.2) 328 (19.0) 356 (19.9) 378 (20.3) 361 (22.4) 449 (25.1) 533 (28.2) 263 (19.4) 2 3575 (29.7) 724 (41.9) 589 (32.8) 467 (25.1) 555 (34.5) 475 (26.5) 409 (21.7) 356 (26.2) 3 2914 (24.2) 438 (25.4) 429 (23.9) 492 (26.4) 265 (16.5) 401 (22.4) 472 (25.0) 417 (30.7) 4 2873 (23.9) 236 (13.7) 419 (23.4) 525 (28.2) 429 (26.6) 466 (26.0) 475 (25.1) 323 (23.8) Comorbidities at decision Major conditions† 2901 (24.1) 373 (21.6) 386 (21.5) 418 (22.4) 368 (22.9) 452 (25.2) 518 (27.4) 386 (28.4) Other conditions‡ 2856 (23.7) 520 (30.1) 496 (27.7) 526 (28.2) 384 (23.9) 379 (21.2) 354 (18.7) 197 (14.5) None 6273 (52.1) 833 (48.3) 911 (50.8) 918 (49.3) 858 (53.3) 960 (53.6) 1017 (53.8) 776 (57.1) Coronary anatomy aﬀected at decision Left main 1780 (14.8) 251 (14.5) 287 (16.0) 299 (16.1) 256 (15.9) 265 (14.8) 284 (15.0) 138 (10.2) Multivessel§ 8715 (72.4) 1361 (78.9) 1407 (78.5) 1418 (76.2) 1202 (74.7) 1320 (73.7) 1274 (67.4) 733 (53.9) Limited∥ 1535 (12.8) 114 (6.6) 99 (5.5) 145 (7.8) 152 (9.4) 206 (11.5) 331 (17.5) 488 (35.9) Table 1 Characteristics of 12,030 wait-list registered patients, who had decision for coronary arte period of decision haracteristics of 12,030 wait-list registered patients, who had decision for coronary artery bypass gra decision had decision for coronary artery bypass grafting in British Columbia 1992–2005, by calendar Sobolev et al. g p y ‡Peripheral vascular disease, cerebrovascular disease, dementia, peptic ulcer disease, hemiplegia, renal disease, or liver disease. §Two or three-vessel disease with stenosis of the proximal left anterior descending (PLAD) artery. ∥ f f 218 patients had unknown values for urgency at decision. †Congestive heart failure, diabetes mellitus, chronic obstructive pulmonary disease, rheumatoid arthritis, or cancer. 201 3/12 Two or three vessel disease with stenosis of the proximal left anterior descending (PLAD) artery. ∥Two-vessel disease with no stenosis of the PLAD artery or one-vessel disease with stenosis of the PLAD artery. †Congestive heart failure, diabetes mellitus, chronic obstructive pulmonary disease, rheumatoid arthritis, or cancer. ‡Peripheral vascular disease, cerebrovascular disease, dementia, peptic ulcer disease, hemiplegia, renal disease, or live § 218 patients had unknown values for urgency at decision. †Congestive heart failure, diabetes mellitus, chronic obstructive pulmonary disease, rheumatoid arthritis, or cancer. 218 patients had unknown values for urgency at decision. † h f l d b ll h b l d h d h §Two or three-vessel disease with stenosis of the proximal left anterior descending (PLAD) artery. ∥Two-vessel disease with no stenosis of the PLAD artery or one-vessel disease with stenosis of the PLAD artery Patients characteristics BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Table 1 Characteristics of 12,030 wait-list registered patients, who had decision for coronary artery bypass grafting in British Columbia 1992–2005, by calendar period of decision (Continued) Time between catheterization and decision for surgical revascularization (weeks) 0–1 6651 (55.3) 726 (42.1) 932 (52.0) 1005 (54.0) 912 (56.6) 1093 (61.0) 1236 (65.4) 747 (55.0) 2–3 2066 (17.2) 422 (24.4) 377 (21.0) 350 (18.8) 269 (16.7) 245 (13.7) 215 (11.4) 188 (13.8) 4–5 1041 (8.7) 209 (12.1) 174 (9.7) 157 (8.4) 150 (9.3) 139 (7.8) 102 (5.4) 110 (8.1) 6–7 642 (5.3) 122 (7.1) 92 (5.1) 123 (6.6) 74 (4.6) 84 (4.7) 66 (3.5) 81 (6.0) ≥8 1630 (13.5) 247 (14.3) 218 (12.2) 227 (12.2) 205 (12.7) 230 (12.8) 270 (14.3) 233 (17.1) 218 patients had unknown values for urgency at decision Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 ait-list registered patients, who had decision for coronary artery bypass grafting in British Columbia 1992–2005, by calendar n for surgical revascularization (weeks) (42.1) 932 (52.0) 1005 (54.0) 912 (56.6) 1093 (61.0) 1236 (65.4) 747 (55.0) (24.4) 377 (21.0) 350 (18.8) 269 (16.7) 245 (13.7) 215 (11.4) 188 (13.8) (12.1) 174 (9.7) 157 (8.4) 150 (9.3) 139 (7.8) 102 (5.4) 110 (8.1) (7.1) 92 (5.1) 123 (6.6) 74 (4.6) 84 (4.7) 66 (3.5) 81 (6.0) (14.3) 218 (12.2) 227 (12.2) 205 (12.7) 230 (12.8) 270 (14.3) 233 (17.1) t decision. onic obstructive pulmonary disease, rheumatoid arthritis, or cancer. ease, dementia, peptic ulcer disease, hemiplegia, renal disease, or liver disease. proximal left anterior descending (PLAD) artery. D artery or one-vessel disease with stenosis of the PLAD artery. Patients characteristics ts, who had decision for coronary artery bypass grafting in British Columbia 1992–2005, by calendar Table 1 Characteristics of 12,030 wait-list registered patients, who had decision for coronary arte period of decision (Continued) ∥Two-vessel disease with no stenosis of the PLAD artery or one-vessel disease with stenosis of the PLAD artery Table 2 Characteristics of 12,818 direct admissions, who had decision for coronary artery bypass grafting in British Columbia 1992–2005, by calendar period of decision All periods 1992–1993 1994–1995 1996–1997 1998–1999 2000–2001 2002–2003 2004–2005 Characteristic (n = 12818) (n = 1204) (n = 1474) (n = 1621) (n = 2089) (n = 2121) (n = 1940) (n = 2369) Age group at decision (years) <50 920 (7.2) 111 (9.2) 97 (6.6) 129 (8.0) 142 (6.8) 157 (7.4) 139 (7.2) 145 (6.1) 50–59 2604 (20.3) 228 (18.9) 275 (18.7) 313 (19.3) 435 (20.8) 432 (20.4) 414 (21.3) 507 (21.4) 60–69 4354 (34.0) 442 (36.7) 569 (38.6) 571 (35.2) 715 (34.2) 676 (31.9) 615 (31.7) 766 (32.3) 70–79 4341 (33.9) 389 (32.3) 481 (32.6) 550 (33.9) 726 (34.8) 747 (35.2) 663 (34.2) 785 (33.1) ≥80 599 (4.7) 34 (2.8) 52 (3.5) 58 (3.6) 71 (3.4) 109 (5.1) 109 (5.6) 166 (7.0) Sex Men 10067 (78.5) 914 (75.9) 1104 (74.9) 1247 (76.9) 1659 (79.4) 1687 (79.5) 1525 (78.6) 1931 (81.5) Women 2750 (21.5) 290 (24.1) 370 (25.1) 374 (23.1) 430 (20.6) 434 (20.5) 415 (21.4) 437 (18.4) Urgency at decision Urgent 5944 (46.4) 727 (60.4) 944 (64.0) 900 (55.5) 1046 (50.1) 887 (41.8) 698 (36.0) 742 (31.3) Semiurgent 6445 (50.3) 453 (37.6) 485 (32.9) 654 (40.3) 906 (43.4) 1148 (54.1) 1211 (62.4) 1588 (67.0) Nonurgent 429 (3.3) 24 (2.0) 45 (3.1) 67 (4.1) 137 (6.6) 86 (4.1) 31 (1.6) 39 (1.6) Institution at decision 1 2437 (19.0) 89 (7.4) 206 (14.0) 295 (18.2) 381 (18.2) 473 (22.3) 427 (22.0) 566 (23.9) 2 2962 (23.1) 342 (28.4) 369 (25.0) 333 (20.5) 466 (22.3) 426 (20.1) 457 (23.6) 569 (24.0) 3 4964 (38.7) 417 (34.6) 623 (42.3) 838 (51.7) 870 (41.6) 799 (37.7) 683 (35.2) 734 (31.0) 4 2455 (19.2) 356 (29.6) 276 (18.7) 155 (9.6) 372 (17.8) 423 (19.9) 373 (19.2) 500 (21.1) Comorbidities at decision Major conditions† 5458 (42.6) 413 (34.3) 552 (37.4) 626 (38.6) 885 (42.4) 949 (44.7) 914 (47.1) 1119 (47.2) Other conditions‡ 6248 (48.7) 708 (58.8) 809 (54.9) 885 (54.6) 1033 (49.4) 1027 (48.4) 857 (44.2) 929 (39.2) None 1112 (8.7) 83 (6.9) 113 (7.7) 110 (6.8) 171 (8.2) 145 (6.8) 169 (8.7) 321 (13.6) Coronary anatomy aﬀected at decision Left main 3184 (24.8) 282 (23.4) 358 (24.3) 377 (23.3) 466 (22.3) 539 (25.4) 534 (27.5) 628 (26.5) Multivessel§ 8855 (69.1) 814 (67.6) 1019 (69.1) 1136 (70.1) 1501 (71.9) 1470 (69.3) 1307 (67.4) 1608 (67.9) Limited∥ 779 (6.1) 108 (9.0) 97 (6.6) 108 (6.7) 122 (5.8) 112 (5.3) 99 (5.1) 133 (5.6) Table 2 Characteristics of 12,818 direct admissions, who had decision for coronary artery bypass decision ecision for coronary artery bypass grafting in British Columbia 1992–2005, by calendar period of eristics of 12,818 direct admissions, who had decision for coronary artery bypass grafting in British C Sobolev et al. 1 direct admission with decision date in 2004–2005 had unknown sex. †Congestive heart failure, diabetes mellitus, chronic obstructive pulmonary disease, rheumatoid arthritis, or cancer. ‡Peripheral vascular disease, cerebrovascular disease, dementia, peptic ulcer disease, hemiplegia, renal disease, or liver disease. § 1 direct admission with decision date in 2004–2005 had unknown sex. 201 3/12 Two or three vessel disease with stenosis of the proximal left anterior descending (PLAD) artery. ∥Two-vessel disease with no stenosis of the PLAD artery or one-vessel disease with stenosis of the PLAD artery. §Two or three-vessel disease with stenosis of the proximal left anterior descending (PLAD) artery. ∥ §Two or three-vessel disease with stenosis of the proximal left anterior descending (PLAD) artery. ∥Two-vessel disease with no stenosis of the PLAD artery or one-vessel disease with stenosis of the PLAD ch 201 963/12 ee-vessel disease with stenosis of the proximal left anterior descending (PLAD) artery. disease with no stenosis of the PLAD artery or one-vessel disease with stenosis of the PLAD artery. , , p y , , ral vascular disease, cerebrovascular disease, dementia, peptic ulcer disease, hemiplegia, renal disease, or liver disease. Patients characteristics BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Sobolev et al. BMC Health Services R http://www.biomedcentral.com/14 Table 2 Characteristics of 12,818 direct admissions, who had decision for coronary artery bypass grafting in British Columbia 1992–2005, by calendar period of decision (Continued) Time between catheterization and decision for surgical revascularization (weeks) 0–1 8576 (66.9) 885 (73.5) 1021 (69.3) 991 (61.1) 1381 (66.1) 1465 (69.1) 1246 (64.2) 1587 (67.0) 2–3 2232 (17.4) 164 (13.6) 215 (14.6) 326 (20.1) 334 (16.0) 372 (17.5) 406 (20.9) 415 (17.5) 4–5 592 (4.6) 46 (3.8) 54 (3.7) 81 (5.0) 87 (4.2) 85 (4.0) 113 (5.8) 126 (5.3) 6–7 364 (2.8) 25 (2.1) 38 (2.6) 45 (2.8) 76 (3.6) 46 (2.2) 58 (3.0) 76 (3.2) ≥8 1054 (8.2) 84 (7.0) 146 (9.9) 178 (11.0) 211 (10.1) 153 (7.2) 117 (6.0) 165 (7.0) 1 direct admission with decision date in 2004–2005 had unknown sex Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 rect admissions, who had decision for coronary artery bypass grafting in British Columbia 1992–2005, by calendar period of n for surgical revascularization (weeks) (73.5) 1021 (69.3) 991 (61.1) 1381 (66.1) 1465 (69.1) 1246 (64.2) 1587 (67.0) (13.6) 215 (14.6) 326 (20.1) 334 (16.0) 372 (17.5) 406 (20.9) 415 (17.5) (3.8) 54 (3.7) 81 (5.0) 87 (4.2) 85 (4.0) 113 (5.8) 126 (5.3) (2.1) 38 (2.6) 45 (2.8) 76 (3.6) 46 (2.2) 58 (3.0) 76 (3.2) (7.0) 146 (9.9) 178 (11.0) 211 (10.1) 153 (7.2) 117 (6.0) 165 (7.0) 005 had unknown sex. onic obstructive pulmonary disease, rheumatoid arthritis, or cancer. ease, dementia, peptic ulcer disease, hemiplegia, renal disease, or liver disease. e proximal left anterior descending (PLAD) artery. D artery or one-vessel disease with stenosis of the PLAD artery. d decision for coronary artery bypass grafting in British Columbia 1992–2005, by calendar period of Table 2 Characteristics of 12,818 direct admissions, who had decision for coronary artery bypass decision (Continued) Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 ∥Two-vessel disease with no stenosis of the PLAD artery or one-vessel disease with stenosis of the PLAD artery Sobolev et al. Patients characteristics After the ﬁrst 10 weeks, the conditional median times for surgery remained relatively constant for period 2004– 2005, but these times increased with longer waits for period 1992–1993. In contrast, in the two periods showing the longest waiting times, namely 1996–1997 and 2002– 2003, the conditional median time decreased with the longer waits, indicating perhaps an active wait-list man- agement in years when the supplementary funding was not provided. Patients characteristics BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Page 9 of 15 Page 9 of 15 Table 3 Prognostic risk of in-hospital death, by calendar period of surgery and type of access Table 3 Prognostic risk of in-hospital death, by calendar period of surgery and type of access Calendar period of surgery Risk∗, % Access type <1.0 1.0–3.0 >3.0 p value† 1992–1993 <0.001 Wait-listed 808 (58.4) 555 (40.1) 20 (1.4) Direct admission 559 (46.2) 581 (48.0) 70 (5.8) 1994–1995 <0.001 Wait-listed 920 (61.3) 549 (36.6) 32 (2.1) Direct admission 694 (46.9) 673 (45.5) 112 (7.6) 1996–1997 <0.001 Wait-listed 929 (58.2) 625 (39.2) 41 (2.6) Direct admission 761 (46.9) 776 (47.9) 84 (5.2) 1998–1999 <0.001 Wait-listed 829 (52.9) 675 (43.1) 62 (4.0) Direct admission 867 (41.2) 984 (46.8) 251 (11.9) 2000–2001 <0.001 Wait-listed 525 (37.6) 724 (51.8) 148 (10.6) Direct admission 413 (19.4) 1119 (52.6) 594 (27.9) 2002–2003 <0.001 Wait-listed 691 (40.6) 876 (51.4) 136 (8.0) Direct admission 458 (23.6) 1037 (53.5) 444 (22.9) 2004–2005 <0.001 Wait-listed 489 (37.6) 692 (53.2) 119 (9.2) Direct admission 461 (19.3) 1242 (52.1) 683 (28.6) . Based on Northern New England Cardiovascular Disease Study Group [22]. †Compares the distribution of risk between wait-listed patients and direct admissions for each period. Based on Northern New England Cardiovascular Disease Study Group [22]. †Compares the distribution of risk between wait-listed patients and direct admissions for each period. Once wait-listed patients had a decision to operate, the average weekly number of operations was lowest in 1996– 1997, 2000–2001, and 2002–2003 (Table 4). The average rate of undergoing surgery from a wait list was about 5 procedures per 100 patient-weeks in these periods with the lowest rates. After adjustment for patient-related fac- tors, compared to the surgery rate in 1996–1997, the weekly odds that a patient would undergo operation were highest in 1992–1993, decreased to a low in 1996–1997, rose in 1998–1999, decreased to another low in 2000– 2001 and 2002–2003, and then rose again in 2004–2005 (Table 4). length of the waiting-time interval required for a speci- ﬁed proportion to undergo the operation. Patients characteristics As measured by the diﬀerence between 90th and 50th percentiles of the wait time distributions, 40% of the 1998, 1999, 2004 and 2005 cohorts underwent surgery within 16 to 20 weeks following the median waiting time (50th percentile), while it took between 27 and 37 weeks for the cohorts in years when supplementary funding was not available (Figure 2). length of the waiting-time interval required for a speci- ﬁed proportion to undergo the operation. As measured by the diﬀerence between 90th and 50th percentiles of the wait time distributions, 40% of the 1998, 1999, 2004 and 2005 cohorts underwent surgery within 16 to 20 weeks following the median waiting time (50th percentile), while it took between 27 and 37 weeks for the cohorts in years when supplementary funding was not available (Figure 2). Periods 1992–1993 and 2004–2005 showed the short- est waiting time and showed a notable diﬀerence in changes of conditional remaining times between the peri- ods (Figure 3). For each period, the remaining time during which half of the patients were expected to access surgery did not change for about the ﬁrst 10 weeks since decision. After the ﬁrst 10 weeks, the conditional median times for surgery remained relatively constant for period 2004– 2005, but these times increased with longer waits for period 1992–1993. In contrast, in the two periods showing the longest waiting times, namely 1996–1997 and 2002– 2003, the conditional median time decreased with the longer waits, indicating perhaps an active wait-list man- agement in years when the supplementary funding was not provided. length of the waiting-time interval required for a speci- ﬁed proportion to undergo the operation. As measured by the diﬀerence between 90th and 50th percentiles of the wait time distributions, 40% of the 1998, 1999, 2004 and 2005 cohorts underwent surgery within 16 to 20 weeks following the median waiting time (50th percentile), while it took between 27 and 37 weeks for the cohorts in years when supplementary funding was not available (Figure 2). Periods 1992–1993 and 2004–2005 showed the short- est waiting time and showed a notable diﬀerence in changes of conditional remaining times between the peri- ods (Figure 3). For each period, the remaining time during which half of the patients were expected to access surgery did not change for about the ﬁrst 10 weeks since decision. Access to surgery through direct admission h h d d Among patients who had a decision to operate for CABG and were directly admitted to hospital, times between decision to operate and surgical revascularization were longest in the periods 1992–1994, but were much shorter than the times for wait-listed patients (Figure 4). During these periods 80% of patients underwent surgery within 7 days. Half of patients underwent surgery within 1 day of decision to operate and 80% within 5 days during periods 1995 and later. Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Page 10 of 15 0 4 8 12 16 20 24 28 32 36 40 44 48 52 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1992−1993 1994−1995 1996−1997 1998−1999 2000−2001 2002−2003 2004−2005 Time since decision (weeks) Probability of surgical revascularization Figure 1 Estimated probability of surgical revascularization within a certain time of decision for surgical revascularization, by calendar period of decision, among patients who accessed surgery through registration on a wait list. 0 4 8 12 16 20 24 28 32 36 40 44 48 52 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1992−1993 1994−1995 1996−1997 1998−1999 2000−2001 2002−2003 2004−2005 Time since decision (weeks) Probability of surgical revascularization Figure 1 Estimated probability of surgical revascularization within a certain time of decision for surgical revascularization, by calendar period of decision, among patients who accessed surgery through registration on a wait list. 92 93 94 95 96 97 98 99 00 01 02 03 04 05 0 4 8 12 16 20 24 28 32 36 40 44 48 52 Calendar period of decision Time between decision and surgery (weeks) Figure 2 Time between decision to operate and surgical revascularization among patients who accessed surgery through registration on a wait list, by calendar period of decision. Bottom of bar = 50th percentile, top of bar = 90th percentile. Figure 1 Estimated probability of surgical revascularization within a certain time of decision for surgical revascularization, by calendar period of decision, among patients who accessed surgery through registration on a wait list. Access to surgery through direct admission h h d d 92 93 94 95 96 97 98 99 00 01 02 03 04 05 0 4 8 12 16 20 24 28 32 36 40 44 48 52 Calendar period of decision Time between decision and surgery (weeks) Figure 2 Time between decision to operate and surgical revascularization among patients who accessed surgery through registration on a wait list, by calendar period of decision. Bottom of bar = 50th percentile, top of bar = 90th percentile. 92 93 94 95 96 97 98 99 00 01 02 03 04 05 0 4 8 12 16 20 24 28 32 36 40 44 48 52 Calendar period of decision Time between decision and surgery (weeks) Figure 2 Time between decision to operate and surgical revascularization among patients who accessed surgery through registration on a wait list, by calendar period of decision. Bottom of bar = 50th percentile, top of bar = 90th percentile. Figure 2 Time between decision to operate and surgical revascularization among patients who accessed surgery through registration on a wait list, by calendar period of decision. Bottom of bar = 50th percentile, top of bar = 90th percentile. Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Page 11 of 15 0 4 8 12 16 20 5 10 15 20 25 30 1992−1993 1994−1995 1996−1997 1998−1999 2000−2001 2002−2003 2004−2005 Time since decision (weeks) Conditional median time for surgical revascularization (weeks) 1992−1993 1994−1995 1996−1997 1998−1999 2000−2001 2002−2003 2004−2005 Figure 3 Conditional median time of surgical revascularization at a certain time since decision for surgical revascularization, by calendar period of decision, among patients who accessed surgery through registration on a wait list. Figure 3 Conditional median time of surgical revascularization at a certain time since decision for surgical revascularization, by calendar period of decision, among patients who accessed surgery through registration on a wait list. Once direct admissions had a decision to operate, the average daily number of operations was lowest in 1992– 1993 at about 23 procedures per 100 patient-days, rose to about 36 procedures per 100 patient-days in 1998– 1999, after which the rate remained stable (Table 5). After adjustment for patient-related factors, compared to the Table 4 Average weekly rate of coronary artery bypass procedure in relation to calendar period of decision to proceed with surgery, for patients registered on a wait list Type of access by No. of No. g p er 100 patient weeks calculated as the number of procedures performed divided by the sum of waiting times. f f f p p p y g ‡As measured by odds ratios derived from discrete-time survival models, adjusting for consecutive weeks of waiting (0.5–52 weeks). p p p y g ‡As measured by odds ratios derived from discrete-time survival models, adjusting for consecutive weeks of waiting (0.5–52 weeks). §Adjusted for sex, age at decision, urgency at decision, institution at decision, comorbidities at decision, coronary anatomy at decision, and time between catheterization and decision p p p y g ‡As measured by odds ratios derived from discrete-time survival models, adjusting for consecutive weeks of waiting (0.5–52 weeks). §Adjusted for sex, age at decision, urgency at decision, institution at decision, comorbidities at decision, coronary anatomy at decision, and time between catheterization and decision. ∥218 patients with unknown values for urgency at decision were excluded. To measure waiting times, patients were followed for a maximum of 52 weeks. p ks calculated as the number of procedures performed divided by the sum of waiting times. To measure waiting times, patients were followed for a maximum of 52 weeks. ; times, patients were followed for a maximum of 52 weeks. j g g , age at decision, urgency at decision, institution at decision, comorbidities at decision, coronary anatomy at decision, and time between nd decision. h unknown values for urgency at decision were excluded. Time between catheterization and surgery We found that for those who underwent CABG, times between catheterization and surgical revascularization were shorter among patients who accessed surgery through direct admission compared to access through wait-list registration. In the urgent group, half of direct admissions underwent surgery within 1 week, whereas half of wait-listed patients underwent surgery within 7 weeks. In addition, half of direct admissions and half of wait-listed patients underwent surgery within 1 and 13 weeks respectively, in the semiurgent group. In the nonurgent group these timeframes were 8 and 23 weeks, respectively. The weekly odds of surgery after catheteriza- tion were 4, 5, and 3 times higher among direct admissions compared to wait-listed patients in the urgent, semiur- gent, and nonurgent groups, respectively, after adjustment for age, sex, hospital at catheterization, mode of admis- sion at catheterization, comorbidity at surgery, hospital at surgery, and coronary anatomy at surgery. We also found that time between catheterization and decision to operate became shorter in the 2000s compared to the 1990s for the Access to surgery through direct admission h h d d of Total Crude rate† Crude OR‡ Adjusted OR ‡§∥ calendar period of decision patients procedures waiting time∗ (95% CI) (95% CI) (95% CI) 1992–1993 1726 1519 20462 7.4 (7.1–7.8) 1.7 (1.6–1.9) 2.0 (1.9–2.2) 1994–1995 1793 1545 26452 5.8 (5.5–6.1) 1.3 (1.2–1.4) 1.4 (1.3–1.5) 1996–1997 1862 1555 35054 4.4 (4.2–4.7) 1.0 1.0 1998–1999 1610 1426 21266 6.7 (6.4–7.1) 1.5 (1.4–1.6) 1.5 (1.4–1.7) 2000–2001 1791 1509 31887 4.7 (4.5–5.0) 1.1 (1.0–1.2) 1.1 (1.0–1.2) 2002–2003 1889 1613 35447 4.6 (4.3–4.8) 1.0 (0.9–1.1) 1.0 (0.9–1.1) 2004–2005 1359 1172 14252 8.2 (7.8–8.7) 1.9 (1.7–2.0) 1.9 (1.8–2.1) All periods 12030 10339 184820 5.6 (5.5–5.7) – – Abb i ti OR dd ti CI ﬁd i t l eekly rate of coronary artery bypass procedure in relation to calendar period of decision to proceed atients registered on a wait list Table 4 Average weekly rate of coronary artery bypass procedure in relation to calendar period of d with surgery, for patients registered on a wait list Abbreviations: OR = odds ratio; CI = conﬁdence interval. Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Page 12 of 15 92 93 94 95 96 97 98 99 00 01 02 03 04 05 0 1 2 3 4 5 6 7 8 Calendar period of decision Time between decision and surgery (days) Figure 4 Time between decision to operate and surgical revascularization among patients who accessed surgery through direct admission, by calendar period of decision. Bottom of bar = 50th percentile, top of bar = 80th percentile. Figure 4 Time between decision to operate and surgical revascularization among patients who accessed surgery through direct admission, by calendar period of decision. Bottom of bar = 50th percentile, top of bar = 80th percentile. surgery rate in 1996–1997, the daily odds of surgery were lower prior to 1996–1997, but higher after this period (Table 5). semiurgent group; the weekly rate of registration was 16% and 25% higher in 2000–2001 and 2002–2003 compared to 1996–1997. In the urgent group, these rates were 2.5 and 1.6 times higher in 2002–2003 and 2004–2005 com- pared to 1996–1997. The weekly rates of decision were not diﬀerent across calendar periods in the nonurgent group, after adjustment. Discussion Table 5 Average daily rate of coronary artery bypass procedure in relation to calendar period of decision to proceed with surgery, for direct admissions Table 5 Average daily rate of coronary artery bypass procedure in relation to calendar period of surgery, for direct admissions e of coronary artery bypass procedure in relation to calendar period of decision to proceed with i aily rate of coronary artery bypass procedure in relation to calendar period of decision to proceed wi admissions arterial lesions on the coronary angiogram, then a car- diac surgeon is consulted to assess the patients’ suitability for CABG. Patients are transferred to an in-patient ward directly from the catheterization laboratory if expedited assessment is necessary and, if deemed suitable, these patients wait for the operation in hospital without regis- tration on a wait list. Alternatively, a consultation with the surgeon is scheduled at a later date. Surgeons register on their wait lists patients who need CABG and for whom the operation can be safely delayed. To address this issue, we studied access to surgical coronary revascularization for non-emergency patients through direct admission to hos- pital at the surgeon’s discretion, and contrasted the total amount of time between cardiac catheterization and sur- gical revascularization for the two pathways: through a wait list and through direct admission. to surgery could have been managed diﬀerently for com- bined procedures than for isolated CABG. As a result, we did not consider data from 312 (2%) wait-listed patients and 211 (1%) directly-admitted patients whose procedure was not isolated CABG. We found that during two periods when supplemen- tary funding was available, 1998–1999 and 2004–2005, the weekly rate of undergoing surgery from a wait list was, respectively, 50% and 90% higher than in 1996–1997, the period with the longest waiting times. We also observed a reduction in the diﬀerence in 90th and 50th percentiles of the waiting-time distributions. Forty percent of patients in the 1998, 1999, 2004 and 2005 cohorts (years when supplementary funding was provided) underwent surgery within 16 to 20 weeks following the median waiting time, while it took between 27 and 37 weeks for the cohorts reg- istered in the years when supplementary funding was not available. Among patients who were directly admitted to hospital, time between decision and surgery was longest in 1992–1993 and then steadily decreased through the late nineties. Discussion Coronary revascularization is indicated to alleviate chest pain and to reduce the risk of death among patients who have limiting angina that persists despite optimal medical treatment and who have coronary anatomy that is suitable for the procedure. However, in healthcare systems that use wait lists to manage access to care, patients requir- ing non-emergency surgical revascularization may have to wait after the decision to operate. In this paper, we sought to determine whether the year of decision to pro- ceed with non-emergency CABG had an eﬀect on time to surgery in a health system with publicly-funded uni- versal coverage for the procedure. We estimated temporal changes in the length of time that patients had to wait between decision to operate and the procedure itself over 14 years that included several years with increases in fund- ing. We focused on isolated CABG surgery because access Page 13 of 15 Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Table 5 Average daily rate of coronary artery bypass procedure in relation to calendar period of decision to proceed with surgery, for direct admissions Type of access by No. of No. of Total Crude rate† Crude OR‡ Adjusted OR‡§∥ calendar period of decision patients procedures waiting time∗ (95% CI) (95% CI) (95% CI) 1992–1993 1204 1003 4228 23.7 (22.3–25.2) 0.7 (0.6–0.7) 0.7 (0.6–0.8) 1994–1995 1474 1270 4489 28.3 (26.7–29.8) 0.8 (0.7–0.9) 0.9 (0.8–0.9) 1996–1997 1621 1442 4173 34.6 (32.8–36.3) 1.0 1.0 1998–1999 2089 1920 5301 36.2 (34.6–37.8) 1.1 (1.0–1.2) 1.2 (1.1–1.3) 2000–2001 2121 1895 5579 34.0 (32.4–35.5) 1.0 (0.9–1.1) 1.2 (1.1–1.3) 2002–2003 1940 1740 5145 33.8 (32.2–35.4) 1.0 (0.9–1.1) 1.1 (1.0–1.2) 2004–2005 2369 2140 6206 34.5 (33.0–35.9) 1.0 (1.0–1.1) 1.2 (1.1–1.3) All periods 12818 11410 35121 32.5 (31.9–33.1) – – Abbreviations: OR = odds ratio; CI = conﬁdence interval. To measure waiting times, patients were followed for a maximum of 7 days. †Per 100 patient days calculated as the number of procedures performed divided by the sum of waiting times. ‡As measured by odds ratios derived from discrete-time survival models, adjusting for consecutive days of waiting (0–7 days). §Adjusted for sex, age at decision, urgency at decision, institution at decision, comorbidities at decision, coronary anatomy at decision, and time between catheterization and decision. ∥1 patient with unknown value for sex was excluded. Competing interests p g 14. Christenson JT, Simonet F, Schmuziger M: The inﬂuence of age on the outcome of primary coronary artery bypass grafting. J Cardiovasc Surg (Torino) 1999, 40(3):333–338. 14. Christenson JT, Simonet F, Schmuziger M: The inﬂuence of age on the outcome of primary coronary artery bypass grafting. J Cardiovasc Surg (Torino) 1999, 40(3):333–338. Discussion The rate of surgery among patients directly admitted to hospital was the highest in 1998–1999, and has not changed afterwards, even in years when supple- mentary funding was provided. In this analysis, a potential concern is the misclassi- ﬁcation of the recorded urgency for treatment, because surgeons may manage access to surgery on the basis of various considerations, such as the best use of operating time or the availability of hospital resources. Therefore, the outcome might have been inﬂuenced by the individual surgeon’s threshold for accepting a patient for nonurgent treatment. It is plausible that the time to surgery may diﬀer between patients treated by surgeons with a high volume of CABG procedures and surgeons who perform a diverse range of cardiac procedures. The most important contribution of this analysis is providing a more complete picture of access times for the patient population requiring surgical revasculariza- tion on a non-emergency basis in a health care system that budgets the number of CABG procedures and uses sup- plementary funding to reduce the number patients who have to wait for the procedure and their waiting times. We contrasted two pathways for accessing CABG. If angio- plasty is not indicated when the cardiologist evaluates the We did not have access to detailed information about physicians’ decision-making on access to the procedure. To explore further the eﬀect of the supplementary fund- ing, we classiﬁed all patients using the algorithm devel- oped by Northern New England Cardiovascular Disease Study Group [22]. The percentage of low risk patients Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Page 14 of 15 Page 14 of 15 accessing surgery through direct admission declined con- siderably in years when supplementary funding was pro- vided. and cardiologists in the participating hospitals for their eﬀorts to ensure the completeness and accuracy of the registry data. and cardiologists in the participating hospitals for their eﬀorts to ensure the completeness and accuracy of the registry data. References CMAJ 2001, 165(4):429–433. 8. Levy A, Sobolev B, Hayden R, Kiely M, FitzGerald M, Schechter M: Time on wait lists for coronary bypass surgery in British Columbia, Canada, 1991 - 2000. BMC Health Services Res 2005, 5:22. 9. Sobolev BG, Levy AR, Kuramoto L, Hayden R: Chances of late surgery in relation to length of wait lists. BMC Health Services Res 2005, 5:63. 10. Sobolev B, Levy A, Hayden R, Kuramoto L: Does wait-list size at registration inﬂuence time to surgery? analysis of a population-based cardiac surgery registry. Health Services Res 2006, 41:23–49. Received: 14 December 2011 Accepted: 30 August 2012 Published: 11 September 2012 Received: 14 December 2011 Accepted: 30 August 2012 Published: 11 September 2012 Since 2002, percutaneous coronary intervention has become an increasingly common method of coronary revascularization, leading to a considerable change in the composition of patient population for both catheter-based and surgical procedures. We only had data for the period before 2005, and therefore our analysis could not adjust for changes in the proportional use of surgical revascular- ization over the past decade. Author’s contributions S d h d BS conceived the study concept and design, participated in analysis and interpretation, and drafted the manuscript. GF participated in data acquisition and critically revised the manuscript. LK participated in analysis and interpretation, and drafted the manuscript. AL participated in data acquisition. RS performed database analysis and has been involved in drafting the manuscript. BR performed statistical analysis and drafted the manuscript. All authors read and approved the ﬁnal manuscript. 15. O’Connor NJ, Morton JR, Birkmeyer JD, Olmstead EM, O’Connor GT: Eﬀect of coronary artery diameter in patients undergoing coronary bypass surgery. Northern New England Cardiovascular Disease Study Group. Circulation 1996, 93(4):652–655. 16. Naylor CD, Baigrie RS, Goldman BS, Basinski A: Assessment of priority for coronary revascularisation procedures. Lancet 1990, 335(8697):1070–1073. 17. DeLong ER, Nelson CL, Wong JB, Pryor DB, Peterson ED, Lee KL, Mark DB, CaliﬀRM, Pauker SG: Using observational data to estimate prognosis: an example using a coronary artery disease registry. Stat Med 2001, 20(16):2505–2532. Conclusions Our study provides evidence that waiting times for non- emergency coronary bypass surgery shortened after sup- plementary funding was granted to increase volume of cardiac surgical care in a health system with publicly- funded universal coverage for the procedure. The eﬀect of the supplementary funding was not uniform for patients that access the services through wait lists and through direct admissions. This might indicate that surgical ser- vices have used supplementary funding and direct admis- sions as two independent mechanisms to provide more treatments without delay. Considering that the hospitals had capacity to increase the number of operations, the supply-side initiatives indeed were eﬀective in reducing waiting times. Perhaps it was an empirical way to ﬁnd the level to budget the number of surgeries in the well deﬁned population. In our view, some further options for improving access to cardiac care should include policies for eﬀective management of patient ﬂow. g 10. Sobolev B, Levy A, Hayden R, Kuramoto L: Does wait-list size at registration inﬂuence time to surgery? analysis of a population-based cardiac surgery registry. Health Services Res 2006, 41:23–49. g 10. Sobolev B, Levy A, Hayden R, Kuramoto L: Does wait-list size at registration inﬂuence time to surgery? analysis of a population-based cardiac surgery registry. Health Services Res 2006, 41:23–49. 11. Chamberlayne R, Green B, Barer ML, Hertzman C, Lawrence WJ, Sheps SB: Creating a population-based linked health database: a new resource for health services research. Can J Public Health 1998, 89(4):270–273. 11. Chamberlayne R, Green B, Barer ML, Hertzman C, Lawrence WJ, Sheps SB: Creating a population-based linked health database: a new resource for health services research. Can J Public Health 1998, 89(4):270–273. y Creating a population-based linked health database: a new resource for health services research. Can J Public Health 1998, 89(4):270–273. 12. Blake JT, Carter MW: Surgical process scheduling: a structured review. J Soc Health Syst 1997, 5(3):17–30. 12. Blake JT, Carter MW: Surgical process scheduling: a structured review. J Soc Health Syst 1997, 5(3):17–30. 12. Blake JT, Carter MW: Surgical process scheduling: a structured review. J Soc Health Syst 1997, 5(3):17–30. 13. Sobolev B, Kuramoto L: Analysis of Waiting-Time Data in Health Services Research. New York, NY: Springer; 2008. y , ( ) 13. Sobolev B, Kuramoto L: Analysis of Waiting-Time Data in Health Services Research. New York, NY: Springer; 2008. 13. Sobolev B, Kuramoto L: Analysis of Waiting-Time Data in Health Services Research. References 1. Rihal CS, Raco DL, Gersh BJ, Yusuf S: Indications for coronary artery bypass surgery and percutaneous coronary intervention in chronic stable angina: review of the evidence and methodological considerations. Circulation 2003, 108(20):2439–2445. 2. Naylor CD: A diﬀerent view of queues in Ontario. Health Aﬀairs 1991, 10(3):110–128. 3. Fierlbeck K: Health Care in Canada: A Citizen’s Guide to Policy and Politics. Toronto: University of Toronto Press; chap. Funding Health Care; 2011: 3–43. 3. Fierlbeck K: Health Care in Canada: A Citizen’s Guide to Policy and Politics. Toronto: University of Toronto Press; chap. Funding Health Care; 2011: 3–43. 4. Naylor CD, Levinton CM, Baigrie RS: Adapting to waiting lists for coronary revascularization. Do Canadian specialists agree on which patients come ﬁrst? Chest 1992, 101(3):715–722. 5. Morgan CD, Sykora K, Naylor CD: Analysis of deaths while waiting for cardiac surgery among 29,293 consecutive patients in Ontario, Canada. Heart 1998, 79(4):345–349. 6. Ray AA, Buth KJ, Sullivan JA, Johnstone DE, Hirsch GM: Waiting for cardiac surgery: results of a risk-stratiﬁed queuing process. Circulation 2001, 104(12 Suppl 1):I92–I98. 7. Sampalis J, Boukas S, Liberman M, Reid T, Dupuis G: Impact of waiting time on the quality of life of patients awaiting coronary artery bypass grafting. CMAJ 2001, 165(4):429–433. 8. Levy A, Sobolev B, Hayden R, Kiely M, FitzGerald M, Schechter M: Time on wait lists for coronary bypass surgery in British Columbia, Canada, 1991 - 2000. BMC Health Services Res 2005, 5:22. 9. Sobolev BG, Levy AR, Kuramoto L, Hayden R: Chances of late surgery in relation to length of wait lists. BMC Health Services Res 2005, 5:63. 10. Sobolev B, Levy A, Hayden R, Kuramoto L: Does wait-list size at registration inﬂuence time to surgery? analysis of a population-based cardiac surgery registry. Health Services Res 2006, 41:23–49. 4. Naylor CD, Levinton CM, Baigrie RS: Adapting to waiting lists for coronary revascularization. Do Canadian specialists agree on which patients come ﬁrst? Chest 1992, 101(3):715–722. 5. Morgan CD, Sykora K, Naylor CD: Analysis of deaths while waiting for cardiac surgery among 29,293 consecutive patients in Ontario, Canada. Heart 1998, 79(4):345–349. 6. Ray AA, Buth KJ, Sullivan JA, Johnstone DE, Hirsch GM: Waiting for cardiac surgery: results of a risk-stratiﬁed queuing process. Circulation 2001, 104(12 Suppl 1):I92–I98. 7. Sampalis J, Boukas S, Liberman M, Reid T, Dupuis G: Impact of waiting time on the quality of life of patients awaiting coronary artery bypass grafting. Conclusions New York, NY: Springer; 2008. Competing interests The authors declare that they have no competing interests. Author details More research is needed to evaluate whether waiting times for non-emergency surgery vary because of chance alone after adjustment for clinical factors and variation in supply. For example, it remains unclear whether directly admitting patients of low risk is done to circumvent long wait lists, or to substitute for cancellations on the operat- ing room schedule. 1School of Population and Public Health, The University of British Columbia, Vancouver, BC, Canada. 2Department of Surgery, The University of British Columbia, Vancouver, BC, Canada. 3Centre for Clinical Epidemiology and Evaluation, Vancouver Coastal Health Research Institute, Vancouver, BC, Canada. 4Utilitas Consulting, Vancouver, BC, Canada. 5Community Health & Epidemiology, Dalhousie University, Halifax, NS, Canada. Received: 14 December 2011 Accepted: 30 August 2012 Published: 11 September 2012 Acknowledgements This study received ﬁnancial support from the Canada Research Chairs Program (BS), the Canada Foundation for Innovation (BS, AL), the Michael Smith Foundation for Health Research (AL), the Vancouver Coastal Health Research Institute (BS, LK), and the St. Paul’s Hospital Foundation (AL, RS). None of the sponsors had a role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the paper for publication. We are indebted to nurses, cardiac surgeons 18. Romano PS, Roos LL, Jollis JG: Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: diﬀering perspectives. J Clin Epidemiol 1993, 46(10):1075–1079. 18. Romano PS, Roos LL, Jollis JG: Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: diﬀering perspectives. J Clin Epidemiol 1993, 46(10):1075–1079. p 19. King KM, Ghali WA, Faris PD, Curtis MJ, Galbraith PD, Graham MM, Knudtson ML: Sex diﬀerences in outcomes after cardiac 19. King KM, Ghali WA, Faris PD, Curtis MJ, Galbraith PD, Graham MM, Knudtson ML: Sex diﬀerences in outcomes after cardiac Page 15 of 15 Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 Sobolev et al. BMC Health Services Research 2012, 12:311 http://www.biomedcentral.com/1472-6963/12/311 catheterization - Eﬀect modiﬁcation by treatment strategy and time. Jama-J Am Med Assoc 2004, 291(10):1220–1225. 20. Bland JM, Altman DG: Survival probabilities (the Kaplan-Meier method). BMJ: Br Med J 1998, 317(7172):1572. 21. Allison PD: Discrete-time methods for the analysis of event histories. Sociological Methodology 1982, 13:61–98. 22. O’Connor GT, Plume SK, Olmstead EM, Coﬃn LH, Morton JR, Maloney CT, Nowicki ER, Levy DG, Tryzelaar JF, Hernandez F: Multivariate prediction of in-hospital mortality associated with coronary artery bypass graft surgery. Northern New England Cardiovascular Disease Study Group. Circulation 1992, 85:2110–2118. doi:10.1186/1472-6963-12-311 Cite this article as: Sobolev et al.: Evaluation of supply-side initiatives to improve access to coronary bypass surgery. BMC Health Services Research 2012 12:311. catheterization - Eﬀect modiﬁcation by treatment strategy and time. Jama-J Am Med Assoc 2004, 291(10):1220–1225. 20. Bland JM, Altman DG: Survival probabilities (the Kaplan-Meier method). BMJ: Br Med J 1998, 317(7172):1572. 21. Allison PD: Discrete-time methods for the analysis of event histories. Sociological Methodology 1982, 13:61–98. 22. O’Connor GT, Plume SK, Olmstead EM, Coﬃn LH, Morton JR, Maloney CT, Nowicki ER, Levy DG, Tryzelaar JF, Hernandez F: Multivariate prediction of in-hospital mortality associated with coronary artery bypass graft surgery. Acknowledgements Northern New England Cardiovascular Disease Study Group. Circulation 1992, 85:2110–2118. doi:10.1186/1472-6963-12-311 Cite this article as: Sobolev et al.: Evaluation of supply-side initiatives to improve access to coronary bypass surgery. BMC Health Services Research 2012 12:311. g gy 22. O’Connor GT, Plume SK, Olmstead EM, Coﬃn LH, Morton JR, Maloney CT, Nowicki ER, Levy DG, Tryzelaar JF, Hernandez F: Multivariate prediction of in-hospital mortality associated with coronary artery bypass graft surgery. Northern New England Cardiovascular Disease Study Group. Circulation 1992, 85:2110–2118. doi:10.1186/1472-6963-12-311 Cite this article as: Sobolev et al.: Evaluation of supply-side initiatives to improve access to coronary bypass surgery. BMC Health Services Research 2012 12:311. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit
https://openalex.org/W2283887526	https://europepmc.org/articles/pmc4664469?pdf=render	English	null	Protocol Improvements for Low Concentration DNA-Based Bioaerosol Sampling and Analysis	PloS one	2,015	cc-by	12,251	RESEARCH ARTICLE Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Funding: This research was funded by the Republic of Singapore's National Research Foundation through a grant to the Berkeley Education Alliance for Research in Singapore (BEARS) for the Singapore Berkeley Building Efficiency and Sustainability in the Tropics (SinBerBEST) Program. BEARS has been established by the University of California, Berkeley as a center for intellectual excellence in research and education in Singapore (http://sinberbest.berkeley. edu/). This research was also funded by Singapore's National Research Foundation, Award Number NRF- CRP8-2011-03; and Ministry of Education - Introduction As bioaerosol research attracts increasing attention, there is a need for additional efforts that focus on method development to deal with different environmental samples. Bioaerosol environmental samples typically have very low biomass concentrations in the air, which often leaves researchers with limited options in choosing the downstream analysis steps, especially when culture-independent methods are intended. Editor: Maosheng Yao, Peking University, CHINA Editor: Maosheng Yao, Peking University, CHINA Received: June 16, 2015 Accepted: October 3, 2015 Published: November 30, 2015 Copyright: © 2015 Luhung et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Editor: Maosheng Yao, Peking University, CHINA Received: June 16, 2015 Accepted: October 3, 2015 Published: November 30, 2015 Copyright: © 2015 Luhung et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Published: November 30, 2015 Published: November 30, 2015 OPEN ACCESS Citation: Luhung I, Wu Y, Ng CK, Miller D, Cao B, Chang VW-C (2015) Protocol Improvements for Low Concentration DNA-Based Bioaerosol Sampling and Analysis. PLoS ONE 10(11): e0141158. doi:10.1371/ journal.pone.0141158 Editor: Maosheng Yao, Peking University, CHINA Received: June 16, 2015 Accepted: October 3, 2015 Published: November 30, 2015 Copyright: © 2015 Luhung et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Citation: Luhung I, Wu Y, Ng CK, Miller D, Cao B, Chang VW-C (2015) Protocol Improvements for Low Concentration DNA-Based Bioaerosol Sampling and Analysis. PLoS ONE 10(11): e0141158. doi:10.1371/ journal.pone.0141158 Irvan Luhung1,2☯, Yan Wu1,2☯, Chun Kiat Ng3,4, Dana Miller3, Bin Cao2,3, Victor Wei-Chung Chang1,2☯* Irvan Luhung1,2☯, Yan Wu1,2☯, Chun Kiat Ng3,4, Dana Miller3, Bin Cao2,3, Victor Wei-Chung Chang1,2☯* 1 SinBerBEST Program, Berkeley Education Alliance for Research in Singapore, Singapore, 2 School of Civil and Environmental Engineering, Nanyang Technological University, Singapore, Singapore, 1 SinBerBEST Program, Berkeley Education Alliance for Research in Singapore, Singapore, 2 School of Civil and Environmental Engineering, Nanyang Technological University, Singapore, Singapore, 3 Singapore Centre on Environmental Life Sciences Engineering, Nanyang Technological University, Singapore, Singapore, 4 Interdisciplinary Graduate School, Nanyang Technological University, Singapore, Singapore 3 Singapore Centre on Environmental Life Sciences Engineering, Nanyang Technological University, Singapore, Singapore, 4 Interdisciplinary Graduate School, Nanyang Technological University, Singapore, Singapore ☯These authors contributed equally to this work. * wcchang@ntu.edu.sg * wcchang@ntu.edu.sg Objectives Copyright: © 2015 Luhung et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. This study investigates the impacts of three important factors that can influence the perfor- mance of culture-independent DNA-based analysis in dealing with bioaerosol environmen- tal samples engaged in this study. The factors are: 1) enhanced high temperature sonication during DNA extraction; 2) effect of sampling duration on DNA recoverability; and 3) an alternative method for concentrating composite samples. In this study, DNA extracted from samples was analysed using the Qubit fluorometer (for direct total DNA measurement) and quantitative polymerase chain reaction (qPCR). Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Introduction Singapore, Award Number MOE2013-T3-1-013. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Interest in bioaerosol research has grown rapidly over the past few decades, spurred by the advent of analytical methods based on DNA. There are numerous reasons for interest in mea- suring bioaerosols. Some research has been motivated by potential human health effects [1,2]. While airborne microorganisms can be benign or even beneficial to humans, [3,4], bioaerosols are also known to cause infections, allergenic responses and toxic effects ranging from rela- tively common asthma and sick building syndrome to epidemic diseases (such as SARS) and bioweapon use (such as Anthrax) in extreme cases [5–8]. Indirect impacts of bioaerosols, such as crops infection [9], biodegradation of building materials [10,11] or how it can be transported widely by the wind, affecting life along the way [12–14], have also attracted equal attention to understand the dynamics of bioaerosols around us. Such interest to measure or identify bioaer- osols in different environments, however, is seldom coupled with satisfactorily robust technical capabilities given the limitations of current culture-based and DNA-based sampling and analy- sis methods for filter and liquid samples. Researchers are faced with problems such as collect- ing enough biomass in a relatively short time (low time resolution)[15], low sampling efficiency or removing unique inhibitors from environmental samples. Competing Interests: The authors have declared that no competing interests exist. Competing Interests: The authors have declared that no competing interests exist. The road to developing the much needed standardized methods for bioaerosol research has been challenging mostly due to major technical limitations. First, bioaerosol concentrations are naturally dilute in the environment [13,16–19]. The low concentrations of interest lead to detection limits and sensitivity problems in subsequent analysis. The second issue is the fact that bioaerosols are highly dynamic in time and space [13,20–23]. This large variability makes it difficult to establish and compare protocols from various studies. Culture-based and microscopic methods dominated early bioaerosol research [2,20,24] and are still used today [25–28] because of their practicality and specificity. However, microscopic methods are generally labour intensive, while the culture based methods might carry a certain level of bias particularly when a broader range of microorganisms are targeted. DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples Singapore, Award Number MOE2013-T3-1-013. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Results and Findings The findings suggest that additional lysis from high temperature sonication is crucial: DNA yields from both high and low biomass samples increased up to 600% when the protocol included 30-min sonication at 65°C. Long air sampling duration on a filter media was shown to have a negative impact on DNA recoverability with up to 98% of DNA lost over a 20-h sampling period. Pooling DNA from separate samples during extraction was proven to be feasible with margins of error below 30%. 1 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 Introduction The culture based methods also require the captured organisms to be both viable and culturable on the selected medium, which only accounts for less than 5% of total organisms in the air [18,29,30]. Furthermore, some sampling approaches may cause some stress to the targeted microorgan- isms and lead them, to completely or partially, lose their viability during or after collection [16,31,32]. To complement culture-based assessments, culture-independent methods have been devel- oped. Several bioaerosol analysis protocols were built based on biological entities such as endo- toxin [33,34] or glucans [35]. While these methods have merit, DNA-based analysis, which uses PCR (polymerase chain reaction) as the core technology has emerged to assume the main role [16,17,36,37] in recent bioaerosol studies. Unlike other biological entities, DNA is ubiqui- tous in all living things. This advantage has driven a variety of DNA-based technology, such as time-saving commercial DNA extraction kits (MOBIO, Qiagen, etc.), DNA quantification devices (Nanodrop, Qubit, qPCR), and advanced DNA sequencing technologies [38–40] to flourish with a trend towards increased technical capabilities and rapidly decreasing unit costs. The focus of this study is to deal with challenges associated with low bioaerosol concentra- tion from complex environmental samples, utilizing culture-independent, DNA-based analysis methods. Previous studies have commonly applied three approaches to optimize DNA yield from low biomass samples, which include: (1) extended sampling with high flowrate or long duration [16,41]; (2) improving the sample extraction process [14,42,43]; and (3) attempting to concentrate samples [8]. However, the use of these approaches is often not complemented with details on to what extent or how these efforts directly affect final results. This study aims to provide a thorough investigation of how the following three factors may improve DNA yield 2 / 18 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples from bioaerosol environmental samples: (i) the effect of thermal sonication on DNA yield; (ii) the effect of long-duration filter sampling on nonviable DNA based analysis; and (iii) an alter- native method to pool several low concentration samples; may improve DNA yield from bioaerosol environmental samples. Sampling We tested the approaches proposed in this study on two types of environmental samples with different biomass loadings. Filter-based ambient air samples were collected to represent low biomass samples. Dust extracted from used mechanical-ventilation-system filters represents high biomass samples. The latter filters were obtained during regular air-handling unit (AHU) filter replacement in buildings of the Nanyang Technological University (NTU), Singapore. All AHU filter samples were collected with permission from NTU facilities management office and the servicing company (SMM Pte. Ltd., Singapore). An open balcony at NTU was chosen as the location for all ambient air sampling in this study. Human contribution at that site is minimal except for brief times when researchers came to conduct sampling activity. Temperature and relative humidity (RH) measurements were routinely recorded along with filter sampling. The time table of ambient air sampling activities can be obtained from S1 File. All samples were collected during summer 2013 by means of filtering air (without any size-cut) onto a poly- ethersulphone (PES) filter membrane (47mm, 0.2 μm pore size, Pall Corporation, USA) using a diaphragm vacuum pump (HCS Scientific, Singapore) with a flow rate of 16 L/min for 8 to 20 hours. Filters from three AHUs at NTU were acquired from regular maintenance activities. These filters were in service for three months prior to acquisition. Details regarding filter collection can also be seen in S1 File. These filters are indoor secondary units, which come after the pri- mary filter that treats only outdoor air. The secondary filter served as a filter for a mixture of both recirculated indoor air plus make-up outdoor air. Materials and Methods Investigations of bioaerosol sampling, DNA extraction and analysis are the key components for this study. First, we examined the effect of additional lysis to enhance DNA extraction. Sec- ond, we investigated the impact of sampling duration on the recoverability of DNA from filters. Third, we investigated a DNA-pooling approach to overcome the analytical limitations of low DNA concentrations in environmental bioaerosol samples. DNA Extraction The DNA extraction of one set of samples followed the original MOBIO PW protocol, while the other used additional high temperature sonication. The final DNA yields were analysed with Qubit and qPCR. fluorometry (Qubit, Life Technologies, USA) and gene copy numbers of bacterial (16S) and fungal (18S) marker genes were estimated by qPCR (LC480, Roche Scientific, Switzerland). W f h i d h ff f i i d h i b i i h We further examined the effect of sonication and heat incubation using the temperature from the previous experiment on two additional sets of environmental samples: ambient air samples (low biomass) and AHU filter samples (high biomass). The DNA extraction of one set of samples followed the original MOBIO PW protocol, while the other used additional high temperature sonication. The final DNA yields were analysed with Qubit and qPCR. Sampling Duration Experiment Two experiments shown in Fig 1 were carried out to assess the effect of sampling duration on extracted DNA results. Fig 1A illustrates the first experiment, which employed a pair of identi- cal sampling trains operated in parallel. With one train, the air was sampled continuously for 24 hours using a single filter. The second train also operated for 24 hours, but the filters were replaced with new ones at eight-hour intervals so that three sequential filters were collected in all. All four filters were then processed. The one 24-hour filter was extracted alone, while the DNA extracted from three 8-hour filters was pooled into a single DNA solution. Qubit mea- surements and qPCR were then applied to the final DNA solutions. The purpose of this investi- gation was to assess whether there is any DNA loss associated with long-duration filter-based sampling. If DNA degrades with extended sampling, then the yield recovered from three sepa- rate filters would be higher in sum than the single 24-h filter, since these sequential filters are each exposed to a lesser duration of air sampling. Fig 1B illustrates another experiment, which was performed utilizing a species of Gram-neg- ative bacteria which belongs to the Phylum Proteobacteria. Green-fluorescent protein (GFP)- tagged Gram-negative bacteria Shewanella oneidensis [44] was chosen for this experiment as this species is unlikely to present in the ambient air. Quanta of these bacteria were spiked onto filter samples at different times during ambient-air bioaerosol sampling, exposing the Fig 1. Sampling duration experiments for microorganisms collected on filters. (a)–Comparison of 1×24 hour versus 3×8 hour parallel sampling. Two sets of sampling trains collected air at the same time and location for 24 hours. The first set sampled the air continuously for 24 hours on a single filter, while the other set’s filter was replaced every 8 hours. DNA extracted from the three 8-h filters was then pooled and the result is compared with the first set which was extracted alone. (b)–Testing loss of GFP-tagged bacteria during sampling. A set of filters sampled air at the same time and location for a total duration of 20 hours. Known quanta of GFP-tagged S. oneidensis cells were spiked onto the filters at staggered timepoints, exposing them to different durations of air sampling. Fig 1. Sampling duration experiments for microorganisms collected on filters. (a)–Comparison of 1×24 hour versus 3×8 hour parallel sampling. DNA Extraction MOBIO Power Water (PW, MOBIO Carlsbad, USA) DNA extraction kits were used for all extractions. The ambient air filters were directly put into the 5-ml bead beating tube of the PW kits to proceed with DNA extraction using flame sterilized tweezers and working under a bio- safety hood to reduce potential contamination. The AHU filters were first cut into smaller pieces (1 cm × 5 cm per piece) and then directly placed in the bead beating tube for extraction. We first conducted an experiment to determine the optimum temperature for water-bath sonication. The DNA extraction followed the original MOBIO PW protocol with an additional 30 minutes of water bath sonication and thermal incubation at different temperatures (Elmaso- nic, SH250EL, Germany) before the bead-beating step. The temperature settings include no- treatment (original MOBIO PW protocol), water-bath sonication only (no heating), 55°C, 60°C, 65°C, 70°C and 75°C sonication. Total extracted DNA was then directly measured by MOBIO Power Water (PW, MOBIO Carlsbad, USA) DNA extraction kits were used for all extractions. The ambient air filters were directly put into the 5-ml bead beating tube of the PW kits to proceed with DNA extraction using flame sterilized tweezers and working under a bio- safety hood to reduce potential contamination. The AHU filters were first cut into smaller pieces (1 cm × 5 cm per piece) and then directly placed in the bead beating tube for extraction. We first conducted an experiment to determine the optimum temperature for water-bath sonication. The DNA extraction followed the original MOBIO PW protocol with an additional 30 minutes of water bath sonication and thermal incubation at different temperatures (Elmaso- nic, SH250EL, Germany) before the bead-beating step. The temperature settings include no- treatment (original MOBIO PW protocol), water-bath sonication only (no heating), 55°C, 60°C, 65°C, 70°C and 75°C sonication. Total extracted DNA was then directly measured by 3 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples fluorometry (Qubit, Life Technologies, USA) and gene copy numbers of bacterial (16S) and fungal (18S) marker genes were estimated by qPCR (LC480, Roche Scientific, Switzerland). We further examined the effect of sonication and heat incubation using the temperature from the previous experiment on two additional sets of environmental samples: ambient air samples (low biomass) and AHU filter samples (high biomass). doi:10.1371/journal.pone.0141158.g001 Sampling Duration Experiment Two sets of sampling trains collected air at the same time and location for 24 hours. The first set sampled the air continuously for 24 hours on a single filter, while the other set’s filter was replaced every 8 hours. DNA extracted from the three 8-h filters was then pooled and the result is compared with the first set which was extracted alone. (b)–Testing loss of GFP-tagged bacteria during sampling. A set of filters sampled air at the same time and location for a total duration of 20 hours. Known quanta of GFP-tagged S. oneidensis cells were spiked onto the filters at staggered timepoints, exposing them to different durations of air sampling. Fig 1. Sampling duration experiments for microorganisms collected on filters. (a)– doi:10.1371/journal.pone.0141158.g001 PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 4 / 18 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples Shewanella cells to airstreams for different sampling durations. Filter 1 was spiked at 0 h (beginning of sampling), filter 2 at 8 h, filter 3 at 14 h, filter 4 at 20 h (i.e., after completion of sampling) and one set of filters was processed without any spike as the blank. Utilizing infor- mation from a preliminary experiment, we spiked the filters with an amount of cells that is comparable to the total DNA mass normally collected during 20 hours air sampling at this par- ticular location, i.e., 106 cells counts. After sampling, DNA was extracted using the same proto- col described previously. A qPCR analysis was also performed using custom designed GFP- sequence primers and probes. Shewanella cells to airstreams for different sampling durations. Filter 1 was spiked at 0 h (beginning of sampling), filter 2 at 8 h, filter 3 at 14 h, filter 4 at 20 h (i.e., after completion of sampling) and one set of filters was processed without any spike as the blank. Utilizing infor- mation from a preliminary experiment, we spiked the filters with an amount of cells that is comparable to the total DNA mass normally collected during 20 hours air sampling at this par- ticular location, i.e., 106 cells counts. After sampling, DNA was extracted using the same proto- col described previously. A qPCR analysis was also performed using custom designed GFP- sequence primers and probes. DNA Concentration / Pooling Approach A concentration approach is proposed in this study to investigate whether more biomass could be gathered from several low concentration samples without necessitating amplification either by PCR or via culturing. Fig 2 illustrates an approach to concentrate DNA in the middle of DNA extraction process. To test the concept, four sets of AHU filters and ambient air samples, consisting of four rep- licates each (32 in total) were prepared and extracted in this manner. For each set of four filters, the DNA solutions from three filters were introduced onto the same spin filter after the lysis and inhibitor removal step, whereas the remaining one filter was extracted alone. The extracted DNA concentrations were used to verify whether the expected 3-to-1 ratios were obtained for both high and low biomass samples. Fig 2. Concentration approach during extraction. The DNA from three filters is pooled during the spin- filter binding step of DNA extraction. The total DNA yield is then compared to the one filter extracted alone to investigate whether the expected 3-to-1 ratio is obtained. doi:10.1371/journal.pone.0141158.g002 Fi 2 C t ti h d i t ti Th DNA f th filt i l d d i th i Fig 2. Concentration approach during extraction. The DNA from three filters is pooled during the spin- filter binding step of DNA extraction. The total DNA yield is then compared to the one filter extracted alone to investigate whether the expected 3-to-1 ratio is obtained. doi:10.1371/journal.pone.0141158.g002 5 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples Results Sonication and Thermal Incubation DNA Measurement: Qubit and qPCR These are sequences from 16S region for general bacteria, 18S region [45] for general fungi and the GFP gene for the experiments involving Shewanella oneidensis Table 1 lists the detailed sequences of the afore- Three sets of primers and probes were used for the qPCR assay. These are sequences from 16S region for general bacteria, 18S region [45] for general fungi and the GFP gene for the experiments involving Shewanella oneidensis. Table 1 lists the detailed sequences of the afore- mentioned primers and probes, each of which was synthesized by TIB Molbiol (Berlin, Germany). Specific to the DNA pooling performance test and the impact of sampling duration on the Gram-negative GFP-tagged bacteria test, part of the data is reported in terms of total recovered DNA (ng or pg) as the purpose of the experiments was to investigate the amount of DNA recovered regardless of concentration. To obtain these measures, the DNA concentrations measured by both Qubit and qPCR (ng/μL and pg/μL) were multiplied by the elution volume used during extraction to obtain the final extracted DNA mass for each sample (ng for Qubit and pg for qPCR). DNA Measurement: Qubit and qPCR DNA fluorometry (Qubit, Life Technologies, USA) and quantitative PCR (Roche LC 480, Basel, Switzerland) were applied to quantify and verify the quality of the extracted DNA solu- tions. The double stranded (ds) DNA high sensitivity kit was used for all of the Qubit assays. Qubit measures total DNA concentration in terms of ng of DNA per μL of elution liquid (ng/μL). The total mass of DNA (ng) is then calculated by multiplying the concentration with the amount of elution liquid used during extraction. For low concentration air samples, the derived DNA concentration is expressed in terms of ng of DNA per volume of air sampled through the 47-mm PES filter (ng/m3 of air). For the higher biomass AHU filter samples, the final total DNA concentration is expressed in terms of total extracted DNA per area of filter segmented for extraction (ng/cm2 of filter). Quantitative PCR was also carried out to estimate concentrations of DNA associated with specific microbiological targets, namely bacteria, fungi and GFP. Quantitative PCR is based on the concept of DNA replication, doubling the number of targeted DNA sequences every cycle. The number of cycles needed for each sample to cross a set point is recorded as the Cp (cross- ing point) value. Roche LC 480 uses second derivative maximum analysis method which rec- ords the point where acceleration of the fluorescence signal is at its highest as its set point to determine the Cp value. The lower the Cp value, the higher the amount of targeted DNA sequence in the sample as fewer cycles of DNA duplication are needed to reach the set point. In addition, each Cp unit reflects a doubling of DNA abundance, approximately a doubling of mass. Standard curves (Cp value vs. DNA mass concentration) were established for general bacte- ria, fungi and GFP based on Qubit measurements of extracted pure bacteria (E. coli), fungus (Aspergillus sp.) and GFP-tagged Shewanella oneidensis, which were chosen as equivalent representatives (S1 File). All the Cp values were then converted to partial DNA concentration (pg/μL) based on their respective standard curves. Similar to the Qubit result calculations, the partial DNA concentrations were finally converted to pg of DNA per m3 of air (pg/m3) for ambient air samples and to pg of DNA per area of filter (pg/cm2) for AHU filter samples. Three sets of primers and probes were used for the qPCR assay. The qPCR result on the same set of samples (Fig 3B) shows a similar trend. The bacterial DNA concentration rose from 51 pg/cm2 to 320 pg/cm2 with incubation at 65°C. The bacterial DNA yield then declined to 262 pg/cm2 with incubation at 70°C and 168 pg/cm2 with incubation at 75°C. Similarly, the highest concentration of fungal DNA (1290 pg/cm2) was also obtained from samples incubated at 65°C. The concentration of fungal DNA remained rela- tively constant for samples incubated at 70°C before decreasing to 914 pg/cm2 at 75°C. Based on these results, 65°C was chosen as the incubation temperature for the remaining investigations in this study. To further evaluate the applicability of sonication and thermal incubation to a broader range of biomass concentrations, these parameters were applied to two sets of environmental samples: a set of AHU filters collected from a different building (high biomass samples) and a set of ambient air filters (low biomass samples). temperature rises. The total DNA yield shown in Fig 3A was 1.48 ng/cm2 with no treatment, increasing to 4.56 ng/cm2 and 5.22 ng/cm2 with incubation treatment at 65°C and 70°C, respectively. The qPCR result on the same set of samples (Fig 3B) shows a similar trend. The bacterial DNA concentration rose from 51 pg/cm2 to 320 pg/cm2 with incubation at 65°C. The bacterial DNA yield then declined to 262 pg/cm2 with incubation at 70°C and 168 pg/cm2 with incubation at 75°C. Similarly, the highest concentration of fungal DNA (1290 pg/cm2) was also obtained from samples incubated at 65°C. The concentration of fungal DNA remained rela- tively constant for samples incubated at 70°C before decreasing to 914 pg/cm2 at 75°C. Based on these results, 65°C was chosen as the incubation temperature for the remaining investigations in this study. To further evaluate the applicability of sonication and thermal incubation to a broader range of biomass concentrations, these parameters were applied to two sets of environmental samples: a set of AHU filters collected from a different building (high biomass samples) and a set of ambient air filters (low biomass samples). The total DNA extracted from both sets of environmental samples increased substantially with the addition of sonication at 65°C, as shown in Fig 4A and 4B. Sonication and Thermal Incubation The effects of sonication and thermal incubation on the final DNA yield are reported in terms of total DNA per cm2 area of filter or per m3 of air (as determined by Qubit), and as targeted DNA sequence per cm2 area of filter or per m3 of air (as determined by qPCR). Fig 3 shows that the DNA yields of the tested AHU filter samples increases gradually as the incubation 6 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples Table 1. Probe-and-primer sets for qPCR assays. Function Sequence Forward primer, 16S (bacteria) 5’–ACTCCTACGGGAGGCAG–3’ BAC338F Reverse primer, 16S (bacteria) 5’–GACTACCAGGGTATCTAATC–3’ BAC805R Taqman Probe, 16S (bacteria) 6FAM–TGCCAGCAGCCGCGGTAATAC–3’–BBQ BAC516F Forward primer, 18S (fungi) 5’–GGRAAACTCACCAGGTCCAG–3’ FungiQuant-F Reverse primer, 18S (fungi) 5’–GSWCTATCCCCAKCACGA–3’ FungiQuant-R Taqman Probe, 18S (fungi) 6FAM–TGGTGCATGGCCGTT–3’–BBQ FungiQuant-PrbLNA Forward primer, GFP gene 5’–ATGGAAACATTCTTGGACACAAATTG–3’ GFP-S Reverse primer, GFP gene 5’–GTTGATAATGGTCTGCTAGTTGAACG–3’ GFP-R Taqman Probe, GFP gene 6FAM–TCCATTCTTTTGTTTGTCTGCCATGATGT–BBQ GFP-TM doi:10 1371/journal pone 0141158 t001 Table 1. Probe-and-primer sets for qPCR assays. temperature rises. The total DNA yield shown in Fig 3A was 1.48 ng/cm2 with no treatment, increasing to 4.56 ng/cm2 and 5.22 ng/cm2 with incubation treatment at 65°C and 70°C, respectively. The qPCR result on the same set of samples (Fig 3B) shows a similar trend. The bacterial DNA concentration rose from 51 pg/cm2 to 320 pg/cm2 with incubation at 65°C. The bacterial DNA yield then declined to 262 pg/cm2 with incubation at 70°C and 168 pg/cm2 with incubation at 75°C. Similarly, the highest concentration of fungal DNA (1290 pg/cm2) was also obtained from samples incubated at 65°C. The concentration of fungal DNA remained rela- i l f l i b d 70°C b f d i 914 / 2 75°C Table 1. Probe-and-primer sets for qPCR assays. Function Sequence Forward primer, 16S (bacteria) 5’–ACTCCTACGGGAGGCAG–3’ BAC338F Reverse primer, 16S (bacteria) 5’–GACTACCAGGGTATCTAATC–3’ BAC805R Taqman Probe, 16S (bacteria) 6FAM–TGCCAGCAGCCGCGGTAATAC–3’–BBQ BAC516F Forward primer, 18S (fungi) 5’–GGRAAACTCACCAGGTCCAG–3’ FungiQuant-F Reverse primer, 18S (fungi) 5’–GSWCTATCCCCAKCACGA–3’ FungiQuant-R Taqman Probe, 18S (fungi) 6FAM–TGGTGCATGGCCGTT–3’–BBQ FungiQuant-PrbLNA Forward primer, GFP gene 5’–ATGGAAACATTCTTGGACACAAATTG–3’ GFP-S Reverse primer, GFP gene 5’–GTTGATAATGGTCTGCTAGTTGAACG–3’ GFP-R Taqman Probe, GFP gene 6FAM–TCCATTCTTTTGTTTGTCTGCCATGATGT–BBQ GFP-TM doi:10.1371/journal.pone.0141158.t001 Table 1. Probe-and-primer sets for qPCR assays. Table 1. Probe-and-primer sets for qPCR assays. Function Sequence Forward primer, 16S (bacteria) 5’–ACTCCTACGGGAGGCAG–3’ BAC338F Reverse primer, 16S (bacteria) 5’–GACTACCAGGGTATCTAATC–3’ BAC805R Taqman Probe, 16S (bacteria) 6FAM–TGCCAGCAGCCGCGGTAATAC–3’–BBQ BAC516F Forward primer, 18S (fungi) 5’–GGRAAACTCACCAGGTCCAG–3’ FungiQuant-F Reverse primer, 18S (fungi) 5’–GSWCTATCCCCAKCACGA–3’ FungiQuant-R Taqman Probe, 18S (fungi) 6FAM–TGGTGCATGGCCGTT–3’–BBQ FungiQuant-PrbLNA Forward primer, GFP gene 5’–ATGGAAACATTCTTGGACACAAATTG–3’ GFP-S Reverse primer, GFP gene 5’–GTTGATAATGGTCTGCTAGTTGAACG–3’ GFP-R Taqman Probe, GFP gene 6FAM–TCCATTCTTTTGTTTGTCTGCCATGATGT–BBQ GFP-TM doi:10 1371/journal pone 0141158 t001 5’–ACTCCTACGGGAGGCAG–3’ BAC338F 5’–GACTACCAGGGTATCTAATC–3’ BAC805R 6FAM–TGCCAGCAGCCGCGGTAATAC–3’–BBQ BAC516F 5’–GGRAAACTCACCAGGTCCAG–3’ FungiQuant-F 5’–GSWCTATCCCCAKCACGA–3’ FungiQuant-R 6FAM–TGGTGCATGGCCGTT–3’–BBQ FungiQuant-PrbLNA 5’–ATGGAAACATTCTTGGACACAAATTG–3’ GFP-S 5’–GTTGATAATGGTCTGCTAGTTGAACG–3’ GFP-R 6FAM–TCCATTCTTTTGTTTGTCTGCCATGATGT–BBQ GFP-TM doi:10.1371/journal.pone.0141158.t001 temperature rises. The total DNA yield shown in Fig 3A was 1.48 ng/cm2 with no treatment, increasing to 4.56 ng/cm2 and 5.22 ng/cm2 with incubation treatment at 65°C and 70°C, respectively. doi:10.1371/journal.pone.0141158.g003 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples Fig 4. Improving DNA yield with additional heat and sonication lysis. Additional sonication and thermal lysis show improved DNA yield for (a)—AHU filter samples and (b)–ambient air samples as measured by the Qubit fluorometer for total DNA (left bar, left axis) and by qPCR for bacterial (middle bar, right axis) and fungal (right bar, right axis) DNA (N = 4). Fig 4. Improving DNA yield with additional heat and sonication lysis. Additional sonication and thermal lysis show improved DNA yield for (a)—AHU filter samples and (b)–ambient air samples as measured by the Qubit fluorometer for total DNA (left bar, left axis) and by qPCR for bacterial (middle bar, right axis) and fungal (right bar, right axis) DNA (N = 4). doi:10.1371/journal.pone.0141158.g004 doi:10.1371/journal.pone.0141158.g004 samples increases from 0.03 ng/m3 to 0.44 ng/m3 (Fig 4B, left bar), more than a ten-fold increase. samples increases from 0.03 ng/m3 to 0.44 ng/m3 (Fig 4B, left bar), more than a ten-fold increase. As also seen in Fig 4A and 4B (middle and right bar), the qPCR results show a similar trend to the Qubit fluorometry data, displaying a substantial jump in DNA concentration after soni- cation and thermal incubation for both fungi and bacteria. Fungal DNA concentrations increased from 5.5 pg/m3 to 187 pg/m3 (34×) and from 557 pg/cm2 to 3280 pg/cm2 (6×) for ambient air and AHU filter samples, respectively. Bacterial DNA concentrations went from 1750 to 4490 pg/cm2 (2.6×) for the AHU filters samples and from 1.9 to 10.3 pg/m3 (5.5×) for the ambient air samples. Thus, sonication and thermal incubation was seen to be very effective for enhancing the amount of DNA extracted from both fungi and bacteria from air. Generally, the results show that such an additional processing step might be crucial for analysing low con- centration bioaerosol samples. PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 For the AHU filter samples, the additional sonication and heat incubation boosts the mean of total DNA yield from 14 to 31 ng/cm2, more than a two-fold increase (Fig 4A, left bar). The DNA yield of the ambient air Fig 3. DNA measurements of AHU filter samples for seven treatment pathways of sonication at varying temperature. (a)—Total DNA yields (Qubit). (b)–Results from bacterial and fungal qPCR. N = 4 for each case. * denotes statistically significant difference to mean of DNA yield extracted with 65°C incubation. Fig 3. DNA measurements of AHU filter samples for seven treatment pathways of sonication at varying temperature. (a)—Total DNA yields (Qubit). (b)–Results from bacterial and fungal qPCR. N = 4 for each case. * denotes statistically significant difference to mean of DNA yield extracted with 65°C incubation. doi:10.1371/journal.pone.0141158.g003 PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 7 / 18 a Time (t) during the 20-h sampling period when a spike of S. oneidensis cells was applied to the ﬁlter. b Mean ± standard deviation reading from qPCR analysis using GFP probe and primers Table 2. qPCR results from a series of 20-h samples illustrating the decay of GFP-tagged Shewanella oneidensis cells. Sample treatment a No. of replicates Estimated mass of GFP DNA recovered (ng) b Exposure to air sampling (hours) % of DNA retained (relative to t = 0) No spike 3 0.02 ± 0.001 - - Spike at t = 0 6 0.13 ± 0.02 20 1.3% Spike at t = 8 h 3 0.73 ± 0.01 12 7.1% Spike at t = 14 h 3 1.13 ± 0.02 6 11% Spike at t = 20 h 6 10.2 ± 0.3 0 100% a Time (t) during the 20-h sampling period when a spike of S. oneidensis cells was applied to the ﬁlter. b Mean ± standard deviation reading from qPCR analysis using GFP probe and primers doi:10.1371/journal.pone.0141158.t002 Effect of Sampling Duration on DNA Yield Two sets of experiments were conducted to investigate the effect of filter sampling on the extracted DNA yield. The first set (Fig 1A) compares the DNA yield from one filter sampled continuously for 24 hours with the yield from three separate filters sampled for 8 hours in series, the DNA from which was then combined into one concentrated sample. Similar to pre- vious experiment, the DNA concentration of the samples are presented in terms of ng of DNA per volume of air sampled (ng/m3 of air) for total DNA measurement by the Qubit and for par- tial DNA concentration (pg/m3 of air) by qPCR. Fig 5 (left bar) shows that—based on DNA fluorometry—there is no distinct difference in total DNA yield between the 1×24 hour samples (0.14 ng/m3 of air) and the 3×8 hour samples (0.13 ng/m3 of air) (Paired t-test, p value = 0.18). This result led us to further analyze the DNA by performing fungal and bacterial qPCR analysis to investigate whether different yield can be seen for different microbiological targets using these two sampling approaches. Bacterial and fungal DNA analysis (Fig 5, middle and right bar) indicates that there is no significant difference in fungal DNA concentration with 77 pg/m3 and 80 pg/m3 for the 1×24 and 3×8 hour samples, respectively (Paired t-test, p value = 0.32). However, there is a clear gap between the two sampling techniques for bacterial DNA (Paired t-test, p value = 0.001). The 1×24 hour samples yield 5.4 pg/m3 of air, whereas the 3×8 hour samples yield 8.2 pg/m3, an increase of more than 50%. This result suggests that bacterial DNA may be more sensitive to 8 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples Fig 5. Comparison of two sampling approaches. Comparison of a sampling approach utilizing a single filter continuously sampled for 24 h (grey bar) and a combined series of three filters, each operated for 8 h (black bar) expressed in terms of total DNA (left bar, left axis) measured by Qubit and in terms of bacterial (middle bar, right axis) and fungal (right bar, right axis) DNA measured by qPCR (N = 3). doi:10.1371/journal.pone.0141158.g005 Fig 5. Comparison of two sampling approaches. Effect of Sampling Duration on DNA Yield Comparison of a sampling approach utilizing a single filter continuously sampled for 24 h (grey bar) and a combined series of three filters, each operated for 8 h (black bar) expressed in terms of total DNA (left bar, left axis) measured by Qubit and in terms of bacterial (middle bar, right axis) and fungal (right bar, right axis) DNA measured by qPCR (N = 3). Fig 5. Comparison of two sampling approaches. Comparison of a sampling approach utilizing a single filter continuously sampled for 24 h (grey bar) and a combined series of three filters, each operated for 8 h (black bar) expressed in terms of total DNA (left bar, left axis) measured by Qubit and in terms of bacterial (middle bar, right axis) and fungal (right bar, right axis) DNA measured by qPCR (N = 3). doi:10.1371/journal.pone.0141158.g005 doi:10.1371/journal.pone.0141158.g005 doi:10.1371/journal.pone.0141158.g005 sampling duration than fungal DNA. If so, then shorter sampling time might be particularly important for preserving bacterial DNA. Fungal DNA appears insensitive to sampling duration on filter media. The second experiment on the influence of sampling duration was performed by spiking the same amount of GFP-tagged S. oneidensis cells on four different filters at different times. These parallel samplings were conducted simultaneously at same location for a period of 20 hours while the DNA spiking was applied at t = 0, 8, 14, and 20 h, as illustrated in Fig 1B. As the main goal of this experiment was to understand the effect of long sampling duration on cells of a known bacterial species, the result in Table 2 is only presented in terms of total DNA mass esti- mated (pg) by GFP-specific qPCR. Table 2. qPCR results from a series of 20-h samples illustrating the decay of GFP-tagged Shewanella oneidensis cells. Sample treatment a No. of replicates Estimated mass of GFP DNA recovered (ng) b Exposure to air sampling (hours) % of DNA retained (relative to t = 0) No spike 3 0.02 ± 0.001 - - Spike at t = 0 6 0.13 ± 0.02 20 1.3% Spike at t = 8 h 3 0.73 ± 0.01 12 7.1% Spike at t = 14 h 3 1.13 ± 0.02 6 11% Spike at t = 20 h 6 10.2 ± 0.3 0 100% a Time (t) during the 20-h sampling period when a spike of S. oneidensis cells was applied to the ﬁlter. Effect of Sampling Duration on DNA Yield b Mean ± standard deviation reading from qPCR analysis using GFP probe and primers doi:10.1371/journal.pone.0141158.t002 Table 2. qPCR results from a series of 20-h samples illustrating the decay of GFP-tagged Shewanella oneidensis cells. Sample treatment a No. of replicates Estimated mass of GFP DNA recovered (ng) b Exposure to air sampling (hours) % of DNA retained (relative to t = 0) No spike 3 0.02 ± 0.001 - - Spike at t = 0 6 0.13 ± 0.02 20 1.3% Spike at t = 8 h 3 0.73 ± 0.01 12 7.1% Spike at t = 14 h 3 1.13 ± 0.02 6 11% Spike at t = 20 h 6 10.2 ± 0.3 0 100% a Time (t) during the 20-h sampling period when a spike of S. oneidensis cells was applied to the ﬁlter. b Mean ± standard deviation reading from qPCR analysis using GFP probe and primers doi:10.1371/journal.pone.0141158.t002 Table 2. qPCR results from a series of 20-h samples illustrating the decay of GFP-tagged Shewanella oneidensis cells. PCR results from a series of 20-h samples illustrating the decay of GFP-tagged Shewanella oneidensis cells. samples illustrating the decay of GFP-tagged Shewanella oneidensis cells. PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 9 / 18 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples Table 2 indicates that there is a diminishing trend of recoverable DNA with longer sampling duration. Up to 98% of DNA loss is shown upon exposure of the spiked filter to 20 hours of subse- quent sampling. For the filters spiked at the end of air sampling, 10.2 ng of GFP gene DNA was recovered, whereas only 0.13 ng (1.3%) of GFP gene DNA was recovered from filters spiked at t = 0. The DNA recovered from the filters spiked at 8 h and 14 h were intermediate between these two limiting values: 0.73 ng (7.1%) and 1.13 ng (11%) respectively. This result suggests that much of the biomass collected at the beginning of an extended sampling period could be lost throughout subsequent sampling, especially for species more susceptible to stress from air sampling. Mid-Extraction Pooling or Concentration Approach The performance of DNA pooling during the extraction step was analysed by means of Qubit measurements and qPCR in terms of total DNA mass extracted (ng or pg). The result compares the DNA yield of one filter and the yield of three replicate filters combined. The DNA is first extracted to a certain extent and combined into one concentrated sample during the spin filter binding step of the MOBIO PW protocol (Fig 2). Table 3 displays the ratios of biomass extracted from one and three filters which range from 1:2.5 to 1:3.9. These ratios are close to the theoretical value for lossless concentration of 1:3. The calculated concentrations per unit volume of air sampled (ng/m3) and per unit area of AHU filters (ng/cm2) in Table 3 are expected to be the same (between one and three filters) as these comparisons were made from replicates of the same samples. The difference of concen- trations calculated from one and three filters are relatively small with a margin of error of 28% (using the results from one filter as the reference). The high concentration AHU filter samples display smaller deviations, ranging from 5 to 14%. Conversely, the low concentration ambient air samples have higher deviation ranging from 17 to 28%. Discussion DNA Extraction: Cell Lysis p b Concentrations calculated from total DNA extracted divided by total amount of air sampled through one and three ﬁlters (for ambient air samples) or total area of AHU ﬁlter extracted (for AHU ﬁlter samples). doi:10.1371/journal.pone.0141158.t003 PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 DNA Extraction: Cell Lysis DNA Extraction: Cell Lysis The extraction step plays a big role in the effectiveness of DNA-based analysis of environmen- tal samples. As DNA-related technology grows, a variety of extraction kits are continuously threshold value data for mid-concentration approach considering both 16S bacterial and 18S fungal probe-a Table 3. qPCR cycle threshold value data for mid-concentration approach considering both 16S bacterial and 18S fungal probe-and-primer sets. Ambient Air Samples Mean Total DNAa (ng) Mean DNA Concentrationb (ng/m3) Mean Bacterial DNAa (pg) Mean DNA Concentrationb (pg/m3) Mean Fungal DNAa (pg) Mean DNA Concentrationb (pg/m3) 1 Filter 3.1 ± 0.9 0.3 167 ± 37 17.4 1481 ± 142 154 3 Filters 11.9 ± 1.9 0.4 414 ± 114 14.4 3780 ± 440 131 Ratio 1: 3.9 1: 2.5 1: 2.6 Concentration Deviation (%) 28% 21% 18% AHU Filter Samples Mean Total DNA (ng) Mean DNA Concentration (ng/cm2) Mean Bacterial DNA (pg) Mean DNA Concentration (pg/cm2) Mean Fungal DNA (pg) Mean DNA Concentration (pg/cm2) 1 Filter 44 ± 6 4.4 570 ± 46 57 2970 ± 410 297 3 Filters 148 ± 12 4.9 1798 ± 124 60 10100 ± 1420 338 Ratio 1: 3.4 1: 3.2 1: 3.4 Concentration Deviation (%) 12% 5% 14% a Mean ± standard deviation for N = 4 samples in each case b Concentrations calculated from total DNA extracted divided by total amount of air sampled through one and three ﬁlters (for ambient air samples) or total area of AHU ﬁlter extracted (for AHU ﬁlter samples). doi:10.1371/journal.pone.0141158.t003 Table 3. qPCR cycle threshold value data for mid-concentration approach considering both 16S bacterial and 18S fungal probe-and-primer sets. or mid-concentration approach considering both 16S bacterial and 18S fungal probe-and-primer sets. Table 3. qPCR cycle threshold value data for mid-concentration approach considering both 16S bacterial an a Mean ± standard deviation for N = 4 samples in each case b Concentrations calculated from total DNA extracted divided by total amount of air sampled through one and three ﬁlters (for ambient air samples) or total area of AHU ﬁlter extracted (for AHU ﬁlter samples). PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 10 / 18 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples improving the efficiency in both DNA yield and extraction time. DNA Extraction: Cell Lysis The MOBIO Power Water (PW) kit was selected in this study mainly due to its frequent use in previous bioaerosol studies [14,42], as well as compatibility with the size of our filter samples. The PW kit uses 5 ml sample tube which allows filter samples to be directly placed inside the tubes (47 mm PES and the 5 cm2 cut AHU filters) without having to scrape, grind, vacuum or dissolve the filters before pro- ceeding with extraction steps. This feature minimizes extraction time and biomass loss. Choos- ing the appropriate kits design and size is important. Other bioaerosol studies have chosen other extraction kits such as the MOBIO Power Max Soil kit for processing larger sample vol- umes in 50 ml tubes [46] or MOBIO Power Soil for more rigorous inhibitor removal when smaller starting tubes (2 ml) can be used [47]. DNA extraction consists of three main steps: 1) cell lysis to expose the intracellular material; 2) isolation of DNA from contaminants; and 3) final elution. For environmental samples, the challenge in DNA extraction often comes from the lysis and isolation steps. Environmental samples start with a mixture of species with different characteristics, which may require differ- ent approaches to effectively lyse all the cells. Moreover, if inhibitors such as humic acid from floor dust or particulate matter suspended in the air remains in the final DNA solution, the integrity of subsequent analysis steps, such as qPCR, can be compromised. As a quality control measure, a dilution series was performed on a separate set of DNA sam- ples extracted from AHU filter and PES ambient air filter to verify potential inhibitor effects. The results suggested no lowered inhibition from the diluted samples. Furthermore, most of our qPCR results correspond well with the Qubit measurements and all of our final DNA sam- ples are clear in colour. Given the fact that some of the samples had high dust contents (e.g. the extracts from AHU filters), it seems the MOBIO PW kit was able to effectively handle the inhibitors in all environmental samples collected in this study. Final DNA yield of an extraction is often related to its lysis. The complex matrix of bioaero- sol samples has forced several studies to apply some improvements to their lysis protocols, such as sonication [14], thermal incubation [42], and chemical addition [17,43]. PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 DNA Extraction: Cell Lysis Three com- monly used cell lysis methods are physical disruption, chemical and enzymatic lysis. This study focused on the impact of a combination of sonication and thermal incubation (physical disrup- tion) because this approach, combined with the physical bead beating in the MOBIO protocol, is known to be effective in lysing cells with tough cell wall/membrane (spores, Gram-positive bacteria) and small bacteria cells[48]. In addition, it is relatively more flexible to be applied to almost all DNA extraction protocols. The impact of chemical and enzymatic lysis on various forms of bioaerosol samples, however, also warrants future investigations. Choosing the optimum temperature for high temperature sonication can be challenging when applied to environmental samples with unknown quantity and composition of biomass. Ideally, one wants the temperature to be high enough to help lyse the cells, but also not too high to avoid DNA denaturation or lysis buffer compatibility issues. Some studies have chosen to incubate their samples at 60°C[49] or 65°C [14,42] without elaborating the underlying rea- sons for the choice. The results from Fig 3A and 3B agree that the highest DNA yields came from filter samples incubated at either 65°C or 70°C. Incubation at 65°C over 70°C was chosen for the rest of the study as it preserves more bacterial DNA (Fig 3B) and displays more consis- tency among the replicates (Fig 3A and 3B). In order to verify the validity of this temperature choice, ANOVA single factor followed by Tukey honest significant difference (HSD) analysis for pairwise comparisons was performed comparing results from samples incubated at 65°C to the other temperatures (65°C to 55°C, 60°C, 70°C, and 75°C) for total, fungal and bacterial DNA. The means of total DNA for all 5 temperatures showed no significant difference from that of 65°C (p-value > 0.05). However, the analysis results of fungal and bacterial DNA provided further insights. For fungal DNA, 11 / 18 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples only comparisons of 65°C to 60°C and 65°C to 70°C have p-values > 0.05. As for bacterial DNA, only comparison between 65°C and 70°C results in p-value > 0.05. The results may explain why different studies mentioned above incubated their samples at different tempera- tures to obtain the required DNA yield. It is likely that in the complex matrix of an environ- mental sample, different target biomass would behave differently under varying temperatures. DNA Extraction: Cell Lysis Therefore, depending on the purpose of a study, various incubation temperatures may be con- sidered. For example, when only rough estimation of total DNA via Qubit is required, any tem- perature from 55°C to 75°C may all produce comparable results. If the focus is on fungal DNA, temperatures from 60°C to 70°C may be chosen. Lastly, when bacterial DNA is targeted, tem- peratures at 65°C or 70°C could be the options as they have equally high yields. In the present study, our focus is to preserve as much DNA as possible for further analysis of all three DNA types. The statistical analysis above shows that there are only two tempera- tures (65°C and 70°C) that are within the range of temperatures that produces the highest yield for all DNA types (total, bacterial and fungal). Between these two temperatures, the lower one (65°C) is chosen as the final incubation temperature because in addition to the higher consis- tency and better preservation of bacterial DNA, lower temperature is also preferred to save energy and further reduce the chance of DNA denaturation during processing. During sonication, the additional force provided by heating [50] and cavitation [51] help disrupt the cell walls and subsequently release more intracellular matter, including DNA. The impact of sonication alone on DNA yield can be seen from the difference between no treatment and sonication only bars in Fig 3A and 3B. The total DNA yield goes up by 23% (Fig 3A) with additional sonication alone while bacterial and fungal DNA yield went up by 43% and 28%, respectively (Fig 3B). The effect of heat incubation is seen to dominate with up to 600% increase in DNA yield as the temperature is increased. The influences of the two approaches appear to be interrelated, with sonication playing a supporting role. It is also worthwhile to note that sonication itself dissipates heat in the water. We’ve observed that when the instru- ment is operated in sonication-only mode, water temperature rises to 35–40°C within the first 5 minutes and then stays in that range for the remainder of the 30-min process. We would like to also highlight that there are indeed limitations associated with studies uti- lizing environmental samples. It can be observed from the results that the impact of one improvement approach (i.e. thermal sonication) varies on different target biomass. PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 DNA Extraction: Cell Lysis Compared to the typical controlled experiments, the complex environmental matrix makes it very difficult to estimate the absolute extraction efficiency from environmental samples. The absolute extrac- tion efficiency can only be calculated when artificial samples with a known starting amount (spiking a clean/PM-loaded filter with a known amount of cells or using nebulizer to mimic bioaerosol sampling) are used. These efficiencies, however, are more relevant for studies target- ing a specific species or a small group of species as they only correspond to the chosen cells that are spiked, which in most cases, does not represent actual environmental conditions. We believe it is beneficial to also report the improvements of an optimization approach (physical disruptions, enzyme, etc.) tested on environmental samples to complement the more readily available studies with known samples in the literature. We further tested the improved protocol with 65°C sonication on two sets of environmental samples with significantly different biomass concentrations. The results displayed in Fig 4A and 4B show that the total DNA yield from both ambient air and AHU filter samples increased significantly with the addition of a 65°C sonication step. Note that two of four replicates for total DNA from ambient air samples without sonication (Fig 4B) were below the detection limit of the Qubit. This result highlights the importance of sonication and heat incubation in DNA extraction because it permits the direct use of DNA fluorometry to quantify total DNA mass for samples that would otherwise be undetectable. DNA fluorometry provides total DNA 12 / 18 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples quantification in a relatively short time, which is an attractive feature in indoor environmental assessment. Total DNA measurement is generally not available in PCR-based techniques as their analysis is limited to the chosen primers. While DNA fluorometry (i.e. Qubit) is consid- ered more accurate than light-absorbance-based methods, such as nanodrop [52], there are only limited reports of direct quantification of DNA in air samples using this method. This is likely due to its relatively higher detection limit (typically 0.0005 ng DNA/μL) compared to PCR-based instruments (capable of detecting DNA concentration up to 10 times lower than that of Qubit with enough duplication cycles). DNA Extraction: Cell Lysis As shown in Fig 4A and 4B, large deficits of biomass for both bacterial and fungal DNA in both ambient air and AHU filter samples suggests that the original MOBIO protocol (with bead beating only) does not effectively lyse all the cells in environmental samples such as these. One can speculate that the higher DNA yield from additional lysis originates from cells with thicker walls, such as fungal spores or Gram-positive bacteria, which might not be effectively disrupted by means of only bead beating. The results of this study indicate that it would be worthwhile to be attentive to lysis as a crit- ical processing step for environmental bioaerosol sample analysis. The total DNA measure- ments of ambient air samples extracted with the unimproved protocol shows that one may miss detection with inadequate lysis. The influence of an improvement approach can also vary among different types of environmental samples (in this case, ambient air filters and AHU fil- ters). More attention towards method development would be beneficial for future bioaerosol studies, including systematic cross-comparison of improvement techniques under a range of environmental sampling conditions. PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples that some more vulnerable species, such as Gram-negative bacteria, may be prone to DNA loss from long duration sampling with filters. The second experiment, illustrated in Fig 1B, was conducted to confirm the above hypothe- sis. There are two major reasons for selecting Shewanella oneidensis [44] for this experiment. First, this species is embodied as Gram-negative bacteria, which are relatively less-protected from air sampling due to their thinner cell membrane. We suspect that if DNA degradation were to happen during air sampling, then less protected cells would be more vulnerable. The second reason is because this GFP-tagged species is rarely found in ambient air. The ability to control the timing of the presence of this species on air filters through the spiking process per- mits us to isolate the impact of air sampling on DNA recoverability. The qPCR result reported in Table 2 indicates a progressive degradation of DNA recover- ability for this specific species in relation to sampling duration. This occurrence leads us to believe that without properly testing the compatibility of our target species and our chosen sampling protocol, the DNA of the cells captured at the beginning of sampling may be falsely undetected in the subsequent analysis steps. The degree of DNA recoverability loss is likely to be species dependent, as the previous results (Fig 5, middle and right bar) show that fungal DNA is less sensitive to air sampling on filter media than bacterial DNA. The two experiments indicate that the duration of filter-based air sampling might have some effects on the recoverability of DNA. Depending on the chosen sampling medium and the characteristics of the targeted species, it would be beneficial to have a preliminary investiga- tion to assess how durable is the biomass DNA during sampling. For instance, in the case where certain Gram-negative bacteria or other more vulnerable species are targeted, shorter sampling duration would be recommended to preserve more DNA. Conversely, in the case when tougher species like fungal spores or Gram-positive bacteria are targeted, longer-term sampling can be utilized as the DNA collected from these species is less likely to degrade from sampling stress. In addition to sampling duration, storage duration may also play a part in DNA preservation. It is recommended to extract the DNA as early as possible after sampling since some DNA on the filter may continue to degrade during storage [48]. PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 Effect of Sampling Duration on DNA Yield To better understand how the sampling duration affects the DNA yield from the filter sample, we have utilized an open (no size-cut) filter-based sampling protocol for its relatively higher collection efficiency (as compared to liquid impinger) [30,53], flexibility (in choosing sample flowrate and duration)[18] and cost. The naturally low bioaerosol concentrations in ambient air often lead to issues of inadequate measurement sensitivity. Therefore, to collect enough bio- mass, long sampling durations are sometimes inevitable for non-viable based analysis. Many studies have utilized relatively long sampling times such as eight hours [42], 20+ hours [54], or even several days for collection on individual filters [16,41]. Concerns about time-dependent decreases of the viability of captured cells during air sam- pling have been raised over the years. Depending on the sampling medium, microorganisms may partially or completely lose their viability owing to stress from different parameters such as shear-stress induced by air flow, osmotic pressure or lack of moisture [16,26]. Of course, DNA-based analyses do not rely on viability as the DNA can be extracted and analysed regard- less of whether the captured microorganisms are viable or culturable. Nevertheless, the general concern remains: do extended sampling periods lead to degradation of previously collected DNA? The purpose of our investigation on this point was to further assess whether long sam- pling duration has a significant adverse impact on DNA analysis yield. The total DNA measurement in Fig 5 (left bar) suggests that culture-independent DNA based analysis is not impeded by long sampling duration on a filter media. The qPCR result (Fig 5, middle and right bar), however, reveals that although there is no notable difference on the fungal DNA, 50% more bacterial DNA was preserved in the DNA solution extracted from the combined 3×8-h filters. As the amount of bacterial DNA is only a small portion of total DNA, which also includes fungal and other non-microbial DNA, the 50% reduction in bacte- rial DNA does not generate a notable change in total DNA. However, this result does suggest PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 13 / 18 Supporting Information S1 File. Sampling Schedule and Standard Curves. Sampling schedules information and stan- dard curves for bacteria, fungus and custom-GFP for qPCR Analysis. (XLSX) DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples because these deviations are not only caused by the pooling method, but also would result from natural variability of the replicate environmental samples. In addition to the reliability, the suggested DNA pooling method is generally applicable to all extraction kits with similar spin filters. It is not limited to bioaerosol samples. Different con- centration approaches generally help gather enough biomass with shorter sampling duration, which could be beneficial to DNA recoverability. Furthermore, this concentration approach allows more flexibility in designing a suitable sampling plan, because other concerns such as noise, sampler durability or site availability often arise when intense sampling activity is needed. Conclusions Collecting enough biomass during for desired analysis methods a short time period is a well- known challenge for the typically low concentrations of bioaerosol levels found in indoor and outdoor air. This study has shown that one can improve DNA recoverability and yield in bioaerosol samples by fine-tuning several parameters such as applying more rigorous cell lysis during DNA extraction or pooling composite samples to obtain enough biomass. In the cases reported here, outdoor air and building AHU filter samples were analyzed using DNA-based measurement methods. Adding enhanced lysis by means of high tempera- ture water bath sonication was proven to be effective with up to 600% increase in DNA yield at the optimum temperature of 65°C. We found that for some more vulnerable species, up to 98% of the captured species’ DNA could degrade over the course of 20 hours of filter sampling. Combining DNA from separate samples during the spin filter binding step of the DNA extrac- tion protocol showed promising results with a margin of error below 30%. We examined the impact of the aforementioned parameters using a commercially available DNA extraction kit, which generally demonstrates good capacity for urban environmental samples. The chosen kit seems to be able to effectively deal with inhibitors of both high biomass (AHU filters used for three months) and low biomass (PES filters applied to sample ambient air for 8–20 h) samples. Incorporating the proposed modifications saved both time and reagents in addition to achieving improvements in DNA yield. Dealing with highly variable bioaerosol environmental samples, researchers are often faced with uncertainty on which approaches may best improve their analytical methods. Therefore, we believe that it is beneficial to have more method development-based bioaerosol studies in the future that assess how efficient different improvement techniques (e.g., enhanced chemical or enzyme lysis) are in dealing with other forms of environmental samples (liquid impinger samples, dust-wipe samples). More such efforts will help reach the goal of more standardized analysis methods for studies of aerosol microbiology. Mid-Extraction Pooling or Concentration Approach Various efforts to concentrate bioaerosol samples have been reported in previous studies to cir- cumvent detection limit issues. For example, Boreson et al. [8] concentrated 5 mL of liquid from a Biosampler by filtering the collected liquid media onto a section of filter membrane before proceeding with extraction. As commercial DNA extraction kits continue to improve, it is now possible to combine DNA recovered from separate samples in the middle of extraction steps, which saves time and, in some cases, reagents. In the case of MOBIO PW kit, the DNA can be pooled by means of introducing extraction liquid (after PW3 addition) from different samples on to the same spin filter. One concern that arises for this concentration approach is the DNA binding capacity of the spin filter itself. According to MOBIO, the PW kit spin filter can bind up to 20000 ng of DNA, which is ample for the level of biomass processed in this study (up to 300–400 ng of DNA from 30 cm2 of AHU filters following 3 months of in-use service). The ratios resulted in Table 3 are all relatively close to 3:1, confirming the suitability of this pooling approach to the range of biomass concentration engaged in this study. As discussed previously, it is sometimes necessary to combine several parallel filters into one concentrated sample to overcome instrument detection limits owing to the low airborne DNA concentra- tion. The approach of combining DNA from three samples onto one spin filter has a margin of error below 30% for the experiments reported here. We consider this outcome to be satisfactory 14 / 18 References 1. Gregory PH. Airborne microbes: their significance and distribution. Proceedings of the Royal Society of London Series B Biological sciences. 1971; 177: 469–483. 2. Haldane J, Anderson A. The carbonic acid, organic matter, and micro-organisms in air, more especially of dwellings and schools. Philosophical Transactions of the Royal Society of London B. 1887: 61–111. 3. Green JL. Can bioinformed design promote healthy indoor ecosystems? Indoor Air. 2014; 24: 113– 115. doi: 10.1111/ina.12090 PMID: 24628783 4. Rintala H, Pitkäranta M, Täubel M. Microbial communities associated with house dust. Advances in Applied Microbiology. 2012; 78: 75–120. doi: 10.1016/B978-0-12-394805-2.00004-X PMID: 22305094 5. Xu Z, Wu Y, Shen F, Chen Q, Tan M, et al. Bioaerosol science, technology, and engineering: Past, present, and future. Aerosol Science and Technology. 2011; 45: 1337–1349. 6. Xu Z, Wei K, Wu Y, Shen F, Chen Q, et al. Enhancing bioaerosol sampling by Andersen impactors using mineral-oil-spread agar plate. PLoS ONE. 2013; 8: e56896. doi: 10.1371/journal.pone.0056896 PMID: 23460818 7. Crook B, Burton NC. Indoor moulds, Sick Building Syndrome and building related illness. Fungal Biol- ogy Reviews. 2010; 24: 106–113. 8. Boreson J, Dillner AM, Peccia J. Correlating bioaerosol load with PM2.5 and PM10cf concentrations: A comparison between natural desert and urban-fringe aerosols. Atmospheric Environment. 2004; 38: 6029–6041. 9. Yoshida K, Schuenemann VJ, Cano LM, Pais M, Mishra B, et al. The rise and fall of the Phytophthora infestans lineage that triggered the Irish potato famine. eLife. 2013; 2: e00731. doi: 10.7554/eLife. 00731 PMID: 23741619 10. Vacher S, Hernandez C, Bärtschi C, Poussereau N. Impact of paint and wall-paper on mould growth on plasterboards and aluminum. Building and Environment. 2010; 45: 916–921. 11. Gravesen S, Nielsen PA, Iversen R, Nielsen KF. Microfungal contamination of damp buildings—Exam- ples of risk constructions and risk materials. Environmental Health Perspectives. 1999; 107: 505–508. PMID: 10347000 12. Kellogg CA, Griffin DW. Aerobiology and the global transport of desert dust. Trends in Ecology and Evolution. 2006; 21: 638–644. PMID: 16843565 13. Burrows SM, Elbert W, Lawrence MG, Pöschl U. Bacteria in the global atmosphere—Part 1: Review and synthesis of literature data for different ecosystems. Atmospheric Chemistry and Physics. 2009; 9: 9263–9280. 14. Smith DJ, Jaffe DA, Birmele MN, Griffin DW, Schuerger AC, et al. Free tropospheric transport of micro- organisms from Asia to North America. Microbial Ecology. 2012; 64: 973–985. doi: 10.1007/s00248- 012-0088-9 PMID: 22760734 15. Nazaroff WW. Indoor bioaerosol dynamics. Indoor Air. Acknowledgments We would like to express our gratitude to Prof. William W Nazaroff from the University of California, Berkeley for his invaluable insights and feedbacks during the writing of this manuscript. 15 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples We would like to also thank researchers from the Advanced Environmental Biotechnology Centre (AEBC) in the Nanyang Environment and Water Research Institute (NEWRI), Singa- pore and the Singapore Centre on Environmental Life Sciences Engineering (SCELSE) in Nanyang Technological University (NTU), Singapore for providing lab facilities for the experi- ments in this study. Author Contributions Conceived and designed the experiments: IL VC BC. Performed the experiments: IL YW CKN DM. Analyzed the data: IL YW VC BC. Contributed reagents/materials/analysis tools: IL CKN DM YW W t th IL VC BC PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 References 2014: doi: 10.1111/ina.12174 PMID: 25483392 16. Peccia J, Hernandez M. Incorporating polymerase chain reaction-based identification, population char- acterization, and quantification of microorganisms into aerosol science: A review. Atmospheric Environ- ment. 2006; 40: 3941–3961. 17. Hospodsky D, Yamamoto N, Peccia J. Accuracy, precision, and method detection limits of quantitative PCR for airborne bacteria and fungi. Applied and Environmental Microbiology. 2010; 76: 7004–7012. doi: 10.1128/AEM.01240-10 PMID: 20817798 16 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples 18. Radosevich JL, Wilson WJ, Shinn JH, DeSantis TZ, Andersen GL. Development of a high-volume aero- sol collection system for the identification of air-borne micro-organisms. Letters in Applied Microbiology. 2002; 34: 162–167. PMID: 11874535 19. Womack AM, Bohannan BJM, Green JL. Biodiversity and biogeography of the atmosphere. Philosophi- cal Transactions of the Royal Society B: Biological Sciences. 2010; 365: 3645–3653. 20. Ariatti A, Comtois P. Louis Pasteur: the first experimental aerobiologist. Aerobiologia. 1993; 9: 5–14. 21. Wang C-C, Fang G-C, Lee L-Y. The study of ambient air bioaerosols during summer daytime and night- time periods in Taichung, Central Taiwan. Environmental Forensics. 2008; 9: 6–14. 22. Fierer N, Liu Z, Rodríguez-Hernández M, Knight R, Henn M, et al. Short-term temporal variability in air- borne bacterial and fungal populations. Applied and Environmental Microbiology. 2008; 74: 200–207. PMID: 17981945 23. Dybwad M, Skogan G, Blatnya JM. Temporal variability of the bioaerosol background at a subway sta- tion: Concentration level, size distribution, and diversity of airborne bacteria. Applied and Environmen- tal Microbiology. 2014; 80: 257–270. doi: 10.1128/AEM.02849-13 PMID: 24162566 24. Carnelley T, Wilson T. A new method of determining the number of micro-organisms in air. Proceedings of the Royal Society of London. 1888; 44: 455–464. 25. Dutkiewicz J. Bacteria and fungi in organic dust as potential health hazard. Annals of Agriculture and Environmental Medicine. 1997; 4: 11–16. 26. Lin X, A. Reponen T, Willeke K, Grinshpun SA, Foarde KK, et al. Long-term sampling of airborne bacte- ria and fungi into a non-evaporating liquid. Atmospheric Environment. 1999; 33: 4291–4298. 27. Thorne PS, Kiekhaefer MS, Whitten P, Donham KJ. Comparison of bioaerosol sampling methods in barns housing swine. Applied and Environmental Microbiology. 1992; 58: 2543–2551. PMID: 1514801 28. Rajasekar A, Balasubramanian R. Assessment of airborne bacteria and fungi in food courts. Building and Environment. 2011; 46: 2081–2087. 29. Toivola M, Alm S, Reponen T, Kolari S, Nevalainen A. PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 References Personal exposures and microenvironmental concentrations of particles and bioaerosols. Journal of Environmental Monitoring. 2002; 4: 166–174. PMID: 11871701 30. Eduard W, Heederik D. Methods for quantitative assessment of airbone levels of noinfectious microor- ganisms in highly contaminated work environment. American Industrial Hygiene Association Journal. 1998; 59: 113–127. PMID: 9487665 31. Henningson EW, Ahlberg MS. Evaluation of microbiological aerosol samplers: A review. Journal of Aerosol Science. 1994; 25: 1459–1492. 32. Jensen PA, Todd WF, Davis GN, Scarpino PV. Evaluation of eight bioaerosol samplers challenged with aerosols of free bacteria. American Industrial Hygiene Association Journal. 1992; 53: 660–667. PMID: 1456208 33. Simon X, Duquenne P, Koehler V, Piernot C, Coulais C, et al. Aerosolisation of Escherichia coli and associated endotoxin using an improved bubbling bioaerosol generator. Journal of Aerosol Science. 2011; 42: 517–531. 34. Duchaine C, Thorne PS, Mériaux A, Grimard Y, Whitten P, et al. Comparison of endotoxin exposure assessment by bioaerosol impinger and filter-sampling methods. Applied and Environmental Microbiol- ogy. 2001; 67: 2775–2780. PMID: 11375194 35. Cyprowski M, Sowiak M, Szadkowska-Stańczyk I. β(1 ! 3)-glucan aerosols in different occupational environments. Aerobiologia. 2011; 27: 345–351. 36. An HR, Mainelis G, White L. Development and calibration of real-time PCR for quantification of airborne microorganisms in air samples. Atmospheric Environment. 2006; 40: 7924–7939. 37. Betelli L, Duquenne P, Grenouillet F, Simon X, Scherer E, et al. Development and evaluation of a method for the quantification of airborne Thermoactinomyces vulgaris by real-time PCR. Journal of Microbiological Methods. 2013; 92: 25–32. doi: 10.1016/j.mimet.2012.10.009 PMID: 23103956 38. Cao C, Jiang W, Wang B, Fang J, Lang J, et al. Inhalable microorganisms in Beijing’s PM2.5 and PM10 pollutants during a severe smog event. Environmental Science & Technology. 2014; 48: 1499–1507. 39. Nonnenmann MW, Bextine B, Dowd SE, Gilmore K, Levin JL. Culture-independent characterization of bacteria and fungi in a poultry bioaerosol using pyrosequencing: A new approach. Journal of Occupa- tional and Environmental Hygiene. 2010; 7: 693–699. doi: 10.1080/15459624.2010.526893 PMID: 21058154 40. Glenn TC. Field guide to next-generation DNA sequencers. Molecular Ecology Resources. 2011; 11: 759–769. doi: 10.1111/j.1755-0998.2011.03024.x PMID: 21592312 17 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 DNA-Based Methods Improvements on Low Concentration Bioaerosol Samples 41. Qian J, Hospodsky D, Yamamoto N, Nazaroff WW, Peccia J. Size-resolved emission rates of airborne bacteria and fungi in an occupied classroom. Indoor Air. 2012; 22: 339–351. doi: 10.1111/j.1600-0668. 2012.00769.x PMID: 22257156 42. PLOS ONE \| DOI:10.1371/journal.pone.0141158 November 30, 2015 References Meadow JF, Altrichter AE, Kembel SW, Kline J, Mhuireach G, et al. Indoor airborne bacterial communi- ties are influenced by ventilation, occupancy, and outdoor air source. Indoor Air. 2013; 24: 41–48. doi: 10.1111/ina.12047 PMID: 23621155 43. Noris F, Siegel JA, Kinney KA. Evaluation of HVAC filters as a sampling mechanism for indoor microbial communities. Atmospheric Environment. 2011; 45: 338–346. 44. Zhang Y, Ng CK, Cohen Y, Cao B. Cell growth and protein expression of Shewanella oneidensis in bio- films and hydrogel-entrapped cultures. Molecular BioSystems. 2014; 10: 1035–1042. doi: 10.1039/ c3mb70520j PMID: 24626808 45. Liu CM, Kachur S, Dwan MG, Abraham AG, Aziz M, et al. FungiQuant: A broad-coverage fungal quanti- tative real-time PCR assay. BMC Microbiology. 2012; 12: 1–11. 46. Yamamoto N, Bibby K, Qian J, Hospodsky D, Rismani-Yazdi H, et al. Particle-size distributions and seasonal diversity of allergenic and pathogenic fungi in outdoor air. ISME J. 2012; 6: 1801–1811. doi: 10.1038/ismej.2012.30 PMID: 22476354 47. Yamamoto N, Shendell DG, Peccia J. Assessing allergenic fungi in house dust by floor wipe sampling and quantitative PCR. Indoor Air. 2011. 48. Robe P, Nalin R, Capellano C, Vogel TM, Simonet P. Extraction of DNA from soil. European Journal of Soil Biology. 2003; 39: 183–190. 49. Schafer MP. Detection and characterization of airborne Mycobacterium tuberculosis H37Ra particles, a surrogate for airborne pathogenic M. tuberculosis. Aerosol Science and Technology. 1999; 30: 161– 173. 50. Pribylka A, Almeida AV, Altmeyer MO, Petr J, Sevcik J, et al. Fast spore breaking by superheating. Lab on a Chip. 2013; 13: 1695–1698. doi: 10.1039/c3lc41305e PMID: 23474861 51. Pilli S, Bhunia P, Yan S, LeBlanc RJ, Tyagi RD, et al. Ultrasonic pretreatment of sludge: A review. Ultra- sonics Sonochemistry. 2011; 18: 1–18. doi: 10.1016/j.ultsonch.2010.02.014 PMID: 20471901 52. Tyler K, Wang G, Tyler S, Johnson W. Factors affecting reliability and reproducibility of amplification- based DNA fingerprinting of representative bacterial pathogens. Journal of Clinical Microbiology. 1997; 35: 339. PMID: 9003592 53. Wu Y, Shen F, Yao M. Use of gelatin filter and BioSampler in detecting airborne H5N1 nucleotides, bac- teria and allergens. Journal of Aerosol Science. 2010; 41: 869–879. 54. Hospodsky D, Qian J, Nazaroff WW, Yamamoto N, Bibby K, et al. Human occupancy as a source of indoor airborne bacteria. PLoS ONE. 2012; 7. 18 / 18
https://openalex.org/W3003920983	https://ieeexplore.ieee.org/ielx7/4609443/8994817/08970288.pdf	English	null	Representation and Classification of Auroral Images Based on Convolutional Neural Networks	IEEE journal of selected topics in applied earth observations and remote sensing	2,020	cc-by	8,820	Representation and Classiﬁcation of Auroral Images Based on Convolutional Neural Networks Qiuju Yang and Penghui Zhou image classiﬁcation and classiﬁed auroras into arcs, patchy auroras, omega bands, and north–south structures according to their shape information. Since then, a lot of automatic auroral studies, including auroral image representation, classiﬁcation, retrieval, and segmentation, have been emerging [9]–[15]. Wang et al. [16] proposed to combine the local binary pattern operator and a delicately designed block partition scheme to characterize auroral morphology (shape and texture). Rao et al. [17] de- scribed the color variants of scale-invariant feature transform features for performing the automated classiﬁcation of all-sky camera images into three mutually exclusive classes: aurora, no aurora, and cloudy. Yang and Hu [18] utilized the Weber local descriptor to represent and classify auroral images into the arc, drapery corona, radial corona, and hotspot. By com- bining multiple handcrafted features (grayscale, structural, and textural features) extracted from auroral images, Zhong et al. [19] proposed an auroral image classiﬁcation method based on multifeature latent Dirichlet allocation. Abstract—Auroral forms are correlated with certain physical processes in the magnetosphere and ionosphere. It is, therefore, desirable to automatically classify the vast amount of observed au- roral images and make large statistical studies. The key problem in classiﬁcation tasks is image representation. In this article, using the adaptive feature learning ability of convolutional neural networks, an end-to-end auroral image classiﬁcation network is proposed, which automatically classiﬁes the auroral images observed at the Chinese Yellow River Station into four classes: arc, drapery corona, radial corona, and hotspot corona. Based on the AlexNet, our method exploits the advanced spatial transformer network (STN) and large margin Softmax (L-Softmax) loss function to extract auroral features. STN is able to learn invariance to translation, scaling, and rotation, whereas L-Softmax increases the difﬁculty of auroral feature learning so that it encourages the intraclass compactness and interclass separability between learned features. The proposed method was validated on the auroral image datasets by supervised classiﬁcation, image retrieval, and statistical anal- ysis of the temporal occurrence distributions of the four auroral categories. Experimental results showed that after trained on the winter auroral observations in 2003, the proposed model achieves an average classiﬁcation accuracy of 93.7% on the auroral data of the following ﬁve winters (2004–2009) while maintaining high efﬁciency, which is superior to the previously reported articles. Above auroral image classiﬁcation techniques include two separate procedures: feature extraction and classiﬁer design. 523 523 IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 Representation and Classiﬁcation of Auroral Images Based on Convolutional Neural Networks The extracted features represent the intensity, shape, and texture of the aurora and play an important role in auroral image clas- siﬁcation. Ideally, they should be distinctive and, at the same time, robust to a variety of possible image transformations (e.g., rotation, scaling, and translation). However, hand designing an effective feature is always a lengthy process that requires considerable expertise to delicately design according to speciﬁc data and tasks. Index Terms—Auroral image classiﬁcation, convolutional neural network (CNN), large margin Softmax (L-Softmax) loss function, spatial transformer network (STN). This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see http://creativecommons.org/licenses/by/4.0/ Manuscript received September 17, 2019; revised December 27, 2019; ac- cepted January 21, 2020. Date of publication January 27, 2020; date of current version February 13, 2020. This work was supported in part by the National Natural Science Foundation of China under Grant 41504122, and in part by the Fundamental Research Funds for the Central Universities under Grant GK201903021. (Corresponding author: Qiuju Yang.) Manuscript received September 17, 2019; revised December 27, 2019; ac- cepted January 21, 2020. Date of publication January 27, 2020; date of current version February 13, 2020. This work was supported in part by the National Natural Science Foundation of China under Grant 41504122, and in part by the Fundamental Research Funds for the Central Universities under Grant GK201903021. (Corresponding author: Qiuju Yang.) The authors are with the School of Physics and Information Tech- nology, Shaanxi Normal University, Xi’an 710119, China (e-mail: yangqiuju@snnu.edu.cn; zhoupenghui@snnu.edu.cn). Digital Object Identiﬁer 10.1109/JSTARS.2020.2969245 The authors are with the School of Physics and Information Tech- nology, Shaanxi Normal University, Xi’an 710119, China (e-mail: yangqiuju@snnu.edu.cn; zhoupenghui@snnu.edu.cn). y gq j p g Digital Object Identiﬁer 10.1109/JSTARS.2020.2969245 IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 524 2) The proposed classiﬁcation model is domain knowledge based and its submodules, STN and L-Softmax loss func- tion, are chosen according to the unique characteristics of the aurora. Yellow River Station (YRS) into the arc, drapery corona, radial corona, and hotspot. Clausen and Nickisch [30] utilized the pretrained Inception-v4 to automatically classify auroral images into clear/no aurora, cloudy, moon, arc, diffuse, and discrete from the Oslo Auroral THEMIS (OATH) dataset. Niu et al. [10] proposed a weakly supervised semantic segmentation method to achieve joint pixel-level localization of the key local structure and image-level classiﬁcation of the auroral images. Han et al. [25] proposed a multisize kernels CNN with eye movement guided task-speciﬁc initialization to classify auroral images into the arc, drapery corona, and radiation corona. 3) In order to meet the actual demands in auroral physics, dif- ferent from previous experimental settings, the proposed model is trained on one-winter auroral observations and tested on the following ﬁve-winter observations. 4) Extensive experiments are carried out to validate the ef- fectiveness of the proposed model, which suggests the practical application value of our method to the automatic classiﬁcation of the huge amount of auroral images. p y However, there is a gap between the above classiﬁcation methods and the actual application. On the one hand, the aurora is rich in spatial variations and often lacking well-deﬁned inter- class boundaries. The extracted CNN features for auroral image classiﬁcation should be rotation invariant and discriminative [42]–[44]. But as shown in [40] and [30], the auroral features obtained by AlexNet and Inception-v4 are not satisfactory. On the other hand, the number of auroral images is very huge and increasing every winter. Therefore, the method for auroral image classiﬁcation must be efﬁcient and easy-to-implement before it has practical value. However, the training of the patch scale model in [10] is a very time-consuming process, and obtaining the eye movement annotation in [25] is extremely expensive. Theremainder of this articleis organizedas follows. SectionII outlines the overall framework of the auroral image classiﬁca- tion model and brieﬂy introduces the STN and L-Softmax loss function. Section III presents the description of auroral image classiﬁcation mechanism and dataset setup. The experiments and results are reported in Section IV. We conclude this article in Section V. II. STN AND L-SOFTMAX-BASED ALEXNET FOR AURORAL IMAGE REPRESENTATION Fig. 1 shows the whole network architecture of the auroral image classiﬁcation model. The spatial transformer network (STN) is embedded in the input layer of the AlexNet for the spatial transformation of auroral images, which enables the network to automatically learn more effective auroral regions in the training process. In addition, the L-Softmax loss function is utilized to optimize the network so that the network can learn more discriminatory auroral features. In the following, we will abbreviate the model to STN-Lsoftmax-AlexNet. y y p This article aims to develop a practical method to classify the vast amount of existing auroral images automatically. A new auroral image classiﬁcation model based on CNN architectures is explored according to the unique characteristics of the aurora. First, we compare the three CNN models (AlexNet, VGG16, and Inception-v4) used in the previous automatic analysis of auroral images [23]–[24] and [30], and the simplest AlexNet is chosen as the backbone for auroral image classiﬁcation in consideration of accuracy and efﬁciency. Second, considering the large spatial difference of auroral images (see Fig. 3), the spatial transformer network (STN) [22] is inserted before AlexNet. STN can adap- tively transform and align the images according to the classi- ﬁcation task (including translation, scaling, rotation, and more generic warping), which can make the model pay more attention to the region of interest (ROI) and improve the classiﬁcation accuracy (such as [41]). Third, as a natural phenomenon, the continuous evolution of the aurora gives it abundant intraclass forms while lacking discriminative interclass boundaries. The most frequently occurring arc auroras last a long time and exhibit high intraclass appearance variance, whereas there is little inter- class difference when auroras transit from one type to another as they are always gradually changing. In view of this, the large margin Softmax (L-Softmax) loss function [27] is adopted to optimize the auroral image classiﬁcation model. L-Softmax loss function simultaneously maximizes the intraclass compactness and interclass separability. We pretrain the classiﬁcation model on ImageNet and ﬁne-tune it with our auroral image data. I. INTRODUCTION A A URORA is a luminous glow of the upper atmosphere causedbyenergeticparticlesthatenterintotheatmosphere from the magnetosphere. Different auroral types are correlated with speciﬁc Earth’s magnetosphere activities, and effective auroral classiﬁcation beneﬁts the prediction of Earth’s magneto- sphere structure and space weather. How to accurately character- ize and reasonably classify aurora has become an indispensable part of the study of space physics [1], [2]. Recently, deep learning based algorithms have quickly domi- nated most vision-based tasks due to their discriminative power [21]. Image representations based on the convolutional neural network (CNN) have attracted increasing attention in the com- munity and demonstrated impressive performance [34], [38], [39]. The main difference between CNNs and traditional ma- chine learning methods is that CNNs learn features directly from the data without an additional manual feature extraction process. So far, several CNN models have been studied for the applications of various computer vision tasks, including OverFeat [28], AlexNet [32], GoogLeNet [35], VGGNet [36], ResNet [37], Inception [20], and their variants. Razavian et al. [29] used pretrained OverFeat to extract features and proved the off-the-shelf features yielded better results than the handcrafted features on various computer vision tasks. Aurora is always analyzed in a “case study” way by space scientists [3]–[6]. These studies, however, are subjective, opera- tor dependent, and very time consuming. Syrjäsuo and Donovan [8] ﬁrst introduced the computer vision technique into auroral These advances have been quickly brought into the ﬁeld of auroral image classiﬁcation. Wang and Yang [40] introduced AlexNet to automatically classify the dayside auroral images of 1) We propose an easy-to-implement yet effective CNN model for the auroral image classiﬁcation, which can help analyze the huge auroral image datasets. A. Spatial Transformer Network The correspondence between (xt i, yt i) and (xs i, ys i ) can be written as B. L-Softmax Loss dimensional matrix transformation parameters θ. The second part grid generator is used to map every coordinate (xt i, yt i) in the target image V to the coordinate (xs i, ys i ) in the source image U. The correspondence between (xt i, yt i) and (xs i, ys i ) can be written as A. Spatial Transformer Network Although CNN is a powerful classiﬁcation model, it is still affected by the spatial diversity of data. Jaderberg et al. [22] proposed a new learning module, spatial transformer network, tosolvethisproblem.STNcanbetrainedtopaymoreattentionto the ROIs. STN does not require the calibration of key points and can adaptively transform the input images according to a certain task (including translation, scaling, rotation, and other geometric transformations to select the most relevant regions). When the spatial variance of the input data is large, STN can be added to the existing CNN architecture to improve the classiﬁcation accuracy of the model. In this article, the STN is inserted in the input layer of the AlexNet network (see Fig. 1), and the auroral images are afﬁne transformed into the new images with strong characterization ability. As shown in Fig. 2, from the input U to the output V, STN consists of the localization network, grid generator, and sampler [22]. The ﬁrst part localization network is a custom CNN architec- ture that computes the generated two-dimensional (2-D) afﬁne transformation parameter θ. We deﬁne it as two convolutional layers, a max-pooling layer, a fully connected layer, and a regression layer. The network takes the input auroral images U ∈RH×W with width H and height W and outputs the 2 × 3 The main contributions of our work can be summarized as follows. 1) We propose an easy-to-implement yet effective CNN model for the auroral image classiﬁcation, which can help analyze the huge auroral image datasets. YANG AND ZHOU: REPRESENTATION AND CLASSIFICATION OF AURORAL IMAGES BASED ON CONVOLUTIONAL NEURAL NETWORKS 525 Fig. 1. Architecture of STN-Lsoftmax-AlexNet. Fig. 2. STN, (1)–(4) are executed sequentially. dimensional matrix transformation parameters θ. The second part grid generator is used to map every coordinate (xt i, yt i) in the target image V to the coordinate (xs i, ys i ) in the source image B. L-Softmax Loss It is important to choose an appropriate loss function for a speciﬁc task Softmax loss has been widely adopted by many Fig. 1. Architecture of STN-Lsoftmax-AlexNet. Fig. 2. STN, (1)–(4) are executed sequentially. Fig. 2. STN, (1)–(4) are executed sequentially. Fig. 2. STN, (1)–(4) are executed sequentially. dimensional matrix transformation parameters θ. The second part grid generator is used to map every coordinate (xt i, yt i) in the target image V to the coordinate (xs i, ys i ) in the source image U. IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 526 Therefore, the original Softmax loss can be written as Therefore, the original Softmax loss can be written as classiﬁed into two broad categories: discrete and diffuse auroras with structured forms or relatively homogenous luminosity, respectively [31]. Discrete auroras can be further classiﬁed into arc and corona according to their spatial morphology. Based on the observations acquired by the three-wavelength (427.8, 557.7, and 630.0 nm) all-sky imageries (ASIs) at YRS in Ny- ˚Alesund, Svalbard, Hu et al. [26] further classiﬁed the day- side coronal auroras into the radial corona, drapery corona, and hotspot. L = −log(oi) = 1 N i Li = 1 N i −log efyi j efj (5) (5) where fj denotes the class score of the jth element (j ∈[1, k], k is the number of classes), and N is the number of training data. If we omit the constant byi in fj, fj, it can also be formulated as fj = \|\|Wj\|\| \|\|xi\|\| cos(θj), 0 ≤θj ≤π (6) (6) Considering that the auroral data used in this article were obtained by the ASI during daytime at YRS, we classify the auroral images according to the classiﬁcation scheme proposed in [26], which was also applied in [13], [16], and [18]. Specif- ically, the dayside discrete auroras were classiﬁed into the arc, drapery corona, radial corona, and hotspot categories according to the spectral and morphological characteristics. Examples and characteristics of each type are given in Fig. 3. where θj is the angular margin. Then, the Softmax loss can further be deﬁned as Li = −log e\|\|Wyi \|\| \|\|xi\|\|ψ(θyi) j e\|\|Wj\|\| \|\|xi\|\| cos(θj) . (7) (7) A standard CNN can be viewed as a convolutional feature learning machine that is supervised by the Softmax loss [27]. Although the Softmax loss function has been widely used, it does not explicitly encourage discriminative learning of fea- tures. To solve this problem, Liu et al. [27] generalized the Softmax loss to a more general large-margin Softmax (L- Softmax) loss in terms of angular similarity, which leads to potentially larger angular separability between learned features and, thus, generates more discriminative features. The intuition behind L-Softmax is simple. Consider a binary classiﬁcation and a sample x from class 1, the original Softmax is to force \|\|W1\|\| \|\|x1\|\| cos(θ1) > \|\|W2\|\| \|\|x2\|\| cos(θ2) to classify x cor- rectly. IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 With larger m, the classiﬁcation margin becomes larger and the learning objective becomes harder. Speciﬁcally, when m = 1, the L-Softmax loss becomes identical to the original Softmax loss. D(θ) is a monotonically decreasing function and D(π/m) = cos(π/m). By adjusting the margin m between classes,arelativelydifﬁcultlearningobjectivewithanadjustable margin will be deﬁned, which can effectively avoid overﬁtting. In order to facilitate the following experiments, all images in the dataset were preprocessed in the same manner as in [16], including subtracting system noise and cutting off the outer regions of images, where signiﬁcant wide-angle distortion hap- pened and might contain YRS lights, and auroral images were cropped from 512 × 512 to 440 × 440 pixels ﬁnally. To beneﬁt the future research work in this ﬁeld, all datasets used in this article are released at https://github.com/wszph/Aurora_data. IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 Instead, L-Softmax requires \|\|W1\|\| \|\|x1\|\| cos(mθ1) > \|\|W2\|\| \|\|x2\|\| cos(θ2), (0 ≤θ1 ≤π/m), where m is a positive integer. Since m is a positive integer, and the cosine function is monotonically decreasing on the intervals [0, π]; thus, cos(mx) is less than cos(x). Therefore, the new classiﬁcation criteria is a stronger requirement to correctly classify x, producing a more rigorous decision boundary for class 1. The optical instruments at YRS capture photoemissions at 427.8, 557.7, and 630.0 nm during the winter season with a time resolution of 10 s. The ASI auroral images observed at YRS are available at http://www.chinare.org.cn/uap/database. In consideration of the image characteristics, we concentrated on the dayside aurora at 557.7 nm from December 2003 to February 2009. To better focus on the study of auroral image classiﬁcation, the images that do not contain auroral structures or were captured under bad weather conditions (e.g., auroral structures were severely covered by clouds) were eliminated by human visual inspection. Speciﬁcally, the dataset consists of three parts. 1) ASI8K contains 8000 images (2913 arc images, 1771 drapery corona images, 1640 radial corona images, and 1676 hotspot images, respectively) from December 2003 to February 2004, which is used to train the classiﬁ- cation network. Speciﬁcally, the ratio of training and validation is 4:1; that is, there are 6400 training images and 1600 validation images, respectively, in each training epoch. g y Formally, the L-Softmax loss is deﬁned as Formally, the L-Softmax loss is deﬁned as L −Softmax = −log e\|\|Wyi \|\| \|\|xi\|\|ψ(θyi) e\|\|Wyi \|\| \|\|xi\|\|ψ(θyi) + j̸=yi e\|\|Wj\|\| \|\|xi\|\| cos(θj) (8) n which ψ(θ) = cos(mθ), 0 < θ ≤π m D(θ), π m < θ ≤π (9) 2) ASI2K contains 2184 images with class labels in which each category has an approximately equal number of images from December 2004 to February 2009 (different winters from the training images). It is used in the testing phase to evaluate the performance of the proposed method. (8) in which in which (9) 3) ASI399K contains 399515 auroral images from 2004 to 2009 without the manual labels. Almost all the ob- served images are selected except those who have no auroral activity or only have diffuse auroras or under cloudy weather. It is used to make a statistical study of the temporal occurrence distributions of the four auroral categories. where m is an integer that controls the classiﬁcation margin [27]. B. L-Softmax Loss It is important to choose an appropriate loss function for a speciﬁc task. Softmax loss has been widely adopted by many CNNs due to its simplicity, probabilistic interpretation, and superior performance. Speciﬁcally, Softmax function is used to transform the prediction of the jth class for the ith input by (xs i, ys i ) = Tθ (Gi) = Aθ ⎛ ⎜ ⎝ xt i yt i 1 ⎞ ⎟ ⎠= θ11 θ12 θ13 θ21 θ22 θ23 ⎛ ⎜ ⎝ xt i yt i 1 ⎞ ⎟ ⎠ (1) σi(fyi) = exp (fyi) k j=1 exp (fyj) , i = 1, 2, . . . , k (2) (2) where Tθ is a 2-D afﬁne transformation function that is com- posed of the transformation parameter θ obtained in the ﬁrst part [22]. Aθ is the transformation matrix, and Gi is the mapping space grid coordinates, where i represents the ith pixel point in the target image V. The third part sampler ﬁlls the pixel value of the coordinate (xt i, yt i) in the target image V using both the pixel values of the source image U and the coordinate (xs i, ys i ) obtained in the second part. Since some of the coordinates mapped to the source image U may be decimals, we estimate the pixel values according to each surrounding pixel by bilinear interpolation. where yi is the sample class label, and fyi is usually the activa- tions of a fully connected layer and can be written as fyi = W T yixi + byi. (3) (3) Softmax turns the predictions into the nonnegative values and normalizes them to get a probability distribution over classes. The probability that data x belongs to class i is called likelihood. Softmax turns the predictions into the nonnegative values and normalizes them to get a probability distribution over classes. The probability that data x belongs to class i is called likelihood. oi = σi(fyi). (4) (4) III. AURORAL CLASSIFICATION MECHANISM AND DATASET SETUP So far, there is no uniform classiﬁcation scheme for auroral images. Generally,opticalaurorasobservedonthegroundcanbe YANG AND ZHOU: REPRESENTATION AND CLASSIFICATION OF AURORAL IMAGES BASED ON CONVOLUTIONAL NEURAL NETWORKS 527 Examples and characteristics of the four auroral types sampled from the dataset. (a) Arc. (b) Drapery Corona. (c) Radial Corona. (d) Hot-spot IV. EXPERIMENTS AND RESULTS is section, auroral image classiﬁcation and retrieval A. Implementation Details In this article, the AlexNet was pretrained on the Im Fig. 3. Examples and characteristics of the four auroral types sampled from the dataset. (a) Arc. (b) Drapery Corona. (c) Radial Corona. (d) Hot-spot. TABLE I COMPARISON OF ALEXNET, VGG16, AND INCEPTION-V4 of the last FC8 layer, including four nodes, can be used as the category labels in the task of auroral image classiﬁcation. Fig. 4 illustrates the diagram of the auroral image feature extraction and classiﬁcation using AlexNet architecture. The network was trained and optimized with the L-Softmax loss function. All weights were learned via backpropagation and stochastic gradient descent. Each minibatch consisted of 64 semantic regions for the training dataset or 16 for the validation dataset, which comprised the predeﬁned four auroral categories randomly. We used an initial learning rate of 0.00003, a momen- tum of 0.9, and a weight decay of 0.0005. The experiments were performed on a PC with a 3.4 GHz Intel i7-6700 CPU under a Linux system. deeper and more complex networks. However, the deeper the network, the more likely it is to have a gradient dispersion problem and more challenging to optimize the model. The clas- siﬁcation accuracy and running time were compared in Table I. The best results are highlighted with bold fonts. It is clear that the AlexNet achieves the best performance in terms of both average accuracy and efﬁciency on datasets ASI8K and ASI2K. Therefore, the following experiments are based on AlexNet. Considering that the L-Softmax loss had difﬁculty in conver- gence, a decay factor λ was added to the learning strategy [27], which was expressed as fyi = λ\|\|Wyi\|\| \|\|xi\|\| cos(θyi) + \|\|Wyi\|\| \|\|xi\|\|ψ(θyi) 1 + λ (10) b) Component ablation studies: In this section, we present the ablation studies to isolate the effect of each module (STN andL-Softmax)oftheproposedSTN-Lsoftmax-AlexNetmodel. Fig. 5 illustrates the comparison of the classiﬁcation results ob- tained by different methods. The dotted green line and the solid mauve line depict the performance of the original AlexNet and combined AlexNet with STN, respectively, the dotted blue lines represent that of the L-Softmax loss based AlexNet, and the solid red lines show the performance of STN-Lsoftmax-AlexNet. The following conclusions can be drawn from Fig. 5. where λ is a large number at the beginning of gradient descent, and it is gradually reduced during iteration. In our experiments, the initial value of λ was set to 10000 and the minimum value was set to 15 by visualizing the convergence performance of the network in the training process. IV. EXPERIMENTS AND RESULTS In this section, auroral image classiﬁcation and retrieval experiments were carried out on ASI8K and ASI2K. The importance of each module of the STN-Lsoftmax-AlexNet model was veriﬁed and the accuracy and efﬁciency of the model were compared with the existing auroral image classi- ﬁcation methods. Besides, a statistical analysis of the temporal occurrence distribution of auroral types was further made on ASI399K. In this article, the AlexNet was pretrained on the ImageNet dataset as a feature extraction tool and all parameter settings were obtained from the Caffe toolbox [33]. Generally, the AlexNet consists of ﬁve convolutional layers (abbreviated as Conv1–Conv5) and three fully connected layers (FC6–FC8). With the increase of the layer number, the outputs represent the higher semantic features. Speciﬁcally, the input image is resized to 256 × 256 before fed into the AlexNet, and the output IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 528 Fig. 4. Block diagram illustrates the feature extraction and classiﬁcation of auroral images using AlexNet architecture. C represents the convolutional layer, S is the max-pooling layer, and F refers to the fully connected layer. Fig. 4. Block diagram illustrates the feature extraction and classiﬁcation of auroral images using AlexNet architecture. C represents the convolutional layer, S is the max-pooling layer, and F refers to the fully connected layer. B. Experiments and Analysis In terms of running time, our method takes 12.43 ms to predict the class label of a testing image, which is 2.27 ms more than the basic AlexNet. This is related to the part of the local network of STN (see Fig. 2). And the difference of m value of the loss function has little effect on the time cost as long as the model structure is the same. Inception-v4 and RSM are much more complex networks than our method, taking 40.41 and 1233.8 ms, respectively, to predict the label of an image, which is much slower than our method. In brief, the proposed method that combines STN and L-Softmax loss function based AlexNet is both effective and efﬁcient for auroral image classiﬁcation. hotspot and drapery coronas, respectively. It also indicates that STN is effective in the spatial transformation of the original auroral images during the training process. 3) The red lines show a higher classiﬁcation rate than the mauve line and the dotted blue lines show a higher classiﬁ- cationratethanthedottedgreenline.Theaverageaccuracy has been improved by about 2%, while radial coronas get an improvement of 5.2% with m = 4. This proves that the L-Softmax loss function can guide the network to learn more distinctive auroral features. However, as m increases, the model performance was not improved continuously. From Fig. 5, we can see that with m = 3, all categories achieve the best classiﬁcation results. The following experiments are, therefore, based on the STN- Lsoftmax-AlexNet with m = 3. 3) The red lines show a higher classiﬁcation rate than the mauve line and the dotted blue lines show a higher classiﬁ- cationratethanthedottedgreenline.Theaverageaccuracy has been improved by about 2%, while radial coronas get an improvement of 5.2% with m = 4. This proves that the L-Softmax loss function can guide the network to learn more distinctive auroral features. However, as m increases, the model performance was not improved continuously. From Fig. 5, we can see that with m = 3, all categories achieve the best classiﬁcation results. The following experiments are, therefore, based on the STN- Lsoftmax-AlexNet with m = 3. c) Comparison with existing methods: We compare the performance of the proposed STN-Lsoftmax-AlexNet model (m = 3) in terms of both accuracy and time costs with those reported in the state-of-the-art on auroral image classiﬁcation using deep learning techniques. B. Experiments and Analysis 1) Supervised Classiﬁcation: The following three super- vised classiﬁcation experiments, including base model selec- tion, component ablation studies, and comparison with existing methods, were all trained on dataset ASI8K and tested on dataset ASI2K. The classiﬁcation rate of each auroral type was calcu- lated by dividing the number of correctly classiﬁed images by the total labeled number of that type, and the average classiﬁcation accuracy was equal to the number of correctly classiﬁed images divided by the total number of testing images. 1) Supervised Classiﬁcation: The following three super- vised classiﬁcation experiments, including base model selec- tion, component ablation studies, and comparison with existing methods, were all trained on dataset ASI8K and tested on dataset ASI2K. The classiﬁcation rate of each auroral type was calcu- lated by dividing the number of correctly classiﬁed images by the total labeled number of that type, and the average classiﬁcation accuracy was equal to the number of correctly classiﬁed images divided by the total number of testing images. 1) The classiﬁcation accuracy of each auroral type indicates that almost all methods are easier to recognize the arc and drapery corona. In contrast, the accuracy of the hotspot aurora is relatively low. This is directly related to the morphologicalcomplexityoftheseauroraltypes,asshown in Fig. 3. 2) The classiﬁcation accuracy shown by the solid lines is higher than that shown by the dotted lines, which indicates that the combination of STN and AlexNet improves the accuracy of the classiﬁcation model. The average accuracy improvement is about 1%–2% over no STN structures, and the improvements are up to 4.4% and 3.5% for the a) Base model selection: We compared the performance of three popular CNN models, AlexNet, VGG16, and Inception- v4, which have been used for the automatic analysis of auroral images before. Compared with AlexNet, VGG16 and Inception- v4 are more powerful for feature extraction because of their YANG AND ZHOU: REPRESENTATION AND CLASSIFICATION OF AURORAL IMAGES BASED ON CONVOLUTIONAL NEURAL NETWORKS 529 Fig. 5. Comparison of the performance of different methods. TABLE II COMPARISON WITH EXISTING METHODS Fig. 5. Comparison of the performance of different methods. Inception-v4, and a region scale model (RSM) applied to the auroral image classiﬁcation in [40], [30], and [10], respectively. Table II depicts the comparison results. Our method achieves much higher classiﬁcation accuracy than the previous methods. B. Experiments and Analysis Speciﬁcally, they are AlexNet, Furthermore, our experimental setting is different from the previous automatic auroral image classiﬁcation works (e.g., [13], [16], [19], and [30]). The training and testing data in these articles came from the same year(s), which were obtained by shufﬂing and randomly dividing the auroral data from one or 530 IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 Some auroral types, such as the drapery and radial coronas, have no distinct shape boundaries to be segmented. Actually, one of the uses of aurora segmentation is to improve the accuracy of aurora classiﬁcation [7], [10]. classiﬁcation accuracy of the four auroral categories indicates that the arc and drapery coronas are more easily recognized, and radial coronas might be classiﬁed as drapery coronas or hotspot. The classiﬁcation accuracy of hotspot auroras is somewhat low and they are likely to be classiﬁed as auroral arcs or radial coronas. The reason is that there are many complex auroral structures in the hotspot displays, such as rays, beams, spots, and irregular patches, as shown in Fig. 3(d). classiﬁcation accuracy of the four auroral categories indicates that the arc and drapery coronas are more easily recognized, and radial coronas might be classiﬁed as drapery coronas or hotspot. The classiﬁcation accuracy of hotspot auroras is somewhat low and they are likely to be classiﬁed as auroral arcs or radial coronas. The reason is that there are many complex auroral structures in the hotspot displays, such as rays, beams, spots, and irregular patches, as shown in Fig. 3(d). g p g 2) Image Retrieval: In recent decades, with the increasing number of auroral images, it is urgent to retrieve the images of in- terest quickly and effectively from massive auroral observations. In this section, the F7 layer of AlexNet was extracted from the trained STN-Lsoftmax-AlexNet model as the feature vector to represent each image, and the Euclidean distance between two features was used to measure the similarity between the two images. The representation ability of the proposed model is g p g 2) Image Retrieval: In recent decades, with the increasing number of auroral images, it is urgent to retrieve the images of in- terest quickly and effectively from massive auroral observations. In this section, the F7 layer of AlexNet was extracted from the trained STN-Lsoftmax-AlexNet model as the feature vector to represent each image, and the Euclidean distance between two features was used to measure the similarity between the two images. IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 TABLE III CONFUSION MATRIX OF SUPERVISED CLASSIFICATION ON TESTING IMAGES WITH STN-LSOFTMAX-ALEXNET MODEL (M = 3) TABLE III CONFUSION MATRIX OF SUPERVISED CLASSIFICATION ON TESTING IMAGES WITH STN-LSOFTMAX-ALEXNET MODEL (M = 3) Retrieval results of four types of auroral images Query images are in the ﬁrst column where the red green yellow and blue boxes denote the arc d TABLE III CONFUSION MATRIX OF SUPERVISED CLASSIFICATION ON TESTING IMAGES WITH STN-LSOFTMAX-ALEXNET MODEL (M = 3) TABLE III 6. Retrieval results of four types of auroral images. Query images are in the ﬁrst column, where the red, green, yellow, and blue boxes denote the arc, al, and hotspot auroras, respectively. The second–eighth columns are the returned retrieved images (ranks 1, 3, 5, 8, 10, 15, and 20, respectively.). . 6. Retrieval results of four types of auroral images. Query images are in the ﬁrst column, where the red, green, yellow, and blue boxes denote the arc, dra ial, and hotspot auroras, respectively. The second–eighth columns are the returned retrieved images (ranks 1, 3, 5, 8, 10, 15, and 20, respectively.). Fig. 6. Retrieval results of four types of auroral images. Query images are in the ﬁrst column, where the red, green, yellow, and blue boxes denote the arc, drapery, radial, and hotspot auroras, respectively. The second–eighth columns are the returned retrieved images (ranks 1, 3, 5, 8, 10, 15, and 20, respectively.). Fig. 6. Retrieval results of four types of auroral images. Query images are in the ﬁrst column, where the red, green, yellow, and blue boxes denote the arc, drapery, radial, and hotspot auroras, respectively. The second–eighth columns are the returned retrieved images (ranks 1, 3, 5, 8, 10, 15, and 20, respectively.). YANG AND ZHOU: REPRESENTATION AND CLASSIFICATION OF AURORAL IMAGES BASED ON CONVOLUTIONAL NEURAL NETWORKS 531 Fig. 7. Temporal occurrence distributions of dayside auroral categories. 7. Temporal occurrence distributions of dayside auroral categories. Fig. 7. Temporal occurrence distributions of dayside auroral categories. several years. The model in this study is trained on one-year auroral data and then tested on the following ﬁve-year auro- ral data according to the practical needs of auroral physics. In addition, although some advanced issues, such as weakly supervised semantic segmentation, have been developed [10], the task of automatic aurora classiﬁcation is more signiﬁcant and practical in auroral physics. V. CONCLUSION 3) Occurrence Distributions: The auroral events caused by similar or repeatable physical processes share the same funda- mental morphology; thus, a statistical study of the occurrence time of auroral types is important to our understanding of magnetospheric dynamics. In this section, the images in dataset ASI399K from 2004 to 2009 were predicted by the trained STN-Lsoftmax-AlexNet model, and the temporal distribution of the four auroral types was drawn in Fig. 7. The temporal axis between 6 and 18 MLT was divided into 240 bins of 3 min duration. At the top of these panels, four active regions proposed in [26] were denoted and partitioned by bold dashed lines, while the two states of the midday gap region were indicated by a thin dashed line. From top to bottom, the ﬁrst panel shows the frame number of all predicted images occurring in each time period. The second–ﬁfth panels show the occurrence ratio of each category, which is the result of dividing the number of images of each type within 3 min by the total number of images in that time period, respectively. Auroral image retrieval experiments were also carried out to visually evaluate the representation ability of the proposed model. The retrieval results are basically in accordance with human visual judgments. In addition, to verify the generalization performance of the proposed model, a larger dataset containing nearly 0.4 million auroral images from 2004 to 2009 was tested and the category of each image was predicted. The tempo- ral occurrence distributions of the dayside auroral types were in accordance with the empirical rules of auroral physics. In summary, the proposed model has the advantages of excellent classiﬁcation accuracy, high efﬁciency, good generalization, and simple implementation, and thus can be truly and widely applied for aurora classiﬁcation. The proposed classiﬁcation model is data-driven and does not rely on handcrafted designs, so it can also be applied to the classiﬁcation of auroral images from other stations or with different classiﬁcation schemes. Just providing a set of labeled data from these new stations according to a speciﬁc classiﬁcation scheme, the model will be retrained and its performance can be assessed with these data. Once the vast amounts of the existing ground-based auroral data are classiﬁed, we will be able to make large statistical studies to analyze the physical mechanisms of each auroral type. As shown in Fig. V. CONCLUSION Fig. 6 shows examples of the query and retrieved images. The ﬁrst column shows the query images, and the second–eighth columns are the most similar (ranks 1, 3, 5, 8, 10, 15, and 20, respectively)retrievedimagesbythetrainedmodel,respectively. In the query images, (a) and (b) with red boxes are auroral arcs, (c) and (d) with green boxes are drapery coronas, (e) and (f) with yellow boxes are radial coronas, and (g) and (h) with blue boxes are hotspot auroras. From the perspective of human vision, the query images and their retrieved results are very similar. Except forafewimagesmarkedwithcolorboxes[thatis,rank3andrank 20inrow(c)andrank15inrow(f)],theretrievedimageshavethe same category labels (manually labeled) as their query images. This proves that the proposed model is able to characterize the key features of auroral images. As a result, we can utilize this model to select the interested auroral images from the massive auroral observations. For the drapery and radial coronas, some retrieved images have different category labels (manual labels) from their query images, such as the images marked with yellow and green boxes in Fig. 6(c) and (f), respectively. This indicates that extracting only one layer (F7 layer) from the classiﬁcation model is not enough to distinguish the two coronal auroras. We developed an automatic representation and classiﬁcation method based on CNNs to extract auroral features and classify the auroral images into the arc, drapery corona, radial corona, and hotspot. Aurora is a natural phenomenon whose morphol- ogy continuously changes over time, and there is no obvious boundary when it gradually transits from one type to another. Therefore,amajorchallengeforauroralimageclassiﬁcationisto maximize the intraclass compactness and interclass separability of auroral features. To tackle this problem, this article combines the STN and AlexNet and optimizes the network by using the L-Softmax loss function. The classiﬁcation results show the following. 1) AlexNet is a good choice for the auroral image classiﬁca- tion in terms of both accuracy and efﬁciency. 2) L-Softmax loss function can guide the network to learn more discriminative auroral features and, thus, improves the classiﬁcation accuracy, especially that of some subcat- egories (e.g., radial coronas get an improvement of 5.2% with m = 4). 3) STNisabletoactivelyspatiallytransformtheinputimages and simplify the subsequent classiﬁcation task, and thus lead to superior classiﬁcation performance (the improve- ments are up to 4.4% and 3.5% for the hotspot and drapery coronas, respectively, as compared with no STN modules). V. CONCLUSION 7, dayside auroras mainly occur before noon, followed by the afternoon, and rarely occurs at noon. More arcs and drapery coronas were observed than radial coronas and hotspot auroras from 2004 to 2009 in YRS. The auroral arc has a double-peak distribution around noon and the forenoon peak is weaker than the postnoon peak. Both drapery coronas and radial coronas predominantly occur before 13 MLT but with different peak positions. The occurrence of these auroral categories dominates the different regions of the dayside oval. The drapery coronas occur more often in region G, whereas the radial coronas occupy in region R, and the hotspot auroras dominate region H. Their global distributions approximately coincide with the multiple-wavelength intensity distribution of IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 532 veriﬁed by image retrieval experiments. Dataset ASI8K from 2003 to 2004 was used as the query images, and the retrieval experiments were carried out on dataset ASI2K from 2004 to 2009. We chose ASI2K, instead of ASI399K, as the retrieved dataset, because all images in it have class labels that could help to evaluate the retrieval results. the dayside aurora presented in [26]. This proves that our method can be used for the automatic classiﬁcation of a huge amount of auroral images in practical applications. IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 The representation ability of the proposed model is d) Classiﬁcation confusion matrix: In order to quantita- tively evaluate the classiﬁcation effectiveness of our proposed STN-Lsoftmax-AlexNet model (m = 3) on each auroral type, we calculated the classiﬁcation rate of each auroral type and the averageclassiﬁcationaccuracyof all testingimages anddepicted the confusion matrix of the classiﬁcation results in Table III. The REFERENCES Yang, J. Liang, Z. Hu, Z. Xing, and H. Zhao, “Automatic recognition of poleward moving auroras from all-sky image sequences based on HMM and SVM,” Planet. Space Sci., vol. 69, no. 1, pp. 40–48, Aug. 2012. [36] K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” in Proc. Int. Conf. Learn. Represent., San Diego, CA, USA, May 7–9, 2015, pp. 1–14. [37] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit., Las Vegas, NV, USA, Jun. 27–30, 2016, pp. 770–778. [15] X. Gao, R. Fu, X. Li, D. Tao, B. Zhang, and H. Yang, “Aurora image segmentation by combining patch and texture thresholding,” Comput. Vision Image Understanding, vol. 115, no. 3, pp. 390–402, Mar. 2011. [38] Y. Yu, Z. Gong, C. Wang, and P. Zhong, “An unsupervised convolu- tional feature fusion network for deep representation of remote sensing images,” IEEE Geosci. Remote Sens. Lett., vol. 15, no. 1, pp. 23–27, Jan. 2018. [16] Q. Wang et al., “Spatial texture based automatic classiﬁcation of dayside aurora in all-sky images,” J. Atmos. Sol.-Terr. Phys., vol. 72, no. 5/6, pp. 498–508, Apr. 2010. [17] J. Rao, N. Partamies, O. Amariutei, M. Syrjäsuo, and K. E. van de Sande, “Automatic auroral detection in color all-sky camera images,” IEEE J. Sel. Topics Appl. Earth Observ. Remote Sens., vol. 7, no. 12, pp. 4717–4725, Dec. 2014. [39] H. Liang, X. Sun, Y. Sun, and Y. Gao, “Text feature extraction based on deep learning: A review,” EURASIP J. Wireless Commun. Netw., vol. 2017, Dec. 2017, Art. no. 211. [18] Q. Yang and Z. Hu, “An automatic auroral classiﬁcation method based on morphological characteristics (in Chinese),” Scientia Sinica Terrae, vol. 47, no. 2, pp. 252–260, Jan. 2017. [40] F. Wang and Q. Yang, “Classiﬁcation of auroral images based on convo- lutional neural network,” Chin. J. Polar Res., vol. 30, no. 2, pp. 123–131, Jun. 2018. [41] R. Q. Charles, H. Su, M. Kaichun, and L. J. Guibas, “PointNet: Deep learning on point sets for 3D classiﬁcation on and segmentation,” in Proc. IEEE Conf. Comput. Vision Pattern Recognit., Honolulu, HI, USA, Jul. 21–26, 2017, pp. 77–85. [19] Y. Zhong, R. Huang, J. Zhao, B. Zhao, and T. Liu, “Aurora image classi- ﬁcation based on multi-feature latent Dirichlet allocation,” Remote Sens., vol. 10, no. 2, pp. 233–249, 2018. [20] C. Szegedy, S. REFERENCES [24] X. Yang, X. Gao, B. Song, N. Wang, and D. Yang, “ASI aurora search: An attempt of intelligent image processing for circular ﬁsheye lens,” Opt. Express, vol. 26, no. 7, pp. 7985–8000, Jun. 2018. [1] T. R. Pedersen and E. A. Gerken, “Creation of visible artiﬁcial optical emissions in the aurora by high-power radio waves,” Nature, vol. 433, no. 7025, pp. 498–500, Feb. 2005. [25] B. Han, F. Chu, X. Gao, and Y. Yan, “A multi-size kernels CNN with eye movement guided taskspeciﬁc initialization for aurora image clas- siﬁcation,” in Proc. CCF Chin. Conf. Comput. Vision, Tianjin, China, Nov. 2017, pp. 533–544. pp [2] S.-I. Akasofu, “The development of the auroral substorm,” Planet. Space Sci., vol. 12, no. 4, pp. 273–282, Apr. 1964. [26] Z.-J. Hu et al., “Synoptic distribution of dayside aurora: Multiple- wavelength all-sky observation at yellow river station in Ny- ˚Alesund, Svalbard,” J. Atmos. Sol.-Terr. Phys., vol. 71, no. 8/9, pp. 794–804, Jun. 2009. [3] D.-S. Han, Y. Nishimura, L. R. Lyons, H.-Q. Hu, and H. Yang, “Throat aurora: The ionospheric signature of magnetosheath particles pene- trating into the magnetosphere,” Geophys. Res. Lett., vol. 43, no. 5, pp. 1819–1827, Feb. 2016. [27] W. Liu, Y. Wen, Z. Yu, and M. Yang, “Large-margin softmax loss for convolutional neural networks,” in Proc. 33rd Int. Conf. Mach. Learn., New York, NY, USA, Jun. 19–24, 2016, pp. 507–516. [4] Z.-J.Huetal.,“Thehemisphericconjugateobservationofpostnoon“bright spots”/auroral spirals,” J. Geophys. Res., Space Phys., vol. 118, no. 4, pp. 1428–1434, Apr. 2013. [5] Q.-H. Zhang et al., “Direct observations of the evolution of polar cap ionization patches,” Science, vol. 339, pp. 1597–1600, Mar. 2013. [28] P. Sermanet, D. Eigen, X. Zhang, M. Mathieu, R. Fergus, and Y. LeCun, “Overfeat: Integrated recognition, localization and detection using convo- lutional networks,” in Proc. 2nd Int. Conf. Learn. Represent., Banff, AB, Canada, Apr. 14–16, 2014, pp. 1–16. [6] Z. Xing et al., “Conjugate observations of the evolution of polar cap arcs in both hemispheres,” J. Geophys. Res., Space Phys., vol. 123, no. 3, pp. 1794–1805 Mar. 2018. [29] A. S. Razavian, H. Azizpour, J. Sullivan, and S. Carlsson, “CNN features off-the-shelf: An astounding baseline for recognition,” in Proc. IEEE Conf. Comput. Vision Pattern Recognit. Workshops, Columbus, OH, USA, Jun. 23–28, 2014, pp. 512–519. [7] Q. Yang, D. Tao, D. Han, and J. Liang, “Extracting auroral key local structures from all-sky auroral images by artiﬁcial intelligence technique,” J. Geophys. REFERENCES Res., Space Phys., vol. 124, pp. 3512–3521, May 2019, doi: 10.1029/2018JA026119. [30] L. B. Clausen and H. Nickisch, “Automatic classiﬁcation of auroral images from the Oslo auroral THEMIS (OATH) data set using ma- chine learning,” J. Geophys. Res., Space Phys., vol. 120, pp. 7447–7465, Sep. 2015. [8] M. T. Syrjäsuo and E. F. Donovan, “Diurnal auroral occurrence statis- tics obtained via machine vision,” Ann. Geophys., vol. 22, no. 4, pp. 1103–1113, Apr. 2004. [9] X. Yang, X. Gao, and Q. Tian, “Polar embedding for aurora image retrieval,” IEEE Trans. Image Process., vol. 24, no. 11, pp. 3332–3344, Nov. 2015. [31] D.-S. Han et al., “An extensive survey of dayside diffuse aurora based on optical observations at yellow river station,” J. Geophys. Res., Space Phys., vol. 120, pp. 7447–7465, Sep. 2015. y pp p [32] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classiﬁcation with deep convolutional neural networks,” in Proc. Adv. Neural Inf. Pro- cess. Syst., Lake Tahoe, NV, USA, Dec. 3–6, 2012, pp. 1097–1105. [10] C. Niu, J. Zhang, Q. Wang, and J. Liang, “Weakly supervised semantic segmentation for joint key local structure localization and classiﬁcation of aurora image,” IEEE Trans. Geosci. Remote Sens., vol. 56, no. 12, pp. 7133–7146, Dec. 2018. [33] Y. Jia et al., “Caffe: Convolutional architecture for fast feature embedding,” in Proc. 22nd ACM Int. Conf. Multimedia, Orlando, FL, USA, Nov. 3– 7, 2014, pp. 675–678. [11] X. Yang, X. Gao, D. Tao, and X. Li, “Improving level set method for fast auroral oval segmentation,” IEEE Trans. Image Process., vol. 23, no. 7, pp. 2854–2865, Jul. 2014. pp [34] Y. Chen, H. Jiang, C. Li, X. Jia, and P. Ghamisi, “Deep feature extrac- tion and classiﬁcation of hyperspectral images based on convolutional neural networks,” IEEE Trans. Geosci. Remote Sens., vol. 54, no. 10, pp. 6232–6251, Oct. 2016. [12] X. Yang, X. Gao, D. Tao, X. Li, B. Han, and J. Li, “Shape-constrained sparse and low-rank decomposition for auroral substorm detection,” IEEE Trans. Neural Netw. Learn. Syst., vol. 27, no. 1, pp. 32–46, Jan. 2016. [13] Q. Yang, J. Liang, Z. Hu, and H. Zhao, “Auroral sequence representation and classiﬁcation using hidden Markov models,” IEEE Trans. Geosci. Remote Sens., vol. 50, no. 12, pp. 5049–5060, Dec. 2012. [35] C. Szegedy et al., “Going deeper with convolutions,” in Proc. IEEE Conf. Comput. Vision Pattern Recognit., Boston, MA, USA, Jun. 8–10, 2015, pp. 1–9. [14] Q. YANG AND ZHOU: REPRESENTATION AND CLASSIFICATION OF AURORAL IMAGES BASED ON CONVOLUTIONAL NEURAL NETWORKS 533 ACKNOWLEDGMENT The authors would like to thank the Polar Research Institute of China for providing the auroral data. REFERENCES Ioffe, V. Vanhoucke, and A. Alemi, “Inception-v4, inception-resnet and the impact of residual connections on learning,” in Proc. 31st AAAI Conf. Artif. Intell., San Francisco, CA, USA, Feb. 4–9, 2017, pp. 4278–4284. [42] G. Cheng, P. Zhou, and J. Han, “RIFD-CNN: Rotation-invariant and ﬁsher discriminative convolutional neural networks for object detection,” in Proc. IEEE Conf. Comput. Vision Pattern Recognit., Las Vegas, NV, USA, Jun. 2016, pp. 2884–2893. [21] L. Zheng, Y. Yang, and Q. Tian, “SIFT meets CNN: A decade survey of instance retrieval,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 40, no. 5, pp. 1224–1244, May 2018. [43] G. Cheng, C. Yang, X. Yao, L. Guo, and J. Han, “When deep learning meets metric learning: Remote sensing image scene classiﬁcation via learning discriminative CNNs,” IEEE Trans. Geosci. Remote Sens., vol. 56, no. 5, pp. 2811–2821, May 2018. [22] M. Jaderberg, K. Simonyan, A. Zisserman, and K. Kavukcuoglu, “Spatial transformer networks,” in Proc. Adv. Neural Inf. Process. Syst.,” Montreal, QC, Canada, Dec. 7–12, 2015, pp. 2017–2025. [44] G. Cheng, J. Han, P. Zhou, and D. Xu, “Learning rotation-invariant and ﬁsher discriminative convolutional neural networks for object de- tection,” IEEE Trans. Image Process., vol. 28, no. 1, pp. 265–278, Jan. 2019. [23] X. Yang, X. Gao, B. Song, and D. Yang, “Aurora image search with contextual CNN feature,” Neurocomputing, vol. 281, no. 15, pp. 67–77, Mar. 2018. IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 534 IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 Penghui Zhou received the B.S. degree in electronic information science and technology from the Baoji University of Arts and Sciences, Baoji, China, in 2017, and the M.S. degree in electronics and com- munication engineering from Shaanxi Normal Uni- versity, Xi’an, China, in 2019. His main research interests include auroral image processing and analysis. Penghui Zhou received the B.S. degree in electronic information science and technology from the Baoji University of Arts and Sciences, Baoji, China, in 2017, and the M.S. degree in electronics and com- munication engineering from Shaanxi Normal Uni- versity, Xi’an, China, in 2019. His main research interests include auroral image processing and analysis. Qiuju Yang received the B.S. degree in electronic information science and technology from Shaanxi Normal University, Xi’an, China, in 2008, and the Ph.D. degree in pattern recognition and intelligence systems from Xidian University, Xi’an, China, in 2013. In 2013, she joined the School of Physics and Information Technology, Shaanxi Normal Univer- sity, where she is currently an Associate Professor in electronics and information systems. From 2018 to 2019, she was a Visiting Scholar with UBTECH Sydney AI Centre, University of Sydney, Sydney, NSW, Australia. Her research interests include machine learning, aurora physics, and computer vision. She has authored or coauthored more than ten papers in referred journals, including IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING and Journal of Geophysical Research: Space Physics. Penghui Zhou received the B.S. degree in electronic information science and technology from the Baoji University of Arts and Sciences, Baoji, China, in 2017, and the M.S. degree in electronics and com- munication engineering from Shaanxi Normal Uni- versity, Xi’an, China, in 2019. Qiuju Yang received the B.S. degree in electronic information science and technology from Shaanxi Normal University, Xi’an, China, in 2008, and the Ph.D. degree in pattern recognition and intelligence systems from Xidian University, Xi’an, China, in 2013. In 2013, she joined the School of Physics and Information Technology, Shaanxi Normal Univer- sity, where she is currently an Associate Professor in electronics and information systems. From 2018 to 2019, she was a Visiting Scholar with UBTECH Sydney AI Centre, University of Sydney, Sydney, NSW, Australia. Her research interests include machine learning, aurora physics, and computer vision. She has authored or coauthored more than ten papers in referred journals, including IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING and Journal of Geophysical Research: Space Physics. IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 13, 2020 His main research interests include auroral image processing and analysis.
https://openalex.org/W4294868092	https://www.dora.lib4ri.ch/psi/islandora/object/psi%3A50948/datastream/PDF/Schobesberger-2023-Airborne_flux_measurements_of_ammonia-%28published_version%29.pdf	English	null	Airborne flux measurements of ammonia over the Southern Great Plains using chemical ionization mass spectrometry	null	2,022	cc-by	26,198	Siegfried Schobesberger1,2,3, Emma L. D’Ambro4,a, Lejish Vettikkat2, Ben H. Lee1, Qiaoyun Peng1, David M. Bell5,6, John E. Shilling5, Manish Shrivastava5, Mikhail Pekour5, Jerome Fast5, and Joel A. Thornton1 Siegfried Schobesberger1,2,3, Emma L. D’Ambro4,a, Lejish Vettikkat2, Ben H. Lee1, Qiaoyun Peng1, David M. Bell5,6, John E. Shilling5, Manish Shrivastava5, Mikhail Pekour5, Jerome Fast5, and Joel A. Thornton1 1Department of Atmospheric Sciences, University of Washington, Seattle, WA, USA 2Department of Technical Physics, University of Eastern Finland, Kuopio, Finland 3Department of Physics, University of Helsinki, Helsinki, Finland 4Department of Chemistry, University of Washington, Seattle, WA, USA 5Atmospheric Sciences and Global Change Division, Paciﬁc Northwest National Laboratory, Richland, WA, USA 6Laboratory of Atmospheric Chemistry, Paul Scherrer Institute, Villigen, Switzerland anow at: Ofﬁce of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC, USA 1Department of Atmospheric Sciences, University of Washington, Seattle, WA, USA 2Department of Technical Physics, University of Eastern Finland, Kuopio, Finland 3Department of Physics, University of Helsinki, Helsinki, Finland 4Department of Chemistry, University of Washington, Seattle, WA, USA 5Atmospheric Sciences and Global Change Division, Paciﬁc Northwest National Laboratory, Richland, WA, USA 6Laboratory of Atmospheric Chemistry, Paul Scherrer Institute, Villigen, Switzerland anow at: Ofﬁce of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC, USA Correspondence: Siegfried Schobesberger (siegfried.schobesberger@uef.ﬁ) Received: 24 August 2022 – Discussion started: 7 September 2022 Revised: 14 December 2022 – Accepted: 22 December 2022 – Published: 19 January 2023 Received: 24 August 2022 – Discussion started: 7 September 2022 Received: 24 August 2022 – Discussion started: 7 September 2022 Received: 24 August 2022 – Discussion started: 7 September 2022 Revised: 14 December 2022 – Accepted: 22 December 2022 – Published: 19 January 2023 Revised: 14 December 2022 – Accepted: 22 December 2022 – Published: 19 January 2023 Abstract. Ammonia (NH3) is an abundant trace gas in the at- mosphere and an important player in atmospheric chemistry, aerosol formation and the atmosphere–surface exchange of nitrogen. The accurate determination of NH3 emission rates remains a challenge, partly due to the propensity of NH3 to interact with instrument surfaces, leading to high detection limits and slow response times. In this paper, we present a new method for quantifying ambient NH3, using chemical ionization mass spectrometry (CIMS) with deuterated ben- zene cations as reagents. The setup aimed at limiting sample– surface interactions and achieved a 1σ precision of 10– 20 pptv and an immediate 1/e response rate of < 0.4 s, which compares favorably to the existing state of the art. The sensi- tivity exhibited an inverse humidity dependence, in particular in relatively dry conditions. Siegfried Schobesberger1,2,3, Emma L. D’Ambro4,a, Lejish Vettikkat2, Ben H. Lee1, Qiaoyun Peng1, David M. Bell5,6, John E. Shilling5, Manish Shrivastava5, Mikhail Pekour5, Jerome Fast5, and Joel A. Thornton1 Background of up to 10 % of the total signal required consideration as well, as it responded on the order of a few minutes. To showcase the method’s capa- bilities, we quantiﬁed NH3 mixing ratios from measurements obtained during deployment on a Gulfstream I aircraft dur- ing the HI-SCALE (Holistic Interactions of Shallow Clouds, Aerosols, and Land-Ecosystems) ﬁeld campaign in rural Ok- lahoma during May 2016. Typical mixing ratios were 1–10 parts per billion by volume (ppbv) for the boundary layer and 0.1–1 ppbv in the lower free troposphere. Sharp plumes of up to tens of ppbv of NH3 were encountered as well. We identiﬁed two of their sources as a large fertilizer plant and a cattle farm, and our mixing ratio measurements yielded up- per bounds of 350 ± 50 and 0.6 kg NH3 h−1 for their respec- tive momentary source rates. The fast response of the CIMS also allowed us to derive vertical NH3 ﬂuxes within the tur- bulent boundary layer via eddy covariance, for which we chieﬂy used the continuous wavelet transform technique. As expected for a region dominated by agriculture, we observed predominantly upward ﬂuxes, implying net NH3 emissions from the surface. The corresponding analysis focused on the most suitable ﬂight, which contained two straight-and-level legs at ∼300 m above ground. We derived NH3 ﬂuxes be- tween 1 and 11 mol km−2 h−1 for these legs, at an effec- tive spatial resolution of 1–2 km. The analysis demonstrated how ﬂux measurements beneﬁt from suitably arranged ﬂight tracks with sufﬁciently long straight-and-level legs, and it ex- plores the detrimental effect of measurement discontinuities. Following ﬂux footprint estimations, comparison to the NH3 area emissions inventory provided by the U.S. Environmen- tal Protection Agency indicated overall agreement but also the absence of some sources, for instance the identiﬁed cattle farm. Our study concludes that high-precision CIMS mea- surements are a powerful tool for in situ measurements of ambient NH3 mixing ratios, and even allow for the airborne mapping of the air–surface exchange of NH3. Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 © Author(s) 2023. This work is distributed under the Creative Commons Attribution 4.0 License. Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 © Author(s) 2023. This work is distributed under the Creative Commons Attribution 4.0 License. 1 Introduction Ammonia (NH3) is the most abundant alkaline gas in the at- mosphere, with mixing ratios ranging from < 10 parts per trillion by volume (pptv) in very remote regions to tens of parts per billion by volume (ppbv) in areas with high anthro- pogenic emissions (e.g., Norman and Leck, 2005; Shephard et al., 2020; Wu et al., 2021; Zhu et al., 2022). Consequently, it plays an important role in atmospheric and environmental chemistry as well as atmosphere–ecosystem relations: from the formation of inorganic and organic aerosol to soil acidi- ﬁcation and nutrient cycles. NH3 is a key player in the atmosphere–ecosystem ex- change and biogeochemical cycling of nitrogen (N). Agri- cultural soils are typically deﬁcient in N as a nutrient for plant growth, leading to the copious use of NH3 as fertil- izer and related N fertilizers such as urea. Volatilization of NH3, in particular from NH4-forming fertilizers, is in turn a major N loss mechanism for agricultural soils (Ma et al., 2021) while constituting a major source of atmospheric NH3. Agricultural activities are indeed the dominant source of at- mospheric NH3. Of particular importance is also livestock farming (in particular pig, cattle, poultry) and manure pro- cessing. On the other hand, ecosystem exposure to and up- take of NH3 (e.g., via dry deposition) are associated with nu- merous adverse environmental effects (e.g., via conversion to nitrous oxide, a greenhouse gas, or nitrate, which may leach into water) and biological effects, in particular on native veg- etation (Krupa, 2003). Critical NH3 thresholds (Cape et al., 2009) are exceeded across Europe (Tang et al., 2021), and NH3 contributes to critical reactive N load exceedances in North America (Walker et al., 2019). p The accurate quantiﬁcation of NH3 emissions and concen- trations is challenging, due to the wide range of ambient mix- ing ratios and its infamous propensity to interact with sam- pling and instrument surfaces, causing losses and slow re- sponse times. A wide variety of techniques have to date been used to quantify NH3 mixing ratios. Typical precisions and detection limits are tens of pptv at best, and time responses are often on the scale of minutes (von Bobrutzki et al., 2010). S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 248 Despite its importance, there are high uncertainties in at- tributing atmospheric NH3 to speciﬁc sources, and current emissions inventories for NH3 are suspected to have large uncertainties (Vonk et al., 2016; Grönroos et al., 2017; EEA, 2019). A major reason is also practical difﬁculties in es- tablishing NH3 emissions by concentration measurements, as bottom-up approaches may not capture all sources, and top-down approaches may not capture the total emissions, especially for outdoor farming activities or naturally ven- tilated buildings (e.g., Calvet et al., 2013; Oliveira et al., 2021). Also, some sources of NH3 may not be understood well enough. For example, for urban environments, catalytic converters in vehicles have been recognized as a source of NH3 that is likely grossly underrepresented in current emis- sions inventories (Sun et al., 2017; Farren et al., 2020). Cor- respondingly, observations tend to indicate that NH3 emis- sions are substantially underestimated, e.g., airborne mea- surements in Utah, where NH4NO3 plays a major role in pol- lution (Franchin et al., 2018; Moravek et al., 2019a). Glob- ally, satellite data have recently revealed hundreds of small- area (< 50 km) or point sources to be mostly underrepre- sented in emissions inventories by even more than an order of magnitude (Van Damme et al., 2018). Strong day-to-day variability was found as well (Fortems-Cheiney et al., 2016). Overall, however, observational data on NH3 concentrations are scarce, limiting evaluations of model simulations, such as models of NH3 emissions and aerosol formation. Verti- cally resolved observational data, ideally using airborne in situ measurements, are sparser still. 1 Introduction Studies over the last 10–15 years have developed continuous wavelet transform (CWT) analysis to calculate spatially resolved ﬂuxes from airborne measurements (e.g., Mauder et al., 2007; Karl et al., 2009; Metzger et al., 2013; Karl et al., 2013; Misztal et al., 2014; Yuan et al., 2015; Wolfe et al., 2015; Vaughan et al., 2017; Sayres et al., 2017; Desjardins et al., 2018; Wolfe et al., 2018; Hannun et al., 2020), including dedicated aircraft campaigns (e.g., BOREAS, CABERNET, CARAFE, OPFUE) and platforms (e.g., FOCAL). Spatial resolutions of a few kilometers are typically achieved with good accuracy. Over the last decade, developments in the application of mass spectrometry for ambient measurements, e.g., using CIMS techniques, have greatly improved our capabilities in identifying and quantifying atmospheric trace gases. State- of-the-art time-of-ﬂight (TOF) mass spectrometers typically feature a versatile atmospheric pressure interface that ef- ﬁciently transmits ions from a high-pressure (up to atmo- spheric) or low-vacuum (> 1 mbar) ion source to the high- vacuum (< 10−5 mbar) TOF region that facilitates identiﬁ- cation and detection (Junninen et al., 2010; Jokinen et al., 2012). The high-pressure ion source may be a simple cham- ber, typically held at > 100 mbar, in which reagent ions are admixed to the analyte sample to chemically ionize target compounds (ion–molecule reaction region, IMR). That setup readily allows for using a variety of different reagent ions, common examples being acetate, iodide and bromide an- ions (e.g., Bertram et al., 2011; Lee et al., 2014; Sanchez et al., 2016) and water, benzene and toluene cations (e.g., Al- jawhary et al., 2013; Kim et al., 2016; Alton and Browne, 2020). For efﬁciently ionized compounds, these TOF-CIMS devices achieve limits of detection down to < 1 pptv and 1 Hz precisions (1σ) < 10 pptv (e.g., Bertram et al., 2011). Con- sequently, they have also been used to measure surface-layer EC ﬂuxes of a variety of compounds (e.g., Nguyen et al., 2015; Schobesberger et al., 2016; Fulgham et al., 2019). A potential key advantage of TOF mass spectrometers is their high acquisition rate of full mass spectra. They can be read- ily recorded at 10 Hz or more, until sensitivity or data storage become practical limitations. By routinely counting a wide range of ions simultaneously, the mixing ratios or EC ﬂuxes of multiple compounds can in principle be quantiﬁed from the same datasets. 1 Introduction Schobesberger et al.: Airborne ﬂux measurements of ammonia tend to achieve similar or better 1 Hz precisions, in particu- lar when extra attention is paid to reducing sample–surface interactions in the sampling setup. For example, Pollack et al. (2019) reported an Allan deviation (1σ) of 60 pptv for their airborne TILDAS with optimized sampling. Airborne CIMS deployments for measuring NH3 have achieved com- parable precision, e.g., 80 pptv (1 Hz, 1σ) was reported for the acetone-CIMS deployments on the WP-3D aircraft of the National Oceanic and Atmospheric Administration (NOAA) (Nowak et al., 2012). times are design elements that reduce interactions between sample and surfaces of the sampling setup and the instru- ment, e.g., a shortened and straightened sampling line, high sampling ﬂow, and shortened reaction chamber in airborne CIMS deployments (Nowak et al., 2010) or a heated high- ﬂow sampling line with active continuous passivation in air- borne TILDAS deployments (Pollack et al., 2019). For measuring vertical ﬂuxes, including surface– atmosphere exchange rates (e.g., emission or dry deposi- tion), EC has been established as one of the most direct techniques. EC relies on measuring both the ﬂuctuations of the vertical wind component caused by the turbulence in the atmospheric boundary layer and the simultaneous ﬂuctuations of a scalar magnitude, such as temperature or a vapor’s mixing ratio. If the surface or possibly the air below the measurement height is a net source or sink for the scalar, these ﬂuctuations will correlate positively or negatively, and their covariance is a direct measure of the vertical ﬂux of the scalar at the measurement height. When the surface is the only net source or sink, that vertical ﬂux is assumed constant with height within the surface layer, i.e., up to ∼100 m, and typically decreases linearly with height (referred to as vertical ﬂux divergence) within the core boundary layer above (e.g., Lenschow et al., 1980). Under these conditions and assumptions, airborne ﬂux measurements can directly infer rates of net emission and dry deposition. Consequently, airborne EC has been applied for more than 30 years (e.g., Lenschow et al., 1980, 1981; Desjardins et al., 1982; Ritter et al., 1990, 1992, 1994; Dabberdt et al., 1993). 1 Introduction Such performance limitations can lead to substantial errors when low or fast-changing concentrations are to be captured accurately (e.g., plumes, mobile deployments, remote loca- tions, free troposphere) or for applying the eddy covariance (EC) method to measure vertical exchange (Moravek et al., 2019b), from which emission rates can be derived. Some op- tical and mass spectrometry techniques have pushed these boundaries and offer fast response while also allowing de- ployment on aircraft or for EC. Most airborne in situ mea- surements of NH3 have been during a number of aircraft campaigns, mostly in the US, that have deployed chemical ionization mass spectrometers (CIMSs; Nowak et al., 2007, 2010, 2012) or infrared laser spectrometry techniques. Ex- amples for the latter are off-axis integrated cavity output spectrometry (off-axis ICOS; Leen et al., 2013) and, pre- dominating more recently, tunable infrared laser differential absorption spectrometry (TILDAS; Moravek et al., 2019a; Pollack et al., 2019). Critical for fast instrument response ( ) In the atmosphere, NH3 contributes to aerosol particle for- mation, in particular by associating with nitric acid (HNO3) to form ammonium nitrate (NH4NO3), which can dominate the inorganic pollution load (Tang et al., 2021; Bressi et al., 2021). NH3 emissions thereby contribute substantially to ﬁne-particle pollution and may make agriculture the leading air pollution source to contribute to premature mortality in Europe and parts of North America and Asia (Lelieveld et al., 2015). Solid NH4NO3 particles have also been detected in the upper troposphere, where they may play an important role as ice nuclei (Höpfner et al., 2019). Furthermore, NH3 is implicated in the ﬁrst steps of new-particle formation, in particular by association with sulfuric acid (Schobesberger et al., 2013), and they are expected to play an important role in organic-poor environments such as the upper troposphere (Dunne et al., 2016). NH3 may limit new-particle forma- tion also in Antarctica (Jokinen et al., 2018), and it is im- plicated in intensive local cluster formation events that were observed over agricultural ﬁelds (Olin et al., 2022). Typical atmospheric aerosol also contains a major, chemically com- plex organic component, which likely also facilitates reactive uptake of NH3, forming salts or N-containing organics (e.g., Liu et al., 2015; Bell et al., 2017; Wu et al., 2021). https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 249 S. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 250 compositions containing reagents, e.g., via adduct forma- tion. We deployed the instrument on the Gulfstream I (G- 1) aircraft of the U.S. Department of Energy’s Atmospheric Radiation Measurement (ARM) Aerial Facility during the Holistic Interactions of Shallow Clouds, Aerosols, and Land- Ecosystems (HI-SCALE) ﬁeld campaign in Oklahoma in 2016 (Fast et al., 2019). This paper hence demonstrates the suitability of our benzene-CIMS setup for airborne measure- ments. Moreover, the datasets proved suitable for calculating vertical ﬂuxes via EC, making use of the 3-D wind data ob- tained by turbulence probes also carried by the G-1. To our knowledge, we are thereby reporting the ﬁrst use of TOF- CIMS for measuring EC ﬂuxes of NH3. We explore that capability of the instrument, including the use of the CWT method, as well as the capability of the airborne eddy ﬂux data to infer area emission rates of NH3 attributed to agricul- ture in rural Oklahoma. if forced; and (iv) the inlet tip was simply cut straight and faced perpendicular to the airﬂow. The aircraft cabin was not pressurized. Setup, ﬂows and pressures of the CIMS inlet system are schematically shown in Fig. 1. Ambient air was sampled at 22 L min−1 through the cabin wall via a straight, 40 cm long, 3/4 in. (1.9 cm) outer diameter (o.d.), 1.6 cm inner diame- ter (i.d.) Teﬂon tube. Most of that inlet ﬂow was provided by a dedicated sample pump (Vaccubrand MD1) and dis- carded; 2000 sccm entered the IMR through the variable ori- ﬁce. The pressure in the IMR was maintained at 100 mbar by a servo-controlled valve, throttling a dedicated scroll pump (Agilent IDP-3). The measured mass ﬂow of the pump ex- haust was used to control the variable oriﬁce to maintain the 2000 sccm sample ﬂow into the IMR. An additional ﬂow of 1500 sccm nitrogen (ultra-high-purity N2, Scott-Marrin or Airgas, UHP grade 5.0) carried methyl iodide (CH3I, from a permeation device) and deuterated benzene (C6D6) through an ionizer (210Po, 10 mCi, NRD) into the IMR. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Nominal ben- zene mixing ratios in the IMR of above 100 parts per mil- lion by volume (ppmv) were desired to obtain high, stable sensitivities in benzene-CI mode (Lavi et al., 2018), which was achieved by diverting 10 sccm out of the 1500 sccm ion- izer ﬂow over the headspace of about a milliliter of C6D6 (Cambridge Isotope Laboratories, Benzene-D6, D, 99.5 %) in a glass test tube. The 10 : 1490 ﬂow ratio would momen- tarily be increased to 100 : 1400 during instrument start-up to speed up conditioning the lines and obtain stable reagent ion counts. A pair of manual shut-off valves isolated the benzene when not in use. A needle valve in its fully open position served as a critical oriﬁce (1 L min−1 at 1 atm) upstream of the CH3I permeation device, keeping the headspace of the benzene at ∼1.5 atm to slow down its vaporization (and in- deed prevent its boiling under otherwise 100 mbar). The ben- zene was not temperature controlled. With the G-1 cabin tem- perature ranging from 15 to 30 ◦C, the calculated C6D6 mix- ing ratios in the IMR ranged from 160 to 360 ppm, nominally, i.e., assuming the headspace was saturated (thermodynamics data from Zhao et al., 2008). The benzene consumption rate appeared to suggest somewhat lower actual mixing ratios, but it was not systematically monitored. IMR background sig- nals (“zeros”) were determined every 42 s by overﬂowing the variable oriﬁce with 2200 sccm of N2 for 6 s (Fig. 1). While the IMR was actively humidiﬁed in iodide-CI mode to reduce variations of sensitivities as a function of ambient humidity (Lee et al., 2014, 2018), that humidiﬁcation was turned off for operations in benzene-CI mode, starting from research ﬂight 5 (RF5). The humidiﬁcation caused spurious spikes in signals of interest (e.g., C10H+ 16) that we could not otherwise dispose of in the ﬁeld. Typical reagent ion count rates were 2–3 × 106 cps of C6D+ 6 and 2–4 × 105 cps of (C6D6)+ 2 . Also observed dur- i i ﬂi ht ti CH I C D+ H O H O+ 1 Introduction Optical devices, on the other hand, may be limited to quantifying single compounds only or a hand- ful at best. Several chemical ionization schemes have been used for detecting NH3 in the past, including acetone-CIMS (Nowak et al., 2012), ethanol-CIMS (You et al., 2014) and water-cluster-CIMS (Zheng et al., 2015; Pfeifer et al., 2020). A key challenge for successful EC ﬂux measurements is the requirements for fast time response and high precision, in order to capture the full range of turbulence timescales, similar to the desired performance of airborne measurements more generally. A sampling rate of 10 Hz is typically de- sired for surface-layer eddies. One way of achieving fast instrument response is the use of open-path sensors, which practically eliminate interactions between sample and instru- ment surfaces altogether. For example, an open-path sensor that measures NH3 via absorption of a quantum cascade in- frared laser offers a precision of 150 pptv at > 1 Hz (1σ) and has been successfully deployed to measure EC ﬂuxes (Miller et al., 2014; Sun et al., 2015), also in low-NH3 en- vironments (Pan et al., 2021). Airborne deployments have so far favored closed-path systems, as introduced above. They In this paper, we present a new TOF-CIMS method for quantifying ambient NH3 mixing ratios, using as reagents deuterated benzene cations (C6D+ 6 ) and their clusters (e.g., dimer cations (C6D6)+ 2 ). The use of benzene-CIMS was mo- tivated by its capability of detecting isoprene and terpenes, important biogenic volatile organic compounds (Leibrock and Huey, 2000; Kim et al., 2016; Lavi et al., 2018), and deuterated benzene was chosen to aid in differentiating ion https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia g Figure 1. Schematic of the sampling setup and the ﬂows into the IMR of the CIMS. The ﬂows into the IMR are 1500 sccm of ultra- high-purity (UHP) N2, set by mass ﬂow controllers, and 2000 sccm of sample, maintained by the variable oriﬁce and controlled by mass ﬂow controllers (MFCs) and mass ﬂow meter (MFM) measure- ments (details in text). IMR pressure is maintained at 100 mbar by the auto-valve in the pumping line. A fraction of the UHP N2 passes over the headspace of a reservoir of liquid C6D6 under ∼1500 mbar due to a critical oriﬁce; all N2 then passes over a CH3I permeation device. Just prior to entering the IMR, a 210Po ionizer provides the primary reagent ions: I−, C6D+ 6 and (C6D6)+ 2 . or 120 km north of Oklahoma City, at an elevation of 314 m above mean sea level (a.m.s.l.). The ﬂight proﬁles consisted mostly of patterns of vertically stacked straight-and-level legs as well as ascending or descending straight transects or spirals. Vertically, the ﬂights focused on the region around shallow clouds, from the middle of the boundary layer to the lower free troposphere, as the campaign goal was to study surface–aerosol–cloud interactions. Ground speeds were typ- ically between 80 and 110 m s−1. Occasionally, straight-and- level legs were ﬂown as low as ∼300 m above ground. Those legs were our focus for eddy covariance analysis, especially when ﬂown broadly perpendicular to the wind. Figure S1 in the Supplement provides a geographical overview of the campaign environment and of the ﬂight portions when oper- ating in benzene-CI mode. The CIMS was installed on the port side, in the front sec- tion of the main cabin, and sampled straight through the port cabin wall. The CIMS was turned on several hours before each research ﬂight to allow for conditions to stabilize before departure and to perform calibrations. Power typically con- tinued to be available for several hours after landing, allow- ing for some more checks and maintenance, but the CIMS needed to be shut down at the conclusion of each workday, as the aircraft remained unattended and unpowered overnight. During research ﬂights, the CIMS recorded full mass spectra at a frequency of 2 Hz. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia A data acquisition rate of 2 Hz (rather than, for example, 10 Hz) was a compromise that reduced re- quirements for data storage and computing times during data processing, and potentially induced errors in obtained ﬂuxes were deemed acceptable (see Sect. 3.8 for details). Figure 1. Schematic of the sampling setup and the ﬂows into the IMR of the CIMS. The ﬂows into the IMR are 1500 sccm of ultra- high-purity (UHP) N2, set by mass ﬂow controllers, and 2000 sccm of sample, maintained by the variable oriﬁce and controlled by mass ﬂow controllers (MFCs) and mass ﬂow meter (MFM) measure- ments (details in text). IMR pressure is maintained at 100 mbar by the auto-valve in the pumping line. A fraction of the UHP N2 passes over the headspace of a reservoir of liquid C6D6 under ∼1500 mbar due to a critical oriﬁce; all N2 then passes over a CH3I permeation device. Just prior to entering the IMR, a 210Po ionizer provides the primary reagent ions: I−, C6D+ 6 and (C6D6)+ 2 . normalized count rate of NH3 · C6D+ 6 adduct ions. Normal- ization was to 106 cps of C6D+ 6 ; that is, measured counts per second (cps) of NH3 · C6D+ 6 would typically be divided by a factor of 2–3 to obtain normalized counts per second (ncps). Only the C6D+ 6 signal was used in normalization, as NH3 · C6D+ 6 responded more directly to changes in C6D+ 6 rather than (C6D6)+ 2 or the sum of both. For the eddy covariance analysis in this study, airborne meteorological data were provided by the Aircraft-Integrated Meteorological Measurement System (AIMMS-20, Aven- tech Research Inc., Canada) that was mounted on the side of the nose of the aircraft. The AIMMS-20 measured tem- perature, relative humidity (RH) and static pressure at nomi- nally 20 Hz, and it calculated 20 Hz 3-D ambient wind based on measurements of the differential pressures from a ﬁve- port hemispheric gust probe, aircraft position, velocity and attitude (Beswick et al., 2008). For ambient temperature and RH, however, we used static air and dew-point temperature data obtained by the Rosemount 102E Pt100 sensor and Gen- eral Eastern 1011-B chilled-mirror hygrometer, respectively. These data were nominally only taken at 1 Hz but exhibited a better real time response than the corresponding AIMMS-20 data. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 2 NH+ 4 was detected as well but with ∼2 orders of mag- nitude lower counts. As its signal clearly covaried with the protonated water cluster signals, NH+ 4 was likely formed by proton transfer from water and of no further interest here. 2.1 Measurement technique We added the benzene-CI capability to an existing CIMS setup designed to use iodide anions as reagents. Iodide- CI remained the instrument’s primary mode of operation. By switching instrument voltage polarities, iodide-CI and benzene-CI could be used alternatingly. At its core, the CIMS consists of a high-resolution TOF mass analyzer and a sequentially pumped atmospheric-pressure interface (APi) to guide and focus ions from a pressure-controlled ion– molecule reaction region (IMR) towards the high vacuum in the TOF region (Aerodyne Research Inc., USA, and Tofwerk AG, Switzerland). Detailed descriptions thereof are found in existing literature (e.g., Junninen et al., 2010; Lee et al., 2014). The speciﬁc CIMS at hand had been conﬁgured to ﬁt into an aircraft rack for deployment on the National Sci- ence Foundation’s C-130 aircraft and featured some modiﬁ- cations to allow for efﬁcient sampling and quantiﬁable mea- surements during airborne deployments. These modiﬁcations are described in detail in Lee et al. (2018). They include a computer-controlled variable oriﬁce to maintain a constant sample mass ﬂow rate into the IMR and a port at the oriﬁce to inject clean gas for determining instrument backgrounds. A high-ﬂow inlet with subsampling from the centerline into a shortened IMR reduced vapor–wall interactions. For de- ployment on the G-1, the conﬁguration differed slightly: (i) a machined PTFE cup was press ﬁt into the IMR to further reduce vapor–wall interactions; (ii) the CIMS was isolated from vibrations and (some) shocks by mounting it on wire- rope isolators inside its rack, whereas most accessories (most electronics, pumps, ﬂow controllers, etc.) as well as the rack itself were not; (iii) the inlet tube was fastened to the air- craft fuselage but connected to the instrument via an Ultra- Torr ﬁtting that allowed for motion relative to the instrument Typical reagent ion count rates were 2–3 × 106 cps of C6D+ 6 and 2–4 × 105 cps of (C6D6)+ 2 . Also observed dur- ing in-ﬂight operation were CH3I · C6D+ 6 , H2O · H3O+, (H2O)2 · H3O+ and (H2O)3 · H3O+, typically below 1, 2, 3 and 0.5 × 104 cps, respectively. NH3 was quantiﬁed from the Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 251 3.1 Calibration In the next set of lab experiments, the humidity dependence was more systematically explored by overﬂowing the inlet with 2.2 slpm of humidiﬁed N2 passing over the permeation source. Measured RH values of < 1 %, 15.5 % and 52.5 % (Vaisala HM34) were obtained by optionally bubbling part of the N2 through water. Later lab experiments using another CIMS device (3) yielded a similar humidity dependence (Fig. 2, blue). Fig- ure 2 shows measured sensitivities against absolute humidity in the IMR, along with empirical exponential ﬁts. Plotting against sample humidity (absolute or relative) would reduce the gap between the curves, as the earlier experiments (or- ange) used a 50 % larger ionizer ﬂow, implying a smaller absolute humidity in the IMR compared to the later exper- iments (blue) for any given sample humidity. As the CIMS devices (blue vs. orange) differed in particular in their IMR geometries, we leave it up to future studies to ascertain if sensitivity to NH3 relates to the humidity of only the sample (absolute or relative) or the humidity in the IMR. 3. Another set of experiments was conducted during spring 2022, to further explore the dependence of sensi- tivity to NH3 on humidity, as well as the possible inﬂu- ences of carrier gas (N2 vs. air) and ionizer ﬂow com- position: C6D6 vs. C6H6 (anhydrous, 99.8 %, Sigma- Aldrich) and CH3I vs. no CH3I. For those experiments, a different CIMS was used. That instrument was largely identical to the CIMS used in 2016. The most important difference was the use of the commercially available stainless-steel IMR (Bertram et al., 2011; Aerodyne Re- search) with a simple 10 mm o.d. stainless-steel inlet port. Inlet and ionizer ﬂows were 2 and 1 slpm, respec- tively; there was no make-up ﬂow. A newly purchased NH3 permeation source was press ﬁt into a hole on the side of a short 10 mm polyoxymethylene tube, which was attached to the CIMS inlet. The source was not heated but subject to the temperature in the lab, which was air-conditioned to 23 ◦C. Its gravimetric permeation rate was 11 ± 3 ng min−1. Various NH3 mixing ratios were delivered to the instrument by overﬂowing with In the later lab experiments (3), we also examined the sen- sitivity to NH3 when using regular, non-deuterated benzene (C6H6) instead of C6D6, when the CH3I permeation source was removed, and both. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia N2 or puriﬁed air (optionally humidiﬁed) and venting either before or after the source. 3.1 Calibration All variations led to slightly lower sensitivities but not signiﬁcantly so, in particular given possi- ble variability in the NH3 permeation rates due to the source not being temperature-controlled. The Supplement contains further details on our calibration experiments, including Fig. S2. 2.3 Calibration setups 1. Initial experiments for determining the CIMS sensitiv- ity to NH3 (along with isoprene, α-pinene and dimethyl sulﬁde) were carried out in the ARM Aerial Facility’s hangar in Pasco, WA, during summer 2016. The in- strument was in the same place and conﬁguration (i.e., in the aircraft) as during the ﬁeld campaign. A wafer- type permeation device was used as the source of NH3 (type 40F3, VICI Metronics Inc., Poulsbo, WA, USA; 41 ± 10 ng min−1 at 30 ◦C). We continuously kept the source at 40 ◦C and gravimetrically measured a perme- ation rate of 46 ± 4 ng min−1. For the experiments in Pasco, a ﬂow of 1 slpm N2 over the NH3 source was di- luted by a larger ﬂow of N2 (up to 10 slpm), and the total was directed at the CIMS inlet outside the aircraft. The sample pump was turned off, which reduced the total in- let ﬂow to 2 slpm, and the inlet was thus overblown by dry N2 containing NH3 as adjusted by the dilution ratio. 1. Initial experiments for determining the CIMS sensitiv- ity to NH3 (along with isoprene, α-pinene and dimethyl sulﬁde) were carried out in the ARM Aerial Facility’s hangar in Pasco, WA, during summer 2016. The in- strument was in the same place and conﬁguration (i.e., in the aircraft) as during the ﬁeld campaign. A wafer- type permeation device was used as the source of NH3 (type 40F3, VICI Metronics Inc., Poulsbo, WA, USA; 41 ± 10 ng min−1 at 30 ◦C). We continuously kept the source at 40 ◦C and gravimetrically measured a perme- ation rate of 46 ± 4 ng min−1. For the experiments in Pasco, a ﬂow of 1 slpm N2 over the NH3 source was di- luted by a larger ﬂow of N2 (up to 10 slpm), and the total was directed at the CIMS inlet outside the aircraft. The sample pump was turned off, which reduced the total in- let ﬂow to 2 slpm, and the inlet was thus overblown by dry N2 containing NH3 as adjusted by the dilution ratio. 2.2 Field campaign The results reported in this paper are taken from measure- ments during the CIMS’ deployment on the ARM Aerial Facility’s G-1 aircraft for the HI-SCALE ﬁeld campaign in 2016. An overview of the HI-SCALE campaign is provided in Fast et al. (2019), including measurement approach and descriptions of the instrumentation deployed on the G-1 be- sides the CIMS. We focus here speciﬁcally on the ﬁrst of the two intensive operating periods (IOP1), during which re- search ﬂights were performed between 24 April and 21 May. The aircraft was based out of Bartlesville Municipal Air- port (airport code KBVO), 60 km north of Tulsa, Okla- homa, USA. The research ﬂights, however, were concen- trated around the ARM Southern Great Plains Central Facil- ity ground site (SGP), located 130 km to the east of KBVO All times in this study are in coordinated universal time (UTC), which was 5 h ahead of the locally observed central daylight time (CDT). Terrain elevation data were retrieved from a digital elevation model (Yamazaki et al., 2017). Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 252 3.1 Calibration The initial calibration experiments, using dry N2 (setup 1), resulted in a measured sensitivity of only ∼0.5 ncps pptv−1, which was at least an order of magnitude lower than roughly expected from the campaign data. However, that result was consistent with follow-up calibrations in dry N2 or dry air (2 + 3). For the follow-up experiments using lab air for dilution, the sampled RH varied between around 20 % and 36 % be- tween individual measurements. The results were consistent with a sensitivity of 4.2 ncps pptv−1, as obtained by a line ﬁt (Fig. S2a); that is, no humidity dependence was apparent. The experiments using humidiﬁed N2 (RH > 15 %) yielded similar sensitivities but again with a much lower sensitivity (0.54 ncps pptv−1) for dry N2. The resulting picture was that of a relatively low humidity dependence for RH > ∼20 % but a sharp drop in sensitivity under dry conditions (Fig. 2, orange). 2. Follow-up experiments took place in the lab (September–October 2016) and used lab air and dry or humidiﬁed N2. First, a N2 ﬂow over the per- meation source of between 0.2 and 1 slpm was simply directed at the IMR oriﬁce via a short 19 mm o.d. Teﬂon line, thereby effectively diluting with lab air (∼40 % RH) at ratios of 1.8 : 0.2 to 1 : 1, which varied NH3 mixing ratios while resulting in an RH range from 20 % to 36 % for that experiment. In the next set of lab experiments, the humidity dependence was more systematically explored by overﬂowing the inlet with 2.2 slpm of humidiﬁed N2 passing over the permeation source. Measured RH values of < 1 %, 15.5 % and 52.5 % (Vaisala HM34) were obtained by optionally bubbling part of the N2 through water. 2. Follow-up experiments took place in the lab (September–October 2016) and used lab air and dry or humidiﬁed N2. First, a N2 ﬂow over the per- meation source of between 0.2 and 1 slpm was simply directed at the IMR oriﬁce via a short 19 mm o.d. Teﬂon line, thereby effectively diluting with lab air (∼40 % RH) at ratios of 1.8 : 0.2 to 1 : 1, which varied NH3 mixing ratios while resulting in an RH range from 20 % to 36 % for that experiment. 3.2 Response times and precision Unlike most ﬂights, the IMR was continuously humidiﬁed via a separate line during RF3 (cyan); the dashed line indicates the underestima- tion in case the IMR humidiﬁcation was entirely ineffective in main- taining sensitivity to NH3. Figure 2. Humidity dependence of the sensitivity of NH3 · C6D+ 6 to the NH3 mixing ratio, shown as a function of water partial pressure in the IMR as calculated from the humidity in the sample ﬂow. Ion counts were normalized to the total reagent ion counts ([C6D+ 6 ]). Orange markers are from calibration experiments at the University of Washington (UW) lab immediately after the HI-SCALE cam- paign. Bars for the lab air result indicate a sample RH range from 16 % to 40 % (i.e., including 4 % uncertainty); sample RH values for results in N2 (stars) were 0 %, 16 % and 52 %. The right-hand scale also reports the corresponding sensitivities obtained from a temperature-controlled NH3 permeation source. Blue markers are from experiments using a largely corresponding setup but with a different CIMS device at the University of Eastern Finland (UEF). Corresponding sample RH values are from 0 % to 60 % in steps of 10 %, i.e., divergent from the UW results due to a smaller (dry) ionizer ﬂow into the IMR. Lines are weighted ﬁts of the form y = 1 −ae−bx. Figure 3. Responses to instrument zeros by overﬂowing dry N2. Blue: a zero at the sampling inlet tip performed on the ground (pre- ﬂight, RF16, 19 May, ambient [NH3] = 2.9 ppb); crosses are nor- malized count rates; the line (panel a only) presents a ﬁt of these data by the sum of two exponential decays with time constants of 0.24 and 6.4 s; the shaded area is for multiplying the ﬁt by up to a factor of 7.7 (except for unity as the upper bound) corresponding to the loss of sensitivity observed in dry N2 vs. ambient air (Figs. S2c and 2). Panel (b) compares these results to examples for in-ﬂight ze- ros at the IMR oriﬁce (magenta and cyan), randomly chosen among momentarily stable [NH3] signals during RF6 (17:24 UTC, 3 May, ambient [NH3] = 3.9 ppb) and RF3 (21:44 UTC, 28 April, ambient [NH3] = 2.5 ppb). 3.2 Response times and precision The right-hand scale also reports the corresponding sensitivities obtained from a temperature-controlled NH3 permeation source. Blue markers are from experiments using a largely corresponding setup but with a different CIMS device at the University of Eastern Finland (UEF). Corresponding sample RH values are from 0 % to 60 % in steps of 10 %, i.e., divergent from the UW results due to a smaller (dry) ionizer ﬂow into the IMR. Lines are weighted ﬁts of the form y = 1 −ae−bx. responding to an ambient NH3 mixing ratio of 2.9 ppbv, d d idl b 90 % ithi 1 t i di t 1/ Figure 3. Responses to instrument zeros by overﬂowing dry N2. Blue: a zero at the sampling inlet tip performed on the ground (pre- ﬂight, RF16, 19 May, ambient [NH3] = 2.9 ppb); crosses are nor- malized count rates; the line (panel a only) presents a ﬁt of these data by the sum of two exponential decays with time constants of 0.24 and 6.4 s; the shaded area is for multiplying the ﬁt by up to a factor of 7.7 (except for unity as the upper bound) corresponding to the loss of sensitivity observed in dry N2 vs. ambient air (Figs. S2c and 2) Panel (b) compares these results to examples for in ﬂight ze Figure 3. Responses to instrument zeros by overﬂowing dry N2. Blue: a zero at the sampling inlet tip performed on the ground (pre- ﬂight, RF16, 19 May, ambient [NH3] = 2.9 ppb); crosses are nor- malized count rates; the line (panel a only) presents a ﬁt of these data by the sum of two exponential decays with time constants of 0.24 and 6.4 s; the shaded area is for multiplying the ﬁt by up to a factor of 7.7 (except for unity as the upper bound) corresponding to the loss of sensitivity observed in dry N2 vs. ambient air (Figs. S2c and 2). Panel (b) compares these results to examples for in-ﬂight ze- ros at the IMR oriﬁce (magenta and cyan), randomly chosen among momentarily stable [NH3] signals during RF6 (17:24 UTC, 3 May, ambient [NH3] = 3.9 ppb) and RF3 (21:44 UTC, 28 April, ambient [NH3] = 2.5 ppb). Crosses are again normalized count rates; lines and shadings indicate underestimations of zero signals due to dry- ing by the N2 overﬂowing the IMR oriﬁce following Fig. 2. 3.2 Response times and precision We investigated the CIMS response to changes in NH3 mix- ing ratios by analyzing “zero” measurements, i.e., transitions from ambient (NH3 levels > 1 ppbv) to instrumental back- grounds. Figure 3 presents three such occasions at 1 s reso- lution. The dark blue crosses are normalized count rates of NH3 · C6D+ 6 before and during overﬂowing the tip of the in- let line with > 26 L min−1 of dry N2. The signal, ﬁrst cor- https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 253 Figure 2. Humidity dependence of the sensitivity of NH3 · C6D+ 6 to the NH3 mixing ratio, shown as a function of water partial pressure in the IMR as calculated from the humidity in the sample ﬂow. Ion counts were normalized to the total reagent ion counts ([C6D+ 6 ]). Orange markers are from calibration experiments at the University of Washington (UW) lab immediately after the HI-SCALE cam- paign. Bars for the lab air result indicate a sample RH range from 16 % to 40 % (i.e., including 4 % uncertainty); sample RH values for results in N2 (stars) were 0 %, 16 % and 52 %. The right-hand scale also reports the corresponding sensitivities obtained from a temperature-controlled NH3 permeation source. Blue markers are from experiments using a largely corresponding setup but with a different CIMS device at the University of Eastern Finland (UEF). Corresponding sample RH values are from 0 % to 60 % in steps of 10 %, i.e., divergent from the UW results due to a smaller (dry) ionizer ﬂow into the IMR. Lines are weighted ﬁts of the form y = 1 −ae−bx. Figure 2. Humidity dependence of the sensitivity of NH3 · C6D+ 6 to the NH3 mixing ratio, shown as a function of water partial pressure in the IMR as calculated from the humidity in the sample ﬂow. Ion counts were normalized to the total reagent ion counts ([C6D+ 6 ]). Orange markers are from calibration experiments at the University of Washington (UW) lab immediately after the HI-SCALE cam- paign. Bars for the lab air result indicate a sample RH range from 16 % to 40 % (i.e., including 4 % uncertainty); sample RH values for results in N2 (stars) were 0 %, 16 % and 52 %. https://doi.org/10.5194/amt-16-247-2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 254 by factors of 9 and 16. That observation is consistent with active IMR humidiﬁcation aiding in maintaining sensitivity to NH3, plus some contaminant NH3. by about 2 orders of magnitude, though about 1 order was attributed to momentarily decreased sensitivities. Between 17:00 and 17:15 UTC, the aircraft made an excursion into the free troposphere, and NH3 mixing ratios dropped down to ∼1 ppbv. After 17:30 UTC, the aircraft twice crossed a plume, during which we observed NH3 mixing ratios up to > 30 ppbv. The background counts also responded to such transitions in ambient NH3 levels. But even in the sharp and drastic plumes, that response was subdued, as expected (cf. Fig. 3), and the immediate instrument response times were thus hardly affected. We examined the instrument’s precision based on the pure background signal during four of the longer zero measure- ments by overﬂowing the inlet tip with dry N2 on the ground (e.g., Fig. 3, dark blue). We obtained 1 Hz precisions of 5 to 11 ncps (1σ) or in terms of mixing ratios 10 to 20 pptv. A cus- tomary deﬁnition of the limit of detection (LOD) is 3 times the 1σ precision, yielding a 1 Hz LOD in the range of 30 to 60 pptv. This is on par with or better than high-performance NH3 detectors previously deployed on aircraft (see Introduc- tion section). Overall, we could conclude that the initially fast instru- ment response, on the order of a second, makes the instru- ment very well-suited for airborne in situ measurements of NH3 mixing ratios. The remaining instrument background responded more slowly. Throughout all ﬂights, the corrected backgrounds remained low enough so that their subtraction did not incur signiﬁcant uncertainties, but they might be- come an issue during sufﬁciently drastic transitions to very high or in particular to very low NH3 mixing ratios. Cor- respondingly, mixing ratios in plumes may be overestimat- ing, as we did not correct for slow background response (see next session for an estimate). Conversely, some of the mix- ing ratios we report here for the lower free troposphere may be underestimating when measured during a climb out of the boundary layer. However, [NH3] only dropped by a fac- tor of more than 5 on one such occasion, speciﬁcally dur- ing RF8 (7 May), with a drop of about an order of magni- tude (Fig. S3). 3.3 Quantiﬁcation The ﬁrst step for obtaining NH3 mixing ratios from the CIMS measurements was normalization of NH3 · C6D+ 6 count rates to C6D+ 6 primary ion count rates. The high-frequency (1 Hz) stability of C6D+ 6 during ﬂights was < 0.3 % and therefore had practically no effect on the signal from ambient NH3. Occasional slower drifts did occur, typically changes in pri- mary ion signal of up to 13 % over 1–3 min. We hypothesize that those drifts were related to temperature changes of the benzene reservoir, as they sometimes appeared to coincide with ∼2 K drifts in cabin air temperature. There was, unfor- tunately, no temperature measurement at the benzene reser- voir. We would generally recommend controlling the ben- zene reservoir’s temperature, although we did not do so. An additional observation was that the primary ion signal often started off about 15 % to 18 % low upon switching from neg- ative to positive polarity. It would take ∼10 min to reach a stable value of 2–3 × 106 cps. We speculate that behavior was due to the re-stabilization of ion guidance elements in the atmospheric-pressure interface of the mass spectrometer. The ﬂight chosen for Fig. 4, RF6, featured both the high- est NH3 mixing ratios observed during HI-SCALE’s IOP1 and the highest NH3 levels on average. An overview of all the mixing ratios obtained during 11 research ﬂights in May 2016 is given in the Supplement, with results presented in Fig. S3. In short, the observed mixing ratios spanned more than 2 orders of magnitude, from 100 pptv to tens of ppbv. Mixing ratios < 1 ppbv were measured either during cloudy days and above cloud base or clearly in the free troposphere, whereas NH3 mixing ratios > 1 ppbv were observed on over- all sunny days, clear of cloud and with good conﬁdence also within the boundary layer, within which NH3 appeared verti- cally well mixed. Climbing into the (lower) free troposphere on these ﬂights, mixing ratios dropped by factors of 3 to 30. Normalized count rates of NH3 · C6D+ 6 were then divided by calculated sensitivity values. We estimated a maximum (high-RH) sensitivity of 4.4 ncps pptv−1 and multiplied it by the relative sensitivity as a function of IMR water pressures (orange in Fig. 2) that were calculated from temperature and humidity measurements throughout each ﬂight. 3.3 Quantiﬁcation Instrument background signals were determined by linearly interpolat- ing between the short and frequent zero measurements and subtracted from the signals observed during ambient sam- pling, thus obtaining the ambient NH3 mixing ratios. We could not rigorously establish the uncertainty in used sen- sitivities (see also Sect. 3.8), but a conservative estimate of ±1 ncps pptv−1 would result in a systematic error for the re- ported mixing ratios of typically 20 % to 30 %. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia If we pessimistically assumed the (humidity- corrected) background signal was 1 order of magnitude too high for the free tropospheric measurements in this case, the reported mixing ratios of ∼200 pptv would be about 25 % too low in this worst-case scenario. 3.2 Response times and precision Crosses are again normalized count rates; lines and shadings indicate underestimations of zero signals due to dry- ing by the N2 overﬂowing the IMR oriﬁce following Fig. 2. Unlike most ﬂights, the IMR was continuously humidiﬁed via a separate line during RF3 (cyan); the dashed line indicates the underestima- tion in case the IMR humidiﬁcation was entirely ineffective in main- taining sensitivity to NH3. responding to an ambient NH3 mixing ratio of 2.9 ppbv, dropped rapidly by 90 % within 1 s, at an immediate 1/e re- sponse rate of 0.25 s. Part of that drop, however, needed to be due to the decreased sensitivity with a dry IMR (cf. Fig. 2). The blue shading indicates the theoretical “worst case” of an immediate sensitivity drop. It implies a somewhat higher than apparent background and a fast drop possibly by only 80 %, though the worst-case immediate 1/e response rate would still be < 0.4 s. Magenta crosses (Fig. 3b) show the re- sponse to an in-ﬂight zero at the IMR oriﬁce, which dropped yet more rapidly and steeply, probably because it was sub- ject to mixing and equilibration processes only in the IMR but not in the 40 cm inlet line. The steeper drop thus suggests that the sampling inlet played a role in the instrument’s re- sponse besides the IMR, at least for timescales longer than 1 s. Most likely, NH3 was partitioning back from the walls of the sampling line, creating an elevated background. The level of that background would then be related to previously sam- pled NH3 mixing ratios. That background level is responding more slowly. For three long zeros at the sampling inlet tip, that slower decay followed a time constant of 4 ± 2 min. Figure 3b also shows an in-ﬂight zero from an earlier ﬂight during which the IMR was actively humidiﬁed (cyan). In- terestingly, the response here was even better, even though the concurrent drop in sensitivity was expected to be much smaller (cf. Fig. 2), We attributed this observation to ﬂight- to-ﬂight variability, as two separate tests of turning the IMR humidiﬁcation on during inlet-tip zeros (dry N2, on the ground) did multiply the NH3 · C6D+ 6 background count rate https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia S. Schobesberger et al.: Airborne ﬂux measurements of ammonia S. Schobesberger et al.: Airborne ﬂux measurements of ammonia S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 255 Figure 4. Example time series showing instrument time response to NH3 in plumes and instrument zeros, based on a segment of RF6 on 3 May. Panel (a) shows the ﬂown altitude proﬁle in meters above mean sea level; horizontal lines mark the elevations of the SGP ground site (dark gray) and KBVO airport (light gray). Panel (b) shows normalized count rates of NH3 · C6D+ 6 (blue, in-ﬂight zeros in red). Sensitivities as per calibration results were applied (cf. Fig. 2), including consideration of their humidity dependence (pink and purple). Insets highlight the response to plumes of high NH3 levels, to measurements in the much cleaner and drier free troposphere, and the respective in-ﬂight zero measurements. Background signals were interpolated between zero measurements (red and pink lines) and subtracted from the ambient signals to obtain ambient NH3 mixing ratios (c). The sensitivity values used here range from 2.8 to 4.2 ncps pptv−1, with a general uncertainty of ±1 ncps pptv−1 estimated based on the calibration tests (indicated by shadings in panel c). Figure 4. Example time series showing instrument time response to NH3 in plumes and instrument zeros, based on a segment of RF6 on 3 May. Panel (a) shows the ﬂown altitude proﬁle in meters above mean sea level; horizontal lines mark the elevations of the SGP ground site (dark gray) and KBVO airport (light gray). Panel (b) shows normalized count rates of NH3 · C6D+ 6 (blue, in-ﬂight zeros in red). Sensitivities as per calibration results were applied (cf. Fig. 2), including consideration of their humidity dependence (pink and purple). Insets highlight the response to plumes of high NH3 levels, to measurements in the much cleaner and drier free troposphere, and the respective in-ﬂight zero measurements. Background signals were interpolated between zero measurements (red and pink lines) and subtracted from the ambient signals to obtain ambient NH3 mixing ratios (c). The sensitivity values used here range from 2.8 to 4.2 ncps pptv−1, with a general uncertainty of ±1 ncps pptv−1 estimated based on the calibration tests (indicated by shadings in panel c). In Eq. (1), h is the boundary-layer depth, and θ the angle between the wind direction and the aircraft ground track while transecting the plume. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Figure 5c shows the resulting amounts of transiting NH3, expressed in kg h−1, for the ﬁve plume transects in RF6 and in addition for an additional tran- sect of a plume from the same plant during RF12 ten days later. We discounted the ﬁrst transect, which gave a much lower advection rate than the other transects, probably due to the proximity to both the plant and the boundary-layer top, suggesting insufﬁcient mixing. It also yielded a high rel- ative uncertainty due to a very small θ. On average, tran- sects 2–5 yielded a source rate of M = 350 ± 50 kg h−1. We had neglected here the response time of the instrument back- grounds, which our experiments and observations suggested would occur on timescales longer than the plume transects (Figs. 3–4). The ensuing error would be an overestimation, though even at worst by less than 10 %. In any case, our result for M is of the order broadly expected, as the U.S. Environ- mental Protection Agency’s (EPA) 2017 National Emissions Inventory (NEI) lists the Koch site at Enid as emitting a to- tal of 1905 t NH3 yr−1, which corresponds to an average of 218 kg h−1. the locations of the plume crossings, a large fertilizer plant (Koch Industries Inc.) was identiﬁed as the NH3 source, lo- cated 35 km SW from the SGP site, near the town of Enid, OK (Fig. 5b). The US Geological Survey’s 2016 Miner- als Yearbook ranks this plant as the fourth largest domestic producer of anhydrous NH3, with a production capacity of 930 000 t yr−1. To estimate the observed source rate, we crudely estimated the total amount of horizontally advected NH3 (M) for each plume transected, based on the assumption that the plume ﬁlled out the full boundary layer in the vertical. For previous instances of that simple mass balance approach, see, for ex- ample, Turnbull et al. (2011) and references therein. The al- titude of the boundary-layer top was estimated based on the characteristic drops in humidity (1700 m around 17:30 UTC, 2000 m around 19:00 UTC, with a conservative uncertainty of ±200 m due to spatiotemporal variabilities). The CIMS measurements provided estimates for the experienced widths of the plumes (w) and plume and background concentrations of NH3 (cp and cbg). 3.4 Fertilizer plant plume transects Figure 4 illustrates the process of obtaining NH3 mixing ratios for a segment of RF6 that featured variations of mixing ratios as they were typically encountered. For most of that segment, the aircraft was ﬂying within the boundary layer, and we observed NH3 levels between 3 and 5 ppbv. As also presented in Fig. 3, dry in-ﬂight zeros dropped count rates As mentioned above (and seen in Fig. 4), RF6 repeatedly crossed a plume of substantially enhanced NH3 levels. In fact, the same plume was fully crossed ﬁve times during that ﬂight, at altitudes between 500 and 1700 m a.m.s.l. (Fig. 5a). Given the prevailing southwesterly winds at the time and Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia S. Schobesberger et al.: Airborne ﬂux measurements of ammonia S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 256 Figure 5. Source rate estimation for a large fertilizer plant (Koch Industries Inc.) in the town of Enid, OK, based on horizontal ﬂuxes calculated from plume transects during RF6 on 3 May. Panel (a) shows the altitude proﬁles ﬂown from 17:15 to 19:15 UTC; panel (b) the corresponding ground tracks in relation to Enid, the fertilizer plant and the SGP ground site (blue square). Circle sizes correspond to measured NH3 mixing ratios; colors correspond to ﬂown altitudes. Arrows in panel (b) indicate the average wind di- rections for respectively co-located plume transects 1–2 and 3–5. Panel (c) presents the mass of NH3 transiting in the total plume, es- timated as per on Eq. (1). The weighted mean using the four most reliable transects amounts to 350 ± 50 kg h−1 (gray line and shad- ings). Also shown is a plume transect (#6) during RF12 on 13 May, for which the same calculation yielded 730 ± 125 kg h−1 for detection of NH3. Therefore, we focus here on one re- search ﬂight, RF13 (14 May), which provided the most suit- able dataset for analyzing eddy covariance (EC) ﬂuxes of NH3, with three straight-and-level legs within the turbu- lently mixed boundary layer, for > 10 min each. Our anal- ysis primarily showcases the capability of the setup to de- rive airborne EC ﬂuxes, and we explore the suitability of our datasets for ensemble average (EA) and continuous wavelet transform (CWT) ﬂux calculation methods. ( ) The three example legs of RF13 occurred in the early afternoon of 14 May 2016. Surface temperatures were be- tween 12 and 16 ◦C with RH between 36 % and 39 %. Con- ditions were generally sunny, with few, occasionally scat- tered clouds with a base above 3000 m a.m.s.l. There was a marked drop in both humidity and NH3 (Fig. 6b), along with a temperature inversion, at 1450 m a.m.s.l. (19:27 UTC) and later at 1650 m a.m.s.l. (20:08 UTC), likely marking the top of the turbulently mixed boundary layer. All three exam- ple legs overﬂew the SGP site around their mid-points. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia The ﬁrst leg was ﬂown between 18:10 to 18:25 UTC, at 320 m above ground, from NE to SW; the second between 19:25 and 19:40 UTC, at 580 m above ground, from NW to SE; the third between 19:40 and 19:55 UTC, at 310 m above ground, from SE to NW. Boundary-layer mixing ratios of NH3 were mostly between 1.5 and 2 ppbv (Fig. 6b), or ∼1.2 µg m−3. The onboard aerosol mass spectrometer observed submicron aerosol loadings between 0.9 and 1.8 µg m−3, thereof 0.16 to 0.26 µg m−3 of particulate ammonium (pNH4), yielding a gas-to-particle partitioning ratio of ∼6 : 1. In the free tropo- sphere, that ratio decreased, as [NH3] dropped to 400 pptv (0.24 µg m−3) but pNH4 only to 0.1 µg m−3 (Fig. 6b and c). This observation may imply that the submicron aerosol was more acidic in the free troposphere than in the boundary layer (Pye et al., 2020). Figure 6 also indicates that particulate NO3 and pNH4 concentrations within the boundary layer tended to increase with altitude, despite broadly constant availability of NH3, which is consistent with lower temperatures favoring NH4NO3 formation. Figure 5. Source rate estimation for a large fertilizer plant (Koch Industries Inc.) in the town of Enid, OK, based on horizontal ﬂuxes calculated from plume transects during RF6 on 3 May. Panel (a) shows the altitude proﬁles ﬂown from 17:15 to 19:15 UTC; panel (b) the corresponding ground tracks in relation to Enid, the fertilizer plant and the SGP ground site (blue square). Circle sizes correspond to measured NH3 mixing ratios; colors correspond to ﬂown altitudes. Arrows in panel (b) indicate the average wind di- rections for respectively co-located plume transects 1–2 and 3–5. Panel (c) presents the mass of NH3 transiting in the total plume, es- timated as per on Eq. (1). The weighted mean using the four most reliable transects amounts to 350 ± 50 kg h−1 (gray line and shad- ings). Also shown is a plume transect (#6) during RF12 on 13 May, for which the same calculation yielded 730 ± 125 kg h−1. We also found that our measured NH3 levels agreed well with Weather Research and Forecasting model (WRF-Chem) predictions. The model was conﬁgured to cover a domain extending > 1000 km from the SGP site in every direction at a grid spacing of 12 km. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia For the ﬂight track and times of RF13, WRF-Chem predicted between 1.2 and 1.4 ppbv for the boundary layer, falling within the range of 1.1 to 2.4 ppbv we measured (Fig. 6b). Further details on the WRF-Chem model conﬁguration, including NH3 emissions, are given in the Supplement. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia The cross-wind component was then used to calculate the horizontal NH3 mass ﬂux (M) perpen- dicular to the plume cross-section: M = cp −cbg · h · w · V · \|sinθ\| (1) (1) Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 3.6 Eddy covariance analysis EA ﬂuxes were determined for time windows of lengths 1 (cyan), 2 (blue), 3 (black) or 10 min (gray), ±50 % each. Note that straight-and-level legs were required for obtaining sensible values for EA ﬂuxes, leading to erratic results during climbs (after 18:25 UTC) and descents (∼19:40 UTC) that did not pass quality checks. Quality-controlled ﬂux values obtained using the continuous wavelet transform method (CWT) are shown as red lines (details in text and Figs. 7 and 8). (b) NH3 mixing ratios as measured (orange) and as in WRF model simulation (purple) along the ﬂight track, together with ﬂight altitude (black), using the same ordinate in units of pptv and m a.m.s.l., respectively. (c) Contributions of nitrate (NO3), sulfate (SO4), ammonium (NH4) and non-refractory organic material to the mass concentration of submicron aerosol particles, as measured by the onboard aerosol mass spectrometer (AMS). our dataset. We refer to W2018 also for a more comprehen- sive discussion of the involved ﬂux methodology, which we largely followed here. generally increase the chances of passing the stationarity test. They here corresponded to a spatial resolution of 6 km along the ﬂight track. More details on this analysis are provided in the Supplement, with Fig. S4 showing lag correlations, co- spectra, and power spectra for NH3 and temperature data. The co-spectra conﬁrm that most ﬂux was indeed carried by eddies observed at periods shorter than 1 min (i.e., frequen- cies > 0.017 Hz). The power spectra exhibit the f −5/3 power dependence towards high frequencies, as theoretically pre- dicted for the inertial subrange (Kaimal and Finnigan, 1994), and suggesting that high-frequency attenuation was negligi- ble up to the Nyquist frequencies (1 Hz for the NH3 data). In the traditional EA method, individual ﬂux values are calculated for a pre-deﬁned time interval each. That ﬂux, FEA, is simply the covariance of the time series for a scalar s (here, NH3 mixing ratio) and the vertical wind speed w during that time interval. We calculated FEA for a variety of interval lengths (from 1 to 10 min), for assessing the general feasibility of the EA method for obtaining EC ﬂuxes from our dataset. The results are presented in Fig. 6a (blues and grays); lines connecting all results, markers being used only if quality checks were passed, in particular the stationarity test (Foken and Wichura, 1996; using three sub-intervals and requiring < 35 % deviation). 3.6 Eddy covariance analysis The HI-SCALE airborne campaign focused on aerosol– cloud interactions, and its ﬂight proﬁles were not designed for quantifying emission ﬂuxes. Furthermore, the CIMS mostly did not operate in the benzene-CI mode that allowed For the EC analysis, we used a custom-made MATLAB tool- box (Wolfe, 2022) that was also used for airborne EC in Wolfe et al. (2018; hereafter referred to as W2018). Only minor modiﬁcations were necessary to adapt the scripts to Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 257 Figure 6. (a) Eddy-covariance NH3 ﬂux calculation results for three legs within the mixed boundary layer during RF13 on 14 May. The ﬁrst and third leg were ﬂown at 310–320 m above ground; the second leg at 580 m and directly above the third one but in opposite direction. Thin lines without markers connect all results using the ensemble-average method (EA); circles mark those results that fulﬁlled the quality criteria (stationarity test, lag correlation, altitude stability). EA ﬂuxes were determined for time windows of lengths 1 (cyan), 2 (blue), 3 (black) or 10 min (gray), ±50 % each. Note that straight-and-level legs were required for obtaining sensible values for EA ﬂuxes, leading to erratic results during climbs (after 18:25 UTC) and descents (∼19:40 UTC) that did not pass quality checks. Quality-controlled ﬂux values obtained using the continuous wavelet transform method (CWT) are shown as red lines (details in text and Figs. 7 and 8). (b) NH3 mixing ratios as measured (orange) and as in WRF model simulation (purple) along the ﬂight track, together with ﬂight altitude (black), using the same ordinate in units of pptv and m a.m.s.l., respectively. (c) Contributions of nitrate (NO3), sulfate (SO4), ammonium (NH4) and non-refractory organic material to the mass concentration of submicron aerosol particles, as measured by the onboard aerosol mass spectrometer (AMS). Figure 6. (a) Eddy-covariance NH3 ﬂux calculation results for three legs within the mixed boundary layer during RF13 on 14 May. The ﬁrst and third leg were ﬂown at 310–320 m above ground; the second leg at 580 m and directly above the third one but in opposite direction. Thin lines without markers connect all results using the ensemble-average method (EA); circles mark those results that fulﬁlled the quality criteria (stationarity test, lag correlation, altitude stability). https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 258 Figure 7. Continuous wavelet transform (CWT) analysis of the covariance between vertical wind (w) and NH3 mixing ratio ([NH3]) for the three selected straight-and-level legs in three columns. Panels in each column share the same abscissa, representing distance along the respective leg. The regular gaps in the data are the zero measurements by the CIMS. Panel (a) shows the normalized time series, i.e., ﬂuctuations around the leg’s mean. Panel (b) presents local co-spectral powers as surface plots, following scale bias correction and normalization; reds for positive (upward) power, blues for negative (downward) power. The darkened lower parts represent the “cone of inﬂuence” (COI) that marks the locations and scales where co-spectral power may be subject to edge effects. Scales are expressed as periods (left ordinate) and lengths (right ordinate); periods correspond to lengths as per the aircraft’s ground speed (90–98 m s−1). Panel (c) shows the resulting ﬂux values, FCWT, orange for the full-scale averages, thin black for averages outside the COI only, and red for the data passing the conservatively chosen quality criteria. For comparison the ensemble-average covariance, FEA, is shown in blue. Figure 7. Continuous wavelet transform (CWT) analysis of the covariance between vertical wind (w) and NH3 mixing ratio ([NH3]) for the three selected straight-and-level legs in three columns. Panels in each column share the same abscissa, representing distance along the respective leg. The regular gaps in the data are the zero measurements by the CIMS. Panel (a) shows the normalized time series, i.e., ﬂuctuations around the leg’s mean. Panel (b) presents local co-spectral powers as surface plots, following scale bias correction and normalization; reds for positive (upward) power, blues for negative (downward) power. The darkened lower parts represent the “cone of inﬂuence” (COI) that marks the locations and scales where co-spectral power may be subject to edge effects. Scales are expressed as periods (left ordinate) and lengths (right ordinate); periods correspond to lengths as per the aircraft’s ground speed (90–98 m s−1). Panel (c) shows the resulting ﬂux values, FCWT, orange for the full-scale averages, thin black for averages outside the COI only, and red for the data passing the conservatively chosen quality criteria. For comparison the ensemble-average covariance, FEA, is shown in blue. covariance applications (Schaller et al., 2017). We applied lag time as obtained by the EA ﬂux calculations (typically < 1 s; zero lag was used for RF13; see Fig. S4). S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Figure 7 presents the ﬂuctuations of the w and s = [NH3] time se- ries, local co-spectra and resulting FCWT time series for the three selected legs. The regular gaps in the data (e.g., white stripes in Fig. 7b) are due to the frequent background deter- minations in the CIMS measurement routine. We dealt with these data gaps by ﬁlling them with covariance-based pro- jected values as suggested in W2018 (“covariance ﬁlling”). In agreement with their work, this method led to apparently smaller artifacts in the vicinity of the gaps than other gap- ﬁlling methods. To be conservative, we anyway discarded results for within the gaps as well as half a gap width on either side, which appeared sufﬁcient even if we instead used the more artifact-prone “stitching” method (which simply removes the times of the gaps, stitching the time series to- gether). The resulting gaps for the ﬂux time series were 15 s wide and occurred every 42 s. Shadings in Fig. 7b also il- lustrate the “cone of inﬂuence” (COI) for each leg, which refers to the scales and locations of the wavelet coefﬁcients (and hence co-spectra) that could be inﬂuenced by data that remained unmeasured before and after the leg. Co-spectral power within the COI may thus be subject to edge artifacts. factor) corresponds to local co-spectra for each point in time (e.g., Fig. 7b). Their scale-weighted sum over all scales yields a time series for the covariance between w and s, i.e., ﬂux (FCWT). The CWT method has important advantages over the traditional EA method, especially when it comes to ﬂux calculations using aircraft data. A major advantage is that stationarity is not required. There is hence also generally no need for detrending the input time series or dividing them into intervals. Instead, one obtains a continuous time series of ﬂuxes, along with time-resolved contributions of scales. The consequent (at least theoretically) high time resolution does not come at the expense of neglecting lower-frequency contributions, which is an inherent tradeoff when going for higher time resolutions using the EA method. For airborne measurements in particular, the higher time resolution cor- responds to a ﬁner spatial resolution, and by not relying on stationarity, heterogeneous conditions (e.g., due to heteroge- neous surface emissions) can be investigated more readily. 3.6 Eddy covariance analysis The results are internally con- sistent, as FEA using shorter time intervals (down to 1 min, corresponding to ∼6 km, cyan) broadly average to the results using longer intervals. This suggests that little ﬂux was “lost” even when limiting the covariance calculation to successions of as small as 1 min intervals. Smaller intervals would also In the CWT method, covariance is analyzed via the con- tinuous wavelet transforms of w and s, Ww and Ws, which are the convolutions of their time series with scaled and translated versions of a time-dependent “mother wavelet” function (Torrence and Compo, 1998; Mauder et al., 2007; W2018). The wavelet “cross-scalogram”, WwW ∗ s (∗denot- ing the complex conjugate), is a function of scale (frequency) and translation (time) and its real part (times a conversion https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia where the ﬁrst term on the right-hand side is storage and the second is horizontal advection, with U the horizontal wind speed in direction x, and the last term is the local net source or sink. Subsidence and horizontal turbulent terms have been neglected in Eq. (2), as they are typically at least an order of magnitude smaller (Karl et al., 2013). Our ﬂight proﬁles did not allow us to assess the terms on the right-hand side of Eq. (2). However, for scalars with no or only slow at- mospheric sources or sinks (e.g., non-reactive species), the ﬂux divergence is expected to be linear throughout most of the boundary layer (Vinuesa and Arellano, 2011). One could also expect such linearity for NH3, as its boundary-layer lifetime against oxidation is weeks to months (Diau et al., 1990), while its gas–particle partitioning can be assumed to be in equilibrium, at least outside plumes. Flux divergence may then be obtained more directly by measuring ﬂuxes at multiple altitudes, as suggested, for example, in W2018. For species with vertically inhomogeneous source or sink rates, however, ﬂux divergence may be non-linear (Wolfe et al., 2015). Indeed, the partitioning of NH3 into the particle phase is expected to be enhanced at lower temperature, which generally decreases with height within the boundary layer (Sect. 3.5, Fig. 6). In any case, we were unfortunately unable to consistently investigate ﬂux divergence for this study, as the only pair of suitable legs (vertically stacked, horizontally co-located and close in time) were the second and third legs of RF13, as shown in Fig. 6, and much of the respective ﬂux time series was riddled with gaps and subject to nearby point sources. For the remaining pairs of ﬂux values, we obtained a median NH3 ﬂux divergence of −0.02 pptv s−1, although with a standard deviation of 0.15 pptv s−1, i.e., at insufﬁcient accuracy and precision. That divergence rate is, however, of the expected order, as it corresponds to ∼1 mol km−2 h−1 per 315 m of height or to a loss of ∼15 % compared to typi- cally measured ﬂuxes at that height. To simplify the remain- der of our footprint analysis, we continued with using data obtained from 315 m a.g.l. and neglected any, presumably slightly low, bias due to ﬂux divergence. The co-spectral powers shown in Fig. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia When we calculated FCWT, we included the COI (orange line in Fig. 7c). But to be conservative again, we discarded any ﬂuxes for which the COI extended to periods < 60 s or for which the ﬂux excluding the COI (thin black line in Fig. 7c) differed by > 50 %. Altogether, almost two-thirds (65 %) of the full FCWT time series were thus discarded: ∼35 % due to the 15 s gaps and another ∼30 % due to the COI ﬁltering. can be used to calculate that net air–surface exchange. With rare exceptions (Crawford et al., 1996; Sayres et al., 2017), it is not feasible to perform research ﬂights within the shallow near-surface “constant ﬂux layer”, which extents roughly up to 10 % of the mixed layer depth. Above that layer, the ﬂux typically decreases with altitude (z), and this vertical ﬂux di- vergence needs to be considered for typical airborne EC mea- surements. In general, ﬂux divergence ∂F/∂z for a scalar s can result from several processes (W2018): g p g The obtained FCWT values compared well overall to the ﬂuxes obtained using the EA method (Fig. 6a), in particu- lar for FEA in 1 min intervals. Using such small intervals, the EA method allowed for retrieving ﬂuxes closer to the edges of each leg. However, that was achieved by excluding any larger-scale ﬂuctuations and covariance a priori, which the CWT method did not. In addition, the CWT procedure acknowledged the possible but unknown inﬂuence of larger scales towards the leg perimeters, via the COI considerations above. Further away from the leg perimeters, fewer FCWT were ﬂagged due to COI. And despite the zeroing gaps, the CWT method clearly achieved a denser coverage here than the EA method, mainly because it did not rely on stationar- ity, while NH3 mixing ratios would experience both gradual changes as well as several sharp plumes of varying intensity. In particular in the presence of strong plumes, the FEA failed their quality checks, whereas the FCWT time series responded with peaks on their own, which we will further explore be- low. Nominally, the CWT method yielded ﬂuxes at the fre- quency of the scalar measurements (2 and 20 Hz), but the “true” time resolution of the ﬂuxes is of course much lower, which was reﬂected by the wider peaks in FCWT as compared to the corresponding peaks in the [NH3] time series. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 7b include a so- called bias correction for wavelet scale (Liu et al., 2007), as also performed for calculating the EC ﬂuxes. With that, they illustrate that most of the power contributing to FCWT was at scales smaller than 1 min. This is also apparent from the leg-wide averages or analogously from the leg-wide frequency-weighted Fourier transform co-spectra (Fig. S6; quality-controlled locations only), and it is in agreement with the co-spectra obtained following the EA method (cf. Fig. S4, center row). Likewise, the wavelet power spectra of the mea- sured time series (Fig. S7) were similar to their Fourier trans- form counterparts (cf. Fig. S4, bottom row). Figure S7 also shows wavelet power spectra for w and ambient temperature (T ) measurements. Turbulence was captured well by the w measurements up to highest frequencies, whereas the Rose- mount sensor’s 1 Hz T measurements were somewhat atten- uated at frequencies > 0.3 Hz. For further discussion of the T power spectra, see the Supplement. 3.7 Vertical ﬂux divergence The motivation of EC ﬂux measurements is often the inves- tigation of atmosphere–surface interactions. If certain con- ditions are fulﬁlled, the measured EC ﬂuxes correspond to net emissions from or net deposition to the surface, or they S. Schobesberger et al.: Airborne ﬂux measurements of ammonia ∂F ∂z = −∂s ∂t −U ∂s ∂x + Q, (2) (2) where the ﬁrst term on the right-hand side is storage and the second is horizontal advection, with U the horizontal wind speed in direction x, and the last term is the local net source or sink. Subsidence and horizontal turbulent terms have been neglected in Eq. (2), as they are typically at least an order of magnitude smaller (Karl et al., 2013). Our ﬂight proﬁles did not allow us to assess the terms on the right-hand side of Eq. (2). However, for scalars with no or only slow at- mospheric sources or sinks (e.g., non-reactive species), the ﬂux divergence is expected to be linear throughout most of the boundary layer (Vinuesa and Arellano, 2011). One could also expect such linearity for NH3, as its boundary-layer lifetime against oxidation is weeks to months (Diau et al., 1990), while its gas–particle partitioning can be assumed to be in equilibrium, at least outside plumes. Flux divergence may then be obtained more directly by measuring ﬂuxes at multiple altitudes, as suggested, for example, in W2018. For species with vertically inhomogeneous source or sink rates, however, ﬂux divergence may be non-linear (Wolfe et al., 2015). Indeed, the partitioning of NH3 into the particle phase is expected to be enhanced at lower temperature, which generally decreases with height within the boundary layer (Sect. 3.5, Fig. 6). In any case, we were unfortunately unable to consistently investigate ﬂux divergence for this study, as the only pair of suitable legs (vertically stacked, horizontally co-located and close in time) were the second and third legs of RF13, as shown in Fig. 6, and much of the respective ﬂux time series was riddled with gaps and subject to nearby point sources. For the remaining pairs of ﬂux values, we obtained a median NH3 ﬂux divergence of −0.02 pptv s−1, although with a standard deviation of 0.15 pptv s−1, i.e., at insufﬁcient accuracy and precision. That divergence rate is, however, of the expected order, as it corresponds to ∼1 mol km−2 h−1 per 315 m of height or to a loss of ∼15 % compared to typi- cally measured ﬂuxes at that height. To simplify the remain- der of our footprint analysis, we continued with using data obtained from 315 m a.g.l. and neglected any, presumably slightly low, bias due to ﬂux divergence. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Due to these advantages, CWT has been applied for calcu- lating EC ﬂuxes from airborne measurements for decades (e.g., Attié and Durand, 2003; Mauder et al., 2007; Karl et al., 2009). In our CWT ﬂux analysis, we used the Morlet wavelet with a wavenumber of 6, which is the standard choice for eddy https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 259 Atmos. Meas. Tech., 16, 247–271, 2023 3.8.1 Systematic errors The largest systematic error was likely due to the limited ac- curacy of the NH3 mixing ratio measurements (SEacc,MR), caused by the uncertainty of the sensitivity values used to convert count rates to mixing ratios (Sect. 3.1 and 3.3). We could only crudely estimate that uncertainty. For the maxi- mum (dry) sensitivity of 4.4 ncps pptv−1, an uncertainty of ±0.2 ncps pptv−1 would correspond to a relative systematic uncertainty of 5 % – likely an optimistic estimate for over- all accuracy, given its humidity dependence (Fig. 2). If we changed to the steeper humidity dependence found later (blue in Fig. 2; instead of orange), NH3 mixing ratios would typi- cally increase by 23 % to 26 % – likely a pessimistic estimate for the instrument’s accuracy. SEacc,MR is of unknown sign and would usually propagate, in relative terms, directly to the derived ﬂuxes. NH3 ﬂux, however, could also be affected by small ﬂuctuations in ambient water vapor, which itself had a consistent upward ﬂux (as typical for the turbulent boundary layer due to evaporation from the surface). These ﬂuctuations were in principle accounted for, as 1 Hz humidity data were used for calculating NH3 mixing ratios, but uncertainty in the humidity dependence could potentially lead to larger errors for NH3 ﬂuxes, SEacc,F (e.g., Fig. S8). Using the same pro- cedure as above, we obtained a high estimate for SEacc,F of 10 % to 28 %, again of unknown sign. The variability of these upper estimates for SEacc,F is illustrated in Fig. S9. , Other systematic error sources for ﬂuxes are undersam- pling of turbulent ﬂuctuations at low as well as high fre- quencies. These errors correspond to low biases in the abso- lute values of the measured ﬂuxes. For airborne CWT ﬂuxes, sampling of low frequencies is primarily limited by the ﬁnite length of the ﬂight leg. An upper limit for the resulting frac- tional systematic error, SEturb, can be estimated as a func- tion of leg length, ﬂight altitude and boundary-layer depth (Lenschow et al., 1980; W2018), yielding between 2 % and 3 %. High-frequency sampling is mainly limited by the re- sponse time of the instrument. Using the worst-case response time of 0.4 s (Sect. 3.2), integration over transfer function- weighted leg-wide co-spectra (Horst, 1997; W2018) yielded systematic error fractions, SERT, of 4 % to 12 %. 3.8.1 Systematic errors Additional undersampling of high frequencies may have occurred as the CIMS data were acquired at only 2 Hz, whereas the stan- dard for EC is 10 Hz. However, the co-spectra (Figs. S4, S6) For spatially resolved FCWT values, this total random error was clearly the dominant uncertainty. It is illustrated in detail in Fig. 8 for the three legs in RF13, and it will be used as the uncertainty estimate for ﬂuxes and net surface exchange in the remainder of this chapter. For simplicity, none of the systematic ﬂux errors are henceforth considered, including those of known sign, which would amount to slight increases in absolute ﬂux values. 3.8.2 Random ﬂux error For assessing total random ﬂux error, we followed the method proposed in W2018. It uses the wavelet coefﬁcients of the measured scalar (i.e., [NH3]) and vertical wind to calculate cross- and auto-covariances across a range of lag times. Total random error (REwave) is then obtained empiri- cally via estimating the variance of the covariance over of a certain range of lags (Finkelstein and Sims, 2001). We used lags ±10 s, so the range would capture the integral timescale (∼3–7 s; Fig. S5) while keeping the contribution of trends in the time series low. Further following previous works (Mauder et al., 2013; W2018), we did not consider frequen- cies lower than fmin = 0.02 Hz (spatial scales > ∼5 km), again to limit the potential inﬂuence of trends (cf. Figs. 7, S6). For 1 Hz ﬂuxes, the resulting typical REwave values were 250 % to 490 %; averaging to 0.1 Hz ﬂuxes (spatial scales of ∼1 km) yielded “more useful” REwave values of typically 79 % to 145 % (Fig. 8; Table 1). Note that the choice of the low-frequency cutoff (fmin) had a marked inﬂuence on that result: the median REwave of 116 % for fmin = 0.02 Hz would increase to 156 % for fmin = 0.01 Hz or decrease to 73 % for fmin = 0.04 Hz or 34 % for fmin = 0.1 Hz. Total random ﬂux error is due to inherent randomness of boundary-layer turbu- lence (REturb) and uncorrelated instrumental noise (REnoise). Theoretical upper bounds to REturb were estimated leg-wise (Lenschow et al., 1994), similar to SEturb, yielding ∼19 %. Noise in the [NH3] time series was estimated from lagged auto-covariances (Langford et al., 2015), yielding REnoise of ∼17 %. The estimated total random errors were thus ∼26 % (or less, as REturb values are estimated upper bounds), which broadly agreed with the leg-averaged REwave of 12 % to 32 % (Table 1). Analogously to W2018, we thus concluded that the calculated REwave values were of the correct order. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 260 extended by some considerations speciﬁc to our case. We start with a compilation of various sources of systematic er- rors for mixing ratios and ﬂuxes, followed by a discussion of random ﬂux errors. Table 1 summarizes the results for each of the three legs in RF13. In the following, typical ranges refer to interquartile ranges. show quickly diminishing ﬂux contributions as frequency in- creases towards 1 Hz. The corresponding ogives (not shown) indicate that 92 % to 96 % of co-spectral power occurred at frequencies < 0.2 Hz, which is a ﬁfth of our Nyquist fre- quency of 1 Hz. With that, we crudely and conservatively estimate that quintupling our sampling rate to 10 Hz would add at most 4 % to 8 % of ﬂux (SESR). The ﬂux errors directly propagate into surface exchange rates, and the scaling to account for ﬂux divergence can intro- duce additional uncertainty (W2018). However, our limited datasets did not allow us to assess these quantities (Sect. 3.7). https://doi.org/10.5194/amt-16-247-2023 3.8 Flux uncertainties We investigated the uncertainties in our ﬂux calculations fol- lowing primarily procedures as outlined in detail in W2018, https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Shadings represent REwave values (±1σ) for those aver- ages. Horizontal cyan bars mark the width of blocks of continuous FCWT (∼2.5 km) and their block-wide averages; vertical cyan bars denote the REwave values for these averages. As above, h is the mixed layer depth, and U is the horizon- tal wind speed; zm is the measurement height, and w∗is the Deardorff convective velocity scale, typically in the range of 1–2 m s−1 (Stull, 1988; Karl et al., 2013). We estimated h like above (1100–1300 m a.g.l.), and calculated w∗based on our best estimate for the sensible temperature ﬂux during RF13 (0.16 K m s−1) yielding ∼1.9 m s−1. We obtained the shape of the 1-D footprint functions from the crosswind-integrated footprint predictions that resulted from the parametrization presented in Kljun et al. (2015), which itself is based on La- grangian stochastic particle dispersion simulations. The pre- dicted shapes are identical when referenced to the horizon- tal distance of the footprint distribution’s median from the measurement location, at least for the range of conditions we encountered. The footprint shape is presented with that refer- ence in Fig. 9. We could thus calculate cross-wind integrated footprint functions as a function of actual distance by scal- ing the general shape according to the half-widths dx0.5 that were calculated as per Eq. (3). Thereby, we obtained a rela- tively robust 1-D footprint for each derived ﬂux, in particular for the full FCWT time series. Errors in the obtained dimen- sions were nominally subject only to uncertainties in w∗and h, whereas uncertainties in friction velocity and lateral wind ﬂuctuations did not need to be considered. The footprint lo- cations were assumed simply upwind according to the con- currently measured wind direction. Figure 8. Total random ﬂux errors along the FCWT time series for the three case-study legs (Figs. 6, 7), estimated empirically us- ing wavelet coefﬁcients (REwave). Quality-controlled FCWT values (red in Figs. 6, 7) are shown (thin gray lines), along with moving averages to 0.1 Hz, corresponding to spatial scales of ∼1 km (thick blue lines). Shadings represent REwave values (±1σ) for those aver- ages. Horizontal cyan bars mark the width of blocks of continuous FCWT (∼2.5 km) and their block-wide averages; vertical cyan bars denote the REwave values for these averages. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 261 Table 1. Estimates for systematic errors (SEs) and random errors (REs) for the case study of ﬂight RF13. Descriptions for each error type are given in the text (Sect. 3.8). For SEs, signs indicate error direction relative to the absolute values of measured ﬂuxes; “±” means the direction could be either side. Ranges of errors, indicated by “. . . ”, correspond to leg-wide interquartile ranges; errors without ranges apply t h ﬂ i th l Table 1. Estimates for systematic errors (SEs) and random errors (REs) for the case study of ﬂight RF13. Descriptions for each error type are given in the text (Sect. 3.8). For SEs, signs indicate error direction relative to the absolute values of measured ﬂuxes; “±” means the direction could be either side. Ranges of errors, indicated by “. . . ”, correspond to leg-wide interquartile ranges; errors without ranges apply to each ﬂux in the leg. Error type Applying to Leg 1 Leg 2 Leg 3 (320 m a.g.l., 55 km) (580 m a.g.l., 69 km) (310 m a.g.l., 55 km) SEacc,MR mixing ratios ±5 % to ±22. . . 25 % ±5 % to ±22. . . 26 % ±5 % to ±24. . . 27 % SEacc,F ﬂuxes, emissions ±5 % to ±6. . . 20 % ±5 % to ±19. . . 31 % ±5 % to ±11. . . 30 % SEturb ﬂuxes, emissions + ≤2 % + ≤3 % + ≤3 % SERT ﬂuxes, emissions +8 % +12 % +4 % SESR ﬂuxes, emissions + < 8 % + < 6 % + < 4 % REwave (0.1 Hz) ﬂuxes, emissions 70 %. . . 161 % 86 %. . . 140 % 74 %. . . 147 % REturb (leg) ﬂuxes, emissions ≤18 % ≤20 % ≤19 % REnoise (leg) ﬂuxes, emissions ∼18 % ∼17 % ∼16 % (RE2 turb + RE2 noise)1/2 ﬂuxes, emissions ≤26 % ≤26 % ≤25 % REwave (leg) ﬂuxes, emissions 32 % 26 % 12 % Flux divergence emissions not determined not determined not determined Figure 8. Total random ﬂux errors along the FCWT time series for the three case-study legs (Figs. 6, 7), estimated empirically us- ing wavelet coefﬁcients (REwave). Quality-controlled FCWT values (red in Figs. 6, 7) are shown (thin gray lines), along with moving averages to 0.1 Hz, corresponding to spatial scales of ∼1 km (thick blue lines). 3.9 Flux footprint analysis The ﬂux footprint is the area of the surface that contributes to the net ﬂux observed at a certain location and height. The https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 262 Figure 9. Cross-wind integrated ﬂux footprint obtained from the footprint parametrization described in Kljun et al. (2015) as a func- tion of horizontal upwind distance from the measurement loca- tion and referenced to the footprint distribution’s median (P50%). Marked in blue is the interquartile range, IQR, spanning from the 25th to the 75th percentiles (P25% and P75%), which we used for illustrating ﬂux footprint locations and dimensions in subse- quent georeferenced ﬁgures. The half-width of the function, dx0.5, is pointed out as well. parison implies that the FCWT observed at ∼300 m a.g.l. cor- responded to the net emission (or deposition) of gaseous NH3 from (to) the surface, i.e., neglecting ﬂux divergence (see above) and potential net uptake or release of NH3 by aerosol. As apparent from Fig. 10, the overﬂown landscape is domi- nated by agricultural land; emissions projected for that land presumably dominate the NEI area emissions for NH3. Fig- ure 11a shows the NEI 2017 area emissions for the day and time of RF13 for a wider area that also encompasses the HI- SCALE research ﬂights. Emissions hotspots in that wider area relate to concentrations of intensive farming, including animal husbandry to the south (southwest of and east of Ok- lahoma City) and north (northeast of Wichita). The inven- tory’s afternoon emissions range from 5 to 12 mol km−2 h−1 for the areas of our ﬂux footprints from RF13 around the SGP site (Fig. 11b). The diel maxima (up to 13 mol km−2 h−1) are reached a bit later in the afternoon (∼21:00 UTC). Note that the NEI map contains some unexpectedly sharp transitions of NH3 emissions along lines (Fig. 11a). These lines coincide with political boundaries rather than changes in land cover or land use, speciﬁcally the boundaries of various counties, and are therefore likely the result of county-level inconsis- tencies in emissions reporting or NEI compilation. Shifts in the overall level of NEI area emissions are also apparent in Fig. 11b, and they correspond to both ﬂight tracks crossing county lines. Figure 9. Cross-wind integrated ﬂux footprint obtained from the footprint parametrization described in Kljun et al. (2015) as a func- tion of horizontal upwind distance from the measurement loca- tion and referenced to the footprint distribution’s median (P50%). y On average, the ﬂuxes measured during the ﬁrst leg of RF13 broadly corresponded to the emissions indicated by the NEI (Fig. 11b, left). Measured NH3 ﬂuxes, slightly averaged to 0.5 Hz, corresponding to spatial scales of ∼200 m, reached up to 18 mol km−2 h−1 but also down to −4 mol km−2 h−1 (i.e., net deposition). Note, however, that averaging to scales of at least 1–2.5 km was necessary to re- duce random ﬂux errors to ∼100 % or less. These more ro- bust averages ranged from 1 to 11 mol km−2 h−1. These re- sults provide only a snapshot of the NH3 emissions in the area but illustrate at least their spatiotemporal variability in the real world. And as expected from the EC analysis, they do so at a much ﬁner resolution than the NEI’s 12 km grid. An overall similar situation was observed for the other low- level leg (Fig. 11b, right), except for the pair of peaks be- tween 19:47 and 19:49 UTC with emission ﬂuxes clearly el- evated above background. Figure 12a is a zoomed-in version of Fig. 10, focusing on the respective geographical area. The ﬂux time series reached up to ∼30 and ∼50 mol km−2 h−1, respectively, for the two peaks, which coincided with the crossings of two NH3 plumes, apparent as sharp peaks in the mixing ratio time series (Figs. 7, 12a). Note that the ﬁrst of the two peaks in the ﬂux time series is missing its maxi- mum due to a close-by zero measurement. The correspond- ing gap in the NH3 mixing ratio time series was < 500 m from the plume (hereafter “plume 1”). To avoid artifacts near data gaps (see above), the resulting gaps in the ﬂux time series were widened, thereby engulﬁng the peak maximum. Noteworthy also are the widths of the respective peaks. The Figure 10. Map of the ﬂux footprint estimates for the two low-level legs of RF13, ﬂown at 310–320 m above the SGP ground site eleva- tion (location marked by blue square). Black crosses mark ﬂux mea- surement (i.e., aircraft) locations at 0.5 Hz resolution, converting to a spacing of 180–195 m. For each cross, a colored stripe marks the location of the interquartile range (IQR) of the respective ﬂux foot- print distribution (cf. Fig. 9), assuming locations directly upwind. That is, winds were northeasterly. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Figure 10 illustrates the locations and dimensions of our 1-D footprint estimates for the two low-level (∼315 m a.g.l.) legs of RF13 (the ﬁrst and the third leg in Figs. 6–7), by over- laying the interquartile range (IQR) of the footprint func- tions as stripes on a satellite map. The IQR stripes were semi-transparently colored by the magnitude of respectively derived ﬂuxes FCWT. To give an idea of scales, the means and standard deviations for dx0.5 were 1.4 ± 0.3 km. IQRs were 3.1 ± 0.6 km long, ranging out to P75 % distances of footprint is a generally two-dimensional function of location and describes how strongly sources and sinks in the area con- tribute to the ﬂux. For our study, we calculated ﬂux foot- prints based on relatively simple parametrizations. Weil and Horst (1992) proposed as a metric the half-width of the hori- zontal (one-dimensional) footprint, dx0.5: footprint is a generally two-dimensional function of location and describes how strongly sources and sinks in the area con- tribute to the ﬂux. For our study, we calculated ﬂux foot- prints based on relatively simple parametrizations. Weil and Horst (1992) proposed as a metric the half-width of the hori- zontal (one-dimensional) footprint, dx0.5: dx0.5 = 0.9Uz2/3 m h1/3 w∗ . (3) (3) https://doi.org/10.5194/amt-16-247-2023 https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Marked in blue is the interquartile range, IQR, spanning from the 25th to the 75th percentiles (P25% and P75%), which we used for illustrating ﬂux footprint locations and dimensions in subse- quent georeferenced ﬁgures. The half-width of the function, dx0.5, is pointed out as well. full width of plume 1 was only ∼200 m (half-width 100 m), whereas the peak in FCWT had a full width of 2–3 km. Like- wise, the wider plume 2 (full/half widths of ∼800/400 m) also left a FCWT peak ∼2–3 km wide. The half-widths of the FCWT peaks were more poorly deﬁned but likely ∼1–2 km. These observations suggest that our application of the CWT technique to derive EC ﬂuxes, speciﬁcally from NH3 mea- surements at 315 m a.g.l., yielded a ﬂux time series able to resolve NH3 emissions at a spatial resolution of ∼1–2 km along the ﬂight track. This heuristic ﬁnding broadly agrees with the results of our error analysis that suggested that aver- aging to scales of ∼1 km or more was typically necessary to reduce random errors to < 100 % (Sect. 3.8). Note that these errors increased markedly in the vicinity of relatively local- ized peaks or dips in the FCWT time series (Fig. 11b; plumes after 19:47 UTC, also ∼18:21–18:22 UTC), further caution- ing against relying on FCWT at too short timescales (or spatial scales). full width of plume 1 was only ∼200 m (half-width 100 m), whereas the peak in FCWT had a full width of 2–3 km. Like- wise, the wider plume 2 (full/half widths of ∼800/400 m) also left a FCWT peak ∼2–3 km wide. The half-widths of the FCWT peaks were more poorly deﬁned but likely ∼1–2 km. These observations suggest that our application of the CWT technique to derive EC ﬂuxes, speciﬁcally from NH3 mea- surements at 315 m a.g.l., yielded a ﬂux time series able to resolve NH3 emissions at a spatial resolution of ∼1–2 km along the ﬂight track. This heuristic ﬁnding broadly agrees with the results of our error analysis that suggested that aver- aging to scales of ∼1 km or more was typically necessary to reduce random errors to < 100 % (Sect. 3.8). Note that these errors increased markedly in the vicinity of relatively local- ized peaks or dips in the FCWT time series (Fig. 11b; plumes after 19:47 UTC, also ∼18:21–18:22 UTC), further caution- ing against relying on FCWT at too short timescales (or spatial scales). As for plume 1, we are conﬁdent that its source was a cattle farm that the G-1 passed about 1 km downwind (Fig. 12). Stripe widths were chosen for visibility; stripe colors correspond to the magnitude of the measured EC ﬂuxes (FCWT). Figure 10. Map of the ﬂux footprint estimates for the two low-level legs of RF13, ﬂown at 310–320 m above the SGP ground site eleva- tion (location marked by blue square). Black crosses mark ﬂux mea- surement (i.e., aircraft) locations at 0.5 Hz resolution, converting to a spacing of 180–195 m. For each cross, a colored stripe marks the location of the interquartile range (IQR) of the respective ﬂux foot- print distribution (cf. Fig. 9), assuming locations directly upwind. That is, winds were northeasterly. Stripe widths were chosen for visibility; stripe colors correspond to the magnitude of the measured EC ﬂuxes (FCWT). 4.1 ± 0.8 km. Note, however, that substantial ﬂux contribu- tions are also expected much closer, namely, 25 % of the to- tal from ranges between ∼400 and 1000 m (cf. Fig. 9). In Fig. 10, those close-in ranges are located between the close end of the stripes and the ﬂight path. The two ﬂux hotspots observed during the third leg (cf. Figs. 6–7) are also conspic- uous in Fig. 10, northwest of the SGP ground site. Next, we compared our derived FCWT to the NH3 area emissions expected as per the EPA’s National Emissions In- ventory (NEI, 2017 data, 12 km grid size). Such direct com- https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 263 Figure 11. (a) Map of the NEI for area emissions of NH3 for 14 May at 19:00 UTC (14:00 local time, LT), using the same color scale as in Fig. 10. The geographic location of Fig. 10 is highlighted in green, and an example of artifacts caused by county lines (details in text) is pointed out. Major towns are indicated in white for reference (OKC = Oklahoma City). (b) Time series of quality-controlled measurement- derived ﬂuxes (FCWT) during the low-level legs of RF13 (gray and blues) are shown compared to the emissions in the NEI at the locations of the IQR of the respective footprints (as shown in Fig. 10). NEI values for 18:00 UTC (left, 13:00 LT) and 20:00 UTC (right, 15:00 LT) are shown in red, diel mean values are shown in brown, and the diel minimum and maximum are shown in yellow. Measured ﬂuxes at 0.5 Hz (∼200 m spatial resolution), as used for Fig. 10, are shown in gray; moving averages to 0.1 Hz (∼1 km) are shown in blue, with shadings representing their random errors (REwave). Cyan crosses indicate block-wise (mostly ∼2.5 km wide) ﬂux averages and respective REwave values. The times of transecting plume 1 and plume 2 (see text and Fig. 12) are marked by arrows. Figure 11. (a) Map of the NEI for area emissions of NH3 for 14 May at 19:00 UTC (14:00 local time, LT), using the same color scale as in Fig. 10. The geographic location of Fig. 10 is highlighted in green, and an example of artifacts caused by county lines (details in text) is pointed out. Major towns are indicated in white for reference (OKC = Oklahoma City). (b) Time series of quality-controlled measurement- derived ﬂuxes (FCWT) during the low-level legs of RF13 (gray and blues) are shown compared to the emissions in the NEI at the locations of the IQR of the respective footprints (as shown in Fig. 10). NEI values for 18:00 UTC (left, 13:00 LT) and 20:00 UTC (right, 15:00 LT) are shown in red, diel mean values are shown in brown, and the diel minimum and maximum are shown in yellow. Measured ﬂuxes at 0.5 Hz (∼200 m spatial resolution), as used for Fig. 10, are shown in gray; moving averages to 0.1 Hz (∼1 km) are shown in blue, with shadings representing their random errors (REwave). Cyan crosses indicate block-wise (mostly ∼2.5 km wide) ﬂux averages and respective REwave values. The times of transecting plume 1 and plume 2 (see text and Fig. 12) are marked by arrows. Choosing a smaller wavenumber for the Morlet mother wavelet (or other wavelets) can improve the localization of the FCWT peaks slightly. Both the standard choice of the Morlet wavelet with wavenumber 6 and the Paul wavelet of order 6 generally led to the best agreement with ﬂuxes ob- tained through the ensemble-average method, and either one therefore appeared to be the best choice overall. The Paul wavelet improved localization but increased locational noise. Once averaging to > 1 km, however, these differences would largely disappear. We leave it up to future studies, for in- stance with a more copious dataset, to explore the beneﬁts of different choices for the mother wavelet in more detail. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Remaining conservative, we estimated the farm contributed 20 mol km−2 h−1 to the total observed FCWT at peak. With that, we obtained 0.17 kg h−1 for our As for plume 2, we were not able to conﬁdently iden- tify a source. The plume’s larger width suggested the source area was larger or farther aﬁeld compared to plume 1. It was observed 5 km downwind from the small town of La- mont, OK, where aerial/satellite (Google and Maxar Tech- nologies; dated July 2015) and street-level imagery (Google; dated May 2013; e.g., at 36.6949◦N, 97.5568◦W; viewing SW) revealed storage facilities for agricultural supplies, in- cluding tanks and tank trailers with “ammonia” labels. We hypothesized that leakage associated with such storage facil- ities contributed to plume 2. Via Eq. (1), we obtained a NH3 source rate of 2.7 ± 0.4 kg h−1. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Figure 12. (a) Close-up view of Fig. 10 for the two NH3 plume transects south of the town of Lamont, OK. As in Fig. 10, the IQR of the ﬂux footprints are overlaid on satellite imagery and color-coded for the measured ﬂuxes using the same color scale. Additionally, measured NH3 mixing ratios are shown by separately color-coded markers along the ﬂight track. Panels (b) and (c) are photographs of the cattle farm identiﬁed as the source of plume 1, made by the nadir and forward cameras on the G-1 aircraft, respectively. (The nadir camera was slightly tilted to starboard, allowing the farm to come into view.) Figure 12. (a) Close-up view of Fig. 10 for the two NH3 plume transects south of the town of Lamont, OK. As in Fig. 10, the IQR of the ﬂux footprints are overlaid on satellite imagery and color-coded for the measured ﬂuxes using the same color scale. Additionally, measured NH3 mixing ratios are shown by separately color-coded markers along the ﬂight track. Panels (b) and (c) are photographs of the cattle farm identiﬁed as the source of plume 1, made by the nadir and forward cameras on the G-1 aircraft, respectively. (The nadir camera was slightly tilted to starboard, allowing the farm to come into view.) cluded as a source of NH3 in the NEI (2017 data) point in- ventory nor resolved in the area inventory. For estimating the NH3 source rate from the farm, we ﬁrst used the same ap- proach as for the fertilizer plant plume transects (Eq. 1), even though mixing throughout the boundary layer might not have occurred in this case, due to the source’s proximity, so the obtained 0.6 ± 0.1 kg h−1 value was likely a high estimate. As described above, the corresponding FCWT peak is likely missing its maximum, while its ﬂanks are subject to large un- certainties. But putting these issues aside at ﬁrst, we demon- strate how the observed ﬂuxes and footprint considerations could be used to construct a reasonable low estimate for the NH3 source rate. For that, we assumed the source area mea- sured 200 m by 200 m (the observed plume width and about half of the farm’s dimensions), and, more importantly, that it contributed maximally to the observed ﬂux. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia To achieve max- imum contribution, we assumed the source area was located optimally near the footprint’s maximum and exactly occu- pied the footprint in the crosswind dimension. The former assumption was not unreasonable given the farm’s actual dis- tance from the ﬂight path vs. IQR locations (cf. Figs. 9 and 12). The latter assumption would also err on the intended side, given the low (> 200 m) spatial resolution we conjec- tured for our FCWT. Remaining conservative, we estimated the farm contributed 20 mol km−2 h−1 to the total observed FCWT at peak. With that, we obtained 0.17 kg h−1 for our low estimate for the NH3 source rate. In combination, these considerations constrained the cattle farm’s momentary NH3 emissions to between 0.2 and 0.6 kg h−1. However, we want to remind readers that the low estimate, in this case, is subject to a ﬂux measurement uncertainty on the order of 100 %. cluded as a source of NH3 in the NEI (2017 data) point in- ventory nor resolved in the area inventory. For estimating the NH3 source rate from the farm, we ﬁrst used the same ap- proach as for the fertilizer plant plume transects (Eq. 1), even though mixing throughout the boundary layer might not have occurred in this case, due to the source’s proximity, so the obtained 0.6 ± 0.1 kg h−1 value was likely a high estimate. As described above, the corresponding FCWT peak is likely missing its maximum, while its ﬂanks are subject to large un- certainties. But putting these issues aside at ﬁrst, we demon- strate how the observed ﬂuxes and footprint considerations could be used to construct a reasonable low estimate for the NH3 source rate. For that, we assumed the source area mea- sured 200 m by 200 m (the observed plume width and about half of the farm’s dimensions), and, more importantly, that it contributed maximally to the observed ﬂux. To achieve max- imum contribution, we assumed the source area was located optimally near the footprint’s maximum and exactly occu- pied the footprint in the crosswind dimension. The former assumption was not unreasonable given the farm’s actual dis- tance from the ﬂight path vs. IQR locations (cf. Figs. 9 and 12). The latter assumption would also err on the intended side, given the low (> 200 m) spatial resolution we conjec- tured for our FCWT. Evaluation of historical satellite imagery, using Google Earth, indicated that most of this speciﬁc farm’s feed- lots were created in 2007/2008 and smaller expansions im- plemented between 2012 and 2015. However, it is not in- https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 264 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia We estimated that at least about half of the random error was due to instru- mental noise, which could be reduced by ﬂying at lower air- speeds. Calibration experiments revealed a humidity dependence of the sensitivity of the NH3 · C6D+ 6 ion counts to NH3 mix- ing ratios, in particular a substantial drop in sensitivity when sampling relatively dry air (RH < 20 % at room temperature). In the atmospheric boundary layer, evaporation from the sur- face typically causes a substantial upward ﬂux of water (of- ten expressed as a latent heat ﬂux). That ﬂux would generally cause a positive bias in derived NH3 ﬂuxes if the sensitiv- ity’s humidity dependence was not considered, and a precise understanding of that dependence becomes even more im- portant. Consequently, when we assessed that understanding rather conservatively, it became the largest source of system- atic uncertainties (Sect. 3.8). We were not able to determine the mechanism behind that humidity dependence, and it re- mained unclear if the source of humidity in the IMR matters. In any case, we suggest careful calibrations prior to future ﬁeld deployments. Active humidiﬁcation of the IMR could be considered while keeping in mind the risk of possibly in- troducing contaminants. Addition of a suitable dopant might also be effective in reducing the humidity dependence. g y p The time response of our setup to changes in NH3 mix- ing ratios was on the order of a second. As we demonstrated, such a quick response makes the instrument very well suited for precise measurements in airborne applications and EC analysis. There was a non-negligible background signal (up to 10 % of the total) that responded more slowly, on the order of a few minutes, which may become an issue in the form of high relative background signal when quickly transitioning from generally high to relatively much lower NH3 mixing ratios. Corrections for that background response time could be considered, e.g., analogous to the time response correction method discussed in Nguyen et al. (2015). For this study, we did not apply such corrections but assessed them to amount to < 10 % of overestimation and < 25 % of underestimation for the worst-case plume transect and climb into a cleaner free troposphere, respectively. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia IMR. To avoid exposure, the instrument exhaust was routed outside of the aircraft cabin or into a fume hood exhaust when in the lab. The risk of spillage remained, especially when re- ﬁlling the benzene reservoir in ﬁeld settings. One could at- tempt to substitute with toluene, which has seen use in lieu of benzene for some applications for that very reason (Alton and Browne, 2020). tive and responsive device for measuring ambient NH3. We demonstrated its capabilities by presenting results from its deployment on a G-1 aircraft during the HI-SCALE ﬁeld campaign. The focus was on analyzing plume transects, as well as on eddy covariance (EC) analysis to derive vertical ﬂuxes that we connected with agricultural NH3 emissions through footprint considerations. For our EC analysis, we largely followed the example set by W2018 and used the CWT method to obtain time series of turbulent NH3 ﬂuxes. It allowed us to constrain the net atmosphere–surface exchange (here, mostly emissions) of NH3 upwind of the ﬂight path and to do so at a higher spatial resolution than what is generally achieved via the more tradi- tional EA method. The CIMS data featured frequent gaps due to fast zero measurements (every 42 s for 6 s), which caused comparatively larger gaps (15 s) in the ﬂux time series. How- ever, frequent zeros are required for accurately and precisely quantifying mixing ratios, especially when experiencing fast changes, which are typical in airborne applications (cf., Lee et al., 2018), and errors in mixing ratios directly propagate into the derived ﬂuxes. The resulting tradeoff between accu- racy and spatial coverage of the ﬂux data requires consid- eration. It has also become clear that longer legs would be beneﬁcial, as many of the ﬂux data ended up ﬂagged and discarded due to possible edge effects of unquantiﬁed larger- scale covariance (“COI”, Fig. 7). It may be useful to estimate that possible error and continue using many of these ﬂagged ﬂuxes, furnished with appropriate uncertainties, rather than just discarding them. Longer legs, as well as lower altitudes, are also expected to reduce systematic and random errors re- lated to (low-frequency) turbulence (Sect. 3.8). Random er- rors were clearly the dominant source of uncertainty for the obtained ﬂux time series. They often dropped below 100 % only after averaging to at least ∼1 km, thereby imposing ef- fective limits to spatially resolving ﬂuxes. Atmos. Meas. Tech., 16, 247–271, 2023 4 Summary and conclusions We have presented a new mass-spectrometry-based tech- nique for detecting and quantifying NH3 mixing ratios, speciﬁcally via chemical ionization using benzene cations. The technique was adapted to a CIMS instrument that had been modiﬁed for airborne measurements, in particular an efﬁcient sampling setup, which resulted in a highly sensi- https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 265 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia D footprints could be calculated, e.g., using the full footprint parametrization proposed by Kljun et al. (2015), for which those ﬂuctuations should lead to a lateral broadening. How- ever, additional input parameters are then required, and the processing of the sum of obtained footprints for each ﬂight leg will be more involved. The 2-D approach was pursued, for example, by Hannun et al. (2020) to attribute airborne CWT ﬂuxes of greenhouse gases to different land classes. have seen that switching back and forth to another reagent ion (or several other reagent ions) works at least for iodide-CI here, multiplying the detectable range of compounds. From the NH3 point of view, interesting compounds detectable by iodide-CI could be HNO3, HONO and other oxidized forms of N (e.g., Lee et al., 2018). Our experience with such a mode of operation was mixed for an airborne deployment, but it is likely more viable during ground-based deployments. g g This paper’s main goal is to introduce (airborne) benzene- CIMS as a method of measuring NH3 mixing ratios and ﬂuxes, but there are also some scientiﬁc takeaways. First, we provided an overview of the variability of NH3 in the lower troposphere, likely representative at least for rural Oklahoma in May (Fig. S3). Boundary-layer mixing ratios spanned over 1 order of magnitude (broadly from 1 to 10 ppbv) but within each ﬂight appeared vertically well mixed. Tens of ppbv were observed in plumes from a large fertilizer plant. Free tropo- spheric NH3 mixing ratios were a factor of 3 to 10 lower and reached down to 100 pptv. A better understanding of the vertical distribution and transfer of NH3 may be desirable, as a substantial fraction of NH4NO3 may actually occur in the cooler upper layers of the atmosphere (e.g., Fig. 6b), in- cluding the free troposphere (Paulot et al., 2016; Höpfner et al., 2019). Airborne in situ measurements as we present here could provide observational constraints, in particular if en- compassing more than occasional stages of research ﬂights and climbing sufﬁciently high. Appropriately planned, com- prehensive NH3 vertical proﬁle measurements could also be used to improve satellite retrievals (Van Damme et al., 2015). Second, our analysis of NH3 plumes and ﬂuxes, which we tied to surface emissions, provided quantiﬁed snapshots of agriculture-related area and point sources of NH3. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia It appeared that at least the slowly responding background was due to repartitioning of NH3 from walls of both the inlet line and the IMR, suggest- ing that attention should be paid to the designs of both the sampling setup and the IMR geometry in view of limiting wall interactions to achieve optimal performance. Also, ac- curacy would be improved if instrument backgrounds were determined by overﬂowing the full inlet instead of only the IMR; again, this is especially relevant for transitions to rela- tively lower mixing ratios. When connecting aircraft-measured EC ﬂuxes to emis- sions from (or deposition to) the ground, also vertical ﬂux divergence needs to be considered, as discussed. Flight pro- ﬁles should be planned to allow for the collected data to con- strain the terms in Eq. 2 as needed (e.g., Karl et al., 2013; W2018). For estimating the surface areas affecting the mea- sured ﬂuxes, the ﬂux footprints must be estimated, which in general are functions of upwind distance and direction, preferably perpendicular to the ﬂight track. Again, suitable ﬂight planning could enable valuable additional constraints, e.g., via parallel legs that lead to partial footprint overlaps. For the footprint estimates in this study, we used a conve- nient mix of established parametrizations to obtain 1-D ﬂux footprint functions. Required inputs, in addition to wind data, were boundary-layer depth and near-surface sensible heat ﬂux. Estimating the latter from airborne data introduced most uncertainty regarding footprint dimensions. Alternatively, 2- A different, practical concern may be the toxicity of ben- zene, which was consumed in substantial quantities in order to achieve mixing ratios of > 100 ppm (at 100 mbar) in the https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 266 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia The re- sults were overall consistent with the NEI inventories, but substantial point sources also seemed to be missing in the NEI. That ﬁnding is in line with recent literature arriving at similar conclusions (see Introduction section). If the focus is on NH3, on the other hand, it may be feasible to use benzene-CI also in a smaller, lighter and cheaper mass spectrometer, e.g., using a residual gas analyzer. During HI- SCALE, NH3 · C6D+ 6 was by far the dominant composition detected at m/z 101, and one could have gone without the high resolution provided by an expensive TOF, as long as sensitivity was preserved. Code and data availability. Data from the HI-SCALE cam- paign are accessible via the ARM Intensive Operating Period (IOP) Data Browser upon free registration (At- mospheric Radiation Measurement User Facility, 2022; https://iop.archive.arm.gov/arm-iop/2016/sgp/hiscale/, last ac- cess: 24 August 2022). National Emissions Inventory 2017 data are available from the U.S. Environmental Protection Agency (2022; https://www.epa.gov/air-emissions-inventories/ 2017-national-emissions-inventory-nei-data, last access: 24 Au- gust 2022). All data used in this study are also available from the authors upon request. We used the Google Maps web mapping platform to access satellite and street-level imagery (Google, 2022; https://www.google.com/maps, last access: 24 August 2022, in- cluding used imagery). The EC ﬂux analysis toolbox is available on GitHub (Wolfe, 2022; https://github.com/AirChem/FluxToolbox, last access: 29 November 2022). S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Bell, D. M., Imre, D., T. Martin, S., and Zelenyuk, A.: The proper- ties and behavior of α-pinene secondary organic aerosol particles exposed to ammonia under dry conditions, Phys. Chem. Chem. Phys., 19, 6497–6507, 2017. Disclaimer. Publisher’s note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. Bertram, T. H., Kimmel, J. R., Crisp, T. A., Ryder, O. S., Yatavelli, R. L. N., Thornton, J. A., Cubison, M. J., Gonin, M., and Worsnop, D. R.: A ﬁeld-deployable, chemical ionization time- of-ﬂight mass spectrometer, Atmos. Meas. Tech., 4, 1471–1479, https://doi.org/10.5194/amt-4-1471-2011, 2011. Acknowledgements. We thank the ARM Aerial Facility team and everyone involved with the HI-SCALE campaign for the outstand- ing support and collaboration. Particular thanks go to payload direc- tor John Hubbe, whose support from the planning phase onwards assured the seamless integration of the CIMS into the G-1. We also appreciate the help of Dennis Canuelle in the machine shop, of Lexie Goldberger during test ﬂights, and valuable discussions re- garding the benzene-CI technique with Tim Bertram and Gordon Novak (University of Wisconsin–Madison). Beswick, K. M., Gallagher, M. W., Webb, A. R., Norton, E. G., and Perry, F.: Application of the Aventech AIMMS20AQ air- borne probe for turbulence measurements during the Convec- tive Storm Initiation Project, Atmos. Chem. Phys., 8, 5449–5463, https://doi.org/10.5194/acp-8-5449-2008, 2008. Bressi, M., Cavalli, F., Putaud, J. P., Fröhlich, R., Petit, J. E., Aas, W., Äijälä, M., Alastuey, A., Allan, J. D., Aurela, M., Berico, M., Bougiatioti, A., Bukowiecki, N., Canonaco, F., Crenn, V., Dusanter, S., Ehn, M., Elsasser, M., Flentje, H., Graf, P., Green, D. C., Heikkinen, L., Hermann, H., Holzinger, R., Hueglin, C., Keernik, H., Kiendler-Scharr, A., Kubelová, L., Lunder, C., Maasikmets, M., Makeš, O., Malaguti, A., Mihalopoulos, N., Nicolas, J. B., O’Dowd, C., Ovadnevaite, J., Petralia, E., Poulain, L., Priestman, M., Riffault, V., Ripoll, A., Schlag, P., Schwarz, J., Sciare, J., Slowik, J., Sosedova, Y., Stavroulas, I., Teine- maa, E., Via, M., Vodiˇcka, P., Williams, P. I., Wiedensohler, A., Young, D. E., Zhang, S., Favez, O., Minguillón, M. C., and Prevot, A. S. H.: A European aerosol phenomenology – 7: High-time resolution chemical characteristics of submicron par- ticulate matter across Europe, Atmos. Environ. X, 10, 100108, https://doi.org/10.1016/j.aeaoa.2021.100108, 2021. We gratefully acknowledge the ﬁnancial support by the Euro- pean Commission, the Academy of Finland, the U.S. Department of Energy (DOE), the Paciﬁc Northwest National Laboratory and the National Science Foundation. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia John E. Shilling and Jerome Fast are supported by the DOE Atmospheric System Research (ASR) pro- gram. Manish Shrivastava was supported by the DOE Ofﬁce of Sci- ence, Ofﬁce of Biological and Environmental Research through the Early Career Research Program. Paciﬁc Northwest National Labo- ratory (PNNL) is operated for DOE by Battelle Memorial Institute under contract DE-AC05-76RL01830. The contributions by PNNL were also funded by the DOE’s Atmospheric Radiation Measure- ment (ARM) User Facility. L., Priestman, M., Riffault, V., Ripoll, A., Schlag, P., Schwarz, Financial support. This research has been supported by the H2020 Marie Skłodowska-Curie Actions (grant no. 701958), the Academy of Finland (grant nos. 310682 and 337550), the Ofﬁce of Science (grant nos. DE-SC0021097 and DE-SC0011791), the Paciﬁc North- west National Laboratory (grant no. 243766 T.O. 276416), and the National Science Foundation (grant no. DGE-1256082). Calvet, S., Gates, R. S., Zhang, G., Estellés, F., Ogink, N. W. M., Pedersen, S., and Berckmans, D.: Measuring gas emissions from livestock buildings: A review on uncertainty analysis and error sources, Biosyst. Eng., 116, 221–231, 2013. Cape, J. N., van der Eerden, L. J., Sheppard, L. J., Leith, I. D., and Sutton, M. A.: Evidence for changing the critical level for am- monia, Environ. Pollut., 157, 1033–1037, 2009. Review statement. This paper was edited by Reem Hannun and re- viewed by Glenn Wolfe and one anonymous referee. Crawford, T. L., Dobosy, R. J., McMillen, R. T., Vogel, C. A., and Hicks, B. B.: Air-surface exchange measurement in heteroge- neous regions: extending tower observations with spatial struc- ture observed from small aircraft, Glob. Chang. Biol., 2, 275– 285, 1996. Supplement. The supplement related to this article is available on- line at: https://doi.org/10.5194/amt-16-247-2023-supplement. Supplement. The supplement related to this article is available on- line at: https://doi.org/10.5194/amt-16-247-2023-supplement. We suggest that setups to measure EC ﬂuxes of NH3 could play an important role in providing top-down observational constraints on NH3 emissions, in particular from agricultural sources. Airborne measurements in particular could help by providing regional coverage across ranges of surface prop- erties and ecosystems. As they can resolve air–surface ex- change at high spatial resolution, they could be used in con- junction with detailed information regarding concurrent agri- cultural practices (e.g., timing and type of fertilizer appli- cation or manure management). Ground-based deployments, on the other hand, would more easily provide longer-term and more continuous information to cover a wider range of environmental conditions and surface activities. Author contributions. JF led the HI-SCALE ﬁeld campaign, in par- ticular the G-1 research ﬂights, and prepared NEI area emissions inventory data. JAT perceived and supervised the deployment of the CIMS using dual reagent ions on the G-1. The CIMS setup and installation were realized by SS and tested and deployed by SS, ELD’A and BHL. DMB and JES helped with its operation, in par- ticular during research ﬂights. Calibration experiments were per- formed and analyzed by SS, ELD’A and LV. SS analyzed the HI- SCALE CIMS data. Eddy-covariance analysis was performed by SS, with assistance from QP and MP. AMS data were provided by JES. MS performed the WRF-Chem model runs and provided re- spective details and ﬁgures. SS wrote the paper. All co-authors par- ticipated in paper-related discussions and commented on the paper. An additional strength of the (TOF-)CIMS method in par- ticular is that a range of other compounds can be quanti- ﬁed independently and at the same time. For benzene-CI, the obvious candidates are isoprene, terpenes and dimethyl sulﬁde (Lavi et al., 2018), as well as a range of many other volatile organics, such as polyaromatic hydrocarbons. We Competing interests. The contact author has declared that none of the authors has any competing interests. Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 267 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia C.: Underestimated Ammonia Emissions from Road Vehicles, Environ. Sci. Technol., 54, 15689–15697, 2020. Horst, T. W.: A simple formula for attenuation of eddy ﬂuxes mea- sured with ﬁrst-order-response scalar sensors, Bound.-Lay. Me- teorol., 82, 219–233, 1997. Fast, J. D., Berg, L. K., Alexander, L., Bell, D., D’Ambro, E., Hubbe, J., Kuang, C., Liu, J., Long, C., Matthews, A., Mei, F., Newsom, R., Pekour, M., Pinterich, T., Schmid, B., Schobes- berger, S., Shilling, J., Smith, J. N., Springston, S., Suski, K., Thornton, J. A., Tomlinson, J., Wang, J., Xiao, H., and Zelenyuk, A.: Overview of the HI-SCALE Field Campaign: A New Per- spective on Shallow Convective Clouds, B. Am. Meteorol. Soc., 100, 821–840, 2019. Jokinen, T., Sipilä, M., Junninen, H., Ehn, M., Lönn, G., Hakala, J., Petäjä, T., Mauldin III, R. L., Kulmala, M., and Worsnop, D. R.: Atmospheric sulphuric acid and neutral cluster measure- ments using CI-APi-TOF, Atmos. Chem. Phys., 12, 4117–4125, https://doi.org/10.5194/acp-12-4117-2012, 2012. Jokinen, T., Sipilä, M., Kontkanen, J., Vakkari, V., Tisler, P., Du- plissy, E.-M., Junninen, H., Kangasluoma, J., Manninen, H. E., Petäjä, T., Kulmala, M., Worsnop, D. R., Kirkby, J., Virkkula, A., and Kerminen, V.-M.: Ion-induced sulfuric acid–ammonia nucle- ation drives particle formation in coastal Antarctica, Sci. Adv., 4, eaat9744, https://doi.org/10.1126/sciadv.aat9744, 2018. Finkelstein, P. L. and Sims, P. F.: Sampling error in eddy correlation ﬂux measurements, J. Geophys. Res.-Atmos., 106, 3503–3509, 2001. Foken, T. and Wichura, B.: Tools for quality assessment of surface- based ﬂux measurements, Agr. Forest Meteorol., 78, 83–105, 1996. Junninen, H., Ehn, M., Petäjä, T., Luosujärvi, L., Kotiaho, T., Kos- tiainen, R., Rohner, U., Gonin, M., Fuhrer, K., Kulmala, M., and Worsnop, D. R.: A high-resolution mass spectrometer to mea- sure atmospheric ion composition, Atmos. Meas. Tech., 3, 1039– 1053, https://doi.org/10.5194/amt-3-1039-2010, 2010. Fortems-Cheiney, A., Dufour, G., Hamaoui-Laguel, L., Foret, G., Siour, G., Van Damme, M., Meleux, F., Coheur, P. F., Clerbaux, C., Clarisse, L., Favez, O., Wallasch, M., and Beekmann, M.: Unaccounted variability in NH3 agricultural sources detected by IASI contributing to European spring haze episode, Geophys. Res. Lett., 43, 5475–5482, 2016. Kaimal, J. C. and Finnigan, J. J.: Atmospheric Boundary Layer Flows: Their Structure and Measurement, Oxford University Press, New York, NY, ISBN 9780195062397, 1994. Franchin, A., Fibiger, D. L., Goldberger, L., McDufﬁe, E. E., Moravek, A., Womack, C. C., Crosman, E. T., Docherty, K. S., Dube, W. P., Hoch, S. W., Lee, B. H., Long, R., Murphy, J. G., Thornton, J. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 268 enlechner, M., Clarke, A., Curtius, J., Dommen, J., Donahue, N. M., Ehrhart, S., Flagan, R. C., Franchin, A., Guida, R., Hakala, J., Hansel, A., Heinritzi, M., Jokinen, T., Kangasluoma, J., Kirkby, J., Kulmala, M., Kupc, A., Lawler, M. J., Lehtipalo, K., Makhmutov, V., Mann, G., Mathot, S., Merikanto, J., Miet- tinen, P., Nenes, A., Onnela, A., Rap, A., Reddington, C. L. S., Riccobono, F., Richards, N. A. D., Rissanen, M. P., Rondo, L., Sarnela, N., Schobesberger, S., Sengupta, K., Simon, M., Sipila, M., Smith, J. N., Stozkhov, Y., Tome, A., Trostl, J., Wagner, P. E., Wimmer, D., Winkler, P. M., Worsnop, D. R., and Carslaw, K. S.: Global atmospheric particle formation from CERN CLOUD measurements, Science, 354, 1119–1124, 2016. Hannun, R. A., Wolfe, G. M., Kawa, S. R., Hanisco, T. F., New- man, P. A., Alﬁeri, J. G., Barrick, J., Clark, K. L., DiGangi, J. P., Diskin, G. S., King, J., Kustas, W. P., Mitra, B., Noormets, A., Nowak, J. B., Thornhill, K. L., and Vargas, R.: Spatial hetero- geneity in CO2, CH4, and energy ﬂuxes: insights from airborne eddy covariance measurements over the Mid-Atlantic region, Environ. Res. Lett., 15, 035008, https://doi.org/10.1088/1748- 9326/ab7391, 2020. Hakala, J., Hansel, A., Heinritzi, M., Jokinen, T., Kangasluoma, J., Kirkby, J., Kulmala, M., Kupc, A., Lawler, M. J., Lehtipalo, Riccobono, F., Richards, N. A. D., Rissanen, M. P., Rondo, L., Sarnela N Schobesberger S Sengupta K Simon M Sipila Höpfner, M., Ungermann, J., Borrmann, S., Wagner, R., Spang, R., Riese, M., Stiller, G., Appel, O., Batenburg, A. M., Bucci, S., Cairo, F., Dragoneas, A., Friedl-Vallon, F., Hünig, A., Johansson, S., Krasauskas, L., Legras, B., Leisner, T., Mahnke, C., Möhler, O., Molleker, S., Müller, R., Neubert, T., Orphal, J., Preusse, P., Rex, M., Saathoff, H., Stroh, F., Weigel, R., and Wohltmann, I.: Ammonium nitrate particles formed in upper troposphere from ground ammonia sources during Asian monsoons, Nat. Geosci., 12, 608–612, 2019. Smith, J. N., Stozkhov, Y., Tome, A., Trostl, J., Wagn E., Wimmer, D., Winkler, P. M., Worsnop, D. R., and Carslaw, K. S.: Global atmospheric particle formation from CERN CLOUD measurements, Science, 354, 1119–1124, 2016. S., Krasauskas, L., Legras, B., Leisner, T., Mahnke, C., Möhler, O., Molleker, S., Müller, R., Neubert, T., Orphal, J., Preusse, P., EEA: EMEP/EEA air pollutant emission inventory guidebook 2019, Luxembourg, Report No. 13/2019, https://doi.org/10.2800/293657, 2019. Farren, N. J., Davison, J., Rose, R. A., Wagner, R. L., and Carslaw, D. References Dabberdt, W. F., Lenschow, D. H., Horst, T. W., Zimmerman, P. R., Oncley, S. P., and Delany, A. C.: Atmosphere-Surface Exchange Measurements, Science, 260, 1472–1481, 1993. Aljawhary, D., Lee, A. K. Y., and Abbatt, J. P. D.: High-resolution chemical ionization mass spectrometry (ToF-CIMS): application to study SOA composition and processing, Atmos. Meas. Tech., 6, 3211–3224, https://doi.org/10.5194/amt-6-3211-2013, 2013. Desjardins, R. L., Brach, E. J., Alvo, P., and Schuepp, P. H.: Aircraft Monitoring of Surface Carbon Dioxide Exchange, Science, 216, 733–735, 1982. Alton, M. W. and Browne, E. C.: Atmospheric Chemistry of Volatile Methyl Siloxanes: Kinetics and Products of Oxidation by OH Radicals and Cl Atoms, Environ. Sci. Technol., 54, 5992–5999, 2020. Desjardins, R. L., Worth, D. E., Pattey, E., VanderZaag, A., Srini- vasan, R., Mauder, M., Worthy, D., Sweeney, C., and Metzger, S.: The challenge of reconciling bottom-up agricultural methane emissions inventories with top-down measurements, Agr. Forest Meteorol., 248, 48–59, 2018. Atmospheric Radiation Measurement User Facility (ARM): Inten- sive Operating Period (IOP) Data Browser, Atmospheric Radi- ation Measurement User Facility [data set], https://iop.archive. arm.gov/arm-iop/2016/sgp/hiscale, last access: 24 August 2022. Diau, E. W. G., Tso, T. L., and Lee, Y. P.: Kinetics of the reaction hydroxyl + ammonia in the range 273–433 K, J. Phys. Chem., 94, 5261–5265, 1990. Attié, J.-L. and Durand, P.: Conditional Wavelet Technique Applied to Aircraft Data Measured in the Thermal Internal Boundary Layer During Sea-Breeze Events, Bound.-Lay. Meteorol., 106, 359–382, 2003. Dunne, E. M., Gordon, H., Kuerten, A., Almeida, J., Duplissy, J., Williamson, C., Ortega, I. K., Pringle, K. J., Adamov, A., Baltensperger, U., Barmet, P., Benduhn, F., Bianchi, F., Breit- Atmos. Meas. Tech., 16, 247–271, 2023 Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 269 Krupa, S. V.: Effects of atmospheric ammonia (NH3) on terrestrial vegetation: a review, Environ. Pollut., 124, 179–221, 2003. Mauder, M., Desjardins, R. L., and MacPherson, I.: Scale anal- ysis of airborne ﬂux measurements over heterogeneous terrain in a boreal ecosystem, J. Geophys. Res.-Atmos., 112, D13112, https://doi.org/10.1029/2006JD008133, 2007. Langford, B., Acton, W., Ammann, C., Valach, A., and Nemitz, E.: Eddy-covariance data with low signal-to-noise ratio: time-lag de- termination, uncertainties and limit of detection, Atmos. Meas. Tech., 8, 4197–4213, https://doi.org/10.5194/amt-8-4197-2015, 2015. Mauder, M., Cuntz, M., Drüe, C., Graf, A., Rebmann, C., Schmid, H. P., Schmidt, M., and Steinbrecher, R.: A strategy for quality and uncertainty assessment of long-term eddy-covariance mea- surements, Agr. Forest Meteorol., 169, 122–135, 2013. Lavi, A., Vermeuel, M. P., Novak, G. A., and Bertram, T. H.: The sensitivity of benzene cluster cation chemical ioniza- tion mass spectrometry to select biogenic terpenes, Atmos. Meas. Tech., 11, 3251–3262, https://doi.org/10.5194/amt-11- 3251-2018, 2018. Metzger, S., Junkermann, W., Mauder, M., Butterbach-Bahl, K., Trancón y Widemann, B., Neidl, F., Schäfer, K., Wieneke, S., Zheng, X. H., Schmid, H. P., and Foken, T.: Spatially ex- plicit regionalization of airborne ﬂux measurements using envi- ronmental response functions, Biogeosciences, 10, 2193–2217, https://doi.org/10.5194/bg-10-2193-2013, 2013. Lee, B. H., Lopez-Hilﬁker, F. D., Mohr, C., Kurtén, T., Worsnop, D. R., and Thornton, J. A.: An Iodide-Adduct High-Resolution Time-of-Flight Chemical-Ionization Mass Spectrometer: Appli- cation to Atmospheric Inorganic and Organic Compounds, Envi- ron. Sci. Technol., 48, 6309–6317, 2014. Miller, D. J., Sun, K., Tao, L., Khan, M. A., and Zondlo, M. A.: Open-path, quantum cascade-laser-based sensor for high- resolution atmospheric ammonia measurements, Atmos. Meas. Tech., 7, 81–93, https://doi.org/10.5194/amt-7-81-2014, 2014. Lee, B. H., Lopez-Hilﬁker, F. D., Veres, P. R., McDufﬁe, E. E., Fibiger, D. L., Sparks, T. L., Ebben, C. J., Green, J. R., Schroder, J. C., Campuzano-Jost, P., Iyer, S., D’Ambro, E. L., Schobes- berger, S., Brown, S. S., Wooldridge, P. J., Cohen, R. C., Fiddler, M. N., Bililign, S., Jimenez, J. L., Kurtén, T., Weinheimer, A. J., Jaegle, L., and Thornton, J. A.: Flight Deployment of a High- Resolution Time-of-Flight Chemical Ionization Mass Spectrom- eter: Observations of Reactive Halogen and Nitrogen Oxide Species, J. Geophys. Res.-Atmos., 123, 7670–7686, 2018. Misztal, P. K., Karl, T., Weber, R., Jonsson, H. H., Guenther, A. B., and Goldstein, A. H.: Airborne ﬂux measurements of bio- genic isoprene over California, Atmos. Chem. Phys., 14, 10631– 10647, https://doi.org/10.5194/acp-14-10631-2014, 2014. Moravek, A., Murphy, J. G., Hrdina, A., Lin, J. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia C., Pennell, C., Franchin, A., Middlebrook, A. M., Fibiger, D. L., Womack, C. C., McDufﬁe, E. E., Martin, R., Moore, K., Baasandorj, M., and Brown, S. S.: Wintertime spatial distribution of ammonia and its emission sources in the Great Salt Lake region, Atmos. Chem. Phys., 19, 15691–15709, https://doi.org/10.5194/acp-19-15691- 2019, 2019a. Leen, J. B., Yu, X.-Y., Gupta, M., Baer, D. S., Hubbe, J. M., Kluzek, C. D., Tomlinson, J. M., and Hubbell, M. R.: Fast In Situ Air- borne Measurement of Ammonia Using a Mid-Infrared Off-Axis ICOS Spectrometer, Environ. Sci. Technol., 47, 10446–10453, 2013. Moravek, A., Singh, S., Pattey, E., Pelletier, L., and Mur- phy, J. G.: Measurements and quality control of ammonia eddy covariance ﬂuxes: a new strategy for high-frequency attenuation correction, Atmos. Meas. Tech., 12, 6059–6078, https://doi.org/10.5194/amt-12-6059-2019, 2019b. Leibrock, E. and Huey, L. G.: Ion chemistry for the detection of isoprene and other volatile organic compounds in ambient air, Geophys. Res. Lett., 27, 1719–1722, 2000. Lelieveld, J., Evans, J. S., Fnais, M., Giannadaki, D., and Pozzer, A.: The contribution of outdoor air pollution sources to premature mortality on a global scale, Nature, 525, 367–371, 2015. Nguyen, T. B., Crounse, J. D., Teng, A. P., St. Clair, J. M., Paulot, F., Wolfe, G. M., and Wennberg, P. O.: Rapid deposition of oxidized biogenic compounds to a temperate forest, P. Natl. Acad. Sci. USA, 112, E392–E401, 2015. Lenschow, D. H., Delany, A. C., Stankov, B. B., and Stedman, D. H.: Airborne measurements of the vertical ﬂux of ozonen in the boundary layer, Bound.-Lay. Meteorol., 19, 249–265, 1980. Norman, M. and Leck, C.: Distribution of marine boundary layer ammonia over the Atlantic and Indian Oceans during the Aerosols99 cruise, J. Geophys. Res.-Atmos., 110, D16302, https://doi.org/10.1029/2005JD005866, 2005. Lenschow, D. H., Pearson Jr., R., and Stankov, B. B.: Estimating the ozone budget in the boundary layer by use of aircraft measure- ments of ozone eddy ﬂux and mean concentration, J. Geophys. Res.-Oceans, 86, 7291–7297, 1981. Nowak, J. B., Neuman, J. A., Kozai, K., Huey, L. G., Tanner, D. J., Holloway, J. S., Ryerson, T. B., Frost, G. J., McKeen, S. A., and Fehsenfeld, F. C.: A chemical ionization mass spectrometry tech- nique for airborne measurements of ammonia, J. Geophys. Res.- Atmos., 112, D10S02, https://doi.org/10.1029/2006JD007589, 2007. Lenschow, D. H., Mann, J., and Kristensen, L.: How Long Is Long Enough When Measuring Fluxes and Other Turbulence Statis- tics?, J. Atmos. Ocean. Technol., 11, 661–673, 1994. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia A., Brown, S. S., Baasandorj, M., and Middlebrook, A. M.: Airborne and ground-based observations of ammonium- nitrate-dominated aerosols in a shallow boundary layer during in- tense winter pollution episodes in northern Utah, Atmos. Chem. Phys., 18, 17259–17276, https://doi.org/10.5194/acp-18-17259- 2018, 2018. Karl, T., Apel, E., Hodzic, A., Riemer, D. D., Blake, D. R., and Wiedinmyer, C.: Emissions of volatile organic compounds in- ferred from airborne ﬂux measurements over a megacity, At- mos. Chem. Phys., 9, 271–285, https://doi.org/10.5194/acp-9- 271-2009, 2009. Karl, T., Misztal, P. K., Jonsson, H. H., Shertz, S., Goldstein, A. H., and Guenther, A. B.: Airborne Flux Measurements of BVOCs above Californian Oak Forests: Experimental Investigation of Surface and Entrainment Fluxes, OH Densities, and Damköhler Numbers, J. Atmos. Sci., 70, 3277–3287, 2013. Fulgham, S. R., Brophy, P., Link, M., Ortega, J., Pollack, I., and Farmer, D. K.: Seasonal Flux Measurements over a Colorado Pine Forest Demonstrate a Persistent Source of Organic Acids, ACS Earth Space Chem., 3, 2017–2032, 2019. Kim, M. J., Zoerb, M. C., Campbell, N. R., Zimmermann, K. J., Blomquist, B. W., Huebert, B. J., and Bertram, T. H.: Revisiting benzene cluster cations for the chemical ionization of dimethyl sulﬁde and select volatile organic compounds, Atmos. Meas. Tech., 9, 1473–1484, https://doi.org/10.5194/amt-9-1473-2016, 2016. Google: Google Maps [web mapping platform], https://www. google.com/maps, last access: 24 August 2022. Grönroos, J., Munther, J., and Luostarinen, S.: Calculation of at- mospheric nitrogen and NMVOC emissions from Finnish agri- culture. Description of the revised model, Reports of the Finnish Environment Institute, vol. 37, 64 pp., ISBN 978-952-11-4902-3, 2017. Kljun, N., Calanca, P., Rotach, M. W., and Schmid, H. P.: A simple two-dimensional parameterisation for Flux Foot- print Prediction (FFP), Geosci. Model Dev., 8, 3695–3713, https://doi.org/10.5194/gmd-8-3695-2015, 2015. https://doi.org/10.5194/amt-16-247-2023 Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 270 nium nitrate formation, Geophys. Res. Lett., 39, L07804, https://doi.org/10.1029/2012GL051197, 2012. Sanchez, J., Tanner, D. J., Chen, D., Huey, L. G., and Ng, N. L.: A new technique for the direct detection of HO2 radi- cals using bromide chemical ionization mass spectrometry (Br- CIMS): initial characterization, Atmos. Meas. Tech., 9, 3851– 3861, https://doi.org/10.5194/amt-9-3851-2016, 2016. Olin, M., Okuljar, M., Rissanen, M. P., Kalliokoski, J., Shen, J., Dada, L., Lampimäki, M., Wu, Y., Lohila, A., Duplissy, J., Sipilä, M., Petäjä, T., Kulmala, M., and Dal Maso, M.: Measurement report: Atmospheric new particle formation in a coastal agricultural site explained with binPMF analysis of ni- trate CI-APi-TOF spectra, Atmos. Chem. Phys., 22, 8097–8115, https://doi.org/10.5194/acp-22-8097-2022, 2022. Sayres, D. S., Dobosy, R., Healy, C., Dumas, E., Kochendorfer, J., Munster, J., Wilkerson, J., Baker, B., and Anderson, J. G.: Arctic regional methane ﬂuxes by ecotope as derived using eddy covari- ance from a low-ﬂying aircraft, Atmos. Chem. Phys., 17, 8619– 8633, https://doi.org/10.5194/acp-17-8619-2017, 2017. Oliveira, M. D., Sousa, F. C., Saraz, J. O., Calderano, A. A., Tinôco, I. F., and Carneiro, A. P.: Ammonia Emission in Poultry Facil- ities: A Review for Tropical Climate Areas, Atmosphere, 12, 1091, https://doi.org/10.3390/atmos12091091, 2021. Schaller, C., Göckede, M., and Foken, T.: Flux calculation of short turbulent events – comparison of three methods, Atmos. Meas. Tech., 10, 869–880, https://doi.org/10.5194/amt-10-869- 2017, 2017. Pan, D., Benedict, K. B., Golston, L. M., Wang, R., Collett, J. L., Tao, L., Sun, K., Guo, X., Ham, J., Prenni, A. J., Schichtel, B. A., Mikoviny, T., Müller, M., Wisthaler, A., and Zondlo, M. A.: Am- monia Dry Deposition in an Alpine Ecosystem Traced to Agri- cultural Emission Hotpots, Environ. Sci. Technol., 55, 7776– 7785, 2021. Schobesberger, S., Junninen, H., Bianchi, F., Lönn, G., Ehn, M., Lehtipalo, K., Dommen, J., Ehrhart, S., Ortega, I. K., Franchin, A., Nieminen, T., Riccobono, F., Hutterli, M., Duplissy, J., Almeida, J., Amorim, A., Breitenlechner, M., Downard, A. J., Dunne, E. M., Flagan, R. C., Kajos, M., Keskinen, H., Kirkby, J., Kupc, A., Kürten, A., Kurtén, T., Laaksonen, A., Mathot, S., Onnela, A., Praplan, A. P., Rondo, L., Santos, F. D., Schallhart, S., Schnitzhofer, R., Sipilä, M., Tomé, A., Tsagkogeorgas, G., Vehkamäki, H., Wimmer, D., Baltensperger, U., Carslaw, K. S., Curtius, J., Hansel, A., Petäjä, T., Kulmala, M., Donahue, N. M., and Worsnop, D. R.: Molecular understanding of atmospheric particle formation from sulfuric acid and large oxidized organic molecules, P. Natl. Acad. Sci. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia USA, 110, 17223–17228, 2013. Paulot, F., Ginoux, P., Cooke, W. F., Donner, L. J., Fan, S., Lin, M.-Y., Mao, J., Naik, V., and Horowitz, L. W.: Sensitivity of ni- trate aerosols to ammonia emissions and to nitrate chemistry: implications for present and future nitrate optical depth, At- mos. Chem. Phys., 16, 1459–1477, https://doi.org/10.5194/acp- 16-1459-2016, 2016. Vehkamäki, H., Wimmer, D., Baltensperger, U., Carslaw, K. S., Pfeifer, J., Simon, M., Heinritzi, M., Piel, F., Weitz, L., Wang, D., Granzin, M., Müller, T., Bräkling, S., Kirkby, J., Curtius, J., and Kürten, A.: Measurement of ammonia, amines and io- dine compounds using protonated water cluster chemical ion- ization mass spectrometry, Atmos. Meas. Tech., 13, 2501–2522, https://doi.org/10.5194/amt-13-2501-2020, 2020. Schobesberger, S., Lopez-Hilﬁker, F. D., Taipale, D., Millet, D. B., D’Ambro, E. L., Rantala, P., Mammarella, I., Zhou, P., Wolfe, G. M., Lee, B. H., Boy, M., and Thornton, J. A.: High upward ﬂuxes of formic acid from a boreal forest canopy, Geophys. Res. Lett., 43, 9342–9351, 2016. Pollack, I. B., Lindaas, J., Roscioli, J. R., Agnese, M., Permar, W., Hu, L., and Fischer, E. V.: Evaluation of ambient ammo- nia measurements from a research aircraft using a closed-path QC-TILDAS operated with active continuous passivation, At- mos. Meas. Tech., 12, 3717–3742, https://doi.org/10.5194/amt- 12-3717-2019, 2019. Shephard, M. W., Dammers, E., Cady-Pereira, K. E., Kharol, S. K., Thompson, J., Gainariu-Matz, Y., Zhang, J., McLinden, C. A., Kovachik, A., Moran, M., Bittman, S., Sioris, C. E., Grif- ﬁn, D., Alvarado, M. J., Lonsdale, C., Savic-Jovcic, V., and Zheng, Q.: Ammonia measurements from space with the Cross- track Infrared Sounder: characteristics and applications, Atmos. Chem. Phys., 20, 2277–2302, https://doi.org/10.5194/acp-20- 2277-2020, 2020. Pye, H. O. T., Nenes, A., Alexander, B., Ault, A. P., Barth, M. C., Clegg, S. L., Collett Jr., J. L., Fahey, K. M., Hennigan, C. J., Her- rmann, H., Kanakidou, M., Kelly, J. T., Ku, I.-T., McNeill, V. F., Riemer, N., Schaefer, T., Shi, G., Tilgner, A., Walker, J. T., Wang, T., Weber, R., Xing, J., Zaveri, R. A., and Zuend, A.: The acid- ity of atmospheric particles and clouds, Atmos. Chem. Phys., 20, 4809–4888, https://doi.org/10.5194/acp-20-4809-2020, 2020. Stull, R. B.: An Introduction to Boundary Layer Me- teorology, Kluwer Academic Publishers, Dordrecht, https://doi.org/10.1007/978-94-009-3027-8, 1988. Sun, K., Tao, L., Miller, D. J., Zondlo, M. A., Shonkwiler, K. B., Nash, C., and Ham, J. M.: Open-path eddy covariance measure- ments of ammonia ﬂuxes from a beef cattle feedlot, Agr. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Liu, Y., San Liang, X., and Weisberg, R. H.: Rectiﬁcation of the Bias in the Wavelet Power Spectrum, J. Atmos. Ocean. Technol., 24, 2093–2102, 2007. Nowak, J. B., Neuman, J. A., Bahreini, R., Brock, C. A., Middlebrook, A. M., Wollny, A. G., Holloway, J. S., Peis- chl, J., Ryerson, T. B., and Fehsenfeld, F. C.: Airborne observations of ammonia and ammonium nitrate formation over Houston, Texas, J. Geophys. Res.-Atmos., 115, D22304, https://doi.org/10.1029/2010JD014195, 2010. Liu, Y., Liggio, J., Staebler, R., and Li, S.-M.: Reactive up- take of ammonia to secondary organic aerosols: kinetics of organonitrogen formation, Atmos. Chem. Phys., 15, 13569– 13584, https://doi.org/10.5194/acp-15-13569-2015, 2015. Nowak, J. B., Neuman, J. A., Bahreini, R., Middlebrook, A. M., Holloway, J. S., McKeen, S. A., Parrish, D. D., Ryer- son, T. B., and Trainer, M.: Ammonia sources in the Cal- ifornia South Coast Air Basin and their impact on ammo- Ma, R., Zou, J., Han, Z., Yu, K., Wu, S., Li, Z., Liu, S., Niu, S., Horwath, W. R., and Zhu-Barker, X.: Global soil-derived ammo- nia emissions from agricultural nitrogen fertilizer application: A reﬁnement based on regional and crop-speciﬁc emission factors, Glob. Chang. Biol., 27, 855–867, 2021. Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia 271 Wolfe, G. M.: FluxToolbox, GitHub [code], https://github.com/ AirChem/FluxToolbox, last access: 29 November 2022. Wolfe, G. M.: FluxToolbox, GitHub [code], https://github.com/ AirChem/FluxToolbox, last access: 29 November 2022. mospheric pollution load, Atmos. Chem. Phys., 21, 875–914, https://doi.org/10.5194/acp-21-875-2021, 2021. mospheric pollution load, Atmos. Chem. Phys., 21, 875–914, https://doi.org/10.5194/acp-21-875-2021, 2021. Torrence, C. and Compo, G. P.: A Practical Guide to Wavelet Anal- ysis, B. Am. Meteorol. Soc., 79, 61–78, 1998. Wolfe, G. M., Hanisco, T. F., Arkinson, H. L., Bui, T. P., Crounse, J. D., Dean-Day, J., Goldstein, A., Guenther, A., Hall, S. R., Huey, G., Jacob, D. J., Karl, T., Kim, P. S., Liu, X., Marvin, M. R., Mikoviny, T., Misztal, P. K., Nguyen, T. B., Peischl, J., Pollack, I., Ryerson, T., St. Clair, J. M., Teng, A., Travis, K. R., Ullmann, K., Wennberg, P. O., and Wisthaler, A.: Quantifying sources and sinks of reactive gases in the lower atmosphere using airborne ﬂux observations, Geophys. Res. Lett., 42, 8231–8240, 2015. Turnbull, J. C., Karion, A., Fischer, M. L., Faloona, I., Guilder- son, T., Lehman, S. J., Miller, B. R., Miller, J. B., Montzka, S., Sherwood, T., Saripalli, S., Sweeney, C., and Tans, P. P.: Assess- ment of fossil fuel carbon dioxide and other anthropogenic trace gas emissions from airborne measurements over Sacramento, California in spring 2009, Atmos. Chem. Phys., 11, 705–721, https://doi.org/10.5194/acp-11-705-2011, 2011. I., Ryerson, T., St. Clair, J. M., Teng, A., Travis, K. R., Ullmann, Wolfe, G. M., Kawa, S. R., Hanisco, T. F., Hannun, R. A., New- man, P. A., Swanson, A., Bailey, S., Barrick, J., Thornhill, K. L., Diskin, G., DiGangi, J., Nowak, J. B., Sorenson, C., Bland, G., Yungel, J. K., and Swenson, C. A.: The NASA Carbon Airborne Flux Experiment (CARAFE): instrumenta- tion and methodology, Atmos. Meas. Tech., 11, 1757–1776, https://doi.org/10.5194/amt-11-1757-2018, 2018. U.S. Environmental Protection Agency (EPA): National Emissions Inventory (NEI) 2017, U.S. Environmental Protection Agency [data set], https://www.epa.gov/air-emissions-inventories/ 2017-national-emissions-inventory-nei-data, last access: 24 August 2022. Van Damme, M., Clarisse, L., Dammers, E., Liu, X., Nowak, J. B., Clerbaux, C., Flechard, C. R., Galy-Lacaux, C., Xu, W., Neu- man, J. A., Tang, Y. S., Sutton, M. A., Erisman, J. W., and Coheur, P. F.: Towards validation of ammonia (NH3) measure- ments from the IASI satellite, Atmos. Meas. Tech., 8, 1575– 1591, https://doi.org/10.5194/amt-8-1575-2015, 2015. Wu, K., Zhu, S., Liu, Y., Wang, H., Yang, X., Liu, L., Dabdub, D., and Cappa, C. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia D.: Modeling Ammonia and Its Uptake by Sec- ondary Organic Aerosol Over China, J. Geophys. Res.-Atmos., 126, e2020JD034109, https://doi.org/10.1029/2020JD034109, 2021. Yamazaki, D., Ikeshima, D., Tawatari, R., Yamaguchi, T., O’Loughlin, F., Neal, J. C., Sampson, C. C., Kanae, S., and Bates, P. D.: A high-accuracy map of global terrain elevations, Geophys. Res. Lett., 44, 5844–5853, 2017. Van Damme, M., Clarisse, L., Whitburn, S., Hadji-Lazaro, J., Hurt- mans, D., Clerbaux, C., and Coheur, P.-F.: Industrial and agri- cultural ammonia point sources exposed, Nature, 564, 99–103, 2018. You, Y., Kanawade, V. P., de Gouw, J. A., Guenther, A. B., Madronich, S., Sierra-Hernández, M. R., Lawler, M., Smith, J. N., Takahama, S., Ruggeri, G., Koss, A., Olson, K., Baumann, K., Weber, R. J., Nenes, A., Guo, H., Edgerton, E. S., Porcelli, L., Brune, W. H., Goldstein, A. H., and Lee, S.-H.: Atmospheric amines and ammonia measured with a chemical ionization mass spectrometer (CIMS), Atmos. Chem. Phys., 14, 12181–12194, https://doi.org/10.5194/acp-14-12181-2014, 2014. Vaughan, A. R., Lee, J. D., Shaw, M. D., Misztal, P. K., Metzger, S., Vieno, M., Davison, B., Karl, T. G., Carpenter, L. J., Lewis, A. C., Purvis, R. M., Goldstein, A. H., and Hewitt, C. N.: VOC emission rates over London and South East England obtained by airborne eddy covariance, Faraday Discuss., 200, 599–620, 2017. Vinuesa, J.-F. and Arellano, J. V.-G. D.: Fluxes and (co-)variances of reacting scalars in the convective boundary layer, Tellus B, 55, 935–949, 2011. von Bobrutzki, K., Braban, C. F., Famulari, D., Jones, S. K., Black- all, T., Smith, T. E. L., Blom, M., Coe, H., Gallagher, M., Gha- laieny, M., McGillen, M. R., Percival, C. J., Whitehead, J. D., El- lis, R., Murphy, J., Mohacsi, A., Pogany, A., Junninen, H., Ranta- nen, S., Sutton, M. A., and Nemitz, E.: Field inter-comparison of eleven atmospheric ammonia measurement techniques, Atmos. Meas. Tech., 3, 91–112, https://doi.org/10.5194/amt-3-91-2010, 2010. Yuan, B., Kaser, L., Karl, T., Graus, M., Peischl, J., Campos, T. L., Shertz, S., Apel, E. C., Hornbrook, R. S., Hills, A., Gilman, J. B., Lerner, B. M., Warneke, C., Flocke, F. M., Ryerson, T. B., Guen- ther, A. B., and de Gouw, J. A.: Airborne ﬂux measurements of methane and volatile organic compounds over the Haynesville and Marcellus shale gas production regions, J. Geophys. Res.- Atmos., 120, 6271–6289, 2015. Zhao, H., Unhannanant, P., Hanshaw, W., and Chickos, J. S.: En- thalpies of Vaporization and Vapor Pressures of Some Deuterated Hydrocarbons. S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Forest Meteorol., 213, 193–202, 2015. Ritter, J. A., Lenschow, D. H., Barrick, J. D. W., Gregory, G. L., Sachse, G. W., Hill, G. F., and Woerner, M. A.: Airborne ﬂux measurements and budget estimates of trace species over the Amazon Basin during the GTE/ABLE 2B expedition, J. Geo- phys. Res.-Atmos., 95, 16875–16886, 1990. Sun, K., Tao, L., Miller, D. J., Pan, D., Golston, L. M., Zondlo, M. A., Grifﬁn, R. J., Wallace, H. W., Leong, Y. J., Yang, M. M., Zhang, Y., Mauzerall, D. L., and Zhu, T.: Vehicle Emissions as an Important Urban Ammonia Source in the United States and China, Environ. Sci. Technol., 51, 2472–2481, 2017. Ritter, J. A., Barrick, J. D. W., Sachse, G. W., Gregory, G. L., Wo- erner, M. A., Watson, C. E., Hill, G. F., and Collins Jr., J. E.: Airborne ﬂux measurements of trace species in an Arctic bound- ary layer, J. Geophys. Res.-Atmos., 97, 16601–16625, 1992. Tang, Y. S., Flechard, C. R., Dämmgen, U., Vidic, S., Djuri- cic, V., Mitosinkova, M., Uggerud, H. T., Sanz, M. J., Sim- mons, I., Dragosits, U., Nemitz, E., Twigg, M., van Dijk, N., Fauvel, Y., Sanz, F., Ferm, M., Perrino, C., Catrambone, M., Leaver, D., Braban, C. F., Cape, J. N., Heal, M. R., and Sut- ton, M. A.: Pan-European rural monitoring network shows dominance of NH3 gas and NH4NO3 aerosol in inorganic at- Ritter, J. A., Barrick, J. D. W., Watson, C. E., Sachse, G. W., Gre- gory, G. L., Anderson, B. E., Woerner, M. A., and Collins Jr., J. E.: Airborne boundary layer ﬂux measurements of trace species over Canadian boreal forest and northern wetland regions, J. Geophys. Res.-Atmos., 99, 1671–1685, 1994. Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023 S. Schobesberger et al.: Airborne ﬂux measurements of ammonia Liquid-Vapor Pressure Isotope Effects, J. Chem. Eng. Data, 53, 1545–1556, 2008. Vonk, J., Bannink, A., van Bruggen, C., Groenestein, C. M., Hui- jsmans, J. F. M., van der Kolk, J. W. H., Luesink, H. H., Oude Voshaar, S. V., van der Sluis, S. M., and Velthof, G. L.: Method- ology for estimating emissions from agriculture in the Nether- lands, Statutory Research Tasks Unit for Nature & the Environ- ment, https://doi.org/10.18174/383679, 2016. Zheng, J., Ma, Y., Chen, M., Zhang, Q., Wang, L., Khalizov, A. F., Yao, L., Wang, Z., Wang, X., and Chen, L.: Measurement of at- mospheric amines and ammonia using the high resolution time- of-ﬂight chemical ionization mass spectrometry, Atmos. Envi- ron., 102, 249–259, 2015. Walker, J. T., Bell, M. D., Schwede, D., Cole, A., Beachley, G., Lear, G., and Wu, Z.: Aspects of uncertainty in total reactive ni- trogen deposition estimates for North American critical load ap- plications, Sci. Total Environ., 690, 1005–1018, 2019. Zhu, S., Wu, K., Nizkorodov, S. A., and Dabdub, D.: Modeling Reactive Ammonia Uptake by Secondary Organic Aerosol in a Changing Climate: A WRF-CMAQ Evaluation, Front. Environ. Sci., 10, 867908, https://doi.org/10.3389/fenvs.2022.867908, 2022. Weil, J. C. and Horst, T. W.: Footprint estimates for atmospheric ﬂux measurements in the convective boundary layer, in: Precipi- tation Scavenging and Atmosphere-Surface Exchange, edited by: Schwartz, S. and Slinn, W. G. N., Hemisphere Publishing Corpo- ration, London, 717–728, ISBN 1-56032-263-2, 1992. Atmos. Meas. Tech., 16, 247–271, 2023 https://doi.org/10.5194/amt-16-247-2023
https://openalex.org/W2972774775	https://bmcpregnancychildbirth.biomedcentral.com/track/pdf/10.1186/s12884-019-2492-3	English	null	Prevalence of gestational diabetes mellitus and associated factors among women attending antenatal care at Gondar town public health facilities, Northwest Ethiopia	BMC pregnancy and childbirth	2,019	cc-by	10,317	Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 https://doi.org/10.1186/s12884-019-2492-3 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 https://doi.org/10.1186/s12884-019-2492-3 (2019) 19:334 Muche et al. BMC Pregnancy and Childbirth https://doi.org/10.1186/s12884-019-2492-3 Open Access Prevalence of gestational diabetes mellitus and associated factors among women attending antenatal care at Gondar town public health facilities, Northwest Ethiopia Prevalence of gestational diabetes mellitus and associated factors among women attending antenatal care at Gondar town public health facilities, Northwest Ethiopia Achenef Asmamaw Muche1,2, Oladapo O. Olayemi3 and Yigzaw Kebede Gete2 Correspondence: ashua2014@gmail.com * Correspondence: ashua2014@gmail.com 1Pan African University Life and Earth Sciences Institute (including health and agriculture), Department of Obstetrics and Gynaecology, College of Medicine, University of Ibadan, Ibadan, Nigeria 2Department of Epidemiology and Biostatistics, Institute of Public Health, University of Gondar, Gondar, Ethiopia Full list of author information is available at the end of the article Abstract Background: Globally, Gestational Diabetes Mellitus (GDM) is rising, but it is a neglected health threat to mothers and their children in low resource countries. Although, GDM is known in Ethiopia, information regarding it remains scarce by recent diagnostic criteria. Therefore, this study aimed to determine the prevalence of GDM and associated factors among women attending antenatal care at Gondar town public health facilities, Northwest Ethiopia. Methods: A cross-sectional study was conducted among 1027 pregnant women selected by the systematic random sampling technique. The universal one-step screening and diagnostic strategy was done using a two-hour 75 g oral glucose tolerance test. GDM was diagnosed using updated diagnostic criteria (2017 American Diabetes Association (ADA) or 2013 World Health Organization (WHO) or modified International Association of the Diabetes and Pregnancy Study Groups diagnostic criteria (IADPSG)). Binary logistic regression model was used to identify factors associated with GDM. Results: Of the total 1027 pregnant women, 12.8% (95% CI: 10.8–14.8) were diagnosed with GDM. Overweight and/or obesity (MUAC ≥28 cm) (AOR = 2.25, 95% CI: 1.18–4.26), previous history of GDM (AOR = 5.82, 95% CI: 2.57–13.18), family history of diabetes (AOR = 4.03, 95% CI: 1.57–10.35), low physical activity (AOR = 3.36, 95% CI: 1.60–7.04), inadequate dietary diversity (AOR = 1.9, 95% CI: 1.02–3.53), and antenatal depression (AOR = 4.12, 95% CI: 1.85–9.20) were significantly associated with GDM. Conclusions: The prevalence of GDM among women attending antenatal care at Gondar town public health facilities was high. Previous history of GDM, antenatal depression, family history of diabetes, low physical activity, overweight and/ or obesity and inadequate dietary diversity were significantly associated with GDM. Routine screening of pregnant women and healthy lifestyle are strongly recommended. Keywords: Gestational diabetes mellitus, Determinants, Overweight, Obesity, Physical activity, Dietary diversity, Antenatal depression Keywords: Gestational diabetes mellitus, Determinants, Overweight, Obesity, Physical activity, Dietary diversity, Antenatal depression Background for GDM screening in the country, most clinicians focus on risk factors to indicate GDM screening. This approach was leaving many pregnant women un- noticed until they develop symptoms of overt diabetes and complications. The World Health Organization (WHO) defined “Gesta- tional Diabetes Mellitus (GDM) as glucose intolerance first detected during pregnancy” [1]. Various adverse maternal and neonatal outcomes were complicated by GDM [2], while its complex care requires risk reduc- tion strategies beyond the control of blood glucose level [3]. Recognizing GDM offers an opportunity to reduce adverse pregnancy outcomes and improve the lifestyle to prevent the occurrence of diabetes in the future [23, 24]. Moreover, as per the knowledge of the principal in- vestigator, no study has been conducted on prevalence of GDM in Ethiopia using the updated international diagnostic criteria. Therefore, the present study aimed to determine the prevalence of GDM and associated factors among women attending antenatal care at se- lected Gondar town public health facilities, Northwest Ethiopia. Globally, GDM affects an estimated 15% of the preg- nant women, 87.6% of the hyperglycemia were in low and middle-income countries. It is one of the challenging health problems of sub-Saharan African countries [4]. p A review indicated that the occurrence of GDM in sub-Saharan Africa was 14% [5] and Middle East and North Africa ranged from 8.4 to 24.5% [6] though the study used different screening and diagnostic criteria. Research findings also showed that the prevalence of GDM varied to a certain extent among regions in Africa. For example, East [7] and West [8] Africa reported 6 and 14%, respectively. Variations were also noted within sub-regions, like Rwanda [9] and Tanzania [10], where the prevalence was 8.3 and 19.5%, respectively. Two decades ago, the prevalence of GDM in the rural area of North Ethiopia was reported as 3.7% [11]. Additionally, a survey done in the same region (North Ethiopia) found a prevalence of 13% among urban women which was higher than that of women in rural areas (5%) [12]. However, the study used only fasting blood glucose test as diagnostic criteria for GDM. © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted 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. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Page 2 of 13 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 Materials and methods Study setting The study was conducted at selected Gondar town pub- lic health facilities. It included one tertiary facility called University of Gondar Comprehensive Specialized Hos- pital (UOGCSH) and four health centers namely, Azezo, Gondar polyclinic, Woleka, and Maraki. Gondar town is located in the Northwestern part of Ethiopia, 747 km from Addis Ababa (the capital of Ethiopia), and 170 km from Bahirdar (the regional capital). It has 12 sub-cities with 12 urban and 10 rural kebeles (the smallest admin- istrative units). Based on the 2014 population projection and census of the Central Statistical Agency (CSA) of Ethiopia [25], the total population of the town was esti- mated to be 306, 246 of whom 149, 970 were men and 156, 276 women. The majority (84.2%) of the people were Orthodox Christians, 11.8% Muslims and 1.1% Protestants. The expected number of pregnant women in the town was 11,225, of which at least 8913 were living in urban kebeles in 2017/18. The town had one comprehensive specialized hospital, eight health centers, and more than 15 private clinics. According several studies, the increasing occurrence of GDM was related to advanced age, family history of diabetes, inactive physical activity, obesity, and risky be- haviors [13–15]. Studies had also recognized that there was association between dietary habits during pregnancy and GDM. However, there had been no concrete con- sensus on the effects specific dietary aspects and the risk for GDM [16–18]. Gestational diabetes mellitus commonly identified dur- ing the second or third trimester of pregnancy as a result of the placental hormone plays an important role in the adverse effect on glucose metabolism [2]. As pregnancy progresses, various hormones such as estrogen, proges- terone, leptin, cortisol, placental lactogen, and placental growth hormone promote a state of insulin resistance [19]. Primarily, human placental lactogen produced by placenta raises maternal blood glucose level and makes a woman’s body less sensitive to insulin leading to a higher-than-normal blood glucose level and perhaps GDM [20]. Participant selection and recruitment The short form International Physical Activity Ques- tionnaire (IPAQ) was used to assessed the physical activ- ities that women do as part of their everyday lives [30]. The IPAQ was suitable for adults between 15 and 69 years of age and implemented in different countries. It was designed to assess specific types of activity such as walking, moderate and vigorous intensity activities done at work, as part of house and yard work, to get place to place, and in spare time for recreation, exercise or sport. Women were asked to recall their activities of the last 7 days preceding the interview. Data was reported as metabolic equivalents (MET-minutes per week) using the IPAQ scoring protocol to considered women into high, moderate and low level of physical activity categor- ies [30]. During the study, 1110 new antenatal care attendants were invited and screened to participate in the study. Women were enrolled if they were aged 18 years or more with 20–23+ 6 weeks of gestational age and single- ton pregnancy. But pregnant women who had pre-exist- ing diabetes mellitus or overt DM, chronic diseases, medications that may affect glucose metabolism such as steroids, β-adrenergic agonists, anti-psychotic drugs [26, 27] were excluded. On their first visit, participants were asked for written consent for their enrolment in the study. Screening at first visit was carried out according to the recommendations of the International Diabetes Feder- ation (IDF) [22], ADA [28], and WHO [29] in order to rule out pre-existing diabetes. However, most pregnant women did not fast in their first visits. For subsequent tests, participants were informed to come fasting on their next appointments. Universal screening for GDM using a two-hour 75 g OGTT was performed for all pregnant women at 24–28 weeks of gestational age. Be- sides, 352(34.3%) women had at least one type of risk factors for GDM (pre pregnancy BMI ≥30 Kg/m2, MUAC ≥28 cm, age ≥35 years, previous macrosomia, glycosuria, history of GDM, family history of diabetes, previous poor pregnancy outcome or developed preg- nancy-related complications) were advised to repeat the test at 32–36 weeks even if their OGTT results were negative at regular tests and GDM diagnosis ascertained by the second test. Dietary diversity was assessed using a 24-h food recall method by the Food and Nutrition Technical Assistance (FANTA) 2016 version woman’s minimum dietary diver- sity measurement tool [31]. Applying the formula Applying the formula Demographic, obstetric, lifestyle, and anthropometric information Demographic, obstetric, lifestyle, and anthropometric information n ¼ z ð Þ2P 1‐P ð Þ d2 n ¼ 1:96 ð Þ2:13 1−:13 ð Þ 0:03 ð Þ2 ¼ 482:7 n ¼ 482:7x2 design effect ð Þ ¼ 965 n ¼ z ð Þ2P 1‐P ð Þ d2 n ¼ 1:96 ð Þ2:13 1−:13 ð Þ 0:03 ð Þ2 ¼ 482:7 n ¼ 482:7x2 design effect ð Þ ¼ 965 A structured and pretested questionnaire was developed and interviewer-administered to all study participants The questionnaire was prepared in English and then translated to Amharic (the local and national language) and then translated back to English by language experts to check its consistency (English version questionnaire attached as Additional file 1). Where, n = the sample size, Z = the desired level of the confidence interval, P = proportion of GDM, and d = margin of error. The minimum required sample size for significant result was 965, however we have added 15% by considered the non-response and attrition rates throughout the study period. A face to face interview was employed to gather infor- mation on residence, age, last normal menstrual period (LNMP), marital status, religion, ethnicity, level of edu- cation, occupational status, average monthly income, family history of diabetes, previous history of GDM and birth weight of the previous child, behavioral and life- style characteristics, such as current exposure to alcohol, coffee, antenatal depressive symptoms, physical activity, and dietary diversity. Gestational age estimations were based on a reliable LNMP combined with first-trimester ultrasonography (if available). Additionally, socio-demo- graphic information (parity, gravidity, gestational age (early fetal ultrasound result), obstetric history, medical history, hemoglobin) were retrieved from ANC cards. Therefore a total of 1110 were recruited for the study. A two-stage sampling technique used to invite pregnant mothers and include them in the cohort. On the first stage, one hospital and four health centers were selected by the simple random sample; on the second stage, pregnant women who fulfilled the inclusion criteria were chosen using the systematic random sampling technique. Study design, sample size, and sampling procedure Study design, sample size, and sampling procedure A cross-sectional study was conducted among a cohort of pregnant women receiving antenatal care at selected public health facilities from March 30, 2018 to January 4, 2019. The required sample size was determined using the single population proportion formula with the fol- lowing statistical assumptions: a survey conducted by Management Sciences for Health (MSF) reported the prevalence of GDM among urban women in Tigray, Northern Ethiopia as 13% [12], a 95% confidence level, 3% marginal of error, a design effect of 2, and consider- ing 15% of non-response and attrition rates. After the pre-existing diabetes (overt diabetes) is ruled out at the first antenatal visit, the recommended oral glucose tolerance test (OGTT) has to be performed at 24–28 weeks of pregnancy [21, 22]. However, since there has been no well-established national guideline Page 3 of 13 Page 3 of 13 Page 3 of 13 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 (2019) 19:334 Participant selection and recruitment It contained a list of ten food groups (starchy staples, nuts and seeds, pulses, dairy, meat, eggs, poultry and fish, dark green leafy vege- tables, other vitamin-A rich fruits and vegetables, other vegetables, and other fruits). The minimum dietary di- versity score (MDDS) was dichotomized on the basis of whether or not women have consumed the list of defined food groups the previous day or night. The MDDS of five and more was categorized as adequate dietary diversity [31]. A mother was asked, “How often have you drunk cof- fee since your pregnancy?” If the answer was “daily” or Page 4 of 13 Page 4 of 13 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 the cuvette holder. After 40–240 s the glucose value of the sample was displayed. The necessary precautions were taken during sample collection. Then, they were given a 75 g of glucose dissolved in 250 mL of water which they drank within 5 min under supervision. Capillary blood samples were taken again at 1 h and 2 h. The whole blood capillary values were converted to plasma venous values by multiplying a constant factor of 1.11 [40]. GDM diag- nosis was made by using the 2017 American Diabetes Association (ADA) [28] or 2013 WHO [41] or modified International Association of the Diabetes and Pregnancy Study Groups (IADPSG) [42] diagnostic criteria. Similar values have been accepted by the WHO though slightly modified as a range [29]. The diagnosis of GDM is made when one or more of the values was met (fasting: ≥92 mg/dL, 1 h: ≥180 mg/dL; 2 h: ≥153 mg/dL). Hemoglobin analysis was also carried out by laboratory technologists in their respective health facilities. Pregnant women with hemoglobin concentration below 11 g/dl were considered as having anaemia [43]. “sometimes in a week” the mother was categorized as exposed to coffee. She was also asked, “How often have you drunk alcohol since your pregnancy?” If the answer was “daily” or “sometimes in a week” the mother was labelled as exposed to alcohol. The mid-upper arm circumference (MUAC) was mea- sured on the left arm using a non-stretchable measuring tape. As the result of most of the pregnant women could not recall their weight before conception, it was difficult to determine BMI. Laboratory assessment Pregnant women scheduled for 2 h-75 g OGTT were called and reminded to complete an overnight fast for 8–12 h. Blood glucose level was analyzed using 5 μl capillary whole blood with HemoCue Glucose B-201+ (A¨ngelholm AB, Sweden). (The specifications of the HemoCue Glucose B-201+ glucometer attached as Additional file 2). Pregnant women were asked to sit and take rest before the first sample was taken using a finger prick at the side of the fingertip with a sterile lancet after cleaning in disinfectant and allowed to dry. A drop of blood was allowed to fill the cuvette after wiping away the first 2 or 3 drops of blood. Wipe off excess blood on the outside of the cuvette tip and place the filled cuvette in Statistical analysis ll d y All data were entered into Epi Info™7 software, then exported to statistical package for the social sciences (SPSS) version 22.0 (IBM, NY, USA) for analysis. Descriptive statistics, like frequency, percentage, and mean, SD, and the range were used for the presentation of variables. Tables and figure were also used for data presentation. Categorical variables were expressed as proportions and chi-square analysis was performed to compare proportions. Student t-test was used to com- pare means for normally distributed variables. Binary logistic regression model was used to identify factors associated with GDM. Variables with P-value of ≤0.20 in the bivariate analysis were exported to the multivariate analysis to control the possible effect of confounders. Model goodness of fit test was checked by the Hosmer- and Lemeshow test (P-value = 0.953). Multicollinearity was checked by using the variance inflation factor (VIF). Adjusted Odds Ratio (AOR) with a 95% confidence level was estimated to show the strength of association, and a P-Value < 0.05 was used to declare statistical significance in the multivariate analysis. Antenatal depression symptoms were measured by using the Edinburgh Postnatal Depression Scale (EPDS) screening tool developed [35] and validated in urban Ethiopia [36]. The tool was used to measure the feelings a mother had experienced in the past week. The tool contains ten specific questions with four Likert scale re- sponse options (most of the time, sometimes, not often, never), scored from 0 to 3 (a higher score indicated more depressive symptoms), which is simple to use, can be scored by simple addition. An EPDS score of 13 and more used like similar studies conducted in Ethiopia and abroad [37–39] to categorize the presence of antenatal depression. Participant selection and recruitment MUAC was as reliable measure due to quite stable during the course of pregnancy and highly correlated to the BMI before conception [32, 33]. Pregnant women with MUAC of ≥28 cm were consid- ered as having overweight and/or obesity [34]. For measuring blood pressure (BP), the pregnant women were asked to take rest for at least 5 min in sitting positions if they were exerted. Then, the pressure was measured on the right arm using normal cuffs fitted for adults with a standard sphygmomanometer placing the stethoscope bell lightly over the brachial artery. The average systolic blood pressure (SBP) and diastolic blood pressure (DBP) were recorded in mmHg after two read- ings were taken with an intermission of 5–10 min. Hypertension was considered to be present if the systolic and diastolic blood pressures were greater than or equal to 140 mmHg and 90 mmHg, respectively. Socio demographic and clinical characteristics of pregnant mothers Out of the 1110 pregnant women invited to participate in the study, 83 mothers (52 did not return for OGTT, 15 did not complete the tests, 9 were diagnosed with overt diabetes, 5 have a medical emergency, and 2 had an abortion before OGTT) were excluded. The charac- teristics of pregnant women who of included and excluded in the final analysis were similar in terms of Page 5 of 13 Page 5 of 13 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 key socio demographic variables such as residence (P = 0.835), maternal age (P = 0.0691), parity (P = 0.544), gra- vidity (P = 0.109), marital status (P = 0.734), MUAC (P = 0.08), employment status (P = 0.056). However women with complete data were on average attended secondary educational level and above and had higher income level than those who were excluded in the analysis. 119.10 (SD ± 19.75) mg/dl, respectively. The plasma glu- cose levels taken at the three time points of the OGTT were significantly different for women with GDM and normal glucose level (p < 0.001). Pregnant women with GDM had the mean fasting, one-hour and two-hour glucose level of 100.81 (SD ±15.88) mg/dl, 163.82 (SD ± 28.42) mg/dl, and 138.76 (SD ±23.40) mg/dl, respect- ively. Whereas, women with normal glucose profiles had the mean fasting, one-hour and two-hour glucose levels of 78.3 (SD ± 9.55) mg/dl, 137.31 (SD ±22) mg/dl, and 116.22 (SD ±17.4) mg/dl, respectively (Table 1). More than half (56.5%) of the women with GDM were diag- nosed on a fasting plasma glucose status only, while 21.4% had high fasting glucose level and one or two extra abnormal levels (Fig. 1). y The remaining 1027 women who participated and completed OGTT with overall response rate of 92.5% were included in the analysis. The mean age of the women was 27.22 (SD ± 5.24) years. Most of the partici- pants (91%) lived in urban areas, 92% were married, and 60.5% unemployed. Five hundred and ninety-eight (58.2%) of women had secondary education and above. The mean of the MUAC was 24.77 (SD ± 3.08) cm, sys- tolic blood pressure was 112.07 (SD ± 11.91) mmHg, and diastolic blood pressure was 71.91 (SD ± 10.22) mmHg. Factors associated with GDM Factors associated with GDM Results of the unadjusted binary logistic regression showed that advanced maternal age, marital status, level of education, employment status, MUAC ≥28 cm, parity, previous history of GDM, family history of DM, history of having macrocosmic baby, history of abortion, history of stillbirth, SBP, DBP, anemia (Hb < 11 g/dl), level of physical activity, dietary diversity, and antenatal depres- sion were associated with GDM. However, on multivari- ate logistic regression, variables GDM was independently associated with being overweight and/or obese women (MUAC ≥28 cm) AOR = 2.25; 95% CI: 1.18–4.26), previ- ous history of GDM (AOR = 5.82; 95% CI: 2.57–13.18), family history of diabetes (AOR = 4.03; 95% CI: 1.57– 10.35), low physical activity (AOR = 3.36; 95% CI: 1.60– 7.04), inadequate dietary diversity (AOR = 1.9; 95% CI: 1.02–3.53), and antenatal depression (AOR = 4.12; 95% CI: 1.85–9.20) were significantly associated with GDM (Table 4). Behavioral and lifestyle characteristics of pregnant mothers Out of total participants, a moderate level of physical activity was reported by 44.8%. Normal glucose level was more common among women having high physical activity level (33.4% vs 17.6%) and adequate dietary diversity (51.5% vs 27.5%) than women with GDM. Con- cerning exposure to drinking alcohol and coffee, no statistically significant difference was observed among women with normal glucose profile and GDM at P = 0.476 and P = 0.311 respectively. Eighty-nine (8.7%) of the pregnant women had antenatal depression symp- toms (Table 2). Obstetric history of pregnant mothers This study was conducted to determine the prevalence of GDM and associated factors among women attending antenatal care at Gondar town public health facilities, Northwest Ethiopia, using the updated international diagnostic criteria. Of the total of 588 pregnant women who had prior preg- nancy history, 71(12.1%) had macrocosmic babies, 166 (28.2%) history of abortion, 38 (6.5%) history of stillbirth, 42 (7.1%) history of preterm labor, 97(16.5%) history of caesarean delivery, and 54 (9.2%) previous GDM (Table 3). In this study, the overall prevalence of GDM among women was 12.8% (95% CI: 10.8–14.8). The result was consistent with the Management Sciences for Health (MSF) reported the prevalence of GDM among urban women in Tigray, Northern Ethiopia (13%), but the study used only a fasting blood glucose test which means it did not consider internationally recommended diag- nostic criteria [12]. The finding was higher than those of studies conducted in Rwanda (8.3%) [9], Tanzania (5.9%) [7], Egypt (8%) [44], and Nigeria (8.6%) [45]. The main reason for the high prevalence of GDM in this study setting might be the fact that the lower cut-off points for Socio demographic and clinical characteristics of pregnant mothers The mean hemoglobin level of the pregnant women with GDM and normal glucose level was 12.39 (SD ± 1.80) g/ dl and 12.64 (SD ± 1.74) g/dl, respectively. Nearly half, 504 (49.1%) of the women were nulliparous. Sixty-two (6%) had a family history of DM (Table 1). Prevalence of GDM A total of 131 women were diagnosed as having GDM according to 2017 ADA and IADPSG diagnostic criteria which resulted in a GDM prevalence of 12.8% (95% CI: 10.7–14.8%); 118 (90%) of them were diagnosed at regular OGTT and the rest 13 (10%) at late OGTT. At regular OGTT test (24–28 weeks), the mean fasting, 1 h, and 2-h glucose levels of the pregnant women were 81.17 (SD ± 12.96) mg/dl, 140.69 (SD ±24.55) mg/dl, and Muche et al. Prevalence of GDM BMC Pregnancy and Childbirth (2019) 19:334 Page 7 of 13 Page 7 of 13 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 Table 1 Socio demographic and clinical characteristics of the study participants attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 (N = 1027) (Continued) Variables n (%) or mean ± SD t/× 2 P value Total participants (n = 1027) Normal (n = 896) GDM (n = 131) 2-h blood glucose (OGTT) (mg/dl) 119.10 ± 19.75 116.22 ± 17.40 138.76 ± 23.40 13.188 < 0.001 Hemoglobin (g/dl)c 12.609 ± 1.75 12.64 ± 1.74 12.39 ± 1.80 −1.545 0.123 Normal (Hb ≥11 g/dl) 874 (86.6) 769 (88.0) 105 (12.0) 1.949 0.163 Anemia (Hb < 11 g/dl) 135 (13.4) 113 (83.7) 22 (16.3) GDM gestational diabetes mellitus, h hour, t student t value, x2 chi square, Hb hemoglobin, g/dl gram per deciliter, mg/dl milligram deciliter, OGTT oral glucose tolerance test, MUAC Mid upper arm circumference aDivorced or widowed bBlood glucose value at regular time (24–28 weeks) c18 participants were missed (N = 1009) Table 1 Socio demographic and clinical characteristics of the study participants attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 (N = 1027) (Continued) Table 1 Socio demographic and clinical characteristics of the study participants attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 (N = 1027) (Continued) Ghana and South Africa which considered obesity using BMI noted its significant association with GDM [14, 15, 48]. Similarly, a review and meta-analysis by Nelson SM et al [49]. revealed that pre-pregnancy BMI was more strongly associated with the risk of GDM. This was happened by the fact that the decreased insulin sensitiv- ity in obese pregnancies, resulted increases the blood glucose level [49, 50]. Overweight or obesity exposed to sedentary life led to obesity due to inactive lifestyle. This cycle also adversely affects the glucose metabolism. Like- wise, we expect a further rise in GDM in the coming years as obesity is emerging as a public health problem. FPG and OGTT were used in the updated diagnostic criteria. On the other hand, the finding is lower than those of studies conducted in Tanzania (19.5%) [10] and South Africa (25.8%) [46] which used similar diagnostic criteria. GDM Gestational Diabetes Mellitus, x2 chi square Prevalence of GDM This evidence indicates that the prevalence of GDM might also be affected not only by different diag- nostic criteria but also by the characteristics of the population [2, 45, 46]. Increased testing for GDM, change in lifestyle, and the rising prevalence of over- weight and obesity might have contributed [47]. FPG and OGTT were used in the updated diagnostic criteria. On the other hand, the finding is lower than those of studies conducted in Tanzania (19.5%) [10] and South Africa (25.8%) [46] which used similar diagnostic criteria. This evidence indicates that the prevalence of GDM might also be affected not only by different diag- nostic criteria but also by the characteristics of the population [2, 45, 46]. Increased testing for GDM, change in lifestyle, and the rising prevalence of over- weight and obesity might have contributed [47]. Women with the mid-upper arm circumference of ≥28 cm were two times more likely to develop GDM than women with MUAC < 28 cm. This finding agreed with findings from Tanzania by Mwanri et al. [7] and Njete et al [10]. .Although, pre-gestational weight was not available in this study, different studies in Egypt, y y g g p p Pregnant women with a previous history of GDM had six fold more increased odds of developing of the prob- lem during the pregnancy at the moment. Prevalence of GDM BMC Pregnancy and Childbirth (2019) 19:334 Page 6 of 13 Table 1 Socio demographic and clinical characteristics of the study participants attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 (N = 1027) Table 1 Socio demographic and clinical characteristics of the study participants attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 (N = 1027) Variables n (%) or mean ± SD t/× 2 P value Total participants (n = 1027) Normal (n = 896) GDM (n = 131) Maternal age (years) 27.22 ± 5.241 26.69 ± 5.047 30.86 ± 5.109 8.827 < 0.001 < 25 323 (31.5) 304 (94.1) 19 (5.9) 61.108 < 0.001 25–29 366 (35.6) 335 (91.5) 31 (8.5) 30–34 214 (20.8) 168 (78.5) 46 (21.5) ≥35 124 (12.1) 89 (71.8) 35 (28.2) Marital status Married 945 (92) 831(87.9) 114 (12.1) 5.094 0.024 Single and othersa 82 (8.0) 65 (79.3) 17 (20.7) Educational level Not formal education 200 (19.5) 168 (84) 32 (16.0)) 5.472 0.065 Primary education 229 (22.3) 194 (84.7) 35 (15.3) Secondary education and above 598 (58.2) 534 (89.3) 64 (10.7) Employment status Unemployed 621 (60.5) 556 (89.5) 65 (10.5) 7.393 0.007 Employed 406 (39.5) 340 (83.7) 66 (16.3) Residence Urban 935 (91) 812 (86.8) 123 (13.2) 1.497 0.221 Rural 92 (9.0) 84 (91.3) 8 (8.7) Monthly income (birr) 3159.7 ± 2973.05 3155.17 ± 2952.66 3190.8 ± 3120.55 0.128 0.898 < 1500 236 (23) 203 (86) 33 (14) 1.641 0.650 1500–2499 257 (25) 230 (89.5) 27 (10.5) 2500–3999 230 (22.4) 199 (86.5) 31 (13.5) ≥4000 304 (29.6) 264(86.8) 40 (13.2) Family history of DM No 965 (94) 862 (89.3) 103 (10.7) 62.264 < 0.001 Yes 62 (6) 34 (54.8) 28 (45.2) Parity 0.92 ± 1.208 0.91 ± 1.227 1.02 ± 1.063 1.013 0.311 Nullipara 504 (49.1) 449 (89.1) 55 (10.9) 5.717 0.057 Primipara 277 (27) 243 (87.7) 34 (12.3) Multipara 246 (24) 204 (82.9) 42 (17.1) Gravidity 2.12 ± 1.354 2.09 ± 1.364 2.32 ± 1.267 1.838 0.066 Primigravida 439 (42.7) 395 (90) 44 (10) 5.146 0.023 Multigravida 588 (57.3) 501 (85.2) 87 (14.8) MUAC 24.77 ± 3.08 24.52 ± 2.837 26.47 ± 4.014 6.93 < 0.001 MUAC < 28 cm 852 (83) 767 (90) 85 (10) 34.700 < 0.001 MUAC ≥28 cm 175 (17) 129 (73.7) 46 (26.3) Blood pressure Systolic blood pressure (mmHg) 112.07 ± 11.91 111.57 ± 11.31 115.46 ± 15.01 3.508 < 0.001 Diastolic blood pressure (mmHg) 71.91 ± 10.22 71.60 ± 9.88 74.05 ± 12.09 2.565 0.010 Blood glucose levelb Fasting blood glucose (mg/dL) 81.17 ± 12.960 78.3 ± 9.55 100.81 ± 15.88 22.779 < 0.001 1-h blood glucose (OGTT) (mg/dl) 140.69 ± 24.55 137.31 ± 22.0 163.82 ± 28.42 12.368 < 0.001 Muche et al. Prevalence of GDM This finding was in line with those of studies in Egypt [48], Nigeria Table 2 Behavioral and life style characteristics of the study participants attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 (N = 1027) Variables Total participants (n = 1027) Normal (n = 896) GDM (n = 131) X2 P value Level of physical activity High 322 (31.4) 299 (92.9) 23 (7.1) 100.841 < 0.001 Moderate 460 (44.8) 429 (93.3) 31 (6.7) Low 245 (23.9) 168 (68.6) 77 (31.4) Dietary diversity status Adequate (≥5) 497 (48.4) 461 (92.8) 36 (7.2) 26.294 < 0.001 Inadequate (< 5) 530 (51.6) 435 (82.1) 95 (17.9) Alcohol intake No 574 (55.9) 497 (86.6) 77 (13.4) 0.508 0.476 Yes 453 (44.1) 339 (88.1) 54 (11.9) Coffee intake No 291 (28.3) 249 (85.6) 42 (14.4) 1.027 0.311 Yes 736 (71.7) 647 (87.9) 89 (12.1) Antenatal depression No 938 (91.3) 839 (89.4) 99 (10.6) 47.128 < 0.001 Yes 89 (8.7) 57 (64) 32 (36) GDM Gestational Diabetes Mellitus, x2 chi square style characteristics of the study participants attending antenatal care at Gondar town public health March 2018–January 2019 (N = 1027) Table 2 Behavioral and life style characteristics of the study participants attending antenatal care at Gondar facilities, Northwest Ethiopia: March 2018–January 2019 (N = 1027) Muche et al. Prevalence of GDM BMC Pregnancy and Childbirth (2019) 19:334 Page 8 of 13 Table 3 Obstetric history of the study participants attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 (N = 588) Variables Total participants (n = 588) Normal (n = 502) GDM (n = 86) X2 P value History of having macrocosmic baby No 517 (87.9) 453 (87.6) 64 (12.4) 17.309 < 0.001 Yes 71 (12.1) 49 (69) 22 (31) History of delivery by caesarean delivery No 491 (83.5) 423 (86.2) 68 (13.8) 1.437 0.231 Yes 97 (16.5) 79 (81.4) 18 (18.6) History of abortion No 422 (71.8) 368 (87.2) 54 (12.8) 4.007 0.045 Yes 166 (28.2) 134 (80.7) 32 (19.3) History of still birth No 550 (93.5) 476 (86.5) 74 (13.5) 9.351 0.002 Yes 38 (6.5) 26 (68.4) 12 (31.6) Preterm labor No 546 (92.9) 469 (85.9) 77 (14.1) 1.676 0.195 Yes 42 (7.1) 33 (78.6) 9 (21.4) Previous history GDM No 534 (90.8) 477 (89.3) 57 (10.7) 72.718 < 0.001 Yes 54 (9.2) 25 (46.3) 29 (53.7) GDM Gestational Diabetes Mellitus, x2 chi square study participants attending antenatal care at Gondar town public health facilities, Northwest GDM Gestational Diabetes Mellitus, x2 chi square [51], Colorado [52], and a systematic review by Catherine Kim et al. [53]. The frequent occurrence of GDM showed the presence of communal risk factors in succeeding pregnancies [54]. Similarly, GDM was four fold higher in women with family history of diabetes. The finding was consistent with an evidence in the US [55] and studies in Iran [56, 57]. This might be the fact that the hypergly- cemia was linked with a genetically dysfunction of beta cell and familial predisposition to insulin secretory defects [58]. Moreover, living standards and lifestyles of families are more likely similar resulting in sharing the related risk factors [59]. Fig. 1 Analysis of the abnormal blood glucose reading among pregnant women with GDM attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 (N = 131) Fig. 1 Analysis of the abnormal blood glucose reading among pregnant women with GDM attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 (N = 131) Fig. 1 Analysis of the abnormal blood glucose reading among pregnant women with GDM attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 (N = 131) Muche et al. Prevalence of GDM BMC Pregnancy and Childbirth (2019) 19:334 Page 10 of 13 Table 4 Factors associated with GDM among pregnant women attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 (Continued) Variables Non-GDM n (%) GDM n (%) COR (95% CI) P -value AOR (95% CI) a P value History of still birth No 476 (86.5) 74 (13.5) 1 1 Yes 26 (68.4) 12 (31.6) 2.97 (1.44, 6.14) 0.003 1.42 (0.48, 4.15) 0.526 Previous GDM No 477 (89.3) 57 (10.7) 1 1 Yes 25 (46.3) 29 (53.7) 9.71 (5.32, 17.71) < 0.001 5.82 (2.57, 13.18) < 0.001 Antenatal depression No 839 (89.4) 99 (10.6) 1 1 Yes 57 (64) 32 (36) 4.76 (2.94, 7.69) < 0.001 4.12 (1.85, 9.20) 0.001 OR odds ratio, CI confidence interval, MUAC mid-upper arm circumference, SBP systolic blood pressure, DBP diastolic blood pressure, 1 Reference aAdjusted for factors including maternal age, marital status, educational level, employment status, MUAC, systolic blood pressure, diastolic blood pressure, anemia, level of physical activity, dietary diversity status, family history of diabetes, parity, history of macrocosmic baby, history of abortion, history of still birth, previous GDM and antenatal depression bDivorced or widowed Table 4 Factors associated with GDM among pregnant women attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 (Continued) ciated with GDM among pregnant women attending antenatal care at Gondar town public health facilities, March 2018–January 2019 (Continued) GDM among pregnant women attending antenatal care at Gondar town public health facilities, OR odds ratio, CI confidence interval, MUAC mid-upper arm circumference, SBP systolic blood pressure, DBP diastolic blood pressure, 1 Reference aAdjusted for factors including maternal age, marital status, educational level, employment status, MUAC, systolic blood pressure, diastolic blood pressure, anemia, level of physical activity, dietary diversity status, family history of diabetes, parity, history of macrocosmic baby, history of abortion, history of still birth, previous GDM and antenatal depression bDivorced or widowed Women with antenatal depression had four times higher GDM compared to women with non-depressive symptoms. The finding is supported studies by Morrison et al. [67] showed antenatal depression as a risk factor for GDM. The finding was in line with that of a study in Chicago which showed that a history of prenatal depres- sion increased the risk for the development of GDM [68]. Prevalence of GDM The association between depression and GDM was bidirectional [69] and could be explained by shared psychosocial and physiological factors for these comor- bid situations. Pregnant women who had depression were more likely to practice unhealthy behaviors like sedentary lifestyle which lead to the risk of developing GDM [70]. Moreover, change in cortisol levels by de- pression have hyperglycemic effects [71, 72]. Low physical activity increases the likelihood of devel- oping GDM at least three times compared to a high level of physical activity during pregnancy. This finding was in line with a study conducted in Vietnam [60], a meta- analysis by Tobias et al. [61], and review by Biase et al. [59]. Increased level of physical activity during preg- nancy reduces glucose levels by preventing weight gain and enhancement of insulin sensitivity [13, 60, 62]. y Pregnant women with inadequate dietary diversity had two times higher GDM than their counterparts. Healthy eating habits and the consumption of food from a variety of dietary groups during pregnancy contributes to the prevention of pregnancy-related complications [63]. Des- pite a piece of contradict evidence regarding the impact of dietary variability during pregnancy and developing GDM, studies revealed the association between dietary diversity and GDM [17, 64–66]. The possible reason for the association between inadequate dietary diversity and GDM might be the fact that the majority of women with GDM were classified as inadequate dietary diversity group. Similarly, a large proportion women relied on the monotonous food group in which cereals were the most common consumed food group. Their diets were likely to have been excessive in refined carbohydrates and sugars. On the other hand, dietary diversity varied across a range of factors related to the demographic and socio- economic status of individuals and households. Further research is required to know whether improving dietary pattern adherence during pregnancy is associated with a lower risk of GDM. Our result suggests that clinical and public health efforts to encourage dietary diversity for women of reproductive age might yield benefits in the reduction of GDM in future pregnancies. p yp g y This study has strengths and limitation. The strength of the study was that it employed recent and universal screening tool to detect GDM and it has been done for all pregnant women at 24–28 weeks gestation. They underwent a two-hour 75 g OGTT, and updated stand- ard reference cutoff values were considered. Prevalence of GDM BMC Pregnancy and Childbirth (2019) 19:334 Page 9 of 13 Table 4 Factors associated with GDM among pregnant women attending antenatal care at Gondar town public health facilities, Northwest Ethiopia: March 2018–January 2019 Variables Non-GDM n (%) GDM n (%) COR (95% CI) P -value AOR (95% CI) a P value Maternal age (years) < 25 304 (94.1) 19 (5.8) 1 1 25–29 335 (91.4) 31 (8.5) 1.48 (0.82, 2.68) 0.194 0.99 (0.32, 3.12) 0.994 30–34 168 (78.5) 46 (21.5) 4.38 (2.49, 7.72) < 0.001 2.24 (0.73, 6.83) 0.156 ≥35 89 (71.8) 35 (28.2) 6.29 (3.43, 11.54) < 0.001 3.05 (0.88, 10.51) 0.077 Marital status Married 831(87.9) 114(12.1) 1 1 Single and othersb 65(79.3) 17(20.7) 1.91 (1.08, 3.37) 0.026 2.04 (0.72, 5.78) 0.179 Educational level Not formal education 168 (84) 32 (16.0) 1 1 Primary education 194 (84.7) 35 (15.3) 0.95 (0.56, 1.60) 0.838 1.82 (0.75, 4.43) 0.184 Secondary education and above 534 (89.3) 64 (10.7) 0.63 (0.40, .99) 0.048 0.90 (0.39, 2.11) 0.815 Employment status Unemployed 556 (89.5) 65 (10.5) 1 1 Employed 340 (83.7) 66 (16.3) 1.66 (1.15, 2.40) 0.007 1.44 (0.71, 2.91) 0.309 MUAC MUAC < 28 cm 767 (90) 85 (10) 1 1 MUAC ≥28 cm 129 (73.7) 46 (26.3) 3.22(2.15, 4.82) < 0.001 2.25 (1.18, 4.26) 0.013 SBP (mmHg) 111.57 ± 11.31 115.46 ± 15.01 1.025(1.01, 1.04) 0.001 1.024 (0.99, 1.06) 0.132 DBP(mmHg) 71.60 + 9.882 74.05 ± 12.09 1.022 (1.01, 1.04) 0.011 0.99(0.96, 1.03) 0.757 Anemic status Normal (Hb ≥11 g/dl) 769 (88.0) 105 (12.0) 1 1 Anemia (Hb < 11 g/dl) 113 (83.7) 22 (16.3) 1.43 (0.86, 2.35) 0.164 1.73 (0.81, 3.70) 0.160 Level of physical activity High 299 (92.9) 23 (7.1) 1 1 Moderate 429 (93.3) 31 (6.7) 0.94 (.054, 1.64) 0.827 0.613 (0.27, 1.38) 0.238 Low 168 (68.6) 77 (31.4) 5.96 (3.60, 9.85) < 0.001 3.36 (1.60, 7.04) 0.001 Dietary diversity status Adequate (≥5) 461 (92.8) 36 (7.2) 1 1 Inadequate (< 5) 435 (82.1) 95 (17.9) 2.80 (1.86, 4.19) < 0.001 1.90 (1.02, 3.53) 0.042 Family history of diabetes No 862 (89.3) 103 (10.7) 1 1 Yes 34 (54.8) 28 (45.2) 6.89 (4.02, 11.83) < 0.001 4.03 (1.57, 10.35) 0.004 Parity Nullipara 449 (89.1) 55 (10.9) 1 1 Primipara 243 (87.7) 34 (12.3) 1.14 (0.72, 1.80) 0.567 0.77 (0.28, 2.16) 0.626 Multipara 204 (82.9) 42 (17.1) 1.68 (1.09, 2.60) 0.019 0.99 (0.34, 2.81) 0.987 History of macrocosmic baby No 453 (87.6) 64 (12.4) 1 1 1 Yes 49 (69) 22 (31) 3.18 (1.80, 5.60) < 0.001 1.54 (0.70, 3.37) 0.283 History of abortion No 368 (87.2) 54 (12.8) 1 1 Yes 134 (80.7) 32 (19.3) 1.63 (1.01, 2.63) 0.047 1.014 (0.51, 2.03) 0.968 Table 4 Factors associated with GDM among pregnant women attending antenatal care at Gondar town public health facilities, N h E hi i M h 2018 J 2019 Table 4 Factors associated with GDM among pregnant women attending antenatal care at Gondar town pu Northwest Ethiopia: March 2018–January 2019 ciated with GDM among pregnant women attending antenatal care at Gondar town public health facilities, March 2018–January 2019 Table 4 Factors associated with GDM among pregnant women attending antenatal care at Gondar town public health facilities, Muche et al. Prevalence of GDM Besides, for pregnant women who had risk factors for GDM and whose OGTT results were negative at the regular test were tested again at late gestational age. WHO recom- mends that in settings where laboratories or proper storages and transport of blood samples is not guaran- teed, which is the case in resource limited countries like Ethiopia, the use of point of care tests may influence the result [23]. However, we used plasma-calibrated hand-held glucometers because of convenience and acceptable reliability. Moreover, due to the nature of study design which restricted causal inference and could not reveal the temporal sequence between the Page 11 of 13 Page 11 of 13 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 factors and the outcome variable which could be limita- tion of the study. Conclusions In conclusion, the overall prevalence of GDM was found to be high and a major public health concern among preg- nant women in Gondar town, Northwest Ethiopia. Previ- ous history of GDM, antenatal depression, family history of diabetes, low physical activity, overweight and/or obese women and inadequate dietary diversity were significantly associated with GDM. Given the imminent burden of obesity, unhealthy eating, and physical inactivity, a sub- stantial threat of GDM is anticipated in Ethiopia. This may be an appropriate time to think about considering the need for routine screening of GDM to reduce the impact of the diseases in the country. In this regard, our study may serve as a baseline for larger studies and back- ground evidence for policy debates for universal GDM screening in the existing health care system. Integrating healthy lifestyle choices into ANC services, such as reduced obesity, healthy eating, improvement in physical activities and effective strategies for coping with depres- sion are strongly recommended. Authors’ contributions AAM conceived and designed the study, analyzed the data and prepared the manuscript. OO & YKG assisted the development of the research idea, the analysis, interpretation and preparation of the manuscript. All authors read and approved the final manuscript. Received: 27 March 2019 Accepted: 5 September 2019 Received: 27 March 2019 Accepted: 5 September 2019 Received: 27 March 2019 Accepted: 5 September 2019 Availability of data and materials h d d d l d d The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Consent for publication Not applicable. Consent for publication Not applicable. Author details 1 1Pan African University Life and Earth Sciences Institute (including health and agriculture), Department of Obstetrics and Gynaecology, College of Medicine, University of Ibadan, Ibadan, Nigeria. 2Department of Epidemiology and Biostatistics, Institute of Public Health, University of Gondar, Gondar, Ethiopia. 3Department of Obstetrics and Gynaecology, College of Medicine, University College Hospital, University of Ibadan, Ibadan, Nigeria. Additional file 2. The specifications of the HemoCue Glucose B-201+ glucometer. Additional file 2. The specifications of the HemoCue Glucose B-201+ glucometer. References 1. World Health Organization (WHO). Diagnostic criteria and classification of hyperglycaemia first detected in pregnancy. 2013. Retrived from: http:// apps.who.int/iris/bitstream/10665/85975/1/WHO_NMH_MND_13.2_eng.pdf. On 24 Feb 2019. 1. World Health Organization (WHO). Diagnostic criteria and classification of hyperglycaemia first detected in pregnancy. 2013. Retrived from: http:// apps.who.int/iris/bitstream/10665/85975/1/WHO_NMH_MND_13.2_eng.pdf. On 24 Feb 2019. 2. Hartling L, Dryden DM, Guthrie A, Muise M, Vandermeer B, Aktary WM, et al. Screening and diagnosing gestational diabetes mellitus. Evidence report/ technology assessment. 2012;210:1–327. 2. Hartling L, Dryden DM, Guthrie A, Muise M, Vandermeer B, Aktary WM, et al. Screening and diagnosing gestational diabetes mellitus. Evidence report/ technology assessment. 2012;210:1–327. 3. Metzger BE, Buchanan TA, Coustan DR, De Leiva A, Dunger DB, Hadden DR, et al. Summary and recommendations of the fifth international workshop- conference on gestational diabetes mellitus. Diabetes Care. 2007; 30(Supplement 2:S251–S60. Competing interests The authors declare that they have no competing interests. e authors declare that they have no competing interests. The authors declare that they have no competing interests. Additional file 1. English version consent form and questionnaire for GDM survey among women attending ANC at Gondar town public health facilities, Northwest Ethiopia. Additional file 1. English version consent form and questionnaire for GDM survey among women attending ANC at Gondar town public health facilities, Northwest Ethiopia. Funding This study was sponsored by the Pan African University (PAU), a continental initiative of the African Union Commission (AU), Addis Ababa, Ethiopia, as part of the Ph.D. Fellowship Program in Reproductive Health Sciences. The funder had no any role in design of the study, data collection, analysis, interpretation of the data and preparation of the manuscript. Supplementary information Supplementary information accompanies this paper at https://doi.org/10. 1186/s12884-019-2492-3. Abbreviations ANC: Antenatal care; AOR: Adjusted odds ratio; BMI: Body mass index; BP: Blood pressure; CI: Confidence interval; CM: Centi meter; DBP: Diastolic blood pressure; DM: Diabetes mellitus; EDPS: Edinburgh Postnatal Depression Scale; FANTA: Food and Nutrition Technical Assistance; FPG: Fasting plasma glucose; GDM: Gestational diabetes mellitus; HAPO: Hyperglycemia and adverse pregnancy outcome; IADPSG: International Association of the Diabetes and Pregnancy Study Group; IDF: International diabetes federation; IGT: Impaired glucose tolerance; IPAQ: International Physical Activity Questionnaire; LMIC: Low and middle-income countries; LNMP: Last normal menstrual period; MDDS: Minimum dietary diversity score; MET: Metabolic equivalent of task; MUAC: Mid-upper arm circumference; NCD: Non- communicable diseases; OGTT: Oral glucose tolerance test; OR: Odds ratio; SBP: Systolic blood pressure; SD: Standard deviation; SPSS: Statistical package for social sciences; WHO: World Health Organization Ethics approval and consent to participate The study was conducted after ethical approval was obtained from the Institutional Review Board (IRB) of the University of Ibadan/University College Hospital (Ref.No; UI/EC/17/0435) and the IRB of the University of Gondar (Ref.No; O/V/P/RCS/05/811/2018). Permission from the Amhara Public Health Institute and the health authorities of the study sites was also received prior to the start of the study. Before enrolment, participants were informed about the study, its aims, effects, and importance of screening. Written consent was obtained by trained data collectors. Then, participants were subjected to full history taking through clinical examinations and laboratory investigations. All information was made anonymous in order to maintain confidentiality. Participants diagnosed with GDM was referred to health providers and get possible treatment options in their respective public health facilities. Acknowledgements 4. Ogurtsova K, da Rocha Fernandes J, Huang Y, Linnenkamp U, Guariguata L, Cho N, et al. IDF diabetes atlas: global estimates for the prevalence of diabetes for 2015 and 2040. Diabetes Res Clin Pract. 2017;128:40–50. We would also like to forward our sincere gratitude to the African Union Commission (AU) for funding this study, the Pan African University Life and Earth Science Institute (PAULESI), and University of Ibadan (UI) for hosting the program and University of Gondar. We would also like to acknowledge the Setema Limited PLC for their material support (Hemocue Glucose 201 with micro cuvette). We appreciate Dr. Mulat Adefris (Associate professor of Obstetrics and Gynecology) at the University of Gondar for his professional assistant and Mr. Demeke Dessu of the University of Gondar for his editing assistance. Last but not least, we would like to extend our thanks to all data collectors and study participants who participated in the study. 5. Mwanri AW, Kinabo J, Ramaiya K, Feskens EJ. Gestational diabetes mellitus in sub-Saharan Africa: systematic review and metaregression on prevalence and risk factors. Tropical Med Int Health. 2015;20(8):983–1002. 6. Zhu Y, Zhang C. Prevalence of gestational diabetes and risk of progression to type 2 diabetes: a global perspective. Curr Diab Rep. 2016;16(1):7. 7. Mwanri AW, Kinabo J, Ramaiya K, Feskens EJ. Prevalence of gestational diabetes mellitus in urban and rural Tanzania. Diabetes Res Clin Pract. 2014; 103(1):71–8. Page 12 of 13 Page 12 of 13 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 8. Kuti MA, Abbiyesuku FM, Akinlade KS, Akinosun OM, Adedapo KS, Adeleye JO, et al. Oral glucose tolerance testing outcomes among women at high risk for gestational diabetes mellitus. Journal of clinical pathology. 2011. https://doi.org/10.1136/jcp.2010.087098. 30. Hagströmer M, Oja P, Sjöström M. The international physical activity questionnaire (IPAQ): a study of concurrent and construct validity. Public Health Nutr. 2006;9(6):755–62. 31. Food and Agriculture Organization (FAO). Minimum dietary diversity for women: a guide for measurement, food and agriculture organization of the United Nations, USAID’s food and nutrition technical assistance III project (FANTA), managed by FHI 360, Rome. 2016. Retrieved from www.fao.org/3/ a-i5486e.pdf. On 12 Dec 2018. 9. Niyibizi JB, Safari F, Ahishakiye JB, Habimana JB, Mapira H, Mutuku NC. Gestational diabetes mellitus and its associated risk factors in pregnant women at selected health facilities in Kigali city, Rwanda. J Diabetes Mellitus. 2016;6(04):269. 32. Acknowledgements Gale CR, Javaid MK, Robinson SM, Law CM, Godfrey KM, Cooper C. Maternal size in pregnancy and body composition in children. J Clin Endocrinol Metab. 2007;92(10):3904–11. 10. Njete H, John B, Mlay P, Mahande M, Msuya S. Prevalence, predictors and challenges of gestational diabetes mellitus screening among pregnant women in northern Tanzania. Tropical Med Int Health. 2018;23(2):236–42. 11. Seyoum B, Kiros K, Haileselase T, Leole A. Prevalence of gestational diabetes mellitus in rural pregnant mothers in northern Ethiopia. Diabetes Res Clin Pract. 1999;46(3):247–51. 33. Ricalde AE, Velásquez-Meléndez G, Tanaka ACA, de Siqueira AA. Mid-upper arm circumference in pregnant women and its relation to birth weight. Rev Saude Publica. 1998;32:112–7. 12. Management Sciences for Health; Gestational Diabetes: A Risk Factor for Maternal Death in Tigray, Ethiopia, 5 March 2015. https://www.msh.org/ sites/msh.org/files/eth_enhat_techbrief10_webv.pdf. On 18 Nov 2018. 34. Oza-Frank R, Ali MK, Vaccarino V, Narayan KV. Asian Americans: diabetes prevalence across US and World Health Organization weight classifications. Diabetes Care. 2009;32(9):1644–6. 35. Murray D, Cox JL. Screening for depression during pregnancy with the Edinburgh Depression Scale (EDDS). J Reprod Infant Psychol. 1990;8(2):99–107. 13. van Poppel MN, Oostdam N, Eekhoff ME, Wouters MG, van Mechelen W, Catalano PM. Longitudinal relationship of physical activity with insulin sensitivity in overweight and obese pregnant women. J Clin Endocrinol Metab. 2013;98(7):2929–35. 36. Tesfaye M, Hanlon C, Wondimagegn D, Alem A. Detecting postnatal common mental disorders in Addis Ababa, Ethiopia: validation of the Edinburgh postnatal depression scale and Kessler scales. J Affect Disord. 2010;122(1–2):102–8. 14. Macaulay S, Ngobeni M, Dunger DB, Norris SA. The prevalence of gestational diabetes mellitus amongst black south African women is a public health concern. Diabetes Res Clin Pract. 2018;139:278–87. 37. Wado YD, Afework MF, Hindin MJ. Effects of maternal pregnancy intention, depressive symptoms and social support on risk of low birth weight: a prospective study from southwestern Ethiopia. PLoS One. 2014;9(5):e96304. 15. Oppong SA, Ntumy MY, Amoakoh-Coleman M, Ogum-Alangea D, Modey- Amoah E. Gestational diabetes mellitus among women attending prenatal care at Korle-Bu teaching hospital, Accra, Ghana. Int J Gynecol Obstet. 2015; 131(3):246–50. 38. Biratu A, Haile D. Prevalence of antenatal depression and associated factors among pregnant women in Addis Ababa, Ethiopia: a cross-sectional study. Reprod Health. 2015;12(1):99. 16. Liang C, DeCourcy K, Prater MR. High–saturated-fat diet induces gestational diabetes and placental vasculopathy in C57BL/6 mice. Metabolism. 2010; 59(7):943–50. 39. Golbasi Z, Kelleci M, Kisacik G, Cetin A. Acknowledgements Prevalence and correlates of depression in pregnancy among Turkish women. Matern Child Health J. 2010;14(4):485–91. 17. Bo S, Menato G, Lezo A, Signorile A, Bardelli C, De Michieli F, et al. Dietary fat and gestational hyperglycaemia. Diabetologia. 2001;44(8):972–8. 40. D’Orazio P, Burnett RW, Fogh-Andersen N, Jacobs E, Kuwa K, Külpmann WR, et al. Approved IFCC recommendation on reporting results for blood glucose (abbreviated). Clin Chem. 2005;51(9):1573–6. 18. Moses RG, Shand JL, Tapsell LC. The recurrence of gestational diabetes: could dietary differences in fat intake be an explanation? Diabetes Care. 1997;20(11):1647–50. 41. World Health Organization (WHO). Diagnostic criteria and classification of hyperglycaemia first detected in pregnancy. 2013. Available from: http:// apps.who.int/iris/bitstream/10665/8597/1/WHO_NMH_MND_13.2_eng.pdf. 19. Catalano PM, Tyzbir ED, Roman NM, Amini SB, Sims EA. Longitudinal changes in insulin release and insulin resistance in nonobese pregnant women. Am J Obstet Gynecol. 1991;165(6 Pt 1):1667–72. 42. Diabetes IAo, Panel PSGC. International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care. 2010;33(3):676–82. 20. Plows JF, Stanley JL, Baker PN, Reynolds CM, Vickers MH. The pathophysiology of gestational diabetes mellitus. Int J Mol Sci. 2018;19(11): 3342. 43. World Health Organization (WHO). Haemoglobin concentrations for the diagnosis of anaemia and assessment of severity. 2011. Retrieved from https://www.who.int/vmnis/indicators/haemoglobin/en/. On 12 Dec 2018. 21. Farrar D, Duley L, Dowswell T, Lawlor DA. Different strategies for diagnosing gestational diabetes to improve maternal and infant health. Cochrane Database Syst Rev. 2017;8:CD007122. 44. Khalil NA, Fathy WM, Mahmoud NS. Screening for gestational diabetes among pregnant women attending a rural family health center-Menoufia governorate-Egypt. J Fam Med Health Care. 2017;3(1):6–11. 22. International Diabetes Federation (IDF). GDM model of care implementation protocol guidelines for healthcare professionals. 2015. Retrieved from https://www.idf.org/.../guidelines/77-idf-gdm-model-of-care-implementation- protocol.pdf. On 27 Nov 2018. 45. Olagbuji BN, Atiba AS, Olofinbiyi BA, Akintayo AA, Awoleke JO, Ade-Ojo IP, et al. Prevalence of and risk factors for gestational diabetes using 1999, 2013 WHO and IADPSG criteria upon implementation of a universal one-step screening and diagnostic strategy in a sub-Saharan African population. Eur J Obstet Gynecol Reprod Biol. 2015;189:27–32. 23. Hod M, Kapur A, Sacks DA, Hadar E, Agarwal M, Di Renzo GC, et al. The International Federation of Gynecology and Obstetrics (FIGO) initiative on gestational diabetes mellitus: a pragmatic guide for diagnosis, management, and care. Int J Gynecol Obstet. 2015;131:S173–211. 46. Adams S, Rheeder P. Acknowledgements Screening for gestational diabetes mellitus in a south African population: prevalence, comparison of diagnostic criteria and the role of risk factors. S Afr Med J. 2017;107(6):523–7. 24. Buchanan TA, Xiang AH, Page KA. Gestational diabetes mellitus: risks and management during and after pregnancy. Nat Rev Endocrinol. 2012;8(11): 639–49. 47. Organization WH. Global report on diabetes: World Health Organization; 2016. 2017. 25. Federal Democratic Republic of Ethiopia Central Statistical Agency Population Projection of Ethiopia for All Regions At Wereda Level from 2014–2017. Addis Ababa: Central Statistical Agency; 2014. 48. El Sagheer GM, Hamdi L. Prevalence and risk factors for gestational diabetes mellitus according to the diabetes in pregnancy study group India in comparison to International Association of the diabetes and pregnancy study groups in El-Minya, Egypt. Egypt J Intern Med. 2018;30(3):131. 26. Pylvanen V, Pakarinen A, Knip M, Isojarvi J. Insulin-related metabolic changes during treatment with valproate in patients with epilepsy. Epilepsy Behav. 2006;8(3):643–8. 49. Nelson SM, Matthews P, Poston L. Maternal metabolism and obesity: modifiable determinants of pregnancy outcome. Hum Reprod Update. 2010;16(3):255–75. 27. Kwon S, Hermayer KL, Hermayer K. Glucocorticoid-induced hyperglycemia. Am J Med Sci. 2013;345(4):274–7. 50. Huda SS, Brodie LE, Sattar N. Obesity in pregnancy: prevalence and metabolic consequences. Elsevier; Seminars in Fetal and Neonatal Medicine. 2010;15(2):70–6. 28. American Diabetes Association. Classification and diagnosis of diabetes. Sec. 2. In standards ofMedical Care in Diabetesd 2017. Diabetes Care. 2017; 40(Suppl. 1):S11–24. 51. Ogonowski J, Miazgowski T, Engel K, Celewicz Z. Birth weight predicts the risk of gestational diabetes mellitus and pregravid obesity. Nutrition (Burbank, Los Angeles County, Calif). 2014;30(1):39–43. 29. Colagiuri S, Falavigna M, Agarwal MM, Boulvain M, Coetzee E, Hod M, et al. Strategies for implementing the WHO diagnostic criteria and classification of hyperglycaemia first detected in pregnancy. Diabetes Res Clin Pract. 2014;103(3):364–72. 52. Dabelea D, Snell-Bergeon JK, Hartsfield CL, Bischoff KJ, Hamman RF, McDuffie RS. Increasing prevalence of gestational diabetes mellitus (GDM) Page 13 of 13 Muche et al. BMC Pregnancy and Childbirth (2019) 19:334 over time and by birth cohort: Kaiser Permanente of Colorado GDM screening program Diabetes Care 2005;28(3):579 84 over time and by birth cohort: Kaiser Permanente of Colorado GDM over time and by birth cohort: Kaiser Permanente of Colorado G screening program. Diabetes Care. 2005;28(3):579–84. screening program. Diabetes Care. 2005;28(3):579–84. 53. Kim C, Berger DK, Chamany S. Recurrence of gestational diabetes mellitus: a systematic review. Diabetes care. 2007;30(5):1314-9. systematic review. Diabetes care. Acknowledgements Byrn M, Penckofer S. The relationship between gestational diabetes and antenatal depression. J Obstet Gynecol Neonatal Nurs. 2015;44(2):246–55. 68. Byrn M, Penckofer S. The relationship between gestational diabetes and antenatal depression. J Obstet Gynecol Neonatal Nurs. 2015;44(2):246–55. 69. Hinkle SN, Buck Louis GM, Rawal S, Zhu Y, Albert PS, Zhang C. A longitudinal study of depression and gestational diabetes in pregnancy and the postpartum period. Diabetologia. 2016;59(12):2594–602. 69. Hinkle SN, Buck Louis GM, Rawal S, Zhu Y, Albert PS, Zhang C. A longitudinal study of depression and gestational diabetes in pregnancy and the postpartum period. Diabetologia. 2016;59(12):2594–602. 70. Byrn MA, Penckofer S. Antenatal depression and gestational diabetes. Nurs Womens Health. 2013;17(1):22–33. 71. Lustman PJ, Clouse RE. Depression in diabetic patients: the relationship between mood and glycemic control. J Diabetes Complicat. 2005;19(2):113–22. 72. Silva N, Atlantis E, Ismail K. A review of the association between depression and insulin resistance: pitfalls of secondary analyses or a promising new approach to prevention of type 2 diabetes? Curr Psychiatry Rep. 2012;14(1): 8–14. 72. Silva N, Atlantis E, Ismail K. A review of the association between depression and insulin resistance: pitfalls of secondary analyses or a promising new approach to prevention of type 2 diabetes? Curr Psychiatry Rep. 2012;14(1): 8–14. Acknowledgements 2007;30(5):1314-9 54. Spong C, Guillermo L, Kuboshige J, Cabalum T. Recurrence of gestational diabetes mellitus: identification of risk factors. Am J Perinatol. 1998;15(01): 29–33. 55. Carr DB, Utzschneider KM, Hull RL, Tong J, Wallace TM, Kodama K, et al. Gestational diabetes mellitus increases the risk of cardiovascular disease in women with a family history of type 2 diabetes. Diabetes Care. 2006;29(9): 2078–83. 56. Keshavarz M, Cheung NW, Babaee GR, Moghadam HK, Ajami ME, Shariati M. Gestational diabetes in Iran: incidence, risk factors and pregnancy outcomes. Diabetes Res Clin Pract. 2005;69(3):279–86. outcomes. Diabetes Res Clin Pract. 2005;69(3):279–86. 57. Garshasbi A, Faghihzadeh S, Naghizadeh MM, Ghavam M. Prevalence and risk factors for gestational diabetes mellitus in Tehran. J Family Reprod Health. 2008;2(2):75–80. 58. Ehrmann DA, Sturis J, Byrne MM, Karrison T, Rosenfield RL, Polonsky KS. Insulin secretory defects in polycystic ovary syndrome. Relationship to insulin sensitivity and family history of non-insulin-dependent diabetes mellitus. J Clin Invest. 1995;96(1):520–7. 59. Ferrannini E, Gastaldelli A, Iozzo P. Pathophysiology of prediabetes. Med Clin. 2011;95(2):327–39. 60. Nguyen CL, Pham NM, Lee AH, Nguyen PTH, Chu TK, Ha AVV, et al. Physical activity during pregnancy is associated with a lower prevalence of gestational diabetes mellitus in Vietnam. Acta Diabetol. 2018;55(9):955–62. 61. Tobias DK, Zhang C, Van Dam RM, Bowers K, Hu FB. Physical activity before and during pregnancy and risk of gestational diabetes mellitus: a meta- analysis. Diabetes Care. 2011;34(1):223–9. 62. Anjana RM, Sudha V, Lakshmipriya N, Anitha C, Unnikrishnan R, Bhavadharini B, et al. Physical activity patterns and gestational diabetes outcomes–the wings project. Diabetes Res Clin Pract. 2016;116:253–62. 63. Cuco G, Fernandez-Ballart J, Sala J, Viladrich C, Iranzo R, Vila J, et al. Dietary patterns and associated lifestyles in preconception, pregnancy and postpartum. Eur J Clin Nutr. 2006;60(3):364. 64. Rifas-Shiman SL, Rich-Edwards JW, Kleinman KP, Oken E, Gillman MW. Dietary quality during pregnancy varies by maternal characteristics in project viva: a US cohort. J Am Diet Assoc. 2009;109(6):1004–11. 65. Saldana TM, Siega-Riz AM, Adair LS. Effect of macronutrient intake on the development of glucose intolerance during pregnancy. Am J Clin Nutr. 2004;79(3):479–86. 66. Wang Y, Storlien LH, Jenkins AB, Tapsell LC, Jin Y, Pan JF, et al. Dietary variables and glucose tolerance in pregnancy. Diabetes Care. 2000;23(4): 460–4. 67. Morrison C, McCook JG, Bailey BA. First trimester depression scores predict development of gestational diabetes mellitus in pregnant rural Appalachian women. J Psychosom Obstet Gynecol. 2016;37(1):21–5. 68. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
https://openalex.org/W4289147353	https://zenodo.org/records/1872237/files/(34-53)PEDICABS_edited%20(1)-format%20(2).pdf	English	null	Design and Modification of Pedicabs and Analysis on Various Factors	Zenodo (CERN European Organization for Nuclear Research)	2,018	cc-by	7,967	Journal of Recent Trends in Mechanics Volume 3 Issue 3 Abstract bst act This research completely based on the design of pedicabs (also known as bicycle rickshaw, in India), many other names were commonly used in all other countries across the world. Calculations of major components such as brakes, suspension, steering, electricals, transmissions etc. are covered in this paper, as per the standards. The main objective of this work is to provide the safety and comfort for both driver and passengers, and to decrease the use of natural resources and to reduces the pollution. For design and analysis of complete project PRO-E and ANSYS is used a tool to maintain the efficiency. Keywords:PEDICABS, rickshaw, transport, velocity Design and Modification of Pedicabs and Analysis on Various Factors DeepjyotiBasak Assistant Professor Department of Mechanical Engineering Parul Institute of Technology, Vadodara, Gujarat, India Email:djeratones43@gmail.com INTRODUCTION replaced less-efficient hand-pulled rickshaw that required the driver to walk or run while pulling the vehicle[8]. Cycle rickshaws are often hailed as environmentally-friendly, inexpensive modes of transport[9]. Many cities in developing countries are highly polluted[10]–[14]. The main reasons are the air and noise pollution caused by transport vehicles, especially petrol- powered two and three wheelers[15]. A cycle rickshaw, also known as a Pedicab[1], velotaxi or trishaw from tricycle rickshaw[2], is a human-powered vehicle for hire[3], it is usually with one or two seats for carrying passengers in addition to the driver[4]. The driver pedals in front of the passenger seat[5], though some vehicles have the driver in the rear[6]. Cycle rickshaws are widely used all over Asia[7], where they have largely Present Frame 34 Page 34-53 © MAT Journals 2018. All Rights Reserved Present Frame A:Present Frame B: Present Frame Fig: 1. Geometry A:Present Frame B: Present Frame Fig: 1. Geometry Page 34-53 © MAT Journals 2018. All Rights Reserved 34 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Proposed Frame Fig: 2.Proposed Frame Fig: 2.Proposed Frame PEDAL FORCE Considering the weight of driver = 75 kg i.e. 75 kg x 5 = 375 w (this is the amount of energy which is to be delivered by the driver for one-hour continuous operation) An elite cyclist produces 5 watts of power for every kilogram of body weight for an hour. (data from vehicle ergonomics by Howard Sutherland)  At present 40 T chain ring and 24 T rear cog is installed. Inturn it provides only single speed ratio of 1.666 Considering the weight of driver = 75 kg i.e. 75 kg x 5 = 375 w (this is the amount of energy which is to be delivered by the driver for one-hour continuous operation) An elite cyclist produces 5 watts of power FINDING Gear ratio x Pedal force x Pedal leverage x Mechanical efficiency = Radius of wheel x Tractive effort  In present scenario only single gear is available for pedicabs driver. The main problem of single speed drive is air drag, the higher air drag the higher effort. PROBLEM An elite cyclist produces 5 watts of power for every kilogram of body weight for an hour. (data from vehicle ergonomics by Howard Sutherland)  Main problem is to achieve or to maintain the proper speed at inclination. Calculating for vehicle velocity at N kmph (where N = vel. In kmph)  Increase in effort.  Increase in effort.  Difficult to overcome the higher air drag. Velocity = N kmph (1000/3600) = ……m/sec  Improper vehicle stability.  Improper vehicle stability. We know that, We know that, Acceleration = velocity/time (assume time by equation of motion) Page 34-53 © MAT Journals 2018. All Rights Reserved Journal of Recent Trends in Mechanics Volume 3 Issue 3 Where, Fr = frictional force μ = friction co-efficient N = is the normal or perpendicular force. I.e. μ = weight of vehicle x frictional coefficient μ = answer/3 (number of tires) μ = _____ Newton (in one tire) Rolling Resistance (Froll) = Cr. m. g Where, Cr = rolling coefficient w= m.g Air Drag: - (Fd) = Where, = density of fluid (air) kg/m^3 v = speed of object related to fluid A = cross sectional area Cd= drag coefficient F (slope) = s m g Where, s = upward slope, dimensionless F (acceleration) = a. m Where a = acceleration in m/s^2 Force = Power generated by driver (i.e. 375 watts)/velocity (m/sec) Force (F) = _____ Newton We know that, Where, = torque I = moment of inertia. = angular acceleration Also, (Both the values from eq. 1) = _________rad/ Now calculating the moment of inertia. I =mr^2 (Moment of inertia will be calculated for all the rotational parts in the vehicle i.e. sprockets, wheels, etc.) Now substituting the value of I and in eq…. = ________ N-m Now, we know that, Force x Radius Force = /Radius Force = _________N (Pedal force without considering the various resistance forces and air drag) Now resistive forces to be calculated as:- Frictional force (Fr) = μN The power required to overcome the total drag is: P = Ftotal x v Where v: velocity in m/s Ftotal = (Froll + Fslope + Facceleration + Fair drag)/h Where h: drivetrain efficiency, dimensionless VEHICLE CONSIDERATION  Unladen weight= 40 kg  Maximum authorized weight (MAM):-  Passenger = 3 nos. (3 x 75kg=225kg)  Driver = 1 (70-75 kg)  Luggage = 30 kg  Total weight = 370 kg  Maximum vehicle speed = 15 kmph  Average speed (in city) = 8 kmph Where, v, u are the initial and final velocity Page 34-53 © MAT Journals 2018. All Rights Reserved 35 RESULT OF PEDAL FORCE AT DIFFERENT SPEED RESULT OF PEDAL FORCE AT DIFFERENT SPEED RESULT OF PEDAL FORCE AT DIFFERENT SPEED Table: 1. Result of Pedal Force at Different Speed SPEED Time assumption to achieve the speed PEDAL FORCE (N) without resistance Fr (N) Froll (N) Facc (N) Power needed to overcome the resistances (watts) At 5 kmph 2 sec 3.885 277.5 1.48 51.5 347.87 At 10kmph 3 sec 5.141 277.5 1.48 102.7 401.76 At 15kmph 5 sec 4.500 277.5 1.48 156.9 458.82 Table: 1. Result of Pedal Force at Different Speed 36 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Journal of Recent Trends in Mechanics Volume 3 Issue 3 rnals 2018. All Rights Reserved B: Pedal Leverage Leverage 48X16 – 3.0 48X20 – 2.4 48X24 – 2.0 FORMULA USED GEAR RATIO = N1/N2 METER DEVELOPMENT FOR PROPOSED GEARS 48X16 – 6.0 48X20 – 4.8 48X24 – 4.0 FORMULA Development = Drive wheel circumference in meters x (N1 / N2) SPEED AT CADENCE Table: 2. Speed at Cadence CADENCE AT SPEED Table: 3. Cadence at Speed FORMULA = A: Pedal Leverage B: Pedal Leverage Fig: 3.Pedal Leverage 48X16 – 3.0 48X20 – 2.4 48X24 – 2.0 Pedal force is calculated according to the height of the driver. Consideration of Pedicab driver Weight = 75 kg height according to the medical survey = 5.8” (i.e. 176.68 mm) therefore leverage will be 176.68 mm Standard Sizes of crank are 160 mm, 165 mm, 172mm, 175mm, 185mm 48X16 – 6.0 48X20 – 4.8 48X24 – 4.0 For this project 175 mm is to be used 37 Page 34-53 © M MATERIAL: - Aluminum alloy GEAR SHIFTING MECHANISM Fig: 4.Gear Shifting Mechanism SPECIFICATION  Weight = 362 grams  Material = ANSI 430, Steel alloy  No. of teeth in both pulley = 10  Screw: i. 5 no. - 2nos ii. 3 no. Page 34-53 © MAT Journals 2018. All Rights Reserved contact with chain) For single teeth = 294 lbs/ 48 = 5.91 lbs = 2.68kg = 26.28 newton on single teeth Weight on pedal (maximum condition taken) = weight of rider x 1.78 Weight on pedal (maximum condition taken) = weight of rider x 1.78 = 293.7 lbs = 294 lbs (force on teeth in contact with chain) For single teeth = 294 lbs/ 48 = 5.91 lbs = 2.68kg = 26.28 newton on single teeth CAD MODEL AND ANALYSIS A: Cad Model B: Equivalent Strain C: Equilavent Stress D: Total Deformation Fig: 6. Teeth Sprocket = 293.7 lbs = 294 lbs (force on teeth in Consideration of Pedicab driver - 2nos  Spring : pitch = 2.8 mm GEAR RATIO PROPOSED GEAR RATIO CHAIN RING – 48 TEETH REAR COG – 16, 20, 24 TEETH MATERIAL: - Aluminum alloy y GEAR SHIFTING MECHANISM Fig: 4.Gear Shifting Mechanism 37 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Journal of Recent Trends in Mechanics Volume 3 Issue 3 SPROCKET DESIGN AND ANALYSIS Fig: 5.Sprocket Design and Analysis P = Chain Pitch N = Number of Teeth Dr = Roller Diameter Ds = (Seating curve diameter) = 1.0005 Dr + 0.003 R = Ds/2 = 0.5025 Dr + 0.0015 A = 35º + 60º/N B = 18º + 56º/N ac = 0.8 x Dr M = 0.8 x Drcos(35º+60º/N) T = 0.8 x Drsin (35º+60º/N) E = 1.3025 Dr + 0.0015 Chordal Length of Arc xy = (2.605 Dr + 0.003) sin (9º-28º/N) yz = Dr [1.4 sin (17º-64º/N)-0.8sin(18º- 56º/N)] ab = 1.4 Dr W = 1.4 Dr Cos 180º/N V = 1.4 Dr Sin 180º/N F = Dr [0.8 cos (18º-56º/n) + 1.4 cos (17º – 64º/N) – 1.3025] – 0.0015 H = S = P/2 cos 180º/N + H sin180º/N PD = Outer Diameter (OD) = P (0.6 + cot 180º/N) (when j = 0.3 P) B = 18º + 56º/N ac = 0.8 x Dr M = 0.8 x Drcos(35º+60º/N) T = 0.8 x Drsin (35º+60º/N) E = 1.3025 Dr + 0.0015 Chordal Length of Arc xy = (2.605 Dr + 0.003) sin (9º-28º/N) yz = Dr [1.4 sin (17º-64º/N)-0.8sin(18º- 56º/N)] ab = 1.4 Dr W = 1.4 Dr Cos 180º/N V = 1.4 Dr Sin 180º/N F = Dr [0.8 cos (18º-56º/n) + 1.4 cos (17º – 64º/N) – 1.3025] – 0.0015 H = S = P/2 cos 180º/N + H sin180º/N PD = Outer Diameter (OD) = P (0.6 + cot 180º/N) (when j = 0.3 P) SPROCKET DESIGN AND ANALYSIS B 18 + 56 /N ac = 0.8 x Dr M = 0.8 x Drcos(35º+60º/N) T = 0.8 x Drsin (35º+60º/N) E = 1.3025 Dr + 0.0015 Chordal Length of Arc xy = (2.605 Dr + 0.003) sin (9º-28º/N) yz = Dr [1.4 sin (17º-64º/N)-0.8sin(18º- 56º/N)] ab = 1.4 Dr W = 1.4 Dr Cos 180º/N V = 1.4 Dr Sin 180º/N F = Dr [0.8 cos (18º-56º/n) + 1.4 cos (17º – 64º/N) – 1.3025] – 0.0015 H = S = P/2 cos 180º/N + H sin180º/N PD = Outer Diameter (OD) = P (0.6 + cot 180º/N) (when j = 0.3 P) Fig: 5.Sprocket Design and Analysis P = Chain Pitch N = Number of Teeth Dr = Roller Diameter Ds = (Seating curve diameter) = 1.0005 Dr + 0.003 R = Ds/2 = 0.5025 Dr + 0.0015 A = 35º + 60º/N CALCULATIVE VALUES OF SPROCKETS Table: 4.Calculative Values of Sprockets P (Chain Pitch) 12.70 12.70 12.70 12.70 N (Number of Teeth) 48 16 20 24 Dr (Roller Diameter) 7.5 7.5 7.5 7.5 PD 194.1803 65.0981 81.1842 97.2985 Ds 7.5405 7.5405 7.5405 7.5405 R 3.7703 3.7703 3.7703 3.7703 A 36.2500 38.7500 38.0000 37.5000 B 16.8333 14.5000 13.2000 15.6667 ac 6.0000 6.0000 6.0000 6.0000 M 4.8387 4.6793 4.7281 4.7601 T 3.5479 3.7555 3.6940 3.6526 E 9.7703 9.7703 9.7703 9.7703 xy 2.8602 2.4660 2.5844 2.6632 yz 1.0979 0.8597 0.9315 0.9792 ab 10.9000 10.9000 10.9000 10.9000 W 10.4775 10.2982 10.3707 10.4102 V 0.6867 2.0484 1.6426 1.3705 F 6.0826 6.2695 6.2183 6.1800 H 4.4469 4.6994 4.6288 4.5793 S 6.6272 7.1448 6.9939 6.8934 FORCES ON SPROCKET TEETH Riders weight = 165 pounds Crank length = 175 mm Radius of front sprocket = 7.6449/2 = 3.82245 inch Crank length / radius of sprocket = 6.837/3.82245 = 1.78 times of pedal force CALCULATIVE VALUES OF SPROCKETS Table: 4 C l l ti V l 38 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Page 34-53 © MAT Journals 2018. All Rights Reserved Page 34-53 © MAT Journals 2018. All Rights Reserved CAD MODEL AND ANALYSIS CAD MODEL AND ANALYSIS CAD MODEL AND ANALYSIS A: Cad Model B: Equivalent Strain CAD MODEL AND ANALYSIS A: Cad Model B: Equivalent Strain B: Equivalent Strain A: Cad Model A: Cad Model C: Equilavent Stress D: Total Deformation Fig: 6. Teeth Sprocket D: Total Deformation C: Equilavent Stress D: Total Deformation Fig: 6. Teeth Sprocket D: Total Deformation D: Total Deformation : Equilavent Stress Fig: 6. Teeth Sprocket A: Cad Model B: Total Deformation B: Total Deformation B: Total Deformation A: Cad Model Page 34-53 © MAT Journals 2018. All Rights Reserved 39 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Journal of Recent Trends in Mechanics Volume 3 Issue 3 D: Equivalent Strain C: Equivalent Stress D: Equivalent Strain Fig: 7.20 Teeth Sprocket D: Equivalent Strain C: Equivalent Stress C: Equivalent Stress D: Equivalent Strain Fig: 7.20 Teeth Sprocket Fig: 7.20 Teeth Sprocket A: Cad Model B: Total Deformation C: Equivalent Strain D: Equivalent Stress Fig: 8.24 Teeth Sprocket A: Cad Model B: Total Deformation B: Total Deformation A: Cad Model B: Total Deformation B: Total Deformation Cad Model C: Equivalent Strain D: Equivalent Stress Fig: 8.24 Teeth Sprocket D: Equivalent Stress C: Equivalent Strain q Fig: 8.24 Teeth Sprocket Fig: 8.24 Teeth Sprocket 40 Journal of Recent Trends in Mechanics Volume 3 Issue 3 A: Cad Model B: Total Deformation C: Equivalent Strain D: 23 EquivalentStress Fig: 9.48 Teeth Sprocket A: Cad Model B: Total Deformation A: Cad Model B: Total Deformation B: Total Deformation A: Cad Model B: Total Deformation B: Total Deformation A: Cad Model A: Cad Model A: C: Equivalent Strain D: 23 EquivalentStress Fig: 9.48 Teeth Sprocket C: Equivalent Strain D: 23 EquivalentStress Fig: 9.48 Teeth Sprocket D: 23 EquivalentStress Rrs = radius of rear sprocket Hence every point on the chain has a linear speed of when engaged with front chain ring. Front sprocket has a linear speed with respect to the rear sprocket i.e. Vfs = 0.27388 m/s, 0.34164 m/sec 0.40620m/sec w.r.t no. of sprocket 16, 20, 24 RPM OF CHAINRING Rrs = radius of rear sprocket Hence every point on the chain has a linear speed of when engaged with front chain ring. Front sprocket has a linear speed with respect to the rear sprocket i.e. Vfs = 0.27388 m/s, 0.34164 m/sec 0.40620m/sec w.r.t no. of sprocket 16, 20, 24 RPM OF CHAINRING Speed = 10 kmph = 2.77m/sec Rear wheel radius = 13 inch (.33 mts) Angular speed of rear wheel ( = 2.77 m/sec/.33mts = 8.417 rad/sec (80.3764 rpm) Therefore angular speed of rear sprocket = 8.417 rad/sec Now, the radius of rear sprocket D = Where P = pitch of roller in chain N = no. of teeth in sprocket Linear speed of rear sprocket = Vrs = Wrs x Rrs Where, Wrs = angular speed of rear sprocket Now for calculating the radius of front sprocket in order to find its angular speed i.e. = Therefore, Wfs = Vfs/Rfs i.e. = Page 34-53 © MAT Journals 2018. All Rights Reserved 41 Journal of Recent Trends in Mechanics Volume 3 Issue 3 RPM OF CHAINRING Table: 5. Rpm ofChainring No. of teeth Diameter (inch) Radius (inch) Radius (mtrs) Vrs= Wrs x Rrs (m/sec) Rfs (mtrs) Wfs (rad/sec) Wfs (revolutions) 16 2.5629 1.2814 0.03254 .27388 .09762 2.8055 28 20 3.1962 1.5981 0.04059 .34164 .09741 3.5072 35 24 3.8300 1.9000 0.04826 .40620 .09652 4.2084 42 RPM OF CHAINRING COMPARISON FOR 15 kmph, 10 kmph& 5 kmph SPEED Table: 6.Comparisonfor 15 kmph, 10 kmph& 5 kmph Speed Speed (kmph) No. Teeth Radius of Sprocket (mtrs) Vrs = Wrs x Rrs (m/sec) Vfs Rfs Wfs = Vfs/Rfs (rad/sec) Wfs (RPM) 15 16 .03254 0.41074 0.41074 0.09762 4.2075 40 15 20 .04059 0.51235 0.51235 0.09741 5.2599 50 15 24 .04826 0.60917 0.60917 0.09652 6.3113 60 10 16 .03254 0.27366 0.27366 0.09762 2.8033 28 10 20 .04059 0.34136 0.34136 0.09741 3.5043 34 10 24 .04826 0.40586 0.40586 0.09652 4.2049 40 5 16 .03254 0.13695 0.13695 0.09762 1.4028 14 5 20 .04059 0.17083 0.17083 0.09741 1.7537 17 5 24 .04826 0.20311 0.20311 0.09652 2.1043 22 COMPARISON FOR 15 kmph, 10 kmph& 5 kmph SPEED Table: 6.Comparisonfor 15 kmph, 10 kmph& 5 kmph S COMPARISON FOR 15 kmph, 10 kmph& 5 kmph SPEED COMPARISON FOR 15 kmph, 10 kmph& 5 kmph SPEED Table: 6 Comparisonfor 15 kmph 10 kmph& 5 COMPARISTION OF THEORITICAL AND PRACTICAL DATA Table: 7.Comparison ofTheoretical and Practical Data Speed No. of teeth Wfs (rad/sec) Wfs (RPM) 15 16 4.2075 40 15 20 5.2599 50 15 24 6.3113 60 10 16 2.8055 28 10 20 3.5072 34 10 24 4.2084 40 5 16 1.4028 14 5 20 1.7537 17 5 24 2.1043 21 COMPARISTION OF THEORITICAL AND PRACTICAL DATA T bl 7 C i fTh ti l d P ti l D t Table: 8.Comparisonof Theoretical and Practical Data Person Age Speed Rpm (average) A 23 16 40 B 24 12 33 C 22 17 41 D 27 13 35 E 32 9 24 42 Page 34-53 © MAT Journals 2018. All Rights Reserved CHAIN DRIVE Chain design for bigger rear cog and chain ring Pitch = 12.7 mm D (chain ring) = 194.1803 mm D (rear cog) = 97.30 Velocity ratio = 3 Average velocity = = .4060m/sec Length of chain (including derailleur) = 140 links = (140 x pitch of chain) = 1778 mm Velocity ratio = 3 Average velocity = = .4060m/sec Length of chain (including derailleur) = 140 links = (140 x pitch of chain) = 1778 mm Page 34-53 © MAT Journals 2018. All Rights Reserved 42 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Journal of Recent Trends in Mechanics Volume 3 Issue 3 For calculating the tension forces Fig: 10. For calculating the tension forces For calculating the tension forces For calculating the tension forces 43 Page 34-53 © MAT Journals 2018. All Rights Reserved Fig: 10. For calculating the tension forces 1, 2 = angle of wrap in smaller & bigger sprocket respectively = angle between horizontal and the last teeth of bigger sprocket engaged with chain According to geometry, Sin = 0.043 1 = 3.05 rad 2 = 3.22 rad No. of teeth in contact = Length of chain = L= kp Where k = no. of links p = pitch of chain K = (T1/T2)/2 + 2x/p + [(T1/T2)/2 ] ^2 x p/x X = p/4[k- (T1/T2)/2 + Average velocity= v= = V (for 48 teeth sprocket) = 0.4063m/sec We know that, Tension in chain is given by, = Also, = R (T1 – T2), Where, = torque T1 (tension at tight side) = 202.4820 N T2 (tension at slag side) = 77.5935 N According to calculation chain no- 428# is to be used Material – ANSI 430 Geometry Fig: 11. Geometry Fig: 10. For calculating the tension forces X = p/4[k- (T1/T2)/2 + 1, 2 = angle of wrap in smaller & bigger sprocket respectively = angle between horizontal and the last teeth of bigger sprocket engaged with chain Average velocity= v= = V (for 48 teeth sprocket) = 0.4063m/sec We know that, According to geometry, According to geometry, We know that, g g y Sin = 0.043 1 = 3.05 rad 2 = 3.22 rad No. of teeth in contact = Length of chain = L= kp Where k = no. of links p = pitch of chain K = (T1/T2)/2 + 2x/p + [(T1/T2)/2 ] ^2 x p/x Sin = 0.043 1 = 3.05 rad 2 = 3.22 rad Tension in chain is given by, = Also, = R (T1 – T2), Where, = torque Length of chain = L= kp p = pitch of chain p/x 43 Page 34-53 © MAT Journals 2018. All Rights Reserved Fig: 11. Geometry Fig: 11. Geometry Page 34-53 © MAT Journals 2018. All Rights Reserved 43 Journal of Recent Trends in Mechanics Volume 3 Issue 3 PROCKET TENSION AND TORQUE CALCULATION AT PROPOSED GEAR ATIOS SPROCKET TENSION AND TORQUE CALCULATION AT PROPOSED GEAR RATIOS Q RATIOS Table: 9. Sprocket Tension And Torque Calculation At Proposed Gear Ratios Sprocket combination Pitch of chain (mm) Length of chain (mm) No. of links No. For calculating the tension forces of teeth in contact Tension (N) Torque (N-m) 16x10x10x48 12.7 1780 137 16 8 T1=205.4280 4.1610 48 25 T2= 77.6 12.3985 10 4 T3=53 2.5564 10 4 T4=53 3.4125 20x10x10x48 12.7 1830 140 20 10 T1=205.4280 5.1771 48 25 T2= 77.6 12.3985 10 4 T3=53 2.5564 10 4 T4=53 3.4426 24x10x10x48 12.7 1880 142 24 12 T1=205.4280 6.1997 48 24 T2= 77.6 12.3985 10 4 T3=53 2.5564 10 4 T4=53 3.4839 A: Cad Model of Chain B: Cad Model of Chain Fig: 12. Chain Model and Analysis A: Cad Model of Chain B: Cad Model of Chain Fig: 12. Chain Model and Analysis SUSPENSION CALCULATION Total weight of vehicle = 275 kg Front weight = 30% of total weight = 82.5 kg Rear weight = 70% of total weight = 192.5 kg Design Of front suspension on 100 kg (18kg factor added) Considering the dynamic or impact loads on spring will be double i.e. 200kg = 1962 N For single spring = 1962/2 = 981 N Total weight of vehicle = 275 kg Total weight of vehicle = 275 kg Bicycle suspension is the system, or systems, used to suspend the rider and bicycle in order to insulate them from the roughness of the terrain. Bicycle suspension is used primarily on mountain bikes, but is also common on hybrid bicycles. kg Design Of front suspension on 100 kg (18kg factor added) Bicycle suspension can be implemented in a variety of ways, and any combination thereof: Considering the dynamic or impact loads on spring will be double i.e. 200kg = 1962 ● Front suspension ● Rear suspension ● Suspension seat post ● Suspension saddle ● Suspension stem (now uncommon) ● Suspension hub For single spring = 1962/2 = 981 N ● Front suspension ● Rear suspension Page 34-53 © MAT Journals 2018. All Rights Reserved Suspension travel Displacement 170mm 170mm 170mm 170 mm 170mm 170mm Length of wire 2160mm 2160mm 2160mm 2160mm 2160mm 2160mm Solid Length 157mm 165mm 160mm 170mm 160mm 172mm Pitch 18.5mm 18.5mm 18.5mm 18.5mm 18.5mm 18.5mm Rise angle of coil 9.74 deg 9.74 deg 9.74 deg 9.74 deg 9.74 deg 9.74 deg Lowest frequency (fres) 8.17 hz 11.8 hz 14 hz 9.18 hz 9.50 hz 9.85Hz Shear modulus of material 0.315 Gpa 0.663 Gpa 0.927 Gpa 0.398 Gpa 0.427 Gpa 0.622 Gpa Mass of spring 0.855 kg 0.855 kg 0.855 kg 0.855 kg 0.855 kg 0.855 kg REAR SUSPENSION DESIGN Table: 11.Rear Suspension Design Oil tampered Music wire ASTM A228 High Carbon Wire AS 8 Diameter of wire 12mm 12mm 12mm 12mm 12mm 12mm Outer dia. Of spring 76mm 76mm 76mm 76mm 76mm 76mm Free length of spring 430mm 430mm 430mm 430mm 430mm 430mm No. of active coils 12 12 12 12 12 12 Young’s modulus (E) 1070 Mpa 2070 Mpa 1725 Mpa 3725 Mpa 1035 Mpa 2060 Mpa Modulus Of Torsion (G) 79 Gpa 79 Gpa 79 Gpa 82 Gpa 79 Gpa 79 Gpa Poisson’s ratio 0.29 0.29 0.33 0.33 0.3 0.3 Density of material 0.284 lbm/in^3 0.284 lbm/in^3 0.284 lbm/in^3 0.284 lbm/in^3 0.284 lbm/in^3 0.284 lbm/in^3 Spring Constant 3.39 x 10^5 N/m 6.56 x 10^5 N/m 5.47 x 10^6 N/m 1.18x10^6 N/m 2.8x10^6 N/m 5.95x10^5 N/m Max. Load possible 8.88 x 10^4 N 1.72 X10^5 N 1.43x10^5 N 3.09 x 10^5 N 7.8 x 10^4 N 1.68 x 10^5 N Max. Shear Stress 1.08 x 10^4 Mpa 2.09 x 10^4 Mpa 1.74 x 10^4 Mpa 1.08 x 10^4 Mpa 0.9 x 10^4 Mpa 1.98 x 10^4 Mpa Max. Page 34-53 © MAT Journals 2018. All Rights Reserved Suspension travel Displacement 262 mm 262 mm 262 mm 262 mm 262 mm 262 mm Length of wire 2850 mm 2850 mm 2850 mm 2850 mm 2850 mm 2850 mm Solid Length 168 mm 168 mm 168 mm 168 mm 168 mm 168 mm Pitch 35.8 deg 35.8 deg 35.8 deg 35.8 deg 35.8 deg 35.8 deg Rise angle of coil 10.1 deg 10.1 deg 10.1 deg 10.1 deg 10.1 deg 10.1 deg Lowest frequency (fres) 183 hz 254 hz 232 hz 341 hz 175 hz 246hz Shear modulus 412 gpa 796 Gpa 232 Gpa 341 Gpa 385 Gpa 720 Gpa Journal of Recent Trends in Mechanics Volume 3 Issue 3 material lbm/in^3 lbm/in^3 lbm/in^3 lbm/in^3 Spring Constant 228N/m 480N/m 671 N/m 288N/m 309 N/m 650N/m Max. Load possible 49 N 38 N 114 N 49N 52.5 N 68.3N Max. Shear Stress 11.4 Mpa 9.03 Mpa 26.5 Mpa 11.4 Mpa 12.2 Mpa 17.8Mpa Max. Displacement 170mm 170mm 170mm 170 mm 170mm 170mm Length of wire 2160mm 2160mm 2160mm 2160mm 2160mm 2160mm Solid Length 157mm 165mm 160mm 170mm 160mm 172mm Pitch 18.5mm 18.5mm 18.5mm 18.5mm 18.5mm 18.5mm Rise angle of coil 9.74 deg 9.74 deg 9.74 deg 9.74 deg 9.74 deg 9.74 deg Lowest frequency (fres) 8.17 hz 11.8 hz 14 hz 9.18 hz 9.50 hz 9.85Hz Shear modulus of material 0.315 Gpa 0.663 Gpa 0.927 Gpa 0.398 Gpa 0.427 Gpa 0.622 Gpa Mass of spring 0.855 kg 0.855 kg 0.855 kg 0.855 kg 0.855 kg 0.855 kg Suspension travel For city bicycles- 100 to 230 mm (data from recent manufactured springs) For city bicycles- 100 to 230 mm (data from recent manufactured springs) Page 34-53 © MAT Journals 2018. All Rights Reserved 44 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Journal of Recent Trends in Mechanics Volume 3 Issue 3 A: Cad Model B: Total Deformation C: Equivalent Elastic Strain D: 31 Equivalent Stress Fig: 13.Cad Model and Analysis of Front Fork Spring Material Calculation Report Table: 10. Material Calculation Report Oil tampered Music wire ASTM A228 High Carbon WireAS 8 Diameter of wire 8mm 8mm 8mm 8mm 8mm 8mm Outer dia. Of spring 42mm 42mm 42mm 42mm 42mm 42mm F l th f A: Cad Model B: Total Deformation B: Total Deformation A: Cad Model C: Equivalent Elastic Strain D: 31 Equivalent Stress Fig: 13.Cad Model and Analysis of Front Fork Spring C: Equivalent Elastic Strain D: 31 Equivalent Stress Fig: 13.Cad Model and Analysis of Front Fork Spring C: Equivalent Elastic Strain D: 31 Equivalent Stress Fig: 13.Cad Model and Analysis of Front Fork Spring Material Calculation Report Table: 10. Material Calculation Report Oil tampered Music wire ASTM A228 High Carbon WireAS 8 Diameter of wire 8mm 8mm 8mm 8mm 8mm 8mm Outer dia. Of spring 42mm 42mm 42mm 42mm 42mm 42mm Free length of spring 330mm 330mm 330mm 330mm 330mm 330mm No. of active coils 18 18 18 18 18 18 Young’s modulus of material (E) 820 Mpa 1725 Mpa 2410 Mpa 1035 Mpa 1110Mpa 1450Mpa Modulus OfTorsion 79 Gpa 79 Gpa 79 Gpa 82 Gpa 79 Gpa 79 Gpa Poisson’s ratio 0.29 0.29 0.33 0.33 0.3 0.3 Density of 0.284 0.284 0.284 lbm/in^3 0.284 0.284 lbm/in^3 0.284 Material Calculation Report T bl Material Calculation Report 45 Journal of Recent Trends in Mechanics Volume 3 Issue 3 material lbm/in^3 lbm/in^3 lbm/in^3 lbm/in^3 Spring Constant 228N/m 480N/m 671 N/m 288N/m 309 N/m 650N/m Max. Load possible 49 N 38 N 114 N 49N 52.5 N 68.3N Max. Shear Stress 11.4 Mpa 9.03 Mpa 26.5 Mpa 11.4 Mpa 12.2 Mpa 17.8Mpa Max. REAR SUSPENSION DESIGN Table: 11.Rear Suspension Design Oil tampered Music wire ASTM A228 High Carbon Wire AS 8 Diameter of wire 12mm 12mm 12mm 12mm 12mm 12mm Outer dia. Of spring 76mm 76mm 76mm 76mm 76mm 76mm Free length of spring 430mm 430mm 430mm 430mm 430mm 430mm No. of active coils 12 12 12 12 12 12 Young’s modulus (E) 1070 Mpa 2070 Mpa 1725 Mpa 3725 Mpa 1035 Mpa 2060 Mpa Modulus Of Torsion (G) 79 Gpa 79 Gpa 79 Gpa 82 Gpa 79 Gpa 79 Gpa Poisson’s ratio 0.29 0.29 0.33 0.33 0.3 0.3 Density of material 0.284 lbm/in^3 0.284 lbm/in^3 0.284 lbm/in^3 0.284 lbm/in^3 0.284 lbm/in^3 0.284 lbm/in^3 Spring Constant 3.39 x 10^5 N/m 6.56 x 10^5 N/m 5.47 x 10^6 N/m 1.18x10^6 N/m 2.8x10^6 N/m 5.95x10^5 N/m Max. Load possible 8.88 x 10^4 N 1.72 X10^5 N 1.43x10^5 N 3.09 x 10^5 N 7.8 x 10^4 N 1.68 x 10^5 N Max. Shear Stress 1.08 x 10^4 Mpa 2.09 x 10^4 Mpa 1.74 x 10^4 Mpa 1.08 x 10^4 Mpa 0.9 x 10^4 Mpa 1.98 x 10^4 Mpa Max. Pedal Force (Fp) = 112.5 N  (According to driver’s ergonomic consideration, driver will exert 1.5W per kg work of his weight by hands)Disk Dimension :- Diameter = 203 mm (8 inch)  Thickness = 3 mm Material Used to manufacture CAST IRON Stopping Distance of the vehicle = Thinking Distance +Braking Distance Fig: 16.Vehicle Brakes Calculation Thinking Distance = u X 0.33 Where, u = Velocity at which vehicle is moving when brakes are applied Assuming, u = 10 kmph = 2.77 m/s (max. vehicle velocity) Thinking Distance = 2.77 X 0.33 = 0.9141 m Braking Distance = (v²-u²)/2a Where, v = final velocity = 0 u = Velocity at which vehicle is moving when brakes are applied d l ti t 0 75 X 9 81 VEHICLE BRAKES CALCULATION  Total Mass of vehicle = 275 kg  Acceleration due to gravity (g) = 9.81 m/s²  Acceleration due to gravity (g) = 9.81 m/s² u = Velocity at which vehicle is moving when brakes are applied 2 All Rights Reserved 47 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Kinetic Energy (E) = 1/2mu2 Where, m = mass of the vehicle with driver u = Velocity at which vehicle is moving when brakes are applied E = 1/2 X 275 X (2.772) E = 1055 J Kinetic Energy (E) = 1/2mu2 Where, m = mass of the vehicle with driver VEHICLE BRAKES CALCULATION  Total Mass of vehicle = 275 kg  Acceleration due to gravity (g) = 9.81 m/s²  Coefficient of friction between road and tyres (µ) = 0.75  Normal force on Wheels (N) = µmg  N = 0.75 X 275 X 9.81  N = 2023.3125 ~ 2024 newton  Force applied to the brake lever by the driver i.e. Pedal Force (Fp) = 112.5 N  (According to driver’s ergonomic consideration, driver will exert 1.5W per kg work of his weight by hands)Disk Dimension :- Diameter = 203 mm (8 inch)  Thickness = 3 mm Material Used to manufacture CAST IRON Stopping Distance of the vehicle = Thinking Distance +Braking Distance Fig: 16.Vehicle Brakes Calculation Thinking Distance = u X 0.33 Where, u = Velocity at which vehicle is moving when brakes are applied Assuming, u = 10 kmph = 2.77 m/s (max. vehicle velocity) Thinking Distance = 2.77 X 0.33 = 0.9141 m Braking Distance = (v²-u²)/2a Where, v = final velocity = 0 u = Velocity at which vehicle is moving when brakes are applied a = deceleration rate = g = 0.75 X 9.81 = 7 35 / VEHICLE BRAKES CALCULATION VEHICLE BRAKES CALCULATION  Total Mass of vehicle = 275 kg  Acceleration due to gravity (g) = 9.81 m/s²  Coefficient of friction between road and tyres (µ) = 0.75  Normal force on Wheels (N) = µmg  N = 0.75 X 275 X 9.81  N = 2023.3125 ~ 2024 newton  Force applied to the brake lever by the driver i.e. rnals 2018. All Rights Reserved E = 1055 J Braking Force = µ X (Kinetic Energy/Braking Distance) Fb = 0.75 X (1055/0.5219) Fb = 1516.129 N Braking Torque = Braking Force X Radius of rotor disc Tb = 1516.129 X 0.1015 Tb = 154 N-m We know that, Braking Force (Bf) = µ X (Kinetic Energy/Braking Distance) Or, Braking force = M.a.g Where, M = total vehicle mass (kg) a = deacceleration (m/sec^2) g = acc. Due to gravity (m/sec^2) Therefore, 1516.1129 = 275 x a x 9.81 a = .56 m/sec^2 Now equation of motion says, v – u = at 2.77 – 0 = a.t 2.77 / .56 = t t = 4.92 sec = 5 sec Now, we know the acceleration we can find the distance we went in those 5 sec, i.e. Distance (S) = u.t + ½ (at^2) S = 2.77 x 5 + ½ (0.56) (5^2) S = 6.925 mts STEERING Steering Geometry Fig: 17. Steering Geometry Braking Force = µ X (Kinetic Energy/Braking Distance) Fb = 0.75 X (1055/0.5219) Fb = 1516.129 N Braking Torque = Braking Force X Radius of rotor disc Tb = 1516.129 X 0.1015 Tb = 154 N-m We know that, Braking Force (Bf) = µ X (Kinetic Energy/Braking Distance) Or, Braking force = M.a.g Where, M = total vehicle mass (kg) a = deacceleration (m/sec^2) g = acc. Due to gravity (m/sec^2) Therefore, 1516.1129 = 275 x a x 9.81 a = .56 m/sec^2 Now equation of motion says v – u = at  Force applied to the brake lever by the driver i.e. Pedal Force (Fp) = 112.5 N  (According to driver’s ergonomic consideration, driver will exert 1.5W per kg work of his weight by hands)Disk Dimension :- Diameter = 203 mm (8 inch) Material Used to manufacture CAST IRON Stopping Distance of the vehicle = Thinking Distance +Braking Distance Fig: 16.Vehicle Brakes Calculation Displacement 262 mm 262 mm 262 mm 262 mm 262 mm 262 mm Length of wire 2850 mm 2850 mm 2850 mm 2850 mm 2850 mm 2850 mm Solid Length 168 mm 168 mm 168 mm 168 mm 168 mm 168 mm Pitch 35.8 deg 35.8 deg 35.8 deg 35.8 deg 35.8 deg 35.8 deg Rise angle of coil 10.1 deg 10.1 deg 10.1 deg 10.1 deg 10.1 deg 10.1 deg Lowest frequency (fres) 183 hz 254 hz 232 hz 341 hz 175 hz 246hz Shear modulus of material 412 gpa 796 Gpa 232 Gpa 341 Gpa 385 Gpa 720 Gpa Mass of spring 2.54 kg 2.54 kg 2.54 kg 2.54 kg 2.54 kg 2.54 kg 46 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Selected Material: Music Wire ASTM A228 (suitable for the spring used in project) A:CAD model B: 33 Total deformation C:Equivalent Strain D: Equivalent Stress Fig: 14.Cad Model and Analysis of Rear Spring Selected Material: Music Wire ASTM A228 (suitable for the spring used in project) elected Material: Music Wire ASTM A228 (suitable for the spring used in project) Selected Material: Music Wire ASTM A228 (suitable for the spring used in project) A:CAD model B: 33 Total deformation Selected Material: Music Wire ASTM A228 (suitable for the spring used in project) A:CAD model B: 33 Total deformation B: 33 Total deformation B: 33 Total deformation B: 33 Total deformation A:CAD model C:Equivalent Strain D: Equivalent Stress Fig: 14.Cad Model and Analysis of Rear Spring D: Equivalent Stress s of Rear Spring D: Equivalent Stress C:Equivalent Strain Fig: 14.Cad Model and Analysis of Rear Spring 47 P 34 53 © MAT J l 2018 All Ri ht R d A: Disc Brake B: Brake Level Fig: 15. Brakes B: Brake Level A: Disc Brake Fig: 15. Brakes Page 34-53 © MAT Journals 2018. All Rights Reserved Page 34-53 © MAT Journals 2018. Page 34-53 © MAT Journals 2018. All Rights Reserved Material Used to manufacture CAST IRON 2.77 – 0 = a.t 2.77 / .56 = t Thinking Distance = u X 0.33 Thinking Distance = u X 0.33 g Where, u = Velocity at which vehicle is moving when brakes are applied Assuming, u = 10 kmph = 2.77 m/s (max. vehicle velocity) vehicle velocity) Thinking Distance = 2.77 X 0.33 = 0.9141 m Braking Distance = (v²-u²)/2a Where, v = final velocity = 0 u = Velocity at which vehicle is moving when brakes are applied a = deceleration rate = g = 0.75 X 9.81 = 7.35 m/s Braking Distance = (2.77)²/ (2 X 7.35) = 0.5219 m Stopping Distance = 0.5219 + 0.9141 = 1.436 m Time to stop vehicle (ts) = (v-u) / a ts = 0.5219/7.35 = 0.07100 seconds Thinking Distance = 2.77 X 0.33 = 0.9141 m Braking Distance = (v²-u²)/2a Where, v = final velocity = 0 u = Velocity at which vehicle is moving when brakes are applied a = deceleration rate = g = 0.75 X 9.81 = 7.35 m/s Braking Distance = (2.77)²/ (2 X 7.35) = 0.5219 m Stopping Distance = 0.5219 + 0.9141 = 1.436 m Time to stop vehicle (ts) = (v-u) / a ts = 0.5219/7.35 = 0.07100 seconds STEERING Steering Geometry TEERING teering Geometry Fig: 17. Steering Geometry Fig: 17. Steering Geometry Page 34-53 © MAT Journals 2018. Material Used to manufacture CAST IRON All Rights Reserved 48 Journal of Recent Trends in Mechanics Volume 3 Issue 3 According to geometry Solving for Beta: Substituting into the equation for fork offset B yields the following quadratic: (R2 + N2) * cos2Beta - 2 * R * B * cos Beta + B2 - N2 = 0 => "minus" for N <0 The steering angle at which trail becomes zero (turns to negative values): If B>0 (fork bent forward): If B<0 (fork bent backward): Project Data Beta = 75 deg B= 2.663 cm N = 4.673 cm Further results from the above steering geometry data: Trail passes zero (turns to negative values) if the steering is turned by 69 degrees Now from geometry, Journal of Recent Trends in Mechanics Volume 3 Issue 3 According to geometry Solving for Beta: Substituting into the equation for fork offset B yields the following quadratic: (R2 + N2) * cos2Beta - 2 * R * B * cos Beta + B2 - N2 = 0 => "minus" for N <0 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Solving for Beta: Substituting into the equation for fork offset B yields the following quadratic: (R2 + N2) * cos2Beta - 2 * R * B * cos Beta + B2 - N2 = 0 "minus" for N <0 The steering angle at which trail becomes zero (turns to negative values): The steering angle at which trail becomes zero (turns to negative values): If B>0 (fork bent forward): If B<0 (fork bent backward): Project Data Beta = 75 deg B= 2.663 cm N = 4.673 cm Further results from the above steering geometry data: If B>0 (fork bent forward): If B<0 (fork bent backward): Project Data Beta = 75 deg B= 2.663 cm N = 4.673 cm Further results from the above steering geometry data: g g y Trail passes zero (turns to negative values) if the steering is turned by 69 degrees Now from geometry, g g y Trail passes zero (turns to negative values) if the steering is turned by 69 degrees Now from geometry, Trail passes zero (turns to negative values) if the steering is turned by 69 degrees Now from geometry, 49 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Volume 3 Issue 3 Fig: 18. Page 34-53 © MAT Journals 2018. All Rights Reserved Material Used to manufacture CAST IRON Project Data Cot ∆= w/l ∆ = cot^-1 (w/l) According to design, W = 44 inch L = 65 inch Therefore, ∆ = 0.80 degree Turning Radius:- Therefore, Turning Radius of vehicle (R) = 1.9177 mt 24 inch wheel used in front side of vehicle26 inch wheel used in rear side of vehicle 50 Page 34-53 © MAT Journals 2018. All Rights Reserved Fig: 18. Project Data Cot ∆= w/l ∆ = cot^-1 (w/l) According to design, W = 44 inch L = 65 inch Therefore, ∆ = 0.80 degree Turning Radius:- Therefore, Turning Radius of vehicle (R) = 1.9177 mt 24 inch wheel used in front side of vehicle26 inch wheel used in rear side of vehicle Fig: 19. Front WheelFig: 20. Rear Wheel Fig: 18. Project Data Fig: 18. Project Data ∆ = 0.80 degree Turning Radius:- Therefore, Turning Radius of vehicle (R) = 1.9177 mt Cot ∆= w/l ∆ = cot^-1 (w/l) According to design, W = 44 inch L = 65 inch Therefore, 4 inch wheel used in front side of vehicle26 inch wheel used in rear side of vehicle 24 inch wheel used in front side of vehicle26 inch wheel used in rear side of vehicle 24 inch wheel used in front side of vehicle26 inch wheel used in rear side of vehicle Fig: 19. Front WheelFig: 20. Rear Wheel Fig: 19. Front WheelFig: 20. Rear Wheel Page 34-53 © MAT Journals 2018. All Rights Reserved 50 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Journal of Recent Trends in Mechanics Volume 3 Issue 3 optional basis i.e. optional basis i.e. 1. Charging Socket for Mobile 1. Charging Socket for Mobile 2. For music player (music player of alternate source of charging can also be used i.e. inbuilt battery system) 51 Page 34-53 © MAT Journals 2018. All Rights Reserved Indicators – 1 watt led, 2nos. for both side Other features can be provided but on be used i.e. Page 34-53 © MAT Journals 2018. All Rights Reserved Journal of Recent Trends in Mechanics Volume 3 Issue 3 Material Used to manufacture CAST IRON inbuilt battery system Fig: 21.Circuit Diagram A: CAD Model of FrameB: total deformation Head lamp Tail Lamp Switch Switch Fuse BATTERY DYNAMO ON/OFF Flash L R Fig: 21.Circuit Diagram Head lamp Tail Lamp Switch Switch Fuse BATTERY DYNAMO ON/OFF Flash L R Tail Lamp Head lamp Switch Switch ON/OFF DYNAMO Fig: 21.Circuit Diagram Fig: 21.Circuit Diagram A: CAD Model of FrameB: total deformation A: CAD Model of FrameB: total deformation A: CAD Model of FrameB: total deformation 51 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Journal of Recent Trends in Mechanics Volume 3 Issue 3 C:Equivalent StressD: Equivalent Strain E: 48 Safety Factor Fig: 22.Frame Model and Analysis C:Equivalent StressD: Equivalent Strain C:Equivalent StressD: Equivalent Strain E: 48 Safety Factor Fig: 22.Frame Model and Analysis E: 48 Safety Factor y Fig: 22.Frame Model and Analysis  Using the parts of proper standards helps in availability of parts and leads to proper maintenance. By implementing the mentioned changes in the vehicle, not only provide employment but reduces the dependences on fossil fuels which directly reduces the major problem of todays society i.e. pollution. The vehicle as made cost effectively and easily maintenance for the riders belongs to low literacy rate. By implementing the mentioned changes in the vehicle, not only provide employment but reduces the dependences on fossil fuels which directly reduces the major problem of todays society i.e. pollution. The vehicle as made cost effectively and easily maintenance for the riders belongs to low literacy rate.  Calculation of braking load and torque helps to calculate the forces acting at the time of braking.  Introducing the suspension leads to minimize the forces and provide the better comfort to the driver as well as passengers Page 34-53 © MAT Journals 2018. All Rights Reserved CONCLUSION The whole project was to develop or upgrade the current pedicabs for both driver as well as passengers. To provide comfort the existing design is developed with ergonomic standard and introduced with gear shifting mechanism, which helps to maintain the speed as well as to carry heavy load even at inclined surfaces. Special compartments under the seat have been provided to keep luggage. Whole vehicle is provided with weather protection unit to avoid excessive sunlight and rain to both passenger as well as rider. Addition features such as battery unit, GPS and electrical system has been installed in the vehicle to make it more efficient. 7. J. Prakash and G. Habib, “A technology-based mass emission factors of gases and aerosol precursor and spatial distribution of emissions from on-road transport sector in India,” Atmos. Environ., vol. 180, pp. 192– 205, 2018. 8. S. Mohanty, S. Bansal, and K. Bairwa, “Effect of integration of bicyclists and pedestrians with transit in New Delhi,” Transp. Policy, vol. 57, no. April 2016, pp. 31–40, 2017. pp 9. I. Mateo-Babiano, S. Kumar, and A. Mejia, “Bicycle sharing in Asia: A stakeholder perception and possible futures,” Transp. Res. Procedia, vol. 25, pp. 4970–4982, 2017. pp 10. Jaiprakash and G. Habib, “On-road assessment of light duty vehicles in Delhi city: Emission factors of CO, CO2and NOX,” Atmos. Environ., vol. 174, no. July 2017, pp. 132–139, 2018. RESULTS  By providing the weather protection helps to increase the usage of vehicle in all seasons  By using standard design of pedicabs we can improve the ergonomic factors, and it reduces the health issues.  Luggage compartment helps to carry minimize luggage.  Proper calculation of forces and examine the forces on riders and vehicle will provide the comfort to both driver and passenger.  Safety features like GPS, Utility box and theft protection helps the vehicle to meets the present context.  Introducing the gear ratios instead of single ratio provides the comfort to the driver.  Extremely Reliable  Simple to maintain  Changing the material leads to minimization on weight. Page 34-53 © MAT Journals 2018. All Rights Reserved 52 Journal of Recent Trends in Mechanics Volume 3 Issue 3 Journal of Recent Trends in Mechanics Volume 3 Issue 3 cities,” J. Transp. Geogr., vol. 58, pp. 40–51, 2017. Page 34-53 © MAT Journals 2018. All Rights Reserved Journal of Recent Trends in Mechanics Volume 3 Issue 3 REFERENCES 1. E. J. Zolnik, A. Malik, and Y. Irvin- Erickson, “Who benefits from bus rapid transit? Evidence from the Metro Bus System (MBS) in Lahore,” J. Transp. Geogr., vol. 71, no. June 2017, pp. 139–149, 2018. 11. S. O. N. Merits, B. Operated, E. Bike, W. Motor, O. Rickshaw, and S. City, “Icmere2015-Pi-225 Study on Merits and Demerits of Two Transport Systems :,” vol. 2015, pp. 26–29, 2015. pp 2. S. Wijaya, M. Imran, and J. McNeill, “Socio-political tensions in Bus Rapid Transit (BRT) development in low- income Asian cities,” Dev. Pract., vol. 25, pp. 5104–5120, 2018. y pp 12. M. M. U. Hasan and J. D. Dávila, “The politics of (im)mobility: Rickshaw bans in Dhaka, Bangladesh,” J. Transp. Geogr., vol. 70, no. May, pp. 246–255, 2018. 3. S. Tabassum, S. Tanaka, F. Nakamura, and A. Ryo, “Feeder Network Design for Mass Transit System in Developing Countries (Case study of Lahore, Pakistan),” Transp. Res. Procedia, vol. 25, pp. 3133–3150, 2017. 13. S. Gupta, “Role of Non -Motorized Transport in Distribution of Goods in the Metropolitan City of Delhi,” Transp. Res. Procedia, vol. 25, pp. 978–984, 2017. 4. N. Singh and V. Vasudevan, “Understanding school trip mode choice – The case of Kanpur (India),” J. Transp. Geogr., vol. 66, no. December 2017, pp. 283–290, 2018. 14. M. R. Hickman, “A Study on Power Assists for Bicycle Rickshaws in India, including Fabrication of Test Apparatus,” p. 47, 2011. 5. R. Sharma and P. Newman, “Can land value capture make PPP’s competitive in fares? A Mumbai case study,” Transp. Policy, vol. 64, no. July 2017, pp. 123–131, 2018. 15. S. Rana, F. Hossain, S. S. Roy, and S. K. Mitra, “Exploring operational Characteristics of Battery operated Auto- Rickshaws in Urban Transportation System,” Am. J. Eng. Res., no. 4, pp. 1–11, 2013. 6. T. M. Rahul and A. Verma, “The influence of stratification by motor- vehicle ownership on the impact of built environment factors in Indian Page 34-53 © MAT Journals 2018. All Rights Reserved 53
https://openalex.org/W4290759722	https://www.frontiersin.org/articles/10.3389/fonc.2022.944589/pdf	English	null	Long-term oncological outcomes of oncoplastic breast-conserving surgery after a 10-year follow-up – a single center experience and systematic literature review	Frontiers in oncology	2,022	cc-by	8,342	TYPE Original Research PUBLISHED 09 August 2022 DOI 10.3389/fonc.2022.944589 TYPE Original Research PUBLISHED 09 August 2022 DOI 10.3389/fonc.2022.944589 OPEN ACCESS OPEN ACCESS EDITED BY Rene´ Aloisio Da Costa Vieira, Barretos Cancer Hospital, Brazil REVIEWED BY Idam Oliveira Junior, Barretos Cancer Hospital, Brazil Fa´ bio Bagnoli, Santa Casa of Sao Paulo, Brazil CORRESPONDENCE Ho Yong Park phy123@knu.ac.kr †These authors have contributed equally to this work and share ﬁrst authorship SPECIALTY SECTION This article was submitted to Surgical Oncology, a section of the journal Frontiers in Oncology RECEIVED 15 May 2022 ACCEPTED 07 July 2022 PUBLISHED 09 August 2022 Jun Xian Hing 1,2,3†, Byeong Ju Kang 1†, Hee Jung Keum 1, Jeeyeon Lee 1, Jin Hyang Jung 1, Wan Wook Kim 1, Jung Dug Yang 1,4, Joon Seok Lee 1,4 and Ho Yong Park 1 Jun Xian Hing 1,2,3†, Byeong Ju Kang 1†, Hee Jung Keum 1, Jeeyeon Lee 1, Jin Hyang Jung 1, Wan Wook Kim 1, Jung Dug Yang 1,4, Joon Seok Lee 1,4 and Ho Yong Park 1* 1Department of Surgery, School of Medicine, Kyungpook National University, Kyungpook National University Chilgok Hospital, Daegu, South Korea, 2Division of Breast Surgery, Department of General Surgery, Changi General Hospital, Singapore, Singapore, 3Singhealth Duke-NUS Breast Centre, Singapore Health Services Pte Ltd, Singapore, Singapore, 4Department of Plastic and Reconstructive Surgery, School of Medicine, Kyungpook National University, Daegu, South Korea Aim: While many studies reported the oncological outcomes of oncoplastic breast-conserving surgery (OBCS), there were inherent differences in the study population, surgeons’ expertise, and classiﬁcations of techniques used. There were also limited studies with long term follow up oncological outcomes beyond 5 years. This current study aimed to compare long-term oncological outcomes of ipsilateral breast tumor recurrence (IBTR) disease-free survival (DFS) and overall survival (OS) following conventional and oncoplastic breast- conserving surgery using volume displacement and replacement techniques. Hing JX, Kang BJ, Keum HJ, Lee J, Jung JH, Kim WW, Yang JD, Lee JS and Park HY (2022) Long-term oncological outcomes of oncoplastic breast-conserving surgery after a 10-year follow-up – a single center experience and systematic literature review. Front. Oncol. 12:944589. doi: 10.3389/fonc.2022.944589 COPYRIGHT Methods: Between 2009 and 2013, 539 consecutive patients who underwent breast conservation surgery including 174 oncoplastic and 376 conventional procedures were analysed. A systematic review of studies with at least ﬁve years of median follow up were performed to compare long term oncological outcomes. © 2022 Hing, Kang, Keum, Lee, Jung, Kim, Yang, Lee and Park. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Results: At a median follow-up of 82.4 months, there were 23 (4.2%) locoregional recurrences, 17 (3.2%) metachronous contralateral breast cancer, 26 (4.8%) distant metastases, and 13 (2.4%) deaths. The hazard ratio of OBCS for IBTR, DFS and OS were 0.78 (95% conﬁdence interval [CI] 0.21– 2.94, p=0.78), 1.59 (95% CI, 0.88 to 2.87, p=0.12), and 2.1 (95% CI, 0.72 to 5.9, p=0.17) respectively. The 10-year IBTR-free, DFS and OS rate were 97.8%, 86.2%, and 95.7% respectively. Conclusion: There remained a dearth in well-balanced comparative studies with sufﬁcient long-term follow-up, and our study reported long-term Frontiers in Oncology Frontiers in Oncology 01 frontiersin.org Hing et al. Hing et al. 10.3389/fonc.2022.944589 oncological outcomes for OBCS which were favourable of either VD or replacement techniques. oncoplastic, breast-conserving surgery, oncological outcomes, volume displacement, volume replacement Introduction 20), this current study aimed to compare long-term oncological safety following conventional BCS (CBCS) and OBCS, focusing on overall survival (OS), ipsilateral breast tumor recurrence (IBTR) rates, and disease-free survival (DFS). We also report the rate of positive margins (PMR) detected in intraoperative frozen sections and eventual rate of conversion to mastectomy (CMR) following BCS during a 10-year follow-up period. A literature review was performed to discuss the available data on long-term oncological outcomes with at least ﬁve years median follow up duration reported to date and how our results compared with those of other centers. Historical data have shown that breast-conserving surgery followed by radiotherapy has equivalent oncological outcomes to those of mastectomy in early breast cancer (1, 2). As long-term survival after breast cancer treatment has become commonplace, more attention has been given to develop oncoplastic techniques to provide better patient and aesthetic satisfaction (3, 4). The primary role of oncoplastic breast-conserving surgery (OBCS) is to achieve oncological safety while minimizing the risk of unacceptable local deformity by allowing reconstruction of the defect and preventing the need for mastectomy (5, 6). Following the inception of tumor-speciﬁc immediate breast reconstruction more than 20 years ago, Werner Audretsch coined the term oncoplastic surgery, and many international experts contributed to the burgeoning ﬁeld of OBCS (5, 7–11). Despite the similarity in rationale behind various oncoplastic techniques, there remained differences across geographical locations in terms of surgeons’ perspectives and practices in deﬁning OBCS (12–14). Clough described a classiﬁcation based on tumor volume, location, and glandular density, while Hoffmann and Wallweiner divided breast cancer surgery into two broad types with six tiers, each of increasing complexity (13, 14). A notable consensus deﬁnition came from the American Society of Breast Surgeons, which stated that OBCS incorporated oncologic partial mastectomy with ipsilateral defect repair using volume displacement (VD) and volume replacement (VR) techniques, with contralateral symmetry surgery as appropriate (11). For small-to-moderate breast volumes, however, there was also a difference in technical considerations compared with those for larger breast volumes, which require signiﬁcantly more VR techniques (6, 15–17). Frontiers in Oncology Methods We analyzed prospectively collected data from 539 consecutive breast cancer patients at Kyungpook National University Chilgok Hospital who underwent breast conservation surgery performed by four breast surgeons between January 2009 and December 2013. Treatment strategy was coordinated at multidisciplinary board discussions, which included breast surgeons, plastic surgeons, radiologists, pathologists, and medical and radiation oncologists. All breast conservation surgeries were performed by the breast surgeons, with oncoplastic techniques performed by either a breast or plastic surgeon. A literature review was performed to summarise suitable studies for comparison of deﬁnitions and reported oncological outcomes (Tables 6, 7) (21–40). A search was conducted through the MEDLINE database using PubMed in March 2022. Our search terms included ‘oncoplastic’ [All Fields] AND (‘breast’ [MeSH Terms] OR ‘breast’ [All Fields]) AND (‘surgery’ [Subheading] OR ‘surgery’ [All Fields] OR ‘surgical procedures, operative’ [MeSH Terms]) AND (‘oncological’ [All Fields] OR ‘outcomes’ [All Fields]). A manual search of bibliographies of relevant articles was performed. Korea had been an early adopter of oncoplastic surgery but long-term follow up data remained limited. As with any surgical procedures, long-term follow up was necessary to establish safety parameters of surgical techniques. Furthermore, locoregional recurrences after breast conservation surgery could occur later than mastectomy, perhaps due to the differences in biology or presentations that led to a decision for mastectomy (18). Having previously examined the short-term oncological safety and patient-reported outcomes of various OBCS techniques (17– We included single center studies reporting on various oncoplastic breast conserving surgery to ensure consistency in the reported surgical procedures. Studies with cohort size less than 50 were deemed too small; similarly, a follow up period less than 60 months inadequate to capture late recurrences and death events and hence excluded. Case series or cohort studies reporting on particular Frontiers in Oncology 02 frontiersin.org frontiersin.org Hing et al. 10.3389/fonc.2022.944589 tumor type, pathological tumor size, pathological tumor, nodal stage, receptor status, grade, presence of neoadjuvant and adjuvant therapy, metachronous contralateral breast cancer, locoregional and distant disease recurrences, and death were recorded (Table 2). surgical techniques were also excluded as they were not generalisable to all oncoplastic breast conserving surgery. A PRISMA ﬂowchart is available as supplementary material. Follow-up CBCS involved a direct skin incision, including use of a parallelogram incision overlying the index tumor to allow direct parenchymal closure. Following excision of primary breast tumors with gross margins, a frozen section of the circumferential margins was processed. The defect was closed primarily without further mobilization. When tumor cells were detected on the frozen section, more extensive resection was performed until negative frozen section results were achieved or no further surgical margins were deemed necessary. A ﬁnal parafﬁn block of the surgical margins was examined by pathologists for the presence of tumor cells, and the presence of no stained tumor cell was deﬁned as a negative resection margin. Patients were followed up after surgery using a standardized protocol. After completing adjuvant treatments, frequency of follow-up was biannually for the ﬁrst 2 years and annually for 5 to 10 years. Locoregional recurrence or distant metastasis was evaluated with clinical examination, blood tests including tumor markers, mammography, breast ultrasonography, with or without magnetic resonance imaging, bone scans, and positron emission tomography/computed tomography. Oncological outcomes OBCS was performed as described previously in detail based on general principles of oncoplastic breast surgery in small-to- moderate-sized breasts (15–17). The oncological outcomes assessed include OS, DFS with disease events deﬁned as local or regional recurrences, distant recurrences, and metachronous contralateral breast cancer. The procedures were divided into VD and replacement techniques. VD techniques included dual-plane glandular ﬂap mobilization-closure, purse string suture closure of central defect, roundblock mastopexy, tennis racket incision, batwing mastopexy, rotating ﬂap, and reduction mammoplasty (Table 1) (15). Statistical analysis Statistical analysis was performed in March 2022 using Stata software, v17.0 (StataCorp); a statistically signiﬁcant difference was concluded when p<0.05. Categorical variables were analyzed using the chi-square or Fisher exact test. Continuous variables were analyzed using the Student’s t-test. The Kaplan–Meier method was used to estimate survival function, and the log- rank test was used to compare survival functions. The univariate Cox proportional hazard regression model was also used to examine the correlation of clinically relevant covariates that were likely to affect oncological outcomes. These included patient age, tumor grade, hormonal proﬁle, pathological tumor stage, nodal disease, and adjuvant therapy received. A multivariate analysis was performed with variables with signiﬁcant p-values in the univariate model. In cases of anticipated signiﬁcant breast volume loss, VR techniques were individualized according to the excised breast volume and tumor location with planned use of either adipofascial ﬂap, lateral thoracodorsal ﬂap (LTD), intercostal artery perforators (ICAP), thoracodorsal artery perforator (TDAP), thoracoepigastric (TE), or latissimus dorsi (LD). LD myocutaneous ﬂaps were preferred for excised specimen >150 g (16, 17). Patient and tumor characteristics Patient demographics, surgical details, clinicopathological characteristics, including clinical tumor size, specimen weight, TABLE 1 Oncoplastic procedures divided into volume displacement and volume replacement techniques (N=174). TABLE 1 Oncoplastic procedures divided into volume displacement and volume replacement techniques (N=174). Volume displacement N=98 Volume replacement N=76 Tennis racket 32 (18.3%) Latissimus dorsi myocutaneous ﬂap 23 (13.2%) Rotating ﬂap 31(17.8%) Intercostal artery perforator ﬂap 20 (11.5%) Reduction mammoplasty 14 (8.0%) Lateral thoracodorsal ﬂap 18 (10.3%) Purse string suture closure 13 (7.5%) Thoracodorsal artery perforator ﬂap 11 (6.3%) Batwing mastopexy 4 (2.3%) Thoracoepigastric ﬂap 2 (1.1%) Glandular ﬂap 4 (2.3%) Adipofascial ﬂap 2 (1.1%) 03 Hing et al. 10.3389/fonc.2022.944589 TABLE 2 Baseline characteristics of patients who underwent conventional and oncoplastic breast conserving surgery (BCS). Patient and tumor characteristics All, N=539 Conventional BCS, N=365 Oncoplastic BCS, N=174 p-value Mean age (years, ± SD) 49.4 ± 9.0 50.7 ± 9.2 46.5 ± 7.5 <0.001 Mean body mass index (kg/m2 ± SD) 23.6 ± 3.3 23.6 ± 3.5 23.5 ± 3.1 0.55 Mean specimen weight ± SD, g 68.1 ± 46.6 53.1 ± 26.8 96.3 ± 60.9 <0.001 Mean clinical tumor size (cm ± SD) 1.7 0.8 1.5 ± 0.7 2.1 ± 1.1 <0.001 Tumor location by quadrant (n, %)* Central Upper outer quadrant Upper inner quadrant Lower inner quadrant Lower outer quadrant Multifocal 34 140 61 16 32 12 9 (6.0%) 87 (58.3%) 36 (24.2%) 5 (3.3%) 10 (6.7) 2 (1.3%) 26 (17.4%) 53 (35.6%) 25 (16.8%) 11 (7.4%) 22 (14.8%) 10 (6.7%) <0.001 Mean pathological tumor size (cm ± SD) 1.5 ± 0.7 1.3 ± 0.6 1.8 ± 0.8 <0.001 Axillary lymph node dissection 108 (20.5%) 66 (18.5%) 42 (24.6%) 0.10 Tumor type DCIS/pleomorphic LCIS Invasive ductal carcinoma Invasive lobular carcinoma Mixed/others 26 (4.8%) 469 (87.0%) 17 (3.1%) 27 (5.0%) 19 (5.2%) 318 (87.1%) 10 (2.7%) 18 (4.9%) 7 (4.0%) 151 (86.8%) 7 (4.0%) 9 (5.2%) 0.81 Pathological tumour staging 0 1 2 3 23 (4.3%) 401 (74.5%) 113 (21.0%) 1 (0.2%) 16 (4.3%) 295 (80.1%) 54 (14.8%) 0 7 (4.0%) 106 (60.9%) 59 (33.9%) 1 (0.5%) <0.001 Pathological nodal staging 0 1 2 3 430 (80.0%) 89 (16.5%) 14 (2.6%) 5 (0.9%) 299 (81.9%) 55 (15.1%) 6 (1.6%) 5 (1.4%) 131 (75.3%) 34 (19.5%) 8 (4.6%) 0 0.03 Pathological TNM stage 0 1 2 3 36 (6.1%) 325 (60.3%) 158 (29.3%) 20 (3.71%) 23 (6.0%) 240 (65.7%) 91 (25.2%) 11 (3.0%) 13 (7.5%) 85 (48.9%) 67 (38.5%) 9 (5.2%) 0.002 Receptor proﬁle HR+ Her2- HR+ Her2+ HR- Her2- HR- Her2+ 362 (67.2%) 57 (10.6%) 86 (16.0%) 25 (4.6%) 249 (68.5%) 38 (10.4%) 57 (15.6%) 15 (4.1%) 113 (64.9%) 19 (10.9%) 29 (16.7%) 10 (5.7%) 0.89 Grade 1 2 3 114 (22.8%) 268 (53.5%) 115 (23.0%) 86 (23.5%) 184 (50.4%) 67 (18.3%) 28 (16.1%) 84 (48.3%) 48 (27.6%) 0.04 Positive frozen margin status 36 (6.8%) 32 (9.0%) 4 (2.3%) 0.04 Neoadjuvant therapy 19 (3.5%) 12(3.2%) 7 (4.0%) 0.67 Adjuvant chemotherapy 308 (57.1%) 196 (53.7%) 112(64.7%) 0.02 Adjuvant radiotherapy 467 (86.6%) 325 (89.0%) 142 (81.6%) 0.02 Adjuvant hormonal therapy 407 (75.5%) 282 (77.3%) 125 (71.8%) 0.17 Contralateral breast cancer 17 (3.2%) 10 (2.7%) 7 (4.0%) 0.42 Ipsilateral breast tumor recurrence 11 (2.0%) 8 (2.1%) 3 (1.7%) 0.78 Locoregional recurrence 23 (4.2%) 14 (3.8%) 9 (5.7%) 0.47 Distant recurrence 26 (4.8%) 16 (4.8%) 10(5.7%) 0.49 Death 13 (2.4%) 7 (1.9%) 6 (3.5%) 0.28 SD = standard deviation; DCIS = ductal carcinoma in situ; LCIS: lobular carcinoma in situ; HR = Hormone receptor (estrogen receptor and progesterone receptor); Her2 = human epidermal growth factor receptor 2. wth factor receptor 2. a from 149 consecutive conventional breast conserving surgery was compared with 147 oncoplastic breast conserving surgery between 2011 and 2013. SD = standard deviation; DCIS = ductal carcinoma in situ; LCIS: lobular carcinoma in situ; HR = Hormone receptor (estrogen receptor and progesterone receptor); Her2 = human d l h f ation; DCIS = ductal carcinoma in situ; LCIS: lobular carcinoma in situ; HR = Hormone receptor (estrogen receptor and progesterone receptor); Her2 = hum actor receptor 2 = standard deviation; DCIS = ductal carcinoma in situ; LCIS: lobular carcinoma in situ; HR = Hormone receptor (estrogen receptor and progesterone recept dermal growth factor receptor 2. Discussion Over the last two decades, oncoplastic breast surgery quickly gained widespread acceptance as a standard of care option that balanced oncological and aesthetic outcomes of oncological resection in breast cancer management (21–27). The main ﬁndings of this study were that there was an overall low rate of IBTR (2.2%) and death (4.3%) observed in this cohort of 539 patients after a median follow-up of 82.4 months. This study had one of the largest single center cohorts with a long follow-up period (Table 7). Like other studies, IBTR rates were estimated to be between 1.4% and 14.6%, and 10-year OS rates were approximately 90.2–100%. Stratiﬁed analysis did not reveal any associated difference in survival outcomes in larger tumors, higher grade disease, or disease with a nodal burden. The observed outcomes could be the result of other factors, such as younger age (mean age <50 year), earlier disease stage (majority stage 1 and 2), favorable histological subtype, and Results Of the 539 patients who were analyzed, 365 (67.7%) patients underwent CBCS while 174 (32.3%) underwent OBCS. Of the 174 cases of OBCS, VR techniques were utilized in 98 (56.3%) cases, while VD techniques were utilized in 76 (43.7%) cases. Table 1 shows the breakdown of oncoplastic procedures in descending order. The most commonly employed techniques among the oncoplastic procedures were tennis racket incision (32), rotating ﬂap (31), and LD myocutaneous ﬂap (23). Five patients who defaulted further clinical visits or transferred care to other hospitals were considered lost to follow up. Patient characteristics Patients who underwent CBCS were older (50.7 vs. 46.5 years old), had smaller clinical tumor size (1.5 cm vs. 2.1cm), smaller specimen weight (53.1 g vs. 96.3 g), and smaller pathological tumor size (1.3 cm vs. 1.8 cm) compared to those who underwent OBCS. In terms of tumor characteristics, patients who underwent CBCS had earlier pathological T and N stage compared to those who underwent OBCS, while there was no statistically signiﬁcant difference in histology subtype, grade, or hormone proﬁle (Table 2) among the two groups. Comparison of our current study with other similar studies reporting long-term oncological outcomes are summarized in Tables 6 and 7. Intraoperatively detected involved margins on frozen section The rate ofintraoperatively detected involvedmarginson frozen section was higher in the CBCS than in the oncoplastic group, and further margins were excised intraoperatively. Three patients requiredcompletion mastectomyforcloseorinvolved ﬁnal margins. Tumor location OBCS was performed on a higher proportion of central (17.4%), lower outer quadrant (14.8%), lower inner quadrant (7.4%), and multifocal tumors (6.7%) than CBCS. The majority of all CBCS was performed on upper outer quadrant tumors (58.5%). Statistical analysis of oncological outcomes The use of oncoplastic surgery was not associated with a higher likelihood of IBTR or death in the Cox regression model analysis (Tables 3, 4, 5). Patients who underwent adjuvant chemotherapy had signiﬁcantly lower IBTR rates, with a hazard ratio of 0.25 (95% CI, 0.07 to 0.98). Regarding OS, higher histological grade was signiﬁcantly associated with higher risk of death, with a hazard ratio of 9.56 (95% CI, 2.41 to 37.86) (Tables 3 and 5). Univariate analysis was performed using the log-rank method stratiﬁed by tumor histological grade, pathological tumor staging, nodal disease, and hormone receptor proﬁle. There was no difference in IBTR-free survival when performing OBCS after stratifying by high-grade tumors; larger tumors (T2/3); and node positive, hormone receptor- positive, or triple negative breast tumors (Figures 1, 2, 3). Patient and tumor characteristics A il bl d f 149 i i l b i d i h 147 l i b i b 2011 d 2013 All, N=539 Conventional BCS, N=365 Oncoplastic BCS, N=174 04 Frontiers in Oncology frontiersin.org frontiersin.org Hing et al. Hing et al. 10.3389/fonc.2022.944589 The study was approved by the Institutional Review Board of Kyungpook National University (2015-05-205) and conducted in compliance with the principles of the Declaration of Helsinki. 11 had ipsilateral breast tumor recurrences, 17 (3.2%) metachronous contralateral breast cancer, 26 (4.8%) distant metastases, and 13 (2.4%) deaths. The hazard ratio of OBCS for IBTR, DFS and OS were 0.78 (95% conﬁdence interval [CI] 0.21–2.94, p=0.78), 1.59 (95% CI, 0.88 to 2.87, p=0.12), and 2.1 (95% CI, 0.72 to 5.9, p=0.17) respectively. The 10-year IBTR- free, DFS and OS rate were 97.8%, 86.2%, and 95.7% respectively. Overall, ﬁve patients underwent mastectomy either from involved margins or disease recurrence, giving a successful BCS rate of 99.1%. Disease recurrence, overall survival, and success of breast conservation surgery at 10 years At a median follow-up of 82.4 months, (range, 1.4–156.7 months) there were 23 (4.2%) locoregional recurrences of which 05 frontiersin.org Frontiers in Oncology Hing et al. 10.3389/fonc.2022.944589 TABLE 3 Univariate and multivariate Cox regression analysis with ipsilateral breast tumor recurrence free survival as an endpoint. No. of cases, N=539 No. of IBTR, N=11 Univariate HR p Multivariate HR p Type of BCS Oncoplastic Conventional 174 365 3 8 0.78 (0.21–2.94) Ref 0.71 0.89 (0.23-3.39) Ref 0.87 ALND Yes No Missing 108 420 11 5 6 2.91 (0.88–9.63) Ref 0.09 – – Age 539 11 1.0 (0.96–1.08) 0.55 – – Histological subtype IDC Others 469 70 11 0 Ref 0.67 (0.17–2.68) 0.55 – – Grade Grade 1/2 Grade 3 Missing 386 115 38 9 2 Ref 0.77 (0.17–3.60) 0.74 – – Tumor stage T1 T2/3 424 115 10 1 Ref 0.34 (0.04–2.69) 0.24 – – Nodal stage Node negative Node positive 430 108 7 4 Ref 2.06 (0.60–7.08) 0.27 – – Hormone receptor Positive Negative Triple negative Yes No 427 111 63 476 7 4 2 9 0.44 (0.21–1.51) Ref 1.75 (0.37–8.10) Ref 0.21 0.50 - - – Adjuvant chemotherapy Yes No 308 231 8 3 0.26 (0.07–0.96) Ref 0.03 0.25 (0.07–0.98) Ref 0.047 Adjuvant radiotherapy Yes No 467 72 9 2 0.71 (0.15–3.28) Ref 0.67 – – Adjuvant hormonal therapy Yes No 407 132 7 4 0.52 (0.15–1.82) Ref 0.32 – – Variables with p-values <0.05 in the univariate analysis were included in the multivariate analysis. No. of cases, N=539 No. of IBTR, N=11 Univariate HR p Multivariate HR p Variables with p-values <0.05 in the univariate analysis were included in the multivariate analysis. Variables with p-values <0.05 in the univariate analysis were included in the multivariate analysis. satisfaction rates given the lower rate of reoperation and conversion to mastectomy (28). generally high uptake rates of adjuvant therapies such as chemotherapy, radiotherapy, and hormonal therapy, when indicated. The proportion of cases with neoadjuvant chemotherapy was lower than expected in current practice; this might be because of the trend of favoring upfront surgery 10 years ago. However, as this was a retrospective cohort analysis, the cumulative incidence of events could also be underestimated because of a loss to follow-up or selection bias. generally high uptake rates of adjuvant therapies such as chemotherapy, radiotherapy, and hormonal therapy, when indicated. Frontiers in Oncology Disease recurrence, overall survival, and success of breast conservation surgery at 10 years The proportion of cases with neoadjuvant chemotherapy was lower than expected in current practice; this might be because of the trend of favoring upfront surgery 10 years ago. However, as this was a retrospective cohort analysis, the cumulative incidence of events could also be underestimated because of a loss to follow-up or selection bias. Our literature review showed that there were several registry studies and meta-analyses published on oncological outcomes of oncoplastic breast surgery (21–42). However, we must caution that conclusions drawn from such meta-analyses or registries have inherent limitations. Many studies have difﬁculties pooling study subjects together due to the heterogeneity of the study population, surgeons’ expertise, and techniques and classiﬁcations used (21–26, 42). Therefore, we analyzed the different deﬁnitions and breakdowns of oncoplastic techniques used across various studies (Table 5). We noted that majority of the studies were small observational studies on speciﬁc techniques, limiting their generalizability and had to be excluded from the meta-analysis. Most had a limited cohort We also observed a similar trend that oncoplastic techniques allowed higher resection volumes for larger tumors and reduced intraoperative positive margin rates. Large systematic reviews showed that oncoplastic surgery was more frequently performed in younger patients who required greater breast volume removal for larger tumors (23, 25, 27). While this may not translate to any survival beneﬁt, there could be improvement in patients’ Frontiers in Oncology 06 frontiersin.org Hing et al. 10.3389/fonc.2022.944589 TABLE 4 Univariate and multivariate Cox regression analysis with disease-free survival+ as an endpoint. No. of cases, N=539 No. Disease recurrence, overall survival, and success of breast conservation surgery at 10 years of recurrences, N=47 Univariate HR p Multivariate HR* p Type of BCS Oncoplastic Conventional 174 365 19 28 1.59 (0.88–2.88) Ref 0.13 1.95 (1.04–3.64) Ref 0.04 ALND Yes No Missing 108 420 11 16 31 1.79 (0.97–3.31) Ref 0.07 – – Age 539 47 1.02 (1.00–1.05) 0.05 1.05 (1.00–1.07) 0.03 Histological subtype IDC Others 469 70 43 4 Ref 2.23 (0.31–16.2) 0.48 – – Grade Grade 1/2 Grade 3 Missing 386 115 38 30 17 2.15 (1.17–3.94) 0.02 1.80 (0.91–3.55) 0.09 Tumor stage T1 T2/3 423 115 35 12 1.21 (0.62–2.34) 0.57 – – Nodal stage Node negative Node positive 430 108 33 14 1.47 (0.78–2.76) Ref 0.24 – – Hormone receptor Positive Negative Triple negative Yes No 427 111 63 476 31 16 6 41 0.49 (0.27–0.91) Ref 1.13 (0.48–2.68) Ref 0.03 0.78 0.70 (0.35–1.39) - 0.31 - Adjuvant chemotherapy Yes No 308 231 26 21 0.84 (0.47–1.51) Ref 0.57 – – Adjuvant radiotherapy Yes No 467 72 37 10 0.62 (0.31–1.25) Ref 0.20 – – Adjuvant hormonal therapy Yes No 407 132 31 16 0.57 (0.31–1.03) Ref 0.07 – – +Disease-free survival events were deﬁned as any ipsilateral or contralateral breast recurrence (invasive or non-invasive) or regional or distant metastases. Variables with p-values <0.05 in the univariate analysis were included in a multivariate analysis. HR = hazard ratio; Ref = Reference; BCS = breast conservation surgery; ALND = axillary lymph node dissection; IDC = invasive ductal carcinoma. No. of cases, N=539 No. of recurrences, N=47 Univariate HR p Multivariate HR +Disease-free survival events were deﬁned as any ipsilateral or contralateral breast recurrence (invasive or non-invasive) or regional or distant metastases Variables with p-values <0.05 in the univariate analysis were included in a multivariate analysis. HR = hazard ratio; Ref = Reference; BCS = breast conservation surgery; ALND = axillary lymph node dissection; IDC = invasive ductal carcinoma. +Disease-free survival events were deﬁned as any ipsilateral or contralateral breast recurrence (invasive or non-invasive) or regional or distant metastases. Variables with p-values <0.05 in the univariate analysis were included in a multivariate analysis. HR = hazard ratio; Ref = Reference; BCS = breast conservation surgery; ALND = axillary lymph node dissection; IDC = invasive ductal carcinoma. size or a barely sufﬁcient follow-up duration to fully capture recurrence or death events. Frontiers in Oncology Disease recurrence, overall survival, and success of breast conservation surgery at 10 years As it would be impossible to conduct any randomized control trial studying conventional and oncoplastic techniques because of ethical considerations, large cohort studies with long-term follow-up could be regarded as the highest level of evidence. than a median of 80 months to allow for more valid capture of long-term outcomes. Next, we examined the most commonly used deﬁnition of the Clough classiﬁcation in the literature. The Clough classiﬁcation of oncoplastic techniques primarily considers the excision volume ratio, requirement of skin excision for reshaping or mammoplasty, and tumor location. However, VR techniques were notably excluded because of their primary use in smaller breasts (13). Similarly, we found that many comparative studies with long-term outcomes reported a disproportionately low number of VD techniques, mainly level 2 oncoplastic mammoplasty with little or no representation of VR techniques. In our and many other East Asian populations, we adopted similar principles of deciding the type of oncoplastic This study generated fresh data on long-term outcomes so as to compare with the reported standards over the last decade. First, the main strength of this study was the clear deﬁnition of procedures performed with balanced representations of both VD and replacement techniques. Second, consistency in surgical standards was maintained in the procedures performed by a dedicated oncoplastic team made up of both breast and plastic surgeons. Third, these 539 patients were followed up for more 07 frontiersin.org frontiersin.org Hing et al. 10.3389/fonc.2022.944589 TABLE 5 Univariate and multivariate Cox regression analysis with overall survival as an endpoint. TABLE 5 Univariate and multivariate Cox regression analysis with overall survival as an endpoint. No. of cases,N=539 No. p y y HR = hazard ratio; Ref = Reference; BCS = breast conservation surgery; ALND = axillary lymph node dissection; IDC = invasive ductal carcinoma. Disease recurrence, overall survival, and success of breast conservation surgery at 10 years 10.3389/fonc.2022.944589 TABLE 6 Continued Study and center Year Cohort size Classiﬁcation of OBCS Percentage of VR among OBCS Clough, Paris Breast Centre, France 30 2017 350 (All OBCS) Clough Bilevel Excluded VR Mansell, Victoria & Western Inﬁrmary, UK 31 2017 666 -108 (OBCS) -558 (CBCS) Clough Bilevel 13.5 De Lorenzi, European Institute of Oncology, Italy 32 2016 1362 -454 (OBCS) -908 (CBCS) Tumor location Includes VD, VR and implant 10.3 Chakravorty, Royal Marsden, UK 34 2012 590 -150 (OBCS) -440 (CBCS) By location and 3 standardized VD Excluded VR Fitoussi, Institut Curie Paris, France 35 2010 540 (All OBCS) Tumor location Aesthetic vs combination Excluded VR No available data on breakdown VD = volume displacement; VR = volume replacement; OBCS = oncoplastic breast-conserving surgery; CBCS = conventional breast-conserving surgery. Year Cohort size No available data on breakdown VD = volume displacement; VR = volume replacement; OBCS = oncoplastic breast-conserving surgery; CBCS = conventional breast-conserving surgery. BLE 7 Retrospective studies showing oncologic outcomes of oncoplastic breast conservation surgery according to surgeons, o d follow-up interval to show directly reported results for local recurrence, disease-free survival, and overall survival. Disease recurrence, overall survival, and success of breast conservation surgery at 10 years of deaths, N=13 Univariable HR p Multivariate HR p Type of BCS Oncoplastic Conventional 174 365 6 7 1.78 (0.60–5.29) Ref 0.31 1.82 (0.55–5.97) 0.33 Age 539 2.61 (0.87–7.82) 0.10 – – Histological subtype IDC Others 469 70 13 0 1.03 (0.98–1.09) Ref 0.29 – – Grade Grade 1/2 Grade 3 Missing 386 115 38 3 10 Ref 1.78 (2.96–39.2) 0.0001 Ref 9.56 (2.41–37.86) 0.001 Tumor stage T1 T2/3 Missing 423 115 1 10 3 Ref 1.01 (0.28–3.68) 0.98 – – Nodal stage Node negative Node positive Missing 430 108 1 9 4 Ref 1.35 (0.41–4.42) 0.62 – – Hormone receptor Positive Negative Triple negative Yes No 427 111 63 476 7 6 3 10 0.31 (0.10–0.91) Ref 2.32 (0.63–8.43) Ref 0.04 0.24 0.80 (0.25–2.56) Ref - 0.70 - Adjuvant chemotherapy Yes No 308 231 10 3 2.21 (0.61–8.06) Ref 0.20 – – Adjuvant radiotherapy Yes No 467 72 11 2 0.92 (0.20-4.14) Ref 0.91 – – Adjuvant hormonal therapy Yes No 407 132 7 6 0.36 (0.12–1.08) Ref 0.08 – – Variables with p-values <0.05 in the univariate analysis were included in the multivariate analysis. HR = hazard ratio; Ref = Reference; BCS = breast conservation surgery; ALND = axillary lymph node dissection; IDC = invasive ductal carcinoma. No. of cases,N=539 No. of deaths, N=13 Univariable HR p Multivariate HR p p y y HR = hazard ratio; Ref = Reference; BCS = breast conservation surgery; ALND = axillary lymph node dissection; IDC = invasive ductal carcinoma. spective studies showing single center studies with large cohort and long term follow up, comparing deﬁnitions of OBCS and ncoplastic procedures by year of published study. TABLE 6 Retrospective studies showing single center studies with large cohort and long term follow up, comparing deﬁnitions of OBCS and breakdown of oncoplastic procedures by year of published study. breakdown of oncoplastic procedures by year of published study. Study and center Year Cohort size Classiﬁcation of OBCS Percentage of VR among OBCS Our study, Kyungpook National University Chilgok Hospital, Korea 2022 539 -174 (OBCS) -365 (CBCS) VD and VR 43.7 Oh, Seoul National University Hospital, Korea 28 2021 742 -371 (OBCS) -371 (CBCS) VD and VR 5.4 Kelemen, National Institute of Oncology, Hungary 29 2019 756 -378 (OBCS) -378 (CBCS) Clough bilevel Excluded VR Calabrese, Sapienza University Italy 36 2018 1024 (All OBCS) VD Excluded VR (Continued) 08 Frontiers in Oncology frontiersin.org Hing et al. Disease recurrence, overall survival, and success of breast conservation surgery at 10 years Study Surgeons Operation period Follow-up, months IBTR rates (%) Disease-free sur- vival, % Overall survival, % Our study Both breast and plastic surgeons 5 years (2009–2013) 82.5 (all) 82.9 (OBCS) 81.4 (CBCS) 2.2 (10 years, all) 1.8 (10 years OBCS) 2.4 (10 years CBCS) 86.2 (10 years, all) 79.7 (10 years, OBCS) 88.5 (10 years, CBCS) 95.7 (10 years, all) 92.6 (10 years, OBCS) 96.8 10 years, CBCS) Oh 28 Not speciﬁed 4 years (2011–2014) 84.4 (OBCS) 87.9 (CBCS) 3.1 (5 years, OBCS) 1.4 (5 years, CBCS) 92.9 (5 years, OBCS) 94.5 (5 years, CBCS) – Kelemen 29 2 breast surgeons 7 years (2010–2017) 51 (OBCS) 52 (CBCS) – 88.5 (5 years, OBCS) 78.2 (5 years, CBCS) 100 (5 years, OBCS) 97.3 (5 years CBCS) Calabrese 36 Breast and plastic surgeons 11 years (2000- 2010) 74.2 (all) 4.7 (all) 95.0 (all) 98.4 Clough 30 Not speciﬁed 13 years (2004–2016) 55 (all) – 84.8 95.1 (5 years) Mansell 31 Either breast or plastic surgeons 4 years (2009–2012) 56.2 (all) 56.8 (OBCS) 57.2 (CBCS) 2 (5 years, OBCS) 3.4 (5 years, CBCS) 90.7 (5 years, OBCS) 93.2 (5 years, CBCS) 98.1 (5 years, OBCS) 95.1 (5 years, CBCS) De Lorenzi 32 Not speciﬁed 9 years (2000–2008) 86.4 6.7 (10 years, OBCS) 4.2 (10 years, CBCS) 69 (10 years, OBCS) 73.1 (10 years, CBCS) 91.4 (10 years, OBCS) 91.3 (10 years, CBCS) Chakravorty 34 2 oncoplastic surgeons 7 years (2003–2010) 28 4.3 (Projected 6 years, OBCS) 3.7 (Projected 6 years, CBCS) – – Fitoussi 35 Not speciﬁed 22 years (1986–2008) 49 6.8 (5 years, all) 87.9 (5 years, all) 92.9 (5 years, all) OBCS = oncoplastic breast-conserving surgery; CBCS = conventional breast-conserving surgery. 6.8 (5 years, all) Frontiers in Oncology Frontiers in Oncology 09 frontiersin.org frontiersin.org Hing et al. 10.3389/fonc.2022.944589 FIGURE 1 (First row) Kaplan–Meier estimates of (Left) ipsilateral breast tumor recurrence (IBTR)-free survival, (Middle) disease-free survival (DFS), and (Right) overall survival (OS) curves (shown with 95% conﬁdence level) for all patients undergoing breast-conserving surgery (BCS) and (second row) by conventional (CBCS) versus oncoplastic breast-conserving surgery (OBCS) group. FIGURE 1 (First row) Kaplan–Meier estimates of (Left) ipsilateral breast tumor recurrence (IBTR)-free survival, (Middle) disease-free survival (DFS), and (Right) overall survival (OS) curves (shown with 95% conﬁdence level) for all patients undergoing breast-conserving surgery (BCS) and (second row) by conventional (CBCS) versus oncoplastic breast-conserving surgery (OBCS) group. Disease recurrence, overall survival, and success of breast conservation surgery at 10 years FIGURE 2 Kaplan–Meier estimates for ipsilateral breast tumor recurrence (IBTR)-free survival by (First Row) pathological tumor stage (ﬁrst row) and (Second Row) nodal stage (second row) showing no difference between oncoplastic and conventional breast-conserving surgery (BCS). i l b (Fi R ) h l i l (ﬁ ) d FIGURE 2 Kaplan–Meier estimates for ipsilateral breast tumor recurrence (IBTR)-free survival by (First Row) pathological tumor stage (ﬁrst row) and (Second Row) nodal stage (second row) showing no difference between oncoplastic and conventional breast-conserving surgery (BCS). 10 Frontiers in Oncology frontiersin.org Hing et al. 10.3389/fonc.2022.944589 FIGURE 3 Kaplan–Meier estimates for ipsilateral breast tumor recurrence (IBTR)-free survival stratiﬁed by (Left) high-grade, (Middle) hormone-positive tumors, and (Right) triple negative breast cancer subtypes (ﬁrst row) and others (second row) showing no difference between oncoplastic and conventional breast-conserving surgery (BCS). FIGURE 3 Kaplan–Meier estimates for ipsilateral breast tumor recurrence (IBTR)-free survival stratiﬁed by (Left) high-grade, (Middle) hormone-positive tumors, and (Right) triple negative breast cancer subtypes (ﬁrst row) and others (second row) showing no difference between oncoplastic and conventional breast-conserving surgery (BCS). establish comparable oncological outcomes of various oncoplastic techniques to reassure patients that oncoplastic breast surgery would not compromise on oncological safety in the long run, and that both aesthetic outcomes and patient satisfaction were equally important performance indicators in the treatment of breast cancer. procedures such as breast-to-tumor volume ratio, and tumor locations, but it was also proposed that an absolute value of tumor volume excised in itself could be an indication for VR techniques in small-to-moderate-sized breasts. These may be due to inherent differences in the patients’ morphometric characteristics or inﬂuenced by different cultural beliefs and resource settings (15, 41). In a smaller native breast with less space for VD maneuvers, a different threshold for VR techniques may apply. Evidence also shows that patients are more accepting of VR options and have good functional outcomes regardless of the VR technique (17). Frontiers in Oncology frontiersin.org Author contributions The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/ fonc.2022.944589/full#supplementary-material All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication. Acknowledgments Our review of existing literature on the oncological outcomes of OBCS highlighted the dearth in well-balanced comparative studies with sufﬁcient long-term follow-up, and reported our center’s own long-term oncological outcomes for OBCS to support the use of either VD or replacement techniques. We would like to thank Editage (www.editage.co.kr) for English language editing. Conﬂict of interest 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. Limitations The main limitations of the study were largely in its retrospective nature, which could lead to underestimated incidence rates due to the nature of selection bias and loss to follow-up. The surgical teams involved a dedicated oncoplastic team including both breast and plastic surgeons; consequently, these ﬁndings may not be logistically reproducible in all centers. We acknowledged that there were many confounding factors that could affect oncological outcomes and tried to address these by adjusting for the variables in the statistical analysis. However, considering the limitations of cohort size and event rates, it would be prudent to avoid generating too many hypotheses regarding secondary analysis ﬁndings but rather appreciate the general theme of oncological safety established across various tumor characteristics and adjuvant therapies provided in our study population. We also noted there was a low percentage of patients treated with neoadjuvant chemotherapy in our cohort. Neoadjuvant chemotherapy has gained much traction in its role in increasing rates of breast conservation; therefore, future research should be directed to study its inﬂuence on long-term oncological outcomes (43). As a result, our percentage of VR performed among OBCS was the highest among the selected studies, with VR techniques accounting for 43% of all oncoplastic procedures. Most of the other studies either had less than 10% of procedures represented by VR, or did not specify the type of reconstruction techniques at all. Our cohort also showed that the LD myocutaneous ﬂap was the most commonly used VR techniques followed by chest wall perforator ﬂaps. This was concordant to our ﬁnding that the LD ﬂap was the largest and the most commonly reported VR technique as a single cohort series in the literature (20, 25, 37–42). However, we did not report and compare the oncological results from these studies that only focused on singular technique such as LD ﬂap or omental ﬂap reconstruction because they would have limited generalizability to other oncoplastic techniques and patient selection (40, 41). We maintained that both VD and VR techniques formed the fundamentals of oncoplastic techniques and would not need to be separately studied from each other. Hence, it remained vital to 11 Frontiers in Oncology frontiersin.org frontiersin.org Hing et al. 10.3389/fonc.2022.944589 Data availability statement The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. Ethics statement All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. The studies involving human participants were reviewed and approved by Kyungpook National University Chilgok Hospital. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. References Evolving trends in surgical management of breast cancer: An analysis of 30 years of practice changing papers. Front Oncol (2021) 11:622621. doi: 10.3389/fonc.2021.622621 12. Lebovic G. Oncoplastic surgery: a creative approach to breast cancer management. Surg Oncol Clin N Am (2010) 19:567–80. doi: 10.1016/j.soc.2010.04.003 4. Kaufman CS. Increasing role of oncoplastic surgery for breast cancer. Curr Oncol Rep (2019) 21(12):111. doi: 10.1007/s11912-019-0860-9 4. Kaufman CS. Increasing role of oncoplastic surgery for breast cancer. Curr Oncol Rep (2019) 21(12):111. doi: 10.1007/s11912-019-0860-9 4. Kaufman CS. Increasing role of oncoplastic surgery for breast cancer. Curr Oncol Rep (2019) 21(12):111. doi: 10.1007/s11912-019-0860-9 13. Clough KB, Kaufman GJ, Nos C, Buccimazza I, Sarfati IM. Improving breast cancer surgery: a classiﬁcation and quadrant per quadrant atlas for oncoplastic surgery. Ann Surg Oncol (2010) 17:1375–91. doi: 10.1245/s10434-009-0792-y 5. Audretsch WP, Rezai M, Kolotas C. Tumor-speciﬁc immediate reconstruction in breast cancer patients. Perspect Plast Surg (1998) 11:71–100. doi: 10.1055/s-2008-1080243 5. Audretsch WP, Rezai M, Kolotas C. Tumor-speciﬁc immediate reconstruction in breast cancer patients. Perspect Plast Surg (1998) 11:71–100. doi: 10.1055/s-2008-1080243 14. Hoffmann J, Wallwiener D. Classifying breast cancer surgery: a novel, complexity-based system for oncological, oncoplastic and reconstructive procedures, and proof of principle by analysis of 1225 operations in 1166 patients. MC Cancer (2009) 9:108. doi: 10.1186/1471-2407-9-108 6. Rainsbury RM. Surgery insight: oncoplastic breast-conserving reconstruction–indications, beneﬁts, choices and outcomes. Nat Clin Pract Oncol (2007) 4:657–64. doi: 10.1038/ncponc0957 6. Rainsbury RM. Surgery insight: oncoplastic breast-conserving reconstruction–indications, beneﬁts, choices and outcomes. Nat Clin Pract Oncol (2007) 4:657–64. doi: 10.1038/ncponc0957 15. Yang JD, Lee JW, Cho YK, Kim WW, Hwang SO, Jung JH, et al. Surgical techniques for personalized oncoplastic surgery in breast cancer patients with small- to moderate-sized breasts (part 1): volume displacement. J Breast Cancer (2012) 15:1–6. doi: 10.4048/jbc.2012.15.1.1 7. Clough KB, Kroll SS, Audretsch W. An approach to the repair of partial mastectomy defects. Plast Reconstr Surg (1999) 104:409–20. doi: 10.1097/ 00006534-199908000-00014 7. Clough KB, Kroll SS, Audretsch W. An approach to the repair of partial mastectomy defects. Plast Reconstr Surg (1999) 104:409–20. doi: 10.1097/ 00006534-199908000-00014 8. Petit JY, Rietjens M, Garusi C. Breast reconstructive techniques in cancer patients: which one, when to apply, which immediate and long-term risks? Crit Rev Oncol Hematol (2001) 38:231–9. doi: 10.1016/S1040-8428(00) 00137-2 8. Petit JY, Rietjens M, Garusi C. Breast reconstructive techniques in cancer patients: which one, when to apply, which immediate and long-term risks? Crit Rev Oncol Hematol (2001) 38:231–9. References 1. Fisher B, Anderson S, Bryant J, Margolese RG, Deutsch M, Fisher ER, et al. Twenty-year follow-up of a randomized trial comparing total mastectomy, lumpectomy, and lumpectomy plus irradiation for the treatment of invasive breast cancer. N Engl J Med (2002) 347:1233–41. doi: 10.1056/NEJMoa022152 9. Silverstein MJ, Mai T, Savalia N, Vaince F, Guerra L. Oncoplastic breast conservation surgery: the new paradigm. J Surg Oncol (2014) 110:82–9. doi: 10.1002/jso.23641 9. Silverstein MJ, Mai T, Savalia N, Vaince F, Guerra L. Oncoplastic breast conservation surgery: the new paradigm. J Surg Oncol (2014) 110:82–9. doi: 10.1002/jso.23641 9. Silverstein MJ, Mai T, Savalia N, Vaince F, Guerra L. Oncoplastic breast conservation surgery: the new paradigm. J Surg Oncol (2014) 110:82–9. doi: 10.1002/jso.23641 10. Macmillan RD, James R, Gale KL, McCulley SJ. Therapeutic mammaplasty. J Surg Oncol (2014) 110:90–5. doi: 10.1002/jso.23659 10. Macmillan RD, James R, Gale KL, McCulley SJ. Therapeutic mammaplasty. J Surg Oncol (2014) 110:90–5. doi: 10.1002/jso.23659 2. Veronesi U, Cascinelli N, Mariani L, Greco M, Saccozzi R, Luini A, et al. Twenty-year follow-up of a randomized study comparing breast-conserving surgery with radical mastectomy for early breast cancer. N Engl J Med (2002) 347:1227–32. doi: 10.1056/NEJMoa020989 2. Veronesi U, Cascinelli N, Mariani L, Greco M, Saccozzi R, Luini A, et al. Twenty-year follow-up of a randomized study comparing breast-conserving surgery with radical mastectomy for early breast cancer. N Engl J Med (2002) 347:1227–32. doi: 10.1056/NEJMoa020989 2. Veronesi U, Cascinelli N, Mariani L, Greco M, Saccozzi R, Luini A, et al. Twenty-year follow-up of a randomized study comparing breast-conserving surgery with radical mastectomy for early breast cancer. N Engl J Med (2002) 347:1227–32. doi: 10.1056/NEJMoa020989 11. Chatterjee A, Gass J, Patel K, Holmes D, Kopkash K, Peiris L, et al. A consensus deﬁnition and classiﬁcation system of oncoplastic surgery developed by the american society of breast surgeons. Ann Surg Oncol (2019) 26:3436–44. doi: 10.1245/s10434-019-07345-4 11. Chatterjee A, Gass J, Patel K, Holmes D, Kopkash K, Peiris L, et al. A consensus deﬁnition and classiﬁcation system of oncoplastic surgery developed by the american society of breast surgeons. Ann Surg Oncol (2019) 26:3436–44. doi: 10.1245/s10434-019-07345-4 3. Keelan S, Flanagan M, Hill ADK. Evolving trends in surgical management of breast cancer: An analysis of 30 years of practice changing papers. Front Oncol (2021) 11:622621. doi: 10.3389/fonc.2021.622621 3. Keelan S, Flanagan M, Hill ADK. 12. Lebovic G. Oncoplastic surgery: a creative approach to breast cancer management. Surg Oncol Clin N Am (2010) 19:567–80. doi: 10.1016/j.soc.2010.04.003 References doi: 10.1016/S1040-8428(00) 00137-2 16. Yang JD, Lee JW, Cho YK, Kim WW, Hwang SO, Jung JH, et al. Surgical techniques for personalized oncoplastic surgery in breast cancer patients with small- to moderate-sized breasts (part 2): volume replacement. J Breast Cancer (2012) 15:7–14. doi: 10.4048/jbc.2012.15.1.7 16. Yang JD, Lee JW, Cho YK, Kim WW, Hwang SO, Jung JH, et al. Surgical techniques for personalized oncoplastic surgery in breast cancer patients with small- to moderate-sized breasts (part 2): volume replacement. J Breast Cancer (2012) 15:7–14. doi: 10.4048/jbc.2012.15.1.7 Frontiers in Oncology 12 frontiersin.org frontiersin.org Hing et al. 10.3389/fonc.2022.944589 retrospective study. Eur J Surg Oncol (2019) 45:118–24. doi: 10.1016/ j.ejso.2018.09.006 17. Lee JW, Kim MC, Park HY, Yang JD. Oncoplastic volume replacement techniques according to the excised volume and tumor location in small- to moderate-sized breasts. Gland Surg (2014) 3:14–21. doi: 10.3978/j.issn.2227- 684X.2014.02.02 31. Clough KB, van la Parra RFD, Thygesen HH, Levy E, Russ E, Halabi NM, et al. Long-term results after oncoplastic surgery for breast cancer: A 10-year follow-up. Ann Surg (2018) 268:165–71. doi: 10.1097/SLA.0000000000002255 18. Sabel MS, Hirsch A, Bellon JR. Clinical manifestations and evaluation of locoregional recurrences of breast cancer (2021). Available at: https://www.uptodate. com/contents/clinical-manifestations-and-evaluation-of-locoregional- recurrences-of-breast-cancer. 32. Mansell J, Weiler-Mithoff E, Stallard S, Doughty JC, Mallon E, Romics L. Oncoplastic breast conservation surgery is oncologically safe when compared to wide local excision and mastectomy. Breast (2017) 32:179–85. doi: 10.1016/ j.breast.2017.02.006 19. Lee J, Jung JH, Kim WW, Hwang SO, Chae YS, Lee SJ, et al. Five-year oncologic outcomes of volume displacement procedures after partial mastectomy for breast cancer. Clin Breast Cancer (2017) 17:70–5. doi: 10.1016/ j.clbc.2016.06.018 33. De Lorenzi F, Hubner G, Rotmensz N, Bagnardi V, Loschi P, Maisonneuve P, et al. Oncological results of oncoplastic breast-conserving surgery: Long term follow-up of a large series at a single institution: A matched-cohort analysis. Eur J Surg Oncol (2016) 42:71–7. doi: 10.1016/j.ejso.2015.08.160 20. Lee J, Jung JH, Kim WW, Hwang SO, Kang JG, Baek J, et al. Oncologic outcomes of volume replacement technique after partial mastectomy for breast cancer: A single center analysis. Surg Oncol (2015) 24:35–40. doi: 10.1016/ j.suronc.2014.12.001 34. Carter SA, Lyons GR, Kuerer HM, Bassett RLJr, Oates S, Thompson A, et al. Operative and oncologic outcomes in 9861 patients with operable breast cancer: Single-institution analysis of breast conservation with oncoplastic reconstruction. Ann Surg Oncol (2016) 23:3190–8. doi: 10.1245/s10434-016-5407-9 21. Kosasih S, Tayeh S, Mokbel K, Kasem A. References Is oncoplastic breast conserving surgery oncologically safe? a meta-analysis of 18,103 patients. Am J Surg (2020) 220:385–92. doi: 10.1016/j.amjsurg.2019.12.019 35. Chakravorty A, Shrestha AK, Sanmugalingam N, Rapisarda F, Roche N, Querci Della Rovere G, et al. How safe is oncoplastic breast conservation? comparative analysis with standard breast conserving surgery. Eur J Surg Oncol (2012) 38:395–8. doi: 10.1016/j.ejso.2012.02.186 22. Raufdeen F, Murphy J, Ahluwalia M, Coroneos CJ, Thoma A. Outcomes in volume replacement and volume displacement techniques in oncoplastic breast conserving surgery: A systematic review. J Plast Reconstr Aesthet Surg (2021) 74:2846–55. doi: 10.1016/j.bjps.2021.06.004 36. Fitoussi AD, Berry MG, Fama F, Falcou MC, Curnier A, Couturaud B, et al. Oncoplastic breast surgery for cancer: analysis of 540 consecutive cases. Plast Reconstr Surg (2010) 125:454–62. doi: 10.1097/PRS.0b013e3181c82d3e 23. Haloua MH, Krekel NM, Winters HA, Rietveld DH, Meijer S, Bloemers FW, et al. A systematic review of oncoplastic breast-conserving surgery: current weaknesses and future prospects. Ann Surg (2013) 257:609–20. doi: 10.1097/ SLA.0b013e3182888782 37. Calabrese C, Casella D, Di Taranto G, Marcasciano M, Kothari A, Sordi S, et al. Oncoplastic conservative surgery for breast cancer: long-term outcomes of our ﬁrst ten years experience. Eur Rev Med Pharmacol Sci (2018) 22(21):7333–42. doi: 10.26355/eurrev_201811_16270 24. Chen JY, Huang YJ, Zhang LL, Yang CQ, Wang K. Comparison of oncoplastic breast-conserving surgery and breast-conserving surgery alone: A meta-analysis. J Breast Cancer (2018) 21:321–9. doi: 10.4048/jbc.2018.21.e36 38. Teoh LY, Lai LL, Hanim Aa A, Teh MS, Jamaris S, Yahya A, et al. Oncological safety and postoperative complications in oncoplastic breast surgery among Asian women: A single institutional review. Breast J (2020) 26(11):2208–12. doi: 10.1111/tbj.14060 25. Losken A, Dugal CS, Styblo TM, Carlson GW. A meta-analysis comparing breast conservation therapy alone to the oncoplastic technique. Ann Plast Surg (2014) 72:145–9. doi: 10.1097/SAP.0b013e3182605598 39. Rietjens M, Urban CA, Rey PC, Mazzarol G, Maisonneuve P, Garusi C, et al. Long-term oncological results of breast conservative treatment with oncoplastic surgery. Breast (2007) 16:387–95. doi: 10.1016/j.breast.2007.01.008 26. Rutherford CL, Barker S, Romics L. A systematic review of oncoplastic volume replacement breast surgery: oncological safety and cosmetic outcome. Ann R Coll Surg Engl (2022) 104:5–17. doi: 10.1308/rcsann.2021.0012 40. Rusby JE, Paramanathan N, Laws SAM, Rainsbury RM. Immediate latissimus dorsi miniﬂap volume replacement for partial mastectomy: use of intra-operative frozen sections to conﬁrm negative margins. Am J Surg (2008) 196:512–18. doi: 10.1016/j.amjsurg.2008.06.026 27. References De La Cruz L, Blankenship SA, Chatterjee A, Geha R, Nocera N, Czerniecki BJ, et al. Outcomes after oncoplastic breast-conserving surgery in breast cancer patients: A systematic literature review. Ann Surg Oncol (2016) 23:3247–58. doi: 10.1245/s10434-016-5313-1 41. Zaha H, Abe N, Sagawa N, Unesoko M. Oncoplastic surgery with omental ﬂap reconstruction: a study of 200 cases. Breast Cancer Res Treat (2017) 162:267– 74. doi: 10.1007/s10549-017-4124-9 28. Gulcelik MA, Dogan L, Karaman N, Bahcecitapar M, Ozaslan C. Oncoplastic level II surgical techniques for breast cancer treatment: Long-term outcomes. Breast Care (Basel) (2022) 17:24–30. doi: 10.1159/000514468 42. André C, Holsti C, Svenner A, Oikonomou I, Appelgren M, Johansson A, et al. Recurrence and survival after standard versus oncoplastic breast-conserving surgery for breast cancer. BJS Open (2021) 5(1):zraa013. doi: 10.1093/bjsopen/ zraa013 29. Oh MY, Kim Y, Kim J, Cheun JH, Jung JG, Kim HK, et al. Comparison of long-term oncological outcomes in oncoplastic breast surgery and conventional breast-conserving surgery for breast cancer: A propensity score-matched analysis. J Breast Cancer (2021) 24:520–30. doi: 10.4048/jbc.2021.24.e52 43. Youssef MMG, Namour A, Youssef OZ, Morsi A. Oncologic and cosmetic outcomes of oncoplastic breast surgery in locally advanced breast cancer after neoadjuvant chemotherapy, experience from a developing country. Indian J Surg Oncol (2018) 9:300–6. doi: 10.1007/s13193-017-0689-3 30. Kelemen P, Pukancsik D, Ú jhelyi M, Sávolt Á , Kovács E, Ivády G, et al. Comparison of clinicopathologic, cosmetic and quality of life outcomes in 700 oncoplastic and conventional breast-conserving surgery cases: A single-centre 13 Frontiers in Oncology frontiersin.org
https://openalex.org/W2775756541	https://europepmc.org/articles/pmc5722819?pdf=render	English	null	Bioinspired graphene membrane with temperature tunable channels for water gating and molecular separation	Nature communications	2,017	cc-by	11,032	1 Laboratory of Bioinspired Smart Interfacial Science and Technology of the Ministry of Education, Beijing Key Laboratory of Bioinspired Energy Materials and Devices, School of Chemistry, Beihang University, Beijing 100901, China. 2 School of Molecular Sciences, The University of Western Australia, 35 Stirling Highway Crawley, Perth, WA 6009, Australia. 3 Laboratory of Polymer Chemistry, Department of Polymer Materials, College of Materials Science and Engineering, Shanghai University, Nanchen Street 333, Shanghai 200444, China. 4 John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA. 5 Department of Physics, Harvard University, Cambridge, MA 02138, USA. Jingchong Liu and Nü Wang contributed equally to this work. Correspondence and requests for materials should be addressed to N.W. (email: wangn@buaa.edu.cn) or to Y.Z. (email: zhaoyong@buaa.edu.cn) Bioinspired graphene membrane with temperature tunable channels for water gating and molecular separation Jingchong Liu1, Nü Wang1, Li-Juan Yu2, Amir Karton2, Wen Li3,4, Weixia Zhang4, Fengyun Guo1,4, Lanlan Hou1, Qunfeng Cheng 1, Lei Jiang1, David A. Weitz4,5 & Yong Zhao1,4 Smart regulation of substance permeability through porous membranes is highly desirable for membrane applications. Inspired by the stomatal closure feature of plant leaves at relatively high temperature, here we report a nano-gating membrane with a negative temperature- response coefﬁcient that is capable of tunable water gating and precise small molecule separation. The membrane is composed of poly(N-isopropylacrylamide) covalently bound to graphene oxide via free-radical polymerization. By virtue of the temperature tunable lamellar spaces of the graphene oxide nanosheets, the water permeance of the membrane could be reversibly regulated with a high gating ratio. Moreover, the space tunability endows the membrane with the capability of gradually separating multiple molecules of different sizes. This nano-gating membrane expands the scope of temperature-responsive membranes and has great potential applications in smart gating systems and molecular separation. 1 Laboratory of Bioinspired Smart Interfacial Science and Technology of the Ministry of Education, Beijing Key Laboratory of Bioinspired Energy Materials and Devices, School of Chemistry, Beihang University, Beijing 100901, China. 2 School of Molecular Sciences, The University of Western Australia, 35 Stirling Highway Crawley, Perth, WA 6009, Australia. 3 Laboratory of Polymer Chemistry, Department of Polymer Materials, College of Materials Science and Engineering, Shanghai University, Nanchen Street 333, Shanghai 200444, China. 4 John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA. 5 Department of Physics, Harvard University, Cambridge, MA 02138, USA. Jingchong Liu and Nü Wang contributed equally to this work. Correspondence and requests for materials should be addressed to N.W. (email: wangn@buaa.edu.cn) or to Y.Z. (email: zhaoyong@buaa.edu.cn) 1 NATURE COMMUNICATIONS\| 8: 2011 \| DOI: 10.1038/s41467-017-02198-5\| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 F or a number of drought-enduring vegetations in nature (such as cacti), the stomata on their leaves close at relatively high temperatures, effectively reducing the water loss in order to sustain life1,2. This biological phenomenon displays a negative temperature-response characteristic from a membrane perspective, that is, the membrane’s permeability decreases as the surrounding temperature increases3–5. Although this negative temperature-responsive phenomenon is prevalent in nature, most man-made temperature-responsive membranes have the opposite responsive trend where water permeability increases with increasing temperature, i.e., they show a positive temperature- response characteristic6–9. This is because most temperature- responsive polymeric materials adopted in gating membranes undergo a coil-globule shrinking transition of their chains at a lower critical solution temperature (LCST)10–12, and as a result, the membrane pores switch from a closed state below the LCST to an open state above the LCST. F their scope of practical applications. Temperature controlled gating membrane systems, which are convenient and less dependent on the chemical environment, have more extensive adaptability36. However, negative temperature-response gra- phene-based membranes that can achieve water gating and molecular separation have never been reported. Here, we construct a negative temperature-response nano- gating membrane by covalently grafting poly(N-iso- propylacrylamide) (PNIPAM) chains on GO sheets. The water permeance of this membrane varies from 12.4 l m−2 h−1 bar−1 at 25 °C to 1.8 l m−2 h−1 bar−1 at 50 °C with a high gating ratio of ~7. Moreover, the tunable lamellar spacing of the GOM enables it to separate small molecules with different sizes by regulating the temperature. This smart membrane shows promise in many ﬁelds, such as ﬂuid transport systems, microﬂuidic chip systems and molecular separation devices. p Graphene oxide (GO) is a two-dimensional material consisting of a carbon lattice and oxygen-containing groups13. Recently, GO membranes (GOMs) have exhibited superior gating and separa- tion performances due to their unique unimpeded two- dimensional nanochannels and nacre like lamellar struc- ture14,15. The ordered brick and mortar assembly of inorganic and/or organic layers in GO-based membranes endows them with strong interfacial interactions and excellent chemical stability16. Therefore, GO-based membranes have been successfully applied to gas, molecular, and organic solvent separations17–30. Moreover, controlling molecule penetration through GO-based membranes has received considerable attention due to the crucial importance of gatekeepers control the movement of substances, mimicking living organisms31,32. Results b Fabrication of the PNIPAM covalently grafted GOMs. The temperature-responsive PNIPAM covalently grafted GO mem- branes(P-GOMs) are based on regulating the layer spacing through tuning the molecular conﬁguration of PNIPAM. The two-step construction process and water gating property of P- GOMs are illustrated in Fig. 1, involving polymerization of N- isopropylacrylamide (NIPAM) on GO sheets to form PNIPAM- grafted GO (P-GO) and assembling P-GO into P-GOMs through pressure-driven ﬁltration. When the tempera- ture is below the LCST of P-GO, the PNIPAM chains on P-GO exhibit a swollen coil structure because of hydrogen bond inter- action between amide group of PNIPAM and water molecules. Under this condition, the space between two adjacent GO sheets provides appropriate nanochannels for water transport. If the temperature increases above its LCST, the PNIPAM chain shrinks because the hydrogen bond interaction between the amide group and water molecule is replaced by the intramolecular/ Gas pressure PNIPAM Open Closed > LCST < LCST CH2 CH x C O N H O H H CH2 CH x C N H O C N H O CH CH2 y Intramolecular/intermolecular hydrogen bonding PNIPAM and water molecule hydrogen bonding Substrate P-GOMs P-GO Polymerization Dispersion GO NIPAM + O N H Fig. 1 Fabrication process of the temperature-responsive membrane and its water gating property. N-isopropylacrylamide (NIPAM) monomers polymerize on graphene oxide (GO) sheets to form PNIPAM-grafted GO (P-GO). Then P-GO assembles into PNIPAM covalently grafted GO membranes (P-GOMs) through pressure-driven ﬁltration of their aqueous dispersion. The water permeance of P-GOMs could be regulated via the environmental temperature (T). P-GOMs have a large water permeance when T is below its lower critical solution temperature (LCST) because of the expanded water ﬂow channel caused by the swollen PNIPAM chains. The water permeance decreases when T is above the LCST because of the narrowed channel of P-GOMs, which is caused by the shrink of PNIPAM chains Gas pressure Substrate P-GOMs P-GO Polymerization Dispersion GO NIPAM + O N H PNIPAM Open CH2 CH x C O N H O H H PNIPAM and water molecule hydrogen bonding PNIPAM Fig. 1 Fabrication process of the temperature-responsive membrane and its water gating property. N-isopropylacrylamide (NIPAM) monomers polymerize on graphene oxide (GO) sheets to form PNIPAM-grafted GO (P-GO). Then P-GO assembles into PNIPAM covalently grafted GO membranes (P-GOMs) through pressure-driven ﬁltration of their aqueous dispersion. The water permeance of P-GOMs could be regulated via the environmental temperature (T). NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 Previous studies have demonstrated that water permeance of different kinds of GO-based membranes can be regulated by external ﬁelds such as pH, solvent, ion con- centration, electric and magnetic ﬁelds20,33–35. However, these systems rely on speciﬁc chemical environments like ionic strength, acid-base reactions and so on, which signiﬁcantly limit ARTICLE ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 d AFM image of a single P-GO sheet with thickness ~5 nm. e ATR-FTIR spectra of GOMs, PNIPAM and P-GOMs, illustrating covalent bound between PNIPAM and GO. f XPS spectrum of P-GO, of which C1s region ﬁtting into ﬁve peaks at 284.6, 285.8, 286.9, 287.8, and 288.6 eV, representing C–C, C–N, C–O, C = O and C(O)O. Inset is N1s region, ﬁtting into two peaks at 399.4 and 400.0 eV, representing amine and amide, respectively, because of the graft of PNIPAM intermolecular hydrogen bonding of PNIPAM37. In this process, the GO sheets are pulled closer because they are covalently tethered to the ends of PNIPAM chains, resulting in a smaller lamellar distance of P-GOMs. Then, the water permeance decreases as water is mostly blocked by P-GOMs. It is thereby feasible to construct P-GOMs with negative coefﬁcient in hydraulic permeability. PNIPAM. During the reaction, NIPAM not only polymerized on GO surfaces but also formed free PNIPAM through self- polymerization42. To eliminate the possibility that the new peaks detected in the ATR-FTIR spectrum of P-GOMs were from the free PNIPAM attached to GO, the ﬁltrate of the P-GOMs construction process was also analyzed by ATR-FTIR (Supple- mentary Fig. 4). There was no absorption peak of PNIPAM in the spectrum of the ﬁltrate, conﬁrming the strong covalent interac- tion between PNIPAM and GO. The X-ray photoelectron spectroscopy (XPS) spectra of GO and P-GO clearly indicate the introduction of nitrogen-containing functional groups after modifying GO with PNIPAM (Fig. 2f and Supplementary Fig. 5). The C1s XPS spectrum of P-GO shows curve ﬁttings with a new peak of C–N at the binding energy of 285.8 eV compared with that of GO43. More characterization including Raman spectra, and 1H nuclear magnetic resonance also demonstrate that PNIPAM chains have been successfully covalently grafted to GO sheets (Supplementary Figs. 6 and 7). y p y The temperature-responsive P-GOMs were obtained by pressure-assisted self-assembly technology38 with the freeze- dried P-GO (Supplementary Figs. 1a and b). The P-GOMs displayed typical brown color with good optical transparency and uniform surfaces without defects (Fig. 2a and Supplementary Fig. 1c), which was similar to GOMs (Supplementary Fig. 1d)19. The surface and cross-section scanning electron microscope images of the P-GOMs (Fig. 2b, c) demonstrate that the P-GOMs have wrinkled surface and stacking microstructure like normal GOMs (Supplementary Figs. 2a and b). NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 a Wavenumbers (cm–1) 2200 2000 1800 1600 1400 1200 1000 800 292 290 288 286 284 282 Binding energy (eV) Transmittance (%) Intensity 10 µm 200 nm C=O C=O N-H GOMs PNIPAM P-GOMs C(O)O(288.6) C=O(287.8) C-O(286.9) C-N(285.8) C-C(284.6) N1s N-C=O(400) N-H (399.4) 403 402 0.00 –2 0 2 4 6 0.05 0.10 0.15 0.20 0.25 0.30 0.35 401 400 399 398 397 Binding energy (eV) C1s CH(CH3)2 a b c d e f 100 nm nm µm N N Fig. 2 Characterization of P-GOMs. a Digital photo of P-GOMs. b, c Top-view and cross-sectional SEM images of P-GOMs showing a typical wrinkle surface and stacking microstructure. The insets are EDS maps indicating that nitrogen is evenly distributed on both surface and cross-section of P-GOMs. d AFM image of a single P-GO sheet with thickness ~5 nm. e ATR-FTIR spectra of GOMs, PNIPAM and P-GOMs, illustrating covalent bound between PNIPAM and GO. f XPS spectrum of P-GO, of which C1s region ﬁtting into ﬁve peaks at 284.6, 285.8, 286.9, 287.8, and 288.6 eV, representing C–C, C–N, C–O, C = O and C(O)O. Inset is N1s region, ﬁtting into two peaks at 399.4 and 400.0 eV, representing amine and amide, respectively, because of the graft of PNIPAM 10 µm b N b a 200 nm c N c f 0.00 –2 0 2 4 6 0.05 0.10 0.15 0.20 0.25 0.30 0.35 d 100 nm nm µm Wavenumbers (cm–1) 2200 2000 1800 1600 1400 1200 1000 800 Transmittance (%) C=O C=O N-H GOMs PNIPAM P-GOMs CH(CH3)2 e 292 290 288 286 284 282 Binding energy (eV) Intensity C(O)O(288.6) C=O(287.8) C-O(286.9) C-N(285.8) C-C(284.6) N1s N-C=O(400) N-H (399.4) 403 402 401 400 399 398 397 Binding energy (eV) C1s f e Wavenumbers (cm–1) 2200 2000 1800 1600 1400 1200 1000 800 292 290 288 286 284 282 Binding energy (eV) Transmittance (%) Intensity C=O C=O N-H GOMs PNIPAM P-GOMs C(O)O(288.6) C=O(287.8) C-O(286.9) C-N(285.8) C-C(284.6) N1s N-C=O(400) N-H (399.4) 403 402 .25 0.30 0.35 401 400 399 398 397 Binding energy (eV) C1s CH(CH3)2 100 nm Intensity Binding energy (eV) Binding energy (eV) Fig. 2 Characterization of P-GOMs. a Digital photo of P-GOMs. b, c Top-view and cross-sectional SEM images of P-GOMs showing a typical wrinkle surface and stacking microstructure. The insets are EDS maps indicating that nitrogen is evenly distributed on both surface and cross-section of P-GOMs. Results b P-GOMs have a large water permeance when T is below its lower critical solution temperature (LCST) because of the expanded water ﬂow channel caused by the swollen PNIPAM chains. The water permeance decreases when T is above the LCST because of the narrowed channel of P-GOMs, which is caused by the shrink of PNIPAM chains NATURE COMMUNICATIONS\| 8: 2011 \| DOI: 10.1038/s41467-017-02198-5\| www.nature.com/naturecommunications 2 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 f The water permeance of P-GOMs decreases when switching on NIR light and returns to initial values when the NIR light is off nanochannels behaves as a viscous ﬂow29,30,47,48, the water permeated through the nanochannels behaves as a viscous ﬂow29,30,47,48, the water permeated through the P-GOMs can be described by the Hagen–Poiseuille equation: nanochannels behaves as a viscous ﬂow29,30,47,48, the water permeated through the nanochannels behaves as a viscous ﬂow29,30,47,48, the water permeated through the P-GOMs can be described by the Hagen–Poiseuille equation: after functionalization and thus became a little hydrophobic. As a result, the LCST of PNIPAM decreased when bound to hydro- phobic groups45. This hypothesis was supported by the results of the XRD, water contact angle, XPS, and Raman characterizations (Supplementary Fig. 10). permeated through the P-GOMs can be described by the Hagen–Poiseuille equation: p g P-GOMs can be described by the Hagen–Poiseuille equation: J ¼ επr2Δp=8ηdτ ð1Þ ð1Þ These results indicate that PNIPAM is successfully grafted onto GO surfaces by covalent bonds, making it feasible to act as water gating by virtue of the temperature tunable lamellar spaces of the P-GOMs. Then we performed pressure-driven hydraulic perme- ability experiments to study the temperature-response behavior of P-GOMs. As controls, GOMs and physically blended GO/ PNIPAM membrane were also compared. As illustrated in Fig. 3b, the P-GOMs exhibited the negative temperature- response characteristic, which was distinctly different with the GOMs and physically blended GO/PNIPAM membrane. The water permeance of GOMs slightly increased as the temperature went up, owing to the decreased water viscosity46. For the physically blended GO/PNIPAM membranes, the permeance had a sharp transition when the temperature changed from 30 °C to 35 °C. Given that the LCST of PNIPAM is ~32 °C, the PNIPAM chains non-covalently attached to GO in the membranes with shrunken conformation expanded the size of the channels at 35 ° C compared with that at 30 °C, thus resulting in a larger permeance increase (Supplementary Fig. 11)36. For our P-GOMs, however, the permeability decreased with the temperature increasing when the temperature was above 30 °C, showing a negative temperature-response characteristic. The P-GOMs (1.1 μm thickness) had an average water permeance of 12.4 l m−2 h−1 bar−1 at 25 °C, and decreased to 1.8 l m−2 h−1 bar−1 when the temperature increased to 50 °C. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 The energy dispersive spectroscopy results show that nitrogen element is evenly distributed in both surface and cross-section of P-GOMs (Fig. 2b, c), whereas almost no nitrogen is detected in pristine GOMs (Supplementary Fig. 2). These are side evidences of the existence of PNIPAM between the GO laminates. After the polymerization of NIPAM on GO sheets, the thickness of a single P-GO sheet increased to ~5 nm as shown by the atomic force microscope (AFM) image in Fig. 2d. Hence the thickness of the PNIPAM layer was ~4 nm considering the 0.8 nm thickness of pristine GO sheet (Supplementary Fig. 3). To further examine the interactions between GO and PNIPAM, attenuated total reﬂectance Fourier transform infrared (ATR-FTIR) measurements of GOMs, pure PNIPAM, and P-GOMs were conducted (Fig. 2e). Besides the stretching peak of oxygen functional groups of GO at 1743 cm−1, several new peaks at 1642, 1540, and 1367~1460 cm−1 appeared after grafting with PNIPAM39. These new peaks are attributed to the C = O stretching (amide I band), the deformation of N-H bond and CH(CH3)2 groups of PNIPAM, respectively40,41. These characteristic peaks coincided with the spectrum of pure Temperature-responsive property of the P-GOMs. As is well known, PNIPAM will undergo a phase transition behavior in water and form aggregates (which makes the solution turbid) when the temperature is above its LCST. This reversible temperature-response property of PNIPAM, as expected, was well retained after it was covalently bound to GO. Figure 3a shows that the P-GO reversibly dispersed and aggregated upon cooling and heating of the aqueous dispersion between 25 °C and 50 °C. This phenomenon was not observed for the GO aqueous dis- persion (Supplementary Fig. 8), indicating that GO does not have any temperature-response property. As reﬂected by the variable- temperature ultraviolet visible (UV-vis) spectrum, the LCST of P- GO was ~30 °C (determined by the temperature point where the absorption value dramatically increased), which was slightly lower than that of pure PNIPAM (~ 32 °C)44 (Supplementary Fig. 9). NATURE COMMUNICATIONS\| 8: 2011 \| DOI: 10.1038/s41467-017-02198-5\| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 This small variation was because GO was partially reduced NATURE COMMUNICATIONS\| 8: 2011 \| DOI: 10.1038/s41467-017-02198-5\| www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 Temperature (°C) Number of cycles Thickness (µm) Permeance (l m–2 h–1 bar –1) Permeance (l m –2 h –1 bar –1) Permeance (l m –2 h –1 bar –1) Permeance (l m–2 h–1 bar –1) Gating ratio (J25/J50) Time (min) 50 °C 25 °C 0 30 60 90 120 150 180 210 a b c d e f NIR ON NIR OFF NIR ON NIR OFF Heating Cooling 1.1 µm OFF OFF ON ON OFF 40 14 12 10 8 6 4 2 0 12 8 4 0 0 1 2 3 4 5 35 30 25 20 15 10 5 0 15 0.3 25 °C 50 °C 800 600 400 200 50 20 15 10 5 0 1000 0.8 1.1 1.6 0 5 10 15 20 25 30 20 25 30 35 40 45 50 55 60 P-GOMs GOMs GO+PNIPAM 25 °C 50 °C Fig. 3 Thermal- and photo-gating performances of P-GOMs. a P-GO sheets re-dispersed and aggregated upon cooling and heating of the aqueous dispersion between 25 °C and 50 °C, indicating the temperature-responsive property of P-GO. b Temperature-dependent water permeance curves of GOMs, P-GOMs and physically blending PNIPAM/GO membranes, P-GOMs show the unique negative temperature-responsive coefﬁcient, which is opposite to the positive coefﬁcient of GOMs and GO/PNIPAM blending membrane. Error bars, s.d. (n = 3). c Reverse stability test of the temperature responsive water gating property of P-GOMs. d Thickness-dependent water permeance and gating ratio of P-GOMs, 1.1 μm thickness is adopted after balancing the water permeance and high gating ratio. Error bars, s.d. (n = 3). e Schematic showing the water permeance of P-GOMs decreased after irradiation of near-infrared (NIR). The bottom pictures are thermographic images of P-GOMs before and after irradiation of NIR. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 f The water permeance o P-GOMs decreases when switching on NIR light and returns to initial values when the NIR light is off a Heating Cooling Temperature (°C) Permeance (l m–2 h–1 bar –1) b 40 35 30 25 20 15 10 5 0 15 20 25 30 35 40 45 50 55 60 P-GOMs GOMs GO+PNIPAM Number of cycles Permeance (l m –2 h –1 bar –1) c 14 12 10 8 6 4 2 0 0 1 2 3 4 5 25 °C 50 °C b a c Temperature (°C) Temperature (°C) Thickness (µm) Permeance (l m–2 h–1 bar –1) Gating ratio (J25/J50) d 1.1 µm 0.3 25 °C 50 °C 800 600 400 200 50 20 15 10 5 0 1000 0.8 1.1 1.6 0 5 10 15 20 25 30 Permeance (l m –2 h –1 bar –1) Time (min) 50 °C 25 °C 0 30 60 90 120 150 180 210 f R ON R OFF R ON R OFF OFF OFF ON ON OFF 12 8 4 0 f 50 °C 25 °C e NIR ON NIR OFF NIR ON NIR OFF f d e Time (min) Fig. 3 Thermal- and photo-gating performances of P-GOMs. a P-GO sheets re-dispersed and aggregated upon cooling and heating of the aqueous dispersion between 25 °C and 50 °C, indicating the temperature-responsive property of P-GO. b Temperature-dependent water permeance curves of GOMs, P-GOMs and physically blending PNIPAM/GO membranes, P-GOMs show the unique negative temperature-responsive coefﬁcient, which is opposite to the positive coefﬁcient of GOMs and GO/PNIPAM blending membrane. Error bars, s.d. (n = 3). c Reverse stability test of the temperature- responsive water gating property of P-GOMs. d Thickness-dependent water permeance and gating ratio of P-GOMs, 1.1 μm thickness is adopted after balancing the water permeance and high gating ratio. Error bars, s.d. (n = 3). e Schematic showing the water permeance of P-GOMs decreased after irradiation of near-infrared (NIR). The bottom pictures are thermographic images of P-GOMs before and after irradiation of NIR. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 The newly formed bond is highlighted in yellow water transport through P-GOMs may expand their applications such as in lab-on-a-chip settings52. 2θ = 5.91°. The smaller lamellar distance at 50 °C proved the movement of GO sheets because of the shrinking of PNIPAM chains, resulting in a decreased water permeance. Besides, the entanglement of PNIPAM chains between two adjacent GO sheets also results in the decreased water permeance53,54. Fur- thermore, the temperature-response behavior of P-GO was characterized by AFM. The pre-heated P-GO dispersion (~ 50 °C) was spin-coated (2000 rpm) onto freshly cleaved mica to obtain the AFM sample55. As shown in Fig. 4c, the cross-sectional proﬁle indicates the aggregation of P-GO sheets with heights of ~20 nm, implying the dense stacking of P-GO sheets at 50 °C. Mechanism of the negative temperature-response behavior of P-GOMs. The unusual negative temperature-response behavior of P-GOMs, which is quite different from the GO/PNIPAM physical blending membranes36, is due to the covalently bound interactions between PNIPAM chains and GO sheets (Supple- mentary Figs. 13 and 14, Supplementary Note 1). When syn- thesizing P-GO during the free-radical polymerization, there were two cases of the PNIPAM chains termination (except the self- polymerization of NIPAM): two ends of a PNIPAM chain ter- mination by connecting two GO sheets and terminating freely at the bulk solution with only one end of the chain anchored to GO sheet. The AFM images of GO and P-GO at 25 °C presented in Supplementary Fig. 15 show that, both single- and multi-layered P-GO sheets exist because of the interactions of PNIPAM chains with GO sheets, whereas almost no stacked sheet is observed for GO sample. Figure 4a demonstrates that, for the ﬁrst case, the covalently linked PNIPAM would pull two adjacent GO sheets closer by molecular chain contraction when the temperature is above its LCST. For the second case, part of the PNIPAM chains bound to adjacent GO sheets would form intermolecular hydrogen bonds when the temperature is above its LCST, which narrows the distance between two adjacent GO sheets because of the entanglement of PNIPAM chains3. Both these two cases result in a tunable GO lamellar space which play a dominant role in the negative temperature-response gating performance. This hypothesis was proved by XRD and AFM. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 The gating ratio of P-GOMs (J25/J50, where JT is the water permeance at temperature T °C) was ~7 and sustained after several cycles (Fig. 3c). Therefore, the P-GOMs can serve as a hydraulic permeability gating membrane with a relatively large negative gating ratio. Furthermore, we studied the thickness effect of P-GOMs on the gating ratio as shown in Fig. 3d. Assuming that water inside the where J is the water permeance, ε is the surface porosity, r is the pore radius, Δp is the applied pressure, η is the water viscosity, d is the membrane thickness and τ is the tortuosity. According to Equation (1), the water permeance is inversely proportional to the membrane thickness. The thickness variations of P-GOMs with different loading amounts of P-GO aqueous dispersion can be acquired from Supplementary Fig. 12. The water permeance decreased at both of 25 °C and 50 °C as the thickness of P-GOMs increased from 0.3 to 1.6 μm, whereas the gating ratio increased gradually. Although the gating ratio of 23.7 was able to reach when the thickness of P-GOMs was 1.6 μm, the low permeance of only 1.9 l m−2 h−1 bar−1 at 25 °C would restrict its scope of applicability. Thus, the P-GOMs with thickness of 1.1 μm were chosen for the temperature-response gating experiments. Besides directed thermal tunable permeation, we realized remote control of the membrane permeability via light because of the superior light-to-heat conversion capability of GO. This would be implemented more conveniently in deserved applica- tions14,32,49. When the P-GOMs were exposed under 808 nm near-infrared (NIR) light at power density of 0.3 W cm−2 in air at an initial temperature of ~25 °C, its temperature rose across the LCST (30 °C) (Fig. 3f). Water has a low absorption at 808 nm NIR while the P-GOMs have a strong absorption, thus the water permeance of the P-GOMs can be regulated by the NIR light (Fig. 3e)50,51. As shown in Fig. 3f, the water permeance was ~11.2 l m−2 h−1 bar−1 at ambient temperature (~ 25 °C) without the NIR light. After irradiation, the water permeance of P-GOMs decreased to ~2.8 l m−2 h−1 bar−1. The cycle of light-induced temperature-responsive gating performance can be repeated as the Fig. 3c shows without obvious decay. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 a Schematic diagram of the channel size change of P-GOMs from D1 to D2 caused by the swelling or shrinking of PNIPAM chains covalent bound to GO sheets. Case 1: PNIPAM chains grow from one GO sheet and terminate at another GO sheet. Intramolecular hydrogen bonding is formed when the temperature is above its LCST. Case 2: PNIPAM chains grow from one GO sheet and terminate freely. Intermolecular hydrogen bonding is formed when the temperature is above its LCST. b The XRD patterns of P-GOMs at 25 °C and 50 °C. The right shift of the diffraction peak indicates the smaller lamellar distance at 50 °C than 25 °C. c The AFM image of P-GO. The height of cross- sectional proﬁle proves that the P-GO sheets would like to stack together at 50 °C. d Optimized geometries of [C36H16-CH3]• and the representative products of reactions (1b) C36H16-CH3-NIPAM-CH3-C36H16, (2b) C36H16-CH3-(NIPAM)2-CH3-C36H16, (3b) C36H16-CH3-(NIPAM)3-CH3-C36H16 and (4b) C36H16-CH3-(NIPAM)4-CH3-C36H16. Others are shown in Supplementary Fig. 16. The graphene sheet is shown in green tubes, the PNIPAM as tubes (atomic color scheme: C, gray; N, blue; O, red; H, white). The newly formed bond is highlighted in yellow Fig. 4 Mechanism of the negative temperature-response behavior of P-GOMs. a Schematic diagram of the channel size change of P-GOMs from D1 to D2 caused by the swelling or shrinking of PNIPAM chains covalent bound to GO sheets. Case 1: PNIPAM chains grow from one GO sheet and terminate at another GO sheet. Intramolecular hydrogen bonding is formed when the temperature is above its LCST. Case 2: PNIPAM chains grow from one GO sheet and terminate freely. Intermolecular hydrogen bonding is formed when the temperature is above its LCST. b The XRD patterns of P-GOMs at 25 °C and 50 °C. The right shift of the diffraction peak indicates the smaller lamellar distance at 50 °C than 25 °C. c The AFM image of P-GO. The height of cross- sectional proﬁle proves that the P-GO sheets would like to stack together at 50 °C. d Optimized geometries of [C36H16-CH3]• and the representative products of reactions (1b) C36H16-CH3-NIPAM-CH3-C36H16, (2b) C36H16-CH3-(NIPAM)2-CH3-C36H16, (3b) C36H16-CH3-(NIPAM)3-CH3-C36H16 and (4b) C36H16-CH3-(NIPAM)4-CH3-C36H16. Others are shown in Supplementary Fig. 16. The graphene sheet is shown in green tubes, the PNIPAM as tubes (atomic color scheme: C, gray; N, blue; O, red; H, white). NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 The photocontrol of NATURE COMMUNICATIONS\| 8: 2011 \| DOI: 10.1038/s41467-017-02198-5\| www.nature.com/naturecommunications 4 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 a b c d 100 nm µm 0.0 –5 5 0 10 15 20 0.1 0.2 0.3 0.4 0.5 0.6 nm (1b): –65.2 kJ mol–1 (2b): –154.6 kJ mol–1 (3b): –254.5 kJ mol–1 (4b): –396.0 kJ mol–1 < LCST > LCST Case 1 Intra- Inter- D2 D1 2 Theta (degree) 5 10 15 25 °C 50 °C Intensity (a.u.) Case 2 Fig. 4 Mechanism of the negative temperature-response behavior of P-GOMs. a Schematic diagram of the channel size change of P-GOMs from D1 to D2 caused by the swelling or shrinking of PNIPAM chains covalent bound to GO sheets. Case 1: PNIPAM chains grow from one GO sheet and terminate at another GO sheet. Intramolecular hydrogen bonding is formed when the temperature is above its LCST. Case 2: PNIPAM chains grow from one GO sheet and terminate freely. Intermolecular hydrogen bonding is formed when the temperature is above its LCST. b The XRD patterns of P-GOMs at 25 °C and 50 °C. The right shift of the diffraction peak indicates the smaller lamellar distance at 50 °C than 25 °C. c The AFM image of P-GO. The height of cross- sectional proﬁle proves that the P-GO sheets would like to stack together at 50 °C. d Optimized geometries of [C36H16-CH3]• and the representative products of reactions (1b) C36H16-CH3-NIPAM-CH3-C36H16, (2b) C36H16-CH3-(NIPAM)2-CH3-C36H16, (3b) C36H16-CH3-(NIPAM)3-CH3-C36H16 and (4b) C36H16-CH3-(NIPAM)4-CH3-C36H16. Others are shown in Supplementary Fig. 16. The graphene sheet is shown in green tubes, the PNIPAM as tubes (atomic color scheme: C, gray; N, blue; O, red; H, white). The newly formed bond is highlighted in yellow b 2 Theta (degree) 5 10 15 25 °C 50 °C Intensity (a.u.) a < LCST > LCST Case 1 Intra- Inter- D2 D1 Case 2 c 100 nm µm 0.0 –5 5 0 10 15 20 0.1 0.2 0.3 0.4 0.5 0.6 nm b a c Intensity (a.u.) 2 Theta (degree) (3b): –254.5 kJ mol–1 (4b): –396.0 kJ mol–1 d (1b): –65.2 kJ mol–1 (2b): –154.6 kJ mol–1 d (1b): –65.2 kJ mol–1 (2b): –154.6 kJ mol–1 (3b): –254.5 kJ mol–1 (3b): –254.5 kJ mol–1 (4b): –396.0 kJ mol–1 Fig. 4 Mechanism of the negative temperature-response behavior of P-GOMs. NATURE COMMUNICATIONS\| 8: 2011 \| DOI: 10.1038/s41467-017-02198-5\| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 c) aqueous solutions from 25 °C to 50 °C. e–h The digital photos of original solutions and the ﬁltrate obtained at 25 °C and 50 °C of [Fe(CN)6]3−, RB, CBB, and Cyt. c aqueous solutions Moreover, we suppose that the P-GOMs could realize gradient separation of multiple molecules with different sizes. (1b), and −154.6 kJ mol−1 (2b). For the consecutive reactions, three and four NIPAM monomers react with one or two functionalized graphene sheets. In this situation, all the reactions are exothermic. However, the reaction enthalpies for the reactions involving one functionalized graphene sheet were −158.3 (3a) and −262.7 kJ mol−1 (4a), whereas the reactions involving two graphene sheets were much more exothermic with reaction enthalpies of −254.5 (3b) and −396.0 kJ mol−1 (4b). The higher exothermicities of the later reactions could be partly attributed to the stronger van der Waals interactions between the graphene sheets and the side chain of the PNIPAM. These computational results indicate that the polymerization reaction between graphene and PNIPAM will prefer to be terminated by another functionalized graphene sheet as case (1) showed in Fig. 4a (Supplementary Note 3). We have shown that this is the case for 1–4 NIPAM molecules, however this trend is expected to continue for larger numbers of NIPAM molecules. (1b), and −154.6 kJ mol−1 (2b). For the consecutive reactions, three and four NIPAM monomers react with one or two functionalized graphene sheets. In this situation, all the reactions are exothermic. However, the reaction enthalpies for the reactions involving one functionalized graphene sheet were −158.3 (3a) and −262.7 kJ mol−1 (4a), whereas the reactions involving two graphene sheets were much more exothermic with reaction enthalpies of −254.5 (3b) and −396.0 kJ mol−1 (4b). The higher exothermicities of the later reactions could be partly attributed to the stronger van der Waals interactions between the graphene sheets and the side chain of the PNIPAM. These computational results indicate that the polymerization reaction between graphene and PNIPAM will prefer to be terminated by another functionalized graphene sheet as case (1) showed in Fig. 4a (Supplementary Note 3). We have shown that this is the case for 1–4 NIPAM molecules, however this trend is expected to continue for larger numbers of NIPAM molecules. The multiple molecules separation schematic diagram is illustrated in Fig. 6b. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 The smallest molecule A is separated out at 50 °C, leaving the medium and largest molecules (B and C) in the retentate. Then the temperature changes to 25 °C to separate the medium size molecule from the large molecule. Thus, we can separate multiple molecules with different sizes using the P-GOMs by simple regulating the temperature. Subsequently, we designed a separation scenario using the mixed molecules solution, containing Cu2+, RB, and Cyt. c (Fig. 6c, d). Figure 6e shows the UV-vis absorption spectra of ﬁltrates obtained at 50 °C and 25 °C and retentate. For the Cu2+, it only existed in the ﬁltrate obtained at 50 °C according to the results of inductively coupled plasma optical emission spectrometry. In the ﬁltrate obtained at 50 °C, there were trace RB and Cyt. c according to its UV-vis absorption spectrum. For RB, it enriched in the ﬁltrate obtained at 25 °C. Therefore, only the largest Cyt. c remained in the retentate obtained at 25 °C. The results veriﬁed the selective separation performance of the P-GOMs to the mixed ions/ molecules solution (Supplementary Fig. 20). Separation performance of the P-GOMs. To validate the results of temperature tunable channel size of P-GOMs discussed above, we designed a separation experiment to test the lamellar space change at different temperatures by using small molecules with different sizes. Here we employed ﬁve ions/molecules, Cu2+ (0.8 nm), [Fe(CN)6]3−(0.9 × 0.9 nm), rhodamine B (RB, 1.8 × 1.4 nm), coomassie brilliant blue (CBB, 2.7 × 1.8 nm), and cytochrome c (Cyt. c, 2.5 × 2.5 × 3.7 nm)22,29,57. The rejection rates of the ﬁve ions/molecules through the P-GOMs from 25 °C to 50 °C were measured. The zeta potential results (Supplemen- tary Fig. 17) demonstrate that the P-GOMs have few anionic sites for adsorption of the positively charged molecules and rejection of the negatively charged molecules because of the relatively low absolute value of the zeta potential. Thus size-dependent rejection rates nearly indicate the nanochannels size change of P-GOMs58,59. The results in Figs. 5 and 6a demonstrate that the nanochannel sizes of the P-GOMs gradually decrease with the increasing temperature (Supplementary Table 2). These results not only veriﬁed the results of quantum chemical calculations, but also demonstrated that the negative temperature-response P-GOMs could be promising for precise molecular separation (Supplementary Figs. 18 and 19, Supplementary Note 4). NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 Original 25 °C 30 °C 35 °C 40 °C 45 °C 50 °C Original 25 °C 30 °C 35 °C 40 °C 45 °C 50 °C Original 25 °C 30 °C 35 °C 40 °C 45 °C 50 °C Original 25 °C 30 °C 35 °C 40 °C 45 °C 50 °C [Fe(CN)6]3– RB CBB Wavelength (nm) Wavelength (nm) Wavelength (nm) Wavelength (nm) Absorbance (a.u.) Absorbance (a.u.) Absorbance (a.u.) Absorbance (a.u.) Cyt. c [Fe(CN)6]3– 25 °C 50 °C 25 °C 50 °C RB CBB 25 °C 50 °C Cyt. c 25 °C 50 °C a b c d e f g h 12.5% 48.2% 68.2% 100.0% 98.0% 100.0% 99.8% 100.0% 250 300 350 400 450 500 300 350 400 450 500 550 400 450 500 550 600 650 700 450 500 550 600 650 700 750 Fig. 5 P-GOMs for single molecule separation. a–d The UV-vis absorption spectra before and after ﬁltering [Fe(CN)6]3−aqueous solutions, rhodamine B (RB) aqueous solutions, coomassie brilliant blue (CBB) aqueous solutions, and cytochrome c (Cyt. c) aqueous solutions from 25 °C to 50 °C. e–h The digital photos of original solutions and the ﬁltrate obtained at 25 °C and 50 °C of [Fe(CN)6]3−, RB, CBB, and Cyt. c aqueous solutions Original 25 °C 30 °C 35 °C 40 °C 45 °C 50 °C [Fe(CN)6]3– Wavelength (nm) Absorbance (a.u.) a 250 300 350 400 450 500 Original 25 °C 30 °C 35 °C 40 °C 45 °C 50 °C RB Wavelength (nm) Absorbance (a.u.) b 400 450 500 550 600 650 700 Original 25 °C 30 °C 35 °C 40 °C 45 °C 50 °C CBB Wavelength (nm) Absorbance (a.u.) c 450 500 550 600 650 700 750 b Original 25 °C 30 °C 35 °C 40 °C 45 °C 50 °C Wavelength (nm) Absorbance (a.u.) Cyt. c d 300 350 400 450 500 550 d c Absorbance (a.u.) Wavelength (nm) Wavelength (nm) Wavelength (nm) Wavelength (nm) [Fe(CN)6]3– 25 °C 50 °C e 12.5% 48.2% f h Cyt. c 25 °C 50 °C h 99.8% 100.0% CBB 25 °C 50 °C g 98.0% 100.0% 25 °C 50 °C RB f 68.2% 100.0% e g Fig. 5 P-GOMs for single molecule separation. a–d The UV-vis absorption spectra before and after ﬁltering [Fe(CN)6]3−aqueous solutions, rhodamine B (RB) aqueous solutions, coomassie brilliant blue (CBB) aqueous solutions, and cytochrome c (Cyt. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 We have reported a smart temperature-response gating membrane composed of GO and PNIPAM for hydraulic permeability applications. The membranes have negative response coefﬁcients with a gating ratio of ~7. What’s more, the gating membranes could realize gradual small/medium/large molecule separation by simple step-wise tuning of the tempera- ture. This study provides insight toward the design of smart gating membranes, showing promise for applications in ﬂuid transport systems, microﬂuidic chip systems, and molecular separation devices. Materials. GO was prepared by a modiﬁed Hummer’s method. Ethanol, 2,2’- Azobis(2-methylpropionitrile) (≥99.0%, recrystallization), N, N-dimethylforma- mide (DMF, ≥99.8%, anhydrous), copper (II) chloride (CuCl2), potassium ferri- cyanide (K3[Fe(CN)6]), rhodamine B (RB), coomassie brilliant blue, cytochrome c (Cyt. c), and rafﬁnose were all purchased from Aladdin Chemistry Co., Ltd. (Shanghai, China) and used as received. Maltopentaose and maltoheptaose (HPLC) were provided by Miragen (USA). NIPAM was provided by Sigma-Aldrich (Shanghai, China) and puriﬁed by recrystallization from n-hexane ( ≥99%, Alad- din). PNIPAM (Mn ~ 20,000–40,000) was provided by Sigma-Aldrich. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 For the P-GOMs in wet state, because PNIPAM was intercalated into the GO sheets to form a layered nanostructure, the diffraction peak of GO (Supplementary Fig. 10) disappeared and a new peak at 2θ = 4.71° appeared. When the temperature increased to 50 °C, the diffrac- tion peak of P-GOMs shifted right to p y g g In order to further understand the mechanism, quantum chemical calculations were performed using density functional theory to elucidate the reaction enthalpy for the formation of the graphene-PNIPAM. In radical polymerization, a polymer forms by the successive addition of radicals. The aim of this computational investigation is to establish whether the reaction of functionalized graphene with NIPAM monomers would terminate composite molecules freely in bulk solution or terminate at another functionalized graphene sheet. The poly- merization reaction mechanism for the two possibilities along with the full computational details are given as Supplementary Note 2. Mylvaganam et al. found that the carbon adjacent to the sp3 carbon is prone to be attacked by an ethylene moiety56. Owing to steric factors the graphene sheet is expected to react with NIPAM molecules on the opposite side of the methyl group. Supplementary Fig. 16 and Fig. 4d show the optimized geometries of these products in reactions (1a–4a) and (1b–4b) (Supplemen- tary Note 2). The calculated reaction enthalpies for these two sets of reactions are given in Supplementary Table 1. An important result is that the ﬁrst two reactions involving one functionalized graphene sheet (1a and 2a) are endothermic, whereas the corresponding reactions including two functionalized graphene sheets (1b and 2b) are exothermic. Speciﬁcally, the reaction enthalpies at 298 K (ΔH298) were + 50.9 (1a), + 20.8 (2a), −65.2 right to 5 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 The average values were obtained by measuring three independent samples for each. What is more, in order to keep the GOM stable in water, a negative pressure was applied to the pressure-driven ﬁltration when added 10 ml DI water above the GOM until putting pressure on the membrane. Permeability performances of P-GOMs, GOMs and GO blending with PNI- PAM membranes. For the temperature-responsive hydraulic permeability experiments, a water bath was used to regulate the temperature of the membrane. The ﬁltration device contained the formed membrane and 10 ml DI water was put into the water bath with desired temperature. After standing for 30 mins, 100 kPa pressure generated by N2 was applied to the membrane. The weight of water penetrated through the membrane was recorded every 30 mins, and the water permeance was calculated using the value which kept constant for >90 mins. For the photo-responsive hydraulic permeability experiments, a vacuum ﬁltration system combined with an 808 nm NIR laser was used. well-dispersed P-GO aqueous suspension during the whole synthesis process except dispersing GO into DMF at ﬁrst. Preparation of GO dispersion. GO (0.009 g) was added into 100 ml DI water followed by stirring for 1 h and ultrasonic treatment for 2 h to obtain a well- dispersed GO aqueous dispersion (0.09 mg ml−1). Molecular separation experiments of P-GOMs. While for the molecule separation experiments, the aqueous solutions of 100 mg l−1 CuCl2, 40 mg l−1 K3[Fe(CN)6], 4 mg l−1 RB (Supplementary Fig. 21), 40 mgl−1 coomassie brilliant blue (Supplementary Fig. 22), 100 mg l−1 Cyt. c, 200 mg l−1 rafﬁnose, 200 mg l−1 maltopentaose and 200 mg l−1 maltoheptaose were used instead of DI water, and the water bath contained the ﬁltration device with a stir bar was placed on a magnetic heated stirrer in case of concentration polarization. After certain time, the ﬁltrates were collected. The rejection rate (R) was calculated using the following equation: Preparation of GO blending with PNIPAM dispersion. PNIPAM (0.021 g) was added into 100 ml GO aqueous dispersion (0.09 mg ml−1) and stirred for 6 h to obtain the homogeneous GO blending with PNIPAM dispersion. Fabrication of P-GOMs, GOMs, and GO blending with PNIPAM membranes. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 50 °C B + C A B C 25 °C Large Medium Small Cu2+ (0.8 nm) 50 °C 25 °C RB (1.8 × 1.4 nm) Cyt. c (2.5 × 2.5 × 3.7 nm) Filtrate at 25 °C RB Cyt. c Mixed molecules Filtrate at 50 °C Retentate Wavelength (nm) Absorbance (a.u.) Temperature (°C) 0 25 30 35 40 45 50 300 400 500 600 700 20 40 60 80 100 Cu2+ RB CBB Cyt. c [Fe(CN)6]3– Rejection rate (%) a b c d e A + B + C Feed solution Mixed Retentate at 50 °C Filtrate at 50 °C Filtrate at 25 °C Retentate at 25 °C RB + Cyt. c Cu2+ RB Cyt. c Fig. 6 P-GOMs for multiple molecules separation. a Temperature-dependent rejection rates of P-GOMs to Cu2+, [Fe(CN)6]3−, RB, CBB, and Cyt. c, implying smaller channel size of the P-GOMs with the increasing temperature. Error bars, s.d. (n = 3). b Schematic diagram of the separation mechanism of P-GOMs towards mixed molecules solution. c Separation of mixed molecules solution, containing Cu2+, RB, and Cyt. c. d The digital photos of the mixed molecules feed solution, retentate at 50 °C, ﬁltrate at 50 °C, ﬁltrate at 25 °C, and retentate at 25 °C. e The UV-vis absorption spectra of pure RB aqueous solution, Cyt. c aqueous solution, mixed molecules solution, ﬁltrate obtained at 50 °C and 25 °C, and retentate at 25 °C Temperature (°C) 0 25 30 35 40 45 50 20 40 60 80 100 Cu2+ RB CBB Cyt. c [Fe(CN)6]3– Rejection rate (%) a b Filtrate at 25 °C RB Cyt. c Mixed molecules Filtrate at 50 °C Retentate Wavelength (nm) Absorbance (a.u.) 300 400 500 600 700 e a e Temperature (°C) 50 °C p ( c Absorbance (a.u.) c d Wavelength (nm) 500 Fig. 6 P-GOMs for multiple molecules separation. a Temperature-dependent rejection rates of P-GOMs to Cu2+, [Fe(CN)6]3−, RB, CBB, and Cyt. c, implying smaller channel size of the P-GOMs with the increasing temperature. Error bars, s.d. (n = 3). b Schematic diagram of the separation mechanism of P-GOMs towards mixed molecules solution. c Separation of mixed molecules solution, containing Cu2+, RB, and Cyt. c. d The digital photos of the mixed molecules feed solution, retentate at 50 °C, ﬁltrate at 50 °C, ﬁltrate at 25 °C, and retentate at 25 °C. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 e The UV-vis absorption spectra of pure RB aqueous solution, Cyt. c aqueous solution, mixed molecules solution, ﬁltrate obtained at 50 °C and 25 °C, and retentate at 25 °C Preparation of P-GO dispersion. The P-GO were synthesized by free-radical polymerization. Typically, GO (0.05 g) was ﬁrst fully dispersed in DMF (50 ml) by stirring for 1 h and ultrasonic treatment for 2 h. Afterwards, NIPAM (50 mmol, 5.659 g) and 2,2’-Azobis(2-methylpropionitrile) (0.056 g) were added to the dis- persion, respectively. The mixture was subsequently transferred to a 100 ml Schlenk ﬂask followed by stirring for 30 mins. Next, the reagent mixture was subjected to three freeze-pump-thaw cycles and was then placed in a 65 °C oil bath. The reaction was terminated after 12 h by immersing the ﬂask into liquid nitrogen. The obtained solution was diluted with 250 ml DI water and transferred into centrifuge tubes followed by 1 h centrifugation at the speed of 10,000 r min−1 (10,744 g). The precipitates located in the bottom of centrifuge tubes after cen- trifugation were collected and washed three times each with 300 ml DI water and 300 ml ethanol. To make sure NIPAM monomer and the PNIPAM not bound to GO were completely separated from the P-GO, the precipitates were puriﬁed through dialysis in DI water for 3 days using dialysis bag (MW cutoff: 30,000). Then the P-GO were obtained after freezing dry of the dispersion for 2 days with a freezing dryer. A brownish transparent, stable P-GO dispersion (0.1 mg ml−1) was obtained by adding 0.01 g of the as-prepared P-GO in 100 ml DI water and stirring for 2 h. Note, no ultrasound treatment was needed to obtain the well-dispersed P-GO aqueous suspension during the whole synthesis process di i GO i DMF ﬁ P-GOM in a wet state. The mass of GO in a P-GOM was ~0.00018 g based on the calculation from the TGA and differential thermogravimetrie curves as described in Supplementary Note 1. Thus, the procedure was basically the same as the P-GOM for the fabrication of GOM or GO blending with PNIPAM membrane, except using 2 ml GO dispersion (0.09 mg ml−1) or 2 ml GO blending with PNIPAM dispersion (containing 0.00018 g GO) instead of the P-GO dispersion. The GO mass in three kinds of membranes were all ﬁxed at ~0.18 mg. Methods Materials G Materials. GO was prepared by a modiﬁed Hummer’s method. Ethanol, 2,2’- Azobis(2-methylpropionitrile) (≥99.0%, recrystallization), N, N-dimethylforma- mide (DMF, ≥99.8%, anhydrous), copper (II) chloride (CuCl2), potassium ferri- cyanide (K3[Fe(CN)6]), rhodamine B (RB), coomassie brilliant blue, cytochrome c (Cyt. c), and rafﬁnose were all purchased from Aladdin Chemistry Co., Ltd. (Shanghai, China) and used as received. Maltopentaose and maltoheptaose (HPLC) were provided by Miragen (USA). NIPAM was provided by Sigma-Aldrich (Shanghai, China) and puriﬁed by recrystallization from n-hexane ( ≥99%, Alad- din). PNIPAM (Mn ~ 20,000–40,000) was provided by Sigma-Aldrich. 6 NATURE COMMUNICATIONS\| 8: 2011 \| DOI: 10.1038/s41467-017-02198-5\| www.nature.com/naturecommunications R ¼ 1 ðcp=coÞ ´ 100% References where cp and co are the concentrations of the molecule in the permeates and original solutions. where cp and co are the concentrations of the molecule in the permeates and original solutions. 1. Collatz, G. J., Ball, J. T., Grivet, C. & Berry, J. A. Physiological and environmental regulation of stomatal conductance, photosynthesis and transpiration: a model that includes a laminar boundary layer. Agric. For. Meteorol. 54, 107–136 (1991). For the mixed molecules separation experiments, 5 ml aqueous solution contained 100 mg l−1 Cu2+, 4 mg l−1 RB, and 100 mg l−1 Cyt. c was used as the feed solution. First, the ﬁltration device was put in the 50 °C water bath, after standing for 30 mins, 100 kPa pressure was applied to the membrane. After collecting ~4.5 ml ﬁltrate, 4.5 ml DI water was added into the ﬁltration device, followed by the second separation at 50 °C. The temperature was altered to 2. Park, C. H. et al. Nanocrack-regulated self-humidifying membranes. Nature 532, 480–483 (2016). 3. Chu, L. Y., Li, Y., Zhu, J. H. & Chen, W. M. Negatively thermoresponsive membranes with functional gates driven by zipper-type hydrogen-bonding interactions. Angew. Chem. Int. Ed. 44, 2124–2127 (2005). 25 °C until no Cu2+ could be detected in the ﬁltrate at 50 °C. Like the separation process at 50 °C, after several separations at 25 °C, the ﬁltrate contained only RB could be obtained. The retentate contained only Cyt. c could be obtained as well in the end. 25 °C until no Cu2+ could be detected in the ﬁltrate at 50 °C. Like the separation process at 50 °C, after several separations at 25 °C, the ﬁltrate contained only RB could be obtained. The retentate contained only Cyt. c could be obtained as well in the end. 4. Liu, Z., Wang, W., Xie, R., Ju, X. J. & Chu, L. Y. Stimuli-responsive smart gating membranes. Chem. Soc. Rev. 45, 460–475 (2016). 5. Schepelina, O. & Zharov, I. PNIPAAM-modiﬁed nanoporous colloidal ﬁlms with positive and negative temperature gating. Langmuir 23, 12704–12709 (2007). Computational details. To simplify the calculations, we used a methyl radical (CH3•) to mimic the experimentally carbon related radical (namely CN(CH3)2C•) as the initiator. References A graphene sheet containing 36 carbon atoms, C36H16, was used to study the polymerization reaction instead of GO used in experiment because the polymer chains were grafted to graphitic surface utilizing the double bonds of surfaces in the free-radical polymerization42. Using graphene sheet did not inﬂu- ence the results, however, substantially saved computational cost. The initiation, propagation, and termination steps of a free-radical polymerization reaction were carried out by using NIPAM as a monomer and CH3• as the initiator. The reaction of a methyl radical with graphene and the subsequent reaction of the newly formed graphene-methyl radical with NIPAM were examined for the initiation step. The subsequent reaction of graphene-methyl- NIPAM radical with another NIPAM monomer was examined for the propagation step. The termination step was examined by considering the reaction of the resultant radical with another methyl functionalized graphene sheet. For every step, geometries of the reactants and products were optimized using density functional theory method with a hybrid functional B3LYP60–62 and a 6–31 G(d, p) basis set. Empirical D3 dispersion corrections are included using the Becke–Johnson damping potential as recom- mended in Ref.63 (denoted by the sufﬁx-D3). Bulk solvent effects of DMF (ε = 37.2) have been included using charge-density based SMD continuum solvation model at the same level of theory64. All calculations were performed by using Gaussian 09 package65. 6. Yameen, B. et al. Ionic transport through single solid-state nanopores controlled with thermally nanoactuated macromolecular gates. Small 5, 1287–1291 (2009). 7. Guo, W. et al. Current rectiﬁcation in temperature-responsive single nanopores. Chem. Phys. Chem. 11, 859–864 (2010). 8. Morones-Ramírez, J. R. Bioinspired synthesis of optically and thermally responsive nanoporous membranes. NPG Asia Mater. 5, e52 (2013). 9. Hou, X. Smart gating multi-scale pore/channel-based membranes. Adv. Mater. 28, 7049–7064 (2016). 10. Halperin, A., Kröger, M. & Winnik, F. M. Poly (N-isopropylacrylamide) phase diagrams: ﬁfty years of research. Angew. Chem. Int. Ed. 54, 15342–15367 (2015). 11. Wang, N., Zhao, Y. & Jiang, L. Low-cost, thermoresponsive wettability of surfaces: poly (n-isopropylacrylamide)/polystyrene composite ﬁlms prepared by electrospinning. Macromol. Rapid Commun. 29, 485–489 (2008). 12. Junk, M. J. et al. EPR spectroscopic characterization of local nanoscopic heterogeneities during the thermal collapse of thermoresponsive dendronized polymers. Angew. Chem. Int. Ed. 49, 5683–5687 (2010). p y g 13. Li, H. et al. Ultrathin, molecular-sieving graphene oxide membranes for selective hydrogen separation. Science 342, 95–98 (2013). 14. Kurapati, R. & Raichur, A. M. References Near-infrared light-responsive graphene oxide composite multilayer capsules: a novel route for remote controlled drug delivery. Chem. Commun. 49, 734–736 (2013). Characterization. Digital photographs were shot by a digital camera. Scanning electron microscope images were obtained on a scanning electron microscopy (JSM-7500F, JEOL, Japan) equipped with a energy dispersive spectroscopy detec- tor. AFM images were carried out on a Bruker Dimension Icon (Japan). Diluted GO or P-GO aqueous dispersion was spin-coated (2000 rpm) on freshly cleaved mica and dried at room temperature to make the AFM samples. Thermographic images were obtained with a FLIR TG165 imaging IR thermometer (FLIR Systems, Inc., OR, USA). The values of contact angle were measured on an OCA20 machine (Dataphysics, Germany) at room temperature. ATR-FTIR spectra were measured by an infrared microspectrometer (Thermo Scientiﬁc Nicolet iN10, USA). XPS characterizations were performed by ESCALAB 250Xi (Thermo Fisher Scientiﬁc, USA) equipped with a monochromatic Al Kα X-ray source. Raman spectroscopy measurements were performed using a Lab-RAM HR800 (Horiba JobinYvon, France) with an incident laser of 514 nm wavelength. The XRD analyses were carried out by an X-ray diffractometer (XRD-6000, Shimadzu, Japan), using Cu Kα radiation (λ = 0.15418 nm). The 1H nuclear magnetic resonance spectra were obtained on a Bruker Avance III 700 MHz NMR spectrometer (Bruker Biospin, Rheinstetten, Germany). TGA was performed on a thermogravimetric analyzer (STA449F3, Netzsch, Germany), with a temperature increase rate of 10 K min−1 under nitrogen from 25 °C to 750 °C. The thicknesses of P-GOMs were measured by Bruker DektakXT Stylus Proﬁler (Germany). The concentrations of Cu2+ in the solution were analyzed using inductively coupled plasma optical emission spec- trometry (ICP-OES, Optima 5300DV, PerkinElmer, USA). The concentrations of rafﬁnose, maltopentaose and maltoheptaose in the solution were analyzed by ion chromatography equipped with a CarboPac PA200 3 × 250 column. UV-vis absorption spectra were obtained with a Shimadzu UV-3600 spectrometer (Japan) equipped with a thermal controller. Zeta potentials were analyzed by DelsaNano C Zeta Potential Analyzer (Beckman Coulter Inc., USA). Tensile mechanical prop- erties were measured at room temperature (25 °C) via a Shimadzu AGS-X Tester (Japan) with a 20 N load cell at a loading rate of 1 mm min−1. The samples were cut into strips with a width of 3 mm and length of 10 mm, and the thicknesses of all samples were calculated by scanning electron microscope thickness measuring instrument. 15. Sun, P., Wang, K. & Zhu, H. ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 For the preparation of a P-GOM, 6 ml (except discussing as variable) P-GO aqu- eous dispersion (0.1 mg ml−1) was ﬁltered through a CA membrane by pressure- driven ﬁltration (Millipore 8010, Millipore, USA) with a ﬁlter area of 2.83 cm2 under the pressure of 100 kPa. Once no P-GO dispersion left above the formed membrane, 10 ml DI water was immediately added to the ﬁltration cell to keep the R ¼ 1 ðcp=coÞ ´ 100% ð2Þ 7 7 NATURE COMMUNICATIONS\| 8: 2011 \| DOI: 10.1038/s41467-017-02198-5\| www.nature.com/naturecommunications Received: 27 June 2017 Accepted: 13 November 2017 ARTICLE 01; (Gaussian: Wallingford, CT, USA, 2009). 39. Moon, I. K., Lee, J., Ruoff, R. S. & Lee, H. Reduced graphene oxide by chemical graphitization. Nat. Commun. 1, 73 (2010). 40. Kundu, A., Nandi, S., Das, P. & Nandi, A. K. Fluorescent graphene oxide via polymer grafting: an efﬁcient nanocarrier for both hydrophilic and hydrophobic drugs. ACS Appl. Mat. Interfaces 7, 3512–3523 (2015). Acknowledgements We are grateful for the ﬁnancial support from the National Natural Science Foundation of China (21774005, 21433012, 21374001, 21222309), the National Youth Talent Support Program, the Program for New Century Excellent Talents in University of China, the Fundamental Research Funds for the Central Universities, the China Scholarship Council (201506025110) and the National Instrumentation Program (2013YQ120355). The work at Harvard was supported by the NSF (DMR-1310266) and the Harvard MRSEC (DMR- 1420570). We are grateful for the ﬁnancial support from the National Natural Science Foundation of China (21774005, 21433012, 21374001, 21222309), the National Youth Talent Support Program the Program for New Century Excellent Talents in University of China the We are grateful for the ﬁnancial support from the National Natural Science Foundation of China (21774005, 21433012, 21374001, 21222309), the National Youth Talent Support Program, the Program for New Century Excellent Talents in University of China, the Fundamental Research Funds for the Central Universities, the China Scholarship Council (201506025110) and the National Instrumentation Program (2013YQ120355). The work at Harvard was supported by the NSF (DMR-1310266) and the Harvard MRSEC (DMR- 1420570). 41. Qi, J., Lv, W., Zhang, G., Zhang, F. & Fan, X. Poly (N-isopropylacrylamide) on two-dimensional graphene oxide surfaces. Polym. Chem. 3, 621–624 (2012). 42. Kan, L., Xu, Z. & Gao, C. General avenue to individually dispersed graphene oxide-based two-dimensional molecular brushes by free radical polymerization. Macromolecules 44, 444–452 (2010). 43. Liu, J. et al. A graphene-like oxygenated carbon nitride material for improved cycle-life lithium/sulfur batteries. Nano. Lett. 15, 5137–5142 (2015). ARTICLE ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-02198-5 31. Hou, X., Hu, Y., Grinthal, A., Khan, M. & Aizenberg, J. Liquid-based gating mechanism with tunable multiphase selectivity and antifouling behaviour. Nature 519, 70–73 (2015). 58. Patil, S., Sandberg, A., Heckert, E., Self, W. & Seal, S. Protein adsorption and cellular uptake of cerium oxide nanoparticles as a function of zeta potential. Biomaterials 28, 4600–4607 (2007). 32. Hu, L. et al. Photothermal-responsive single-walled carbon nanotube-based ultrathin membranes for on/off switchable separation of oil-in-water nanoemulsions. ACS Nano 9, 4835–4842 (2015). 59. Zhu, S. et al. Strongly green-photoluminescent graphene quantum dots for bioimaging applications. Chem. Commun. 47, 6858–6860 (2011). 60. Lee, C., Yang, W. & Parr, R. G. Development of the Colle-Salvetti correlation- energy formula into a functional of the electron density. Phys. Rev. B 37, 785 (1988). 33. Huang, H. et al. Salt concentration, pH and pressure controlled separation of small molecules through lamellar graphene oxide membranes. Chem. Commun. 49, 5963–5965 (2013). 61. Becke, A. D. Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 98, 5648–5652 (1993). 34. Liu, R. et al. Graphene oxide membrane for liquid phase organic molecular separation. Carbon N. Y. 77, 933–938 (2014). 62. Stephens, P., Devlin, F., Chabalowski, C. & Frisch, M. J. Ab initio calculation of vibrational absorption and circular dichroism spectra using density functional force ﬁelds. J. Phys. Chem. 98, 11623–11627 (1994). 35. Sun, P. et al. Electro-and magneto-modulated ion transport through graphene oxide membranes. Sci. Rep. 4, 6798 (2014). 36. Wang, Y. et al. Graphene-directed supramolecular assembly of multifunctional polymer hydrogel membranes. Adv. Funct. Mater. 25, 126–133 (2015). 63. Grimme, S., Ehrlich, S. & Goerigk, L. Effect of the damping function in dispersion corrected density functional theory. J. Comput. Chem. 32, 1456–1465 (2011). 37. Dong, J., Weng, J. & Dai, L. The effect of graphene on the lower critical solution temperature of poly (N-isopropylacrylamide). Carbon N. Y. 52, 326–336 (2013). 64. Marenich, A. V., Cramer, C. J. & Truhlar, D. G. Universal solvation model based on solute electron density and on a continuum model of the solvent deﬁned by the bulk dielectric constant and atomic surface tensions. J. Phys. Chem. B 113, 6378–6396 (2009). 38. Hung, W. S. et al. Cross-linking with diamine monomers to prepare composite graphene oxide-framework membranes with varying d-spacing. Chem. Mater. 26, 2983–2990 (2014). Chem. B 113, 6378–6396 (2009). 65. Frisch, M. et al. Gaussian 09, Revision D. Competing interests: The authors declare no competing ﬁnancial interests. 49. Acik, M. et al. Unusual infrared-absorption mechanism in thermally reduced graphene oxide. Nat. Mater. 9, 840–845 (2010). Reprints and permission information is available online at http://npg.nature.com/ reprintsandpermissions/ Reprints and permission information is available online at http://npg.nature.com/ reprintsandpermissions/ 50. Zhong, Y. et al. Elimination of photon quenching by a transition layer to fabricate a quenching-shield sandwich structure for 800 nm excited upconversion luminescence of Nd3+-sensitized nanoparticles. Adv. Mater. 26, 2831–2837 (2014). Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. 51. Zong, L., Li, M. & Li, C. Bioinspired coupling of inorganic layered nanomaterials with marine polysaccharides for efﬁcient aqueous exfoliation and smart actuating hybrids. Adv. Mater. 29, 1604691 (2017). 52. Lv, Ja et al. Photocontrol of ﬂuid slugs in liquid crystal polymer microactuators. Nature 537, 179–184 (2016). Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, 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. The images or other third party material in this article are included in the article’s Creative Commonslicense, unless indicated otherwise in a credit line to the material. If material is not included in the article’sCreative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. 53. Chu, L. Y., Niitsuma, T. & Yamaguchi, T. & Nakao, S. I. Thermoresponsive transport through porous membranes with grafted PNIPAM gates. AlChE J. 49, 896–909 (2003). 54. Li, Y. et al. Thermoresponsive gating characteristics of poly (N- isopropylacrylamide)-grafted porous poly (vinylidene ﬂuoride) membranes. Ind. Eng. Chem. Res. 43, 2643–2649 (2004). 55. Li, W. et al. Low toxic, thermoresponsive dendrimers based on oligoethylene glycols with sharp and fully reversible phase transitions. Chem. Commun. 0, 5948–5950 (2008). 56. Mylvaganam, K. & Zhang, L. In situ polymerization on graphene surfaces. J. Phys. Chem. C. 117, 2817–2823 (2013). 57. Tansel, B. Signiﬁcance of thermodynamic and physical characteristics on permeation of ions during membrane separation: Hydrated radius, hydration free energy and viscous effects. Author contributions 44. Kim, Y. S. et al. Thermoresponsive actuation enabled by permittivity switching in an electrostatically anisotropic hydrogel. Nat. Mater. 14, 1002–1007 (2015). Y.Z. conceived the study and designed the experiments. J.C.L. performed most of the experiments. L.J.Y and A.K. performed the density functional theory calculations. W.L., W.X.Z., and Q.F.C. provided some valuable scientiﬁc suggestions about the temperature- responsive property of the P-GOMs. F.Y.G., and L.L.H. performed the tensile experi- ments and analyzed the data. J.C.L. and N.W. wrote the text. Y.Z., N.W., L.J., and D.A.W revised the paper. 45. Pan, Y., Bao, H., Sahoo, N. G., Wu, T. & Li, L. Water-soluble poly (N- isopropylacrylamide) -graphene sheets synthesized via click chemistry for drug delivery. Adv. Funct. Mater. 21, 2754–2763 (2011). 46. Bergman, T. L., Incropera, F. P., DeWitt, D. P. & Lavine, A. S. Fundamentals of heat and mass transfer (Hobeken, NJ, USA: Wiley, 2011). 47. Wei, N., Peng, X. & Xu, Z. Breakdown of fast water transport in graphene oxides. Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 89, 012113 (2014). References Recent developments in graphene-based membranes: structure, mass-transport mechanism and potential applications. Adv. Mater. 28, 2287 (2016). 16. Zhao, J. et al. Fabricating graphene oxide-based ultrathin hybrid membrane for pervaporation dehydration via layer-by-layer self-assembly driven by multiple interactions. J. Membr. Sci. 487, 162–172 (2015). 17. Koenig, S. P., Wang, L., Pellegrino, J. & Bunch, J. S. Selective molecular sieving through porous graphene. Nat. Nanotechnol. 7, 728–732 (2012). 18. Kim, H. W. et al. Selective gas transport through few-layered graphene and graphene oxide membranes. Science 342, 91–95 (2013). 19. Joshi, R. et al. Precise and ultrafast molecular sieving through graphene oxide membranes. Science 343, 752–754 (2014). 20. Huang, L., Li, Y., Zhou, Q., Yuan, W. & Shi, G. Graphene oxide membranes with tunable semipermeability in organic solvents. Adv. Mater. 27, 3797–3802 (2015). 21. Huang, L. et al. Reduced graphene oxide membranes for ultrafast organic solvent nanoﬁltration. Adv. Mater. 28, 8669–8674 (2016). 22. Gao, S. J., Qin, H., Liu, P. & Jin, J. SWCNT-intercalated GO ultrathin ﬁlms for ultrafast separation of molecules. J. Mater. Chem. A 3, 6649–6654 (2015). 23. Liu, N. et al. Ultralight free-standing reduced graphene oxide membranes for oil-in-water emulsion separation. J. Mater. Chem. A 3, 20113–20117 (2015). 24. Mi, B. Graphene oxide membranes for ionic and molecular sieving. Science 343, 740–742 (2014). 25. Abraham, J. et al. Tunable sieving of ions using graphene oxide membranes. Nat. Nanotechnol. 12, 546–550 (2017). 26. Han, Y., Xu, Z. & Gao, C. Ultrathin graphene nanoﬁltration membrane for water puriﬁcation. Adv. Funct. Mater. 23, 3693–3700 (2013). 27. Long, Y. et al. Molecule channels directed by cation-decorated graphene oxide nanosheets and their application as membrane reactors. Adv. Mater. 29, 1606093 (2017). 28. Fathizadeh, M., Xu, W. L., Zhou, F., Yoon, Y. & Yu, M. Graphene oxide: a novel 2-dimensional material in membrane separation for water puriﬁcation. Adv. Mater. Interfaces 4, 1600918 (2017). Data availability. All data generated or analyzed during this study are present in the main text and the Supplementary Information. Additional data are available from the authors on reasonable request. 29. Huang, H. et al. Ultrafast viscous water ﬂow through nanostrand-channelled graphene oxide membranes. Nat. Commun. 4, 2979 (2013). Received: 27 June 2017 Accepted: 13 November 2017 30. Abozar, A. et al. Large-area graphene-based nanoﬁltration membranes by shear alignment of discotic nematic liquid crystals of graphene oxide. Nat. Commun. 7, 10891 (2016). NATURE COMMUNICATIONS\| 8: 2011 \| DOI: 10.1038/s41467-017-02198-5\| www.nature.com/naturecommunications 8 © The Author(s) 2017 Additional information Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- 017-02198-5. Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- 017-02198-5. 48. Nair, R., Wu, H., Jayaram, P., Grigorieva, I. & Geim, A. Unimpeded permeation of water through helium-leak–tight graphene-based membranes. Science 335, 442–444 (2012). Competing interests: The authors declare no competing ﬁnancial interests. Competing interests: The authors declare no competing ﬁnancial interests. Competing interests: The authors declare no competing ﬁnancial interests. Sep. Purif. Technol. 86, 119–126 (2012). © The Author(s) 2017 NATURE COMMUNICATIONS\| 8: 2011 \| DOI: 10.1038/s41467-017-02198-5\| www.nature.com/naturecommunications 9
https://openalex.org/W4316810851	https://bmcoralhealth.biomedcentral.com/counter/pdf/10.1186/s12903-023-02725-5	English	null	Structural validity and reliability of the “Oral Health Assessment Tool” applied by speech-language therapists in a population of older Chilean people	BMC oral health	2,023	cc-by	8,262	© The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, 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 licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativeco mmons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Abstract Background A good state of oral health allows people to communicate and eat. This topic is relevant in older people given its close relationship with their general health. At present, health challenges are directed at detecting and pre- venting oral disorders and are seen to exclusively by dentists, because the existing instruments can only be applied by them. However, speech-language therapists undergo similar training, which would allow them to collaborate in these processes. In this context, the Oral Health Assessment Tool (OHAT) is a detection instrument with good psychometric properties that is currently available for non-dental use. The objective of this study is the translation into Chilean Span- ish of the OHAT and a structural validation of that version for application by these professionals. Materials and methods A mixed qualitative-quantitative study was carried out. The OHAT instrument was adapted to Chilean Spanish and subsequently subject to structural validity and evaluation of internal consistency reliability, as well as a valuation of its reproducibility in 286 older people (166 female, 120 male) from different health contexts. Results The cultural adaptation of the instrument proved to be semantically consistent with the original instrument. Its application was considered to be speedy and simple in the pre-test. The confirmatory factor analysis evidenced the unidimensionality of the OHAT. In addition, the instrument shows good internal consistency and test–retest reliability. Conclusions The OHAT instrument was considered to possess adequate structural validity and test–retest reliability properties. Its usefulness in the context of oral health disorders of this population in Chile is discussed. unidimensionality of the OHAT. In addition, the instrument shows good internal consistency and test–retest reliability. Conclusions The OHAT instrument was considered to possess adequate structural validity and test–retest reliability properties. Its usefulness in the context of oral health disorders of this population in Chile is discussed. Keywords Older people, Oral health, Validity, Reliability, Validation, Speech-language therapist, Chile Morales et al. BMC Oral Health (2023) 23:24 https://doi.org/10.1186/s12903-023-02725-5 Morales et al. BMC Oral Health (2023) 23:24 https://doi.org/10.1186/s12903-023-02725-5 BMC Oral Health Open Access Camilo Morales1, Felipe Henríquez1 and Sergio Muñoz2 Camilo Morales1, Felipe Henríquez1 and Sergio Muñoz2 Background Oral health is defined as the ability to speak, smile, smell, taste, touch, chew and swallow, as well as to transmit emotions through facial expressions with confidence, without pain, discomfort and/or craniofacial disorders [1]. It allows people to communicate and feed themselves effectively [2]. This construct is particularly important given its implication and close relationship with over- all health [3]; for this reason, poor oral health (usually expressed by the presence of caries, periodontal diseases, Correspondence: Camilo Morales camilo.morales@uct.cl 1 Departamento de Procesos Terapéuticos, Facultad de Ciencias de La Salud, Universidad Católica de Temuco, Manuel Montt No. 056, Campus San Francisco, edificio B, Temuco, Chile 2 Departamento de Salud Pública, Facultad de Medicina‑CIGES, Universidad de Frontera, Temuco, Chile Correspondence: Camilo Morales camilo.morales@uct.cl 1 Departamento de Procesos Terapéuticos, Facultad de Ciencias de La Salud, Universidad Católica de Temuco, Manuel Montt No. 056, Campus San Francisco, edificio B, Temuco, Chile 2 Departamento de Salud Pública, Facultad de Medicina‑CIGES, Universidad de Frontera, Temuco, Chile © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, 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 licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativeco mmons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Open Access T Additionally, pathologies such as diabe- tes and respiratory ailments can be related to poor oral health conditions [11]; Chalmers [10], reported that state of dental health, loss of teeth and temporomandibular disorders are associated with auditory impairment. These various relationships acquire greater relevance because older people seem prone to present oral health problems [12]. The data indicate that the elderly population tends to have poor oral health largely due to dental care defi- ciencies during their entire lifetime. Elderly people with some degree of dependence or limited autonomy tend to present worse oral conditions [13]. Therefore, the risk of developing these problems in older people with attention needs is high, particularly for those with severe depend- ence problems living in nursing homes or who are hos- pitalized. This risk is also related to social patterns in the older population, such as income level, knowledge regarding oral health care or access to health facilities [14], and therefore its prevalence varies depending on these variables. Although several studies have tried to look into and establish oral health intervention programs for older people, this is still described as insufficient or jeopardized [3, 14]. Although the detection of alterations or their oversight In this sense, at an international level there are tools available to evaluate and detect oral health problems, such as the Oral Health Assessment developed by the World Health Organization (WHO) [19] in its version for adults; the Geriatric Oral Health Assessment Index (GOHAI) [20] specifically targeted at older people, and the Oral Health Impact Profile (OHIP) [5]. However, these last two measure the perception of individuals regarding their own oral health [21], with their respective limitations. Conversely, reports in the literature point to the Oral Health Assessment Tool (OHAT) as an instrument that measures oral health intended as an interdisciplinary valuation of this condition. That is to say, its application by other professionals, specifically nurses and speech- language therapists has been signaled as feasible [17, 18], making it possible for this tool not to be linked solely to dentists. The OHAT consists of eight categories aimed at identifying oral health impairments as well as pinpoint- ing the need for prevention actions or referrals for dental intervention, making it a useful instrument for the detec- tion of possible disorders and their early management in elderly adults, whether or not dependent. Open Access T Morales et al. BMC Oral Health (2023) 23:24 Page 2 of 12 oral pain or cancer) affects the self-perception of a per- son both in terms of self-esteem and self-confidence [4].h resources, even in high-income countries [4, 15]. There- fore, alternative and less resource-demanding approaches are needed. In this context, it is important to mention the existence of self-report questionnaires, associated mainly with the oral health dimension in relation to quality of life. However, because it is based on the ability of a per- son to report any adverse dental symptom, it increases the risk of bias, especially in people with some kind of cognitive impairment [16]. Moreover, the majority of clinical instruments or oral health indices are designed to be used by dentists and dental hygienists, but they are not suitable for use by non-dental professionals [17, 18], even though, given their disciplinary similarity, speech- language therapists would be suitable for this purpose. Therefore, the availability of valid and reliable instru- ments enabling the evaluation of oral health through the observation of structures by trained professionals would be especially relevant. i The progressive understanding of the consequences associated with this construct has given rise to the imple- mentation of oral health promotion plans and programs at a local and international level. At the end of the twen- tieth century, modest reductions in the prevalence of dental caries were achieved in children [5, 6]; however, focus on older people is still incipient. Considering cur- rent demographic changes at a global level, promotion and prevention efforts vis-à-vis this latter group has become increasingly relevant. As people reach older age, their needs, including oral health issues, require contin- ual attention. The aging process and associated changes affecting the population pose major challenges to main- tain an optimal state of health throughout the lives of individuals and populations [5, 7]. In this context, sev- eral research efforts have sounded the alarm regarding the risks that poor oral health and mouth diseases have on general health, particularly in older people [3, 8, 9]. The literature describes a link between oral health and systemic diseases. For example, an association between the number of missing teeth with heart disease has been reported; periodontal disorders have been related to car- diovascular disease, atherosclerosis, subclinical lower artery disease, strokes, metabolic and lipid disorders and obesity [3, 10]. Recruitment of participantsh The participants were people aged 60 years or more, pro- ceeding from residences or groups for the elderly, health care groups for prostrated and hospitalized patients from different institutions in the La Araucanía region of Chile, during years 2019 and 2020. Excluded from the study were older people presenting difficulties regarding responsiveness or the ability to follow simple instructions in the context of the application of the instrument, given a situation of severe dependence, dementia or impaired level of consciousness, which were determined via clini- cal records, the application of the Barthel Index and the Pfeiffer Short Portable Mental State Questionnaire (SPMSQ), respectively. Instruments (a) For the sociodemographic characterization and subsequent descriptive statistical analysis, a general information record of the participants was drawn up, which included gender, age, diagnosis of pathol- ogies such as dementia, diabetes or hypertension, level of dependence and type of nutrition, among others. Materials and methods Study designh This study has a mixed qualitative-quantitative design, using a methodological approach to validating a meas- urement instrument. Qualitative, because it aims to establish the cultural validation of the OHAT question- naire, contemplating phases of translation and evaluation of the coverage and façade of the scale. The quantitative approach is analytical and relational. Sample sizeh The estimation of the sample size required to achieve structural validity was conducted following the criteria proposed by Streiner et al. [29], consisting of 10 individu- als participating per item of the measurement instru- ment, with a minimum of 200 persons when the number of items is small. Consequently, a non-probabilistic con- venience sampling was implemented, with a total of 286 participants recruited. The test–retest reliability estimation was carried out as proposed by Donner et al. [30], with an Intraclass Cor- relation Coefficient (ICC) of 0.6 as the acceptable mini- mum and 0.8 expected, a significance level of α = 0.05 and a power (1 − β) of 80%, for two measurements with a dropout rate of 10%. In this case, the calculated mini- mum sample was 49 persons. In this study, 76 persons were finally included for the purpose of this analysis. Open Access T On the basis of the original (Australian) instrument, validation studies of similar instruments have been conducted in Germany, Japan, Brazil, Indonesia, the Netherlands and Turkey [22–27]. At present, the OHAT has not been translated or validated in the Spanish language in any country, including Chile. Given the relevance and usefulness that this instrument represents for health policies and clinical approaches regarding oral health of the older adult population in different contexts, the objective of this study is to deter- mine the structural validity and reliability of the OHAT instrument in the Chilean older adult population. We Although the detection of alterations or their oversight are carried out through clinical examinations performed by dentists, these methods are increasingly more dif- ficult to use due to the high cost and scarcity of human Morales et al. BMC Oral Health (2023) 23:24 Morales et al. BMC Oral Health (2023) 23:24 Page 3 of 12 dependent, 45 to 55 points moderately dependent and 60 or more points mildly dependent [28]. dependent, 45 to 55 points moderately dependent and 60 or more points mildly dependent [28]. hypothesize the OHAT correctly adjust to a one-factor model, consistent with the “Disease and condition status” sub-dimension of the oral health framework proposed by the World Dental Federation (FDI), defined as the thresh- old of severity or level of progression of a possible oral health pathology, which also includes pain or discomfort [1]. Moreover, we hypothesize that the instrument has adequate internal consistency and test–retest reliability. (c) (c) Oral Health Assessment Tool (OHAT) measurement instrument, which on this occasion was adminis- tered by a speech-language therapist. It is made up of eight items (lips, tongue, gums and tissues, saliva, natural teeth, dentures, oral cleanliness and dental pain), the answers to which are organized according to a Likert scale from 0 to 2 points, where 0 indi- cates absence of oral health issues (healthy) and 2 suggests possible disease (unhealthy). The score of each item reflects a description of the observed structures (healthy, signs of possible disease and unhealthy). For its application, a professional with competence in the discipline or formal training is required. Flashlight, gloves and mask should be used in the case of hospitalized patients or people with poor oral hygiene of orofacial structures. The instrument can also be applied to persons with cog- nitive alterations but who are responsive to simple instructions [18]. Procedure for obtaining evidence of reliabilityi After the first application of the OHAT, the participants in the study were asked to answer it on a second occa- sion, within a maximum period of 7 days, to determine the test–retest reliability. Procedure Thi d This study was carried out in two phases. In the first place, the original version of the OHAT was trans- lated into Chilean Spanish and adapted to its culture. In the second phase, its psychometric properties were assessed in a sample of older people. This study received the approval of the scientific ethics committee of the Universidad Católica de Temuco, under resolution No. 40/20. The participation of the subjects was completely voluntary or authorized by a family member or tutor, and they were at liberty to drop out of the study without this involving any detriment to the daily care provided in (b) Barthel Index used to measure the level of depend- ency of the participants with the objective of com- plementing the characterization of the population under study. It assesses a person’s ability to perform 10 activities of daily living, for example: eating, per- sonal hygiene, going up and down stairs, among others. The total score ranges from 0 to 100 points, being classified within the following ranges: < 20 points: fully dependent; 20 to 40 points severely Morales et al. BMC Oral Health (2023) 23:24 Page 4 of 12 (d) Pre-test three speech-language therapists applied the OHAT to a group of 30 persons. After this they were asked for feedback, to allow them to identify difficulties experienced with regard to understand- ing some item(s) of the instrument, or aspects related to the instructions, semantics, grammar or comprehension regarding the type of answer required. their respective facilities. Consequently, each participant or tutor signed an informed consent form, evidencing their free and voluntary participation in accordance with the principles of the Helsinki Decalogue [31]. their respective facilities. Consequently, each participant or tutor signed an informed consent form, evidencing their free and voluntary participation in accordance with the principles of the Helsinki Decalogue [31]. Cultural adaptation First of all, authorization to adapt the instrument was requested from the Iowa Geriatric Education Center by e-mail. The cultural adaptation consisted of the following [32]: Collection of data for structural validity (a) Direct translation undertaken by two independent bilingual translators whose mother tongue is Chil- ean Spanish. The first addressed the study blind and the second was informed of its objective. In addi- tion to translating, they identified comprehension and translation problems arising from semantic ele- ments that were difficult to understand or confus- ing. After this, the translators got together to ana- lyze their texts, detect discrepancies between them and produce the consensus version. They were also asked to maintain the conceptual equivalence of terms rather than a literal translation, when neces- sary. To carry out this procedure, the participants were first required to answer a brief questionnaire in order to col- lect information regarding their sociodemographic background. After this, the Short Portable Mental Ques- tionnaire and the Barthel index were used to comple- ment the general data. Once the base characteristics and eligibility of the subjects had been checked, the OHAT instrument was applied. Given the diversity of contexts, its application took place in a speech-language attention booth in the respective physical space of the participating centers or in the residences themselves, safeguarding the lighting conditions and absence of distractors, and the delivery of clear (protocolized) instructions by the evalu- ator, who was trained for this purpose. (b) Reverse translation translation of the consensus version produced by the bilingual translators back to the original language. This was carried out by a bilingual speech-language therapist whose mother tongue is English. This person rated each of the translated items in terms of (1) Semantic/concep- tual equivalence (maintaining most of the linguis- tic-semantic terms as expressed in the original translation); (2) Functional equivalence (grammati- cal modification of the original idea, maintaining the conceptual equivalence) and (3) Non-evident equivalence (major departure from the concept). Whenever any translated word or phrase fell into the third category, an alternative tending toward equivalence or a justification for the change was reached by consensus by the experts.i Cultural adaptation of OHATfi In general terms, no difficulties were encountered in the translation of the original instrument. Some conceptual terms were modified by consensus, among which were the concepts ‘patchy’, ‘swollen’, ‘rope-like’, and the concept ‘changes’ (second category of the instrument), which was replaced by “signs of possible disease”.h The reverse translation report contemplated a com- parison with the original instrument subsequent to the process, together with a valuation of the linguistic equivalence of the translation produced. In this context, a majority of the elements of the OHAT were categorized within a semantic-conceptual equivalence framework, except for the items tongue, gums and tissues and den- tures, which were categorized as functional equivalences given the previously reported modifications. With regard to nutrition, 77.6% (n = 222) of the partici- pants eat normally, 11.2% (n = 32) eat pureed food and 11.2% are subject to some modification in their nutrition as a result of a medical or speech-language indication (only liquids, only solids, chopped food, etc.). None of the participants were absolutely restricted from ingesting orally, with 5.2% (n = 15) currently using a feeding tube to complement their nutrition process and 1% (n = 3) fed via gastrostomy. i In the case of dentures, the content of the text was modified at the time of producing the final version of the instrument, where the experts analyzed whether to include “name on dentures” in all categories of the item, because this is not usual practice in Chile in any context. It was decided to maintain the descriptor, but in conditional form (in contexts where it might be nec- essary). Figures 1 and 2 show the original version of the OHAT and the final consensus version translated into Spanish, respectively. The pre-test of the instru- ment was undertaken by three selected evaluators, speech-language therapists by profession, experienced in attending to older people for this purpose. Each one was interviewed to get to know their opinion about Statistical analysis A i d A sociodemographic characterization of the study popu- lation was carried out, using descriptive statistics, spe- cifically central tendency measures and dispersion for quantitative variables, and absolute and relative frequen- cies for categorical variables. To achieve the objective regarding the structural validity, a confirmatory factor analysis (CFA) [33] was carried out to corroborate the factor structure of the sub-dimension “Disease and condition status”, which is part of oral health as reported by the literature [1]. In this case, given that the instrument variables are ordinal in nature, the diagonally weighted least squares (DWLS) estimator was used, since the data are ordinal and do not meet the assumption of multivariate normality [34]. For model assessment, several model fit indices were cal- culated, including the comparative fit index (CFI), the Tucker-Lewis index (TLI), the root mean square error of approximation (RMSEA), in addition to the standardized root mean square error of approximation (SRMSEA). For CFI and TLI, values greater than or equal to 0.95 were (c) (c) Consolidation and final production of the instru- ment a committee was formed made up of the translators who generated the consensus version, a speech-language therapist trained in geronto-ger- iatrics and a dentist, who were presented the two initial versions provided by the translators, the con- sensus version and the reverse translation submit- ted by the speech-language therapist. Discrepancies regarding the translation of the instrument were discussed in terms of quality of translation, main- tenance of the linguistic or functional equivalence and the modifications associated with contextual pertinence made to arrive at the final instrument. Morales et al. BMC Oral Health (2023) 23:24 Page 5 of 12 considered acceptable [35]. In the case of RMSEA, values less than 0.06 and for SRMSEA, less than 0.08 [36]. the general and conceptual comprehensibility of the test and its applicability within the context of speech and language assessment. After this, they were asked to evaluate a total of 30 persons (10 each). Consulted regarding their experience using OHAT, all the profes- sionals stated that the application of the instrument is not difficult in general, the description of the items is clear and leaves no room for doubt, and reported an estimated average duration of 7 min in its application. To obtain evidence of test–retest reliability, the intra- class correlation coefficient (ICC) was estimated with its respective 95% reliability interval, in order to determine the degree of consistency between measurements. Results The average age of the participants in the study was 75.01 ± 9.4 years. With regard to a background of pathol- ogies or associated conditions, of the total number of participants 68.5% have arterial hypertension (n = 196), 27.3% type 2 diabetes (n = 78), 17.8% stroke (n = 51), 4.9% concussion (n = 14), 5.2% some type of cancer (n = 15, more specifically prostate and stomach), 12.6% hypo or hyperthyroidism (n = 36), 24.5% hearing loss (n = 70), 11.5% present dementia (n = 33) and 5.9% have Parkin- son’s disease (n = 17). Descriptive analysis Th l i i i l The total initial population consisted of 293 partici- pants. Of these, seven persons were excluded from the study—three for not signing the informed consent form and four because they presented dementia as a base condition, which did not allow them to compre- hend or follow simple instructions in the application of the instrument—leaving 286 persons as the total end sample. All the participants fully completed the evalu- ation. No missing item responses reported. The charac- teristics of the population participating in the study are described in Table 1.h fi The internal consistency reliability was obtained through ordinal Cronbach’s alpha and McDonald’s omega [38]. Values higher than 0.7 were considered acceptable for the factor hypothesized. The above-described processing and subsequent sta- tistical analysis were carried out using Rstudio software, version 1.4.1743–4. Statistical analysis A i d In this context, a two-way mixed-effects model was used to achieve measures of “absolute agreement”. Although no consensus was reached regarding the interpretation of this coefficient, some guidelines were established. For the purposes of this study, values higher than 0.75 were con- sidered acceptable reliability values. Values between less than 0.75 and 0.5 were considered moderately reliable and those below 0.5 insufficiently reliable [37].h Reliability with the original instrument, safeguarding that any modification made of terms that were difficult to understand or not appropriate for implementation in the Chilean context in general, and more specifically in older adults, was as close as possible to the origi- nal meaning. Following this precept, the only concept that differed from the original text was the presence (or absence) of dentures marked with the name of the users, a usual practice in geriatric residences of other countries but currently not applicable to the Chilean reality. Thus, the general valuation of the instrument was not limited solely to the situation of institutional- ized older people but to any evaluation context. Internal consistency was carried out calculating the ordi- nal Cronbach’s alpha (α) coefficient, which resulted in a value of 0.82 for the evaluated dimension. In the same way, the McDonald’s Omega (ω) was 0.87. Regarding the test–retest reliability A sample of 76 persons was contemplated for the analysis, and was sub- ject to a second valuation that took place 7 days after the first application. The ICC thus obtained was 0.81 (IC95% 0.74–0.87; p < 0.001). Study of psychometric properties Validity Figure 3 shows results of the confirmatory factor analy- sis (CFA) in which the eight items of OHAT loaded on a single latent variable. This model provided a satisfactory adjustment as we hypothesized: DWLS X2 = 43.88 (20 degrees of freedom), p = 0.002; CFI = 0.98; TLI = 0.97; RMSEA = 0.05 and SRMSEA = 0.07. From these results it is possible to point out that the model fits the data well. Page 6 of 12 Morales et al. BMC Oral Health (2023) 23:24 Fig. 1 Original version of the OHAT instrument Page 7 of 12 Morales et al. BMC Oral Health (2023) 23:24 Final version of the OHAT instrument translated and adapted to Chilean Spanish Morales et al. BMC Oral Health (2023) 23:24 Page 8 of 12 g. 2 continued Discussionh The translated and adapted Chilean version of the Oral Health Assessment Tool is presented initially as a tool that is easy to understand and apply. In general terms, the main goal was to achieve conceptual equivalence In terms of the validity of the instrument and as we hypothesized, the CFA of the OHAT confirmed the Page 9 of 12 Morales et al. BMC Oral Health (2023) 23:24 Table 1 Sociodemographic characterization of the study population Characteristic n (%) Gender Female 166 (58) Male 120 (42) Provenance Urban 224 (78.3) Rural 62 (21.7) Type of residence Hospitalized 36 (12.6) Geriatric residence (State-funded/Private) 59 (20.6) Residential care for prostrated patients 74 (25.9) Older persons group 43 (15.0) Day center 61 (21.3) Home visit 13 (4.5) Dependence level Independent 50 (17.5) Mildly dependent 134 (46.9) Moderately dependent 15 (5.2) Severely dependent 27 (9.4) Fully dependent 60 (21.0) Cognitive performance No deterioration 111 (38.8) Mild/moderate deterioration 110 (38.5) Severe deterioration 65 (22.7) one-dimension solution. This is the first study that considered the structural validity of the instrument through confirmatory factor analysis. The factor load- ings of the instrument were all greater or equal to 0.4 demonstrating adequate correlation between item and latent variable [39], as well as the previously reported model fit indices. In other validation research such as the Dutch study [23], the number of dimensions extracted did not concur with the dimension hypoth- esized in the model of the World Dental Federation (FDI) [1], but it is important to stress that the latent variables or constructs to be evaluated tend to differ from culture to culture and thus some variation is to be expected considering the modifications and contexts to which the instrument has been subject [29]. In addi- tion, it is important to mention that these studies used the exploratory factor analysis procedure and that the sample size used was smaller than that recommended by the literature [29, 40]. As mentioned above, the OHAT measures a sub- dimension of oral health (disease and condition status). However, the concept of oral health is complex and mul- tidimensional and not only involves identifying affected structures or functions but also the perceptions of the individuals themselves regarding their health status, or the context that favors or obstructs maintaining their health [1]. In this sense the OHAT is an instrument Fig. Discussionh 3 Confirmatory factor analysis on the 8-item OHAT Confirmatory factor analysis on the 8-item OHAT Fig. 3 Confirmatory factor analysis on the 8-item OHAT Morales et al. BMC Oral Health (2023) 23:24 Page 10 of 12 Page 10 of 12 to compare performance in terms of detecting the OHAT using a previously established criterion (such as, for example, the considered opinion or valuation of a profes- sional dentist), its concurrent and discriminant validity, or likewise its predictive validity. From a methodological viewpoint, a prospective (cohort) study would be useful to test these hypotheses. Notwithstanding the above, pre- vious systematic reviews have reported that for general geriatric assessment, screening or triage carried out by non-dentist professionals, the OHAT is suggested [48]. In this sense, one of the strengths of this research is that it contributes with antecedents regarding the construct to be assessed at the time of evaluation, adding to the exist- ing evidence of content and criterion validity [48]. whose clinical usefulness lies in the observation of struc- tures or consequences in oral functioning that increase risk or detect possible alterations that can lead to disease as a more objective measure than just self-perception or self-report. Consequently, the results obtained in the pre- sent research are sufficient to consider that the OHAT has structural validity. y In terms of internal consistency reliability, the reported ordinal Cronbach’s alpha and Mcdonald’s Omega coef- ficient were good (α = 0.82; ω = 0.87) for the entire one- dimension instrument, similar results than obtained by Kuwamura et al. [41] but differing from the values obtained by Mello et al. [26], who evaluated the inter- nal consistency of the instrument applied by a group of nurses and dentists to 50 older people, obtaining alpha values considered to be low. In this respect, the following should be considered: first of all, it is important to note the definition of this measure, which is frequently consid- ered to refer to the degree in which all elements of a test or instrument measure the same attribute or dimension [42]. In this sense, according to Bonett [43] and Charter [44], sample sizes of more than 250 subjects are required to obtain appropriate and precise values with this coef- ficient, both for determining the coefficient of the instru- ment and to validate any comparisons. Discussionh In this study, the calculation of the coefficient was carried out in compli- ance with this condition. Although there is evidence available regarding the clin- ical usefulness of the OHAT in other contexts, especially in terms of its application by speech-language therapists and nurses, no certainty exists regarding the feasibility of its use by the latter professionals given that the training and application of this study was carried out solely by the former. It would be a mistake to assume that the forma- tive processes or conceptual contents of nurses are iden- tical to their similar counterparts in other countries or contexts. It is therefore considered relevant to undertake additional research efforts that will evaluate the feasibil- ity of the use of the OHAT by these evaluators. Similarly, given variable conditions for assessment, it is necessary to consider future validation studies con- templating more assessment contexts. Currently, the resources of tele-assessment or tele-care associated with the virtualization of oral health care are of interest and growing acceptability given the evidence of its benefits, especially in terms of promotion and prevention related to access to dental or oral health professionals (shorten- ing distances, facilitating access to care) [49, 50], so, hav- ing an instrument such as the OHAT adapted for remote application could be an interesting projection regarding these new trends. In the same vein, another objective of the study was to determine the test–retest reliability or reproducibility of the instrument. This provides an indicator of the stability of the test measurements as a function of a specific inter- val between two evaluations taken in different time peri- ods [45], which is also of clinical interest. In this study the intraclass correlation coefficient value was 0.81 (IC95% 0.74 – 0.87), considered to be good. Similar values were obtained in the original validation [17] of the OHAT and the English validation directed at evaluators who were speech-language therapists by profession [18]. With regard to the limitations of this study, the aspect of representativeness is mentioned. Although the calcula- tion of the sample size is based on the number of meas- ures per item, it is also necessary to safeguard that, from an epidemiological point of view, there are sufficient con- texts available to encompass greater variability. Discussionh In this sense, access to the population of older people receiving home care (n = 13) was lower, and therefore it is likely that such variability was not fully covered. Received: 21 July 2022 Accepted: 9 January 2023 Received: 21 July 2022 Accepted: 9 January 2023 Funding This work was partially supported by Universidad Católica de Temuco, Chile. 16. Feasley JC. Health related quality of life. In: health outcomes for older people: questions for the coming decade. Washington D.C: National Academy of Sciences; 1996. pp. 17–9. 16. Feasley JC. Health related quality of life. In: health outcomes for older people: questions for the coming decade. Washington D.C: National Academy of Sciences; 1996. pp. 17–9. Availability of data and materials The datasets generated during and analyzed during the current study are not publicly available due to data protection regulations and ethical concerns but are available from the corresponding author on reasonable request. 17. Chalmers JM, King PL, Spencer AJ, Wright FAC, Carter KD. The oral health assessment tool–validity and reliability. Aust Dent J. 2005;50(3):191–9.f 18. Simpelaere IS, Van Nuffelen G, Vanderwegen J, Wouters K, De Bodt M. Oral health screening: feasibility and reliability of the oral health assess- ment tool as used by speech pathologists. Int Dent J. 2016;66(3):178–89. Acknowledgements 11. Amar S, Han X. The impact of periodontal infection on systemic diseases. Med Sci Monit. 2003;9(12):291–9. g we thank all the participants who agreed to take part in this study and the Iowa Geriatric Education Center for kindly authorizing the use of the instru- ment for the purposes of this research. 12. Mankekar G. Swallowing–physiology, disorders, diagnosis and therapy. 1st ed. New Delhi: Springer; 2015. p. 189–229. 1st ed. New Delhi: Springer; 2015. p. 189–229. 13. Klotz AL, Hassel AJ, Schröder J, Rammelsberg P, Zenthöfer A. Oral health- related quality of life and prosthetic status of nursing home residents with or without dementia. Clin Interv Aging. 2017;12:659–65. Author contributions Conceptualization, CM, and SM; patient recruitment and sampling, CM and FH; CM and SM; formal analysis; writing—original draft preparation, CM, FH and SM; writing—review and editing CM, FH and SM; supervision, CM; project administration, CM and FH; funding acquisition, CM and FH All authors have read and agreed to the published version of the manuscript. 14. Petersen PE, Kjøller M, Christensen LB, Krustrup U. Changing dentate sta- tus of adults, use of dental health services, and achievement of National Dental Health goals in Denmark by the year 2000. J Public Health Dent. 2004;64:127–35. 15. Levine R, Stillman-Lowe C. The scientific basis of oral health education. 8th ed. Community Dental Health. West Yorkshire: Springer Nature Swit- zerland; 2019. pp. 1–96. Declarations 19. World Health Organization. Oral health surveys-basic methods. 5th ed. Who Library Cataloguing-in-Publication Data, editor. Numerical methods and optimization in Finance. Paris: World Health Organization; 2011. pp. 1–125 Competing interests Competing interests The authors declare that they have no competing interests. Conclusions In light of the foregoing, it is considered that the Oral Health Assessment Tool has structural validity and possesses adequate properties of internal consistency and test–retest reliability for the population under study. Its orientation is initially clinical, aimed at favor- ing the detection of possible oral health problems in Chilean older people and referring them in a timely fashion to a professional dentist for their optimal care. The information provided could be useful also in possi- ble actions of health promotion and disease prevention, not only directed at older adults but also their caretak- ers and treatment team, and could have an impact on indicators of oral health and particularly in the quality Another limitation has to do with the process itself. International guidelines on the validation of instruments contemplate other types of validation that are useful when considering whether or not an instrument is valid (and also reliable) [46, 47]. In this sense, it would be use- ful to develop other studies that would make it possible Page 11 of 12 Morales et al. BMC Oral Health (2023) 23:24 Morales et al. BMC Oral Health (2023) 23:24 Abbreviations 6. Ministerio de Salud. Gobierno de Chile. Plan Nacional de Salud Bucal 2018–2030. Chile; 2017. Available from: https://diprece.minsal.cl/wrdpr ss_minsal/wp-content/uploads/2018/03/PLAN-NACIONAL-DE-SALUD- BUCAL-2018-2030.pdf 7. Zarebski GCME. Cuestionario Mi Envejecer: un instrumento psicogeron- tológico para evaluar la actitud frente al propio envejecimiento. 1a ed. Buenos Aires: Paidós; 2014. 8. Ghezzi EM, Ship JA. Systemic diseases and their treatments in the elderly impact on oral health. J Public Health Dent. 2000;60(4):289–96. 9. Strömberg E, Hagman-Gustafsson ML, Holmén A, Wårdh I, Gabre P. Oral status, oral hygiene habits and caries risk factors in home-dwelling elderly dependent on moderate or substantial supportive care for daily living. Commun Dent Oral Epidemiol. 2012;40(3):221–9. 9. Strömberg E, Hagman-Gustafsson ML, Holmén A, Wårdh I, Gabre P. Oral status, oral hygiene habits and caries risk factors in home-dwelling elderly dependent on moderate or substantial supportive care for daily living. Commun Dent Oral Epidemiol. 2012;40(3):221–9. 10. Chalmers JM. Oral health promotion for our ageing Australian popula- tion. Aust Dent J. 2003;48(1):2–9. Competing interests The authors declare that they have no competing interests. of life of older people. In turn, it is considered as an alternative or complement to the valuation of speech and language structures, and even the valuation of the swallowing process by the speech-language thera- pists, who could use the instrument in future lines of research to determine its contribution to the diagnostic or decision-making process regarding the treatment of swallowing disorders in the elderly. 20. Sánchez-García S, Heredia-Ponce E, Juárez-Cedillo T, Gallegos-Carrillo K, Espinel-Bermúdez C, De La Fuente-Hernández J, et al. Psychometric properties of the General Oral Health Assessment Index (GOHAI) and Ethics approval and consent to participatei This study received the approval of the scientific ethics committee of the Universidad Católica de Temuco, under resolution No. 40/20. 20. Sánchez-García S, Heredia-Ponce E, Juárez-Cedillo T, Gallegos-Carrillo K, Espinel-Bermúdez C, De La Fuente-Hernández J, et al. Psychometric properties of the General Oral Health Assessment Index (GOHAI) and 20. Sánchez-García S, Heredia-Ponce E, Juárez-Cedillo T, Gallegos-Carrillo K, Espinel-Bermúdez C, De La Fuente-Hernández J, et al. Psychometric properties of the General Oral Health Assessment Index (GOHAI) and References 1. Glick M, Williams DM, Kleinman DV, Vujicic M, Watt RG, Weyant RJ. A new definition for oral health developed by the FDI World Dental Federa- tion opens the door to a universal definition of oral health. Int Dent J. 2016;66(6):322–4. 1. Glick M, Williams DM, Kleinman DV, Vujicic M, Watt RG, Weyant RJ. A new definition for oral health developed by the FDI World Dental Federa- tion opens the door to a universal definition of oral health. Int Dent J. 2016;66(6):322–4. 1. Glick M, Williams DM, Kleinman DV, Vujicic M, Watt RG, Weyant RJ. A new definition for oral health developed by the FDI World Dental Federa- tion opens the door to a universal definition of oral health. Int Dent J. 2016;66(6):322–4. Finally, it is important to undertake further studies of different types of validity and reliability in order to collect more information about the psychometric prop- erties of the instrument and extend or project its use- fulness into the future. By the same token, it is useful to stress that no instrument is definitive in terms of its psychometric properties, thus making it advisable to review and enrich the instrument over time in order to improve its characteristics. 2. Watt RG. Strategies and approaches in oral disease prevention and health promotion. Bull World Health Organ. 2005;83(9):711–8. 3. Petersen PE, Yamamoto T. Improving the oral health of older people: the approach of the WHO Global Oral Health Programme. Commun Dent Oral Epidemiol. 2005;33(2):81–92. 4. Registered Nurses’ Association of Ontario. Oral health : nursing assess- ment and interventions. Nursing best practice guideline. Toronto, Canada: Registered Nurses’ Association of Ontario; 2008. Available from: www.rnao.org/bestpractices 5. León S, Correa-Beltrán G, De Marchi RJ, Giacaman RA. Ultra-short version of the oral health impact profile in elderly Chileans. Geriatr Gerontol Int. 2017;17(2):277–85. Abbreviations CFI Comparative fit index CFA Confirmatory factor analysis DWLS Diagonally weighted least squares DCS Disease and condition status FDI World dental federation GOHAI Geriatric oral health assessment index ICC Intraclass correlation coefficient OHAT Oral Health Assessment Tool OHIP Oral health impact profile RMSEA Root mean square error of approximation SPMSQ Short portable mental state questionnaire SRMSEA Standardized root mean square error of approximation TLI Tucker-Lewis index WHO World health organization Consent for publication Consent for publication Not applicable. Not applicable. Morales et al. BMC Oral Health (2023) 23:24 Page 12 of 12 Page 12 of 12 45. Koo TK, Li MY. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. J Chiropr Med. 2016;15(2):155–63. https://doi.org/10.1016/j.jcm.2016.02.012. dental status of an elderly Mexican population. J Public Health Dent. 2010;70(4):300–7. dental status of an elderly Mexican population. J Public Health Dent. 2010;70(4):300–7. 21. Campos JADB, Zucoloto ML, Geremias RF, Nogueira SS, Maroco J. Valida- tion of the geriatric oral health assessment index in complete denture wearers. J Oral Rehabil. 2015;42(7):512–20. 46. Mokkink LB, Terwee CB, Patrick DL, Alonso J, Stratford PW, Knol DL, et al. The COSMIN study reached international consensus on taxonomy, ter- minology, and definitions of measurement properties for health-related patient-reported outcomes. J Clin Epidemiol. 2010;63(7):737–45. https:// doi.org/10.1016/j.jclinepi.2010.02.006. 22. Gita F, Wiryasmoro T, Sagala DM, Hendry HE, Kusdhany LS. Oral health status and oral health-related quality of life in Indonesian elderly (analysis using the Indonesian version of the oral health assessment tool and the geriatric oral health assessment index). J Int Dent Med Res. 2017;10(Specialissue):533–9. 47. Mokkink LB, Prinsen CAC, Bouter LM. The COnsensus-based Standards for the selection of health Measurement INstruments ( COSMIN ) and how to select an outcome measurement instrument. 2016;20(2):105–13. p 23. Harderwijk H. Translating and validating the Dutch version of the oral health assessment tool for older. University of Utrecht; 2019. Available from: https://dspace.library.uu.nl/handle/1874/384287 48. Everaars B, Weening-verbree LF, Jerkovi K, Schoonmade L, Bleijenberg N, De WNJ, et al. Measurement properties of oral health assessments for non-dental healthcare professionals in older people : a systematic review. BMC Geriatr. 2020;20(4):1–18. 24. Klotz AL, Zajac M, Ehret J, Hassel AJ, Rammelsberg P, Zenthöfer A. Devel- opment of a German version of the oral health assessment tool. Aging Clin Exp Res. 2019. https://doi.org/10.1007/s40520-019-01158-x. 49. Howell SEI, Fukuoka B. Teledentistry for patient-centered screening and assessment. Dent Clin North Am. 2022;66(2):195–208. 25. Matsuo K, Nakagawa K. Reliability and validity of the Japanese version of the oral health assessment tool (OHAT-J). J Jpn Soc Disabil Oral Heal. 2016;37(1):1–7. 50. Fernández CE, Maturana CA, Coloma SI, Carrasco-Labra A, Giacaman RA. Teledentistry and mHealth for promotion and prevention of oral health: a systematic review and meta-analysis. J Dent Res. 2021;100(9):914–27. 26. de Mello ALSF, Zimermann K, Gonçalves LHT. Avaliação da saúde bucal de idosos por enfermeiros: validade e confiabilidade do instrumentO ASBTO. Publisher’s Note 27. Şahin NE, Jablonski RA. Psychometric properties of the oral health assess- ment tool Turkish version. Cukurova Med J. 2019;44(Suppl 1):513–20. Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. 28. María Trigás-Ferrín, Lucía Ferreira-González, Meijide-Míguez H. Escalas de valoración funcional en el anciano. Galicia Clin [Internet]. 2011;72(1):11–6. Available from: http://galiciaclinica.info/pdf/11/225.pdf 29. Streiner DL, Norman G, Cairney J. Health measurement scales: a practical guide to their development and use. 5th ed. Oxford: Oxford University Press; 2015. 30. Donner A, Eliasziw M. Sample size requirements for reliability studies. Stat Med. 1987;6(4):441–8. 31. Asociación Médica Mundial. Declaración de Helsinki de la AMM–Prin- cipios éticos para las investigaciones Médicas en seres humanos. 2013. Available from: https://www.wma.net/es/policies-post/declaracion-de- helsinki-de-la-amm-principios-eticos-para-las-investigaciones-medic as-en-seres-humanos/ 32. Ramada-rodilla JM, Serra-pujadas C, Delclós-Clanchet GL. Adaptación cul- tural y validación de cuestionarios de salud : revisión y recomendaciones metodológicas. Salud Públ Méx. 2013;55(1):57–66. 33. Norman G, Streiner DL. Biostatistics. The bare essentials. 3rd ed. Hamilton, Ontario: BC Decker; 2008. 34. Li C. Confirmatory factor analysis with ordinal data : Comparing robust maximum likelihood and diagonally weighted least squares. Behav Res. 2015;48:936–49. 34. Li C. Confirmatory factor analysis with ordinal data : Comparing robust maximum likelihood and diagonally weighted least squares. Behav Res. 2015;48:936–49. 35. Hu L, Bentler PM, Hu L. Cutoff criteria for fit indexes in covariance struc- ture analysis: conventional criteria versus new alternatives. Struct Equ Model A Multidiscip J. 1999;6(1):1–55. 35. Hu L, Bentler PM, Hu L. Cutoff criteria for fit indexes in covariance struc- ture analysis: conventional criteria versus new alternatives. Struct Equ Model A Multidiscip J. 1999;6(1):1–55. 36. Boateng GO, Neilands TB, Frongillo EA. Best practices for developing and validating scales for health, social, and Behavioral research: a primer. Front Public Heal. 2018;6(June):1–18. 37. Portney LG, Watkins MP. Foundations of clinical research. Applications to practice. 3rd ed. Philadelphia: F.A Davis Company; 2012. 38. Elosua P, Zumbo BD. Coeficientes de fiabilidad para escalas de respuesta categórica ordenada. Psicothema. 2008;20(4):896–901. 38. Elosua P, Zumbo BD. Coeficientes de fiabilidad para esca 38. Elosua P, Zumbo BD. Coeficientes de fiabilidad para escalas de respuesta categórica ordenada. Psicothema. 2008;20(4):896–901. categórica ordenada. Psicothema. 2008;20(4):896–901. 39. Pituch KA, Steevens JP. Consent for publication Rev Gaúcha Enferm. 2012;33(2):36–44. Publisher’s Note Applied multivariate statistics for the social sci- ences. 6th ed. New York, NY: Routledge; 2016. 39. Pituch KA, Steevens JP. Applied multivariate statistics for the social sci- ences. 6th ed. New York, NY: Routledge; 2016. • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: 40. Terwee CB, Mokkink LB, Knol DL, Ostelo RWJG. Rating the methodologi- cal quality in systematic reviews of studies on measurement properties : a scoring system for the COSMIN checklist. Qual Life Res. 2012;21:651–7. 41. Kuwamura Y, Sumikawa M, Tanioka T, Nagata T, Sakamoto E, Murata H, et al. Development of the diabetes oral health assessment tool; for nurses. Health. 2015;07(12):1710–20. 42. Cronbach LJ, Gleser GC. Interpretation of reliability and validity coef- ficients: remarks on a paper by Lord. J Educ Psychol. 1959;50(5):230–7. 42. Cronbach LJ, Gleser GC. Interpretation of reliability and validity coef- ficients: remarks on a paper by Lord. J Educ Psychol. 1959;50(5):230–7. 43. Bonett DG. Sample size requirements for testing and estimating coef- ficient alpha. J Educ Behav Stat. 2002;27(4):335–40. i p p y y ; ( ) 43. Bonett DG. Sample size requirements for testing and estimating coef- ficient alpha. J Educ Behav Stat. 2002;27(4):335–40. i y y 43. Bonett DG. Sample size requirements for testing and estimating coef- ficient alpha. J Educ Behav Stat. 2002;27(4):335–40. i 44. Charter RA. Study samples are too small to produce sufficiently precise reliability coefficients. J Gen Psychol. 2003;130(2):117–29. 44. Charter RA. • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: Publisher’s Note Study samples are too small to produce sufficiently precise reliability coefficients. J Gen Psychol. 2003;130(2):117–29.
https://openalex.org/W4281629346	https://zenodo.org/record/6581898/files/Ta%27lim%20fidoyilari-22.05.2022%20%284%29-298-300.pdf	English	null	IMPORTANCE OF LANGUAGE IN WRITING FOR NEWSPAPERS	Zenodo (CERN European Organization for Nuclear Research)	2,022	cc-by	1,276	IMPORTANCE OF LANGUAGE IN WRITING FOR NEWSPAPERS This article is about a number specific characteristic of the language which is used for printed media. It is given very important points of used language and given some suggestions for young journalists. Key words: language, journalism, mass media, journalistic language, med y g g , j , , j g g , g g Language is especially important in journalism. Language is involved with the precise presentation of words and phrases in writing. It contains abbreviations, capitalization compounding and other printed language features Language is especially important in journalism. Language is involved with the precise presentation of words and phrases in writing. It contains abbreviations, capitalization, compounding, and other printed language features. The language used in journalism to present the most recent and up-to-date facts to the public in the simplest, easiest, and most unbiased manner is referred to as journalistic language. Journalism is a crucial kind of communication in which we spread news, information, and knowledge to a wide number of people. It is vital to employ straightforward and basic language for this goal. That is why it is also known as people's language and mass communication language. Here is given some important characteristics of journalistic language. Language is especially important in journalism. Language is involved with the precise presentation of words and phrases in writing. It contains abbreviations, capitalization, compounding, and other printed language features. The language used in journalism to present the most recent and up-to-date facts to the public in the simplest, easiest, and most unbiased manner is referred to as journalistic Language is especially important in journalism. Language is involved with the precise presentation of words and phrases in writing. It contains abbreviations, capitalization, compounding, and other printed language features. Th l d i j li h d d f h capitalization, compounding, and other printed language features. The language used in journalism to present the most recent and up-to-date facts to the public in the simplest, easiest, and most unbiased manner is referred to as journalistic language. Journalism is a crucial kind of communication in which we spread news, information, and knowledge to a wide number of people. It is vital to employ straightforward and basic language for this goal. That is why it is also known as people's language and mass communication language. Here is given some important characteristics of journalistic language. newspapers, newspapers cover all themes like as political, social, religious, showbiz, criminal, business, and so on, but other languages just cover a subset of their concerns.  Seriousness-unlike in literary language, facts are expressed in journalistic language in a serious manner.  Seriousness-unlike in literary language, facts are expressed in journalistic language in a serious manner.  Awareness- the primary goal of reporting language is to transmit and impart knowledge and information to a big group of individuals in their current state. As a result, the quality of reporting language raises public awareness and encourages individuals to address their own issues.  Awareness- the primary goal of reporting language is to transmit and impart knowledge and information to a big group of individuals in their current state. As a result, the quality of reporting language raises public awareness and encourages individuals to address their own issues.  Communication Area-because journalistic language is intended for the entire public, it has a limitless communication area. As a result, it is straightforward and easy to grasp for the general public.  Objectivity- to be objective, explain an event or item "as it is," without embellishment or extraneous details. This is also an essential aspect of journalistic language.  Media Language- journalistic language is intended for the general public. Mass communication, on the other hand, refers to communicating with a huge number of individuals. In summary, journalistic language is an essential source of mass communication, which is why it is referred to as the language of media  Media Language- journalistic language is intended for the general public. Mass communication, on the other hand, refers to communicating with a huge number of individuals. In summary, journalistic language is an essential source of mass communication, which is why it is referred to as the language of media  Realism-realism refers to genuineness and actuality. Things are reported in a realistic manner in journalistic terminology.  Journalistic language represents all social classes. It also reflects all sorts of individuals because newspapers are read by people of different socioeconomic backgrounds.  Politeness- in journalism, politeness of language is required in order to catch the attention of the general public. Otherwise, people would not pay attention to complex words.  Politeness- in journalism, politeness of language is required in order to catch the attention of the general public. Otherwise, people would not pay attention to complex words.  Simple and Easy- simple and easy language is required for journalism since the main objective of journalism is to enlighten the general public, literate and illiterate alike, which cannot be accomplished without simple and easy language, which is also a key quality of journalistic language.  Unbiased- in journalistic language, we impart knowledge and send data without any expression or personal sentiments; for example, news on television is delivered without the newscaster's own perspective (who reads the news in news bulletin) Reporters at newspapers do not include their sentiments and feelings when writing a report on an incident.  Brevity- Brevity denotes "shortness." We compose news and reports in journalistic style to be succinct and to the point, with no extraneous information.  Broadness-the broadness of journalistic language distinguishes it from other languages. This indicates that journalistic language may describe all themes and their difficulties, whether they are social, religious, political, or educational. As we can see in [298]  Avoid Difficult and Unfamiliar Words- we do not employ complicated and new words in journalistic language for public comprehension.  Avoid Difficult and Unfamiliar Words- we do not employ complicated and new words in journalistic language for public comprehension.  Avoid Difficult Terms- journalism requires plain and straightforward language. And a journalist takes care of this element by using simple language and avoiding difficult phrases while producing a report since he or she realizes that the report is for everyone, not just one group, so he or she transforms sophisticated and tough terminology into simple ones.  Avoid Difficult Terms- journalism requires plain and straightforward language. And a journalist takes care of this element by using simple language and avoiding difficult phrases while producing a report since he or she realizes that the report is for everyone, not just one group, so he or she transforms sophisticated and tough terminology into simple ones.  Short Phrases- because it is intended for everyone, journalistic language is built on short and simple sentences. To sum up, the language used to present the most recent and up-to-date facts to the public in the simplest, easiest, and most unbiased manner is referred to as journalistic language. Language is involved with the precise presentation of words and phrases in writing. Journalism is a crucial kind of communication in which we spread news, information, and knowledge to a wide number of people.  Grammar-correct grammar usage is essential in journalistic language; otherwise, it will leave a negative impression on readers. Proper and accurate use of tenses such as past, present, and future is also essential. [299] LIST OF REFERENCES: 1. H.D. Zavala González, The importance of journalistic language features and functions in print media, 2010 1. H.D. Zavala González, The importance of journalistic language features and functions in print media, 2010 2. Galperin, I. R. Stylistics, Moscow: Higher School, 1981 3. https://dailyasianage.com 4. https://writingcenter.uagc.edu/journalistic-writing [300]
https://openalex.org/W3122026157	https://opus.lib.uts.edu.au/bitstream/10453/129631/1/Vibration_Stender_et_al_2019.pdf	English	null	Recovery of Differential Equations from Impulse Response Time Series Data for Model Identification and Feature Extraction	null	2,018	cc-by	13,561	Received: 6 December 2018; Accepted: 7 January 2019; Published: 10 January 2019 Abstract: Time recordings of impulse-type oscillation responses are short and highly transient. These characteristics may complicate the usage of classical spectral signal processing techniques for (a) describing the dynamics and (b) deriving discriminative features from the data. However, common model identiﬁcation and validation techniques mostly rely on steady-state recordings, characteristic spectral properties and non-transient behavior. In this work, a recent method, which allows reconstructing differential equations from time series data, is extended for higher degrees of automation. With special focus on short and strongly damped oscillations, an optimization procedure is proposed that ﬁne-tunes the reconstructed dynamical models with respect to model simplicity and error reduction. This framework is analyzed with particular focus on the amount of information available to the reconstruction, noise contamination and nonlinearities contained in the time series input. Using the example of a mechanical oscillator, we illustrate how the optimized reconstruction method can be used to identify a suitable model and how to extract features from uni-variate and multivariate time series recordings in an engineering-compliant environment. Moreover, the determined minimal models allow for identifying the qualitative nature of the underlying dynamical systems as well as testing for the degree and strength of nonlinearity. The reconstructed differential equations would then be potentially available for classical numerical studies, such as bifurcation analysis. These results represent a physically interpretable enhancement of data-driven modeling approaches in structural dynamics. Keywords: signal processing; sparse regression; system identiﬁcation; impulse response; optimization; feature generation; structural dynamics; time series classiﬁcation Merten Stender 1,2* , Sebastian Oberst 2 and Norbert Hoffmann 1,3 1 Department of Mechanical Engineering, Hamburg University of Technology, 21073 Hamburg, Germany; n.hoffmann@imperial.ac.uk 2 Centre for Audio Acoustics and Vibration, University of Technology Sydney, 2007 Broadway, Australia; sebastian.oberst@uts.edu.au 1 Department of Mechanical Engineering, Hamburg University of Technology, 21073 Hamburg, Germany; n.hoffmann@imperial.ac.uk 2 Centre for Audio Acoustics and Vibration, University of Technology Sydney, 2007 Broadway, Australia; sebastian.oberst@uts.edu.au 2 Centre for Audio Acoustics and Vibration, University of Technology Sydney, 2007 Broadway, Australia sebastian.oberst@uts.edu.au 3 Department of Mechanical Engineering, Imperial College London, London SW7 2AZ, UK * Correspondence: m.stender@tuhh.de; Tel.: +49-40-42878-3120 3 Department of Mechanical Engineering, Imperial College London, London SW7 2AZ, UK * Correspondence: m.stender@tuhh.de; Tel.: +49-40-42878-3120 Vibration 2019, 2, 25–46; doi:10.3390/vibration2010002 Recovery of Differential Equations from Impulse Response Time Series Data for Model Identiﬁcation and Feature Extraction Merten Stender 1,2* , Sebastian Oberst 2 and Norbert Hoffmann 1,3 1. Introduction Impact and impulse measurements represent a major challenge for feature extraction as the dynamics are highly transient and short, in some cases causing subsequent values to differ in orders of magnitudes [4,6]. Hence, discriminative features cannot always be derived from classical methods like the Fourier transform for those signals. While each of these objectives has unique features, specialized methods and particularities, there are common themes throughout all purposes for data acquisition in a mechanical engineering environment. Typically, time series data are collected, which is a number of values arranged in a sequence prescribed by time, and is then post-processed to yield quantities of interest. These quantities are the input values for the methods used and thus must be discriminative for the respective purpose. This work aims to derive new features for short and highly transient vibration time series data. Particularly, we propose to derive governing differential equations directly from the vibration data [7]. Along with increased data accessibility, a rise in data-driven methods can be observed. Recently, several methodologies have been proposed that enable the extraction of differential equations from time series data. Mostly driven by ﬂuid dynamics applications, nonlinear system identiﬁcation and control, different core methodologies can be observed: • advanced signal processing for updating gray-box models [1,2,4,8,9], g p g p g al network based methods [10,11], • compressive sensing approaches [12–14], • and sparse regression approaches [15,16]. In this work, we employ the Sparse Identiﬁcation of Nonlinear Dynamics (SINDy) approach recently proposed by Brunton et al. [15]. Using SINDy, a set of ordinary differential equations that models the input time series data can be obtained. Because not all states can be measured during experiments, state space reconstructions are utilized to ﬁnd the missing active degrees of freedom that are encoded in a measurement. Then, the sparse regression allows for ﬁnding a set of coefﬁcients that represents a set of ordinary differential equations (ODEs) in a prescribed space of admissible ansatz functions. The core contribution of this work is similar to that of Oberst et al. [17] but is more comprehensive and general in that it constitutes an extension of SINDy to a higher degree of automation. Here, the SINDy ODE approximation is subjected to a constrained nonlinear optimization scheme to deﬁne equations that better replicate the input data and favor simplicity of the reconstructed model. The approximation quality is assessed by means of time series comparison measures. 1. Introduction As measurement equipment and sensors have become less expensive and computationally more powerful in recent years, the amount and quality of data acquired in various engineering applications have steadily improved. While data have been collected and analyzed in classical laboratory testing ever since, online measurements of machines during operation are becoming increasingly popular. Revisiting the main reasons for data collection on mechanical structures, at least three fundamental motives can be identiﬁed: 1. Model identiﬁcation, model validation and model updating for control and numerical simulation, mostly performed during the product development phase [1–4]. 1. Model identiﬁcation, model validation and model updating for control and numerical simulation, mostly performed during the product development phase [1–4]. Vibration 2019, 2, 25–46; doi:10.3390/vibration2010002 www.mdpi.com/journal/vibration Vibration 2019, 2 26 2. State observation and supervision, especially for quality assurance and human–machine interaction purposes during operation. 2. State observation and supervision, especially for quality assurance and human–machine interaction purposes during operation. p p g p 3. Future state and fault prediction, commonly known as structural health monitoring [5]. p p g p 3. Future state and fault prediction, commonly known as structural health monitoring [5]. While each of these objectives has unique features, specialized methods and particularities, there are common themes throughout all purposes for data acquisition in a mechanical engineering environment. Typically, time series data are collected, which is a number of values arranged in a sequence prescribed by time, and is then post-processed to yield quantities of interest. These quantities are the input values for the methods used and thus must be discriminative for the respective purpose. For example, a trend of changing feature values might indicate and predict a fault in a machine or its malfunction—a situation commonly studied in the ﬁeld of structural health monitoring. While some of the post-processing methods do not rely on a sequential character, such as statistical moments, others, such as spectral analysis, are deﬁned based on the temporal information of sequence data (frequencies). In both cases, the sample of values drawn from measurement are assumed to (a) represent the underlying distribution well and (b) remain stationary for the time of observation. However, real vibration data from complex operating structures are often non-stationary, leading to less descriptive classical signal processing features as the data become decreasingly stationary. 1. Introduction Hence, the proposed approach represents an integrated, data-driven model identiﬁcation methodology that is speciﬁcally developed for short and transient signals [17]. The structure of the extracted set of ordinary differential equations (ODEs) can be used for several purposes, such as model identiﬁcation, model updating, and as a feature engineering proxy for machine learning techniques. In the following, we illustrate the SINDy algorithm and related techniques in detail. Then, a single degree-of-freedom (DOF) damped oscillator is studied as a reference in various linear and nonlinear conﬁgurations. The complexity of the ODE identiﬁcation is gradually increased by removing information. Opportunities and limitations of the proposed method are discussed with regards to model identiﬁcation and feature engineering purposes. 27 Vibration 2019, 2 27 Vibration 2019, 2 2. Methods In the following, we introduce the data-driven reconstruction of differential equations as well as several techniques that are required to yield an optimal reconstruction given time-limited and noise-contaminated data. 2.1. Sparse Identiﬁcation of Non-Linear Dynamics—SINDy We apply the methodology proposed by Brunton et al. [15] to ﬁnd differential equations that describe measured time signals. The sparse identiﬁcation of nonlinear dynamics (SINDy) makes use of a sparsity assumption on the observed dynamics: most physical phenomena can be described by governing equations that are rather sparse in the huge space of possible nonlinear functions. For instance, the Lorenz system consists of three ordinary differential equations with only two multiplicative terms of the system’s states. Hence, this system can be considered very sparse in the possible function space of polynomials, irrational functions and others. However, the Lorenz system exhibits regular as well as irregular dynamics and is able to describe some features of atmospheric convection, which is inarguably complex, involving a plethora of qualitatively different dynamics. Next, the SINDy framework, as described in detail by Brunton et al. [15], is brieﬂy introduced here, with focus on its interface to nonlinear optimization techniques. The procedure of SINDy is depicted in Figure 1. Here, only time-continuous physical ﬂows of the type ˙x(t) = f (x (t)) (1) (1) are considered. The system states x ∈Rm as well as their temporal derivatives ˙x are assumed to be accessible, while the underlying system is unknown. For instance, the system states can be measured during vibration experiments and their derivatives can be approximated numerically. Then, a library of nonlinear functions Θ is compiled. The central objective of the reconstruction is to approximate f from Equation (1) with as few terms as possible from the functions library, which is to ﬁnd a sparse representation in the space of possible candidate functions. The state measurements and derivatives for time instances t1, . . . , tN are collected in matrices X = x⊤ 1 (t) , . . . , x⊤ m (t) and ˙X = ˙x⊤ 1 (t) , . . . , ˙x⊤ m (t) , respectively. To reconstruct the underlying system of differential equations, the over-determined system of equations ˙X = Θ (X) Ξ (2) (2) ˙X = Θ (X) Ξ is evaluated using the function library Θ for each system state. The central idea of SINDy is then to solve the resulting system of equations using a sparse regression approach. 2.1. Sparse Identiﬁcation of Non-Linear Dynamics—SINDy In contrast to classical solution schemes, such as least-squares algorithms, this procedure promotes sparsity in the coefﬁcient matrix Ξ, which provides the best solution to the over-determined system. As a result, f, which is the underlying dynamical system, is approximated by only a few functions from the library of ansatz functions. Hence, the solution to the reconstruction problem is sparse with respect to many possible candidate functions to approximate the nonlinear relation between system states and derivatives. The reconstructed set of differential equations can be read directly from the non-zero entries in the solution Ξ ⊤ ˙˜x = Ξ⊤ Θ x⊤⊤ ≈f (x) , (3) (3) where Θ (x) denotes the evaluation of the symbolic library of nonlinear functions and ˜x the reconstructed state. Essentially, each coefﬁcient vector ξ deﬁnes the linear combination of nonlinear functions for each state of the system: where Θ (x) denotes the evaluation of the symbolic library of nonlinear functions and ˜x the reconstructed state. Essentially, each coefﬁcient vector ξ deﬁnes the linear combination of nonlinear functions for each state of the system: ˙˜x1 = ξ1,1 · θ1 (x) + ξ1,2 · θ2 (x) + · · · + ξ1,k · θk (x) . (4) (4) Here, θ denotes one of the k ansatz functions, such as a quadratic polynomial (·)2, in Θ—see also the example in Figure 1. The reconstruction result, i.e., the set of governing differential equations, is Here, θ denotes one of the k ansatz functions, such as a quadratic polynomial (·)2, in Θ—see also the example in Figure 1. The reconstruction result, i.e., the set of governing differential equations, is 28 28 Vibration 2019, 2 built from the sparse coefﬁcient matrix and the library of nonlinear functions. This system of equations can then be subjected to further analysis, such as forward time integration or bifurcation analysis. In this work, we use the time integration of the reconstructed equations to compare the reconstructed dynamics ˜x (t) to the observed dynamics x (t), i.e., the input of SINDy, to deﬁne a measure for the reconstruction error estimate Γ. time = ˙x ˙y 1 x y x2 y3 ξ1 ξ2 library of nonlinear functions sparse coeﬃcients non-zero entries ˜x(t) ˜y(t) input: x(t), y(t) . . . 2.1. Sparse Identiﬁcation of Non-Linear Dynamics—SINDy from sparse regression ˙x(t), ˙y(t) time ˙˜x = ξ1,3 · y = ˙x ˙y 1 x y x2 y3 ξ1 ξ2 library of nonlinear functions sparse coeﬃcients non-zero entries ˙˜y = ξ2,2 · x + ξ2,3 · y + ξ2,7 · x3 ˙x = y ˙y = y x x3 time integration ˜x(t) ˜y(t) reconstructed diﬀerential equations input: x(t), y(t) . . . from sparse regression ⇒ ˙x(t), ˙y(t) ξ1 ξ2 Figure 1. Schematic illustration of the SINDy procedure, closely related to the illustration given in [15]. System states and their derivatives are required as time series input. Then, a library of nonlinear candidate functions is compiled. The resulting over-determined system of equations is solved by sparsity-promoting regression techniques to yield a sparse reconstruction of the underlying dynamical system in terms of ordinary differential equations. non-zero entries from sparse regression time sparse coeﬃcients ˙˜x = ξ1,3 · y ˙˜y = ξ2,2 · x + ξ2,3 · y + ξ2,7 · x3 ˙x = y ˙y = y x x3 time integration reconstructed diﬀerential equations ⇒ ξ1 ξ2 ξ1 reconstructed diﬀerential equations Figure 1. Schematic illustration of the SINDy procedure, closely related to the illustration given in [15]. System states and their derivatives are required as time series input. Then, a library of nonlinear candidate functions is compiled. The resulting over-determined system of equations is solved by sparsity-promoting regression techniques to yield a sparse reconstruction of the underlying dynamical system in terms of ordinary differential equations. 2.2. Time-Delay Embedding In measurements, data is limited to a ﬁnite number of signals points. Hence, only a fraction of the relevant dynamical states of a dynamical system can be measured in laboratory experiments. Time delay embedding allows for the reconstruction of the phase space of a deterministic motion that is observed in a uni-variate fashion. Takens’s theorem [18] guarantees that, for strictly deterministic processes, the trajectories in the reconstructed space have the same geometrical and dynamical properties as in the true phase space. Hence, time-delay embedding enables us to unfold the dynamics on an attractor, the invariant set, that is encoded in a single scalar time measurement. Then, the reconstructed attractor’s properties, so-called invariant measures, can be studied, which include so-called static quantities like the correlation dimension D2 or dynamic quantities like the Lyapunov spectrum Λ [19,20]. Time-delay embedding, given by x(tn) = h sn−(m−1)τ, sn−(m−2)τ, . . . , sn−τ, sn i , (5) (5) essentially re-arranges the scalar time series s(tn) via sampling into a m-variate time series x(tn). Here, τ denotes the delay, which measures the distance of successive entries that are arranged to the Vibration 2019, 2 29 instantaneous state vector of embedding dimension m—see Figure 2. The embedding parameters are obtained sequentially. First, the optimal delay value is derived from the temporal correlations present in the time series s(t). The aim is to ﬁnd temporally maximal de-correlated state space vectors. Hence, common approaches for deriving τ involve the ﬁrst zero of the signal’s auto-correlation function (ACF) or the ﬁrst minimum of the auto-mutual information (AMI) function which takes also nonlinear correlations into account. The embedding dimension must be chosen to be larger than twice the attractor’s dimension. For example, the motion on a limit cycle, i.e., a one-dimensional attractor, has to be embedded into two dimensions to fully unfold the attractor. While there exist multiple approaches to estimate the embedding dimension, the false nearest neighbor (FNN) algorithm is the standard tool for deriving the global dimension m [20–22]. The FNN algorithm iteratively increases the embedding dimension and at the same time tracks neighborhood relations in the m-dimensional phase space. False nearest neighbors are deﬁned as states that are neighbors in dimension m but lose their neighborhood relation once the embedding dimension is increased to m + 1. The fraction of false nearest neighbors measures the overall share of those points on all state space vectors. 2.2. Time-Delay Embedding Once the attractor is fully unfolded, states remain neighbors even if the dimension is further increased. Hence, the required embedding dimension is characterized by a low fraction of false neighbors in the m-dimensional reconstructed phase space. In this work, we employ time-delay embedding to reconstruct multivariate trajectories from uni-variate time series measurements. Additionally, the SINDy-reconstructed dynamics can be compared to measured dynamics by employing embedding parameters as error metric, cf. [23]. (a) s(t) t ∆t delay τ = 3 dimension m = 4 τ τ τ xn = sn−(m−1)τ, sn−(m−2)τ, . . . , sn−τ, sn t1 t2 t3 t10 t5 x1 =   s(t1) s(t4) s(t7) s(t10)   \| {z } □ , x2 =   s(t2) s(t5) s(t8) s(t11)   \| {z } △ , x3 =   s(t3) s(t6) s(t9) s(t12)   \| {z } ⃝ , . . . (b) τ = 3 m = 4 0 1 Fraction FNN Dimension Delay Mutual information 0 0 8 1 2 Figure 2. (a) schematic illustration of the time delay embedding procedure using the embedding parameters τ and m that were obtained from (b), the ﬁrst minimum of the auto-mutual information function and the dimension with vanishing fraction of false nearest neighbors, respectively. (a) (b) τ = 3 m = 4 0 1 Fraction FNN Dimension Delay Mutual information 0 0 8 1 2 (b) s(t) t ∆t delay τ = 3 dimension m = 4 τ τ τ xn = sn−(m−1)τ, sn−(m−2)τ, . . . , sn−τ, sn t1 t2 t3 t10 t5 x1 =   s(t1) s(t4) s(t7) s(t10)   \| {z } □ , x2 =   s(t2) s(t5) s(t8) s(t11)   \| {z } △ , x3 =   s(t3) s(t6) s(t9) s(t12)   \| {z } ⃝ , . . . Figure 2. (a) schematic illustration of the time delay embedding procedure using the embedding parameters τ and m that were obtained from (b), the ﬁrst minimum of the auto-mutual information function and the dimension with vanishing fraction of false nearest neighbors, respectively. 2.3. Uncertainty Suppressing Numerical Differentiation Besides all states, SINDy also requires the states’ time derivatives, either measured or generated numerically. Classical numerical differentiation schemes, such as ﬁnite differences, exhibit noise ampliﬁcation—as a rule of thumb, accuracy decreases by one order of magnitude per differentiation [24]. As measurements are always contaminated with noise, a differentiation scheme is required that is particularly robust to account for this additional loss of accuracy. We employ Total Variation Regularized Numerical Differentiation, TVRegDiff, as introduced by [25] and proposed in [15], to be used within SINDy. The central idea of this differentiation scheme is to regularize the differentiation itself by balancing (a) the irregularity introduced by the derivative and (b) the error Vibration 2019, 2 30 introduced by the smoothing property using a regularization parameter. Hence, this regularization parameter, denoted as α, plays a crucial role. In the following, we propose an integrated and completely data-driven approach for choosing an appropriate α for the speciﬁc data under study. In fact, the regularized derivative can serve as a ﬁltering method. The regularization can then be utilized to also dampen noise in the differentiation step. The smoothed derivative is then integrated in time to yield a qualitatively enhanced (ﬁltered) version of the input signal—see illustrative examples in Appendix A. In this work, we use TVRegDiff to compute the numerical derivatives of states x(t) with respect to time. In order to arrive at a self-consistent procedure, we substitute the original states with the time-integrated derivatives for usage in the SINDy procedure as explained before. This approach ensures that states and derivatives are matching. Here, we propose to deﬁne a time series difference measure that compares the raw input signal to the one obtained through numerical derivation and time-integration similar to Oberst et al. [17]. The search of the difference minimum will yield the optimal regularization parameter—see Figure 3. In order to reduce boundary effects at the beginning and end of each time series, the ﬁve ﬁrst and last samples are discarded after differentiation. TVRegDiﬀ error measure →α R ˙xdt Figure 3. Integrated approach for ﬁnding the optimal regularization parameter α for the numerical differentiation using TVRegDiff: the time integration of the numerical derivative allows for deriving an error measure that can be used to adapt the regularization parameter for noise contaminated input time series. TVRegDiﬀ Figure 3. 2.3. Uncertainty Suppressing Numerical Differentiation Integrated approach for ﬁnding the optimal regularization parameter α for the numerical differentiation using TVRegDiff: the time integration of the numerical derivative allows for deriving an error measure that can be used to adapt the regularization parameter for noise contaminated input time series. 2.5. Constrained Nonlinear Optimization The sparse regression solution of the over-determined SINDy system of equations is not continuous in the sense of the coefﬁcient matrix Ξ, which is a function of the sparsity parameter λ. This means that a ﬁxed set of coefﬁcients will result from a range of sparsiﬁcation levels λ. If λ is increased, at some point, an additional coefﬁcient entry will be set to zero. Until then, the coefﬁcients remain constant. As a result, SINDy may be able to ﬁnd the correct locations of non-zero coefﬁcient entries, but not their optimal absolute value. We propose to use a nonlinear optimization scheme to ﬁne-tune the entries in the sparse coefﬁcient matrix found by SINDy. In this work, the Matlab routine fmincon is employed to run the optimization using the sequential quadratic programming (sqp) method. For each non-zero coefﬁcient, we prescribe a lower and upper bound. Then, the optimizer ﬁnds the optimal set of coefﬁcients with respect to the cost function, which measures the difference between input signal and reconstruction signal. Hence, the SINDy results represent the ﬁrst approximation of the reconstruction and the nonlinear optimization ﬁne-tunes the model for the best ﬁt. Depending on the speciﬁc reconstruction objective, the cost function may be formulated such that it penalizes complexity and nonlinearities, promotes matching spectral properties [30], or improves other characteristics of the reconstructed model. 2.4. Time Series Comparison Measures This work extracts differential equations from time series input data. Hence, appropriate error measures have to be chosen to compare the reconstructed dynamics with the input data and thus quantify the reconstruction quality. Particularly, the set of reconstructed ODEs can be integrated in time (setting the initial value to the ﬁrst input data entry) to create the reconstructed time series. Depending on the qualitative character of the time series and the speciﬁc reconstruction purpose, the reconstruction error needs to be deﬁned. This objective represents an analogon to research ﬁelds of time series classiﬁcation and time series feature engineering, where ﬁnding discriminative features forms the central methodical challenge [26–28]. The types of features, and thus possible candidates for deﬁnition of the reconstruction error, can be clustered as follows: • Instance-based schemes that compare contemporaneous pairs of time series instances. In the simplest case, sequences are subtracted from each other. Modiﬁcations and advanced approaches include warping methods that add more ﬂexibility and the ability to also take into account phase shifts. • Instance-based schemes that compare contemporaneous pairs of time series instances. In the simplest case, sequences are subtracted from each other. Modiﬁcations and advanced approaches include warping methods that add more ﬂexibility and the ability to also take into account phase shifts. 31 Vibration 2019, 2 • Higher-level features based on transforms, sampling strategies or correlation measures. Examples include statistical moments and linear transforms, such as the Fourier transform. Time series comparison is then performed based on features extracted from the transforms, such as the standard deviation or major periodicities. j p • Quantiﬁers for qualitative behavior, mostly borrowed from nonlinear time series analysis and complexity sciences [29]. Here, the actual shape of the sequence is rather irrelevant. Instead, the qualitative nature of the dynamics encoded in the time series is of interest: dynamical invariants quantify the degree of regularity, entropy, or fractal properties of the sequence when studied in a dynamical framework, cf. [17]. Depending on the speciﬁc signal, reconstruction purpose and availability of data, any of the aforementioned features may be chosen to quantify the reconstruction error between the input and SINDy-reconstructed time series. In this work, we use the cumulative difference between reconstructed and original signal as a measure for instance-based reconstruction error due to its higher robustness, computational efﬁciency, and simplicity of implementation compared to e.g., recurrence plot-based measures [17]. 2.6. Models Used We employ a linear, 1-DOF, viscously damped mechanical oscillator to illustrate the automation and optimization extensions of SINDy. Special focus is put on the short and highly transient nature of ring-down data, i.e., impulse-response-type time series data. As a reference, the oscillator is deﬁned as shown in Figure 4. Owing to high damping levels, there are usually only a few signiﬁcant oscillations. Hence, determining classical features of ring-down data, such as the decay rate, may be challenging and prone to errors. In the following, we discuss the reconstruction in the phase space, which is represented by a ﬁrst-order description of the form ˙x1 = x2 (6) ˙x2 = −c M x2 −k M x1 . (6) This dynamical system is integrated in time for 4 s using a sampling time of ∆t = 0.01 s. If not denoted differently, a unit initial deﬂection is prescribed, i.e., x(t = 0) = [1, 0]⊤. For the sake of clearer notation, we discard the physical units in the following discussions. Vibration 2019, 2 32 M k c x1, x2 M k c x1, x2 0 1 2 3 4 5 −1 −0.5 0 0.5 1 time position x1 M = 1, k = 64, c = 3.2 0 1 2 3 4 5 −5 0 5 time velocity x2 Figure 4. Linear 1-DOF viscously damped oscillator and time traces of position x1 and velocity x2 for an initial deﬂection of x1 (t = 0) = 1. 0 1 2 3 4 5 −1 −0.5 0 0.5 1 time position x1 M = 1, k = 64, c = 3.2 0 1 2 3 4 5 −5 0 5 time velocity x2 Figure 4. Linear 1-DOF viscously damped oscillator and time traces of position x1 and velocity x2 for an initial deﬂection of x1 (t = 0) = 1. 3. Results We elaborate on the proposed SINDy extensions using the example of the mechanical oscillator that oscillates freely. In the ﬁrst step, the linear damped oscillator is subjected to the SINDy reconstruction for various degrees of input information level. The complexity of the reconstruction task is increased incrementally. First, the analytical derivatives are replaced by the regularized derivatives. Then, only the ﬁrst state is recorded, which requires reconstructing the second state using time delay embedding. Secondly, we illustrate how the optimization procedure can improve both the reconstruction quality and the simplicity of the reconstructed models. Finally, a cubic stiffness term is introduced and identiﬁed by the proposed methods. These cascades of decreasing amounts of information available to SINDy are illustrated in Figure 5. 33 33 Vibration 2019, 2 Embedding + TVRegDiﬀ dx/dt = f (x) x1 x2 ˙x1 ˙x2 SINDy 1 optimization dx/dt = f (x) x1 x2 ˙x1 ˙x2 SINDy 2 dx/dt = f (x) x1 SINDy 4 TVRegDiﬀ dx/dt = f (x) x1 x2 SINDy 3 optimization optimization TVRegDiﬀ dx/dt = f (x) x1 x2 SINDy 5 optimization nonlinear system linear system Figure 5. Overview of the studies performed using the 1-DOF oscillator. The reconstruction task becomes increasingly challenging due to decreasing amounts of information available and added nonlinearity. Figure 5. Overview of the studies performed using the 1-DOF oscillator. The reconstruction task becomes increasingly challenging due to decreasing amounts of information available and added nonlinearity. 3.1. Instance-Based Error Measure As the error measure, we employ the cumulative absolute Euclidean distance, given by As the error measure, we employ the cumulative absolute Euclidean distance, given by Γ = DOF ∑ i=1 N ∑ j=1 xtrue,i(tj) max (\|xtrue,i\|) − xrecon,i(tj) max (\|xtrue,i\|) (7) (7) throughout the following discussion. For each DOF, the difference is computed between the true xtrue and reconstructed xrecon states per time step tj. Then, all differences are added up to form the scalar error measure Γ. To achieve a balanced error measure, each time series of length N is normalized by the respective maximum of the true states per DOF. Hence, differences in position and velocity are weighted equally in the error measure and do not depend on the actual value range per DOF. 3.1.1. Providing Full Information: All States and Analytical Derivatives 3.1.1. Providing Full Information: All States and Analytical Derivatives Initially, all information is provided for SINDy: position and velocity signals as well as their derivatives are provided from the time integration of the equations of motion. As we provide a library of polynomials up to order p = 3, the exact coefﬁcients should be found by SINDy. However, as the parameter remains the sparsiﬁcation level, which has to be deﬁned according to the length, the sampling time, the nonlinear function library and the qualitative nature of the dynamics. Following the previous illustrations, the sparsiﬁcation parameter λ is varied in the full range between non-zero entries (NZE) of 100% to NZE = 0%. Meanwhile, the error measure is computed from the time domain solutions of instantaneous SINDy results. Then, the optimal sparsiﬁcation parameter is chosen such that it minimizes the error measure. The resulting SINDy reconstruction is integrated in time and compared to the input signal—see Figure 6. The result of SINDy matches perfectly with the analytical expression for the oscillator. Even though polynomials up to order pmax = 3 were permitted, 34 34 Vibration 2019, 2 Vibration 2019, 2 the optimal sparsiﬁcation level yields the correct ﬁrst-order differential equations ˙x1 = 1 · x2 and ˙x2 = −64 · x1 −3.2 · x2 with coefﬁcients equaling zero for all other entries, thus an NZE fraction of 3/20. As the error is close to machine precision, we do not need to further tune the non-zero coefﬁcients, regardless of whether or not the true solution of the input signal is known. (b) −1 0 1 2 y raw reconstructed 0 1 2 3 4 −5 0 5 time ˙y (a) 10−18 10−13 10−8 10−3 102 10−12 10−8 10−4 100 104 sparsiﬁcation λ error measure 10−18 10−13 10−8 10−3 102 0 0.2 0.4 0.6 0.8 1 fraction NNZ λopt = 1.4175 · 10−16 (b) −1 0 1 2 y raw reconstructed 0 1 2 3 4 −5 0 5 time ˙y Figure 6. (a) Finding the optimal sparsiﬁcation level as minimum of the error measure; (b) comparison of input signal to SINDy reconstruction with maximal polynomial order pmax = 3. (a) 10−18 10−13 10−8 10−3 102 10−12 10−8 10−4 100 104 sparsiﬁcation λ error measure 10−18 10−13 10−8 10−3 102 0 0.2 0.4 0.6 0.8 1 fraction NNZ λopt = 1.4175 · 10−16 (b) (a) Figure 6. 3.1.1. Providing Full Information: All States and Analytical Derivatives (a) Finding the optimal sparsiﬁcation level as minimum of the error measure; (b) comparison of input signal to SINDy reconstruction with maximal polynomial order pmax = 3. 3.1.2. Providing Less Information: All States and Numerical Derivatives Analytical derivatives will hardly be available for an analysis as proposed here. In other words, if states and derivatives were available, there would be different solutions to ﬁnd the best system of differential equations that links states with derivatives. Hence, in this scenario, we record only the states x1 and x2 and compute their derivatives using TVRegDiff. Therefore, the regularization parameter α, has to be set prior to applying SINDy. Here, we consider noise-free signals, which makes the selection of α rather straightforward. Following the description illustrated in Figure 3, we adapt the regularization such that it minimizes the difference between an input state and the state which is obtained after time integration of the numerical derivative—see Figure 7. When dealing with noise-contaminated signals, this is a semi-automated procedure that must be supervised by the user in order to make use of the ﬁltering property of regularized derivatives, depicted in Appendix A. (a) 10−9 10−7 10−5 10−3 10−1 101 102 103 regularization α error measure minimum αopt = 1.87 · 10−9 (b) 0 1 2 3 4 −1 −0.5 0 0.5 1 time t s(t) sin sf 3.6 3.9 0.35 0.5 Figure 7. Comparison of the input time signal sin and the ﬁltered one sf obtained through time integration of the regularized numerical derivative. (a) Error measure as a function of the regularization parameter α and (b) the resulting signal for an optimal regularization αopt. (a) 10−9 10−7 10−5 10−3 10−1 101 102 103 regularization α error measure minimum αopt = 1.87 · 10−9 (b) (a) 0 1 2 3 4 −1 −0.5 0 0.5 1 time t ( ) sin sf 3.6 3.9 0.35 0.5 Figure 7. Comparison of the input time signal sin and the ﬁltered one sf obtained through time integration of the regularized numerical derivative. (a) Error measure as a function of the regularization parameter α and (b) the resulting signal for an optimal regularization αopt. The ﬁltered states, that is the time integrated numerical derivatives, are fed as input xin to the SINDy procedure along with the numerical derivatives. Using a maximal polynomial order p = 3, the sparse reconstruction is obtained following the procedure elaborated before. The best reconstruction is obtained for a sparsiﬁcation parameter λ = 0.3583. The reconstructed time signals for x1 and x2 match Vibration 2019, 2 35 well with the input signals—see Figure 8a. 3.1.2. Providing Less Information: All States and Numerical Derivatives Out of the 20 possible coefﬁcients, there are four non-zero entries: ˙x1 = 0.999 · x2, ˙x2 = −64.8124 · x2 −2.9204 · x1 −0.5125 · x3 1. The cubic term represents an artefact that was introduced by SINDy: compared to the true system of equations, one can observe that the linear system’s eigenfrequency is slightly too large ( √ 64.81 > √ 64.0), while the damping term is slightly smaller (2.92 < 3.2), which explains the additional cubic stiffness term. This cubic term creates larger restoring forces and thus compensates for the lack of damping in the system. As the reconstruction error is rather small, and the reconstructed dynamics are sparse in the admissible nonlinear function library space, one may ﬁnish the reconstruction at this point. However, we strive for an even smaller reconstruction error, fewer model terms and a test for nonlinearity. In particular, the latter objective may be of interest for various studies in system identiﬁcation scenarios. In structural dynamics modeling, the degree of linearity is a point of crucial importance. Since linear systems obey the superposition principle, analysis of these systems can be performed with ease using a variety of established methods. Strongly nonlinear systems, on the other hand, require more complex analytical methods and can exhibit a profusion of qualitatively different dynamics that typically have not been understood as thoroughly as those of linear systems. Hence, in the current case, it might be interesting to see whether the cubic stiffness term is a decisive model parameter, or if the system can be modeled in linear fashion instead. A central contribution of our work is the extension of the SINDy reconstruction with a nonlinear optimization procedure. The optimization problem is formulated such that the non-zero coefﬁcients are subjected to a bounded optimization with respect to a global cost function that measures the difference of input and reconstructed time signals. The non-zero entries (NZE) in Ξ found by SINDy represent the starting point for the optimization, which is permitted to change the NZE values within prescribed boundaries ξ−and ξ+. The optimization boundaries are set in relative fashion, i.e., allowing for a certain fraction of change with respect to the initial value. However, to promote sparsiﬁcation during the optimization, we set one of the boundaries to zero. 3.1.2. Providing Less Information: All States and Numerical Derivatives This approach, which allows for further elimination of model parameters, corresponds to the dropout or pruning procedure, a well-known concept of complexity reduction in artiﬁcial neural networks [31]. Here, we allow a maximal relative change of 50% according to ξi < 0 →ξ+ −= [0, ξi · γ], γ = 1.5, ξi < 0 →ξ+ −= [ξi · γ, 0]. (8) (8) In our case, the non-zero entries in ξ obtained by SINDy are the coefﬁcients of x2 in the ﬁrst differential equation and those of x1, x2 and x3 1 in the second differential equation. Figure 8b depicts the optimization iterations of the bounded nonlinear optimization. The error measure decays quickly within the ﬁrst ten iterations. Only slight adaptations are made in the following iterations before the optimization algorithm ﬁnishes at 48 iterations. When observed in the time domain, the optimization iterations are not visible in the full-scale time signal but can be identiﬁed through high magniﬁcation of the peak regions—see Figure 8c. The ﬁnal minimisation result is then compared to the input signal, i.e., the reconstruction aim, and the SINDy result in Figure 8d. Through optimization, the reconstruction error Γ has been reduced by 81.95% to 2.345. While the reconstruction has been enhanced, the coefﬁcient matrix Ξ has also become sparser thanks to this method: The auxiliary cubic stiffness term that was computed by SINDy is erased by the optimization, which makes use of the dropout option discussed before. The reconstructed set of differential equations after optimization is very similar to the analytical model which was used to create the input data—see Figure 8e. 3.1.3. Providing Less Information: State Space Reconstruction and Numerical Derivatives Vibration 2019, 2 36 Vibration 2019, 2 Vibration 2019, 2 We mimic this situation by restricting the input data to a single scalar time series, e.g., x1 of the oscillator. The missing information about other relevant degrees of freedom in the true phase space has to be reconstructed using time delay embedding. Then, TVRegDiff allows for computing the corresponding derivatives that can be used for the SINDy reconstruction with subsequent coefﬁcient optimization. We mimic this situation by restricting the input data to a single scalar time series, e.g., x1 of the oscillator. The missing information about other relevant degrees of freedom in the true phase space has to be reconstructed using time delay embedding. Then, TVRegDiff allows for computing the corresponding derivatives that can be used for the SINDy reconstruction with subsequent coefﬁcient optimization. y q p The embedding parameters τ = 18, m = 2 are found for the x1 time series by the ﬁrst zero of the auto-correlation function and by the false nearest neighbor algorithm—see Figure 9. The resulting trajectories q1, q2 in the reconstructed phase space and their respective numerical derivatives are used as inputs to SINDy. (a) −1 0 1 x1 input SINDy 0 1 2 3 4 −5 0 5 time x2 (b) 0 10 20 30 40 50 5 10 optimization iteration cost value (c) 0 1 2 3 4 0 2 time x1 SINDy iteration final 0.7 0.8 0.9 0.1 0.2 0.3 0.4 (d) 0 1 2 3 4 −1 0 1 2 3 time x1 input SINDy ﬁnal 1.4 2.2 0 0.1 (e) ˙x1 ˙x2 1 [0] [0] x1 [0] [−64] x2 [1] [−3.2] x1x1 [0] [0] x1x2 [0] [0] x2x2 [0] [0] x1x1x1 [0] [0] x1x1x2 [0] [0] x1x2x2 [0] [0] x2x2x2 [0] [0] ˙x1 ˙x2 1 [0] [0] x1 [0] [−64.81] x2 [0.99] [−2.92] x1x1 [0] [0] x1x2 [0] [0] x2x2 [0] [0] x1x1x1 [0] [−0.51] x1x1x2 [0] [0] x1x2x2 [0] [0] x2x2x2 [0] [0] ˙x1 ˙x2 1 [0] [0] x1 [0] [−64.43] x2 [0.98] [−3.19] x1x1 [0] [0] x1x2 [0] [0] x2x2 [0] [0] x1x1x1 [0] [0] x1x1x2 [0] [0] x1x2x2 [0] [0] x2x2x2 [0] [0] analytical model SINDy result optimization result Figure 8. Vibration 2019, 2 Vibration 2019, 2 Two-step reconstruction procedure: (a) comparison of input data and SINDy reconstruction that used regularized numerical derivatives; (b,c) iterations of the bounded optimizer on the non-zero coefﬁcients to reduce the difference between input and reconstruction signal; (d) result of the two-step reconstruction and (e) evolution of the coefﬁcient matrices ξ obtained from SINDy and the optimization procedure in comparison to the analytical oscillator model that generated the initial input signals. (a) −1 0 1 x1 input SINDy 0 1 2 3 4 −5 0 5 time x2 (b) 0 10 20 30 40 50 5 10 optimization iteration cost value (b) (a) (c) 0 1 2 3 4 0 2 time x1 SINDy iteration final 0.7 0.8 0.9 0.1 0.2 0.3 0.4 (d) 0 1 2 3 4 −1 0 1 2 3 time x1 input SINDy ﬁnal 1.4 2.2 0 0.1 (d) (c) (e) ˙x1 ˙x2 1 [0] [0] x1 [0] [−64] x2 [1] [−3.2] x1x1 [0] [0] x1x2 [0] [0] x2x2 [0] [0] x1x1x1 [0] [0] x1x1x2 [0] [0] x1x2x2 [0] [0] x2x2x2 [0] [0] ˙x1 ˙x2 1 [0] [0] x1 [0] [−64.81] x2 [0.99] [−2.92] x1x1 [0] [0] x1x2 [0] [0] x2x2 [0] [0] x1x1x1 [0] [−0.51] x1x1x2 [0] [0] x1x2x2 [0] [0] x2x2x2 [0] [0] ˙x1 ˙x2 1 [0] [0] x1 [0] [−64.43] x2 [0.98] [−3.19] x1x1 [0] [0] x1x2 [0] [0] x2x2 [0] [0] x1x1x1 [0] [0] x1x1x2 [0] [0] x1x2x2 [0] [0] x2x2x2 [0] [0] analytical model SINDy result optimization result (e) ˙x1 ˙x2 1 [0] [0] x1 [0] [−64] x2 [1] [−3.2] x1x1 [0] [0] x1x2 [0] [0] x2x2 [0] [0] x1x1x1 [0] [0] x1x1x2 [0] [0] x1x2x2 [0] [0] x2x2x2 [0] [0] analytical model (e) ˙x2 [0] −64] −3.2] [0] [0] [0] [0] [0] [0] [0] ˙x1 ˙x2 1 [0] [0] x1 [0] [−64.81] x2 [0.99] [−2.92] x1x1 [0] [0] x1x2 [0] [0] x2x2 [0] [0] x1x1x1 [0] [−0.51] x1x1x2 [0] [0] x1x2x2 [0] [0] x2x2x2 [0] [0] ˙x1 ˙x2 1 [0] [0] x1 [0] [−64.43] x2 [0.98] [−3.19] x1x1 [0] [0] x1x2 [0] [0] x2x2 [0] [0] x1x1x1 [0] [0] x1x1x2 [0] [0] x1x2x2 [0] [0] x2x2x2 [0] [0] del SINDy result optimization result Figure 8. 3.1.3. Providing Less Information: State Space Reconstruction and Numerical Derivatives In the next step, the amount of information provided by the signal input is further reduced. In real measurements, only a few measurement points are available for data acquisition. Therefore, only a fraction of the active degrees of freedom will be measured and available to the reconstruction procedure. Vibration 2019, 2 Two-step reconstruction procedure: (a) comparison of input data and SINDy reconstruction that used regularized numerical derivatives; (b,c) iterations of the bounded optimizer on the non-zero coefﬁcients to reduce the difference between input and reconstruction signal; (d) result of the two-step reconstruction and (e) evolution of the coefﬁcient matrices ξ obtained from SINDy and the optimization procedure in comparison to the analytical oscillator model that generated the initial input signals. Figure 8. Two-step reconstruction procedure: (a) comparison of input data and SINDy reconstruction that used regularized numerical derivatives; (b,c) iterations of the bounded optimizer on the non-zero coefﬁcients to reduce the difference between input and reconstruction signal; (d) result of the two-step reconstruction and (e) evolution of the coefﬁcient matrices ξ obtained from SINDy and the optimization procedure in comparison to the analytical oscillator model that generated the initial input signals. Figure 8. Two-step reconstruction procedure: (a) comparison of input data and SINDy reconstruction that used regularized numerical derivatives; (b,c) iterations of the bounded optimizer on the non-zero coefﬁcients to reduce the difference between input and reconstruction signal; (d) result of the two-step reconstruction and (e) evolution of the coefﬁcient matrices ξ obtained from SINDy and the optimization procedure in comparison to the analytical oscillator model that generated the initial input signals. 37 Vibration 2019, 2 (a) 0 10 20 30 40 −1 0 1 delay ACF 1 2 3 4 0 1 dimension fraction FNN τ = 18 m = 2 (b) 0 1 2 3 4 −0.6 −0.4 −0.2 0 0.2 0.4 time amplitude q1 q2 Figure 9. (a) Determining the embedding parameters as a ﬁrst zero of the ACF and vanishing fraction of the FNN algorithm for x1; (b) resulting trajectories q1 and q2 in the reconstructed phase space. b) 0 1 2 3 4 −0.6 −0.4 −0.2 0 0.2 0.4 time amplitude q1 q2 (a) 0 10 20 30 40 −1 0 1 delay ACF τ = 18 (b) (a) Figure 9. (a) Determining the embedding parameters as a ﬁrst zero of the ACF and vanishing fraction of the FNN algorithm for x1; (b) resulting trajectories q1 and q2 in the reconstructed phase space. The optimized SINDy reconstruction is computed for polynomials varying from degrees p = 1 to p = 5. For each conﬁguration, the sparsity as well as the reconstruction error estimate Γ between trajectories q(t) and their reconstruction is computed. Vibration 2019, 2 The corresponding sets of ODEs are listed in Table 1. Generally speaking, the reconstructed ODEs will not mimic the analytical oscillator as a result of the embedding step. The reconstruction of the trajectories q(t) maps the input x1(t) to another space and, therefore, the resulting states cannot be compared directly to the analytical model. In contrast, the general structure of the dynamical system as well as nonlinearity serve as viable parameters for comparison. Table 1. Coefﬁcient matrices for optimized SINDy reconstruction for maximal polynomial degrees p and resulting reconstruction error Γ. The states q1 and q2 were taken from the time delay embedded signal x1 and used as input signals—see Figure 9. Terms that are not displayed became zero in this analysis. The reconstruction for p = 5 failed with an error of Γ = 159.66 and 13 NZE after optimization. 1 q1 q2 q1q2 q2 1q2 q3 2 Other Terms p = 1, Γ = 1.88 ˙q1 0 −0.74 −5.84 - - - - ˙q2 −0.008 10.60 −2.45 - - - - p = 2, Γ = 1.82 ˙q1 0 −0.75 −5.82 −0.06 - - - ˙q2 −0.007 10.62 −2.45 0 - - - p = 3, Γ = 1.63 ˙q1 0 −0.81 −5.84 −0.13 0.51 0.16 - ˙q2 0 10.60 −2.39 0 0 0 - p = 4, Γ = 1.64 ˙q1 0 −0.85 −5.86 0 0.7080 0 −0.96 q3 1q2 −1.44 q1q3 2 ˙q2 0 10.54 −2.36 0 0 0 0 The previous study involving both true model states x exhibits a reconstruction error of 2.345. In the current study with less information, reconstruction errors of the same order of magnitude can be observed. As the polynomial degree is increased, the reconstruction error decreases slightly before it saturates for p ≥3. The coefﬁcient matrices exhibit predominantly constant entries beginning with the simplest model, where p = 1. As a result of this trend and the low reconstruction error, one can conclude that a reconstruction using only linear combinations of the monomials q1 and q2 is sufﬁcient. However, since the original analytical model in this investigation was the simple linear oscillator, the Vibration 2019, 2 38 resulting linear reconstruction is not very surprising. Still, the processes of embedding and numerical differentiation could have introduced artefacts that would have increased the model complexity after reconstruction. 3.2. Balancing Reconstruction Error and Model Complexity As illustrated in the previous study, reconstruction error and model complexity represent conﬂicting objectives when dealing with incomplete or noisy data. We classify those conﬂicting objectives in the framework of under- and over-ﬁtting models, which is a classical framework in system identiﬁcation and machine learning research. The number of non-zero coefﬁcients in the reconstructed set of ODEs is a proxy for the model complexity. Hence, each SINDy model can be located in the plane spanned by model complexity and reconstruction error—see Figure 10a. Re-visiting the assumption of sparse dynamics, models of low complexity and low reconstruction error are the optimal solution to the time series based reconstruction challenge. A simple model with high reconstruction error does not capture all of the salient dynamics of the input and is thus under-ﬁtted. On the contrary, if high complexity is required to obtain a low reconstruction error, the model may over-ﬁt the dynamics and is unlikely to yield a generally applicable result. To display a sample case for selecting an optimal trade-off between model complexity and reconstruction error, we revisit the reconstruction from limited input data. In the previous study, the time delay embedding was computed from the ﬁrst state x1 of the analytical oscillator. In measurements, one will most likely measure a combination of the system’s states rather than a single active degree of freedom. Thus, we study how the composition of the measured signal impacts the reconstruction error. The input s(t) to the time delay embedding is now an aggregate of both states x1 and x2 according to s(t) = βx1 + (1 −β) x2, where the scaling parameter β prescribes the mixing ratio of both states. To study the effect of different input information, polynomial orders up to p = 4 are studied and β is varied in the range [1, 0] with a step size of 0.05. The resulting reconstruction results are displayed in Figure 10b. The best reconstructions, consisting of low complexity models with low error, are found for p = 3. Lower polynomial orders result in simple but imprecise reconstructions, while p = 4 is prone to over-ﬁtting, i.e., generating models of high complexity that reproduce the given data well, but generalize weakly. Depending on the speciﬁc purpose of the reconstruction, either conﬁguration may be favorable. Vibration 2019, 2 Depending on a speciﬁc objective, one may use the linear, or also the nonlinear models that arise from this study. The latter models show a smaller reconstruction error, while the number of non-zero entries increases. This means that SINDy arrives at better reconstructions with an increasing number of terms and, as a result, greater model complexity. Thus, both the reconstruction error and the model complexity have to be balanced. For a speciﬁc application, either of those objectives may be favored. In conclusion, SINDy is able to reconstruct dynamical models that reproduce the observed dynamics to a high degree, even if only a limited amount of input information is provided. In this case, a single state of the mechanical oscillator sufﬁces to reconstruct a dynamical model that captures the salient feature of the dynamics and exhibits low reconstruction errors for a linear model. This result is promising for the application of the proposed methods to real experimental impulse data, which offer limited access to the states of the system and thus require embedding approaches. 3.3. Model Identiﬁcation Studies For illustrative purposes, we simulate three qualitatively different model identiﬁcation scenarios. First, data is generated for different system conﬁgurations driven by a control parameter. This control parameter is then identiﬁed in the reconstructed models. Second, reconstruction models are built for data that stems from the same system, but for different initial conditions. This case is designed to represent multiple measurements that contain different amounts of information about a given dynamical system. Finally, data are generated for a linear and weakly nonlinear system conﬁguration. Here, we illustrate the potential of the proposed methods to serve as a test for nonlinearity. 3.2. Balancing Reconstruction Error and Model Complexity Still, the illustration of the resulting models in the complexity-error plane helps to ﬁnd the optimal choice of parameters such as the number of candidate functions in the library Θ. 39 Vibration 2019, 2 (a) optimum overﬁtting underﬁtting sparsiﬁcation level reconstruction error (b) 0 10 20 30 0 1 2 3 4 number NZE error measure Γ p = 1 p = 2 p = 3 p = 4 Figure 10. (a) Balancing model complexity and reconstruction error: depending on a speciﬁc parameter selection, the reconstructed equations may represent under-, well-, and over-ﬁtting models, as depicted in this schematic representation; (b) depicts the resulting model characteristics for various input conﬁgurations and varying polynomial orders of the nonlinear functions library. In this case, most of the p = 4 conﬁgurations result in very complex, i.e., unsatisfactory, reconstructions. (a) optimum overﬁtting underﬁtting sparsiﬁcation level reconstruction error (b) (a) ) 0 10 20 30 0 1 2 3 4 number NZE error measure Γ p = 1 p = 2 p = 3 p = 4 Figure 10. (a) Balancing model complexity and reconstruction error: depending on a speciﬁc parameter selection, the reconstructed equations may represent under-, well-, and over-ﬁtting models, as depicted in this schematic representation; (b) depicts the resulting model characteristics for various input conﬁgurations and varying polynomial orders of the nonlinear functions library. In this case, most of the p = 4 conﬁgurations result in very complex, i.e., unsatisfactory, reconstructions. 3.3.2. Robust Model Parameter Identiﬁcation Imagine testing a dynamical system using hammer impacts. Naturally, each excitation is different, but the dynamical system remains the same. We try to simulate this situation by varying the initial conditions for creating various input time series of the oscillator. Each realization stems from the same dynamical system, and thus we look for one model reconstructed from all data. For each piece of input data, an optimized SINDy model is computed and coefﬁcient values are recorded. Figure 12 depicts histograms of the relevant, i.e., non-zero, coefﬁcient matrix entries. In order to deﬁne the optimal model, one would select representative coefﬁcient values, such as the most frequent or median values, from these histograms and form the ﬁnal coefﬁcient matrix. Here, the ﬁnal model with ξ1,3 = 0.983, ξ2,2 = −64.6 and ξ2,3 = −3.2 would result from the most frequent coefﬁcient values, which is a good approximation of the true analytical model used to generate the data (ξ1,3 = 1, ξ2,2 = −64, ξ2,3 = −3.2). This procedure is different from the one proposed by Brunton et al. [15], where all data are stacked into a large data matrix and subjected to a single SINDy reconstruction. Due to the highly transient nature and resulting spectra of amplitudes of the data recorded in this work, we suggest deriving multiple models instead. 0.97 0.98 0.99 1 1.01 0 0.1 0.2 0.3 ξ1,3 probability −65.5 −65 −64.5 −64 −63.5 −63 0 0.1 0.2 0.3 ξ2,2 probability −3.21 −3.2 −3.19 −3.18 −3.17 0 0.1 0.2 0.3 0.4 0.5 ξ2,3 probability Figure 12. Model identiﬁcation from various measurements: Histograms of the non-zero coefﬁcient matrix entries resulting from 36 realizations of the same linear oscillator for all possible combinations of the initial conditions [x1(0), x2(0)] = [−1, −0.6, −0.2, 0.2, 0.6, 1] with highest polynomial order p = 1. For model coefﬁcient identiﬁcation, the most frequent or median coefﬁcient value may be selected from the histograms. The underlying analytical model is given by ξ1,3 = 1, ξ2,2 = −64, ξ2,3 = −3.2. −65.5 −65 −64.5 −64 −63.5 −63 0 0.1 0.2 0.3 ξ2,2 probability −3.21 −3.2 −3.19 −3.18 −3.17 0 0.1 0.2 0.3 0.4 0.5 ξ2,3 probability 0.97 0.98 0.99 1 1.01 0 0.1 0.2 0.3 ξ1,3 probability Figure 12. 3.3.1. Identiﬁcation of Parameter Dependencies In this study, the damping term δ is varied as a control parameter in the range of δ ∈[0.1, 0.3]. Optimized reconstruction models are created from each input time series and the non-zero coefﬁcient matrix entries are recorded. As the reference system conﬁguration, the standard value δ = 0.2 is selected. Figure 11 depicts the relative change ∆ξ = ξ/ξref of each coefﬁcient relative to the reference model. The two states x1, x2 are used with polynomial degree p = 1 such that six ODE coefﬁcients result. It can be observed that ﬁve out of those six coefﬁcients remain constant throughout the control parameter variation, while only ξ2,3 changes signiﬁcantly. Hence, the control parameter variation can be traced back to a single term in the reconstructed models. Consequently, this model parameter has to be studied in order to investigate control parameter changes in simulations. Of course, this scenario is well-constructed to illustrate a simple parameter dependency. However, the concept of tracking model parameters along with control parameters can be transferred directly to more complicated and more realistic parameter identiﬁcation challenges. Moreover, this procedure can be used in an inverse sense to identify the model parameters that are invariant under parameter changes. Vibration 2019, 2 40 0.1 0.15 0.2 0.25 0.3 −50 0 50 damping term δ rel. diﬀ. ∆ξ1,1 [%] 0.1 0.15 0.2 0.25 0.3 −50 0 50 damping term δ rel. diﬀ. ∆ξ1,2 [%] 0.1 0.15 0.2 0.25 0.3 −50 0 50 damping term δ rel. diﬀ. ∆ξ1,3 [%] 0.1 0.15 0.2 0.25 0.3 −50 0 50 damping term δ rel. diﬀ. ∆ξ2,1 [%] 0.1 0.15 0.2 0.25 0.3 −50 0 50 damping term δ rel. diﬀ. ∆ξ2,2 [%] 0.1 0.15 0.2 0.25 0.3 −50 0 50 damping term δ rel. diﬀ. ∆ξ2,3 [%] Figure 11. Relative deviation of the reconstructed ODE coefﬁcients ξi from the reference conﬁguration as function of the control parameter δ. As a reference, δ = 0.2 is chosen to identify those terms that are invariant under the parameter change. 0.1 0.15 0.2 0.25 0.3 −50 0 50 damping term δ rel. diﬀ. ∆ξ1,2 [%] 0.1 0.15 0.2 0.25 0.3 −50 0 50 damping term δ rel. diﬀ. ∆ξ1,3 [%] 0.25 0.3 m δ 0.1 0.15 0.2 0.25 0.3 −50 0 50 damping term δ rel. diﬀ. ∆ξ2,2 [%] 0.1 0.15 0.2 0.25 0.3 −50 0 50 damping term δ rel. diﬀ. 3.3.1. Identiﬁcation of Parameter Dependencies ∆ξ2,3 [%] Figure 11. Relative deviation of the reconstructed ODE coefﬁcients ξi from the reference conﬁguration as function of the control parameter δ. As a reference, δ = 0.2 is chosen to identify those terms that are invariant under the parameter change. 3.3.2. Robust Model Parameter Identiﬁcation Model identiﬁcation from various measurements: Histograms of the non-zero coefﬁcient matrix entries resulting from 36 realizations of the same linear oscillator for all possible combinations of the initial conditions [x1(0), x2(0)] = [−1, −0.6, −0.2, 0.2, 0.6, 1] with highest polynomial order p = 1. For model coefﬁcient identiﬁcation, the most frequent or median coefﬁcient value may be selected from the histograms. The underlying analytical model is given by ξ1,3 = 1, ξ2,2 = −64, ξ2,3 = −3.2. Vibration 2019, 2 Vibration 2019, 2 Vibration 2019, 2 41 3.3.3. Test for Nonlinearity 3.3.3. Test for Nonlinearity 0 1 2 3 4 −2 0 2 time x1 linear nonlinear 0 1 2 3 4 −10 0 10 time x2 linear nonlinear Figure 13. Comparison of linear and nonlinear 1-DOF oscillators in time domain (top) and in terms of the optimized SINDy reconstruction models (bottom). The analytical model parameters are indicated by the coefﬁcient values given in brackets in the table. Only non-vanishing terms are displayed here for a maximal polynomial order of p = 3. Figure 13. Comparison of linear and nonlinear 1-DOF oscillators in time domain (top) and in terms of the optimized SINDy reconstruction models (bottom). The analytical model parameters are indicated by the coefﬁcient values given in brackets in the table. Only non-vanishing terms are displayed here for a maximal polynomial order of p = 3. 3.3.3. Test for Nonlinearity In order to extend the ﬁndings from the given mechanical oscillator to nonlinear dynamics, we introduce an additional cubic stiffness term knl such that the governing equations read ˙x1 = x2, (9) ˙x2 = −cx2 −kx1 −knlx3 1. (9) ˙x1 = x2, ˙x2 = −cx2 −kx1 −knlx3 1. ˙x2 = −cx2 −kx1 −knlx3 1. The reference conﬁguration is given by k = 64, c = 3.2 and, in the nonlinear conﬁguration, knl = 10, if not denoted differently. Figure 13 displays the analytical states x1, x2 for both the linear as well as nonlinear conﬁguration of the 1-DOF oscillator. For small initial values, the nonlinearity is practically inactive and only minimal differences to the linear conﬁguration can be observed in the time domain. However, the SINDy reconstruction clearly identiﬁes a non-zero entry for the ansatz polynomial x3 1 and thus allows for testing for nonlinearity, even if the time traces are barely distinguishable in the time domain. However, while the cubic stiffness parameter is underestimated, the linear stiffness is overestimated. This result matches well with the understanding of linearization of the cubic stiffness for small displacement values. As the initial displacement is increased, the nonlinearity becomes active and visible in time domain and the SINDy model exhibits a cubic stiffness term that is closer to that of the analytical model. 0 1 2 3 4 −2 0 2 time x1 linear nonlinear 0 1 2 3 4 −10 0 10 time x2 1 x1 x2 . . . x3 1 . . . linear conﬁg. knl = 0, x1(0) = 1 ˙x1 0 (0) 0 (0) 0.98 (1) 0 0 (0) 0 ˙x2 0 (0) -64.60 (-64) -3.19 (-3.2) 0 0 (0) 0 nonlinear conﬁg. knl = 10, x1(0) = 1 ˙x1 0 (0) 0 (0) 0.98 (1) 0 0 (0) 0 ˙x2 0 (0) -65.32 (-64) -3.19 (-3.2) 0 -5.41 (-10) 0 nonlinear conﬁg. knl = 10, x1(0) = 2 ˙x1 0 (0) 0 (0) 0.98 (1) 0 0 (0) 0 ˙x2 0 (0) -65.31 (-64) -3.19 (-3.2) 0 -8.86 (-10) 0 Figure 13. Comparison of linear and nonlinear 1-DOF oscillators in time domain (top) and in terms of the optimized SINDy reconstruction models (bottom). The analytical model parameters are indicated by the coefﬁcient values given in brackets in the table. Only non-vanishing terms are displayed here for a maximal polynomial order of p = 3. 3.4. Feature Generation for Unsupervised Time Series Classiﬁcation Tasks Finally, we illustrate a sample case for using the SINDy reconstruction as a feature generator. In various ﬁelds of research, the discrimination of time series data forms the central challenge for both supervised as well as unsupervised classiﬁcation tasks. To mimic the situation of an unsupervised classiﬁcation task, we compute various ring-down time series data from varying model conﬁgurations. We then compute the reconstruction and employ the resulting ODE coefﬁcients as features for Vibration 2019, 2 42 describing the input data. These features are then used to solve the classiﬁcation task in this setting. Particularly, the linear stiffness is varied over a wide range of values k ∈[36, 400], while the nonlinear stiffness is kept constant for two different scenarios. First, the oscillator is considered in the linear conﬁguration, i.e., knl = 0. Then, a constant cubic stiffness of knl = 50 is considered. For each of those conﬁgurations, 40 time-series realizations (t ∈[0, 4] s) with varying linear stiffness terms are created. Then, optimized SINDy reconstruction models are computed for each input. The highest polynomial order is set to p = 3 and both states x1, x2 from the analytical model are used as input to the reconstruction. The ﬁrst epochs of all the 80 input time series of state x1 are depicted in Figure 14. While it may seem easy to distinguish some of the linear system responses from the nonlinear responses, most of the time traces overlap. Due to rather small amplitudes and strong damping, classical methods for describing and discriminating those dynamics may fail to cluster the data into two distinct groups representing the linear and nonlinear conﬁgurations of the underlying model, as shown in the previous discussion. In particular, the nonlinearity may be negligible for higher linear stiffness terms, which makes it nearly impossible to distinguish the system response from the raw time series data stemming from a purely linear system. We employ the proposed methods to derive optimized SINDy reconstructions and then collect the resulting ODE coefﬁcients Ξ into a matrix. In ﬁve cases, the SINDy ODE reconstruction was unstable, and was hence discarded. The resulting matrix has 75 rows, each corresponding to one reconstruction, and 20 columns, each of which result from the coefﬁcients ξ for the library of polynomials up to the third order for two states. 3.4. Feature Generation for Unsupervised Time Series Classiﬁcation Tasks For each observation, we treat the coefﬁcients as discriminative features for describing the input time series data. The 20-dimensional feature space is reduced to two dimensions by principle component analysis (PCA). Hence, for each observation, two generalized features are given by the projection of the features onto the ﬁrst and second principle components (PC)—see Figure 14. In this feature space, two distinct clusters form, which, since we know the label for each observation, can be traced back to the underlying dynamical systems. The features created by our methods exhibit variance in mainly one direction for the linear oscillator, which corresponds to the variation of the linear stiffness term. On the contrary, the features derived for the signals stemming from the nonlinear systems show additional variance in the second principle component, which makes them distinguishable from the linear system features. We note that the linear stiffness has been varied over one order of magnitude for both cases in order to complicate the classiﬁcation task. Still, our results illustrate that the unsupervised time series classiﬁcation task can be solved using the SINDy reconstruction models as feature generator in a highly automated fashion. The gaps in both clusters at PC1 ≈−250 stem from the ﬁve invalid SINDy reconstructions. Naturally, this approach can be employed equally for supervised classiﬁcation tasks, by assigning a label to an unknown input according to a-priori knowledge that was gathered from labeled training data. 0 0.2 0.4 0.6 0.8 −0.5 0 0.5 1 time [s] x1 linear nonlinear −400 −300 −200 −100 0 −60 −40 −20 0 20 principle component 1 principle component 2 linear system nonlinear system PCA (Ξ) (a) (b) Figure 14. (a) 40 time series realizations for various linear stiffness terms k ∈[36, 400] for the linear (knl = 0) and the nonlinear conﬁguration (knl = 50). The resulting SINDy coefﬁcient entries Ξ are used as descriptive features for classiﬁcation of the input data; (b) using the principle component analysis, the feature space is reduced to two dimensions to ﬁnd two clusters that can be traced back to the character of the underlying dynamical model that generated the input data. −400 −300 −200 −100 0 −60 −40 −20 0 20 principle component 1 principle component 2 linear system nonlinear system (Ξ) (b) 0 0.2 0.4 0.6 0.8 −0.5 0 0.5 1 time [s] x1 linear nonlinear P (a) principle component 1 Figure 14. 3.4. Feature Generation for Unsupervised Time Series Classiﬁcation Tasks (a) 40 time series realizations for various linear stiffness terms k ∈[36, 400] for the linear (knl = 0) and the nonlinear conﬁguration (knl = 50). The resulting SINDy coefﬁcient entries Ξ are used as descriptive features for classiﬁcation of the input data; (b) using the principle component analysis, the feature space is reduced to two dimensions to ﬁnd two clusters that can be traced back to the character of the underlying dynamical model that generated the input data. Vibration 2019, 2 Vibration 2019, 2 43 4. Conclusions In this work, we propose several extensions to the SINDy method, which allows for reconstructing differential equations from time series input data. The reconstruction is sparse in the space of possible candidate functions and thus represents a minimal model for the observed dynamics. The main contribution of this work comprises of a highly automated parameter selection approach and of a sophisticated optimization procedure to ﬁne-tune reconstructed models. Special focus is put on short and highly transient time series data as commonly obtained from impulse response and ring-down vibration measurements. Employing a 1-DOF oscillator, we illustrate the proposed extensions as well as common use-cases and challenges that may also arise and be observed in real-life laboratory data. These situations include limited data availability, noise contamination, and uncertain parameters of the dynamical system under study. The main ﬁndings and features of the proposed framework are summarized as follows: • Reconstruction of dynamic minimal models: The sparse regression reconstructs systems of differential equations from time series data. Hence, these equations can be studied and analyzed by classical methods and provide detailed insight into the governing dynamics underlying an observation. • Model reconstruction for limited input data: The proposed framework automates and optimizes the model reconstruction procedure while being suited well for accommodating limited data quality resulting from the amount of information, noise contamination, and unknown model dimensions. • Test for nonlinearity: The qualitative character of the underlying dynamical system can be estimated in terms of linearity and type and degree of nonlinearity by inspecting the set of reconstructed differential equations. • Model identiﬁcation and model updating methods: The optimized reconstruction allows for identiﬁcation of terms that depend explicitly on parameters that are prescribed or measured during experimentation. After identifying those terms in the reconstructed ODEs, uncertainty and bifurcation studies can be used in predictive modeling approaches to design safe and efﬁcient structures without extensive testing. • Time series feature generation for classiﬁcation and regression tasks: The reconstructed models represent features that are discriminative and possibly superior to classical time series features for uni-variate, short and highly transient input data. • Time series feature generation for classiﬁcation and regression tasks: The reconstructed models represent features that are discriminative and possibly superior to classical time series features for uni-variate, short and highly transient input data. Appendix A. The Filtering Property of TVRegDiff Numerical Derivation Schemes Appendix A. The Filtering Property of TVRegDiff Numerical Derivation Schemes Figure A1 illustrates the ﬁltering property of the TVRegDiff scheme with appropriate choice of the regularization parameter for different signal-to-noise ratios. (a) −2 0 2 s(t) SNR = 100, α =1 noisy sn clean sc ﬁltered sf −2 −1 0 1 2 sn(t) −sf(t) 0 2 4 6 8 10 12 14 16 18 20 10−6 10−3 100 time \|sc(t) −sf(t)\| (b) −2 0 2 s(t) SNR = 10, α =1 −2 −1 0 1 2 sn(t) −sf(t) 0 2 4 6 8 10 12 14 16 18 20 10−6 10−3 100 time \|sc(t) −sf(t)\| (c) −2 0 2 s(t) SNR = 5, α =1 −2 −1 0 1 2 sn(t) −sf(t) 0 2 4 6 8 10 12 14 16 18 20 10−6 10−3 100 time \|sc(t) −sf(t)\| (d) −2 0 2 s(t) SNR = 1, α =10 −2 −1 0 1 2 sn(t) −sf(t) 0 2 4 6 8 10 12 14 16 18 20 10−6 10−3 100 time \|sc(t) −sf(t)\| Figure A1. Filtering by integration of regularized derivatives: different noise levels applied to sinusoidal oscillation s(t) = sin(t). Clean signal is indicated by sc, noisy (signal-to-noise ratio SNR) signal by sn and ﬁltered signal, which is obtained from time integration of the numerical derivative, sf. Regularization parameters α were employed for computing derivatives using TVRegDiff, and SNR values decrease from (a) to (d) (a) −2 0 2 s(t) SNR = 100, α =1 noisy sn clean sc ﬁltered sf −2 −1 0 1 2 sn(t) −sf(t) (b) −2 0 2 s(t) SNR = 10, α =1 −2 −1 0 1 2 sn(t) −sf(t) 0 2 4 6 8 10 12 14 16 18 20 10−6 10−3 100 time \|sc(t) −sf(t)\| (b) (a) 0 2 4 6 8 10 12 14 16 18 20 10−6 10−3 100 time \|sc(t) −sf(t)\| time (d) −2 0 2 s(t) SNR = 1, α =10 −2 −1 0 1 2 sn(t) −sf(t) 0 2 4 6 8 10 12 14 16 18 20 10−6 10−3 100 time \|sc(t) −sf(t)\| (d) (c) −2 0 2 s(t) SNR 5, α 1 −2 −1 0 1 2 sn(t) −sf(t) 0 2 4 6 8 10 12 14 16 18 20 10−6 10−3 100 time \|sc(t) −sf(t)\| Figure A1. Filtering by integration of regularized derivatives: different noise levels applied to sinusoidal oscillation s(t) = sin(t). 4. Conclusions Future research will use the proposed framework of methods to identify dynamic minimal models and change the system underlying these minimal models taking real-life vibration measurements, such as hammer-impact testing during modal analysis of mechanical structures. Future research will use the proposed framework of methods to identify dynamic minimal models and change the system underlying these minimal models taking real-life vibration measurements, such as hammer-impact testing during modal analysis of mechanical structures. Author Contributions: Conceptualization, M.S., S.O. and N.H.; Software and Investigation, M.S.; Writing—Original Draft Preparation, M.S.; Writing—Review and Editing, S.O. and N.H.; Supervision and Project Administration, S.O. and N.H. Author Contributions: Conceptualization, M.S., S.O. and N.H.; Software and Investigation, M.S.; Writing—Original Draft Preparation, M.S.; Writing—Review and Editing, S.O. and N.H.; Supervision and Project Administration, S.O. and N.H. Funding: This research was funded by Deutsche Forschungsgemeinschaft (DFG) within the priority program ’calm, smooth, smart’ under grant HO 3852/21-1. M.S. acknowledges the ﬁnancial support S.O. over the UTS Centre for Audio, Acoustics and Vibration (CAAV) international visitor funds. Funding: This research was funded by Deutsche Forschungsgemeinschaft (DFG) within the priority program ’calm, smooth, smart’ under grant HO 3852/21-1. M.S. acknowledges the ﬁnancial support S.O. over the UTS Centre for Audio, Acoustics and Vibration (CAAV) international visitor funds. Acknowledgments: The Matlab code provided by Steven Brunton was used to generate the SINDy reconstruction. The authors would like to thank Edgar Hoover for his assistance in improving the clarity of the manuscript. Acknowledgments: The Matlab code provided by Steven Brunton was used to generate the SINDy reconstruction. The authors would like to thank Edgar Hoover for his assistance in improving the clarity of the manuscript. Conﬂicts of Interest: The 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. Conﬂicts of Interest: The 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. 44 44 Vibration 2019, 2 Vibration 2019, 2 Abbreviations The following abbreviations are used in this manuscript: DOF Degree Of Freedom NZE Non-Zero Entries ODE Ordinary Differential Equation SINDy Sparse Identiﬁcation of Nonlinear Dynamics TVRegDiff Total Variation Regularized Numerical Differentiation gDiff Total Variation Regularized Numerical Differentiation References Identiﬁcation of distributed parameter systems: A neural net based approach. Comput. Chem. Eng. 1998, 22, S965–S968. [CrossRef] 11. Raissi, M.; Karniadakis, G.E. Hidden physics models: Machine learning of nonlinear partial differential equations. J. Comput. Phys. 2018, 357, 125–141. [CrossRef] 12. Gouesbet, G.; Letellier, C. Global vector-ﬁeld reconstruction by using a multivariate polynomial L2 approximation on nets. Phys. Rev. E 1994, 49, 4955–4972. [CrossRef] 13. Liu, W.-D.; Ren, K.F.; Meunier-Guttin-Cluzel, S.; Gouesbet, G. Global vector-ﬁeld reconstruction of nonlinear dynamical systems from a time series with SVD method and validation with Lyapunov exponents. Chin. Phys. 2003, 12, 1366. 14. Wang, W.X.; Yang, R.; Lai, Y.C.; Kovanis, V.; Grebogi, C. Predicting catastrophes in nonlinear dynamical systems by compressive sensing. Phys. Rev. Lett. 2011, 106, 154101. [CrossRef] [PubMed] 15. Brunton, S.L.; Proctor, J.L.; Kutz, J.N. Discovering governing equations from data by sparse identiﬁcation of nonlinear dynamical systems. Proc. Natl. Acad. Sci. USA 2016, 113, 3932–3937. [CrossRef] [PubMed] 16. Rudy, S.H.; Brunton, S.L.; Proctor, J.L.; Kutz, J.N. Data-driven discovery of partial differential equations. Sci. Adv. 2017, 3, e1602614. [CrossRef] 17. Oberst, S.; Stender, M.; Baetz, J.; Campbell, G.; Lampe, F.; Morlock, M.; Lai, J.C.; Hoffmann, N. Extracting differential equations from measured vibro-acoustic impulse responses in cavity preparation of total hip arthroplasty. In Proceedings of the 15th Experimental Chaos and Complexity Conference, Madrid, Spain, 4–7 June 2018. 18. Takens, F. Detecting strange attractors in turbulence. In Dynamical Systems and Turbulence; Springer: Berlin/Heidelberg, Germany, 1981; pp. 366–381. 9. Kantz, H.; Schreiber, T. Nonlinear Time Series Analysis; Cambridge University Press: Cambridge, UK, 20 19. Kantz, H.; Schreiber, T. Nonlinear Time Series Analysis; Cambridge University Press: Cambridge, UK, 2003. 20. Oberst, S.; Lai, J.C. A statistical approach to estimate the Lyapunov spectrum in disc brake squeal. J. Sound Vib. 2015, 334, 120–135. [CrossRef] 20. Oberst, S.; Lai, J.C. A statistical approach to estimate the Lyapunov spectrum in disc brake squeal. J. Sound Vib. 2015, 334, 120–135. [CrossRef] 21. Abarbanel, H.D.I.; Brown, R.; Sidorowich, J.J.; Tsimring, L.S. The analysis of observed chaotic data in physical systems. Rev. Mod. Phys. 1993, 65, 1331–1392. [CrossRef] 22. Kennel, M.B.; Brown, R.; Abarbanel, H.D.I. Determining embedding dimension for phase-space reconstruction using a geometrical construction. Phys. Rev. A 1992, 45, 3403–3411. [CrossRef] 23. Oberst, S.; Baetz, J.; Campbell, G.; Lampe, F.; Lai, J.C.; Hoffmann, N.; Morlock, M. Vibro-acoustic and nonlinear analysis of cadavric femoral bone impaction in cavity preparations. Int. J. Mech. Sci. References 1. Ondra, V.; Sever, I.A.; Schwingshackl, C.W. A method for non-parameteric identiﬁcation of nonlinear vibration systems with asymmetric restoring forces from a resonant decay response. Mechan. Syst. Signal Process. 2019, 114, 239–258. [CrossRef] 1. Ondra, V.; Sever, I.A.; Schwingshackl, C.W. A method for non-parameteric identiﬁcation of nonlinear vibration systems with asymmetric restoring forces from a resonant decay response. Mechan. Syst. Signal Process. 2019, 114, 239–258. [CrossRef] 2. Ondra, V.; Sever, I.A.; Schwingshackl, C.W. A method for detection and characterisation of structural nonlinearities using the Hilbert transform. Mechan. Syst. Signal Process. 2017, 83, 210–227. [CrossRef] 2. Ondra, V.; Sever, I.A.; Schwingshackl, C.W. A method for detection and characterisation of structural nonlinearities using the Hilbert transform. Mechan. Syst. Signal Process. 2017, 83, 210–227. [CrossRef] 3. Pesaresi, L.; Stender, M.; Rufﬁni, V.; Schwingshackl, C.W. DIC Measurement of the Kinematics of a Friction Damper for Turbine Applications. In Dynamics of Coupled Structures, Volume 4, Conference Proceedings of the Society for Experimental Mechanics Series; Springer: Cham, Switzerland, 2017; pp. 93–101. 3. Pesaresi, L.; Stender, M.; Rufﬁni, V.; Schwingshackl, C.W. DIC Measurement of the Kinematics of a Friction Damper for Turbine Applications. In Dynamics of Coupled Structures, Volume 4, Conference Proceedings of the Society for Experimental Mechanics Series; Springer: Cham, Switzerland, 2017; pp. 93–101. . Kurt, M.; Chen, H.; Lee, Y.; McFarland, M.; Bergman, L.; Vakakis, A. Nonlinear system identiﬁcation of dynamics of a vibro-impact beam: Numerical results. Arch. Appl. Mech. 2012, 82, 1461–1479. [CrossRef 5. Farrar, C.R.; Worden, K. An introduction to structural health monitoring. Philos. Trans. R. Soc. A 2007, 365, 303–315. [CrossRef] [PubMed] 5. Farrar, C.R.; Worden, K. An introduction to structural health monitoring. Philos. Trans. R. Soc. A 2007, 365, 303–315. [CrossRef] [PubMed] 6. Nuawi, M.Z.; Bahari, A.R.; Abdullah, S.; Arifﬁn, A.K.; Nopiah, Z.M. Time Domain Analysis Method of the Impulse Vibro-acoustic Signal for Fatigue Strength Characterisation of Metallic Material. Procedia Eng. 2013, 66, 539–548. [CrossRef] 7. Crutchﬁeld, J.P.; McNamara, B.S. Equations of motion from a data series. Complex Syst. 1987, 1, 121. 8. Kurt, M.; Eriten, M.; McFarland, M.; Bergman, L.; Vakakis, A. Methodology for model updating of mechanical components with local nonlinearities. J. Sound Vib. 2015, 357, 331–348. [CrossRef] 9. Moore, K.; Kurt, M.; Eriten, M.; McFarland, M.; Bergman, L.; Vakakis, A. Direct detection of nonlinear modal interactions from time series measurements. Mechan. Syst. Signal Process. 2018. [CrossRef] 10. González-García, R.; Rico-Martínez, R.; Kevrekidis, I.G. Appendix A. The Filtering Property of TVRegDiff Numerical Derivation Schemes Clean signal is indicated by sc, noisy (signal-to-noise ratio SNR) signal by sn and ﬁltered signal, which is obtained from time integration of the numerical derivative, sf. Regularization parameters α were employed for computing derivatives using TVRegDiff, and SNR values decrease from (a) to (d) 45 Vibration 2019, 2 References 2018, 144, 739–745. [CrossRef] Vibration 2019, 2 46 24. Gilmore, R.; Lefranc, M.; Tuﬁllaro, N.B. The Topology of Chaos. Am. J. Phys. 2003, 71, 508–510. [CrossRef] 24. Gilmore, R.; Lefranc, M.; Tuﬁllaro, N.B. The Topology of Chaos. Am. J. Phys. 2003, 71, 508–510. [CrossRef] 25. Chartrand, R. Numerical Differentiation of Noisy, Nonsmooth Data. ISRN Appl. Math. 2011, 2011, 1–11. 25. Chartrand, R. Numerical Differentiation of Noisy, Nonsmooth Data. ISRN Appl. Math. 2011, 2011, 1–11. [CrossRef] 25. Chartrand, R. Numerical Differentiation of Noisy, Nonsmooth Data. ISRN Appl. Math. 2011, 2011, 1 11. [CrossRef] 26. Fulcher, B.D.; Jones, N.S. Highly Comparative Feature-Based Time-Series Classiﬁcation. IEEE Trans. Knowl. 6. Fulcher, B.D.; Jones, N.S. Highly Comparative Feature-Based Time-Series Classiﬁcation. IEEE Trans. Kn Data Eng. 2014, 26, 3026–3037. [CrossRef] 27. Bagnall, A.; Lines, J.; Hills, J.; Bostrom, A. Time-Series Classiﬁcation with COTE: The Collective of Transformation-Based Ensembles. IEEE Trans. Knowl. Data Eng. 2015, 27, 2522–2535. [CrossRef] 28. Fulcher, B.D.; Jones, N.S. hctsa: A Computational Framework for Automated Time-Series Phenotyping Using Massive Feature Extraction. Cell Syst. 2017, 5, 527–531.e3. [CrossRef] [PubMed] 29. Marwan, N.; Romano, M.C.; Thiel, M.; Kurths, J. Recurrence plots for the analysis of complex systems. Phys. Rep. 2007, 438, 237–329. [CrossRef] 30. Stender, M.; Papangelo, A.; Allen, M.; Brake, M.; Schwingshackl, C.; Tiedemann, M. Structural Design with Joints for Maximum Dissipation. In Shock & Vibration, Aircraft/Aerospace, Energy Harvesting, Acoustics & Optics; Springer: Cham, Switzerland, 2016; Volume 9, pp. 179–187. 31. Srivastava, N.; Hinto, G.; Krizhevsky, A.; Sutskever, I.; Salakhutdinov, R. Dropout: A simple way to prevent neural networks from overﬁtting. J. Mach. Learn. Res. 2014, 15, 1929–1958. c⃝2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W1974508979	https://europepmc.org/articles/pmc6271126?pdf=render	English	null	Diversity-Oriented Synthesis as a Tool for Chemical Genetics	Molecules/Molecules online/Molecules annual	2,014	cc-by	9,981	Elena Lenci, Antonio Guarna and Andrea Trabocchi * Department of Chemistry “Ugo Schiff”, University of Florence, via della Lastruccia 13, I-50019 Sesto Fiorentino, Florence, Italy * Author to whom correspondence should be addressed; E-Mail: andrea.trabocchi@unifi.it; Tel.: +39-055-457-3507; Fax: +39-055-457-4913. * Author to whom correspondence should be addressed; E-Mail: andrea.trabocchi@unifi.it; Tel.: +39-055-457-3507; Fax: +39-055-457-4913. Received: 22 August 2014; in revised form: 30 September 2014 / Accepted: 1 October 2014 / Published: 14 October 2014 Received: 22 August 2014; in revised form: 30 September 2014 / Accepted: 1 October 2014 / Published: 14 October 2014 Abstract: Chemical genetics is an approach for identifying small molecules with the ability to induce a biological phenotype or to interact with a particular gene product, and it is an emerging tool for lead generation in drug discovery. Accordingly, there is a need for efficient and versatile synthetic processes capable of generating complex and diverse molecular libraries, and Diversity-Oriented Synthesis (DOS) of small molecules is the concept of choice to give access to new chemotypes with high chemical diversity. In this review, the combination of chemical genetics and diversity-oriented synthesis to identify new chemotypes as hit compounds in chemical biology and drug discovery is reported, giving an overview of basic concepts and selected case studies. Keywords: diversity-oriented synthesis; chemical genetics; drug discovery; chemical biology; small molecules; high-throughput screening Molecules 2014, 19, 16506-16528; doi:10.3390/molecules191016506 Molecules 2014, 19, 16506-16528; doi:10.3390/molecules191016506 1. Introduction Drug discovery is a key field of pharmaceutical industries. After an impressive growth at the end of the last century, the number of new molecular entities launched on the market dramatically decreased in recent years, so that it has been claiming that the “ice age” of drug discovery is approaching [1]. Although target-based drug discovery approach remains the “gold standard” in hit identification, a fundamental challenge has emerged. Many disorders, such as cancer and neurodegenerative diseases, are often associated with complex interactions, such as those involving transcription factors, protein-protein interactions (PPIs) and DNA-protein interactions [2]. As a consequence of that, such Molecules 2014, 19 Molecules 2014, 19 16507 targets have been termed “undruggable”, due to the difficulty in being applied to typical drug-screening programs [3]. In this context, chemical genetics [4–6] is a brilliant example of a methodological development in lead generation. It uses small molecules to perturb the function of gene products, thus facilitating the dissection of biological processes. Similarly to genetics, chemical genetics can be divided into “forward” and “reverse” approach (Figure 1). In forward chemical genetics, a small molecule eliciting a desired phenotype is identified, and its protein partner is discovered subsequently following a deconvolution method. These studies (from molecule to protein to phenotype) are exploited when the aim of the investigation is the identification of molecules able to induce a specific biological effect. On the other hand, reverse chemical genetics approaches are fundamental to validate a known target. These studies involve functional small-molecule assays targeted directly to a protein of interest. Once a compound targeting a given protein is identified, the challenge is to discover if the small molecule has an effect in a cellular context. In fact, there is no guarantee that the compounds will affect a protein in a way that results in a functional phenotypic outcome in the cell. Indeed, targeting a specific protein may not result in giving the desired therapeutic consequence, and a considerable effort is paid to validate the target prior to executing full-scale screening. Nevertheless, forward chemical genetics is the most attractive and the most used approach in chemical biology, because it allows for the discovery of both new targets and new lead compounds with potential therapeutic applications [7–9]. Figure 1. (a) Forward chemical genetics approach; (b) Reverse chemical genetics approach. 1.1. 1. Introduction Plasmodium Falciparum: An Example of the Application of Both Forward and Reverse Chemical Genetics Studies (a) (b) (a) (a) (b) 1.1. Plasmodium Falciparum: An Example of the Application of Both Forward and Reverse Chemical Genetics Studies The application of chemical genetics in drug discovery programs was demonstrated being particularly powerful in the search for new antimalarials [10–13]. Guy and co-workers have done a significant contribution for new chemotypes discovery in this area, using both forward and reverse chemical genetics approaches (Figure 2) [14]. The authors screened a library of 309474 compounds (including drugs, enzyme inhibitors, natural products, etc.) against Plasmodium falciparum strains, according to a forward chemical genetics approach. This primary screen revealed 1130 compounds able to induce more 16508 Molecules 2014, 19 than 80% of growth inhibition. From this set, 228 small molecules were re-purchased and re-tested by three independent laboratories, providing 172 cross-validated hits. than 80% of growth inhibition. From this set, 228 small molecules were re-purchased and re-tested by three independent laboratories, providing 172 cross-validated hits. Figure 2. Illustration of forward and chemical genetics approaches used by Guy and co-workers for malaria drug discovery [14]. Figure 2. Illustration of forward and chemical genetics approaches used by Guy and co-workers for malaria drug discovery [14]. Figure 2. Illustration of forward and chemical genetics approaches used by Guy and co-workers for malaria drug discovery [14]. These active compounds were then subjected to a reverse chemical genetics study, using 66 potential antimalarial targets. In particular, the authors investigated the inhibition activity of these compounds against dihydroorotate dehydrogenase (Pf-DHOD) [15], falcipain-2 cystein protease (Pf-FP2) [16] and dihydrofolate reductase (Pf-DHFR). Additionally, they evaluated heme detoxification and hemozoin formation inhibition, using heme crystallization assay [17,18]. To expand the pool of potential targets, these 172 compounds were then studied for binding to 61 recombinant malarial proteins, using thermal-melt shift assays. In this way, 19 new inhibitors of malaria drug targets were discovered. Molecules 2014, 19 Molecules 2014, 19 Molecules 2014, 19 16509 In this context, Diversity-Oriented Synthesis (DOS) [25,26], which aims to synthesize the largest number of structurally complex small molecules, was conceived as a novel concept for the construction of libraries mainly addressing drug discovery issues. DOS has been defined by Spring as “the deliberate, simultaneous and efficient synthesis of more than one target compound in a diversity-driven approach” [27]. Since Schreiber’s seminal paper [28], several DOS strategies have been proposed in order to achieve three different types of molecular diversity, namely appendage, stereochemical and skeletal diversity [29,30]. Such strategies are usually divided in: (1) reagent-based approaches, where many distinct compounds are obtained subjecting a starting molecule to different reaction conditions; and (2) substrate-based approaches, where different starting materials, containing pre-encoded skeletal information, are transformed to distinct products using the same reaction conditions. Additionally, a very useful strategy, proposed by Schreiber and Nielsen, is the build/couple/pair approach [31]. Building blocks are assembled in polyfunctional intermediates, which can undergo different intramolecular cyclizations, thus achieving complex scaffolds in a divergent fashion. This approach exploits efficient and modular synthetic routes, typically composed by no more than five steps, so once interesting activities of some compounds are ascertained, focused libraries around such hit structures can be conveniently generated in a follow-up process. Nevertheless, the chemical genetics approach combined with DOS synthetic strategy is currently somewhat limited. The emergence of international screening initiatives, such as The Society for Laboratory Automation and Screening (SLAS), EU-OPENSCREEN or ChemBioNet, has increased significantly the number of high-throughput screening studies conducted on DOS libraries. Most drug discovery initiatives rely, however, on screening compound collections for their activity against known biological targets [32]. For example, Schreiber and co-workers discovered a potent sonic hedgehog inhibitor with a KD value of 3.1 µM, robotnikinin (Figure 3a) from the screening of a DOS library of 2070 amino alcohol-derived macrocycles with the Small Molecule Microarray (SMM) technology [33]. The Spring’s group, instead, screened a DOS library of 223 compounds, based on 30 distinct molecular scaffolds, using high-throughput phenotypic assays, thus identifying emmacin (Figure 3b) as a potent antibacterial compound showing growth inhibitory activity against epidemic methicillin-resistant strains of S. aureus (EMRSA-15 and -16) with MIC50 values of 9 µg/mL [34,35]. Figure 3. 1.2. Generating Libraries for Chemical Genetics Studies. The Importance of Diversity-Oriented Synthesis 1.2. Generating Libraries for Chemical Genetics Studies. The Importance of Diversity-Oriented Synthesis There is no technology enabling a chemical genetics approach to be applied generally. A chemical genetics study requires: (1) the design and synthesis of a chemical library; (2) the screening of the library in the system of choice; (3) target identification/validation. Thus, in the first step of the chemical-genetics process, it is necessary to assemble a library of different and potentially bioactive small molecules [19]. During the last decades, organic chemists have taken advantage of a significant number of high-throughput synthesis methods, such as solid-phase techniques [20,21] and combinatorial chemistry [22,23]. Unfortunately, despite the success in several drug discovery programs, the combinatorial chemistry approach has not fulfilled the desired expectations in difficult-to-target areas, such as those addressing protein-protein and DNA-protein interactions. Most of the combinatorial libraries are prepared through the functionalization of a common skeleton, so the compounds therein generated, often possessing flat structures and few stereocenters, showed limited structural complexity and diversity. The structural features common to diverse classes of bioactive natural products confirm that three-dimensional complexity is necessary for interacting with biological macromolecules. Furthermore, libraries that interrogate larger areas of chemical space increase the chance of identifying novel lead compounds [24]. Molecules 2014, 19 16510 On the contrary, there are only few examples of reverse and forward chemical genetics programs based on DOS libraries. In order to make new chemical genetics approaches time- and money-saving a double effort is required: the set-up of an easy, rapid and versatile chemical synthesis and the development of a rapid, efficient and economical screening process. These tools are used mostly for drug development, rather than on the dissection of biological systems, especially in pharmaceutical industries. However, these technologies have become sufficiently robust and wide-spread, that we can now conceive their use to study virtually any biological process. Dissecting biological systems with small molecule probes became increasingly popular, as it holds the promise of discovering new targets outside the “druggable” genome. In the present work, we describe three selected case studies that illustrate the power of diversity-oriented synthesis as a tool for chemical genetics, focusing on the challenges of designing and manufacturing compound libraries. Molecules 2014, 19 (a) Robotnikinin, a low µM sonic hedgehog inhibitor discovered by Schreiber’s group [33]; (b) Emmacin, a potent antibacterial compound with MIC50 on the low µg/mL range discovered by Spring’s group [34,35]. HN O O N H O O Robotnikinin (a) Emmacin (b) Cl Cl OH N H N EtO O NH2 Cl HN O O N H O O Robotnikinin (a) Cl (b) (a) Emmacin (b) Cl Cl OH N H N EtO O NH2 Robotnikinin Emmacin 2. Case Study 1: Pioneering Work of Stuart L. Schreiber The pioneering work of Schreiber, published in 1999, still remains one of the most important examples of a striking DOS synthesis for chemical genetics and cell-based assays [36,37]. Starting from the simple shikimic acid derivative 1, Schreiber and co-workers were able to generate the tetracyclic natural product-like scaffold 2, using a tandem acylation/1,3-dipolar cycloaddition with an array of nitrone-carboxylic acids [38] (Scheme 1). Scheme 1. Synthesis of the tetracyclic scaffold 2 with tandem acylation/1,3-dipolar cycloaddition [36]. As shown in Scheme 2, the tetracyclic template 2 is a rigid and densely functionalized compound, which can be decorated with a variety of appendages. The introduction of benzyl iodide substituents on the isoxazoline nitrogen allows for the application of a wide number of different palladium-catalyzed reactions (e.g., cross coupling, amination, etherification, carbonylation), even though the authors demonstrated that only Suzuki [39], Stille [40] and Sonogashira/Castro-Stephens [41,42] reactions were successfully performed on template 3 (Scheme 2a). Additionally, the electrophilic lactone and epoxide moieties can react with nucleophiles like amines and nitriles (Scheme 2b,c), unmasking alcohol functionalities suitable for further reactions with acid chlorides, anhydrides, sulfonyl chlorides, chloroformates, carbamoyl chlorides and isocyanates (Scheme 2d). Finally, reductive N-O bond cleavage provides two additional orthogonal sites for functionalization, although this opportunity was not exploited. Molecules 2014, 19 16511 Scheme 2. Potential sites of functionalization and representative examples of compounds derived from template 3 by (a) cross-coupling reactions, (b) aminolysis of lactone moiety, (c) epoxide ring opening and (d) alcohol esterification [36,37]. Scheme 2. Potential sites of functionalization and representative examples of compounds derived from template 3 by (a) cross-coupling reactions, (b) aminolysis of lactone moiety, (c) epoxide ring opening and (d) alcohol esterification [36,37]. Starting from these possibilities, the authors chose the most efficient reactions compatible with solid-phase split-and-pool technique as a synthetic strategy of choice to provide the greatest number of small molecules as possible [43,44]. According to this technique, the starting materials are linked to the solid support and divided into different batches (split); each batch is allowed reacting with the subsequent building blocks and the products are mixed again (pool). In this way, the number of products increases exponentially, affording thousands of structures, as required for the successful outcome in chemical genetics screening programs. 2. Case Study 1: Pioneering Work of Stuart L. Schreiber Although this technique presented some difficulties in purification processes and in developing a validation protocol, Schreiber and co-workers were able to construct a DOS library of more than 2 million distinct, spatially separated, chemical entities [36,37]. As shown in Scheme 3, the resin was split into three portions and labelled with three different spacers (ω-aminocaproic acid, glycine and “no spacer”). The functionalized resin 4 was pooled, mixed and split in two equal portions. Each enantiomer of epoxycyclohexenol was coupled to each portion, thus resulting in six structures for 1. The inclusion of o-, m-, p-iodobenzyl nitrone carboxylic acids led to 18 tetracyclic templates for 3. Then, a Sonogashira/Castro-Stephens coupling reaction employing 30 different commercially available alkynes, afforded 558 compounds for 5. Lactone aminolysis with 62 different 16512 Molecules 2014, 19 amines resulted in 35,154 structures for 6, and finally, alcohol acylation with 62 different carboxylic acids led to a DOS library of 2,180,106 distinct molecules of general formula 7. Scheme 3. Split-and-pool DOS synthesis of the library based on template 3 [37]. Si O i-Pr i-Pr X N O Ph Ph H O HN I O O O (1) (2) (3) R1 R4 NCO template 8 Si O i-Pr i-Pr X N O Ph Ph H O N O NR2(R3) O R2 H N R3 O R4 R1 3520 Library Members Chemical genetic screening N O Ph Ph H O N O O 3-C Williams Reaction HN N O OH O O MeO 9 Enhance effect of Latrunculin B on yeast growth with EC50 = 550 nM step (c) step (b) step (a) Scaffold 8, similarly to the previous template 3, is a densely functionalized structure that allows f e application of several approaches for appendage’s decorations. The authors obtained a library Si O i-Pr i-Pr X H O HN O O Ph Ph N H O I O O Mg(ClO4)2 pyridine HC(OMe)3 toluene, r.t. Si O i-Pr i-Pr X N O Ph Ph H O HN I O O O template 8 3-C Williams Reaction step (a) Si O i-Pr i-Pr X H O HN O O Ph Ph N H O I O O Ph template 8 (1) (2) (3) R1 R4 NCO R2 H N R3 step (b) (2) (3) R4 NCO Si O i-Pr i-Pr X N O Ph Ph H O N O NR2(R3) O R2 H N R3 O R4 R1 3520 Library Members Chemical genetic screening N O Ph Ph H O N O O HN N O OH O O MeO 9 Enhance effect of Latrunculin B on yeast growth with EC50 = 550 nM step (c) step (b) step (c) Chemical genetic screening Enhance effect of Latrunculin B on yeast growth with EC50 = 550 nM Enhance effect of Latrunculin B on yeast growth with EC50 = 550 nM Scaffold 8, similarly to the previous template 3, is a densely functionalized structure that allows for the application of several approaches for appendage’s decorations. The authors obtained a library of 3520 members using palladium-catalyzed coupling reactions (1), amide bond formations on the unmasked alcohol (2), and N-acylations of the γ-lactam (3) (Scheme 4b). A cell-based screening was then performed on this compounds collection to identify enhancers of latrunculin B, an actin polymerization inhibitor, which induces yeast growth arrest [52]. Scheme 3. Split-and-pool DOS synthesis of the library based on template 3 [37]. Through follow-up synthesis, the hit compound 9 as in Scheme 4, (step c) was confirmed and found to be a latrunculin B enhancer with an EC50 value in the sub-micromolar range. Scheme 3. Split-and-pool DOS synthesis of the library based on template 3 [37]. Scheme 3. Split-and-pool DOS synthesis of the library based on template 3 [37]. Scheme 3. Split-and-pool DOS synthesis of the library based on template 3 [37]. The solid-phase technique can be arranged efficiently with cytoblot assays [45] and small molecule The solid-phase technique can be arranged efficiently with cytoblot assays [45] and small molecule microarray systems [46,47] for forward and reverse chemical genetics studies, respectively. In order to take full advantage of this feature, assay formats must be developed using a controlled release from the individual supports into distinct plate wells. For example, the authors developed instrumentations and robotics to array efficiently synthesis beads into distinct wells containing mink lung cells. With this miniaturized cytoblot assay, they found that all the library compounds showed a significant inhibitory effect on cell proliferation [37]. Additionally, in a preliminary study, they discovered that several compounds activate a TGF-β (Transforming growth factor β) responsive reporter gene in mammalian cells [48]. TGF-β induces cell cycle arrest in G1, so new compounds targeting TGF-β signaling components may be considered as hit compounds for the discovery of anti-cancer or anti-angiogenic agents. In a further study, Schreiber and co-workers repeated the same synthetic strategy [49]. They reported the split-and-pool synthesis of more than 3000 spirooxindol compounds, exploiting a three-component Williams’ coupling reaction [50,51] to assemble the central core 8 (Scheme 4, step a). Molecules 2014, 19 165 Molecules 2014, 19 16513 Scheme 4. (a) Stereoselective synthesis of the spirooxindole scaffold 8; (b) Scaffold decoration by cross-coupling reactions (1), amide formations (2), N-acylations (3); (c) From this library, a potent enhancer of latrunculin B was discovered [49]. Scheme 4. (a) Stereoselective synthesis of the spirooxindole scaffold 8; (b) Scaffold decoration by cross-coupling reactions (1), amide formations (2), N-acylations (3); (c) From this library, a potent enhancer of latrunculin B was discovered [49]. Si O i-Pr i-Pr X H O HN O O Ph Ph N H O I O O Mg(ClO4)2 pyridine HC(OMe)3 toluene, r.t. Si O i-Pr i-Pr X N O Ph Ph H O HN I O O O template 8 3-C Williams Reaction step (a) Si O i-Pr i-Pr X H O HN O O Ph Ph N H O I O O Mg(ClO4)2 pyridine HC(OMe)3 toluene, r.t. 3. Case Study 2: Contribution of David R. Spring to the Discovery of Antimitotic Agents through Chemical Genetics Screening of DOS Libraries Recently, the group of David Spring reported a striking diversity-oriented synthesis combined to a forward chemical genetics study [53]. In this work, they obtained a DOS library of 35 compounds based on 10 distinct molecular scaffolds exploiting rhodium carbenoid chemistry. As shown in Scheme 5a, Molecules 2014, 19 16514 phenyldiazoester 10 reacts with terminal alkynes and alkenes providing cyclopropene (12) and cyclopropane structures (13), respectively, which can be further functionalized into scaffold 14 and scaffold 15. The styril diazoester derivative 11 (Scheme 5b) revealed to be more interesting to access diversity in the DOS library through a tandem cyclopropanation-Cope rearrangement reaction [54], thus providing the multivalent bicyclo[3.2.1]octadiene 16. In fact, the presence of both an electron-deficient and an electron-neutral double bond moiety, together with the proximal aryl bromide and a carboxylic acid ester, gave the opportunity to modify regioselectively the scaffold in a multidirectional divergent approach (Scheme 6). Scheme 5. Scaffold generated from phenyldiazoester compounds 10 (a) and 11 (b) [53]. Scheme 5. Scaffold generated from phenyldiazoester compounds 10 (a) and 11 (b) [53]. Scheme 5. Scaffold generated from phenyldiazoester compounds 10 (a) and 11 (b) [53]. N2 O O 10 • Ph MeO2C Ph CO2Me R MeO2C Ph 12 14 NO2 R 13 Heck-type cross- coupling with p-nitroiodobenzene 1) iodolactonization 2) radical deiodination O O R H 15 (a) (b) 11 Br N2 O O Cyclopropanation Cope Rearrangement with cyclopentadiene Br CO2Me 16 (b) 11 Br N2 O O N2 O O 10 R (a) (a) (b) 10 11 1 • CO2Me R R 13 R Ph MeO2C Ph 12 Cyclopropanation Cope Rearrangement with cyclopentadiene 12 13 s- ene Br CO2Me 16 Heck-type cross- coupling with p-nitroiodobenzene 1) iodolactonization 2) radical deiodination MeO2C Ph 14 NO2 O O R H 15 MeO2C Ph 14 NO2 16 14 15 For example, compound 16 was converted to epoxide 17 by a stereoselective olefin epoxidation (Scheme 6a), whereas stereoselective dihydroxylation of the electron-neutral double bond led to cis-diol 18, which can be further transformed to acetals or sulfoxides (Scheme 6b). Additionally, the electron-neutral alkene moiety in 16 was exploited for ring opening metathesis with terminal olefins, affording cyclohexene structures such as compound 19 (Scheme 6c). The application of 16 to one-pot dihydroxylation/oxidative cleavage methodology [55], followed by reductive amination conditions of the dialdehyde intermediate, resulted to a new ring system. 3. Case Study 2: Contribution of David R. Spring to the Discovery of Antimitotic Agents through Chemical Genetics Screening of DOS Libraries In particular, when 16 was left reacting with dimethylamine (Scheme 6d), the cyclohexene scaffold 20 was synthesized, whereas the use of primary amines resulted in a 6-6 fused ring system (compound 21) as a consequence of double reductive amination (Scheme 6e). Additionally, the dialdehyde intermediate was selectively reduced to diol and cyclized under transesterification reaction conditions, thus obtaining the lactone 22 in just one synthetic step (Scheme 6f). Finally, the ortho-bromoaryl substituent of 16 provided a functional handle for Suzuki Molecules 2014, 19 16515 cross-coupling reactions (Scheme 6g), introducing aryl substituents around the core scaffold (compounds 23 and 24). Scheme 6. Library synthesis from the key intermediate 16. The exploration of the chemical space around scaffold 16 was achieved using (a) epoxidation, (b) dihydroxylation, (c) ring opening metathesis, (d) dihydroxylation/oxidative cleavage + reductive amination with dimethylamine, (e) dihydroxylation/oxidative cleavage + double reductive amination with primary amines, (f) dihydroxylation/oxidative cleavage + esterification, (g) Suzuki reactions [53]. p y ( ) y y g (g) reactions [53]. A very useful tool for describing quantitatively the molecular diversity of DOS library compounds in the chemical space is the application of chemoinformatics [56–58]. Chemical space, as described by Dobson, is “the total descriptor space that encompasses all the small carbon-based molecules that could in principle be created” [59]. To computationally assess the structural diversity of a library and its ability to interrogate the chemical space, two comparative statistical approaches are typically employed: the principal component analysis (PCA) [60–62] and the principal moment of inertia (PMI) analysis [63,64]. PCA utilizes a defined number of descriptors (such as molecular weight, logP values, number of H-bond donors or acceptors) to represent each molecule as two-dimensional vectors, which can be plotted to give a graphic representation of the library diversity. Spring and co-workers analysed 15 physicochemical properties of the compounds by comparing them with two reference compound A very useful tool for describing quantitatively the molecular diversity of DOS library compounds in the chemical space is the application of chemoinformatics [56–58]. Chemical space, as described by Dobson, is “the total descriptor space that encompasses all the small carbon-based molecules that could in principle be created” [59]. To computationally assess the structural diversity of a library and its ability to interrogate the chemical space, two comparative statistical approaches are typically employed: the principal component analysis (PCA) [60–62] and the principal moment of inertia (PMI) analysis [63,64]. 3. Case Study 2: Contribution of David R. Spring to the Discovery of Antimitotic Agents through Chemical Genetics Screening of DOS Libraries PCA utilizes a defined number of descriptors (such as molecular weight, logP values, number of H-bond donors or acceptors) to represent each molecule as two-dimensional vectors, which can be plotted to give a graphic representation of the library diversity. Spring and co-workers analysed 15 physicochemical properties of the compounds by comparing them with two reference compound Molecules 2014, 19 16516 collections, a pool of 40 top-selling brand-name drugs and a pool of 60 bioactive natural products [62]. PCA graphs (Figure 4a–c) showed that the DOS library overlaps considerably with the chemical space covered by the top-selling drugs, indicating the potential drug-likeness of these scaffolds. Figure 4. (a–c) Comparative PCA plots of DOS library compounds (red, circles) vs. top-selling brand-name drugs (purple, squares) and natural products (blue, triangles); (d) PMI plot illustrating the molecular shape diversity of the DOS library (red, circles) and lowest-energy conformation of three representative DOS library members (15, 19, 21). Adapted by permission from Macmillan Publishers Ltd: ref. 53, copyright (2014). Normalized ratios of principal moments of inertia (PMI) represent an intuitive way to descri ecular shape. The authors plotted the PMI normalized ratio of the minimum energy conformation h library member into a triangular graph. The obtained PMI graph, with its vertices representing t e extremes of molecular 3D geometry (rod, disc, sphere), showed that the DOS library has a ve d level of shape diversity (Figure 4d). Additionally, DOS compounds tend towards the spheric ner more than drugs, suggesting that their scaffolds are more three-dimensional in character th (a) (b) (c) (d) (b) (a) (c) (d) ) Normalized ratios of principal moments of inertia (PMI) represent an intuitive way to describe molecular shape. The authors plotted the PMI normalized ratio of the minimum energy conformation of each library member into a triangular graph. The obtained PMI graph, with its vertices representing the three extremes of molecular 3D geometry (rod, disc, sphere), showed that the DOS library has a very good level of shape diversity (Figure 4d). Additionally, DOS compounds tend towards the spherical corner more than drugs, suggesting that their scaffolds are more three-dimensional in character than drugs. This is a very interesting feature, because an increase in the three-dimensional and stereochemical Normalized ratios of principal moments of inertia (PMI) represent an intuitive way to describe molecular shape. 3. Case Study 2: Contribution of David R. Spring to the Discovery of Antimitotic Agents through Chemical Genetics Screening of DOS Libraries The authors plotted the PMI normalized ratio of the minimum energy conformation of each library member into a triangular graph. The obtained PMI graph, with its vertices representing the three extremes of molecular 3D geometry (rod, disc, sphere), showed that the DOS library has a very good level of shape diversity (Figure 4d). Additionally, DOS compounds tend towards the spherical corner more than drugs, suggesting that their scaffolds are more three-dimensional in character than drugs. This is a very interesting feature, because an increase in the three-dimensional and stereochemical Molecules 2014, 19 16517 complexity enhances the chance of modulating challenging targets, such as protein-protein and protein-DNA interactions [65]. The DOS library was then assayed in a microscopy-based phenotypic screening [66] for mitotic arrest, thus finding that compounds 23 and 24 induce a marked cell arrest in the mitosis. Starting from these results, the authors synthesized a partially saturated analogue of compound 23, named dosabulin (Figure 5), discovering that all the activity was located in the (S)-enantiomer. Furthermore, according to a forward chemical genetics approach, the authors tried to identify the target responsible for the mitotic arrest induced by (S)-dosabulin. Since the tubulin network was demolished by treatment with (S)-dosabulin, they suggested that dosabulin targets tubulin itself, as several antimitotic drugs do [67]. Figure 5. The antimitotic compound, (S)-Dosabulin, discovered with forward chemical genetics approach [53]. This work revealed an additional advantage of the forward chemical genetics approach. In fact, it is often possible to identify compounds that perturb other cell functions, as compared to the phenotype of study, and in this case the authors identified two other compounds, based on the scaffold 21 (Scheme 6), able to decrease cell viability, without inducing a mitotic arrest. Their identification is a further demonstration of the ability of DOS to produce useful bioactive scaffolds with potential therapeutic applications. 4. Case Study 3: Our Contribution to the Chemical Genetics Screening of Peptidomimetic DOS Libraries towards Yeast Deletant Strains From the current literature, it’s worth noting that the rate-limiting step of DOS-chemical genetics studies is the development of an easy and economical screening process rather than the design of a DOS library [68]. In this context, our group, in collaboration with the research group of Duccio Cavalieri, found that a useful tool for forward chemical genetics programs is the use of Saccharomyces cerevisiae [69]. Thanks to its easy manipulation and its rapid life cycle, this model organism can be successfully used to identify novel compounds inducing cell growth inhibition [70]. Furthermore, in addition to the high degree of conservation with mammalian cells, S. cerevisiae is particularly suited for dissecting biological pathways and for identifying novel targets. Once active compounds have been selected, the screen of yeast mutant strains, showing hypersensitivity or hyper-resistance to each molecule, allows for the identification of the target responsible for the observed phenotype [71–74]. Specifically, the deletant strain that is not perturbed by the presence of such a compound is the one that lacks the gene encoding for the target of the molecule. Using this approach, together with Cavalieri’s group, we developed forward chemical genetics methods using DOS libraries of morpholine-based peptidomimetics [75–77]. A peptidomimetic 16518 Molecules 2014, 19 compound may be defined as a molecule having a secondary structure similar to the parent peptide, such that it binds to enzymes or receptors with higher affinity than the starting peptide. These compounds, able to perturb protein-protein interactions, have attracted wide interest in chemical genetics studies for discovering protein function and identifying novel ligands [78,79]. In this context, our group reported several efficient synthetic strategies for the production of peptidomimetic templates characterized by polyfunctional and heterocyclic structures, starting from amino acid and sugar derivatives [80–83]. 4.1. Chemical Genetics of a Bicyclic Peptidomimetic DOS Library In a first work [75], we developed a chemical genetics study based on a library of bicyclic peptidomimetics characterized by a 6,8-dioxa-3-azabicyclo[3.2.1]octane core. This collection of compounds was generated, as shown in Figure 6a, through two key steps consisting of a coupling reaction between two building blocks followed by an intramolecular cyclization [80,81]. By tuning the starting materials it was possible to generate a large array of different scaffolds (Figure 6b). These templates show structural similarity to dipeptides through an atom-by-atom correlation, in fact the side chain can be placed in the same position as found in native dipeptide sequences (Figure 6c). Additionally these compounds satisfy all the requirements needed for the peptidomimetic chemistry, as they are easily synthesized and decorated with functional groups, and are well-suited for solid-phase synthesis. These scaffolds revealed to be active hit compounds towards different targets, as they have been used as aspartyl protease inhibitors [84,85] and NGF-agonists [86]. Figure 6. (a) Strategic approaches for the synthesis of peptidomimetic scaffolds starting from sugar and amino acids derivatives; (b) Representative examples of bicyclic rigid scaffolds obtained with this strategy; (c) 6,8-dioxa-3-azabicyclo[3.2.1]octane scaffold as a constrained dipeptide isostere [80–83]. (a) (b) (c) (a) In a preliminary chemical genetics study, a library composed of 140 bicyclic peptidomimetics (Figure 7) was tested on a panel of S. cerevisiae wild-type strains, using cell growth as the phenotype of study. Sixteen compounds resulted in a decrease of more than 10% of the growth rate. In order to gain insight 16519 Molecules 2014, 19 into their mode of action, these selected compounds were re-tested on a panel of mutant strains, harbouring a HAP1 gene deletion that resulted in limited respiratory ability [87]. This process enabled the selection of compound 25 (Figure 7), which was identified as the molecule inducing the most intense cell growth decrease. This peptidomimetic was more active on the wild-type strain than on the HAP1-deleted one, thus giving an indication of the involvement of the respiratory metabolism in response to perturbation operated by compound 25. Figure 7. From a library 140 bicyclic peptidomimetics, a forward chemical genetics study allowed for the identification of 25 as a hit antifungal compound [75]. The pathway signature-based approach [88,89], which compares the transcriptional profile of wild-type and mutant cells, has proven to be a very useful method to validate drug targets and to identify secondary effects. 4.1. Chemical Genetics of a Bicyclic Peptidomimetic DOS Library According to this approach, the pathway analysis of gene expression levels of mutant and wild-type yeasts confirmed HAP1 as one of the targets of compound 25. Furthermore, the use of pool of strains bearing deletion of cell wall and multidrug resistance genes (ERG6, SNQ2 and PDR3) indicated that 25 induces other biological effects, such as the generation of an important stress condition. Additionally, these data enabled the localization of the first molecular target of compound 25 at the cell wall level. Since cell wall is a feature of yeast, compound 25 can be considered a potential selective antifungal agent. 4.2. Chemical Genetics of Morpholine-Based DOS Library In a further chemical genetics work [76], we generated a DOS library of morpholine-based compounds to be screened towards yeast strains. According to the build/could/pair approach, we developed a DOS strategy for the generation of morpholine scaffolds, taking advantage of a two-steps process. The strategic coupling of the four derivatives A–D as shown in Scheme 7, resulted to a highly functionalized acyclic precursor. This intermediate revealed to be very useful for accessing diversity around the morpholine nucleus. In fact, the selection of building blocks and the choice of cyclization conditions, allowed for obtaining several skeletally different scaffolds [90]. For example, compounds 26 and 27 gave bicylic lactones 31 and 32, respectively, upon treatment with a methanolic SOCl2 solution. Instead, compound 28, as a consequence of steric bias, could not cyclize under the same reaction conditions, and gave the polyfunctional morpholine scaffold 33. Furthermore, starting from compounds 29 and 30, containing the protected amino group, oxazine 34 and the bicylic scaffold 35 were obtained (Scheme 7). Molecules 2014, 19 16521 The effects of these 48 library members were tested on S. Cerevisiae wild-type strain. In order to improve the profile of phenotype screening, yeast cell growth was evaluated in both exponential and stationary phases. This screening allowed for the selection of 21 molecules inducing a variation in the O.D.650 value, which is correlated to the number of cells in the stationary phase, and/or in GenT, which is the time responsible for cell division in the initial phase. Such compounds were thus tested on both wild-type (WT) and mutant strains bearing deficiencies in genes involved in cell wall remodelling (∆erg6) and multidrug resistance (∆snq2 and ∆pdr3), to gain insight into their mode of action, as described above (Figure 8). Figure 8. Bidimensional scatter plots showing effects of library members on cell growth of selected deletant strains and structures of compounds 36 and 37 [76]. Cell growth inhibition on (a) wild-type strain; (b) ∆snq2 deletant strain; (c) ∆erg6 deletant strain; (d) ∆pdr3 deletant strain. The bidimensional scatter plots, as showed in Figure 8, allowed to rapidly identify eight compounds displaying the most prominent activity. Specifically, while compound 36 showed modest inhibition on both phases of cell growth (upper-left quadrant of Figure 8a), and compound 37 gave modest activation of cell growth and no inhibition of the stationary phase (Figure 8a), the activity on mutant strains resulted in dramatic inhibitory effects. Compound 36 proved to reduce strongly (about tenfold) the initial phase of cell growth (Figure 8b), which is related to cell replication in the ∆snq2 deletant strain having defects in the extrusion pathways, as Snq2p is a drug-efflux pump ABC (ATP-Binding Cassette) transporter conferring resistance to drugs and oxygen radicals. Compound 37 showed 100% and 200% reduction of cell growth in the initial phase (see vertical axis of Figure 8c,d) in both ∆erg6 and ∆pdr3 deletant strains, bearing deficiencies in genes involved in cell wall and MDR (Multi Drug Resistance), respectively. Such The bidimensional scatter plots, as showed in Figure 8, allowed to rapidly identify eight compounds displaying the most prominent activity. Specifically, while compound 36 showed modest inhibition on both phases of cell growth (upper-left quadrant of Figure 8a), and compound 37 gave modest activation of cell growth and no inhibition of the stationary phase (Figure 8a), the activity on mutant strains resulted in dramatic inhibitory effects. Molecules 2014, 19 Molecules 2014, 19 16520 From scaffold 31 and 33, a library of 48 distinct compounds, embedding 2,5-diketopiperazine, 2-oxopiperazine and 1,4-dihydrooxazine heterocycles, was generated (Scheme 8). Such skeletal diversity was accessed exploiting the reactivity of morpholine acetals as a function of the reaction methodology [91]. The compounds of this library contain interesting biochemical features, such as the morpholine moiety, which is present in several bioactive molecules [92,93], and the 2,5-diketopiperazine nucleus, which is considered a privileged scaffold in medicinal chemistry [94,95]. Scheme 7. Selected building blocks for the coupling step and skeletal diversity resulting from the cyclization step [90]. Scheme 8. Representative examples of library members obtained (a) from scaffold 31 and (b) from scaffold 33 [76]. Scheme 8. Representative examples of library members obtained (a) from scaffold 31 and (b) from scaffold 33 [76]. N H O R4 R3 CO2Me MeO 31 N O N R4 R3 O O R5 R6 R7 HN N O R3 CO2Me O R5 R6 N O CO2Me O H N Fmoc (a) N H O R4 R3 CO2Me MeO 31 N O N R4 R3 O O R5 R6 R7 HN N O R3 CO2Me O R5 R6 N O CO2Me O H N Fmoc N H O O O MeO 33 N O N O O O O ( )12 N N O O O O R5 R6 R7 N O N MeO O O R7 R5 R6 OR8 (a) (b) N H O O O MeO 33 N O N O O O O ( )12 N N O O O O R5 R6 R7 N O N MeO O O R7 R5 R6 OR8 (b) (b) (a) 5. Conclusions The application of chemical genetics in drug discovery research is becoming increasingly popular, as it holds the promise of discovering both new targets and new lead compounds. In order to make this kind of study valid, a suitable phenotype screening and versatile synthetic methods are needed. Diversity-oriented synthesis, as shown by the herein presented case studies, contributed significantly to the development of large small molecule collections. Nevertheless, greater efficiency is still required, especially for purifying and arraying library compounds in an arrangement suitable for screening processes. This interdisciplinary profile is striking, as it consists of harmonizing synthetic methodology with new analytical and biological methods. On the other hand, the whole screening apparatus does not satisfy yet the chemical genetics requirements. Considerable efforts are needed for developing new deconvolution methods, in order to decipher the mode of action of active compounds and to identify novel molecular targets. If the difficulties regarding the screenings will be solved, chemical genetics will expand significantly our knowledge of the druggable genome, opening the way to better chemotherapeutic treatments. Acknowledgments Financial support from University of Florence, MIUR PRIN2010-2011, cod. 2010NRREPL is acknowledged. Molecules 2014, 19 Compound 36 proved to reduce strongly (about tenfold) the initial phase of cell growth (Figure 8b), which is related to cell replication in the ∆snq2 deletant strain having defects in the extrusion pathways, as Snq2p is a drug-efflux pump ABC (ATP-Binding Cassette) transporter conferring resistance to drugs and oxygen radicals. Compound 37 showed 100% and 200% reduction of cell growth in the initial phase (see vertical axis of Figure 8c,d) in both ∆erg6 and ∆pdr3 deletant strains, bearing deficiencies in genes involved in cell wall and MDR (Multi Drug Resistance), respectively. Such mensional scatter plots, as showed in Figure 8, allowed to rapidly identify eight compoun The bidimensional scatter plots, as showed in Figure 8, allowed to rapidly identify eight compounds displaying the most prominent activity. Specifically, while compound 36 showed modest inhibition on both phases of cell growth (upper-left quadrant of Figure 8a), and compound 37 gave modest activation of cell growth and no inhibition of the stationary phase (Figure 8a), the activity on mutant strains resulted in dramatic inhibitory effects. Compound 36 proved to reduce strongly (about tenfold) the initial phase of cell growth (Figure 8b), which is related to cell replication in the ∆snq2 deletant strain having defects in the extrusion pathways, as Snq2p is a drug-efflux pump ABC (ATP-Binding Cassette) transporter conferring resistance to drugs and oxygen radicals. Compound 37 showed 100% and 200% reduction of cell growth in the initial phase (see vertical axis of Figure 8c,d) in both ∆erg6 and ∆pdr3 deletant strains, bearing deficiencies in genes involved in cell wall and MDR (Multi Drug Resistance), respectively. Such Molecules 2014, 19 Molecules 2014, 19 16522 effects gave initial clues on the role of compound 37 in targeting the ergosterol biosynthesis pathway downstream Erg6p. Following the screening of eight selected compounds, including compounds 36 and 37, for the mitochondrial membrane potential activation and peroxisomal proliferation, the two hit compounds 36 and 37 were successively validated. Compound 36, may act as a starvation inductor causing respiration and β-oxidation decrease. On the other hand, compound 37, inducing the lowest peroxisomal proliferation is a promising molecular probe to dissect the drug resistance mechanism in yeast and mammalian cells. Conflicts of Interest The authors declare no conflict of interest. Author Contributions A.T. and E.L. devised the review content, E.L. wrote the paper, A.T. and A.G. reviewed the final manuscript. References 1. Paul, S.M.; Mytelka, D.S.; Dunwiddie, C.T.; Persinger, C.C.; Munos, B.H.; Lindborg, S.R.; Schacht, A.L. How to improve R&D productivity: The pharmaceutical industry’s grand challenge. Nat. Rev. Drug Discov. 2010, 6, 203–214. 1. Paul, S.M.; Mytelka, D.S.; Dunwiddie, C.T.; Persinger, C.C.; Munos, B.H.; Lindborg, S.R.; Schacht, A.L. How to improve R&D productivity: The pharmaceutical industry’s grand challenge. Nat. Rev. Drug Discov. 2010, 6, 203–214. Molecules 2014, 19 16523 2. Ryan, D.P.; Matthews, J.M. Protein-protein interactions in human disease. Curr. Opin. Struct. Biol. 2005, 15, 441–446. 3. Altmann, K.H.; Buchner, J.; Kessler, H.; Diederich, F.; Krautler, B.; Lippard, S.; Liskamp, R.; Muller, K.; Nolan, E.M.; Samorì, B.; et al. The state of the art of chemical biology. ChemBioChem 2009, 10, 16–29. 4. Stockwell, B.R. Chemical genetics: Ligand-based discovery of gene function. Nat. Rev. Genet. 2000, 1, 116–125. 5. Stockwell, B.R. Exploring biology with small organic molecules. Nature 2004, 432, 847– 6. Walsh, D.P.; Chang, Y.-T. Chemical genetics. Chem. Rev. 2006, 106, 2476–2530. 7. O’Connor, C.J.; Laraia, L.; Spring, D.R. Chemical genetics. Chem. Soc. Rev. 2011, 40, 4332–4345. 8. Schreiber, S.L. The small-molecule approach to biology. Chem. Eng. News 2003, 81 S.L. The small-molecule approach to biology. Chem. Eng. News 2003, 81, 51–61. 9. Spring, D.R. Chemical genetics to chemical genomics: Small molecules offer big insights. Chem. Soc. Rev. 2005, 34, 472–482. 10. Yeoh, S.; O’Donnell, R.A.; Koussis, K.; Dluzewski, A.R.; Ansel, K.H.; Osborne, S.A.; Hackett, F.; Withers-Martinez, C.; Mitchell, G.H.; Bannister, L.H.; et al. Subcellular discharge of a serine protease mediates release of invasive malaria parasites from host erythrocytes. Cell 2007, 131, 1072–1083. 11. Arastu-Kapur, S.; Ponder, E.L.; Fonović, U.P.; Yeoh, S; Yuan, F.; Fonović, M.; Grainger, M.; Phillips, C.; Powers, J.C.; Bogyo, M. Identification of proteases that regulate erythrocyte rupture by the malaria parasite Plasmodium falciparum. Nat. Chem. Biol. 2008, 4, 203–213. 12. Gamo, F.J.; Sanz, L.M.; Vidal, J.; de Cozar, C.; Alvarez, E.; Lavandera, J.L., Vanderwall, D.E.; Green, D.V.; Kumar, V.; Hasan, S.; et al. Thousands of chemical starting points for antimalarial lead identification. Nature 2010, 465, 305–310. 13. Rottmann, M.; McNamara, C.; Yeung, B.K.; Lee, M.C.; Zou, B.; Russell, B.; Seitz, P.; Plouffe, D.M.; Dharia, N.V.; Tan, J.; et al. Spiroindolones, a potent compound class for the treatment of malaria. Science 2010, 329, 1175–1180. 14. Guiguemde, W.A.; Shelat, A.A.; Bouck, D.; Duffy, S.; Crowther, G.J.; Davis, P.H.; Smithson, D.C.; Connelly, M.; Clark, J.; Zhu, F.; et al. Molecules 2014, 19 Molecules 2014, 19 16524 19. Kugawa, F.; Watanabe, M.; Tamanoi, F. Chemical biology/chemical genetics/chemical genomics: Importance of chemical libraries. Chem-Bio Inform. J. 2007, 3, 49–68. 20. Bunin, B.A.; Ellman, J.A. A general and expedient method for the solid-phase synthesis of 1,4-benzodiazepine derivatives. J. Am. Chem. Soc. 1992, 114, 10997–10998. 21. De Witt, S.H.; Kiely, J.S.; Stankovic, C.J.; Schroeder, M.C.; Reynolds Cody, D.M.; Pavia, M.R. “Diversomers”: An approach to nonpeptide, nonoligomeric chemical diversity. Proc. Natl. Acad. Sci. USA 1993, 90, 6909–6913. 22. Balkenholh, F.; von dem Bussche-Hunnefeld, C.; Lansky, A.; Zechel, C. Combinatorial synthesis of small organic molecules. Angew. Chem. Int. Ed. 1996, 35, 2288–2337. 23. Plumkett, M.J.; Ellman, J.A. Solid-phase synthesis of structurally diverse 1,4-benzodiazepine derivatives using the Stille coupling reaction. J. Am. Chem. Soc. 1995, 117, 3306–3307. 24. Haggarty, S.J. The principle of complementarity: Chemical versus biological space. Curr. Opin. Chem. Biol. 2005, 9, 296–303. 25. Trabocchi, A., Ed. Diversity-Oriented Synthesis: Basics and Applications in Organic Synthesis, Drug Discovery and Chemical Biology, 1st ed.; John Wiley and Sons: Hoboken, NJ, USA, 2013; p. 664. 26. Galloway, W.R.J.D.; Isidro-Llobet, A.; Spring, D.R. Diversity-oriented synthesis as a tool for the discovery of novel biologically active small molecules. Nat. Commun. 2010, 1, 80–93. 27. Spring, D.R. Diversity-oriented synthesis; a challenge for synthetic chemists. Org. Biomol. Chem. 2003, 1, 3867–3870. 28. Schreiber, S.L. Target-oriented and diversity-oriented organic synthesis in drug discovery. Science 2000, 287, 1964–1968. 29. Burke, M.D.; Schreiber, S.L. A planning strategy for diversity-oriented synthesis. Angew. Chem. Int. Ed. 2004, 43, 46–58. 30. Spandl, R.J.; Bender, A.; Spring, D.R. Diversity-oriented synthesis: A spectrum of approach and results. Org. Biomol. Chem. 2007, 6, 1149–1158. 31. Nielsen, T.E.; Schreiber, S.L. Towards the optimal screening collection: A synthesis strategy. Angew. Chem. Int. Ed. 2008, 47, 48–56. 32. O’Connor, C.J.; Beckmann, H.S.G.; Spring, D.R. Diversity-oriented synthesis: Producing chemical tools for dissecting biology. Chem. Soc. Rev. 2012, 41, 4444–4456. 33. Stanton, B.Z.; Peng, L.F.; Maloof, N.; Nakai, K.; Wang, X.; Duffner, J.L.; Taveras, K.M.; Hyman, J.M.; Lee, S.W.; Koehler, A.N.; et al. A small molecule that binds Hedgehog and block its signaling in human cells. Nat. Chem. Biol. 2009, 5, 154–156. 34. Wyatt, E.E.; Fergus, S.; Galloway, W.R.J.D.; Bender, A.; Fox, D.J.; Plowright, A.T.; Jessiman, A.S.; Welch, M.; Spring, D.R. Skeletal diversity construction via a branching synthetic strategy. Chem. Commun. 2006, 31, 3296–3298. 35. References Chemical genetics of Plasmodium falciparum. Nature 2010, 465, 311–315. 15. Gujjar, R.; Marwaha, A.; el Mazouni, F.; White, J.; White, K.L.; Creason, S.; Shackleford, D.M.; Baldwin, J.; Charman, W.N.; Buckner, F.S.; et al. Identification of a metabolically stable triazolopyrimidine-based dihydroorotate dehydrogenase inhibitor with antimalarial activity in mice. J. Med. Chem. 2009, 52, 1864–1872. 16. Sijwali, P.S.; Rosenthal, P.J. Gene disruption confirms a critical role for the cysteine protease falcipain-2 in hemoglobin hydrolysis by Plasmodium falciparum. Proc. Natl. Acad. Sci. USA 2004, 101, 4384–4389. 17. Pisciotta, J.M.; Coppens, I.; Tripathi, A.K.; Scholl, P.F.; Shuman, J.; Bajad, S.; Shulaev, V.; Sullivan, D.J. The role of neutral lipid nanospheres in Plasmodium falciparum haem crystallization. Biochem J. 2007, 402, 197–204. 18. Weissbuch, I.; Leiserowitz, L. Interplay between malaria, crystalline hemozoin formation, and antimalarial drug action and design. Chem. Rev. 2008, 108, 4899–4914. Molecules 2014, 19 Wyatt, E.E.; Galloway, W.R.J.D.; Thomas, G.L.; Welch, M.; Loiseleur, O.; Plowright, A.T.; Spring, D.R. Identification of an anti-MRSA dihydrofolate reductase inhibitor from a diversity-oriented synthesis. Chem. Commun. 2008, 40, 4962–4964. 36. Tan, D.S.; Foley, M.A.; Shair, M.D.; Schreiber, S.L. Stereoselective synthesis of over two million compounds having structural features both reminiscent of natural products and compatible with miniaturized cell-based assays. J. Am. Chem. Soc. 1998, 120, 8565–8566. Molecules 2014, 19 16525 37. Tan, D.S.; Foley, M.A.; Stockwell, B.R.; Shair, M.D.; Schreiber, S.L. Synthesis and preliminary evaluation of a library of polycyclic small molecules for use in chemical genetics assays. J. Am. Chem. Soc. 1999, 121, 9073–9087. 37. Tan, D.S.; Foley, M.A.; Stockwell, B.R.; Shair, M.D.; Schreiber, S.L. Synthesis and preliminary evaluation of a library of polycyclic small molecules for use in chemical genetics assays. J. Am. Chem. Soc. 1999, 121, 9073–9087. 38. Tamura, O.; Okabe, T.; Yamaguchi, T.; Gotanda, K.; Noe, K.; Sakamoto, M. Studies on tandem transesterification and intramolecular cycloaddition of nitrones. 1. Sequential bicyclization of α-methoxycarbonylnitrones with allyl alcohols. Tetrahedron 1995, 51, 107–118. 39. Miyaura, N.; Suzuki, A. Palladium-catalyzed cross-coupling reactions of organoboron compounds. Chem. Rev. 1995, 95, 2457–2483. 40. Stille, J.K. The palladium-catalyzed cross-coupling reactions of organotin reagents with organic electrophiles. Angew. Chem. Int. Ed. Engl. 1986, 25, 508–524. 41. Sonogashira, K.; Tohda, Y.; Hagihara, N. A convenient synthesis of acetylenes: Catalytic substitutions of acetylenic hydrogen with bromoalkenes, iodoarenes and bromopyridines. Tetrahedron Lett. 1975, 50, 4467–4470. 42. Stephens, R.D.; Castro, C.E. The substitution of aryl iodides with cuprous acetylides. A synthesis of tolanes and heterocyclics. J. Org. Chem. 1963, 28, 3313–3315. 43. Lam, K.S.; Salmon, S.E.; Hersh, E.M.; Hruby, V.J.; Kazmierski, W.M.; Knapp, R.J. A new type of synthetic peptide library for identifying ligand-binding activity. Nature 1991, 354, 82–84. 44. Furka, A.; Sebestyen, F.; Asgedom, M.; Dibo, G. General method for rapid synthesis of multicomponent peptide mixtures. Int. J. Pept. Protein Res. 1991, 37, 487–493. 45. Stockwell, B.R.; Haggarty, S.J.; Schreiber, S.L. High-throughput screening of small molecules in miniaturized mammalian cell-based assays involving post-translational modifications. Chem. Biol. 1999, 6, 71–83. 46. MacBeath, G.; Koehler, A.N.; Schreiber, S.L. Printing small molecules as microarrays and detecting protein-ligand Interactions en masse. J. Am. Chem. Soc. 1999, 121, 7967–7968. 47. Vegas, A.J.; Fuller, J.H.; Koehler, A.N. Small-molecule microarrays as tools in ligand discovery. Chem. Soc. Rev. 2008, 37, 1385–1394. 48. Stockwell, B.R.; Schreiber, S.L. Molecules 2014, 19 Molecules 2014, 19 16526 54. Davies, H.M.L.; Stafford, D.G.; Doan, B.D.; Houser, J.H. Tandem asymmetric cyclopropanation/Cope rearrangement. A highly diastereoselective and enantioselective method for the construction of 1,4-cycloheptadienes. J. Am. Chem. Soc. 1998, 120, 3326–3331. 54. Davies, H.M.L.; Stafford, D.G.; Doan, B.D.; Houser, J.H. Tandem asymmetric cyclopropanation/Cope rearrangement. A highly diastereoselective and enantioselective method for the construction of 1,4-cycloheptadienes. J. Am. Chem. Soc. 1998, 120, 3326–3331. 55. Nicolaou, K.C.; Adsool, V.A.; Hale, C.R.H. An expedient procedure for the oxidative cleavage of olefinic bonds with PhI(OAc)2, NMO and catalytic OsO4. Org. Lett. 2010, 12, 1552–1555. 56. Engel, T. Basic overview of chemoinformatics. J. Chem. Inf. Model. 2006, 46, 2267 57. Medina-Franco, J.L. Chemoinformatic characterization of the chemical space and molecular diversity of compound libraries. In Diversity-Oriented Synthesis: Basics and Applications in Organic Synthesis, Drug Discovery and Chemical Biology, 1st ed.; Trabocchi, A., Ed.; John Wiley and Sons: Hoboken, NJ, USA, 2013; pp. 326–352. 58. Varnek, A.; Baskin, I.I. Chemoinformatics as a theoretical chemistry discipline. Mol. Inform. 2011, 30, 20–32. 59. Dobson, C.M. Chemical space and biology. Nature 2004, 432, 824–828. 60. Tan, D.S. Diversity-oriented synthesis: Exploring the intersections between chemistry and biology. Nat. Chem. Biol. 2005, 1, 74–84. 61. Joliffe, I.T. Principal Component Analysis, 2nd ed.; Springer-Verlag: New York, NY, USA, 2002. 61. Joliffe, I.T. Principal Component Analysis, 2nd ed.; Springer-Verlag: New York, NY, USA, 2002. 62. Kopp, F.; Stratton, C.F.; Akella, L.B.; Tan, D.S. Diversity-oriented synthesis approach to 62. Kopp, F.; Stratton, C.F.; Akella, L.B.; Tan, D.S. Diversity-oriented synthesis approach to macrocycles via oxidative ring expansion. Nat. Chem. Biol. 2012, 8, 358–365. 63. Sauer, W.H.B.; Schwarz, M.K. Molecular shape diversity of combinatorial libraries: A prerequisite for broad bioactivity. J. Chem. Inf. Comput. Sci. 2003, 43, 987–1003. 64. Hung, A.W.; Ramek, A.; Wang, Y.; Kaya, T.; Wilson, J.A.; Clemons, P.A.; Young, D.M. Route to three-dimensional fragments using diversity-oriented synthesis. Proc. Natl. Acad. Sci. USA 2011, 108, 6799–6804. 65. Lovering, F.; Bikker, J.; Humblet, C. Escape from flatland: increasing saturation as an approach to improving clinical success. J. Med. Chem. 2009, 52, 6752–6756. 66. Young, D.W.; Bender, A.; Hoyt, J.; McWhinnie, E.; Chirn, G.W.; Tao, C.Y.; Tallarico, J.A.; Labow, M.; Jenkins, J.L.; Mitchison, T.J.; et al. Integrating high-content screening and ligand-target prediction to identify mechanism of action. Nat. Chem. Biol. 2008, 4, 59–68. 67. Stanton, R.A.; Gernert, K.M.; Nettles, J.H.; Aneja, R. Drugs that target dynamic microtubules: A new molecular perspective. Med. Molecules 2014, 19 Probing the role of homomeric and heteromeric receptor interactions in TGF-β signaling using small molecule dimerizers. Curr. Biol. 1998, 8, 761–770. 49. Lo, M.M.-C.; Neumann, C.S.; Nagayama, S.; Perlstein, E.O.; Schreiber, S.L. A library of spirooxindoles based on a stereoselective three-component coupling reaction. J. Am. Chem. Soc. 2004, 126, 16077–16086. 50. Sebahar, P.R.; Williams, R.M. The asymmetric total synthesis of (+)- and (−)-spirotryprostatin B. J. Am. Chem. Soc. 2000, 122, 5666–5667. 51. Sebahara, P.R.; Osadab, H.; Usuib, T.; Williams, R.M. Asymmetric, stereocontrolled total synthesis of (+)- and (−)-spirotryprostatin B via a diastereoselective azomethine ylide [1,3]-dipolar cycloaddition reaction. Tetrahedron 2002, 58, 6311–6322. 52. Spector, U.; Shocket, N.R.; Kashman, Y.; Groweiss, A. Latrunculins: Novel marine toxins that disrupt microfilament organization in cultured cells. Science 1983, 219, 493–395. 53. Ibbeson, B.M.; Laraia, L.; Alza, E.; O’Connor, C.J.; Tan, Y.S.; Davies, H.M.L.; McKenzie, G.; Venkitaraman, A.R.; Spring, D.R. Diversity-oriented synthesis as a tool for identifying new modulators of mitosis. Nat. Commun. 2014, 5, 3155–3163. Molecules 2014, 19 Res. Rev. 2011, 31, 443–481. 68. Das, R.K.; Samanta, A.; Ghosh, K.; Zhai, D.; Xu, W.; Su, D.; Leong, C.; Chang, T.Y. Target identification: A challenging step in forward chemical genetics. Interdiscip. Bio Cent. 2011, 3, 1–18. identification: A challenging step in forward chemical genetics. Interdiscip. Bio Cent. 2011, 3, 1–18. 69. Resnick, M.A.; Cox, B.S. Yeast as an honorary mammal. Mutat. Res. 2000, 451, 1–11. 69. Resnick, M.A.; Cox, B.S. Yeast as an honorary mammal. Mutat. Res. 2000, 451, 1–11. C. Chemical genomics in yeast. Genome Biol. 2004, 5, doi:10.1186/gb-2004-5-9-240. 70. Brenner, C. Chemical genomics in yeast. Genome Biol. 2004, 5, doi:10.1186/gb-200 71. Giaever, G.; Flaherty, P.; Kumm, J.; Proctor, M.; Nislow, C.; Jaramillo, D.F.; Chu, A.M.; Jordan, M.I.; Arkin, A.P.; Davis, R.W. Chemogenomic profiling: Identifying the functional interactions of small molecules in yeast. Proc. Natl. Acad. Sci. USA 2004, 101, 793–798. 72. Lum, P.Y.; Armour, C.D.; Stepaniants, S.B.; Cavet, G.; Wolf, M.K.; Butler, J.S.; Hinshaw, J.C.; Garnier, P.; Prestwich, G.D.; Leonardson, A.; et al. Discovering modes of action for therapeutic compounds using a genome-wide screen of yeast heterozygotes. Cell 2004, 116, 121–137. Molecules 2014, 19 16527 73. Parsons, A.B.; Brost, R.L.; Ding, H.; Li, Z.; Zhang, C.; Sheikh, B.; Brown, G.W.; Kane, P.M.; Hughes, T.R.; Boone, C. Integration of chemical-genetics and genetics interaction data links bioactive compounds to cellular target pathways. Nat. Biotechnol. 2004, 22, 62–69. 74. Parsons, A.B.; Lopez, A.; Givoni, I.E.; Williams, D.E.; Gray, C.A.; Porter, J.; Chua, G.; Sopko, R.; Brost, R.L.; Ho, C.H.; et al. Exploring the mode-of-action of bioactive compounds by chemical-genetics profiling in yeast. Cell 2006, 126, 611–625. 75. Stefanini, I.; Trabocchi, A.; Marchi, E.; Guarna, A.; Cavalieri, D. A systems biology approach to dissection of the effects of small bicyclic peptidomimetics on a panel of Saccharomyces cerevisiae mutants. J. Biol. Chem. 2010, 285, 23477–23485. 76. Trabocchi, A.; Stefanini, I.; Morvillo, M.; Ciofi, L.; Cavalieri, D.; Guarna, A. Chemical genetics approach to identify new small molecule modulators of cell growth by phenotypic screening of Saccharomyces cerevisiae strains with a library of morpholine-derived compounds. Org. Biol. Chem. 2010, 8, 5552–5557. 77. Trabocchi, A.; Cavalieri, D.; Guarna, A. Chemical genetics approach to drug discovery by diversity-oriented synthesis (DOS) of peptidomimetics. Pure Appl. Chem. 2011, 83, 687–698. 78. Gante, J. Peptidomimetics—Tailored enzyme inhibitors. Angew. Chem. Int. Ed. 1994, 3 78. Gante, J. Peptidomimetics—Tailored enzyme inhibitors. Angew. Chem. Int. Ed. 1994, 33, 1699–1720. Molecules 2014, 19 i i l id i i f li d di d l d di l 78. Gante, J. Peptidomimetics—Tailored enzyme inhibitors. Angew. Chem. Int. Ed. 1994, 33, 1699–1720. 79. Giannis, A.; Kolter, T. Peptidomimetics for receptor ligands discovery, development and medical 78. Gante, J. Peptidomimetics—Tailored enzyme inhibitors. Angew. Chem. Int. Ed. 1994, 33, 1699–1720. 79. Giannis, A.; Kolter, T. Peptidomimetics for receptor ligands discovery, development and medical perspectives. Angew. Chem. Int. Ed. 1993, 32, 1244–1267. 79. Giannis, A.; Kolter, T. Peptidomimetics for receptor ligands discovery, development and medical perspectives. Angew. Chem. Int. Ed. 1993, 32, 1244–1267. 80. Trabocchi, A.; Menchi, G.; Guarna, F.; Machetti, F.; Scarpi, D.; Guarna, A. Design, synthesis and applications of 3-aza-6,8-dioxa-bicyclo[3.2.1]octane-based scaffolds for peptidominetic chemistry. Synlett 2006, 3, 331–353. 81. Guarna, A.; Guidi, A.; Machetti, F.; Menchi, G.; Occhiato, E.G.; Scarpi, D.; Sisi, S.; Trabocchi, A. Synthesis and reactivity of bicycles derived from tartaric acids and α-amino acids: A novel class of conformationally constrained dipeptide isosteres based upon enantiopure 3-aza-6,8-dioxabicyclo [3.2.1]octane-7-carboxylic acid. J. Org. Chem. 1999, 64, 7347–7364. 82. Guarna, A.; Bucelli, I.; Machetti, F.; Menchi, G.; Occhiato, E.G.; Scarpi, D.; Trabocchi, A. Synthesis of new enantiopure bicycles γ/δ-amino acid (BTKa), derived from tartaric acid and α-aminoacetophenone. Tetrahedron 2002, 58, 9865–9870. 83. Trabocchi, A.; Potenza, D.; Occhiato, E.G.; Guarna, A. Synthesis and conformational analysis of small peptides containing 6-endo-BT(t)L scaffolds as reverse turn mimetics. J. Org. Chem. 2002, 67, 7483–7492. 84. Calugi, C.; Guarna, A.; Trabocchi, A. Identification of constrained peptidomimetic chemotypes as HIV protease inhibitors. Eur. J. Med. Chem. 2014, 84, 444–453. 85. Trabocchi, A.; Mannino, C.; Machetti, F.; de Bernardis, F.; Arancia, S.; Cauda, R.; Cassone, A.; Guarna, A. Identification of inhibitors of drug-resistant Candida albicans strains from a library of bicyclic peptidomimetic compounds. J. Med. Chem. 2010, 53, 2502–2509. 86. Scarpi, D.; Cirelli, D.; Matrone, C.; Castronovo, G.; Rosini, P.; Occhiato, E.G.; Romano, F.; Bartali, L.; Clemente, A.M.; Bottegoni, G.; et al. Low molecular weight, non-peptidic agonists of TrkA receptor with NGF-mimetic activity. Cell Death Dis. 2012, 3, doi:10.1038/cddis.2012.80. 87. Gaisne, M.; Becam, A.M.; Verdiere, J.; Herbert, C.J. A ‘natural’ mutation in Saccharomyces cerevisiae strains derived from S288c affects the complex regulatory gene HAP1 (CYP1). Curr. Genet. 1999, 36, 195–200. Molecules 2014, 19 16528 88. Beltrame, L.; Rizzetto, L.; Paola, R.; Rocca-Serra, P.; Gambineri, L.; Battaglia, C.; Cavalieri, D. Using pathway signatures as means of identifying similarities among microarray experiments. © 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/). Molecules 2014, 19 PLoS One 2009, 4, doi:10.1371/journal.pone.0004128. 88. Beltrame, L.; Rizzetto, L.; Paola, R.; Rocca-Serra, P.; Gambineri, L.; Battaglia, C.; Cavalieri, D. Using pathway signatures as means of identifying similarities among microarray experiments. PLoS One 2009, 4, doi:10.1371/journal.pone.0004128. 89. Stefanini, I.; de Filippo, C.; Cavalieri, D. Yeast as a model in high-throughput screening of small-molecule libraries. In Diversity-Oriented Synthesis: Basics and Applications in Organic Synthesis, Drug Discovery and Chemical Biology, 1st ed.; Trabocchi, A., Ed.; John Wiley and Sons: Hoboken, NJ, USA, 2013; pp. 455–482. 90. Lalli, C.; Trabocchi, A.; Sladojevich, F.; Menchi, G.; Guarna, A. Diversity-oriented synthesis of morpholine-containing molecular scaffolds. Chem. Eur. J. 2009, 15, 7871–7875. 91. Ciofi, L.; Morvillo, M.; Sladojevich, F.; Guarna, A.; Trabocchi, A. Skeletal diversity by sequential one-pot and stepwise routes using morpholine ester scaffolds. Tetrahedron Lett. 2010, 51, 6282–6285. 92. Levin, J.I.; Chen, J.M.; Laakso, L.M.; Du, M.; Du, X.; Venkatesan, A.M.; Sandanayaka, V.; Zask, A.; Xu, J.; Xu, W.; et al. Acetylenic TACE inhibitors. Part 2: SAR of six-membered cyclic sulfonamide hydroxamates. Bioorg. Med. Chem. Lett. 2005, 15, 4345–4349. 93. Almstead, N.G.; Bradley, R.S.; Pikul, S.; De, B.; Natchus, M.G.; Taiwo, Y.O.; Gu, F.; Williams, L.E.; Hynd, B.A.; Janusz, M.J.; et al. Design, synthesis, and biological evaluation of potent thiazine- and thiazepine-based matrix metalloproteinase inhibitors. J. Med. Chem. 1999, 42, 4547–4562. 94. Martins, M.B.; Carvalho, I. Diketopiperazines: biological activity and synthesis. Tetrahedron 2007, 63, 9923–9932. 95. Nicholson, B.; Lloyd, G.K.; Miller, B.R.; Palladino, M.A.; Kisob, Y.; Hayashib, Y.; Neuteboom, S.T.B. NPI-2358 is a tubulin-depolymerizing agent: In vitro evidence for activity as a tumor vascular-disrupting agent. Anticancer Drugs 2006, 17, 25–31. © 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/). © 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W4200173961	https://zenodo.org/records/5761119/files/7.pdf	English	null	The Effect of Factors on Operational Efficiency and the Relationship between Operational Efficiency and Solvency of Logistics Joint-Stock Companies	International journal of multidisciplinary research and analysis	2,021	cc-by	7,064	INTERNATIONAL JOURNAL OF MULTIDISCIPLINARY RESEARCH AND ANALYSIS ISSN(print): 2643-9840, ISSN(online): 2643-9875 Volume 04 Issue 12 December 2021 DOI: 10.47191/ijmra/v4-i12-07, Impact Factor: 6.072 Page No.- 1809-1819 1. INTRODUCTION In the world’s current trend, logistics has become an economic industry - an essential service not only in Vietnam but also in the world. According to Council of Supply Chain Management Professionals (CSCMP), logistics which is an important sector in the overall structure of national economy, can bridge the gaps among other economic sectors, as well as promote the overall national economy. Therefore, this industry can solve the problem of optimizing unnecessary costs. Not falling out of this trend, top logistics enterprises in Vietnam have equipped themselves with sustainable resources in order to adapt to this change and serve full potential domestic market. In this context, managers need to understand the financial structure of their businesses to make appropriate adjustments and decisions to ensure operational efficiency and enhance development. However, Vietnamese researchers have put too much emphasis on capital structure (the relationship between debt and equity) and they have only paid little attention to solvency, which is also a very important issue in financial management. According to statistics of Hanoi Stock Exchange (HNX) and Ho Chi Minh City Stock Exchange (HOSE), the amount of capital listed on the trading floor of Vietnamese logistics enterprises is still moderate. There exist significant gaps in the amount of limited capital among internal enterprises in the same industry. According to the Vietnam Business Association, there are 1,200 logistics enterprises operating in the public sector. Apart from state-owned enterprises, most of them are small and medium-sized companies with limited resources. Therefore, these private companies cannot guarantee the ability to pay their own debts and ensure operational efficiency at the same time. This current situation generates deep challenges in management, investment, and production. All of these give rise to challenges for logistics JSCs in their business operation, in which, OE and solvency are not exceptional. The link between OE with the solvency of these companies is still controversial. Up to now, there is generally hardly any domestic concrete research to clarify the relationship between OE and solvency of enterprises and especially among logistics JSCs in Vietnam. As a result, this article will concentrate on identifying determinants of OE and the tie between OE and solvency of logistics JSCs so as to give a supported clue for this problem. Long Thanh Le1, Ha Thi Thu Nguyen2 1,2Advanced Education Program Institute, National Economics University ABSTRACT: Operational Efficiency (OE) and solvency play an important role in the success of any enterprises. Therefore, studying the relationship between OE and solvency needs to be taken comprehensively and continuously in order to find out long –term solutions for increasing business effectiveness. This paper examines the determinants of OE and the relationship between OE and solvency of the logistics joint-stock companies (JSCs) listed on the Vietnam Stock market with answers for the above - mentioned issues. This study mainly investigates 30 listed logistics firms from 2014 to 2018. The findings of this study suggested that the relationship between OE and solvency of logistics JSCs is inversely correlation. KEY WORDS: Operational Efficiency, Solvency, Logistics JSCs. Long Thanh Le1, Ha Thi Thu Nguyen2 The Effect of Factors on Operational Efficiency and the Relationship between Operational Efficiency and Solvency of Logistics Joint-Stock Companies including service quality, still have to be achieved in order to keep existing customers and maintain high revenue. Also, the study of Dennis Hartman (2016) pointed out a method of measuring OE by evaluating how well a business managed its resources and utilized them to generate profits. A theory which was developed by Neil Kokemuller (2016) focused on OE in terms of encompassing several strategies and techniques used to accomplish the basic goal of delivering quality goods to customers in the most cost- effective and timely manner; and on OE involving performing similar activities in more efficient ways than their competitors. Moreover, Subha Varadan (2016) proposed that “OE is a critical system that can keep a company in business or close it down”. In terms of typical domestic researches, Nguyen Van Cong (2019) pointed out that the OE of a company reflected operation results that a company possibly got when it used its input for business operation. Basically, OE is an indicator of the efficiency of using the input elements of business operation and solvency. In terms of measuring OE, the study of Bernstein (1988) identified that OE which presented firm performance can be evaluated by financial ratios such as solvency, capital structure, profitability, and turnover. In addition, Feng (2000) did also take financial ratios into consideration on carrying out firms’ performance evaluation process for airlines by using grey relation analysis and TOPSIS method to overcome the problem of small samples and outranking of airlines. According to Hobarth (2006), he showed the correlation between financial indicators and OE of listed firms in the USA for a19- year period by using 17 financial indicators and 3 variables to measure firm’s performance, namely market performance (measured by changes in stock market value), cash flow performance (solvency), and profitability (ROI). Sandstrom (2007) presented a historical review and some pragmatic solutions to Solvency. More specifically he discussed four fundamental issues needed for constructing solvency systems, i.e. valuation of assets and liabilities, risk margin for uncertainty in liabilities and assets and risk measures and modeling (risk categories, risk mitigation, diversification, etc.). Regarding solvency, there is an array of different definitions. First of all, according to Pentikäinen, T. (1952), the definition of solvency was the difference between assets and liabilities. Benjamin (1977) refered to the Oxford Dictionary, which defined solvency as "having enough money to meet all liabilities". H1: Scale of company has a positive impact on Operational Efficiency Degree of financial independence: The degree of financial independence is presented by the proportion of owner equity out of total equity (abbreviated as ER). The more debts a company owed, the lower degree of financial independence of this company is. According to the theory of Modigliani and Miller, when a business starts to borrow, the business will have advantages of tax shield. The low cost of debt combined with the tax shield advantage will lead to a decrease in total cost. Therefore, business performance is enhanced. Specifically, if the companies kept the degree of financial independence at a low level, its OE would increase. In contrast, the empirical research results of Zeitun and Tian (2007), Neil Nagy (2009); Fozia Memon, Niaz Ahmed Bhutto, Ghulam Abbaas (2012) showed that debt ratio had a negative and significant effect on business performance of the enterprise. The Effect of Factors on Operational Efficiency and the Relationship between Operational Efficiency and Solvency of Logistics Joint-Stock Companies Developing the theory of Pentikäinen, T, Jackson et al. (2002) defined a firm's solvency as when its total assets were higher than current liabilities. Langiemer (2004), solvency measured the amount of debt and other cost obligations used in a company's business compared to the amount of equity invested in an enterprise. Ibenta (2005) argued that solvency was a firm's ability to meet capital needs. Scale of company: Usually, large-scale businesses will have incessant and sustainable resources. With financial strength these businesses will easily minimize input costs and increase output efficiency. Research by Gleason, L. K Mathur and I.Mathur (2000), Onaolapo and Kajola (2010) showed that firms’ size had a positive and significant impact on business performance, i.e.the bigger a business is, the efficiency it could generate. However, according to the results of other studies from such researchers as Mudambi and Nicosia (1998), Lauterbach and Vaninsky (1999), Durand and Coeuderoy (2001), Tzelepis and Skuras (2004), Zeitun and Tian (2007), the scale had positive impact but not significant on business performance of the enterprise. H2: Degree of financial independence has a positive impact on Operational Efficiency Total assets: Empirical research results of Rami Zeitun and Gary Gang Tian (2007); Onaolapo and Kajola (2010); Marian Siminica, Daniel Circiumaru, Dalia Simion (2011); Fozia Memon, Niaz Ahmed Bhutto and Ghulam Abbas (2012) showed that total assets were significantly negative impacts on business performance of the enterprise. According to the research of Marian Siminica, Daniel Circiumaru, Dalia Simion (2011), total assets also had a positive and significant effect on the OE of the company. 2. LITERATURE REVIEW AND HYPOTHESIS There are many concepts of Operational Efficiency (OE) from different researchers both domestically and internationally and below are some typical ones. Operational efficiency can be formally defined as the ability of an enterprise to ensure the high quality of its products and services while delivering products or services to its customers in the most cost-effective manner possible (Beal, 2016). In addition, according to Matthew Burrows (2016), OE is not just about reducing costs; other business objectives IJMRA, Volume 4 Issue 12 December 2021 IJMRA, Volume 4 Issue 12 December 2021 www.ijmra.in Page 1809 www.ijmra.in H3: Total assets has a positive impact on Operational Efficiency Sales: The revenue will reflect the capacity of sales activities. An increase in goods sold, can positively affect business performance of a business. The study of Nguyen Quoc Nghi (2010) proved that there existed a direct correlation between total sales and OE. H4: Total sales has a positive impact on Operational Efficiency IJMRA, Volume 4 Issue 12 December 2021 Page 1810 www.ijmra.in The relationship between OE and solvency After considering different concepts of OE and solvency, one question has appeared: Is there any relationships between them? In fact, this topic has not attracted many studies from domestic researchers except for some foreign ones. On the one hand, there exists a positive relationship between OE and solvency.Lambery and Valming (2009) evaluated and compared the solvency management of companies listed on the Stockholm Stock Exchange over two times to see how solvency’s changes affecting firms’ performance. The results showed that solvency maintenance strategies did not have any significant impact on operational efficiency. A study of Victor Chukwunwieke (2014) determined the relationship between quick ratio and profitability ratios of listed companies on the Nigerian stock exchange NSE. The author concluded that there was a positive correlation between quick ratio and profitability ratios. Rafid Hamad (2016) conducted an assessment of the relationship between solvency and performance of Standard Chartered Bank in Pakistan. This study showed that there was a positive relationship between solvency and OE of enterprises. On the other hand, some researchers announced a negative correlation between OE and solvency. Jennifer Muthio Kyule (2015) conducted a research on the relationship of OE and solvency for financial activities of listed companies on Nairobi stock exchange, Kenya. The dependent variable presented for OE was ROA, while the independent variables were: solvency, liquidity, financial leverage, minimum capital adequacy ratio, and size of businesses (total assets). The author concluded that solvency was negatively correlated with ROA which presented for OE. Evengi Raykov (2017) explored a trade-off between solvency and profitability in the event of financial crisis. Analysis data were collected by the authors from quarterly financial reports of 20 companies selected from different economies in the Bulgarian economy during 2007-2015. Through empirical testings of many hypotheses, the work came to a conclusion that during the 2007 crisis, financial managers successfully separated profits from payables when the correlation between them was negative. H5: Operational efficiency has a significant and positive relationship on Solvency H5: Operational efficiency has a significant and positive relationship on Solvency p y g p p y Fig 1. Conceptual Model 3. METHODOLOGY Evaluation indicators Operational Efficiency Equity ER Total assets Sales Operational efficiency (OE) Solvency Operational efficiency (OE) Solvency Sales Fig 1. Conceptual Model 4. RESULT There are many listed firms providing logistic services in Vietnam and their stocks are listed on the two main securities exchanges of Vietnam, the HOSE and HNX. In order to have an overview of financial situation of all firms in this industry, this paper has selected listed firms on both HNX and HOSE. There are thirty Logistics JSCs which were divided equally in both HNX and HOSE (14 firms listed on the HNX and 16 firms listed on the HOSE). Of these, Vietnam Airlines corporation (HVN) has the highest authorized-capital with nearly 82,000 billion Vietnam Dong (VND), which is nearly two times bigger than the second highest authorized-capital (VJC – 36,000 billion VND and three times bigger than the HNX highest authorized-capital (PHP), while the smallest authorized-capital is PRC with only 12 billion VND. Concretely, eight companies including CDN, DL1, PHP, VJC, GMD, HVN, PVT, and VOS have their scale of capital from over 1000 billion VND. In this paper, all surveyed firms shall be mentioned by their coded stocks instead of their names. Method In order to examine the OE of researched enterprises, there are six variables used as follows. Two dependent variables which present for OE are Equity Turnover (ET) and Total Assets Turnover (TAT) and four other independent variables: Assets (presenting total assets of a company), Equity (presenting scale of a company), Equity Ratio (ER = 𝑂𝑤𝑛𝑒𝑟𝑠 𝑒𝑞𝑢𝑖𝑡𝑦 𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠, presenting liquidity of companies) and Sales (presenting the result of selling process). After that, in order to measure the solvency of listed logistics firms, there are two dependent variables: Quick ratio (QUICK = 𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑎𝑠𝑠𝑒𝑡𝑠 − 𝑖𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑦 𝐶𝑢𝑟𝑒𝑛𝑡 𝑙𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑖𝑒𝑠 ) and Overall ratio (OVERALL = 𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠 𝑇𝑜𝑡𝑎𝑙 𝑙𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑖𝑒𝑠) three other controlling variables, including TAT, ER, and ET. The study applies both qualitative and quantitative approaches. For a qualitative approach, the study takes a comparative and analytical method in order to assess the current situation of OE and solvency as well as detect the relationship between them. Theory frame is based on a fundamental base about a system of ratios which reflect the OE (ET and TAT) and Solvency (QUICK and OVERALL) of a company. In addition, in order to strengthen the reliability of this study’s result, this paper also uses a quantitative approach by running a regression model of Ordinary Least Square (OLS) with the above-mentioned variables. The OLS’s first aim is to investigate how many factors impact OE and what they are. The second purpose is to forecast the link between OE and solvency. This paper uses the statistic software Stata 15 to run the regression to answer these questions. The use of both qualitative and quantitative approaches aims to strengthen the reliability of the analyses and judgments because it collects evidence from different sources and creates a multi-directional vision of an issue. This combination also helps the result to satisfy planned purposes better and answer the research questions clearly as well as lead to conclusions which ensure a scientific base and feasibility. Data Data used in this study was collected from financial statements, annual reports of logistic firms listed in two major State Securities Commission of Vietnam: HNX and HOSE for the period from 2014 to 2018. These data were audited by the world famous auditing firms (e.g.: E&Y, KPMG, A&C…) and downloaded from the reliable websites of the HNX and the HOSE in this survey. These logistic firms with their data lead to a research sample with 150 observations during this period. In this case, the above-mentioned data were transferred into Excel and encoded as variables. After that, they become inputs for running regression Method Evaluation indicators Evaluation indicators In a study on the valuation of operating efficiency in Egypt firms, Armer, Mostafa and Eldomiaty (2011) found asset quality, capital adequacy, credit risk and liquidity as major determinants of OE in highly competitive companies. Based on the results found by Armer, Mustapha and Eldomiaty (2011) and a theory developed by Hobarth (2006) about using financial ratios to evaluate OE, Nguyen Van Cong (2019) measured OE via: ● Total asset turnover (TAT) presented for asset quality. ● Equity turnover (ET) presented for capital adequacy. ● Equity Ratio (ER) presented for liquidity. Solvency The dimensions’ choice was influenced by the work of Nguyen Van Cong (2019).Based on this scale, Solvency can be measured via two ratios: ● Overall Ratios presented for the ability to repay all companies’ debts in overall. ● Quick Ratios presented for the ability to secure short-term debts. IJMRA, Volume 4 Issue 12 December 2021 IJMRA, Volume 4 Issue 12 December 2021 www.ijmra.in Page 1811 Page 1811 www.ijmra.in Operational efficiency Firstly, a company’s capital scale is not directly proportional to its OE. More clearly, despite its highest capital scale at nearly 82,000 billion VND, circulating turnover of total assets in HVN only ranks kin the eleventh place at 0.75times, which is lower than the lowest authorized-capital firm VNLas shown in Table 1. Firstly, a company’s capital scale is not directly proportional to its OE. More clearly, despite its highest capital scale at nearly 82,000 billion VND, circulating turnover of total assets in HVN only ranks kin the eleventh place at 0.75times, which is lower than the lowest authorized-capital firm VNLas shown in Table 1. This conclusion is also strengthened when VJC stands in the second place of capital scale (at 36,000 billion VND) but at the tenth place of OE in terms of total assets turnover (TAT). On the other hand, in this period, VNL and PRC have the lowest capital scale but express its graduation in circulating turnover of total assets among the first and can be seen clearly in Table 1. IJMRA, Volume 4 Issue 12 December 2021 www.ijmra.in Page 1812 Table 1. Circulating turnover of total assets (TAT). Operational efficiency JSCs 2014 2015 2016 2017 2018 Average 01.VNL 2.706 2.451 2.231 2.567 2.851 2.561 02.PRC 3.174 2.371 1.922 1.846 1.598 2.182 Page 1812 Page 1812 www.ijmra.in e Effect of Factors on Operational Efficiency and the Relationship between Operational Efficiency and Solvency Logistics Joint-Stock Companies 03.GSP 1.073 1.371 2.407 2.572 2.579 2.000 04.STG 3.959 2.144 0.602 0.499 0.554 1.552 05.SFI 1.748 1.363 1.277 1.493 1.680 1.512 06.CAG 0.627 0.571 4.995 0.431 0.435 1.412 07.VNS 1.676 1.604 1.487 0.945 0.705 1.283 08.VJC 1.305 1.117 1.037 0.950 1.029 1.088 09.MAC 1.002 0.858 0.851 0.832 0.748 0.858 10.TTZ 0.389 0.777 0.971 0.358 1.720 0.843 11.HVN 0.865 0.532 0.663 0.762 0.926 0.750 12.NAP 0.644 0.511 0.624 0.641 0.682 0.620 13.BSC 0.591 0.496 0.961 0.495 0.528 0.614 14.HHG 0.583 0.613 0.602 0.604 0.469 0.574 15.HTV 0.587 0.590 0.524 0.495 0.571 0.553 16.VGP 1.031 1.397 0.234 0.104 0.000 0.553 17.HAH 0.568 0.509 0.463 0.525 0.505 0.514 18.CDN 0.226 0.505 0.558 0.516 0.477 0.456 19.VSC 0.574 0.512 0.337 0.334 0.391 0.430 20.TMS 0.506 0.461 0.423 0.343 0.300 0.407 21.CLL 0.453 0.440 0.414 0.391 0.316 0.403 22.VOS 0.369 0.314 0.260 0.399 0.450 0.358 23.PVT 0.283 0.338 0.349 0.370 0.438 0.356 24.DL1 0.452 0.378 0.235 0.046 0.019 0.226 25.PHP 0.155 0.255 0.251 0.227 0.218 0.221 26.ACV 0.211 0.255 0.000 0.289 0.316 0.214 27.HMH 0.203 0.283 0.164 0.150 0.136 0.187 28.PGT 0.261 0.072 0.043 0.043 0.239 0.131 29.GMD 0.161 0.152 0.139 0.092 0.033 0.115 30.TCO 0.124 0.088 0.057 0.059 0.060 0.078 From the above analysis, it can be said that a big capital scale is a convenient condition for a company to increase its OE but Effect of Factors on Operational Efficiency and the Relationship between Operational Effici ogistics Joint-Stock Companies on Operational Efficiency and the Relationship between Operational Efficiency and Solvency k Companies rational Efficiency and the Relationship between Operational Efficiency and Solvency From the above analysis, it can be said that a big capital scale is a convenient condition for a company to increase its OE but whether a company is able to explore this advantage or not, it is quite a different case. From the above analysis, it can be said that a big capital scale is a convenient condition for a company to increase its OE but whether a company is able to explore this advantage or not, it is quite a different case. Operational efficiency To strengthen the founded result, our research team uses statistical software STATA 15 to model a regression with the least squares regression method (OLS). After running model with the dependent variable of (TAT) as well as the three independent variables including Equity, ER and Sales (Independent variables are under logarithm),the results are expressed in Table 2. To strengthen the founded result, our research team uses statistical software STATA 15 to model a regression with the least squares regression method (OLS). After running model with the dependent variable of (TAT) as well as the three independent variables including Equity, ER and Sales (Independent variables are under logarithm),the results are expressed in Table 2. To strengthen the founded result, our research team uses statistical software STATA 15 to model a regression with the least squares regression method (OLS). After running model with the dependent variable of (TAT) as well as the three independent variables including Equity, ER and Sales (Independent variables are under logarithm),the results are expressed in Table 2. TAT= 0.20 + 1.57ER - 0.70log(Equity) + 0.67log(Sales) From the above-mentioned results, it can be seen that independent variables of Sales and the ER maintain direct proportion with TAT and have at least 99% statistical meaning. This explains that enterprises with a larger proportion of equity accompanied with higher revenue from sales of goods and services will have significant increases in their OE. In other words, the bigger the companies are, the larger their OE is and vice versa. This result is consistent with the initial judgment of the group. Secondly, logistics firms’ degree of financial independence is not directly proportional to their OE. The percentage of owners’ equity in total capital (ER) is the most important ratio to express a company’s degree of financial independence. Again, PRC and VNL are still leading companies in circulating turnover of owners’ equity at 4,827 and 4,578 times respectively while these firm’s percentage of owners’ equity is only with its arithmetical mean of 20% during five years. DL1is an enterprise that has the lowest circulating turnover of owners’ equity (ET) with its arithmetical mean for the surveyed period of 0.255 times only. This can be expressed in Table 3. Table 3. Circulating turnover of total equity (ET). le 3. Circulating turnover of total equity (ET). Table 2. Regression TAT with Equity, ER and Sales g q y IJMRA, Volume 4 Issue 12 December 2021 Page 1813 www.ijmra.in To strengthen the founded result, our research team uses statistical software STATA 15 to model regression with the least squares regression method (OLS). After running model with dependent variable of (ET) as well as three independent variables including Asset, ER, and Sales (Independent variables are under logarithm),the results are expressed in Table 4. TAT= 0.20 + 1.57ER - 0.70log(Equity) + 0.67log(Sales) JSCs 2014 2015 2016 2017 2018 Average 01.VJC 9.022 8.748 6.801 5.174 5.673 7.083 02.PRC 6.563 5.200 4.578 4.243 3.551 4.827 03.VNL 4.518 4.588 4.168 4.619 4.996 4.578 04.HVN 5.340 3.166 3.919 3.814 4.020 4.052 05.STG 6.040 6.232 1.345 0.847 0.779 3.048 06.GSP 1.302 1.654 3.060 3.380 3.918 2.663 07.VNS 2.991 3.032 2.945 1.745 1.152 2.373 08.SFI 2.383 1.749 1.628 1.938 2.160 1.972 09.VOS 1.522 1.337 1.428 2.581 2.648 1.903 10.CAG 0.668 0.615 5.374 0.456 0.725 1.568 11.MAC 1.434 1.157 1.094 1.093 0.987 1.153 12.TTZ 0.531 1.051 1.309 0.454 2.065 1.082 13.HHG 1.235 1.090 1.027 1.011 0.740 1.020 14.VGP 1.625 1.617 0.245 1.188 0.000 0.935 15.BSC 0.747 0.654 1.251 0.624 0.650 0.785 16.HAH 0.850 0.864 0.786 0.752 0.619 0.775 17.PVT 0.703 0.761 0.716 0.708 0.839 0.745 18.NAP 0.775 0.600 0.709 0.704 0.706 0.699 19.HTV 0.760 0.700 0.588 0.568 0.675 0.658 20.TMS 0.720 0.704 0.717 0.577 0.508 0.645 21.CDN 0.325 0.654 0.696 0.711 0.663 0.610 22.VSC 0.689 0.606 0.386 0.370 0.427 0.495 23.CLL 0.614 0.576 0.508 0.432 0.328 0.492 24.ACV 0.434 0.552 0.000 0.535 0.562 0.416 25.PHP 0.203 0.348 0.343 0.292 0.285 0.294 26.DL1 0.522 0.440 0.247 0.047 0.019 0.255 27.HMH 0.224 0.295 0.173 0.157 0.140 0.198 28.GMD 0.215 0.205 0.196 0.122 0.041 0.156 29.PGT 0.268 0.073 0.043 0.044 0.282 0.142 30.TCO 0.133 0.091 0.060 0.063 0.064 0.082 To strengthen the founded result, our research team uses statistical software STATA 15 to model regression with the l ares regression method (OLS). After running model with dependent variable of (ET) as well as three independent varia uding Asset, ER, and Sales (Independent variables are under logarithm),the results are expressed in Table 4. To strengthen the founded result, our research team uses statistical software STATA 15 to model regression with the least squares regression method (OLS). After running model with dependent variable of (ET) as well as three independent variables including Asset, ER, and Sales (Independent variables are under logarithm),the results are expressed in Table 4. To strengthen the founded result, our research team uses statistical software STATA 15 to model regression with the least squares regression method (OLS). After running model with dependent variable of (ET) as well as three independent variables including Asset, ER, and Sales (Independent variables are under logarithm),the results are expressed in Table 4. ET = 2.75 – 2.86ER – 0.89log(Asset) + 1.00log(Sales) With the above-mentioned results, it can be seen that two independent variables Asset (Asset) and ER maintain an inverse relationship with dependent variable ET and have at least 99% statistical meaning. This shows that OE of logistic JSCs will decrease when it keeps total assets and its financial independence at a high level. This result is not consistent with the initial judgment of the group. TAT= 0.20 + 1.57ER - 0.70log(Equity) + 0.67log(Sales) IJMRA, Volume 4 Issue 12 December 2021 Page 1814 www.ijmra.in The Effect of Factors on Operational Efficiency and the Relationship between Operational Efficiency and Solvency of Logistics Joint-Stock Companies Table 4. Regression ET with Asset, ER and Sales The Effect of Factors on Operational Efficiency and the Relationship between Operational Efficiency and Solvency of Logistics Joint-Stock Companies The Effect of Factors on Operational Efficiency and the Relationship between Operational Efficiency and Solvency of Logistics Joint-Stock Companies IJMRA, Volume 4 Issue 12 December 2021 IJMRA, Volume 4 Issue 12 December 2021 www.ijmra.in Solvency Regarding the solvency of joint - stock companies listed on Vietnam's stock market, our research group uses two common indicators namely Overall ratio and Quick ratio to measure, evaluate, and our team draws the following conclusions. Regarding the solvency of joint - stock companies listed on Vietnam's stock market, our research group uses two common indicators namely Overall ratio and Quick ratio to measure, evaluate, and our team draws the following conclusions. Firstly, through Table 5, we can see that the order of all surveyed enterprises has completely changed when compared to OE. In more detail, DL1 is a low-performing enterprise, and its asset and equity turnover ratios are significantly smaller than other firms in the same industry. However, this company has sufficient resources to maintain the overall ratio at 72.02 times, which is the highest among all surveyed firms. In contrast, HVN and VJC are two enterprises with a large capital scale and good operating efficiency, but the solvency of these businesses is limited when both ratios are ranked at the last level. Table 5. Solvency Overall ratio JSCs 2014 2015 2016 2017 2018 Average 01.DL1 5.842 9.060 34.112 169.439 141.661 72.023 02.PGT 61.533 68.433 43.347 32.582 4.025 41.984 03.HMH 21.033 26.170 16.623 32.103 36.246 26.435 04.TCO 38.102 24.741 18.057 14.192 21.755 23.369 05.CAG 16.710 12.195 17.009 20.076 24.507 18.099 06.CLL 3.985 4.458 6.896 25.520 30.494 14.271 07.VGP 4.510 24.747 18.660 1.051 0.000 9.794 08.NAP 6.403 7.149 10.092 12.917 10.324 9.377 09.VSC 5.913 7.062 8.818 12.457 11.546 9.159 10.HTV 4.906 8.872 9.520 6.659 6.355 7.262 11.BSC 4.507 3.830 5.081 4.671 6.262 4.870 12.TTZ 3.966 3.719 3.820 5.663 6.070 4.648 13.SFI 4.619 4.451 4.815 4.008 5.089 4.596 14.GSP 6.605 5.276 4.243 4.128 2.406 4.532 15.GMD 4.013 3.759 3.154 5.328 5.108 4.272 16.CDN 3.645 5.575 4.632 3.090 4.042 4.197 17.PHP 3.897 3.609 4.496 4.542 3.993 4.107 18.MAC 3.090 4.545 4.451 3.944 4.326 4.071 19.HAH 2.516 2.369 2.491 5.403 5.450 3.646 20.STG 2.530 1.327 2.110 2.843 4.490 2.660 Page 1815 www.ijmra.in The Effect of Factors on Operational Efficiency and the Relationship between Operational Efficiency and Solvency of Logistics Joint-Stock Companies 21.TMS 3.293 2.648 2.303 2.650 2.286 2.636 22.HHG 2.020 2.563 2.340 2.651 2.827 2.481 23.VNS 2.171 2.083 1.967 2.497 2.661 2.276 24.VNL 2.186 2.112 2.193 2.311 2.347 2.230 25.PVT 1.753 1.846 2.074 2.119 2.072 1.973 26.PRC 1.868 1.811 1.695 1.863 1.775 1.802 27.ACV 1.902 1.839 0.000 2.241 2.330 1.663 28.VOS 1.344 1.275 1.172 1.195 1.215 1.240 29.HVN 1.185 1.177 1.230 1.272 1.330 1.239 30.VJC 1.138 1.152 1.199 1.242 1.206 1.187 In terms of quick ratios, the order is still unchanged. Enterprises have low operational efficiency will have high solvency and vice versa. Specifically, PGT holds the highest position for quick ratio (36.07 times), followed by DL1 with the second highest solvency at 15.91 times, but the performance of both these businesses is at a low level. The Effect of Factors on Operational Efficiency and the Relationship between Operational Ef of Logistics Joint-Stock Companies on Operational Efficiency and the Relationship between Operational Efficiency and Solvency k Companies In terms of quick ratios, the order is still unchanged. Enterprises have low operational efficiency will have high solvency and vice versa. Specifically, PGT holds the highest position for quick ratio (36.07 times), followed by DL1 with the second highest solvency at 15.91 times, but the performance of both these businesses is at a low level. In terms of quick ratios, the order is still unchanged. OVERALL = -23.96 – 12.91TAT + 5.44ET+51.77ER Moreover, our research team also used statistical software STATA 15 to run model with the dependent variable of (QUICK) as well as the three independent variables including TAT, ER and ET, the results are expressed in Table 8. Solvency Enterprises have low operational efficiency will have high solvency and vice versa. Specifically, PGT holds the highest position for quick ratio (36.07 times), followed by DL1 with the second highest solvency at 15.91 times, but the performance of both these businesses is at a low level. e 6. Quick ratio JSCs 2014 2015 2016 2017 2018 Average 01.PGT 61.386 68.433 27.680 20.599 2.243 36.068 02.DL1 5.378 9.151 9.423 30.266 25.315 15.907 03.BSC 3.392 6.087 13.228 11.270 14.157 9.627 04.TCO 7.554 7.366 11.247 6.418 9.738 8.465 05.HMH 5.728 7.320 4.440 8.441 13.070 7.800 06.CAG 7.068 4.711 5.925 7.422 9.392 6.904 07.CLL 0.693 2.721 4.767 10.468 13.255 6.381 08.HTV 3.454 7.012 7.846 3.632 3.950 5.179 09.VGP 1.761 13.535 8.838 1.036 0.000 5.034 10.GSP 5.821 4.000 3.618 3.790 2.445 3.935 11.VSC 2.455 3.113 3.184 3.185 3.319 3.051 12.VNL 1.835 3.060 3.294 3.244 2.691 2.825 13.CDN 1.618 4.920 3.465 1.253 2.331 2.717 14.PHP 1.559 1.857 3.410 3.204 2.239 2.454 15.SFI 2.176 2.302 2.904 2.252 2.376 2.402 16.ACV 1.758 2.694 0.000 3.519 4.032 2.401 17.NAP 1.641 2.233 3.814 1.818 2.150 2.331 18.HAH 1.863 1.772 1.590 1.786 3.137 2.030 19.MAC 1.429 2.214 2.171 1.876 2.033 1.945 20.PVT 1.935 1.625 1.979 1.949 1.665 1.831 21.GMD 3.870 2.161 0.537 1.261 1.010 1.768 22.TMS 1.981 2.305 1.203 1.083 1.469 1.608 23.STG 1.965 1.122 1.011 1.106 0.659 1.173 24.TTZ 1.209 0.765 0.913 1.116 1.165 1.034 25.HHG 0.314 1.206 0.757 1.407 1.258 0.988 26.VJC 0.973 0.876 0.806 0.914 0.800 0.874 27.VOS 1.161 1.059 0.758 0.619 0.665 0.852 28.PRC 0.884 0.924 0.688 0.828 0.793 0.823 29.VNS 0.767 0.633 0.385 0.511 0.566 0.572 30.HVN 0.577 0.584 0.617 0.476 0.486 0.548 IJMRA, Volume 4 Issue 12 December 2021 IJMRA, Volume 4 Issue 12 December 2021 www.ijmra.in Page 1816 www.ijmra.in Relationship between operational efficiency and solvency After considering both OE and solvency of logistics JSCs listed on the Vietnamese Stocks market, this study draws some findings as follows. Firstly, there is an inverse correlation between OE and solvency of enterprises. A business with good performance always has not enough resources to maintain a stable solvency. This is explained by the fact that VNL and PRC are always the two leading companies in terms of operating efficiency, however, they show solvency at a low level. Not only does it occur in small and medium-sized companies, but this relationship also exists in large companies. Through Table 6 and Table 6, despite their largest scale ad high financial independence, HVN and VJC only have enough resources to ensure the overall solvency and cannot guarantee the ability to pay short-term debts. The results are similar to the results of the previous studies. According to (Kyule, 2015), the solvency of companies was negatively correlated with OE. This has been validated for all surveyed companies by our research team, especially VNL, PRC, HVN, VJC and DL1. In addition, (Raykov, 2017), companies that reduced their solvency and raise more debts would boost their efficiency of the companies' operations. This is a negative correlation. To strengthen the founded result, our research team used statistical software STATA 15 to model regression with the least squares regression method (OLS). After running model with the dependent variable of (OVERALL) as well as the three independent variables including TAT, ER and ET, the results are expressed in Table 7. IJMRA, Volume 4 Issue 12 December 2021 5. CONCLUSION AND RECOMMENDATION By analyzing the relationship between OE and solvency of logistics JSCs listed on the Vietnam Stock market in the 2014-2018 period, this study draws the conclusion that OE and solvency exist an inversely proportional relationship. In other words, a strong OE possibly causes low solvency and vice versa. From the above-mentioned results, it is possible to come up with some policy implications for logistics JSCs. Firstly, these surveyed firms should restructure their total assets. This movement aims at controlling assets turnover and raising these companies’ solvency. A suitable structure for assets shows a reasonable usage of capital and helps a company save its costs of mobilizing capital and guarantee their solvency. This also means a company should promote mobilized capital for its business operation or expands its scale of capital and assets as well. In other words, a reasonable structure of assets is a necessary condition. Secondly, low OE firms should raise more debts to balance between their performance and solvency. Raising more debts helps these companies tighten their cost by tax shield and keep their business performance. This would help companies improve their operational activities considerably and stabilize their profitability. By researching the relation between OE with solvency of logistics JSCs listed on the Vietnam Stock market, this paper contributes both certain theoretical (clarifying their links) and practical content (giving solutions to increase both OE and solvency). However, as the data of these logistics JSCs being are only take from a five-year period, it is not long enough to have a large research sample. A related study in the future may be undertaken including more industries rather than the logistics field only and a study over a longer period of time would give out a more precise prediction the relationship between OE and solvency, and more elements which affect them in companies. QUICK = -8.99 – 4.53TAT + 1.98* ET + 20.71ER QUICK = -8.99 – 4.53TAT + 1.98* ET + 20.71ER QUICK = -8.99 – 4.53TAT + 1.98* ET + 20.71*ER Table 7 and Table 8 show that the independent variable TAT is inversely proportional to both solvency ratios: Overall ratio and Quick ratio and has at least 99% statistical meaning. This result is consistent with the previous research conclusion of (Raykov, 2017) and (Kyule, 2015). If a logistics company increased the turnover of its total assets to enhance its OE, this firm would have to bear the risk of reducing its solvency. In other words, the more efficient in the use of assets is, the more limited its ability to pay debts and vice versa. Secondly, equity turnover (ET) and ER are necessary yet sufficient conditions to increase solvency. Generally, there is a direct proportion between ET, ER and solvency; or big scales of capital and highly independent JSCs are premises for the creation of a high solvency. This is proven when two independent variables ET and ER are directly proportional with both solvency ratios and have at least 99% statistical meaning. However as analyzed above, increasing turnover of the total assets reduced the solvency of the enterprises significantly. For logistics JSCs, despite having a large capital scale and high degree of financial independence, these enterprises allocated a large amount of companies’ resources on enhancing OE by promoting total assets turnover. Although the efficiency of operations increases, the solvency decreases due to over-utilized resources. This gives rise to payment risk for these businesses. From the results of qualitative and quantitative approach in thirty logistics JSCs, it can be concluded that high performance can lead to low solvency and vice versa. Table 9. Hypothesis test’s results Hypothesis Contents Results H1 Scale of company has a positive impact on Operational efficiency Accepted H2 Degree of financial independence has a positive impact on Operational efficiency Rejected H3 Total assets has a positive impact on Operational efficiency Accepted H4 Total sales has a positive impact on Operational efficiency Accepted H5 Operational efficiency has a significant and positive relationship on Solvency Accepted 5. CONCLUSION AND RECOMMENDATION This research is funded by National Economics University, Hanoi, Vietnam. Table 8. Regression QUICK with TAT, ER and ET Table 8. Regression QUICK with TAT, ER and IJMRA, Volume 4 Issue 12 December 2021 Page 1817 The Effect of Factors on Operational Efficiency and the Relationship between Operational Efficiency and Solvency of Logistics Joint-Stock Companies 6. ACKNOWLEDGEMENT This research is funded by National Economics University, Hanoi, Vietnam. IJMRA, Volume 4 Issue 12 December 2021 IJMRA, Volume 4 Issue 12 December 2021 www.ijmr Page 1818 Page 1818 www.ijmra.in REFERENCES The interrelationship between culture, capital structure, and performance: evidence from European retailers. Journal of business research, 50(2), 185-191. 9) Hartman, D. How to Calculate the Maximization of Operational Efficiency. cited 30/07/2016 from http://smallbusiness. chron. com/calculate-maximization-operational-efficiency-35448. html. 06). Modeling the relationship between financial indicators and company performance. An empirical study 10) Höbarth, L. L. (2006). Modeling the relationship between financial indicators and company performance. An empirical study for US-listed companies (Doctoral dissertation, WU Vienna University of Economics and Business). 11) Kokemuller, N. Operational Effectiveness Vs. Strategic Positioning. cited 30/07/2016 from http://www. ehow. com/info_8447934_operational-effectiveness-vs-strategic-positioning. html. 12) Kyule, J. (2015). Impact of Liquidity And Solvency On Financial Performance Of Firms Listed At The Nairobi Securities Exchange. Unpublished MSc Finance Project, University of Nairobi, 6(3), 526-533. 13) Lauterbach, B., & Vaninsky, A. (1999). Ownership structure and firm performance: Evidence from Israel. Journal of Management and Governance, 3(2), 189-201. 14) Memon, F., Bhutto, N. A., & Abbas, G. (2012). Capital structure and firm performance: A case of textile sector of Pakistan. Asian Journal of Business and Management Sciences, 1(9), 9-15. 15) Mudambi, R., & Nicosia, C. (1998). Ownership structure and firm performance: evidence from the UK financial services industry. Applied Financial Economics, 8(2), 175-180. 16) Nguyen Van Cong, Business Operation Analysis Textbook, National Economics Publisher, Hanoi, 2009 17) Onaolapo, A. A., & Kajola, S. O. (2010). Capital structure and firm performance: Evidence from Nigeria. European Journal of Economics, Finance and Administrative Sciences, 25(1), 70-82. 18) Pentikäinen, T. (1952). On the net retention and solvency of insurance companies. Scandinavian Actuarial Journal, 1952(1- 2), 71-92. 19) Raykov, E. (2017). The liquidity-profitability trade-off in Bulgaria in terms of the changed financial management functions during crisis. Management: journal of contemporary management issues, 22(1), 135-156. 20) Sandström, A. (2007). Solvency–a historical review and some pragmatic solutions. Mitteilungen der Schweizerischen Aktuarvereinigung, 1, 11-34 21) Siminica, M., Circiumaru, D., & Simion, D. (2012). The correlation between the return on assets and the measures of financial balance for Romanian companies. International journal of mathematical models and methods in applied sciences, 6(2), 232-253. 22) Tzelepis, D., & Skuras, D. (2004). The effects of regional capital subsidies on firm performance: an empirical study. Journal of Small Business and Enterprise Development. REFERENCES EFERENCES 1) Ahmad, R. (2016). A study of relationship between liquidity and profitability of standard charterd bank Pakistan: Analysis of financial statement approach. Global Journal of Management and Business Research. 1) Ahmad, R. (2016). A study of relationship between liquidity and profitability of standard charterd bank Pakistan: Analysis of financial statement approach. Global Journal of Management and Business Research. 1) Ahmad, R. (2016). A study of relationship between liquidity and profitability of standard charterd bank Pakistan: Analysis of financial statement approach. Global Journal of Management and Business Research. 2) Amer, H. H., Moustafa, W., & Eldomiaty, T. (2011, June). Determinants of operating efficiency for lowly and highly competitive banks in Egypt. In Cambridge Business and Economics Conference (Vol. 1, pp. 1-35). 2) Amer, H. H., Moustafa, W., & Eldomiaty, T. (2011, June). Determinants of operating efficiency for lowly and highly competitive banks in Egypt. In Cambridge Business and Economics Conference (Vol. 1, pp. 1-35). 3) Beal, V. (2016). Operational Efficiency. cited 30/07/2016 from http://www. webopedia. 3) Beal, V. (2016). Operational Efficiency. cited 30/07/2016 from http://www. webopedia. com/TERM/O/operational_efficiency. html. 4) Bernstein, L.A. (1988). Financial statement analysis, theory, application, and interpretation, Journal of Accounting Research, Vol. 3, 61-80. 4) Bernstein, L.A. (1988). Financial statement analysis, theory, application, and interpretation, Journal of Accounting Research, Vol. 3, 61-80. 5) Burrows, M. (2012). Operational Efficiency-It’s not just about Cost Cutting. cited 30/07/2016 from http://www. bsmreview. com/oppseff. shtml. 5) Burrows, M. (2012). Operational Efficiency-It’s not just about Cost Cutting. cited 30/07/2016 from http://www. bsmreview. com/oppseff. shtml. 6) Ehiedu, V. C. (2014). The impact of liquidity on profitability of some selected companies: The financial statement analysis (FSA) approach. Research Journal of Finance and Accounting, 5(5), 81-90. 6) Ehiedu, V. C. (2014). The impact of liquidity on profitability of some selected companies: The financial statement analysis (FSA) approach. Research Journal of Finance and Accounting, 5(5), 81-90. 7) Feng, C. M., & Wang, R. T. (2000). Performance evaluation for airlines including the consideration of financial ratios. Journal of Air Transport Management, 6(3), 133-142. 7) Feng, C. M., & Wang, R. T. (2000). Performance evaluation for airlines including the consideration of financial ratios. Journal of Air Transport Management, 6(3), 133-142. 8) Gleason, K. C., Mathur, L. K., & Mathur, I. (2000). The interrelationship between culture, capital structure, and performance: evidence from European retailers. Journal of business research, 50(2), 185-191. 8) Gleason, K. C., Mathur, L. K., & Mathur, I. (2000). REFERENCES 23) Varadan S., “What Is the Meaning of Operational Efficiency?”, 2016, cited 30/07/2016 from h // h / b l ff h l 23) Varadan S., “What Is the Meaning of Operational Efficiency?”, 2016, cited 30/07/2016 from http://www.ehow.com/about_6495198_meaning-operational-efficiency_.html>. 24) Zeitun, R., & Tian, G. G. (2007). Does ownership affect a firm's performance and default risk in Jordan?. Corporate Governance: The international journal of business in society. IJMRA, Volume 4 Issue 12 December 2021 Page 1819 IJMRA, Volume 4 Issue 12 December 2021 www.ijmra.in Page 1819 www.ijmra.in
https://openalex.org/W3200437266	https://europepmc.org/articles/pmc8484650?pdf=render	English	null	Acid-Sensing Ion Channels: Expression and Function in Resident and Infiltrating Immune Cells in the Central Nervous System	Frontiers in cellular neuroscience	2,021	cc-by	16,355	Abbreviations: AD, Alzheimer’s disease; ASIC, acid-sensing ion channel; BBB, blood–brain barrier; BCSFB, blood- cerebrospinal ﬂuid barrier; BMM, bone marrow-derived macrophage; BSCB, blood-spinal cord barrier; CCL-2, C-C motif chemokine ligand 2; CNS, central nervous system; CRAC, calcium release-activated calcium channel; CXCL-8, interleukin- 8; DAMP, damage-associated molecular pattern; DC, dendritic cell; Deg, degenerin; DRG, dorsal root ganglion; DS, lactate-based dialysis solution; EAE, experimental autoimmune encephalomyelitis; ENaC, epithelial sodium channel; HD, Huntington’s disease; ICAM-1, intercellular adhesion molecule 1; i.c.v., intracerebroventricular; IFN−γ, interferon gamma; IL, interleukin; LPS, lipopolysaccharide; MCT, monocarboxylate transporter; MHC, major histocompatibility complex; MPO, myeloperoxidase; MS, multiple sclerosis; NHE-1, sodium-hydrogen antiporter 1; NK, natural killer; NO, nitric oxide; NOS, nitric oxide synthase; OLC, oligodendrocyte lineage cell; PECAM-1, platelet endothelial cell adhesion molecule 1; PAMP, pathogen associated molecular pattern; PD, Parkinson’s disease; PNS, peripheral nervous system; PRR, pattern recognition receptor; RAW 264.7, mouse macrophage-like cell line; RIPK1, receptor-interacting serine/threonine-protein kinase 1; ROS, reactive oxygen species; SCI, spinal cord injury; SLE, systemic lupus erythematosus; SNP, single nucleotide polymorphism; TBI, traumatic brain injury; TCR, T cell receptor; TLE, temporal lobe epilepsy; TLR, Toll-like receptor; TNF, tumor necrosis factor; TRP, transient receptor potential. Acid-Sensing Ion Channels: Expression and Function in Resident and Inﬁltrating Immune Cells in the Central Nervous System Victoria S. Foster1, Lachlan D. Rash2, Glenn F. King1,3* and Michelle M. Rank4 1 Institute for Molecular Bioscience, The University of Queensland, St Lucia, QLD, Australia, 2 School of Biomedical Sciences, The University of Queensland, St Lucia, QLD, Australia, 3 Australian Research Council Centre of Excellence for Innovations in Peptide and Protein Science, The University of Queensland, St Lucia, QLD, Australia, 4 Anatomy and Physiology, Medicine, Dentistry and Health Sciences, The University of Melbourne, Melbourne, VIC, Australia Peripheral and central immune cells are critical for ﬁghting disease, but they can also play a pivotal role in the onset and/or progression of a variety of neurological conditions that affect the central nervous system (CNS). Tissue acidosis is often present in CNS pathologies such as multiple sclerosis, epileptic seizures, and depression, and local pH is also reduced during periods of ischemia following stroke, traumatic brain injury, and spinal cord injury. These pathological increases in extracellular acidity can activate a class of proton-gated channels known as acid-sensing ion channels (ASICs). ASICs have been primarily studied due to their ubiquitous expression throughout the nervous system, but it is less well recognized that they are also found in various types of immune cells. In this review, we explore what is currently known about the expression of ASICs in both peripheral and CNS-resident immune cells, and how channel activation during pathological tissue acidosis may lead to altered immune cell function that in turn modulates inﬂammatory pathology in the CNS. We identify gaps in the literature where ASICs and immune cell function has not been characterized, such as neurotrauma. Knowledge of the contribution of ASICs to immune cell function in neuropathology will be critical for determining whether the therapeutic beneﬁts of ASIC inhibition might be due in part to an effect on immune cells. Keywords: acid-sensing ion channel (ASIC), central nervous system, immune cell, acidosis, neuropathology, neuroimmunology, ion channel Edited by: Enrique Soto, Meritorious Autonomous University of Puebla, Mexico Reviewed by: Zhigang Xiong, Morehouse School of Medicine, United States Marcelo D. Carattino, University of Pittsburgh, United States Correspondence: Glenn F. King glenn.king@imb.uq.edu.au Correspondence: Glenn F. King glenn.king@imb.uq.edu.au Specialty section: This article was submitted to Cellular Neurophysiology, a section of the journal Frontiers in Cellular Neuroscience Specialty section: This article was submitted to Cellular Neurophysiology, a section of the journal Frontiers in Cellular Neuroscience Received: 08 July 2021 Accepted: 30 August 2021 Published: 17 September 2021 Citation: Foster VS, Rash LD, King GF and Rank MM (2021) Acid-Sensing Ion Channels: Expression and Function in Resident and Inﬁltrating Immune Cells in the Central Nervous System. Front. Cell. Neurosci. 15:738043. doi: 10.3389/fncel.2021.738043 Keywords: acid-sensing ion channel (ASIC), central nervous system, immune cell, acidosis, neuropathology, neuroimmunology, ion channel Keywords: acid-sensing ion channel (ASIC), central nervous system, immune cell, acidosis, neuropathology, neuroimmunology, ion channel Received: 08 July 2021 Accepted: 30 August 2021 Published: 17 September 2021 REVIEW published: 17 September 2021 doi: 10.3389/fncel.2021.738043 published: 17 September 2021 doi: 10.3389/fncel.2021.738043 INFLAMMATION There is a growing body of literature suggesting that ASICs may contribute to immune cell function and neuroinﬂammation during CNS pathology. Inﬂammation is a complex biological response to an insult, which may be pathogenic or self-derived (i.e., induced by trauma, ischemia, or autoimmune processes). It is driven by the innate immune system, dependent on immune cell inﬁltration, and results in the release of plethora of inﬂammatory mediators including histamines, bradykinins, arachidonic acid, leukotrienes, prostaglandins, cytokines and chemokines. Plasma leakage from capillary beds causes swelling, alongside extravasation of granulocytes into the tissue which is facilitated by P-selectin and platelet endothelial cell adhesion molecule 1 (PECAM-1) (Woodﬁn et al., 2007). Chronic inﬂammation causes long term alterations to cell populations, such as an increase in the number of white blood cells such as lymphocytes. This inﬁltration and persistent involvement of immune cells results in both healing alongside further damage. Neuroinﬂammation is an inﬂammatory response within the CNS, leading to immune cell inﬁltration (of CNS or peripheral origin) and increased levels of cytokines, chemokines, and reactive oxygen species (ROS). It is worth noting that inﬂammation can be “sterile,” occurring without a pathogenic external source [e.g., pathogen associated molecular patterns (PAMPs)] and instead resulting from damage-associated molecular patterns (DAMPs) triggered by trauma, ischemia or other environmental factors (e.g., ultraviolet radiation) (Rock et al., 2010; Feldman et al., 2015). ASICs are expressed in a range of tissues and have been associated with diverse pathologies, including diabetes, stroke, myocardial infarction, and epilepsy (see Figure 1; Xiong et al., 2004; Lv et al., 2011; Radu et al., 2014; Chassagnon et al., 2017; Redd et al., 2021). Most ASIC subtypes are present in the peripheral nervous system (PNS). However, ASIC1a is expressed at high levels and is the dominant subtype in both the human and rodent central nervous system (CNS) (Sluka et al., 2003; Wemmie et al., 2003; Baron et al., 2008), and it has been implicated in a variety of CNS disorders such as neurodegenerative diseases, Neuroinﬂammation has been implicated in degenerative and traumatic conditions and even mental health disorders such as depression and anxiety. Despite robust intrinsic CNS barriers, resident peripheral immune cells frequently cross into the CNS parenchyma when the blood-brain barrier (BBB), blood- spinal cord barrier (BSCB), or blood-cerebrospinal ﬂuid barrier (BCSFB) become porous. Citation: Foster VS, Rash LD, King GF and Rank MM (2021) Acid-Sensing Ion Channels: Expression and Function in Resident and Inﬁltrating Immune Cells in the Central Nervous System. Front. Cell. Neurosci. 15:738043. doi: 10.3389/fncel.2021.738043 September 2021 \| Volume 15 \| Article 738043 1 Frontiers in Cellular Neuroscience \| www.frontiersin.org ASICs in CNS Immune Cells Foster et al. ACID-SENSING ION CHANNELS depression, epilepsy, and ischemia-induced injury of the brain and spinal cord (Xiong et al., 2004; Friese et al., 2007; Wong et al., 2008; Ziemann et al., 2008; Coryell et al., 2009; Hu et al., 2011; Lv et al., 2011; Radu et al., 2014; Koehn et al., 2016; Chassagnon et al., 2017). Acid-sensing ion channels (ASICs) are proton-gated ion channels that are permeable to Na+ (Waldmann et al., 1997), and they constitute a subfamily of the epithelial sodium channel/degenerin (ENaC/Deg) superfamily (Kellenberger and Schild, 2015). ENaCs facilitate Na+ reabsorption in the kidney, and they regulate the volume in the ﬂuid/cilia interface in both the lung and colonic epithelial cells (Garty and Palmer, 1997; Enuka et al., 2012; Hanukoglu and Hanukoglu, 2016). ASICs have ∼30% sequence identity to ENaCs, and both groups are inhibited by the diuretic drug amiloride (Paukert et al., 2004). However, in contrast to ENaCs, ASICs appear to have evolved earlier, possibly ﬁrst appearing in deuterostomes ∼600 million years ago (Lynagh et al., 2018). Humans possess four ASIC-coding genes (ASIC1–ASIC4), three of which (ASIC1, ASIC2 and ASIC3) are alternatively spliced, to produce six main subunits with diﬀering properties: ASIC1a, ASIC1b, ASIC2a, ASIC2b, ASIC3a, and ASIC4 (Table 1; Wemmie et al., 2013). ASIC3a is the major splice isoform of ASIC3, with little known about the function of ASIC3b and ASIC3c (Delaunay et al., 2012). ASICs form homotrimeric or heterotrimeric channels (Jasti et al., 2007) that have diﬀerent pH thresholds for channel activation (see Table 1), diﬀerent physiological and pathological roles, and diﬀerent tissue distribution (Figure 1; Waldmann et al., 1997, 1999; Chen et al., 1998). ASIC1a homomers, ASIC1a/2b heteromers (but not ASIC1a/2a), and human ASIC1b mediate the inﬂux of Ca2+ in addition to Na+, although they are all more permeable to Na+ (Waldmann et al., 1997; Olena et al., 2004; Hoagland et al., 2010; Sherwood et al., 2011). INFLAMMATION The BBB, BSCB, and BCSFB are often disrupted and left partially open for weeks after a physical injury such as spinal cord injury (SCI) and traumatic brain injury (TBI), or after ischemic insults such as stroke (Sinescu et al., 2010; Chodobski et al., 2011; Suzuki et al., 2016). Degenerative disorders such as Alzheimer’s disease (AD), Parkinson’s disease (PD), and multiple sclerosis (MS) also cause pervasive disruption to CNS barriers that results in CNS invasion by peripheral immune cells (Ffrench-Constant, 1994; Town et al., 2005; Polman et al., 2006). Breakdown of the BBB is also thought to contribute to the etiology of epilepsy (Marchi et al., 2012). TABLE 1 \| ASIC subtypes and their pH sensitivity. Gene Subtype Alternative names pH sensitivity (pH50)1 ASIC1 ASIC1a ASIC, ASICα, BNaC2(α) 5.8–6.8 ASIC1b ASICβ, BNaC2(β) 6.1–6.2 ASIC2 ASIC2a BNaC1(α), MDEG, BNC(1a) 4.5–4.9 ASIC2b BNaC1(β), MDEG2 Does not form pH-sensitive homomers; associates with other ASICs to form pH-sensitive channels ASIC3 ASIC3 DRASIC, TNaC1, SLNAC1 6.4–6.6 ASIC4 ASIC4 BNAC4, ACCN4, SPASIC Does not form pH-sensitive homomeric channels 1The pH sensitivities of the homomeric ASIC channels are from Wemmie et al. (2013). Neutrophils, macrophages, T cells, and dendritic cells (DCs) have each been shown to occupy the CNS after an insult. Neutrophil chemoattractants such as PGF2α, complement component C5a, and interleukin-8 (CXCL-8) are produced by September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org 2 ASICs in CNS Immune Cells Foster et al. FIGURE 1 \| Schematic showing tissues in which ASICs are expressed (green boxes) and their putative role in various pathologies (orange boxes) in (A) rodents and (B) humans. Data summarized here are drawn from Supplementary Table 1, with expression sites and pathology based on studies that used antibody staining, transcriptomic data, proteomics and/or functional methods. Additional information is available in the Human Protein Atlas (www.proteinatlas.org). References supporting the putative roles of ASICs in human pathologies are as follows: 1Brain: ASIC1a detected using RT-PCR and Western blot; acidosis-mediated damage to cultured human brain neurons rescued by ASIC1a inhibition with PcTx1 (Li et al., 2010). Signiﬁcant association between ASIC1a SNPs, amygdala volume, and panic disorder in humans (Smoller et al., 2014). Elevated levels of ASIC1a (Yang et al., 2016) and ASIC3 (Cao et al., 2016) in brains of patients with temporal lobe epilepsy shown using immunostaining and Western blot. INFLAMMATION 8Heart: Transcriptomics revealed presence of ASIC1a in human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs); ASIC1a inhibition with Hi1a or PcTx1 improved viability of hiPSC-CMs under conditions of hypoxia and acidosis, whereas viability was further reduced in the presence of the ASIC1a activator MitTx (Redd et al., 2021). 9Eye: ASIC1a detected in retinal pigment epithelial cells using RT-PCR and Western blotting; PcTx1 protects cells from oxidative stress (Tan et al., 2013). ASIC3 mRNA detected using RT-PCR on human retina samples (Maubaret et al., 2002). 10Ear: Presence of ASIC4 shown using RT-PCR and Northern blot (Gründer et al., 2000). ischemic tissues, suggesting that these leukocytes do not just passively ﬁlter through the disrupted BBB/BSCB but are actively recruited to the CNS following the insult (Kilgore et al., 1994; Arnould et al., 2001; De Bondt et al., 2020). Consistent with this, neutrophil migration is altered by acidity, which is discussed further below (Rotstein et al., 1988). There is conﬂicting evidence as to whether neutrophils are beneﬁcial or detrimental once they have inﬁltrated the CNS, which may be due to the heterogenous nature of neutrophil activation states (e.g., diﬀerential expression of proteins), but they are accepted to contribute to SCI severity (Taoka et al., 1998; Neirinckx et al., 2014; Deniset and Kubes, 2018). Inﬁltrating macrophages/monocytes and possibly neutrophils initiate demyelination of CNS neurons through phagocytic and inﬂammatory processes (Ajami et al., 2011; Yamasaki et al., 2014; Stranahan et al., 2016; De Bondt et al., 2020). T cells that cross a porous BBB contribute signiﬁcantly to MS pathology by causing central demyelination through attack and degradation of the myelin sheath (Bielekova et al., 2004). During experimental autoimmune encephalomyelitis (EAE), an animal model of MS, DCs also breach the CNS and subsequently prime T cells to exacerbate the autoimmune inﬂammation (Karman et al., 2006; Sagar et al., 2012). In summary, recruitment of peripheral immune cells into the CNS during inﬂammation contributes to the wave of secondary damage that emanates from the initial site of injury. INFLAMMATION ASIC expression in CNS; treatment with amiloride alleviated patient symptoms of MS (Arun et al., 2013). ASIC1 and ASIC3 found in glioblastoma stem cell lines using RT-PCR and Western blot; microarray data revealed that ASIC expression is associated with improved survival in glioma patients (Tian et al., 2017). 2Lung: ASICs expressed in human lung cancer cell line A549 determined using RT-PCR, immunoﬂuorescence, and Western blot. Proliferation and migration promoted by overexpression of ASIC1a and inhibited by PcTx1 (Wu et al., 2017). 3Spinal cord: ASICs in human spinal cord, stained with ASIC1 antibody (Arun et al., 2013). 4Bone: ASIC expression in human skeleton shown using RT-PCR and antibody staining performed on human chondrocytes (Jahr et al., 2005). 5Testis: ASIC3 in human testis based on Northern blot (Ishibashi and Marumo, 1998). 6Gut: ASIC2 measured using RT-PCR and immunostaining in human colorectal cancer cells, with ASIC2 promoting cell invasion and proliferation in xenografts (worsened via overexpression and impeded with ASIC2 knockout) (Zhou et al., 2017). 7Kidney: ASICs found in human proximal tubular cell line using RT-PCR and Western blotting; apoptosis of cells due to ischemia-reperfusion injury reduced by ASIC1a inhibition using PcTx1 (Song et al., 2019). ASIC1, ASIC2 and ASIC3 protein expression in patients with Henoch-Schönlein purpura nephritis; ASIC blockade with amiloride reduced expression of damage marker proteins (Yuan et al., 2010). 8Heart: Transcriptomics revealed presence of ASIC1a in human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs); ASIC1a inhibition with Hi1a or PcTx1 improved viability of hiPSC-CMs under conditions of hypoxia and acidosis, whereas viability was further reduced in the presence of the ASIC1a activator MitTx (Redd et al., 2021). 9Eye: ASIC1a detected in retinal pigment epithelial cells using RT-PCR and Western blotting; PcTx1 protects cells from oxidative stress (Tan et al., 2013). ASIC3 mRNA detected using RT-PCR on human retina samples (Maubaret et al., 2002). 10Ear: Presence of ASIC4 shown using RT-PCR and Northern blot (Gründer et al., 2000). FIGURE 1 \| Schematic showing tissues in which ASICs are expressed (green boxes) and their putative role in various pathologies (orange boxes) in (A) rodents and (B) humans. Data summarized here are drawn from Supplementary Table 1, with expression sites and pathology based on studies that used antibody staining, transcriptomic data, proteomics and/or functional methods. Additional information is available in the Human Protein Atlas (www.proteinatlas.org). INFLAMMATION References supporting the putative roles of ASICs in human pathologies are as follows: 1Brain: ASIC1a detected using RT-PCR and Western blot; acidosis-mediated damage to cultured human brain neurons rescued by ASIC1a inhibition with PcTx1 (Li et al., 2010). Signiﬁcant association between ASIC1a SNPs, amygdala volume, and panic disorder in humans (Smoller et al., 2014). Elevated levels of ASIC1a (Yang et al., 2016) and ASIC3 (Cao et al., 2016) in brains of patients with temporal lobe epilepsy shown using immunostaining and Western blot. ASIC expression in CNS; treatment with amiloride alleviated patient symptoms of MS (Arun et al., 2013). ASIC1 and ASIC3 found in glioblastoma stem cell lines using RT-PCR and Western blot; microarray data revealed that ASIC expression is associated with improved survival in glioma patients (Tian et al., 2017). 2Lung: ASICs expressed in human lung cancer cell line A549 determined using RT-PCR, immunoﬂuorescence, and Western blot. Proliferation and migration promoted by overexpression of ASIC1a and inhibited by PcTx1 (Wu et al., 2017). 3Spinal cord: ASICs in human spinal cord, stained with ASIC1 antibody (Arun et al., 2013). 4Bone: ASIC expression in human skeleton shown using RT-PCR and antibody staining performed on human chondrocytes (Jahr et al., 2005). 5Testis: ASIC3 in human testis based on Northern blot (Ishibashi and Marumo, 1998). 6Gut: ASIC2 measured using RT-PCR and immunostaining in human colorectal cancer cells, with ASIC2 promoting cell invasion and proliferation in xenografts (worsened via overexpression and impeded with ASIC2 knockout) (Zhou et al., 2017). 7Kidney: ASICs found in human proximal tubular cell line using RT-PCR and Western blotting; apoptosis of cells due to ischemia-reperfusion injury reduced by ASIC1a inhibition using PcTx1 (Song et al., 2019). ASIC1, ASIC2 and ASIC3 protein expression in patients with Henoch-Schönlein purpura nephritis; ASIC blockade with amiloride reduced expression of damage marker proteins (Yuan et al., 2010). 8Heart: Transcriptomics revealed presence of ASIC1a in human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs); ASIC1a inhibition with Hi1a or PcTx1 improved viability of hiPSC-CMs under conditions of hypoxia and acidosis, whereas viability was further reduced in the presence of the ASIC1a activator MitTx (Redd et al., 2021). 9Eye: ASIC1a detected in retinal pigment epithelial cells using RT-PCR and Western blotting; PcTx1 protects cells from oxidative stress (Tan et al., 2013). ASIC3 mRNA detected using RT-PCR on human retina samples (Maubaret et al., 2002). 10Ear: Presence of ASIC4 shown using RT-PCR and Northern blot (Gründer et al., 2000). INFLAMMATION FIGURE 1 \| Schematic showing tissues in which ASICs are expressed (green boxes) and their putative role in various pathologies (orange boxes) in (A) rodents and (B) humans. Data summarized here are drawn from Supplementary Table 1, with expression sites and pathology based on studies that used antibody staining, transcriptomic data, proteomics and/or functional methods. Additional information is available in the Human Protein Atlas (www.proteinatlas.org). References supporting the putative roles of ASICs in human pathologies are as follows: 1Brain: ASIC1a detected using RT-PCR and Western blot; acidosis-mediated damage to cultured human brain neurons rescued by ASIC1a inhibition with PcTx1 (Li et al., 2010). Signiﬁcant association between ASIC1a SNPs, amygdala volume, and panic disorder in humans (Smoller et al., 2014). Elevated levels of ASIC1a (Yang et al., 2016) and ASIC3 (Cao et al., 2016) in brains of patients with temporal lobe epilepsy shown using immunostaining and Western blot. ASIC expression in CNS; treatment with amiloride alleviated patient symptoms of MS (Arun et al., 2013). ASIC1 and ASIC3 found in glioblastoma stem cell lines using RT-PCR and Western blot; microarray data revealed that ASIC expression is associated with improved survival in glioma patients (Tian et al., 2017). 2Lung: ASICs expressed in human lung cancer cell line A549 determined using RT-PCR, immunoﬂuorescence, and Western blot. Proliferation and migration promoted by overexpression of ASIC1a and inhibited by PcTx1 (Wu et al., 2017). 3Spinal cord: ASICs in human spinal cord, stained with ASIC1 antibody (Arun et al., 2013). 4Bone: ASIC expression in human skeleton shown using RT-PCR and antibody staining performed on human chondrocytes (Jahr et al., 2005). 5Testis: ASIC3 in human testis based on Northern blot (Ishibashi and Marumo, 1998). 6Gut: ASIC2 measured using RT-PCR and immunostaining in human colorectal cancer cells, with ASIC2 promoting cell invasion and proliferation in xenografts (worsened via overexpression and impeded with ASIC2 knockout) (Zhou et al., 2017). 7Kidney: ASICs found in human proximal tubular cell line using RT-PCR and Western blotting; apoptosis of cells due to ischemia-reperfusion injury reduced by ASIC1a inhibition using PcTx1 (Song et al., 2019). ASIC1, ASIC2 and ASIC3 protein expression in patients with Henoch-Schönlein purpura nephritis; ASIC blockade with amiloride reduced expression of damage marker proteins (Yuan et al., 2010). Immune Cell Response to Tissue Acidiﬁcation Although often referred as neuronal ion channels, ASICs are expressed in numerous types of immune cells (Huang et al., 2010; Tong et al., 2011; Kong et al., 2013; Yu et al., 2015) and they are potentiated by molecules associated with neuroinﬂammation such as arachidonic acid, histamine, lactate, and nitric oxide (Immke and McCleskey, 2001; Cadiou et al., 2007; Smith et al., 2007; Rash, 2017; Shteinikov et al., 2017). Acidity can provoke diﬀerent responses in immune cells depending on the source of the acidosis, as exempliﬁed in Table 2 for macrophages. Therefore, whether or how inﬂammation is triggered by acidity depends on the source of the lowered pH. As proton-coupled monocarboxylate transporters (MCTs) shuttle lactate in a 1:1 ratio with H+ (Hertz and Dienel, 2005), altering lactate levels will have secondary eﬀects on the H+ concentration. Acidiﬁcation is not solely linked to traumatic and ischemic insults. Extracellular acidosis has been reported in spinal cord tissue of EAE mice, and increased levels of tissue lactate have also been reported in human MS brain tissue, suﬃcient to activate ASIC1a (Bitsch et al., 1999; Friese et al., 2007). In both animal models of Huntington’s disease (HD) and in patients with HD, a signiﬁcant build-up of lactic acid has been observed in the brain (Tsang et al., 2006; Josefsen et al., 2010). Neutrophils exposed to acidic pH undergo various functional changes, including signiﬁcantly impaired migration, defective chemotaxis, reduced speed of apoptosis, increased CD18 expression, and increased phagocytosis but with decreased bactericide action (Rotstein et al., 1988; Trevani et al., 1999; Tissue acidosis and inﬂammation inﬂuence one another and, as shown by research within an intensive care unit, alterations in the acid/base status of patients contribute to diﬀerences in their interleukin (IL) and cytokine proﬁle (Zampieri et al., 2014). We have highlighted thus far that both inﬂammation and acidity contribute to tissue damage in pathological CNS conditions, with the former moderated by immune cells and the latter by receptors such as ASICs. Immune cells themselves are also modulated by acidity, and in the following sections we present an overview of immune cells, their response to acidic conditions, and the potential role of immune-cell ASICs in neuroinﬂammation. TABLE 2 \| The effects of pH on inﬂammatory mediators in macrophages (Kellum et al., 2004). TISSUE ACIDIFICATION Vascular damage after SCI, hemorrhagic stroke, or occlusion of cerebral arteries during an ischemic stroke, causes a marked reduction in blood ﬂow to the injured region. Neurons and supporting glial cells in the center of the ischemic territory (i.e., where blood ﬂow is lowest) are rapidly, and perhaps irreparably, September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org Frontiers in Cellular Neuroscience \| www.frontiersin.org 3 ASICs in CNS Immune Cells Foster et al. damaged. At the periphery of this ischemic core (the penumbra), an expanding wave of secondary damage (i.e., progression of injury beyond the initial insult, as outlined in Figure 2) develops more slowly because supplementary blood ﬂow from adjacent regions maintains perfusion above the threshold for immediate cell death (Moskowitz et al., 2010). However, in the absence of adequate blood ﬂow, oxygen deprivation (hypoxia) in areas surrounding the primary ischemic insult forces neurons to resort to anaerobic glycolysis for their energy needs, which causes lactic acidosis and acidiﬁcation of the tissue (Rehncrona, 1985). The resultant acidosis contributes to secondary damage through activation of ASICs (e.g., ASIC1a), which is discussed further below. Patients admitted with trauma often exhibit a decrease in blood pH as part of the “triad of death”—systemic hypothermia, acidosis, and coagulopathy—which contributes to mortality and is also a reminder that in trauma, acidiﬁcation is not limited to the CNS or just the site of damage (Mikhail, 1999; Mitra et al., 2012). and BCSFB allow the CNS to maintain its own regulatory environment with a distinct set of immune cells comprising the neuroglia (microglia, astrocytes, oligodendrocytes) and ependymal cells (Figure 4). Immune Cell Response to Tissue Acidiﬁcation Acid pHo1 Cells LPS2 Effect References HCl 6.5 Alveolar macrophages + ↑TNF3 mRNA Heming et al., 2001a HCl 5.5 Alveolar macrophages + ↑TNF mRNA ↓TNF secretion Heming et al., 2001a HCl 5.5 RAW 264.7 cell line4 + No 1TNF mRNA ↓TNF secretion Heming et al., 2001b HCl 7.0 Alveolar macrophages + ↓TNF secretion Bidani et al., 1998 HCl 7.0 Peritoneal macrophages – ↑NO5, ↑TNF6, ↑NF-κB7 Bellocq et al., 1998 HCl 7.2 RAW 264.7 cell line + ↑NO Huang et al., 2002 Lactate 6.7 Peritoneal macrophages + ↑TNF mRNA ↑TNF secretion Jensen et al., 1990 DS8 6.0 Peritoneal macrophages + ↓TNF mRNA ↓TNF secretion Jörres et al., 1995 DS8 6.5 Human blood-borne macrophages + ↓TNF mRNA ↓NF-κB Douvdevani et al., 1995 1pHo, extracellular pH. 2LPS, lipopolysaccharide. 3TNF, tumor necrosis factor. 4RAW 264.7, mouse macrophage-like cell line. 5NO, nitric oxide. 6TNF not measured directly. 7NF-κB, nuclear factor-κB. 8DS, lactate-based dialysis solution. TABLE 2 \| The effects of pH on inﬂammatory mediators in macrophages (Kellum et al., 2004). Frontiers in Cellular Neuroscience \| www.frontiersin.org September 2021 \| Volume 15 \| Article 738043 OVERVIEW OF IMMUNE CELLS White blood cell lineages (peripheral immune cells) arise from hemopoietic stem cells in the bone marrow (Figure 3), in particular the common myeloid progenitor and the common lymphoid progenitor cells. Common myeloid progenitors give rise to granulocytes (mast cells, eosinophils, basophils and neutrophils) and monocytes (which can diﬀerentiate into macrophages and DCs), while the common lymphoid progenitor produces various classes of lymphocytes: natural killer (NK) cells, T cells and B cells (Kondo et al., 2003; Figure 3). T cells are produced in the bone marrow and mature in the thymus, and they can be further subdivided into memory, cytotoxic, regulatory, and helper T cells. T cells express T cell receptors (TCRs) on their surface to recognize antigens. Peripheral immune cells are separated from CNS immune cells by two distinct barriers that divide the central and peripheral vascular systems: the BBB/BSCB and the BCSFB. The BBB/BSCB Frontiers in Cellular Neuroscience \| www.frontiersin.org September 2021 \| Volume 15 \| Article 738043 4 ASICs in CNS Immune Cells Foster et al. FIGURE 2 \| Progression of secondary damage after physical damage is inﬂicted on the spinal cord. Blood ﬂow is restricted as a result of damaged blood vessels, causing tissue acidiﬁcation and subsequent activation of ASICs. FIGURE 3 \| Immune cell lineages in the periphery. Cao et al., 2015). The combination of acidosis-induced decreases in migration chemotaxis and apoptosis together with increased (MPO) produced by neutrophils enhances conversion of H2O2 to HOCl and thus contributes to the secondary damage after FIGURE 2 \| Progression of secondary damage after physical damage is inﬂicted on the spinal cord. Blood ﬂow is restricted as a result of damaged blood vessels, causing tissue acidiﬁcation and subsequent activation of ASICs. ndary damage after physical damage is inﬂicted on the spinal cord. Blood ﬂow is restricted as a result of damaged blood vessels, subsequent activation of ASICs. FIGURE 2 \| Progression of secondary damage after physical damage is inﬂicted on the spinal cord. Blood ﬂow is restricted as a result of damaged blood vessels, causing tissue acidiﬁcation and subsequent activation of ASICs. \| Progression of secondary damage after physical damage is inﬂicted on the spinal cord. Blood ﬂow is restricted as a result of dam sue acidiﬁcation and subsequent activation of ASICs. FIGURE 3 \| Immune cell lineages in the periphery. FIGURE 3 \| Immune cell lineages in the periphery. FIGURE 3 \| Immune cell lineages in the periphery. OVERVIEW OF IMMUNE CELLS Cao et al., 2015). The combination of acidosis-induced decreases in migration, chemotaxis and apoptosis together with increased phagocytic activity might suggest that once neutrophils enter a region of low pH, they become trapped there with enhanced lifespan and inﬂammatory activity. CD18 is a cell-surface molecule expressed by neutrophils that promotes cell adhesion, a process that allows neutrophils to aid tissue repair in the periphery. Altered expression of CD18 is believed to facilitate neutrophil breach into the CNS (Serrano et al., 1996; Neri et al., 2018). The acidic blood pH after trauma (as low as 6.6) likely induces altered CD18 expression on peripheral neutrophils which then undergo facilitated extravasation into the CNS, where they accumulate in acidic regions (Mitra et al., 2012). Furthermore, in patients with SCI, neutrophils exhibit a higher “respiratory burst” in the blood (Kanyilmaz et al., 2013). The myeloperoxidase (MPO) produced by neutrophils enhances conversion of H2O2 to HOCl, and thus contributes to the secondary damage after SCI (Kubota et al., 2012). Neutrophils in the blood of SCI patients in rehabilitation have decreased ability to phagocytose Escherichia coli and Staphylococcus aureus (Campagnolo et al., 1997; Kanyilmaz et al., 2013). These data suggest that, at least for neutrophils, typical immune responses are altered in the presence of acidity and this may explain why neutrophils can have a detrimental impact in the CNS after neurotrauma. Vermeulen et al. (2004) and Martínez et al. (2007) found that murine and human DCs mature upon exposure to acidity. Acidosis modiﬁes the inﬂammatory proﬁle of monocytes and macrophages, and is proposed to induce an inﬂammatory state in the latter (Riemann et al., 2016). Microglial motility decreases after exposure to acidic pH (Faﬀand Nolte, 2000). September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org Frontiers in Cellular Neuroscience \| www.frontiersin.org 5 ASICs in CNS Immune Cells Foster et al. microglia in CNS pathologies is an emergent ﬁeld and has been recently reviewed (Norris and Kipnis, 2019; Butler et al., 2021). Rat macrophages exposed to acidic environments (pH below 7.4) upregulate nitric oxide synthase (NOS), and thus increase production of NO (Bellocq et al., 1998). Excess amounts of NO are detrimental in neurodegenerative pathologies such as HD and after striatal lesions, causing DNA damage and oxidative stress to cells (Schulz et al., 1996; Browne and Beal, 2006; Calabrese et al., 2007). OVERVIEW OF IMMUNE CELLS These responses are discussed further below when considering the role of ASICs. y p It has been known for some time that the severity of disease progression in MS patients is linked to microglial activation (Heppner et al., 2005; Jack et al., 2005; Geladaris et al., 2021), and it seems that microglial phagocytosis of stressed neurons contributes to neurodegeneration in AD, PD, ischemic stroke, CNS viral infections, and aging (Butler et al., 2021). Rat microglia have been shown to express ASIC1, ASIC2a and ASIC3. In cultured primary rat microglia stimulated with lipopolysaccharide (LPS), a known immune activator, ASIC1 and ASIC2a expression increased both at the surface and intracellularly (Yu et al., 2015) and ASIC-speciﬁc inward currents could be recorded using whole-cell patch- clamp electrophysiology. Both PcTx1 and amiloride reduced the amplitude of these currents and reduced the expression of inﬂammatory cytokines (Yu et al., 2015). Treatment of cultured primary microglia with amiloride also decreased the extent of phagocytosis under acidic conditions (Vergo et al., 2011). A decrease in phagocytosis was also seen in microglia cultured from ASIC1−/−mice, with no change upon the addition of amiloride, indicating that the eﬀect of amiloride is ASIC1- mediated. Decreased extracellular pH, therefore, can modify the function of immune cells by altering motility, impacting phagocytosis, aiding CNS invasion, and increasing the production of potentially neurodamaging agents such as NO. Although the function of neutrophils is modiﬁed by acidity, there are, to date, no studies on whether they express ASICs. Thus, the role of ASICs in acidosis-induced responses of neutrophils is currently unknown. Receptors and Ion Channels on Immune Cells The key immune cell receptors involved in the innate immune response are complement and cytokine/chemokine receptors and pattern recognition receptors (PRRs) such as Toll-like receptors (TLRs). The PRRs recognize both PAMPs and also DAMPs which are produced by dying or damaged cells and trigger sterile inﬂammatory processes. As part of the adaptive immune response, the TCRs and antibody-binding receptors are involved in cascades that facilitate the immune cell response. Immune cells also possess ion channels that are important for their function, including the voltage-gated potassium channel Kv1.3, the calcium-activated potassium channel KCa3.1, calcium release- activated calcium channels (CRAC), and transient receptor potential (TRP) channel M7 (Cadiou et al., 2007; Chandy and Norton, 2017; Chiang et al., 2017; Nadolni and Zierler, 2018; Clemens and Lowell, 2019). The function of these ion channels in immune cells and their contribution to pathology have been reviewed (Feske et al., 2015; Vaeth and Feske, 2018) but neither of these reviews discuss the presence or role of ASICs in immune cells. These data collectively show that ASICs, and ASIC1a speciﬁcally, are present in microglia and contribute to their function in acidic microenvironments. In turn, this indicates that microglial ASIC1a contributes to microglia function, aﬀecting release of inﬂammatory cytokines and phagocytosis, potentially impacting the immune response in the CNS during MS and many other CNS pathologies. Acid-Sensing Ion Channels in Astrocytes Astrocytes fulﬁll a variety of roles in the CNS, such as regulating the formation and maintenance of the BBB (Abbott et al., 2006). After brain injury and disease, astrocytes can become reactive, participating in repair cascades and gene upregulation, and forming an astroglial scar (Zamanian et al., 2012; Anderson et al., 2016). Reactive astrocytes are key players in CNS disease, an area reviewed recently by Dossi et al. (2018). Stimulation of astrocytes by cytokines such as interferon gamma (IFN−γ) leads to expression of the antigen-presenting major histocompatibility complex II (MHC II) on the cell surface alongside other costimulatory molecules such as intercellular adhesion molecule 1 (ICAM-1) (Wong et al., 1984; Lee et al., 2000). MHC II expression occurs in MS plaques and may contribute to the inﬂammatory cascade through antigen presentation to T cells (Zeinstra et al., 2000). Ponath et al. (2018) recently reviewed the diﬀering roles of astrocytes in MS. Frontiers in Cellular Neuroscience \| www.frontiersin.org d-Sensing Ion Channels in Microglia Acid Sensing Ion Channels in Microglia Microglia are the main immune cell type in the CNS. Although considered CNS-resident macrophages, microglia branch oﬀ earlier in development from yolk sac precursors (Ginhoux et al., 2010; Schulz et al., 2012; Kierdorf et al., 2013). In non-pathological conditions, microglia are involved in clearing tissue debris and surveying the environment (Nimmerjahn et al., 2005). During neuropathology, microglia can enter an activated state and release enzymes such as NADPH oxidase, which causes neuronal damage. Microglia can additionally modify myelination, aﬀecting developmental myelination as well as beneﬁcial re-myelination processes (Popovich et al., 2002; Butovsky et al., 2006). The divergent context-speciﬁc roles of ASIC1, ASIC2a, ASIC3 are expressed in the nucleus of rat astrocytes, and astrocytes express an ASIC-like current that is blocked by amiloride (Huang et al., 2010; Yu et al., 2015). Yang et al. (2016) noted increased ASIC1a expression in the membrane and cytoplasm of reactive astrocytes from the hippocampi of deceased TLE patients, ﬁndings which were replicated in a mouse model of chronic epileptogenesis. To explore normal astrocytic function, the group used primary cultures from wild-type mice. At 24 h after stimulation with LPS, astrocytes upregulated ASIC1a ASIC1, ASIC2a, ASIC3 are expressed in the nucleus of rat astrocytes, and astrocytes express an ASIC-like current that is blocked by amiloride (Huang et al., 2010; Yu et al., 2015). Yang et al. (2016) noted increased ASIC1a expression in the membrane and cytoplasm of reactive astrocytes from the hippocampi of deceased TLE patients, ﬁndings which were replicated in a mouse model of chronic epileptogenesis. To explore normal astrocytic function, the group used primary cultures from wild-type mice. At 24 h after stimulation with LPS, astrocytes upregulated ASIC1a September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org Frontiers in Cellular Neuroscience \| www.frontiersin.org 6 ASICs in CNS Immune Cells Foster et al. FIGURE 4 \| Immune cell lineages in the CNS. FIGURE 4 \| Immune cell lineages in the CNS. cytokines such as C-C motif chemokine ligand 2 (CCL-2) and IL- 8 in response to neuroborreliosis, the neurological manifestation of Lyme’s disease (Ramesh et al., 2012). IL-8 and CCL-2 are chemotaxic for a number of peripheral immune cells, such as neutrophils and T cells respectively (Turner et al., 2014). and exhibited increased Ca2+ inﬂux upon exposure to pH 6.0, which was signiﬁcantly reduced by PcTx1. Yang et al. d-Sensing Ion Channels in Microglia (2016) showed that seizure frequency in TLE mice is reduced by knockdown of astrocytic ASIC1a, with restoration of the ASIC1 gene increasing the frequency of spontaneous seizures. Thus, in contrast to what is believed regarding epilepsy and the beneﬁcial function of ASIC1a (likely via neuronal expression), these data suggest that activation of astrocytic ASIC1a causes pathological Ca2+ inﬂux and contributes to the pathogenesis of TLE. Oligodendrocytes are especially sensitive to ischemic insults (Bradl and Lassmann, 2010), and resident ASICs may contribute to ischemia-induced injury of these cells. ASIC1 is upregulated in oligodendrocytes in chronic brain lesions of patients with MS, and use of amiloride in patients provided neuroprotection at primary stages of the disease (Arun et al., 2013). ASIC1a, ASIC2a/b (non-selective primers were used, and thus ASIC2a and ASIC2b could not be distinguished), and ASIC4 mRNA were found in oligodendrocytes, and ASIC-speciﬁc inward currents recorded from oligodendrocyte lineage cells (OLCs) were inhibited by PcTx1 (Feldman et al., 2008). Cultured mouse oligodendrocytes express ASIC1 (ASIC1a and ASIC1b were not separable) and they were protected from acidosis-induced damage by PcTx1, as were oligodendrocytes derived from ASIC1 KO mice (Vergo et al., 2011). Thus, the combined data suggest Acid-Sensing Ion Channels in T Cells g T cells play a major role in coordinating and executing multiple functions of the adaptive immune response. Autoreactive T cells that have lost their ability to diﬀerentiate self (i.e., host cells and antigens) from non-self (i.e., foreign/pathogenic cells and antigens) are major players in autoimmune diseases such as MS. In the ﬁrst study that identiﬁed a role for ASIC1a in the neuropathology of MS, ASIC1b, ASIC3 and ASIC4 were identiﬁed at the mRNA level in mouse T cells, with the presence of ASIC1 protein conﬁrmed in these immune cells via Western blot (Friese et al., 2007). In this study, no obvious function for ASIC1 in T cells was observed based on responses of wild- type and ASIC1 knockout T cells in terms of proliferation and cytokine secretion. Thus, the role of ASICs in T cells remains to be determined, but they do not seem to play a role in these cells in the context of MS, at least not in the mouse EAE model. Data from the human binary protein interactome (Luck et al., 2020) indicate an interaction between ASIC1a and the killer cell immunoglobulin-like receptor 3DL3, a key regulator of NK cell function. Reduced levels of NK cells have also been implicated in depression (in which immune responses tend to be impaired). Brain acidity is altered in several mental health conditions (Coryell et al., 2009; Suzuki et al., 2017), but it is not yet clear whether this impacts on NK levels or function or whether ASICs have any role in these changes. Acid-Sensing Ion Channels in Oligodendrocytes Oligodendrocytes deliver critical trophic support to neurons, producing the proteins and lipids that comprise the myelin sheath that wrap around and insulate axons, and they engage in signiﬁcant cross-talk with microglia during the process of myelination (Peferoen et al., 2014). Oligodendrocytes secrete September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org 7 ASICs in CNS Immune Cells Foster et al. that ASIC1 activation contributes to oligodendrocyte injury during and after periods of tissue acidosis. that ASIC1 activation contributes to oligodendrocyte injury during and after periods of tissue acidosis. of proinﬂammatory IL-12 and the authors suggest that this triggers a bias toward a proinﬂammatory Th1 response (Martínez et al., 2007). The acidosis-induced increase in cell-surface marker expression was replicated using cultured DCs derived from mouse bone marrow, a response that was blocked by amiloride (Tong et al., 2011). Using non-speciﬁc antibodies for ASIC subtypes (i.e., not distinguishing between splice isoforms), Tong et al. (2011) found that mouse DCs express ASIC protein in the cytoplasm (ASIC1 and ASIC2), on the plasma membrane (ASIC2), in the endoplasmic reticulum and perinuclear regions (ASIC1), and in the mitochondria (ASIC3). Functional surface expression of ASICs was conﬁrmed using patch-clamp electrophysiology experiments and the observation of amiloride-sensitive, acidosis-induced inward currents with a pH50 of ∼6.0. Acidosis (pH 6.5) increased the antigen-presenting ability of DCs as assessed by increased ability to stimulate T cell proliferation, and this eﬀect was blocked by amiloride (Tong et al., 2011). Acid-Sensing Ion Channels in Macrophages Macrophages are key players in inﬂammation and autoimmune disease. They produce NO under acidic conditions and promote demyelination in the CNS (Bellocq et al., 1998; Yamasaki et al., 2014; Li and Barres, 2018). RT-PCR and Western blots were used to demonstrate expression of ASIC1 and ASIC3 in the cytoplasm of macrophages (Kong et al., 2013). When exposed to acidic conditions (pH 6.5), bone marrow-derived macrophages (BMMs) increase their rate of phagocytosis (Kong et al., 2013). Extracellular acidosis also causes an upregulation of key cell surface markers related to BMM maturation (e.g., CD80, CD86 and MHC II). These eﬀects were blocked by application of amiloride before exposure to tissue acidosis. Interestingly, expression of IL-10—thought to be a beneﬁcial anti-inﬂammatory cytokine—was increased in macrophages after acid exposure. These results strongly suggest that ASIC activation on DCs enhances expression of cell-surface proteins involved in antigen presentation and subsequent T cell activation and proliferation. As T cells can be highly destructive to healthy tissue in chronic inﬂammatory conditions, acidosis-enhanced communication between these two CNS-invading peripheral immune cells may exacerbate CNS damage in disease (Korn and Kallies, 2017). Thus, the limited available data suggest that ASICs can modulate the macrophage response to acidic microenvironments. CNS macrophages/monocytes are believed to be a primary contributor to MS demyelination, and ASICs may contribute to their response in this pathology (Yamasaki et al., 2014). Acid-Sensing Ion Channels in Dendritic Cells It has been suggested that in a neurotypical system, seizures should be self-limiting due to the ASIC1a response to increased acidity at seizure onset; in support of this hypothesis, a single nucleotide polymorphism (SNP) in the ASIC1 gene is associated with TLE (Lv et al., 2011; Wemmie et al., 2013). Mental Health Patients with schizophrenia exhibit lower pH and increased lactate in the cerebellum (Halim et al., 2008), whereas patients with bipolar disorder show decreased pH in the dorsolateral prefrontal cortex and increased lactate in gray matter regions of the brain (Dager et al., 2004; Sun et al., 2006). The frontal cortex and brain homogenate of rodent models replicated these ﬁndings with both increased lactate and reduced pH observed (Halim et al., 2008; Hagihara et al., 2017). ASIC1a is abundantly expressed in the amygdala, a brain region that has been described as a chemosensor that promotes fear behavior after detection of CO2 and/or acidosis (Wemmie et al., 2003; Coryell et al., 2007; Ziemann et al., 2009). Activation or inhibition/disruption of ASIC1a and altered expression of the ASIC1 gene all modify fear-related behaviors. Coryell et al. (2007) found that loss of the ASIC1 gene aﬀected neuronal activity in the amygdala after exposure to fear-inducing odors. Not only do ASIC1 KO mice exhibit reduced fear, overexpression of the channel increases fear responses (Wemmie et al., 2002, 2004). In humans, two SNPs in the ASIC1 gene are associated with anxiety, linking with amygdala structure and function and also risk of panic disorder (Smoller et al., 2014). Lastly, Coryell and colleagues used a mouse model to examine the link between ASIC1a and depression. They found that pharmacological inhibition and or genetic ablation of ASIC1a reduced depression like-symptoms, while restoring the gene to the amygdala returned responses back to baseline (Coryell et al., 2009). In a PD model, knockout of ASIC1a made no diﬀerence to the number of dopaminergic neurons, the key subset of neurons that are lost in this pathology and lead to motor impairment (Komnig et al., 2016). A signiﬁcant build-up of lactate has been observed across all brain regions in animal models of HD, and in the frontal cortex of HD patients (Harms et al., 1997; Dautry et al., 1999; Tsang et al., 2006). Administration of an amiloride derivative in a HD model caused a reduction in polyQ aggregation, one potential eﬀector in this pathology, both in vivo and in vitro (Wong et al., 2008). Although the authors state that this indicates a role for ASIC1a in HD, their use of a non-speciﬁc ASIC inhibitor means that one cannot rule out the potential involvement of other channels/transporters in this neurodegenerative disease. Acid-Sensing Ion Channels in Dendritic Cells ASIC1a is expressed in the cell body, dendritic arbor and postsynaptic dendritic spines of brain neurons (Wemmie et al., 2002; Alvarez de La Rosa et al., 2003; Vukicevic and Kellenberger, 2004; Zha et al., 2006). The brain can suﬀer signiﬁcant pH reductions during CNS pathologies, falling to as low as 6.0 during severe cerebral ischemia (Rehncrona, 1985). Given that the pH50 for activation of ASIC1a in human cortical neurons is 6.6 (Li et al., 2010), such drops in brain pH are suﬃcient to robustly activate ASIC1a. ASIC1a expression is upregulated in both dorsal root ganglia (DRG) and spinal dorsal horn neurons in response to inﬂammation, and in DRGs this upregulation is suppressed by ASIC inhibitors (Voilley et al., 2001; Duan et al., 2007). Prevention of ASIC1a endocytosis caused elevated death of Mature DCs are professional antigen presenting cells that constitute a key part of the inﬂammatory cascade, forming a major link between the innate and adaptive immune systems. The expression of MHC II and co-stimulatory molecules such as CD86 on their cell surface allows them to present antigens to T cells leading to T cell activation and proliferation (Reis e Sousa, 2006; Dudek et al., 2013). When mature DCs are exposed to acidic conditions (pH 6.5) for a moderate time period (∼4 h), the expression of their speciﬁc cell-surface markers is increased (Vermeulen et al., 2004). Human DCs exposed to acidosis increase production September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org 8 ASICs in CNS Immune Cells Foster et al. cortical neurons exposed to acidosis in wild-type, but not ASIC1a knockout, mice (Zeng et al., 2013). inhalation requires ASIC1a to interrupt induced seizures. This may be due to activation of ASICs causing generation of action potentials in inhibitory interneurons, although ASICs are also expressed in excitatory pyramidal neurons (Cho and Askwith, 2008; Ziemann et al., 2008; Weng et al., 2010). Another study found that amiloride suppressed pilocarpine-induced seizures, but the use of this pan-ASIC inhibitor makes it diﬃcult to relate this outcome to speciﬁc involvement of ASIC1a (Liang et al., 2015). Consistent with ASIC1a playing a role in seizure termination, hippocampal levels of ASIC1a are elevated in patients with temporal lobe epilepsy (TLE) and in epileptic mice (Yang et al., 2016). Multiple Sclerosis and Other Neurodegenerative Diseases Axonal degeneration plays a major role in MS, with associated inﬂammation causing dysfunctional activity of mitochondria (Waxman, 2006). Using the EAE model of MS, it was found that mice in which the ASIC1 gene was genetically inactivated or the channel was inhibited with amiloride exhibited marked axonal preservation (Friese et al., 2007). Consistent with the hypothesis that ASIC1a is activated during MS, increased levels of lactate are found in brain lesions of MS patients (Bitsch et al., 1999). ASICs are also expressed in microglia, astrocytes and oligodendrocytes, major players in MS that engage in crosstalk (Domingues et al., 2016). While microglia are the primary phagocytotic cells in the CNS, astrocytes also possess the ability to phagocytose neuronal debris, axonal mitochondria, and pathological protein aggregates (Lee and Chung, 2021). In aging mice, microglia appear to accumulate myelin debris at the same time as myelin degeneration (Hill et al., 2018). Indeed, Popovich et al. (2002) demonstrated that direct activation of “CNS macrophages” (covering both invading peripheral macrophages and CNS resident microglia) results in axonal damage and demyelination. This conclusion has gained strong support over the past 20 years and activated microglia in particular seem to represent promising targets in MS (Geladaris et al., 2021). As described above, ASICs are functionally expressed in both macrophages and microglia, and contribute to their activation during periods of acidosis. Thus, the protection aﬀorded by ASIC1a inhibition in the EAE mouse model of MS is possibly due to decreased macrophage/microglial activation as well as direct neuroprotection. Seizures As described earlier, vascular disruption due to trauma can result in acidiﬁcation of the brain, and the resultant drop in pH is often suﬃcient to activate neuronal ASIC1a (Xiong et al., 2004; Li et al., 2010; Chassagnon et al., 2017). It was originally envisaged that neuronal ASIC1a might exacerbate ischemia-induced brain injury by contributing to excitotoxicity by virtue of its ability to mediate ﬂux of Ca2+ into neurons in addition to Na+ (Xiong et al., 2004; Yermolaieva et al., 2004). However, recent studies suggests that, at least in the brain, activation of ASIC1a Epileptic seizures are associated with increased acidity in the brain, and upon CO2 administration the resultant hypercapnia allows for seizure termination through lowering of brain pH (Lennox, 1928; Wang and Sonnenschein, 1955; Miller, 2011). Overexpression of ASIC1a in mice appears to limit the duration and progression of chemoconvulsant-induced seizures, but not the number, and knockout of the gene has the converse eﬀect (Ziemann et al., 2008). Furthermore, the beneﬁcial eﬀect of CO2 September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org 9 Foster et al. ASICs in CNS Immune Cells FIGURE 5 \| Function, ASIC proﬁle, and putative role of ASICs in CNS immune cells. FIGURE 5 \| Function, ASIC proﬁle, and putative role of ASICs in CNS immune cells. of necroptosis (Wang et al., 2015; Redd et al., 2021), apoptosis (Song et al., 2019), and the stimulation of cell executioners such as calcium-activated proteases, endonucleases, and PLA2 due to the increased levels of intracellular Ca2+ (Yermolaieva et al., 2004; Xiong et al., 2007). also leads to recruitment and activation of receptor-interacting serine/threonine-protein kinase 1 (RIPK1), a key mediator of necroptosis (Wang et al., 2015, 2020). Remarkably, this activation of RIPK1 occurs independently of ion ﬂow through the channel (Wang et al., 2015). How does ASIC1a contribute to ischemic injury when it is subject to rapid steady-state desensitisation (SSD) when exposed to sustained acidic pH in vitro (Gründer and Chen, 2010), and is distinguished from other ASICs by a reduced responsiveness to successive acid stimulations, a phenomenon known as tachyphylaxis (Chen and Grunder, 2007)? One might assume that these properties would limit the persistence of ASIC1a currents during sustained periods of tissue acidosis in vivo. However, a variety of ensuing biochemical events act to persistently activate ASIC1a during a sustained drop in tissue pH. Seizures First, ASIC1a currents are potentiated by several ischemia-related factors, including: (i) extracellular lactate which increases from a basal level of 1–2 mM to 12–20 mM during ischemia (Allen and Attwell, 2002; Gonzalez-Inchauspe et al., 2020); (ii) membrane stretch resulting from the rise in extracellular [K+] (Allen and Attwell, 2002); and (iii) CaMKII phosphorylation of ASIC1a (Gao et al., 2005). Second, during ischemic stroke, the increase in extracellular spermine prolongs ASIC1a currents and enhances recovery from SSD (Duan et al., 2011). Third, the increase in intracellular Ca2+ inhibits tachyphylaxis. Fourth, arachidonic acid, which is elevated during stroke due to activation of phospholipase A2 (PLA2), potentiates ASIC1a currents and enhances the sustained component of the current (Allen and Attwell, 2002). Finally, and perhaps most importantly, the train of cell-death signaling set oﬀ by the initial activation of ASIC1a continues regardless of the subsequent state of the channel. This includes activation Thus, despite its susceptibility to desensitization (which has largely been studied under control conditions in in vitro assays), activation of ASIC1a during sustained tissue acidosis is a major contributor to ischemic injury, as evidenced by the fact that genetic ablation or speciﬁc pharmacological inhibition of ASIC1a greatly reduces the tissue damage caused by ischemic stroke (Xiong et al., 2004; McCarthy et al., 2015; Chassagnon et al., 2017). Genetic knockout or knockdown of ASIC1a, as well as pharmacological inhibition of the channel, also provide neuroprotective eﬀects post-SCI (Hu et al., 2011; Koehn et al., 2016). However, it should be noted that these studies have focused exclusively on the role of ASIC1a in CNS neurons. Despite the critical role of immune cells in damage progression after CNS trauma, very few studies have explored whether the resident population of ASIC1a, or other ASICs, in immune cells contributes to ischemic jury of the CNS. CONCLUSION Here we have described what is currently known about the expression and role of ASICs in immune cells known to contribute to a wide range of neuropathologies. We highlight the fact that ASICs are expressed in many key CNS-resident and inﬁltrating immune cells involved in these pathologies and are not only expressed in neurons (Figure 5). Thus, it is important to consider whether the therapeutic beneﬁt provided by ASIC September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org Frontiers in Cellular Neuroscience \| www.frontiersin.org 10 ASICs in CNS Immune Cells Foster et al. inhibitors in some of these pathologies is exerted not only via eﬀects on neurons, but also on immune cells. There are some studies where subtype-speciﬁc inhibitors or speciﬁc genetic ablation has been used to deﬁnitively demonstrate a role for ASICs in CNS pathologies, such as ischemic stroke (Pignataro et al., 2007; McCarthy et al., 2015; Chassagnon et al., 2017). However, even in these instances, the relative contribution of neuronal and immune-cell ASICs was not considered. Our understanding of how immune-cell ASICs contribute to neuropathology would be signiﬁcantly enhanced by utilizing technologies that allow direct study of these channels in speciﬁc immune cells. This was achieved in an eloquent study by Yang et al. (2016) through knockdown and then restoration of ASIC1a solely on astrocytes in a mouse model of TLE. Such techniques, particularly in combination with selective pharmacological tools, will allow determination of whether pan-inhibition of speciﬁc ASIC subtypes or their inhibition only in speciﬁc cell types is likely to be therapeutically useful. A common theme across immune cells is alterations to phagocytosis after exposure to acidity. Ineﬀective phagocytosis contributes to autoimmune diseases such as systemic lupus erythematosus (SLE) and polyarthritis (Nagata et al., 2010). Whether ASICs are the primary mediators of acidity-induced alterations in phagocytosis remains to be determined. Their role may depend on the type and severity of the pathology, in turn leading to beneﬁcial or negative outcomes, as seen for other receptors such as P2X7 (Savio et al., 2018). The CNS is heterogenous in its way of dealing with acidity (Chesler, 2003) and immune cells also respond diﬀerently to the “type” of acidity (Kellum et al., 2004). The resultant complexity of pH-induced eﬀects results in a multitude of unique outcomes dependent on CNS region and immune cell type, as well as the ways in which acidity is induced. FUNDING We acknowledge funding from the Australian Research Council (Centre of Excellence Grant CE200100012 to GK), Australian National Health and Medical Research Council (Principal Research Fellowship APP1136889 to GK and Project Grant APP1154622 to LR), and The University of Queensland (International Ph.D. Scholarship to VF). We acknowledge funding from the Australian Research Council (Centre of Excellence Grant CE200100012 to GK), Australian National Health and Medical Research Council (Principal Research Fellowship APP1136889 to GK and Project Grant APP1154622 to LR), and The University of Queensland (International Ph.D. Scholarship to VF). SUPPLEMENTARY MATERIAL The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fncel. 2021.738043/full#supplementary-material AUTHOR CONTRIBUTIONS VF conceived the study and wrote the ﬁrst draft of the manuscript. LR, MR, and GK revised and streamlined the manuscript. MR and GK contributed funding and mentored VF. All authors contributed to the article and approved the submitted version. CONCLUSION To unravel the role of immune-cell ASICs in neuropathology will likely require standardization of models, including the method of acidosis induction. In conclusion, the role of ASIC in immune cells is an exciting frontier in neurological research. Merging CNS immune function studies with cutting-edge molecular techniques will provide greater insight into whether immune-cell ASICs are likely to be useful drug targets for CNS disorders. g Another caveat of current research on the role of ASICs in immune-cell function is the widespread use of amiloride as a broad-spectrum ASIC inhibitor. Unfortunately, amiloride has targets outside of the ASIC family, including sodium-hydrogen antiporter 1 (NHE-1), which is involved in acid-sensing and pH regulation by immune cells. Amiloride can also inhibit TRPP3 channels (Dai et al., 2007), which are thought to be involved in acid sensing in mouse spinal cord and alkaline sensing in lamprey (Huang et al., 2006; Dai et al., 2007; Jalalvand et al., 2016). This raises the question of whether the beneﬁcial eﬀects of amiloride in human MS patients is solely due to inhibition of ASICs (Arun et al., 2013). Much more potent and speciﬁc ASIC inhibitors are available, such as Hi1a (Chassagnon et al., 2017) and PcTx1 (Escoubas et al., 2000) for ASIC1a, mambalgins for rodent ASIC1a and ASIC1b (Diochot et al., 2012) (although they potentiate human ASIC1b under the levels of acidosis likely encountered in vivo; Cristofori-Armstrong et al., 2021) and the sea-anemone peptide APETx2 for ASIC3 (Diochot et al., 2004), and they should be used in preference to amiloride wherever possible. Unfortunately, there are no speciﬁc inhibitors available for ASIC1b and ASIC2a. A ﬁnal concern with investigations of ASIC protein expression is ongoing challenges with the speciﬁcity of available ASIC antibodies. As highlighted by Lin et al. (2015), examples of lack of speciﬁcity include “positive” ASIC1a staining in a pan-ASIC1a knockout mouse (Vig et al., 2014). Thus, due to the limitations of currently available commercial antibodies, antibody staining data should always be combined with proteomics and/or functional data using subtype-speciﬁc inhibitors to provide conﬁrmation of ASIC localization. Alvarez de La Rosa, D., Krueger, S. R., Kolar, A., Shao, D., Fitzsimonds, R. M., and Canessa, C. M. (2003). Distribution, subcellular localization and ontogeny of ASIC1 in the mammalian central nervous system. J. Physiol. 546, 77–87. Frontiers in Cellular Neuroscience \| www.frontiersin.org REFERENCES Allen, N. J., and Attwell, D. (2002). Modulation of ASIC channels in rat cerebellar Purkinje neurons by ischaemia-related signals. J. Physiol. 543, 521–529. doi: 10.1113/jphysiol.2002.020 297 Abbott, N., Rönnbäck, L., and Hansson, E. (2006). Astrocyte-endothelial interactions at the blood-brain barrier. Nat. Rev. Neurosci. 7, 41–53. Ajami, B., Bennett, J., Krieger, C., McNagny, K., and Rossi, F. M. (2011). Inﬁltrating monocytes trigger EAE progression, but do not contribute to the resident microglia pool. Nat. Neurosci. 14, 1142–1149. doi: 10.1038/nn.2887 Alvarez de La Rosa, D., Krueger, S. R., Kolar, A., Shao, D., Fitzsimonds, R. M., and Canessa, C. M. (2003). Distribution, subcellular localization and ontogeny of ASIC1 in the mammalian central nervous system. J. Physiol. 546, 77–87. Ajami, B., Bennett, J., Krieger, C., McNagny, K., and Rossi, F. M. (2011). Inﬁltrating monocytes trigger EAE progression, but do not contribute to the resident microglia pool. Nat. Neurosci. 14, 1142–1149. doi: 10.1038/nn.2887 September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org Frontiers in Cellular Neuroscience \| www.frontiersin.org 11 Foster et al. ASICs in CNS Immune Cells Anderson, M., Burda, J., Ren, Y., Ao, Y., O’Shea, T., Kawaguchi, R., et al. (2016). Astrocyte scar formation aids central nervous system axon regeneration. Nature 532, 195–200. doi: 10.1038/nature17623 Chesler, M. (2003). Regulation and modulation of pH in the brain. Physiol. Rev. 83, 1183–1221. doi: 10.1152/physrev.00010.2003 Chiang, E. Y., Li, T., Jeet, S., Peng, I., Zhang, J., Lee, W. P., et al. (2017). Potassium channels Kv1.3 and KCa3.1 cooperatively and compensatorily regulate antigen- speciﬁc memory T cell functions. Nat. Commun. 8:14644. doi: 10.1038/ ncomms14644 Arnould, T., Thibaut-Vercruyssen, R., Bouaziz, N., Dieu, M., Remacle, J., and Michiels, C. (2001). PGF2α, a prostanoid released by endothelial cells activated by hypoxia, is a chemoattractant candidate for neutrophil recruitment. Am. J. Pathol. 159, 345–357. doi: 10.1016/S0002-9440(10)61701-4 Cho, J.-H., and Askwith, C. (2008). Presynaptic release probability is increased in hippocampal neurons from ASIC1 knockout mice. J. Neurophysiol. 99, 426–441. Arun, T., Tomassini, V., Sbardella, E., de Ruiter, M., Matthews, L., Leite, M., et al. (2013). Targeting ASIC1 in primary progressive multiple sclerosis: evidence of neuroprotection with amiloride. Brain 136, 106–115. doi: 10.1093/brain/ aws325 Chodobski, A., Zink, B. J., and Szmydynger-Chodobska, J. (2011). Blood–brain barrier pathophysiology in traumatic brain injury. Transl. Stroke Res. 2, 492– 516. doi: 10.1007/s12975-011-0125-x Baron, A., Voilley, N., Lazdunski, M., and Lingueglia, E. (2008). Acid sensing ion channels in dorsal spinal cord neurons. J. Neurosci. REFERENCES 28, 1498–1508. doi: 10.1523/jneurosci.4975-07.2008 Clemens, R. A., and Lowell, C. A. (2019). CRAC channel regulation of innate immune cells in health and disease. Cell Calcium 78, 56–65. doi: 10.1016/j.ceca. 2019.01.003 Bellocq, A., Suberville, S., Philippe, C., Bertrand, F., Perez, J., Fouqueray, B., et al. (1998). Low environmental pH is responsible for the induction of nitric-oxide synthase in macrophages. Evidence for involvement of nuclear factor-kappaB activation. J. Biol. Chem. 273, 5086–5092. Coryell, M. W., Wunsch, A. M., Haenﬂer, J. M., Allen, J. E., Schnizler, M., Ziemann, A. E., et al. (2009). Acid-sensing ion channel-1a in the amygdala, a novel therapeutic target in depression-related behavior. J. Neurosci. 29, 5381–5388. doi: 10.1523/jneurosci.0360-09.2009 Bidani, A., Wang, C. Z., Saggi, S. J., and Heming, T. A. (1998). Evidence for pH sensitivity of tumor necrosis factor-alpha release by alveolar macrophages. Lung 176, 111–121. doi: 10.1007/pl00007593 Coryell, M. W., Ziemann, A. E., Westmoreland, P. J., Haenﬂer, J. M., Kurjakovic, Z., Zha, X.-M., et al. (2007). Targeting ASIC1a reduces innate fear and alters neuronal activity in the fear circuit. Biol. Psychiatry 62, 1140–1148. doi: 10.1016/ j.biopsych.2007.05.008 Bielekova, B., Sung, M.-H., Kadom, N., Simon, R., McFarland, H., and Martin, R. (2004). Expansion and functional relevance of high-avidity myelin-speciﬁc CD4+ T cells in multiple sclerosis. J. Immunol. 172, 3893–3904. Cristofori-Armstrong, B., Budusan, E., and Rash, L. D. (2021). Mambalgin- 3 potentiates human acid-sensing ion channel 1b under mild to moderate acidosis: implications as an analgesic lead. Proc. Natl. Acad. Sci. USA 118:e2021581118. doi: 10.1073/pnas.2021581118 Bitsch, A., Bruhn, H., Vougioukas, V., Stringaris, A., Lassmann, H., Frahm, J., et al. (1999). Inﬂammatory CNS demyelination: histopathologic correlation with in vivo quantitative proton MR spectroscopy. Am. J. Neuroradiol. 20, 1619–1627. Dager, S. R., Friedman, S. D., Parow, A., Demopulos, C., Stoll, A. L., Lyoo, I. K., et al. (2004). Brain metabolic alterations in medication-free patients with bipolar disorder. Arch. Gen. Psychiatry 61, 450–458. doi: 10.1001/archpsyc.61.5.450 Bradl, M., and Lassmann, H. (2010). Oligodendrocytes: biology and pathology. Acta Neuropathol. 119, 37–53. doi: 10.1007/s00401-009-0601-5 Dai, X. Q., Ramji, A., Liu, Y., Li, Q., Karpinski, E., and Chen, X. Z. (2007). Inhibition of TRPP3 channel by amiloride and analogs. Mol. Pharmacol. 72, 1576–1585. doi: 10.1124/mol.107.037150 Browne, S. E., and Beal, M. F. (2006). Oxidative damage in Huntington’s disease pathogenesis. Antioxid. Redox Signal. 8, 2061–2073. Butler, C. A., Popescu, A. S., Kitchener, E. J. A., Allendorf, D. H., Puigdellivol, M., and Brown, G. C. (2021). REFERENCES Microglial phagocytosis of neurons in neurodegeneration, and its regulation. J. Neurochem. 158, 621–639. doi: 10. 1111/jnc.15327 Dautry, C., Condé, F., Brouillet, E., Mittoux, V., Beal, M. F., Bloch, G., et al. (1999). Serial 1H-NMR spectroscopy study of metabolic impairment in primates chronically treated with the succinate dehydrogenase inhibitor 3- nitropropionic acid. Neurobiol. Dis. 6, 259–268. doi: 10.1006/nbdi.1999.0244 Butovsky, O., Ziv, Y., Schwartz, A., Landa, G., Talpalar, A., Pluchino, S., et al. (2006). Microglia activated by IL-4 or IFN-gamma diﬀerentially induce neurogenesis and oligodendrogenesis from adult stem/progenitor cells. Mol. Cell. Neurosci. 31, 149–160. doi: 10.1016/j.mcn.2005.10.006 De Bondt, M., Hellings, N., Opdenakker, G., and Struyf, S. (2020). Neutrophils: underestimated players in the pathogenesis of Multiple Sclerosis (MS). Int. J. Mol. Sci. 21:4558. doi: 10.3390/ijms21124558 Delaunay, A., Gasull, X., Salinas, M., Noël, J., Friend, V., Lingueglia, E., et al. (2012). Human ASIC3 channel dynamically adapts its activity to sense the extracellular pH in both acidic and alkaline directions. Proc. Natl. Acad. Sci. USA 109, 13124–13129. doi: 10.1073/pnas.1120350109 Cadiou, H., Studer, M., Jones, N. G., Smith, E. S. J., Ballard, A., McMahon, S. B., et al. (2007). Modulation of acid-sensing ion channel activity by nitric oxide. J. Neurosci. 27, 13251–13260. doi: 10.1523/JNEUROSCI.2135-07.2007 Calabrese, V., Mancuso, C., Calvani, M., Rizzarelli, E., Butterﬁeld, D., and Giuﬀrida Stella, A. (2007). Nitric oxide in the central nervous system: neuroprotection versus neurotoxicity. Nat. Rev. Neurosci. 8, 766–775. doi: 10.1038/nrn2214 Deniset, J. F., and Kubes, P. (2018). Neutrophil heterogeneity: Bona ﬁde subsets or polarization states? J. Leukocyte Biol. 103, 829–838. doi: 10.1002/JLB.3RI0917- 361R Campagnolo, D. I., Bartlett, J. A., Keller, S. E., Sanchez, W., and Oza, R. (1997). Impaired phagocytosis of Staphylococcus aureus in complete tetraplegics. Am. J. Phys. Med. Rehabil. 76, 276–280. doi: 10.1097/00002060-199707000-00005 Diochot, S., Baron, A., Rash, L. D., Deval, E., Escoubas, P., Scarzello, S., et al. (2004). A new sea anemone peptide, APETx2, inhibits ASIC3, a major acid-sensitive channel in sensory neurons. EMBO J. 23, 1516–1525. doi: 10.1038/sj.emboj. 7600177 Cao, Q., Wang, W., Gu, J., Jiang, G., Wang, K., Xu, Z., et al. (2016). Elevated expression of acid-sensing ion channel 3 inhibits epilepsy via activation of interneurons. Mol. Neurobiol. 53, 485–498. doi: 10.1007/s12035-014-9014-0 Diochot, S., Baron, A., Salinas, M., Douguet, D., Scarzello, S., Dabert-Gay, A. S., et al. (2012). Black mamba venom peptides target acid-sensing ion channels to abolish pain. Nature 490, 552–555. REFERENCES doi: 10.1038/nature11494 Cao, S., Liu, P., Zhu, H., Gong, H., Yao, J., Sun, Y., et al. (2015). Extracellular acidiﬁcation acts as a key modulator of neutrophil apoptosis and functions. PLoS One 10:e0137221. doi: 10.1371/journal.pone.0139500 Domingues, H. S., Portugal, C. C., Socodato, R., and Relvas, J. B. (2016). Oligodendrocyte, astrocyte, and microglia crosstalk in myelin development, damage, and repair. Front. Cell Dev. Biol. 4:71. doi: 10.3389/fcell.2016.0 0071 Chandy, K. G., and Norton, R. S. (2017). Peptide blockers of Kv1.3 channels in T cells as therapeutics for autoimmune disease. Curr. Opin. Chem. Biol. 38, 97–107. doi: 10.1016/j.cbpa.2017.02.015 Chassagnon, I. R., McCarthy, C. A., Chin, Y. K., Pineda, S. S., Keramidas, A., Mobli, M., et al. (2017). Potent neuroprotection after stroke aﬀorded by a double- knot spider-venom peptide that inhibits acid-sensing ion channel 1a. Proc. Natl. Acad. Sci. USA 114, 3750–3755. doi: 10.1073/pnas.1614728114 Dossi, E., Vasile, F., and Rouach, N. (2018). Human astrocytes in the diseased brain. Brain Res. Bull. 136, 139–156. doi: 10.1016/j.brainresbull.2017.02.001 Douvdevani, A., Abramson, O., Tamir, A., Konforty, A., Isakov, N., and Chaimovitz, C. (1995). Commercial dialysate inhibits TNFα mRNA expression and NF-κB DNA-binding activity in LPS-stimulated macrophages. Kidney Int. 47, 1537–1545. doi: 10.1038/ki.1995.217 Chen, C. C., England, S., Akopian, A. N., and Wood, J. N. (1998). A sensory neuron-speciﬁc, proton-gated ion channel. Proc. Natl. Acad. Sci. USA 95, 10240–10245. doi: 10.1073/pnas.95.17.10240 Duan, B., Wang, Y.-Z., Yang, T., Chu, X.-P., Yu, Y., Huang, Y., et al. (2011). Extracellular spermine exacerbates ischemic neuronal injury through sensitization of ASIC1a channels to extracellular acidosis. J. Neurosci. 31, 2101–2112. doi: 10.1523/JNEUROSCI.4351-10.2011 Chen, X. M., and Grunder, S. (2007). Permeating protons contribute to tachyphylaxis of the acid-sensing ion channel (ASIC) 1a. J. Physiol. 579, 657– 670. doi: 10.1113/jphysiol.2006.120733 September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org 12 ASICs in CNS Immune Cells Foster et al. Duan, B., Wu, L. J., Yu, Y. Q., Ding, Y., Jing, L., Xu, L., et al. (2007). Upregulation of acid-sensing ion channel ASIC1a in spinal dorsal horn neurons contributes to inﬂammatory pain hypersensitivity. J. Neurosci. 27, 11139–11148. doi: 10.1523/ JNEUROSCI.3364-07.2007 spectroscopy study. J. Neurol. Neurosurg. Psych. 62, 27–30. doi: 10.1136/jnnp. 62.1.27 Heming, T. A., Davé, S. K., Tuazon, D. M., Chopra, A. K., Peterson, J. W., and Bidani, A. (2001a). Eﬀects of extracellular pH on tumour necrosis factor-alpha production by resident alveolar macrophages. Clin. Sci. 101, 267–274. Dudek, A. M., Martin, S., Garg, A. REFERENCES 26, Jahr, H., van Driel, M., van Osch, G. J. V. M., Weinans, H., and van Leeuwen, J. P. T. M. (2005). Identiﬁcation of acid-sensing ion channels in bone. Biochem. Biophys. Res. Commun. 337, 349–354. Gonzalez-Inchauspe, C., Gobetto, M. N., and Uchitel, O. D. (2020). Modulation of acid sensing ion channel dependent protonergic neurotransmission at the mouse calyx of Held. Neuroscience 439, 195–210. doi: 10.1016/j.neuroscience. 2019.04.023 p y Jalalvand, E., Robertson, B., Tostivint, H., Wallén, P., and Grillner, S. (2016). The spinal cord has an intrinsic system for the control of pH. Curr. Biol. 26, 1346–1351. doi: 10.1016/j.cub.2016.03.048 Gründer, S., and Chen, X. (2010). Structure, function, and pharmacology of acid- sensing ion channels (ASICs): focus on ASIC1a. Int. J. Physiol. Pathophysiol. Pharmacol. 2, 73–94. Jasti, J., Furukawa, H., Gonzales, E. B., and Gouaux, E. (2007). Structure of acid- sensing ion channel 1 at 1.9 Å resolution and low pH. Nature 449:316. doi: 10.1038/nature06163 Gründer, S., Geisler, H., Bässler, L., and Ruppersberg, J. P. (2000). Cloning of ASIC 4, a new member of acid-sensing ion channels from pituitary gland. Eur. J. Neurosci. 12:22. Jensen, J. C., Buresh, C., and Norton, J. A. (1990). Lactic acidosis increases tumor necrosis factor secretion and transcription in vitro. J. Surg. Res. 49, 350–353. doi: 10.1016/0022-4804(90)90036-2 Hagihara, H., Catts, V. S., Katayama, Y., Shoji, H., Takagi, T., Huang, F. L., et al. (2017). Decreased brain pH as a shared endophenotype of psychiatric disorders. Neuropsychopharmacology 43, 459–468. doi: 10.1038/npp.2017.167 Jörres, A., Gahl, G. M., and Frei, U. (1995). In vitro studies on the eﬀect of dialysis solutions on peritoneal leukocytes. Peritoneal Dialysis Int. 15, S41–S45. Josefsen, K., Nielsen, S., Campos, A., Seifert, T., Hasholt, L., Nielsen, J., et al. (2010). Reduced gluconeogenesis and lactate clearance in Huntington’s disease. Neurobiol. Dis. 40, 656–662. doi: 10.1016/j.nbd.2010.08.009 Halim, N. D., Lipska, B. K., Hyde, T. M., Deep-Soboslay, A., Saylor, E. M., Herman, M. M., et al. (2008). Increased lactate levels and reduced pH in postmortem brains of schizophrenics: medication confounds. J. Neurosci. Methods 169, 208–213. doi: 10.1016/j.jneumeth.2007.11.017 Kanyilmaz, S., Hepguler, S., Atamaz, F., Gokmen, N., Ardeniz, O., and Sin, A. (2013). Phagocytic and oxidative burst activity of neutrophils in patients with spinal cord injury. Arch. Phys. Med. Rehabil. 94, 369–374. Hanukoglu, I., and Hanukoglu, A. (2016). Epithelial sodium channel (ENaC) family: Phylogeny, structure–function, tissue distribution, and associated inherited diseases. Gene 579, 95–132. REFERENCES D., and Agostinis, P. (2013). Immature, semi- mature, and fully mature dendritic cells: toward a DC-cancer cells interface that augments anticancer immunity. Front. Immunol. 4:438. doi: 10.3389/ﬁmmu. 2013.00438 Heming, T. A., Tuazon, D. M., Davé, S. K., Chopra, A. K., Peterson, J. W., and Bidani, A. (2001b). Post-transcriptional eﬀects of extracellular pH on tumour necrosis factor-alpha production in RAW 246.7 and J774 A.1 cells. Clin. Sci. 100, 259–266. Enuka, Y., Hanukoglu, I., Edelheit, O., Vaknine, H., and Hanukoglu, A. (2012). Epithelial sodium channels (ENaC) are uniformly distributed on motile cilia in the oviduct and the respiratory airways. Histochem. Cell Biol. 137, 339–353. doi: 10.1007/s00418-011-0904-1 Heppner, F. L., Greter, M., Marino, D., Falsig, J., Raivich, G., Hovelmeyer, N., et al. (2005). Experimental autoimmune encephalomyelitis repressed by microglial paralysis. Nat. Med. 11:146. Escoubas, P., De Weille, J. R., Lecoq, A., Diochot, S., Waldmann, R., Champigny, G., et al. (2000). Isolation of a tarantula toxin speciﬁc for a class of proton-gated Na+ channels. J. Biol. Chem. 275, 25116–25121. doi: 10.1074/jbc.M003643200 Hertz, L., and Dienel, G. A. (2005). Lactate transport and transporters: general principles and functional roles in brain cells. J. Neurosci. Res. 79, 11–18. doi: 10.1002/jnr.20294 Hill, R. A., Li, A. M., and Grutzendler, J. (2018). Lifelong cortical myelin plasticity and age-related degeneration in the live mammalian brain. Nat. Neurosci. 21, 683–695. doi: 10.1038/s41593-018-0120-6 Faﬀ, L., and Nolte, C. (2000). Extracellular acidiﬁcation decreases the basal motility of cultured mouse microglia via the rearrangement of the actin cytoskeleton. Brain Res. Bull. 853, 22–31. doi: 10.1016/S0006-8993(99)02221-0 Hoagland, E. N., Sherwood, T. W., Lee, K. G., Walker, C. J., and Askwith, C. C. (2010). Identiﬁcation of a calcium permeable human acid-sensing ion channel 1 transcript variant. J. Biol. Chem. 285, 41852–41862. doi: 10.1074/jbc.M110. 171330 Feldman, D., Horiuchi, M., Keachie, K., McCauley, E., Bannerman, P., Itoh, A., et al. (2008). Characterization of acid-sensing ion channel expression in oligodendrocyte-lineage cells. Glia 56, 1238–1249. doi: 10.1002/glia.20693 Feldman, N., Rotter-Maskowitz, A., and Okun, E. (2015). DAMPs as mediators of sterile inﬂammation in aging-related pathologies. Ageing Res. Rev. 24, 29–39. doi: 10.1016/j.arr.2015.01.003 Hu, R., Duan, B., Wang, D. S., Yu, Y., Li, W. G., Luo, H. S., et al. (2011). Role of acid-sensing ion channel 1a in the secondary damage of traumatic spinal cord injury. Ann. Surg. 254, 353–362. doi: 10.1097/SLA.0b013e31822645b4 Feske, S., Wulﬀ, H., and Skolnik, E. Y. (2015). Ion channels in innate and adaptive immunity. Annu. Rev. Immunol. REFERENCES 33, 291–353. doi: 10.1146/annurev-immunol- 032414-112212 Huang, A. L., Chen, X., Hoon, M. A., Chandrashekar, J., Guo, W., Tränkner, D., et al. (2006). The cells and logic for mammalian sour taste detection. Nature 442, 934–938. doi: 10.1038/nature05084 Ffrench-Constant, C. (1994). Pathogenesis of multiple sclerosis. Lancet 343, 271– 275. Huang, C. J., Haque, I. U., Slovin, P. N., Nielsen, R. B., Fang, X., and Skimming, J. W. (2002). Environmental pH regulates LPS-induced nitric oxide formation in murine macrophages. Nitric Oxide 6, 73–78. doi: 10.1006/niox.2001. 0391 Friese, M. A., Craner, M. J., Etzensperger, R., Vergo, S., Wemmie, J. A., Welsh, M. J., et al. (2007). Acid-sensing ion channel-1 contributes to axonal degeneration in autoimmune inﬂammation of the central nervous system. Nat. Med. 13, 1483–1489. doi: 10.1038/nm1668 Huang, C., Hu, Z.-L., Wu, W.-N., Yu, D.-F., Xiong, Q.-J., Song, J.-R., et al. (2010). Existence and distinction of acid-evoked currents in rat astrocytes. Glia 58, 1415–1424. doi: 10.1002/glia.21017 Gao, J., Duan, B., Wang, D. G., Deng, X. H., Zhang, G. Y., Xu, L., et al. (2005). Coupling between NMDA receptor and acid-sensing ion channel contributes to ischemic neuronal death. Neuron 48, 635–646. doi: 10.1016/j.neuron.2005. 10.011 Immke, D., and McCleskey, E. (2001). Lactate enhances the acid-sensing Na+ channel on ischemia-sensing neurons. Nat. Neurosci. 4, 869–870. doi: 10.1038/ nn0901-869 Garty, H., and Palmer, L. (1997). Epithelial sodium channels: Function, structure, and regulation. Physiol. Rev. 77, 359–396. doi: 10.1152/physrev.1997.77.2.359 Ishibashi, K., and Marumo, F. (1998). Molecular cloning of a DEG/ENaC sodium channel cDNA from human testis. Biochem. Biophys. Res. Commun. 245, 589–593. doi: 10.1006/bbrc.1998.8483 Geladaris, A., Hausler, D., and Weber, M. S. (2021). Microglia: the missing link to decipher and therapeutically control MS progression? Int. J. Mol. Sci. 22:3461. doi: 10.3390/ijms22073461 Jack, C., Ruﬃni, F., Bar-Or, A., and Antel, J. P. (2005). Microglia and multiple sclerosis. J. Neurosci. Res. 81, 363–373. doi: 10.1002/jnr.20482 Ginhoux, F., Greter, M., Leboeuf, M., Nandi, S., See, P., Gokhan, S., et al. (2010). Fate mapping analysis reveals that adult microglia derive from primitive macrophages. Science 330, 841–845. doi: 10.1126/science.1194637 Jahr, H., van Driel, M., van Osch, G. J. V. M., Weinans, H., and van Leeuwen, J. P. T. M. (2005). Identiﬁcation of acid-sensing ion channels in bone. Biochem. Biophys. Res. Commun. 337, 349–354. Jalalvand, E., Robertson, B., Tostivint, H., Wallén, P., and Grillner, S. (2016). The spinal cord has an intrinsic system for the control of pH. Curr. Biol. REFERENCES Immune cells and CNS physiology: Microglia and beyond. J. Exp. Med. 216, 60–70. doi: 10.1084/jem.20180199 Lee, S. J., Drabik, K., Van Wagoner, N. J., Lee, S., Choi, C., Dong, Y., et al. (2000). ICAM-1-induced expression of proinﬂammatory cytokines in astrocytes: involvement of extracellular signal-regulated kinase and p38 mitogen-activated protein kinase pathways. J. Immunol. 165, 4658–4666. Olena, Y., Leonard, A. S., Mikael, K. S., Francois, M. A., and Michael, J. W. (2004). Extracellular acidosis increases neuronal cell calcium by activating acid-sensing ion channel 1a. Proc. Natl. Acad. Sci. USA 101, 6752–6757. doi: 10.1073/pnas. 0308636100 Lee, S. Y., and Chung, W. S. (2021). The roles of astrocytic phagocytosis in maintaining homeostasis of brains. J. Pharmacol. Sci. 145, 223–227. doi: 10. 1016/j.jphs.2020.12.007 Paukert, M., Babini, E., Pusch, M., and Grunder, S. (2004). Identiﬁcation of the Ca2+ blocking site of acid-sensing ion channel (ASIC)1: implications for channel gating. J. Gen. Physiol. 124, 383–394. doi: 10.1085/jgp.200308973 Lennox, W. G. (1928). The eﬀect on epileptic seizures of varying the composition of the respired air. J. Clin. Invest. 6, 23–24. Peferoen, L., Kipp, M., Valk, P., Noort, J. M., and Amor, S. (2014). Oligodendrocyte-microglia cross-talk in the central nervous system. Immunology 141, 302–313. doi: 10.1111/imm.12163 Li, M., Inoue, K., Branigan, D., Kratzer, E., Hansen, J. C., Chen, J. W., et al. (2010). Acid-sensing ion channels in acidosis-induced injury of human brain neurons. J. Cerebral Blood Flow Metab. 30, 1247–1260. doi: 10.1038/jcbfm.2010.30 Pignataro, G., Simon, R. P., and Xiong, Z.-G. (2007). Prolonged activation of ASIC1a and the time window for neuroprotection in cerebral ischaemia. Brain 130, 151–158. doi: 10.1093/brain/awl325 Li, Q., and Barres, B. (2018). Microglia and macrophages in brain homeostasis and disease. Nat. Rev. Immunol. 18, 225–242. doi: 10.1038/nri.2017.125 Liang, J.-J., Huang, L.-F., Chen, X.-M., Pan, S.-Q., Lu, Z.-N., and Xiao, Z.-M. Liang, J.-J., Huang, L.-F., Chen, X.-M., Pan, S.-Q., Lu, Z.-N., and Xiao, Z.-M. (2015). Amiloride suppresses pilocarpine-induced seizures via ASICs other than NHE in rats. Int. J. Clin. Exp. Path. 8, 14507–14513. g g (2015). Amiloride suppresses pilocarpine-induced seizures via ASICs other than NHE in rats. Int. J. Clin. Exp. Path. 8, 14507–14513. Polman, C. H., O’Connor, P. W., Havrdova, E., Hutchinson, M., Kappos, L., Miller, D. H., et al. (2006). A randomized, placebo-controlled trial of natalizumab for relapsing multiple sclerosis. New Engl. J. Med. 354, 899–910. doi: 10.1056/ NEJMoa044397 Lin, S. H., Sun, W. H., and Chen, C. C. (2015). REFERENCES doi: 10.1016/j.gene.2015.12.061 Karman, J., Chu, H., Co, D., Seroogy, C. M., Sandor, M., and Fabry, Z. (2006). Dendritic cells amplify T cell-mediated immune responses in the central nervous system. J. Immunol. 177, 7750–7760. Harms, L., Meierkord, H., Timm, G., Pfeiﬀer, L., and Ludolph, A. C. (1997). Decreased N-acetyl-aspartate/choline ratio and increased lactate in the frontal lobe of patients with Huntington’s disease: a proton magnetic resonance Kellenberger, S., and Schild, L. (2015). International union of basic and clinical pharmacology. XCI. structure, function, and pharmacology of acid-sensing September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org 13 ASICs in CNS Immune Cells Foster et al. ion channels and the epithelial Na+ channel. Pharmacol. Rev. 67, 1–35. doi: 10.1124/pr.114.009225 ion channels and the epithelial Na+ channel. Pharmacol. Rev. 67, 1–35. doi: 10.1124/pr.114.009225 cells and the production of IL-12. J. Immunol. 179, 1950–1959. doi: 10.4049/ jimmunol.179.3.1950 cells and the production of IL-12. J. Immunol. 179, 1950–1959. doi: 10.4049/ jimmunol.179.3.1950 Kellum, J. A., Song, M., and Li, J. (2004). Science review: extracellular acidosis and the immune response: clinical and physiologic implications. Crit. Care 8, 331–336. doi: 10.1186/cc2900 Maubaret, C., Delettre, C., Sola, S., and Hamel, C. P. (2002). Identiﬁcation of preferentially expressed mRNAs in retina and cochlea. DNA Cell Biol. 21, 781–791. doi: 10.1089/104454902320908432 Kierdorf, K., Erny, D., Goldmann, T., Sander, V., Schulz, C., Perdiguero, E., et al. (2013). Microglia emerge from erythromyeloid precursors via Pu.1- and Irf8-dependent pathways. Nat. Neurosci. 16, 273–280. doi: 10.1038/nn.3318 McCarthy, C. A., Rash, L. D., Chassagnon, I. R., King, G. F., and Widdop, R. E. (2015). PcTx1 aﬀords neuroprotection in a conscious model of stroke in hypertensive rats via selective inhibition of ASIC1a. Neuropharmacology 99, 650–657. doi: 10.1016/j.neuropharm.2015.08.040 Kilgore, K. S., Friedrichs, G. S., Homeister, J. W., and Lucchesi, B. R. (1994). The complement system in myocardial ischaemia/reperfusion injury. Cardiovasc. Res. 28, 437–444. doi: 10.1093/cvr/28.4.437 650–657. doi: 10.1016/j.neuropharm.2015.08.040 Mikhail, J. (1999). The trauma triad of death: hypothermia, acidosis, and coagulopathy. AACN Clin. Issues 10, 85–94. Miller, J. W. (2011). Stopping seizures with carbon dioxide. Epilepsy Curr. 11, 114–115. doi: 10.5698/1535-7511-11.4.114 Koehn, L. M., Noor, N. M., Dong, Q., Er, S.-Y., Rash, L. D., King, G. F., et al. (2016). Selective inhibition of ASIC1a confers functional and morphological neuroprotection following traumatic spinal cord injury. F1000Research 5:1822. doi: 10.12688/f1000research.9094.2 Mitra, B., Tullio, F., Cameron, P. A., and Fitzgerald, M. (2012). REFERENCES Trauma patients with the ’triad of death’. Emerg. Med. J. 29, 622–625. doi: 10.1136/emj.2011. 113167 Komnig, D., Imgrund, S., Reich, A., Grunder, S., and Falkenburger, B. H. (2016). ASIC1a deﬁcient mice show unaltered neurodegeneration in the subacute MPTP model of Parkinson disease. PLoS One 11:e0165235. doi: 10.1371/ journal.pone.0165235 Moskowitz, M., Lo, E., and Iadecola, C. (2010). The science of stroke: mechanisms in search of treatments. Neuron 67, 181–198. doi: 10.1016/j.neuron.2010.07.002 Nadolni, W., and Zierler, S. (2018). The channel-kinase TRPM7 as novel regulator of immune system homeostasis. Cells 7:109. doi: 10.3390/cells7080109 Kondo, M., Wagers, A. J., Manz, M. G., Prohaska, S. S., Scherer, D. C., Beilhack, G. F., et al. (2003). Biology of hematopoietic stem cells and progenitors: implications for clinical application. Annu. Rev. Immunol. 21, 759–806. doi: 10.1146/annurev.immunol.21.120601.141007 Nagata, S., Hanayama, R., and Kawane, K. (2010). Autoimmunity and of dead cells. Cell 140, 619–630. doi: 10.1016/j.cell.2010.02.014 Nagata, S., Hanayama, R., and Kawane, K. (2010). Autoimmunity and the clearance Nagata, S., Hanayama, R., and Kawane, K. (2010). Autoimmunity and the clearance of dead cells Cell 140, 619–630 doi: 10 1016/j cell 2010 02 014 Nagata, S., Hanayama, R., and Kawane, K. (2010). Autoimmunity an of dead cells. Cell 140, 619–630. doi: 10.1016/j.cell.2010.02.014 of dead cells. Cell 140, 619–630. doi: 10.1016/j.cell.2010.02.01 Neirinckx, V., Coste, C., Franzen, R., Gothot, A., Rogister, B., and Wislet, S. (2014). Neutrophil contribution to spinal cord injury and repair. J. Neuroinﬂamm. 11:150. doi: 10.1186/s12974-014-0150-2 Kong, X., Tang, X., Du, W., Tong, J., Yan, Y., Zheng, F., et al. (2013). Extracellular acidosis modulates the endocytosis and maturation of macrophages. Cell. Immunol. 281, 44–50. doi: 10.1016/j.cellimm.2012.12.009 Neri, T., Scalise, V., Passalacqua, I., Giusti, I., Lombardi, S., Balia, C., et al. (2018). CD18-mediated adhesion is required for the induction of a proinﬂammatory phenotype in lung epithelial cells by mononuclear cell-derived extracellular vesicles. Exp. Cell Res. 365, 78–84. doi: 10.1016/j.yexcr.2018.02.023 Korn, T., and Kallies, A. (2017). T cell responses in the central nervous system. Nat. Rev. Immunol. 17:179. doi: 10.1038/nri.2016.144 Kubota, K., Saiwai, H., Kumamaru, H., Maeda, T., Ohkawa, Y., Aratani, Y., et al. (2012). Myeloperoxidase exacerbates secondary injury by generating highly reactive oxygen species and mediating neutrophil recruitment in experimental spinal cord injury. Spine 37, 1363–1369. doi: 10.1097/BRS.0b013e31824b9e77 Nimmerjahn, A., Kirchhoﬀ, F., and Helmchen, F. (2005). Resting microglial cells are highly dynamic surveillants of brain parenchyma in vivo. Science 308, 1314–1318. doi: 10.1126/science.1110647 Norris, G. T., and Kipnis, J. (2019). REFERENCES Circulation doi: 10.1161/ CIRCULATIONAHA.121.054360 frontal cortex of subjects with bipolar disorder. J. Psychiatry Neurosci. 31, 189–196. Suzuki, H., Savitz, J., Kent Teague, T., Gandhapudi, S. K., Tan, C., Misaki, M., et al. (2017). Altered populations of natural killer cells, cytotoxic T lymphocytes, and regulatory T cells in major depressive disorder: Association with sleep disturbance. Brain Behav. Immun. 66, 193–200. doi: 10.1016/j.bbi.2017. 06.011 Rehncrona, S. (1985). Brain acidosis. Ann. Emerg. Med. 14, 770–776. doi: 10.1016/ S0196-0644(85)80055-X Reis e Sousa, C. (2006). Dendritic cells in a mature age. Nat. Rev. Immunol. 6, 476–483. doi: 10.1038/nri1845 Suzuki, Y., Nagai, N., and Umemura, K. (2016). A review of the mechanisms of blood-brain barrier permeability by tissue-type plasminogen activator treatment for cerebral ischemia. Front. Cell. Neurosci. 10:2. doi: 10.3389/fncel. 2016.00002 Riemann, A., Wußling, H., Loppnow, H., Fu, H., Reime, S., and Thews, O. (2016). Acidosis diﬀerently modulates the inﬂammatory program in monocytes and macrophages. Biochim. Biophys. Acta 1862, 72–81. doi: 10.1016/j.bbadis.2015. 10.017 Tan, J., Xu, Y. P., Liu, G. P., and Ye, X. H. (2013). Involvement of acid-sensing ion channel 1a in functions of cultured human retinal pigment epithelial cells. J. Huazhong Univ. Sci. Technol. Med. Sci. 33, 137–141. doi: 10.1007/s11596-013- 1086-y Rock, K. L., Latz, E., Ontiveros, F., and Kono, H. (2010). The sterile inﬂammatory response. Ann. Rev. Immunol. 28, 321–342. doi: 10.1146/annurev-immunol- 030409-101311 Rotstein, O. D., Fiegel, V. D., Simmons, R. L., and Knighton, D. R. (1988). The deleterious eﬀect of reduced pH and hypoxia on neutrophil migration in vitro. J. Surg. Res. 45, 298–303. doi: 10.1016/0022-4804(88)90079-0 Taoka, Y., Okajima, K., Murakami, K., Johno, M., and Naruo, M. (1998). Role of neutrophil elastase in compression-induced spinal cord injury in rats. Brain Res. Bull. 799, 264–269. doi: 10.1016/S0006-8993(98)00459-4 Tian, Y., Bresenitz, P., Reska, A., El Moussaoui, L., Beier, C. P., and Gründer, S. (2017). Glioblastoma cancer stem cell lines express functional acid sensing ion channels ASIC1a and ASIC3. Sci. Rep. 7:13674. doi: 10.1038/s41598-017- 13666-9 Sagar, D., Lamontagne, A., Foss, C. A., Khan, Z. K., Pomper, M. G., and Jain, P. (2012). Dendritic cell CNS recruitment correlates with disease severity in EAE via CCL2 chemotaxis at the blood–brain barrier through paracellular transmigration and ERK activation. J. Neuroinﬂamm. 9:245. doi: 10.1186/1742- 2094-9-245 Tong, J., Wu, W.-N., Kong, X., Wu, P.-F., Tian, L., Du, W., et al. (2011). Acid- sensing ion channels contribute to the eﬀect of acidosis on the function of dendritic cells. J. Immunol. REFERENCES 186, 3686–3692. Savio, L. E. B., de Andrade Mello, P., da Silva, C. G., and Coutinho-Silva, R. (2018). The P2X7 receptor in inﬂammatory diseases: angel or demon? Front. Pharmacol. 6:52. doi: 10.3389/fphar.2018.00052 Town, T., Tan, J., Flavell, R. A., and Mullan, M. (2005). T-cells in Alzheimer’s disease. Neuro Mol. Med. 7, 255–264. doi: 10.1385/NMM:7:3:255 Schulz, C., Chorro, L., Szabo-Rogers, H., Cagnard, N., Kierdorf, K., Prinz, M., et al. (2012). A lineage of myeloid cells independent of myb and hematopoietic stem cells. Science 336, 86–90. doi: 10.1126/science.1219179 Trevani, A. S., Andonegui, G., Giordano, M., López, D. H., Gamberale, R., Minucci, F., et al. (1999). Extracellular acidiﬁcation induces human neutrophil activation. J. Immunol. 162, 4849–4857. Schulz, J. B., Huang, P. L., Matthews, R. T., Passov, D., Fishman, M. C., and Beal, M. F. (1996). Striatal malonate lesions are attenuated in neuronal nitric oxide synthase knockout mice. J. Neurochem. 67, 430–433. Tsang, T. M., Woodman, B., McLoughlin, G. A., Griﬃn, J. L., Tabrizi, S. J., Bates, G. P., et al. (2006). Metabolic characterization of the R6/2 transgenic mouse model of Huntington’s disease by high-resolution MAS 1H NMR spectroscopy. J. Proteome Res. 5, 483–492. Serrano, C., Fraticelli, A., Paniccia, R., Teti, A., Noble, B., Corda, S., et al. (1996). pH dependence of neutrophil-endothelial cell adhesion and adhesion molecule expression. Am. J. Physiol. Cell Physiol. 271, C962–C970. Turner, M. D., Nedjai, B., Hurst, T., and Pennington, D. J. (2014). Cytokines and chemokines: at the crossroads of cell signalling and inﬂammatory disease. Biochim. Biophys. Acta 1843, 2563–2582. Sherwood, T. W., Lee, K. G., Gormley, M. G., and Askwith, C. C. (2011). Heteromeric acid-sensing ion channels (ASICs) composed of ASIC2b and ASIC1a display novel channel properties and contribute to acidosis-induced neuronal death. J. Neurosci. 31, 9723–9734. doi: 10.1523/JNEUROSCI.1665-11. 2011 Vaeth, M., and Feske, S. (2018). Ion channelopathies of the immune system. Curr. Opin. Immunol. 52, 39–50. doi: 10.1016/j.coi.2018.03.021 Vergo, S., Craner, M., Etzensperger, R., Attﬁeld, K., Friese, M., Newcombe, J., et al. (2011). Acid-sensing ion channel 1 is involved in both axonal injury and demyelination in multiple sclerosis and its animal model. Brain 134, 571–584. doi: 10.1093/brain/awq337 Shteinikov, V. Y., Korosteleva, A. S., Tikhonova, T. B., Potapieva, N. N., and Tikhonov, D. B. (2017). Ligands of histamine receptors modulate acid-sensing ion channels. Biochem. Biophys. Res. Commun. 490, 1314–1318. doi: 10.1016/j. bbrc.2017.07.019 Vermeulen, M. E., Gamberale, R., Trevani, A. S., Martinez, D., Ceballos, A., Sabatte, J., et al. REFERENCES (2004). The impact of extracellular acidosis on dendritic cell function. Crit. Rev. Immunol. 24, 363–384. doi: 10.1615/CritRevImmunol.v24.i5.40 Sinescu, C., Popa, F., Grigorean, V. T., Onose, G., Sandu, A., Popescu, M., et al. (2010). Molecular basis of vascular events following spinal cord injury. J. Med. Life 3, 254–261. Vig, P. J. S., Hearst, S. M., Shao, Q. M., and Lopez, M. E. (2014). Knockdown of Acid-Sensing Ion Channel 1a (ASIC1a) suppresses disease phenotype in SCA1 mouse model. Cerebellum 13, 479–490. doi: 10.1007/s12311-014-0563-6 Sluka, K., Price, M. P., Breese, N., Stucky, C., Wemmie, J., and Welsh, M. (2003). Chronic hyperalgesia induced by repeated acid injections in muscle is abolished by the loss of ASIC3, but not ASIC1. Pain 106, 229–239. doi: 10.1016/S0304- 3959(03)00269-0 Voilley, N., de Weille, J., Mamet, J., and Lazdunski, M. (2001). Nonsteroid anti- inﬂammatory drugs inhibit both the activity and the inﬂammation-induced expression of acid-sensing ion channels in nociceptors. J. Neurosci. 21, 8026– 8033. doi: 10.1523/JNEUROSCI.21-20-08026.2001 Smith, E. S., Cadiou, H., and McNaughton, P. A. (2007). Arachidonic acid potentiates acid-sensing ion channels in rat sensory neurons by a direct action. Neuroscience 145, 686–698. doi: 10.1016/j.neuroscience.2006.12.024 Vukicevic, M., and Kellenberger, S. (2004). Modulatory eﬀects of acid-sensing ion channels on action potential generation in hippocampal neurons. Am. J. Physiol. Cell Physiol. 287, C682–C690. Smoller, J. W., Gallagher, P. J., Duncan, L. E., McGrath, L. M., Haddad, S. A., Holmes, A. J., et al. (2014). The human ortholog of acid-sensing ion channel gene ASIC1a is associated with panic disorder and amygdala structure and function. Biol. Psychiat. 76, 902–910. doi: 10.1016/j.biopsych.2013.12.018 Waldmann, R., Champigny, G., Bassilana, F., Heurteaux, C., and Lazdunski, M. (1997). A proton-gated cation channel involved in acid-sensing. Nature 386, 173. doi: 10.1038/386173a0 Song, N., Lu, Z., Zhang, J., Shi, Y., Ning, Y., Chen, J., et al. (2019). Acid-sensing ion channel 1a is involved in ischaemia/reperfusion induced kidney injury by increasing renal epithelia cell apoptosis. J. Cell. Mol. Med. 23, 3429–3440. doi: 10.1111/jcmm.14238 Waldmann, R., Champigny, G., Lingueglia, E., De Weille, J. R., Heurteaux, C., and Lazdunski, M. (1999). H+-gated cation channels. Ann. N.Y. Acad. Sci. 868, 67–76. j Stranahan, A. M., Hao, S., Dey, A., Yu, X., and Baban, B. (2016). Blood-brain barrier breakdown promotes macrophage inﬁltration and cognitive impairment in leptin receptor-deﬁcient mice. J. Cerebral. Blood Flow Metab. 36, 2108–2121. doi: 10.1177/0271678X16642233 Wang, J.-J., Liu, F., Yang, F., Wang, Y.-Z., Qi, X., Li, Y., et al. (2020). REFERENCES Genetic exploration of the role of acid-sensing ion channels. Neuropharmacology 94, 99–118. doi: 10.1016/j. neuropharm.2014.12.011 Ponath, G., Park, C., and Pitt, D. (2018). The role of astrocytes in multiple sclerosis. Front. Immunol. 9:217. doi: 10.3389/ﬁmmu.2018.00217 Luck, K., Kim, D. K., Lambourne, L., Spirohn, K., Begg, B. E., Bian, W., et al. (2020). A reference map of the human binary protein interactome. Nature 580, 402–408. doi: 10.1038/s41586-020-2188-x Popovich, G. P., Guan, F. Z., McGaughy, M. V., Fisher, M. L., Hickey, M. W., and Basso, M. D. (2002). The neuropathological and behavioral consequences of intraspinal microglial/macrophage activation. J. Neuropathol. Exp. Neurol. 61, 623–633. doi: 10.1093/jnen/61.7.623 Lv, R. J., He, J. S., Fu, Y. H., Zhang, Y. Q., Shao, X. Q., Wu, L. W., et al. (2011). ASIC1a polymorphism is associated with temporal lobe epilepsy. Epilepsy Res. 96, 74–80. doi: 10.1016/j.eplepsyres.2011.05.002 Radu, B. M., Dumitrescu, D. I., Marin, A., Banciu, D. D., Iancu, A. D., Selescu, T., et al. (2014). Advanced type 1 diabetes is associated with ASIC alterations in mouse lower thoracic dorsal root ganglia neurons. Cell Biochem. Biophys. 68, 9–23. doi: 10.1007/s12013-013-9678-5 Lynagh, T., Mikhaleva, Y., Colding, J. M., Glover, J. C., and Pless, S. A. (2018). Acid- sensing ion channels emerged over 600 Mya and are conserved throughout the deuterostomes. Proc. Natl. Acad. Sci. USA 115, 8430–8435. doi: 10.1073/pnas. 1806614115 Ramesh, G., Benge, S., Pahar, B., and Philipp, M. T. (2012). A possible role for inﬂammation in mediating apoptosis of oligodendrocytes as induced by the Lyme disease spirochete Borrelia burgdorferi. J. Neuroinﬂamm. 9, 72. doi: 10. 1186/1742-2094-9-72 Marchi, N., Granata, T., Ghosh, C., and Janigro, D. (2012). Blood-brain barrier dysfunction and epilepsy: pathophysiologic role and therapeutic approaches. Epilepsia 53, 1877–1886. doi: 10.1111/j.1528-1167.2012.03637.x Rash, L. D. (2017). Acid-sensing ion channel pharmacology, past, present, and future. Adv. Pharmacol. 79, 35–66. doi: 10.1016/bs.apha.2017.02.001 Martínez, D., Vermeulen, M., von Euw, E., Sabatté, J., Maggíni, J., Ceballos, A., et al. (2007). Extracellular acidosis triggers the maturation of human dendritic September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org 14 ASICs in CNS Immune Cells Foster et al. frontal cortex of subjects with bipolar disorder. J. Psychiatry Neurosci. 31, 189–196. Redd, M. A., Scheuer, S. E., Saez, N. J., Yoshikawa, Y., Chiu, H. S., Gao, L., et al. (2021). Therapeutic inhibition of acid sensing ion channel 1a recovers heart function after ischemia-reperfusion injury. REFERENCES Disruption of auto-inhibition underlies conformational signaling of ASIC1a to induce neuronal necroptosis. Nat. Commun. 11:475. doi: 10.1038/s41467-019-13873-0 Wang, R. I., and Sonnenschein, R. R. (1955). pH of cerebral cortex during induced convulsions. J. Neurophysiol. 18, 130–137. doi: 10.1152/jn.1955.18. 2.130 Sun, X., Wang, J.-F., Tseng, M., and Young, L. T. (2006). Downregulation in components of the mitochondrial electron transport chain in the postmortem September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org 15 Foster et al. ASICs in CNS Immune Cells Wang, Y.-Z., Wang, J.-J., Huang, Y., Liu, F., Zeng, W.-Z., Li, Y., et al. (2015). Tissue acidosis induces neuronal necroptosis via ASIC1a channel independent of its ionic conduction. eLife 4:e05682. doi: 10.7554/eLife.05682 Yu, X.-W., Hu, Z.-L., Ni, M., Fang, P., Zhang, P.-W., Shu, Q., et al. (2015). Acid- sensing ion channels promote the inﬂammation and migration of cultured rat microglia. Glia 63, 483–496. doi: 10.1002/glia.22766 f Waxman, S. G. (2006). Axonal conduction and injury in multiple sclerosis: the role of sodium channels. Nat. Rev. Neurosci. 7, 932–941. doi: 10.1038/nrn2023 Yuan, F. L., Chen, F. H., Lu, W. G., Li, X., Li, J. P., Li, C. W., et al. (2010). Inhibition of acid-sensing ion channels in articular chondrocytes by amiloride attenuates articular cartilage destruction in rats with adjuvant arthritis. Inﬂamm. Res. 59, 939–947. doi: 10.1007/s00011-010-0206-4 Wemmie, J. A., Chen, J., Askwith, C. C., Hruska-Hageman, A. M., Price, M. P., Nolan, B. C., et al. (2002). The acid-activated ion channel ASIC contributes to synaptic plasticity, learning, and memory. Neuron 34, 463–477. doi: 10.1016/ s0896-6273(02)00661-x Zamanian, J., Xu, L., Foo, L. C., Nouri, N., Zhou, L., Giﬀard, R. G., et al. (2012). Genomic analysis of reactive astrogliosis. J. Neurosci. 32, 6391–6410. doi: 10. 1523/JNEUROSCI.6221-11.2012 Wemmie, J. A., Coryell, M. W., Askwith, C. C., Lamani, E., Leonard, A. S., Sigmund, C. D., et al. (2004). Overexpression of acid-sensing ion channel 1a in transgenic mice increases acquired fear-related behavior. Proc. Natl. Acad. Sci. USA 101, 3621–3626. doi: 10.1073/pnas.0308753101 Zampieri, F. G., Kellum, J. A., Park, M., Ranzani, O. T., Barbeiro, H. V., de Souza, H. P., et al. (2014). Relationship between acid-base status and inﬂammation in the critically ill. Crit. Care 18:R154. doi: 10.1186/cc13993 Zeinstra, E., Wilczak, N., Streeﬂand, C., and De Keyser, J. (2000). Astrocytes in chronic active multiple sclerosis plaques express MHC class II molecules. NeuroReport 11, 89–91. REFERENCES doi: 10.1097/00001756-200001170-0 0018 Wemmie, J., Askwith, C., Lamani, E., Cassell, M. D., Freeman, J., and Welsh, M. (2003). Acid-sensing ion channel 1 is localized in brain regions with high synaptic density and contributes to fear conditioning. J. Neurosci. 23, 5496–5502. Wemmie, J., Taugher, R., and Kreple, C. (2013). Acid-sensing ion channels in pain and disease. Nat. Rev. Neurosci. 14, 461–471. doi: 10.1038/nrn3529 Zeng, W., Liu, D., Duan, B., Song, X. L., Wang, X., Wei, D., et al. (2013). Molecular mechanism of constitutive endocytosis of acid-sensing ion channel 1a and its protective function in acidosis-induced neuronal death. J. Neurosci. 33, 7066–7078. doi: 10.1523/JNEUROSCI.5206-12.2013 Weng, J. Y., Lin, Y. C., and Lien, C. C. (2010). Cell type-speciﬁc expression of acid- sensing ion channels in hippocampal interneurons. J. Neurosci. 30, 6548–6558. doi: 10.1523/jneurosci.0582-10.2010 Zha, X. M., Wemmie, J. A., Green, S. H., and Welsh, M. J. (2006). Acid-sensing ion channel 1a is a postsynaptic proton receptor that aﬀects the density of dendritic spines. Proc. Natl. Acad. Sci. USA 103, 16556–16561. doi: 10.1073/ pnas.0608018103 Wong, G. H. W., Bartlett, P. F., Clark-Lewis, I., McKimm-Breschkin, J. L., and Schrader, J. W. (1984). Interferon-γ induces the expression of H-2 and Ia antigens on brain cells. J. Neuroimmunol. 7, 255–278. doi: 10.1016/S0165- 5728(84)80026-0 Zhou, Z.-H., Song, J.-W., Li, W., Liu, X., Cao, L., Wan, L.-M., et al. (2017). The acid- sensing ion channel, ASIC2, promotes invasion and metastasis of colorectal cancer under acidosis by activating the calcineurin/NFAT1 axis. J. Exp. Clin. Cancer Res. 36:130. doi: 10.1186/s13046-017-0599-9 Wong, H. K., Bauer, P. O., Kurosawa, M., Goswami, A., Washizu, C., Machida, Y., et al. (2008). Blocking acid-sensing ion channel 1 alleviates Huntington’s disease pathology via an ubiquitin-proteasome system-dependent mechanism. Hum. Mol. Genet. 17, 3223–3235. doi: 10.1093/hmg/ddn218 Ziemann, A. E., Allen, J. E., Dahdaleh, N. S., Drebot, I. I., Coryell, M. W., Wunsch, A. M., et al. (2009). The amygdala is a chemosensor that detects carbon dioxide and acidosis to elicit fear behavior. Cell 139, 1012–1021. doi: 10.1016/j.cell.2009. 10.029 Woodﬁn, A., Voisin, M.-B., and Nourshargh, S. (2007). PECAM-1: a multi- functional molecule in inﬂammation and vascular biology. Arterioscler. Thromb. Vasc. Biol. 27, 2514–2523. doi: 10.1161/ATVBAHA.107.15 1456 Ziemann, A. E., Schnizler, M. K., Albert, G. W., Severson, M. A., Howard, M. A., Welsh, M. J., et al. (2008). Seizure termination by acidosis depends on ASIC1a. Nat. Neurosci. 11, 816–822. doi: 10.1038/nn.2132 Wu, Y., Gao, B., Xiong, Q. J., Wang, Y. REFERENCES C., Huang, D. K., and Wu, W. N. (2017). Acid-sensing ion channels contribute to the eﬀect of extracellular acidosis on proliferation and migration of A549 cells. Tumour Biol. 39:1010428317705750. doi: 10.1177/1010428317705750 Conﬂict of Interest: 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. Xiong, Z. G., Chu, X. P., and Simon, R. P. (2007). Acid sensing ion channels— novel therapeutic targets for ischemic brain injury. Front. Biosci. 12, 1376–1386. doi: 10.2741/2154 Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their aﬃliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Xiong, Z.-G., Zhu, X.-M., Chu, X.-P., Minami, M., Hey, J., Wei, W.-L., et al. (2004). Neuroprotection in ischemia: blocking calcium-permeable acid-sensing ion channels. Cell 118, 687–698. Yamasaki, R., Lu, H., Butovsky, O., Ohno, N., Rietsch, A. M., Cialic, R., et al. (2014). Diﬀerential roles of microglia and monocytes in the inﬂamed central nervous system. J. Exp. Med. 211, 1533–1549. doi: 10.1084/jem.20132477 Copyright © 2021 Foster, Rash, King and Rank. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Yang, F., Sun, X., Ding, Y., Ma, H., Yang, T. O., Ma, Y., et al. (2016). Astrocytic acid-sensing ion channel 1a contributes to the development of chronic epileptogenesis. Sci. Rep. 6:31581. doi: 10.1038/srep31581 Yermolaieva, O., Leonard, A. S., Schnizler, M. K., Abboud, F. M., and Welsh, M. J. (2004). Extracellular acidosis increases neuronal cell calcium by activating acid-sensing ion channel 1a. Proc. Natl. Acad. Sci. USA 101, 6752–6757. September 2021 \| Volume 15 \| Article 738043 Frontiers in Cellular Neuroscience \| www.frontiersin.org 16
https://openalex.org/W3210540855	https://econpapers.repec.org/scripts/redir.pf?u=https%3A%2F%2Fredfame.com%2Fjournal%2Findex.php%2Faef%2Farticle%2Fdownload%2F5369%2F5598;h=repec:rfa:aefjnl:v:8:y:2021:i:6:p:10-20	English	null	How Poverty, Residence Status and Health Insurance Predict Unmet Healthcare Needs among Chinese Elders?	Applied economics and finance	2,021	cc-by	7,444	1. Introduction China is facing a growing aging population. In 2000, when China stepped into the aging society, the percentage of people over the age of 65 was 7%. That percentage rose to 11.9% in 2018 (National Bureau of Statistics of China) and is expected to reach 22.6% by 2050 (United Nation, 2002). Since healthcare needs tend to increase with advanced age, a rapidly aging population inevitably leads to greater needs for healthcare services. Such needs, however, cannot be met with current levels of social security benefits and institutional care. This is an undesirable feature of the Chinese healthcare and social systems. The latest statistics from the Chinese Health Ministry show that one third of older patients in China have unmet healthcare needs, meaning they are not receiving needed healthcare (Center for Health Statistics and Information, 2014). Since ensuring everyone has access to necessary health services is required under an equal healthcare system, it is important to understand why some fail to obtain the healthcare they need. In this paper, we assess unmet healthcare needs of Chinese elders and investigate factors associated with these unmet needs. Unmet health care needs, defined as the difference between the health care services deemed necessary and the actual services received, represent a measure of access to healthcare. There is a dearth of empirical research aimed at identifying the factors that contribute to unmet healthcare needs among Chinese elders. However, theoretical research in developed countries offers a number of possibilities. In general, unmet healthcare needs depend on the particular circumstances of the individuals seeking care and the characteristics of the healthcare system. In this study, we draw on an existing dataset and evaluate the empirical validity of these hypotheses in a sample of elders in China. We examine socioeconomic vulnerabilities in old adults, looking especially at their heterogeneity. We also investigate how the distribution of economic and social resources among elders affects the degree of their unmet healthcare needs. Although socioeconomic and institutional mechanisms have long been the focus of public health research on healthcare needs which are met (or healthcare utilization), little is known about how these are associated with potential or unmet healthcare needs in old age. As for socioeconomic vulnerabilities, poverty and rural residence may be potential risk factors for elders’ unmet healthcare needs in China. How Poverty, Residence Status and Health Insurance Predict Unmet Healthcare Needs among Chinese Elders? Lili Wu1, Chunyin Wang1 & Jiayun Kou1 1 School of Economics and Management, China University of Petroleum-Beijing, Beijing, China Correspondence: Lili Wu, School of Economics and Management, China University of Petroleum-Beijing, Beijing, China. Received: October 3, 2021 doi:10.11114/aef.v8i6.5369 Accepted: October 22, 2021 Available online: October 26, 2021 URL: https://doi.org/10.11114/aef.v8i6.5369 Received: October 3, 2021 Accepted: October 22, 2021 Available online: October 26, 2021 doi:10.11114/aef.v8i6.5369 URL: https://doi.org/10.11114/aef.v8i6.5369 Received: October 3, 2021 doi:10.11114/aef.v8i6.5369 Abstract This study focuses on the variability in unmet healthcare needs among vulnerable Chinese elders and the degree to which these unmet needs are associated with socioeconomic disadvantages. We use the 2013 wave of China Health and Retirement Longitudinal Study (CHARLS) and a multinomial logistic model to investigate how poverty, residence status and particular health insurance schemes influence unmet healthcare needs independently and in combination. Our results show that poverty and rural residence are strong predictors of unmet healthcare needs due to financial and non-financial constraints, respectively. Although health insurance can reduce financial barriers, its influence varies with different insurance schemes, thus generating unequal healthcare access among heterogeneous vulnerable subgroups of elders and putting poor rural migrants at the highest risk for unmet healthcare needs. Our findings direct attention to the differences in resources available to various subgroups of elders and the importance of social stratification in predicting unmet health care needs. Keywords: poverty, residence status, health insurance, unmet health care needs, elders Applied Economics and Finance Vol. 8, No. 6; November 2021 ISSN 2332-7294 E-ISSN 2332-7308 Published by Redfame Publishing URL: http://aef.redfame.com Applied Economics and Finance Vol. 8, No. 6; November 2021 ISSN 2332-7294 E-ISSN 2332-7308 Published by Redfame Publishing URL: http://aef.redfame.com Applied Economics and Finance Vol. 8, No. 6; November 2021 ISSN 2332-7294 E-ISSN 2332-7308 Published by Redfame Publishing URL: http://aef.redfame.com 1. Introduction According to the WHO World Health Survey 2015 (World Health Organization, 2015), more than 60% of elderly people in China have unmet healthcare needs due to poverty, in contrast to less than 16% in 10 Applied Economics and Finance Vol. 8, No. 6; 2021 developed countries. In China, poverty is also related to unmet needs as it may reflect how place of residence makes accessing healthcare services more difficult. Evidence shows that the majority of poor elders live in remote rural areas, where the lack of higher quality medical personnel and the low efficiency of medical services leave much of the population with health service utilization shortages (Zheng, Gong, & Zhang, 2019). Elders living in rural areas with low socioeconomic status may face multiple risk factors for unmet healthcare needs, but to date, little attention has been paid to examining to what extent elders’ poverty and residence status influence their unmet healthcare needs independently and in combination. Any analysis of unmet healthcare needs among Chinese elders must also consider equality and efficiency within the healthcare system, as well as the mechanism for distributing healthcare benefits. During the past decades, the Chinese healthcare system has undergone a series of reforms, with the goal of ensuring equal access to needed healthcare. The state managed to achieve near-universal health insurance coverage by 2011 (Cheng & T.-M., 2012; Zhang, Nikoloski, & Mossialos, 2017). This remarkable progress has attracted a large body of research into the association between China’s providing health insurance and the healthcare subsequently received (Huang & Gan, 2015; H. Liu & Zhao, 2014; Q. Liu, Wang, Kong, & Cheng, 2011; Ye, Shia, Fang, & Lee, 2019; Zhang et al., 2017). But few studies have investigated how effective China’s health insurance has been in reducing the possibility of not receiving needed healthcare. In addition to this uncertainty surrounding health insurance’s overall effects, the distributional effects of China’s health insurance on unmet healthcare needs are also unclear. In China, the basic insurance system consists of three schemes: Urban Employee Basic Medical Insurance (UEBMI), Urban Resident Basic Medical Insurance (URBMI), and New Rural Cooperative Medical Scheme (NRCMS). These schemes target different populations based on their residence registration (hukou) and/or employment status. With different benefit packages being provided between urban and rural hukou holders (Jin, Hou, Zhang, & Zhang, 2016), fragmented health insurance schemes have been suggested to be inextricably linked with inequity in healthcare access. 1. Introduction Although there is abundant evidence showing the correlation between socioeconomic disadvantage, as well as institutional constraints, and unmet healthcare needs, current knowledge is only partial due to considering only a limited number of factors. It was well-known that the dual economy and society in China have caused a distinct welfare system, under which its health system also features an ―urban–rural duality.‖ In this system, inequalities and discrimination that stem from hukou restrictions still exist and exacerbate social differentiation. The discrimination has led to unequal outcomes associated with access to healthcare services. Moreover, the hardships of socioeconomic disadvantages and institutional discrimination may cluster, potentially making those caught in the ―urban–rural duality‖ gap the most vulnerable subgroup not offered strong institutional protection against unmet healthcare needs. However, the relationship between socioeconomic disadvantage and unmet healthcare needs — and how the association varies by health institutional arrangements — is still unknown. Therefore, in addition to studying variables such as poverty and residence status, it is necessary to examine the relationship between institutional discrimination and unmet needs for healthcare. Further, there is a need for more evidences with which to assess the effectiveness of the various health-focused interventions being offered across vulnerable population subgroups. The aim of this paper is to bridge the above research gaps. Empirically, we apply the 2013 China Health and Retirement Longitudinal Study (CHARLS) to capture updated evidence on health utilization gaps among the older people in China. We pose these research questions: Has health insurance reduced unmet healthcare needs of the elderly? Have differences in insurance schemes contributed to inequality in older people’s potential healthcare utilization? What are the effects of institutional and socioeconomic characteristics on older people’s access to healthcare services? By answering these questions, we generate important insights to better understand healthcare inequality in recent years. Additionally, in order to deepen our understanding of the unique resource allocation mechanisms in China, and the resulting patterns of healthcare inequality, we develop a framework for analyzing the Chinese healthcare system. To this end, we study both financial and non-financial barriers to health service access, and conceptualize two distinct sets of indicators related to inequality in health service access. Through this investigation, we can reveal not only the overall consequences but also the distributional effects of Chinese health reforms in terms of unmet healthcare needs. 2. Background In China, there is no specific medical coverage scheme targeting old people. Instead, the elderly population participates in a universal basic health insurance system. This insurance system covers 95% of the entire population in China, and tends to create more equitable access to healthcare by reducing income related barriers and out-of-pocket payments for healthcare service. It mainly consists of three different schemes: the Urban Employee Basic Medical Insurance (UEBMI), the Urban Resident Basic Medical Insurance (URBMI), and the New Rural Cooperative Medical Scheme (NRCMS). These insurance schemes are designed to cover different people: UEBMI is intended to cover urban employees and retirees in formal sectors; URBMI provides insurance to informal-sector employees, the self-employed and unemployed; and NRCMS aims to cover the registered rural population. Clearly, these three schemes are classified 11 Applied Economics and Finance Vol. 8, No. 6; 2021 mainly by hukou status, which has long been regarded as one of the most significant institutions in China in defining persons’ access to social welfare. Although insurance should offer protection for vulnerable groups such as the poor and older adults, inequalities in schemes arising from hukou policy may give rise to disparities in healthcare access and utilization. These three insurance schemes have different regulations, premiums and coverage amounts. Generally speaking, UEBMI provides the most generous benefit packages, whereas NRCMS is usually regarded as the most rudimentary type of insurance (Jin et al., 2016), with the highest deductibles, lowest reimbursement rates, and highest coinsurance rates (Lei & Lin, 2009; Wagstaff, Lindelow, Gao, Xu, & Qian, 2009). Many services, particularly outpatient care, are not covered, or only partially covered, by NRCMS. An individual’s ability to upgrade insurance schemes is tightly constrained because one’s hukou status is difficult to change (e.g. from rural to urban) (Jin et al., 2016). Meanwhile, China’s long-lasting urban-rural structure results in enormous differences between urban and rural areas in terms of the economy, government subsidies, infrastructure construction and so on. These gaps might be associated with urban–rural medical resources. In China’s rural regions, there are three types of healthcare institutions: village clinics, township hospitals, and county (and higher level) hospitals. Since NRCMS has higher reimbursement rates for health services provided by local healthcare institutions, rural patients are encouraged to seek healthcare firstly from primary village clinics and only go to higher level care facilities if needed (Xi Li, Lu, Hu, Cheng, & Hu, 2017). 3.1 Data We draw our data from the 2013 wave of the China Health and Retirement Longitudinal Study (CHARLS), a broad-purposed social science and health survey of residents aged 45 or above in continental China. As a high-quality nationally representative survey, CHARLS 2013 adopted multi-stage stratified Probability Proportional to Size (PPS) sampling and covered 18,621 individuals scattered over 450 villages/urban communities in 28 provinces. A major element of this survey was devoted to recording health status, health-related behaviors, healthcare utilization and health expenditures. It also included questions pertaining to unmet healthcare needs of the population. CHARLS asked respondents whether or not they have needed healthcare but did not receive the needed services. Those answering yes, were also asked to provide their reasons for not receiving the services. In addition, the survey contained rich information on demographic background, residence, household income, and the type(s) of health insurance. Therefore, CHARLS offers a strong dataset with which to study the distribution of unmet healthcare needs across different socioeconomic subgroups of elderly Chinese citizens. In this study, 8973 elders were selected who were 60 and older. After dropping non-responses to healthcare-related questions (0.2%) and covariates (5.0%), the selected sample includes 8000 individuals. 2. Background However, the quality of primary care is poor. Most village doctors have less than a high school education, and few of them have received professional training and supervision (Yang, 2018). With reference to a survey conducted among rural residents in three provinces in China, around 30% of residents rated quality of healthcare provided by village clinics as poor or very poor (Ratigan, 2015). Therefore, some patients bypass primary care facilities and go directly to higher level hospitals (Fe, Powell-Jackson, & Yip, 2016), while some forgo or postpone the utilization of healthcare. In addition to insurance and residence disadvantage, economic status is another potential risk factor for old people’s unmet healthcare needs. A strong association between low income status and unmet healthcare need has been consistently found in previous studies (S. Kim & Huh, 2011; Park, Kim, & Kim, 2016). However, there has been no research focusing on the combined effects of poverty and residence status. Under China’s dual structure, rural residents are more likely to be poor and less likely to have access to government-related income and social programs than urban residents (Pei & Xu, 2011). Thus, their access to health services is more likely to be affected by poverty because they receive fewer health insurance benefits. Moreover, health insurance is tied to the local hukou and managed by different government agencies (Xin Li & Zhang, 2013), which usually designate local health care facilities as their coverage network. This means that health services received non-locally might not be reimbursed by local insurance (Jin et al., 2016). This regional segregation poses a significant institutional barrier to rural migrants. With hukou registered at a rural county, migrants can have healthcare in their places of residence, but their NRCMS are not eligible for enrollment in urban destinations. Meanwhile, they are generally not eligible for enrollment in local urban insurance (except that URBMI is offered to migrants in some cities (Yip et al., 2012)). It follows that this situation would be most detrimental to poor rural migrants. With a number of urban hospitals refusing NRCMS, poverty would make rural migrants more sensitive to healthcare costs and more likely to forgo healthcare services. 3.3 Statistical Methods We applied multinomial logistic in a multilevel framework for two dependent variables — “not receiving needed outpatient services‖ and ―not receiving needed inpatient services‖ — to estimate the relative risk ratios of experiencing unmet healthcare needs. The possibility of unmet needs for healthcare was estimated separately because both socioeconomic disadvantage and health insurance could have different effects on these two unmet needs components. We added variables in a model-building process sequentially. Model 1 tested the main influences of poverty status and residence status on unmet need for healthcare, with predisposing and need factors added as covariates. In Model 2, health insurance factors were added to examine their effects on unmet needs. Model 3 estimated the joint effects by adding interaction terms between poverty, rural residence and insurance status. Unmet need variables Unmet need variables In this study unmet healthcare needs were measured by respondents’ subjective self-assessments that they had not 12 Applied Economics and Finance Vol. 8, No. 6; 2021 received the care that they needed. It included two aspects: ―not receiving needed outpatient services‖ and ―not receiving needed inpatient services.‖ In CHARLS data, ―not receiving needed outpatient services‖ was measured by respondents’ not visiting doctors although they had been ill during the previous 4 weeks; ―not receiving needed inpatient services‖ was defined as not getting hospitalized after receiving a doctor’s recommendation for hospital admission during the previous year. This subjective self-reported evaluation is helpful because respondents were able to assess their health status (Idler & Benyamini, 1997) and were more able to identify deficiencies in their healthcare experiences. In addition, the survey respondents’ reasons for forgoing medical care provide possible targets for policy actions (J. Kim, Kim, Park, & Cho, 2015). received the care that they needed. It included two aspects: ―not receiving needed outpatient services‖ and ―not receiving needed inpatient services.‖ In CHARLS data, ―not receiving needed outpatient services‖ was measured by respondents’ not visiting doctors although they had been ill during the previous 4 weeks; ―not receiving needed inpatient services‖ was defined as not getting hospitalized after receiving a doctor’s recommendation for hospital admission during the previous year. This subjective self-reported evaluation is helpful because respondents were able to assess their health status (Idler & Benyamini, 1997) and were more able to identify deficiencies in their healthcare experiences. In addition, the survey respondents’ reasons for forgoing medical care provide possible targets for policy actions (J. Kim, Kim, Park, & Cho, 2015). In CHARLS, the reasons for forgoing medical services grouped into several categories, including financial constraints, not having a serious illness, inconvenient traffic, no available treatment, poor healthcare service, and lack of time. Among these, financial burden dwarfed the other explanations for unmet healthcare need. For our analysis, we planned to focus mostly on respondents who reported financial barriers. Thus, we distinguished the types of unmet needs by coding no unmet needs as “0,” finance induced unmet needs as “1” and unmet needs due to any other reason as “2.” I d d t i bl Independent variables Poverty status. We used China’s national poverty line to measure poverty. This measure was chosen because it is adjusted according to household size and the consumer price index each year. This threshold was divided by disposable income to calculate income-to-needs ratio (INR). Following the previous research (Park et al., 2016), we classified poverty status into three categories. A respondent was categorized poor if the per-capita household income was lower than the poverty line, i.e. the INR fell below 1; near-poor if the INR ranked between 1 and 1.2; non-poor if the INR was higher than 1.2. Thus three dichotomous variables are generated. Residence status. The second key independent variable was the place of residence (current location). It was a binary indicator of whether a respondent currently resides in a rural or urban area (rural residence=1, urban residence=0). Covariates In addition to identifying two main independent variables, we controlled for a comprehensive set of covariates that could have potential associations with unmet healthcare needs. Following previous studies, we grouped these covariates into three categories: predisposing, enabling and need. Predisposing variables, indicating individuals’ propensity to use healthcare services, included age measured in 3 categories (60-69, 70-79 and 80 plus), gender, marital status (married or partnered, separated or divorced or widowed, single), education (illiterate, literate or primary school, junior high and above) and number of people within the household. Enabling variables included income and health insurance. The health insurance variables were represented by several dummy variables capturing UEBMI, URBMI, NCMS, GMI, and other insurance. We also used county fixed effect dummies to capture regional disparity within each province. Need factors reflecting the need for healthcare included self-reported health status, chronic conditions, functional impairments involving activities of daily living (ADLs) and instrumental ADLs (IALDs). Self-rated health was defined by good, fair and poor. Chronic condition was a dummy variable which equaled 1 if the respondent had any self-reported chronic diseases such as hypertension, dyslipidemia, diabetes, chronic lung diseases and so on. We also used a dummy variable to indicate whether the respondent has ADL difficulties. We measured IADLs based on 3 categories according to numbers of activities with difficulties (0, 1-3, or 4 and above). 4. Results 4.1 Sample Characteristics 4.1 Sample Characteristics Unmet healthcare needs were reported by 29.3% (95% CI: 27.6%-31.2%) of the study population. Table 1 shows the 13 Applied Economics and Finance Vol. 8, No. 6; 2021 descriptive characteristics of the sample, as well as their bivariate associations between poverty and residence status. Of 8000 participants, 51.3% of the sample was poor and 3.7% near poor; 60.3% lived in rural areas. Elders who were poor, near poor or living in rural areas were more likely to have financially-induced unmet healthcare needs. descriptive characteristics of the sample, as well as their bivariate associations between poverty and residence status. Of 8000 participants, 51.3% of the sample was poor and 3.7% near poor; 60.3% lived in rural areas. Elders who were poor, near poor or living in rural areas were more likely to have financially-induced unmet healthcare needs. Those who were poor were more likely to be female and less educated; they were also more likely to have poor self-rated health status, more chronic diseases, and more disabilities. Near poor elders had the biggest family size and were most likely to rate their health as fair. Poor (89.6%) and near poor elders (72.9%) were more likely to be insured under the NRCMS. Very few poor elders enjoyed the UEBMI (1.1%) or the URBMI (1.5%). The majority (77.8%) of the poor lived in rural regions. Compared to those who lived in urban areas, rural elders tended to be less educated and to have bigger families. They had poorer self-rated health status and a higher likelihood of having ADL and IADL difficulties. Most rural elders (88.47%) were insured by the NRCMS, and most of them (66.32%) were poor. Table 1. Sample descriptive at by poverty status and residence status of Chinese elders Table 1. 4. Results Sample descriptive at by poverty status and residence status of Chinese elders Poor (51.34%) Near poor (3.73%) Non-poor (44.93%) Rural (60.31%) Urban (39.69%) Unmet healthcare (%)  2 4  =28.66***  2 2  =3.29 Economic reason 5.75 4.17 1.99 4.51 3.23 Non-economic reason 23.15 32.29 27.31 24.66 26.42 Predisposing Age groups (%)  2 4  =5.23  2 2  =5.97* 60-69 61.20 70.83 62.92 63.75 60.18 70-79 32.00 21.88 29.73 30.07 31.41 80 and over 6.81 7.29 7.35 6.18 8.41 Female (%) 55.07 54.17 51.60  2 2  =3.00 54.02 52.64  2 1  =0.47 Marital status (%)  2 4  =10.90  2 2  =5.24* Married/partnered 77.61 76.04 81.59 78.82 80.14 Divorced /separated/widowed 21.18 20.83 17.72 19.77 19.37 Single 1.21 3.13 0.69 1.42 0.49 Education (%)  2 4  =278.76*  2 2  =245.03* Illiterate 46.07 35.42 23.42 43.14 23.87 Literate /primary school 45.31 43.75 43.04 46.17 41.29 Junior high and above 8.62 20.83 33.54 10.69 34.83 Family size (M) 4.77 5.70 5.29 F(1,7998)=42.66* 5.14 4.89 F(1,7998)=9.27* Health need Self-rated health (%)  2 4  =54.54*  2 2  =42.44* Poor 52.95 41.67 38.98 51.38 38.45 Fair 35.85 50.00 49.09 38.76 47.75 Good 11.20 8.33 11.93 9.85 13.80 Chronic numbers (M) 2.28 2.19 2.42 F(1,7998)=3.87 2.26 2.45 F(1,799)=7.57* ADLs difficulties (%) 13.09 8.33 10.80  2 2  =4.27 12.04 11.64  2 1  =0.09 IADLs difficulties (%)  2 4  =47.17*  2 2  =36.07* 0 activity 50.61 57.29 64.22 52.22 64.19 1-3 activities 38.73 35.42 28.35 37.80 28.08 4 or above activities 10.67 7.29 7.43 9.98 7.73 Enabling Health insurance (%)  2 8  =652.18*  2 4  =714.62* UEBMI only 1.56 4.17 30.08 2.96 31.31 URBMI only 1.51 8.33 7.69 1.16 9.69 NRCMS only 89.64 72.92 46.76 88.47 41.29 14 Vol. 8, No. 4. Results 6; 2021 Applied Economics and Finance Multiple or other 3.63 5.21 11.67 3.73 12.72 No insurance 4.16 9.38 3.80 3.67 4.99 Rural residence (%) 77.84 51.04 40.62  2 2  =359.81* Poverty status (%)  2 2  =355.39* Poor 66.32 28.57 Near poor 3.16 4.60 Non-poor 30.52 66.83 4.2 Multivariate Results Effects of health insurance and other factors on unmet outpatient needs 4.2 Multivariate Results Effects of health insurance and other factors on unmet outpatient needs Table 2 presents the results of the multinomial logistic analysis. Being poor was significantly associated with an increased risk of having unmet needs because of economic reasons. Regarding residence status, compared to elders living in urban areas, rural elders were more likely to have unmet needs due to non-economic reasons. Variables related to health insurance were included in Model 2. This model showed that the implementation of the multiple basic insurance programs significantly moderated the influences of poverty on unmet needs, with relative risk ratio (RRR) reduced from 2.32 to 1.94. Moreover, it showed that the insurance schemes have been at least partly effective in reducing unmet outpatient needs. Study elders with health insurance of any kind had a lower possibility of unmet needs than those who were not insured. However, different schemes led to varying outcomes: elders with UEBMI were significantly less likely to have unmet needs compared with those with NRCMS. In addition, although the multiple insurance schemes reduced financial barriers and lowered the risk of not receiving needed health services, insurance schemes were not associated with unmet needs due to non-financial reasons. An unequal distribution of medical care resources still existed, and rural residence significantly accounted for increasing risk of unmet needs. In Model 3, interaction terms between poverty, residence status and health insurance were included in order to examine the extent to which particular health insurance schemes (arising from hukou status) boosted the main effects of poverty and residence status on unmet needs. The results showed poor rural elders covered under NRCMS were more likely to have unmet needs due to financial reasons (RRR=1.5, p<0.1). The situation was more serious among poor urban residents insured by NRCMS (RRR=5.37, p<0.01). 92% of them were rural hukou holders who had migrated from a rural to an urban destination. 4. Results Effects of health insurance on unmet inpatient needs Among the elders who experienced unmet inpatient needs, 51% cited financial problems as the main reason for not accessing inpatient care. Within this group that cited financial obstacles, 66% came from poor of near poor families. Factors associated with unmet inpatient needs as revealed by the multinomial logistic analysis are presented in Table 3. After controlling for elders’ level of ill health, being poor was the main risk factor for unmet inpatient needs, with a more than 3-fold increase in poor elders compared with non-poor ones. Although health insurance coverage was protective against the risk of not receiving appropriate outpatient care for economic reasons, it did not significantly associate with reducing unmet needs of more expensive care, such as hospitalization. The Chinese insurance system was not effective enough to protect vulnerable elders from high healthcare expenses. By including interaction terms in Model 3, this result was even more obvious. Poor rural residents and rural-urban migrants had a significantly higher likelihood of experiencing unmet needs due to financial reasons (RRR=1.44, and 1.91 respectively). Coverage under NRCMS did not offer them enough protection from hospitalization. Table 2. Multinomial logistic model of unmet healthcare needs for Chinese elders (Outpatient) Financial reason Non- financial reason Financial reason Non- financial reason Financial reason Predisposing Age groups (ref.=60-69) 70-79 1.097 (0.26) 0.914 (0.10) 1.118 (0.26) 0.895 (0.10) 1.143 (0.27) 80 and over 0.572 (0.28) 0.916 (0.17) 0.635 (0.31) 0.896 (0.17) 0.675 (0.33) Female (%) 1.458 (0.34) 0.889 (0.09) 1.408 (0.33) 0.883 (0.09) 1.369 (0.32) Marital status (ref.=single) e 2. Multinomial logistic model of unmet healthcare needs for Chinese elders (Outpatient) 15 Applied Economics and Finance Vol. 8, No. 4. Results 6; 2021 pp , ; Married/partnered 0.639 (0.51) 0.748 (0.33) 0.802 (0.66) 0.656 (0.29) 0.835 (0.68) Divorced /separated/widowed 0.437 (0.36) 0.782 (0.35) 0.516 (0.44) 0.701 (0.32) 0.540 (0.46) Education (ref.= illiterate) Literate /primary school 1.463 (0.34) 1.054 (0.12) 1.542 (0.36) 1.022 (0.12) 1.509* (0.36) Junior high and above 0.338 (0.17) 0.737 (0.11) 0.426* (0.22) 0.699** (0.11) 0.401* (0.21) Family size (M) 0.953 (0.06) 1.025 (0.02) 0.938 (0.06) 1.032 (0.03) 0.918 (0.06) Health need Self-rated health (ref.=good) Poor 2.379* (1.27) 0.692 (0.11) 2.313 (1.24) 0.699 (0.11) 2.267 (1.22) Fair 1.899 (1.03) 1.000 (0.14) 1.917 (1.04) 0.994 (0.14) 1.918 (1.04) Chronic numbers 1.166* (0.07) 0.984 (0.03) 1.180* (0.07) 0.974 (0.03) 1.178* (0.07) ADLs difficulties 1.265 (0.37) 0.958 (0.16) 1.256 (0.37) 0.960 (0.16) 1.307 (0.39) IADLs difficulties (ref.= 0 activity) 1-3 activities 1.618 (0.40) 0.958 (0.10) 1.607** (0.40) 0.961 (0.11) 1.606* (0.40) 4 or above activities 3.203*** (1.09) 1.376* (0.26) 3.112*** (1.07) 1.403* (0.27) 2.986* (1.02) Enabling Poverty status (ref.=non-poor) Poor 2.316* (0.62) 0.845 (0.09) 1.938 (0.54) 0.902 (0.10) Near poor 2.136 (1.22) 1.259 (0.29) 1.674 (0.97) 1.327 (0.31) Rural residence 0.805 (0.19) 2.869* (0.10) 0.739 (0.18) 2.782*** (0.12) Health insurance (ref.= NRCMS only) UEBMI only 0.342* (0.23) 1.217 (0.21) 0.352* (0.23) URBMI only 0.741 (0.48) 2.032 (0.44) 0.892 (0.57) Multiple or other 0.536 (0.33) 1.028 (0.20) 0.638 (0.40) No insurance 1.810 (0.70) 0.589 (0.16) 2.413** (0.99) Poorrural 1.501 (0.39) Poorurban NRCMS 5.600* (3.43) Constant 0.013* (0.01) 0.013* (0.01) 0.646 (0.32) 0.015* (0.01) LR chi2 125.55 149.96 149.56 Log -1796.921 -1784.714 -1784.966 Pseudo R2 0.034 0.040 0.040 < 0.1; < 0.05; < 0.01. 16 Vol. 8, No. 6; 202 17 care needs for Chinese elders (Inpatient) Model 1 Model 2 Model 3 0.565* (0.11) 0.570* (0.11) 0.575* (0.11) 0.313* (0.14) 0.320* (0.14) 0.328* (0.14) 0.855 (0.15) 0.848 (0.15) 0.848 (0.15) 0.772 (0.51) 0.805 (0.53) 0.794 (0.52) 0.891 (0.60) 0.933 (0.63) 0.940 (0.64) 0.928 (0.18) 0.929 (0.18) 0.934 (0.18) 0.732 (0.21) 0.763 (0.22) 0.760 (0.22) 0.990 (0.04) 0.988 (0.04) 0.989 (0.04) 2.643 (1.08) 2.584 (1.05) 2.577 (1.05) 1.782 (0.73) 1.770 (0.73) 1.783 (0.74) 1.284* (0.06) 1.285* (0.06) 1.284 (0.06) 1.038 (0.25) 1.031 (0.25) 1.034 (0.25) 1.613* (0.31) 1.599* (0.31) 1.594 (0.30) 2.314* (0.65) 2.290* (0.65) 2.256* (0.64) 1.351* (0.25) 1.301 (0.27) 1.093 (0.50) 1.053 (0.49) 0.949 (0.31) 1.005 (0.33) 1.095 (0.45) 1.213 (0.51) 0.407* (0.21) 0.464 (0.25) 1.103 (0.42) 1.339 (0.53) 1.440* (0.29) 1.908* (1.23) 0.022* (0.02) 0.023* (0.02) 0.022* (0.02) Vol. 8, No. 4. Results 6; 2021 care needs for Chinese elders (Inpatient) Model 1 Model 2 Model 3 0.565* (0.11) 0.570* (0.11) 0.575* (0.11) 0.313* (0.14) 0.320* (0.14) 0.328* (0.14) 0.855 (0.15) 0.848 (0.15) 0.848 (0.15) 0.772 (0.51) 0.805 (0.53) 0.794 (0.52) 0.891 (0.60) 0.933 (0.63) 0.940 (0.64) 0.928 (0.18) 0.929 (0.18) 0.934 (0.18) 0.732 (0.21) 0.763 (0.22) 0.760 (0.22) 0.990 (0.04) 0.988 (0.04) 0.989 (0.04) 2.643 (1.08) 2.584 (1.05) 2.577 (1.05) 1.782 (0.73) 1.770 (0.73) 1.783 (0.74) 1.284* (0.06) 1.285* (0.06) 1.284 (0.06) 1.038 (0.25) 1.031 (0.25) 1.034 (0.25) 1.613* (0.31) 1.599* (0.31) 1.594 (0.30) 2.314* (0.65) 2.290* (0.65) 2.256* (0.64) 1.351* (0.25) 1.301 (0.27) 1.093 (0.50) 1.053 (0.49) 0.949 (0.31) 1.005 (0.33) 1.095 (0.45) 1.213 (0.51) 0.407* (0.21) 0.464 (0.25) 1.103 (0.42) 1.339 (0.53) 1.440* (0.29) 1.908* (1.23) 0.022* (0.02) 0.023* (0.02) 0.022* (0.02) Vol. 8, No. 6; 2021 Vol. 8, No. 6; 2021 Applied Economics and Finance Table 3. Multinomial logistic model of unmet healthcare needs for Chinese elders (Inpatient) Model 1 Model 2 Model 3 Predisposing Age groups (ref.=60-69) 70-79 0.565* (0.11) 0.570* (0.11) 0.575* (0.11) 80 and over 0.313* (0.14) 0.320* (0.14) 0.328* (0.14) Female (%) 0.855 (0.15) 0.848 (0.15) 0.848 (0.15) Marital status (ref.=single) Married/partnered 0.772 (0.51) 0.805 (0.53) 0.794 (0.52) Divorced /separated/widowed 0.891 (0.60) 0.933 (0.63) 0.940 (0.64) Education (ref.= illiterate) Literate /primary school 0.928 (0.18) 0.929 (0.18) 0.934 (0.18) Junior high and above 0.732 (0.21) 0.763 (0.22) 0.760 (0.22) Family size 0.990 (0.04) 0.988 (0.04) 0.989 (0.04) Health need Self-rated health (ref.=good) Self-rated health (ref.=good) Poor 2.643 (1.08) 2.584 (1.05) 2.577 (1.05) Fair 1.782 (0.73) 1.770 (0.73) 1.783 (0.74) Chronic numbers 1.284* (0.06) 1.285* (0.06) 1.284 (0.06) ADLs difficulties 1.038 (0.25) 1.031 (0.25) 1.034 (0.25) IADLs difficulties (ref.= 0 activity) 1-3 activities 1.613* (0.31) 1.599* (0.31) 1.594 (0.30) 4 or above activities 2.314* (0.65) 2.290* (0.65) 2.256* (0.64) Enabling Poverty status (ref.=non-poor) Poor 1.351* (0.25) 1.301 (0.27) Near poor 1.093 (0.50) 1.053 (0.49) Health insurance (ref.= NRCMS only) UEBMI only 0.949 (0.31) 1.005 (0.33) URBMI only 1.095 (0.45) 1.213 (0.51) Multiple or other 0.407* (0.21) 0.464 (0.25) No insurance 1.103 (0.42) 1.339 (0.53) Poorrural 1.440 (0.29) 1 908* Table 3. 5. Discussion The elders with unmet needs usually had worse health status, suffered with a higher number of chronic diseases, and experienced more limitations in IADL. Their healthcare needs were more urgent than the average elderly individual’s, but their access to healthcare was not sufficient. Other than ―need factors,‖ we observed socioeconomic and institutional predictors of unmet healthcare needs. If using healthcare services was mainly associated with other factors, rather than need for care, inequality of access to health emerged. With reference to the health inequality literature, our study examined relationships between socioeconomic disadvantage and unmet healthcare needs among elders in China by controlling for predisposing and health need factors. In detail, we investigated to what extent elders’ low-income status and residence status influence their unmet healthcare needs, and how this association varied under China’s multiple health insurance schemes. To the best of our knowledge, this is the first study focusing on poverty and residence status to explore unequal influence of insurance schemes on unmet needs for healthcare. Our findings demonstrated a pattern of association between socioeconomic disadvantages and unmet healthcare needs. Poor elders generally had a high likelihood of experiencing unmet healthcare needs due to financial constraints, even after accounting for poor health status that was more prevalent in populations with socioeconomic disadvantages, indicating that economic barriers to healthcare access still existed. Poverty could also be related to unmet needs as it may reflect residence areas that make accessing healthcare services more difficult. In our study population, 77% of poor elders lived in the vast, remote rural areas. Although a primary goal of China's New Healthcare Reform was developing rural healthcare and medical systems, and narrowing the urban-rural gap, healthcare facilities in rural China are still generally equipped with a less qualified workforce, and they provided less comprehensive services compared to secondary or tertiary hospitals (Xi Li et al., 2017), which are mostly concentrated in urban areas (Wang, Yang, Duan, & Pan, 2018). The unbalanced distribution of health resources between rural and urban areas has widened the gap in healthcare availability, with poor rural elders more likely to suffer an increased risk of unmet needs due to non-financial constraints. Therefore, implementing equal and efficient health insurance should focus on benefiting the poor, especially the poor in rural areas to avoid the regressive benefit on the urban rich (Pan, Xu, & Meng, 2016). 4. Results Multinomial logistic model o Predisposing Age groups (ref.=60-69) 70-79 80 and over Female (%) Marital status (ref.=single) Married/partnered Divorced /separated/widowed Education (ref.= illiterate) Literate /primary school Junior high and above Family size Health need Self-rated health (ref.=good) Self-rated health (ref.=good) Poor Fair Chronic numbers ADLs difficulties IADLs difficulties (ref.= 0 activity) 1-3 activities 4 or above activities Enabling Poverty status (ref.=non-poor) Poor Near poor Health insurance (ref.= NRCMS only) UEBMI only URBMI only Multiple or other No insurance Poorrural Poorurban* NRCMS Constant 17 Vol. 8, No. 6; 2021 Applied Economics and Finance LR chi2 182.24 193.04 198.68 Log likelihood -1131.189 -1125.790 -1122.970 Pseudo R2 0.075 0.079 0.081 < 0.1; < 0.05; *< 0.01. 182.24 193.04 198.68 -1131.189 -1125.790 -1122.970 0.075 0.079 0.081 LR chi2 Log likelihood Pseudo R2 < 0.1; < 0.05; *< 0.01. 5. Discussion 6; 2021 redistributing healthcare resources towards the rural areas, and integrating the urban and rural insurance schemes and making a unified insurance system managed by one government agency. Future research will focus on how these efforts influence the association between socioeconomic disadvantage, unmet healthcare need, and health insurance. And more waves of CHARLS will provide longitudinal data and a time-adjusted understanding for further study as well. redistributing healthcare resources towards the rural areas, and integrating the urban and rural insurance schemes and making a unified insurance system managed by one government agency. Future research will focus on how these efforts influence the association between socioeconomic disadvantage, unmet healthcare need, and health insurance. And more waves of CHARLS will provide longitudinal data and a time-adjusted understanding for further study as well. 5. Discussion Our findings, however, indicate that although health insurance has reduced income-related barriers and offered protection for vulnerable older adults, its multiple schemes have created unequal results. In contrast to urban-based insurance schemes, NRCMS did not exhibit statistically significant effects on improving unmet healthcare needs among the Chinese elderly. This could be due to the nature of NRCMS coverage: the deductibles of NRCMS were generally high; the reimbursement of NRCMS was typically low; and NRCMS appeared to make insured patients more likely to use lower-level providers. Altogether, the insurance scheme designed for rural hukou holders is relatively meagre when compared with the schemes for urban employees or urban hukou holders. This study also examined the effects of NRCMS on the unmet needs of poor elders migrating from rural to urban destinations. We found that these migrants hardly benefited from health insurance even if they were technically covered by one of the schemes. Although residing in urban areas, the majority of migrants (92%) with hukou registered in their home counties were on NRCMS. Since the public funding for insurance schemes was managed by local governments to serve local residents (Zhu & Österle, 2017), rural migrants could have healthcare in their places of residence, but under NRCMS, they were generally not eligible for enrollment in urban destinations. Only those with local insurance account, which was mainly driven by having more than two or other insurances, resulted in a higher likelihood of using health services in urban destinations (Zhang et al., 2017); otherwise they were limited to healthcare facilities in rural regions (Y. Li et al., 2016). Additionally, low-income migrants were not able to satisfy their needs by means of private services. Therefore, they had a higher possibility of forgoing or postponing healthcare utilization. These disadvantages among rural migrants highlight institutional barriers in the current health insurance system. Such institutional discrimination has exacerbated the adverse effects of socioeconomic disadvantages on unmet healthcare needs. Although insurance schemes in China have achieved almost universal coverage for elders in both rural and urban areas, the poor, rural residents, and migrants are still marginalized by the system. To eliminate these socioeconomic disparities in access to health services, the government should put more effort into reforming the current healthcare system by 18 Applied Economics and Finance Vol. 8, No. 6. Conclusion Previous studies on health service utilization among elders had largely focused on factors associated with realized healthcare utilization, i.e. why patients choose to visit doctors and how much they spend. This study is the first to focus on elders’ lack of healthcare utilization. Our findings suggest that those at the lower end of the socioeconomic distribution can be identified as a vulnerable subgroup in China, despite the country’s near-universal health insurance system. Their lower socioeconomic standing will not only contribute to their greater healthcare needs but also contribute to their not having these needs met. Adding health insurance coverage to the socio-demographic variables can moderate the predictive power of socioeconomic disadvantage. However, the effects of insurance vary with the schemes. Integrating the relatively meagre NRCMS scheme with poor socioeconomic status, our findings showed that poor rural migrating elders are the most vulnerable subgroup, experiencing the greatest unmet healthcare need. These findings provide needed evidence on the vulnerability of this subgroup, which is proven to be influenced by both socio-demographic and institutional arrangements. From a public policy perspective, these findings direct attention to the importance of social stratification and the differences in resources available to elders of different status. unding Funding This work was supported by the Natural Science Foundation of China (NSFC) under Grant 71603276. This work was supported by the Natural Science Foundation of China (NSFC) under Grant 71603276. References Center for Health Statistics and Information, M. o. H. o. t. P. s. R. o. C. (2014). An analysis report of China National Health Services Survey in 2013. Beijing. Cheng, T. M. (2012). Early results of China's historic health reforms: the view from minister Chen Zhu. Health Affairs, 31(11), 2536-2544. https://doi.org/10.1377/hlthaff.2012.1017 Fe, E., Powell-Jackson, T., & Yip, W. (2016). Doctor competence and the demand for healthcare: evidence from rural China. Health Economics, 26, 1177-1190. https://doi.org/10.1002/hec.3387 Huang, F., & Gan, L. (2015). The impacts of China's urban employee basic medical insurance on healthcare expenditures and health outcomes. Health Economics, 26(2), 149-163. https://doi.org/10.1002/hec.3281 Idler, E. L., & Benyamini, Y. (1997). Self-Rated Health and Mortality: A Review of Twenty-Seven Community Studies. Journal of Health & Social Behavior, 38(1), 21. https://doi.org/10.2307/2955359 Jin, Y., Hou, Z., Zhang, D., & Zhang, H. (2016). Determinants of health insurance coverage among people aged 45 and over in China: who buys public, private and multiple insurance. Plos One, 11(8), e0161774. https://doi.org/10.1371/journal.pone.0161774 Kim, J., Kim, T. H., Park, E. C., & Cho, W. H. (2015). Factors Influencing Unmet Need for Health Care Services in Korea. Asia Pacific Journal of Public Health, 27(2), 2555-2569. https://doi.org/10.1177/1010539513490789 Kim, S., & Huh, S. (2011). Financial burden of health care expenditures and unmet needs by socioeconomic status. The Korean Journal of Health Eco-nomics and Policy, 17(1), 47-70. Lei, X., & Lin, W. (2009). The New Cooperative Medical Scheme in rural China: does more coverage mean more service and better health? Health Economics, 18(S2), S25-S46. https://doi.org/10.1002/hec.1501 Li, X., Lu, J., Hu, S., Cheng, K. K., & Hu, S. (2017). The primary health-care system in China. Lancet, 390(10112), 2584-2594. https://doi.org/10.1016/S0140-6736(17)33109-4 Li, X., & Zhang, W. (2013). The impacts of health insurance on health care utilization among the older people in China. Social Science & Medicine, 85, 59-65. https://doi.org/10.1016/j.socscimed.2013.02.037 Li, Y., Sun, Y., Zhang, Y., Yi, D., Ma, C., & Ma, S. (2016). Rural–urban disparity in health care: observations from Suzhou, China. Public Health, 138, 164-167. https://doi.org/10.1016/j.puhe.2016.03.026 Liu, H., & Zhao, Z. (2014). Does health insurance matter? Evidence from China’s urban resident basic medical insurance. Journal of Comparative Economics, 42(4), 1007-1020. https://doi.org/10.1016/j.jce.2014.02.003 Liu, Q., Wang, B., Kong, Y., & Cheng, K. (2011). China's primary health-care reform. Lancet, 377(9783), 2064-2066. https://doi.org/ 10.1016/S0140-6736(11)60167-0 Pan, X. F., Xu, J., & Meng, Q. (2016). Integrating social health insurance systems in China. Lancet, 387(10025), 19 Vol. 8, No. 1274-1275. https://doi.org/ 10.1016/S0140-6736(16)30021-6 1274-1275. https://doi.org/ 10.1016/S0140-6736(16)30021-6 Park, S., Kim, B., & Kim, S. (2016). Poverty and working status in changes of unmet health care need in old age. Health Policy, S0168851016300495. https://doi.org/10.1016/j.healthpol.2016.03.004 Pei, X., & Xu, Q. (2011). Old age security, inequality, and poverty. In S. Chen, J. Powell & J. P. S. Chen (Eds.), Aging in perspective and the case of China (pp. 133-149). New York: Nova science publishers, Inc. Ratigan, K. (2015). Too little, but not too late? Health reform in rural China and the limits of experimentalism. Journal of Asian Public Policy, 8(1), 1-19. https://doi.org/ 10.1080/17516234.2015.1009407 Wagstaff, A., Lindelow, M., Gao, J., Xu, L., & Qian, J. (2009). Extending health insurance to the rural population : an impact evaluation of China's New Cooperative Medical Scheme. Journal of Health Economics, 28(1), 1-19. https://doi.org/10.1016/j.jhealeco.2008.10.007 Wang, X., Yang, H., Duan, Z., & Pan, J. (2018). Spatial accessibility of primary health care in China: A case study in Sichuan Province. Social Science & Medicine, 209, 14-24. https://doi.org/10.1016/j.socscimed.2018.05.023 World Health Organization. (2015). China country assessment report on ageing and health: WHO. Yang, M. (2018). Demand for social health insurance: Evidence from the Chinese New Rural Cooperative Medical Scheme. China Economic Review, 56, 126-135. https://doi.org/10.1016/j.chieco.2018.06.004 Ye, L., Shia, B. C., Fang, Y., & Lee, T. S. (2019). Heterogeneous health profiles and healthcare utilization of the middle-aged and elderly with multiple health insurance schemes in China. Public Health, 170, 61-69. https://doi.org/10.1016/j.puhe.2019.01.011 Yip, W. C., Hsiao, W. C., Chen, W., Hu, S., Ma, J., & Maynard, A. (2012). Early appraisal of China's huge and complex health-care reforms. Lancet, 379(9818), 833-842. https://doi.org/10.1016/S0140-6736(11)61880-1 Zhang, A., Nikoloski, Z., & Mossialos, E. (2017). Does health insurance reduce out-of-pocket expenditure? Heterogeneity among China's middle-aged and elderly. Social Science & Medicine, 190, 11-19. https://doi.org/ 10.1016/j.socscimed.2017.08.005 Zheng, D., Gong, J., & Zhang, C. (2019). Efficiency of medical service systems in the rural areas of Mainland China: a comparative study from 2013 to 2017. Public Health, 171, 139-147. https://doi.org/ 10.1016/j.puhe.2019.04.003 Zhu, Y., & Österle, A. (2017). Rural-urban disparities in unmet long-term care needs in China: The role of the hukou status. Social Science & Medicine, 191, 30-37. https://doi.org/ 10.1016/j.socscimed.2017.08.025 References 6; 2021 Applied Economics and Finance 1274-1275. https://doi.org/ 10.1016/S0140-6736(16)30021-6 Copyrights y g yright for this article is retained by the author(s), with first publication rights granted to the journal. Copyright for this article is retained by the author(s), with first publication rights granted to the journal. This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 20
https://openalex.org/W4304782088	https://link.springer.com/content/pdf/10.1007/978-3-031-18326-3_10.pdf	English	null	Classification and Detection of Malicious Attacks in Industrial IoT Devices via Machine Learning	Lecture notes in mechanical engineering	2,022	cc-by	4,170	Mohammad Shahin(B), F Chen, Hamed Bouzary, Ali Hosseinzadeh, and Rasoul Rashidifar Mohammad Shahin(B), F Chen, Hamed Bouzary, Ali Hosseinzadeh, and Rasoul Rashidifar The University of Texas at San Antonio, San Antonio, TX 78249, USA mshahin86@ymail.com, FF.Chen@utsa.edu Abstract. The term “the Industrial Internet of Things” has become increasingly more pervasive in the context of manufacturing as digitization has become a busi- ness priority for many manufacturers. IIoT refers to a network of interconnected industrial devices, resulting in systems that can monitor, collect, exchange, ana- lyze, and deliver valuable data and new insights. These insights can then help drive smarter, and faster business decisions for manufacturers. However, these beneﬁts have come at the cost of creating a new attack vector for the malicious agents that aim at stealing manufacturing trade secrets, blueprints, or designs. As a result, cybersecurity concerns have become more relevant across the ﬁeld of man- ufacturing. One of the main tracks of research in this ﬁeld deals with developing effective cyber-security mechanisms and frameworks that can identify, classify, and detect malicious attacks in industrial IoT devices. In this paper, we have developed and implemented a classiﬁcation and detection framework for address- ing cyber-security concerns in industrial IoT which takes advantage of various machine learning algorithms. The results prove the satisfactory performance and robustness of the approach in classifying and detecting the attacks. Keywords: Malicious attacks · Industrial IoT · Machine learning · Classiﬁcation and detection Keywords: Malicious attacks · Industrial IoT · Machine learning · Classiﬁcation and detection © The Author(s) 2023 K.-Y. Kim et al. (Eds.): FAIM 2022, LNME, pp. 99–106, 2023. https://doi.org/10.1007/978-3-031-18326-3_10 1 Introduction Cyber-Physical Systems (CPS) are deﬁned as systems in which a tight integration between the real-world and cyberspace exists [1]. Cyberspace is the virtual medium responsible for facilitating interconnections between users through telecommunications and computers to store, modify, or exchange data [2]. Once a CPS device is connected to the internet, it is referred to as the Internet of Things (IoT) [3]. IoT allows the inter- action and cooperation of inter-networked physical objects to collect and exchange data over the Internet [4]. Advancements in IoT devices are urging traditional manufacturing systems to be integrated into cyberspace to take advantage of this emerging interaction and cooperation [5]. These systems are then can be replaced by a geographically dis- persed network of services that are connected to the shop ﬂoor through the power of IoT. This spread or decentralization in manufacturing systems can help with providing © The Author(s) 2023 K.-Y. Kim et al. (Eds.): FAIM 2022, LNME, pp. 99–106, 2023. https://doi.org/10.1007/978-3-031-18326-3_10 M. Shahin et al. 100 more ﬂexibility, agility, and adaptivity through a faster responsivity in processing shop ﬂoor data and thus can effectively overcome the challenges corresponding with tradi- tional manufacturing systems. However, this higher connectivity can come at the cost of an increase in the number of cyber-attacks [6–8]. These attacks showed that given enough resources, all systems can be breached, with manufacturing systems being no exception with one in every three cyber-physical attacks happening in the manufactur- ing sector according to the Industrial Control Systems Monito Newsletter issued by the U.S. Department of Homeland Security [9, 10]. The rapid occurrence of such attacks on manufacturing and business operations and their information systems and the resulting damages and costs associated with them have urged scholars to consider new ways of detecting such attacks [11]. As the continuation of such efforts, we intend to show how appropriate machine learning approaches can be utilized to enhance the deterrence level of malicious attacks in industrial IoT devices in manufacturing. To this end, we have implemented a set of preprocessing and data analytics techniques on a new dataset in which various cyber-security attacks have been successfully detected via classiﬁcation algorithms. 2 Background Machine learning methods have been applied in many aspects of today’s manufacturing enterprises. Many scholars are now focusing on the use of these techniques to improve cybersecurity by monitoring and conducting surveillance of real-time network streams and real-time detection of threat patterns [12]. These methods can learn from historical data and train a model to correlate events, identify patterns, and detect anomalous behav- ior. Apart from the algorithm implementation and development, various efforts have been put forward by researchers in this ﬁeld to simulate breach scenarios and record the sub- sequent data. These studies have resulted in a variety of data sets existing in the ﬁeld within each different pre-processing technique have been coupled. As a result, a detailed literature review is needed to summarize the state-of-the-art of the ﬁeld and identify the potential areas of improvement. The following paragraphs summarize the most notable research works done in this ﬁeld to date. Terzi, Terzi & Sagiroglu [13] have used an unsupervised anomaly detection approach and Principal Component Analysis (PCA) to identify anomalies in public big network data to understand network behavior to distinguish cyber-attacks and to provide better detection in the future. Autoencoder has been used with dimension reduction to detect cyber-attack anomalies [14]. In another study, Wan et al. [15] showed that using Wavelet NeuralNetwork(WNN)todetectanomaliesinindustrialcontrolcommunicationsystems canleadtobetteraccuracycomparedtousingBackPropagationNeuralNetwork(BPNN) in addition to being more adequate in real-time analysis. The denial of service category (DoS) in KDD CUP 1999 (KDD) and CSE-CIC- IDS2018 data sets have been used by Kim et al. [15] to develop Convolutional Neural Network (CNN) models to detect DoS intrusion attacks resulting in a high accuracy detection that ranged between 89%–99%. Wang et al. [16], McLaughlin et al.[17], and Gibert [18] have also used a CNN approach to detect malware. The latter evaluated their technique using the MalImg dataset and the Microsoft Malware Classiﬁcation Challenge Classiﬁcation and Detection of Malicious Attacks 101 dataset and managed to outperform other methods in terms of accuracy and classiﬁcation time. Deep Neural Network (DNN) has been deployed to detect malware [19] on large scales data sets such as the Internal Microsoft dataset with over 2.6 million labeled samples with results for a two-class error rate of 0.49% for a single neural network and 0.42% for an ensemble of neural networks [20]. Xu et al. [21] combined DNN with Multiple Kernel Learning (MKL) to detect malware in applications run by users of Android devices. 2 Background Aside from the aforementioned studies, there exist other studies that attempt to address the problem from aspects other than algorithm development. For instance, Elhabashy et al. [9] have proposed an attack taxonomy to better understand the relationships between quality control systems, manufacturing systems, and cyber- physical attacks. In another study, Wu et al. [22] have utilized anomaly detection and Random Forest algorithm to detect 3D printing and CNC milling machine malicious attacks. 3 Dataset and Methodology In this paper, we used a dataset called “N-BaIoT” that was initially generated by Meidan et al. from network trafﬁc patterns [23]. The initial data was gathered from nine com- mercial IoT devices infected by two different botnets. They have deployed two of the most common IoT botnet families namely, Gafgyt and Mirai, and collected trafﬁc data before and after the infection. Gafgyt (also known as BASHLITE, Q-Bot, Torlus, Lizard- Stresser, and Lizkebab) is one of the most infamous types of IoT botnets. To launch an attack, the botnet infects Linux-based IoT devices by brute-forcing default credentials of devices with open Telnet ports. Mirai is the second botnet that has been deployed in this isolated network. The experimental setup included a C&C server and a server with a scanner and loader. The scanner and loader components were responsible for scan- ning and identifying vulnerable IoT devices, and loading the malware to the vulnerable IoT devices detected. Once a device was infected, it automatically started scanning the network for new victims while waiting for instructions from the C&C server [23]. In our analysis, we only use seven of the devices out of the nine that exist in this data set. We have implemented and chosen the most effective classiﬁers for this speciﬁc data set which turned out to be KNN, DT, and RF. A brief description of these algorithms is described below: 1. K-Nearest Neighbors (KNN): KNN is a supervised machine learning algorithm that can be used to solve both classiﬁcation and regression problems. KNN assumes that similar data points exist nearby. In other words, similar data points are near to each other. KNN searches the entire data set for the k number of most neighbors and calculates distances for proximities before sorting the calculated distances in ascending order from smallest to largest and picking the ﬁrst K with its feature that is associated with the smallest distance. KNN uses a large amount of training data, where data points are plotted in a high-dimensional space, where each axis in the space corresponds to an individual variable that characterizes that data point [24]. KNN has been used in intelligent mechanical systems to detect online fraud [25] and has been successfully implemented in a large number of business problems [26, 27]. 1. K-Nearest Neighbors (KNN): KNN is a supervised machine learning algorithm that can be used to solve both classiﬁcation and regression problems. 3 Dataset and Methodology Other minor differences exist between these four main DT algorithms such as, how to deal with missing value, variable selection, capacity to handle a huge number of classes in variables, and pruning methods [28–30]. DT has been used in phishing detection [31] and Adversarial detection [32]. p g 3. Random Forest (RF): RF is a type of ensemble learning method that have been widely used in many ﬁelds, such as computer vision and data mining. MRF performs very well with a large data set in a short time compared with other techniques. MRF is easy to interpret and understand, can handle both numerical and categorical data. MRF consists of a large number of individual decision trees that operate as a group producing a single effect (ensemble). Each decision tree is built by randomly selecting observations and speciﬁc features and averaging the results at the end. Thus, allowing it to limit overﬁtting without a substantial increase in the generalization error [33, 34]. RF has been used to detect ransomware and achieved a high accuracy level of 97.74% in detecting ransomware [35]. At the same time, RF was used as a feature selection tool when building an Auto-Encoder Intrusion Detection System (AE-IDS). The results showed that using RF helped in reducing the detection time and effectively improved the prediction accuracy [36]. 3. Random Forest (RF): RF is a type of ensemble learning method that have been widely used in many ﬁelds, such as computer vision and data mining. MRF performs very well with a large data set in a short time compared with other techniques. MRF is easy to interpret and understand, can handle both numerical and categorical data. MRF consists of a large number of individual decision trees that operate as a group producing a single effect (ensemble). Each decision tree is built by randomly selecting observations and speciﬁc features and averaging the results at the end. Thus, allowing it to limit overﬁtting without a substantial increase in the generalization error [33, 34]. RF has been used to detect ransomware and achieved a high accuracy level of 97.74% in detecting ransomware [35]. At the same time, RF was used as a feature selection tool when building an Auto-Encoder Intrusion Detection System (AE-IDS). The results showed that using RF helped in reducing the detection time and effectively improved the prediction accuracy [36]. 3 Dataset and Methodology KNN assumes that similar data points exist nearby. In other words, similar data points are near to each other. KNN searches the entire data set for the k number of most neighbors and calculates distances for proximities before sorting the calculated distances in ascending order from smallest to largest and picking the ﬁrst K with its feature that is associated with the smallest distance. KNN uses a large amount of training data, where data points are plotted in a high-dimensional space, where each axis in the space corresponds to an individual variable that characterizes that data point [24]. KNN has been used in intelligent mechanical systems to detect online fraud [25] and has been successfully implemented in a large number of business problems [26, 27]. 102 M. Shahin et al. 2. Decision Tree (DT): DT is a set of rules for dividing a large heterogeneous popula- tion into smaller, more homogeneous groups concerning a particular output feature. DT is one of the most common Data Mining (DM) techniques that is widely being used for both classiﬁcation and regression analysis. DT comes in many types of decision algorithms, some of which are binary trees that always produce two cat- egories (binary-split) at any level of the tree-like CART and QUEST. Others like CHAID and C5.0 are non-binary trees that often produce more than two categories at any level in the tree. Other minor differences exist between these four main DT algorithms such as, how to deal with missing value, variable selection, capacity to handle a huge number of classes in variables, and pruning methods [28–30]. DT has been used in phishing detection [31] and Adversarial detection [32]. 2. Decision Tree (DT): DT is a set of rules for dividing a large heterogeneous popula- tion into smaller, more homogeneous groups concerning a particular output feature. DT is one of the most common Data Mining (DM) techniques that is widely being used for both classiﬁcation and regression analysis. DT comes in many types of decision algorithms, some of which are binary trees that always produce two cat- egories (binary-split) at any level of the tree-like CART and QUEST. Others like CHAID and C5.0 are non-binary trees that often produce more than two categories at any level in the tree. 4 Results and Discussion A 90/10 split has been used to form the training and test data sets considering the large scale of the data set. Also, in all of the experiments, a 5-fold cross-validation has been used for model validation. The accuracy results for each of these classiﬁers can be found in Fig. 1. As one can see from Fig. 1, the algorithms have been implemented on three different IoT devices (Ecobee Thermostat, Philips B120N10 Baby Monitor, and Provision PT737E Security Camera) compromised by two different bots (Mirai, and Gafgyt). The results indicate that the determining factor in the ﬁnal accuracy of attack classiﬁcation is the type of bot rather than the device type. In other words, the accuracy results show a similar pattern among three different devices compromised by a similar bot. According to the results, for devices attacked by Mirai bot, RF algorithm delivers the highest accuracy followed by the DT, and KNN. In particular, the accuracy achieved by the KNN algorithm dealing with the Thermostat compromised by the Mirai bot is the lowest among any other scenarios as this algorithm is only capable of accurately classifying the data in 0.755426 of the test data instances. This translates to a signiﬁcant number of misclassiﬁcation instances (12846 out of the 52525 instances in the test dataset) which underlines the poor performance of this algorithm in this speciﬁc scenario. On the other hand, for the Gafgyt bot, RF outperforms the other two algorithms while DT performs worst among them. As opposed to the left-hand side scenarios corresponding Classiﬁcation and Detection of Malicious Attacks 103 with the Mirai bot, even the worst-performing algorithm dealing with the Gafgyt bot (DT) is capable of accurately classifying the attacks in more than 0.99 of the test data instances. It is important to note that even though the accuracy values for different algorithms look reasonably close, they translate to a signiﬁcantly different number of misclassiﬁca- tions due to the large size of the dataset. This can be very critical in real-world scenarios as even a single cyber-security breach can result in a signiﬁcant amount of loss from security and/ or economic points of view. The corresponding misclassiﬁcation values can be found in Table 1. Fig. 1. Accuracy results for three algorithms detecting six different device and bot type combinations. Fig. 1. Accuracy results for three algorithms detecting six different device and bot type combinations. Table 1. Misclassiﬁcation results. Table 1. Misclassiﬁcation results. Device/Attack type KNN DT RF EcobeeThermostat_Mirai 12846/52525 5/52525 0 EcobeeThermostat_Gafgyt 90/32374 109/32374 36/32374 BabyMonitor_Mirai 474/78595 4/78595 0 BabyMonitor_ Gafgyt 118/43786 321/43786 105/43786 SecurityCamera_Mirai 230/49816 8/49816 1/49816 SecurityCamera_ Gafgyt 106/39225 579/39225 221/39225 104 M. Shahin et al. 5 Conclusion We proposed a machine learning-based framework for attack classiﬁcation and detection in IIoT devices. The experiments have shown the successful adoption of artiﬁcial intelli- gence to cybersecurity, which has led to an effective and robust approach for identifying, classifying, and detecting two different types of botnet attacks compromising three dif- ferent IIoT devices. The evaluation process has employed accuracy as a performance metric to show the effectiveness of this approach. The experiments have demonstrated that a combination of various machine learning algorithms is capable of accurately detecting and classifying the attacks in more than 99.9% of the instances in the test data set employed. Future endeavors can focus on enhancing our approach by developing deep neural network-based models and also taking advantage of other emerging IIoT data sets. Future work can also attempt to develop more effective feature engineering methods that can transform the raw network data into richer input sources for building learning methods. References https://doi.org/10.1109/ EMR.2019.2927559 Classiﬁcation and Detection of Malicious Attacks 105 12. Mahmood, T., Afzal, U.: Security analytics: big data analytics for cybersecurity: a review of trends, techniques and tools. In: 2013 2nd National Conference on Information Assurance (NCIA), pp. 129–134 (2013). https://doi.org/10.1109/NCIA.2013.6725337 12. Mahmood, T., Afzal, U.: Security analytics: big data analytics for cybersecurity: a review of trends, techniques and tools. In: 2013 2nd National Conference on Information Assurance (NCIA), pp. 129–134 (2013). https://doi.org/10.1109/NCIA.2013.6725337 12. Mahmood, T., Afzal, U.: Security analytics: big data analytics for cybersecurity: a review of trends, techniques and tools. In: 2013 2nd National Conference on Information Assurance (NCIA), pp. 129–134 (2013). https://doi.org/10.1109/NCIA.2013.6725337 13. Terzi,D.S.,Terzi,R.,Sagiroglu,S.:Bigdataanalyticsfornetworkanomalydetectionfromnet- ﬂow data. In: 2017 International Conference on Computer Science and Engineering (UBMK), pp. 592–597 (2017). https://doi.org/10.1109/UBMK.2017.8093473 14. Gaggero, G.B., Rossi, M., Girdinio, P., Marchese, M.: Neural network architecture to detect system faults/cyberattacks anomalies within a photovoltaic system connected to the grid. In: 2019 International Symposium on Advanced Electrical and Communication Technologies (ISAECT), pp. 1–4 (2019). https://doi.org/10.1109/ISAECT47714.2019.9069683 15. Wan, M., Song, Y., Jing, Y., Wang, J.: Function-aware anomaly detection based on wavelet neural network for industrial control communication. Secur. Commun. Netw. (2018). https:// doi.org/10.1155/2018/5103270 16. Wang, W., Zhu, M., Zeng, X., Ye, X., Sheng, Y.: Malware trafﬁc classiﬁcation using con- volutional neural network for representation learning. In: 2017 International Conference on Information Networking (ICOIN), pp. 712–717. IEEE (2017).https://doi.org/10.1109/ICOIN. 2017.7899588 17. McLaughlin, N., et al.: Deep Android Malware Detection, pp. 301–308 (2017). https://doi. org/10.1145/3029806.3029823 18. Gibert, D., Mateu, C., Planes, J., Vicens, R.: Using convolutional neural networks for classiﬁ- cation of malware represented as images. J. Comput. Virol. Hack. Tech. 15(1), 15–28 (2018). https://doi.org/10.1007/s11416-018-0323-0 19. Grosse, K., Papernot, N., Manoharan, P., Backes, M., McDaniel, P.: Adversarial perturbations against deep neural networks for malware classiﬁcation (2016). arXiv:1606.04435 [cs], http:// arxiv.org/abs/1606.04435. Accessed 18 Jun 2020 20. Dahl, G.E., Stokes, J.W., Deng, L., Yu, D.: Large-scale malware classiﬁcation using ran- dom projections and neural networks. In: 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, pp. 3422–3426, May 2013. https://doi.org/10.1109/ICASSP. 2013.6638293 21. Xu, L., Zhang, D., Jayasena, N., Cavazos, J.: HADM: hybrid analysis for detection of malware. In: Bi, Y., Kapoor, S., Bhatia, R. (eds.) IntelliSys 2016. LNNS, vol. 16, pp. 702–724. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-56991-8_51 22. Wu, M., Song, Z., Moon, Y.B.: Detecting cyber-physical attacks in CyberManufacturing systems with machine learning methods. J. Intell. Manuf. 30(3), 1111–1123 (2017). https:// doi.org/10.1007/s10845-017-1315-5 23. References 1. Chhetri, S.R., Rashid, N., Faezi, S., Al Faruque, M.A.: Security trends and advances in manu- facturing systems in the era of industry 4.0. In: 2017 IEEE/ACM International Conference on Computer-Aided Design (ICCAD), pp. 1039–1046 (2017). https://doi.org/10.1109/ICCAD. 2017.8203896 2. Koppisetty, H., Potdar, K., Jain, S.: Cyber-crime, forensics and use of data mining in cyber space: a survey. In: 2019 International Conference on Smart Systems and Inventive Tech- nology (ICSSIT), Smart Systems and Inventive Technology (ICSSIT), pp. 722–727 (2019). https://doi.org/10.1109/ICSSIT46314.2019.8987921 p g 3. Jazdi, N.:Cyber physical systems in the context of industry 4.0. In: 2014 IEEE International Conference on Automation, Quality and Testing, Robotics, pp. 1–4, May 2014. https://doi. org/10.1109/AQTR.2014.6857843 p g 3. Jazdi, N.:Cyber physical systems in the context of industry 4.0. In: 2014 IEEE International Conference on Automation, Quality and Testing, Robotics, pp. 1–4, May 2014. https://doi. org/10.1109/AQTR.2014.6857843 g Q 4. Atzori, L., Iera, A., Morabito, G.: The Internet of Things: a survey. Comput. Netw. 54(15), 2787–2805 (2010). https://doi.org/10.1016/j.comnet.2010.05.010 g Q 4. Atzori, L., Iera, A., Morabito, G.: The Internet of Things: a survey. Comput. Netw. 54(15), 2787–2805 (2010). https://doi.org/10.1016/j.comnet.2010.05.010 5. Shahin, M., Chen, F.F., Bouzary, H., Krishnaiyer, K.: Integration of lean practices and Industry 4.0 technologies: smart manufacturing for next-generation enterprises. Int. J. Adv. Manufact. Technol. 107(5–6), 2927–2936 (2020). https://doi.org/10.1007/s00170-020-05124-0 6. Rauch, E., Dallasega, P., Matt, D.T.: Distributed manufacturing network models of smart and agile mini-factories. Int. J. Agile Syst. Manage. 10(3–4), 185–205 (2017) 7. Elhabashy, A.E., Wells, L.J., Camelio, J.A.: Cyber-physical security research efforts in man- ufacturing - a literature review. Procedia Manufact. 34, 921–931 (2019). https://doi.org/10. 1016/j.promfg.2019.06.115 8. Shahin, M., Chen, F.F., Bouzary, H., Zarreh, A.: Frameworks proposed to address the threat of cyber-physical attacks to lean 4.0 systems. Procedia Manufact. 51, 1184–1191 (2020). https:// doi.org/10.1016/j.promfg.2020.10.166 9. Elhabashy, A.E., Wells, L.J., Camelio, J.A., Woodall, W.H.: A cyber-physical attack taxonomy for production systems: a quality control perspective. J. Intell. Manuf. 30(6), 2489–2504 (2018). https://doi.org/10.1007/s10845-018-1408-9 p g 10. ICS Monitor Newsletters \| CISA. https://www.us-cert.gov/ics/monitors. Accessed 20 Oct. 20 10. ICS Monitor Newsletters \| CISA. https://www.us-cert.gov/ics/monitors. Accessed 20 Oct. 20 11. Culot, G., Fattori, F., Podrecca, M., Sartor, M.: Addressing industry 4.0 cybersecurity chal- l IEEE E M R 47(3) 79 86 hi d (2019) h //d i /10 1109/ 11. Culot, G., Fattori, F., Podrecca, M., Sartor, M.: Addressing industry 4.0 cybersecurity chal- lenges. IEEE Eng. Manage. Rev. 47(3), 79–86, thirdquarter (2019). References Meidan, Y., et al.: N-BaIoT: network-based detection of IoT botnet attacks using deep autoen- coders. IEEE Pervasive Comput. 17(3), 12–22 (2018). https://doi.org/10.1109/MPRV.2018. 03367731 24. Samui, P., Sekhar, S., Balas, V.E.: Handbook of Neural Computation. Elsevier (2017). https:// doi.org/10.1016/C2016-0-01217-2 25. Kannagi, A., Mohammed, J.G., Murugan, S.S.G., Varsha, M.: Intelligent mechanical systems and its applications on online fraud detection analysis using pattern recognition K-nearest neighbor algorithm for cloud security applications. Mater. Today: Proc. (2021). https://doi. org/10.1016/j.matpr.2021.04.228 26. Greenwell, B.B.B.: Hands-On Machine Learning with R. 2020. https://bradleyboehmke.git hub.io/HOML/knn.html. Accessed 17 Jun 2020 27. Cahyani, D.E., Nuzry, K.A.P.: Trending topic classiﬁcation for single-label using multino- mial naive bayes (MNB) and multi-label using k-nearest neighbors (KNN). In: 2019 4th International Conference on Information Technology, Information Systems and Electrical M. Shahin et al. 106 Engineering (ICITISEE), Information Technology, Information Systems and Electrical Engi- neering (ICITISEE), 2019 4th International Conference on, pp. 547–552 (2019). https://doi. org/10.1109/ICITISEE48480.2019.9003944 Engineering (ICITISEE), Information Technology, Information Systems and Electrical Engi- neering (ICITISEE), 2019 4th International Conference on, pp. 547–552 (2019). https://doi. org/10.1109/ICITISEE48480.2019.9003944 28. Kass, G.V.: An exploratory technique for investigating large quantities of categorical data. J. R. Stat. Soc. Ser. C (Appl. Stat.) 29(2), 119–127 (1980). https://doi.org/10.2307/2986296 29. Salzberg, S.L.: C4.5: programs for machine learning by J. Ross Quinlan. Morgan Kaufmann Publishers Inc, 1993. Mach Learn 16(3), 235–240 (1994). https://doi.org/10.1007/BF0099 3309 30. Loh, W.-Y., Shih, Y.-S., Loh, W.-Y., Shih, Y.-S.: Split selection methods for classiﬁcation trees. 7(4) (1997). http://www3.stat.sinica.edu.tw/statistica/j7n4/j7n41/j7n41.htm . Accessed 17 Jun 2020 31. Zhu, E., Ju, Y., Chen, Z., Liu, F., Fang, X.: DTOF-ANN: an artiﬁcial neural network phishing detection model based on decision tree and optimal features. Appl. Soft Comput. 95, 106505, (2020). https://doi.org/10.1016/j.asoc.2020.106505 32. Appiah, B., Qin, Z., Abra, A.M., Kanpogninge, A.J.A.: Decision tree pairwise metric learning against adversarial attacks. Comput. Secur. 106, 102268, (2021). https://doi.org/10.1016/j. cose.2021.102268 33. Bai, J., Li, Y., Li, J., Yang, X., Jiang, Y., Xia, S.-T.: Multinomial random forest. Pattern Recogn. 122, 108331, (2022). https://doi.org/10.1016/j.patcog.2021.108331 34. Breiman, L.: Random forests. Mach. Learn. 45(1), 5–32 (2001). https://doi.org/10.1023/A: 1010933404324 35. Khammas, B.M.: Ransomware detection using random forest technique. ICT Express 6(4), 325–331 (2020). https://doi.org/10.1016/j.icte.2020.11.001 36. Li, X., Chen, W., Zhang, Q., Wu, L.: Building auto-encoder intrusion detection system based on random forest feature selection. Comput. Secur. 95, 10185110 (2020). https://doi.org/10. References 1016/j.cose.2020.101851 Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, 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. The images or other third party material in this chapter are included in the chapter’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
https://openalex.org/W4285590449	https://www.frontiersin.org/articles/10.3389/fimmu.2022.963444/pdf	English	null	Cardiac Mast Cells: A Two-Head Regulator in Cardiac Homeostasis and Pathogenesis Following Injury	Frontiers in immunology	2,022	cc-by	8,613	REVIEW published: 15 July 2022 doi: 10.3389/fimmu.2022.963444 REVIEW published: 15 July 2022 doi: 10.3389/fimmu.2022.963444 Reviewed by: Reviewed by: Kun Yang, East Tennessee State University, United States Yonggang Ma, University of South Florida, United States Lu Wang, Renmin Hospital of Wuhan University, China Reviewed by: Kun Yang, East Tennessee State University, United States Yonggang Ma, University of South Florida, United States Lu Wang, Renmin Hospital of Wuhan University, China *Correspondence: Zhaoliang Su szl30@ujs.edu.cn †These authors have contributed equally to this work Reviewed by: Kun Yang, East Tennessee State University, United States Yonggang Ma, University of South Florida, United States Lu Wang, Renmin Hospital of Wuhan University, China Keywords: cardiac mast cells, cardiac development, cardiac aging, cardiac injury, inﬂammation INTRODUCTION Specialty section: This article was submitted to Inﬂammation, a section of the journal Frontiers in Immunology Received: 07 June 2022 Accepted: 24 June 2022 Published: 15 July 2022 Mast cells, innate non-circulating immune cells, exist in almost all tissues and play key roles in allergic disease and host defense, including the heart. Mast cells are highly heterogeneous, express a range of receptors on their surface and generate a variety of mediators to involve in extensive inﬂammation and immune regulation through degranulation. Therefore, mast cells are called “sentinels” in harmful conditions, with the ability to rapidly perceive invasion and initiate immune defense and different biochemical programs of homeostasis in time (1). Cardiac mast cells (CMCs) are present in the heart at a low density at homeostasis and is generally detected in the epicardium, endocardium, and myocardium of ventricle and atrium in mice, rats, and humans. Published data show that CMCs density is <1 cell/mm2 in mouse heart (2). Our unpublished data also demonstrate that CMCs account for <3% of CD45+ cells in mouse heart. Therefore, the present review will summarize the novels progress of CMCs on origin, development, replenishment, especially on cardiac development, function and ageing under physiological conditions as well as the roles of CMCs on inﬂammatory progression and resolution. Received: 07 June 2022 Accepted: 24 June 2022 Published: 15 July 2022 Edited by: Guo-Chang Fan, University of Cincinnati, United States Cardiac mast cells (CMCs) are multifarious immune cells with complex roles both in cardiac physiological and pathological conditions, especially in cardiac ﬁbrosis. Little is known about the physiological importance of CMCs in cardiac homeostasis and inﬂammatory process. Therefore, the present review will summarize the recent progress of CMCs on origin, development and replenishment in the heart, including their effects on cardiac development, function and ageing under physiological conditions as well as the roles of CMCs in inﬂammatory progression and resolution. The present review will shed a light on scientists to understand cardioimmunology and to develop immune treatments targeting on CMCs following cardiac injury. ORIGIN, DEVELOPMENT AND SURVIVAL OF CMCS and death. Christine Moller and his colleagues are the ﬁrst to directly demonstrate that both Bcl-x and Bcl-2 are essential for keeping mast cells survival during late development. Nevertheless, the upregulation of Bcl-XL and Bcl-2 by IgE is eliminated for bone marrow derived mast cells in IL-3–/–mice. IL- 3 regulates pro-survival Bcl-2 family members (24, 25) and SCF suppresses pro-apoptotic Bim (23). The survival of mast cells depends on the ratio and interaction of anti-apoptotic and pro- apoptotic factors (26). Additionally, ﬁbroblasts can also promote human mast cells survival (27). For example, it has been proved that mouse skin 3T3 ﬁbroblasts can sustain the primary human lung mast cells for 13 days in the absence of exogenous growth factors (28). Although all the mast cells seemed not to have tight junctions and proliferation, these cells still maintained the general morphology, granule morphology and mediator content (29). Furthermore, the co-culture of gut-derived human mast cells with gut ﬁbroblasts has the similar phenomena, and it is also indicated that human ﬁbroblasts promote survival of human mast cells independent of SCF, IL-3, IL-4, and nerve growth factor (NGF). That’s because ﬁbroblasts can release a soluble heat-sensitive molecule that downregulates apoptosis without promoting cell proliferation (30). In addition, an increase of mast cells is noted in the healing myocardium, and their progenitors are also found in the infarcted area, which is related to the activation and proliferation of ﬁbroblasts following cardiac injury or not? The contribution of ﬁbroblasts proliferation cannot be clariﬁed, while the chemotaxis of circulating precursors to the heart may be the main mechanism leading to the accumulation of mast cells in ischemic heart (31). We at least know that ﬁbroblasts contribute to mast cell survival. Other chemokines and cytokines involving in mast cells growth and survival need to be further conﬁrmed in future (Figure 1) (32). Mast cells have been always thought to originate from hematopoietic stem cells (HSCs) in bone marrow (3). However, recent data indicate that mast cells probably derive from three embryonic hematopoietic waves: early and late erythron-myeloid progenitors (EMPs) from yolk sac, and deﬁnitive HSCs from the aorta, gonads, and mesonephros region (4). Mast cells derived from different hematopoietic waves have different tissue preferences, for example, from the early EMPs distribute in adipose tissue, from late EMPs widely distribute in most connective tissues, and from fetal HSCs are the main cells group in mucosa. Citation: Jin J, Jiang Y, Chakrabarti S and Su Z (2022) Cardiac Mast Cells: A Two-head Regulator in Cardiac Homeostasis and Pathogenesis Following Injury. Front. Immunol. 13:963444. doi: 10.3389/fimmu.2022.963444 July 2022 \| Volume 13 \| Article 963444 Frontiers in Immunology \| www.frontiersin.org CMCs and Cardiac Homeostasis, Pathogenesis Jin et al. ORIGIN, DEVELOPMENT AND SURVIVAL OF CMCS It is also suggested that bone marrow derived mast cells mainly replenish the mucosal mast cells (MMCs) after birth (4–6). Mast cells from the embryonic stage are thought to have reached peripheral tissues and matured into resident mast cells before birth, which possess tissue and function heterogeneity. After birth, mast cells precursors from bone marrow need to be released into the bloodstream and recruited by various mediators before entering the peripheral tissues. It is known that plenty of biologic agents, including growth factors, integrins, chemokines and adenosine nucleotides contribute to this recruitment process (7–10). Different mast cells subsets express different receptors which may contribute to their movement into speciﬁc tissues. For example, the recruitment to intestine requires a4b7 integrin and chemokine receptor CXCR2 expressed on mast cells progenitors (MCps), accompanying with mucosal addressin cellular adhesion molecule-1 and vascular cell adhesion molecule-1 on intestinal endothelium (11, 12). Furthermore, a4b7 and vascular cell adhesion molecule-1 are also required for the recruitment of mast cells precursors to the lungs (13). However, it is unclear that CMCs are derived from early and late EMPs, or maturation from MCp (14, 15). If it is the later, which factor can mediate the speciﬁc homing or recruitment of MCps to heart? Frontiers in Immunology \| www.frontiersin.org MAST CELLS SUBSETS CMCs and Cardiac Homeostasis, Pathogenesis FIGURE 1 \| CMCs origin, development and survival. Most of CMCs in the physiological state of the heart come from embryonic stage, and only a small part come from bone marrow. The increase of CMCs can be differentiated from MCp or through self-proliferation. Under pathological conditions, MCps are recruited from bone marrow. The density can also be increased by the relocation of CMCs in non-injured sites. In addition, some cytokine chemokines, cardiac ﬁbroblast derived growth factors and Bcl-2 family can promote the survival of CMCs. FIGURE 1 \| CMCs origin, development and survival. Most of CMCs in the physiological state of the heart come from embryonic stage, and only a small part come from bone marrow. The increase of CMCs can be differentiated from MCp or through self-proliferation. Under pathological conditions, MCps are recruited from bone marrow. The density can also be increased by the relocation of CMCs in non-injured sites. In addition, some cytokine chemokines, cardiac ﬁbroblast derived growth factors and Bcl-2 family can promote the survival of CMCs. and CTMCs appear to differ in many other aspects of biochemistry, histochemical characteristics, function and roles in inﬂammation and immunity (44–46). Similarly, human mast cells also differ in various aspects of their phenotype, just like morphologic characteristics, histochemistry, contents of proteases and sensitivity to stimulation by secretogogues (47–51). Notably, it has been suggested that the phenotype of mast cells, such as mediator contents or responsive abilities to speciﬁc stimuli, can be regulated, at least in some cases reversibly, by microenvironmental signals such as cytokines and growth factor (52). In fact, many potential variations in microenvironment may affect phenotype. The anatomical location is the ﬁrst factor that affects the phenotype. For example, when cultured mast cells in vitro were transferred into different locations in vivo, which can give the chance to develop into CTMCs or MMCs, depending on local signals (53, 54). Secondly, inﬂammatory or immune processes may also cause transient changes of mast cells phenotype. For instance, the number of CMCs increase in cardiomyopathy compared to normal myocardium, and a second increase occurred after long- term mechanical support, but the phenotype is conversion from MCTCs into MCTs with the decrease of cardiac ﬁbrosis (55). MAST CELLS SUBSETS The classic classiﬁcation of mast cells is based on their tissue distribution and granule content. According to the proteases they contained, mast cells are divided into mainly containing tryptase (MCTs) or chymase (MCCs) or both (MCTCs) (33). In human, almost 90% CMCs are MCTCs (34). MCTs is usually localized to mucosal surfaces and closely related to T cells, especially Th2- type. MCTCs, on the other hand, includes tryptase, chymase, carboxypeptidase and cathepsin G. It mainly exists in the gastrointestinal tract, skin, synovium and subcutaneous tissues. But the proportion and distribution of the two subsets may change in pathological states. For example, the number of MCTCs is increased in ﬁbrotic diseases, but relatively unchanged in allergic or parasitic diseases. Therefore, MCTCs may be biased towards tissue remodeling and angiogenesis, and MCTs contribute to inﬂammation (35). Besides, mouse mast cells also can be divided into two lineages: inducible bone marrow–mucosal mast cells (MMCs) and constitutive embryonic-derived connective tissue mast cells (CTMCs) (36). Phenotypic differences between CTMCs and MMCs are acquired during the local tissue development, rather than determined by In both mouse and human, obtaining the cell surface and intracellular characteristics of fully differentiated mature mast cells requires a gradual process, that can be regulated by different cytokines, in which stem cell factor (SCF) and IL-3 may play a major role. SCF, the ligand of c-kit, not only facilitates cells migration, but also contributes to their development (16, 17). IL- 3 can beneﬁt the multiple hematopoietic lineage differentiation into mast cells in vitro (18, 19). However, IL-3 is not necessary for the generation of mast cells at homeostasis, it does beneﬁt to increase the number (20). Like IL-3, the other cytokines, such as IL-4, IL-9, IL-10 and IL-13 can also synergistically promote mast cell proliferation and differentiation (21). Then, how are CMCs maintained and renewed? Tissue mast cells are known to be long-lived cells and even after degranulation they can re-granulate and continue to survive, which is dependent on the local SCF levels (22). Because SCF can inactivate FOXO3a, a fork-head transcription factor, and down-regulate and phosphorylate its target Bim (a Bcl-2 homology 3-only proapoptotic protein) which promote mast cells survival (23). Bcl-2 family, well-known proteins, are critical for cells survival July 2022 \| Volume 13 \| Article 963444 Frontiers in Immunology \| www.frontiersin.org 2 Jin et al. Jin et al. MAST CELLS SUBSETS Furthermore, similar switch also exists in other speciﬁc conditions, for example, T cell-dependent response may contribute to mast cells proliferation or maturation/differentiation, high concentrations of eosinophils may beneﬁt the switch of mast cells from MMCs to MCTCs (42). Finally, mast cells may also participate in the regulation of their numbers and phenotypes, especially during inﬂammation or disease, by autocrine or paracrine or other potential mechanisms (56). For instance, IL-4 possessing growth factor activity for mast cells in mice can promote phenotypic conversion into CTMCs with IL-3 (57). The more detailed the genetic composition of their MCp or their different location in connective or mucosa tissue (37). In addition, like neutrophils (N1 and N2) and macrophages (M1 and M2), the complex biochemical environment of the tumor may promote mast cells differentiation into anti-tumor MC1 or pro-tumor MC2 (38). On the one hand, mast cells can generate excessive functionally active ROS which may induce cytotoxic effects that can promote tumor regression (39). On the other hand, large amounts of ROS exceed the capacity of cellular DNA repair systems, that may foster the occurrence of tumors. In addition, many other mast cell-derived mediators can also play distinct or even opposite roles in tumorigenesis (40). Frontiers in Immunology \| www.frontiersin.org CMCS DISTRIBUTION IN HEART AND THEIR FUNCTIONS In mice, CMCs are mostly distributed in the epicardium (50%) or myocardium (45%), and a fraction is distributed in the endocardium (5%) (58). Similarly, the most CMCs of human are located in the interstitium and in the epicardium (59). Mast cells and their mediators are generally thought to participate in allergic diseases, however, increasing evidences suggest that mast cells may also play protective roles in several other pathological or physiological processes (59, 60). Single-Cell Sequencing shows that CMCs are existence in myocardium and epicardium, and activated and expanding in pressure overload-driven heart failure mouse model (61), furthermore, CMCs inﬁltration increase atrial ﬁbrillation susceptibility following atrial burst stimulus (62). CMCs increase has been implicated in the chronic volume overload secondary to mitral regurgitation and aorto-caval ﬁstula (63). Furthermore, mast cells in different site may possess the functional heterogeneity (60), for example, tryptase can activate protease-activated receptor 2 (PAR-2) located on cardiomyocytes, which may play a protective role during myocardial infarction (64). Moreover, PAR-2 on nerve ﬁbers and myoﬁbroblasts can also be activated by tryptase, which stimulates the release of substance P from sensory nerve ﬁbers, which in turn activates MRGPRX2 receptors, a family of mas- related G-protein-coupled receptors, on human CMCs (1). The renin and chymase derived from the activated MRGPRX2 receptor, then respectively remove angiotensinogen and angiotensin I (Ang I) to form Ang II. The co-expression of renin and chymase by CMCs is very important for regulating the homeostasis of the cardiac renin-angiotensin system (59). Additionally, immunologic stimuli, bacterial and viral superantigens can activate primary human CMCs to release angiogenic (VEGF-A) and lymphangiogenic (VEGF-C) factors (1, 60, 65). Besides VEGF-A promoting angiogenesis, VEGF-C can also stablize blood pressure, promote lipid metabolism, and coronary artery development (60, 66–68). Multiple evidences indicate that mast cells may be involved in the development of the heart, so whether it has an impact on the aging of the heart? Although there is no direct evidence that mast cells are involved in heart aging, we can speculate from the effect of mast cells on the aging of other tissues and organs. Firstly, the number of mast cells in the mesenteric lymphatic vessels is 27% higher and in the mesentery is 400% higher of the older rats (24 months) compared with the younger rats (9 months) (76). CMCS DISTRIBUTION IN HEART AND THEIR FUNCTIONS In healthy elderly (≥75 years old), the mast cells in the skin increased by 40% compared with the biopsy of young people (≤30 years old) (77). Furthermore, although the liver has only a slight aging process compared to other organs, mast cells also play an important role in this process (32). One study has demonstrated that inhibiting SCF/c-Kit signaling pathway can reduce biliary senescence, with decrease mast cells activation and hepatic damage (78). In conclusion, the increase of mast cells can be detected in a variety of aging organs, so we suspect that they may play a role in the process of organ aging. The effect of mast cells on cardiac senescence can be reﬂected from two aspects: structure and function (79). It is normally assumed that the damage and apoptosis caused by mast cells to cardiomyocytes will eventually lead to cardiac dysfunction, a manifestation of cardiac aging. Co-culture of mast cells with cardiomyocytes promotes signiﬁcant cardiomyocytes apoptosis for possibly the exposure to mast cell granules (80). It has been suggested that chymase derived from CMCs may induce myosin degradation in cardiomyocytes (81). Furthermore, activation of CMCs is pro- inﬂammatory and not only induces apoptosis, but also leads to extracellular matrix degradation, which may lead to eventual myocardial dysfunction (Figure 2) (82). These data suggest that CMCs can induce heart aging, but a more detailed mechanism remains to be explored. THE MICROENVIRONMENT PROMOTES FORMATION OF SPECIFIC MAST CELLS PHENOTYPE Heterogeneity is a major feature of mast cells, reﬂecting the complex interaction between different microenvironmental signals transmitted by tissues and the differentiation programs that determine their phenotypes (41). However, how mast cells form a highly heterogeneous phenotype affected by microenvironment in peripheral tissues has rarely been mentioned. Generally, cells in a given population show heterogeneity, which means that once they show a certain minimum level of variation in one or more characteristics (42). The preliminary studies demonstrate that mast cells at different anatomical positions have signiﬁcant morphological differences (43). Other studies also show that in addition to differing in morphologic, rat and mouse MMCs July 2022 \| Volume 13 \| Article 963444 Frontiers in Immunology \| www.frontiersin.org 3 CMCs and Cardiac Homeostasis, Pathogenesis Jin et al. speculate that the CMCs may contribute to cardiac development. Similarly, CMCs density in children was low under the two years old, but the number of CMCs ﬁrstly increases and then decreases continuously with age (72). The rapid increase of CMCs density in the early postnatal period accompanies angiogenesis. Furthermore, the corneal mast cells promote corneal angiogenesis (73). All these data suggest that CMCs may play physiological roles on cardiac growth and development. In addition, immune-activated human CMCs can also produce VEGF-A and VEGF-C to induce the formation of new blood vessels and lymphatics, while the similar function in physiological state has not been conﬁrmed (60, 68). CMCs exist not only around the cardiac vessels of neonatal mouse, but also around the nerve ﬁbers. So CMCs might also have a positive effect on nerve development in heart (74, 75). mechanisms of the microenvironment on phenotype still need to be explored. Single cell RNA sequencing data of mast cells will provide further insights into heterogeneity as well as clear views of differences between and within different tissues. CMCS ON CARDIAC DEVELOPMENT, SENESCENCE AND FUNCTION Most reviews focus on the roles of mast cells in pathological conditions, the present review focuses on the physiological roles. A few studies have suggested that mast cells may participate in the morphogenesis of some mouse organs, such as the mammary glands (69) and corneal (70). The published data demonstrated that the density and number of CMCs are dynamically changed with age in rat (71). Our unpublished data also demonstrate that CMCs exist at embryonic stage in mouse heart. Therefore, we CMCS IN CARDIAC INFLAMMATION AND FUNCTIONAL REMODELING FOLLOWING INJURY inﬂux lead to mast cell degranulation (83). Currently, there are at least three ways in which mast cells release intracellular mediators, namely kiss-and-run, piecemeal, and compound exocytosis (84). In IgE-mediated allergic reactions, almost all vesicles are released from mast cells within minutes to hours. However, IgE is not the only trigger that stimulates mast cell degranulation, and activation induced by different components also leads to release of different mediators. There are numerous stimulants such as IgG, neuropeptides, cytokines, chemokines, TLR ligands, complements and other inﬂammatory products, that can directly cause mast cells to degranulate and selectively release mediators to stimulate proliferation, differentiation and migration (85). Mast cells subsets functions are different, not only because of the mediators produced, but also because of different sensitivities to stimulus. In addition to endogenous stimulus, some exogenous molecules can also directly activate mast cells, manifested as drug side effects or aggravating individual allergic state (86). It is worth mentioning that the process of mast cell degranulation in ﬁbrosis is different from that in allergic reactions, and the release of mast cell vesicles may be more frequent and accompany with more subtle symptoms. It can occur by a slow process called piecemeal degranulation, and the vesicles can travel through the lymphatic vessels across the interstitial space to distant lymph nodes. Additionally, less discussed mechanism is the direct penetration of mast cell vesicles into another cell via intercellular contact, known as the transgranulation (83). The growing evidence shows that CMCs plays an important role in the occurrence and development of cardiovascular diseases (87). After myocardial infarction, CMCs density increase, rapidly degranulate, release a large number of bioactive mediators and initiate a cascade of cytokines to promote early inﬂammatory healing (88). CMCs play an undeniable role in the cardiac inﬂammation initiation and resolution. Because optimal healing requires inhibition of chemokine and cytokine synthesis, this leads to regression of inﬂammation and collagen deposition (31). However, the inﬂuence of CMCs on ﬁbrosis remains a focus. CMCs produce a variety of growth factors, angiogenic factors and extracellular matrix regulators. All the products can affect matrix remodeling, promote granulation and scar formation, and have an important role on cardiac remodeling. Inﬂammatory Development and Resolution The association of inﬂammation with myocardial infarction has been perceived for more than a century and inﬂammation is properly considered part of the healing process. THE TRIGGER OF CMCS ACTIVATION AND DEGRANULATION Degranulation is considered to be the main way of mast cells playing physiological and pathological roles with IgE as the main trigger. As well known, mast cells express a large number of FcϵRI receptors, once IgE receptor cross-linking and calcium July 2022 \| Volume 13 \| Article 963444 Frontiers in Immunology \| www.frontiersin.org 4 Jin et al. Jin et al. CMCs and Cardiac Homeostasis, Pathogenesis FIGURE 2 \| The roles of CMCs in cardiac development and aging. VEGF-A, VEGF-C, NGF and neurotrophin from CMCs contribute to cardiac development through beneﬁting the formation of blood vessels and lymphatic vessels, and the development of cardiac nerves (neonatal stage). CMCs produce Chymase, TNF and IL-1b to degrade myosin or to damage cardiomyocytes. CMCs’ activation following cardiac injury causes inﬂammatory response, and then lead to the structural damage and cardiac dysfunction (Old Stage). The red arrow means speculation. FIGURE 2 \| The roles of CMCs in cardiac development and aging. VEGF-A, VEGF-C, NGF and neurotrophin from CMCs contribute to cardiac development through beneﬁting the formation of blood vessels and lymphatic vessels, and the development of cardiac nerves (neonatal stage). CMCs produce Chymase, TNF and IL-1b to degrade myosin or to damage cardiomyocytes. CMCs’ activation following cardiac injury causes inﬂammatory response, and then lead to the structural damage and cardiac dysfunction (Old Stage). The red arrow means speculation. Frontiers in Immunology \| www.frontiersin.org CMCS IN CARDIAC INFLAMMATION AND FUNCTIONAL REMODELING FOLLOWING INJURY Eventually it mediates the inﬁltration of inﬂammatory cells, such as neutrophils, basophils, monocytes/ macrophages, lymphocytes, etc (85) (89). Inﬂammation beneﬁts to cardiac repair, but this effect does not last(88).Thereleaseofcytokinesandinﬂammatorycellsinﬁltration directly or indirectly induced by CMCs are signiﬁcant events in the progression of myocardial infarction, which play a key role in phagocytosis and clearance of dead cells and debris. Nevertheless, this acute inﬂammatory response is transient and then disappears (31, 90). This may be related to some anti-inﬂammatory mediators secreted by CMCs, such as IL-10 and IL-13, which can limit the expansion of inﬂammatory response and protect non-infarcted cardiomyocytes. IL-10 restrains the inﬂammatory response by inhibiting the production of IL-1a, IL-1b, TNF-a, IL-6, and IL-8 through lipopolysaccharide-activated monocytes (91). This can be demonstrated by the obvious inﬂammatory response of IL-10 knockout mice after myocardial infarction, which is characterized by increased neutrophil inﬁltration and elevated blood TNF-a levels (92). The importance of IL-13 on CMCs needs to be further investigated as it is not only derived from CMCs, but also secreted bymanyothercellsinthemicroenvironment.Inaddition,mastcells can also exert anti-inﬂammatory or immunosuppressive effects by releasing mediators that degrade proinﬂammatory molecules (52). Mast cell proteinase 4 has been shown to degrade mast cell-derived TNF in mice in vitro, and it also can reduce TNF levels in vivo and limit inﬂammation (93). Besides, IL-37 is an important regulatory cytokine that inhibits inﬂammation, and mast cells can modulate the anti-inﬂammatory activity of IL-37 by trypsin-like action, resulting in the more biologically active form of IL-37 (94). Notably, VEGF-C is a major lymphangiogenic factor produced by humanCMCs(95,96),whichhasapotentialcardioprotectiveeffect, as cardiac lymphatic activation contributes to inﬂammation resolution and plays a crucial role in ﬁghting myocardial edema (60, 97). Furthermore, mast cells can also inhibit inﬂammation through activation of PAR-2 on cardiomyocytes (64). Timely suppression of the inﬂammatory mediators such as chemokines and cytokines in healing infarction is critical to the repair process and can inhibit the continuous recruitment of inﬂammatory cells (31). More detailed anti-inﬂammatory mechanisms of mast cells remain to be studied. If we can ﬁnd out the speciﬁc mechanism of the occurrence and resolution of cardiac inﬂammation regulated by CMCs, and then identify clinically appropriate targets, it may bring great improvement to the treatment of cardiovascular disease. CMCS IN CARDIAC INFLAMMATION AND FUNCTIONAL REMODELING FOLLOWING INJURY The involvement of mast cells in inﬂammation has traditionally been thought to be only one aspect of the allergic response, but this does not seem to be the case. Following cardiac injury, the internal and external factors mentioned above can induce CMCs degranulation, and July 2022 \| Volume 13 \| Article 963444 Frontiers in Immunology \| www.frontiersin.org 5 CMCs and Cardiac Homeostasis, Pathogenesis Jin et al. Firstly, ﬁbrosis is necessary for proper wound healing which can restores function to damaged tissue after myocardial injury, such as myocardial infarction or hypertension-induced stretch injury. Chymase and tryptase in CMCs have pro-ﬁbrotic properties which are well-known ﬁbroblast activity promoters, can mediate the activation of TGF-b and Ang II. However, ﬁbrotic deposits are essential to restore normal heart function, but excessive remodeling can reduce contractile force and heart function, resulting in chronic heart failure (15). Additionally, CMCs can also secrete some anti- ﬁbrotic mediators, suchas IL-10, IL-13, CXCL-10 and VEGF, which have their own anti-ﬁbrotic pathways, respectively (98). For example, IL-10 can reduce ﬁbrotic remodeling by decreasing IL-1b and TNF levels, as well as MMP-9 expression and activity, and by increasing capillary density (99). CMCs-derived IL-13 can induce macrophages with an M2c phenotype, which is associated with reduced ﬁbrosis. Moreover, VEGF-A can increase capillary density in damaged tissues and promote proper repair of cardiac ﬁbrosis (93). At last, CXCL10 has been proved that it can inhibit the migration of ﬁbroblasts to myocardium and delay their phenotypic differentiation into ﬁbrogenic myoﬁbroblasts. (Figure 3) (15). It cannot be ignored that CMCs have signiﬁcant pro-ﬁbrotic and anti-ﬁbrotic effects, several studies have drawn controversial conclusions and described possible implications for thisphenomenon, including harmful,neutral,orprotectiveeffectsin cardiac remodeling (100). These conﬂicting conclusions are attributed to the failure to ensure a strictly correct clinical environment and the selection of appropriate animal models (27). Different culture systems, primary cell sources and even the initial cell number used in the experiment are also critical, and subtle differences may lead to different or even contradictory conclusions. To clarify these contradictory results, it is signiﬁcant to correctly understand the characteristics of each in vitro and in vivo system used to culture mast cells, which can help us understand the real function of CMCs in the heart (21). their derived histamine and TNF activate microvascular endothelium, up-regulate P and E-selectin, respectively, as well as adhesion molecules such as ICAM-1, which affect vascular tension and permeability. FUTURE PERSPECTIVE Like macrophages and dendritic cells, CMCs are highly heterogeneous population of innate immune cells, with different morphological functions, mediator contents and surface receptors. The origin and differentiation of the different subsets remain unclear. CMCs are strategically located in close proximity to cardiomyocytes, coronary microvessels, nerves, and lymphatic vessels. Understanding the speciﬁc roles of CMCs in different sites of the heart in pathological and physiological processes will lead to a breakthrough in the treatment of cardiovascular diseases. Although they are distributed in small numbers and proportions within the steady-state heart, we reasonably suspect they are linked to the cardiac development and function, even the aging process. In a word, CMCs are a double-edged sword that may have potentially beneﬁcial or harmful effects. The detailed roles of CMCs in cardiac development and injury remain controversial and contradictory, thus, several key questions about them remain unanswered. For example, the mechanisms about migration and differentiation of CMCs remain to be conﬁrmed: whether CMCs precursors are regulated by speciﬁc mediators during migration to FUNDING This work was supported by National Natural Science Foundation of China (Grant No. 81871244), Primary Research & Development Plan of Jiangsu Province (BE2019700), Jiangsu Province “333” project (BRA2018016), Six talent peaks project in Jiangsu Province (2019-WSN-122), Projects of International Cooperation from Jiangsu (BX2019100), and international cooperation and exchange from Zhenjiang (GJ2020010). CMCs: Pro-Fibrosis and Anti-Fibrosis Although the obvious inﬂammation-related properties of CMCs, its main function in cardiac remodeling is related to the regulation of ﬁbrous tissue metabolism. Cardiac ﬁbrosis is actually an accumulation of the extracellular matrix, such as collagen (89). However, current studies have found that CMCs are double-edged sword in inducing cardiac remodeling, which can not only stimulate collagen synthesis and lead to ﬁbrosis, but also induce matrix metalloproteinase activation and collagen degradation, with ultimately ventricular dilation (87). July 2022 \| Volume 13 \| Article 963444 Frontiers in Immunology \| www.frontiersin.org 6 Jin et al. CMCs and Cardiac Homeostasis, Pathogenesis Jin et al. FIGURE 3 \| CMCs play an important role in cardiac remodeling following injury. Following cardiac injury, ﬁrstly, CMCs are activated, release a large number of inﬂammatory mediators, such as histamine, TNF-a and IL-1 to change vascular permeability, recruit a lot of inﬂammatory cells (basophils, monocytes, neutrophils, etc.), and induce early inﬂammation. Secondly, CMCs also mediate the inﬂammatory resolution by secreting some anti-inﬂammatory factors such as IL-10, IL-13 and IL-37. Finally, a series of pro-ﬁbrotic and anti-ﬁbrotic effects coexist and lead to cardiac remodeling. FIGURE 3 \| CMCs play an important role in cardiac remodeling following injury. Following cardiac injury, ﬁrstly, CMCs are activated, release a large number of inﬂammatory mediators, such as histamine, TNF-a and IL-1 to change vascular permeability, recruit a lot of inﬂammatory cells (basophils, monocytes, neutrophils, etc.), and induce early inﬂammation. Secondly, CMCs also mediate the inﬂammatory resolution by secreting some anti-inﬂammatory factors such as IL-10, IL-13 and IL-37. Finally, a series of pro-ﬁbrotic and anti-ﬁbrotic effects coexist and lead to cardiac remodeling. AUTHOR CONTRIBUTIONS heart, whether CMCs proliferate and renew according to the pathway we mentioned above in both presence and absence of pathological injury, and whether cardiac-resident and recruited mast cells play divergent roles during homeostasis. Do different CMCs subsets have the same origin and developmental process, and whether their different phenotypes are changed by their microenvironment or driven by their designated progenitor cells? Speciﬁc mechanisms of CMCs on the development and function of the heart remain in the speculative stage. Its effect on the aging of the heart is only inferred from the performance of other organs. Some direct evidence is still lacking. With the growing understanding to CMCs, the other function may be demonstrated in future except for pro-inﬂammation and pro- ﬁbrosis in cardiac injury. However, the dispute as to whether they perform harmful, neutral or protective activities has also not been resolved. JJ and YJ collected the material and draw Figures. JJ and ZS wrote the draft. ZS provides ideas and grant. SC revised the manuscript. All authors contributed to the article and approved the submitted version. Visions to Precision Medicine Concepts. Theranostics (2020) 10(23):10743–68. doi: 10.7150/thno.46719 4. Li Z, Liu S, Xu J, Zhang X, Han D, Liu J, et al. Adult Connective Tissue-Resident Mast Cells Originate From Late Erythro-Myeloid Progenitors. Immunity (2018) 49(4):640–53.e5. doi: 10.1016/j.immuni.2018.09.023 5. Kitamura Y, Shimada M, Hatanaka K, Miyano Y. Development of Mast Cells From Grafted Bone Marrow Cells in Irradiated Mice. Nature (1977) 268 (5619):442–3. doi: 10.1038/268442a0 6. Dahlin JS, Maurer M, Metcalfe DD, Pejler G, Sagi-Eisenberg R, Nilsson G. The Ingenious Mast Cell: Contemporary Insights Into Mast Cell Behavior and Function. Allergy. (2022) 77(1):83–99. doi: 10.1111/all.14881 5. Kitamura Y, Shimada M, Hatanaka K, Miyano Y. Development of Mast Cells From Grafted Bone Marrow Cells in Irradiated Mice. Nature (1977) 268 (5619):442–3. doi: 10.1038/268442a0 Visions to Precision Medicine Concepts. Theranostics (2020) 10(23):10743–68. doi: 10.7150/thno.46719 6. Dahlin JS, Maurer M, Metcalfe DD, Pejler G, Sagi-Eisenberg R, Nilsson G. The Ingenious Mast Cell: Contemporary Insights Into Mast Cell Behavior and Function. Allergy. (2022) 77(1):83–99. doi: 10.1111/all.14881 4. Li Z, Liu S, Xu J, Zhang X, Han D, Liu J, et al. Adult Connective Tissue-Resident Mast Cells Originate From Late Erythro-Myeloid Progenitors. Immunity (2018) 49(4):640–53.e5. doi: 10.1016/j.immuni.2018.09.023 REFERENCES J Immunol (2004) 172(1):260–7. doi: 10.4049/ jimmunol.172.1.260 12. Gurish MF, Tao H, Abonia JP, Arya A, Friend DS, Parker CM, et al. Intestinal Mast Cell Progenitors Require CD49dbeta7 (Alpha4beta7 Integrin) for Tissue-Speciﬁc Homing. J Exp Med (2001) 194(9):1243–52. doi: 10.1084/ jem.194.9.1243 13. Abonia JP, Hallgren J, Jones T, Shi T, Xu Y, Koni P, et al. Alpha-4 Integrins and VCAM-1, But Not MAdCAM-1, are Essential for Recruitment of Mast Cell Progenitors to the Inﬂamed Lung. Blood (2006) 108(5):1588–94. doi: 10.1182/blood-2005-12-012781 31. Frangogiannis NG. The Mechanistic Basis of Infarct Healing. Antioxid Redox Signal (2006) 8(11-12):1907–39. doi: 10.1089/ars.2006.8.1907 32. Jarido V, Kennedy L, Hargrove L, Demieville J, Thomson J, Stephenson K, et al. The Emerging Role of Mast Cells in Liver Disease. Am J Physiol Gastrointest Liver Physiol (2017) 313(2):G89–101. doi: 10.1152/ ajpgi.00333.2016 14. Dahlin JS, Hallgren J. Mast Cell Progenitors: Origin, Development and Migration to Tissues. Mol Immunol (2015) 63(1):9–17. doi: 10.1016/ j.molimm.2014.01.018 15. Legere SA, Haidl ID, Legare JF, Marshall JS. Mast Cells in Cardiac Fibrosis: New Insights Suggest Opportunities for Intervention. Front Immunol (2019) 10:580. doi: 10.3389/ﬁmmu.2019.00580 33. Balakumar P, Singh AP, Ganti SS, Krishan P, Ramasamy S, Singh M. Resident Cardiac Mast Cells: Are They the Major Culprit in the Pathogenesis of Cardiac Hypertrophy? Basic Clin Pharmacol Toxicol (2008) 102(1):5–9. doi: 10.1111/ j.1742-7843.2007.00147.x 16. Tsai M, Takeishi T, Thompson H, Langley KE, Zsebo KM, Metcalfe DD, et al. Induction of Mast Cell Proliferation, Maturation, and Heparin Synthesis by the Rat C-Kit Ligand, Stem Cell Factor. Proc Natl Acad Sci USA (1991) 88 (14):6382–6. doi: 10.1073/pnas.88.14.6382 34. Sperr WR, Bankl HC, Mundigler G, Klappacher G, Grossschmidt K, Agis H, et al. The Human Cardiac Mast Cell: Localization, Isolation, Phenotype, and Functional Characterization. Blood (1994) 84(11):3876–84. doi: 10.1182/ blood.V84.11.3876.bloodjournal84113876 17. Cho KA, Park M, Kim YH, Woo SY. Th17 Cell-Mediated Immune Responses Promote Mast Cell Proliferation by Triggering Stem Cell Factor in Keratinocytes. Biochem Biophys Res Commun (2017) 487(4):856–61. doi: 10.1016/j.bbrc.2017.04.141 35. Krishnaswamy G, Kelley J, Johnson D, Youngberg G, Stone W, Huang SK, et al. The Human Mast Cell: Functions in Physiology and Disease. Front Biosci (2001) 6:D1109–27. doi: 10.2741/krishnas 18. Ihle JN, Keller J, Oroszlan S, Henderson LE, Copeland TD, Fitch F, et al. Biologic Properties of Homogeneous Interleukin 3. I. Demonstration of WEHI-3 Growth Factor Activity, Mast Cell Growth Factor Activity, P Cell- Stimulating Factor Activity, Colony-Stimulating Factor Activity, and Histamine-Producing Cell-Stimulating Factor Activity. REFERENCES 1. Varricchi G, Pecoraro A, Loffredo S, Poto R, Rivellese F, Genovese A, et al. Heterogeneity of Human Mast Cells With Respect to MRGPRX2 Receptor Expression and Function. Front Cell Neurosci (2019) 13:299. doi: 10.3389/ fncel.2019.00299 2. Gersch C, Dewald O, Zoerlein M, Michael LH, Entman ML, Frangogiannis NG. Mast Cells and Macrophages in Normal C57/BL/6 Mice. Histochem Cell Biol (2002) 118(1):41–9. doi: 10.1007/s00418-002-0425-z 3. Valent P, Akin C, Hartmann K, Nilsson G, Reiter A, Hermine O, et al. Mast Cells as a Unique Hematopoietic Lineage and Cell System: From Paul Ehrlich's July 2022 \| Volume 13 \| Article 963444 Frontiers in Immunology \| www.frontiersin.org 7 CMCs and Cardiac Homeostasis, Pathogenesis Jin et al. 7. Liu C, Liu Z, Li Z, Wu Y. Molecular Regulation of Mast Cell Development and Maturation. Mol Biol Rep (2010) 37(4):1993–2001. doi: 10.1007/s11033-009- 9650-z 26. Moller C, Karlberg M, Abrink M, Nakayama KI, Motoyama N, Nilsson G. Bcl- 2 and Bcl-XL are Indispensable for the Late Phase of Mast Cell Development From Mouse Embryonic Stem Cells. Exp Hematol (2007) 35(3):385–93. doi: 10.1016/j.exphem.2006.11.008 8. Meininger CJ, Yano H, Rottapel R, Bernstein A, Zsebo KM, Zetter BR. The C- Kit Receptor Ligand Functions as a Mast Cell Chemoattractant. Blood (1992) 79 (4):958–63. doi: 10.1182/blood.V79.4.958.bloodjournal794958 27. Bradding P, Pejler G. The Controversial Role of Mast Cells in Fibrosis. Immunol Rev (2018) 282(1):198–231. doi: 10.1111/imr.12626 9. Gruber BL, Marchese MJ, Kew R. Angiogenic Factors Stimulate Mast-Cell Migration. Blood. (1995) 86(7):2488–93. doi: 10.1182/blood.V86.7.2488.2488 28. Levi-Schaffer F, Austen KF, Caulﬁeld JP, Hein A, Gravallese PM, Stevens RL. Co-Culture of Human Lung-Derived Mast Cells With Mouse 3T3 Fibroblasts: Morphology and IgE-Mediated Release of Histamine, Prostaglandin D2, and Leukotrienes. J Immunol (1987) 139(2):494–500. 10. Taub D, Dastych J, Inamura N, Upton J, Kelvin D, Metcalfe D, et al. Bone Marrow-Derived Murine Mast Cells Migrate, But Do Not Degranulate, in Response to Chemokines. J Immunol (1995) 154(5):2393–402. 29. Levi-Schaffer F, Kelav-Appelbaum R, Rubinchik E. Human Foreskin Mast Cell Viability and Functional Activity is Maintained Ex Vivo by Coculture With Fibroblasts. Cell Immunol (1995) 162(2):211–6. doi: 10.1006/cimm.1995.1071 11. Hallgren J, Gurish MF. Mast Cell Progenitor Trafﬁcking and Maturation. Adv Exp Med Biol (2011) 716:14–28. doi: 10.1007/978-1-4419-9533-9_2 30. Sellge G, Lorentz A, Gebhardt T, Levi-Schaffer F, Bektas H, Manns MP, et al. Human Intestinal Fibroblasts Prevent Apoptosis in Human Intestinal Mast Cells by a Mechanism Independent of Stem Cell Factor, IL-3, IL-4, and Nerve Growth Factor. REFERENCES J Immunol (1983) 131(1):282–7. 36. Derakhshan T, Samuchiwal SK, Hallen N, Bankova LG, Boyce JA, Barrett NA, et al. Lineage-Speciﬁc Regulation of Inducible and Constitutive Mast Cells in Allergic Airway Inﬂammation. J Exp Med (2021) 218(1):e20200321. doi: 10.1084/jem.20200321 37. Gurish MF, Austen KF. Developmental Origin and Functional Specialization of Mast Cell Subsets. Immunity (2012) 37(1):25–33. doi: 10.1016/ j.immuni.2012.07.003 19. Rennick DM, Lee FD, Yokota T, Arai KI, Cantor H, Nabel GJ. A Cloned MCGF cDNA Encodes a Multilineage Hematopoietic Growth Factor: Multiple Activities of Interleukin 3. J Immunol (1985) 134(2):910–4. 38. Varricchi G, de Paulis A, Marone G, Galli SJ. Future Needs in Mast Cell Biology. Int J Mol Sci (2019) 20(18):4397. doi: 10.3390/ijms20184397 20. Lantz CS, Boesiger J, Song CH, Mach N, Kobayashi T, Mulligan RC, et al. Role for Interleukin-3 in Mast-Cell and Basophil Development and in Immunity to Parasites. Nature. (1998) 392(6671):90–3. doi: 10.1038/32190 39. Swindle EJ, Metcalfe DD, Coleman JW. Rodent and Human Mast Cells Produce Functionally Signiﬁcant Intracellular Reactive Oxygen Species But Not Nitric Oxide. J Biol Chem (2004) 279(47):48751–9. doi: 10.1074/ jbc.M409738200 21. Hu ZQ, Zhao WH, Shimamura T. Regulation of Mast Cell Development by Inﬂammatory Factors. Curr Med Chem (2007) 14(28):3044–50. doi: 10.2174/ 092986707782793998 40. Kessenbrock K, Plaks V, Werb Z. Matrix Metalloproteinases: Regulators of the Tumor Microenvironment. Cell (2010) 141(1):52–67. doi: 10.1016/ j.cell.2010.03.015 22. Mekori YA, Oh CK, Metcalfe DD. The Role of C-Kit and its Ligand, Stem Cell Factor, in Mast Cell Apoptosis. Int Arch Allergy Immunol (1995) 107(1- 3):136–8. doi: 10.1159/000236955 41. Frossi B, Mion F, Sibilano R, Danelli L, Pucillo CEM. Is it Time for a New Classiﬁcation of Mast Cells? What do We Know About Mast Cell Heterogeneity? Immunol Rev (2018) 282(1):35–46. doi: 10.1111/imr.12636 23. Möller C, Alfredsson J, Engström M, Wootz H, Xiang Z, Lennartsson J, et al. Stem Cell Factor Promotes Mast Cell Survival via Inactivation of FOXO3a- Mediated Transcriptional Induction and MEK-Regulated Phosphorylation of the Proapoptotic Protein Bim. Blood (2005) 106(4):1330–6. doi: 10.1182/ blood-2004-12-4792 42. Galli SJ. New Insights Into "the Riddle of the Mast Cells": Microenvironmental Regulation of Mast Cell Development and Phenotypic Heterogeneity. Lab Invest (1990) 62(1):5–33. doi: 10.1007/978-1-4612-0485-5_5 43. Hardy WB, Wesbrook FF. The Wandering Cells of the Alimentary Canal. J Physiol (1895) 18(5-6):490–i3. doi: 10.1113/jphysiol.1895.sp000582 24. Kohno M, Yamasaki S, Tybulewicz VL, Saito T. REFERENCES Rapid and Large Amount of Autocrine IL-3 Production is Responsible for Mast Cell Survival by IgE in the Absence of Antigen. Blood (2005) 105(5):2059–65. doi: 10.1182/blood-2004- 07-2639 44. Mayrhofer G, Fisher R. Mast Cells in Severely T-Cell Depleted Rats and the Response to Infestation With Nippostrongylus Brasiliensis. Immunology (1979) 37(1):145–55. 45. Jarrett EE, Haig DM. Mucosal Mast Cells In Vivo and In Vitro. Immunol Today (1984) 5(4):115–9. doi: 10.1016/0167-5699(84)90046-X 25. Karlsson R, Engström M, Jönsson M, Karlberg P, Pronk CJ, Richter J, et al. Phosphatidylinositol 3-Kinase is Essential for Kit Ligand-Mediated Survival, Whereas Interleukin-3 and Flt3 Ligand Induce Expression of Antiapoptotic Bcl-2 Family Genes. J Leukoc Biol (2003) 74(5):923–31. doi: 10.1189/ jlb.0403142 46. Befus AD, Pearce FL, Gauldie J, Horsewood P, Bienenstock J. Mucosal Mast Cells. I. Isolation and Functional Characteristics of Rat Intestinal Mast Cells. J Immunol (1982) 128(6):2475–80. July 2022 \| Volume 13 \| Article 963444 Frontiers in Immunology \| www.frontiersin.org CMCs and Cardiac Homeostasis, Pathogenesis Jin et al. 47. Razin E, Stevens RL, Akiyama F, Schmid K, Austen KF. Culture From Mouse Bone Marrow of a Subclass of Mast Cells Possessing a Distinct Chondroitin Sulfate Proteoglycan With Glycosaminoglycans Rich in N- Acetylgalactosamine-4,6-Disulfate. J Biol Chem (1982) 257(12):7229–36. doi: 10.1016/S0021-9258(18)34561-7 64. Ngkelo A, Richart A, Kirk JA, Bonnin P, Vilar J, Lemitre M, et al. Mast Cells Regulate Myoﬁlament Calcium Sensitization and Heart Function After Myocardial Infarction. J Exp Med (2016) 213(7):1353–74. doi: 10.1084/ jem.20160081 65. Varricchi G, Loffredo S, Borriello F, Pecoraro A, Rivellese F, Genovese A, et al. Superantigenic Activation of Human Cardiac Mast Cells. Int J Mol Sci (2019) 20(8):1828. doi: 10.3390/ijms20081828 48. Kawanami O, Ferrans VJ, Fulmer JD, Crystal RG. Ultrastructure of Pulmonary Mast Cells in Patients With Fibrotic Lung Disorders. Lab Invest (1979) 40(6):717–34. 66. Machnik A, Neuhofer W, Jantsch J, Dahlmann A, Tammela T, Machura K, et al. Macrophages Regulate Salt-Dependent Volume and Blood Pressure by a Vascular Endothelial Growth Factor-C-Dependent Buffering Mechanism. Nat Med (2009) 15(5):545–52. doi: 10.1038/nm.1960 49. Ruitenberg EJ, Gustowska L, Elgersma A, Ruitenberg HM. Effect of Fixation on the Light Microscopical Visualization of Mast Cells in the Mucosa and Connective Tissue of the Human Duodenum. Int Arch Allergy Appl Immunol (1982) 67(3):233–8. doi: 10.1159/000233024 67. Martel C, Li W, Fulp B, Platt AM, Gautier EL, Westerterp M, et al. Lymphatic Vasculature Mediates Macrophage Reverse Cholesterol Transport in Mice. J Clin Invest (2013) 123(4):1571–9. doi: 10.1172/JCI63685 50. REFERENCES Int J Immunopathol Pharmacol (2014) 27(1):15–9. doi: 10.1177/039463201402700103 55. Akgul A, Skrabal CA, Thompson LO, Loebe M, Lafuente JA, Noon GP, et al. Role of Mast Cells and Their Mediators in Failing Myocardium Under Mechanical Ventricular Support. J Heart Lung Transplant (2004) 23 (6):709–15. doi: 10.1016/j.healun.2003.06.006 76. Chatterjee V, Gashev AA. Aging-Associated Shifts in Functional Status of Mast Cells Located by Adult and Aged Mesenteric Lymphatic Vessels. Am J Physiol Heart Circ Physiol (2012) 303(6):H693–702. doi: 10.1152/ ajpheart.00378.2012 56. Galli SJ. Mast Cells and Basophils. Curr Opin Hematol (2000) 7(1):32–9. doi: 10.1097/00062752-200001000-00007 57. Hamaguchi Y, Kanakura Y, Fujita J, Takeda S, Nakano T, Tarui S, et al. Interleukin 4 as an Essential Factor for In Vitro Clonal Growth of Murine Connective Tissue-Type Mast Cells. J Exp Med (1987) 165(1):268–73. doi: 10.1084/jem.165.1.268 77. Pilkington SM, Barron MJ, Watson REB, Grifﬁths CEM, Bulfone-Paus S. Aged Human Skin Accumulates Mast Cells With Altered Functionality That Localize to Macrophages and Vasoactive Intestinal Peptide-Positive Nerve Fibres. Br J Dermatol (2019) 180(4):849–58. doi: 10.1111/bjd.17268 58. Ingason AB, Mechmet F, Atacho DAM, Steingrı́msson E, Petersen PH. Distribution of Mast Cells Within the Mouse Heart and its Dependency on Mitf. Mol Immunol (2019) 105:9–15. doi: 10.1016/j.molimm.2018.11.009 78. Meadows V, Kennedy L, Hargrove L, Demieville J, Meng F, Virani S, et al. Downregulation of Hepatic Stem Cell Factor by Vivo-Morpholino Treatment Inhibits Mast Cell Migration and Decreases Biliary Damage/Senescence and Liver Fibrosis in Mdr2-/- Mice. Biochim Biophys Acta Mol Basis Dis (2019) 1865(12):165557. doi: 10.1016/j.bbadis.2019.165557 59. Varricchi G, Marone G, Kovanen PT. Cardiac Mast Cells: Underappreciated Immune Cells in Cardiovascular Homeostasis and Disease. Trends Immunol (2020) 41(8):734–46. doi: 10.1016/j.it.2020.06.006 79. Steenman M, Lande G. Cardiac Aging and Heart Disease in Humans. Biophys Rev (2017) 9(2):131–7. doi: 10.1007/s12551-017-0255-9 60. Varricchi G, Rossi FW, Galdiero MR, Granata F, Criscuolo G, Spadaro G, et al. Physiological Roles of Mast Cells: Collegium Internationale Allergologicum Update 2019. Int Arch Allergy Immunol (2019) 179(4):247–61. doi: 10.1159/ 000500088 80. Hara M, Matsumori A, Ono K, Kido H, Hwang MW, Miyamoto T, et al. Mast Cells Cause Apoptosis of Cardiomyocytes and Proliferation of Other Intramyocardial Cells In Vitro. Circulation (1999) 100(13):1443–9. doi: 10.1161/01.CIR.100.13.1443 61. Martini E, Kunderfranco P, Peano C, Carullo P, Cremonesi M, Schorn T, et al. Single-Cell Sequencing of Mouse Heart Immune Inﬁltrate in Pressure Overload-Driven Heart Failure Reveals Extent of Immune Activation. Circulation (2019) 140(25):2089–107. doi: 10.1161/CIRCULATIONAHA. 119.041694 81. REFERENCES Leung KB, Flint KC, Brostoff J, Hudspith BN, Johnson NM, Lau HY, et al. Effects of Sodium Cromoglycate and Nedocromil Sodium on Histamine Secretion From Human Lung Mast Cells. Thorax (1988) 43(10):756–61. doi: 10.1136/thx.43.10.756 68. Chen HI, Poduri A, Numi H, Kivela R, Saharinen P, McKay AS, et al. VEGF-C and Aortic Cardiomyocytes Guide Coronary Artery Stem Development. J Clin Invest (2014) 124(11):4899–914. doi: 10.1172/JCI77483 51. Lawrence ID, Warner JA, Cohan VL, Hubbard WC, Kagey-Sobotka A, Lichtenstein LM. Puriﬁcation and Characterization of Human Skin Mast Cells. Evidence for Human Mast Cell Heterogeneity. J Immunol (1987) 139 (9):3062–9. 69. Lilla JN, Werb Z. Mast Cells Contribute to the Stromal Microenvironment in Mammary Gland Branching Morphogenesis. Dev Biol (2010) 337(1):124–33. doi: 10.1016/j.ydbio.2009.10.021 70. Liu J, Fu T, Song F, Xue Y, Xia C, Liu P, et al. Mast Cells Participate in Corneal Development in Mice. Sci Rep (2015) 5:17569. doi: 10.1038/srep17569 52. Galli SJ, Kalesnikoff J, Grimbaldeston MA, Piliponsky AM, Williams CM, Tsai M. Mast Cells as "Tunable" Effector and Immunoregulatory Cells: Recent Advances. Annu Rev Immunol (2005) 23:749–86. doi: 10.1146/ annurev.immunol.21.120601.141025 71. Yong LC, Watkins SG, Boland JE. The Mast Cell: III. Distribution and Maturation in Various Organs of the Young Rat. Pathology (1979) 11 (3):427–45. doi: 10.3109/00313027909059020 53. Nakano T, Sonoda T, Hayashi C, Yamatodani A, Kanayama Y, Yamamura T, et al. Fate of Bone Marrow-Derived Cultured Mast Cells After Intracutaneous, Intraperitoneal, and Intravenous Transfer Into Genetically Mast Cell- Deﬁcient W/Wv Mice. Evidence That Cultured Mast Cells can Give Rise to Both Connective Tissue Type and Mucosal Mast Cells. J Exp Med (1985) 162 (3):1025–43. doi: 10.1084/jem.162.3.1025 72. Hellstrom B, Holmgren H. Numerical Distribution of Mast Cells in the Human Skin and Heart. Acta Anat (Basel) (1950) 10(1-2):81–107. doi: 10.1159/000140456 73. Rakusan K, Sarkar K, Turek Z, Wicker P. Mast Cells in the Rat Heart During Normal Growth and in Cardiac Hypertrophy. Circ Res (1990) 66(2):511–6. doi: 10.1161/01.RES.66.2.511 54. Otsu K, Nakano T, Kanakura Y, Asai H, Katz HR, Austen KF, et al. Phenotypic Changes of Bone Marrow-Derived Mast Cells After Intraperitoneal Transfer Into W/Wv Mice That are Genetically Deﬁcient in Mast Cells. J Exp Med (1987) 165(3):615–27. doi: 10.1084/jem.165.3.615 74. Blennerhassett MG. Nerve and Mast Cell Interaction: Cell Conﬂict or Information Exchange? Prog Clin Biol Res (1994) 390:225–41. 75. Kritas SK, Caraffa A, AntinolﬁP, Saggini A, Pantalone A, Rosati M, et al. Nerve Growth Factor Interactions With Mast Cells. REFERENCES Powell PC, Wei CC, Fu L, Pat B, Bradley WE, Collawn JF, et al. Chymase Uptake by Cardiomyocytes Results in Myosin Degradation in Cardiac Volume Overload. Heliyon (2019) 5(4):e01397. doi: 10.1016/ j.heliyon.2019.e01397 62. Liao CH, Akazawa H, Tamagawa M, Ito K, Yasuda N, Kudo Y, et al. Cardiac Mast Cells Cause Atrial Fibrillation Through PDGF-A-Mediated Fibrosis in Pressure-Overloaded Mouse Hearts. J Clin Invest (2010) 120(1):242–53. doi: 10.1172/JCI39942 82. Kritikou E, Kuiper J, Kovanen PT, Bot I. The Impact of Mast Cells on Cardiovascular Diseases. Eur J Pharmacol (2016) 778:103–15. doi: 10.1016/ j.ejphar.2015.04.050 83. Hugle T. Beyond Allergy: The Role of Mast Cells in Fibrosis. Swiss Med Wkly (2014) 144:w13999. doi: 10.4414/smw.2014.13999 63. Janicki JS, Brower GL, Gardner JD, Forman MF, Stewart JAJr., Murray DB, et al. Cardiac Mast Cell Regulation of Matrix Metalloproteinase-Related Ventricular Remodeling in Chronic Pressure or Volume Overload. Cardiovasc Res (2006) 69(3):657–65. doi: 10.1016/j.cardiores.2005.10.020 84. Sahid MNA, Kiyoi T. Mast Cell Activation Markers for In Vitro Study. J Immunoassay Immunochem (2020) 41(4):778–816. doi: 10.1080/ 15321819.2020.1769129 July 2022 \| Volume 13 \| Article 963444 Frontiers in Immunology \| www.frontiersin.org 9 CMCs and Cardiac Homeostasis, Pathogenesis Jin et al. 85. McNeil HP, Gotis-Graham I. Human Mast Cell Subsets–Distinct Functions in Inﬂammation? Inﬂammation Res (2000) 49(1):3–7. doi: 10.1007/PL00012386 96. Karaman S, Detmar M. Mechanisms of Lymphatic Metastasis. J Clin Invest (2014) 124(3):922–8. doi: 10.1172/JCI71606 86. Yu Y, Blokhuis BR, Garssen J, Redegeld FA. Non-IgE Mediated Mast Cell Activation. Eur J Pharmacol (2016) 778:33–43. doi: 10.1016/ j.ejphar.2015.07.017 97. Vieira JM, Norman S, Villa Del Campo C, Cahill TJ, Barnette DN, Gunadasa- Rohling M, et al. The Cardiac Lymphatic System Stimulates Resolution of Inﬂammation Following Myocardial Infarction. J Clin Invest (2018) 128 (8):3402–12. doi: 10.1172/JCI97192 87. Janicki JS, Brower GL, Levick SP. The Emerging Prominence of the Cardiac Mast Cell as a Potent Mediator of Adverse Myocardial Remodeling. Methods Mol Biol (2015) 1220:121–39. doi: 10.1007/978-1-4939-1568-2_8 98. Kologrivova I, Shtatolkina M, Suslova T, Ryabov V. Cells of the Immune System in Cardiac Remodeling: Main Players in Resolution of Inﬂammation and Repair After Myocardial Infarction. Front Immunol (2021) 12:664457. doi: 10.3389/ﬁmmu.2021.664457 88. Shao Z, Nazari M, Guo L, Li SH, Sun J, Liu SM, et al. The Cardiac Repair Beneﬁts of Inﬂammation do Not Persist: Evidence From Mast Cell Implantation. J Cell Mol Med (2015) 19(12):2751–62. doi: 10.1111/ jcmm.12703 99. Verma SK, Garikipati VNS, Krishnamurthy P, Schumacher SM, Grisanti LA, Cimini M, et al. REFERENCES Interleukin-10 Inhibits Bone Marrow Fibroblast Progenitor Cell-Mediated Cardiac Fibrosis in Pressure-Overloaded Myocardium. Circulation (2017) 136(10):940–53. doi: 10.1161/CIRCULATIONAHA. 117.027889 89. Legere SA, Haidl ID, Castonguay MC, Brunt KR, Legare JF, Marshall JS, et al. IncreasedMastCellDensityisAssociatedWithDecreasedFibrosisinHumanAtrial Tissue. J Mol Cell Cardiol (2020) 149:15–26. doi: 10.1016/j.yjmcc.2020.09.001 90. Dewald O, Ren G, Duerr GD, Zoerlein M, Klemm C, Gersch C, et al. Of Mice and Dogs: Species-Speciﬁc Differences in the Inﬂammatory Response Following Myocardial Infarction. Am J Pathol (2004) 164(2):665–77. doi: 10.1016/S0002-9440(10)63154-9 100. Baci D, Bosi A, Parisi L, Buono G, Mortara L, Ambrosio G, et al. Innate Immunity Effector Cells as Inﬂammatory Drivers of Cardiac Fibrosis. Int J Mol Sci (2020) 21(19):7165. doi: 10.3390/ijms21197165 Conﬂict of Interest: 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. 91. Pereira Beserra F, Fernando Sérgio Gushiken L, Fernanda Hussni M, Helena Pellizzon C. Regulatory Mechanisms and Chemical Signaling of Mediators Involved in the Inﬂammatory Phase of Cutaneous Wound Healing. Wound Healing - Curr Perspect (2019) 5–22. doi: 10.5772/intechopen.81731 Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. 92. Yang Z, Zingarelli B, Szabó C. Crucial Role of Endogenous Interleukin-10 Production in Myocardial Ischemia/Reperfusion Injury. Circulation. (2000) 101(9):1019–26. doi: 10.1161/01.CIR.101.9.1019 93. Mukai K, Tsai M, Saito H, Galli SJ. Mast Cells as Sources of Cytokines, Chemokines, and Growth Factors. Immunol Rev (2018) 282(1):121–50. doi: 10.1111/imr.12634 Copyright © 2022 Jin, Jiang, Chakrabarti and Su. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 94. Theoharides TC, Tsilioni I, Conti P. Mast Cells May Regulate The Anti- Inﬂammatory Activity of IL-37. Int J Mol Sci (2019) 20(15):3701. doi: 10.3390/ ijms20153701 95. Dieterich LC, Detmar M. Tumor Lymphangiogenesis and New Drug Development. July 2022 \| Volume 13 \| Article 963444 Frontiers in Immunology \| www.frontiersin.org REFERENCES Adv Drug Delivery Rev (2016) 99(Pt B):148–60. doi: 10.1016/ j.addr.2015.12.011 July 2022 \| Volume 13 \| Article 963444 Frontiers in Immunology \| www.frontiersin.org 10
https://openalex.org/W1922878222	https://europepmc.org/articles/pmc4414178?pdf=render	English	null	miR-191 promotes tumorigenesis of human colorectal cancer through targeting C/EBPβ	Oncotarget	2,014	cc-by	9,707	Correspondence to: Xiao-Dong Zhang, e-mail: zhangxd@whu.edu.cn Correspondence to: Xiao-Dong Zhang, e-mail: zhangxd@whu.edu.cn Keywords: microRNA-191, colorectal cancer, apoptosis, C/EBPβ, tumorigenesis y , , p p , / β, g Received: August 28, 2014 Accepted: December 07, 2014 Published: December 27, 2014 Published: December 27, 2014 Received: August 28, 2014 Accepted: December 07, 2014 ABSTRACT MicroRNA-191 (miR-191), a small non-coding RNA, is involved in disease development and cancer diagnosis and prognosis. However, how miR-191 functions in colorectal cancer remains largely unclear. In this study, we show that miR-191 is highly expressed in colon tumor tissues, and that inhibition of miR-191 leads to decreased cell growth, proliferation and tumorigenicity in a xenograft model. Overexpression of miR-191 in colorectal cancer cell lines alters cell cycle progression and cell resistance to 5-Fu induced cell apoptosis. Mechanistic studies demonstrated that miR-191 directly binds to the 3’UTR of the C/EBPβ mRNA and mediates a decrease in the mRNA and protein expression of C/EBPβ. We further showed that C/EBPβ induces growth arrest in a colorectal cancer cell line and that its expression is negatively correlated with the miR-191 level in patient samples. Our findings suggest that miR-191 may be a potential gene therapy target for the treatment of colorectal cancer. Oncotarget, Vol. 6, No.6 Oncotarget, Vol. 6, No.6 www.impactjournals.com/oncotarget/ Xiao-Fei Zhang1, Ke-ke Li1, Lu Gao3, Shang-Ze Li1, Ke Chen4, Jun-Bin Zhang5, Di Wang6, Rong-Fu Tu1, Jin-Xiang Zhang2, Kai-Xiong Tao2, Guobin Wang2,, Xiao-Dong Zhang1, Xiao-Fei Zhang1, Ke-ke Li1, Lu Gao3, Shang-Ze Li1, Ke Chen4, Jun-Bin Zhang5, Di Wang6, Rong-Fu Tu1, Jin-Xiang Zhang2, Kai-Xiong Tao2, Guobin Wang2,, Xiao-Dong Zhang1, 1College of Life Sciences, Wuhan University, Wuhan 430072, PR China 1College of Life Sciences, Wuhan University, Wuhan 430072, PR China 2Department of General Surgery, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430000, PR China Department of General Surgery, Union Hospital, Tongji Medical College, Huazhong University of S Wuhan 430000, PR China 2Department of General Surgery, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430000, PR China 3Institute of Cardiovascular Disease, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430000, PR China 3Institute of Cardiovascular Disease, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430000, PR China 4Institute of Urology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430000, PR China 4Institute of Urology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430000, PR China 5Department of Emergency, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430000, PR China 5Department of Emergency, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430000, PR China 6 Department of Gastrointestinal Surgery, Union Hospital, Tongji Medical College, Huazhong Un Technology, Wuhan 430000, PR China 6Department of Gastrointestinal Surgery, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430000, PR China 6Department of Gastrointestinal Surgery, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430000, PR China These authors have contributed equally to this work These authors have contributed equally to this work www.impactjournals.com/oncotarget [2, 3]. Aberrant miRNA expression has been frequently found in many cancers and multiple miRNA-expression profiles of human tumors uncovered the correlation between miRNA expression patterns and the tumor type and stage [4, 5]. Depending on the mRNA targets that they regulate, miRNAs can act as oncogenes (e.g., miR-17-92) or as tumor suppressor genes (e.g., miR-15a/16-1, let-7 and miR-34) [6–9]. INTRODUCTION [2, 3]. Aberrant miRNA expression has been frequently found in many cancers and multiple miRNA-expression profiles of human tumors uncovered the correlation between miRNA expression patterns and the tumor type and stage [4, 5]. Depending on the mRNA targets that they regulate, miRNAs can act as oncogenes (e.g., miR-17-92) or as tumor suppressor genes (e.g., miR-15a/16-1, let-7 and miR-34) [6–9]. MicroRNAs (miRNAs) are small noncoding RNAs (18~22 nucleotides) that play important regulatory roles in plants and animals by repressing the translation of proteins from mRNAs (messenger RNAs) or by increasing the degradation of mRNAs through binding to their 3’ untranslated regions (3’ UTRs) [1]. Each miRNA regulates the expression of tens or hundreds of proteins and plays critical roles in most biological processes such as cell proliferation, cell survival and inflammatory responses Colorectal cancer (CRC) is the second leading cause of cancer death for both males and females in the United States [10]. Increasing evidence indicates that deregulation www.impactjournals.com/oncotarget Oncotarget 4144 non-tumor colon tissues by way of real-time PCR. Our results demonstrated that miR-191 was up-regulated in the majority of examined tumor tissues, with 10 of 16 (62.5%) tumor tissues displaying a more than 38% increase, which suggested a probable ‘oncomiR’ role of miR-191 in colorectal cancer. Notably, there were three colon cancer tissues displayed a 3-fold down-regulation of miR-191 expression. (Figure 1A) We next examined miR-191 expression in five human colorectal cancer cell lines (HCT116, RKO, HT29, SW480, DLD1 and Lovo) and HEK293T cells by quantitative PCR. miR-191 was expressed in all six cell lines, and HCT116 displayed a higher expression level of miR-191 (Figure 1B). So, we used HCT116 cells as a model to investigate the effect of miR191 on cell growth and proliferation. of miRNAs affects cell growth and development of colorectal cancer [11–13]. Specific miRNAs could serve as useful clinical biomarkers and potential therapeutic targets for colorectal carcinoma [14, 15]. Previous work has demonstrated that miR-191 was deregulated in a wide range of human cancers, including breast cancer [16], hepaotocellular carcinoma [17], thyroid follicular tumors [18] and acute myeloid leukemia [19]; and this deregulation may be associated with clinical stage, patient survival and disease prognosis. Interestingly, miR-191 exerts diverse and often conflict biological effects, which are always cell-type and context specific. INTRODUCTION In hepatocellular carcinoma, miR-191 was identified as an oncogene and its inhibition led to decreased cell proliferation and induced apoptosis in vitro and in vivo [17]. miR-191 functions as an estrogen inducible oncomiR in breast cancer, and mediate enhanced cell proliferation and migration by targeting SATB1 [20]. In contrast, Di Leva G et al. reported that activation of the miR-191/425 cluster reduced proliferation and impaired tumorigenesis in breast cancer cells [21]. In addition, miR-191 reduced growth and cell migration by targeting CDK6 in thyroid follicular tumor [18]. Several studies have provided strong evidence that miR-191 is overexpressed in human CRC [22–24]. Recently, Dong et al. reported that high miR-191 expression was associated with CRC tumor invasion by directly targeting tissue inhibitor of metalloprotease 3 (TIMP3), a pro-apoptotic gene in various cancers and diseases [25]. However, the molecular mechanism by which miR-191 functions in CRC remains largely unknown. Therefore, identification of the effects of miR-191 and its targets in CRC may lead to new perspectives for gene therapy clinical trials. miR-191 promotes the tumorigenic features of colorectal cancer cells To assess the role of miR-191 in the growth of CRC, stable miR-191-expressing cell lines were prepared using lenti-virus-mediated gene transfer, wherein the plemiR-191 and sponge-miR-191 were used as mediators for gain- and loss- of -function studies, respectively. The levels of miR-191 in the stable cell lines were determined by quantitative PCR, and our results demonstrated the effectiveness of this transfection (Figure 2A). Cell viability was measured using CCK8 assays, and we observed that plemiR-191-transfected HCT116 and RKO cells have a viability advantage over time, when compared with cells transfected with the plemiR-control. On the contrary, miR- 191 inhibition by sponge-miR-191 decreased cell viability in HCT116, RKO, HT29 and SW480 cells (Figure 2B, Supplementary Figure 1A and Figure 2A). To examine the effect of miR-191 upon proliferation, the cells were seeded in 6-well plates (500 cells/well) for 14 days. The colony formation assay demonstrated a significant increase in the number of colony-forming units in the miR-191-transfected cells and a decrease in the number of clonies formed from sponge-miR-191-transfected cells (Figure 2C). In the current study, we examined the expression of miR-191 in different human colorectal cancer cell lines and tissues. We showed that miR-191 expression was significantly up-regulated in colon cancer tissues compared to adjacent non-cancerous lung tissues. Sustained miR-191 overexpression was associated with increased viability, cell proliferation and tumorigenicity in vivo, and we further revealed that miR-191 promoted cell resistance to chemotherapeutic agents such as 5-Fu. Mechanistically, we found that miR-191 functioned as an ‘oncomiR’ by directly targeting the tumor suppressor C/ EBPβ and that there was a negative correlation between miR-191 and C/EBPβ expression. In addition, C/EBPβ overexpression partially abolished the effects of miR-191 in CRC cells. ( g ) Next, a BALB/c nude mouse xenograft model was applied to evaluate the effect of miR-191 on tumorigenicity. The plemiR-191/plemiR-control or sponge-miR-191/sponge-control stable cell lines derived from HCT116 cells were subcutaneously injected into either the flank or forelimb armpits of nude mice. The tumor volume was measured regularly and tumor weights were recorded. Compared with the control group, miR- 191-transfected cells revealed an advanced tumor formation and a significantly increase in tumor size and tumor weight. miR-191 inhibition effectively suppressed tumor growth in nude mice, as determined by the retarded tumor growth rate, reduced tumor volume and decreased tumor weight compared with the negative control (Figure 2D, 2E, 2F). www.impactjournals.com/oncotarget miR-191 is upregulated in colon cancers Several reports indicated that miR-191 is up- regulated in human colorectal cancers by using high throughput sequencing [22–24]. Here, miR-191 expression was further analyzed in 16 paired colon and adjacent www.impactjournals.com/oncotarget Oncotarget 4145 Figure 1: miR-191 is up-regulated in colon cancers. Stem-loop RT-PCR analysis of the miR-191 level in tissues and cell lines. (A) Relative miR-191 expression in 16 paired colon cancer tissues and adjacent non-tumor tissues (upper panel). miR-191 expression values were expressed as ratios with U6 snRNA (×10). (lower panel, P = 0.011) The statistical significance was evaluated by paired-samples T test. (B) Relative miR-191 expression in five human colorectal cancer cell lines (HCT116, RKO, HT29, SW480, DLD1 and Lovo) and human embryonic kidney 293T cells. U6 snRNA was used as an internal control. The data represents the means ± SDs. Figure 1: miR-191 is up-regulated in colon cancers. Stem-loop RT-PCR analysis of the miR-191 level in tissues and cell lines. (A) Relative miR-191 expression in 16 paired colon cancer tissues and adjacent non-tumor tissues (upper panel). miR-191 expression values were expressed as ratios with U6 snRNA (×10). (lower panel, P = 0.011) The statistical significance was evaluated by paired-samples T test. (B) Relative miR-191 expression in five human colorectal cancer cell lines (HCT116, RKO, HT29, SW480, DLD1 and Lovo) and human embryonic kidney 293T cells. U6 snRNA was used as an internal control. The data represents the means ± SDs. miR-191 induces the G1-to-S cell-cycle transition Cell cycle progression was controlled by a series of genes, including cell cycle progression regulators and cell cycle inhibitors such as p21, p15 and p16. Here, the mRNA and protein levels of cell cycle-related factors were detected by quantitative PCR and western blotting, respectively. We found that miR-191 induced the expression of CDK4, a key regulator in the G1-to-S cell cycle transition; however, the changes in the cyclin D1 and cyclin E levels were not significant. Notably, the levels of p15, p16, p21, p27 and p57, which are cyclin-dependent kinase inhibitors, were significantly decreased in the plemiR-191-transfected stable cell line (Figure 3D, 3E). In contrast, miR-191 inhibition by sponge-miR-191 led to a decrease in the level of CDK4 and an increase in the levels of p15, p16, p27 and p57 (Figure 3F, 3G). To elucidate the effect of miR-191 on cell cycle regulation, HCT116 cells transfected with miR-191 mimic or inhibitor were subjected to flow cytometry. First, we determined the miR-191 level to verify the effectiveness of the transfection. Quantitative PCR analysis showed that the expression of miR-191 increased by 3.6-fold and decreased by 12.8-fold following miR-191 mimic or inhibitor transfection, respectively. (Figure 3A) Cell cycle analysis showed a distinct decrease in the G1- phase cell population (66.48% vs. 58.62%) and an increase in the S- phase cell population (13.56% vs. 18.93%) in miR- 191-mimic transfectants compared with the mimic control (Figure 3B). On the contrary, the miR-191 inhibitor restrained the G1-to-S cell cycle transition. (Figure 3C) www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget Oncotarget 4146 / t t promotes cell viability and proliferation. (A) Confirmation of the level of miR-191 i RT-PCR; cells transfected with the empty vector were used as a negative control. (B) The cell v by CCK8 assays after transfection of plemiR-191(left, B) and sponge-miR-191(right, B), at 24 olony formation units of plemiR-191- (left, C) and sponge-miR-191-transfected (right, C) stable the quantification of the indicated relative colony-forming units (n = 3, P < 0.05 versus plemiR-c of plemiR-191-/sponge-miR-191-treated mice 23 days after transplantation. (E) Representative im d tumors from injected mice (n = 6 – 8). The data represents the means ± SDs. Figure 2: miR-191 promotes cell viability and proliferation. (A) Confirmation of the level of miR-191 in stably transfected HCT116 cell lines by RT-PCR; cells transfected with the empty vector were used as a negative control. miR-191 induces the G1-to-S cell-cycle transition (B) The cell viability of HCT116 cells was determined by CCK8 assays after transfection of plemiR-191(left, B) and sponge-miR-191(right, B), at 24, 48, 72 96 and 120 hours. (C) Relative colony formation units of plemiR-191- (left, C) and sponge-miR-191-transfected (right, C) stable HCT116 cells. The lower panels indicate the quantification of the indicated relative colony-forming units (n = 3, P < 0.05 versus plemiR-ctrl and sponge-ctrl). (D) The tumor volume of plemiR-191-/sponge-miR-191-treated mice 23 days after transplantation. (E) Representative images and (F) tumor weights of the isolated tumors from injected mice (n = 6 – 8). The data represents the means ± SDs. miR-191 promotes cell resistance to 5-Fu To explore the molecular mechanism responsible for the function of miR-191 in CRC, we used three publicly available databases (TargetScan, picTar and miRanda) to search for predicted direct target genes of miR-191. The predicted targets were arranged according to the binding probability score. The targets with high score and shared by the three databases were chosen. Seven candidate targets of miR-191 (unpublished and we are doing further research) were screened. C/EBPβ was chosen for further analysis, because previous reports have shown that C/ EBPβ is an important regulator of cell growth [26, 27], cell apoptosis [28] and tumorigenicity [29]. miR-191 has conserved binding sites in the 3’UTRs of C/EBPβ of different species (Figure 5A). The wild-type (WT) or mutant 3’ UTR, in which the seed region was mutated to abolish miR-191 binding, was cloned into the psiCheck-2 plasmid (Figure 5B), and a dual-luciferase reporter system was employed to verify whether C/EBPBβ is a direct target of miR-191. The luciferase activity of the reporter containing the WT 3’ UTR of C/EBPβ was decreased in cells transfected with plemiR-191, whereas the activity of the mutant reporter was not significantly altered following plemiR-191 transfection (Figure 5C). 5-Fu is a widely used anticancer drug, but due to its high-dose regimen in the clinic, the drug’s side effects were observed. In our drug screening assay, we found that endogenous miR-191 was modulated by various drugs. Interestingly, although the changes of cell viability were similar when HCT116 cells were treated with different drugs, the alteration of the level of miR-191 was the most apparent in cells treated with 5-Fu (Figure 4A, 4B). We further demonstrated that 5-Fu decreased the endogenous miR-191 level in a dose-dependent manner (Figure 4C). The induction of the pro-apoptotic pathway by 5-Fu is crucial for its anticancer role, so we hypothesized that miR-191 might be involved in the 5-Fu induced cell apoptotic pathway. HCT116 cells were transfected with miR-191 mimic or inhibitor for 24 hours and then treated with 25 μg/ml 5-Fu for another 24 hours, and the cell viability was measured by CCK8. We observed an almost 50% decrease in cell viability when cells were treated with 5-Fu. Notably, when exposed to 5-Fu, cells transfected with the miR-191 mimic exhibited higher cell viability compared to cells transfected with the mimic control. miR-191 promotes cell resistance to 5-Fu The introduction of miR-191 inhibitor led to a significant decrease in cell viability when compared with the inhibitor control (Figure 4D). We further detected the expression of C/EBPβ in stable transfected HCT116 and RKO cells using quantitative PCR and western blotting. The results indicated that the mRNA and protein levels of C/ EBPβ were impaired under conditions of miR-191 overexpression. Conversely, inhibition of miR-191 resulted in the up-regulation of the levels of C/EBPβ (Figure 5D, 5E and Supplementary Figure 1B and Figure 2B). ( g ) To investigate the role of miR-191 in 5-Fu induced cell apoptosis, we performed quantitative PCR and western blotting to detect the mRNA and protein levels of pro-apoptotic Bax and anti-apoptotic Bcl-2, the markers of cell apoptosis. The mRNA level of Bax decreased in cells transfected with the miR-191-mimic and increased by 50% in cells transfected with the miR-191-inhibitor when compared with the control group. In contrast, the Bcl-2 level was higher in cells transfected with the miR- 191-mimic and lower in cells transfected with the miR- 191-inhibitor when compared with the corresponding control group (Figure 4E). The protein levels of Bax and Bcl-2 were consistent with the mRNA levels. Caspase-3 is an important downstream effecter of the cell apoptotic pathway and cleavage of caspase-3 was observed in HCT116 cells exposed to 5-Fu. Overexpression of miR- 191 inhibited the cleavage of caspase-3, and suppression of miR-191 overtly induced the cleavage of caspase-3 (Figure 4F). To determine the correlation between miR- 191expression and the C/EBPβ level, we analyzed the mRNA and protein levels of C/EBPβ in the same set of specimens shown in Figure 1A, and the results showed that the miR-191 level was inversely correlated with C/ EBPβ expression (Figure 5F–5H). These data suggested that C/EBPβ is a direct target of miR-191 in CRC. www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget Oncotarget 4147 Oncota 4148 mpactjournals.com/oncotarget re 3: Effects of miR-191 on the cell cycle distribution of HCT116 cells. (A) RT-PCR analysis confirmed the express 191 in miR-191 mimic-/inhibitior- transfected HCT116 cells; cells transfected with the corresponding control oligos were use ive control. (B) Cell cycle analysis of HCT116 cells transfected with the miR-191 mimic oligo. (C) Cell cycle analysis of HC transfected with the miR-191 inhibitor oligo. (D–G) The mRNA levels of cyclin D1, cyclin E, CDK4, p15, p16, p21, p27 an ) and the protein levels of cyclin D1, CDK4, p21 and p27 (E, G) were assayed in plemiR-ctrl-/plemiR-191-transfected and sp sponge-miR-191-transfected HCT116 cells, respectively. GAPDH served as a loading control. The data represents the means ± Figure 3: Effects of miR-191 on the cell cycle distribution of HCT116 cells. (A) RT-PCR analysis confirmed the expression of miR-191 in miR-191 mimic-/inhibitior- transfected HCT116 cells; cells transfected with the corresponding control oligos were used as a negative control. (B) Cell cycle analysis of HCT116 cells transfected with the miR-191 mimic oligo. (C) Cell cycle analysis of HCT116 cells transfected with the miR-191 inhibitor oligo. (D–G) The mRNA levels of cyclin D1, cyclin E, CDK4, p15, p16, p21, p27 and p57 (D, F) and the protein levels of cyclin D1, CDK4, p21 and p27 (E, G) were assayed in plemiR-ctrl-/plemiR-191-transfected and sponge- ctrl-/sponge-miR-191-transfected HCT116 cells, respectively. GAPDH served as a loading control. The data represents the means ± SDs. www.impactjournals.com/oncotarget Oncotarget 4148 C/EBPβ inhibits cell proliferation and induces cell cycle arrest in CRC cell lines Growth arrest induced by C/EBPβ is highly context specific [30–32]. To investigate the role of C/EBPβ in CRC, we performed loss- and gain-of-function studies for C/EBPβ. LAP2, the best studied and most transcriptionally active isoform of C/EBPβ, was given particular attention for its role of suppress cell cycle progression. [33, 34] HCT116 cells were transfected with pCMV-flag-LAP2 or sh-C/EBPβ-1/-2 and immunoblotting results showed the effectiveness of up-regulation and down-regulation of C/EBPβ by flag-LAP2 or sh-C/EBPβ-1 (Figure 6A, 6B). Ectopic expression of C/EBPβ diminished the role of In addition, HCT116 cells transfected with the miR-191-mimic or miR-191-inhibitor were subjected to flow cytometry analysis. The apoptotic rate of cells transfected with miR-191-mimic was approximately 5.45%, which was significantly lower than of cells transfected with mimic-control oligo, which showed 10.1% apoptotic rates when exposed to 5-Fu. Compared with the inhibitor control, inhibition of miR-191 clearly increased the apoptotic rates of HCT116 cells treated with 5-Fu (Figure 4G). Taken together, miR-191 decreased the sensitivity of HCT116 cells to 5-Fu. www.impactjournals.com/oncotarget Oncotarget 4149 4: The involvement of miR-191 in the 5-Fu-induced cell apoptotic pathway in HCT116 cells. HCT116 cells w with various commonly-used chemotherapeutic drugs, including 5-fluorouracil (5-Fu, 10 μg/ml), indomethacin (Indo, 10 μg/m n (Cis, 10 μg/ml) and etoposide (Eto, 2 μM). (A) Cell viability was assessed by CCK8 assays. (B) RT-PCR analysis of the rela on of miR-191. (C) 5-Fu down-regulated miR-191 in a dose-dependent manner. HCT116 cells were treated with the indica ration of 5-Fu for 48 hours, total mRNA was isolated, and the miR-191 level was analyzed by RT-PCR (P < 0.05 versus DM cells). HCT116 cells were transfected with the indicated oligos for 24 hours and then treated with 5-Fu (25 μg/ml) for ano s. (D) Cell viability was assessed by CCK8 assays (P < 0.05 versus 5-Fu treated mimic-ctrl/inhibitor-ctrl cells). (E) The mR f Bax and Bcl-2 was determined by RT-PCR (P < 0.05 versus DMSO treated mimic-ctrl/inhibitor-ctrl cells, #P < 0.05 versus 5 mimic-ctrl/inhibitor-ctrl cells). (F) The protein levels of Bax, Bcl-2, caspase-3 and cleaved-caspase-3 were detected by wes analysis in HCT116 cells transfected with mimic-ctrl/miR-191-mimic (left panel) and inhibitor-ctrl/miR-191-inhibitor (right pa MSO or 5-Fu treatment. (mc, mimic ctrl; m, miR-191-mimic; ic, inhibitor ctrl; i, miR-191-inhibitor) (G) Cell apoptosis analysi ted cells after treatment. GAPDH served as a loading control. The data represents the means ± SDs. C/EBPβ inhibits cell proliferation and induces cell cycle arrest in CRC cell lines vement of miR-191 in the 5-Fu-induced cell apoptotic pathway in HCT116 cells. H mmonly-used chemotherapeutic drugs, including 5-fluorouracil (5-Fu, 10 μg/ml), indomethacin ) and etoposide (Eto, 2 μM). (A) Cell viability was assessed by CCK8 assays. (B) RT-PCR ana . (C) 5-Fu down-regulated miR-191 in a dose-dependent manner. HCT116 cells were treated or 48 hours, total mRNA was isolated, and the miR-191 level was analyzed by RT-PCR (P < 0 cells were transfected with the indicated oligos for 24 hours and then treated with 5-Fu (25 ility was assessed by CCK8 assays (P < 0.05 versus 5-Fu treated mimic-ctrl/inhibitor-ctrl cell was determined by RT-PCR (P < 0.05 versus DMSO treated mimic-ctrl/inhibitor-ctrl cells, #P bitor-ctrl cells). (F) The protein levels of Bax, Bcl-2, caspase-3 and cleaved-caspase-3 were d Figure 4: The involvement of miR-191 in the 5-Fu-induced cell apoptotic pathway in HCT116 cells. HCT116 cells were treated with various commonly-used chemotherapeutic drugs, including 5-fluorouracil (5-Fu, 10 μg/ml), indomethacin (Indo, 10 μg/ml), cisplatin (Cis, 10 μg/ml) and etoposide (Eto, 2 μM). (A) Cell viability was assessed by CCK8 assays. (B) RT-PCR analysis of the relative expression of miR-191. (C) 5-Fu down-regulated miR-191 in a dose-dependent manner. HCT116 cells were treated with the indicated concentration of 5-Fu for 48 hours, total mRNA was isolated, and the miR-191 level was analyzed by RT-PCR (P < 0.05 versus DMSO treated cells). HCT116 cells were transfected with the indicated oligos for 24 hours and then treated with 5-Fu (25 μg/ml) for another 24 hours. (D) Cell viability was assessed by CCK8 assays (P < 0.05 versus 5-Fu treated mimic-ctrl/inhibitor-ctrl cells). (E) The mRNA levels of Bax and Bcl-2 was determined by RT-PCR (P < 0.05 versus DMSO treated mimic-ctrl/inhibitor-ctrl cells, #P < 0.05 versus 5-Fu treated mimic-ctrl/inhibitor-ctrl cells). (F) The protein levels of Bax, Bcl-2, caspase-3 and cleaved-caspase-3 were detected by western blotting analysis in HCT116 cells transfected with mimic-ctrl/miR-191-mimic (left panel) and inhibitor-ctrl/miR-191-inhibitor (right panel) after DMSO or 5-Fu treatment. (mc, mimic ctrl; m, miR-191-mimic; ic, inhibitor ctrl; i, miR-191-inhibitor) (G) Cell apoptosis analysis of transfected cells after treatment. GAPDH served as a loading control. The data represents the means ± SDs. www.impactjournals.com/oncotarget Oncotarget 4150 iR-191 down-regulates the expression of C/EBPβ. (A) The predicted binding sites of miR-191 in the 3’ UTRs of ecies. (B) Schematic description of wild type (WT) and mutated 3’ UTRs of the C/EBPβ mRNA. C/EBPβ inhibits cell proliferation and induces cell cycle arrest in CRC cell lines The WT and mutated 00 bp) were cloned into the psiCHECK2 vector. (C) Luciferase analysis was used to detect the reporter activity. HCT tes were co-transfected with 200 ng plemiR/plemiR-191 or psiCHECK2-WT 3’ UTR/psiCHECK2-mutated 3’ UTR (10 hen subjected to luciferase assays according to the Materials and Methods. (D-E) miR-191 inhibits the mRNA and prote HCT116 cells. HCT116 cells were transfected with the miR-191 mimic/mimic control (D) or miR-191 inhibitor/inhibito urs and total mRNA and protein were extracted and analyzed by RT-PCR and western blotting analysis, respectively. ading control. (F) RT-PCR analysis of the relative expression of C/EBPβ in 16 colon cancer tissues. (G) Western blotting 16 colon cancer tissues. GAPDH served as a loading control. (H) Inverse correlation between miR-191 and C/EBPβ . miR-191 level was normalized to the U6 level, and the C/EBPβ level was normalized to the GAPDH level. Statistical d using Person’s correlation coefficient analysis. The data represents the means ± SDs. Figure 5: miR-191 down-regulates the expression of C/EBPβ. (A) The predicted binding sites of miR-191 in the 3’ UTRs of C/EBPβ of different species. (B) Schematic description of wild type (WT) and mutated 3’ UTRs of the C/EBPβ mRNA. The WT and mutated 3’ UTR sequences (~400 bp) were cloned into the psiCHECK2 vector. (C) Luciferase analysis was used to detect the reporter activity. HCT116 cells in 24 well plates were co-transfected with 200 ng plemiR/plemiR-191 or psiCHECK2-WT 3’ UTR/psiCHECK2-mutated 3’ UTR (100 ng) for 48 hours and then subjected to luciferase assays according to the Materials and Methods. (D-E) miR-191 inhibits the mRNA and protein levels of C/EBPβ in HCT116 cells. HCT116 cells were transfected with the miR-191 mimic/mimic control (D) or miR-191 inhibitor/inhibitor control (E) for 48 hours and total mRNA and protein were extracted and analyzed by RT-PCR and western blotting analysis, respectively. GAPDH served as a loading control. (F) RT-PCR analysis of the relative expression of C/EBPβ in 16 colon cancer tissues. (G) Western blotting analysis of C/EBPβ in 16 colon cancer tissues. GAPDH served as a loading control. (H) Inverse correlation between miR-191 and C/EBPβ in colon cancer tissues. miR-191 level was normalized to the U6 level, and the C/EBPβ level was normalized to the GAPDH level. Statistical analysis was performed using Person’s correlation coefficient analysis. The data represents the means ± SDs. www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget Oncotarget 4151 Figure 6: C/EBPβ is involved in miR-191 induced cell growth advantage. HCT116 cells were transfected with flag-LAP2 or sh-C/EBPβ-1/2 for 48 hours and then analyzed by western blotting. Endogenous C/EBPβ expression was efficiently induced or repressed by transfection with flag-LAP2 (A) or sh-C/EBPβ-1/2. Right panel: Quantification of C/EBPβ levels (B). GAPDH served as a loading control. (C) Ectopic expression of C/EBPβ abrogated the cell growth advantage caused by miR-191. HCT116 cells were transfected with the indicated expression vector for 48 hours and then analyzed by CCK8 assays (P < 0.05 versus plemiR-191+pCMV-flag-LAP2 transfected cells). (D) The effects of C/EBPβ on the expression of p15, p16, p21, p27 and p57. Quantification of the mRNA levels of p15, p16, p21, p27 and p57 in HCT116 cells transfected with the indicated expression vectors for 48 hours. (P < 0.05 versus pCMV-ctrl, #P < 0.05 versus scramble) (E) Cell cycle analysis of HCT116 cells transfected with the miR-191 mimic/mimic control or miR-191 inhibitor/ inhibitor control for 48 hours. The data represents the means ± SDs. Figure 6: C/EBPβ is involved in miR-191 induced cell growth advantage. HCT116 cells were transfected with flag-LAP2 or sh-C/EBPβ-1/2 for 48 hours and then analyzed by western blotting. Endogenous C/EBPβ expression was efficiently induced or repressed by transfection with flag-LAP2 (A) or sh-C/EBPβ-1/2. Right panel: Quantification of C/EBPβ levels (B). GAPDH served as a loading control. (C) Ectopic expression of C/EBPβ abrogated the cell growth advantage caused by miR-191. HCT116 cells were transfected with the indicated expression vector for 48 hours and then analyzed by CCK8 assays (P < 0.05 versus plemiR-191+pCMV-flag-LAP2 transfected cells). (D) The effects of C/EBPβ on the expression of p15, p16, p21, p27 and p57. Quantification of the mRNA levels of p15, p16, p21, p27 and p57 in HCT116 cells transfected with the indicated expression vectors for 48 hours. (*P < 0.05 versus pCMV-ctrl, #P < 0.05 versus scramble) (E) Cell cycle analysis of HCT116 cells transfected with the miR-191 mimic/mimic control or miR-191 inhibitor/ inhibitor control for 48 hours. The data represents the means ± SDs. Figure 6: C/EBPβ is involved in miR-191 induced cell growth advantage. HCT116 cells were DISCUSSION miR-191 on cell viability (Figure 6C), and LAP2 transfection induced the expression of p16, p15 and p57. Although the alterations in the p21 and p27 levels were unconspicuous, the inhibition of C/EBPβ by shRNA led to decreased expression of p16, p15 and p57 (Figure 6D). In addition, the effect of C/EBPβ on the cell cycle transition was analyzed by flow cytometry. C/EBPβ overexpression led to G1 cell cycle arrest, while inhibition of endogenous C/EBPβ by sh-C/EBPβ-1 resulted in an increase in S phase (Figure 6E). These results further indicated that C/EBPβ is a direct target of miR-191. In the present report, we showed that up-regulation of miR-191 was a frequent event in colon cancers and that this up-regulation increased cell viability and promoted cell proliferation and tumorigenicity of HCT116 cells. Anticancer reagents, such as 5-Fu and etoposide, inhibited the expression of miR-191 and sustained up-regulation of miR1-191 reduced cell susceptibility to 5-Fu. C/EBPβ was identified as a target of miR-191 and ectopic expression of C/EBPβ impaired the effect of miR-191 overexpression www.impactjournals.com/oncotarget Oncotarget 4152 on CRC by activating expression of downstream genes, including p15, p16 and p57 (Figure 7). Our results suggest a crucial role of miR-191 in tumorigenesis of CRC and provide a therapeutic approach for CRC treatment. miR-191-mediated down-regulation of CDK6 led to reduced cell proliferation [16, 20]. So, we determined the mRNA levels of CDK6 in HCT116 cells transfected with plemiR-191 or sponge-miR-191 and found that miR-191 also decreased the level of CDK6 in HCT116 cells (Supplementary Figure 3A). We considered that the tumor promoting role of miR-191 in CRC was primarily due to the suppression of cell cycle inhibitors. Further investigation is needed to establish the involvement of miR-191 in the alteration of the balance between cyclin/ CDK complexes and CKIs. p p pp Previous reports have shown that miR-191 is deregulated in various cancers and diseases and depending on the tissue and context conditions, miR- 191 may exhibit specific functions [16–20]. Our study demonstrated a higher expression of miR-191 in colon cancer patients, which is consistent with previous reports [22]. Ectopic overexpression of miR-191 promoted cell proliferation in HCT116 cells and tumorigenicity in vivo in a nude mouse model. Furthermore, miR-191 mimic transfected cells displayed more DNA synthesis, while miR-191 inhibition delayed the G1-to-S cell cycle transition. Cyclin/cyclin-dependent kinases (Cyclin-CDK complexes) and CDK inhibitors (CKIs) coordinately control cell-cycle progression [35]. DISCUSSION Our data showed that the level of CDK4, the key regulator of cell cycle progression, was altered when miR-191 was up- or down- regulated. Notably, the levels of p16, p15 and p57, which are important cell cycle inhibitors, were decreased when miR-191 was overexpressed. Di Leva et al. reported that cell cycle regulatory proteins such as CDK6 and CCND2 were established as targets of miR-191 in thyroid carcinoma and aggressive breast cancer and showed that p It has been reported that miR-191 is a regulator of cell fate that inhibits cell apoptosis in colorectal carcinoma and hepatocellular carcinoma and miR-191 can be induced by various stimuli such as hypoxia and serum starvation [17, 20, 25]. 5-Fu is a widely used antimetabolite drug used for the treatment of digestive tumors, particularly for colorectal cancer. Our study revealed that endogenous miR- 191 was regulated by 5-Fu in a dose-dependent manner, which is consistent with results of a previous study [36]. Our further study showed that miR-191 was involved in the 5-Fu-induced cell apoptotic pathway and that miR-191 overexpression increased cell resistance to 5-Fu-induced cell apoptosis. These results were partially consistent with the previous reports which showed that miR-191 inhibition induced cell apoptosis by upregulation of pro-apoptotic Figure 7: Schematic model of miR-191-mediated promotion of tumorigencity of colorectal cancer cells. In colorectal cancer cells, various anticancer drugs regulate the expression of miR-191. miR-191 causes a number of genes to respond and adapt. On one hand, miR-191 induces the expression of CDK4 to increase cell growth. On the other hand, miR-191 suppresses the level of C/EBPβ, a tumor suppressor gene functions as a transcriptional activator of p15, p16 and p57, which are the key regulators of cell cycle and cell survival. As a result, miR-191 induces the cell cycle progression and tumor cell resistance to various stimuli. Figure 7: Schematic model of miR-191-mediated promotion of tumorigencity of colorectal cancer cells. In colorectal cancer cells, various anticancer drugs regulate the expression of miR-191. miR-191 causes a number of genes to respond and adapt. On one hand, miR-191 induces the expression of CDK4 to increase cell growth. On the other hand, miR-191 suppresses the level of C/EBPβ, a tumor suppressor gene functions as a transcriptional activator of p15, p16 and p57, which are the key regulators of cell cycle and cell survival. DISCUSSION As a result, miR-191 induces the cell cycle progression and tumor cell resistance to various stimuli. www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget Oncotarget 4153 TIMP3. Interestingly, we found that the level of miR- 191 was down-regulated by etoposide, a commonly used anticancer agent and an inducer of DNA double-strand breaks and cell apoptosis. Xi et al. reported that miR- 191 was up-regulated in tumors with a p53 deletion [22]; therefore, it is possible that inhibition of miR-191 renders the cells more susceptible to p53-dependent stress responses. Considering the high mutation rate of p53 in human CRC, it would be intresteing to see whether introduction of miR- 191 into CRC cells can modulate p53 activity and uncover the mechanism of miR-191 action. Here, we conclude that miR-191 is an anti-apoptotic gene in a colorectal carcinoma, in 5-Fu-induced cell apoptotic model. of patients limitations in our study, cancer or disease specific mouse models, such as tissue-specific transgenic or knockout mice, would be extremely helpful to verify miR-191 therapeutic potential. Cell culture and reagents The human embryonic kidney 293T cell line (HEK293T) and six human colorectal cancer cell lines (HCT116, RKO, HT29, SW480, DLD1 and Lovo) were purchased from American Type Culture Collection (ATCC, Manassas, VA). The cell lines HCT116, RKO, HT29, SW480, DLD1 and Lovo were grown in McCoy’s 5A medium (AppliChem, A1324, 9050, Darmstadt, Germany) supplemented with 10% fetal bovine serum (FBS, Gibco, Carlsbad, CA, USA) and 1% penicillin– streptomycin (HyClone, Logan, Utah, USA) at 37°C, in 5% CO2. The HEK293T cell line was maintained in Dulbecco’s modified Eagle’s medium (DMEM, HyClone, Logan, Utah, USA) supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin at 37°C in 5% CO2. 5-Fluorouracil (5-Fu), indomethacin (Indo), cisplatin (Cis) and etoposide (Eto) were obtained from Sigma Chemical Co. (St. Louis, MO, USA) Tissue specimens CRC and adjacent non-tumor colon tissues were collected from patients undergoing resection of CRC. Tissue samples were immediately frozen in liquid nitrogen and stored until total RNAs or proteins were extracted. Informed consent was obtained at the Tongji Medical College Huazhong University of Science & Technology in Wuhan, China. None of the patients received chemotherapy prior to colectomy. All patients were unrelated ethnic Han Chinese who lived in Southeast China; the tissues were characterized using immunochemistry methods. The relevant characteristics of the studied subjects are shown in Supplementary Table 1. p p CCAAT/enhancer-binding protein β (C/EBPβ) is a member of the C/EBP family of transcription factors [37, 38]. C/EBPβ has been shown to regulate cell proliferation, differentiation and cell apoptosis in a variety of cellular systems [39–41]. In breast cancer, C/ EBPβ was identified as a direct target of miR-155, which is an oncogenic microRNA involved in cell proliferation and the inflammation response [42]. In this study, we demonstrated that miR-191 could directly bind to the 3’UTR region of C/EBPβ and decrease its mRNA and protein levels, which suggests a novel signaling mechanism wherein miR-191 functions as an ‘onco-miR’. Forced expression of exogenous C/EBPβ can induce cell apoptosis in various malignant cells and growth arrest induced by C/EBPβ is highly context-specific [40, 43, 44]. Overexpression of C/EBPβ arrests cells at or near the G1-S boundary in HepG2 and epidermal keratinocytes [30, 45]. While, in several cases, C/EBPβ displays growth-promoting activity [31, 32]. p21 and p57 were reported as regulated targets of C/EBPβ in MCF7 and chondrocytes [43, 46]. Our findings provide evidence that C/EBPβ expression was low in colon tumor patients and that it functioned as a tumor suppressor in HCT116 cells by suppressing of cell viability and inducing of cell apoptosis, which is consistent with results of a previous study showing that administration of full-length wild- type C/EBPβ significantly suppressed the growth of colon tumors by inducing tumor-cell apoptosis in a nude mouse model [29]. Furthermore, the expression of miR-191 was negatively related to the level of C/EBPβ in 16-paired tumor and non-tumor patient samples. Collectively, this evidence strongly suggests that C/EBPβ is regulated by miR-191 in human colorectal cancer, although the signal transduction is still not well understood. Tumorigenicity assays in nude mice All experimental procedures involving animals were performed in accordance with the Guide for the Care and Use of Laboratory Animals (NIH publications Nos. 80–23, revised 1996) and were approved by the Animal Care and Use Committee of Wuhan University. Female athymic nude mice were bought from (HFKBio, Peking, China). PlemiR-191 or pSicoR-sponges-miR-191 transfected HCT116 cells and the corresponding control cells (5 × 106) were suspended in 200 μl PBS and then injected subcutaneously into either the flank or forelimb armpits of the same female, 5−6-week-old BALB/c athymic nude mouse (HFKBio, Peking, China). Seven to ten female mice were included in the experiments. Tumor growth was examined every three days, and the tumors were removed and weighed 23 days after injection. The tumor volume (V) was monitored using measuring the length (L) and width (W) of the tumor with calipers and was calculated using the formula V = (L × W2) × 0.5. Lentiviral transduction According to the Addgene protocol (http://www. addgene.org/tools/protocols/plko/), human 293T cells cultured in a 6 cm dish were cotransfected with the lentiviral packaging vectors 1 μg of psPAX2 (Addgene, #12260), 2 μg of pMD2.G (Addgene, #12259), and 3 μg of the indicated plasmids to produce lentiviral particles. Sixty hours after transfection, the media were collected and filtered with a 0.45 μm filter (Nalgene, Rochester, NY, USA). Viral supernatant mixed with 5 μg/mL polybrene (Sigma Chemical Co., St. Louis, Mo.) was used to infect HCT116 cells for 12 hours, and was then replaced with fresh media. Cell cycle and cell apoptosis analysis HCT116 cells were transfected with the indicated oligos or plasmids in 6-well plates. After 48 hours of transfection, the cells were collected and washed with PBS. For cell cycle analysis, the cells were fixed with 70% ethanol overnight at −20°C, washed with PBS, resuspended with 400 μl PBS and then incubated with 100 μg/ml RNaseA (Sigma) for 30 minutes at 37°C and with 50 μg/ml propidium iodide (PI) (Sigma, St Louis, MO) for another 10 minutes. After incubation, the cells were subjected to DNA content analysis using a FACSCalibur (Beckman Coulter, Fullerton, CA) and the results were analyzed with the Summit v4.3 software. Apoptosis was evaluated by Annexin V-FITC (BD Biosciences, 556419) and PI (BD Biosciences, 556463) staining according to the manufacturer’s protocol, followed by flow cytometry analysis. In brief, HCT116 Plasmids Has-miR-191-5p mimic, has-miR-191-5p inhibitor and the corresponding mimic/inhibitor control oligo were purchased from Guangzhou RiboBio Co., LTD. (Guangzhou, China) The human miRNA expression vectors plemiR and plemiR-191 were kindly provided by Dr. Yendamuri, Sai (Department of Thoracic Surgery, Roswell Park Cancer Institute, Buffalo, USA) [47]. An miR-191 sponge(sponge-miR-191) was constructed using the methods described by Margaret S Ebert [48]. The sponge sequence, encoding ten repeats of reverse complementary sequence of mature miR-191 was synthesized by Sunny Biotechnology Co. (Shanghai, The present study provides evidence indicating that inhibition of miR-191 suppresses the proliferation of colorectal cells and tumorigenicity in vivo, and that miR-191-down-regulated cells are more sensitive to 5-Fu-induced cell apoptosis. Due to the regulation of C/ EBPβ and downstream signaling by miR-191, the specific inhibition of miR-191 or when in combination with 5-Fu may be useful for the treatment of human colorectal cancer. However, considering technology and the number www.impactjournals.com/oncotarget Oncotarget 4154 China) and then the sequence was ligated to the pSicoR vector after digestion by BamH1 and Xho1. The human C/EBPβ expression vector pCMV-flag-LAP2 was purchased from Addgene (ID 15738). The shRNA sequences against C/EBPβ were as follows: sequence 1, CCCGTGGTGTTATTTAAAGAA; sequence 2, CCTGCCTTTAAATCCATGGAA, and the scramble shRNA sequence was AATTCTCCGAACGTGTCACGT. shRNAs were ligated into the pLKO.1 puro vector (Addgene, # 8453). All plasmids were sequenced to verify the inserted targets. HCT116 cells were transfected with the oligos (100 nmol/L) or indicated plasmids, using Lipofectamine2000TM (Invitrogen, Carlsbad, CA, USA) in OptiMEM medium according to the manufacturer’s instructions. The OptiMEM media (Invitrogen, Carlsbad, CA, USA) was replaced with cell growth media after 8 hours of transfection. cells (1 × 105) were collected, washed with ice-cold PBS and resuspended in 100 μl binding buffer. Then, 2 μl of Annexin V-FITC and 5 μl of PI were added to the cells, the cells were incubated for 15 minutes at RT in the dark, and an additional 400 μl of binding buffer was added to the reaction prior to analysis. The results were analyzed with the Summit v4.3 software. Cell viability and colony formation assay Cell Counting Kit-8 (CCK8) assays were performed to analyze cellular proliferation and activity. HCT116 cells transfected with plemiR-191 or pSicoR-sponge- miR-191 (100 μL) were seeded at a density of 2 × 103 cells/well in 96-well plates. At 24, 48, 72, 96 and 120 hours, 10 μL of CCK8 solution (Dojindo Laboratories, Kumamoto, Japan) was added to each well of the plate and incubated for 1 hour. Cell viability was then determined using a spectrophotometer set (ELx800, BioTek, USA) at a wavelength of 450 nm. For colony formation assays, plemiR-191- or pSicoR-sponge-miR-191-transfected HCT116 cells were cultured in a 6-well plate at 500 cells per well and grown for 10-14 days. After fixation by 4% paraformaldehyde (Electron Microscopy Sciences 16% Paraformaldehyde Cat.15700, diluted into PBS) for 30 minutes, the colonies were stained with 0.1% Crystal Violet for 15 minutes and washed. The colonies were then photographed and counted using the Image J software. RNA quantification Total RNA was extracted using TRIzol reagent (Takara Biotechnology, Dalian, China) according to the manufacturer’s protocol, and the concentrations of the RNA were determined with a NanoDrop instrument (NanoDrop Technologies). Complementary DNA was prepared with the use of a First Strand cDNA Synthesis Kit (Roche Diagnostics, Mannheim, Germany) and stem- loop RT for mature miRNAs was performed according to the manufacturer’s protocol. (Roche) All reagents for stem- loop RT were obtained from Roche (Roche Diagnostics, Mannheim, Germany) and RiboBio (Guangzhou, China). The U6 snRNA was used as an internal control. Real- time PCR was performed using the FastStart Universal SYBR Green Master protocol (ROCHE, 04913850001) with the ABI PRISM 7500 system (Applied Biosystems, Forster City, Calif). For microRNA PCR, the reactions were incubated in a 96-well plate at 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. All reactions were run in triplicate, and the relative gene expression was calculated using the comparative threshold cycle (Ct) method (relative gene expression = 2−(ΔCtsample-ΔCtcontrol)). The following primers were used for qPCR: C/EBPβ sense, 5′-TTCAAGCAGCTGCCCGAGCC-3′; C/ EBPβ antisense, 5′-GCCAAGTGCCCCAGTGCCAA-3′ (Refer to Gibellini et al [49]); cyclin D1 sense, 5′-CCGTCCATGCGGAAGATC-3′; cyclin D1 antisense, 5′-CCTCCTCCTCGCACTTCTGT-3′; cyclin E1 sense, 5′-TTTCTTGAGCAACACCCT-3′; cyclin E1 antisense, 5′-GTCACATACGCAAACTGG-3′; CDK4 sense, 5′-GAAACTCTGAAGCCGACCAG-3′; CDK4 antisense, 5′-AGGCAGAGATTCGCTTGTGT-3′; CDK6 sense, 5′-CCGTGGATCTCTGGAGTGTT-3′; CDK6 antisense, 5′-CTCAATTGGTTGGGCAGATT-3′; p15 sense, 5′-ATGCGCGAGGAGAACAAG-3′; p15 antisense, 5′-CTCCCGAAACGGTTGACTC-3′; p16 sense, 5′-CTTCCTGGACACGCTGGT-3′; p16 antisense, 5′-ATCTATGCGGGCATGGTTAC-3′; p21 sense, 5′-GAGCGATGGAACTTCGACTT-3′; p21 antisense, 5′-CAGGTCCACATGGTCTTCCT-3′; p27 sense, 5′-GGTCTGCAAGTGGATGATGA-3′; p27 antisense, 5′-ATAATGTTGCAGCCCAGCAG-3′; p57 sense, 5′-CACGATGGAGCGTCTTGTC-3′; p57 antisense, 5′-CCTGCTGGAAGTCGTAATCC-3′; Bax sense, 5′-GAGGATGATTGCCGCCGTGGACA-3′; Bax antisense, 5′-GGTGGGGGAGGAGGCTTGAGG-3′; Bcl-2 sense, 5′-ATGTGTGTGGAGAGCGTCAACC-3′; Bcl-2 antisense, Total RNA was extracted using TRIzol reagent (Takara Biotechnology, Dalian, China) according to the manufacturer’s protocol, and the concentrations of the RNA were determined with a NanoDrop instrument (NanoDrop Technologies). Complementary DNA was prepared with the use of a First Strand cDNA Synthesis Kit (Roche Diagnostics, Mannheim, Germany) and stem- loop RT for mature miRNAs was performed according to the manufacturer’s protocol. (Roche) All reagents for stem- loop RT were obtained from Roche (Roche Diagnostics, Mannheim, Germany) and RiboBio (Guangzhou, China). The U6 snRNA was used as an internal control. Real- time PCR was performed using the FastStart Universal SYBR Green Master protocol (ROCHE, 04913850001) with the ABI PRISM 7500 system (Applied Biosystems, Forster City, Calif). Reporter assays For the C/EBPβ 3’UTR reporter assays, experiments were performed on the HCT116 cell line. The 3’UTR fragment of C/EBPβ was amplified from human genomic DNA and subcloned into the XhoI and NotI sites of the psiCHECK2 vector (Promega, Madison, WI).The sequences of the primers used for PCR amplification were as follows: forward primer: 5′-CCCTCGAG GCAACCCACGTGTAACTGTC-3′ and reverse primer 5′-ATAAGAATGCGGCCGCCACCC AACCACCAAAACCTC-3′. Mutagenesis of the miR- 191 seed sequence was performed using the following www.impactjournals.com/oncotarget Oncotarget 4155 Statistical analysis The SPSS 19 software was used for statistical analysis, and the data are presented as the means ± SDs. Differences between groups were assessed by one-way ANOVA, and statistical significance was determined by Student’s t-test in some experiments. Differences with P values of less than 0.05 were considered significant. 5′-TGAGCAGAGTCTTCAGAGACAGCC-3′; GAPDH sense, 5′-GAAGGTGAAGGTCGGAGTC-3′; and GAPDH antisense, 5′-GAAGATGGTGATGGGATT TC-3′. primers: forward primer: 5′-AAGGGAATCTTTCT GCACTCAAGCAT-3′ and reverse primer 5′-ATGC TTGAGTGCAGAAAGATTCCCTT-3′. HCT116 cells were transfected with the reporter plasmid (100 ng) and plemiR-191/plemiR-ctrl (200 ng), and reporter assays were performed 24 hours after transfection using the Dual Luciferase kit (Promega, Madison, WI). RNA quantification For microRNA PCR, the reactions were incubated in a 96-well plate at 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. All reactions were run in triplicate, and the relative gene expression was calculated using the comparative threshold cycle (Ct) method (relative gene expression = 2−(ΔCtsample-ΔCtcontrol)). The following primers were used for qPCR: C/EBPβ sense, 5′-TTCAAGCAGCTGCCCGAGCC-3′; C/ EBPβ antisense, 5′-GCCAAGTGCCCCAGTGCCAA-3′ (Refer to Gibellini et al [49]); cyclin D1 sense, 5′-CCGTCCATGCGGAAGATC-3′; cyclin D1 antisense, 5′-CCTCCTCCTCGCACTTCTGT-3′; cyclin E1 sense, 5′-TTTCTTGAGCAACACCCT-3′; cyclin E1 antisense, 5′-GTCACATACGCAAACTGG-3′; CDK4 sense, 5′-GAAACTCTGAAGCCGACCAG-3′; CDK4 antisense, 5′-AGGCAGAGATTCGCTTGTGT-3′; CDK6 sense, 5′-CCGTGGATCTCTGGAGTGTT-3′; CDK6 antisense, 5′-CTCAATTGGTTGGGCAGATT-3′; p15 sense, 5′-ATGCGCGAGGAGAACAAG-3′; p15 antisense, 5′-CTCCCGAAACGGTTGACTC-3′; p16 sense, 5′-CTTCCTGGACACGCTGGT-3′; p16 antisense, 5′-ATCTATGCGGGCATGGTTAC-3′; p21 sense, 5′-GAGCGATGGAACTTCGACTT-3′; p21 antisense, 5′-CAGGTCCACATGGTCTTCCT-3′; p27 sense, 5′-GGTCTGCAAGTGGATGATGA-3′; p27 antisense, 5′-ATAATGTTGCAGCCCAGCAG-3′; p57 sense, 5′-CACGATGGAGCGTCTTGTC-3′; p57 antisense, 5′-CCTGCTGGAAGTCGTAATCC-3′; Bax sense, 5′-GAGGATGATTGCCGCCGTGGACA-3′; Bax antisense, 5′-GGTGGGGGAGGAGGCTTGAGG-3′; Bcl-2 sense, 5′-ATGTGTGTGGAGAGCGTCAACC-3′; Bcl-2 antisense, Western blotting analysis The cells were lysed in RIPA buffer (1% v/v NP40, 0.5% w/v sodiumdeoxycholate, 0.1% w/v SDS) containing complete protease inhibitors (Roche Applied Sciences) in an ice bath for 30 minutes, and then centrifuged at 13,000 × g for 30 minutes at 4°C. The protein concentration was then determined using the Pierce® BCA Protein Assay Kit (Pierce, 23225). Equal amounts of protein (30-150 ug) were separated by 12% SDS-PAGE and transferred to PVDF membranes (PVDF, Millipore, cat# IPVH00010, Merck KgaA, Darmstadt, Germany). The membranes were blocked with 5% non-fat dry milk in TBST (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Tween-20) for 1 hour at room temperature and then incubated with primary antibody on a rocking platform overnight at 4°C, followed by incubation with a horseradish peroxidase-conjugated secondary antibody. Protein bands were visualized by incubating the membranes with the SuperSignal chemiluminescence kit (Merck Millipore). Finally, Bio- Rad’s ChemiDoc XRS+ imaging system was used for signal detection. Protein expression levels were normalized to GAPDH as a loading control. The antibodies used in this study were as follows: Rabbit monoclonal antibodies against Bax (#5023), Bcl-2 (#4223), caspase 3 (#9665), cleaved-caspase-3(#9661), p21 (#2947) and GAPDH (#5174) and; rabbit polyclonal antibody against C/EBPβ (#3082) were obtained from Cell Signaling Technology (Beverly, MA). Rabbit anti-p27 (sc-528), anti-CDK4 (sc- 260), and anti-cyclin D (sc-753) were from Santa Cruz Biotechnology. Peroxidase-conjugated Goat Anti-Mouse IgG (H+L (115-035-003 and Peroxidase-conjugated Goat Anti-Rabbit IgG (H+L) (111-035-003) were purchased from Jackson ImmunoResearch. (Jackson, West Baltimore Pike West Grove, PA, USA) REFERENCES 15. Hrasovec S, Glavac D. MicroRNAs as Novel Biomarkers in Colorectal Cancer. Front Genet. 2012; 3:180. 1. Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell. 2004; 116:281–97. 1. Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell. 2004; 116:281–97. 16. Hui AB, Shi W, Boutros PC, Miller N, Pintilie M, Fyles T, McCready D, Wong D, Gerster K, Waldron L, Jurisica I, Penn LZ, Liu FF. Robust global micro-RNA profiling with formalin-fixed paraffin-embedded breast cancer tissues. Lab Invest. 2009; 89:597–606. 2. Baek D, Villen J, Shin C, Camargo FD, Gygi SP, Bartel DP. The impact of microRNAs on protein output. Nature. 2008; 455:64–71. 3. Carthew RW. Gene regulation by microRNAs. Curr Opin Genet Dev. 2006; 16:203–8. 17. Elyakim E, Sitbon E, Faerman A, Tabak S, Montia E, Belanis L, Dov A, Marcusson EG, Bennett CF, Chajut A, Cohen D, Yerushalmi N. hsa-miR-191 Is a Candidate Oncogene Target for Hepatocellular Carcinoma Therapy. Cancer Res. 2010; 70:8077–8087. 4. Calin GA, Croce CM. MicroRNA signatures in human can cers. Nat Rev Cancer. 2006; 6:857–66. 5. Lu J, Getz G, Miska EA, Alvarez-Saavedra E, Lamb J, Peck D, Sweet-Cordero A, Ebert BL, Mak RH, Ferrando AA, Downing JR, Jacks T, Horvitz HR, et al. MicroRNA expression profiles classify human cancers. Nature. 2005; 435:834–8. 18. Colamaio M, Borbone E, Russo L, Bianco M, Federico A, Califano D, Chiappetta G, Pallante P, Troncone G, Battista S, Fusco A. miR-191 Down-Regulation Plays a Role in Thyroid Follicular Tumors through CDK6 Targeting. J Clin Endocr Metab. 2011; 96:1915–1924. 6. Lee YS, Dutta A. The tumor suppressor microRNA let-7 represses the HMGA2 oncogene. Genes Dev. 2007; 21:1025–30. 19. Garzon R, Volinia S, Liu CG, Fernandez-Cymering C, Palumbo T, Pichiorri F, Fabbri M, Coombes K, Alder H, Nakamura T, Flomenberg N, Marcucci G, Calin GA, et al. MicroRNA signatures associated with cytogenetics and prognosis in acute myeloid leukemia. Blood. 2008; 111:3183–9. 7. Bonci D, Coppola V, Musumeci M, Addario A, Giuffrida R, Memeo L, D’Urso L, Pagliuca A, Biffoni M, Labbaye C, Bartucci M, Muto G, Peschle C, et al. The miR- 15a-miR-16-1 cluster controls prostate cancer by targeting multiple oncogenic activities. Nat Med. 2008; 14:1271–7. 20. Nagpal N, Ahmad HM, Molparia B, Kulshreshtha R. MicroRNA-191, an estrogen-responsive microRNA, func tions as an oncogenic regulator in human breast cancer. Carcinogenesis. 2013; 34:1889–1899. 8. ACKNOWLEDGEMENTS This work was supported by grants from the National Basic Research Program of China (2011CB944404), the National Natural Science Foundation of China (81270306, 81470375), the National Science and Technology Support Project (2012BAI39B02, 2012BAI39B03, www.impactjournals.com/oncotarget Oncotarget 4156 13. Hara T, Jones MF, Subramanian M, Li XL, Ou O, Zhu Y, Yang Y, Wakefield LM, Hussain SP, Gaedcke J, Ried T, Luo J, Caplen NJ, et al. Selective targeting of KRAS-mutant cells by miR-126 through repression of multiple genes essential for the survival of KRAS-mutant cells. Oncotarget. 2014; 5:7635–50. 2014BAI02B01), the Trans-Century Training Programme Foundation for the Talents by the State Education Commission (NCET-10-0655), and Fundamental Research Funds for the Central Universities (204275771). CONFLICT OF INTEREST 14. Perilli L, Vicentini C, Agostini M, Pizzini S, Pizzi M, D’Angelo E, Bortoluzzi S, Mandruzzato S, Mammano E, Rugge M, Nitti D, Scarpa A, Fassan M, et al. Circulating miR-182 is a biomarker of colorectal adenocarcinoma pro gression. Oncotarget. 2014; 5:6611–9. The authors declare no conflicts of interest. REFERENCES He L, Thomson JM, Hemann MT, Hernando-Monge E, Mu D, Goodson S, Powers S, Cordon-Cardo C, Lowe SW, Hannon GJ, Hammond SM. A microRNA polycistron as a potential human oncogene. Nature. 2005; 435:828–833. 21. Di Leva G, Piovan C, Gasparini P, Ngankeu A, Taccioli C, Briskin D, Cheung DG, Bolon B, Anderlucci L, Alder H, Nuovo G, Li M, Iorio MV, et al. Estrogen mediated- activation of miR-191/425 cluster modulates tumorigenicity of breast cancer cells depending on estrogen receptor status. PLoS Genet. 2013; 9:e1003311. 9. He L, He X, Lim LP, de Stanchina E, Xuan Z, Liang Y, Xue W, Zender L, Magnus J, Ridzon D, Jackson AL, Linsley PS, Chen C, et al. A microRNA component of the p53 tumour suppressor network. Nature. 2007; 447:1130–4. 10. Siegel R, Naishadham D, Jemal A. Cancer statistics. Ca-Cancer J Clin. 2013; 63:11–30. 22. Xi Y, Formentini A, Chien M, Weir DB, Russo JJ, Ju J, Kornmann M, Ju J. Prognostic Values of microRNAs in Colorectal Cancer. Biomarker insights. 2006; 2:113–121. 11. Zhang J, Fei B, Wang Q, Song M, Yin Y, Zhang B, Ni S, Guo W, Bian Z, Quan C, Liu Z, Wang Y, Yu J, et al. MicroRNA-638 inhibits cell proliferation, invasion and regulates cell cycle by targeting tetraspanin 1 in human colorectal carcinoma. Oncotarget. 2014; [Epub ahead of print]. 23. Volinia S, Calin GA, Liu CG, Ambs S, Cimmino A, Petrocca F, Visone R, Iorio M, Roldo C, Ferracin M, Prueitt RL, Yanaihara N, Lanza G, et al. A microRNA expression signature of human solid tumors defines cancer gene targets. Proc Natl Acad Sci U S A. 2006; 103:2257–61. 12. Fang L, Li H, Wang L, Hu J, Jin T, Wang J, Yang BB. MicroRNA-17-5p promotes chemotherapeutic drug resis tance and tumour metastasis of colorectal cancer by repress ing PTEN expression. Oncotarget. 2014; 5:2974–87. 24. Cummins JM, He Y, Leary RJ, Pagliarini R, Diaz LA Jr, Sjoblom T, Barad O, Bentwich Z, Szafranska AE, Labourier E, Raymond CK, Roberts BS, Juhl H, et al. The www.impactjournals.com/oncotarget Oncotarget 4157 colorectal microRNAome. Proc Natl Acad Sci U S A. 2006; 103:3687–92. IL-6 expression (NF-IL6) is a member of a C/EBP family. EMBO J. 1990; 9:1897–906. 38. Ramji DP, Foka P. CCAAT/enhancer-binding proteins: structure, function and regulation. Biochem J. 2002; 365:561–75. 25. Qin S, Zhu Y, Ai F, Li Y, Bai B, Yao W, Dong L. REFERENCES MicroRNA-191 correlates with poor prognosis of colorectal carcinoma and plays multiple roles by targeting tissue inhibitor of metalloprotease 3. Neoplasma. 2014; 61:27–34. 39. Grimm SL, Rosen JM. The role of C/EBPbeta in mammary gland development and breast cancer. J Mammary Gland Biol. 2003; 8:191–204. 26. Luedde T, Duderstadt M, Streetz KL, Tacke F, Kubicka S, Manns MP, Trautwein C. C/EBP beta isoforms LIP and LAP modulate progression of the cell cycle in the regenerat ing mouse liver. Hepatology. 2004; 40:356–65. 40. Lekstrom-Himes J, Xanthopoulos KG. Biological role of the CCAAT/enhancer-binding protein family of transcrip tion factors. J Biol Chem. 1998; 273:28545–8. 27. Pal R, Janz M, Galson DL, Gries M, Li S, Johrens K, Anagnostopoulos I, Dorken B, Mapara MY, Borghesi L, Kardava L, Roodman GD, Milcarek C, et al. C/EBPbeta regulates transcription factors critical for proliferation and survival of multiple myeloma cells. Blood. 2009; 114:3890–8. 41. Umek RM, Friedman AD, McKnight SL. CCAAT-enhancer binding protein: a component of a differentiation switch. Science. 1991; 251:288–92. 42. Johansson J, Berg T, Kurzejamska E, Pang MF, Tabor V, Jansson M, Roswall P, Pietras K, Sund M, Religa P, Fuxe J. MiR-155-mediated loss of C/EBPbeta shifts the TGF-beta response from growth inhibition to epithelial-mesenchymal transition, invasion and metastasis in breast cancer. Oncogene. 2013; 32:5614–24. 28. Shimizu Y, Kishimoto T, Ohtsuka M, Kimura F, Shimizu H, Yoshidome H, Miyazaki M. CCAAT/enhancer binding protein-beta promotes the survival of intravascular rat pan creatic tumor cells via antiapoptotic effects. Cancer Sci. 2007; 98:1706–13. 43. Pal P, Lochab S, Kanaujiya J, Sanyal S, Trivedi AK. Ectopic expression of hC/EBPs in breast tumor cells induces apoptosis. Mol Cell Biochem. 2010; 337:111–8. 29. Sun L, Fu BB, Liu DG. Systemic delivery of full-length C/ EBP beta/liposome complex suppresses growth of human colon cancer in nude mice. Cell Res. 2005; 15:770–6. 44. Zhu MS, Liu DG, Cheng HQ, Xu XY, Li ZP. Expression of exogenous NF-IL6 induces apoptosis in Sp2/0-Ag14 myeloma cells. DNA Cell Biol. 1997; 16:127–35. 30. Buck M, Turler H, Chojkier M. LAP (NF-IL-6), a tissue- specific transcriptional activator, is an inhibitor of hepatoma cell proliferation. EMBO J. 1994; 13:851–60. 45. Zhu S, Yoon K, Sterneck E, Johnson PF, Smart RC. CCAAT/ enhancer binding protein-beta is a mediator of keratinocyte survival and skin tumorigenesis involving oncogenic Ras sig naling. Proc Natl Acad Sci U S A. 2002; 99:207–12. 31. Bundy LM, Sealy L. REFERENCES CCAAT/enhancer binding protein beta (C/EBPbeta)-2 transforms normal mammary epithelial cells and induces epithelial to mesenchymal transition in culture. Oncogene. 2003; 22:869–83. 46. Hirata M, Kugimiya F, Fukai A, Ohba S, Kawamura N, Ogasawara T, Kawasaki Y, Saito T, Yano F, Ikeda T, Nakamura K, Chung UI, Kawaguchi H. C/EBPbeta Promotes transition from proliferation to hypertrophic dif ferentiation of chondrocytes through transactivation of p57. PloS one. 2009; 4:e4543. 32. Wessells J, Yakar S, Johnson PF. Critical prosurvival roles for C/EBP beta and insulin-like growth factor I in macro phage tumor cells. Mol Cell Biol. 2004; 24:3238–50. 33. Calkhoven CF, Muller C, Leutz A. Translational control of C/EBPalpha and C/EBPbeta isoform expression. Genes Dev. 2000; 14:1920–32. 47. Patnaik SK, Kannisto E, Yendamuri S. Overexpression of microRNA miR-30a or miR-191 in A549 lung cancer or BEAS-2B normal lung cell lines does not alter phenotype. PloS one. 2010; 5:e9219. 34. Nerlov C. Transcriptional and translational control of C/ EBPs: the case for "deep" genetics to understand physi ological function. Bioessays. 2010; 32:680–6. 48. Ebert MS, Neilson JR, Sharp PA. MicroRNA sponges: competitive inhibitors of small RNAs in mammalian cells. Nat Methods. 2007; 4:721–6. 35. Besson A, Dowdy SF, Roberts JM. CDK inhibitors: cell cycle regulators and beyond. Dev Cell. 2008; 14:159–69. 36. Zhou J, Zhou Y, Yin B, Hao W, Zhao L, Ju W, Bai C. 5-Fluorouracil and oxaliplatin modify the expression pro files of microRNAs in human colon cancer cells in vitro. Oncol Rep. 2010; 23:121–8. 49. Gibellini D, Alviano F, Miserocchi A, Tazzari PL, Ricci F, Clo A, Morini S, Borderi M, Viale P, Pasquinelli G, Pagliaro P, Bagnara GP, Re MC. HIV-1 and recombinant gp120 affect the survival and differentiation of human ves sel wall-derived mesenchymal stem cells. Retrovirology. 2011; 8:40. 37. Akira S, Isshiki H, Sugita T, Tanabe O, Kinoshita S, Nishio Y, Nakajima T, Hirano T, Kishimoto T. A nuclear factor for www.impactjournals.com/oncotarget Oncotarget 4158
https://openalex.org/W4293677294	https://bmcendocrdisord.biomedcentral.com/counter/pdf/10.1186/s12902-022-01131-2	English	null	Coping strategies and their association with diabetes specific distress, depression and diabetes self-care among people living with diabetes in Zambia	BMC endocrine disorders	2,022	cc-by	9,219	© The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, 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 licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativeco mmons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Abstract Objectives: Utilising coping strategies to reduce and manage the intensity of negative and distressing emotions caused by diabetes is essential. However, little is known about the use of coping strategies among people living with diabetes in Sub-Saharan African countries like Zambia. This study investigates coping strategies used by people with diabetes in Zambia and how these are associated with diabetes-specific emotional distress, depression and diabetes self-care. Methods: Cross-sectional data from 157 people with diabetes aged between 12 and 68 years were collected. Of the 157, 59% were people with type 1 diabetes and 37% with type 2 diabetes. About 4% had missing information in their record but had either type 1 or type 2 diabetes. Coping styles were measured using the Brief Version of the Coping Orientation to Problems Experienced (Brief COPE), diabetes specific-distress using the Problem Areas in Diabetes, depression using the Major Depression Inventory and self-care using the Diabetes Self-Care scale. Results: Data showed that adaptive coping strategies such as religious coping, acceptance among others, were the most frequently used coping strategies among Zambian individuals with diabetes. Maladaptive coping strategies e.g., self-blame and self-distraction were related to increased diabetes specific-distress and depression. Emotional support was related to better diabetes self-care, while self-blame was related to poor diabetes self-care. Conclusion: There is need to help individuals with diabetes identify adaptive strategies that work best for them in order to improve their quality of life. Keywords: Coping strategies, Diabetes specific- distress, Depressive symptoms, Diabetes self-care, Type 1 and Type 2 diabetes blood glucose monitoring and insulin adjustments and administration, engaging in physical activities, consid- eration of nutrition, managing sleep duration and blood glucose level. In addition to the medical, social, financial demands of optimal daily self-management, emotional problems are common and contribute to the overall bur- den of living with diabetes [4, 5]. Therefore, the goal for the treatment of diabetes is to prevent acute and chronic complications while preserving a better quality of life and Coping strategies and their association with diabetes specific distress, depression and diabetes self‑care among people living with diabetes in Zambia Given Hapunda* Given Hapunda* Background Diabetes self-care can be very complex, challenging and stressful [1, 2], especially in developing countries where health care and formal support systems are still under- developed [3]. Daily self-management involves frequent Correspondence: given.hapunda@gmail.com; given.hapunda@unza.zm Department of Psychology, School of Humanities and Social Sciences, University of Zambia, P.O Box 32379, Lusaka, Zambia Department of Psychology, School of Humanities and Social Sciences, University of Zambia, P.O Box 32379, Lusaka, Zambia © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, 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 licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativeco mmons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Hapunda BMC Endocrine Disorders (2022) 22:215 https://doi.org/10.1186/s12902-022-01131-2 Open Access . Open Access T d i di ib An adaptive coping strategy is a technique that an individual uses to help him or her to adjust adequately or appropriately to the situation or stressor that requires one to manage, and the opposite of this is what is known as maladaptive [14]. Adaptive coping strategies such as acceptance and active coping can help maintain good health outcomes, such as glycaemic control [15]. In con- trast, maladaptive strategies such as wishful thinking and avoidant can affect metabolic control and psychosocial outcomes such as quality of life and depressive symptoms [15–17]. Adaptive and maladaptive coping strategies can either be problem-focused or emotional-focused. Problem-focused coping refers to efforts directed toward rational management of a problem, and it is aimed at changing the situation causing distress whereas emotion- focused coping pertains to efforts to reduce emotional distress caused by the stressful event and manage or reg- ulate emotions that might accompany or result from the stressor. Behavioural coping strategies are used by a per- son to manage a stressful event or situation by modifying his or her actions [18].i A stressor’s effect on a person is based on that per- son’s feeling of threat, vulnerability, and ability to cope than the stressful event itself [24]. How one manages the stressor can be adaptive or maladaptive. Coping is mala- daptive when an individual applies techniques that are not adequate or appropriate to the situation or stressor that requires one to manage [14]. Greater engagement in maladaptive coping is associated with anxiety, depressive symptoms and poorer quality of life [25, 26]. Evidence from qualitative data among Zambian adolescents with type 1 diabetes showed young people used maladaptive coping strategies such as avoidance. Specifically, evidence showed that young people avoiding injecting themselves with insulin was common [27]. The problem-focused and the emotional- focused coping strategies are the two common types of personal coping strategies. The prob- lem-focused coping strategies are aimed at changing the source of stress, while emotional focused are oriented towards managing the emotions that accompany the per- ceived stressor. Both are effective in making the stressed individual feel better but are not equally effective in managing stress. In diabetes, individuals with emotion- focused, social-support focused and problem-focused coping have higher levels of positive diabetes self-care Diabetes specific psychosocial challenges are preva- lent in people with type 1 and type 2 diabetes [16, 19]. . Open Access T d i di ib Hapunda BMC Endocrine Disorders (2022) 22:215 Page 2 of 11 Page 2 of 11 psychological well-being [6]. The burden that comes with diabetes management is often stressing. Therefore, it is not surprising that diabetes management is associated with diabetes specific-distress. [7–10] Diabetes specific- distress is defined as emotional distress associated with the ongoing worries, burdens and concerns that occur when managing a demanding chronic disease like diabe- tes over time [11]. anxiety and disease-related distress are known to be problematic for people with type 1 and type 2 diabe- tes [16]. In Zambia, levels of diabetes-specific distress (33.8 ± 27.2) and severe depressive symptoms (10.8%) have been reported among type 1 and type 2 diabetes in Zambia [21]. On its own, diabetes can be an unpredict- able and stressful disease [16]. Data from a 14-country study with about 10 coun- tries sampled from low- and middle-income countries (LMICs), of which two (Uganda and Kenya) were from SSA, showed that in many of these LMICs, diabetes care was not comprehensive. Most services lacked appropriate identification and care processes for psychological and psychiatric problems [22]. This is because, in developing countries, the medical care value chain does not include mental health specialists due to a shortage of such exper- tise. This is despite the evidence that psychosocial issues contribute to the development of diabetes, especially type 2 and that diabetes itself contributes to the devel- opment of psychosocial problems in individuals with diabetes. Thus, if service care providers simultaneously care for psychological well-being and medical outcomes, their patients’ outcomes will also get better [23]. Despite this evidence in Sub Saharan African, little is known on the use of coping strategies in individuals with diabetes. Thus, coping strategies remain an important study area, especially in developing countries with unique social-cul- tural factors, suboptimal diabetes care and management systems. Diabetes is a complex disease to manage. Comorbid complications and conditions such as hypoglycaemia, hyperglycaemia, neuropathy, nephropathy, and retin- opathy are also common in people with diabetes. Hence, there is need for people with diabetes to find optimal ways to manage the condition. [4, 12]. Such conditions negatively affect the quality of life of diabetes people. Already, evidence in Zambia shows that healthy controls (22.42 vs 18.58) had better scores on life satisfaction than young people with type 1 diabetes [13]. To help people with diabetes cope, adaptive coping strategies are impor- tant. . Open Access T d i di ib For instance, the 2018 clinical consensus guidelines on paediatric diabetes highlight that children and ado- lescents with type 1 diabetes experience depression, diabetes-specific distress, stress, cognitive and school performance challenges, eating disorders, low general and diabetes-specific quality of life compared to their healthy counterparts [20]. Evidence from a systematic review also indicated that anxiety, depression, stress and diabetes-specific distress are the key influential psycho- social factors that determine the psychological wellbe- ing of people living with type 2 diabetes [16]. Further, Hapunda BMC Endocrine Disorders (2022) 22:215 Page 3 of 11 version of the COPE developed by Carver, Scheiver & Weinbraub [29]. The 28 items self-report scale assesses coping styles or strategies on two main dimensions – problems focused and emotional focused coping styles. The scale consists of 14 domains/subscales (self-distrac- tion, active coping, denial, substance use, use of emo- tional support, use of instrumental support, behavioural disengagement, venting, positive reframing, planning, humour, acceptance, religion, self-blame) of two items each. Participants respond to statements on a four Likert scale (1 = I have not been doing this, to 4 = I have been doing this). Higher scores indicate greater use of a par- ticular coping strategy. The full scale has a good Cron- bach’ alpha of 0.84 [29]. In the current study, Cronbach alphas were 0.81, and Lambda 2 was 0.82. The two-item based 14 sub-scales have previously been documented to have low reliabilities [30, 31]. In the current study, inter- nal consistency was low, ranging for most scales 0.34–71. Alpha for subscales was as follows: self-distraction = 0.42; active coping = . 56; denial = 0.52; substance use = 0.71; use of emotional support 0.46; use of instrumental sup- port = 0.55; behavioural disengagement = 0.34; vent- ing = 0.40; positive reframing = 0.50; planning = 0.54; humour = 54; acceptance = 0.48; religion = 0.59; self- blame = 0.34. Nevertheless, the overarching copying style i.e., Problem focused, emotional focused and avoidant reliabilities are acceptable [32] and it is widely used for clinical practice.i activities while individuals with avoidance-focused coping had negative diabetes self-care activities [15]. Although some emotional and behavioural strategies are considered maladaptive, in situations where the stressor is unavoidable, such as diabetes self-care, maladaptive strategies, whether emotional or behavioural, can be an option [28]. Methods Study aim Th d This study aimed to identify which coping strategies are used by Zambian with type 1 and type 2 diabetes and examine how these strategies are associated with diabe- tes-specific distress, depression, and diabetes self-care activities. Design Diabetes-specific-distress was assessed using the Prob- lem Areas in Diabetes (PAID). The PAID is a 20-item self-report measure used to assess diabetes-specific distress, including a range of feelings such as diabetes- related anger, fear, depression, worry and guilt. Items can be responded to on a scale from 0 (not a problem) to 4 (serious problem). An overall score for the PAID can be calculated by adding all of the item scores and multiply- ing them by 1.25, which gives a total score ranging from 0–100. Higher scores indicate more diabetes specific- distress. Reported Cronbach’s alphas for the PAID ranges from 0.84 to 0.96 [33, 34]. In the current study the alpha was 0.88 (Lambda2 = 0.89). A cross-sectional study was conducted among outpa- tients with type 1 or 2 diabetes mellitus. Participants were from four major hospitals in Zambia, with different age groups and socioeconomic backgrounds. . Open Access T d i di ib For instance, distraction and religion as cop- ing strategies do not remove the stressor but contribute to higher wellbeing and positive emotions [28]. Still, it can be helpful in diabetes as it empowers an individual and can lead to finding meaning and purpose in disor- der. In light of the Brief Coping Orientation to Problems Experienced (Brief-COPE), the primary tool of this study, planning, active coping, positive reframing, acceptance, are among those considered adaptive coping strategies. Self-distraction, denial, behavioural disengagement, venting, humour, self-blame and substance use are exam- ples of maladaptive coping strategies [29]. Participants and sitesh The study participants were individuals with type 1 and type 2 diabetes recruited from four major hospitals in Zambia, namely Lusaka, Kitwe, Ndola and Livingstone. The sample included both adolescents according to WHO definition (Adolescence is the phase of life between child- hood and adulthood, from ages 10 to 19) and adults. Pur- posive sampling was used to recruit respondents as long as they met the inclusion criteria of being at least 12 years or older, having been diagnosed with diabetes for at least six months, and using oral medication or insulin therapy. The questionnaires were administered during the routine appointments individuals with type 1 and 2 diabetes had with their respective health care providers. Data were collected using researcher assisted questionnaires from these four city hospitals over one year. Diabetes Self-care was assessed using the Self-Care Inventory. The 13 item Self-Care Inventory (SCI) is a self-report measure used to assess people’ perceptions of their adherence to diabetes self-care recommendations over the past month. Individuals rate themselves on a 5-point Likert scale that reflects how well they followed recommendations for self-care during the past month (i.e., 1 = “never do it” to 5 = “always do this as recom- mended, without fail”). Higher scores indicate more opti- mal diabetes self-care [35]. Cronbach’s alpha for the SCI was 0.84 (Lambda2 = 0.85) for type 1 diabetes and 0.85 (lambda2 = 0.86) for type 2 diabetes. Instruments Generic coping strategies (not diabetes-specific) were assessed using the brief COPE. The brief COPE is a short Hapunda BMC Endocrine Disorders (2022) 22:215 Page 4 of 11 Hapunda BMC Endocrine Disorders (2022) 22:215 Table 1 Demographic and clinical characteristics of 157 participants with type 1 and type 2 diabetes Depression was assessed using the Major Depression Inventory (MDI). The MDI is a 12-item self-report ques- tionnaire used to assess depression. Items of the MDI ask individuals to rate how long in the past two weeks each of the depressive symptoms was present on a six-point scale ranging from 0 “not at all” to 5 ”all time. There- fore, a higher score indicates the presence of depressive symptoms. It can be used as an instrument measuring the severity of depression with a range from 0—60. In previous studies, the MDI had excellent internal consist- ency with Cronbach alphas ranging from 0.89 to 0.94 [36, 37]. In the current study Cronbach alpha was 0.80 (Lambda2 = 0.81). pa t c pa ts t type a d type d abetes Sex, n (%) Females 80 (51%) Age, mean (SD) 39 ± 17 Age range 12–68 years Location of patients Lusaka 48 (31%) Kitwe 60 (38%) Ndola 35 (22%) Livingstone 14 (9%) Age Category n (%) Adolescents 42 (27%) Adults 115 (73%) Educational levels n (%) Adolescents (42) 5-7th Grade (Primary school) 14 (31%) 8-12th Grade (Secondary school) 16 (38%) Missing 14 (31%) Adults (115) Primary education 10 (9%) Secondary education 29 (25%) Tertiary education 22 (19%) Missing 54 (47%) Marital status (Adults/115) n (%) Single 6 (5%) Married 80 (70%) Missing 29 (25%) Type of diabetes Type 1 93(59) Type 2 58 (37) Missing (either type 1or 2) 6 (4) BMI mean (SD) 25 (5) kg/m2 Males 25 (5) kg/m2 Females 26 (5) kg/m2 Adolescents 22 (4) kg/m2 Adults 27 (5) kg/m2 Data analysis Descriptive statistics involving means (standard devia- tion) and frequency percentages were computed to iden- tify the coping strategies frequently used by people in this study. Coping strategies used 50% or more were consid- ered frequently (commonly) used. An independent t-test was computed to test whether the use of coping strategies differed between males versus females, type 1 versus type 2 diabetes, adolescents versus adults. Further, Pearson Product correlation was used to examine the association between coping strategies, diabetes-specific—distress, depression and diabetes self-care. All analysis was done in IBM SPSS version 23. Statistical significance was set at p < 0.05. Results Data were collected from a total of 157 individuals with diabetes from four main hospitals in Zambia, namely Lusaka 48 (31%), Kitwe 60 (38%), Ndola 35 (22%) and Livingstone 14 (9%). One hundred and fifty-seven (157) respondents, of which 93 (59%) had type 1 diabetes mel- litus, 80 (51%) females, and 42 (27%) were adolescents. The average age of the respondents was 39 ± 17 years. See Table 1 for the demographic characteristics of the study sample. The most frequently used coping strategies among Zambian individuals with diabetes are presented in Table 2. In general, adaptive coping strategies (67% reli- gious coping; 58% acceptance, 57% seeking instrumental support and 56% using active coping) were commonly used among Zambian individuals with diabetes. For details on strategies applied by males and females with type 1 and type 2, see Table 3. the use of religion between adults [70% (6.00 ± 1.90)] and adolescents [53% (5.18 ± 2.50)], p < 0.05. The other coping strategy that was statistically significant was the use of self-blame with [11% (2.75 ± 2.53)] of adoles- cents vs [5% (1.96 ± 1.97)], p < 0.05 of adults reporting using it. Use of positive reframing was also significantly different between adolescents [18% (3.45 ± 2.19)] vs [40% (4.33 ± 2.30)], p < 0.05. Overall, a significant dif- ference in the use of emotional strategies emerged between individuals with type 1 diabetes (25.60 ± 8.03) and individuals with type 2 diabetes (28.09 ± 6.30), p < 0.05. Further, there was a significant difference in the use of behavioural strategies between adolescents Although there were differences in the use of cop- ing strategies between type 1 and type 2 diabetes indi- viduals, adolescents and adults, and females and males, most of these differences were not statistically signifi- cant. Results The only coping strategy that was significant was Hapunda BMC Endocrine Disorders (2022) 22:215 Page 5 of 11 Table 2 Proportion of individuals with diabetes using each of the fourteen coping strategies Proportion ≥ 50% were considered common coping strategies among diabetes individuals Coping strategy Type of Diabetes Age Category Sex All types Type 1 Type 2 Adolescents Adults Males Females Adaptive coping strategies Active coping 84(56%) 47(53%) 34(62%) 15(38%) 69(63%) 43(61%) 40(52%) Emotional use 70(46%) 36(40%) 31(53%) 16(41%) 54(48%) 40(53%) 30(39%) Instrumental use 87(57%) 47(52%) 36(63%) 20(50%) 67(60%) 40(53%) 47(62%) Positive reframing 52(34%) 28(32%) 21(36%) 7(18%) 45(40%) 29(38%) 13(31%) Planning 69(47%) 36(42%) 30(53%) 10(29%) 59(52%) 38(69%) 30(40%) Acceptance 85(58%) 50(59%) 41(56%) 23(64%) 62(56%) 43(60%) 41(56%) Religion 101(66%) 56(62%) 41(71%) 21(53%) 80(70%) 49(64%) 51(66%) Maladaptive coping strategies Self-blame 12(8%) 9 (10%) 6(10%) 6(11%) 6(5%) 3(4%) 8(10%) Self-distraction 49(33%) 28(32%) 21(37%) 12(31%) 37(34%) 23(31%) 26(34%) Denial 27(18%) 13(14%) 13(22%) 5(12%) 22(19%) 12(16%) 5(19%) Substance use 5(3%) 3(3%) 2(3%) 2(5%) 3((3%) 4(5%) 1(1%) Behavioural disengagement 27(18%) 16(18%) 11(20%) 5(13%) 22(20%) 17(24%) 10(13%) Venting 27(19%) 19(22%) 8(15%) 11(28%) 16(15%) 9(13%) 18(25%) Humour 14(10%) 8(9%) 6(11%) 6(15%) 8(8%) 3(4%) 11(15%) Table 2 Proportion of individuals with diabetes using each of the fourteen coping strategies (24.21 ± 9.61) vs adults with diabetes (29.13 ± 9.08), p < 0.01. For details see Table 3. of diabetes specific-distress (r = 0.430, p < 0.001), and high use of emotional focused strategies was also associ- ated with increase report of specific-distress (r = 0.374, P < 0.001) and with increased report of depressive symp- toms (r = 0.475, p < 0.001). p Overall, Zambian individuals with diabetes reported using more behavioural focused strategies 20.08 ± 9.28 than emotional strategies 26.60 ± 7.46 (t = 2.25, df = 106, p < 0.05). Among adult people, behavioural strategies were more used 29.13 ± 9.08 compared to emotional strategies 26.67 ± 7.11 (t = 2.88, df 82, < 0.01) while for adolescents, emotional strategies were more likely to be used (25.29 ± 8.76) compared to behavioural strategies (24.21 ± 9.61). However, these differences were not sta- tistically significant, t = -0.78, df, 23, p > 0.05. Regarding the type of diabetes, behavioural strategies were likely to be used 27.10 ± 9.47 compared to emotional 25.60 ± 8.03 among individuals with type 1 diabetes, although this dif- ference was not statistically significant t = 1.66, df = 57, p > 0.05. Results For type 2 diabetes, behaviour strategies were more likely to be used 29.20 ± 9.52 compared to emo- tional strategies 28.09 ± 6.30, although this difference was also not statistically significant. For males, behaviour strategies were more likely to be used 29.79 ± 10.48 com- pared to emotional strategies 26.25 ± 8.06 and this differ- ence was statistically significant t = 2.16, df = 52, p < 0.01. For females, there was no difference in the use of behav- ioural strategies (26.30 ± 7.88) and emotional strategies 26.48 ± 6.94), t = -0.20, df = 53, p > 0.05. Adaptive coping strategies including active coping (r = 0.247, p < 0.01) and positive reframing (r = 0.278, p < 0.01) were associated with increased specific-distress while use of emotional support was only associated with improved diabetes self-care (r = 0.263, p < 0.01). p p Maladaptive coping strategies including self-dis- traction (r = 0.370, p < 0.001); behavioural disengage- ment (r = 0.448, p < 0.001); denial (r = 0.402, p < 0.001); substance use (r = 0.193, p < 0.05); humour (r = 0.173, p < 0.05); self-blame (r = 0.238, p < 0.01) and; venting (r = 0.326, p < 0.001) were associated with increased specific-distress. In addition, Denial (r = 0.191, p < 0.05); substance use (r = 0.177, p < 0.05); self-blame (r = 0.230, p < 0.01) and; venting (r = 0.206, p < 0.05) were associ- ated with increased depressive symptoms. Furthermore, use of self-blame (r = -0.171, p < 0.05) was associated with poor diabetes self-care. Discussionh This study aimed to identify coping strategies mostly used among individuals with diabetes in Zambia and examine how these strategies are associated with diabe- tes specific-distress, depression, and diabetes self-care. This study showed that the most frequently used coping strategies were religion, active coping, instrumental use Correlations between coping strategies, diabetes spe- cific-distress, depression and diabetes self-care are pre- sented in Table 4. Discussionh Results showed that use of behavioural focused strategies was associated with increased report Hapunda BMC Endocrine Disorders (2022) 22:215 Page 6 of 11 Table 3 Mean differences of participant characteristic on different coping strategies Coping strategy Type of Diabetes P Value Age Category P Value Sex P Value T1D T2D Adolescents Adults Males Females Brief cope M 53.40 57.18 0.21 50.65 55.95 0.14 56.15 53.40 0.35 SD 15.96 13.77 16.65 14.39 17.09 12.47 Behavioural strategies M 27.10 29.20 0.25 24.21 29.13 0.01 29.79 26.30 0.42 SD 9.47 9.52 9.61 9.08 10.48 7.88 Emotional strategies M 25.60 28.09 0.04 25.29 26.67 0.09 26.25 26.48 0.23 SD 8.03 6.30 8.76 7.11 8.06 6.94 Adaptive coping strategies Active coping M 5.31 5.44 0.75 4.85 5.55 0.07 5.63 5.12 0.14 SD 2.20 1.99 2.28 2.03 2.07 2.15 Emotional support M 4.48 4.95 0.21 4.41 4.76 0.40 4.91 4.49 0.25 SD 2.13 2.27 2.31 2.21 2.11 2.32 Instrumental support M 5.24 5.74 0.15 5.20 5.54 0.36 5.40 5.51 0.73 SD 2.13 2.27 2.31 2.21 2.11 2.32 Positive reframing M 3.98 4.17 0.61 3.45 4.33 0.04 4.28 3.96 0.40 SD 2.29 2.26 2.19 2.30 2.42 2.18 Planning M 4.61 4.91 0.45 4.15 4.96 0.08 5.02 4.49 0.17 SD 2.31 2.35 2.55 2.26 2.35 2.34 Acceptance M 5.76 5.61 0.67 5.72 5.70 0.96 5.75 5.63 0.72 SD 2.03 2.11 2.21 2.03 1.97 2.17 Religion M 5.60 6.03 0.22 5.18 6.00 0.03 5.46 6.09 0.06 SD 2.13 1.98 2.50 1.90 2.20 1.93 Maladaptive coping strategies Denial M 2.76 3.29 0.19 2.54 3.05 0.25 2.85 3.01 0.69 SD 2.40 2.45 2.46 2.45 2.51 2.40 Substance use M 0.91 0.81 0.75 0.73 0.92 0.56 1.10 0.66 0.13 SD 1.90 1.72 1.96 1.77 2.07 1.53 Behavioural disengage M 2.95 3.09 0.74 3.67 3.04 0.41 3.11 2.79 0.42 SD 2.39 2.38 2.31 2.41 2.68 2.09 Venting M 3.26 3.53 0.51 3.40 3.23 0.69 3.18 3.37 0.63 SD 2.45 2.08 2.63 2.21 2.29 2.39 Humour M 1.90 1.62 0.45 1.95 1.67 0.48 1.55 1.97 0.23 SD 2.15 2.09 2.56 1.91 1.86 2.33 Self-blame M 2.44 1.90 0.13 2.73 1.96 0.04 2.16 2.08 0.81 SD 2.34 1.81 2.53 1.97 1.91 2.28 Table 3 Mean differences of participant characteristic on different coping strategies and acceptance. These strategies are considered adaptive in light of people with chronic diseases such as diabetes and non-clinical samples. Discussionh These findings are similar to a Turkish study of type 1 and 2 diabetes except for active coping. The differences included positive reframing, self- distraction, and venting common in a Turkish [6] com- pared to the Zambian sample. These findings could be different from people in Western developed countries. going with hope, similar to findings in Iran [38]. Consist- ent with other chronic illnesses such as HIV, there are reports of individuals relying on inner strength supplied by their Christian faith to cope with HIV [39]. In addition to frequently using adaptive coping strat- egies, Zambian individuals with diabetes were more likely to use behavioural than emotional strategies. Behavioural coping strategies are overt physical activi- ties aimed at removing or averting the stressor. In con- trast, emotional strategies aim to reduce and manage the intensity of the negative and distressing emotions caused by a stressful situation rather than solving the f The highest proportion of individuals used religious coping. Zambia is a very religious country. Therefore, most Christians hope God to heal them or ameliorate their suffering by relying on inner strength to keep them Hapunda BMC Endocrine Disorders (2022) 22:215 Page 7 of 11 Table 4 Correlation matrix for coping strategies, diabetes specific-distress, depression and diabetes self-care Diabetes Specific Distress Depression Self-Care Brief Cope (Total Score) 0.475** 0.246* 0.007 Behavioural focused 0.430* 0.175 0.062 Emotional focused 0.374* 0.321 0.003 Adaptive coping strategies Active coping 0.247 0.108 0.078 Use of emotional sup- port 0.079 0.141 0.263 Use of instrumental support 0.159 -0.039 0.118 Positive reframing 0.278 0.142 0.104 Planning 0.120 -0.001 -0.017 Acceptance -0.053 0.003 -0.016 Religion 0.122 0.102 0.129 Maladaptive coping strategies Self-blame 0.238** 0.150 -0.171* Self- distraction 0.370* 0.137 0.017 Denial 0.402* 0.191* -0.071 Substance use 0.193* 0.177* -0.066 Behavioural disengage- ment 0.448* 0.170 -0.023 Venting 0.326* 0.206* 0.068 Humour 0.173* 0.230** -0.016 p = < 0.05 p = < 0.01 **p = < 0.001 Table 4 Correlation matrix for coping strategies, diabetes specific-distress, depression and diabetes self-care of problems, recognise and change unhealthy emo- tional reactions, and prevent adverse effects on the body. According to Carver, people who use adaptive strategies (active coping, use of informational support, planning, and positive reframing) use them to change the stressful situation [29]. For example, individuals with diabetes may apply adaptive strategies to change treatment and food related problems that they might find distressing. Discussionh Surprisingly, active coping where one is intentionally or is goal-directed to minimise the physical, psychological or social stressor was associated with increased diabetes specific-distress. It could be that some of the stressors, such as deprivation of food or being overwhelmed with diabetes regimens, cannot be removed because they are beyond their capabilities. It can be exasperating to actively cope with what one cannot change (or reduce) given the high poverty and weak health system these people find themselves in. This could explain why many individuals use emotional focused coping, which gets utilised when the problem is out of one’s control. Unlike the adaptive (problem) focused strategies (venting, use of emotional support, humour, acceptance, self-blame, and religion), emo- tional strategies aim to regulate emotions associated with the stressful situation [29]. Distressing emotions such as worrying about low sugar levels, fear of future complications and feeling guilty when off-track with diabetes management are common in patients, and the use of emotional strategies to cope with such emotions is high, as found by this study. Further, in the face of insufficient manageability of a stressor, some individu- als may decide to use maladaptive strategies instead because the conventional resources of help seem to be exhausted [42]. Thus, a person needs to cope and accept the situation beyond their capability. problematic situation itself [40]. Thus, emotional strate- gies make a person feel better but don’t solve the source of one’s emotional distress. Emotional focused coping often gets utilised when the problem is out of control or in situations where the stressor is unavoidable. For example, having a chronic and terminal illness or sudden death, and one needs to cope and accept the situation [40]. Furthermore, adults were more likely to use behav- ioural strategies than adolescents to use behavioural strategies. It is under this strategy that problem solving skills like planning have been found to be a common skill taught in a variety of patient education and self-man- agement interventions [41] Thus, teaching coping and problem-solving skills may improve quality of life and diabetes management [41]. problematic situation itself [40]. Thus, emotional strate- gies make a person feel better but don’t solve the source of one’s emotional distress. Emotional focused coping often gets utilised when the problem is out of control or in situations where the stressor is unavoidable. Discussionh For example, having a chronic and terminal illness or sudden death, and one needs to cope and accept the situation [40]. Furthermore, adults were more likely to use behav- ioural strategies than adolescents to use behavioural strategies. It is under this strategy that problem solving skills like planning have been found to be a common skill taught in a variety of patient education and self-man- agement interventions [41] Thus, teaching coping and problem-solving skills may improve quality of life and diabetes management [41]. Religion and spirituality are commonly employed as coping mechanisms in individuals with diabetes [43]. Although religion was the most frequently used coping strategy, it was not associated with diabetes specific- distress, depression and diabetes self-care. Our find- ings contrast with a sample of Nigerian individuals with diabetes whose high intrinsic and extrinsic religiosities were associated with positive coping skills and better treatment outcomes in people with depression or dia- betes [44]. The Nigerian findings confirm that religion is an adaptive coping strategy for diabetes. In our study, we only found a significant difference in the use of reli- gion as a coping strategy between adolescents and adults. This finding is consistent with the trends worldwide that shows that young people are less religious than adults [45]. This finding suggests that adults compared to ado- lescents use emotional strategies to regulate emotions associated with their diabetes stressful situations. Mixed results were found regarding the association between coping strategies and diabetes specific-dis- tress, depression and self-care. Although individuals with diabetes used adaptive coping strategies more than maladaptive, most of them were not associated with psychological outcomes except for active coping and emotional support. Adaptive coping strategies confront problems directly, make practically realistic evaluations Page 8 of 11 Hapunda BMC Endocrine Disorders (2022) 22:215 Hapunda BMC Endocrine Disorders (2022) 22:215 Regarding diabetes self-care, only emotional sup- port was associated with better diabetes care. Emo- tional and psychological support has been documented to improve people’s ability to adjust or take adequate responsibility in diabetes self-management [46]. Evi- dence from an international observation study, con- sistent with the current study showed that increased emotional support was associated with better diabe- tes self-care [47]. In chronic illnesses such as diabetes, emotional support is considered an adaptive coping strategy. Discussionh Specifically, it positively impacts a healthy diet, increased perceived support, higher self-efficacy, improved psychological well-being, and better gly- caemic control, according to data from a recent sys- tematic review [48]. Although emotional support as a coping strategy was not significantly different between adults and adolescents, adults used emotional support more than adolescents to cope with stressors. While both adolescents and adults need emotional support, emotional support must not be seen as controlling or intrusive in adolescents. Self-blame was associated with reduced diabetes self-care. Consistent with other stud- ies, self-blame has been associated with negative health outcomes [49]. As observed earlier, self-blame is related to emotional distress, affecting diabetes self-care [21, 50]. This is because self-criticism or blame is associ- ated with low resilience to adhere to diabetes care [51] hence considered a maladaptive coping strategy. Furthermore, denial, substance use, venting, self-distrac- tion, and humour increased depressive symptoms. About 18% (14% type 1 and 22% type 2) used denial. Refusing to accept a problem or reality can interfere with one’s ability to tackle challenges simply because one cannot acknowl- edge the problem and downplay consequences which can, in the long run, increase anxiety if the problem is not going away. Moreover, people with depression in stressful situations often use strategies based on denial and avoid- ance and have more difficulties finding positive aspects of stressful events [53]. With regard to substance use, our study is consistent with findings of meta-analysis studies that have shown that depression is associated with concur- rent alcohol use, drug use and impairment in clinical and community samples [54, 55]. Although the percentage of users was small, more education is needed. Only 3% (3% type 1 and 3% type 2) used this strategy. Humour was asso- ciated with increased depressive symptoms. About 10% (9% type 1 and 11% type 2) used this strategy. One possible explanation for this association is that humour is likely to increase depression if it is targeted at mocking a stressor, simply because humour works well as a coping strategy if the event or interaction is pleasurable [56]. Our findings are consistent with Kuiper and colleagues who showed that the maladaptive components of humour that are self-focused (e.g., self-defeating and belaboured humour) predict detrimental effects on poorer self-esteem, greater depression and anxiety [57]. Discussionh Evidence from health care professionals’ interviews also suggests that humour in a therapeutic situation should be used in moderation and under certain socio-cultural conditions if it is to be effec- tive [58]. In a disease like diabetes, it might not work. As with self-distraction, it was expected that this relationship would be observed. As expected, most maladaptive coping strategies were associated with negative outcomes. For instance, some scales within behavioural focus such as self-distraction, substance use, and behavioural disengagement were associated with increased diabetes specific distress because they all temporary address the stressor. These do not take the problem away. These are also known as avoidant strategies in which physical or cognitive effort to disengage from the stressor is applied [29]. For exam- ple, with behavioural disengagement, individuals reduce their efforts to deal with the stressor, which maintains or increases diabetes specific-distress. Further, emotion- ally focused strategies were associated with increased diabetes specific-distress. This domain is a facet that has several maladaptive strategies such as denial, venting or humour which could worsen diabetes specific-distress. Therefore, it was not surprising that self-blame, humour, venting, and denial were associated with increased dia- betes specific-distress. For instance, a systematic review of chronic conditions including diabetes found that self- blame for the onset for the diseases was associated with increased emotional distress [52]. Generally, our data is similar to the Turkish data that showed that problem- focused and emotional-focused strategies were used in both type 1 and type 2 diabetes [6]. Distracting oneself from a stressor is a temporal solu- tion because the stressor does not go away with this strat- egy. Therefore, by temporarily distracting yourself, you may give the emotion some time to decrease in intensity, but it will still emerge later, which can be depressing. We observed that a good percentage of our sample utilises this strategy (35% [32% type 1 and 37% type). This obser- vation is consistent with our initial qualitative study, which showed adolescents used distraction as a coping strategy [27]. Thus, the current study validates the previ- ous qualitative study. Venting was also associated with increased depressive symptoms. About 19% (22% type 1 and 15% type 2) used this strategy. Venting refers to stating unpleasant feel- ings or expressing one’s negative feelings [29] and is a maladaptive strategy. This coping strategy is similar to the explosiveness of speech, one of the Types A behav- iour pattern characteristics. Funding The author discloses receipt of the ISSBD-JF mentored fellowship granted the author during PhD studies. Availability of data and materials The datasets used and/or analysed during the current study available from the corresponding author on reasonable request. Data is not currently public because more analysis is still being conducted on the data set. References 1. Chan J, DeMelo M, Gingras J, Gucciardi E. Challenges of diabetes self- management in adults affected by food insecurity in a large urban centre of Ontario Canada. Int J Endocrinol. 2015;1:2015. Consent for publication Not applicable. Consent for publication Not applicable. Competing interests The author declared no potential financial and non-financial conflicts of inter- est concerning this article’s research, authorship, and publication. Received: 15 February 2022 Accepted: 16 August 2022 Received: 15 February 2022 Accepted: 16 August 2022 Received: 15 February 2022 Accepted: 16 August 2022 Abbreviations B f COP B f Brief COPE: Brief Coping Orientation to Problems Experienced; HIV: Human Immunodeficiency Virus; LMIC: Low- and middle-income country; MDI: Major Depression Inventory; PAID: Problem Areas in Diabetes; SCI: Self Care Inven- tory; T1D: Type 1 Diabetes; T2D: Type 2 Diabetes. Authors’ contributions GH ll d d l i Moreover, emotional oriented coping appears to play a role in developing depressive symptoms, anxiety and diabetes specific-distress [61]. Equally, behavioural- focused strategies were associated with increased expe- rience of diabetes specific-distress. Again, we believe that individuals using behavioural focused strategies apply coping flexibility (i.e., an individual’s ability to modify and change coping strategies depending on the context). The availability of numerous coping strategies if one has mastered their effectiveness may be an impor- tant precursor to coping flexibility. Coping flexibility can only be exercised if an individual can access and use different coping strategies [62] but may be ineffective if they do not master the strategies that work for them.h GH collected data, analysed and wrote the manuscript. The author(s) read and approved the final manuscript. Discussionh The Type A behaviour pat- tern is another response to a stressor characterised by Hapunda BMC Endocrine Disorders (2022) 22:215 Hapunda BMC Endocrine Disorders (2022) 22:215 Page 9 of 11 Zambian individuals with diabetes. The difference only lay in the use of religion, with more adults likely to use it as a coping strategy. Mixed results were found on the use of coping strategies and how they are associated with diabetes-specific distress, depression and diabetes self- care. Some people used maladaptive coping strategies that affected their psychological well-being and diabe- tes management. There is need to help people use more behavioural but adaptive strategies in order to improve their quality of life. extreme hostility, competitiveness, hurry, impatience, restlessness, aggressiveness, and explosiveness of speech [59]. This kind of behaviour is likely to increase sadness and loneliness, precursors of depression. extreme hostility, competitiveness, hurry, impatience, restlessness, aggressiveness, and explosiveness of speech [59]. This kind of behaviour is likely to increase sadness and loneliness, precursors of depression. All coping scales (total score) were related to increased diabetes specific-distress and increased depressive symptoms. This finding remains unclear why. However, it could be that using multiple coping strate- gies can sometimes be frustrating when a person has not mastered the ones that work well for them, hence the increase in experience of diabetes-related stress and depressive symptoms. Consequently, using multiple styles that a person does not find effective may lead to a pessimistic outlook on finding a solution [60]. One has to identify what works to be consistently used and pro- duce results. Using multiple coping strategies can lead to trouble organising thoughts on what works and does not work and may lead individuals to keep switching strategies to find what work. This process may be frus- trating and increase diabetes specific-distress. Acknowledgements h h ld l k The author would like to thank all the participants for actively participating in this study, and the staff at the University Teaching Hospital. Ethics approval and consent to participate Informed consent for participants above 18 years and assent following informed consent from guardians for participants below 18 years was obtained. All procedures performed in studies involving human participants adhered to 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. The ethics committee of the School of Humanities and Social Sciences, University of Zambia, approved the study on 29th April 2011 (Reference number IRB: 00006464, IORG: 005376). This study has some limitations. Firstly, we did not have data on diabetes biological markers such as HbA1c and specific medication the people were using. Secondly, our sample size was small. However, this study compared strategies used between type 1 and type 2 people, which most studies tend to report separately, making it diffi- cult to make comparisons. Further, to the author’s best knowledge, this is the first study to investigate coping strategies that people with diabetes use in Zambia. Thus, this data on different coping strategies used between dif- ferent age groups, types of diabetes, sex, and how they are used on different psychological challenges they face may be important for diabetes care and education in Zambia and other Sub-Saharan African countries. 1. Chan J, DeMelo M, Gingras J, Gucciardi E. Challenges of diabetes self- management in adults affected by food insecurity in a large urban centre of Ontario Canada. Int J Endocrinol. 2015;1:2015. 2. Debussche X, Balcou-Debussche M, Besançon S, Traore SA. Challenges to diabetes self-management in developing countries. Diabetes Voice. 2009;54(Special Issue):12–4. 2. Debussche X, Balcou-Debussche M, Besançon S, Traore SA. Challenges to diabetes self-management in developing countries. Diabetes Voice. 2009;54(Special Issue):12–4. 1. Chan J, DeMelo M, Gingras J, Gucciardi E. Challenges of diabetes self- management in adults affected by food insecurity in a large urban centre of Ontario Canada. Int J Endocrinol. 2015;1:2015. Conclusion Diabetes-specific emo- tional distress among adolescents: feasibility, reliability, and validity of the problem areas in diabetes-teen version. Pediatric Diabetes. 2011;12(4pt1):341–4. 13. Malipa MN, Menon JA. The relationship between compliance and quality of life among adolescents with diabetes mellitus type1. Med J Zambia. 2013;40(3):93–103. 35. Weinger K, Butler HA, Welch GW, La Greca AM. Measuring diabetes self-care: a psychometric analysis of the self-care inventory-revised with adults. Diabetes Care. 2005;28(6):1346–52. 14. Kirby R, Shakespeare-Finch J, Palk G. Adaptive and maladaptive coping strategies predict posttrauma outcomes in ambulance personnel. Trau- matology. 2011;17(4):25–34. 36. Bech P, Rasmussen NA, Olsen LR, Noerholm V, Abildgaard W. The sensitiv- ity and specificity of the major depression inventory, using the present state examination as the index of diagnostic validity. J Affect Disord. 2001;66(2–3):159–64. 15. Albai A, Sima A, Papava I, Roman D, Andor B, Gafencu M. Association between coping mechanisms and adherence to diabetes-related self-care activities: a cross-sectional study. Patient Prefer Adherence. 2017;11:1235. 37. Cuijper P, Dekker J, Noteboom A, Smit N, Peen J. The sensitivity and speci- ficity of the Major Depression Inventory in outpatients. BMC Psychiatry, 2007; 7(39). https://doi.org/10.1186/1471-244X-7-39. 16. McCoy MA, Theeke LA. A systematic review of the relationships among psychosocial factors and coping in adults with type 2 diabetes mellitus. Int J Nurs Sci. 2019;6(4):468–77. 38. Shamsalinia A, Pourghaznein T, Parsa M. The relationship between hope and religious coping among patients with type 2 diabetes. Global J Health Sci. 2016;8(1):208. 17. Grey M. Coping skills training for youths with diabetes. Diabetes Spec- trum. 2011;24(2):70–5. 39. Chinouya M, O’Keefe E. God will look after us: Africans, HIV and religion in Milton Keynes. Diversi Health Soc Care. 2005;2(3):177–86. 18. David R, Weber JN. Diabetes Mellitus in Children. In: Kliegman RM, Geme JWS, Blum NJ, Shah SS, Tasker RC, Wilson KM, editors. Nelson textbook of pediatrics. 21th ed. Philadelphia: Elsevier; 2020. p. 3019–52. 40. Galor S. Emotional coping strategies. Amsterdam; 2012. Available from: https://drsharongalor.wordpress.com/2012/03/31/emotion-focus ed-coping-strategies/. [Updated 2012; cited 2021 November 7]. 19. Young-Hyman D, De Groot M, Hill-Briggs F, Gonzalez JS, Hood K, Peyrot M. Psychosocial care for people with diabetes: a position statement of the American Diabetes Association. Diabetes Care. 2016;39(12):2126–40. 41. Fisher EB, Thorpe CT, McEvoy DeVellis B, DeVellis RF. Healthy cop- ing, negative emotions, and diabetes management. Diabetes Educ. 2007;33(6):1080–103. 20. Delamater AM, de Wit M, McDarby V, Malik JA, Hilliard ME, Northam E, Acerini CL. Conclusion The relationship between anxiety, coping strategies and characteristics of patients with diabetes. Health Qual Life Outcomes. 2008;6(1):1–9. 7. Delahanty L, Grant RW, Wittenberg E, Bosch JL, Wexler DJ, Cagliero E, Meigs JB. Association of diabetes-related emotional distress with diabetes treatment in primary care patients with type 2 diabetes. Diabet Med. 2007;24(1):48–54. 28. Waugh CE, Shing EZ, Furr RM. Not all disengagement coping strategies are created equal: positive distraction, but not avoidance, can be an adaptive coping strategy for chronic life stressors. Anxiety Stress Coping. 2020;33(5):511–29. 8. Strandberg RB, Graue M, Wentzel-Larsen T, Peyrot M, Rokne B. Relation- ships of diabetes-specific emotional distress, depression, anxiety, and overall well-being with HbA1c in adult persons with type 1 diabetes. J Psychosom Res. 2014;77(3):174–9. 29. Carver CS. You want to measure coping but your protocol’too long: Consider the brief cope. Int J Behav Med. 1997;4(1):92–100. 30. Snell DL, Siegert RJ, Hay-Smith EJ, Surgenor LJ. Factor structure of the Brief COPE in people with mild traumatic brain injury. J Head Trauma Rehabil. 2011;26(6):468–77. y 9. Bronner MB, Peeters MA, Sattoe JN, van Staa A. The impact of type 1 diabetes on young adults’ health-related quality of life. Health Qual Life Outcomes. 2020;18(1):1–7. 31. Yusoff N, Low WY, Yip CH. Reliability and validity of the Brief COPE Scale (English version) among women with breast cancer undergoing treat- ment of adjuvant chemotherapy: a Malaysian study. Med J Malaysia. 2010;65(1):41–4. 10. Holt RI, DeVries JH, Hess-Fischl A, Hirsch IB, Kirkman MS, Klupa T, Ludwig B, Nørgaard K, Pettus J, Renard E, Skyler JS. The management of type 1 diabetes in adults. A consensus report by the American Diabetes Asso- ciation (ADA) and the European Association for the Study of Diabetes (EASD). Diabetes Care. 2021;44(11):2589–625. 32. Poulus D, Coulter TJ, Trotter MG, Polman R. Stress and coping in esports and the influence of mental toughness. Front Psychol. 2020;11:628. https://doi.org/10.3389/fpsyg.2020.0062. 11. Fisher L, Gonzalez JS, Polonsky WH. The confusing tale of depression and distress in patients with diabetes: a call for greater clarity and precision. Diabet Med. 2014;31(7):764–72. 33. Welch GW, Jacobson AM, Polonsky WH. The problem areas in diabetes scale: an evaluation of its clinical utility. Diabetes Care. 1997;20(5):760–6. 12. Hema DA, Roper SO, Nehring JW, Call A, Mandleco BL, Dyches TT. Daily stressors and coping responses of children and adolescents with type 1 diabetes. Child Care Health Dev. 2009;35(3):330–9. 34. Weissberg-Benchell J, Antisdel-Lomaglio J. Conclusion In conclusion, adaptive coping strategies such as reli- gion, acceptance, instrumental support and active cop- ing are the most frequently used coping strategies among 2. Debussche X, Balcou-Debussche M, Besançon S, Traore SA. Challenges to diabetes self-management in developing countries. Diabetes Voice. 2009;54(Special Issue):12–4. Page 10 of 11 Page 10 of 11 Hapunda BMC Endocrine Disorders (2022) 22:215 Hapunda BMC Endocrine Disorders (2022) 22:215 3. Pastakia SD, Pekny CR, Manyara SM, Fischer L. Diabetes in sub-Saharan Africa–from policy to practice to progress: targeting the existing gaps for future care for diabetes. Diabetes Metab Syndr Obes. 2017;10:247. 3. Pastakia SD, Pekny CR, Manyara SM, Fischer L. Diabetes in sub-Saharan Africa–from policy to practice to progress: targeting the existing gaps for future care for diabetes. Diabetes Metab Syndr Obes. 2017;10:247. 24. Brannon L, Feist J. Introduction to health psychology. India Edition, Thompson Wadsworth. 2007;9:111–35. 25. Knowles SR, Apputhurai P, O’Brien CL, Ski CF, Thompson DR, Castle DJ. Exploring the relationships between illness perceptions, self-efficacy, coping strategies, psychological distress and quality of life in a cohort of adults with diabetes mellitus. Psychol Health Med. 2020;25(2):214–28. 4. Hilliard ME, De Wit M, Wasserman RM, Butler AM, Evans M, Weissberg- Benchell J, Anderson BJ. Screening and support for emotional burdens of youth with type 1 diabetes: strategies for diabetes care providers. Pediatr Diabetes. 2018;19(3):534–43. 26. Yasui-Furukori N, Murakami H, Otaka H, Nakayama H, Murabayashi M, Mizushiri S, Matsumura K, Tanabe J, Matsuhashi Y, Yanagimachi M, Naka- mura K. Coping behaviors and depressive status in individuals with type 2 diabetes mellitus. Ann Gen Psychiatry. 2019;18(1):1–8. 5. Sumlin LL, Garcia TJ, Brown SA, Winter MA, Garcia AA, Brown A, Cuevas HE. Depression and adherence to lifestyle changes in type 2 diabetes: a systematic review. Diabetes Educ. 2014;40(6):731–44. 5. Sumlin LL, Garcia TJ, Brown SA, Winter MA, Garcia AA, Brown A, Cuevas HE. Depression and adherence to lifestyle changes in type 2 diabetes: a systematic review. Diabetes Educ. 2014;40(6):731–44. 6. Tuncay T, Musabak I, Gok DE, Kutlu M. The relationship between anxiety, coping strategies and characteristics of patients with diabetes. Health Qual Life Outcomes. 2008;6(1):1–9. 27. Hapunda G, Abubakar A, Van de Vijver F, Pouwer F. Living with type 1 diabetes is challenging for Zambian adolescents: qualitative data on stress, coping with stress and quality of care and life. BMC Endocr Disord. 2015;15(1):1–2. 6. Tuncay T, Musabak I, Gok DE, Kutlu M. Conclusion ISPAD Clinical Practice Consensus Guidelines 2018: psycho- logical care of children and adolescents with type 1 diabetes. Pediatr Diabetes. 2018;19:237–49. 42. Büssing A, Ostermann T, Neugebauer EA, Heusser P. Adaptive coping strategies in patients with chronic pain conditions and their interpreta- tion of disease. BMC Public Health. 2010;10(1):1. 43. Sridhar GR. Diabetes, religion and spirituality. Int J Diabetes Dev Ctries. 2013;33(1):5–7. 21. Hapunda G, Abubakar A, Pouwer F, Van de Vijver F. Validity and reliability of the Zambian version of the Problem Areas in Diabetes (PAID) scale: a triangulation with cognitive interviews. South Afr J Diabetes Vasc Dis. 2016;13(1):4–12. 44. Amadi KU, Uwakwe R, Odinka PC, Ndukuba AC, Muomah CR, Ohaeri JU. Religion, coping and outcome in out-patients with depression or diabe- tes mellitus. Acta Psychiatr Scand. 2016;133(6):489–96. 22. Lloyd CE, Nouwen A, Sartorius N, Ahmed HU, Alvarez A, Bahendeka S, Basangwa D, Bobrov AE, Boden S, Bulgari V, Burti L. Prevalence and corre- lates of depressive disorders in people with Type 2 diabetes: results from the International Prevalence and Treatment of Diabetes and Depression (INTERPRET-DD) study, a collaborative study carried out in 14 countries. Diabet Med. 2018;35(6):760–9. 45. Pew Research Centre. The age gap in religion around the world. Wash- ington DC; 2018. Available from https://www.pewforum.org/2018/06/13/ the-age-gap-in-religion-around-the-world/. [Updated 2018; cited 2021 November 7]. 46. Kalra S, Jena BN, Yeravdekar R. Emotional and psychological needs of people with diabetes. Indian J Endocrinol Metab. 2018;22(5):696. 23. Clay RA. More is needed to treat diabetes. Monit Psychol. 2017;48(6):36. Page 11 of 11 Hapunda BMC Endocrine Disorders (2022) 22:215 Hapunda BMC Endocrine Disorders (2022) 22:215 47. Koetsenruijter J, van Eikelenboom N, van Lieshout J, Vassilev I, Lionis C, Todorova E, Portillo MC, Foss C, Gil MS, Roukova P, Angelaki A. Social support and self-management capabilities in diabetes patients: an inter- national observational study. Patient Educ Couns. 2016;99(4):638–43. 48. Pamungkas RA, Chamroonsawasdi K, Vatanasomboon P. A systematic review: family support integrated with diabetes self-management among uncontrolled type II diabetes mellitus patients. Behav Sci. 2017;7(3):62. 49. Harry KM. The Cardiac Self-Blame Attributions scale as a predictor of physical and mental health outcomes in underrepresented patients with cardiovascular disease. University of Missouri-Kansas City; 2018. 50. Khan A, Choudhary P. Investigating the association between diabe- tes distress and self-management behaviors. J Diabetes Sci Technol. 2018;12(6):1116–24. 51. Yi-Frazier JP, Yaptangco M, Semana S, Buscaino E, Thompson V, Cochrane K, Tabile M, Alving E, Rosenberg AR. Conclusion The association of personal resilience with stress, coping, and diabetes outcomes in adolescents with type 1 diabetes: Variable-and person-focused approaches. J Health Psychol. 2015;20(9):1196–206. 52. Callebaut L, Molyneux P, Alexander T. The relationship between self- blame for the onset of a chronic physical health condition and emo- tional distress: A systematic literature review. Clin Psychol Psychother. 2017;24(4):965–86. 53. Orzechowska A, Zajączkowska M, Talarowska M, Gałecki P. Depression and ways of coping with stress: a preliminary study. Med Science Monit. 2013;19:1050. 54. Conner KR, Pinquart M, Gamble SA. Meta-analysis of depression and sub- stance use among individuals with alcohol use disorders. J Subst Abuse Treat. 2009;37(2):127–37. 55. Conner KR, Pinquart M, Duberstein PR. Meta-analysis of depression and substance use and impairment among intravenous drug users (IDUs). Addiction. 2008;103(4):524–34. 56. Nezlek JB, Derks P. Use of humor as a coping mechanism, psychological adjustment, and social interaction. Humor. 2001;14(4):395–413. 57. Kuiper NA, Grimshaw M, Leite C, Kirsh G. Humor is not always the best medicine: Specific components of sense of humor and psychological well-being. 2004;17(12):135–168. https://doi.org/10.1515/humr.2004.002. 58. de Almeida CV, Nunes C. Humour is important in healthcare relationship? Open Access Libr J. 2020;7:e6372. 59. Ito M, Matsushima E. Presentation of coping strategies associated with physical and mental health during health check-ups. Community Ment Health J. 2017;53(3):297–305. 60. Sahler OJ, Carr JE. Coping strategies. In Carrey WB, editor. Developmental behavioral pediatrics. Philadelphia: W.B. Saunders; 2009. p. 491–6. 61. Burns RJ, Deschênes SS, Schmitz N. Associations between coping strate- gies and mental health in individuals with type 2 diabetes: Prospective analyses. Health Psychol. 2016;35(1):78. 61. Burns RJ, Deschênes SS, Schmitz N. Associations between coping strate- gies and mental health in individuals with type 2 diabetes: Prospective analyses. Health Psychol. 2016;35(1):78. 62. Heffer T, Willoughby T. A count of coping strategies: a longitudinal study investigating an alternative method to understanding coping and adjust- ment. PLoS One. 2017;12(10):e0186057. 62. Heffer T, Willoughby T. A count of coping strategies: a longitudinal study investigating an alternative method to understanding coping and adjust- ment. PLoS One. 2017;12(10):e0186057. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from:
https://openalex.org/W1967786809	https://advancesincontinuousanddiscretemodels.springeropen.com/counter/pdf/10.1186/1687-1847-2014-171	Latin	null	Numerical solution of a singularly perturbed Volterra integro-differential equation	Advances in difference equations	2,014	cc-by	7,527	R ES EA RCH Open Access Abstract We study the convergence properties of a diﬀerence scheme for singularly perturbed Volterra integro-diﬀerential equations on a graded mesh. We show that the scheme is ﬁrst-order convergent in the discrete maximum norm, independently of the perturbation parameter. Numerical experiments are presented, which are in agreement with the theoretical results. MSC: 45J05; 65R20; 65L11 Keywords: singular perturbation; Volterra integro-diﬀerential equations; diﬀerence scheme; uniform convergence; graded mesh Sebaheddin ¸Sevgin* Correspondence: ssevgin@yahoo.com Department of Mathematics, Faculty of Sciences, Yuzuncu Yil University, Van, 65080, Turkey ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 ¸ g q , http://www.advancesindifferenceequations.com/content/2014/1/171 Numerical solution of a singularly perturbed Volterra integro-differential equation Sebaheddin ¸Sevgin © 2014 ¸Sevgin; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons Attribu- tion License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 1 Introduction Singularly perturbed Volterra integro-diﬀerential equations arise in many physical and biological problems. Among these are diﬀusion-dissipation processes, epidemic dynam- ics, synchronous control systems, and ﬁlament stretching problems (see, e.g., [–]). For extensive reviews, see [, –]. Singularly perturbed diﬀerential equations are typically characterized by a small param- eter ε multiplying some or all of the highest-order terms in the diﬀerential equations. The diﬃculties arising in the numerical solutions of singularly perturbed problems are well known. A comprehensive review of the literature on numerical methods for singularly perturbed diﬀerential equations may be found in [–]. This paper is concerned with the following singularly perturbed Volterra integro- diﬀerential equation: Lu := εu′(t) + f t,u(t) + t  K t,s,u(s) ds = , t ∈I := [,T], (.) u() = A, (.) Lu := εu′(t) + f t,u(t) + t  K t,s,u(s) ds = , t ∈I := [,T], (.) (.) Lu := εu′(t) + f t,u(t) +  K t,s,u(s) ds = , t ∈I := [,T], (.) u() = A, (.)  u() = A, (.) (.) u() = A, (.) where < ε ≪is the perturbation parameter, f (t,u) ((t,u) ∈I ×R) and K(t,s,u) ((t,s,u) ∈ I × I × R) are suﬃciently smooth functions, A is a given constant and ∂f ∂u ≥α > . By substituting ε = in (.), we obtain the reduced equation f t,ur(t) + t  K t,s,ur(s) ds = , © 2014 ¸Sevgin; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons Attribu- tion License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 Page 2 of 15 which is a Volterra integral equation of the second kind. The singularly perturbed nature of (.) occurs when the properties of the solution with ε > are incompatible with those when ε = . The interest here is in those problems which do imply such an incompatibility in the behavior of u in a neighborhood of t = . This suggests the existence of an initial layer near the origin where the solution undergoes a rapid transition. A special class of singularly perturbed integro-diﬀerential-algebraic equations and sin- gularly perturbed integro-diﬀerential systems has been solved by Kauthen [, ] by im- plicit Runge-Kutta methods. A survey of the existing literature on a singularly perturbed Volterra integral and integro-diﬀerential equations is given by Kauthen []. The expo- nential scheme that has a fourth-order accuracy when the perturbation parameter ε is ﬁxed is derived and a stability analysis of this scheme is discussed in []. The numerical discretization of singularly perturbed Volterra integro-diﬀerential equations and Volterra integral equations by tension spline collocation methods in certain tension spline spaces are considered in []. For the numerical solution of singularly perturbed Volterra integro- diﬀerential equations, we have studied the following articles: [–]. Our goal is to construct an ε-numerical method for solving (.)-(.), by which we mean a numerical method which generates ε-uniformly convergent numerical approximations to the solution. For this, we use a ﬁnite diﬀerence scheme on an appropriate graded mesh which are dense in the initial layer. Graded meshes are dependent on ε and mesh points have to be condensed in a neighborhood of t = in order to resolve the initial layer. In graded meshes, basically half of the mesh points are concentrated in a O(ε\|lnε\|) neigh- borhood of the point t = and the remaining half forms a uniform mesh on the rest of [,T] (see [, , ]). In [], the authors gave a uniformly convergent numerical method with respect to ε on a uniform mesh for the numerical solution of a linear singularly perturbed Volterra integro-diﬀerential equation. vi = v(ti), for any continuous function v(t). In our estimates, we use the maximum norm given by In our estimates, we use the maximum norm given by ∥v∥∞= max [,T] v(t) . ∥v∥∞= max [,T] v(t) . For any discrete function vi, we also deﬁne the corresponding discrete norm by ∥v∥∞,ωh ≡∥v∥∞= max ≤i≤N \|vi\|. ∥v∥∞,ωh ≡∥v∥∞= max ≤i≤N \|vi\|. Throughout the paper, C will denote a generic positive constant that is independent of ε and the mesh parameter. However, in this study, we will derive a uniformly convergent ε-numerical method on a graded mesh for the numerical solution of a nonlinear singularly perturbed Volterra integro-diﬀerential equation. This is the aspect of the problem of this paper that is diﬀerent from [] and the others. The outline of the paper is as follows: In Section , the properties of the problem (.), (.) are given. In Section , the diﬀerence scheme constructed on the non-uniform mesh for the numerical solution (.), (.) is presented and graded mesh is introduced. Stabil- ity and convergence of the diﬀerence scheme are investigated in Section and error of the diﬀerence scheme is evaluated in Section . Finally numerical results are presented in Section . Let us now introduce some notation. Let ωN = {< t< t< ··· < tN–< tN = T}, ϖN = ωN ∪{t = }, be the non-uniform mesh on [,T]. For each i ≥we set the step size hi = ti – ti–. be the non-uniform mesh on [,T]. For each i ≥we set the step size hi = ti – ti–. Here and throughout the paper we use the notation be the non-uniform mesh on [,T]. For each i ≥we set the step size hi = ti – ti–. Here and throughout the paper we use the notation v¯t,i = vi – vi– hi , vi = v(ti), for any continuous function v(t). vi = v(ti), for any continuous function v(t). ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 3 of 15 http://www.advancesindifferenceequations.com/content/2014/1/171 Page 3 of 15 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 2 The continuous problem In this section, we study the behavior of the solution of (.)-(.) and its ﬁrst derivative which are required for the analysis of the remainder term in the next sections when the error of the diﬀerence scheme is analyzed. Lemma .Suppose that f (t,u) and K(t,s,u) have continuous partial derivatives with respect to u, respectively, on I × R and I × I × R and have uniformly bounded ﬁrst partial derivatives in ε. Then the solution u(x) of problem (.)-(.) satisﬁes the inequalities ∥u∥∞≤C, (.) u′(t) ≤C +  ε exp –αt ε , t ∈I. (.) (.) ∥u∥∞≤C, (.) Proof The analysis of the convergence properties of numerical method that will be ob- tained and the study of the behavior of the solution of (.)-(.) with its ﬁrst derivative will necessarily involve the linearization of the given problem using the mean value theo- rem for several variables (see []). Hence, we obtain εu′(t) + p(t)u(t) + t  G(t,s)u(s)ds = q(t), t ∈I, u() = A, (.) (.) where p(t) = ∂ ∂uf (t,θu), < θ < , G(t,s) = ∂ ∂uK(t,s,γ u), < γ <  and q(t) = –f (t,) – t  K(t,s,)ds. q(t) = –f (t,) – t  K(t,s,)ds. We show the validity of (.). 2 The continuous problem For the solution of the problem (.), we have u(t) = u()exp –  ε t  p(η)dη +  ε t  q(ξ)exp –  ε t ξ q(η)dη dξ –  ε t  ξ  G(ξ,s)u(s)ds exp –  ε t ξ q(η)dη dξ ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 4 of 15 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 4 of 15 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 4 of 15 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸ g q http://www.advancesindifferenceequations.com/content/2014/1/171 and from this we can write and from this we can write u(t) ≤ u() exp –  ε t  p(η)dη +  ε t  q(ξ) exp –  ε t ξ p(η)dη dξ +  ε t  ξ  G(ξ,s) u(s) ds exp –  ε t ξ q(η)dη dξ. u(t) ≤ u() exp –  ε t  p(η)dη +  ε t  q(ξ) exp –  ε t ξ p(η)dη dξ +  ε t  ξ  G(ξ,s) u(s) ds exp –  ε t ξ q(η)dη dξ. If M = maxI×I \|G(t,s)\|, then it follows that u(t) ≤\|A\|exp –αt ε + α–∥q∥∞ – exp –αt ε +  ε M t  ξ  u(s) ds exp –α(t – ξ) ε dξ ≤\|A\|exp –αt ε + α–∥q∥∞ – exp –αt ε + Mα– – exp –αt ε t  u(s) ds ≤\|A\| + α–∥q∥∞+ α–M t  u(s) ds. Then, applying the Gronwall inequality to the last estimate, we obtain u(t) ≤ \|A\| + α–∥q∥∞ exp α–Mt , Then, applying the Gronwall inequality to the last estimate, we obtain Then, applying the Gronwall inequality to the last estimate, we obtain u(t) ≤ \|A\| + α–∥q∥∞ exp α–Mt , which proves (.). which proves (.). 2 The continuous problem To prove (.), diﬀerentiating equation (.) we have εv′(t) + b(t)v(t) = c(t), (.) (.) εv′(t) + b(t)v(t) = c(t), εv′(t) + b(t)v(t) = c(t), where where u′(t) = v(t), b(t) = ∂ ∂uf (t,u) u′(t) = v(t), b(t) = ∂ ∂uf (t,u) and c(t) = – ∂ ∂uf t,u(t) – K t,t,u(t) – t  ∂ ∂t K t,s,u(s) ds. By using (.), we can obtain By using (.), we can obtain By using (.), we can obtain v() = u′() ≤ε–f (,A) ≤Cε–. (.) v() = u′() ≤ε–f (,A) ≤Cε–. (.) It follows from (.) that It follows from (.) that v(t) = v()exp –  ε t  b(η)dη +  ε t  c(ξ)exp –  ε t ξ b(η)dη dξ. (.) (.) ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 5 of 15 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 5 of 15 http://www.advancesindifferenceequations.com/content/2014/1/171 Obviously, if f (t,u) and K(t,s,u) has continuous partial derivatives in u, respectively, on I × R and I × I × R, then  ε t  c(ξ)exp –  ε t ξ b(η)dη dξ ≤C. Hence, we can conclude that (.) is a direct consequence of (.), (.). □ Hence, we can conclude that (.) is a direct consequence of (.), (.). □ Hence, we can conclude that (.) is a direct consequence of (.), (.). □ □ 3 Discretization and mesh (.) (.) ce Equations 2014, 2014:171 Page 6 of erenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 6 of 15 ce Equations 2014, 2014:171 Page 6 of 15 erenceequations.com/content/2014/1/171 g q , p://www.advancesindifferenceequations.com/content/2014/1/171 http://www.advancesindifferenceequations.com/content/2014/1/171 Neglecting Ri in (.), we may suggest the following diﬀerence scheme for approximating (.), (.): LhUi ≡εU¯t,i + f (ti,Ui) + i m= hi  K(ti,tm,Um) + K(ti,tm–,Um–) = , i = ,...,N, (.) U= A. (.) LhUi ≡εU¯t,i + f (ti,Ui) + i m= hi  K(ti,tm,Um) + K(ti,tm–,Um–) = , i = ,...,N, = , i = ,...,N, (.) (.) For the diﬀerence scheme (.), (.) to be ε-uniform convergent, we will use a mesh that is graded inside the initial layer region. For an even number N, the graded mesh takes N/+points in the interval [,τ] and also N/+points in the interval [τ,T], where the transition point τ, which separates the ﬁne and coarse portions of the mesh, is obtained by taking τ = min T/,α–ε\|lnε\| . (.) τ = min T/,α–ε\|lnε\| . (.) In practice one usually has τ ≪T, so the mesh is ﬁne on [,τ] and coarse on [τ,T]. We shall consider a mesh ωN which is equidistant in [τ,T] but graded in [,τ] by a logarithmic mesh generating function (see [, ]). The corresponding mesh points are as follows: if τ < T/, ti = –α–ε ln[– (– ε)i/N], i = ,...,N/, τ + (i – N/)h, i = N/+ ,...,N (.) (.) and if τ = T/, ti = –α–ε ln[– (– exp(– αT ε ))i/N], i = ,...,N/, τ + (i – N/)h, i = N/+ ,...,N, (.) (.) where h = (T – τ)/N. where h = (T – τ)/N. where h = (T – τ)/N. We only consider the graded mesh deﬁned by (.)-(.) in the remainder of the paper. 3 Discretization and mesh To obtain an approximation for (.), we integrate (.) over (ti–,ti): h– i ti ti– Ludt = . (.) h– i ti ti– Ludt = . (.) Using the quadrature rules in [], we have Using the quadrature rules in [], we have εu¯t,i + f (ti,ui) + ti  K ti,s,u(s) ds + R() i + R() i = , (.) (.) where R() i = –h– i ti ti– (ξ – ti–) d dξ f ξ,u(ξ) dξ R() i = –h– i ti ti– (ξ – ti–) d dξ f ξ,u(ξ) dξ and R() i = –h– i ti ti– (ξ – ti–) d dξ ξ  K ξ,s,u(s) ds dξ. Applying also (.) in [] for σ =  to the integral in (.), we obtain ti  K ti,s,u(s) ds = i m= hi  K(ti,tm,um) + K(ti,tm–,um–) + R() i , (.) (.) where R() i = i m= ti ti– (tm–  – ξ) d dξ K ti,ξ,u(ξ) dξ. It is clear from (.) and (.) that It is clear from (.) and (.) that εu¯t,i + f (ti,ui) + i m= hi  K(ti,tm,um) + K(ti,tm–,um–) + Ri = , i = ,...,N, (.) u= A, (.) (.) u= A, where the remainder term is where the remainder term is Ri = –h– i ti ti– (ξ – ti–) d dξ f ξ,u(ξ) dξ – h– i ti ti– (ξ – ti–) d dξ ξ  K ξ,s,u(s) ds dξ Ri = –h– i ti ti– (ξ – ti–) d dξ f ξ,u(ξ) dξ – h– i ti ti– (ξ – ti–) d dξ ξ  K ξ,s,u(s) ds dξ + i ti (tm–  – ξ) d dξ K ti,ξ,u(ξ) dξ. (.) + i m= ti ti– (tm–  – ξ) d dξ K ti,ξ,u(ξ) dξ. (.) + i m= ti ti– (tm–  – ξ) d dξ K ti,ξ,u(ξ) dξ. 4 Stability and convergence of the difference scheme Lemma .Let the diﬀerence operator Lemma .Let the diﬀerence operator ℓUi ≡AiUi – BiUi–, ≤i ≤N, (.) ℓUi ≡AiUi – BiUi–, ≤i ≤N, ℓUi ≡AiUi – BiUi–, ≤i ≤N, (.) be given, where Ai > and Bi > . Then we have the following: (i) For the diﬀerence operator (.), the discrete maximum principle holds: If ℓUi ≥, i ≥and U≥, then Ui ≥, i ≥. (i) For the diﬀerence operator (.), the discrete maximum principle holds: If ℓUi ≥, i ≥and U≥, then Ui ≥, i ≥. (ii) If Ai – Bi ≥α > , then the solution of the diﬀerence initial value problem ℓUi = Fi, i ≥, ℓUi = Fi, i ≥, U= μ U= μ s the estimate satisﬁes the estimate ∥U∥∞≤\|μ\| + α–max ≤i≤N \|Fi\|. (.) ∥U∥∞≤\|μ\| + α–max ≤i≤N \|Fi\|. (.) Page 7 of 15 ¸ g q http://www.advancesindifferenceequations.com/content/2014/1/171 (iii) If Fi ≥is nondecreasing and Ai – Bi ≥α > , then (iii) If Fi ≥is nondecreasing and Ai – Bi ≥α > , then \|Ui\| ≤\|μ\| + α–Fi, i ≥. (.) Proof See []. Lemma .Under condition α + hi  ∂ ∂U K(ti,ti,γ Ui) ≥α∗> , i = ,,...,N, for the diﬀerence operator for the diﬀerence operator ℓhUi := εU¯t,i + ∂ ∂U f (ti,γ Ui)Ui + hi  ∂ ∂U K(ti,ti,γ Ui)Ui (.) we have ∥U∥∞≤\|U\| + α– ∗max ≤i≤N ℓhUi . (.) ℓhUi := εU¯t,i + ∂ ∂U f (ti,γ Ui)Ui + hi  ∂ ∂U K(ti,ti,γ Ui)Ui (.) we have we have ∥U∥∞≤\|U\| + α– ∗max ≤i≤N ℓhUi . ∥U∥∞≤\|U\| + α– ∗max ≤i≤N ℓhUi . (.) Proof Diﬀerence expression (.) can be rewritten as Proof Diﬀerence expression (.) can be rewritten as Proof Diﬀerence expression (.) can be rewritten as Proof Diﬀerence expression (.) can be rewritten as ℓUi ≡AiUi – BiUi–, where where Ai = ε hi + ∂ ∂U f (ti,γ Ui) + hi  ∂ ∂U K(ti,ti,γ Ui), Bi = ε hi . It is easy to see that Ai ≥ε hi + α + hi  ∂ ∂U K(ti,ti,γ Ui) >  and Bi = ε hi > . Since Bi = ε hi > . Since Since Ai – Bi = ∂ ∂U f (ti,γ Ui) + hi  ∂ ∂U K(ti,ti,γ Ui) >  Ai – Bi = ∂ ∂U f (ti,γ Ui) + hi  ∂ ∂U K(ti,ti,γ Ui) >  Ai – Bi = ∂ ∂U f (ti,γ Ui) + hi  ∂ ∂U K(ti,ti,γ Ui) >  by (.), (.) follows in view of (.). □ by (.), (.) follows in view of (.). □ by (.), (.) follows in view of (.). □ □ Now we will show stability for the diﬀerence problem (.)-(.). Lemma .Let the diﬀerence operator ℓhUi be deﬁned by (.). Then for the diﬀerence problem (.)-(.) we have ℓhUi ≤∥f ∥∞+ C i m= hi\|Um–\|, ≤i ≤N. ℓhUi ≤∥f ∥∞+ C i m= hi\|Um–\|, ≤i ≤N. (.) (.) quations 2014, 2014:171 Page 8 of 15 nceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 8 of 15 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 8 of 15 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 8 of 15 Page 8 of 15 http://www.advancesindifferenceequations.com/content/2014/1/171 Proof From (.) we have Proof From (.) we have Proof From (.) we have ℓhUi ≤ f (ti,) + i m= hi K(ti,tm,) + i– m= hi  ∂ ∂U K(ti,tm,γ Um)Um + i m= hi  K(ti,tm–,) + ∂ ∂U K(ti,tm–,γ Um–)Um– . ℓhUi ≤ f (ti,) + i m= hi K(ti,tm,) + i– m= hi  ∂ ∂U K(ti,tm,γ Um)Um If we take into consideration that the kernel K(t,s,u) is bounded, it can be concluded that the estimate (.) holds. □ Lemma .We assume that the condition (.) holds. Then for the solution of diﬀerence scheme (.)-(.), we have Lemma .We assume that the condition (.) holds. Proof Diﬀerence expression (.) can be rewritten as Then for the solution of diﬀerence scheme (.)-(.), we have \|Ui\| ≤ α– ∗∥f ∥∞+ \|A\| exp α– ∗Cti , ≤i ≤N. (.) (.) Proof Let Proof Let Vi = i– m=hi\|Um–\|, i ≥, , i = , Vi = i– m=hi\|Um–\|, i ≥, , i = , where where V¯t,i = \|Ui–\|. V¯t,i = \|Ui–\|. Thus, from the inequality (.), we have the following diﬀerence inequality: Thus, from the inequality (.), we have the following diﬀerence inequality: ℓhUi ≤CVi + ∥f ∥∞, U= A. ℓhUi ≤CVi + ∥f ∥∞, U= A. U= A. Using the discrete maximum principle, we have Using the discrete maximum principle, we have Using the discrete maximum principle, we have \|Ui\| ≤Wi, where wi is the solution of the problem where wi is the solution of the problem ℓhWi = CVi + ∥f ∥∞, W= \|A\|. In view of (.), it follows that In view of (.), it follows that In view of (.), it follows that (.) \|Ui\| ≤\|Wi\| ≤α– ∗CVi + α– ∗∥f ∥∞+ \|A\| (.) and and V¯t,i = \|Ui–\| ≤α– ∗CVi–+ α– ∗∥f ∥∞+ \|A\|. ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 9 of 15 ¸ g q http://www.advancesindifferenceequations.com/content/2014/1/171 http://www.advancesindifferenceequations.com/content/2014/1/171 Then application of the diﬀerence analog of the diﬀerential inequality gives Vi ≤ α– ∗∥f ∥∞+ \|A\| α∗C– exp α– ∗Cti – , which together with (.) proves (.). □ □ which together with (.) proves (.). □ which together with (.) proves (.). 5 Uniform error estimates (.) R() i = i m= ti ti– (tm–  – ξ) d dξ K ti,ξ,u(ξ) (.) ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 10 of 15 h // d i diff i / /2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 10 of 15 ¸ g q http://www.advancesindifferenceequations.com/content/2014/1/171 http://www.advancesindifferenceequations.com/content/2014/1/171 In view of Lemma ., for an arbitrary mesh, it follows from (.), (.), and (.) that In view of Lemma ., for an arbitrary mesh, it follows from (.), (.), and (.) that R() i ≤h– i ti ti– (ξ – ti–) d dξ f ξ,u(ξ) dξ = h– i ti ti– (ξ – ti–) ∂ ∂ξ f ξ,u(ξ) + ∂ ∂uf ξ,u(ξ) u′(ξ) dξ ≤Ch– i ti ti– (ξ – ti–) + u′(ξ) dξ ≤C hi + ε– ti ti– u′(ξ) dξ , i = ,...,N, (.) R() i ≤h– i ti ti– (ξ – ti–) d dξ ξ  K ξ,s,u(s) ds dξ ≤h– i ti ti– (ξ – ti–) K ξ,ξ,u(ξ) + ξ  ∂ ∂ξ K ξ,s,u(s) ds dξ ≤Chi, i = ,...,N, (.) ≤C hi + ε– ti ti– u′(ξ) dξ , i = ,...,N, (.) R() i ≤h– i ti ti– (ξ – ti–) d dξ ξ  K ξ,s,u(s) ds dξ (.) R() i ≤h– i ti ti– (ξ – ti–) d dξ ξ  K ξ,s,u(s) ds dξ ≤h– i ti ti– (ξ – ti–) K ξ,ξ,u(ξ) + ξ  ∂ ∂ξ K ξ,s,u(s) ds dξ (.) R() i ≤ i m= ti ti– (tm–  – ξ) d dξ K ti,ξ,u(ξ) dξ ≤ i m= ti ti– (tm–  – ξ) ∂ ∂ξ K ti,ξ,u(ξ) + ∂ ∂uK ti,ξ,u(ξ) u′(ξ) dξ ≤ i m= ti ti– (tm–  – ξ) + u′(ξ) dξ ≤C max ≤m≤ihm ti  + u′(ξ) dξ ≤C max ≤m≤ihm, i = ,,...,N, (.) R() i ≤ i m= ti ti– (tm–  – ξ) d dξ K ti,ξ,u(ξ) dξ ≤ i m= ti ti– (tm–  – ξ) ∂ ∂ξ K ti,ξ,u(ξ) + ∂ ∂uK ti,ξ,u(ξ) u′(ξ) dξ ≤ m= i ti– (tm–  – ξ) ∂ ∂ξ K ti,ξ,u(ξ) + ∂ ∂uK ti,ξ,u(ξ) u′(ξ) dξ ≤ i m= ti ti– (tm–  – ξ) + u′(ξ) dξ ≤C max ≤m≤ihm ti  + u′(ξ) dξ ≤C max ≤m≤ihm, i = ,,...,N, (.) ≤C max ≤m≤ihm, i = ,,...,N, (.) (.) respectively, where respectively, where respectively, where hi = –α–ε ln[– (– ε) i N ] + α–ε ln[– (– ε) (i–) N ], i = ,...,N/, T–τ N/, i = N/+ ,...,N. 5 Uniform error estimates To investigate the convergence of the method, note that the error function zi = Ui – ui, ≤i ≤N, is the solution of the discrete problem Lhzi = εz¯t,i + f (ti,Ui) – f (ti,ui) + hi  K(ti,ti,Ui) – K(ti,ti,ui) + hi  K(ti,ti–,Ui–) – K(ti,ti–,ui–) Lhzi = εz¯t,i + f (ti,Ui) – f (ti,ui) + hi  K(ti,ti,Ui) – K(ti,ti,ui) + hi  K(ti,ti–,Ui–) – K(ti,ti–,ui–) + Ki = Ri, i = ,...,N, (.) + Ki = Ri, i = ,...,N, (.) (.) + Ki = Ri, i = ,...,N, i i, , , , ( ) = , (.) = , (.) z= , (.) where Ri is given by (.) and where Ri is given by (.) and where Ri is given by (.) and Ki = ⎧ ⎪⎨ ⎪⎩ , i = , i– m= hi {[K(ti,tm,Um) – K(ti,tm,um)] + [K(ti,tm–,Um–) – K(ti,tm–,um–)]}, i > . Lemma .Under the condition of Lemma ., for the remainder term Ri of the scheme (.)-(.), the estimate Lemma .Under the condition of Lemma ., for the remainder term Ri of the scheme () () the estimate Lemma .Under the condition of Lemma ., for the remainder term Ri of the scheme (.)-(.), the estimate (.)-(.), the estimate ∥R∥∞,ωN ≤CN– (.) ∥R∥∞,ωN ≤CN– (.) holds. ∥R∥∞,ωN ≤CN– (.) holds. Proof The remainder term of the scheme (.) can be rewritten as Ri = R() i + R() i + R() i , where Ri = R() i + R() i + R() i , where R() i = –h– i ti ti– (ξ – ti–) d dξ f ξ,u(ξ) dξ, (.) R() i = –h– i ti ti– (ξ – ti–) d dξ ξ  K ξ,s,u(s) ds dξ (.) (.) (.) and R() i = i m= ti ti– (tm–  – ξ) d dξ K ti,ξ,u(ξ) dξ. 5 Uniform error estimates First, we consider the τ < T/and estimate Ri on [,τ] and [τ,T] separately. Then τ = –α–ε lnε. In the layer region [,τ], we get First, we consider the τ < T/and estimate Ri on [,τ] and [τ,T] separately. Then τ = –α–ε lnε. In the layer region [,τ], we get R() i ≤C hi + α– exp –αti– ε + exp –αti ε , i = ,...,N/, (.) (.) by (.). Since by (.). Since hi = ti – ti–= –α–ε ln – (– ε)i N + α–ε ln – (– ε)(i – ) N ≤α–(– ε)N– ≤α–(– ε)N– ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 11 of 15 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 11 of 15 http://www.advancesindifferenceequations.com/content/2014/1/171 Page 11 of 15 http://www.advancesindifferenceequations.com/content/2014/1/171 http://www.advancesindifferenceequations.com/content/2014/1/171 and and exp –αti– ε + exp –αti ε = (– ε)N–, exp –αti– ε + exp –αti ε = (– ε)N–, exp –αti– ε + exp –αti ε = (– ε)N– it follows from (.) that it follows from (.) that R() i ≤α–CN–, i = ,...,N/. (.) R() i ≤α–CN–, i = ,...,N/. (.) It follows from (.) and (.) that It follows from (.) and (.) that It follows from (.) and (.) that R() i ≤Ch ≤Cα–(– ε)N–≤α–CN–, i = ,...,N/ (.) and R() i ≤Chi + Cε–hi ≤CN–, i = ,...,N/, (.) respectively. From (.), (.), and (.) for the region [,τ] we get \|Ri\| ≤CN–, i = ,...,N/. (.) \|Ri\| ≤CN–, i = ,...,N/. In the layer region [τ,T], \|u′(x)\| ≤C (or ε–exp(–αx/ε) ≤) by (.), and In the layer region [τ,T], \|u′(x)\| ≤C (or ε–exp(–αx/ε) ≤) by (.), and R() i ≤Ch, i = N/+ ,...,N. In view of the above discussion, we get In view of the above discussion, we get R() i ≤TCN–, i = N/+ ,...,N. (.) (.) Similarly, it is clear that Similarly, it is clear that R() i ≤TCN–, i = N/+ ,...,N (.) R() i ≤TCN–, i = N/+ ,...,N (.) and and R() i ≤TCN–, i = N/+ ,...,N. (.) (.) Combining the estimates (.), (.), and (.) for the region [τ,T], we get \|Ri\| ≤CN–, i = N/+ ,...,N. 5 Uniform error estimates (.) \|Ri\| ≤CN–, i = N/+ ,...,N. (.) Now we consider the case τ = T/. In this case, T/< –α–ε lnε. Therefore, for ti ∈[,τ] with (.), we can obtain similar results to that obtained above. For ti ∈(τ,T], since hi = h = (T – τ)/N = T/N, max i=N/+,...,N ti ti– ε–exp –αt ε dx ≤hε–exp –αT ε ≤ α–exp(–)h /T = Cα–exp(–)N–. ≤ α–exp(–)h /T = Cα–exp(–)N–. ce Equations 2014, 2014:171 Page 12 of 15 erenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 ce Equations 2014, 2014:171 Page 12 of 15 erenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 Page 12 of 15 erence Equations 2014, 2014:171 Page 12 of ndifferenceequations.com/content/2014/1/171 http://www.advancesindifferenceequations.com/content/2014/1/171 It follows from (.), (.), and (.) that It follows from (.), (.), and (.) that It follows from (.), (.), and (.) that R() i ≤C T + exp(–)α– N–, i = N/+ ,...,N, (.) R() i ≤Ch ≤CN–, i = N/+ ,...,N, (.) R() i ≤Ch + Cε–h ≤CN–, i = N/+ ,...,N, (.) R() i ≤C T + exp(–)α– N–, i = N/+ ,...,N, (.) R() i ≤Ch ≤CN–, i = N/+ ,...,N, (.) R() i ≤Ch + Cε–h ≤CN–, i = N/+ ,...,N, (.) R() i ≤C T + exp(–)α– N–, i = N/+ ,...,N, (.) R() i ≤Ch ≤CN–, i = N/+ ,...,N, (.) respectively. When we combine the estimates (.), (.), and (.), we get respectively. When we combine the estimates (.), (.), and (.), we get \|Ri\| ≤CN–, i = N/+ ,...,N. (.) From (.), (.), and (.), it is easy to see that (.) holds. □ Lemma .Under condition (.) and Lemma ., the solution zi of problem (.)-(.) satisﬁes Lemma .Under condition (.) and Lemma ., the solution zi of problem (.)-(.) satisﬁes ∥z∥∞,ϖN ≤max ≤i≤N \|Ri\|. (.) (.) ∥z∥∞,ϖN ≤max ≤i≤N \|Ri\|. Combining the two previous lemmas gives us the following main result. Theorem .Suppose that the conditions of Lemma .and (.) are satisﬁed and u is the solution of problem (.), (.). Then the following ε-uniform convergence result holds for the solution U of the diﬀerence problem (.), (.) on the mesh (.)-(.): Theorem .Suppose that the conditions of Lemma .and (.) are satisﬁed and u is the solution of problem (.), (.). Then the following ε-uniform convergence result holds for the solution U of the diﬀerence problem (.), (.) on the mesh (.)-(.): ∥U – u∥∞,ϖN ≤CN–. ∥U – u∥∞,ϖN ≤CN–. 5 Uniform error estimates Proof Using intermediate value theorem for the problem (.)-(.), we get Lhzi = εz¯t,i + ∂ ∂uf (ti,ui + γ zi)zi + hi  ∂ ∂uK(ti,ti,ui + γ zi)zi + hi  ∂ ∂uK(ti,ti–,ui–+ γ zi–)zi–+ K = Ri, i = ,,...,N, (.) z=  () = Ri, i = ,,...,N, (.) = Ri, i = ,,...,N, = , (.) where K = , i = , i– m= hi [ ∂ ∂uK(ti,tm,um + γ zm)zm + ∂ ∂uK(ti,tm–,um–+ γ zm–)zm–], i > . If we apply Lemma .to (.)-(.), then we see the validity of the inequality (.). □ If we apply Lemma .to (.)-(.), then we see the validity of the inequality (.). □ Combining the two previous lemmas gives us the following main result. 6 Numerical results In this section, we test the performance of the diﬀerence problem (.), (.). It is clear that the diﬀerence problem (.), (.) is a nonlinear problem. When we solve such problems, nonlinear equations arise in each step. There are several methods for solving these kinds ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 Page 13 of 15 of nonlinear equations. One of these methods is quasi-linearization. Quasi-linearization is a method like Newton’s method (see, e.g., []). This method amounts to linearizing the nonlinear terms in the nonlinear problems. A quasi-linearization procedure deﬁnes a sequence of linear problems whose solutions converge to that of the given nonlinear problems. For convergence of this method, one can refer to [, ]. If we use this method for the diﬀerence problem (.), (.), we obtain εU(n) ¯t,i + f ti,U(n–) i + ∂ ∂U f ti,U(n–) i U(n) i – U(n–) i + i m= hi  K ti,tm,U(n–) m + ∂ ∂U K ti,tm,U(n–) m U(n) m – U(n–) m + K ti,tm–,U(n–) m– + ∂ ∂U K ti,tm–,U(n–) m– U(n) m–– U(n–) m– (.) = , i = ,...,N, = , i = ,...,N, (.) U(n) = A. (.) U(n) = A. Acknowledgements h h d b Acknowledgements The author are indebted to Professor Gabil M Amiraliyev for various valuable suggestions and constructive criticism. Moreover, the author wishes to thank the anonymous referees for their very useful comments and suggestions. Received: 20 May 2013 Accepted: 4 June 2014 Published: 23 June 2014 Competing interests p g The author declares that he has no competing interests. p g The author declares that he has no competing interests. 6 Numerical results (.) (.) Here, we obtain the following iteration process: U(n) i = AiU(n) i–+ BiU(n–) i – Ci – Ki ε hi + ∂ ∂U f (ti,U(n–) i ) + hi  ∂ ∂U K(ti,ti,U(n–) i ) , i = ,,...,N, (.) U(n) = A, (.) U(n) i = AiU(n) i–+ BiU(n–) i – Ci – Ki ε hi + ∂ ∂U f (ti,U(n–) i ) + hi  ∂ ∂U K(ti,ti,U(n–) i ) , i = ,,...,N, (.) (.) (.) where Ai = ε hi , Bi = ∂ ∂U f ti,U(n–) i + ∂ ∂U K ti,ti,U(n–) i , Ci = f ti,U(n–) i + hi  K ti,ti,U(n–) i + K ti,ti–,U(n) i– , Ki = , i = , i– m= hi [K(ti,tm,U(n) m ) + K(ti,tm–,U(n) m–)], i > , Ai = ε hi , Bi = ∂ ∂U f ti,U(n–) i + ∂ ∂U K ti,ti,U(n–) i , Ci = f ti,U(n–) i + hi  K ti,ti,U(n–) i + K ti,ti–,U(n) i– , Ki = , i = , i– m= hi [K(ti,tm,U(n) m ) + K(ti,tm–,U(n) m–)], i > , and U() i is given. We apply the diﬀerence scheme (.), (.) to the following Volterra integro-diﬀerential equation: εu′(t) + u(t) + u(t) + t  u(s)ds = e– t ε –  εe– t ε + e– t ε +  ε with u() = . The exact solution of the equation is u(t) = e– t ε . Some computational results are presented in Table . 6 Numerical results We also calculate the experimental rate of uniform convergence p as follows: pN = ln(eN/eN) ln , Page 14 of 15 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 Table 1 Approximate errors eN and computed orders of convergence pN on ωN for vario values of ε and N ε N = 16 N = 32 N = 64 N = 128 N = 256 N = 512 10–1 0.048432 0.025428 0.013244 0.006847 0.003533 0.001797 0.93 0.94 0.95 0.95 0.97 10–2 0.049452 0.015768 0.013189 0.006691 0.003367 0.001691 0.94 0.97 0.98 0.99 0.99 10–3 0.050015 0.026007 0.013277 0.006723 0.003381 0.001697 0.94 0.97 0.98 0.99 0.99 10–4 0.050137 0.026058 0.013295 0.006730 0.003384 0.001698 0.94 0.97 0.98 0.99 0.99 10–5 0.050146 0.026062 0.013297 0.006733 0.003385 0.001698 0.94 0.97 0.98 0.99 0.99 10–6 0.050149 0.026063 0.013297 0.006734 0.003385 0.001698 0.94 0.97 0.98 0.99 0.99 Table 1 Approximate errors eN and computed orders of convergence pN on ωN for various values of ε and N Table 1 Approximate errors eN and computed orders of convergence pN on ωN for various values of ε and N where The obtained results show that the convergence rate of the diﬀerence scheme (.), (.) is essentially in accord with the theoretical analysis. The obtained results show that the convergence rate of the diﬀerence scheme (.), (.) is essentially in accord with the theoretical analysis. References 1. Angell, JS, Olmstead, WE: Singularly perturbed Volterra integral equations II. SIAM J. Appl. Math. 47, 1-14 (1987) 2. Hoppensteadt, FC: An algorithm for approximate solutions to weakly ﬁltered synchronous control systems and nonlinear renewal processes. SIAM J. Appl. Math. 43, 834-843 (1983) 1. Angell, JS, Olmstead, WE: Singularly perturbed Volterra integral equations II. SIAM J. Appl. Ma 1. Angell, JS, Olmstead, WE: Singularly perturbed Volterra integral equations II. SIAM J. Appl. Math. 47, 1 14 (1987) 2. Hoppensteadt, FC: An algorithm for approximate solutions to weakly ﬁltered synchronous control systems and nonlinear renewal processes. SIAM J. Appl. Math. 43, 834-843 (1983) 2. Hoppensteadt, FC: An algorithm for approximate solutions to weakly ﬁltered syn nonlinear renewal processes. SIAM J. Appl. Math. 43, 834-843 (1983) 3. Jordan, GS: A nonlinear singularly perturbed Volterra integrodiﬀerential equation of nonconvolution type. Proc. R. Soc. Edinb., Sect. A 80, 235-247 (1978) 4. Lodge, AS, McLeod, JB, Nohel, JA: A nonlinear singularly perturbed Volterra integrodiﬀerential equation occurring in polymer rheology. Proc. R. Soc. Edinb., Sect. A 80, 99-137 (1978) 4. Lodge, AS, McLeod, JB, Nohel, JA: A nonlinear singularly perturbe polymer rheology. Proc. R. Soc. Edinb., Sect. A 80, 99-137 (1978) p y gy , , ( ) 5. Angell, JS, Olmstead, WE: Singular perturbation analysis of an integrodiﬀerential equation modelling ﬁlament stretching. Z. Angew. Math. Phys. 36, 487-490 (1985) y y 5. Angell, JS, Olmstead, WE: Singular perturbation analys stretching. Z. Angew. Math. Phys. 36, 487-490 (1985) 5. Angell, JS, Olmstead, WE: Singular perturbation analys stretching. Z. Angew. Math. Phys. 36, 487-490 (1985) stretching. Z. Angew. Math. Phys. 36, 487-490 (1985) 6. Angell, JS, Olmstead, WE: Singularly perturbed Volterra integral equations. SIAM J. Appl. Math. 47, 1150-1162 (1987) 7. Bijura, AM: Singularly perturbed Volterra integro-diﬀerential equations. Quaest. Math. 25, 229-248 (2002) 6. Angell, JS, Olmstead, WE: Singularly perturbed Volterra integral equations. SIAM J. Appl. Math. 47, 1150-1162 (1987) 8. Doolan, EP, Miller, JJ, Schilders, WHA: Uniform Numerical Methods for Problems with Initial and Boundary Layer. Boole, Dublin (1980) 9. Farrel, PA, Hegarty, AF, Miller, JJH, O’Riordan, E, Shishkin, GI: Robust Computational Techniques for Boundary Layers. Chapman & Hall/CRC, Boca Raton (2000) 10. Miller, JJH, O’Riordan, E, Shishkin, GI: Fitted Numerical Methods for Singular Perturbation Problems. World Scientiﬁc, Singapore (1996) 11. Roos, HG, Stynes, M, Tobiska, L: Numerical Methods for Singularly Perturbed Diﬀerential Equation: Convection-Diﬀusion and Flow Problems. Springer, Berlin (1996) 11. 7 Conclusion A nonlinear Volterra integro-diﬀerential equation was considered. We solved this equa- tion by using a ﬁnite diﬀerence scheme on an appropriate graded mesh which is dense in the initial layer. We showed that the method shows uniform convergence with respect to the perturbation parameter for the numerical approximation of the solution. Numerical results which support the theoretical results were presented. ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 doi:10.1186/1687-1847-2014-171 Cite this article as: ¸Sevgin: Numerical solution of a singularly perturbed Volterra integro-differential equation. Advances in Diﬀerence Equations 2014 2014:171. References Roos, HG, Stynes, M, Tobiska, L: Numerical Methods for Singularly Perturbed Diﬀerential Equation: 11. Roos, HG, Stynes, M, Tobiska, L: Numerical Methods for Singularly Convection-Diﬀusion and Flow Problems. Springer, Berlin (1996) Convection-Diﬀusion and Flow Problems. Springer, Berlin (1996) Page 15 of 15 ¸Sevgin Advances in Diﬀerence Equations 2014, 2014:171 http://www.advancesindifferenceequations.com/content/2014/1/171 12. Roos, HG, Stynes, M, Tobiska, L: Robust Numerical Methods for Singularly Perturbed Diﬀerential Equations. Springer, Berlin (2008) 12. Roos, HG, Stynes, M, Tobiska, L: Robust Numerical Methods for Singularly Perturbed Diﬀerential Equations. Springer, Berlin (2008) 13. Kauthen, JP: Implicit Runge-Kutta methods for some integrodiﬀerential-algebraic equations. Appl. Numer. Math. 13, 125-134 (1993) 14. Kauthen, JP: Implicit Runge-Kutta methods for singularly perturbed integro-diﬀerential systems. Appl. Numer. Math. 18, 201-210 (1995) 15. Kauthen, JP: A survey of singularly perturbed Volterra equations. Appl. Numer. Math. 24, 95-114 (1997) 16. Salama, AA, Evans, DJ: Fourth order scheme of exponential type singularly perturbed Volterra integro-diﬀerential equations. Int. J. Comput. Math. 77, 153-164 (2001) 16. Salama, AA, Evans, DJ: Fourth order scheme of expon 16. Salama, AA, Evans, DJ: Fourth order scheme of expo equations. Int. J. Comput. Math. 77, 153-164 (2001) equations. Int. J. Comput. Math. 77, 153-164 (2001) 17. Horvat, V, Rogina, M: Tension spline collocation methods for singularly perturbed Volterra integro-diﬀerential and V lt i t l ti J C t A l M th 140 381 402 (2002) 17. Horvat, V, Rogina, M: Tension spline collocation methods for singularly p Volterra integral equations. J. Comput. Appl. Math. 140, 381-402 (2002) 17. Horvat, V, Rogina, M: Tension spline collocation met 17. Horvat, V, Rogina, M: Tension spline collocation methods for singularly perturbed Volterra integro-diﬀerential and Volterra integral equations. J. Comput. Appl. Math. 140, 381-402 (2002) olterra integral equations. J. Comput. Appl. Math. 140, 381-402 (200 Volterra integral equations. J. Comput. Appl. Math. 140, 18. Parand, K, Rad, JA: An approximation algorithm for the solution of the singularly perturbed Volterra 18. Parand, K, Rad, JA: An approximation algorithm for the solution of the singularly perturbed V integro-diﬀerential and Volterra integral equations. Int. J. Nonlinear Sci. 12, 430-441 (2011) , , , pp g g y p integro-diﬀerential and Volterra integral equations. Int. J. Nonlinear Sci. 12, 430-441 (2011) 19. Ramos, JI: Piecewise-quazilinearization techniques for singularly perturbed Volterra integro-diﬀe Appl. Math. Comput. 188, 1221-1233 (2007) 19. Ramos, JI: Piecewise-quazilinearization techni Appl. Math. Comput. 188, 1221-1233 (2007) 19. References Ramos, JI: Piecewise-quazilinearization techniques for singularly perturbed Volterra integro-diﬀerential equations. Appl. Math. Comput. 188, 1221-1233 (2007) Appl. Math. Comput. 188, 1221-1233 (2007) 20. Salama, AA, Bakr, AA: Diﬀerence schemes of exponential type for singularly perturbed Volterra integro-diﬀerential problems. Appl. Math. Model. 31, 866-879 (2007) problems. Appl. Math. Model. 31, 866-879 (2007) p pp , ( ) 21. Wu, S, Gan, S: Errors of linear multistep methods for singularly perturbed Volterra delay-integro-diﬀerential equations. M th C t Si l 79 3148 3159 (2009) p pp 21. Wu, S, Gan, S: Errors of linear multistep methods for singularly perturbed Volterra delay-integro-diﬀerential equat Math. Comput. Simul. 79, 3148-3159 (2009) 21. Wu, S, Gan, S: Errors of linear multistep methods for singularly perturbed Volterra delay-in Math. Comput. Simul. 79, 3148-3159 (2009) 21. Wu, S, Gan, S: Errors of linear multistep meth Math. Comput. Simul. 79, 3148-3159 (2009) Math. Comput. Simul. 79, 3148-3159 (2009) p 22. Linss, T: Layer-Adapted Meshes for Reaction-Convection-Diﬀu y 23. Amiraliyev, GM, ¸Sevgin, S: Uniform diﬀerence method for singularly perturbed Volterra integro-diﬀeren Appl. Math. Comput. 179, 731-741 (2006) 23. Amiraliyev, GM, ¸Sevgin, S: Uniform diﬀerence method for singularly perturbed Volterra integro-diﬀerential equa , ¸Sevgin, S: Uniform diﬀerence method for singularly perturbed Volt erential and Integral Calculus, vol. II. Interscience, New York (1936) 24. Courant, R: Diﬀerential and Integral Calculus, vol. II. Interscienc 25. Amiraliyev, GM, Mamedov, YD: Diﬀerence schemes on the uniform mesh for singularly perturbed pseudo-parabolic equations. Turk. J. Math. 19, 207-222 (1995) 25. Amiraliyev, GM, Mamedov, YD: Diﬀerence sc equations. Turk. J. Math. 19, 207-222 (1995) 26. Amiraliyev, GM, Kudu, M, Duru, H: Uniform diﬀerence method for a parameterized singular perturbation problem. Appl. Math. Comput. 175, 89-100 (2006) 27. Amiraliyev, GM: The convergence of a ﬁnite diﬀerence method on layer-adapted mesh for a singularly perturbed system. Appl. Math. Comput. 162, 1023-1034 (2005) 28. Ascher, UM, Mattheij, RM, Russell, RD: Numerical Solution of Boundary Value Problems for Ordinary Diﬀerential Equations. SIAM, Philadelphia (1995) 29. Cakir, M, Amiraliyev, GM: Numerical solution of a singularly perturbed three-point boundary value problem. Int. J. Comput. Math. 84, 1465-1481 (2007)
https://openalex.org/W3110919169	https://periodicos.ufpb.br/index.php/moringa/article/download/56517/32079	Portuguese	null	PROCEDIMENTOS EM SALA DE AULA	Moringa	2,020	cc-by	5,196	Keywords: Theatre game; Pedagogy; Acting. Procedures in the classroom: theatre games and pedagogy Daves Otani Raíra Rosenkjar Escola Superior de Artes Célia Helena - ESCH Daves Otani Raíra Rosenkjar Escola Superior de Artes Célia Helena - ESCH Resumo: O artigo reflete sobre o jogo teatral a partir de procedimentos na disciplina “Jogos e Pedagogia Teatral”, da “Escola Superior de Artes Célia Helena”, recorrendo a Kusnet, Huizinga e Bogart para pensar o intérprete e o olhar deste para o próprio percurso formativo e para a coordenação de processos pedagógicos em teatro. Palavras-chave: Jogos teatrais; Pedagogia; Atuação. Palavras-chave: Jogos teatrais; Pedagogia; Atuação. Abstract: The article reflects on the theatre games takes as a starting point the procedures in the discipline “Theatre Games and Pedagogy” at the “Escola Superior de Artes Célia Helena”. It reflects upon Kusnet, Huizinga and Bogart identifying ways of conducting the formative process of the interpreter and the latter's look on formative route, as well as for the coordination of pedagogical processes in theatre. Keywords: Theatre game; Pedagogy; Acting. João Pessoa, V. 11 N. 2 jul-dez/2020 João Pessoa, V. 11 N. 2 jul-dez/2020 141 Daves Otani • Raíra Rosenkjar A atividade teatral é transformadora por natureza, coloca o participante em relação a si mesmo e em relação ao outro, essa premissa é, a nosso ver, o que constrói uma atitude de observação que irá detonar processos de descoberta de si e das relações em grupo. Tal atitude pode tornar-se formativa, desde que partindo da intenção de fazer da atividade um mecanismo de busca. Busca por conceitos fundamentais e por analogias, por “aberturas” e contradições. O atuante – do jogo teatral – se coloca na posição de agente e material, ele é o que faz e o que permite “ser feito”. A ideia de regra – premissa do jogo – é tensionada pela ideia de criatividade – igualmente central no jogo –, e essa tensão, que para Huizinga funciona como impulso do jogar, pode levar o atuante a perceber que a presença cênica está condicionada a esse conflito: estrutura e subjetividade, regra e criatividade. Como explica Huizinga (2007, p. 4): “ultrapassa os limites da atividade puramente física e biológica. É uma função significante, isto é, encerra um determinado sentido. No jogo, existe alguma coisa “em jogo” que transcende as necessidades imediatas da vida e confere sentido à ação”. Contudo, o jogo carrega sobretudo relações com o espaço e com o tempo, pilares da teatralidade. O espaço do indivíduo, o espaço do coletivo, o tempo do indivíduo, o tempo do coletivo, o espaço do imaginário, o espaço real e concreto, o tempo da ficção e o tempo do relógio. O real e o ficcional, duas camadas sobrepostas que se articulam pela energia motora da ludicidade. Mas essa ideia de ficção não é clara, ou seja, literal, emoldurada numa história/estória; ela deve nascer da conexão do indivíduo e do grupo com a possibilidade da imaginação, ela – a ficção – é um espaço aberto pela atitude de imaginar. Nas aulas de “Jogos e Pedagogia Teatral”, na “Escola Superior de Artes Célia Helena”, começamos sempre explorando o ritmo e a ocupação do espaço em jogos simples. Exercícios estruturados, em um primeiro momento, para preparar as articulações e o corpo de forma geral, e, posteriormente, envolvendo também “músculos imaginativos”. Para Peter Book, a função da arte é produzir a partir da matéria potências abstratas, que nos coloquem em contato com as nossas subjetividades, com o inexplicável e invisível (BROOK, 2015). João Pessoa, V. 11 N. 2 jul-dez/2020 Como exemplo: uma roda, os atuantes passam uma bolinha imaginária, uma palma passa a “bolinha”, João Pessoa, V. 11 N. 2 jul-dez/2020 142 Procedimentos em sala de aula: jogos e pedagogia teatral em direção ao outro, como uma flecha, olho no olho; se a bola “para”, qualquer um a “pega” e continua. Com esse exercício, começamos a construir relações, criar linhas de tensões imaginárias de ação com intuito de afetar o atuante/ jogador. A “bola” é passada tal como foi recebida, de forma direta, receber já é passar para outro. O jogo assume um lugar de metáfora sobre a ação teatral, a construção de relação e a mudança de estados através da ação. Na mesma roda, os atuantes trocam de lugar no espaço um a um, um vai até o outro, olho no olho, quem vai ser substituído – vai ceder o seu lugar na roda – antecipa a saída, e troca com outro, sucessivamente. Primeiramente andando, depois, correndo. Primeiramente um a um e depois acumulando, quatro, cinco trocas simultâneas. Depois, girando no meio do espaço enquanto faz o percurso de ida até o novo lugar, proporcionando certo risco. Depois, saltando e batendo as mãos ao saltar, antes da troca; posteriormente, colocando sonoridades vocais no deslocamento e junto das palmas. Cada elemento novo vai sendo incorporado e o que se instaura é uma atmosfera de prontidão que vai se reorganizando a partir de novos “desafios”. A partir do momento em que ocorre o acúmulo de elementos ao até então “apenas aquecimento”, a “temperatura” do processo é crescente, ou seja, passa-se a construir uma atmosfera coletiva de concentração e envolvimento psicofísico com a atividade. A concentração de energia é pulsante conforme a divisão de foco, de linhas e tensões (BOGART, 2009), provenientes das relações construídas em estado de jogo. O exercício pedagógico exige que o mediador esteja atento às etapas do processo, não é possível começar a sequência mencionada acima correndo e saltando. Andar, olhar o lugar para o qual se vai e firmar cumplicidades – momento onde os atuantes se olham e um cede lugar para o outro – são etapas fundamentais na construção da confiança necessária para alcançar a ludicidade, qualidade que possibilita aos atuantes se conectarem com aspectos instintivos, como a intuição e o impulso. João Pessoa, V. 11 N. 2 jul-dez/2020 Nessa sequência inicial de jogos espaço-rítmico- temporais trabalhamos atitude atenção e linhas de tensão imaginárias entre os – momento onde os atuantes se olham e um cede lugar para o outro – são etapas fundamentais na construção da confiança necessária para alcançar a ludicidade, qualidade que possibilita aos atuantes se conectarem com aspectos instintivos, como a intuição e o impulso. Nessa sequência inicial de jogos espaço-rítmico- temporais, trabalhamos atitude, atenção e linhas de tensão imaginárias entre os atuantes: – momento onde os atuantes se olham e um cede lugar para o outro – são etapas fundamentais na construção da confiança necessária para alcançar a ludicidade, qualidade que possibilita aos atuantes se conectarem com aspectos instintivos, como a intuição e o impulso. Nessa sequência inicial de jogos espaço-rítmico- temporais, trabalhamos atitude, atenção e linhas de tensão imaginárias entre os atuantes: 143 Daves Otani • Raíra Rosenkjar Uma atitude é uma energia direcionada ao exterior. Se me sinto atraída por alguém, minha energia direcionada ao exterior é bastante específica e minhas respostas flutuam de acordo com a maneira que esta pessoa se relaciona comigo. Todas as minhas escolhas físicas, vocais e temporais são feitas em relação ao meu objeto de interesse. Quando o interesse acaba, a atitude muda. Em qualquer ocasião, minha atitude revela intenção e finalidade (…) No palco, o espaço entre os atores também deveria ser continuamente dotado de qualidade, potencial e até mesmo perigo. (BOGART, 2009, p. 30) Os exercícios em questão não são criados especialmente para a disciplina, é uma rede apreendida, transformada, colhida ao longo de experiências diversas tanto por vivências como por observações, curados para potencializar fatores pedagógicos de encontro com o conceito de “espect-ator” proposto por Augusto Boal para nomear o Homem que em sua natureza observa e age: “(...) teatro é a capacidade dos seres humanos, ausente nos animais, de se observarem a si mesmos em ação. Os humanos são capazes de se ver no ato de ver, capazes de pensar suas emoções e de se emocionar com seus pensamentos” (BOAL, 2015, prefácio). A observação e a ação assumem o caráter da experiência interligada, negando a segregação das mesmas, o atuante gere a ação e em paradoxo é gerido pela mesma, enquanto observa a ação gerada. Trabalhamos situações rudimentares e, progressivamente, estabelecemos relações com o fazer teatral. A construção da qualidade da ludicidade é alicerçada pela rotina e pelo hábito. João Pessoa, V. 11 N. 2 jul-dez/2020 Para Walter Benjamin, o “jogo” está ligado com “repetição”, diz ele: a essência do brincar não é um “fazer como se”, mas um “fazer sempre de novo”, transformação da experiência mais comovente em hábito (BENJAMIN, 2007). Na simplicidade das tarefas, nas regras fáceis de serem seguidas, o atuante pode encontrar espaços de subjetividade: cada giro é um giro, cada salto é um salto, cada sonoridade é um instante criativo potencial. Trabalha-se a atenção às relações: “atenção significa tensão – uma tensão entre um objeto e um observador ou tensão entre pessoas. É um modo de escutar. Atenção é uma tensão sobre o tempo” (BOGART, 2009, p. 31). 31). João Pessoa, V. 11 N. 2 jul-dez/2020 144 Procedimentos em sala de aula: jogos e pedagogia teatral Paralelamente, um vínculo se constrói de modo límpido, o indivíduo percebe- se atuando em uma atmosfera coletiva, comum. Assim como o coletivo incorpora as características e decisões individuais. Andamos depois simplesmente pelo espaço, formamos duplas, sem interromper o fluxo, sem parar para escolher, sem impor vontades e impulsos unicamente individuais, formam-se duplas, trios, quadras no andar. A escolha se faz ao acaso do espaço-tempo, fazem-se e desfazem-se grupos; desenvolve-se a ideia de fluxo, de ação contínua1 – conceito do sistema de Stanislavski (KUSNET, 1987). Um campo de atuação, sob regras específicas, delimitado no espaço-tempo. Para Huizinga, a delimitação de um espaço de jogo define o caráter do próprio, por isso a atividade ou ação de “andar pelo espaço” é proposta: é o momento que o jogador reconhece os limites espaciais onde o jogo ocorrerá (HUIZINGA, 2007). A ideia de simplicidade (rudimentaridade) atua sob a perspectiva do sentido de jogo enquanto elemento da cultura (Johan Huizinga), primordial na constituição do indivíduo. Acreditamos que, pela compreensão das noções básicas, fundantes, como coletividade, individualidade, e ação sobre o espaço e o tempo, é que podemos fundar uma experiência integrada com o jogo como propulsor da expressividade. Numa tentativa de resumir as características formais do jogo, poderíamos considerá-lo uma atividade livre, tomada como não séria e exterior à vida habitual, mas ao mesmo tempo capaz de absorver o jogador de maneira intensa e total. É uma atividade desligada de todo e qualquer interesse material, com a qual não se pode obter lucro, praticada dentro de limites espaciais e temporais próprios, segundo uma certa ordem e certas regras. Numa tentativa de resumir as características formais do jogo, poderíamos considerá-lo uma atividade livre, tomada como não séria e exterior à vida habitual, mas ao mesmo tempo capaz de absorver o jogador de maneira intensa e total. É uma atividade desligada de todo e qualquer interesse material, com a qual não se pode obter lucro, praticada dentro de limites espaciais e temporais próprios, segundo uma certa ordem e certas regras. Promove a formação de grupos sociais com tendência ao segredo e a sublinharem sua diferença em relação ao resto do mundo por meio de disfarces ou outros meios semelhantes. (HUIZINGA, 2007, p. 16) João Pessoa, V. 11 N. 2 jul-dez/2020 Promove a formação de grupos sociais com tendência ao segredo e a sublinharem sua diferença em relação ao resto do mundo por meio de disfarces ou outros meios semelhantes. (HUIZINGA, 2007, p. 16) Dois aspectos gostaríamos de ressaltar neste momento: a absorção “do jogador de maneira intensa e total” e a “formação de grupos sociais com tendência 1 Eugênio Kusnet “Ator método” se refere ao conceito apontando que nunca deixamos de agir, toda a ação presente é justificada por ações passadas e tem objetivos futuros. 1 Eugênio Kusnet “Ator método” se refere ao conceito apontando que nunca deixamos de agir, toda a ação presente é justificada por ações passadas e tem objetivos futuros. João Pessoa, V. 11 N. 2 jul-dez/2020 João Pessoa, V. 11 N. 2 jul-dez/2020 João Pessoa, V. 11 N. 2 jul-dez/2020 145 Daves Otani • Raíra Rosenkjar ao segredo”. Aspectos também relacionados com a potência cultural do jogar. O estado de jogo é uma fissura no cotidiano, por isso é experienciado de maneira intensa, carregando em si força ficcional e imaginativa responsável na criação de imagens, para isso, se organiza (normalmente) de maneira coletiva, buscando cúmplices que funcionem como aliados. Encontrar essas qualidades no princípio de um olhar mais cuidadoso para a formação do intérprete tem grande potencial para a compreensão das necessidades deste na cena. O jogo, na sua simplicidade, proporciona o encontro, de maneira evidente, na experiência prática e real, o atuante, conduzido ao pensamento sobre sua experiência, rapidamente registra o entendimento dos dois referidos aspectos (envolvimento intenso e tendência ao segredo). Por isso, falar sobre o exercício realizado é importante para a apreensão do conhecimento e investigação da experiência, falar, discutir, problematizar e compartilhar “segredos” de jogador. A leitura do artigo “Seis coisas que sei sobre o treinamento de atores”, de Anne Bogart, já citado, auxilia na condução das discussões e torna muito palpáveis os conteúdos práticos da disciplina, colaborando na construção de uma terminologia comum. Passamos, ainda no início do percurso, por um jogo de duplas, a partir do estímulo ação e reação. Um jogo essencialmente físico, sem parâmetros de situações, apenas agir e reagir aos estímulos próprios e do outro, resultantes do encontro no espaço-tempo da atualidade. Situações podem surgir ou não, e essa diferença da experiência também evidencia a qualidade de envolvimento (absorção intensa e total) ou não do atuante. Trocamos a dupla, a troca mostra que ação e reação estão diretamente ligados ao jogador que joga no tempo presente. Voltamos à primeira dupla e, agora, estimulamos a construção sequencial das estruturas que se repetem no jogo, construir um ciclo, dar forma aos materiais. Na sequência, pedimos que cada dupla mostre seu ciclo, seu formato. Sair do jogo livre, da improvisação livre e sem preocupação com sequência e repetição, e entrar no terreno da estrutura proporciona um “choque”. Tudo que naturalmente surgiu pode se perder na tentativa de cercear o material. Acrescente-se a dificuldade de ser visto, ser olhado, ter plateia. A presença de quem assiste aciona, mesmo que por intuição, a elaboração de discursos, linhas narrativas. A moldura cria a forma do João Pessoa, V. 11 N. João Pessoa, V. 11 N. 2 jul-dez/2020 2 jul-dez/2020 oão Pessoa, V. 11 N. 2 jul dez/202 146 Procedimentos em sala de aula: jogos e pedagogia teatral que é mostrado, nesse aspecto, a presença do outro que observa funciona como colaborador para a construção de sentidos. Há, no ato do sequenciamento, uma violência necessária ao ato criativo (BOGART, 2009, p. 32): Somente quando algo foi decidido que o trabalho pode começar. A determinação e a crueldade, que extinguiram a espontaneidade do momento, exigem que o ator comece um trabalho extraordinário: ressuscitar os mortos. O ator deve agora encontrar uma espontaneidade nova e mais profunda dentro desta forma estabelecida. Para mim, é por isto que os atores são heróis. Eles aceitam esta violência e trabalham com ela, trazendo habilidade e imaginação à arte da repetição. Por meio de exercícios que estimulam reflexões, levamos ao atuante, de forma simples, as questões que somam complexidade ao trabalho do intérprete. Entendemos que agir pedagogicamente passa por fundar dúvidas e nomear procedimentos e conceitos. O iniciante fascina-se com a violência necessária, o conceito é rapidamente apreensível no fazer. Torna-se bastante evidente a questão quando temos que, na hora, em jogo de improvisação, rapidamente transformar o jogo em estrutura e mostrá-lo aos colegas, a mudança de qualidade é brusca e perceptível. A sensação de “violentar-se”, e da necessidade de que seja assim, se revela. O uso do termo violência, obviamente negativo na sua acepção comum, causa estranheza e, assim, contraditoriamente, elucida um aspecto importante do fazer artístico. É preciso certo esforço para alterar as formas estratificadas de lidar com a criatividade, a liberdade criativa necessita ser direcionada num jogo de relações: consigo mesmo, como o parceiro de cena, com o diretor, com a plateia, com o espaço e com o tempo. Esse direcionamento tolhe a espontaneidade, violenta a posição de comodidade e coloca o atuante em relação ao jogo da cena, em direção a uma forma de agir coletiva e expressiva, que, ao fim de tudo, abre para profundas transformações. Mexer com o ritmo próprio e os hábitos cotidianos é algo violento, mas, sobretudo, transformador. “Paradoxalmente, são as restrições, a precisão e a exatidão que permitem a possibilidade de liberdade” (BOGART, 2009, p. 32). 147 Daves Otani • Raíra Rosenkjar Os jogos em sala de aula procuraram suscitar a percepção para o interesse, ferramenta fundamental no processo criativo (BOGART, 2009, p. 39). João Pessoa, V. 11 N. 2 jul-dez/2020 Como exemplo: é designado um número para cada aluno; seguindo a sequência numérica, um passa uma bolinha para o outro. Andando pelo espaço, o aluno deve estar atento ao número anterior e ao posterior ao seu, para que, no momento em que estiver chegando seu número, a linha de tensão imaginária entre ele e os outros jogadores esteja estabelecida. O atuante antecipa as suas ações e acumula atenções. Neste exercício, abordamos dois conceitos de Stanislavski trabalhados por Eugênio Kusnet: a ação contínua e o círculo de atenção. “Os atores devem se preocupar muito menos com a ação do momento que com a ação anterior e a ação posterior porque a ação do momento se realiza automaticamente se o ator realmente exerce a ação contínua” (KUSNET, 1987, p. 22). É um apontamento para proceder a antecipação das ações e reforçar o presente, dentro de uma regra de jogo simples. Essa atenção ao número anterior e ao número posterior, acumulada à atenção ao espaço do jogo (muitos atuantes se movendo ao mesmo tempo), cria camadas de atenção. Ainda que concentrado, ora no passar, ora no receber a bolinha, o atuante não deixa de se relacionar com os outros elementos, embora priorize sua atenção em determinadas tarefas – como na cena teatral, faz escolhas de acordo com a circunstância de cada momento. “Essa redução do quadro geral em apenas alguns detalhes e, vice-versa, a ampliação do campo de visualização, são exercidas no nosso trabalho através do uso de mais um elemento do método, denominado CÍRCULOS DE ATENÇÃO” (KUSNET, 1987, p. 49). Os círculos de atenção são, portanto, estratégias de concentração, o atuante se apega a detalhes, ou eventualmente a aspectos mais gerais, dependendo das necessidades que se apresentam, tanto nas circunstâncias da peça e das personagens, bem como nas circunstâncias da atualidade da cena. Ao exercitar o foco em “para quem passa a bolinha e de quem recebe”, de maneira rudimentar, o jogador trabalha com círculos de atenção. Para tal, o ator cria diferentes graus de interesse. No jogo da bolinha, propomos que os jogadores, em um segundo momento, escolham ou se aproximar ou se distanciar da bolinha; o jogador apenas João Pessoa, V. 11 N. 2 jul-dez/2020 148 Procedimentos em sala de aula: jogos e pedagogia teatral se movimenta quando a bolinha está no ar, passando de um jogador a outro. João Pessoa, V. 11 N. 2 jul-dez/2020 Mas é uma decisão antecipada, previamente tomada e que permanece contida antes de acontecer, trabalha-se tanto o interesse quanto o controle físico e a expansão das emoções (BOGART, 2009). Neste jogo, ao previamente escolher se distanciar-se ou se aproximar-se da bolinha – “segurando”, portanto, a ação decidida enquanto o outro não faz a passagem da bolinha –, o atuante está trabalhando tanto a ideia de antecipação quanto de controle e expansão. A decisão, já tomada, pulsa internamente até o momento de ser liberada e expandir. O talento mais especial de um ator é a capacidade de resistir, de conter, de domar, de conservar a energia em si, de concentrar. Com esta compressão, o ator brinca com a sensibilidade sinestésica dos espectadores e evita que eles prevejam o que está para acontecer. A cada instante, o objetivo é esconder do espectador a estrutura predeterminada e o desfecho. (BOGART, 2009, p. 34) O talento mais especial de um ator é a capacidade de resistir, de conter, de domar, de conservar a energia em si, de concentrar. Com esta compressão, o ator brinca com a sensibilidade sinestésica dos espectadores e evita que eles prevejam o que está para acontecer. A cada instante, o objetivo é esconder do espectador a estrutura predeterminada e o desfecho. (BOGART, 2009, p. 34) O talento mais especial de um ator é a capacidade de resistir, de conter, de domar, de conservar a energia em si, de concentrar. Com esta compressão, o ator brinca com a sensibilidade sinestésica dos espectadores e evita que eles prevejam o que está para acontecer. A cada instante, o objetivo é esconder do espectador a estrutura predeterminada e o desfecho. (BOGART, 2009, p. 34) É parte do treinamento a “chamada”, do coordenador do processo, para a atenção necessária a essas decisões que vão sendo tomadas no andamento do jogo, de não desistir, não se entregar a uma simples mecânica de repetição, a atitude de manter a linha de tensão, tecendo relações da ação pedagógica simples com práticas expressivas mais complexas. “Nesses jogos está implicado o processo de constantes decisões, seja para escolher os novos rumos do jogo, ou para avaliar os jogos anteriores, o que desenvolve nos participantes o interesse por cooperar e produzir em conjunto” (DESGRANGES, 2001, p. 89). Trabalhamos os princípios nomeados por Anne Bogart no artigo e também princípios – como citado – de Stanislavski e Kusnet. João Pessoa, V. 11 N. 2 jul-dez/2020 São guias, termos e conceitos que organizam nosso vocabulário, e sobre os quais procuramos entender os exercícios e jogos e fundar bases para o pensamento sobre o trabalho do intérprete. Variamos as estratégias e “iluminamos” a prática com tais conceitos-guias. Prática usual, a corda, nos ajuda a revisitar algumas destas bases. Seguindo a lógica do acúmulo de dificuldades e de procedimentos que buscam desenvolver a reflexão por meio da prática. Primeiro passar livremente pela corda, depois em intervalos regulares (até sem intervalo), então saltando e regularizando o intervalo, até uma João Pessoa, V. 11 N. 2 jul-dez/2020 149 Daves Otani • Raíra Rosenkjar sequência de saltar duas vezes, ora com um, ora com outro, na sequência da fila. Observa-se que, apesar dos procedimentos serem os mesmos de regularizar os intervalos, a sequência de saltar com um e depois com outro, desequilibra e desorienta o atuante. Posteriormente, invertemos o lado em que a corda se movimenta – novo desequilíbrio, nova desorientação. E, de novo, acumulando dificuldades com entradas pelos dois lados na corda invertida: “quando se acolhe o desequilíbrio, você entrará imediatamente em um território novo e desconhecido, onde se sentirá pequeno e inadequado em relação à tarefa por fazer. Mas os frutos deste compromisso serão abundantes” (BOGART, 2009, p. 38). O desequilíbrio e a desorientação aparecem ainda nos nossos exercícios de sonoridade e de toque (tato). Um exercício do comando pelos sons, em duplas (um de olhos fechados segue o som do parceiro), proporciona que o atuante se concentre no som de seu parceiro-comandante em meio a um “mar de sons” (toda a sala está simultaneamente neste exercício). Ao trabalhar o círculo de atenção ao som de seu parceiro, o atuante pode perceber que sua atenção, direcionada pelo seu interesse, separa o seu campo de escuta para o ponto necessário. Constrói-se gradativamente a linha de tensão entre a dupla, começamos bem próximos e com movimentos lentos e pausados em direções previsíveis, e ganhamos em distância e variações de ritmo, velocidade e volume. Experimenta-se uma desorientação inicial que se desconstrói em um processo de atenção e vínculo com o outro. Compreende-se que o comprometimento do coletivo desemboca no ritmo do jogo, e que a compreensão do ritmo coletivo determina o sucesso dos desafios individuais. Prosseguimos com outros jogos que abordam de forma diversa os conceitos que participam de nosso vocabulário comum. João Pessoa, V. 11 N. 2 jul-dez/2020 Em uma roda relativamente pequena – de oito a dez pessoas –, mantemos uma pessoa ao centro de olhos fechados. Ao toque dos outros, a pessoa ao centro deve reagir – deve tentar pegar quem a toca. Cria-se uma avalanche de toques. Observamos uma atitude competitiva e de poder sobre “o mais fraco”. Quem está ao centro de olhos fechados é um campo frágil para a “crueldade” do grupo. Questiona-se: porque tendemos a explorar a fragilidade, a nos divertir com a situação de desequilíbrio em que se encontra o João Pessoa, V. 11 N. 2 jul-dez/2020 João Pessoa, V. 11 N. 2 jul-dez/2020 oão Pessoa, V. 11 N. 2 jul dez/202 150 Procedimentos em sala de aula: jogos e pedagogia teatral outro? Estimula-se o princípio da oferenda de uma experiência: ofereça ao outro um desafio, jogue com sua desorientação e proporcione uma experiência de descoberta. Ainda que suscetível a certa crueldade – perceba essa faceta em sua personalidade que joga –, o que está à frente é a intenção do aprendizado, por meio de uma vivência lúdica e prazerosa, ainda que desafiadora: 1. Atitude, 2. Atenção, 3. Violência Necessária, 4. Controle Físico e Expansão das Emoções, 5. Desequilíbrio e Desorientação, 6. Interesse (BOGART, 2009). “Um desafio de quem coordena o processo é manter a tensão entre divertimento e aprendizagem” (DESGRANGES, 2001). A temática do som e do toque-tato se estende a outros encontros. Em um próximo exercício, exemplar, um dos alunos começa um movimento com som, uma estrutura simples que se repete em um ciclo contínuo. E os outros vão se encaixando, criando uma “engrenagem” sonora. Exercício conhecido e muito utilizado, a “máquina sonora” que se forma, a partir da entrada de uns nos movimentos dos outros, por acúmulo gradual, cria uma cumplicidade que pode evoluir para uma harmonia sonora e física potente. A simples experiência de que a proposta individual cria a rede de intersecções coletivas é esclarecedora acerca do papel da parte no todo, e vice-versa. Os exercícios são realizados sem fala, busca-se a observação de um ritmo comum, da sonoridade proveniente de maneira orgânica e natural. Consideramos sonoridades a partir do conceito de José Batista Dal Farra Martins: A concretização como som, ruído ou silêncio, nos movimentos de pulso e tom que imprimem e impulsionam imagens e conceitos, paixões e pensamentos, desejos e reflexões. João Pessoa, V. 11 N. 2 jul-dez/2020 A voz existe como potência mesmo quando se cala, pedir que ela não seja usada é pedir que o círculo de atenção seja expandido englobando o outro e as sonoridades provenientes desse encontro também. (DAL FARRA MARTINS, 2007, p. 9) A escolha pedagógica pela ausência do verbo é uma opção que busca a criação de uma atmosfera extraordinária (não cotidiana) de concentração que serve como amplificadora das atenções. João Pessoa, V. 11 N. 2 jul-dez/2020 151 Daves Otani • Raíra Rosenkjar Posteriormente, o percurso da disciplina avança para improvisações sobre temas simples. Primeiramente, apenas uma situação e a sua livre exploração em duplas. Duas cadeiras, dois intérpretes, imaginamos que um bateu atrás do carro do outro. O que acontece a partir daí, é cena, está nas mãos dos improvisadores. Na sequência, trabalhamos a improvisação pré-estruturada da mesma situação; ou seja, escolhe-se alguns alicerces que possam direcionar o acontecimento, ainda que guardando incertezas no percurso. Por exemplo: conhecemos-nos, namoramos e nunca mais nos vimos. Ao final, depois da conversa sobre o acidente, marcaremos um jantar para acertar a questão. O que acontece entre esses dois pontos? Nesse momento, já estamos abordando o jogo dentro de uma situação evidentemente teatral, cênica. A busca passa a ser a abordagem dos conceitos levantados ao longo dos jogos do primeiro período da disciplina dentro da relação do jogo na cena e no palco-plateia. Os conceitos ganham mais força – a cena traz mais clareza para conceitos que, de qualquer forma, estão voltados para o aprendizado específico da arte de apresentar conflitos no palco. Ao observar os conceitos relacionados à cena, o aluno pode estabelecer relações de seu aprendizado com a experiência do outro no âmbito da performance cênica. Assim, a experiência pedagógica com a disciplina “Jogos e Pedagogia Teatral”, na ESCH, busca apontar alguns aspectos fundamentais da atuação que serão desenvolvidos ao longo da graduação. “Abre-se a porta” da relação com o outro e com o espaço-tempo, fundam-se territórios, nos quais a experiência da atuação e seus aspectos formativos podem potencializar-se. Os conceitos apresentados são elementos que despertam a observação, possibilitando ao atuante ser o material que apresenta e o curador desse material que ele mesmo emoldura. Prepara-se, desta forma, o atuante para se debruçar sobre os processos formativos da atuação a partir de alguns princípios norteadores, e proporciona a criação de olhares, escutas, sensações e memórias de quem faz aprendendo e aprende fazendo. João Pessoa, V. 11 N. 2 jul-dez/2020 Recebido em 30/07/2019 Aceito em 26/05/2020 Recebido em 30/07/2019 Aceito em 26/05/2020 João Pessoa, V. 11 N. 2 jul-dez/2020 João Pessoa, V. 11 N. 2 jul-dez/2020 152 Procedimentos em sala de aula: jogos e pedagogia teatral BENJAMIN, Walter. Reflexões sobre a criança, o brinquedo e a educação. 2007. 3. Ed. São Paulo, Editora 34. BENJAMIN, Walter. Reflexões sobre a criança, o brinquedo e a educação. 2007. 3. Ed. São Paulo, Editora 34. BOAL, Augusto. Jogos para atores e não atores. 2015, São Paulo, Cosac Naify. BOGART, Anne. Seis coisas que sei sobre o treinamento de atores. Trad. Carolina Paganine. In: Urdimento. nº 12. Florianópolis. Programa de Pós- graduação em Teatro CEART/UDESC, 2009. BROOK, Peter. O espaço vazio. 2015, Rio de Janeiro, Apicuri. DAL FARRA MARTINS, José Batista. Percursos poéticos da voz. In: Sala Preta. nº 7 . São Paulo. USP, 2007. DESGRANGES, Flávio. Pedagogia do teatro: provocação e dialogismo. Hucitec, 2001. HUIZINGA, Johan. Homo Ludens. São Paulo, Perspectiva, 2007. KUSNET, Eugênio. Ator e método. 3. Ed. Rio de Janeiro, Mec – Serviço Nacional de Teatro, 1987. BENJAMIN, Walter. Reflexões sobre a criança, o brinquedo e a educação. 2007. 3. Ed. São Paulo, Editora 34. BOAL, Augusto. Jogos para atores e não atores. 2015, São Paulo, Cosac Naify. BOGART, Anne. Seis coisas que sei sobre o treinamento de atores. Trad. Carolina Paganine. In: Urdimento. nº 12. Florianópolis. Programa de Pós- graduação em Teatro CEART/UDESC, 2009. BOAL, Augusto. Jogos para atores e não atores. 2015, São Paulo, Cosac Naify. BOGART, Anne. Seis coisas que sei sobre o treinamento de atores. Trad. Carolina Paganine. In: Urdimento. nº 12. Florianópolis. Programa de Pós- graduação em Teatro CEART/UDESC, 2009. g ç BROOK, Peter. O espaço vazio. 2015, Rio de Janeiro, Apicuri. , p ç , , p DAL FARRA MARTINS, José Batista. Percursos poéticos da voz. In: Sala Preta. nº 7 . São Paulo. USP, 2007. p ç p DAL FARRA MARTINS, José Batista. Percursos poéticos da voz. In: Sala Preta. nº 7 . São Paulo. USP, 2007. , DESGRANGES, Flávio. Pedagogia do teatro: provocação e dialogismo. Hucitec, 2001. DESGRANGES, Flávio. Pedagogia do teatro: provocação e dialogismo. H 2001. HUIZINGA, Johan. Homo Ludens. São Paulo, Perspectiva, 2007. HUIZINGA, Johan. Homo Ludens. São Paulo, Perspectiva, 2007. KUSNET, Eugênio. Ator e método. 3. Ed. Rio de Janeiro, Mec – Serviço Nacional de Teatro, 1987. p KUSNET, Eugênio. Ator e método. 3. Ed. Rio de Janeiro, Mec – Serviço Nacional de Teatro, 1987. João Pessoa, V. 11 N. 2 jul-dez/2020
https://openalex.org/W4309796583	https://www.researchsquare.com/article/rs-1588156/latest.pdf	English	null	Antileishmania and immunomodulatory potential of cashew nut shell liquid and cardanol	Toxicology in vitro	2,023	cc-by	11,263	Antileishmania and cashew nut shell li Iuliana Marjory Martins Ribeiro Universidade Federal do Piauí Valéria Carlos de Sousa Universidade Federal do Piauí Rita de Cássia Viana de Carvalho Universidade Federal do Piauí Maisa de Sousa dos Santos Universidade Federal do Piauí José Arimatéia de Oliveira Nery Neto Universidade Federal do Piauí Danielly Silva de Melo Universidade Federal do Piauí Letícia Soares de Araújo Teixeira Universidade Federal do Piauí Maria das Graças Lopes Citó Universidade Federal do Piauí Arkellau Kenned Silva Moura Universidade Federal do Piauí Fernando Aécio de Amorim Carvalho Universidade Federal do Piauí Michel Muálem Moraes Alves (  mu Universidade Federal do Piauí Ivete Lopes de Mendonça Universidade Federal do Piauí Antileishmania and cashew nut shell li Iuliana Marjory Martins Ribeiro Universidade Federal do Piauí Valéria Carlos de Sousa Universidade Federal do Piauí Rita de Cássia Viana de Carvalho Universidade Federal do Piauí Maisa de Sousa dos Santos Universidade Federal do Piauí José Arimatéia de Oliveira Nery Neto Universidade Federal do Piauí Danielly Silva de Melo Universidade Federal do Piauí Letícia Soares de Araújo Teixeira Universidade Federal do Piauí Maria das Graças Lopes Citó Universidade Federal do Piauí Arkellau Kenned Silva Moura Universidade Federal do Piauí Fernando Aécio de Amorim Carvalho Universidade Federal do Piauí Michel Muálem Moraes Alves (  mu Universidade Federal do Piauí Ivete Lopes de Mendonça Universidade Federal do Piauí License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Research Article Page 1/28 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Page 2/28 Abstract Leishmaniasis is a neglected disease caused by protozoa of the genus Leishmania, endemic in 98 countries, constituting a major public health problem. Conventional treatments cause serious adverse effects, poor tolerance, development of resistant strains, and are costly. Natural products have been investigated in the search for therapeutic alternatives. The cashew nut shell liquid (CNSL) is a natural source of phenolic compounds, showing antioxidant, anti-inflammatory, antimicrobials, antitumors, larvicides and insecticides, with cardanol (CN) being considered one of the most important and promising technical components. This study aimed to evaluate antileishmania, cytotoxic and immunomodulatory activities of CNSL and CN. The substances showed antileishmania potential, with values of mean inhibitory concentration (IC50) of CNSL and CN against Leishmania infantum: 148.12 and 56.74 µg/mL; against Leishmania braziliensis: 85.71 and 64.28 µg/ml; against Leishmania major: 153.56 and 122.31 µg/mL, respectively. The mean cytotoxic concentrations (CC50) of CNSL and CN were 37.51 and 31.44 µg/mL, respectively. CNSL and CN significantly reduced the percentage of infected macrophages, with a selectivity index (SI) > 20 for CN. CNSL and cardanol caused an increase in phagocytic capacity and lysosomal volume, however, they did not exhibit significant induction of nitric oxide synthesis. Survival rates of Zophobas morio larvae at doses of 3; 30 and 300 mg/Kg were: 85%, 75% and 60% in contact with CNSL and 85%, 60% and 40% in contact with CN, respectively. There was a significant difference between the survival curves of larvae when treated with CN, demonstrating a significant acute toxicity for this substance. Additional investigations are needed to evaluate these substances in the in vivo experimental infection model. Introduction Leishmaniasis is a neglected disease caused by protozoa of the Leishmania genus, constituting a major public health problem due to its high incidence and lethality (Alvar et al. 2012; Benitez et al. 2018). In the year 2020, leishmaniasis was endemic in 98 of the 200 countries and territories that reported it to the WHO, and it is estimated that between 700,000 to 1 million new cases of cutaneous leishmaniasis and 50,000 to 90,000 new cases of visceral leishmaniasis occur worldwide each year (WHO 2021). The etiological agent of the disease is more than 20 species of the genus Leishmania, classified in the subgenus Viannia and Leishmania, transmitted by approximately 70 different species of sandflies of the genera Lutzomyia and Phlebotomus. The disease exhibits a variety of clinical manifestations, the two main ones being visceral leishmaniasis (VL) and tegumentary leishmaniasis (TL), and this, in turn, is divided into four subtypes: localized cutaneous leishmaniasis (LCL), diffuse cutaneous leishmaniasis (DCL), mucocutaneous leishmaniasis (MCL) and disseminated leishmaniasis (DL) (Akhoundi et al. 2016; Akhoundi et al. 2017). The treatment of leishmaniasis is performed with pentavalent antimonials, amphotericin B, The treatment of leishmaniasis is performed with pentavalent antimonials, amphotericin B, paromomycin, miltefosine, sodium stibogluconate, meglumine antimoniate and pentamidine (Blanco e Nascimento-Júnior 2017; Kevric et al. 2015). However, the use of these drugs requires long periods of paromomycin, miltefosine, sodium stibogluconate, meglumine antimoniate and pentamidine (Blanco e Nascimento-Júnior 2017; Kevric et al. 2015). However, the use of these drugs requires long periods of Page 3/28 administration, resulting in serious adverse effects, low tolerance and development of resistant strains to the treatment, contributing to the ineffectiveness of therapeutic regimens, in addition to presenting high costs (Bapela et al. 2017; Souza-Silva et al. 2015). The investigation of the pharmaceutical potential of natural products is the main strategy for discovering new drugs that can be less costly and less toxic than conventional drugs (Tiuman et al. 2011). Plants are considered an important source of natural products that can be successfully exploited for the development of new drugs with antileishmania activity (Funari et al. 2016; González-Coloma et al. 2011; Machado et al. 2014; Mansour et al. 2013; Torres et al. 2014). Compounds such as alkaloids, phenolics, terpenoids and flavonoids have been extensively studied for their potential antileishmania activity (Silva et al. 2014). Introduction Material And Methods Substances used Introduction Among the widely used medicinal plants, cashew (Anacardium occidentale L.) stands out and one of its main by-products, cashew nut shell liquid (CNSL), has been used for decades in traditional medicine in countries in South America, Africa and Asia (Ayyanar and Ignacimuthu 2009; Da Silva et al. 2018; Kudi et al. 1999). CNSL is a product with little commercial value, but with high technological potential due to its phenolic constitution and its various biological characteristics, such as anti-inflammatory, antimicrobial, antioxidant, antitumor, larvicides and insecticides, exhibiting great therapeutic potential (Kubo et al. 1993a; Oliveira et al. 2010; Wu et al. 2011), constituting a natural source of phenolic compounds, such as anacardic acid, cardanol, cardol and 2-methylcardol (Mazzetto et al. 2009). CNSL is a product with little commercial value, but with high technological potential due to its phenolic constitution and its various biological characteristics, such as anti-inflammatory, antimicrobial, antioxidant, antitumor, larvicides and insecticides, exhibiting great therapeutic potential (Kubo et al. 1993a; Oliveira et al. 2010; Wu et al. 2011), constituting a natural source of phenolic compounds, such as anacardic acid, cardanol, cardol and 2-methylcardol (Mazzetto et al. 2009). CNSL derivatives are also widely explored in isolation, mainly for their properties such as antibacterial, antioxidant, antifungal, antitumor, among others (Amorati et al. 2001; Chen et al. 1998; Hemshekhar et al. 2012; Mazzetto et al. 2009; Muroi e Kubo 1996). Cardanol (CN) is the main constituent of technical CNSL and is considered one of the most important and promising components. As it is a by-product of the nut industry, any improvement in the CN, whether in concentration and/or separation, is effectively characterized as a technological innovation (Mazzetto et al. 2009). Thus, the CNSL demonstrates wide utility for pharmacology, and therefore further studies of its isolated metabolites, their mechanisms of action, as well as toxicological aspects are needed (Araújo et al. 2020). Given the above, considering the impact of leishmaniasis mainly in developing countries, the urgency for more effective compounds with fewer adverse effects, and the broad therapeutic potential of CNSL, is strategic to invest in the search for new pharmacological properties of this product and its metabolites, as a potential source of medication for the treatment of leishmaniasis. Thus, the objective of this study was to evaluate the immunomodulatory and antileishmania activities of the CNSL and its main constituent, the CN, in order to propose a search for a new therapeutic strategy. Material And Methods Substances used Substances used Page 4/28 Schneider’s culture medium, DMEM® medium, FBS, MTT, resazurin, acridine orange dye, propidium iodide and the antibiotics Penicillin and Streptomycin from Sigma Chemical (Sigma-Aldrich Brazil). Amph B (90%), fast panoptic® was acquired from Cristália (São Paulo, SP). CNSL and CN were from the Organic Geochemistry Laboratory (LAGO/UFPI) and were diluted in DMSO at a concentration of 80 mg/mL for the experiments. Schneider’s culture medium, DMEM® medium, FBS, MTT, resazurin, acridine orange dye, propidium iodide and the antibiotics Penicillin and Streptomycin from Sigma Chemical (Sigma-Aldrich Brazil). Amph B (90%), fast panoptic® was acquired from Cristália (São Paulo, SP). CNSL and CN were from the Organic Geochemistry Laboratory (LAGO/UFPI) and were diluted in DMSO at a concentration of 80 mg/mL for the experiments. Parasites and cells Strains of Leishmania infantum (MHOM / 5745), Leishmania braziliensis (10CL566) and Leishmania major (MHOM / IL / 80 / Friendlin) were obtained from the Medicinal Plants Research Center of Federal University of Piauí. Parasites were grown in supplemented Schneider’s medium (10% heat-inactivated FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin at 26 °C) (Carneiro et al. 2012; Valadares et al. 2011). Murine macrophages were collected from the peritoneal cavities of male and female BALB/c mice (4–5 weeks old; Medicinal Plants Research Center, UFPI, Brazil), and cultivated in RPMI 1640 medium (10% heat-inactivated FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C and 5% CO2). All protocols were approved by the Animal Use Ethics Committee (CEUA-PI nº 640/2019). Invertebrates Zophobas morio larvae used to assess acute toxicity came from the Physiology Laboratory of the Department of Veterinary Morphophysiology/UFPI. The larvae were kept in plastic organizer boxes measuring 60 x 40 x 80 cm with water and feed ad libidum, at an ambient temperature of 25°C ± 2°C. Larvae weighing between 100 and 200 mg, light and uniform in color, without signs of melanization were selected for the experiment. Investigation of CNSL- and CN-induced activity on macrophages infected by Leishmania infantum and the calculation of selective index As L. infantum is a species that causes zoonotic leishmaniasis and that develops the visceral form of the disease, considered the most severe form (Steverding 2017), the activity of CNSL and CN on L. infantum promastigotes was evaluated. Macrophages (2 × 105 cells/mL) were harvested in 24-well plates containing sterile round coverslips at 13mm and supplemented RPMI 1640 medium (10% inactivated FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin). Culture plates were incubated at 37 °C and 5% of CO2 for 3h. Adhered macrophages were then incubated with a new medium containing axenic amastigotes at a ratio of 10 amastigotes per 1 macrophage at 5% CO2 and 37 °C for 4h. The medium was subsequently aspirated in order to remove non-internalized parasites, and the wells were washed with 0.01 M phosphate buffered saline (PBS). The infected macrophages were then incubated with CNSL and CN at 3.125; 6.25 and 12.5 μg/mL (non-toxic concentrations on host cells) or Amph B at 0,5; 1 and 2 μg/mL. After this period, the coverslips were removed and stained with Panoptic staining kit. For each treatment, the number of infected macrophages and the parasite load (survival index, obtained by counting the number of parasites in 100 macrophages) were counted using optical microscopy (Carneiro et al. 2012). Furthermore, the selectivity indexes of CNSL, CN, and Amph B were determined by the ratio of the mean CC50 against macrophages to the mean IC50 against macrophage-internalized amastigote forms of L. infantum (Oliveira et al. 2017). Investigation of activity against promastigote forms of L. infantum, L. braziliensis andL. major Promastigote forms of L. infantum, L. braziliensis andL. major in the late log phase (1 × 106 leishmania/100 μL of medium) were plated in 96-well culture plates containing supplemented Schneider’s medium. Then, CNSL and CN (6.25, 12.5, 25, 50, 100, 200; 400 and 800 μg/mL) were added, and the plates were incubated during 48h in a BOD (biochemical oxygen demand) incubator at temperature of 26 °C. Remaining 6h to the end of this period, 20 μL of resazurin (1 × 10−3 mol/L) was added. Afterwards, the absorbances were read in a BioTek microplate reader (model ELx800) at a wavelength of 550 nm. The results were expressed as inhibition of parasite growth (%) (Soares et al. 2007; Valadares et al. 2011). The negative control was the Schneider’s medium with promastigotes (1 × 106 cells/well) and for the positive control, amphotericin B (2 μg/mL). The cell viability was considered as 100% for the parasite. The blank was read for each concentration and control in order to avoid interference of absorbance of medium of other compounds (Soares et al. 2007; Valadares et al. 2011). Page 5/28 Page 5/28 Evaluation of cytotoxicity in macrophages Cytotoxicity evaluation was carried out in 96-well plates using the MTT assay. Macrophages (2 × 105 per well) were incubated in 100 μL of supplemented RPMI 1640 medium at 37 °C and 5% CO2 for 4h. Non- adherent cells were removed by washing with RPMI 1640 medium. Then, CNSL and CN were diluted in supplemented RPMI 1640 medium, and added at concentrations of 6.25, 12.5, 25, 50, 100, 200; 400 and 800 μg/mL, followed by incubation at 37 °C with 5% CO2 for 2 days. The cytotoxicity of Amph B was assessed at concentration of 0.2 μg/mL. Afterwards, cytotoxicity was assessed by adding MTT (5 mg/mL). The supernatant was discarded, and the formazan crystals were dissolved by addition of 100 μL of DMSO. Finally, absorbance at 550 nm was measured using a BioTek (ELx800) plate reader (Oliveira et al. 2017). Investigation of acute toxicity on Z. morio larvae To assess acute toxicity, 10 larvae of Z. moriowere used for each concentration tested, in triplicate. CNSL and CN were applied to the larvae in a volume of 10 μL at doses of 3; 30 and 300 mg/kg with the aid of a Hamilton syringe (701 N, manometer 26, Capacity 100 μL) in the hemocoel, in the second or third sternite visible above the legs, in the ventral portion. Larvae were incubated at room temperature in petri dishes containing breeding diet and the number of dead larvae was evaluated after 48h. To establish the death of the larvae, a visual check of each individual was performed for the presence of myelination and/or response to physical stimuli after gentle touch (de Souza et al. 2015). Nitrite measurement Non-infected macrophages or infected by L. infantum were obtained as described previously, and then incubated with CNSL and CN (3.125; 6.25 e 12.5 μg/mL), or Amph B (0.2 μg/mL) at 37 °C and 5% de CO2 for 24h. The lipopolysaccharide (LPS) from E. coli (2 μg/mL) was used as positive control. The standard curve was prepared with sodium nitrite in RPMI medium at varying concentrations of 1, 5, 10, 25, 50, 75, 100, and 150 μM diluted in RPMI 1640 medium. After 24h, the supernatants were transferred, and then incubated with equal parts of Griess reagent. Thereafter, the absorbances were read in a BioTek (ELx800) plate reader at 550 nm (Soares et al. 2007). Phagocytic capability Macrophages (2 × 105/well) were incubated with CNSL and CN (6.25; 12.5; 25; 50 e 100 μg/mL), or Amph B (0.2 μg/mL) in a 96-well plate for 48 h at 37 °C and 5% de CO2. After 48h, 10 μL of zymosan-stained NR solution was added for 30 min. Next, the phagocytic process was interrupted adding 100 μL of Baker’s fixative solution during 30 min. Then, the wells were washed with 0.9% saline, and 100 μL of extractive solution were added. After solubilization in a Kline shaker, the absorbances were read in a BioTek (ELx800) plate reader at 550 nm (Grando et al. 2009). Lysosomal activity Page 6/28 Macrophages (2 × 105/well) were incubated with CNSL and CN (6.25; 12.5; 25; 50 e 100 μg/mL), or Amph B (0.2 μg/mL) in a 96-well plate at 37 °C and 5% de CO2. After 48h, 10 μL of neutral red stock solution were added for 30 min. Then, the supernatant was discarded, the wells were washed with 0.9% saline at 37 °C, and 100 μL of extractive solution were added in order to solubilize the neutral red present within the Page 6/28 Page 6/28 lysosomal secretory vesicles. After 30 min on a Kline shaker, the absorbances were read in a BioTek (ELx800) plate reader at 550 nm (Bonatto et al. 2004). Statistical analyses All tests were performed in triplicate in three independent experiments. The mean inhibitory concentration (IC50) and the mean cytotoxic concentration (CC50) with confidence limits of 95% were determined by regression of probits using the software SPSS 13.0. The selectivity index was calculated as the ratio between CC50 and IC50 of internalized amastigotes. The survival curve of Z. morio larvae was plotted by Kaplan-Meier analysis and the results were analyzed by the Log Rank test to assess the level of relationship between the substances regarding the acute toxicity parameter. Analysis of variance ANOVA followed by Bonferroni’s test was performed, taking the value of p < 0.05 as the maximum level of statistical significance. Page 7/28 Page 7/28 Antileishmania activity assay CNSL and CN demonstrated antileishmania potential, with their action dependent on concentration. It was possible to observe activity against all Leishmania species tested, where at the concentration of 800 μg/mL there was 100% growth inhibition of the promastigote forms of L. infantum, L. braziliensis and L. major in the presence of CNSL and L. infantum and L. braziliensis in the presence of CN. CN showed about 90% growth inhibition of promastigote forms of L. major at a concentration of 800 μg/mL. Amphotericin B inhibited the growth of promastigote forms of L. braziliensis by about 90% and by L. infantum and L. major by about 80%, both at a concentration of 2 μg/mL. The IC50 values of CNSL and CN against L. infantum, L. braziliensis and L. major are shown in Table 1. Table 1 Antileishmania activity of CNSL, cardanol (CN) and amphotericin B (Amph B). Antileishmania activity of CNSL, cardanol (CN) and amphotericin B (Amph B ileishmania activity of CNSL, cardanol (CN) and amphotericin B (Amph B). y , ( ) p ( p ) Compounds L. infantum L. braziliensis L. major IC50 µg/mL a IC50 µg/mL a IC50 µg/mL a CNSL 148.12±0.08 85.71±0.11 153.56±0.08 CN 56.74±0.06 64.28±0.06 122.31±0.05 Amph B 2.46±0.07 2.94±0.05 0.59±0.13 a Inhibition Concentration 50 Citotoxicity assessment When acting on macrophages, CNSL demonstrated significant cytotoxicity from a concentration of 12.5 μg/mL, with a CC50 value of 37.51 μg/mL, while CN significantly reduced macrophage viability from the concentration of 6.25 μg/mL, resulting in a CC50 value of 31.44 μg/ml. Amph B showed high toxicity on murine macrophages, with a CC50 of 8.75 μg/mL (Fig. 1; Table 2). Effects of CNSL and CN against infection of macrophages by L. infantum and Selectivity Index (SI) Table 2 Cytotoxic effect on mammalian cells and calculated selectivity index values for CNSL, cardanol (CN) and amphotericin B (Amph B). Compounds Macrophages Intramacrophagic amastigotes (L. infantum) SIm c L. infantum CC50 µg/mLa IC50 µg/mL b CNSL 37,51±0,04 4,63 8,10 CN 31,44±0,04 1,42 22,14 Amph B 8,75±0,02 d 0,68 12,86 c Selectivity index for amastigotes internalized in macrophages (CC50/IC50) d Alves et al. (2017) c Selectivity index for amastigotes internalized in macrophages (CC50/IC50) d Alves et al. (2017) Effects of CNSL and CN against infection of macrophages by L. infantum and Selectivity Index (SI) In evaluating the activity of CNSL and CN against the intracellular amastigote form, they presented IC50 of 4.63 and 1.42 μg/ml, respectively. Amph B presented IC50 of 0.68 μg/ml on amastigotes internalized in macrophages (Table 2). CNSL and CN were able to reduce the percentage of macrophages parasitized by L. infantum (Fig. 2). Page 8/28 The negative control obtained approximately 87% of parasitized cells, while concentrations of 0.5; 1.0 and 2.0 μg/mL of Amph B reduced this number by about 73, 70 and 50%, respectively. The reduction of parasitized cells treated with CNSL and CN was dependent on the concentration used. Cells treated with CNSL showed parasitism around 79, 77 and 64%, and with CN the parasitism was approximately 78, 79 and 71%, after being treated at concentrations corresponding to 3.125; 6.25 and 12.5 μg/ml, respectively (Fig. 3a). and 71%, after being treated at concentrations corresponding to 3.125; 6.25 and 12.5 μg/ml, respectively (Fig. 3a). Analyzing the survival index, the control obtained an average of 9.0 amastigotes/macrophage. Amph B at concentrations of 0.5; 1.0 and 2.0 μg/mL reduced this amount of parasites to approximately 4.8; 3.9 and 2.1 amastigotes/macrophage, respectively. In the treatment with CNSL, the amount of amastigotes reduced, depending on the concentration, to 5.2 amastigotes/macrophages at a concentration of 3.125 μg/mL; at the concentration of 6.25 μg/mL, this amount decreased to approximately 3.9 amastigotes/macrophages; when treated with 12.5 μg/mL, the survival index was reduced to 2.3 amastigotes/macrophages. For CN treatment, the amount of amastigotes/macrophage corresponding to concentrations of 3.125; 6.25 and 12.5 µg/ml was 3.4; 3.1 and 2.3 amastigotes/macrophage, respectively (Fig. 3b). Cytotoxicity in murine peritoneal macrophages and activity against the intracellular amastigote form of L. infantum were used to determine the selectivity index (SI), whose value represents how much the substance is more toxic to the parasite than to macrophages. CNSL, CN and Anf B showed more selectivity for L. infantum forms and promastigotes than for murine macrophages. The results obtained indicate that CNSL is approximately 8.1 times more selective for protozoa than for mammalian cells, while CN showed 22 times greater selectivity for L. infantum than for macrophages, this value being higher than that presented by Amph B, which proved to be 12.86 times more selective for the parasite (Table 2). Table 2 Cytotoxic effect on mammalian cells and calculated selectivity index values for CNSL, cardanol (CN) and amphotericin B (Amph B). Determination of lysossomal activity and phagocytic capability Page 9/28 Page 9/28 The results regarding macrophage activation parameters, such as lysosomal volume increase and phagocytosis were evaluated based on the retention of neutral red and Zymozan particles by macrophages. CN was able to significantly increase the lysosomal volume of macrophages at concentrations of 3.125 μg/mL, 6.25 and 12.5 μg/mL, while in CNSL there was no retention of neutral red in the secretory vesicles of macrophages (Fig. 4a). In evaluating the phagocytosis capacity of Zymozan, the tested substances significantly induced an increase in phagocytic capacity, this induction being at concentrations of 6.25 and 12.5 μg/mL for CNSL and 12.5 μg/mL for CN (Fig. 4b). Measurement of nitrite production The production of nitric oxide (NO), a macrophage activation index, was quantified by measuring nitrite concentrations by incubating macrophages with CNSL and CN in the absence and presence of promastigote forms of L. infantum. As a result, none of the substances demonstrated a significant increase in nitrite synthesis in the absence (Fig. 5a) and presence (Fig. 5b) of Leishmania at all concentrations tested. The bacterial lipopolysaccharide from Escherichia coli (LPS) was used as a positive control, demonstrating a high capacity in inducing nitrite synthesis. Acute toxicity on Z. morio larvae The survival profiles of Z. morio larvae when subjected to contact with CNSL and CN can be seen in Fig. 6, demonstrating concentration-dependent action. The results showed, after 48 h, that the larvae survival rates against CNSL were approximately 85%, 75% and 60% at doses of 3, 30 and 300 mg/kg, respectively (Fig. 6a), while the survival rates against to CN were approximately 85%, 60% and 40% at doses of 3, 30 and 300 mg/kg, respectively (Fig. 6b). After the toxicity test, it was found that there is a statistically significant difference (p < 0.05) between the CN survival curves, demonstrating a significant acute toxicity of this substance. Discussion CNSL is a natural source of phenolic compounds, which have a long aliphatic chain of fifteen carbons, in the meta position in relation to the hydroxyl, which can be saturated (C15H31) and/or unsaturated with one (C15H29), two (C15H27) and three (C15H25) unsaturations. CNSL is classified as natural (extracted by solvent), consisting of anacardic acid (60–65%), cardol (15–20%), cardanol (10%), and traces of methyl cardol, and technical (submitted to high temperatures), composed mainly of cardanol (60–65%), cardol (15–20%), polymeric material (10%), and traces of methyl cardol (Mele and Vasapollo 2008). CN is the main constituent of technical CNSL, and its derivatives have demonstrated antibacterial, antioxidant, antifungal and antitumor activities, in addition to hydrophobicity (Amorati et al. 2001; Chen et al. 1998; Mazzetto et al. 2009). Page 10/28 CNSL and CN showed significant activity against promastigote forms of L. infantum, L. braziliensis and L. major after 48 hours of incubation. Although pharmacological data regarding the antileishmania activity of CNSL and its main constituents are scarce, other parts of A. occidentale have demonstrated antileishmania activity, corroborating our study. Results from França et al. (1993), showed in vitro activity of the hydroaucolic extract of the stem bark (at concentrations of 7.5 and 15 mg/mL) of A. occidentale on promastigote forms of L. braziliensis. In a study using ethanol extract from the leaves of A. occidentale, activity on promastigotes and amastigotes of L. amazonensis was demonstrated, showing growth inhibition of 5.4% and 32.3%, respectively, both at a concentration of 100 µg / mL (Luize et al. 2005). This study showed better results than those by Braga et al. (2007) using the stem bark extract of A. occidentale, where no activity against promastigote forms of L. amazonenses and L. infantum was found, resulting in IC50 values >200 µg/mL for both. In a study with compounds isolated from the leaves of Schinus terebinthifolius, a medicinal plant native to South America belonging to the Anacardiaceae family, rich in phenolic compounds, it showed efficacy against L. infantum promastigotes, with an IC50 value of 57.82 µg/mL (Moraes et al. 2014), a result similar to the CN growth inhibition value on promastigotes of the same Leishmania species demonstrated in the present study. Discussion This study showed better results when compared to those of Dibyendu and Chakraborti (2014), in which the pentavalent antimonials, pentamidine and paramomycin, used for the treatment of CL and VL, showed a lack of response against promastigote forms of Leishmania. In a study with compounds isolated from the leaves of Schinus terebinthifolius, a medicinal plant native to South America belonging to the Anacardiaceae family, rich in phenolic compounds, it showed efficacy against L. infantum promastigotes, with an IC50 value of 57.82 µg/mL (Moraes et al. 2014), a result similar to the CN growth inhibition value on promastigotes of the same Leishmania species demonstrated in the present study. This study showed better results when compared to those of Dibyendu and Chakraborti (2014), in which the pentavalent antimonials, pentamidine and paramomycin, used for the treatment of CL and VL, showed a lack of response against promastigote forms of Leishmania. The performance of in vitro tests, through cell viability assays, is the first step to assess the biological compatibility of a given substance, providing important data on the analysis of biocompatibility between different materials (Rogero et al. 2003). For natural products to be used as alternative therapies in the treatment of leishmaniasis, cytotoxicity tests in mammalian cells are needed (Brenzan et al. 2007), being essential in macrophages, as they are part of the life cycle of Leishmania in the vertebrate host, seen that the differentiation of promastigote forms into amastigotes and subsequent multiplication occurs within these cells (koutsoni et al. 2014; Liu and Uzonna, 2012). Since CNSL and CN presented significant IC50 values on promastigote forms of the three Leishmania species, it was necessary to investigate the cytotoxicity of these compounds on macrophages. Both showed significant toxicity on these cells, with CNSL and CN showing similar results between them, but less toxic when compared to Amph B, used as a positive control. Despite their significant cytotoxicity, when the selectivity index was determined based on the ratio of CC50 over IC50 in intracellular amastigotes, it was observed that CNSL and CN were more selective for the parasite, with the CN selectivity index value being above 20, corroborating the literature data, where this index must present a value close to or greater than 20 for amastigotes internalized in macrophages (Osorio et al. 2007). Page 11/28 In a study by Mesquita et al. Discussion (2014), miltefosine, used as a standard drug, had a CC50 of 49.72 µg/mL, a result close to the values of the substances tested in this study, and presented selectivity index with a value of 7 on L. infantum, a value similar to that of the CNSL, however much lower than the selectivity index of CN. This substance was also more selective for the parasite than the Amph B used as a positive control in this study. The experimental model of amastigotes internalized in macrophages is the one that best represents the way in which the infection occurs in the host (Carneiro et al. 2012). Thus, substances that are able to reduce the percentage of parasitized macrophages and the survival rate of amastigotes in macrophages are considerably promising to be tested in vivo (Alves et al. 2017). When tested against intracellular amastigotes, there was a significant reduction both in the IC50 values of CNSL and CN, in the percentage of murine macrophages experimentally infected by L. infantum and in the survival rate of amastigotes inside the macrophages, at the three concentrations tested of the compounds. This action was concentration-dependent, with better results being observed at a concentration of 12.5 µg/mL for CNSL and CN compared to Amph B. Lower results were found by Moraes et al. (2014), who tested three natural derivatives of Schinus terebinthifolius (Anacardiaceae) leaves, and found IC50 values of 66.59 µg/Ml, 64.90 µg/mL and 28.95 µg/mL, respectively, against internalized amastigotes L. infantum. Assessing the images taken by microscopy, it is observed that in the control group there is a large concentration of amastigotes around the parasitophorous vacuole, while in the groups treated with CNSL and CN there was a decrease in the agglomeration of parasites, demonstrating the reduction of the parasite load inside murine macrophages, which reinforces the antileishmania potential of these substances. This effect can be attributed to a possible activation of macrophages, which may also have promoted oxygen potentiation and an increase in cytokine regulation by them (Reimão et al. 2010). The activity of the compounds against promastigote and amastigote forms may differ, depending on the sites of antileishmania action, which contributes to their being selective for one of the two forms of development. Discussion The phagosome formed shortly after pathogen endocytosis fuses, followed by fusion with lysosomes to produce a phagolysosome (Niedergang and Chavrier 2004). The Innate immunity plays an important role in infection control through mechanisms of action such as phagocytosis and lysosomal activity, which provide for antigen activation and pathogen elimination (Harrison et al. 2003). The phagosome formed shortly after pathogen endocytosis fuses, followed by fusion with lysosomes to produce a phagolysosome (Niedergang and Chavrier 2004). The phagolysosome is a structure filled with acid hydrolases and reactive oxygen species in which most of the degradation of the contents involved takes place, and finally, the phagocytosed pathogens are killed within the phagolysosome (Lee et al. 2003; Lopes et al. 2006). Thus, macrophage activation parameters lead to conformational changes to increase the performance of their functions, such as locomotion and phagocytosis (Petropolis et al. 2014). Innate immunity plays an important role in infection control through mechanisms of action such as phagocytosis and lysosomal activity, which provide for antigen activation and pathogen elimination (Harrison et al. 2003). The phagosome formed shortly after pathogen endocytosis fuses, followed by fusion with lysosomes to produce a phagolysosome (Niedergang and Chavrier 2004). The phagolysosome is a structure filled with acid hydrolases and reactive oxygen species in which most of A primary resistance mechanism to Leishmania infection is the production of nitric oxide (NO) by infected macrophages (Lima-Júnior et al. 2013). Inside the phagolysosome, NO reacts with O2− 2, forming a reactive oxygen species, peroxynitrite. From there, nitrate and nitrite are formed as the final product, which act as microbicidal agents (Bogdan; Rollinghoff 1998; Ueda-Nakamura et al. 2006). In the present study, it was observed that CNSL and CN did not exhibit significant induction of NO synthesis, either in infected or uninfected macrophages. Likewise, treatment with Amph B in infected macrophages did not induce significant NO production. Similar results were found by Abas et al. (2006) testing A. occidentale leaf extract, where there was no significant NO release in murine macrophages. These data reinforce that the tested compounds use another pathway to enhance their leishmanicidal activity, not depending on NO synthesis, considering that the production of this molecule is not the only way to control intracellular Leishmania infection in macrophages (Singh et al. al. 2012). Discussion The susceptibility of both evolutionary forms to the compound can be explained by the different biochemical characteristics between them, as well as by the chemical profile of the substance, such as its solubility in lipids (Athayde-Filho et al. 2000). The better antileishmania activity of CN when compared to CNSL can probably be due to the large amount of compounds present in it, which can have synergistic or antagonistic effects. With regard to the cellular immune response against Leishmania infections, the Th1 type is desirable, which provides cure or protection, such as increased phagocytic capacity, lysosomal volume, nitric oxide synthesis, among others, through mechanisms of activation of macrophages. In the investigation of new therapeutic alternatives for the treatment of leishmaniasis, drugs that have, in addition to activity on the parasite, immunomodulatory capacity, in order to prevail the Th1 host immune response (Islamuddin et Page 12/28 Page 12/28 al. 2015; Roy et al. 2014). Therefore, the activation parameters of macrophages with microbicidal capacity were evaluated. The results of the lysosomal activity and phagocytic capacity assays demonstrated that the treatment of macrophages with CNSL and CN obtained significant immunomodulation results. CN retained neutral red particles, characterized by a significant increase in lysosomal activity, which may suggest an increase in the defense of these cells. Zymosan induces stimulation of defense cells to produce a response, causing an increase in IFN production and phagocytic capacity (Wei et al. 2011), and this increase was demonstrated by the substances tested. These results corroborate the study carried out by Alves et al. (2017), in which gallic and ellagic acids, naturally occurring phenolic compounds found in some plants, including A. occidentale, showed immunomodulation results in the determination of phagocytic capacity and lysosomal activity at concentrations of 6.25 µg/mL and 3.125 µg/mL, similar to those in the present study. Innate immunity plays an important role in infection control through mechanisms of action such as phagocytosis and lysosomal activity, which provide for antigen activation and pathogen elimination (Harrison et al. 2003). The phagosome formed shortly after pathogen endocytosis fuses, followed by fusion with lysosomes to produce a phagolysosome (Niedergang and Chavrier 2004). The Innate immunity plays an important role in infection control through mechanisms of action such as phagocytosis and lysosomal activity, which provide for antigen activation and pathogen elimination (Harrison et al. 2003). Discussion As there was no significant induction of NO production, it is possible that the reduction in infected macrophages and the survival rate is due to the phagocytic capacity and lysosomal activity by macrophages stimulated by the substances. Page 13/28 The use of Z. morio larvae for animal and human commercialization and consumption has increased in recent times, due to their nutritional composition, ease of handling and rearing and short life cycle (Han et al. 2010). In addition, given the ethical conflict and social aspects, we seek to increasingly prioritize the use of alternative models in experimental research, such as Z. morio, considering that in insects the innate immune system is evolutionarily conserved (Canteri et al. 2018). Due to these characteristics, interest in the use of this insect as an alternative model for in vivo studies to assess various activities such as antimicrobial (Morey et al. 2016), toxicity (Van Der Valk and Meijden, 2014) and insecticide (Wang) has increased et al. 2015) and pathogenicity (McGonigle et al. 2016). The main objective of toxicological studies is to predict the possible adverse effects that a substance may cause when exposed to human or animal, whether it is a candidate for a drug, pesticide, industrial chemical agent, among others (Koeter 1993; Stokes 2002; Meyer 2003). In this context, the in vivo test was carried out using Z. morio as a model organism, and the survival profiles in the larval forms of the insect were determined, in order to assess the acute toxicity of CNSL and CN. As in the cytotoxicity test on macrophages, CNSL had lower toxicity on Z. morio than CN, suggesting a synergistic effect of the compounds present in the CNSL, giving it less toxicity when compared to CN alone. These results corroborate some studies that showed that the isolated pure substances were more toxic than their source fractions (Silva et al. 2007; Simas et al. 2004). Similarly, Guissoni et al. (2013) demonstrated that CNSL was less toxic, with a higher mean lethal concentration (LC50) than its fractions, in larvicidal activity against Aedes aegypti. Assessing the acute oral toxicity of CNSL in Rattus norvegicus, LD50 > 2000 mg/kg was demonstrated, not causing any significant signs of intoxication in animals (Guissoni et al. 2013). Discussion According to the classification of the Organization for Economic Co-operation and Development (OECD, 2002), this value allows classifying the CNSL in class 5 (LD50 2000–5000 mg/Kg), indicating a substance with low toxicity. Also according to the OECD, CN presents non-toxicity as one of its main characteristics. Tests carried out by this organization showed the following results regarding BC ecotoxicity: biodegradability − 96% (28 days) - (OECD-302C, 2009); solubility in water equal to 1.0 g/L; ecotoxicity (96 h): fish < 11 g/L; daphnia < 66 g/L; algae < 1 g/L - (OECD-425, 2008) and regarding genotoxicity with using the Ames Salmonella test, it was negative. According to the Safety Data Sheets (SDS, 2017), the oral LD50 is 500 mg/kg (rats) and the dermal LD50 is > 2,000 mg/kg (Daphnia magna). The non-toxicity of CN reported in the literature differs from the result of CC50 and acute toxicity for this substance in the present study, however, according to the OECD, this toxicity, even if low, can be improved with encapsulation and controlled release for use as a biopharmaceutical (OECD, 2002). In conclusion, CNSL and CN were able to show antileishmania potential against L. infantum, L. braziliensis and L. major, acting also through macrophage activation pathways, such as increased Page 14/28 Page 14/28 Page 14/28 phagocytic capacity, increased lysosomal volume. CNSL and CN were able to reduce the percentage of murine macrophages infected by L. infantum and the survival rate of internalized amastigote forms. It was shown that Z. morio larvae are an alternative invertebrate model suitable for analyzing the acute toxicity of the tested substances. The results are promising and serve as a starting point for further research aimed at evaluating the in vivo leishmanicidal potential. Author Contributions Material preparation, collection and data analysis were carried out by IMMR, VCS, RCVC, MSS, JAONN, DSM, MGLC, AKSM, FAAC, MMMA and ILM. The graphics were created by VCS, MSS and MMMA, and the tables were made by IMMR. The translation of the manuscript into English was carried out by IMMR and LSAT. The first draft of the manuscript was written by IMMR and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Consent for publication Not applicable. Funding Author IMMR. has received research support from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – CAPES (process number 88887.489421/2020-00), as a Master's scholarship. Availability of data and materials All data generated or analysed during this study are included in this published article [and its supplementary information files]. Acknowledgements Not applicable. Ethics approval and consent to participate All protocols were approved by the Animal Ethics Committee from Federal University of Piaui, Brazil (nº 640/2019). Competing interests The authors declare that they have no competing interests. References Page 15/28 Page 15/28 1. Abas F, Lajis NH, Israf DA, Khozirah S, Kalsom YU (2006) Antioxidant and nitric oxide inhibition activities of selected Malay traditional vegetables. Food Chem 95: 566–73. https://doi.org/10.1016/j.foodchem.2005.01.034 2. Akhoundi M, Downing T, Votýpka J et al (2017) Leishmania infections: Molecular targets and 3. diagnosis. Molecular Aspects of Medicine 57: 1–29. https://doi.org/10.1016/j.mam.2016.11.012 3. diagnosis. Molecular Aspects of Medicine 57: 1–29. https://doi.org/10.1 4. Akhoundi M, Kuhls K, Cannet A, Votýpka J, Marty P, Delaunay P, Sereno D (2016) A Historical Overview of the Classification, Evolution, and Dispersion of Leishmania Parasites and Sandflies. PLOS Neglected Tropical Diseases 10(6): e0004770. https://doi.org/10.1371/journal.pntd.0004770 5. Alvar J, Vélez ID, Bern C, Herrero M, Desjeux P, Cano J, Jannin J, Den Boer M. (2012) Leishmaniasis Worldwide and Global Estimates of Its Incidence. PLoS One 7(5): e35671. https://doi.org/10.1371/journal.pone.0035671 6. Alves MMM, Brito LM, Souza AC et al (2017) Gallic and ellagic acids: two natural immunomodulator compounds solve infection of macrophages by Leishmania major. Naunyn Schmiedebergs Arch Pharmacol 390: 893–903. https://doi.org/10.1007/s00210-017-1387-y 6. Alves MMM, Brito LM, Souza AC et al (2017) Gallic and ellagic acids: two natural immunomodulator compounds solve infection of macrophages by Leishmania major. Naunyn Schmiedebergs Arch Pharmacol 390: 893–903. https://doi.org/10.1007/s00210-017-1387-y 7. Amorati R, Pedulli GF, Valgimigli L, Attanasi OA, Filippone P, Fiorucci C, Saladino R (2001) Absolute rate constants for the reaction of peroxyl radicals with cardanol derivatives. J Chem Soc 2: 2142– 2146. https://doi.org/10.1039/B105079F 8. Araújo JMD, Silva AP, Cândido, MB, Silva, TWM, Andrade Júnior FP (2020) Ethnopharmacological study of Anacardium occidentale L.: a brief review. Res Soc Dev 9(8): e487985802. http://doi.org/10.33448/rsd-v9i8.8502 8. Araújo JMD, Silva AP, Cândido, MB, Silva, TWM, Andrade Júnior FP (2020) Ethnopharmacological study of Anacardium occidentale L.: a brief review. Res Soc Dev 9(8): e487985802. http://doi.org/10.33448/rsd-v9i8.8502 9. Athayde-Filho PF, Miller J, Simasc AM (2000) Mesoionic Compounds: Amphiphilic Heterocyclic Betaines. Synthesis 11: 1565–1568. http://doi.org/10.1055/s-2000-7606 9. Athayde-Filho PF, Miller J, Simasc AM (2000) Mesoionic Compounds: Amphiphilic Heterocyclic Betaines. Synthesis 11: 1565–1568. http://doi.org/10.1055/s-2000-7606 10. Ayyanar M, Ignacimuthu S (2009) Herbal medicines for wound healing among tribal people in southern India: ethnobotanical and scientific evidences. Int J App Res Nat Prod 2(3): 29–42. Available online http://www.healthy-synergies.com 10. Ayyanar M, Ignacimuthu S (2009) Herbal medicines for wound healing among tribal people in southern India: ethnobotanical and scientific evidences. Int J App Res Nat Prod 2(3): 29–42. Available online http://www.healthy-synergies.com 11. References Bapela MJ, Kaiser M, Meyer JJM (2017) Anti-Leishmanial activity of selected South African plant species. S Afr J Bot 108: 342–345. https://doi.org/10.1016/j.sajb.2016.08.014 11. Bapela MJ, Kaiser M, Meyer JJM (2017) Anti-Leishmanial activity of selected South African plant species. S Afr J Bot 108: 342–345. https://doi.org/10.1016/j.sajb.2016.08.014 12. Balgude D, Sabnis AS (2014) CNSL: an environment friendly alternative for the modern coating industry. J Coating Tech Res 11:169–183. https://doi.org/10.1007/s11998-013-9521-3 12. Balgude D, Sabnis AS (2014) CNSL: an environment friendly alternative for the modern coating industry. J Coating Tech Res 11:169–183. https://doi.org/10.1007/s11998-013-9521-3 13. Benitez ADN, Martins FDC, Mareze M et al (2018) Spatial and simultaneous seroepidemiology of anti-Leishmania spp. antibodies in dog owners and their dogs from randomly selected households in a major city of southern Brazil. Prev Vet Med 154: 47–53. https://doi.org/10.1016/j.prevetmed.2018.02.011 13. Benitez ADN, Martins FDC, Mareze M et al (2018) Spatial and simultaneous seroepidemiology of anti-Leishmania spp. antibodies in dog owners and their dogs from randomly selected households in a major city of southern Brazil. Prev Vet Med 154: 47–53. https://doi.org/10.1016/j.prevetmed.2018.02.011 14. Berman HM, Battistuz T, Bhat TN et al (2000) The Protein Data Bank. Nucleic Acids Res 28: 235– 242. https://doi.org/10.1107/S0907444902003451 14. Berman HM, Battistuz T, Bhat TN et al (2000) The Protein Data Bank. Nucleic Acids Res 28: 235– 242. https://doi.org/10.1107/S0907444902003451 Page 16/28 15. Blanco VR, Nascimento-Junior NM (2017) Leishmaniose: Aspectos Gerais Relacionados com a Doença, o Ciclo do Parasita, Fármacos Disponíveis, Novos Protótipos e Vacinas. Rev Virtual Quim 9 (3): 861–876. http://dx.doi.org/10.21577/1984-6835.20170055 15. Blanco VR, Nascimento-Junior NM (2017) Leishmaniose: Aspectos Gerais Relacionados com a Doença, o Ciclo do Parasita, Fármacos Disponíveis, Novos Protótipos e Vacinas. Rev Virtual Quim 9 (3): 861–876. http://dx.doi.org/10.21577/1984-6835.20170055 Page 16/28 16. Bogdan C, Rollinghoff M (1998) The immune response to Leishmania: mechanisms of parasite control and evasion. Int J Parasitol 28: 121–134. https://doi.org/10.1016/S0020-7519(97)00169-0 17. Bonatto SJR, Folador A, Aikawa J et al (2004) Lifelong exposure to dietary fish oil alters macrophage responses in Walker 256 tumor-bearing rats. Cell Immunol 231: 56–62. https://doi.org/10.1016/j.cellimm.2004.12.001 18. Braga FG, Bouzada MLM, Fabri RL, Matos MO, Moreira FO, Scio E, Coimbra ES (2007) Anti- Leishmanial and antifungal activity of plants used in traditional medicine in Brazil. J Ethnopharmacol 111: 396–402. https://doi.org/10.1016/j.jep.2006.12.006 19. Brenzan MA, Nakamura CV, Dias Filho BP, Ueda-Nakamura T, Young MC, Cortez DA (2007) Antileishmanial activity of crude extract and coumarin from Calophyllum brasiliensis leaves against Leishmania amazonensis. Parasitol Res 101: 715–722. https://doi.org/10.1007/s00436-007-0542-7 20. References Canteri SP, Caloni CC, Wilson D, Almeida SR (2018) An invertebrate host to study fungal infections, mycotoxins and antifungal drugs: Tenebrio molitor. J Fungi 4(4): 125. https://doi.org/10.3390/jof4040125 21. Carneiro SM, Carvalho FA, Santana LC, Sousa AP, Neto JM, Chaves MH (2012) The cytotoxic and antileishmanial activity of extracts and fractions of leaves and fruits of Azadirachta indica (A Juss.) Biol Res 45:111–116. https://doi.org/10.4067/S0716-97602012000200002 22. Chen J, Zhang JH, Wang LK, Sucheck SJ, Snow AM, Hecht SM (1998) Inhibitors of DNA polymerase β from Schoepfia californica. J Chem Soc Chem Commun 24:2769–2770. https://doi.org/10.1039/A807053I 23. Cohen NC (1996) Guidebook on Molecular Modeling in Drug Design. San Diego, USA https://doi.org/10.1016/B978-0-12-178245-0.X5000-8 24. da Silva DPB, Florentino IF, da Silva Moreira LK et al (2018) Chemical characterization and pharmacological assessment of cashew gum extract (Anacardium occidentale L.), polysaccharide free. J Ethnopharmacol 213: 395–402. https://doi.org/10.1016/j.jep.2017.11.021 25. de Souza PC, Morey AT, Marcondes G et al (2015) Tenebrio molitor (Coleoptera: Tenebrionidae) as an alternative host to study fungal infections. J Microbiol Methods 118: 182–6. https://doi.org/10.1016/j.mimet.2015.10.004 26. Dibyendu P, Partha D, Tripti D, Tapati (2014) In vitro anti-leishmanial efficacy of potato tuber extract (PTEx): Leishmanial serine protease(s) as putative target. Exp. Parasitol 146: 11–19. https://doi.org/10.1016/j.exppara.2014.08.009 27. França F, Cuba CA, Moreira EA, Almeida M, das Virgens ML, Marsden PD (1993) Avaliação do efeito do extrato de casca de cajueiro-branco (Anacardium occidentale L.) sobre a infecção por Leishmania (viannia) brasiliensis. Rev. Soc. Bras Med Trop 26 (3): 151–155. https://doi.org/10.1590/S0037- 86821993000300004 28. Funari CS, Almeida L, Passalacqua TG, Martinez I, Ambrósio DL, Cicarelli RMB, Silva DHS, Graminha MAS (2016) Oleanonic acid from Lippia lupulina (Verbenaceae) shows strong in vitro anti- Page 17/28 Page 17/28 Leishmanial and antitrypanosomalactivity. Acta Amaz 46 (4): 411–416. https://doi.org/10.1590/1809-4392201600204 Leishmanial and antitrypanosomalactivity. Acta Amaz 46 (4): 411–416. https://doi.org/10.1590/1809-4392201600204 29. Guissoni ACP, Silva IG, Geris R, Cunha LC, Silva HHG (2013) Atividade larvicida de Anacardium occidentale como alternativa ao controle de Aedes aegypti e sua toxicidade em Rattus norvegicus. Rev Bras Pl Med 15: 363–367. https://doi.org/10.1590/S1516-05722013000300008 30. González-Coloma A, Reina M, Sáenz C, Lacret R, Ruiz-Mesia L, Arán VJ, Sanz J, Martínez-Díaz RA (2011) Anti-Leishmanial, antitrypanosomal, and cytotoxicscreening of ethnopharmacologically selected Peruvian plants. Parasitol Res 110: 1381–1392. https://doi.org/10.1007/s00436-011-2638- 3 31. Goodsell DS (2009) Computational docking of biomolecular complexes with autodock. Cold Spring Harbor Protocols 4(5). https://doi.org/10.1101/pdb.prot5200 31. Goodsell DS (2009) Computational docking of biomolecular complexes with autodock. Cold Spring Harbor Protocols 4(5). https://doi.org/10.1101/pdb.prot5200 32. References Goodsell DS, Morris GM, Olson AJ (1996) Automated docking of flexible ligands: applications of autodock. J Mol Recognit 9: 1–5. https://doi.org/10.1002/(SICI)1099-1352(199601)9:1<1::AID- JMR241>3.0.CO;2-6 33. Grando FCC, Felício CA, Twardowschy A, Paula FM, Batista VG, Fernandes LC, Curi R, Nishiyama A (2009) Modulation of peritoneal macrophage activity by the saturation state of the fatty acid moiety of phosphatidylcholine. Braz J Med Biol Res 42(7): 599–605. https://doi.org/10.1590/S0100- 879%C3%972009005000003 33. Grando FCC, Felício CA, Twardowschy A, Paula FM, Batista VG, Fernandes LC, Curi R, Nishiyama A (2009) Modulation of peritoneal macrophage activity by the saturation state of the fatty acid moiety of phosphatidylcholine. Braz J Med Biol Res 42(7): 599–605. https://doi.org/10.1590/S0100- 879%C3%972009005000003 34. Han SR, Lee BS, Jung KJ et al (2016) Safety assessment of freeze-dried powdered Tenebrio molitor larvae (yellow mealworm) as novel food source: Evaluation of 90-day toxicity in Sprague-Dawley rats. Regul Toxicol Pharm 77: 206–212. https://doi.org/10.1016/j.yrtph.2016.03.006 34. Han SR, Lee BS, Jung KJ et al (2016) Safety assessment of freeze-dried powdered Tenebrio molitor larvae (yellow mealworm) as novel food source: Evaluation of 90-day toxicity in Sprague-Dawley rats. Regul Toxicol Pharm 77: 206–212. https://doi.org/10.1016/j.yrtph.2016.03.006 35. Harrison RE, Bucci C, Vieira OV, Schroer TA, Grinstein S (2003) Phagosomes fuse with late endosomes and/or lysosomes by extension of membrane protrusions along microtubules: role of Rab7 and RILP. Mol Cell Biol 23: 6494–6506. https://doi.org/10.1128/MCB.23.18.6494-6506.2003 36. Hemshekhar M, Santhosh MS, Kemparaju K, Girish KS (2012) Emerging roles of anacardic acid and its derivatives: a pharmacological overview. Basic Clin Pharmacol Toxicol 110: 122–132. https://doi.org/10.1111/j.1742-7843.2011.00833.x 37. Islamuddin M, Chouhan G, Farooque A, Dwarakanath BS, Sahal D, Afrin F (2015) Th1-biased immunomodulation and therapeutic potential of Artemisia annua in murine visceral Leishmaniasis. PLoS Negl Trop Dis 9(1): e3321. https://doi.org/10.1371/journal.pntd.0003321 38. Kevric I, Cappel MA, Keeling JH (2015) New world and old world Leishmania infections: a practical review. Dermatol Clin 33 (3): 579–593. https://doi.org/10.1016/j.det.2015.03.018 39. Koeter HB (1993) Test guideline development and animal welfare: regulatory acceptance of in vitro studies. Reprod Toxicol 7: 117–123, https://doi.org/10.1016/0890-6238(93)90077-K 40. Koutsoni O, Barhoumi M, Guizani I, Dotsika E (2014) Leishmania eukaryotic initiation factor (LeIF) inhibits parasite growth in murine macrophages. PLoS One 9(5): e97319. https://doi.org/10.1371/journal.pone.0097319 Page 18/28 41. Kubo I, Muroi H, Himejima M, Yamagiwa Y, Mera H, Tokushima K, Ohta S, Kamikawa T. (1993a) Structure-Antibacterial activity relationships of anacardic acids. J Agric Food Chem 41: 1016–1019. http://dx.doi.org/10.1021/jf00030a036 42. References https://doi.org/10.2174/157019308785161611 56. Mesquita JT, Costa-Silva TA, Borborema SET, Tempone AG (2014) Activity of imidazole compounds on Leishmania (L.) infantum chagasi: reactive oxygen species induced by econazole. Mol Cell Biochem 389: 293–300. https://doi.org/10.1007/s11010-013-1954-6 56. Mesquita JT, Costa-Silva TA, Borborema SET, Tempone AG (2014) Activity of imidazole compounds on Leishmania (L.) infantum chagasi: reactive oxygen species induced by econazole. Mol Cell Biochem 389: 293–300. https://doi.org/10.1007/s11010-013-1954-6 57. MEYER O (2003) Testing and assessment strategies, including alternative and new approaches. Toxicol Lett 140: 21–30. https://doi.org/10.1016/S0378-4274(02)00492-7 57. MEYER O (2003) Testing and assessment strategies, including alternative and new approaches. Toxicol Lett 140: 21–30. https://doi.org/10.1016/S0378-4274(02)00492-7 58. Morais TR, da Costa-Silva TA, Tempone AG et al (2014) Antiparasitic activity of natural and semi- synthetic tirucallane triterpenoids from Schinus terebinthifolius (Anacardiaceae): structure/activity relationships. Molecules 19(5): 5761–5776. https://doi.org/10.3390/molecules19055761 58. Morais TR, da Costa-Silva TA, Tempone AG et al (2014) Antiparasitic activity of natural and semi- synthetic tirucallane triterpenoids from Schinus terebinthifolius (Anacardiaceae): structure/activity relationships. Molecules 19(5): 5761–5776. https://doi.org/10.3390/molecules19055761 59. Morey AT, de Souza FC, Santos JP et al (2016) Antifungal Activity of Condensed Tannins from Stryphnodendron adstringens: Effect on Candida tropicalis Growth and Adhesion Properties. Curr Pharm Biotechnol 17(4): 365–375. https://doi.org/10.2174/1389201017666151223123712 59. Morey AT, de Souza FC, Santos JP et al (2016) Antifungal Activity of Condensed Tannins from Stryphnodendron adstringens: Effect on Candida tropicalis Growth and Adhesion Properties. Curr Pharm Biotechnol 17(4): 365–375. https://doi.org/10.2174/1389201017666151223123712 60. Muroi H, Kubo I (1996) Antibacterial activity of anacardic acid and totarol, alone and in combination with methicillin, against methicillinresistant Staphylococcus aureus. J Appl Bacteriol 80: 387–394. https://doi.org/10.1111/j.1365-2672.1996.tb03233.x 60. Muroi H, Kubo I (1996) Antibacterial activity of anacardic acid and totarol, alone and in combination with methicillin, against methicillinresistant Staphylococcus aureus. J Appl Bacteriol 80: 387–394. https://doi.org/10.1111/j.1365-2672.1996.tb03233.x iedergang F, Chavrier P (2004) Signaling and membrane dynamics during 61. Niedergang F, Chavrier P (2004) Signaling and membrane dynamics duri 61. Niedergang F, Chavrier P (2004) Signaling and membrane dynamics during 62. phagocytosis: many roads lead to the phagos(R)ome. Curr Opin Cell Biol 16: 422–428. https://doi.org/10.1016/j.ceb.2004.06.006 62. phagocytosis: many roads lead to the phagos(R)ome. Curr Opin Cell Biol 16: 422–428. https://doi.org/10.1016/j.ceb.2004.06.006 63. OECD - Organization for Economic Co-operation and Development (2002), Test No. 423: Acute Oral toxicity - Acute Toxic Class Method, OECD Guidelines for the Testing of Chemicals, Section 4, OECD Publishing, Paris. https://doi.org/10.1787/9789264071001-en. 63. OECD - Organization for Economic Co-operation and Development (2002), Test No. References Kubo I, Ochi M, Vieira PC, Komatsu S (1993b) Antitumor agents from the cashew (Anacardium occidentale L.) apple juice. J Agric Food Chem 41: 1012–1015. https://doi.org/10.1021/jf00030a035 43. Kudi C, Umoh JU, Eduvie LO, Gefu J (1999) Screening of some Nigerian medicinal plants for antibacterial activity. J Ethnopharmacol 67: 225–228. https://doi.org/10.1016/S0378- 8741(98)00214-1 44. Lee WL, Harrison RE, Grinstein S (2003) Phagocytosis by neutrophils. Microbes Infect 5: 1299–1306. https://doi.org/10.1016/j.micinf.2003.09.014 45. Lemes LFN, Ramos GA, de Oliveira AS et al (2016) Cardanol-derived AChE inhibitors: Towards the development of dual binding derivatives for Alzheimer’s disease. Eur J Med Chem 108: 687–700. https://doi.org/10.1016/j.ejmech.2015.12.024 46. Lopes L, Godoy LMF, de Oliveira CC, Gabardo J, Schadeck RJG, de Freitas Buchi D. (2006) Phagocytosis, endosomal/lysosomal system and other cellularaspects of macrophage activation by Canova medication. Micron 37: 277–287. https://doi.org/10.1016/j.micron.2005.08.005 47. Lima-Junior D, Costa D, Carregaro V et al (2013) Inflammasome-derived IL-1ß production induces nitric oxide-mediated resistance to Leishmania. Nat Med 19: 909–915. https://doi.org/10.1038/nm.3221 48. Liu D, Uzonna JE (2012) The early interaction of Leishmania with macrophages and 49. dendritic cells and its influence on the host immune response. Front Cell Infect Microbiol 2(83): 1–8. https://doi.org/10.3389/fcimb.2012.00083 50. Luize PS, Tiuman TS, Maza MLG, PK Ueda-Nakamura T, Filho BPD, Cortez DAG, de Mello JCP, Nakamura CV (2005) Effects of medicinal plant extracts on growth of Leishmania (L.) amazonensis and Trypanosoma cruzi. Braz J Pharm Sci 41(1): 85–94. https://doi.org/10.1590/S1516- 93322005000100010 51. Machado M, Dinis AM, Santos-Rosa M, Alves V, Salgueiro L, Cavaleiro C, Sousa MC (2014) Activity of Thymus capitellatus volatile extract, 1.8-cineole and borneol against Leishmania species. Vet Parasitol 200: 39–49. http://dx.doi.org/10.1016/j.vetpar.2013.11.016 52. Mansour R, Haouas N, Ben Kahla-Nakbi A, Hammami S, Mighri Z, Mhenni F, Babba H (2013) The effect of Vitis vinifera L. leaves extract on Leishmania infantum in vitro. Iran J Pharm Res 12(3): 349–355. https://doi.org/10.22037/IJPR.2013.1333 53. Mazzetto SE, Lomonaco D, Mele G (2009) Óleo da castanha de caju: oportunidades e desafios no contexto do desenvolvimento e sustentabilidade industrial. Quim Nova 32: 732–741. https://doi.org/10.1590/S0100-40422009000300017 54. McGonigle JE, Purves J, Rolff J (2016) Intracellular survival of Staphylococcus aureus during persistent infection in the insect Tenebrio molitor. Dev Comp Immunol 59: 34–38. Page 19/28 Page 19/28 https://doi.org/10.1016/j.dci.2016.01.002 https://doi.org/10.1016/j.dci.2016.01.002 55. Mele G, Vasapollo G (2008) Fine Chemicals and New Hybrid Materials From Cardanol. Mini Rev Org Chem 5(3): 243–253. https://doi.org/10.2174/157019308785161611 55. Mele G, Vasapollo G (2008) Fine Chemicals and New Hybrid Materials From Cardanol. Mini Rev Org Chem 5(3): 243–253. References Rogero SO, Lugão AB, Ikeda TI, Cruz ÁS (2003) Teste in vitro de citotoxicidade: estudo comparativo entre duas metodologias. Mater Res 6(3): 317–320. https://doi.org/10.1590/S1516- 14392003000300003 72. Roy N, Chakraborty S, Paul Chowdhury B, Banerjee S, Halder K, Majumder S, Majumdar S, Sen PC (2014) Regulation of PKC mediated signaling by calcium during Visceral Leishmaniasis. PLoS One 9(10): e110843. https://doi.org/10.1371/journal.pone.0110843 72. Roy N, Chakraborty S, Paul Chowdhury B, Banerjee S, Halder K, Majumder S, Majumdar S, Sen PC (2014) Regulation of PKC mediated signaling by calcium during Visceral Leishmaniasis. PLoS One 9(10): e110843. https://doi.org/10.1371/journal.pone.0110843 73. Sanner MF (1999) Python: a programming language for software integration and development. J Mol Graph Model 17(1): 57–61. PMID: 10660911 73. Sanner MF (1999) Python: a programming language for software integration and development. J Mol Graph Model 17(1): 57–61. PMID: 10660911 74. Scotti L, Mendonça Junior FJB, Moreira DRM, Silva MS, Pitta IR, Scotti MT (2012) SAR, QSAR and docking of anticancer flavonoids and variants: a review. Curr Top Med Chem 12(24):2785–2809. https://doi.org/10.2174/1568026611212240007 74. Scotti L, Mendonça Junior FJB, Moreira DRM, Silva MS, Pitta IR, Scotti MT (2012) SAR, QSAR and docking of anticancer flavonoids and variants: a review. Curr Top Med Chem 12(24):2785–2809. https://doi.org/10.2174/1568026611212240007 75. Silveira, ES, Rodrigues NLDC, Machado, NMJ, Fonseca, FRM, Teixeira, MJ, Leal LKAM (2020) Medicinal Plants Containing Coumarin or Essential Oils from the Brazilian Biome May be New Option for Treating Leishmaniasis? Pharmacogn Rev 14(27): 53–61. https://doi.org/10.5530/phrev.2020.14.9 76. Simas NK, Lima EC Conceição SR, Kuster RM, Oliveira Filho M, Lage CLS (2004). Produtos naturais para o controle da transmissão da dengue atividade larvicida de Myrozylon balsamum (óleo vermelho) e de terpenóides e fenilpropanóides. Quim Nova 27(1): 46–9. https://doi.org/10.1590/S0100-40422004000100009 77. Singh N, Kumar, M, Singh RK (2012) Leishmaniasis: Current status of available drugs and new potential drug targets. Asian Pac J Trop Med 5: 485–497. https://doi.org/10.1016/S1995- 7645(12)60084-4 78. Silva HHG, Geris R, Rodrigues Filho E, Rocha C, Silva IG (2007) Larvicidal activity of oil-resin fractions from the Brazilian medicinal plant Copaifera reticulata Ducke (Leguminosae-Caesalpinoideae) against Aedes aegypti (Diptera, Culicidae). Rev Soc Bras Med Trop 40(3): 264–67. https://doi.org/10.1590/S0037-86822007000300002 79. Silva RAO, Santos MRMC, Filho FCS, Silva, MCC Oliveira, GAL, Silva MGO, Marques, LGA (2014) Prospecção tecnológica de fitoterápico (Euphorbia tirucalli L.) utilizado no tratamento de neoplasias e outras doenças. Cadernos de Prospecção 6(4): 490–499. https://doi.org/10.9771/cp.v6i4.11394 80. References 423: Acute Oral toxicity - Acute Toxic Class Method, OECD Guidelines for the Testing of Chemicals, Section 4, OECD Publishing, Paris. https://doi.org/10.1787/9789264071001-en. 64. OECD - Organization for Economic Co-operation and Development (2008), Test No. 425: Acute Oral Toxicity: Up-and-Down Procedure, OECD Guidelines for the Testing of Chemicals, Section 4, OECD Publishing, Paris. https://doi.org/10.1787/9789264071049-en. 64. OECD - Organization for Economic Co-operation and Development (2008), Test No. 425: Acute Oral Toxicity: Up-and-Down Procedure, OECD Guidelines for the Testing of Chemicals, Section 4, OECD Publishing, Paris. https://doi.org/10.1787/9789264071049-en. 65. OECD - Organization for Economic Co-operation and Development (2009), Test No. 302C: Inherent Biodegradability: Modified MITI Test (II), OECD Guidelines for the Testing of Chemicals, Section 3, OECD Publishing, Paris. https://doi.org/10.1787/9789264070400-en. 66. Oliveira LGC, Brito LM, Alves MMA et al (2017) In vitro effects of the Neolignan 2,3- Dihydrobenzofuran against Leishmania amazonensis. Basic Clin Pharmacol Toxicol 01: 1–7. https://doi.org/10.1111/bcpt.12639 67. Oliveira MS, Morais SM, Magalhaes DV et al (2010) Antioxidant, larvicidal and antiacetylcholinesterase activities of cashew nut shell liquid constituents. Acta Trop 117: 165–170,. https://doi.org/10.1016/j.actatropica.2010.08.003 67. Oliveira MS, Morais SM, Magalhaes DV et al (2010) Antioxidant, larvicidal and antiacetylcholinesterase activities of cashew nut shell liquid constituents. Acta Trop 117: 165–170,. https://doi.org/10.1016/j.actatropica.2010.08.003 68. Osorio E, Arango GJ, Jiménez N et al (2007) Antiprotozoal and cytotoxic activities in vitro of Colombian Annonaceae. J Ethnopharmacol 111: 630–5. https://doi.org/10.1016/j.jep.2007.01.015 68. Osorio E, Arango GJ, Jiménez N et al (2007) Antiprotozoal and cytotoxic activities in vitro of Colombian Annonaceae. J Ethnopharmacol 111: 630–5. https://doi.org/10.1016/j.jep.2007.01.015 68. Osorio E, Arango GJ, Jiménez N et al (2007) Antiprotozoal and cytotoxic activities in vitro of Colombian Annonaceae. J Ethnopharmacol 111: 630–5. https://doi.org/10.1016/j.jep.2007.01.015 Page 20/28 69. Petropolis DB, Rodrigues JCF, Viana NB, Pontes B, Pereira CFA, Silva-Filho FC (2014) Leishmania amazonensis promastigotes in 3D Collagen I culture: an in vitro physiological environment for the study of extracellular matrix and host cell interactions. PeerJ 2: e317. https://doi.org/10.7717/peerj.317 70. Reimão JQ, Scotti MT, Tempone AG (2010) Anti-Leishmanial and anti-trypanosomal activities of 1, 4- dihydropyridines: In vitro evaluation and structure-activity relationship study. Bioorg Med Chem 18: 8044–8053. https://doi.org/10.1016/j.bmc.2010.09.015 70. Reimão JQ, Scotti MT, Tempone AG (2010) Anti-Leishmanial and anti-trypanosomal activities of 1, 4- dihydropyridines: In vitro evaluation and structure-activity relationship study. Bioorg Med Chem 18: 8044–8053. https://doi.org/10.1016/j.bmc.2010.09.015 71. Rogero SO, Lugão AB, Ikeda TI, Cruz ÁS (2003) Teste in vitro de citotoxicidade: estudo comparativo entre duas metodologias. Mater Res 6(3): 317–320. https://doi.org/10.1590/S1516- 14392003000300003 71. References Soares DC, Pereira CG, Meireles MA, Saraiva EM (2007) Leishmanicidal activity of a supercritical fluid fraction obtained from Tabernaemontana catharinensis. Parasitol Int 56(2): 135–139. Page 21/28 Page 21/28 https://doi.org/10.1016/j.parint.2007.01.004 https://doi.org/10.1016/j.parint.2007.01.004 81. Solis FJ, Wets RJB (1981) Minimization by Random Search Techniques. Math Oper Res 6(1): 19–30. https://doi.org/10.1287/moor.6.1.19 82. Souza-Silva F, Bourguignon SC Pereira BAS et al (2015) Epoxy-α-lapachone has in vitro and in vivo anti-Leishmania (Leishmania) amazonensis effects and inhibits eerine proteinaseactivity in this parasite. Antimicrob Agents Chemother 59(4): 1910–1918. https://doi.org/10.1128/AAC.04742-14 83. Steverding D (2017) The history of Leishmaniasis. Parasit Vectors 10(82): 1–10. https://doi.org/10.1186/s13071-017-2028-5 84. Stokes WS (2002) Humane endpoint for laboratory animals used in regulatory testing. ILAR J 43: 31–38. https://doi.org/10.1093/ilar.43.Suppl_1.S31 85. SDS - Safety Data Sheet (2017) According to Regulation (EC) n º 1907/2006 and its amendments. p. 8. https://nonkorservice.pl/wp-content/uploads/2018/05/C-Pox-ST180-AL-7N180-ms3.pdf Accessed in Jun 20, 2021. 86. Torres FAE, Passalacqua TG, Velásque AMA, de Souza RA, Colepicolo P, Graminha MAS (2014) New drugs with antiprotozoal activity from marine algae: a review. Rev Bras Farmacogn 24(3): 265–276. https://doi.org/10.1016/j.bjp.2014.07.001 87. Tiuman TS, Santos AO, Ueda-Nakamura T, Filho BP, Nakamura CV (2011) Recent advances in Leishmaniasis treatment. Int J Infect Dis 15:e525-e532. https://doi.org/10.1016/j.ijid.2011.03.021 88. Tiwari N, Gedda MR, Tiwari VK Singh SP, Singh RK (2018) Limitations of current therapeutic options, possible drug targets and scope of natural products in control of Leishmaniasis. Mini Rev Med Chem 18(1): 26–41 https://doi.org/10.2174/1389557517666170425105129 89. Ueda-Nakamura T, Mendonça-Filho RR, Morgado-Díaz JA et al (2010) Anti-Leishmanial activity of Eugenol-rich essential oil from Ocimum gratissimum. Parasitol Int 55: 99–105. https://doi.org/10.1016/j.parint.2005.10.006 90. Valadares DG, Duarte MC, Oliveira JS et al (2011) Leishmanicidal activity of the Agaricus blazei Murill in different Leishmania species. Parasitol Int 60(4): 357–63. https://doi.org/10.1016/j.parint.2011.06.001 91. Van Der Valk T, Van Der Meijden A (2014) Toxicity of scorpion venom in chick embryo and mealworm assay depending on the use of the soluble fraction versus the whole venom. Toxicon 88: 38–43. https://doi.org/10.1016/j.toxicon.2014.06.007 92. Wang X, Hao Q, Chen Y, Jiang S, Yang Q, Li Q (2015) The effect of chemical composition and bioactivity of several essential oils on Tenebrio molitor (Coleoptera: Tenebrionidae). J Insect Sci 15 (1): 116. https://doi.org/10.1093/jisesa/iev093 92. Wang X, Hao Q, Chen Y, Jiang S, Yang Q, Li Q (2015) The effect of chemical composition and bioactivity of several essential oils on Tenebrio molitor (Coleoptera: Tenebrionidae). J Insect Sci 15 (1): 116. https://doi.org/10.1093/jisesa/iev093 93. References Wei WC, Su YH, Chen SS, Sheu JH, Yang NS (2011) GM-CSF plays a key role in zymosan-stimulated human dendritic cells for activation of Th1 and Th17 cells. Cytokine 55: 79–89. https://doi.org/10.1016/j.cyto.2011.03.017 94. WHO - World Health Organization (2021) Leishmaniasis. https://www.who.int/en/news-room/fact- sheets/detail/Leishmaniasis. Accessed June 15, 2021. Page 22/28 95. Wu K, Liu J, Tseng SF et al (2011) Anacardic Acid (6-Pentadecylsalicylic Acid) Inhibits Tumor Angiogenesis by Targeting Src/FAK/Rho GTPases Signaling Pathway. J Pharmacol Exp. Ther 339(2): 403–411. https://doi.org/10.1124/jpet.111.181891 95. Wu K, Liu J, Tseng SF et al (2011) Anacardic Acid (6-Pentadecylsalicylic Acid) Inhibits Tumor Angiogenesis by Targeting Src/FAK/Rho GTPases Signaling Pathway. J Pharmacol Exp. Ther 339(2): 403–411. https://doi.org/10.1124/jpet.111.181891 339(2): 403–411. https://doi.org/10.11 Figures Figure 1 Cytotoxic effects of CNSL (a) and CN (b) ag and red blood cells were incubated with CN evaluated using tetrazolium salt (MTT) test Figures Figure 1 Cytotoxic effects of CNSL (a) and CN (b) against BALB/c murine peritoneal macrophages. Macrophage and red blood cells were incubated with CNSL, CN, or Amph B for 48 h. The macrophage viability was evaluated using tetrazolium salt (MTT) test. Data are presented as mean ± SEM of three experiments performed in triplicate p < 0.05; p < 0.01; *p < 0.001 when compared with control (C) or Amph B 95. Wu K, Liu J, Tseng SF et al (2011) Anacardic Acid (6-Pentadecylsalicylic Acid) Inhibits Tumor Angiogenesis by Targeting Src/FAK/Rho GTPases Signaling Pathway. J Pharmacol Exp. Ther 339(2): 403–411. https://doi.org/10.1124/jpet.111.181891 Figure 1 Cytotoxic effects of CNSL (a) and CN (b) against BALB/c murine peritoneal macrophages. Macrophages and red blood cells were incubated with CNSL, CN, or Amph B for 48 h. The macrophage viability was evaluated using tetrazolium salt (MTT) test. Data are presented as mean ± SEM of three experiments performed in triplicate p < 0.05; p < 0.01; p < 0.001 when compared with control (C) or Amph B Page 23/28 Figure 2 Macrophages experimentally infected with Leishmania infantum. Control (a). Amphotericin as a positive control at a concentration of 0.5 (b); 1.0 (c) and 2.0 (d) µg/ml. For the treatme concentrations of 3.125 (e) were used; 6.25 (f) and 12.5 (g) µg/ml. CN was used at concen 3.125 (h); 6.25 (i) and 12.5 (j) µg/ml. The arrows indicate macrophage-internalized amastig L. infantum Figure 2 Macrophages experimentally infected with Leishmania infantum. Control (a). Amphotericin B was used as a positive control at a concentration of 0.5 (b); 1.0 (c) and 2.0 (d) µg/ml. For the treatment with CNSL, concentrations of 3.125 (e) were used; 6.25 (f) and 12.5 (g) µg/ml. CN was used at concentrations of 3.125 (h); 6.25 (i) and 12.5 (j) µg/ml. The arrows indicate macrophage-internalized amastigote forms of L. infantum Page 24/28 Page 24/28 Figure 3 Effects of CNSL, CN and amphotericin B on infected m murine macrophages infected with Leishmania infant were treated with CNSL, CN or Amph B for 48h. Data a performed in triplicate. p < 0.05; p < 0.01; p < 0.0 Figure 3 Effects of CNSL, CN and amphotericin B on infected macrophages (a) and survival index (b) of BALB/c murine macrophages infected with Leishmania infantum, considering the 87% infection control. Cells were treated with CNSL, CN or Amph B for 48h. Data are presented as mean ± SEM of three experiments performed in triplicate. p < 0.05; p < 0.01; p < 0.001 when compared with control (C) Figure 3 Effects of CNSL, CN and amphotericin B on infected macrophages (a) and survival index (b) of BALB/c murine macrophages infected with Leishmania infantum, considering the 87% infection control. Cells were treated with CNSL, CN or Amph B for 48h. Data are presented as mean ± SEM of three experiments performed in triplicate. p < 0.05; p < 0.01; p < 0.001 when compared with control (C) Page 25/28 Page 25/28 Figure 4 Effects of CNSL, CN, and Amph B on lysosomal activity (a) and phagocytic capability (b). Murine peritoneal macrophages were treated at ranging concentrations for 48h. Data are presented as mean ± SEM of three experiments performed in triplicate. p < 0.05; p < 0.01; p < 0.001 when compared with control (C) Figure 4 Effects of CNSL, CN, and Amph B on lysosomal activity (a) and phagocytic capability (b). Murine peritoneal macrophages were treated at ranging concentrations for 48h. Data are presented as mean ± SEM of three experiments performed in triplicate. p < 0.05; p < 0.01; *p < 0.001 when compared with control (C) Page 26/28 Figure 5 Nitrite measurement in non-infected (a) or infected (b) BALB/c murine peritoneal macrophages treated with CNSL or CN and Amph B for 24 h. The culture supernatant was mixed in equal parts with the Griess reagent. LPS (lipopolysaccharide from Escherichia coli; 2 μg/mL) was used as positive control. Data are presented as mean ± SEM of three experiments performed in triplicate. p < 0.05 when compared with no infected (−) or infected (+) macrophages from control group; p < 0.05 when compared with non-infected (−) or infected (+) macrophages from the LPS group HNS71387 Figure 5 Nitrite measurement in non-infected (a) or infected (b) B Figure 5 Nitrite measurement in non-infected (a) or infected (b) BALB/c murine peritoneal macrophages treated with CNSL or CN and Amph B for 24 h. The culture supernatant was mixed in equal parts with the Griess reagent. LPS (lipopolysaccharide from Escherichia coli; 2 μg/mL) was used as positive control. Data are presented as mean ± SEM of three experiments performed in triplicate. p < 0.05 when compared with non- infected (−) or infected (+) macrophages from control group; p < 0.05 when compared with non-infected (−) or infected (+) macrophages from the LPS group HNS71387 Page 27/28 Figure 6 Percentage of survival of Z. morio larvae exposed to CNSL (a) and CN (b). Larvae were incubated for 48 h in the presence of different concentrations. 0.2% DMSO was used as a control and 100% DMSO was used to cause 100% larval mortality. Results represent mean ± S.E.M of treatments performed in triplicate. Figure 6 Figure 6 Percentage of survival of Z. morio larvae exposed to CNSL (a) and CN (b). Larvae were incubated for 48 h in the presence of different concentrations. 0.2% DMSO was used as a control and 100% DMSO was used to cause 100% larval mortality. Results represent mean ± S.E.M of treatments performed in triplicate. Page 28/28 Page 28/28
https://openalex.org/W4248837353	https://www.researchsquare.com/article/rs-10855/v1.pdf	English	null	Multivariate Network Meta-Analysis Incorporating Class Effects	Research Square (Research Square)	2,020	cc-by	968	Preprint: Please note that this article has not completed peer review. Preprint: Please note that this article has not completed peer review. Multivariate Network Meta-Analysis Incorporating Class Effects CURRENT STATUS: UNDER REVIEW Rhiannon Kate Owen University of Leicester Department of Health Sciences rko4@le.ac.ukCorresponding Author ORCiD: https://orcid.org/0000-0001-5977-376X Sylwia Bujkiewicz University of Leicester Department of Health Sciences Douglas G Tincello University of Leicester Department of Health Sciences Keith R Abrams University of Leicester Department of Health Sciences DOI: 10.21203/rs.2.20243/v1 SUBJECT AREAS Health Economics & Outcomes Research KEYWORDS multivariate, network meta-analysis, mixed treatment comparisons, meta-analysis, class effect Multivariate Network Meta-Analysis Incorporating Class Effects CURRENT STATUS: UNDER REVIEW 1 1 Abstract Background: Network meta-analysis synthesises data from a number of clinical trials in order to assess the comparative efficacy of multiple healthcare interventions in similar patient populations. In situations where clinical trial data are heterogeneously reported i.e. data are missing for one or more outcomes of interest, synthesising such data can lead to disconnected networks of evidence, increased uncertainty, and potentially biased estimates which can have severe implications for decision-making. To overcome this issue, strength can be borrowed between outcomes of interest in multivariate network meta-analyses. Furthermore, in situations where there are relatively few trials informing each treatment comparison, there is a potential issue with the sparsity of data in the treatment networks, which can lead to substantial parameter uncertainty. A multivariate network meta-analysis approach can be further extended to borrow strength between interventions of the same class using hierarchical models. Methods: We extend the trivariate network meta-analysis model to incorporate the exchangeability between treatment effects belonging to the same class of intervention to increase precision in treatment effect estimates. We further incorporate a missing data framework to estimate uncertainty in trials that fail to report measures of variability in order to maximise the use of all available information for healthcare decision-making. The methods are applied to a motivating dataset in overactive bladder syndrome. The outcomes of interest were mean change from baseline in incontinence, voiding and urgency episodes. All models were tted using Bayesian Markov Chain Monte Carlo (MCMC) methods in WinBUGS. Results: All models (univariate, multivariate, and multivariate models incorporating class effects) produced similar point estimates for all treatment effects. Incorporating class effects in multivariate models often increased precision in treatment effect estimates. Conclusions: Multivariate network meta-analysis incorporating class effects allowed for the comparison of all interventions across all outcome measures to ameliorate the potential impact of outcome reporting bias, and further borrowed strength between interventions belonging to the same class of treatment to increase the precision in treatment effect estimates for healthcare policy and 2 2 decision-making. Keywords: multivariate; network meta-analysis; mixed treatment comparisons; meta-analysis; class effect Due to technical limitations, full-text HTML conversion of this manuscript could not be completed. However, the manuscript can be downloaded and accessed as a PDF. Abstract Figure 1 Network of evidence for univariate network meta-analysis for incontinence episodes Figure 1 Figure 1 Network of evidence for univariate network meta-analysis for incontinence episodes Network of evidence for univariate network meta-analysis for incontinence episodes Network of evidence for univariate network meta-analysis for incontinence episodes 3 Figure 2 Network of evidence for univariate network meta-analysis for voiding episodes Figure 2 Figure 2 Network of evidence for univariate network meta-analysis for voiding episodes Network of evidence for univariate network meta-analysis for voiding episodes Network of evidence for univariate network meta-analysis for voiding episodes Network of evidence for univariate network meta-analysis for voiding episodes Figure 3 Network of evidence for univariate network meta-analysis for urgency episodes Figure 3 Figure 3 Network of evidence for univariate network meta-analysis for urgency episodes Network of evidence for univariate network meta-analysis for urgency episode Network of evidence for univariate network meta-analysis for urgency episodes 5 5 Figure 4 Classification of interventions Figure 4 Classification of interventions 6 6 Figure 5 Network of evidence for multivariate network meta-analysis Figure 5 Network of evidence for multivariate network meta-analysis 7 Figure 6 Figure 6 Comparison of the estimated posterior median difference (and 95% credible intervals) in change from baseline in incontinence episodes relative to placebo between individual- intervention, hierarchical, and multivariate hierarchical NMA models for the top 10 interventions Comparison of the estimated posterior median difference (and 95% credible intervals) in change from baseline in incontinence episodes relative to placebo between individual- intervention, hierarchical, and multivariate hierarchical NMA models for the top 10 interventions Comparison of the estimated posterior median difference (and 95% credible intervals) in change from baseline in incontinence episodes relative to placebo between individual- intervention, hierarchical, and multivariate hierarchical NMA models for the top 10 interventions 8 Figure 7 Figure 7 Figure 7 Figure 7 Comparison of the estimated posterior median difference (and 95% credible intervals) in change from baseline in voiding episodes relative to placebo between individual- intervention, hierarchical, and multivariate hierarchical NMA models for the top 10 interventions Comparison of the estimated posterior median difference (and 95% credible intervals) in change from baseline in voiding episodes relative to placebo between individual- intervention, hierarchical, and multivariate hierarchical NMA models for the top 10 interventions Comparison of the estimated posterior median difference (and 95% credible intervals) in change from baseline in voiding episodes relative to placebo between individual- intervention, hierarchical, and multivariate hierarchical NMA models for the top 10 interventions 9 Figure 8 Figure 9 Heatmap of intervention profiles for the cardinal symptoms of OAB Supplementary Files Figure 8 Figure 8 Comparison of the estimated posterior median difference (and 95% credible intervals) in change from baseline in urgency episodes relative to placebo between individual- intervention, hierarchical, and multivariate hierarchical NMA models for the top 10 interventions 10 10 10 11 Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. AddFile1_TrtCodes.pdf AddFile7_ConvDiag.pdf AddFile4_WinBUGS code.pdf AddFile5_UnivariateUI.pdf Reference PDF.pdf MVNMA paper tex file.tex AddFile6_UnivTrtProfiles.pdf AddFile8_SensAnalysis.pdf MVNMA_bib.bib 12
https://openalex.org/W4225684891	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0263113&type=printable	English	null	Role of urinary H2O2, 8-iso-PGF2α, and serum oxLDL/β2GP1 complex in the diabetic kidney disease	PloS one	2,022	cc-by	8,238	Data Availability Statement: All relevant data are within the manuscript and its Supporting Information files. Data Availability Statement: All relevant data are within the manuscript and its Supporting Information files. Funding: This study was supported by Q1Q2 Grant from Directorate of Research and Community Engagement, Universitas Indonesia (NKB- 0201/ UN2.R3.1/HKP.05.00/2019). The awardees is Rani Sauriasari. The url of funder website is https:// research.ui.ac.id/. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Rani SauriasariID1, Afina Irsyania Zulfa1, Andisyah Putri Sekar1, Nuriza Ulul Azmi1, Xian Wen Tan2, Eiji Matsuura2,3,4 1 Faculty of Pharmacy, Universitas Indonesia, Depok, Indonesia, 2 Department of Cell Chemistry, Okayama University Graduate School of Medicine, Dentistry, and Pharmaceutical Sciences, Okayama, Japan, 3 Collaborative Research Center (OMIC), Okayama University Graduate School of Medicine, Dentistry, and Pharmaceutical Sciences, Okayama, Japan, 4 Neutron Therapy Research Center, Okayama University, Okayama, Japan a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 rani@farmasi.ui.ac.id * rani@farmasi.ui.ac.id * rani@farmasi.ui.ac.id Abstract Oxidant species is reported as a major determinant in the pathophysiology of diabetic kidney disease. However, reactive oxygen species (ROS) formation in the initial phase and pro- gressing phase of diabetic kidney disease remains unclear. Therefore, we conducted this study to find out what ROS and their modified product are associated with eGFR in type 2 diabetes mellitus (T2DM) patients. A cross-sectional study was performed on 227 T2DM patients. The study subjects were divided into three groups based on their eGFR stage (Group 1, eGFR > 89 ml/min/1.73 m2; Group 2, eGFR = 60–89 ml/min/1.73 m2; and Group 3, eGFR < 60 ml/min/1.73 m2). Enzyme-linked immunosorbent assay (ELISA) was used to measure serum oxLDL/β2GPI complex and urinary 8-iso-PGF2α, while ferrous ion oxidation xylenol orange method 1 (FOX-1) was used to measure urinary hydrogen peroxide (H2O2). H2O2 significantly decreased across the groups, whereas OxLDL/β2GPI complex increased, but not significant, and there was no trend for 8-iso-PGF2α. Consistently, in the total study population, only H2O2 showed correlation with eGFR (r = 0.161, p = 0.015). Multiple linear regression analysis showed that significant factors for increased eGFR were H2O2, diastolic blood pressure, and female. Whereas increased systolic blood pressure and age were sig- nificant factors affecting the decrease of eGFR. We also found that urinary H2O2 had corre- lation with serum oxLDL/β2GPI complex in total population. This finding could lead to further research on urinary H2O2 for early detection and research on novel therapies of diabetic kid- ney disease. PLOS ONE RESEARCH ARTICLE OPEN ACCESS Citation: Sauriasari R, Zulfa AI, Sekar AP, Azmi NU, Tan XW, Matsuura E (2022) Role of urinary H2O2, 8-iso-PGF2α, and serum oxLDL/β2GP1 complex in the diabetic kidney disease. PLoS ONE 17(4): e0263113. https://doi.org/10.1371/journal. pone.0263113 Editor: Jaap A. Joles, University Medical Center Utrecht, NETHERLANDS Editor: Jaap A. Joles, University Medical Center Utrecht, NETHERLANDS Received: June 1, 2021 Accepted: January 13, 2022 Published: April 5, 2022 Editor: Jaap A. Joles, University Medical Center Utrecht, NETHERLANDS Received: June 1, 2021 Accepted: January 13, 2022 Published: April 5, 2022 Copyright: © 2022 Sauriasari et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. PLOS ONE PLOS ONE Introduction The unique characteristic of H2O2 is its ability to diffuse into the cell, forming hydroxyl radicals, and has the potential to attack biological molecules such as DNA, proteins, and lipids leading to cell damage [11, 12]. Lipid peroxidation has been considered to be a key mechanism for the development of ath- erosclerosis and vascular damage [7]. 8-iso-PGF2α are prostaglandin-like compounds pro- duced from the oxidation of arachidonic acid through non-enzymatic processes that are catalyzed by free radicals such as H2O2. The 8-iso-PGF2α production will circulate in blood vessels and be excreted in the urine of humans and animals [13, 14]. 8-iso-PGF2α is also capa- ble of activating platelets through its ability to bind to thromboxane prostanoid receptors [15]. p p y p erosclerosis and vascular damage [7]. 8-iso-PGF2α are prostaglandin-like compounds pro- duced from the oxidation of arachidonic acid through non-enzymatic processes that are catalyzed by free radicals such as H2O2. The 8-iso-PGF2α production will circulate in blood vessels and be excreted in the urine of humans and animals [13, 14]. 8-iso-PGF2α is also capa- ble of activating platelets through its ability to bind to thromboxane prostanoid receptors [15]. Another oxidatively modified-lipid product is the OxLDL/β2GP1 complex. OxLDL/β2GP1 complex is an oxidized product of LDL cholesterol in which some studies showed that OxLDL binds to endogenous β2GPI to form oxLDL/β2GP1 complex that can be found in the blood- stream of patients with chronic renal disease and diabetes mellitus [16]. The oxLDL/β2GP1 complex represents an important pathogenic event in the progression of atherosclerotic lesions and can be used to represent a substitute marker for oxidative inflammation in diabetes melli- tus [17]. In chronic renal disease, an increase in oxidized LDL (OxLDL) participates in the development of glomerulosclerosis and interstitial fibrosis [18, 19]. Another oxidatively modified-lipid product is the OxLDL/β2GP1 complex. OxLDL/β2GP1 complex is an oxidized product of LDL cholesterol in which some studies showed that OxLDL binds to endogenous β2GPI to form oxLDL/β2GP1 complex that can be found in the blood- stream of patients with chronic renal disease and diabetes mellitus [16]. The oxLDL/β2GP1 complex represents an important pathogenic event in the progression of atherosclerotic lesions and can be used to represent a substitute marker for oxidative inflammation in diabetes melli- tus [17]. In chronic renal disease, an increase in oxidized LDL (OxLDL) participates in the development of glomerulosclerosis and interstitial fibrosis [18, 19]. Introduction Diabetes mellitus is one of the most common non-communicable diseases. According to the International Diabetes Federation (IDF) Atlas 9th edition, in 2019 463 million people were liv- ing with diabetes and about three in four of those people (352 million) were of working age 1 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 PLOS ONE Oxidative stress and diabetic kidney disease Competing interests: The authors have declared that no competing interests exist. (i.e. 20–64 years old). This number is expected to increase to 417 million by 2030 and 486 mil- lion by 2045. In 2019, Indonesia was ranked seventh in the world for the highest prevalence of diabetes in the world and is predicted to remain so until 2030 [1]. Competing interests: The authors have declared that no competing interests exist. Abbreviations: BMI, body mass index; CG, cockroft-gault; CI, confidence interval; CKD-EPI, chronic kidney disease epidemiology collaboration; DA, delta absorbance; DNA, deoxyribonucleic acid; eGFR, estimated glomerular filtration rate; ELISA, enzyme-linked immunosorbent assay; ESRD, end- stage renal disease; FOX-1, ferrous ion oxidation xylenol orange method 1; HbA1c, hemoglobin A1c; HRP-SA, horseradish peroxidase-conjugated streptavidin; H2O2, Hydrogen peroxide; IDF, International Diabetes Federation; LDL, low-density lipoprotein; MDRD, modification of diet in renal disease; OR, odds ratio; OxLDL/β2GPI complex, oxidized low-density lipoprotein/beta-2- glycoprotein-I; ROS, reactive oxygen species; SOD, superoxide dismutases; TGF-β, transforming growth factor beta; TMB, tetramethylbenzidine; T2DM, type 2 diabetes mellitus; 8-iso-PGF2α, 8- isoprostaglandin F2α. Diabetes mellitus is associated with a high incidence of atherosclerotic complications that result from chronic metabolic abnormalities such as hyperglycemia and hyperlipidemia, in type 2 diabetes mellitus (T2DM) may enhance systemic oxidative stress, resulting in the exces- sive production of lipid peroxides and subsequently contributing to the pathogenesis of ather- othrombosis and microangiopathy, which is the common pathogenetic mechanism underlying diabetic vascular complications [2–4]. Diabetic nephropathy is one of the chronic complications in the form of decreased renal function characterized by proteinuria, hyperten- sion, and the declining glomerular filtration rate (GFR) [5]. Continuous hyperglycemia and hyperlipidemia induce overproduction of reactive oxygen species (ROS), resulting in renal cell apoptosis, which plays an important role in the molecular mechanism of renal damage in diabetes [6, 7]. Peroxides, free radicals, protein-, lipid- and DNA-modified substances are oxidative stress-related markers that have a pivotal role in cell damage [7, 8]. Hydrogen peroxide (H2O2) is one of a few peroxides used as a biomarker of oxi- dative stress [9, 10]. Introduction Although oxidative stress is known to contribute to diabetic kidney disease, it is still unknown whether oxidative stress is associated with renal hyperfiltration in the initial phase (G1 stage) of chronic kidney disease or associated with renal hypofiltration in the progressing phase (G3a-G5 stage). The purpose of this study was to determine whether oxidative stress markers such as H2O2, 8-iso-PGF2α, and oxLDL/β2GP1 complex affected renal function as shown by eGFR level in T2DM patients. In this study, we used the CKD-EPI equation to deter- mine the patient’s eGFR level, since the CKD-EPI equation shows less bias, greater accuracy, more precision, and also commonly used in clinical practice settings [20]. Materials and apparatus This study used the following apparatus: microplate reader (SUNRISE; Serial number: 1001006648; Firmware: V 3.32 08/07/08; XFLUOR4 Version: V 4.51); Multi-mode Microplate Reader FlexStation1 (Molecular Devices, USA); plate shaker (Heidolph Unimax 1010); micro- centrifuge (Allegra 64 R); AfinionTM AS100 Analyzer (Abbott); and ultrafreezer -80˚C (Bio- medical, Lab Tech). Chemicals used in this study were creatinine detection kit (Enzo Life Sciences, Cat. No. ADI 907–030 A); 8-iso-PGF2α ELISA kit (Enzo Life Sciences, Cat. No. ADI-900-010); AtherOx1 (oxLDL/β2GPI antigen complexes) kit (Corgenix, USA); and H2O2 reagents (xylenol orange, sorbitol, catalase enzyme, 30% H2O2, sulfonic acid, Fe(NH4)2(SO4) 6H2O, NaH2PO4.H2O, Na2HPO42H2O, and deionized water). Collection of urine and serum sample First morning urine samples (30 mL) were collected from each subject. Samples were stored at -80˚C until analysis. Urine samples were divided into 3–5 aliquots and were used for the deter- mination of urinary creatinine, albumin, H2O2, and 8-iso-PGF2α. Blood samples were col- lected from each participant through the vena mediana cubiti into vacutainer tubes. The blood was centrifuged at 1500–2000 rpm for 10 minutes, separated and stored at -80˚C until required for analysis. The sera of the blood samples were used for the determination of oxLDL/β2GP1 complex and serum creatinine. Measurement of eGFR Based on the results of serum creatinine measurements, the eGFR values were calculated using the CKD-EPI equation [22]. Study design and study population A multicenter study was conducted at RSK Sitanala Tangerang and Community Health Cen- ter, Pasar Minggu District, South Jakarta, Indonesia during 2015, 2016, and 2019. The study was approved by the Ethical Committee, Faculty of Medicine Universitas Indonesia of Cipto Mangunkusumo Hospital (No.76/UN2.F1/ETIK/2015 and 0222/UN2.F1/ETIK/2018) and PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 2 / 14 PLOS ONE Oxidative stress and diabetic kidney disease RSK Sitanala Ethical Committee (No. DM.04.04/1101663/2015). Written informed consent was obtained from all subjects. This study was performed on 227 T2DM patients who were divided based on their eGFR value. Group 1 included patients in G1 stage (eGFR > 89 mL/ min/1.73 m2) (n = 121); Group 2 included patients in G2 stage (eGFR 60–89 ml/min/1.73 m2) (n = 74), and Group 3 included patients in G3a to G5 stages (eGFR < 60 ml/min/1.73 m2) (n = 32). Inclusion criteria in this study were T2DM-diagnosed patients aged 36 years (late adulthood, according to age classification by Ministry of Health, Republic of Indonesia), fast- ing for at least eight hours before sampling, and willingness to give informed consent. The exclusion criteria were patients who could not provide blood and/or urine specimens or did not have complete clinical data; pregnant and breastfeeding women; patients with severe ane- mia and/or those receiving blood transfusions; patients suffering from heart disease, stroke, impaired liver function, and infectious disease (e.g. tuberculosis); and patients in kidney failure and/or currently undergoing renal replacement therapy. Measurement of serum creatinine Blood samples (10 mL) were taken from the subjects by certified phlebotomists from Prodia laboratory, an accredited clinical laboratory. Measurement of serum creatinine level in the sample was carried out by Prodia Laboratory by the enzymatic colorimetry method using crea- tininase, creatinase, and sarcosinase [21]. PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 Measurement of UACR UACR was measured using AfinionTM AS100 analyzer. The assay analyzes both the albumin and creatinine in a spot urine sample simultaneously within a single device. Albumin is 3 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 PLOS ONE Oxidative stress and diabetic kidney disease quantified using a immunometric membrane flow through assay, using monoclonal antibody- coated membrane and monoclonal antibodies conjugated to colloid gold. Creatinine is quanti- fied using an enzymatic colorimetris tests involving 4 enzymatic steps [23]. Measurement of serum OxLDL/β2GPI complex Serum oxLDL/β2GPI complex were measured by AtherOx ELISA kits (Corgenix, USA). The test was performed using a sandwich ELISA method. Diluted serum samples, calibrators, and controls were incubated in microwells coated with purified anti-human monoclonal antibody directed only to complexed β2GPI. A 100 μL of 1:50 dilution serum samples were incubated in microwells at room temperature for one hour. The microwells were washed four times with phosphate-buffered saline containing 0.05% polysorbate-20. After washing, anti-human apoB100 (LDL) monoclonal antibody conjugated to biotin was added to form complexes with the bound antigen, followed by 30 minutes’ incubation at room temperature. Following fur- ther washing, a horseradish peroxidase-conjugated streptavidin (HRP-SA) was added to form complex with the bound biotin-conjugated antibody and incubated for 30 minutes. Following further washing, the bound HRP-SA conjugate was assayed by the addition of tetramethylben- zidine (TMB) and H2O2 chromogenic substrate and incubated for 30 minutes. The color developed in the wells was directly proportional to the serum concentration of OxLDL/β2GPI antigen complex. The reaction was stopped with 0.36 N sulfuric acid. The optical density was read at a 450 nm (650 nm as reference). Measurement of urinary H2O2 Urinary H2O2 measurement was conducted by ferrous ion oxidation xylenol orange method 1 (FOX-1) assay [24]. This method is based on the oxidation of the Fe2+ reagent to Fe3+ by an oxidizing agent (H2O2). The oxidized Fe3+ will bind to ethylene xylenol orange (XO), generat- ing a color complex that has maximum absorption at 560 nm. Then, 20 μL of the urine samples were incubated with 20 μL of catalase solution (2200 U/ml in 25 mM phosphate buffer, pH 7.0). The samples were reacted with 160 μL of FOX-1 reagent pH 1.7–1.8 by addition of Na2HPO4 at room temperature for 30 minutes. Absorbance was measured with a microplate reader at 560 nm. Urinary H2O2 concentration was determined by calculating the absorbance difference in the samples with and without catalase [11]. Measurement of urinary 8-iso-PGF2α The analysis of 8-iso-PGF2α was carried out using competitive ELISA with commercial immu- noassay kits for 8-iso-PGF2α (catalog #ADI-900-010, Enzo Life Sciences, Farmingdale, NY, USA). The kit used a polyclonal antibody to bind to 8-iso-PGF2α in a competitive manner, with 8-iso-PGF2α in a sample or with an alkaline phosphatase molecule which has 8-iso- PGF2α covalently attached to it. After simultaneous incubation at room temperature, the excess reagents were washed away and substrate was added. After a short incubation time, the enzyme reaction stopped and the yellow color generated was read on a microplate reader at 405 nm. The intensity of yellow color was inversely proportional to the concentration of 8-iso- PGF2α in either the standards or the samples. The measured optical density was used to calcu- late the concentration of 8-iso-PGF2α. Statistical analysis Comparison among groups was carried out using the chi-square test for categorical variables. Kruskall–Wallis test followed by Mann–Whitney test were used to analyze nonparametric 4 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 PLOS ONE Oxidative stress and diabetic kidney disease data, and one-way ANOVA testing was used for parametric data. Correlation between vari- ables was analyzed using Spearman or Pearson tests, depended on data normality. A p-value of less than 0.05 was considered statistically significant. Multivariate analysis was also used to find the patterns and relationships between multiple variables, enabling factors affecting the results to be identified. All analysis was performed using IBM SPSS Statistics 22 software. Data were expressed in n (%), Mean ± SEM. Notes: p < 0.05 is considered statistically significant; eGFR, estimated glomerular filtration rate; UACR, urine albumin to creatinine ratio; SD, standard deviation; a: Kruskal–Wallis H test; b: one-way ANOVA test; c: chi-square test breviations: eGFR, estimated glomerular filtration rate; HbA1c, hemoglobin A1c. g g n (%), Mean ± SEM. dered statistically significant; eGFR, estimated glomerular filtration rate; UACR, urine albumin to creatinine ratio; SD, stand n (%), Mean ± SEM. dered statistically significant; eGFR, estimated glomerular filtration rate; UACR, urine albumin to creatinine ratio; SD, stand timated glomerular filtration rate; HbA1c, hemoglobin A1c. https://doi.org/10.1371/journal.pone.0263113.t001 Notes: p < 0.05 is considered statistically significant; eGFR, estimated glomerular filtration rate; UACR, urine albumin to creatinine ratio; SD, standard deviation; a: Kruskal–Wallis H test; b: one-way ANOVA test; c: chi-square test. ± SEM. ically significant; eGFR, estimated glomerular filtration rate; UACR, urine albumin to creatinine ratio; SD, standard deviatio ssed in n (%), Mean ± SEM. d d ll f d l l f l lb Abbreviations: eGFR, estimated glomerular filtration rate; HbA1c, hemoglobin A1c. a were expressed in n (%), Mean ± SEM. Kruskal–Wallis H test; b: one-way ANOVA test; c: chi-square test. onsidered statistically significant; eGFR, estimated glomerular filtration rate; UACR, urine albumin to creatinine ratio; SD, s st; b: one-way ANOVA test; c: chi-square test. Abbreviations: eGFR, estimated glomerular filtration rate; HbA1c, hemoglobin A1c. merular filtration rate; HbA1c, hemoglobin A1c. Characteristics of the study groups Data were obtained from 227 subjects (47 men and 180 women) consisted of 121 subjects with eGFR > 89 mL/min/1.73 m2, 74 subjects with eGFR 60–89 mL/min/1.73 m2, and 32 subjects with eGFR < 60 mL/min/1.73 m2 (Table 1). There were no significant differences in the pro- portion of gender, weight, height, body mass index (BMI), exercise habit, smoking habit, blood pressure, HbA1c, urine creatinine, and albuminuria status in the three groups of study subjects (Table 1). All groups mainly consisted of female subjects. The mean age of the study Table 1. Basic characteristics of study subjects. Table 1. Basic characteristics of study subjects. Trend of oxidative stress markers among groups Table 2 shows that there were no significant differences (p>0.05) in the levels of oxLDL/β2GPI complex and 8-iso-PGF2α among the groups. The results showed that oxLDL/β2GPI complex was increased linearly from group 1 (0.48 ± 0.03 units/mL) to group 2 (0.54 ± 0.05 units/mL), and then to Group 3 (0.63 ± 0.08 units/mL). In contrary, H2O2 was shown to be decreased from group 1 (89.99 ± 27.12 μmol/mg creatinine) to group 2 (45.73 ± 11.66 μmol/mg creati- nine), and to group 3 (32.83 ± 4.09 μmol/mg creatinine). H2O2 levels were found to be signifi- cantly different among the groups (p = 0.003). A unique pattern for 8-iso-PGF2α was found, in which there was an elevation in group 2 (47,356.27 ± 31,982.65 pg/mg creatinine), and then decrease in group 3 (13,172.56 ± 3,335.76 pg/mg creatinine) (Table 2). Characteristics of the study groups Characteristic of study subjects Group 1 Group 2 Group 3 p eGFR > 89 mL/min/1.73 m2 (n = 121) eGFR 60–89 mL/min/1.73 m2 (n = 74) eGFR < 60 mL/min/1.73 m2 (n = 32) Mean (%) or Mean ± SEM Gender (%) Men 19 (15.7) 18 (24.3) 10 (31.2) 0.100c Women 102 (84.3) 56 (75.7) 22 (68.8) Age (years) 56.79 ± 0.59 60.95 ± 0.82 60.91 ± 1.51 <0.001b Weight (kg) 61.14 ± 0.99 62.83 ± 1.25 58.17 ± 1.50 0.114b Height (cm) 152.60 ± 0.65 153.21 ± 0.93 152.75 ± 1.27 0.977a Body mass index (kg/m2) (n = 220) 26.28 ± 0.41 26.72 ± 0.51 24.97 ± 0.64 0.055a Exercise routine (%) (n = 166) Exercise 59 (66.3) 31 (56.4) 14 (63.6) 0.486c Do not exercise 30 (33.7) 24 (43.6) 8 (36.4) Smoking habit (%) (n = 167) Smoking 1 (1.1) 2 (3.6) 1 (4.5) 0.501c Not smoking 88 (98.9) 54 (96.4) 21 (95.5) Blood pressure Systole (mmHg) 128.91 ± 1.86 128.57 ± 2.55 133.13 ± 3.78 0.397a Diastole (mmHg) 79.78 ± 0.92 77.91 ± 0.99 77.23 ± 1.86 0.215a HbA1c (%) 9.41 ± 0.57 8.24 ± 0.20 7.98 ± 0.28 0.140a Urinary albumin (mcg/dL) (n = 195) 6054.82 ± 1996.65 4339.73 ± 793.70 8483.68 ± 4267.09 0.315a Urinary creatinine (mg/dL) (n = 227) 84.59 ± 5.15 86.05 ± 7.13 90.40 ± 11.86 0.808a Serum creatinine (mg/dL) 0.61 ± 0.01 0.90 ± 0.02 1.70 ± 0.16 <0.001a Albuminuria UACR (mcg/mg Cre) (n = 195) 105.017 ± 28.584 88.675 ± 26.007 119.600 ± 67.434 0.211a Normoalbuminuria (%) 74 (71.2) 42 (62.7) 17 (70.8) 0.488c Albuminuria (%) 30 (28.8) 25 (37.3) 7 (29.2) PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 5 / 14 PLOS ONE Oxidative stress and diabetic kidney disease subjects in all groups were above 55 years and subjects in group 1 were younger than group 2 and group 3 (p < 0.001). Correlations between markers Bivariate correlation between two oxidative stress markers was conducted using Spearman’s Rho Test. Significant correlations was only found in group 2 (eGFR 60–89 ml/min/1.73 m2). H2O2 was correlated with serum oxLDL/β2GPI complex (r = 0.247; p = 0.034) and also with 8-iso-PGF2α (r = 0.244; p = 0.036). In total population, a significant weak correlation was found between oxLDL/β2GPI complex and H2O2 (r = 0.145, p = 0.029). Correlations between eGFR with oxidative stress markers Spearman’s Rho correlation test was conducted to determine the correlation between eGFR and H2O2, 8-iso-PGF2α, and oxLDL/β2GPI complex. There was no correlation between mark- ers with eGFR in each group. However, in total population, H2O2 showed a weak positive cor- relation with eGFR (r = 0.161, p = 0.015) (Fig 1). Even after excluding one outlier, the correlation was still significant with the r being increased to 0.162. Neither 8-iso-PGF2α levels (r = 0.047; p = 0.483) nor oxLDL/β2GPI complex levels (r = -0.093; p = 0.164) showed signifi- cant correlation with eGFR in total population. Multivariate analysis Multiple linear regression analysis showed that significant factors for increased eGFR were H2O2 (standardized β = 0.157, p = 0.036), diastolic blood pressure (standardized β = 0.201, Table 2. Oxidative stress markers in study subjects. Characteristic of study subjects Group 1 Group 2 Group 3 p eGFR > 89 mL/min/1.73 m2 (n = 121) eGFR 60–89 mL/min/1.73 m2 (n = 74) eGFR < 60 mL/min/1.73 m2 (n = 32) Mean ± SE H2O2 (μmol/mg creatinine) 89.99 ± 27.12 45.73 ± 11.66 32.83 ± 4.09 0.003a 8-iso-PGF2α (pg/mg creatinine) 13,342.32 ± 2,346.87 47,356.27 ± 31,982.64 13,172.56 ± 3,335.76 0.883a OxLDL/β2GPI complex (units/ mL) 0.48 ± 0.03 0.54 ± 0.05 0.63 ± 0.08 0.121a Abbreviations: eGFR, estimated glomerular filtration rate; 8-iso-PGF2α, 8-isoprostaglandin F2α; oxLDL/β2GPI complex, oxidized low-density lipoprotein/beta- 2-glycoprotein-I. Data were expressed in n (%), mean ± SEM. Notes: p = significance (p < 0.05 is considered statistically significant); SEM = standard error mean; a = Kruskal–Wallis H tests. https://doi.org/10.1371/journal.pone.0263113.t002 Table 2. Oxidative stress markers in study subjects. Characteristic of study subjects Group 1 Group 2 Group 3 p eGFR > 89 mL/min/1.73 m2 (n = 121) eGFR 60–89 mL/min/1.73 m2 (n = 74) eGFR < 60 mL/min/1.73 m2 (n = 32) Mean ± SE H2O2 (μmol/mg creatinine) 89.99 ± 27.12 45.73 ± 11.66 32.83 ± 4.09 0.003a 8-iso-PGF2α (pg/mg creatinine) 13,342.32 ± 2,346.87 47,356.27 ± 31,982.64 13,172.56 ± 3,335.76 0.883a OxLDL/β2GPI complex (units/ mL) 0.48 ± 0.03 0.54 ± 0.05 0.63 ± 0.08 0.121a Abbreviations: eGFR, estimated glomerular filtration rate; 8-iso-PGF2α, 8-isoprostaglandin F2α; oxLDL/β2GPI complex, oxidized low-density lipoprotein/beta- Table 2. Oxidative stress markers in study subjects. (pg/ g ) , , , , , , OxLDL/β2GPI complex (units/ mL) 0.48 ± 0.03 0.54 ± 0.05 0.63 ± 0.08 0.121a Abbreviations: eGFR, estimated glomerular filtration rate; 8-iso-PGF2α, 8-isoprostaglandin F2α; oxLDL/β2GPI complex, oxidized low-density lipoprotein/beta- 2-glycoprotein-I. Data were expressed in n (%), mean ± SEM. Abbreviations: eGFR, estimated glomerular filtration rate; 8-iso-PGF2α, 8-isoprostaglandin F2α; oxLDL/β2GPI complex, oxidized low-density lipoprotein/beta- 2-glycoprotein-I. d ( ) Abbreviations: eGFR, estimated glomerular filtration rate; 8-iso-PGF2α, 8-isoprostaglandin F2α; oxLDL/β2GPI complex, oxidized low-density lipoprotein/beta- 2-glycoprotein-I p = significance (p < 0.05 is considered statistically significant); SEM = standard error mean; a = Kruskal–Wallis H tests. PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 6 / 14 PLOS ONE Oxidative stress and diabetic kidney disease Fig 1. Multivariate analysis Scatter plot for the correlation between eGFR with urine H2O2 in total population (n = 227). https://doi.org/10.1371/journal.pone.0263113.g001 Fig 1. Scatter plot for the correlation between eGFR with urine H2O2 in total population (n = 227). https://doi.org/10.1371/journal.pone.0263113.g001 Fig 1. Scatter plot for the correlation between eGFR with urine H2O2 in total population (n = 227). https://doi.org/10.1371/journal.pone.0263113.g001 https://doi.org/10.1371/journal.pone.0263113.g001 p = 0.035), and female (standardized β = 0.191, p = 0.005). Whereas increased systolic blood pressure (standardized β = -0.247, p = 0.011) and age (standardized β = -0.268, p<0.001) were significant factors affecting the decrease of eGFR (Table 3). The model was developed by adjusting UACR, 8-iso-PGF2α, OxLDL/β2GPI complex, HbA1c, and BMI. p = 0.035), and female (standardized β = 0.191, p = 0.005). Whereas increased systolic blood pressure (standardized β = -0.247, p = 0.011) and age (standardized β = -0.268, p<0.001) were significant factors affecting the decrease of eGFR (Table 3). The model was developed by adjusting UACR, 8-iso-PGF2α, OxLDL/β2GPI complex, HbA1c, and BMI. PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 PLOS ONE Oxidative stress and diabetic kidney disease Table 3. Factors affecting the eGFR level in T2DM patients (n = 189). Dependent variable Adjusted R square Predictor Standardized Coefficients β p-value eGFR (mL/min/1.73 m2) 0.182 H2O2 (μmol/mg Cre) 0.157 0.036 8-iso-PGF2α (pg/mg creatinine) 0.018 0.791 OxLDL/β2GPI complex (units/mL) -0.109 0.119 UACR (mcg/mg Cre) -0.061 0.402 HbA1c (%) 0.113 0.098 Systole (mmHg) -0.247 0.011 Diastole (mmHg) 0.201 0.035 Age (years) -0.268 <0.001 Gender (female) 0.191 0.005 BMI (kg/m2) 0.057 0.404 Abbreviations: CI, confidence interval; eGFR, estimated glomerular filtration rate; UACR, urine albumin to creatinine ratio; H2O2, hydrogen peroxide; 8-iso-PGF2α, 8-isoprostaglandin F2α; OxLDL/β2GPI complex, oxidized low-density lipoprotein/beta-2-glycoprotein-I; HbA1c, hemoglobin A1c; BMI, body mass index; statistically significant (p<0.05); statistically significant (p<0.001). Table 3. Factors affecting the eGFR level in T2DM patients (n = 189). Dependent variable Adjusted R square Predictor Standardized Coefficients β p-value eGFR (mL/min/1.73 m2) 0.182 H2O2 (μmol/mg Cre) 0.157 0.036 8-iso-PGF2α (pg/mg creatinine) 0.018 0.791 OxLDL/β2GPI complex (units/mL) -0.109 0.119 UACR (mcg/mg Cre) -0.061 0.402 HbA1c (%) 0.113 0.098 Systole (mmHg) -0.247 0.011 Diastole (mmHg) 0.201 0.035 Age (years) -0.268 <0.001 Gender (female) 0.191 0.005 BMI (kg/m2) 0.057 0.404 Abbreviations: CI, confidence interval; eGFR, estimated glomerular filtration rate; UACR, urine albumin to creatinine ratio; H2O2, hydrogen peroxide; 8-iso-PGF2α, 8-isoprostaglandin F2α; OxLDL/β2GPI complex, oxidized low-density lipoprotein/beta-2-glycoprotein-I; HbA1c, hemoglobin A1c; BMI, body mass index; statistically significant (p<0.05); Table 3. Factors affecting the eGFR level in T2DM patients (n = 189). Abbreviations: CI, confidence interval; eGFR, estimated glomerular filtration rate; UACR, urine albumin to creatinine ratio; H2O2, hydrogen peroxide; 8-iso-PGF2α, 8-isoprostaglandin F2α; OxLDL/β2GPI complex, oxidized low-density lipoprotein/beta-2-glycoprotein-I; HbA1c, hemoglobin A1c; BMI, body mass index; statistically significant (p<0.05); statistically significant (p<0 001) Abbreviations: CI, confidence interval; eGFR, estimated glomerular filtration rate; UACR, urine albumin to creatinine ratio; H2O2, hydrogen peroxide; 8-iso-PGF2α, 8-isoprostaglandin F2α; OxLDL/β2GPI complex, oxidized low-density lipoprotein/beta-2-glycoprotein-I; HbA1c, hemoglobin A1c; BMI, body mass index; statistically significant (p<0.05); Compared to other creatinine based-eGFR equations, CKD EPI is considered to be the most accurate and precise equation to diagnose CKD, especially for higher GFR values. Therefore, the American Diabetes Association recommends the use of the CKD-EPI equation to estimate the value of the glomerular filtration rate in patients with type 1 or type 2 diabetes mellitus [20, 22]. For UACR measurement, first-morning urine sample was used to avoid biological varia- tions in this study. https://doi.org/10.1371/journal.pone.0263113.t003 Discussion There was no significant difference in the proportion of male/female, subjects who routine in exercise, and subjects with smoking habit among three groups. There was also no significant difference in the mean BMI, blood pressure (systole and diastole), HbA1c, urine albumin, urine creatinine, and UACR level. However, there was significant differences in age and of course, serum creatinine. Creatinine clearance and glomerular filtration rate decreases with age [25]. This explains why the prevalence of chronic kidney disease is generally higher in the older age groups. The eGFR was defined by the volume of filtration entering the Bowman’s capsules per unit of time. The influencing factors of the GFR including capillary blood pressure, interstitial fluid pressure, plasma osmotic pressure, interstitial fluid osmotic pressure, and glomerular surface area. In this study we used the CKD-EPI equation to determine the patient’s eGFR level. Serum creatinine-based equations used to estimate eGFR values include age, sex, race, or weight variables to reflect creatinine formation from muscle mass or diet. There are three known equations, namely Cockroft-Gault (CG), The Modification of Diet in Renal Disease (MDRD), or The Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI). 7 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 PLOS ONE O2 −, superoxide anion radical; H2O2, hydrogen peroxide; OH, hydroxyl radical; DNA, deoxyribonucleic acid; oxLDL, oxidized Low Density Lipoprotein; β2GPI, β-2-Glycoprotein-I; TGF-β1, transforming growth factor β1; 8-iso-PGF2α, 8-isoprostaglandin F2α; TP receptor, thromboxane-prostanoid receptor; GFR reduction, Glomerular Filtration Rate reduction. Fig 2. Proposed mechanism of oxidative stress in diabetic kidney disease. O2 −, superoxide anion radical; H2O2, hydrogen peroxide; OH, hydroxyl radical; DNA, deoxyribonucleic acid; oxLDL, oxidized Low Density Lipoprotein; β2GPI, β-2-Glycoprotein-I; TGF-β1, transforming growth factor β1; 8-iso-PGF2α, 8-isoprostaglandin F2α; TP receptor, thromboxane-prostanoid receptor; GFR reduction, Glomerular Filtration Rate reduction. https://doi.org/10.1371/journal.pone.0263113.g002 that localized oxidative stress in renal tissue is a key component in the development of diabetic nephropathy [29]. However, it remains controversy, as to whether this is an early link between hyperglycemia and renal disease or develops because of other primary pathogenic mechanisms [29]. Our study showed that the urinary H2O2 decreased as CKD stage increased. The produc- tion of H2O2 is a continuation process of intramitochondrial superoxide anion radical (O2 −) production which can further initiate a range of damaging reactions through the production of hydroxyl radical (OH) through Fenton and Haber-Weiss reaction catalyzed by iron [8]. It is also reported that H2O2 increases extracellular matrix mRNA through TGF-β in human mesangial cells [31]. Further measurement of serum or tissue H2O and OH could help to explain this phenomenon. We did not found a significant positive correlation between eGFR and 8-iso-PGF2α, as we reported previously [14]. This could be due to the differences in the nature of the subjects between both studies. Most of the study subjects in previous study was normoalbuminuria G1 and G2 stage, whereas the present study consisted of G1, G2, and G3 stage patients with nor- moalbuminuria and albuminuria. 8-iso-PGF2α is a stable marker of a lipid peroxidation prod- uct with a prostaglandin-like structure that is mainly produced in vivo from esterification of arachidonic acid in tissues through non-enzymatic reactions catalyzed by free radicals such as O2 −and OH [32]. A meta-analysis also showed that 8-iso-PGF2α could be specifically used as a marker of oxidative stress [31]. Arachidonic acid is one of the phospholipid components that compose cell membranes, including those found in glomerular mesangial cells. Previous cell- culture studies found an increase of isoprostane synthesis in mesangial and endothelial glo- merular cells in a high glucose state [33]. 8-iso-PGF2α is known to have biological activity as a strong vasoconstrictor [14, 34]. PLOS ONE For diagnosing purpose, this method is more reliable compared to spot urine samples measurement [26]. Diabetic kidney disease known as diabetic nephropathy is a major chronic microvascular complication in long-standing type 1 and type 2 diabetic patients which can lead to end-stage renal disease (ESRD). The pathophysiological mechanisms in the development of diabetic kid- ney disease are known to be multifactorial [5, 27, 28]. Hyperglycemia is the initiating event that causes structural and functional changes such as glomerular hyperfiltration, glomerular and tubular epithelial hypertrophy, and microalbuminuria. This process is followed by the development of glomerular basement membrane thickening, accumulation of mesangial matrix, overt proteinuria, and eventually glomerulosclerosis (Fig 2) [27, 28]. Our novel finding in this study was the significant correlation between urinary H2O2 and eGFR. It means, the concentration of urinary H2O2 decreased as the CKD stage increased. We also found that urinary H2O2 had correlation with serum oxLDL/β2GPI complex in total pop- ulation. Moreover, for each group analysis, we only found significant correlation in Group 2 (eGFR 60–89 ml/min/1.73 m2). H2O2 has positive correlation with serum oxLDL/β2GPI com- plex (r = 0.247; p = 0.034) and also with 8-iso-PGF2α (r = 0.244; p = 0.036). However, there could be a bias in Group 2, since this group not only consisted with T2DM patients in G2 stage (also known as silent stage), but also T2DM patients with normal eGFR. In diabetic kidney disease, a number of sources of ROS are produced through enzymatic and non-enzymatic pathways [29, 30]. H2O2 is a product of oxidative metabolism as one of the ROS produced by superoxide dismutases (SODs) catalyzation [30]. Several functional enzymes within the mitochondria are particularly susceptible to ROS, leading to altered ATP synthesis, cellular calcium dysregulation, and induction of mitochondrial permeability transition, all of which predispose the cell, including renal cell, to necrosis or apoptosis [29]. It is postulated PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 8 / 14 PLOS ONE Oxidative stress and diabetic kidney disease PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 9 / 14 PLOS ONE Oxidative stress and diabetic kidney disease Oxidative stress and diabetic kidney disease Fig 2. Proposed mechanism of oxidative stress in diabetic kidney disease. PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 PLOS ONE The chemical composition of LDL makes these particles susceptible to oxidation by lipid oxidants [16, 17, 43]. Increased level of circulating LDL, especially in oxidized form, is associated with an increased risk of atherosclerosis, includ- ing glomerular atherosclerosis [18, 41, 42]. The mechanism of the relationship of hyperglycemia in T2DM to oxidative stress and the forming of atherosclerosis relates to the activity of glucose in monocyte activation. Monocyte activation in high glucose concentrations will increase expression of cytokines, interleukin-1β, and interleukin-6 and increase the release of O2 −, which can play a role in glucose-mediated oxidative stress [46, 47]. In addition, hyperglycemia can also cause a decrease in nitric oxide which leads to an increase in oxidative stress [47, 48]. This study has several strengths, one of which is the measurement of several variables simulta- neously, such as H2O2, 8-iso-PGF2α, and oxLDL/β2GPI complex to seek for oxidative stress bio- markers for early detection of renal damage. We also used a relatively large sample size in various analysis (n = 227). Every analysis was executed as similar with each other as possible by using the same equipment and conditions according to predetermined standard protocols. Due to the unstability of biomarkers in long-term storage, we performed the overall analysis of the biomark- ers in less than 1 month after sample collection. Furthermore, samples were stored in tempera- tures of -20˚C and -80˚C to maintain their stability. For all biomarkers, we used commercial kit and comply to the instruction guidelines. The data was conducted at minimum duplicate with % CV less than 20. If more than 20, we rerun the analysis. All biomarkers were evaluated for intra assay in previous studies (unpublished) and both quality control low sample (QCL) and low limit of quantification (LLOQ) have shown a %CV less than 15%. In addition, since there was no rou- tine renal function evaluation (eGFR and UACR) for T2DM patients in primary health care, this study is expected to help in examining renal function by measuring eGFR and UACR. However, there were also limitations in this study. The questionnaire given to subjects was not accurate enough to describe the real conditions. Another limitation of this study was the unavailability of treatment data that might affect markers. However, the study subjects were national health coverage (BPJS) outpatients in primary health care or polyclinic, in which their comorbidities and the treatments were not too varied. PLOS ONE Increasing synthesis of 8-iso-PGF2α in the kidneys was reported as increasing the activation of kidney TGF-β in rats with type 1 diabetes [33]. TGF-β is a major pro-fibrotic factor in diabetic nephropathy that could mediate an increase in glo- merular permeability to proteins, including albumin [19, 33]. 8-iso-PGF2α can interact with the thromboxane prostanoid (TP) receptor of platelets, leading to proteinuria as one of its pathological consequences [35]. In this study, we found that H2O2 was correlated with 8-iso- PGF2α and also with serum oxLDL/β2GPI complex in group 2. An old study but with relevant findings reported that H2O2 induces 21-aminosteroid-inhibitable F2-isoprostane production in renal epithelial cells that supports the in vivo report that its levels are elevated in ROS-linked renal injury models [36]. Besides direct cell injury, lipid peroxidation by generating F2-iso- prostanes, such as 8-iso-PGF2α, may further contribute to renal dysfunction through its vaso- constrictive mechanism [14, 36]. In total population, a significant weak correlation was found between oxLDL/β2GPI com- plex and H2O2. Research related to the role of oxLDL/β2GPI complex in various diseases are still lacking. Previously, the serum oxLDL/β2GPI complex were reported to be significantly higher in patients with chronic renal failure, chronic nephritis, and diabetes mellitus than those in healthy individuals [37]. However, in this study, even there was trend of increased oxLDL/β2GPI complex as CKD stage increases, the association was not significant. The associ- ation of serum oxLDL was formerly reported to have significant contribution to cardiovascular complications in diabetes [38–40]. Recently, Roumeliotis et al. also reported that circulating oxLDL levels amplified the magnitude of the association between proteinuria and progression of diabetic kidney disease [41]. They also found that oxLDL outperformed several confounders 10 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 PLOS ONE Oxidative stress and diabetic kidney disease and had better accuracy to predict deterioration of eGFR over time than baseline proteinuria [42]. The concepts of oxidative stress and the role of oxLDL in atherosclerosis as a chronic inflammatory process are widely known [38, 40, 43]. However, interestingly, it was reported that oxLDL/β2GPI complex, but not free oxidized LDL, was associated with the presence and severity of coronary artery disease [44]. In this study, we found that oxLDL/β2GPI complex level was affected by systolic and dia- stolic blood pressure. Lipid changes and lipoprotein oxidation are known to be related to oxi- dative stress in diabetes mellitus [39, 40, 45]. PLOS ONE Another limitation in this study was the varied age among subjects. To overcome this, we also performed a multivariate analysis. In addition, to minimize bias, we excluded patients who pregnant and breastfeeding women; patients with severe anemia and/or those receiving blood transfusions; patients suffering from heart disease, stroke, impaired liver function, and infectious disease (e.g. tuberculosis); and patients in kidney failure and/or currently undergoing renal replacement therapy. Despite the role of oxidative stress in early stage of diabetic kidney diseases is still unclear, our findings provide new insight and could lead to further research on the role of H2O2 in dia- betic kidney disease. S1 File. Raw data. S1 File. Raw data. (PDF) S1 File. Raw data. (PDF) Acknowledgments The authors would like to grateful to RSK Sitanala Tangerang and Pasar Minggu Primary Health Center to serve as study sites, and to Nadya Religiane Aretha, Nisrina Nurfitria, Fatma- wati Fadlin, Raja Andriany, Wilzar Fachri, and Rahmaningtyas Nurifahmi for their technical support. Funding acquisition: Rani Sauriasari. Investigation: Rani Sauriasari, Andisyah Putri Sekar, Eiji Matsuura. Methodology: Rani Sauriasari, Afina Irsyania Zulfa, Eiji Matsuura. Methodology: Rani Sauriasari, Afina Irsyania Zulfa, Eiji Matsuura. Project administration: Rani Sauriasari, Nuriza Ulul Azmi. Project administration: Rani Sauriasari, Nuriza Ulul Azmi. Supervision: Rani Sauriasari, Nuriza Ulul Azmi, Xian Wen Tan, Eiji Matsuura. Validation: Rani Sauriasari, Andisyah Putri Sekar, Nuriza Ulul Azmi, Xian Wen Tan. Writing – original draft: Afina Irsyania Zulfa, Andisyah Putri Sekar. Writing – review & editing: Rani Sauriasari, Afina Irsyania Zulfa, Andisyah Putri Sekar, Nur- iza Ulul Azmi, Xian Wen Tan, Eiji Matsuura. Conclusion Among three oxidative stress biomarkers studied, only urinary H2O2 had significant positive correlation with eGFR, even after adjusted by confounding variables. In total population, PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 11 / 14 PLOS ONE Oxidative stress and diabetic kidney disease urinary H2O2 showed correlation with serum oxLDL/β2GPI complex. This finding could lead to further research on urinary H2O2 for early detection and research on novel therapies of dia- betic kidney disease. urinary H2O2 showed correlation with serum oxLDL/β2GPI complex. This finding could lead to further research on urinary H2O2 for early detection and research on novel therapies of dia- betic kidney disease. Author Contributions Conceptualization: Rani Sauriasari, Afina Irsyania Zulfa. Data curation: Rani Sauriasari, Afina Irsyania Zulfa, Andisyah Putri Sekar, Xian Wen T Data curation: Rani Sauriasari, Afina Irsyania Zulfa, Andisyah Putri Sekar, Xian Wen Tan. Formal analysis: Afina Irsyania Zulfa, Andisyah Putri Sekar, Nuriza Ulul Azmi, Xian Wen Tan. Formal analysis: Afina Irsyania Zulfa, Andisyah Putri Sekar, Nuriza Ulul Azmi, Xian Wen Tan. Funding acquisition: Rani Sauriasari. PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 References 1. Saeedi P, Petersohn I, Salpea P, et al. Global and regional diabetes prevalence estimates for 2019 and projections for 2030 and 2045: Results from the International Diabetes Federation Diabetes Atlas, 9th edition. Diabetes Res Clin Pract. 2019; 157:107843. https://doi.org/10.1016/j.diabres.2019.107843 PMID: 31518657 2. Egan AM, Dinneen SF. What is diabetes? Medicine (Baltimore). 2018:1–4. https://doi.org/10.1016/j. mpmed.2018.10.002 3. Singh N, Singh N, Singh SK, Agarwal N, Kafle D. Effect of hyperglycemia on LDL oxidation in type 2 dia- betic patients. J Med Nutr Nutraceuticals. 2012; 1(2):111–114. https://doi.org/10.4103/2278-019X. 101302 4. Rahimi-Madiseh M, Malekpour-Tehrani A, Bahmani M, Rafieian-Kopaei M. The research and develop- ment on the antioxidants in prevention of diabetic complications. Asian Pac J Trop Med. 2016; 9 (9):825–831. https://doi.org/10.1016/j.apjtm.2016.07.001 PMID: 27633293 5. Bandeira E, Queiroz D. Diabetic nephropathy. Endocrinol Diabetes A Probl Approach. 2014;9781461486:455–466. https://doi.org/10.1007/978-1-4614-8684-8_36 6. Sabatini S, Kurtzman N. Role of Hyperfiltration in the Pathogenesis of Diabetic Nephropathy. Adv Pathog Diabet Nephrop. 2010. 12 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 PLOS ONE Oxidative stress and diabetic kidney disease 7. Kim J-Y, Lee J-W, Youn Y-J, et al. Urinary Levels of 8-Iso-Prostaglandin F2α and 8-Hydroxydeoxygua- nine as Markers of Oxidative Stress in Patients With Coronary Artery Disease. Korean Circ J. 2012; 42 (9):614. https://doi.org/10.4070/kcj.2012.42.9.614 PMID: 23091506 8. Halliwell B, Long LH, Yee TP, Lim S, Kelly R. Establishing Biomarkers of Oxidative Stress: The Mea- surement of Hydrogen Peroxide in Human Urine. Curr Med Chem. 2012; 11(9):1085–1092. https://doi. org/10.2174/0929867043365404 PMID: 15134507 9. Sauriasari R, Mahani RM, Sekar AP, Azizahwati. Urinary Hydrogen Peroxide and Renal Function Parameter of Type 2 Diabetes Mellitus Patients Consuming Metformin and Metformin-Sulfonylurea Rani. J Young Pharm. 2017; 9(1):5–8. https://doi.org/10.5530/jyp.2017 10. Yang H, Jin X, Lam CWK, Yan S. Oxidative stress and diabetes mellitus. Clin Chem Lab Med. 2011; 49 (11):1773–1782. https://doi.org/10.1515/CCLM.2011.250 PMID: 21810068 11. Sato Y, Ogino K, Sakano N, et al. Evaluation of urinary hydrogen peroxide as an oxidative stress bio- marker in a healthy Japanese population. Free Radic Res. 2013; 47(3):181–191. https://doi.org/10. 3109/10715762.2012.759218 PMID: 23249116 12. Wang DH, Ogino K, Sato Y, et al. Urinary Hydrogen Peroxide as Biomarker. Gen Methods Biomark Res their Appl. 2015:313–331. https://doi.org/10.1007/978-94-007-7696-8 13. McCullough PA. Oxidative stress reflected by increased F2-isoprostanes is associated with increasing urinary 11-dehydro thromboxane B2 levels in patients with coronary artery disease. Thromb Res. 2016; 148:85–88. https://doi.org/10.1016/j.thromres.2016.10.022 PMID: 27815971 14. Sauriasari R, Andrajati R, Azizahwati, et al. References Marker of lipid peroxidation related to diabetic nephropathy in Indonesian type 2 diabetes mellitus patients. Diabetes Res Clin Pract. 2015; 108(1):193–200. https:// doi.org/10.1016/j.diabres.2014.12.016 PMID: 25726335 15. Davı` G, Ciabattoni G, Consoli A, et al. In vivo formation of 8-iso-prostaglandin f2alpha and platelet acti- vation in diabetes mellitus: effects of improved metabolic control and vitamin E supplementation. Circu- lation. 1999; 99(2):224–229. https://doi.org/10.1161/01.cir.99.2.224 PMID: 9892587 16. Matsuura E, Lopez LR. Are Oxidized LDL/β2-glycoprotein I Complexes Pathogenic Antigens in Autoim- mune-mediated Atherosclerosis? Clin Dev Immunol. 2004; 11(2):103–111. https://doi.org/10.1080/ 10446670410001722186 PMID: 15330445 17. Berger JS, Rockman CB, Guyer KE, Lopez LR. Proatherogenic Oxidized Low-Density Lipoprotein/β2- Glycoprotein I Complexes in Arterial and Venous Disease. J Immunol Res. 2014. https://doi.org/10. 1155/2014/234316 PMID: 25405208 18. Lee HS, Song CY. Oxidized Low-Density Lipoprotein and Oxidative Stress in the Development of Glo- merulosclerosis. Am J Nephrol. 2009; 29:62–70. https://doi.org/10.1159/000151277 PMID: 18689980 19. Jiang Z, Seo JY, Ha H, et al. Reactive oxygen species mediate TGF-b1-induced plasminogen activator inhibitor-1 upregulation in mesangial cells. Biochem Biophys Res Commun. 2003; 309(4):961–966. https://doi.org/10.1016/j.bbrc.2003.08.102 PMID: 13679067 20. Jessani S, Levey AS, Bux R, et al. Estimation of GFR in South Asians: A Study From the General Popu- lation in Pakistan. Am J Kidney Dis. 2014; 63(1):49–58. https://doi.org/10.1053/j.ajkd.2013.07.023 PMID: 24074822 21. Suzuki M, Yoshida M. A new enzymatic serum creatinine measurement based on an endogenous crea- tine-eliminating system. Clin Chim Acta. 1984 Nov 15; 143(2):147–55. https://doi.org/10.1016/0009- 8981(84)90222-5 PMID: 6509773. 22. Levey AS, Stevens LA, Schmid CH, et al. A New Equation to Estimate Glomerular Filtration Rate Andrew. Ann Intern Med. 2009; 150(9):604–612. https://doi.org/10.7326/0003-4819-150-9-200905050- 00006 PMID: 19414839 23. Kvam C, Dworsky E, Campbell AT, et al. Development and Performance of an Albumin-Creatinine Ratio Assay on the Afinion AS100 Analyzer. Point Care. 2009; 8(1):16–20 https://doi.org/10.1097/POC. 0b013e3181971d57 24. Banerjee D, Jacob J, Kunjamma G, Madhusoodanan UK, Ghosh S. Measurement of urinary hydrogen peroxide by FOX-1 method in conjunction with catalase in diabetes mellitus-a sensitive and specific approach [2]. Clin Chim Acta. 2004; 350(1–2):233–236. https://doi.org/10.1016/j.cccn.2004.06.007 PMID: 15530455 25. Tziomalos K, Athyros VG. Diabetic nephropathy: New risk factors and improvements in diagnosis. Rev Diabet Stud. 2015; 12(1):110–118. https://doi.org/10.1900/RDS.2015.12.110 PMID: 26676664 26. Teo BW, Loh PT, Wong WK, et al. Spot urine estimations are equivalent to 24-hour urine assessments of urine protein excretion for predicting clinical outcomes. Int J Nephrol. 2015; 2015:156484. https://doi. org/10.1155/2015/156484 PMID: 25649135 27. Vinod PB. Pathophysiology of diabetic nephropathy. Clin Queries Nephrol. 2012; 1(2):121–126. References https:// doi.org/10.1016/S2211-9477(12)70005-5 13 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022 PLOS ONE Oxidative stress and diabetic kidney disease 28. Lin YC, Chang YH, Yang SY, Wu KD, Chu TS. Update of pathophysiology and management of diabetic kidney disease. J Formos Med Assoc. 2018; 117(8):662–675. https://doi.org/10.1016/j.jfma.2018.02. 007 PMID: 29486908 29. Forbes JM, Coughlan MT, Cooper ME. Oxidative Stress as a Major Culprit in Kidney Disease in Diabe- tes. Diabetes. 2008; 57(6):1446–1454. https://doi.org/10.2337/db08-0057 PMID: 18511445 30. Iglesias-de la Cruz MC, Ruiz-Torres P, Alcamı´ J, et al. Hydrogen peroxide increases extracellular matrix mRNA through TGF-β in human mesangial cells. Kidney Int. 2001; 59(1):87–95. https://doi.org/10. 1046/j.1523-1755.2001.00469.x PMID: 11135061 31. Rains JL, Jain SK. Oxidative stress, insulin signaling, and diabetes. Free Radic Biol Med. 2011; 50 (5):567–575. https://doi.org/10.1016/j.freeradbiomed.2010.12.006 PMID: 21163346 32. Erve TJ van ‘t, Kadiiska MB, London SJ, Mason RP. Classifying oxidative stress by F2-isoprostane lev- els across human diseases: A meta-analysis. Redox Biol. 2017; 12:582–599. https://doi.org/10.1016/j. redox.2017.03.024 PMID: 28391180 33. Montero A, Munger KA, Khan RZ, et al. F2-isoprostanes mediate high glucose-induced TGF-β synthe- sis and glomerular proteinuria in experimental type I diabetes. Kidney Int. 2000; 58:1963–1972. https:// doi.org/10.1111/j.1523-1755.2000.00368.x PMID: 11044216 34. Daray FM, Minvielle AI, Puppo S, Rothlin RP. Vasoconstrictor effects of 8-iso-prostaglandin E2 and 8- iso-prostaglandin F(2alpha) on human umbilical vein. Eur J Pharmacol. 2004 Sep 19; 499(1–2):189– 95. https://doi.org/10.1016/j.ejphar.2004.07.100 PMID: 15363966. 35. Bauer J, Ripperger A, Frantz S, Ergu¨n S, Schwedhelm E, Benndorf RA. Pathophysiology of isoprostanes in the cardiovascular system: implications of isoprostane-mediated thromboxane A2 receptor activation. Br J Pharmacol. 2014; 171(13):3115–3131. https://doi.org/10.1111/bph.12677 PMID: 24646155 36. Salahudeen A, Badr K, Morrow J, Roberts J 2nd. Hydrogen peroxide induces 21-aminosteroid-inhibita- ble F2-isoprostane production and cytolysis in renal tubular epithelial cells. J Am Soc Nephrol. 1995 Oct; 6(4):1300–3. https://doi.org/10.1681/ASN.V641300 PMID: 8589301. 37. Kasahara J, Kobayashi K, Maeshima Y, et al. Clinical Significance of Serum Oxidized Low-Density Lipoprotein/ß2-Glycoprotein I Complexes in Patients with Chronic Renal Diseases. Nephron Clin Pract. 2004; 98:c15–c24. https://doi.org/10.1159/000079923 PMID: 15361700 38. Steinberg D, Parthasarathy S, Carew TE, Khoo JC, Witztum JL. Beyond Cholesterol: modifications of low-density lipoprotein that increase its atherogenicity. N Engl J Med. 1989; 18(320):915–924. https:// doi.org/10.1056/NEJM198904063201407 PMID: 2648148 39. Chait A, Bornfeldt KE. Diabetes and atherosclerosis: is there a role for hyperglycemia? J Lipid Res. 2009; 50(Supplement):S335–S339. https://doi.org/10.1194/jlr.r800059-jlr200 PMID: 19029122 40. Kuyvenhoven JP, Meinders AE. Oxidative stress and diabetes mellitus Pathogenesis of long-term com- plications. Eur J Intern Med. 1999; 10(1):9–19. https://doi.org/10.1016/S0953-6205(99)00009-6 41. References Roumeliotis S, Georgianos PI, Roumeliotis A, Eleftheriadis T, Stamou A, Manolopoulos VG, et al. Oxi- dized LDL Modifies the Association between Proteinuria and Deterioration of Kidney Function in Protei- nuric Diabetic Kidney Disease. Life (Basel). 2021 May 29; 11(6):504. https://doi.org/10.3390/ life11060504 PMID: 34072583 42. Roumeliotis S, Roumeliotis A, Georgianos PI, Stamou A, Manolopoulos VG, Panagoutsos S, et al. Oxi- dized LDL Is Associated with eGFR Decline in Proteinuric Diabetic Kidney Disease: A Cohort Study. Oxid Med Cell Longev. 2021 Oct 19; 2021:2968869. https://doi.org/10.1155/2021/2968869 PMID: 34712380 43. Matsuura E, Kobayashi K, Inoue K, Lopez LR, Shoenfeld Y. Oxidized LDL/β2-glycoprotein I complexes: new aspects in atherosclerosis. Lupus. 2005; 14:736–741. https://doi.org/10.1191/ 0961203305lu2211oa PMID: 16218478 44. Bliden KP, Chaudhary R, Lopez LR, et al. Oxidized Low-Density Lipoprotein-β2-Glycoprotein I Complex But Not Free Oxidized LDL Is Associated With the Presence and Severity of Coronary Artery Disease. Am J Cardiol. 2016; 118(5):P673–678. https://doi.org/10.1016/j.amjcard.2016.06.023 PMID: 27401271 45. Lopez LR, Torre IG-D La, Matsuura E, Ames PRJ. Pro-Atherogenic Oxidized LDL/β2-Glycoprotein I Complexes in Diabetes Mellitus: Antioxidant Effect of Statins. J Cardiol Cardiovasc Ther. 2017; 4(5). https://doi.org/10.19080/JOCCT.2017.04.555649 46. Dasu MR, Devaraj S, Jialal I. High glucose induces IL-1β expression in human monocytes: mechanistic insights. Am J Physiol Metab. 2007; 293(1):E337–E346. https://doi.org/10.1152/ajpendo.00718.2006 PMID: 17426109 47. Matough FA, Budin SB, Hamid ZA, Alwahaibi N, Mohamed J. The Role of Oxidative Stress and Antioxi- dants in Diabetic Complications. Sultan Qaboos Univ Med J. 2012; 12(1):5–18. https://doi.org/10. 12816/0003082 PMID: 22375253 48. Yoshikawa T, Naito Y. What Is Oxidative Stress? Japan Med Assoc J. 2002; 124(11):1549–1553. 14 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0263113 April 5, 2022
https://openalex.org/W2328451093	https://sh.diva-portal.org/smash/get/diva2:916382/FULLTEXT01	English	null	Is It Possible To Govern Foreign Investments? Balancing Between Klondike And Poltava	Journal of business case studies	2,016	cc-by	11,196	Is It Possible To Govern Foreign Investments? Balancing Between Klondike And Poltava Ol G l b Ph D Söd ö U i i S d Olga Golubeva, Ph.D., Södertörn University, Sweden ABSTRACT FDI (foreign direct investment) of Swedbank in Ukraine is an example of unsuccessful investment in transition economies. The Case Study is presented in relation with Swedbank’s internationalization strategy and rapidly changing investment environment in transition economies and globally. Learning objectives include helping students to develop analytical skills in order to understand how political, economic, financial and social factors effect internalization strategy through FDI. The Case Study should help students to understand the importance of an appropriate long-term strategy of a firm entering transition economies, understand the investment environment of a foreign country and choose the best course of action for a distressed firm considering alternative scenarios. Lessons learned from the Case Study can be beneficial for students studying international business, but also for future decision-makers that would be acting in complex environments under rapidly changing situations. The author developed the Case from secondary sources: Swedbank’s annual reports and press-releases, information published by multilateral organizations and government agencies, research from investment banking houses and reputable news agencies. This Case is written solely for educational purposes and is not intended to analyze successful or unsuccessful internalization strategy through FDI in transition economies. Keywords: Case Study; Internalization Strategy; FDI (Foreign Direct Investments); Transition Economies http://www.diva-portal.org http://www.diva-portal.org This is the published version of a paper published in Journal of Business Case Studies. Citation for the original published paper (version of record): Access to the published version may require subscription. N.B. When citing this work, cite the original published paper. Permanent link to this version: http://urn.kb.se/resolve?urn=urn:nbn:se:sh:diva-29799 Volume 12, Number 2 Journal of Business Case Studies – Second Quarter 2016 INTRODUCTION n July 2007, Swedbank acquired the Ukrainian bank TAS from a local oligarch Sergej Tigipko for about 800 million USD (Swedbank, 2007, July 9). Growth prospects were enormous and many foreign banks were streaming to Ukraine, an investment Klondike at that time. Only 5 years later since arrival to Ukraine, in April 2013, Swedbank informed the market that it would totally discontinue its operations in Ukraine (Swedbank, 2013, April 1). Some journalists compared this Swedbank’s investment with a historical loss of the Swedish army during Poltava battle. How it could go so wrong so quickly? Is it possible to govern foreign investments? Or is it random probability that determines the outcome of acquisitions abroad? I I Copyright by author(s); CC-BY K’S INTERNALIZATION STRATEGY IN BALTIC COUNTRIES, RUSSIA & UKRAINE Swedbank, one of the four major Swedish banks, has its roots in the Swedish savings bank tradition dating back to 1820. In 2014 Swedbank had about 8 million private customers, more than 600,000 companies and organizations, over 14,000 employees and around 500 branches. Swedbank’s total income in 2014 exceeded 5 billion USD, annual profit was approximately 2.15 billion USD, total assets were 274 billion USD and lending to the public amounted to 171 billion USD1 (Swedbank, Annual report 2014). Swedbank is used to apply a traditional banking model focused on close customer relationships and personal advice to private individuals as well as small- and medium-sized companies. When reforms in Central and Eastern Europe (CEE) took place, the interest was turned to the new emerging markets, geographically closely located to Sweden – the Baltic countries and Russia, often labeled as transition economies. The Baltic economies had experienced an extended period of economic growth. Average annual real GDP growth was 8% in 2000–2006, which made the Baltic countries the fastest-growing region in the European The Clute Institute Copyright by author(s); CC-BY Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 Union (EU) (Swedbank, Annual report 2007). Among the key reasons for the strong growth was the Baltics’ inclusion in the EU in 2004, which led to harmonization of legislation and state institutions with the rest of Europe. At the same time, the three Baltic countries were harvesting the fruits of the radical reforms implemented in the 1990s. All three ranked high in terms of economic freedom, transparency and innovation. Rapid growth in the Baltic countries was also fueled by the consumer spending, which has been driven by higher disposable incomes and available credit from abroad. During that time, the Baltic countries were in an expansive period of the credit cycle. Mortgage loans represented approximately 20–30% of GDP in the Baltic countries, compared with 40–60% in the EU. Russia also enjoyed strong GDP growth, approximately 6.6% in 2006 (Swedbank, Annual report 2007). The financial sector in Russia, both the corporate and private markets, developed well. A large and rapidly growing Baltic and Russian financial markets created opportunities for foreign investors. In 1999, Swedbank obtained over a 50% stake in Hansabank, the leading financial institution in the Baltic countries (Swedbank, Annual report 1999). In 2002, a leasing company Hansa Leasing Russia was established in Russia (Swedbank, Annual report 2002). And, finally, in 2005 Hansabank successfully completed the acquisition of the Moscow-based OAO Kvest bank in Russia. The same year Swedbank purchased 100% of Hansabank’s shares and Hansabank become entirely part of Swedbank Group (Swedbank, Annual report 2005). Estonia, Lithuania and Latvia started to account for a growing share of Swedbank’s earnings. In 2008, for example, Swedish banking with 75% of Swedbank’s total lending generated only 52% of the total profit, while Baltic banking with 17% of lending generated 25% of total Swedbank’s profit (Swedbank, Annual report 2008). Swedbank’s lending in the Baltics in local currency grew 59% in 2006 and 33% in 2007 (Swedbank, Annual report 2007). Russia also started to roll out with new branches being opened in St. Petersburg and Kaliningrad (besides Moscow) and retail banking services being launched in addition to corporate banking. In 2007, Swedbank expected that the nominal annual growth in Baltics and Russia in the next three to five years would be about 10% (Swedbank, Annual report 2007). The financial sector would grow even faster, as penetration for banking products in the Baltic countries and Russia was still low. Copyright by author(s); CC-BY 83 Journal of Business Case Studies – Second Quarter 2016 Journal of Business Case Studies – Second Quarter 2016 Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 According to unaudited IFRS Management Accounts, TAS had total assets of 1,142 million USD and loans of 834 million USD as per 2006 year-end. Profit for 2006 amounted to modest 10 million USD, reflecting heavy investments in TAS’ branch network expansion. At year-end 2007, Swedbank in Ukraine had 3,433 employees and more than 125,000 private customers and 16,000 corporate customers. The head office situated in Kiev, and there were 191 more offices around Ukraine, an increase from 95 offices in 2006 (Swedbank, 2007, February 7). Lehman Brothers acted as a financial advisor to Swedbank (Forsberg, 2009). In 2006 TAS had equity of USD 127 million (Swedbank, 2007, February 7). The purchased price of 800 million USD paid by Swedbank was not a cheap deal even in 2007. Goodwill, i.e. difference between net worth of TAS banks and the acquisition price was about 673 million USD. According to Swedbank, goodwill was justified by long-term growth possibilities in the Ukrainian banking sector. “Through Swedbank in the Ukraine, we are now one of the 15 largest banks in the country. The Ukrainian banking market is still in its infancy, but substantial growth is expected to continue for many years to come”, - commented Jan Lidén, CEO of Swedbank at that time in the press-release from 9th of July 2007. Swedbank’s own risk department was asked to prepare a report summarizing risks associated with the FDI in the Ukrainian bank (Forsberg, 2009). Report highlighted both political and economic risks associated with such an investment, including internal political instability and on-going conflict between Ukraine and Russia regarding gas supply and transmission. Besides, the report pointed on the underdeveloped legal system, risk of devaluation of hryvna (UAH), Ukrainian national currency, and corruption. On the anti-corruption organization Transparency International's list of 168 countries, Ukraine is currently on a 130th place.2 These warning conclusions, however, could not stop the acquisition. An audit company Ernst & Young performed the so called ‘due diligence’ report in connection with the acquisition of TAS banks (Forsberg, 2009). Ernst & Young found that TAS, which was owned by Ukrainian oligarch Sergej Tigipko, provided loans to different companies belonging to Tigipko’s own business imperium. Swedbank, therefore, through acquisition of TAS banks, would inherit loans related to Tigipko’s companies. Journal of Business Case Studies – Second Quarter 2016 One might question if the Western investor should buy a bank with dependency on the business of the former owner. On the other hand, many banks in transition economies were founded as a ‘pocket banks’ of local oligarchs servicing their financial needs. The TAS’ dependency on companies belonging to Tigipko’s imperium was not a unique, but a standard practice in the Ukrainian business environment. An impressive profit exceeding 2.3 billion USD earned by Swedbank in 2007 (among others, due to contribution from the Baltic countries) was a good supporting argument for the expansion strategy (Swedbank, Annual report 2007). And Swedbank saw its leading position in the Baltic region as a springboard to the larger Ukrainian economy. "The current situation in Ukraine is similar to what we saw in the Baltics for 8-10 years ago. Possibilities are high in future", - explained Carl Eric Stålberg, at that time the Chairman of the Board of Swedbank AB, to Swedish business newspaper “Dagens industi” on the 7th of February 2007. Acquisition of TAS banks was finally approved by the Board of Directors of Swedbank AB, and no objections from the members of the Board were reported (Swedbank, Annual report 2007). Swedbank was not alone to be interested in the Ukrainian banking market, often described as an investment Klondike at that time. Swedish SEB was the first to establish in Ukraine via purchase of 90% of Agio Bank for 29 million USD. In 2007 SEB bought another bank, Factorial, with 53 branches and assets of 452 million USD for the price of approximately 116 million USD (Realtid Näringsliv, 2007). SEB’s establishment in Ukraine was in line with strategy to grow in the Eastern Europe. Even East Capital, Swedish assets’ management company, was drawn to the banking sector in Ukraine. In August 2006 East Capital purchased about 7% of Nadra bank for approximately 48 million USD; and in January 2008 almost 10% of Pivdennyi Bank was bought by East Capital for approximately 76 million USD. The P/BV (Price for Book Value ratio) at purchase was 3.6 for Nadra Bank and 4.5 – for Pivdennyi Bank (Öman Fondkomission, 2008). Swedish banks were accompanied by other foreign banks. In 2007, at least 10 Ukrainian banks found themselves in the hands of foreign owners (Tigipko, 2008; Hrebeniuk, 2012). Good deals turned Ukrainian oligarchs into dollar billionaires overnight – mostly because of the overheated banking sector. Journal of Business Case Studies – Second Quarter 2016 The banking market in Sweden is a mature one, characterized by a tough competition between banks. A question emerged, if Swedbank’s strategy should be changed from being primarily a Swedish bank to an international banking group with Swedish roots. “Swedbank’s aim is to grow. European countries with lower than-average GDP per capita and penetration of financial services, particularly in Eastern Europe, are considered to offer the greatest opportunities for long-term growth”, - stated Swedbank in the Annual report for 2007. Stable earnings in Sweden and strong balance sheet could support such a growing internationalization strategy. Ukraine, with its 47 million populations, was one of the countries that experienced the greatest economic decline after the fall of communism, by 60 percent between 1990 and 1999. However, between 2000 and 2006, GDP rose 52%, or 7.4% per year (Swedbank, Annual report 2007). Household consumption was growing strongly thanks to a robust job market and falling unemployment as well as double-digit increases in real disposable income. From the early days of the Orange Revolution in 2004 to the market’s peak at the end of 2007, the Ukrainian First Stock Trade System index boasted an impressive increase exceeding 500%. In 2007, the Ukrainian stock market was ranked the best performing stock market in the world, which explained the interest of foreign financial investors (Zelenyuk, 2007). Rapid credit growth in Ukraine was contributing to higher spending and investments. Credit growth reached 76% on an annual basis in July 2007. Bank loans accounted for over 15% of financing for corporate investments, compared with 8% in 2003–2004. However, forward-looking potential was attractive: GDP per capita in Ukraine was only 25% of the OECD average, leaving plenty of room for a "catch-up" growth (Swedbank, Annual report 2007). In early July 2007, Swedbank acquired the Ukrainian bank TAS-Kommerzbank with its subsidiary TAS-Investbank from the local oligarch Sergej Tigipko for about 800 million USD. An additional payment of up to 250 million USD should be payable in three years, subject to growth performance (Swedbank, 2007, July 9). Swedbank’s acquisition of TAS-Kommertzbank and TAS-Investbank was part of the internationalization strategy in growth outside the group’s home markets. After acquisition in Ukraine, the amount of Russian-speaking employees at Swedbank Group (including Baltic countries) had exceeded the amount of Swedish-speaking persons. The Clute Institute Copyright by author(s); CC-BY 84 Journal of Business Case Studies – Second Quarter 2016 Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 banks rapidly expanded regional network and volume of assets. Risks, performance and strategy were often subordinated to growth and marketing targets. The main drive for local oligarchs was to get a good price to be paid by foreign investors. To justify the gargantuan prices, the purchased bank was expected to provide sustainable rapid growth in assets at 10-12% annually, about 20% annual ROE (return on equity) and equity value growth in the range of 13-16% annually (Hrebeniuk, 2012). That was an approximate financial target of foreign investors regarding their acquisitions in the Ukrainian banking sector at that time. According to the National Bank of Ukraine (NBU), only 7.7% of the cumulative FDI into the share capital of Ukrainian companies went to the financial sector in 2004. In 2008, however, the figure was almost 30%. Over 2006- 2008, nearly 42% of 26 billion USD of FDI in Ukraine went to the financial sector as foreign interest in Ukrainian banks peaked (Hrebeniuk, 2012). There have been several positive impacts on the Ukrainian economy from European banking groups’ FDI inflow. Foreign banks introduced the new standards of corporate governance, risk management and customer service. Besides, most European banking groups are public companies, their operations closely watched by shareholders and transparency of operations is high. The statements of their Ukrainian subsidiaries, therefore, started to be prepared and audited under international financial reporting standards (IFRS) accounting rules, which increased the transparency of the Ukrainian banking sector. In late 2007, a strategy for Ukraine was formulated by Swedbank to further expand in the corporate market by offering companies an improved range of financial products as well as in private segments including the growing middle class and the mass market. Substantial volume growth and branch expansion were declared to be strategic cornerstones. It was decided that Swedbank in Ukraine should advance from 15th place to top ten banks in the country (Swedbank, Annual report 2007). Swedbank acquired two banks from Sergej Tigipko - TAS-Kommerzbank and TAS-Investbank. These two banks were re-named to Opened Joint Stock Company (OJSC) Swedbank and Closed Joint Stock Company (CJSC) Swedbank Invest respectively (Swedbank, Annual report 2007). The smaller bank, CJSC Swedbank Invest, had a niche offering loans primarily for investment projects and high net worth individuals, while the larger OJSC Swedbank was a universal bank. Journal of Business Case Studies – Second Quarter 2016 However, the banks’ product offerings partly overlap and banks even compete with each other for customers and projects. It was decided, however, to continue to run these two banks as separate units. It was also agreed that TAS’ former owner, Sergej Tigipko, would continue to act as CEO and would have overall responsibility for the bank’s business and development. Mr. Tigipko at that time had over 14 years of top level banking experience, including top positions as the Governor of Ukraine’s Central Bank as well as Minister of Economy and Vice Prime Minister (Swedbank, 2007, February 7; Waltersson, 2009). In connection with acquisition, it was agreed with Tigipko about additional payment of up to 250 million USD that could be payable in three years, subject to growth performance. Top management of TAS banks remained and continued to work for Swedbank (Swedbank, 2007, July 9). Journal of Business Case Studies – Second Quarter 2016 For this purpose, owners of Ukrainian Copyright by author(s); CC-BY The Clute Institute 85 Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 Journal of Business Case Studies – Second Quarter 2016 company. However, Bear Stearns could not be saved and was sold to JP Morgan Chase (Board of Governors of the Federal Reserve System, 2013). The collapse of the company was a prelude to the risk management meltdown of the investment banking industry. On September 15, 2008, Lehman Brothers, the fourth-largest investment bank in the US, filed for Chapter 11 bankruptcy protection following drastic losses in its stock and devaluation of its assets by credit rating agencies (CNBC, 2008). Swedbank was exposed to Lehman Brothers through some loans, but fortunately they were secured by real estates located in the US (Swedbank, 2008, October 7; Almgren, 2008, September 25). It is interesting to note that Lehman Brothers acted as a financial advisor to Swedbank in their acquisition of TAS banks in Ukraine. A downturn in economic activity started to lead to the 2008–2012’ global recession. Strong credit growth in the Baltic countries and interest rates that were too low relative to economic growth had begun to lead to an imbalance in prices, which was particularly noticeable in the real estate market. During 2007, when the real estate bubble burst after financing terms tightened, it started to be obvious that the rapid growth in the Baltic countries was not sustainable. The current account deficit, which had swelled in pace with rapidly growing imports, reached 20–30% of GDP in the Baltic countries. In December 2008, inflation was 7% in Estonia and 8.5% in Lithuania. Latvia’s inflation rate was 10.5% in December 2008, declined from slightly over 17% in mid-2008. In 2008, Latvia's GDP shrank by minus 4.6% and Estonia's by minus 3.6%, while Lithuania's slowed down to 3% (Swedbank, Annual report 2008). As the crisis swept across CEE, the economic reversal intensified: Estonia's GDP dropped by minus 16.2% year-on-year, Latvia’s by minus 19.6% and Lithuania’s by minus 16.8%. By mid-2009, all three countries experienced one of the deepest recessions in the world (Swedbank, Annual report 2009). Due to the growing global financial crisis, the small but open economy of Ukraine also started to face turbulence. From Q4 2008 until Q1 2009, the economic situation in Ukraine became intolerable. The local currency declined from UAH 5 to a low of 9.5 to the USD before stabilizing at around 8 to the USD. Journal of Business Case Studies – Second Quarter 2016 GDP declined by a ruinous 15% during the whole of 2009 and the stock market, by the end of Q1 2009, had given up all of its gains of the previous 3 years and returned to pre-Orange Revolution levels (Swedbank, Annual report 2009). Demand for Ukrainian production, especially metals, fell in connection with the global recession, and external capital flows essentially came to a halt. Since the debts of domestic banks were largely denominated in foreign currency, depreciation of UAH led to severe banking crises. At the same time, the Ukrainian economy was weighed down by a long-standing political crisis, which was delaying required structural reforms. In November 2008, the International Monetary Fund (IMF) offered help and approved a Stand-By Arrangement for 16.4 billion USD to stabilize the economy of Ukraine (IMF, 2008, November 5). During 2008, Swedbank started to be affected by rapidly changing macroeconomic conditions, particularly in the Baltic countries and Ukraine, where Swedbank operated. Lending growth in the Baltic countries slowed from 33% in 2007 to 6% in 2008. Loan losses increased to 230 million USD. Although the share of impaired loans was only slightly over 2%, it was 5 times higher than in the previous year (Swedbank, Annual report 2008). It became obvious that Swedbank should revise the strategy in the Baltics and Ukraine and break the loan growth. The expansion continued too long and employees were rewarded for growth. For 2009, Swedbank’s profit before credit impairments for the Baltic countries was 630 million USD. Credit impairments for Baltic banking, however, reached 2 billion USD. Profit in Ukraine for 2009 was about 68 million USD, while credit impairments of 890 million USD accompanied with a goodwill’ depreciation resulted in an operating loss of 1 billion USD. For comparison, operating profit for the Swedish banking’ business area during 2009 reached 1.1 billion USD while Swedbank Markets, an investment banking unit, generated profit of 452 million USD. Profit earned by the ‘Swedish part’ of the bank and its investment unit was not enough to off-set losses in Baltics and Ukraine. In 2009 Swedbank Group showed a total loss exceeding 1.3 billion USD (Swedbank, Annual report 2009). Swedbank’s share development at that time reflected worries and concerns from investors and external community. GLOBAL FINANCIAL CRISES AND ITS IMPACT ON SWEDBANK During 2000s, the global economy was stimulated by low interest rates, rising asset prices and credit expansion. At the same time, imbalances increased and debt levels rose to ever higher levels in both OECD countries and transition economies. 2008 saw the start of the most serious international financial crisis since the Great Depression of the 1930s. It resulted in the threat of collapse of large financial institutions, the bailout of banks by national governments, and downturns in stock markets around the world. In many countries, the housing market also suffered resulting in the subprime mortgage crisis, reintroducing the world to an era of bank failures, a credit crunch, private defaults and massive layoffs. The active phase of the crisis, which manifested as a liquidity crisis, can be dated from August 9, 2007, when BNP Paribas terminated withdrawals from three hedge funds citing a ‘complete evaporation of liquidity’ (Gangahar and Jones, 2007). In March 2008, the Federal Reserve Bank of New York provided an emergency loan to The Bear Stearns, a New York based global investment bank, try to avert a collapse of the The Clute Institute Copyright by author(s); CC-BY 86 Journal of Business Case Studies – Second Quarter 2016 During 2008, the OMX Stockholm 30 Index fell by 37 percent and the price of the Swedbank Class A share fell by Copyright by author(s); CC-BY The Clute Institute 87 Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 approximately 75 percent. The share reached a high for the year of SEK 182, on 3 January, and a low of SEK 42, on 22 December.3 Swedbank started to face one of the biggest challenges in the bank’s history since 1820s. It was a question of survival, and external help was needed to overcome the crises. In October 2008, Riksbank (Central Bank of Sweden) initiated a series of liquidity measures, lending money to Swedish banks in order to avoid a more serious credit crunch. The government decided at the same time to introduce a financial stability plan, part of which included a guarantee programme to refinance Swedish banks. The Swedish National Debt Office granted Swedbank’s application to participate in the Swedish state’s guarantee programme (Swedbank, Annual report 2008). During 2008 and 2009, Swedbank executed two additional rights issues that gave Swedbank about 3.7 billion USD (Swedbank, Annual reports 2008 and 2009). The rationale behind the rights issue was to take proactive and decisive action in this volatile and uncertain market environment. An important objective of the rights issue was to gradually leave the guarantee programme. Several projects in 2008 and 2009 at Swedbank centered on further enhancing the efficiency and quality of credit processes, from credit assessment to distressed debt management. In general, more stringent lending criteria have been imposed prioritizing to identify and target customers with strong loan servicing ratios. In addition, more resources were assigned to the team responsible for handling distressed debt. Being a traditional Swedish bank, Swedbank had limited experience to govern its international operations. It was decided to consolidate all operations outside Swedbank’s home markets in the strategic business area International Banking. International Banking’s organizational structure was settled on 1 July 2008 to improve support for international operations, especially in steering and risk control (Swedbank, Annual report 2008). Swedbank also started to change top management. The new President and CEO Michael Wolf took up his new position on 1 March 2009 (Swedbank, Annual report 2009). Since that time, many key managers were gradually replaced. Copyright by author(s); CC-BY Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 Cheap funding available from Sweden in USD, aimed to boost the growth strategy in Ukraine, had a negative impact after UAH depreciated by 60%. Such depreciation was especially suffered by private persons and companies that took loans in hard currency without having matching revenues in the same currency. Due to uncertainty on the market and global turbulence, Swedbank started to revise the growth targets downward. In 2008, it was decided that at least throughout 2009, the operations would focus on quality control rather than growth. New lending was almost abolished and Swedbank placed high quality requirements on new lending to small and medium-sized companies. In order to improve governance, risk management, transparency and achieve economies of scale, Swedbank announced in June 2009 that it would merge two Ukrainian banks acquired by Swedbank - OJSC Swedbank and CJSC Swedbank Invest (Swedbank, Annual report 2009). According to the NBU, OJSC Swedbank was Ukraine’s 15th largest bank by assets (15 billion UAH) at the time of a merger, and Swedbank-Invest – number 30th (with assets of 5.3 billion UAH). Following severe losses and changes in expectations for Ukraine, operations were adapted to lower business volumes. The number of branches has been reduced from over 200 in early 2009 to 92 at year-end 2010. The number of employees was reduced during the same year by 1 326 people to 1 554 (Swedbank, Annual report 2010). Swedbank’s special risk team for financial restructuring and recovery, FR&R, was established in Ukraine. The team was going through all sizable impaired loans, helping the bank either to restructure or to recover the loans. ‘Restructuring’ and ‘recovery’ became the major daily concepts replacing previously applied ‘growth strategy’ and ‘market share targets’. Hundreds of corporate loans and thousands of private credits had been worked out, which finally contributed to the better outcome. As a result, restructured loans in Swedbank Ukraine exceeded 288 million USD in 2009 (Swedbank, Annual report 2009). Although Swedbank started to adjust the strategy to the worsening market conditions, the move was obviously not enough quick to reflect the dramatic changes in the business environment. At the same time, Swedbank discovered that protection of creditor rights was extremely weak due to inadequate legislation and high corruption level. Swedbank decided to write off the remaining goodwill in Ukrainian investment in 2009 (Swedbank, Annual report 2009). Journal of Business Case Studies – Second Quarter 2016 The management team was strengthened with increased focus on risk management, communication and implementation of a common business model for the bank’s domestic and international markets. The reduction of the Group’s lending outside Sweden was one of the measures introduced in 2009 in order to reduce risks in the Swedbank Group. At the end of 2009, for example, the share of lending in the Baltic countries decreased to 12 percent of the total lending, down from 16 percent in 2008. The share in Ukraine and Russia declined to 1 percent, compared to 3 percent in 2008 (Swedbank, Annual report 2009). When the financial crisis reached Ukraine, the problems of two Ukrainian subsidiaries of Swedbank became obvious and logical – the consequences of the mass distribution of loans, including those in the overheated real estate market. For 2008, Swedbank Ukraine’s loans were 2.4 billion USD and gross impaired loans amounted to almost 128 million USD (Swedbank, Annual report 2008). For 2009, loan portfolio in Ukraine was down to approximately 1.2 billion USD, while gross impaired loans exceeded 1 billion USD (Swedbank, Annual report 2009). Performing portfolio, generating income for Swedbank Ukraine, was shrinking while non-performing (impaired portfolio) had grown, requiring more and more provisioning for bad debts. Swedbank Ukraine's rating was cut by the rating agency Fitch to 'E'-level on January, 29th, 2009. Fitch’s individual rating ‘E’ indicated that a bank had very serious problems, which either required or would likely to require external support. Fitch Ratings explained that it downgraded Swedbank Ukraine to reflect the bank's increased credit risk. According to Fitch, Swedbank's loan portfolio grew 170% over nine months of 2008, which was at very high speed taking into account the turbulent market situation. Besides, Swedbank had a high concentration of foreign-currency loans, higher than the majority of its peer banks. At the end of 2008, about 80% of Swedbank’s total loans in Ukraine were in foreign currency (Concorde Capital, 2009, January 29). Copyright by author(s); CC-BY The Clute Institute 88 Journal of Business Case Studies – Second Quarter 2016 Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 Journal of Business Case Studies – Second Quarter 2016 time acted as a CEO of Swedbank Ukraine. Lending the substantial amount to a related party would be prohibited under the Swedish Corporate Governance Code, and this is the case for other European banks (Rolander, 2009). The fact that one billion Swedish krona was not loaned to Sergej Tigipko himself, but to his companies, was not enough to explain and justify the situation to auditors, investors, customers and Swedish public (Mellqvist, 2009). In 2009, CEO of Swedbank Ukraine (and former owner of TAS banks) Sergey Tigipko resigned from his operational position. He was offered, however, the position of a member of the Council of Swedbank Ukraine (Hermele, 2009). The new CEO Reiner Müller-Hanke was appointed in Swedbank Ukraine (Swedbank, 2009, April 29). Reiner Müller-Hanke previously served as CEO of KMB BANK, owned by Intesa Sanpaolo, and hold several top management positions in the financial sector in Germany, Russia and South America. The famous Ukrainian banker and financial oligarch Tigipko was re-placed by a western style banker with broad international experience. The whole Tigipko’s management team was gradually changed to enhance top management under crises. In line with other prudent measures undertaken by Swedbank during Ukrainian crises, change of management from people, loyal to Ukrainian oligarch Tigipko, to western style top professionals, was a necessary step in execution of corporate governance. It was a lot of critic in press, however, that such change should occur much earlier. Already in September 2011, Swedbank had altered its strategy for the Ukrainian market and started exit the retail segment in order to focus solely on the corporate customers. “The change is in line with our strategy to offer universal banking on our four home markets and have more focused offering in niche markets”, - commented Swedbank in a press-release from 20th of September 2011. The performing mortgage portfolio amounting to 275 million USD was sold to Ukrainian Delta Bank at the end of May 2012. In November 2012, the sale of the retail secured non-performing loans (NPL) in the amount of 229 million USD to collection company Ukrborg and Russian Alfa bank was finalized (Swedbank, Annual report 2012). Sales of the retail portfolio occurred with sufficient discounts reflecting limited interest of investors in the Ukrainian assets at that time. Journal of Business Case Studies – Second Quarter 2016 Only 5 years later since arrival to Ukraine, in April 2013, Swedbank informed the market that it would totally discontinue its operations in Ukraine. A major part of the remaining portfolio was sold to a local oligarch – Mr. Mykola Lagun, the majority owner in the Ukrainian Delta Bank (Swedbank, 2013, April 1). Swedbank Ukraine was renamed to Omega Bank, which occupied the 59th place in the ranking of the NBU on assets at that time. It is interesting to note that the fourth biggest bank in Ukraine, Delta Bank, went bankrupt during 2015 after months of non-compliance, according to the National Bank of Ukraine. Two smaller banks in the Delta Banking group, including Omega Bank that was acquired from Swedbank, were categorized insolvent along with Delta Bank (RT News, 2015). 2013 would be known in Swedbank’s history as a year when Swedbank sold its subsidiary in Ukraine. Besides, operations in Russia were discontinued. In 2014 Swedbank also winded down the Russian subsidiary, which resulted in a loss from discontinued operations of 29 million USD (Swedbank, Annual report 2014). Copyright by author(s); CC-BY Journal of Business Case Studies – Second Quarter 2016 Goodwill of approximately 673 million USD gained by Swedbank as a result of acquisition of TAS banks disappeared in 2 years after investment was made (Waltersson, 2009). Investment banking company BG Capital Research estimated that Swedbank also injected about 450 million USD in equity and subordinated debt into its Ukrainian subsidiary in order to survive the crises situation (Vavryshchuk, 2010). According to NBU, the exposure to a single customer at the Ukrainian banks should not exceed 25% of the bank’s capital base, the same rules that apply in Sweden. Swedbank Ukraine inherited loans to Sergei Tigipko’s companies exceeding 160 million USD, which resulted in breaking the rules of NBU about lending limitation to a single customer (Avanza research, 2009; Forsberg 2009). In order to comply with NBU regulations, Swedbank either should minimize the exposure against Tigipko’s companies or inject additional capital. During financial crises, the difference between European and Ukrainian banks had escalated. The interests of owners of European banks lie in the orbit of the banking business. It is hard to imagine a situation where a parent bank based in Europe instructs its Ukrainian subsidiary to issue a loan to a specific company or to overlook the rules for issuing loans to one borrower or associated entities. Unlike them, most bank owners in Ukraine have other primary businesses and use their banks as donors more than anything else in times of crisis. Officially, they comply with the NBU’s restriction on lending more than 25% of the regulatory capital to one borrower. Yet, some analysts suggest that the real level of insider lending in some captive banks may exceed 50% of the total loan portfolio (Hrebeniuk, 2012). The regulatory authority does not monitor the entities of big business group owners deeply enough to determine this. In Sweden, criticism was expressed by several external parties, including Aktiespararna, Swedish Shareholders Association. Aktiespararna questioned how such a big loan could be provided to a single borrower, who at the same Copyright by author(s); CC-BY The Clute Institute 89 Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 Volume 12, Number 2 Performance of Swedbank’s share price reflected normalization of situation in the company. In 2013, the OMX Stockholm 30 Index rose by 21 per cent and OMX Stockholm Banks Index increased by 42 per cent. Swedbank’s A share gained 43 per cent during the year. 3 In line with other banks, Swedbank is focusing on structural changes undergoing in banking industry including broad usage of digital channels - telephone, internet and mobile. In addition to improvements for customers, digitalization creates opportunities for banks, which can profit from efficiency gains such as reduced use of cash in society. On the other hand, competition in banking increases as new niche players’ stream into the industry offering loans and providing banking services. “We want to be a modern and attractive bank in the future”, - writes Swedbank in the Annual Report for 2013. Swedbank also plans to maintain a robust balance sheet that can withstand economic swings. Priorities for 2015 and nearest future include improvement of customer value and increase of efficiency (Swedbank, Annual report 2014). The financial sector is currently undergoing significant change, including at a regulatory level within the EU. The annual stress tests of major Swedish banks conducted by the Riksbank and the Swedish Financial Supervisory Authority (SFSA) in 2014 showed Swedbank’s resilience. During 2014, the Common Equity Tier 1 capital ratio increased to 21.2 per cent, which is a very high level (Swedbank, Annual report 2014). The global recovery in 2014 was unbalanced. The US continued to grow, while the Eurozone was threatened by slower activity and deflation. Shaky global economic conditions could impact the relatively strong growth in Sweden and the Baltic countries. Increasing international competition and digitalization contribute to low Swedish inflation, which coupled with a low interest rate environment, is pinching bank profits and credit demand in the industrial sector. Urbanization and population growth, together with a shortage of housing construction in major Swedish cities, are driving credit demand in the real estate sector, but could mean even higher and more risky levels of household debt. Credit demand in Swedbank’s home remains uncertain. Swedbank’s lending rose by modest 2.6% in 2013; in Baltic Banking, the lending portfolio grew slightly in Estonia and Lithuania, but decreased in Latvia. The discontinuation of the Russian and Ukrainian operations reduced lending volume by 638 million USD in 2013 (Swedbank, Annual report 2013). SWEDBANK AND INVESTMENT CLIMATE IN UKRAINE: POST GLOBAL FINANCIAL CRISES In 2010 the global economy recovered more quickly than expected from the financial crisis. After major production losses in 2009, the Swedish and Baltic economies have begun to improve. The strongest recovery in 2010 was in Sweden, where GDP rose by over 5%. Swedbank’s profit amounted to 1 billion USD for 2010, an increase of 2.6 billion USD compared to 2009. The improvement was mainly due to significantly lower credit impairments in the Baltic countries, Russia and Ukraine (Swedbank, Annual report 2010). The profit for 2014 exceeded 2.1 billion USD, due to stable income, stable expenses and low credit impairments. The share of impaired loans was only 0.41 percent in 2014, which suggests that the credit quality in the loan portfolio is currently good (Swedbank, Annual report 2014). The Board of Directors of Swedbank adopted new goals for the bank’s risk appetite and risk tolerance, which serve as cornerstones to ensure that Swedbank remains a robust bank going forward. The new decentralized organization in Sweden was introduced by Swedbank, giving more decision-making power to managers working close to the clients. Decentralization should both improve customer satisfaction and allow for closer risk monitoring. Copyright by author(s); CC-BY The Clute Institute 90 Journal of Business Case Studies – Second Quarter 2016 Copyright by author(s); CC-BY Journal of Business Case Studies – Second Quarter 2016 Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 The US and EU will probably encourage private companies to invest in Ukraine with new technologies, plants, factories, businesses and joint ventures, similar to what they already did in the 1990’s with respect to the newly democratic countries in Eastern Europe. Such governmental encouragement will likely include direct grants, loan guarantees, insurance packages etc. International Monetary Fund (IMF) approved a two-year Stand-By Arrangement for Ukraine at the end of April 2014. The arrangement of 17 billion USD should aim to restore macroeconomic stability, strengthen economic governance and transparency, and launch sound and sustainable economic growth (IMF, 2014, April 30). The program remains highly challenging and continues to hinge crucially on the assumption that the political situation will stabilize and authorities will strongly perform and strict to the planned reforms. Foreign investors, however, need two critical conditions to invest in any emerging economy: political stability and economic predictability (Golubeva, 2001). These two conditions, unfortunately, are not in place in Ukraine in 2015. Political stability, self-sustained energy sector, comprehensive structural reforms, reducing corruption and improving business climate are needed to secure sustainable growth in Ukraine (USAID, 2010). The financial sector is typically a barometer for changes in long-term investment expectations for Ukraine. The escalation of the European debt crisis in 2011-2012 and limited ability of European banks to support their unprofitable foreign subsidiaries have forced many of them to review their international strategies. High administrative pressure, ad hoc monetary regulation and corruption are obstacles making the operating environment extremely difficult and unpredictable even for the local Ukrainian banks. But this investment environment is completely non-transparent and not-understandable for foreigners. Besides, Fitch credit rating agency estimated that around 60 percent of lending in Ukraine is still done in foreign currency, therefore, miss-matching of currencies remains a problem. Banks have been also cautioned by the U.S. Treasury about potentially suspicious transfers of financial assets by Yanukovich or members of his inner circle (Davies, 2014, February 27). Swedbank’s decision to leave Ukraine was in line with many other foreign banks, they continue to clean up their balance sheets, shrink assets and search for potential buyers to exit the Ukrainian market. The large supply of banks of different sizes and portfolios, compared to demand, led to a considerable decline in the prices for banks. Journal of Business Case Studies – Second Quarter 2016 The prices dropped to a half of the equity, ten times less than in pre-crisis level during 2006-2007 (Rekrut, 2013). Despite that, the list of potential buyers is restricted to local oligarchs and, in the best case, some Russian banks. The details on sale of Bank Forum from Commerzbank to Ukraine’s Smart-Holding in 2012 were not disclosed. Ukrainian credit agency estimated, however, that the deal did not exceed 100-200 million USD, whereas Commerzbank had invested 1 billion USD in its Ukrainian daughter over the five years previous to sale. Kreditprompank, owned by a pool of foreign investors, was sold in 2012 for the symbolic amount of one dollar (!) to the Ukrainian businessman Mr. Mykola Lagun, who later also acquired Swedbank Ukraine (Rekrut, 2013). As mentioned earlier, Lagun’s Delta Bank group became insolvent in 2015. Since 2009, more and more European banks have left Ukraine, including ING Bank, Erste Bank, Home Credit Group, Сredit Europe Bank, Societe Generale, Universal, Intesa Sanpaolo and Volksbank International (Hrebeniuk, 2012). While other foreign lenders have cut their Ukraine exposure in five years since Lehman collapsed - to 20 percent of Ukraine banking sector assets in 2012 from 40 percent in 2008, Russian banks still account for 12 percent. In 2014 they had around 28 billion USD of exposure. Credit agencies have said that Russian banks should be able to cover losses with earnings or by getting government support (Reuters, 2014, February 26). However, even Russian banks almost seized new lending waiting for economic and political stabilization. Journal of Business Case Studies – Second Quarter 2016 During 2014, lending volumes grew to 171 billion USD, mainly attributable to raise in mortgage and acquisition of Sparbanken Öresund (Swedbank, Annual report 2014). Sweden, Estonia, Latvia and Lithuania are Swedbank’s home markets. To support customers’ businesses, Swedbank also has operations in Norway, Finland, Denmark, the US, China and Luxembourg. In 2014 Swedbank decided to open a representative office in Johannesburg, South Africa, in order to explore business opportunities in the rapidly growing African market (Fond & Bank, 2014, June 9). This is, however, a minor investment in an office with few employees. Limited growth possibilities seem to remain a certain problem for Swedbank Group. The same challenge Swedbank faced before the global financial crises, when a step was taken to explore transition economies in Baltics, Russia and Ukraine. FDI’ stream into Ukraine stopped due to the global recession and the severe economic crisis affecting the country. State Statistics Service of Ukraine estimated that after partial recovery in 2010 (almost 24% increase compared to 2009), the FDI flux dried again, declining from 4.13 billion USD in 2012 to 2.86 billion USD in 2013 (Interfax- Ukraine, 2014). Apart from the economic downturn, the inefficient and corrupted legal system, complexity of legislation and regulation, poor contract enforcement and poor governance constitute serious impediments to investment. This is so despite the fact that the country has its strengths: a large domestic market, proven agricultural potential, energy and mineral resources and a strategic geographic location which makes it a transit hub and a gate to Europe. Russia’s annexation of Crimea, military operations in the Eastern Ukraine and escalation of the dispute between Ukraine's gas company Naftogaz and Gazprom of Russia led to a volatile political situation. The Clute Institute Copyright by author(s); CC-BY 91 Copyright by author(s); CC-BY Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 Volume 12, Number 2 First of all, the global macroeconomic environment should be taken into consideration: year 2008 saw the start of the most serious international financial crisis since 1930s. At the same time, in 2007 we saw the highest peak of prices for Ukrainian banks. “It is always easy to have comments in retrospect. Our investment in the Baltics in 1996 was strategically right then, and is right now, even if the countries are in trouble. The problems will be solved in time and there will be an interesting market in our region to remain active on”, - commented Swedbank's former Chairman Carl Eric Stålberg in an interview in 2008, just in the middle of the crises. He continued: “If you put everything in a retrospective, the timing on the Bank's entry into Ukraine could have been better… It is clear that it would have been fun to make this investment at a different time”. At the same time, Stålberg emphasized that Ukrainian exposure accounted only for 1.5 percent of total lending (Joons, 2008). But was it only the timing that was wrong? Swedbank’s CEO Michael Wolf stressed in the Annual report for 2010 that “an important lesson from the financial crisis for Swedbank and other banks is to maintain full control over liquidity risks and ensure long-term financing, rather than maximise profitability from a short-term perspective”. That is an important advice for banks when they considering FDI abroad. Swedbank’s experience is Ukraine also shows the importance of investment environment for decision-makers. As long as changes in the environment are smooth and predictable, the firm can use its knowledge to invest and reach an expected outcome. But the critical problem is when the environmental changes are not predictable, than the company should be able to adjust the strategy quickly to address the new non-expected challenges. It is impossible to answer the question if the decisions to leave Ukraine by foreign banks, literally at any cost, are right or wrong. What is obvious, however, that foreign banks entering transition economies should not limit their business strategy to growth plans based upon cheap forex loans. They should have a long-term strategy operating through the economic cycle, including proper risk management in place, clear corporate governance procedure, well-functioning steering of foreign activities by owners and possibilities to quickly adapt to radical changes. CONCLUSIONS: LESSONS LEARNED? All in, Swedbank’s investment in Ukraine can be ranked as one of the biggest investment failures that happened recently in Sweden. Some journalists compared this Swedbank’s investment with a historical loss of the Swedish army during Poltava battle. How it could go so wrong so quickly? Is it possible to govern foreign investments? Or is it random probability that determines the outcome of acquisitions abroad? Copyright by author(s); CC-BY The Clute Institute 92 Journal of Business Case Studies – Second Quarter 2016 Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 Although certain measures have been introduced, a lot of challenges remain to be addressed including high household debt in Sweden, risk weights5 for private mortgage loans and necessity to extend the maturity for the banks’ funding (Finansinspektionen, 2014). It is also important to decide upon clear responsibility structure among the different bodies steering the banking system. The report from 2010 from Riksdagen, Swedish Parliament, pointed that there is no formal mechanism through which SFSA and Riksbank can discuss their opinions on stability issues (Goodhart and Rochet, 2011). There is also no formal mechanism for documenting these opinions, whether you agree or not. When Riksbank did have a concern about Swedish banks' exposure to transition economies, there is actually no clear documentation on SFSA’s reaction about Riksbank's concern, and different people remember the situation in different ways. If (and when) the warning signals would appear, powerful action plan with clear responsibilities is needed, preventing the banking system from possible troubles. Lessons learned from unsuccessful investment cases deserve attention from both researchers and practitioners. Serious work is required both on a company and on institutional level in Sweden in order to avoid future Poltava losses. ACKNOWLEDGMENTS The author is grateful for helpful comments from the anonymous referees. Proofreading assistance of Dr. Alan Wood (UK) is acknowledged. Journal of Business Case Studies – Second Quarter 2016 According to Business Sweden4, as of 2014, about 100 Swedish companies are established in Ukraine and approximately 400 - have close business relationships. Case of Swedbank in Ukraine might help not only to understand why certain firms choose to invest in Ukraine, but also why many others did not invest, and eventually, should not invest. Former CEO Jan Lidén and Carl Eric Stålberg, Chairman of the Board of Swedbank AB at the time of crises, were often blamed for Swedbank’s expensive internalization strategy. There are some indications, however, that the Swedish lending crisis in the transition economies was not a mistake of few managers, but a systemic error. In 2006, Swedish banks had over 53.4 billion USD exposure to the Baltic states, which was also characterized by a mismatch of currencies. Mismatch of currencies happens if borrowing and lending amounts are distributed in different currencies. Already in 2006, Riksbank warned that this development could become unsustainable (Sveriges Riksbank, 2006). Sweden's financial supervisory authority (SFSA), Finansinspektionen, was aware of the risks, but took no countermeasures (Finansinspektionen, 2006). Global financial crises highlighted the urgency of better control over the financial system. The EU’s Capital Requirements Regulation (CRR) and Capital Requirements Directive IV (CRD IV), which contain new rules on capital adequacy, liquidity and corporate governance of banks, were approved in June 2013 and took effect on the 1st of January 2014. The Swedish banking system is dependent on market funding, therefore, is sensitive to adverse developments abroad. As decided by SFSA, Swedish banks must hold larger capital and liquidity buffers than the minimum requirements imposed by the EU. A part of the new Swedish requirements is that banks must set aside extra capital of 1% from September 2015 in what is known as the countercyclical capital buffer. This figure should be raised to 1.5% from 27th of June 2016, given the present economic conditions (Finansinspektionen, 2015, May 26). Copyright by author(s); CC-BY Copyright by author(s); CC-BY 93 The Clute Institute 93 AUTHOR BIOGRAPHY Olga Golubeva, Ph.D., is a senior lecturer at the Department of Business Administration, Södertörn University, is the contacting author. Dr. Golubeva’s academic career is combined with over 18 years of management experience in the investment and corporate banking. Dr. Golubeva’s research interests include international management, FDI, credit decisions and valuation of companies, financial and banking systems, Public-Private Partnership, sustainable investments. (Corresponding address: Dr. Olga Golubeva, Södertörn högskola, SE-14189 Huddinge, Sweden. olga.golubeva@sh.se) REFERENCES (2008, September 25). Swedbanks ordförande talar ut om krisen. Svenska Dagbladet. Retrieved from: Almgren, J. (2008, September 25). Swedbanks ordförande talar ut om krisen. Svenska Dagbladet. Retrieved from: http://www.svd.se/swedbanks-ordforande-talar-ut-om-krisen. (English translation: Chairman of the Board of Swedbank speaks about crises.) http://www.svd.se/swedbanks-ordforande-talar-ut-om-krisen. (English translation: Chairman of the Board of Swedbank speaks about crises.) p ) Avanza research (2009, December 16). Lån i Ukraina kan kosta Swedbank banktillstånd. Retrieved from: https://www.avanza.se/placera/redaktionellt/2009/12/16/lan-i-ukraina-kan-kosta-swedbank-banktillstand.html. (English translation: The loan in Ukraine can cost Swedbank a banking licence.) Avanza research (2009, December 16). Lån i Ukraina kan kosta Swedbank banktillstånd. Retrieved from: https://www.avanza.se/placera/redaktionellt/2009/12/16/lan-i-ukraina-kan-kosta-swedbank-banktillstand.html. (Eng p p translation: The loan in Ukraine can cost Swedbank a banking licence.) The loan in Ukraine can cost Swedbank a banking licen Board of Governors of the Federal Reserve System. Bear Stearns, JPMorgan Chase, and Maiden Lane LLC. Last updated August Board of Governors of the Federal Reserve System. Bear Stearns, JPMorgan Chase, and Maiden Lane LLC. Last updated August 2, 2013, Retrieved from: http://www.federalreserve.gov/newsevents/reform bearstearns.htm. , , p g _ CNBC (2008, September 15). Lehman Brothers Files For Bankruptcy, Scrambles to Sell Key Business. Wall street in crisis – a CNBC special report. Retrieved from: http://www.cnbc.com/id/26708143 p g _ CNBC (2008, September 15). Lehman Brothers Files For Bankruptcy, Scrambles to Sell Key Business. Wall street in crisis – a CNBC i l R i d f h // b /id/26708143 p g _ CNBC (2008, September 15). Lehman Brothers Files For Bankruptcy, Scrambles to Sell Key Business. Wall street in crisis – a CNBC l R i d f h // b /id/26708143 p g _ CNBC (2008, September 15). Lehman Brothers Files For Bankruptcy, Scrambles to Sell Key Business. W CNBC (2008, September 15). Lehman Brothers Files For Bankruptcy, Scrambles to S CNBC special report. Retrieved from: http://www.cnbc.com/id/26708143 CNBC (2008, September 15). Lehman Brothers Files For Bankruptcy, Scrambles to S CNBC special report. Retrieved from: http://www.cnbc.com/id/26708143 port. Retrieved from: http://www.cnbc.com/id/2670814 Concorde Capital (2009, January 29). Swedbank Ukraine's rating cut by Fitch to 'E'. Retrieved from: http://concorde.ua/en/research/daily/6621/#sthash.BeUw524K.dpuf Davies, M. (2014, February 27). Foreign banks in Ukraine face choice – hunker down or cut losses. Reuters news, Moscow. Finansinspektionen (2006). Finanssektorns Stabilitet, 2006:14. Retrieved from: http://www.fi.se/Tillsyn/Rapporter/Listan/Finanssektorns-stabilitet-2006-200614/. (English translation: Financial sector’s stability). Finansinspektionen (2014, June 12). Household debt and bank’s market funding – risks to stability. Press release. Retrieved from http://www.fi.se/Folder-EN/Startpage/Press/Press-releases/Listan/Household-debt-and-banks-market-funding--risks-to- stability/ Finansinspektionen (2014, June 12). REFERENCES Household debt and bank’s market funding – risks to stability. Press release. Retrieved from: http://www.fi.se/Folder-EN/Startpage/Press/Press-releases/Listan/Household-debt-and-banks-market-funding--risks-to- stability/ Finansinspektionen, (2015, May 26). Proposal to amend the regulations on the countercyclical capital buffer. Press release. Retrieved from: http://www.fi.se/Folder-EN/Startpage/Supervision/Miscellaneous/Listan/Proposal-to-amend-the- regulations-on-the-countercyclical-capital-buffer/ Fond & Bank (2014, June 9). Swedbank öppnar kontor I Johannesburg. Retrieved from: http://www.fondochbank.se/Nyhetsarkivet/Fond--Bank-Online/2014/06/09/Swedbank-startar-kontor-i-Johannesburg/ (English translation: Swedbank opens a representative office in Johannesburg.) Fond & Bank (2014, June 9). Swedbank öppnar kontor I Johannesburg. Retrieved from: http://www.fondochbank.se/Nyhetsarkivet/Fond--Bank-Online/2014/06/09/Swedbank-startar-kontor-i-Johannesburg/ (English translation: Swedbank opens a representative office in Johannesburg.) Copyright by author(s); CC-BY The Clute Institute 94 Journal of Business Case Studies – Second Quarter 2016 Retrieved from: http://www ohman se/upload/ALL/NewYorkPresentation/9%20%20East%20Capital%20Explorer%20Financial pdf Öman Fondkomission (2008, May 7). East Capital Explorer. Financial Institutions Fund. Presentation. Retriev http://www.ohman.se/upload/ALL/NewYorkPresentation/9%20%20East%20Capital%20Explorer% http://www.ohman.se/upload/ALL/NewYorkPresentation/9%20%20East%20Capital%20Explorer%20F Realtid Näringsliv (2007, November 16). SEB Köper bank i Ukraina. Realtid Näringsliv. Retrieved from: http://www.realtid.se/ArticlePages/200711/16/20071116143956_Realtid157/20071116143956_Realtid157.dbp.asp (English translation: SEB buys a bank in Ukraine.) Realtid Näringsliv (2007, November 16). SEB Köper bank i Ukraina. Realtid Näringsliv. Retrieved from: http://www.realtid.se/ArticlePages/200711/16/20071116143956_Realtid157/20071116143956_R Realtid Näringsliv (2007, November 16). SEB Köper bank i U http://www.realtid.se/ArticlePages/200711/16/2007 (English translation: SEB buys a bank in Ukraine.) http://www.realtid.se/ArticlePages/200711/16/20071116143956_Realtid157/20071116143956_Realtid157.dbp.asp y ) 3, January 1). What is making foreigners to sell Ukrainian banks for $1? Credit rating agency of Ukraine. ( g y ) Rekrut, S. (2013, January 1). What is making foreigners to sell Ukrainian banks for $1? Credit rating agency of Ukraine. Retrieved from: http://www.credit-rating.ua/en/analytics/opinion/13282/ Rekrut, S. (2013, January 1). What is making foreigners to sell Ukrainian banks for $1? Credit rating agency of Ukraine. Retrieved from: http://www.credit-rating.ua/en/analytics/opinion/13282/ Reuters (2014, February 26). 3-Russia's largest banks halt new loans in Ukraine eyeing political risk. Retrieved fr http://www.reuters.com/article/ukraine-crisis-idUSL6N0LV40J20140226 Reuters (2014, February 26). 3-Russia's largest banks halt new loans in Ukraine eyeing political risk. Retrieved from: http://www.reuters.com/article/ukraine-crisis-idUSL6N0LV40J20140226 Rolander, D. (2009). Ukraina svider för Swedbank. Aktiespararen, 10. Retrieved from: http://www.aktiespararna.se/analys swedbank0910 (English translation: Ukraine Rolander, D. (2009). Ukraina svider för Swedbank. Aktiespararen, 10. Retrieved from: http://www.aktiespararna.se/analys_swedbank0910 (English translation: Ukraine hurts Swedbank.) ( ) p http://www.aktiespararna.se/analys_swedbank0910 (English translation: Ukraine hurts Swedbank.) RT News (2015, March 15). Ukraine’s fourth largest lender Delta Bank insolvent – national bank. Retrieved from: https://www.rt.com/business/237161-ukraine-nbu-deltabank-bankruptcy/ RT News (2015, March 15). Ukraine’s fourth largest lender Delta Bank insolvent – national bank. Retrieved from https://www rt com/business/237161 ukraine nbu deltabank bankruptcy/ Sveriges Riksbank (2006). Finansiell stabilitet 2006:1 and 2006:2. Retrieved from: http://www.riksbank.se/sv/Press-och- publicerat/Publicerat-fran-Riksbanken/Finansiell-stabilitet/Rapporten-Finansiell-stabilitet/2006/ (English translation Financial stability 2006:1 and 2006:2.) Sveriges Riksbank (2006). Finansiell stabilitet 2006:1 and 2006:2. Retrieved from: http://www.riksbank.se/sv/Press-och- bli t/P bli t f Rik b k /Fi i ll t bilit t/R t Fi i ll t bilit t/2006/ (E li h t l ti Sveriges Riksbank (2006). Finansiell stabilitet 2006:1 and 2006:2. Retrieved from: http://www.riksbank.se/sv/Press-och- publicerat/Publicerat-fran-Riksbanken/Finansiell-stabilitet/Rapporten-Finansiell-stabilitet/2006/ (English translation Sveriges Riksbank (2006). Finansiell stabilitet 2006:1 and 2006:2. Retrieved from: http://www.riksbank.se/sv/Press-och- publicerat/Publicerat-fran-Riksbanken/Finansiell-stabilitet/Rapporten-Finansiell-stabilitet/2006/ (English translation Financial stability 2006:1 and 2006:2.) Swedbank AB. Journal of Business Case Studies – Second Quarter 2016 Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 Forsberg, B. (2009, May 29). Guldjakten i Ukraina. Affärsvärlden. Retrieved from: http://www.affarsvarlden.se/hem/nyheter/article2591729.ece (English translation: Chasing gold in Ukraine.) berg, B. (2009, May 29). Guldjakten i Ukraina. Affärsvärlden. Retrieved from: http://www.affarsvarlden.se/hem/nyheter/article2591729.ece (English translation: Chasing gold in Ukrain Gangahar, A. & Jones, A.(2007, August 9). BNP Paribas investment funds hit by volatility. Financial times. Retrieved from: http://www.ft.com/intl/cms/s/0/9a4cabc4-464d-11dc-a3be-0000779fd2ac.html#axzz3z2o8Zi14 nd Rochet, J.-C. (2011). Utvärdering av Riksbankens penningpolitik och arbete med finansiell stabilitet 2005– Goodhart, C. and Rochet, J.-C. (2011). Utvärdering av Riksbankens penningpolitik och arbete med finansiell stabilitet 2005– 2010, Rapport från Riksdagen 2010/11:RFR5, Riksdagstryckeriet, Stockholm. (English translation: Evaluation of Central Bank’ of Sweden monetary policy and work with the financial stability in 2005–2010). Goodhart, C. and Rochet, J.-C. (2011). Utvärdering av Riksbankens penningpolitik och arbete med finansiell stabilitet 2005– 2010, Rapport från Riksdagen 2010/11:RFR5, Riksdagstryckeriet, Stockholm. (English translation: Evaluation of Golubeva, O. (2001). Foreign Investment Decision-Making in Transition Economies, PhD thesis, School of Business, Stockholm University, Sweden. Hermele, B. (2009, April 29). Swedbank offrar Ukraina-chefen. Realtid. Retrieved from: http://www.realtid.se/ArticlePages/200904/29/20090429084605_Realtid591/20090429084605_Realtid591.dbp.asp. (English translation: Swedbank offers the Ukrainian Manager.) (English translation: Swedbank offers the Ukrainian Manager.) Hrebeniuk, Y. (2012, October 23). Foreign Banks Flee Ukraine. The Ukrainian Week International Edition. Retrieved from: http://ukrainianweek.com/Economics/63007 MF (2008, November 5). IMF Approves US$16.4 Billion Stand-By Arrangement for Ukraine. Press Release N Retrieved from: https://www.imf.org/external/np/sec/pr/2008/pr08271.htm p g p p p IMF (2014, April 30). Executive Board Approves 2-Year US$17.01 Billion Stand-By Arrangement for Ukraine, US$3.19 Billion for immediate Disbursement. Press Release No. 14/189. Retrieved from: https://www.imf.org/external/np/sec/pr/2014/pr14189.htm IMF (2014, April 30). Executive Board Approves 2-Year US$17.01 Billion Stand-By Arrangement for Ukraine, US$3.19 Billion for immediate Disbursement. Press Release No. 14/189. Retrieved from: Interfax-Ukraine (2014, February 17). Foreign direct investment in Ukraine in 2013 shrinks to $2.8 billion. KyivPost. Retrieved from: http://www.kyivpost.com/content/business/foreign-direct-investment-in-ukraine-in-2013-shrinks-to-28-billion- 336936.html Joons, A. (2008, December 4). Stålberg: Tajming i Ukraina kan diskuteras. Svd Näringsliv. Retrieved from: http://www.svd.se/stalberg-tajming-i-ukraina-kan-diskuteras (English translation: Stålberg: Time of entry into Ukraine can be discussed.) ) Mellqvist, G. (2009, April 24). Swedbank-direktör har jättelån i Ukraina. Dagens industri. (English translation: Swedbank’s Director has huge loans in Ukraine”.) g ) Öman Fondkomission (2008, May 7). East Capital Explorer. Financial Institutions Fund. Presentation. Retr http://www.ohman.se/upload/ALL/NewYorkPresentation/9%20%20East%20Capital%20Explore g ) ondkomission (2008, May 7). East Capital Explorer. Financial Institutions Fund. Presentation. Retrieved from Öman Fondkomission (2008, May 7). East Capital Explorer. Financial Institutions Fund. Presentation. Copyright by author(s); CC-BY concentrate the power.) p ) Zelenyuk, V. (2007). The Ukrainian Stock Market. Beyond Transition Newsletter, the World Bank's Europe and Central Asia Department. Retrieved from: http://web.worldbank.org/WBSITE/EXTERNAL/NEWSLETTERS/EXTTRANSITION/EXTDECBEYTRANEWLET/ 0,,contentMDK:21491843~pagePK:64168445~piPK:64168309~theSitePK:1542353,00.html p ) 7). The Ukrainian Stock Market. Beyond Transition Newsletter, the World Bank's Europe and Central Asia ent. Retrieved from: Journal of Business Case Studies – Second Quarter 2016 Retrieved from: http://www.swedbank.com Swedbank, Annual report 2002 Swedbank, Annual report 2005 Swedbank, Annual report 2007 Swedbank, Annual report 2008 Swedbank, Annual report 2009 Swedbank, Annual report 2010 Swedbank, Annual report 2012 Swedbank, Annual report 2013 Swedbank, Annual report 2014 Swedbank, press-release, 2007, February 7 Swedbank, press-release, 2007, July 9 Swedbank, press-release, 2008, October 7 Swedbank, press-release, 2009, April 29 Swedbank, press-release, 2013, April 1 Swedbank AB. Retrieved from: http://www.swedbank.com: Swedbank, Annual report 2005 Swedbank, press-release, 2007, February 7 Swedbank, press-release, 2007, July 9 Swedbank, press-release, 2009, April 29 Swedbank, press-release, 2013, April 1 95 The Clute Institute The Clute Institute The Clute Institute Journal of Business Case Studies – Second Quarter 2016 Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 Tidningarnas Telegrambyrå (2007, February 7). Swedbank köper ukrainska TAS. Dagens industri. Retrieved from: http://www.di.se/artiklar/2007/2/7/swedbank-koper-ukrainska-tas/ (English translation: Swedbank buys Ukrainian TAS.) Tigipko, S. (2008, March 5). Presentation during Capital market days. Retrieved from: https://www.swedbank.com/idc/groups/public/@i/@sbg/@gs/@ir/documents/presentation/cid Tigipko, S. (2008, March 5). Presentation during Capital market days. Retrieved from: https://www.swedbank.com/idc/groups/public/@i/@sbg/@gs/@ir/documents/presentation/cid_007052.pdf https://www.swedbank.com/idc/groups/public/@i/@sbg/@gs/@ir/documents/presentation/cid_007052.pdf USAID (United States Agency International Development) (2010) A Review and Assessment of the Ukrainian Financial Sector USAID (United States Agency International Development) (2010). A Review and Assessment of the Ukrai Retrieved from: https://www.usaid.gov. d States Agency International Development) (2010). A Review and Assessment of the Ukrainian Financial Sec eved from: https://www.usaid.gov. Vavryshchuk, V. (2010, June 15). Slow and Steady Wins the Race. Ukrainian Banks Initiating Coverage, BG Capital Research. Retrieved from: http://bgcapital.ge/img/st_img/researches/2010/BGC%20Ukrainian%20Banks%20Initiating%20Coverage%20Jun15% 202010.pdf capital.ge/img/st_img/researches/2010/BGC%20Ukrainian%20Banks%20Initiating%20Coverage%20Jun15% df p Waltersson, L. (2009, October 26). Swedbanks presidentkandidat: Vi måste koncentrera makten. Nyhetsportalen. Retrieved from: http://www.lasarnasfria.se/artikel/109231 (English translation: Swedbank’s candidate for President: we must The Clute Institute DISCUSSION QUESTIONS 1) Why transition economies became attractive for foreign investors after the fall of the Soviet Union? What were the main risks associated with FDI into transition economies? 1) Why transition economies became attractive for foreign investors after the fall of the Soviet Union? What were the main risks associated with FDI into transition economies? 2) Describe the investment environment of the host country (Sweden) and recipient country (Ukraine) in 2007, when the investment was made. 3) What were the main reasons for Swedbank to make FDI in Ukraine? Summarize the strategy of Swedbank in Ukraine. 4) What possible measures, in your view, could be done additionally to enhance governance and risk control of the subsidiary in Ukraine? 5) Describe the impact of financial crises on the Ukrainian market. 6) Describe and analyze exit strategy of Swedbank from the Ukrainian market. What additional measures can you suggest to improve the recovery of the loan portfolio and decrease losses? 7) What was the impact of this particular investment in Ukraine on the whole Swedbank Group? 8) What measures undertaken by the Swedish state, Swedbank’s shareholders and management helped Swedbank to survive? 9) What type of measures have been suggested and undertaken by regulating authorities in order to improve financial stability in the society after financial crises 2008? Choose a country and elaborate the issue with application to the country of your choice. pp y y 10) What measures could improve the image of Ukraine for the foreign investors? 11) Find an example of successful and un-successful foreign investments abroad. Why some companies became successful while others failed? 12) How important is to view particular investment cases in the relevant institutional, economic, legal, political, social and cultural environments? Explain, based upon materials from the Case Study and from your own example. 13) Companies often face a dilemma: either to insure growth by accepting higher risks or to prioritize the lower risks at the expense of growth. What are the pros and cons of both strategies? Try to formulate your own recommendations on what risks companies, searching for growth, could accept and what risks they should strictly avoid. 14) Some researchers observe that certain opportunities are rooted outside the firm’s relationships and networks. Opportunity creation is instead an outcome of discovery, luck and serendipity. Analyze this hypothesis in relationship with the presented Case Study and your own example. Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 FURTHER READINGS AVAILABLE THROUGH INTERNET Deloitte (2011). Banking and insurance in Ukraine. Industry overview. Report. Deloitte (2011). Banking and insurance in Ukraine. Industry overview. Report. http://www investin if ua/doc/pub/Overview Banking-and-insurance W http://www.investin.if.ua/doc/pub/Overview_Banking-and-insurance_WWW.pdf http://www.investin.if.ua/doc/pub/Overview_Bank Government of Ukraine portal (2014, August 11). Foreign banks in Ukraine. Expectations an http://about-ukraine.net/foreign-banks-in-ukraine-expectations-and-reality.html Government of Ukraine portal (2014, August 11). Foreign banks in Ukraine. Expectations and reality. Retrieved from: http://about-ukraine.net/foreign-banks-in-ukraine-expectations-and-reality.html Invest in Ukraine. Retrieved from: http://www.investukraine.net/ OECD (2011). Ukraine. Investment Policy Reviews. Retrieved from: http://www.oecd-ilibrary.org/finance-and investment policy reviews ukraine 2011 9789264113503 en OECD (2011). Ukraine. Investment Policy Reviews. Retrieved from: http://www.oecd-ilibrary.org/finance-and-investment/oecd- investment-policy-reviews-ukraine-2011_9789264113503-en p y _ PWC (PriceWaterhouseCoopers) (2013). Doing business and investing in Ukraine. Report. Re https://www.pwc.com/ua/en/survey/2013/assets/ukraine_doingbusiness_2013.pdf PWC (PriceWaterhouseCoopers) (2013). Doing business and investing in Ukraine. Report. R https://www.pwc.com/ua/en/survey/2013/assets/ukraine_doingbusiness_2013.pdf State Statistics Service of Ukraine http://www.ukrstat.gov.ua/ U.S. Department of State (2013). Investment Climate Statement – Ukraine. Bureau of Economic and Business Affairs, Report, February 2013. Retrieved from: http://www.state.gov/e/eb/rls/othr/ics/2013/204754.htm U.S. Department of State (2013). Investment Climate Statement – Ukraine. Bureau of Economic and Business Affairs, Report, February 2013. Retrieved from: http://www.state.gov/e/eb/rls/othr/ics/2013/204754.htm Copyright by author(s); CC-BY Copyright by author(s); CC-BY 96 The Clute Institute The Clute Institute 96 Journal of Business Case Studies – Second Quarter 2016 Journal of Business Case Studies – Second Quarter 2016 Volume 12, Number 2 ENDNOTES Translations of currencies in the Case are done according to following exchange rates: Translations of currencies in the Case are done according to following exchange rates: 1 SEK=0.129 USD (2014); 1 SEK=0.152 USD (2013); 1SEK=0.153 USD (2012); 1 SEK=0.147 (2010); 1 SEK=0.137 (2009); 1SEK=0.128 USD (2008); 1SEK=0.149 USD (2007); 1 SEK = 0.146 USD (2006); 1 USD=0.79 EUR (2006); 1 EUR=1.41 USD (2008). ( ); ( ) As a reference, one Swedish crooner (SEK) is approximately 0.118951 USD as of 2016, February 4. 2 See Corruption Perceptions Index http://www.transparency.org/cpi2015 DISCUSSION QUESTIONS 15) Several authors who studied the process of decision-making suggested that changes in the turbulent and uncertain environment are bottom-up driven, emergent, and incremental, rather than formulated and implemented as assumed in the traditional strategic theory. Analyze bottom-up decision-making versus top down based upon empirical data from presented Case Study and your own example. 15) Several authors who studied the process of decision-making suggested that changes in the turbulent and uncertain environment are bottom-up driven, emergent, and incremental, rather than formulated and implemented as assumed in the traditional strategic theory. Analyze bottom-up decision-making versus top down based upon empirical data from presented Case Study and your own example. 15) Several authors who studied the process of decision-making suggested that changes in the turbulent and uncertain environment are bottom-up driven, emergent, and incremental, rather than formulated and implemented as assumed in the traditional strategic theory. Analyze bottom-up decision-making versus top down based upon empirical data from presented Case Study and your own example. Copyright by author(s); CC-BY Copyright by author(s); CC-BY 97 The Clute Institute The Clute Institute Volume 12, Number 2 Journal of Business Case Studies – Second Quarter 2016 Copyright by author(s); CC-BY 2 See Corruption Perceptions Index http://www.transparency.org/cpi2015 3 See http://www.nasdaqomxnordic.com/ The Clute Institute 3 See http://www.nasdaqomxnordic.com/ 4 Business Sweden was founded on the first of January, 2013, by a merger of the Swedish Trade Council (Exportrådet) and Invest Sweden. Business Sweden is owned by the Swedish Government and the industry, a partnership that provides access to contacts and networks at all levels. Business Sweden has offices in nearly 50 of the world's markets. http://www.business-sweden.se/ 5 Risk-weighted asset is a bank's assets, weighted according to risk. Risk weights are used in determining the capital requirement or Capital Adequacy Ratio (CAR) for a financial institution. Copyright by author(s); CC-BY Copyright by author(s); CC-BY 98 The Clute Institute
https://openalex.org/W2790282567	https://europepmc.org/articles/pmc5861146?pdf=render	English	null	Whole-Inactivated Influenza Virus Is a Potent Adjuvant for Influenza Peptides Containing CD8+ T Cell Epitopes	Frontiers in immunology	2,018	cc-by	8,897	Abbreviations: WIV, whole-inactivated influenza virus; CTLs, cytotoxic CD8+ T cells; APCs, antigen-presenting cells; PAMPs, pathogen-associated molecular patterns; IFA, incomplete Freund’s adjuvant; CPLs, chemically altered peptide ligands; DCs, dendritic cells; GIL, GILGFVFTL (M158–66) peptide; FMY, FMYSDFHFI (PA46–54) peptide; NML, NMLSTVLGV (PB1413–421) peptide; SFU, spot-forming units. Whole-inactivated influenza Virus is a Potent adjuvant for influenza Peptides containing cD8+ T cell epitopes Peter C. Soema1,2‡, Sietske K. Rosendahl Huber3‡, Geert-Jan Willems1, Ronald Jacobi3, Marion Hendriks3, Ernst Soethout1†, Wim Jiskoot2, Jørgen de Jonge3, Josine van Beek3, Gideon F. A. Kersten1,2 and Jean-Pierre Amorij1 We furthermore show that WIV may adjuvate T cell responses to a variety of peptides, using pools of either conserved wild-type influenza peptides or chemically altered peptide ligands. This study shows the potential of WIV as an adjuvant for influenza peptides. The simple formulation process and the solid safety record of WIV make this an attractive adjuvant for T cell peptides, and may also be used for non-influenza antigens. Reviewed by: Michael Schotsaert, Icahn School of Medicine at Mount Sinai, United States Raffael Nachbagauer, Icahn School of Medicine at Mount Sinai, United States Correspondence: Peter C. Soema peter.soema@intravacc.nl †Present address: Ernst Soethout, Virtuvax, Odijk, Netherlands ‡These authors have contributed equally to this work. ‡These authors have contributed equally to this work. Specialty section: This article was submitted to Vaccines and Molecular Therapeutics, a section of the journal Frontiers in Immunology Received: 09 January 2018 Accepted: 28 February 2018 Published: 14 March 2018 Specialty section: This article was submitted to Vaccines and Molecular Therapeutics, a section of the journal Frontiers in Immunology Keywords: whole-inactivated influenza virus, adjuvant, T cell peptide, CTL, design of experiments, formulation, peptide vaccine Received: 09 January 2018 Accepted: 28 February 2018 Published: 14 March 2018 Keywords: whole-inactivated influenza virus, adjuvant, T cell peptide, CTL, design of experiments, formulation, peptide vaccine Original Research published: 14 March 2018 doi: 10.3389/fimmu.2018.00525 1 Intravacc (Institute for Translational Vaccinology), Bilthoven, Netherlands, 2 Division of Drug Delivery Technology, Cluster BioTherapeutics, Leiden Academic Centre for Drug Research (LACDR), Leiden University, Leiden, Netherlands, 3 Centre for Infectious Disease Control Netherlands, National Institute for Public Health and the Environment (RIVM), Bilthoven, Netherlands Edited by: Urszula Krzych, Walter Reed Army Institute of Research, United States Influenza peptide antigens coding for conserved T cell epitopes have the capacity to induce cross-protective influenza-specific immunity. Short peptide antigens used as a vaccine, however, often show poor immunogenicity. In this study, we demonstrate that whole-inactivated influenza virus (WIV) acts as an adjuvant for influenza peptide antigens, as shown by the induction of peptide-specific CD8+ T cells in HLA-A2.1 transgenic mice upon vaccination with the influenza-M1-derived GILGFVFTL peptide (GIL), formulated with WIV. By screening various concentrations of GIL and WIV, we found that both components contributed to the GIL-specific T cell response. Whereas co-localization of the peptide antigen and WIV adjuvant was found to be important, neither physical association between peptide and WIV nor fusogenic activity of WIV were relevant for the adjuvant effect of WIV. We furthermore show that WIV may adjuvate T cell responses to a variety of peptides, using pools of either conserved wild-type influenza peptides or chemically altered peptide ligands. This study shows the potential of WIV as an adjuvant for influenza peptides. The simple formulation process and the solid safety record of WIV make this an attractive adjuvant for T cell peptides, and may also be used for non-influenza antigens. Influenza peptide antigens coding for conserved T cell epitopes have the capacity to induce cross-protective influenza-specific immunity. Short peptide antigens used as a vaccine, however, often show poor immunogenicity. In this study, we demonstrate that whole-inactivated influenza virus (WIV) acts as an adjuvant for influenza peptide antigens, as shown by the induction of peptide-specific CD8+ T cells in HLA-A2.1 transgenic mice upon vaccination with the influenza-M1-derived GILGFVFTL peptide (GIL), formulated with WIV. By screening various concentrations of GIL and WIV, we found that both components contributed to the GIL-specific T cell response. Whereas co-localization of the peptide antigen and WIV adjuvant was found to be important, neither physical association between peptide and WIV nor fusogenic activity of WIV were relevant for the adjuvant effect of WIV. Formulation of Vaccinesl Influenza A/PR/8/34 (H1N1) virus was propagated on fertilized eggs and inactivated with β-propiolactone as described before (12, 13), which yielded PR8 WIV bulk vaccine. For certain experiments, WIV was subsequently fusion-inactivated by lower- ing the buffer pH to 4.5 for 15 min at 37°C with a pre-titrated volume of 1 M HCl (Sigma-Aldrich) and was then brought back to physiological pH by dialyzing overnight against PBS pH 7.2 (Life Technologies). Membrane fusion capacity was subsequently determined by a hemolysis assay as described previously (14).h Short linear peptides that represent conserved T cell epitopes can be used as antigens to induce influenza virus-specific CTLs (6). However, peptides are generally not that immunogenic, as they are inefficiently delivered to antigen-presenting cells (APCs), lack pathogen-associated molecular patterns (PAMPs) to trigger the immune system, and are rapidly degraded. Inefficient delivery of an antigenic peptide and lack of PAMPs can be overcome by formulation of the peptide with an appropriate adjuvant (7). The Netherlands Cancer Institute kindly provided the HLA-A2.1-restricted influenza GILGFVFTL (GIL, M158–66), FMYSDFHFI (FMY, PA46–54), and NMLSTVLGV (NML, PB1413–421) peptides, and chemically altered CPLs [am-phg]ILGFVFTL (G1), [4-FPHE]MYSDFHF[2-AOC] (F5), and N[NLE]LSTVLGV (N53), as published by Rosendahl Huber et al. (11) (Figure S1 in Supplementary Material). Water-in-oil emulsions such as incomplete Freund’s adjuvant (IFA) are effective adjuvants for peptides, but are associated with substantial adverse events such as lesion formation at the site of injection, making them undesirable for use in humans for vac- cination (6). Thus, alternative adjuvants for peptide antigens are highly sought after.l Influenza PR8 WIV and peptide antigens were formulated in PBS pH 7.2 at various concentrations (concentration of WIV is total protein content). Where mentioned, 50 µg of CpG ODN 1826 (Invivogen) or 50% (v/v) IFA (Sigma-Aldrich) were added to the formulation. Whole-inactivated influenza virus (WIV) possesses an innate adjuvant capability in the form of viral single-stranded RNA (ssRNA), which is a potent TLR7 agonist (8). Furthermore, WIV contains CD4+ epitopes, which are favorable for the induction of functional CD8+ T cell and B cell responses (9). WIV is also more reactogenic than split and subunit influenza vaccines, which could be linked to its increased immunogenicity. A recent study by Babb et al. described that gamma-irradiated WIV can act as an adjuvant for a Semliki Forest virus (SFV) vaccine, which significantly increased the induction of SFV-specific antibodies in mice (10). Formulation of Vaccinesl However, it has not been studied whether WIV can be used as an adjuvant for T cell-inducing peptide-based vaccines. Citation: Current seasonal influenza vaccines exert their protective effect mainly through the induction of virus-specific neutralizing antibodies directed against the surface proteins of influenza virus (1). However, changes in the influenza virus surface proteins caused by antigenic drift or shift allow influenza viruses to evade these antibodies. Therefore, influenza vaccines require regular updates Soema PC, Rosendahl Huber SK, Willems G-J, Jacobi R, Hendriks M, Soethout E, Jiskoot W, de Jonge J, van Beek J, Kersten GFA and Amorij J-P (2018) Whole-Inactivated Influenza Virus Is a Potent Adjuvant for Influenza Peptides Containing CD8+ T Cell Epitopes. Front. Immunol. 9:525. doi: 10.3389/fimmu.2018.00525 March 2018 \| Volume 9 \| Article 525 Frontiers in Immunology \| www.frontiersin.org 1 Whole-Inactivated Influenza as an Influenza T Cell Adjuvant Soema et al. or need to be completely renewed, to be able to protect against circulating influenza viruses. membrane fusion activity on the adjuvant activity of WIV. Finally, we investigated the adjuvant activity of WIV on influenza pep- tides coding for conserved WT T cell epitopes and CPL variants. l By contrast, cellular immune components such as CD8+ and CD4+ T cells often recognize conserved epitopes of internal influenza virus proteins. This allows these T cells to cross-react with various influenza strains or even subtypes, and provides the host with a better protection against drifting or shifting influenza virus strains (2–4). Cytotoxic CD8+ T cells (CTLs) can clear virus-infected host cells, thereby controlling influenza virus infections by inhibiting viral replication and potentially limiting viral spread. The induction of influenza virus-specific T cells in addition to neutralizing antibodies may, therefore, increase the effectiveness of influenza vaccines. Currently, mul- tiple approaches for the induction of influenza-specific T cell responses by vaccination are being investigated (5, 6). Frontiers in Immunology \| www.frontiersin.org Hemolysis Assay anti-mouse CD4-PE (BD Biosciences), and Live-dead-Aqua (Invitrogen), fixated with fixation/permeabilization buffer (BD Biosciences), and washed with permeabilization wash buffer (BD Biosciences). Finally, cells were stained intracellularly with anti-mouse IFN-γ-APC (BD Biosciences), and IFN-γ+ CD8+ T cells were quantified on a FACS Canto II flow cytometer (BD Biosciences). Acquired data were analyzed with FlowJo version 10 for Mac OSX (TreeStar Inc.). Gating strategy for the quantification of CD8+ IFN-γ+ T cells is shown in Figure S2 in Supplementary Material. Virosome fusion activity was determined by using a hemolysis assay as described previously (16). Formulations were mixed with human blood erythrocytes and 0.1 M 2-(N-morpholino) ethanesulfonic acid (MES) buffer with pH ranging from 4.5 to 5.5, and incubated at 37°C for 30 min. The released hemoglobin was quantified in the supernatant after centrifugation by read- ing absorbance at 540 nm using a Synergy Mx reader (Biotek). Hemoglobin release from erythrocytes mixed with water was set as maximal hemolysis (100%). Determination of Association Between Peptides and WIVh The association of peptides to WIV particles was studied by quan- tification of unassociated peptide in a mixture of peptides and WIV. Peptides were mixed with WIV in similar concentrations as used in the animal studies. WIV particles were subsequently spun down by ultracentrifugation for 2 h at 30,000 × g. Supernatant was collected and analyzed for peptides by mass spectrometry on a nanoscale LC-MS system, essentially as described by Meiring et al. (15), comprising a 100- and 50-µm internal diameter in house-packed Reprosil-Pur C18-AQ trapping and analytical column, respectively. Peptides were trapped in 100% of solvent A (water + 5% dimethylsufoxide + 0.1% formic acid) for 10 min. The linear gradient for the separation ranged from 15 to 90% of solvent B (acetonitrile + 5% DMSO + 0.1% FA) in 25 min at a flow rate of 125 nL/min. The column effluent was directly electro-sprayed into the MS using a gold-coated fused silica- tapered tip of 3.5 µm internal diameter. High-resolution MS1 data were acquired on an LTQ-Orbitrap XL mass spectrometer (Thermo Scientific, San Jose, CA, USA) at a resolution of 60,000 FWHM. Peptide identity was confirmed by their CID MS/MS fragmentation spectra, acquired on-the-fly in the LTQ mass analyzer on the singly or doubly charged ions of the targeted peptides only. The percentage of unassociated peptide was cal- culated by comparing peptide content in supernatants of peptide mixed with WIV to peptide content in supernatants collected from solutions without WIV. Another method to increase the immunogenicity of peptides is to increase their binding affinity to MHC class I molecules by chemical modification. These so-called, chemically altered peptide ligands (CPLs) derived from HLA-A2.1-restricted epitopes were shown to possess a higher affinity to HLA-A2.1, and induced more IFN-γ positive splenocytes compared to wild- type (WT) epitopes in HLA-A2.1 transgenic mice (11). Chemical modification of subdominant peptide epitopes might, therefore, increase the breadth of peptide-specific T cell responses.if In the current study, first we investigated the effect of the addi- tion of WIV to the GILGFVFTL (GIL, M158–66) peptide, which is the dominant CD8+ T cell epitope in HLA-A2.1-restricted individuals. Next, we performed a dose-finding study to deter- mine the optimal WIV and peptide antigen concentrations for the induction of peptide-specific T cells. Determination of Association Between Peptides and WIVh Furthermore, we studied the effect of WIV-peptide co-localization, association, and WIV March 2018 \| Volume 9 \| Article 525 2 Whole-Inactivated Influenza as an Influenza T Cell Adjuvant Soema et al. Animal Studies Animal studies were conducted according to the guidelines provided by the Dutch Animal Protection Act and were approved by the Committee for Animal Experimentation (DEC) of the PD-Alt campus (Bilthoven, The Netherlands) under protocol number PO201400134. Eight- to ten-week- old female transgenic C57BL/6-Tg(HLA-A2.1)1Enge/J mice (Jackson Laboratory, maintained in-house) were used in all studies. Peptide-specific, IFN-γ-producing T cells were determined in splenocytes by an IFN-γ ELISpot. 96-wells Multiscreen PVDF filter plates (Millipore) were activated by adding 25 µL 70% ethanol for 2 min and washed three times with PBS. Plates were coated overnight with anti-mouse IFN-γ antibodies (U-Cytech) at 4°C, washed three times, and blocked with 5% FCS (Hyclone, Thermo Scientific) in RPMI medium for 1 h at 37°C. Subsequently, 4×105 isolated splenocytes resuspended in RPMI medium, 5% FCS was added to each well with or without 100 ng of relevant peptide, and incubated overnight at 37°C, 5% CO2. Next, filter plates were washed five times and IFN-γ was detected using biotinylated anti-mouse antibodies (U-Cytech) and 100 µL BCIP/NBT reagent (Thermo Scientific) per well. Plates were washed with tap water and dried. Spots were counted using an A.EL.VIS ELISpot reader (Aelvis). The number of IFN-γ-producing cells in antigen-stimulated splenocytes was counted and corrected for background by subtracting the number of spots produced by splenocytes incubated with medium only. In the proof-of-principle study, mice (three per group) received immunizations subcutaneously (s.c.) in alternating flanks at days 0 and 21, containing either PBS, 50 µg WIV, 1 nmol GIL peptide adjuvanted with 50 µg WIV or 100 nmol GIL adjuvanted with 50 µg CpG in a volume of 100 µL.i For the dose-finding study, formulations consisting of various doses of WIV and GIL peptide were administered s.c. in alternat- ing flanks of mice (six per group) at days 0 and 21.f To study the effect of adjuvant co-localization, mice (six per group) were immunized at days 0 and 21 either s.c. in one flank with PBS or 100 nmol GIL peptide adjuvanted with 25 µg WIV, or s.c. in separate flanks with 100 nmol GIL peptide in one flank and 25 µg WIV adjuvant in the opposite flank.hf Statistics Results were statistically analyzed with a one-way ANOVA followed by a Tukey post test for multiple comparisons. All statistical analyzes were performed using GraphPad Prism 6.04 for Windows (GraphPad Software Inc.). The effect of membrane fusion activity was assessed by immu- nizing mice (six per group) s.c. in alternating flanks at days 0 and 21 with 100 nmol GIL peptide adjuvanted with either 25 µg of fusion-active WIV or fusion-inactive WIV.hf The adjuvant effect of WIV on a mix of multiple peptides was assessed with either a WT peptide pool (GIL, FMY, and NML; 100 nmol each) or a modified peptide pool (G1, F5, and N53; 100 nmol each). Mice (six per group) received an s.c. immuni- zation in the flank at days 0 and 21 containing either PBS, WT peptide pool adjuvanted with 5 µg WIV or IFA, CPL peptide pool adjuvanted with 5 µg WIV or IFA, or only 5 µg WIV. In all studies, animals were sacrificed at day 35. Addition of WIV to GIL Peptide Enhances GIL-Specific T Cell Responses (A) Splenocytes were re-stimulated with GIL peptide and stained for intracellular cytokine IFN-γ. The percentages or IFN-γ+ CD8+ T cells acquired by FACS are plotted. Data are shown as mean ± SD of three mice per group, each circle is a single replicate and data are from a single experiment representative of two individual experiments. p < 0.01, p < 0.001 (one-way ANOVA). (B) Representative FACS plot displaying GIL-specificity of IFN-γ+ CD8+ T cells in splenocytes from mice immunized with GIL peptide formulated with WIV and re-stimulated with GIL peptide. Figure 1 \| Comparison of CD8+ T cell responses induced by 1 nmol GIL peptide adjuvanted with either 50 µg CpG or 50 µg whole-inactivated influenza virus (WIV). HLA-A2.1 transgenic mice were injected twice, 3 weeks apart with PBS (negative control), WIV or GIL peptide with indicated adjuvant. Shown are the responses 2 weeks after the final immunization. (A) Splenocytes were re-stimulated with GIL peptide and stained for intracellular cytokine IFN-γ. The percentages for IFN-γ+ CD8+ T cells acquired by FACS are plotted. Data are shown as mean ± SD of three mice per group, each circle is a single replicate and data are from a single experiment representative of two individual experiments. p < 0.01, p < 0.001 (one-way ANOVA). (B) Representative FACS plot displaying GIL-specificity of IFN-γ+ CD8+ T cells in splenocytes from mice immunized with GIL peptide formulated with WIV and re-stimulated with GIL peptide. Figure 1 \| Comparison of CD8+ T cell responses induced by 1 nmol GIL peptide adjuvanted with either 50 µg CpG or 50 µg whole-inactivated influenza virus (WIV). HLA-A2.1 transgenic mice were injected twice, 3 weeks apart with PBS (negative control), WIV or GIL peptide with indicated adjuvant. Shown are the responses 2 weeks after the final immunization. (A) Splenocytes were re-stimulated with GIL peptide and stained for intracellular cytokine IFN-γ. The percentages for IFN-γ+ CD8+ T cells acquired by FACS are plotted. Data are shown as mean ± SD of three mice per group, each circle is a single replicate and data are from a single experiment representative of two individual experiments. p < 0.01, **p < 0.001 (one-way ANOVA). (B) Representative FACS plot displaying GIL-specificity of IFN-γ+ CD8+ T cells in splenocytes from mice immunized with GIL peptide formulated with WIV and re-stimulated with GIL peptide. Both GIL and WIV Contribute to the Induction of GIL-Specific T Cell Responses contains approximately 535 pmol of GIL peptide; a negligible amount compared to the 1–100 nmol GIL peptide. These findings imply that the addition of WIV provides a synergistic, adjuvant- like effect for the GIL peptide. To investigate the relative contributions of WIV on the induc- tion of GIL-specific T cell responses, different combinations of amounts of peptide and WIV were tested in a mouse model by determining IFN-γ responses with FACS and ELISpot analyses. The peptide antigen dose ranged from 1 to 100 nmol, and the WIV dose ranged from 1 to 25 µg. Addition of WIV to GIL Peptide Enhances GIL-Specific T Cell Responses Figure 2 \| Induction of GIL-specific T cell responses by different concentrations of GIL and whole-inactivated influenza virus (WIV). HLA-A2.1 transgenic mice were vaccinated twice, 3 weeks apart with indicated amounts of GIL peptide and WIV. (A) Splenocytes were re-stimulated with GIL peptide and stained for intracellular cytokine IFN-γ. The percentages for IFN-γ+ CD8+ T cells acquired by FACS are plotted. (B) Splenocytes of immunized mice were re-stimulated with GIL peptide. The number of CD8+ IFN-γ-secreting cells was subsequently determined with ELISpot. (A,B) Data are shown as mean ± SD of three mice per group, each dot is a single replicate and data are from a single experiment. p < 0.05, *p < 0.01 (one-way ANOVA). Figure 2 \| Induction of GIL-specific T cell responses by different concentrations of GIL and whole-inactivated influenza virus (WIV). HLA-A2.1 transgenic mice were vaccinated twice, 3 weeks apart with indicated amounts of GIL peptide and WIV. (A) Splenocytes were re-stimulated with GIL peptide and stained for intracellular cytokine IFN-γ. The percentages for IFN-γ+ CD8+ T cells acquired by FACS are plotted. (B) Splenocytes of immunized mice were re-stimulated with GIL peptide. The number of CD8+ IFN-γ-secreting cells was subsequently determined with ELISpot. (A,B) Data are shown as mean ± SD of three mice per group, each dot is a single replicate and data are from a single experiment. p < 0.05, *p < 0.01 (one-way ANOVA). Addition of WIV to GIL Peptide Enhances GIL-Specific T Cell Responses As shown previously, 100 nmol GIL peptide, without adjuvant, was unable to induce detectable GIL-specific T cell responses in HLA-A2.1 transgenic mice (14). The addition of the adju- vant CpG, a TLR9 agonist, was also insufficient to induce GIL-specific CD8+ T cells (Figure 1). By contrast, the addition of WIV to the GIL peptide resulted in the induction of high levels of GIL-specific CD8+ T cell responses. Importantly, WIV alone was able to induce low levels of GIL-specific CD8+ T cells, which is caused by processing of the GIL epitope present in the internal M1 protein of WIV. The combination of GIL and WIV, however, induced high levels of GIL-specific T cells as compared to WIV only. Intracellular Staining and Flow Cytometry T cell populations were analyzed by flow cytometry. In short, single-cell suspensions of splenocytes were plated at a concen- tration of 2×106 cells in a 48-well plate in RPMI medium (Life Technologies) with 10% Hyclone fetal calf serum (FCS, Thermo Scientific), and stimulated overnight with either medium, 50 ng peptide or 50 ng PR8 WIV. Cytokine transport was blocked by incubating with Golgi-plug (BD Biosciences) for the last 4 h. Cells were stained with anti-mouse CD8-FITC (BD Biosciences), March 2018 \| Volume 9 \| Article 525 Frontiers in Immunology \| www.frontiersin.org 3 Whole-Inactivated Influenza as an Influenza T Cell Adjuvant Soema et al. Figure 2 \| Induction of GIL-specific T cell responses by different concentrations of GIL and whole-inactivated influenza virus (WIV). HLA-A2.1 transgenic mice were vaccinated twice, 3 weeks apart with indicated amounts of GIL peptide and WIV. (A) Splenocytes were re-stimulated with GIL peptide and stained for intracellular cytokine IFN-γ. The percentages for IFN-γ+ CD8+ T cells acquired by FACS are plotted. (B) Splenocytes of immunized mice were re-stimulated with GIL peptide. The number of CD8+ IFN-γ-secreting cells was subsequently determined with ELISpot. (A,B) Data are shown as mean ± SD of three mice per group, each dot s a single replicate and data are from a single experiment. p < 0.05, p < 0.01 (one-way ANOVA). Figure 1 \| Comparison of CD8+ T cell responses induced by 1 nmol GIL peptide adjuvanted with either 50 µg CpG or 50 µg whole-inactivated influenza virus WIV). HLA-A2.1 transgenic mice were injected twice, 3 weeks apart with PBS (negative control), WIV or GIL peptide with indicated adjuvant. Shown are the esponses 2 weeks after the final immunization. Frontiers in Immunology \| www.frontiersin.org Co-Localization of GIL and WIV Is Necessary to Obtain the Adjuvant Effect of WIVh The previous results show that, aside from providing additional GIL-epitopes, WIV acts as an adjuvant when combined with GIL. Next, the influence of co-localization of GIL (100 nmol) and WIV (25 µg) was investigated. When GIL and WIV were injected separately at two different sites (Figure 3A), only mod- erate GIL-specific T cell responses were obtained. By contrast, when injected at the same site, a significant induction of T cell responses was observed (Figures 3B,C), proving that the T cell response is not simply a sum of the amount of GIL present in the formulations. Thus, co-localization is necessary to obtain the adjuvant effect of WIV for the GIL peptide. A combination of 1 nmol GIL peptide and 1 µg WIV induced very low responses (Figures 2A,B). Increasing the GIL dose to 100 nmol led to a significant induction of GIL-specific CD8+ T cells. This effect was also observed when the peptide dose was increased from 1 to 100 nmol combined with a dose of 25 µg WIV. The increase of WIV from 1 to 25 µg also had a marked effect on the induction T cells, both with 1 and 100 nmol GIL. These data show that GIL and WIV induce the highest T cell responses when combined, whereas GIL or WIV alone induce considerably less T cells. It should be noted that it is estimated that 25 µg WIV March 2018 \| Volume 9 \| Article 525 Frontiers in Immunology \| www.frontiersin.org 4 Whole-Inactivated Influenza as an Influenza T Cell Adjuvant Soema et al. Table 1 \| Association between GIL peptide and whole-inactivated influenza virus (WIV). GIL peptide (μg) WIV (μg) Unassociated GIL peptide (%) 1 1 112 ± 10 100 1 111 ± 6 50 13 96 ± 5 1 25 77 ± 8 100 25 92 ± 20 1 50 87 ± 9 Indicated amounts of GIL peptide and WIV were mixed and subsequently separated by ultracentrifugation. The residual, unassociated peptide in the supernatant was quantified by using mass spectrometry. Data are shown as mean ± SD from three individual experiments. Table 1 \| Association between GIL peptide and whole-inactivated influenza virus (WIV). Figure 3 \| Influence of co-localization on the immunogenicity of GIL (100 nmol) and whole-inactivated influenza virus (WIV, 25 µg). The Adjuvant Effect of WIV Is Not Caused by Physical GIL Peptide–WIV Interaction For some adjuvants, direct physical association (such as adsorp- tion) between adjuvant and antigen is required for optimal antigen delivery to APCs (14, 17). The adjuvant effect observed by WIV might also rely on physical association between WIV and GIL peptide, since co-localization was shown to be required. The association between GIL and WIV was, therefore, determined by using mass spectrometry after separation of non-associated peptide from WIV by centrifugation (Table 1). Overall, little to no association was found between GIL peptide and WIV at various concentrations. Some association was observed with higher concentrations of WIV (25–50 µg), but not at 1 µg WIV. However, no direct correlation can be seen between peptide–WIV association and the immunogenicity of the mixtures, indicating that physical interaction between GIL and WIV is not the predominant mechanism behind the adjuvant activity of WIV. Figure 4 \| Effect of membrane fusion activity on whole-inactivated influenza virus (WIV) adjuvant activity. (A) WIV was fusion-inactivated by brief exposure to a buffer with pH 4.5 at 37°C. Hemolysis was performed by mixing either WIV (“active” WIV) or fusion-inactivated WIV (“inactive” WIV) with human blood erythrocytes at various pHs. Fusion-mediated hemoglobin release was subsequently determined by spectrophotometric analysis. Data are shown as mean ± SD of three individual experiments. (B) HLA-A2.1 transgenic mice were vaccinated twice, 3 weeks apart with GIL peptide (100 µg) and either fusion-active or fusion-inactive WIV (25 µg). Splenocytes were re-stimulated with GIL peptide and stained for intracellular cytokine IFN-γ. The percentages for IFN-γ+ CD8+ T cells acquired by FACS are plotted. (C) Splenocytes were re-stimulated with GIL peptide. The number of IFN-γ-secreting cells was subsequently determined with ELISpot. (B,C) Data are shown as mean ± SD of six mice per group, each dot is a single replicate and data are from a single experiment. n.s. = not significant (one-way ANOVA). responses by WIV (18). Therefore, we investigated the influence of the membrane fusion activity of WIV on the immunogenicity of the GIL and WIV combination. WIV was fusion inactivated by exposure to acidic pH, and loss of membrane fusion activity was confirmed by a hemolysis assay (Figure 4A). A mixture of GIL peptide and fusion-inactivated WIV was able to induce Membrane Fusion Activity Is Not Required for WIV Adjuvant Activity Co-Localization of GIL and WIV Is Necessary to Obtain the Adjuvant Effect of WIVh (A) HLA-A2.1 transgenic mice were vaccinated twice, 3 weeks apart with GIL peptide and WIV either combined in one single flank (mixed) or separate in opposite flanks (separate). (B) Splenocytes were re-stimulated with GIL peptide and stained for intracellular cytokine IFN-γ. The percentages for IFN-γ+ CD8+ T cells acquired by FACS are plotted. (C) Splenocytes were re-stimulated with GIL peptide. The number of IFN-γ-secreting cells was subsequently determined with ELISpot. (A,B) Data are shown as mean ± SD of six mice per group, each dot is a single replicate and data are from a single experiment. p < 0.01, *p < 0.001 (one-way ANOVA). Indicated amounts of GIL peptide and WIV were mixed and subsequently separated by ultracentrifugation. The residual, unassociated peptide in the supernatant was quantified by using mass spectrometry. Data are shown as mean ± SD from three individual experiments. Figure 4 \| Effect of membrane fusion activity on whole-inactivated influenza virus (WIV) adjuvant activity. (A) WIV was fusion-inactivated by brief exposure to a buffer with pH 4.5 at 37°C. Hemolysis was performed by mixing either WIV (“active” WIV) or fusion-inactivated WIV (“inactive” WIV) with human blood erythrocytes at various pHs. Fusion-mediated hemoglobin release was subsequently determined by spectrophotometric analysis. Data are shown as mean ± SD of three individual experiments. (B) HLA-A2.1 transgenic mice were vaccinated twice, 3 weeks apart with GIL peptide (100 µg) and either fusion-active or fusion-inactive WIV (25 µg). Splenocytes were re-stimulated with GIL peptide and stained for intracellular cytokine IFN-γ. The percentages for IFN-γ+ CD8+ T cells acquired by FACS are plotted. (C) Splenocytes were re-stimulated with GIL peptide. The number of IFN-γ-secreting cells was subsequently determined with ELISpot. (B,C) Data are shown as mean ± SD of six mice per group, each dot is a single replicate and data are from a single experiment. n.s. = not significant (one-way ANOVA). Figure 3 \| Influence of co-localization on the immunogenicity of GIL Figure 3 \| Influence of co-localization on the immunogenicity of GIL (100 nmol) and whole-inactivated influenza virus (WIV, 25 µg). (A) HLA-A2.1 transgenic mice were vaccinated twice, 3 weeks apart with GIL peptide and WIV either combined in one single flank (mixed) or separate in opposite flanks (separate). (B) Splenocytes were re-stimulated with GIL peptide and stained for intracellular cytokine IFN-γ. The percentages for IFN-γ+ CD8+ T cells acquired by FACS are plotted. (C) Splenocytes were re-stimulated with GIL peptide. Co-Localization of GIL and WIV Is Necessary to Obtain the Adjuvant Effect of WIVh The number of IFN-γ-secreting cells was subsequently determined with ELISpot. (A,B) Data are shown as mean ± SD of six mice per group, each dot is a single replicate and data are from a single experiment. p < 0.01, **p < 0.001 (one-way ANOVA). Membrane Fusion Activity Is Not Required for WIV Adjuvant Activity Budimir et al. have shown that the membrane fusion activity of WIV was necessary for the induction of cross-reactive T cell March 2018 \| Volume 9 \| Article 525 Frontiers in Immunology \| www.frontiersin.org 5 Whole-Inactivated Influenza as an Influenza T Cell Adjuvant Soema et al. peptide-specific T cell responses similar to those induced by GIL peptide with fusion-active WIV (Figures 4B,C), indicating that fusion activity is not critical for the adjuvant activity of WIV. Figure 5 \| T cell responses against wild-type and modified peptides adjuvanted with whole-inactivated influenza virus (WIV). HLA-A2.1 transgenic mice were vaccinated twice, 3 weeks apart with peptide pools containing either wild-type (WT) peptides (GIL, FMY, and NML, 100 nmol each) or modified (mod.) peptides (G1, F5, and N53, 100 nmol each) adjuvanted with either WIV or 50% (v/v) incomplete Freund’s adjuvant (IFA). (A) Splenocytes of immunized mice were re-stimulated with either GIL or G1 peptide. (B) Splenocytes were re-stimulated with either FMY or F5 peptide. (C) Splenocytes were re-stimulated with either NML or N53 peptide. (A–C) The number of IFN-γ-secreting cells was subsequently determined with ELISpot. Data are shown as mean ± SD of six mice per group, each dot is a single replicate and data are from a single experiment. p < 0.05, p < 0.01, p < 0.001 (one-way ANOVA). WIV Can Act as an Adjuvant for Multiple Combined Peptidesfi (A) Splenocytes of immunized mice were re-stimulated with either GIL or G1 peptide. (B) Splenocytes were re-stimulated with either FMY or F5 peptide. (C) Splenocytes were re-stimulated with either NML or N53 peptide. (A–C) The number of IFN-γ-secreting cells was subsequently determined with ELISpot. Data are shown as mean ± SD of six mice per group, each dot is a single replicate and data are from a single experiment. p < 0.05, p < 0.01, p < 0.001 (one-way ANOVA). Figure 5 \| T cell responses against wild-type and modified peptides adjuvanted with whole-inactivated influenza virus (WIV). HLA-A2.1 transgenic mice were vaccinated twice, 3 weeks apart with peptide pools containing either wild-type (WT) peptides (GIL, FMY, and NML, 100 nmol each) or modified (mod.) peptides (G1, F5, and N53, 100 nmol each) adjuvanted with either WIV or 50% (v/v) incomplete Freund’s adjuvant (IFA). (A) Splenocytes of immunized mice were re-stimulated with either GIL or G1 peptide. (B) Splenocytes were re-stimulated with either FMY or F5 peptide. (C) Splenocytes were re-stimulated with either NML or N53 peptide. (A–C) The number of IFN-γ-secreting cells was subsequently determined with ELISpot. Data are shown as mean ± SD of six mice per group, each dot is a single replicate and data are from a single experiment. p < 0.05, p < 0.01, p < 0.001 (one-way ANOVA). Finally, the NML peptide and the modified N53 were unable to induce any significant T cell responses, regardless of adjuvant (Figure 5C). IFA-adjuvanted peptides showed incidental T cell responses in some animals, but these did not significantly differ from the WIV groups.hi In conclusion, two distinct effects were observed. First, we observed that WIV has an adjuvating effect on several peptides, including GIL, G1, FMY, and F5, indicating that WIV can also have an immunostimulating effect when combined with peptides other than GIL. Second, these data show that the F5 peptide was able to induce T cells which are cross-reactive with their wild-type analogs, which was also found in a previous study by Rosendahl Huber et al. (11). The individual peptides in both WT and modified pools did not show significant association with the WIV particles, similar to the previous observations with the GIL peptide alone (Table S1 in Supplementary Material). Thus, it is unlikely that differ- ences in induced immune responses by the peptide vaccines were caused by differences in association between peptide and WIV. Frontiers in Immunology \| www.frontiersin.org WIV Can Act as an Adjuvant for Multiple Combined Peptidesfi p To increase the efficacy and broadness of the peptide-induced immune response, multiple conserved epitopes should be included into a peptide-based influenza vaccine. To investigate whether WIV also acts as an adjuvant for other peptides, two subdominant conserved human HLA-A2.1-restricted influenza epitopes, FMYSDFHFI (FMY, PA46–54) and NMLSTVLGV (NML, PB1413–421), were added to the mixture of GIL and WIV. Another method to improve the immunogenicity of peptide antigens is by increasing the binding affinity to MHC I through chemical modification (11, 19). To examine whether this in com- bination with the improved immunogenicity of WIV adjuvant would further boost the immune response, chemically altered variants of the GIL, FMY, and NML peptide ligands, being [am-phg]ILGFVFTL (G1), [4-FPHE]MYSDFHF[2-AOC] (F5), and N[NLE]LSTVLGV (N53), were combined with WIV. As a reference control, peptides were also adjuvanted with the water- in-oil-based IFA. As described above, the GIL peptide in the peptide pool was able to induce GIL-specific T cell responses when mixed with WIV (Figure 5A), while IFA-adjuvanted GIL peptide induced significantly lower T cell responses. The modified G1 peptide was able to induce some GIL-specific responses, regardless of adjuvant; however, these responses were not statistically higher than the negative control group. The G1 peptide adjuvanted with either WIV or IFA did induce a G1-specific T cell response, indicating that while the modified peptide was immunogenic in combination with an adjuvant, it failed to induce high responses that were cross-reactive with the wild-type (WT) analog.hi The WT FMY peptide was able to induce FMY-specific T cell responses in combination with either WIV or IFA (Figure 5B). Interestingly, the modified F5 peptide was able to induce signifi- cantly higher FMY-specific responses compared to the WT FMY peptide when adjuvanted with WIV. F5 peptide adjuvanted with IFA did not show such an increase, indicating that WIV is a more potent adjuvant than IFA for the F5 peptide. This difference was also observed for the F5-specific responses; F5 peptide induced significantly higher F5-specific T cell responses when adjuvanted with WIV than with IFA.i Figure 5 \| T cell responses against wild-type and modified peptides adjuvanted with whole-inactivated influenza virus (WIV). HLA-A2.1 transgenic mice were vaccinated twice, 3 weeks apart with peptide pools containing either wild-type (WT) peptides (GIL, FMY, and NML, 100 nmol each) or modified (mod.) peptides (G1, F5, and N53, 100 nmol each) adjuvanted with either WIV or 50% (v/v) incomplete Freund’s adjuvant (IFA). DISCUSSION confirmed that immature murine DCs matured after exposure to WIV; CD40, CD80, and CD86 surface expression increased significantly after 24 h (27). Tapia-Calle et al. found similar surface expression patterns in human DCs, and additionally showed that IRF7, STAT1, and MyD88, which all play a role in the TLR7 signaling pathway, were significantly increased in DCs upon stimulation with WIV (28).f Short peptides covering conserved internal influenza T cell epitopes can induce cross-reactive T cell responses. However, the peptides themselves are not immunogenic enough to elicit strong T cell responses. In recent years, multiple adjuvants have been utilized to improve the immunogenicity of such peptides, including Pam2Cys and CpG-ODN (20, 21), delivery systems, such as liposomes and virosomes (22, 23) and combinations thereof (14, 24). However, such combinations often require complex multi-step formulation procedures, which are difficult to produce on a large scale. In addition, antigen encapsulation and/or adsorption is sometimes very low, hampering large-scale production of such formulations. Furthermore, some of the adjuvants have not received an approval for human use, which hinders the evaluation of such vaccine formulations in human trials.l Another contributing factor for the adjuvant effect of WIV might be related to an increased infiltration of inflammatory cells at the site of injection. Particulates such as WIV can activate innate cells, which is accompanied by the release of chemokines and cytokines (29). Particulate adjuvants can also activate NALP3 inflammasomes, which might enhance innate and cellular immu- nity (30). Inflammasomes are also known to be activated upon influenza virus infection (31). Whether the TLR7 pathway or the inflammasome pathway contributes to the adjuvanticity of WIV could be investigated in a future study using TLR7−/− and ASC−/− knockout mice, respectively. In this study, we improved the immunogenicity of influenza T cell epitopes by simply adding WIV to the peptides. We showed that the combination of WIV and GIL peptide induced signi ficantly higher T cell responses than the individual components. By screening various concentrations of GIL and WIV, we found that both GIL and WIV synergistically contributed to the GIL- specific T cell response. This confirms earlier findings that WIV alone is able to induce influenza-specific T cell responses in mice (25, 26). In addition, WIV contains multiple other T cell antigens, which may induce CD4+ T helper cells, that in turn can aid CD8+ CTL priming (32). DISCUSSION In addition to studies with the GIL peptide, we combined WIV with a mixture of dominant and subdominant epitopes, either in wild-type or CPL form. CPLs were selected based on their improved binding affinity on MHC-I molecules (19), which generally enhances immunogenicity (11). WIV was able to improve the homologous T cell responses of both GIL and G1 peptides. However, the modified G1 peptide did not induce higher T cell responses against the wild-type GIL epitope. By contrast, the modified F5 peptide was able to induce significantly higher FMY-specific T cell responses compared to WT FMY peptide in the presence of WIV. T cell responses against NML or N53 were almost not detectable; it is likely that the subdominant nature of the peptide was the cause of this. For the peptides GIL and F5, WIV might be able to replace IFA, which is commonly used as an adjuvant for peptide antigens.h While these results showed that peptide and WIV together elicit strong T cell responses, the exact mode of action of the observed adjuvant effect of WIV is still unclear. Similar particulate delivery systems, such as virosomes, facilitate the escape of peptide antigens from the endosomal compartment into the cytosol, enabling processing of the peptide in the MHC-I pathway (14). This endosomal escape is mediated by the membrane fusion capabilities of influenza surface proteins. We showed, however, that endosomal escape due to membrane fusion of the WIV particle is most likely not the mechanism of action, since inactivation of the membrane fusion capabilities of WIV had no effect on its adjuvant activity. In addition, we showed a lack of physical association between WIV and pep- tide, which indicates that it is unlikely that WIV functions as a particulate carrier system. This may be confirmed in a future study by conjugating the peptide with WIV and determine the immunogenicity of this construct. It is possible that the GIL peptides are cross-presented by the dendritic cells (DCs) via the vacuolar pathway, which does not require endosomal escape of the antigen, but does require the DCs to display a mature and activated phenotype. To acquire this mature phenotype, interac- tion with PAMPs by pattern recognition receptors such as TLRs on the DC is critical. WIV Can Act as an Adjuvant for Multiple Combined Peptidesfi March 2018 \| Volume 9 \| Article 525 6 Whole-Inactivated Influenza as an Influenza T Cell Adjuvant Soema et al. Frontiers in Immunology \| www.frontiersin.org March 2018 \| Volume 9 \| Article 525 ETHICS STATEMENT The Supplementary Material for this article can be found online at https://www.frontiersin.org/articles/10.3389/fimmu.2018.00525/ full#supplementary-material. Animal studies were conducted according to the guidelines pro- vided by the Dutch Animal Protection Act, and were approved by ACKNOWLEDGMENTS The authors thank Hugo Meiring from Intravacc for his assis tance with mass spectrometry, and Jolien Luimstra and Huib Ovaa from the Netherlands Cancer Institute for supplying the peptides. This work was supported by the Center for Translatio nal Molecular Medicine grant AMPVACS. DISCUSSION It is thus most likely that the presence of viral ssRNA in WIV provides strong TLR7 signaling (8), which activates the DCs and enables subsequent induction of T cell responses. Indeed, co-localization of the WIV particles with the peptide antigen was necessary to increase T cell responses; without WIV, T cell responses were low. This supports the hypothesis that local maturation of DCs through TLR7 activa- tion might be required for efficient cross presentation of the peptide antigen and subsequent T cell induction. Stoel et al. The peptides used in this study are all HLA-A2.1 binding peptides, which is one of the most frequently occurring alleles in the Caucasian population. However, allele frequencies differ between ethnicities, and thus multiple peptide covering multiple alleles should be combined to ensure complete coverage in the human population. Such approaches using peptide pools have already reached phase IIb clinical studies (33–36), with positive results. These T cell-inducing influenza vaccines would be ideal to reduce the morbidity and mortality in vulnerable people exposed to highly pathogenic influenza strains in the event of a pandemic, when a strain-specific vaccine is not yet available.l i A few comparable studies with HLA-A2.1 influenza peptides have been conducted previously in the literature. Recently, Herrera-Rodriguez et al. showed that a peptide pool consist- ing of four modified peptides adjuvanted with ISA-51 induced potent T cell responses in mice (37). These T cell responses were able to reduce weight loss and mortality in H1N1pdm-infected HLA-A2.1 transgenic mice. Our group previously showed that a single T cell peptide delivered by CpG-adjuvanted virosomes was also able to reduce weight loss in H3N2-infected HLA-A2.1 transgenic mice (14). Both these studies showed that vaccinated mice had approximately 200 IFN-γ-producing spot-forming March 2018 \| Volume 9 \| Article 525 7 Whole-Inactivated Influenza as an Influenza T Cell Adjuvant Soema et al. the Committee for Animal Experimentation (DEC) of the PD-Alt campus (Bilthoven, The Netherlands). units (SFU)/106 cells, whereas our current study shows that mice vaccinated with (modified) peptides adjuvanted with WIV reach about 400–900 IFN-γ-producing SFU/106 cells. This may be an indication that our WIV-adjuvanted peptides also offer some protection to heterologous influenza infection in mice.if AUTHOR CONTRIBUTIONS PS and SRH have designed and performed the experiment, as well as written the manuscript. G-JW, RJ, and MH have assisted with the immunological experiments. ES, WJ, JJ, JB, GK and J-PA have assisted with the experimental design, and assisted writing and reviewing the manuscript. l Our current study is the first to show the adjuvant effect of WIV on peptide-induced CD8+ T cell responses. Earlier it was shown that WIV can also boost humoral responses; Babb et al. showed that gamma-irradiated WIV was able to boost specific IgG and IgG2c responses to a gamma-irradiated SFV vaccine (10), but did not report any effect on T cell responses. Our data show that T cell responses are also boosted by WIV. Future studies are required to determine the minimum amount of WIV required to induce this adjuvant effect, since the concentration of WIV used in this study are higher than the dose required for a standard influenza vaccination. Taken together, these data show that inactivated whole influenza virus is an effective adjuvant for CD8+ T cell peptide antigens, which is easy to formulate and has a long track record with regard to usage and safety. REFERENCES 11. Rosendahl Huber SK, Luimstra JJ, van Beek J, Hoppes R, Jacobi RH, Hendriks M, et al. Chemical modification of influenza CD8+ T-cell epitopes enhances their immunogenicity regardless of immunodominance. PLoS One (2016) 11(6):e0156462. doi:10.1371/journal.pone.0156462 11. Rosendahl Huber SK, Luimstra JJ, van Beek J, Hoppes R, Jacobi RH, Hendriks M, et al. Chemical modification of influenza CD8+ T-cell epitopes enhances their immunogenicity regardless of immunodominance. PLoS One (2016) 11(6):e0156462. doi:10.1371/journal.pone.0156462 1. Soema PC, Kompier R, Amorij JP, Kersten GF. Current and next generation influenza vaccines: formulation and production strategies. Eur J Pharm Biopharm (2015) 94:251–63. doi:10.1016/j.ejpb.2015.05.023 1. Soema PC, Kompier R, Amorij JP, Kersten GF. Current and next generation influenza vaccines: formulation and production strategies. Eur J Pharm Biopharm (2015) 94:251–63. doi:10.1016/j.ejpb.2015.05.023 12. Hendriks J, Holleman M, de Boer O, de Jong P, Luytjes W. An international technology platform for influenza vaccines. Vaccine (2011) 29(Suppl 1):A8–11. doi:10.1016/j.vaccine.2011.04.124 2. Sridhar S, Begom S, Bermingham A, Hoschler K, Adamson W, Carman W, et al. Cellular immune correlates of protection against symptomatic pan- demic influenza. Nat Med (2013) 19(10):1305–12. doi:10.1038/nm.3350 2. Sridhar S, Begom S, Bermingham A, Hoschler K, Adamson W, Carman W, et al. Cellular immune correlates of protection against symptomatic pan- demic influenza. Nat Med (2013) 19(10):1305–12. doi:10.1038/nm.3350 13. Kon TC, Onu A, Berbecila L, Lupulescu E, Ghiorgisor A, Kersten GF, et al. Influenza vaccine manufacturing: effect of inactivation, splitting and site of manufacturing. comparison of influenza vaccine production processes. PLoS One (2016) 11(3):e0150700. doi:10.1371/journal.pone.0150700 3. Wang Z, Wan Y, Qiu C, Quiñones-Parra S, Zhu Z, Loh L, et al. Recovery from severe H7N9 disease is associated with diverse response mechanisms dominated by CD8(+) T cells. Nat Commun (2015) 6:6833. doi:10.1038/ ncomms7833 14. Soema PC, Rosendahl Huber SK, Willems GJ, Jiskoot W, Kersten GF, Amorij JP. Influenza T-cell epitope-loaded virosomes adjuvanted with CpG as a potential influenza vaccine. Pharm Res (2015) 32(4):1505–15. doi:10.1007/ s11095-014-1556-3 4. Wilkinson TM, Li CK, Chui CS, Huang AK, Perkins M, Liebner JC, et al. Preexisting influenza-specific CD4+ T cells correlate with disease protec- tion against influenza challenge in humans. Nat Med (2012) 18(2):274–80. doi:10.1038/nm.2612 15. Meiring HD, van der Heeft E, ten Hove GJ, de Jong APJM. Nanoscale LC–MS(n): technical design and applications to peptide and protein anal- ysis. J Sep Sci (2002) 25(9):557–68. doi:10.1002/1615-9314(20020601)25: 9<557::AID-JSSC557>3.0.CO;2-F 5. Soema PC, van Riet E, Kersten G, Amorij JP. Development of cross-protective influenza A vaccines based on cellular responses. REFERENCES Trends Immunol (2011) 32(1):34–41. doi:10.1016/j.it. 2010.11.004 32. Schoenberger SP, Toes RE, van der Voort EI, Offringa R, Melief CJ. T-cell help for cytotoxic T lymphocytes is mediated by CD40-CD40L interactions. Nature (1998) 393(6684):480–3. doi:10.1038/31002 22. Matsui M, Kohyama S, Suda T, Yokoyama S, Mori M, Kobayashi A, et al. A CTL-based liposomal vaccine capable of inducing protection against heterosubtypic influenza viruses in HLA-A0201 transgenic mice. Biochem Biophys Res Commun (2010) 391(3):1494–9. doi:10.1016/j.bbrc.2009. 12.100 33. Pleguezuelos O, Robinson S, Stoloff GA, Caparrós-Wanderley W. Synthetic Influenza vaccine (FLU-v) stimulates cell mediated immunity in a dou- ble-blind, randomised, placebo-controlled phase I trial. Vaccine (2012) 30(31):4655–60. doi:10.1016/j.vaccine.2012.04.089 23. Arkema A, Huckriede A, Schoen P, Wilschut J, Daemen T. Induction of cyto- toxic T lymphocyte activity by fusion-active peptide-containing virosomes. Vaccine (2000) 18(14):1327–33. doi:10.1016/S0264-410X(99)00404-1 j 34. Pleguezuelos O, Robinson S, Fernández A, Stoloff GA, Mann A, Gilbert A, et al. A synthetic influenza virus vaccine induces a cellular immune response that correlates with reduction in symptomatology and virus shedding in a randomized phase Ib live-virus challenge in humans. Clin Vaccine Immunol (2015) 22(7):828–35. doi:10.1128/CVI.00098-15 24. Nagata T, Toyota T, Ishigaki H, Ichihashi T, Kajino K, Kashima Y, et al. Peptides coupled to the surface of a kind of liposome protect infection of influenza viruses. Vaccine (2007) 25(26):4914–21. doi:10.1016/j.vaccine.2007. 04.010 35. van Doorn E, Pleguezuelos O, Liu H, Fernandez A, Bannister R, Stoloff G, et al. Evaluation of the immunogenicity and safety of different doses and formulations of a broad spectrum influenza vaccine (FLU-v) developed by SEEK: study protocol for a single-center, randomized, double-blind and placebo-controlled clinical phase IIb trial. BMC Infect Dis (2017) 17(1):241. doi:10.1186/s12879-017-2341-9 25. Bodewes R, Kreijtz JH, Hillaire ML, Geelhoed-Mieras MM, Fouchier RA, Osterhaus AD, et al. Vaccination with whole inactivated virus vaccine affects the induction of heterosubtypic immunity against influenza virus A/H5N1 and immunodominance of virus-specific CD8+ T-cell responses in mice. J Gen Virol (2010) 91(Pt 7):1743–53. doi:10.1099/vir.0.020784-0 26. Furuya Y, Chan J, Regner M, Lobigs M, Koskinen A, Kok T, et al. Cytotoxic T cells are the predominant players providing cross-protective immunity induced by {gamma}-irradiated influenza A viruses. J Virol (2010) 84(9): 4212–21. doi:10.1128/JVI.02508-09 36. Francis JN, Bunce CJ, Horlock C, Watson JM, Warrington SJ, Georges B, et al. A novel peptide-based pan-influenza A vaccine: a double blind, randomised clinical trial of immunogenicity and safety. Vaccine (2015) 33(2):396–402. doi:10.1016/j.vaccine.2014.06.006 27. REFERENCES Front Immunol (2015) 6:237. doi:10.3389/fimmu.2015.00237 16. Stegmann T, Morselt HW, Booy FP, van Breemen JF, Scherphof G, Wilschut J. Functional reconstitution of influenza virus envelopes. EMBO J (1987) 6(9):2651–9. i 6. Rosendahl Huber S, van Beek J, de Jonge J, Luytjes W, van Baarle D. T cell responses to viral infections – opportunities for peptide vaccination. Front Immunol (2014) 5:171. doi:10.3389/fimmu.2014.00171 17. Hem SL, Hogenesch H. Relationship between physical and chemical proper- ties of aluminum-containing adjuvants and immunopotentiation. Expert Rev Vaccines (2007) 6(5):685–98. doi:10.1586/14760584.6.5.685 i 7. Foged C, Hansen J, Agger EM. License to kill: formulation requirements for optimal priming of CD8(+) CTL responses with particulate vaccine delivery systems. Eur J Pharm Sci (2012) 45(4):482–91. doi:10.1016/j.ejps. 2011.08.016 18. Budimir N, Huckriede A, Meijerhof T, Boon L, Gostick E, Price DA, et al. Induction of heterosubtypic cross-protection against influenza by a whole inactivated virus vaccine: the role of viral membrane fusion activity. PLoS One (2012) 7(1):e30898. doi:10.1371/journal.pone.0030898 8. Budimir N, de Haan A, Meijerhof T, Waijer S, Boon L, Gostick E, et al. Critical role of TLR7 signaling in the priming of cross-protective cytotoxic T lympho- cyte responses by a whole inactivated influenza virus vaccine. PLoS One (2013) 8(5):e63163. doi:10.1371/journal.pone.0063163 j 19. Hoppes R, Oostvogels R, Luimstra JJ, Wals K, Toebes M, Bies L, et al. Altered peptide ligands revisited: vaccine design through chemically modi- fied HLA-A2-restricted T cell epitopes. J Immunol (2014) 193(10):4803–13. doi:10.4049/jimmunol.1400800 9. Wiesel M, Oxenius A. From crucial to negligible: functional CD8(+) T-cell responses and their dependence on CD4(+) T-cell help. Eur J Immunol (2012) 42(5):1080–8. doi:10.1002/eji.201142205 10. Babb R, Chan J, Khairat JE, Furuya Y, Alsharifi M. Gamma-irradiated influenza A virus provides adjuvant activity to a co-administered poorly immunogenic SFV vaccine in mice. Front Immunol (2014) 5:267. doi:10.3389/ fimmu.2014.00267 20. Tan AC, Deliyannis G, Bharadwaj M, Brown LE, Zeng W, Jackson DC. The design and proof of concept for a CD8(+) T cell-based vaccine inducing cross-subtype protection against influenza A virus. Immunol Cell Biol (2013) 91(1):96–104. doi:10.1038/icb.2012.54 March 2018 \| Volume 9 \| Article 525 Frontiers in Immunology \| www.frontiersin.org 8 Whole-Inactivated Influenza as an Influenza T Cell Adjuvant Soema et al. 21. Ichihashi T, Yoshida R, Sugimoto C, Takada A, Kajino K. Cross-protective peptide vaccine against influenza A viruses developed in HLA-A*2402 human immunity model. PLoS One (2011) 6(9):e24626. doi:10.1371/journal. pone.0024626 31. Pang IK, Iwasaki A. Inflammasomes as mediators of immunity against influenza virus. Frontiers in Immunology \| www.frontiersin.org REFERENCES Stoel M, Pool J, de Vries-Idema J, Zaaraoui-Boutahar F, Bijl M, Andeweg AC, et al. Innate responses induced by whole inactivated virus or subunit influ- enza vaccines in cultured dendritic cells correlate with immune responses in vivo. PLoS One (2015) 10(5):e0125228. doi:10.1371/journal.pone. 0125228 37. Herrera-Rodriguez J, Meijerhof T, Niesters HG, Stjernholm G, Hovden AO, Sørensen B, et al. A novel peptide-based vaccine candidate with protective efficacy against influenza A in a mouse model. Virology (2018) 515:21–8. doi:10.1016/j.virol.2017.11.018 28. Tapia-Calle G, Stoel M, de Vries-Idema J, Huckriede A. Distinctive responses in an in vitro human dendritic cell-based system upon stimulation with different influenza vaccine formulations. Vaccines (Basel) (2017) 5(3):E21. doi:10.3390/vaccines5030021 Conflict of Interest Statement: PS, G-JW, GK, and J-PA are or were employed by Intravacc. ES is owner of the company Virtuvax. All other authors declare no competing interests. Copyright © 2018 Soema, Rosendahl Huber, Willems, Jacobi, Hendriks, Soethout, Jiskoot, de Jonge, van Beek, Kersten and Amorij. This is an open-access article dis- tributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribu- tion or reproduction is permitted which does not comply with these terms. 29. Kuroda E, Coban C, Ishii KJ. Particulate adjuvant and innate immunity: past achievements, present findings, and future prospects. Int Rev Immunol (2013) 32(2):209–20. doi:10.3109/08830185.2013.773326 30. Sharp FA, Ruane D, Claass B, Creagh E, Harris J, Malyala P, et al. Uptake of particulate vaccine adjuvants by dendritic cells activates the NALP3 inflammasome. Proc Natl Acad Sci U S A (2009) 106(3):870–5. doi:10.1073/ pnas.0804897106 March 2018 \| Volume 9 \| Article 525 Frontiers in Immunology \| www.frontiersin.org 9
https://openalex.org/W4313574501	https://jurnal-stidnatsir.ac.id/index.php/dakwah/article/download/140/258	English	null	COMMUNICATION STRATEGY IN DELIVERING DA'WAH MESSAGES AND INCREASING THE NUMBER OF FOLLOWERS	Jurnal Da'wah	2,023	cc-by	6,358	ABSTRACT This study aims to describe how the communication strategy of an Instagram account @kajianislam is in conveying da'wah messages and in increasing its followers. This descriptive research was carried out in 2022 by selecting the object of study on the social media Instagram account @kajianislam using qualitative research methods. The results of this study indicate that the Instagram account @kajianislam is viewed from a communication strategy according to Onong Uchjana Effendy, namely identifying communication goals, selecting communication media, assessing communication objectives and source credibility. He also took steps to increase his followers so that he was successfully followed by 2.8 million followers. y y Keywords:Communication Strategy, Da'wah, Social Media, Instagram, Followers 1Erwin Jusuf Thaib, Problems of Da'wah on Social Media, Solok: Independent Scholars, 2021, p. 4. Da'wah: Risalah Merintis, Da'wah Melanjutkan are licensed under a Lisensi Creative Commons Atribusi 4.0 Internasional 57 \| Dakwah \| Vol 5 \| No. 2 \| 2022 INTRODUCTION Developments in Information and Communication Technology (ICT) have provided acceleration and convenience for both conveying and receiving information. So that the presence of the internet in modern human life has also presented a new model of social interaction, namely interaction using social media. It is known that almost all people at this time knew and even had social media accounts such as Instagram, Facebook, Twitter, YouTube, and various other social media platforms.1 Based on British media reports, on February 11 2021, We Are Social, in collaboration with Hootsuite, released the title "Digital 2021: The Latest Insights Into The State of Digital". The report presents the results of research on patterns of social media use in a number of countries, including Indonesia. It is known that the average Indonesian spends 3 hours 14 minutes a day accessing social media. As of January 2021, out of a total of 274.9 million Indonesians, 170 million or 61.8% are active users of social media. When compared Da'wah: Risalah Merintis, Da'wah Melanjutkan are licensed under a Lisensi Creative Commons Atribusi 4.0 Internasional 57 \| Dakwah \| Vol 5 \| No. 2 \| 2022 Da'wah: Risalah Merintis, Da'wah Melanjutkan are licensed under a Lisensi Creative Commons Atribusi 4.0 Internasional 57 \| Dakwah \| Vol 5 \| No. 2 \| 2022 to the previous year, the growth of social media users has increased by 10 million or around 6.3%.2 The growth in the number of social media users above shows that humans in the era of ICT development are dependent on the internet. Erwin Jusuf Thaib said, "taking advantage of advances in information technology is a necessity that cannot be negotiated."3Erwin's opinion is proven by the many various institutions such as government, education, economic, political, social and proselytizing institutions that have used ICT media to receive and convey information. This shows that the development of ICT has provided convenience and acceleration in the flow of receiving and delivering information. As in preaching, many da'wah activists use social media, one of the social media used is Instragram. Instagram is a social media platform that is currently popular with people around the world. Likewise with the people of Indonesia, Instagram users were recorded in July 2022 reaching 91.77 million people. Communication Strategy \| 58 2Conney Stephanie, How Long Do Indonesians Access the Internet and Social Media Every Day, Accessed April 6, 2022,https://amp.kompas.com/tekno/read/2021/02/23/11320087/berapa-lama-orang- indonesia-access-internet-dan-medsos-everyday. 3Erwin Jusuf Thaib, Problems of Da'wah on Social Media, Solok: Independent Scholars, 2021, p. 4. 4Monavia Ayu Rizaty, This is the country with the most Instagram users, what order is Indonesia?, accessed on 09 April 2022,https://databoks.katadata.co.id/datapublish/2021/08/03/inilah-negara- pengguna-instagram-terbanyak-indonesia-urutan-berapa. 5Zida Zakiyatul Husna and Abdul Muhid, “Development of Da'wah Through Social Media Instagram (Literature Review), Ath-Thariq, V, 02, 2021, p. 206-207. 6KumparanNews, 7 Famous Ustaz on Instagram: Abdul Somad to Hanan Attaki, Accessed 20 April 2022, https://m.kumparan.com/amp/kumparannews/7-ustaz-yang-kondang-di-instagram-abdul-somad- until-hanan-attaki 7@kajianislam, Accessed June 17, 2022,https://instagram.com/kajianislam?igshid=YmMyMTA2M2Y= INTRODUCTION This number puts Instagram in third place as a social media platform that is often used by Indonesians after YouTube and WhatsApp.4 Indonesia has also been named the largest Instagram user in Asia Pacific. The existence of Instagram as a social media that is widely accessed by the public has the opportunity to preach in it. Because in Instagram social media users can take advantage of various interesting features including communication, content and collaboration.5 Now it can be said that conveying information through social media Instagram is classified as easier and more practical. However, very few da'wah activists use Instagram as a media for proselytizing. This is evident from the many ustadz who actively preach offline, but only a few preach on social media, especially Instagram, with millions of followers. According to data from coilNews, there are 7 (seven) ustadz who preach on Instagram and have millions of followers.6Even so, there is also an Instagram account that in its posts has the value of da'wah messages and followers of up to millions, namely the Instagram account @kajianislam. The owner of the @kajianislam Instagram account is Abdullah. In his daily activities as a preacher he actively fills in the ta'lim. The Instagram account that he manages already has 2.8 million followers, approximately 16 thousand posts about da'wah messages, both in the form of posters and videos.7What's interesting about this @kajianislam account is that da'wah messages in the form of posters and videos are the result of reposts from other accounts. From the re-uploading activities carried out by the @kajianislam account, it turns out that the number of likes and video viewers is higher, when compared to the content creator's original account. 17, 58 Communication Strategy \| 58 Da'wah content on the Instagram account @kajianislam contains motivation, advice and discussions about fiqh and aqidah. The presentation does not seem monotonous. This can be seen from the number of likes which reached 15 thousand and its viewers reached 298 thousand. Based on the achievements of the @kajianislam Instagram account in conveying da'wah messages, the purpose of this study is first, to find out and understand how the communication strategy of the @kajianislam account is in conveying da'wah messages. Second, to find out and understand how the Instagram account @kajianislam has increased its followers. The theory used in this study uses Onong Uchjana Effendy's communication strategy theory8. 8Prof. Drs. Onong Uchjana Effendy, MA is one of the figures in communication science in Indonesia. He served as Dean of Fikom Unpad for the period 1975-1982 and has written 9 books on communication. 9Onong Uchjana Effendy, Theory and Practice of Communication Studies, Bandung: Rosdakarya Youth, 2017, Cet. XXVIII, p. 35. 10Edi Suryadi, Communication Strategy: A Theoretical and Practical Analysis in the Global Era, Bandung: PT Remaja Rosdakarya, 2021, Cet. II, p. 5. 11Afrizal, Qualitative Research Methods: An Effort to Support the Use of Qualitative Research in Various Disciplines. Depok: Rajawali Press, 2017, Cet., IV, p. 13. 12Sanapiah Faisal, Social Research Formats, Jakarta: PT RajaGrafindo Persada, 2001, Cet., V, p. 20. INTRODUCTION According to him, there are 4 (four) communication components that become supporters and obstacles that must be considered in developing a communication strategy, including: identifying communication goals, selecting communication media, assessing communication objectives and source credibility.9 According to Effendy as quoted by Edi Suryadi, that "communication strategy is a guide for planning (communications management) to achieve a goal". In order to achieve this goal, Effendy argues that there are two aspects that must be studied, namely, the strategy is interpreted macro (Planned multimedia strategy) and micro (single medium communication strategy). Both of these aspects have their functions, namely first, socializing communication messages in an informative, persuasive and instructive manner in order to obtain optimal results. Second, bridging the cultural gap, for example a communication message that has other cultural values will be better when it is applied and used in certain cultures so that they pay attention to developing strategies to package that information in an educational perspective.10 The research method in this study used a qualitative research approach. Namely social science research that collects and analyzes data in the form of words (oral and written) and human actions. In qualitative research, the data collected by researchers will not be counted as in quantitative research.11However, the format of this research design is to use a descriptive design format. A research for the exploration and clarification of phenomena or social reality, by way of describing a number of variables on the problems and units studied.12 p p j , , , , p 12Sanapiah Faisal, Social Research Formats, Jakarta: PT RajaGrafindo Persada, 2001, Cet., V, p. 20. 11Afrizal, Qualitative Research Methods: An Effort to Support the Use of Qualitative Research in Various Disciplines. Depok: Rajawali Press, 2017, Cet., IV, p. 13. 10Edi Suryadi, Communication Strategy: A Theoretical and Practical Analysis in the Global Era Bandung: PT Remaja Rosdakarya, 2021, Cet. II, p. 5. Prof. Drs. Onong Uchjana Effendy, MA is one of the figures in communication science in Indonesia He served as Dean of Fikom Unpad for the period 1975-1982 and has written 9 books on communication. 9Onong Uchjana Effendy, Theory and Practice of Communication Studies, Bandung: Rosdakarya Youth, 2017, Cet. XXVIII, p. 35. 13Riinawati, Introduction to Communication and Organizational Management Theory, Yogyakarta: Pustaka Baru Press, 2019, p. 12. 14Conney Stephanie, How Long Do Indonesians Access the Internet and Social Media Every Day, Accessed April 6, 2022,https://amp.kompas.com/tekno/read/2021/02/23/11320087/berapa-lama-orang- indonesia-access-internet-dan-medsos-everyday. 15Monavia Ayu Rizaty, This is the country with the most Instagram users, what order is Indonesia?, accessed on 09 April 2022,https://databoks.katadata.co.id/datapublish/2021/08/03/inilah-negara- pengguna-instagram-terbanyak-indonesia-urutan-berapa. 16Zida Zakiyatul Husna and Abdul Muhid, “Development of Da'wah Through Social Media Instagram (Literature Review), Ath-Thariq, V, 02, 2021, p. 206-207. RESULTS AND DISCUSSION To achieve an effective communication is not easy, because it is necessary to pay attention to the process in preparing the communication strategy. When going to communicate a communicant is advised to first make a plan by considering both supporting and inhibiting factors. This is very important so that in the communication process a communicant can do it effectively so as to achieve optimal results. 59 \| Dakwah \| Vol 5 \| No. 2 \| 2022 The process of communication is essentially the process of conveying a thought or feeling from a communicator to a communicant.13Thus, to achieve maximum results from a communication process requires planning. In the Instagram account @kajianislam the application of a communication strategy in conveying da'wah messages has gone through the following planning stages: a. Identify Communication Goals In developing a communication strategy, planning is needed by considering the condition and situation of the audience as the recipient of the message. This is because the conditions and situations of each communication target are sometimes different. These conditions require a communicator to recognize the target in communicating. Thus, the intent or content of the communication can be conveyed and well received by the communicant. The Instagram account @kajianislam found alignment with the strategy proposed by Onong Uchjana Effendy. In this case, it can be seen from the purpose of the account @kajianislam in conveying his da'wah messages motivated by the following words of Rasulullah ﷺ: ر )من دلّ على خري فله مثل أجر فاعله. (رواه مسلم Meaning: "Whoever shows (man) to goodness, he will get a reward according to the reward of the one who does it." (HR. Muslim). Instagram account @kajianislam, trying to get to know the goals of communication in general. As it is known that social media in adulthood has become a popular application, plus increasingly advanced and sophisticated technology has made it easier to access it. As data from British media We Are Social in collaboration with Hootsuite, which was published on 11 February 2021, that the average Indonesian spends 3 hours and 14 minutes a day accessing social media. And of the total population in Indonesia, which amounts to 274.9 million people, 170 million of them are active users of social media.14 Of all existing social media platforms, there were 91.77 million Instagram users in Indonesia in July 2022, which means Instagram ranks third as the most used social media platform in Indonesia after YouTube and WhatsApp.15Indonesia has even been named the largest Instagram user in Asia Pacific.16 Judging from the large number and diversity of Instagram users, the Instagram account manager @kajianislam, tries to identify communication targets before formulating what da'wah messages will be conveyed, namely heterogeneous general audiences, who have frames of reference and very different situations. diverse and currently in the midst of the onslaught of information flows that are so strong and so free. a. Identify Communication Goals As Wiwit Hardi, the 60 Communication Strategy manager of the Instagram account @kajianislam revealed that "...I am looking for general posts that can be accepted by many people..."17 As for the efforts of the Instagram account @kajianislam, so that the process of conveying the messages of da'wah can run effectively, namely by conveying material that is easily accepted and understood by Muslims as a whole without any groupings. As according to Wiwit Hardi, the Instagram account manager @kajianislam revealed that "...In general, there are no specific da'wah messages for certain groups, but are intended for all people, both general and ordinary people, with material that is easier to understand."18 As explained by Onong Uchjana Effendy, recognizing the frame of reference of communicants in mass communication is the most difficult level, so that the communication delivered must be informative and general in nature, which can be understood by everyone.19 Likewise with the Instragram account @kajianislam, in conveying his da'wah messages paying attention to the situation and conditions, to achieve success in conveying his da'wah messages. This really needs to be considered, because Instagram is a social media in which there are users who can send and receive information at any time. In fact, sometimes information becomes viral or popular among social media users, because the message conveyed has value packaged in a simple and attractive way. g y p g p y Wiwit Hardi, manager of the Instagram account @kajianislam, when asked to respond about a message that went viral on social media, he said that "if you really need comments, I will comment on them, this is only as needed..."20Looking at Wiwit Hardi's answer, it provides information that the Instagram account @kajianislam always pays attention to the situation and conditions of its communication targets. This is done to be careful in conveying messages so that there are no misunderstandings by social media users, especially followers of the Instagram account @kajianislam. 17Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 18Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 19Onong Uchjana Effendy, Theory and Practice of Communication Studies, Bandung: Rosdakarya Youth, 2017, Cet. XXVIII, p. 35. 20Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 21Sri Ayu Rayhaniah, et. al., Organizational Ethics and Communication, Aceh: Muhammad Zaini Publishers Foundation, 2021, Cet., p. 143. 22Onong Uchjana Effendy, Communication Science Theory and Practice, p. 37. 23Ibid.,matter. 37. 19Onong Uchjana Effendy, Theory and Practice of Communication Studies, Bandung: Rosdakarya Youth, 2017, Cet. XXVIII, p. 35. 20Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 21Sri Ayu Rayhaniah, et. al., Organizational Ethics and Communication, Aceh: Muhammad Zaini Publishers Foundation, 2021, Cet., p. 143. 22Onong Uchjana Effendy, Communication Science Theory and Practice, p. 37. 23Ibid.,matter. 37. 17Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 18Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 19Onong Uchjana Effendy, Theory and Practice of Communication Studies, Bandung: Rosdakarya Youth, 2017, Cet. XXVIII, p. 35. b. Selection of Communication Media Communication media is a means used as a messenger from the communicator to the communicant.21Communication media are very diverse and numerous, both traditional and modern, such as drums, kentongan, letters, newspapers, magazines, telephone, radio, television, internet, and so on. It is generally known that these media can be classified as written, printed, visual, aural, and audio-visual media.22 To reach communication goals, the communicator can choose one of the media, from several existing media in achieving communication goals. It also depends on the goals to be achieved. It is necessary to pay attention to the messages to be conveyed and the techniques to be used, because each communication medium has its own advantages and disadvantages.23 g One of the most efficient and popular communication media today is social media. The presence of the internet in modern human life has also presented a new model of 61 \| Dakwah \| Vol 5 \| No. 2 \| 2022 social interaction among modern society. That is interaction using social media. It is known that almost all people at this time knew and even had social media accounts such as Instagram, Facebook, Twitter, Youtube, and various other social media platforms.24 Among the social media applications that are quite popular among the public is the Instagram application, even among foreign artists without exception.25The Instagram application is an application intended for sharing pictures and videos.26And among the characteristics of this application is that the uploaded images must be beautiful and the videos are of short duration with the shortest possible caption.27Likewise with the @kajianislam account, packaging its da'wah messages in the form of pictures and videos of short duration. This is one of the reasons for selecting the Instagram application as a da'wah medium by the Instagram account manager @kajianislam. As Wiwit Hardi, revealed the reason for using Instagram media, he said “because there are many Instagram users, and until now there are still many who access Instagram almost the same as Facebook and Tik-Tok. It's just that Tik-Tok has drawbacks, the first is that it's difficult to create content, because you have to use videos. While Instagram, the most important thing is the picture, which at this time can upload up to ten picture slides..”28 p p p Instagram is an application whose usage always increases every year. As data from Business of Apps shows that Instagram users globally have reached 1.96 billion users. 24Erwin Jusuf Thaib, Erwin Jusuf Thaib, Problems of Da'wah on Social Media, Solok: Independent Scholars, 2021, p. 4. 25Dhifa Nabila, et., al., Civilization of Social Media in the Industrial Age 4.0, Malang: Communication Studies Study Program, Muhammadiyah Malang University in Collaboration with Media Intelligence, 2020, p. 74. 26Akhmad Muwafik Saleh et., al., Communipreneur: Creative Communication Models in the Industry 4.0 Era, Malang: Media Intelligence, 2020, Cet., p. 52. 27Dhifa Nabila, et., al., Social Media Civilization in the Industrial Age 4.0, p. 74. 28Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 29Monavia Ayu Rizaty, Increased Again, This is the Number of Instagram Users as of the First Quarter of 2022, accessed on 09 April 2022, https://databoks.katadata.co.id/datapublish/2022/06/17/berplus-lagi-ini-nomor- instagram-users-as-i- quarter-2022 30Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 24Erwin Jusuf Thaib, Erwin Jusuf Thaib, Problems of Da'wah on Social Media, Solok: Independent lars, 2021, p. 4. 25Dhif N bil l Ci ili i f S i l M di i h I d i l A 4 0 M l C i i 30Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 24Erwin Jusuf Thaib, Erwin Jusuf Thaib, Problems of Da'wah on Social Media, Solok: Independent Scholars, 2021, p. 4. 25Dhifa Nabila, et., al., Civilization of Social Media in the Industrial Age 4.0, Malang: Communication Studies Study Program, Muhammadiyah Malang University in Collaboration with Media Intelligence, 2020, p. 74. 26Akhmad Muwafik Saleh et., al., Communipreneur: Creative Communication Models in the Industry 4.0 Era, Malang: Media Intelligence, 2020, Cet., p. 52. 27Dhifa Nabila, et., al., Social Media Civilization in the Industrial Age 4.0, p. 74. 28Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 29 A d A b f f 26Akhmad Muwafik Saleh et., al., Communipreneur: Creative Communication Models in the Industry 4.0 Era, Malang: Media Intelligence, 2020, Cet., p. 52. 27Dhifa Nabila, et., al., Social Media Civilization in the Industrial Age 4.0, p. 74. 28Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 29Monavia Ayu Rizaty, Increased Again, This is the Number of Instagram Users as of the First Quarter of 2022, accessed on 09 April 2022, https://databoks.katadata.co.id/datapublish/2022/06/17/berplus-lagi-ini-nomor- instagram-users-as-i- quarter-2022 30Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 25Dhifa Nabila, et., al., Civilization of Social Media in the Industrial Age 4.0, Malang: Communication Studies Study Program, Muhammadiyah Malang University in Collaboration with Media Intelligence, 2020 p. 74. 26Akhmad Muwafik Saleh et al Communipreneur: Creative Communication Models in the Industry 31Onong Uchjana Effendy, Communication Science Theory and Practice, p. 37. 32Ibid.,matter. 37-38. 33Alvian Hardianto, et., al., Business Communication: Communication Concepts & Practices, Surabaya: Scopinda Media Pustaka, 2020, Cet., I, p. 101. 34Sunarno Sastro Atmodjo, Intercultural Communication, Bandung: Indonesian Science Media, 2021, p. 72. 35Sayyidatul Khoiridah (ed), et., al., Padamu Negeri Kami Berbad (Education, Technology, Health, Agriculture, Fisheries and Food), Surabaya: Unitomo Press, 2022, Cet., I, p. 25. b. Selection of Communication Media In everyday life we can find lots of examples of the contents of communication messages conveyed to the communicant by using a combination of several symbols, such as communication messages through newspapers, films or television.32 The techniques and forms of messages conveyed by the Instagram account @kajianislam are as follows: b. Selection of Communication Media As of the first quarter of 2022, this number has increased by 1.67% from the previous quarter, which had 1.92 billion users. Meanwhile, when compared to the first quarter of 2021, the growth reached 4.42% within a year.29 The growth of Instagram users has both good and bad sides. On the bright side, that the opportunities for da'wah messages to be delivered are wider. The downside is that it is balanced with da'wah messages that can be increasingly widespread. Likewise, negative information is increasingly widespread, such as Islamophobic, liberal and secular thoughts. g y p p g In response to this, as explained by Wiwit Hardi, "liberal people are like that, they promote it like that, right, this has to be countered."30The answer shows that social media such as Instagram is an effective medium to serve as a means of proselytizing in today's global era. Only in choosing the form of communication media must be considered carefully and then packaged in an attractive and easy way to be understood by the general public, especially audiences on social media (cyberspace). c. Communication Objectives Assessment Communication messages must have a goal to be achieved. Because the purpose of the communication message will determine the appropriate technique that will be used by the communicator when he wants to carry out the communication process. Onong Uchjana Effendy, divides the techniques that can be used by communicators when carrying c. Communication Objectives Assessment Communication messages must have a goal to be achieved. Because the purpose of the communication message will determine the appropriate technique that will be used by the communicator when he wants to carry out the communication process. Onong Uchjana Effendy, divides the techniques that can be used by communicators when carrying 62 Communication Strategy out the communication process into three techniques, namely: information techniques, persuasion techniques and instruction techniques.31 Onong Uchjana Effendy also divides communication messages into two functions, namely the content of the message and symbols. The content of the communication message can be one, but the symbol used can be varied. Symbols that can be used to convey the content of communication can be in the form of language, images, colors, gestures and so on. 1) Information Engineering Information engineering is a message aimed at broadening public insight and awareness, which is usually a simpler and clearer message to make it easier for the communication target to understand.33Among the examples of applying da'wah messages with information techniques, namely uploads or posts that provide information about recommended study places. Like a book recommendation; motivation about the enormity of a practice if implemented; and the like. The contents of this message are posted on the Instagram account @kajianislam in the form of writing, illustrated pictures, and audiovisual which are informative in nature. 2) Persuasion Techniques Persuasion technique is the process of conveying messages by the communicator to the communicant with the hope that the attitudes, actions, opinions and behavior of the communicant change with their own awareness.34Among the examples of applying da'wah messages with persuasion techniques from the Instagram account @kajianislam, are uploads or da'wah posts that are motivational or solicitation. Like an invitation to do sunnah prayers that are equipped with information will get a reward when doing it, and posts like that. Then the da'wah messages are uploaded in either the form of writing, illustrations, pictures or audiovisual form. 2) Persuasion Techniques Persuasion technique is the process of conveying messages by the communicator to the communicant with the hope that the attitudes, actions, opinions and behavior of the communicant change with their own awareness.34Among the examples of applying da'wah messages with persuasion techniques from the Instagram account @kajianislam, are uploads or da'wah posts that are motivational or solicitation. Like an invitation to do sunnah prayers that are equipped with information will get a reward when doing it, and posts like that. Then the da'wah messages are uploaded in either the form of writing, illustrations, pictures or audiovisual form. 3) Instruction Technique Communication techniques Instruction is a directive to do something, the purpose of giving instructions is to make it clear what is expected and how to process it.35Among the application of instructional techniques on the Instagram account @kajianislam is a video on how to perform the id prayer, or something similar. Then the forms of these messages are posted on the Instagram account @kajianislam in the form of writing, picture illustrations and audiovisuals. Judging from the explanation above, there is a pattern of packaging of da'wah message content carried out by the Instagram account @kajianislam when practicing the three communication techniques, namely da'wah content packaged in writing, illustrated 63 \| Dakwah \| Vol 5 \| No. 2 \| 2022 63 images, audiovisual and videovisual. This is done as an effort to convey da'wah messages effectively and produce optimal results. d. The Role of the Communicator in Communication There are two important factors that must be present in the communicator in facilitating communication, namely source attractiveness and source credibility. The meaning of source appeal is when the communicant feels that the communicator is participating with him, which is when the communicant feels there are similarities between himself and the communicator, then the communicant will be willing to obey the contents of the message conveyed by the communicator.36 himself and the communicator, then the communicant will be willing to obey the contents of the message conveyed by the communicator.36 While what is meant by source credibility is the communicant's trust in the communicator. This belief has a lot to do with the profession or expertise possessed by the communicator. Someone who works as an ustadz, then of course he will gain trust if he explains issues related to religion.37 While what is meant by source credibility is the communicant's trust in the communicator. This belief has a lot to do with the profession or expertise possessed by the communicator. Someone who works as an ustadz, then of course he will gain trust if he explains issues related to religion.37 At this last stage of communication strategy theory, the @kajianislam Instagram account has its own way. That is, the identity of the Instagram account @kajianislam was created anonymously. Namely, the identity of the account manager is kept secret from the public. 36Onong Uchjana Effendy, Communication Science Theory and Practice, p. 37. 37Ibid.,matter. 39 38Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 39Ahmad Zaki Abdul Aziz, "Optimizing Instagram as a Media for Delivering Da'wah Messages", Imperatif Syntax Journal, I, 04, 2020, p., 218. 3) Instruction Technique As Wiwit Hardi said, "...convincing people is difficult, isn't it, because of the people, isn't it the @kajianislam account, the person is anonymous, right..."38 40Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 41M. Suyanto, Start a Business from Zero-Learn from the World's Top Entrepreneurs \| Smart in Entrepreneurship, Yogyakarta: Andi, 2019, p. 171 42Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 43Rachmad Igen, Online Selling Success Strategies, Jakarta: Mediakita, 2010, cet., I, p. 7. 44Rahman Sidiq, Selling Techniques with Digital Marketing, Udacoding, p. 47. 45Interview with Wiwit Hardi P, Instagram Account Admin @kajianislam, Interview, Yogyakarta, 4 August 2022. 46Rahman Sidiq, Selling Techniques with Digital Marketing, p. 47. 1. @kajianislam Instagram Account Strategy to Increase Followers At present it can be said that preaching through social media, especially on Instagram, is classified as easier and also practical, this is none other than because the coverage is almost evenly distributed throughout Indonesia and is supported by an application that has various interesting features. However, it turns out that on the other hand, what is becoming a difficulty for da'wah activists at this time is how to use social media such as Instagram as a medium for proselytizing. Da'wah messages delivered through Instagram media can be conveyed effectively and the Instagram account can develop properly. p p p Currently there are many Instagram accounts that post da'wah content. However, it looks less effective and only runs in place. The reasons are varied, from the presentation of unattractive content to the lack of knowledge of the manager. Such as about what content is needed by the audience (mad'u da'wah), so that engagement from delivering da'wah never increases.39 It should be noted that there are supports for the effectiveness of proselytizing on Instagram media, namely followers. Where with the increasing number of da'wah objects, it has the potential to spread da'wah messages and even become viral and popular. So to answer this problem, the researcher conducted an in-depth study in this study. This is important to research because the Instagram account @kajianislam has followers that have been increasing every year since 2013 until now. Meanwhile, the Instagram account @kajianislam has 2.8 million followers. Instagram account manager @kajianislam, Wiwit Hardi, said: “...In the past, I had a lot of followers because few people knew Instagram, so if you want to do a search in the explore column about something that has Islamic roots, then most likely what will appear is the Instagram account @kajianislam. This was due 64 Communication Strategy to the lack of alternatives to follow da'wah accounts at that time, in contrast to the current situation there are more..."40 As for the results of research on how the Instagram account @kajianislam increases their followers, they are committed to being the first account. As M. Suyanto quoted Jack Trout saying, that one of the essence of strategy is different. 1. @kajianislam Instagram Account Strategy to Increase Followers So to be different can be done by being the first.41It is known that exactly three years after the launch of the Instagram application, the Instagram account @kajianislam was created.42At that time, there were still very few accounts spreading Islamic da'wah content. This is what makes the explore results from searching for Islamic content always point to the Instagram account @kajianislam, so that the followers of this account automatically increase and continue to grow. The next strategy implemented by the Instagram account @kajianislam is by giving an account name that is common and easy to remember. The name of this account has the keywords "Kajian and Islam". The choice of these two words makes the Instagram account @kajianislam always appear as an option when searching for the word "Kajian or Islam" on the search page on Instagram. Rachmad Igen explained that names that are short and easy to remember are meant to be easily memorized by everyone. As for the short name in the case of the internet, it can give a pretty good contribution and promotion.43 The Instagram account @kajianislam, on the content uploading strategy side, always tries to apply consistency in posting content, namely six times a day. This is also in line with Rahmat Sidiq's explanation in his book, Selling Techniques with Digital Marketing. He explained that posting regularly at least three times a day would help content marketing on Instagram. This is none other than because the Instagram system can assess dynamic active users on Instagram. So that the post will be assisted by the system, to appear on the newsfeed and on the Instagram search page.44 Meanwhile, in terms of posting time, the Instagram account @kajianislam determines the posting schedule, namely: after dawn and Duha time with two content postings; noon after noon one post; in the evening after asr once post; and at the time of Isha once post.45The regular posting of Instagram content, as explained by Rahmat Sidiq, should make posts at times when users are active and access social media. Namely in the morning, during the afternoon break, afternoon time, and at night break time. Because usually many users of social media like Instagram actively access it at these times.46 The strategy implemented by the Instagram account @kajianislam in terms of content selection is to choose an interesting image. So that it can invite lots of interactions like the content. 1. @kajianislam Instagram Account Strategy to Increase Followers The more interesting the images posted, the more likes they will receive and the wider their distribution will be. This is because posting on Instagram generates an 65 \| Dakwah \| Vol 5 \| No. 2 \| 2022 average engagement of 4.21% per follower. In this respect it is 58 times bigger than Facebook and 120 times bigger than Twitter.47 Another strategy implemented by the Instagram account @kajianislam is to post more video reels. Namely short video clips that contain da'wah messages. As in line with the explanation from Novy Khusnul Khotimah, that video reels have a wider reach and have more viewers.48 Thus, several strategies or stages implemented by the Instagram account @kajianislam in an effort to increase its followers have been carried out since 2013 until now. CONCLUSION Based on the results of the research described above, the conclusion of this study is that there is a relationship between the application of communication strategies to increasing the number of followers of the Instagram account @kajianislam. Where in conveying his da'wah messages he makes the general public the target of his da'wah. Choose da'wah materials in the form of da'wah messages that are easily understood and accepted by the general public. Choose a name for the account by choosing a name that is common and easy to remember. Make attractive illustrations in conveying the messages of da'wah, both in the form of posters and video reels. In increasing its followers, the Instagram account @kajianislam, established this account as the earliest proselytizing account on the Instagram application. Then choose a name for the account by choosing a name that is common and easy to remember. Make interesting illustrations in conveying the messages of da'wah. Da'wah content that is uploaded in the form of videos, implemented by posting more videos in the form of video reels. The Instagram account @kajianislam when choosing the role of communicator in communication, gives an anonymous name. This is an attempt when in conveying his da'wah messages what is seen is the content or content of the da'wah message, not to the person. 47Mira R., and Linda Irawati (ed.), 99 Steps to Success in E-Commerce Business @Politwika, Jakarta: Gramedia Widiasarana Indonesia, 2017, p. 177. p 48Novy Khusnul Khotimah, Utilization of Social Media by Government Public Relation Practitioners in Indonesia, Yogyakarta: Jejakpustaka, 2021, Cet., I, p. 21. REFERENCE Afrizal, Metode Penelitian Kualitatif: Sebuah Upaya Mendukung Penggunaan Penelitian Kualitatif dalam Berbagai Disiplin Ilmu. Depok: Rajawali Pers, Cet., IV., 2017. g p p j Atmodjo, Sunarno Sastro, Komunikasi Antar Buday, Bandung: Media Sains Indonesia, 2021 j , , y, g , Aziz, Ahmad Zaki Abdul, ”Optimasi Instagram Sebagai Media Penyampaian Pesan Dakwah”, Jurnal Syntax Imperatif, I, 04, 2020. J y p f Effendy, Onong Uchjana, Ilmu Komunikasi Teori dan Praktek, Bandung: Remaja Rosdakarya, Cet. XXVIII., 2017. Faisal, Sanapiah, Format-Format Penelitian Sosial,Jakarta: PT RajaGrafindo Persada, Cet. V., 2001. Hardianto, Alvian, et., al.,, Business Communication: Konsep & Praktek Berkomunikasi, Surabaya: Scopinda Media Pustaka, Cet., I., 2020. 47Mira R., and Linda Irawati (ed.), 99 Steps to Success in E-Commerce Business @Politwika, Jakarta: Gramedia Widiasarana Indonesia, 2017, p. 177. Communication Strategy \| 66 66 Husna, Zida Zakiyatul dan Abdul Muhid, “Perkembangan Dakwah Melalui Media Sosia Instagram (Literatur Review), Ath-Thariq, V, 02, 2021. g ( ) q Igen, Rachmad, Strategi Sukses Berjualan Online, Jakarta: Mediakita, , cet., I., 2010. Khoiridah, Sayyidatul, (ed), et., al., Padamu Negeri Kami Mengbad (Pendidikan, Teknologi, Kesehatan, Pertanian, Perikanan, dan Pangan), Surabaya: Unitomo Press, Cet., I., 2022. Khotimah, Novy Khusnul, Pemanfaatan Media Sosial Oleh Praktisi Humas Pemerintah d Indonesia, Yogyakarta: Jejakpustaka, Cet., I., 2021. Mira R., dan Linda Irawati (ed.), 99 Langkah Sukses Berbisni E-Commerce @Politwika, Jakarta: Gramedia Widiasarana Indonesia, 2017. Nabila, Dhifa, et., al., Peradaban Media Sosial di Era Industri 4.0, Malang: Prodi Ilmu Komunikasi Univesitas Muhammadiya Malang Bekerjasama dengan Intelegensia Media, 2020. Rayhaniah, Sri Ayu, et. al., Etika dan Komunikasi Organisasi, Aceh: Yayasan Penerbit Muhammad Zaini, 2021. Riinawati, Pengantar Teori Manajemen Komunikasi dan Organisasi, Yogyakarta: Pustaka Baru Press, 2019. Saleh, Akhmad Muwafik, et., al.,Communipreneur: Model-model Komunikasi Kreatif di Era Industru 4.0, Malang: Intelegensia Media, 2020. g g Sidiq, Rahman, Teknik Berjualan dengan Digital Marketing, Udacoding. Sidiq, Rahman, Teknik Berjualan dengan Digital Marketing, Udacoding. Suryadi, Edi, Strategi Komunikasi: Sebuah Analisis Teori dan Praktis di Era Global, Bandung: PT Remaja Rosdakarya, Cet. II., 2021. j Suyanto, M., Memulai Bisnis dari Nol-Beajar dari Entrepreneur Top Dunia \| Smart in Entrepreneurship, Yogyakarta: Andi, 2019. Thaib, Erwin Jusuf, Problematika Dakwah di Media Sosial, Solok: Insan Cendikia Mandiri, 2021. Wawancara dengan Wiwit Hardi P, Admin Akun Instagram @kajianislam, Wawancara Yogyakarta, 4 Agustus 2022. @kajianislam, Diakses pada 17 Juni 2022, https://Instagram.com/kajianislam?igshid=YmMyMTA2M2Y= Conney Stephanie, Berapa Lama Orang Indonesia Akses Internet dan Medsos Setiap Hari, Diakses pada 6 April 2022, https://amp.kompas.com/tekno/read/2021/02/23/11320087/berapa-lama- orang-indonesia-akses-internet-dan-medsos-setiap-hari. REFERENCE KumparanNews, 7 Ustaz yang Kondang di Intagram: Abdul Somad Hingga Hanan Attaki, Diakses 20 April 2022, https://m.kumparan.com/amp/kumparannews/7-ustaz- yang-kondang-di-Instagram-abdul-somad-hingga-hanan-attaki. y g g g gg Monavia Ayu Rizaty, Inilah Negara Pengguna Intagram Terbanyak, Indonesia Urutan Berapa?, diakses pada 09 April 2022, https://databoks.katadata.co.id/datapublish/2021/08/03/inilah-negara- pengguna-Instagram-terbanyak-indonesia-urutan-berapa. 67 \| Dakwah \| Vol 5 \| No. 2 \| 2022
https://openalex.org/W4247704465	https://tuprints.ulb.tu-darmstadt.de/17838/1/09252867.pdf	English	null	<i>RESCUE</i>: A Resilient and Secure Device-to-Device Communication Framework for Emergencies	IEEE Transactions on Dependable and Secure Computing/IEEE transactions on dependable and secure computing	2,022	cc-by	14,463	RESCUE: A Resilient and Secure Device-to-Device Communication Framework for Emergencies Milan Stute, Florian Kohnh¨auser, Lars Baumg¨artner, Lars Almon, Matthias Hollick, Stefan Katzenbeisser, and Bernd Freisleben ! ! operations. However, panic reactions and physical damage often lead to inoperable local communication infrastruc- tures, e.g., during the 2013 Typhoon Haiyan [1]. To provide an alternative to expensive satellite-based communication, many researchers have proposed to leverage the ad hoc and disruption-tolerant networking (DTN) capabilities of ubiquitous mobile devices, where all devices relay messages for others in the network [2], [3]. Nevertheless, due to its cooperative, distributed, and resource-constrained nature, DTN communication is vulnerable to many attacks [4]. Adversaries, e.g., terrorists, may exploit such vulnerabilities to subvert the communication system and disrupt disaster relief operations. Furthermore, people in panic may spam the network with messages and thereby unintentionally jeopardize availability. Existing DTN security solutions at- tempt to identify and then blacklist attackers [5]. However, such approaches can exhibit false positives, causing a valid user to be mistakenly identiﬁed as an adversary and be excluded from communication. Furthermore, Sybil attacks, where an adversary operates under multiple identities, have not been addressed in previous DTN research. Abstract—During disasters, existing telecommunication infrastructures are often congested or even destroyed. In these situations, mobile devices can form a backup communication network for civilians and emergency services using disruption-tolerant networking (DTN) princi- ples. Unfortunately, such distributed and resource-constrained networks are particularly susceptible to a wide range of attacks such as terror- ists trying to cause more harm. In this paper, we present RESCUE, a resilient and secure device-to-device communication framework for emergency scenarios that provides comprehensive protection against common attacks. RESCUE features a minimalistic DTN protocol that, by design, is secure against notable attacks such as routing manipulations, dropping, message manipulations, blackholing, or impersonation. To further protect against message ﬂooding and Sybil attacks, we present a twofold mitigation technique. First, a mobile and distributed certiﬁcate infrastructure particularly tailored to the emergency use case hinders the adversarial use of multiple identities. Second, a message buffer management scheme signiﬁcantly increases resilience against ﬂooding attacks, even if they originate from multiple identities, without introduc- ing additional overhead. Finally, we demonstrate the effectiveness of RESCUE via large-scale simulations in a synthetic as well as a realistic natural disaster scenario. Our simulation results show that RESCUE achieves very good message delivery rates, even under ﬂooding and Sybil attacks. In this paper, we present RESCUE, a resilient and secure device-to-device communication framework for emergency scenarios which is the ﬁrst work to provide comprehensive attack protection. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing epted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing 1 RESCUE: A Resilient and Secure Device-to-Device Communication Framework for Emergencies Milan Stute, Florian Kohnh¨auser, Lars Baumg¨artner, Lars Almon, Matthias Hollick, Stefan Katzenbeisser, and Bernd Freisleben RESCUE’s basic communication protocol relies on epidemic routing, authenticated and immutable messages, and an effective acknowledgment processing. This way, common attacks, such as message or routing manipulation, blackholing, or impersonation, are already prevented. Yet, as in today’s Internet infrastructure [6], the key challenge is to defend against Denial of Service (DoS) attacks originating from individuals as well as multiple identities (Sybil attack) that ﬂood the network. • M. Stute, L. Almon, and M. Hollick are with Secure Mobile Networking Lab, Technische Universit¨at Darmstadt, Germany. E-mails: {mstute,lalmon,mhollick}@seemoo.tu-darmstadt.de • F. Kohnh¨auser is with Security Engineering Group, Technische Univer- sit¨at Darmstadt, Germany. E-mails: kohnhaeuser@seceng.informatik.tu-darmstadt.de • L. Baumg¨artner is with Software Technology Group, Technische Univer- sit¨at Darmstadt, Germany. E-mail: baumgaertner@cs.tu-darmstadt.de • S. Katzenbeisser is with Faculty of Computer Science and Mathematics, University of Passau, Germany. E-mail: Stefan.Katzenbeisser@uni-passau.de • B. Freisleben is with Department of Mathematics & Computer Science, Philipps-Universit¨at Marburg, Germany. E-mail: freisleb@informatik.uni-marburg.de This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ • M. Stute, L. Almon, and M. Hollick are with Secure Mobile Networking Lab, Technische Universit¨at Darmstadt, Germany. E-mails: {mstute,lalmon,mhollick}@seemoo.tu-darmstadt.de • F. Kohnh¨auser is with Security Engineering Group, Technische Univer- sit¨at Darmstadt, Germany. E-mails: kohnhaeuser@seceng.informatik.tu-darmstadt.de • L. Baumg¨artner is with Software Technology Group, Technische Univer- sit¨at Darmstadt, Germany. E-mail: baumgaertner@cs.tu-darmstadt.de • S. Katzenbeisser is with Faculty of Computer Science and Mathematics, f f p • B. Freisleben is with Department of Mathematics & Computer Science, Philipps-Universit¨at Marburg, Germany. E-mail: freisleb@informatik.uni-marburg.de • M. Stute, L. Almon, and M. Hollick are with Secure Mobile Networking Lab, Technische Universit¨at Darmstadt, Germany. E-mails: {mstute,lalmon,mhollick}@seemoo.tu-darmstadt.de h h h h h y f y E-mail: Stefan.Katzenbeisser@uni-passau.de 1 INTRODUCTION During ﬂoods, hurricanes, earthquakes, nuclear accidents, or terrorist attacks, fast disaster response can save human life, limit environmental damage, and reduce economic loss. Communication technologies are integral to disaster relief • M. Stute, L. Almon, and M. Hollick are with Secure Mobile Networking Lab, Technische Universit¨at Darmstadt, Germany. E-mails: {mstute,lalmon,mhollick}@seemoo.tu-darmstadt.de • F. Kohnh¨auser is with Security Engineering Group, Technische Univer- sit¨at Darmstadt, Germany. E-mails: kohnhaeuser@seceng.informatik.tu-darmstadt.de • L. Baumg¨artner is with Software Technology Group, Technische Univer- sit¨at Darmstadt, Germany. E-mail: baumgaertner@cs.tu-darmstadt.de • S. Katzenbeisser is with Faculty of Computer Science and Mathematics, University of Passau, Germany. E-mail: Stefan.Katzenbeisser@uni-passau.de • B. Freisleben is with Department of Mathematics & Computer Science, Philipps-Universit¨at Marburg, Germany. E-mail: freisleb@informatik.uni-marburg.de y For this purpose, RESCUE pursues a twofold mitigation technique. First, certiﬁcates are used to cryptographically bind users to network identiﬁers, which hinders the ad- versary from assuming multiple identities. Since traditional static certiﬁcate infrastructures may be unavailable in the disaster area, we propose a novel distributed approach that enables new users to obtain certiﬁcates from mobile authorities during crisis. Second, RESCUE applies a novel buffer management scheme called source-elastic buckets (SEB) This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing 2 that substantially increases message delivery rates, i. e., availability, in the presence of ﬂooding attacks, both from individuals and multiple identities. At its core, SEB isolates authenticated messages and allocates buffer capacity fairly to all source nodes, effectively mitigating ﬂooding attacks from individual nodes. We presented SEB’s initial design in [7]. In this work, we extend SEB with priority sets to provide protection against Sybil attackers as well. As SEB relies on node-local decisions rather than a (complex) dis- tributed protocol, it provides a minimal surface to attacks and causes no network overhead by design. In addition, instead of identifying and excluding misbehaving users from the network, our scheme provides a fair allocation of available resources to all users. Hence, RESCUE does not suffer from false positives, where a valid user is mistakenly excluded from the emergency communication. We make the following contributions: appropriate actions to exclude them from the network. One approach exploits the users’ social networks [14], [15]. 3 SYSTEM MODEL Communication Model. We support a wide range of com- munication models that are reasonable during emergencies, including one-to-one (contact with friends or family), many- to-many (within task forces or departments), or one-to-many (emergency notiﬁcation broadcasts). Due to the inherent delay of DTN-based communication and our focus on emer- gency communication, we consider mainly small messages, such as text and distress messages (including additional information, such as GPS location of the sender), serving a similar purpose as the classic 112 or 911 emergency call. Compared to rich media (images, voice, video), information in text messages is more compact, thus, more suitable for DTN communication. • An evaluation of RESCUE using large-scale network simulations in both, synthetic and realistic disaster scenarios, demonstrating that RESCUE maintains very good delivery rates even under attack (Section 8). Furthermore, Section 2 reviews related work, Section 3 de- picts our system model, Section 4 describes our minimalistic communication protocol, and Section 9 concludes the paper and outlines areas for future work. 1 INTRODUCTION How- ever, this requires communication between peers, which is feasible for online peer-to-peer systems but not for DTN sce- narios. In [16], Sybil identities are detected at direct neighbor nodes. This approach is suitable for proximity services, but not for DTNs where Sybil nodes might be multiple hops away. In [17], nodes bootstrap trust relationships randomly and then collaboratively ﬁlter bogus messages. In contrast to existing work, we do not try to identify attackers but include their presence in our protocol design. Secure Routing. Encounter-based routing in DTNs is used to intelligently select forwarding nodes based on their con- tact history, which works well assuming repetitive mobility patterns. At the same time, it makes the network suscepti- ble to blackhole attacks, where an attacker lies about past encounters to appear as a strong forwarder. Previous works have proposed to use signed encounter tickets that are exchanged upon contact [18], [19], [20], [21]. Unfortunately, exchanging and verifying these tickets introduces commu- nication and computational overhead. We use epidemic rout- ing to thwart all routing attacks and mitigate the problems of increased message replicas using effective buffer management and prioritization. • A mobile distributed certiﬁcate infrastructure tailored to disaster scenarios that hinders an adversary from assuming multiple identities to perform Sybil attacks (Section 5). • A fair buffer management scheme that mitigates the effect of ﬂooding attacks by individuals (Section 6). • An extension to our buffer management scheme by pri- ority sets that increases resilience against Sybil attacks, guarantees at least the performance of direct message delivery, and supports unregistered users without a certiﬁcate (Section 7). 2 RELATED WORK Since our communication framework can withstand a vari- ety of attacks (see Table 1), it supports and complements several existing secure opportunistic communication sys- tems [8], [9], [10]. In particular, it complements a recently proposed framework for anonymous routing in DTNs [11]. We now review related work on network attacks and possi- ble countermeasures. Adversary Model. We consider an adversary Adv who can mount network attacks and compromise network entities. Speciﬁcally, Adv can eavesdrop, manipulate, forge, or drop messages. Furthermore, Adv can assume a limited number of entities, either by compromising or stealing devices or by registering multiple times in our system. Unlike the classic Dolev–Yao adversary model, Adv controls only a part of the communication channel and a fraction of all network entities. Moreover, Adv cannot break cryptographic primitives or tamper with the root authority (see Section 5.1). In Table 1, we summarize well-known attacks [4] that Adv can mount, and list RESCUE’s countermeasures to prevent them. Flooding Attacks in DTNs. Denial-of-Service (DoS) attacks on unauthenticated DTNs have been discussed in the liter- ature, but contrary to previous ﬁndings [12], we show that authentication is essential for reliable operation (Section 8). In authenticated networks, [5] proposes to enforce rate limits hard-coded in certiﬁcates, using an active distributed pro- tocol. Nodes exceeding their rate limit are blacklisted and excluded from the network. [13] proposes a similar scheme which also allows for bursty trafﬁc. In contrast to all previous works, we implicitly solve the problem of ﬂooding attacks using a fair and elastic buffer management scheme which has the beneﬁts of not requiring any pre-deﬁned (and possibly arbitrary) limits and avoiding additional overhead in form of encounter records. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ 4.3 Authentic Acknowledgments RESCUE uses acknowledgments (ACKs) for one-to-one com- munication. Previous work [22] has shown that epidemic routing greatly beneﬁts from ACKs since they free up buffer capacity for other messages. Upon receiving a message, the destination creates an ACK as a reply and forwards it with the same mechanism used for relaying regular messages. The ACK contains only the message ID m and a signature from the destination σd: ACK = (m, σd). (3) (3) Upon receiving and verifying an ACK, intermediate nodes can safely remove the acknowledged message payload from their buffers. The ACK is stored until the corresponding message has expired. Attackers cannot forge ACKs, since they are cryptographically signed and, hence, cannot purge valid undelivered messages from the network. Since ACKs are small, they present a potential attack vector: by creating a large number of bogus ACKs, an attacker can exhaust the computational resources of the receiving nodes, because they have to verify each signature, leading to a DoS. To solve this problem, nodes in RESCUE only accept and process ACKs for messages they currently carry. This stops the spreading of bogus ACKs at the ﬁrst valid node. 4.1 Epidemic Routing TTL = t + ∆t −tnow, (2) (2) Instead of relying on infrastructure, DTN-enabled devices exchange messages directly using Wi-Fi or Bluetooth. DTNs exploit user mobility to increase coverage. To this end, devices act as “data mules” that store their messages as well as messages from other users, carry them, and ﬁnally forward them to the destination upon contact. When a device is in communication range of another device, both devices replicate and transmit all messages not yet received by the other device. Therefore, every message is ﬂooded to every node that comes into contact, spreading like an epidemic. We use epidemic routing [22] because there are no routing control messages, thus, mitigating all types of routing manipulation attacks. In addition, message drop- ping attacks have no effect, since messages are replicated to all available neighbors. Carried messages are stored in the node’s buffer that we protect against ﬂooding and Sybil attacks, as detailed in Sections 6 and 7, respectively. When two devices discover each other via Bluetooth or Wi-Fi beacon frames, they initiate a handshake. As part of the handshake, both devices ﬁrst exchange metadata about carried messages and then start transferring messages that the other device is missing. However, due to limited buffer capacity and short contact times (e. g., two cars passing each other), not all messages might be exchanged. A message pri- oritization scheme (Section 6.3) determines which messages are exchanged ﬁrst upon contact. where tnow is the current time. Nodes regularly remove expired messages (negative TTL) from their buffers. where tnow is the current time. Nodes regularly remove expired messages (negative TTL) from their buffers. 4 MINIMALISTIC COMMUNICATION PROTOCOL This section describes RESCUE’s communication protocol, i.e., its routing protocol, message format, and acknowledg- ment processing. By employing a simple routing mechanism and a minimalistic frame format, RESCUE is immune to a large set of common attacks on DTN protocols (see Table 1). Sybil Attacks in Peer-to-Peer and Mobile Networks. Pre- vious works try to identify Sybil identities and then take This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing epted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing 3 Table 1: Attack Resilience of RESCUE ATTACK COUNTERMEASURE SECTION Routing manipulation ) Epidemic routing 4.1 Message dropping Blackholing Message modiﬁcation o Authentic immutable messages 4.2 Impersonation ACK ﬂooding ACKs only for messages in buffer 4.3 Sybil attack n User registration 5.3 Priority sets 7.3 Message ﬂooding Source-based elastic buckets 6.2 TTL spooﬁng Source-based elastic buckets 6.2 Table 1: Attack Resilience of RESCUE Devices verify messages at each hop by checking the mes- sage signature and, if available, the source’s identity cer- tiﬁcate. Devices discard messages if a check fails, so that corrupted messages do not propagate in the network. We further deﬁne the message ID m as a hash h over all message ﬁelds: m = h(s, d, t, ∆t, P). The signature σs is then calculated on m. We note that all header ﬁelds are immutable, that is, they are not changed in transit, which would be required for time-to-live (TTL) ﬁelds. Immutable ﬁelds allow the signature to protect the entire message and, thus, they prevent all message modiﬁcation and imperson- ation attacks. Assuming that the clocks of all valid nodes are roughly synchronized,1 the TTL of MSG can be locally computed by each node with: 4.2 Authentic Immutable Messages Each user possesses a unique signature key pair generated during initialization. The public signature key serves as a unique addressable network identiﬁer similar to an IP ad- dress. The private signature key is used to sign outgoing messages. A message MSG contains the source network identiﬁer s, the signature σs, and, if available, the identity certiﬁcate C. The identity certiﬁcate is issued by a certﬁcate authority and contains the network identiﬁer as well as an identiﬁcation token T that we explain in Section 5. Identity certiﬁcates can be cached and only transmitted on demand to reduce overhead. In addition, a message MSG contains the destination network identiﬁer d (the public signature key of the destination), the creation time stamp t, the message lifetime ∆t, and an optionally encrypted payload P, resulting in the tuple: This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ 4.4 Storage Overhead For each message, a node stores meta information m, s, d, t, and ∆t as well as P and σs in its buffer. After receiving and verifying an ACK, the node deletes P and σs and replaces them with σd. In comparison with an insecure scheme, the storage overhead is the signature σs or σd per message. Additionally, a node stores the identity certiﬁcate C for each other node s that it carries messages for. If it no longer carries messages for s, a node can decide to delete C or keep it to speed up future transfers for messages from s. 1. Mobile nodes can synchronize their clocks via GPS or a cellular network. Even if synchronization opportunities are no longer available after a disaster, we can neglect small-scale clock drift that might occur since TTL is in the order of hours. (1) MSG = (s, d, t, ∆t, P, σs, C). This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing his article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing 4 Figure 1: Illustration of our mobile distributed certiﬁcate infrastructure. The authorization level (number of stars) decreases from left to right. Registered users have a higher authorization level than unregistered users. Citizens are depicted in a box with a separate authorization level. Figure 1: Illustration of our mobile distributed certiﬁcate infrastructure. The authorization level (number of stars) decreases from left to right. Registered users have a higher authorization level than unregistered users. Citizens are depicted in a box with a separate authorization level. 5 IN-THE-FIELD USER REGISTRATION WITH MOBILE AUTHORITIES Identity certiﬁcates bind the public signing keys of users, which function as their unique network identiﬁers (see Section 4), to user properties. Important properties in the emergency setting are the afﬁliation (e. g., UN, red cross), user role (e. g., citizen, physician) and authorization level, which indicate a user’s permission level and trustworthiness. Fig- ure 1 exempliﬁes the identity certiﬁcate of a UN team leader, and depicts the authorization level of entities by their position on the x-axis as well as stars in the certiﬁcate. To establish trust relationships, we deploy so-called identity certiﬁcates that bind important properties in the emergency context (e. g., user role or afﬁliation) to the network identi- ﬁer of users. In this section, we ﬁrst describe our backbone certiﬁcate infrastructure. Next, we extend the backbone in- frastructure by mobile authorities, enabling their operation during disasters in the ﬁeld, where the backbone infras- tructure is unavailable. Finally, we propose multiple user identity veriﬁcation methods that hamper fake registrations with the certiﬁcate infrastructure. This way, an adversary is prevented from obtaining multiple identities, i.e., cer- tiﬁed network identiﬁers, that could be used to perform distributed DoS attacks. Fig. 1 illustrates our certiﬁcate infrastructure. p We further consider certiﬁcate revocation to defend against an adversary who obtains identity certiﬁcates by compromising user devices or inﬁltrating authorities. Other reasons for certiﬁcate revocation are, for instance, struc- tural changes within organizations or users who quit or- ganizations. RESCUE implements certiﬁcate revocation via certiﬁcate revocation lists (CRLs) that are broadcasted in the network. We distinguish between two different entities: authorities and users. An authority A can revoke an entity E if A has a higher authorization level than E, and there is a certiﬁcate chain (i.e., a chain of trust) between A and E. Upon the revocation of an authority, all certiﬁcates issued by the authority are regarded as invalid, depriving its power. In case a user identity certiﬁcate is revoked, the certiﬁcate is considered invalid and the respective user loses his or her role, authorization level, and any message transmission privileges (see Section 7). 5.1 Static Authorities The backbone certiﬁcate infrastructure consists of multiple hierarchically organized static Certiﬁcate Authorities (CAs). The root of these CAs constitutes a dedicated authority named Root Authority (RA) that serves as a trust anchor and its certiﬁcate is pre-deployed on all RESCUE-enabled devices. Before the actual crisis, the RA establishes rela- tionships with organizations or governments that want to participate as authorities in the certiﬁcate infrastructure. All authorities initially undergo a rigorous audit, since their authenticity and trustworthiness are crucial to the overall security, and negotiate user roles as well as preconﬁgured user groups that the authority introduces to the network. For instance, in Fig. 1, the United Nations (UN) added the user roles team leader and ofﬁcial, and arranged a preconﬁgured user group United Nations, so users can speciﬁcally address all UN members when sending a message. Organizations manage their own certiﬁcate infrastructure and, therefore, maintain one or multiple, potentially hierarchically orga- nized, CAs. On the lowest hierarchical CA level, CAs issue identity certiﬁcates to staff members. Furthermore, the over- all infrastructure contains at least one authority that issues identity certiﬁcates to regular users, i.e., citizens. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ 5.2 Mobile Authorities In a disaster area, infrastructure-based communication is mostly unavailable and, thus, users rarely have a connection to the static authorities in the backbone infrastructure. This is not an issue when established user properties are re- trieved, since identity certiﬁcates can be veriﬁed and trans- mitted between users on demand (see Section 4). Neverthe- less, operations that inevitably involve CAs, such as issuing new certiﬁcates or revoking existing certiﬁcates, cannot be performed when static authorities are unavailable. Therefore, we propose that special privileged users em- ploy their mobile devices to serve as mobile authorities (MAs) during a crisis. Since it is easier for an adversary to com- promise MAs than static authorities, MAs have restricted This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing 5 discuss later. To enforce a method-agnostic format for T, we apply a globally known hash function in the following way: discuss later. To enforce a method-agnostic format for T, we apply a globally known hash function in the following way: capabilities. In detail, they can only issue identity certiﬁcates for citizens, but not for speciﬁc user groups like red cross staff members. We argue that this is not a restriction, since professional emergency workers typically set up their sys- tems prior to the disaster or outside the disaster area, where a connection to static authorities is available. Additionally, MAs are allowed to revoke identity certiﬁcates. An MA M can revoke a user U, if U is a citizen, or if U has a lower authorization level than M and both belong to the same afﬁliation (Fig. 1). T = hash (“Identiﬁcation Method” \|\| “Identiﬁer”) , (4) (4) where \|\| is the concatenation operand. For example, an iden- tiﬁcation token resulting for a SIM-based registration would be calculated as TSIM = hash(“SIM” \|\| “+491234567890”). This approach effectively and elegantly thwarts adversaries that attempt to receive multiple identities by registering with different MAs. Since the identiﬁcation token for the same device and method is constant, all issued identity certiﬁcates would contain the same identiﬁcation token and could, thus, be mapped to the same identity. 5.3 Secure Identity Veriﬁcation Methods In the following, we present methods that enable static and mobile authorities to identify registering entities based on hard-to-forge identiﬁcation tokens. The identiﬁcation to- ken and method are part of the identity certiﬁcate and effectively hinder individuals from registering repeatedly and obtaining multiple identity certiﬁcates. Our goal is to increase the cost for fake registrations, such that bypassing our subsequently presented ﬂooding and Sybil mitigations (see Sections 6 and 7) becomes uneconomical for an ad- versary. We assume that organizations and governments are already able to supply each staff member with exactly one identity certiﬁcate, e. g., by handing out preconﬁgured devices. Therefore, we focus on fake registrations of users with the authorization level ”citizen”. Citizens that employ stronger authentication methods during registration with CAs are considered more trustworthy, indicated by a higher authorization level in their identity certiﬁcate. Messages from users with high authorization levels are transmitted preferentially (see Section 6) to encourage citizens to (i) obtain identity certiﬁcates, and (ii) use strong identiﬁcation methods during registration. Typically, as summarized in Table 2, the stronger the identity proof is, the more restric- tive (i. e., less applicable) and time-consuming the veriﬁca- tion process gets, resulting in limited practical usability. In the following, we ﬁrst deﬁne identiﬁcation tokens and then discuss the identiﬁcation methods in detail. Identiﬁcation via SIM. The subscriber identity module (SIM) card is used as an identiﬁcation token by requiring the user to enter a nonce that is sent via a call or SMS to the user’s device during registration. Upon successful registration, the authority creates the identiﬁcation token T using the user’s phone number as the identiﬁer. The ap- proach provides excellent usability and applicability, since it requires the user to take a minimum effort and SIM cards are available in many mobile devices. Nevertheless, an adversary can create fake users by using anonymous prepaid SIM cards. Also, since the approach requires a functioning cellular network, it is unsuitable for registering new users during a crisis. Identiﬁcation via Remote Attestation. Authorities can per- form a remote attestation [24] with devices of registering users. This way, authorities obtain an attestation report that is signed with a device-unique secret attestation key and a certiﬁcate that testiﬁes the validity of the attesta- tion key. The public attestation key serves as the identiﬁer for the identiﬁcation token. 5.2 Mobile Authorities Since malicious MAs can seriously harm the network, only privileged and trustworthy users with devices that satisfy certain security requirements are permitted to be- come MAs. In the initialization phase, each organization negotiates the maximum permitted number of MAs they introduce to the network, and then carefully selects those users qualifying to become MAs. MA users have a high social trust level and protect their devices using security mechanisms like a trusted execution environment, full disk encryption, and strong passwords. These mechanisms have shown to signiﬁcantly increase the effort for physical at- tacks [23], giving MA users enough time to report and revoke stolen MA devices. Identiﬁcation via Physical Presence. This method con- stitutes a fallback solution only used by MAs during a crisis if all other veriﬁcation methods are inapplicable. For this method, the identiﬁcation token is generated from the devices’s network identiﬁer. While the identiﬁer is not hard to forge (an adversary could easily generate several public keys to assume multiple identities), the identiﬁca- tion method requires a user to physically approach an au- thority during registration, and thus spend physical effort. By spending this effort, we naturally limit the number of identity certiﬁcates that an adversary could receive using different network identiﬁers. MAs assert physical proximity of users by employing short range communication channels (e. g., QR codes, NFC, or Bluetooth) to transmit identity cer- tiﬁcates. Furthermore, MAs manually conﬁrm the issuance of each identity certiﬁcate to prevent an adversary from obtaining multiple certiﬁcates at once. The method has a weak identiﬁcation strength, since an attacker can simply approach different MAs or, with some delay, the same MA repeatedly. Also, usability is poor, since users cannot register remotely. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ 6.2 Source-based Elastic Buckets (SEB) Identiﬁcation via eIDs. This method uses national elec- tronic ID cards, which often provide identiﬁcation capabili- ties, to identify a user. As an example, the eIDAS regulation deﬁnes electronic identiﬁcation services in the entire Euro- pean Union [32]. eIDAS speciﬁes the restricted identiﬁcation (RI) protocol, e. g., implemented in the German identity card since 2010. RI allows a service provider (SP) terminal to rec- ognize an eID chip based on a chip-unique pseudonym. The chip-unique pseudonym acts as the identiﬁer for the identi- ﬁcation token. Using mobile devices as local terminals [33], authorities can act as SP terminals and securely identify eID cards of registering users. The approach provides a strong proof of identity as it is hard to forge eID cards or to obtain multiple valid eID cards including their PIN. Since MAs can in principle act as SP terminals, the approach is applicable during a crisis. As a downside, users must initially activate their eID cards and have them at hand. We now present our novel buffer management strategy Source-based Elastic Buckets (SEB) that, by design, prevents valid messages from being purged from the network during ﬂooding attacks. The basic idea is that all messages from a source s are placed in an isolated bucket Bs. All buckets are stored as a map B that uses identiﬁcation tokens as its key. Bs = B(Ts). (6) (6) Since RESCUE uses identiﬁcation tokens as authenticated identiﬁers and ensures message authenticity through digital signatures, an adversary cannot forge messages in a way that they occupy buckets of valid users. SEB is fair in the sense that each bucket Bs has a guaranteed capacity of Cn = C/n, with n being the number of currently allocated buckets (number of source nodes that a respective node currently carries messages from). The occupancy of a single source bucket O(b) is a non-negative number and is subject to X b∈B O(b) ≤ X b∈B Cn = C. (7) 5.3 Secure Identity Veriﬁcation Methods To date, applicability is good, as recent Samsung [25], Windows [26], and Android [24] devices provide remote attestation capabilities. Addition- ally, remote attestation will become increasingly widespread with upcoming technologies [27], [28] and MAs could act as veriﬁers and thus identify new users during a crisis. Usually, a backbone in the form of stable connections, e. g., Identiﬁcation Tokens. RESCUE uses identiﬁcation tokens to uniquely identify network nodes and implement its ﬂood- ing protection mechanism (see Section 6). An identiﬁcation token T is calculated based on an identiﬁer speciﬁc to one of the identiﬁcation methods. For example, the identiﬁer for the SIM-based identiﬁcation method would be the phone number. Other methods use different identiﬁers, which we This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing 6 Table 2: Overview of identity veriﬁcation methods. Ratings scale from poor (⋆) to excellent (⋆⋆⋆⋆⋆). Table 2: Overview of identity veriﬁcation methods. Ratings scale from poor (⋆) to excellent (⋆⋆⋆⋆⋆). Identiﬁcation Method Strength Usability Applicability Available in Crisis Physical Presence ⋆ ⋆⋆ ⋆⋆⋆⋆⋆ SIM ⋆⋆⋆ ⋆⋆⋆⋆⋆ ⋆⋆⋆⋆ Remote Attestation ⋆⋆⋆⋆ ⋆⋆⋆⋆⋆ ⋆⋆⋆ eID ⋆⋆⋆⋆⋆ ⋆⋆⋆ ⋆⋆ 6.1 Security Requirements and Design Poor buffer management schemes can expose a network to ﬂooding attacks. For example, malicious nodes can exploit trivial schemes such as ﬁrst-in ﬁrst-out (FIFO) queues to replace valid messages with bogus ones [34]. The goals of our buffer management scheme are: (i) a single attacker can only occupy a “fair share” (we formally discuss this in sec:seb:basic) of available buffer space; (ii) maximization of buffer utilization to increase message redundancy and, thus, the delivery rate; and (iii) a MA compromise does not compromise the network (Sybil attacks are discussed in Section 7). To reach these goals, we apply a locality principle [35] to allow nodes to decide locally and independently which messages to store. Hence, nodes do not need to trust and verify third-party information, which keeps the attack surface small. Furthermore, bandwidth efﬁciency is increased, since control messages need not be exchanged. cable or satellite uplink, between the different systems that are responsible for the remote attestation is needed. For increased resilience and to keep functioning even under severely challenging network conditions, it is desirable to not rely on such a backbone but explore new decentralized and federated solutions [29], [30], [31]. In practice, though, remote attestation without backbone access has not yet been implemented. cable or satellite uplink, between the different systems that are responsible for the remote attestation is needed. For increased resilience and to keep functioning even under severely challenging network conditions, it is desirable to not rely on such a backbone but explore new decentralized and federated solutions [29], [30], [31]. In practice, though, remote attestation without backbone access has not yet been implemented. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ 7.1 Secure Copies In direct delivery (DD) forwarding, nodes do not carry messages for others, but only deliver their own messages when they actually encounter the destination. Obviously, this diminishes the advantage of having “data mules,” but DD has a desirable security property: it is inherently immune to ﬂooding attacks even from Sybil attackers since each node simply does not carry messages for other nodes. In other words, DD ensures that there is always one copy of every message in the network, namely in the buffer of the source node. We call this a secure copy, i. e., a message copy that an attacker cannot remove or replace. In the case of DD, the number of secure copies per message is exactly one. Though this is a very simple strategy, other buffer management schemes fail to achieve this guarantee. For example, when using FIFO, the node’s own messages might be replaced by more recently received messages of other nodes. In the following, we increase the number of secure copies to improve delivery reliability. 7 LOCAL PRIORITY SETS Until now, we have assumed that MAs behave correctly and cannot be compromised by an adversary. We recognize that this is a strong assumption, since MA devices may be stolen or infected with malware. However, if we lift our assump- tion on secure MAs, our buffer management presented in Section 6 is vulnerable to Sybil attacks. This is because an adversary that gets hold of an MA can generate as many certiﬁcates as it wants. Since SEB allocates buffer resources fairly among all nodes, a Sybil attacker would receive an unfair amount of buffer space. In this section, we secure RESCUE even against such Sybil attackers by leveraging the concept of secure message copies using priority sets where nodes prioritize messages originating from a certain set of other nodes in the network. g 7: B′ := B′ \ MSG′ {Drop MSG′ from bucket} 8: end while 7: B′ := B′ \ MSG′ {Drop MSG′ from bucket} 8: end while rank function in Section 6.3. Ties are broken at random if there are two or more elements with the smallest message rank. To assert that the buffer converges to a stable state, tie breaking needs to be consistent, i. e., the same tie needs to be broken consistently at a single node. We implement this by comparing the salted hashes [36] of node’s identiﬁcation tokens, while the salt is drawn at random once by each node. A tie is then broken by the smaller hash value. SEB’s robustness relies on the fact that messages are source-authenticated, and on the high costs of registering multiple identities in RESCUE. Without the latter costs, an attacker could assume multiple identities, ﬂood the network with messages and, thus, hijack a disproportional amount of buffer capacity. In addition, SEB mitigates TTL spooﬁng attacks where an attacker sets excessively high values for ∆t to maximize the lifetime of its messages: by separating messages of different sources, an attacker would only be able to replace its own messages. In the following, we ﬁrst introduce the concept of secure copies, explain the workings of priority sets, and discuss Sybil-secure ﬁll strategies. Finally, we explain how priority sets also help to securely support unregistered users. Algorithm 1 Message Insertion using SEB Algorithm 1 Message Insertion using SEB Input: MSG {The message to be inserted} Input: B {The set of all buckets} Input: C {Total capacity} 1: Ts := source’s identiﬁcation token from MSG 2: Bs := B(Ts) {Select source bucket} 3: Bs := Bs ∪MSG {Add new MSG to the source bucket} 4: while P b∈B O(b) > C do 5: B′ := arg minb∈BS(b) {Bucket with smallest surplus} 6: MSG′ := arg minm∈B′MR(m) {message with the lowest message rank} 7: B′ := B′ \ MSG′ {Drop MSG′ from bucket} 8: end while Input: MSG {The message to be inserted} Input: B {The set of all buckets} Input: C {Total capacity} 6 LOCAL BUFFER MANAGEMENT (7) Within the buffer, a node stores unacknowledged and un- expired messages. If there are many such messages, a node might not have the resources to store them all: at some point, it needs to decide which messages to keep and which to drop. Given a set of messages and a buffer with a node- deﬁned capacity C, the buffer management has to decide which messages to store in the buffer without exceeding its capacity to enforce: We further deﬁne the surplus S(b) as the (possibly nega- tive) difference between the occupancy and the guaranteed capacity: We further deﬁne the surplus S(b) as the (possibly nega- tive) difference between the occupancy and the guaranteed capacity: S(b ∈B) = Cn −O(b). (8) (8) If s does not exhaust its guaranteed capacity (S(b) > 0), because it has not sent “enough” messages, S(b) is provided to other buckets requiring it, which means that their surplus can become negative. However, when s sends a message at a later point, overdrawn buckets (S(b) < 0) are emptied ﬁrst. These elastic quotas allow full exploitation of the local buffer capacity, while maintaining strict message separation of dif- ferent sources. Algorithm 1 shows SEB’s message insertion procedure. We deﬁne the argument of the minimum func- tion as arg minx∈Sf(x) = {x ∈S : f(x) = miny∈S f(y)}. Algorithm 1 ﬁrst inserts a new message in the correspond- ing source bucket Bs (l. 3). Until the total occupancy meets C to satisfy Eq. (5) (l. 4), the algorithm drops messages (l. 7) from the bucket with the smallest surplus (l. 5) in order of the messages’ ranks (l. 6). We expain the message X MSG \|MSG\| ≤C. (5) (5) Besides this hard constraint, buffer management can have multiple optimization goals, e. g., throughput maximization, delay minimization, or delivery reliability. In this work, we are concerned with security, in particular resilience against DoS attacks. To this end, we ﬁrst motivate the need for secu- rity mechanisms in the buffer management, deﬁne security requirements, and then present a novel secure buffer man- agement strategy, Source-based Elastic Buckets (SEB), which achieves protection against ﬂooding attacks. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. 6.3 Multi-Factor Message Rank Within each bucket, SEB uses Message Rank (MR) for priori- tization. MR prioritizes: (i) acknowledgments, (ii) messages with the largest TTL, and (iii) messages with the smallest payload size. Carrying messages with a large TTL increases the probability that they will be delivered before expira- tion (we conﬁrm this in Section 8), while small messages take less time for transmission, and help to prevent buffer fragmentation. Upon device contact, messages exchanged ﬁrst have a higher chance of actually being transmitted to the next hop and eventually reaching their destination. A sending node transfers messages in its buffer to a receiving node R in the following order: (i) messages destined for R, (ii) own messages, (iii) messages from other registered users, (iv) all other messages. Messages in each category are sorted by MR. MR only relies on ﬁelds in the message header. Since they are immutable, MR results in the same order independent of the order in which messages were received, which means that the buffers of two nodes will converge to a stable state if the contact duration is long enough. This is a problem that has been ignored by the research community and is reﬂected by the fact that the most popular network simulator for DTN research, ONE [37], only implements non-converging random and FIFO-based dropping strategies. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ 6 LOCAL BUFFER MANAGEMENT Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing Table 3: All priority sets used in RESCUE. LEVEL CONTAINED NODES PURPOSE SECTION 0 Local (own) DD lower bound 7.2 1 Social net./by MA Sybil protection 7.3 2 Other registered Flooding protection 6.2 3 Unregistered Best effort 7.4 Table 3: All priority sets used in RESCUE. Algorithm 1 Message Insertion using SEB 7.2 Priority Sets Overview We apply the idea of secure copies to SEB to mitigate Sybil at- tacks. In particular, we propose to prioritize certain buckets in SEB (e. g., the node’s own bucket) such that they are emp- tied last when the buffer capacity is exceeded. We model the general assignment of node identities (and their buckets) to priorities via priority sets (PS). How we implement the assignment of node identities to PS, i. e., deciding which nodes’ messages should be prioritized, is crucial for the system’s security and is the core question that we address This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing epted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing 8 in this section. Making individual local decisions makes it hard for an attacker to appear in all PS, thus preventing that their messages ﬁll the buffers of all other nodes. By using PS, each node essentially gains a number of secure relays that prioritize messages for it, effectively increasing the number of secure copies. Next, we show how the PS concept integrates with SEB (Section 6), and discuss how PS can be used to protect against Sybil attacks in Section 7.3. via a central server similar to secure messaging applications (e. g., Signal). For all users that registered during a disaster with an MA, a different method is required. Post-registration: MA-assigned. We assume that most users will only register post-disaster and, thus, cannot use social contacts to ﬁll their PS. However, we can leverage the trustworthiness of MAs by letting MAs suggest identities for the priority set during registration. We propose that the suggested identities are those that recently registered with the MA.2 Since an attacker is not able to manipulate the PS of nodes that registered with the MA before be- ing compromised, only nodes that registered after an MA compromise may be affected by PS manipulations.3 The latter may experience a decreased service quality since their identities will not appear in the PS of other nodes. 7.4 Supporting Unregistered Users Apart from Sybil attacks, PS enable us to solve another remaining problem: secure support for unregistered nodes. To this end, we simply introduce another PS with a priority level higher than the one used in Section 7.3. This essentially assigns all remaining buffer capacity to unregistered nodes. These nodes will consequently receive the lowest quality- of-service level since their messages will be dropped ﬁrst. However, we show later that in case the network is not fully congested (i. e., not during a ﬂooding attack), unregistered users receive a quality-of-service level similar to that of reg- istered users. Even when the network is under a ﬂooding at- tack, performance never drops below the DD lower bound. For unregistered nodes, the identiﬁcation token Ts required for Algorithm 1 is derived from the network identiﬁer. 7.2 Priority Sets Overview This effectively reduces the number of secure copies for their messages to one, which is the same for other unregistered users. Other than that, they are not affected negatively by registering with a compromised MA. To integrate PS in SEB, we need to ensure that buckets of nodes in a PS are emptied last. We generalize this approach by allowing an arbitrary number of PS levels l. For the remainder of this work, we denote Sl as the priority set at level l. The set with the lowest level has the highest priority. Formally, Sl are pairwise disjoint subsets of B: [ l=0,... Sl = B and Sl ∩Sk = ∅, l ̸= k [ l=0,... Sl = B and Sl ∩Sk = ∅, l ̸= k To use priority sets with SEB, we adapt Algorithm 1 to start removing messages from the buckets with the lowest priority. In particular, we change Line 5 to ﬁrst select the non-empty PS with the lowest priority, i. e., largest level l. Then, we select the bucket with the smallest surplus as follows: For both strategies, we need to determine the size of S1 (number of contacts in the users’ address books and size of MA-suggested list, respectively) in order to be effective against Sybil attackers. In Section 8, we empirically show that a small PS size is sufﬁcient to withstand Sybil attacks. 5a: l′ := maxl {l : Sl ̸= ∅} 5b: B′ := arg minb∈Sl′ S(b) 5a: l′ := maxl {l : Sl ̸= ∅} 5a: l′ := maxl {l : Sl ̸= ∅} 5b: B′ := arg minb∈Sl′ S(b) To achieve the performance of DD (at least one secure copy per message), S0 only includes the local node, such that messages of the local node are removed last. We explain more PS levels in the following sections and we summarize all PS levels used in RESCUE in Table 3. 2. An MA will consequently not assign any identities to the S1 set of the ﬁrst user that registers. To also assign identities to the ﬁrst users, an MA could use a set of known identities that registered pre-disaster. 7.3 A Sybil-secure Priority Set How to select the nodes that are to be put in S1 is key to achieving protection against Sybil attacks. The selection strategy needs to ensure that the nodes residing within each set are (preferably) different for each node and it is hard for the Sybil attacker to become a member of many of those sets. In the following, we discuss two PS ﬁll strategies that are secure against Sybil attacks and can be practically used in emergency scenarios. The members of this set are selected either (i) by exploiting social relationships between the nodes, e. g., by leveraging phone numbers in the address book of the users’ smart phones, or (ii) by letting the MA assign the set during registration. We now present both approaches in detail. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ g p 3. Using this method, an adversary may be present in the PS of legitimate nodes by registering at an MA during disasters. However, the adversary may only register few identities this way, since it is hard to register multiple times at an MA (see Section 5.3). 8.1 Scenarios We consider two scenarios as detailed in Table 4: a synthetic scenario to isolate the effect of two distinct attacks on the network (Sections 8.2 and 8.3) and a Typhoon Haiyan sce- nario to assess performance under more realistic conditions (Section 8.4). In both scenarios, we consider three different node classes: pre-registered, post-registered, and unregistered. During the course of the simulation, all post-registered nodes start unregistered, and become registered until the end of the simulation linearly over time. All nodes listed as un- registered in Table 4 remain unregistered. For simplicity, we only evaluate two authorization levels (Section 5): registered and unregistered. In the Typhoon Haiyan scenario, we have additional roles such as injured and healthy citizens, urban search-and-rescue teams (USRT), and UN ofﬁcials, which all have distinct mobility patterns (refer to [38] for details). We use epidemic routing and compare four different buffer management strategies: (FIFO) uses a ﬁrst-in ﬁrst-out queue as prioritization and drop strategy, (MR) uses Message Rank instead of a FIFO queue, (SEB) employs our Source-based Elastic Buckets and uses all priority set levels except for S1, and (PS) makes use of all priority set levels. For the eval- uation of the Sybil attack scenario, we additionally include the direct delivery (DD) buffer management strategy as a benchmark. We choose a small buffer capacity to exaggerate the effect of the different buffer management strategies. The message size is ﬁxed to avoid fragmentation effects in the buffers. The simulation parameters in the Typhoon Haiyan scenario were chosen in accordance to Stute et al. [38] for comparison reasons. We use the ONE simulator v1.6.0 [37] for our experiments and, unless stated otherwise, show the average over ten runs with different seeds. Importance of ACKs. In a benign setting, ACKs help to keep buffers clean (SEB and MR). Once a message is delivered (Fig. 2a), the number of copies in the network reduces about as quickly as they increased (Fig. 3a) yielding perfect (i. e., 100 %) delivery rates. Importance of ACKs. In a benign setting, ACKs help to keep buffers clean (SEB and MR). Once a message is delivered (Fig. 2a), the number of copies in the network reduces about as quickly as they increased (Fig. 3a) yielding perfect (i. e., 100 %) delivery rates. FIFO vs. MR. 8 EXPERIMENTAL EVALUATION Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing 9 and focus on evaluating RESCUE’s novel mechanisms to protect against ﬂooding and Sybil attacks. injecting bogus messages at a high rate in the synthetic scenario. The attackers maximize message lifetime of their messages: they address their messages to nonexistent desti- nations, so that acknowledgments are never returned; and they set the message lifetime to a value that is larger than the simulation time to keep their messages persistent unless they are dropped by the buffer management. Valid users choose the destination randomly among all other valid users. Figure 3 shows the overall delivery rate and delay for different attacker injection rates Rϵ as a function of the valid users’ injection rate Rn. Note that Rϵ and Rn are aggregate rates, e. g., Rϵ = 1Rn means that all attackers inject as many messages as all valid users combined. To better understand the results, Fig. 3 differentiates between registered and un- registered users, and includes the network-wide copies per message during the attack. We make multiple observations. 8.1 Scenarios Using FIFO as a buffer management strategy does not yield satisfactory results even in a benign sce- nario because buffer states do not converge: FIFO always accepts an incoming message even if it has previously been dropped. Performance further decreases as the attackers’ injection rate is increased (Fig. 2c). On the other hand, prior- itizing messages by deadline is apparently a very effective metric to achieve 100 % message delivery in less than 1 hour (MR in Fig. 2a) but, at the same time, is tremendously susceptible to ﬂooding attacks, as it uses the TTL for prioriti- zation that the attacker manipulates by setting an arbitrarily large message lifetime, thus, reducing the message delivery rate to about 5 % (Fig. 2c). This occurs since the attacker is able to quickly remove virtually all valid message copies in the network (Fig. 3b). 8.2 Flooding Attack SEB Mitigates Flooding Attacks for All Registered Users. SEB uses MR as a secondary metric within each bucket. Therefore, SEB can achieve the same delivery rate as stan- dalone MR (plots in Fig. 2a overlap), but maintains a high delivery rate of more than 90 % even under the ﬂooding attack (Fig. 2c). In fact, as we vary the ﬂooding injection rate, performance only decreases for unregistered users (Fig. 2g), while all other groups (Figs. 2d to 2f) remain unaffected. Since SEB’s delivery rate did not change from Rϵ = 1Rn to 2Rn, we abstain from studying further increased Rϵ values. We ﬁrst evaluate the resilience of different buffer manage- ment strategies against a small number (5 %) of attackers Table 4: Simulation Settings Scenario Synthetic 2013 Typhoon Haiyan Mobility Map RWP (Helsinki) Natural Disaster [38] Speed 90 % ped., 10 % car 100 % pedestrian Duration 12 h (+5 h cool down) 168 h = 7 days Dimensions 4500 × 3400 m2 5000 × 7000 m2 Total Nodes 1000 500 (7 roles [38]) Unregistered — injured citizens Post-registered 800 healthy citizens Pre-registered 200 all others Message Rate Rn 0.1 s−1 (10 s interval) 0.1 s−1 (10 s interval) Message Size 25 KB 25 KB Message Lifetime 5 h 12 h Buffer Capacity 5 MB 20 MB Buffer Mgmt. FIFO, MR, SEB, PS, DD Routing Epidemic Radio Link Bluetooth (2 Mbit/s at 10 m range) Table 4: Simulation Settings Table 4: Simulation Settings Scenario Synthetic 2013 Typhoon Haiyan Mobility Map RWP (Helsinki) Natural Disaster [38] Speed 90 % ped., 10 % car 100 % pedestrian Duration 12 h (+5 h cool down) 168 h = 7 days Dimensions 4500 × 3400 m2 5000 × 7000 m2 Total Nodes 1000 500 (7 roles [38]) Unregistered — injured citizens Post-registered 800 healthy citizens Pre-registered 200 all others Message Rate Rn 0.1 s−1 (10 s interval) 0.1 s−1 (10 s interval) Message Size 25 KB 25 KB Message Lifetime 5 h 12 h Buffer Capacity 5 MB 20 MB Buffer Mgmt. FIFO, MR, SEB, PS, DD Routing Epidemic Radio Link Bluetooth (2 Mbit/s at 10 m range) This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ 8 EXPERIMENTAL EVALUATION In this section, we evaluate the behavior of our security mechanisms in large networks using the Opportunistic Net- work Environment (ONE) simulator [37] which is a well- accepted tool in the DTN research community. We ﬁrst de- scribe our evaluation scenario and present the performance results of RESCUE under ﬂooding and Sybil attacks. Finally, we repeat the experiments under an accurate realistic mo- bility model for large-scale natural disasters. RESCUE is resilient against several attacks, since it leverages concepts from related work that have been shown to be secure. There- fore, we refrain from studying those attacks via experiments Pre-registration: Social Networks. Social networks have been considered as a solution for effectively detecting Sybil identities [14]. While the particular method [14] is unprac- tical in DTNs (it needs to perform online veriﬁcations), we borrow the idea of using social networks as a defense against Sybil attacks. We exploit the user’s social network to prioritize messages from direct neighbors in the user’s social graph, e. g., those nodes that are in the local node’s address book. Since the attacker has no control over uncom- promised devices, they cannot forcibly add themselves to others’ address books, thus, they will not be able to appear in the PS of legitimate nodes. This option is only available for users that were able to register with an authority before the disaster, e. g., using their SIM card (Table 2), and were able to resolve the phone number to the RESCUE public key 2. An MA will consequently not assign any identities to the S1 set of the ﬁrst user that registers. To also assign identities to the ﬁrst users, an MA could use a set of known identities that registered pre-disaster. 3. Using this method, an adversary may be present in the PS of legitimate nodes by registering at an MA during disasters. However, the adversary may only register few identities this way, since it is hard to register multiple times at an MA (see Section 5.3). This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing epted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. 8.3 Sybil Attack In our next experiment, we evaluate the impact of a Sybil attacker on RESCUE, again in the synthetic scenario. The Sybil attacker is a single node (physical position) that can generate an unlimited number of registered identities, e. g., by compromising an MA. To cause maximum harm, the attacker executes a ﬂooding attack as in Section 8.2 but uses a new identity for each injected message, therefore un- dermining our fair buffer management. Below, we compare the results of DD, SEB, and PS. To isolate the effect of our Sybil-secure priority set, we include an additional strategy PSo which only uses S0 and S1 (i. e., this strategy does not relay messages for unregistered nodes and registered This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing s article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing 10 10 0 1 2 3 4 Delivery Delay [h] 0 25 50 75 100 Delivered Msgs. [%] SEB MR FIFO (a) Rϵ = 0 (no attack) 0 1 2 3 4 5 Delivery Delay [h] 0 25 50 75 Delivered Msgs. [%] SEB MR FIFO (b) Rϵ = 1Rn 0 1 2 3 4 5 Delivery Delay [h] 0 25 50 75 Delivered Msgs. [%] SEB MR FIFO (c) Rϵ = 2Rn . 0 1 2 3 4 5 Delivery Delay [h] 0 25 50 75 100 Delivered Msgs. [%] SEB MR FIFO (e) pre-registered (ped.) 0 1 2 3 4 5 Delivery Delay [h] 0 25 50 75 100 Delivered Msgs. [%] SEB MR FIFO (d) pre-registered (car) 0 1 2 3 4 5 Delivery Delay [h] 0 25 50 75 100 Delivered Msgs. [%] SEB MR FIFO (f) post-registered 0 1 2 3 4 5 Delivery Delay [h] 0 20 40 60 80 Delivered Msgs. [%] SEB MR FIFO (g) unregistered Figure 2: Flooding attack. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ 8.3 Sybil Attack Delivered messages over the delivery delay for different attacker injection rates Rϵ (top) and user groups at Rϵ = 1Rn (bottom) Figure 2: Flooding attack. Delivered messages over the delivery delay for different attacker injection rates Rϵ (top) and user groups at Rϵ = 1Rn (bottom) 0.0 0.1 0.2 0.3 0.4 Time [h] 0 1 2 3 4 Copies per Message (b) MR at Rϵ = 1Rn 0.00 0.25 0.50 0.75 1.00 1.25 Time [h] 0 200 400 600 800 Copies per Message (a) SEB at Rϵ = 1Rn 0 1 2 3 4 5 Time [h] 0 20 40 60 80 Copies per Message (c) FIFO at Rϵ = 1Rn Figure 3: Flooding attack. Message copies over time of the pre-registered (pedestrian) node group at Rϵ = 1Rn for simulation run 1. The center line shows the median, the shaded areas the 10th, 30th, 70th, and 90th percentiles. 0.0 0.1 0.2 0.3 0.4 Time [h] 0 1 2 3 4 Copies per Message (b) MR at Rϵ = 1Rn 0.00 0.25 0.50 0.75 1.00 1.25 Time [h] 0 200 400 600 800 Copies per Message (a) SEB at Rϵ = 1Rn 0 1 2 3 4 5 Time [h] 0 20 40 60 80 Copies per Message (c) FIFO at Rϵ = 1Rn Figure 3: Flooding attack. Message copies over time of the pre-registered (pedestrian) node group at Rϵ = 1Rn for simulation run 1. The center line shows the median, the shaded areas the 10th, 30th, 70th, and 90th percentiles. Figure 3: Flooding attack. Message copies over time of the pre-registered (pedestrian) node group at Rϵ = 1Rn for simulation run 1. The center line shows the median, the shaded areas the 10th, 30th, 70th, and 90th percentiles. 0 1 2 3 4 5 Delivery Delay [h] 0 25 50 75 100 Delivered Msgs. [%] PSo PS SEB DD (a) Delivered messages 0 1 2 3 4 5 Time [h] 0 2 4 6 8 Copies per Message (b) Message copies for PS pre-registered (pedestrian) 0 1 2 3 4 5 Delivery Delay [h] 0 20 40 60 Delivered Msgs. [%] PSo PS SEB DD (c) Delivered messages 0 1 2 3 4 5 Time [h] 0 2 4 6 8 Copies per Message (d) Message copies for PS unregistered Figure 4: Sybil attack. Delivered messages over the delivery delay and message lifetime at Rϵ = 1Rn. 8.3 Sybil Attack Since cars might not be usable due to blocked roads, we focus on pedestrians only. Still, PS achieves signiﬁcantly better results than FIFO. This suggests that PS is applicable even to benign settings. the underlying mobility, e. g., some nodes (injured citizens) do not move at all. In addition, the synthetic scenario features fast-moving cars and a higher node density. Since cars might not be usable due to blocked roads, we focus on pedestrians only. Still, PS achieves signiﬁcantly better results than FIFO. This suggests that PS is applicable even to benign settings. Direct Delivery Deﬁnes the Lower Bound Performance. In Fig. 4c, we show the performance for unregistered users. We see that the performance (even with PS) is signiﬁcantly lower than for registered users. However, all of our buffer management strategies perform better than DD, asserting our claim of achieving a lower bound performance. SEB performs slightly better as it can take advantage of the epidemic ﬂooding at the beginning of the simulation when the attacker’s messages have not yet fully saturated the net- work. PS performs better because some of the unregistered nodes become registered while their messages traverse the network and, thus, are prioritized by other nodes which also receive the new identity certiﬁcate. Figure 4d shows the average message copies which increase towards the end of the message lifetime. Effectively No Service for Unregistered Citizens. Un- registered users () are almost completely denied service under attack, experiencing a delivery rate of 19 % at most (Fig. 6). Those citizens already receive reduced service in a benign scenario (Fig. 5). The direct-delivery performance that is achieved under attack does not sufﬁce to maintain a reasonable delivery rate. This discrepancy emphasizes the importance of user registration especially under a Sybil attack. Impact of PS Size. Under attack, delivery rates drop sig- niﬁcantly for all groups (Fig. 6a). While well-connected registered groups can still achieve inter-group delivery rates above 90 %, unregistered groups are effectively cut off from communication. The situation for registered nodes improves when we increase \|S1\| from 10 up to 50 (Figs. 6b and 6c): at \|S1\| = 50, the delivery rate is only 3–12 % lower than in a benign setting. However, as expected, increasing \|S1\| has no effect on unregistered users. 8.4 2013 Typhoon Haiyan Scenario To evaluate how RESCUE would perform in a disaster sce- nario, we repeat our experiments under an accurate human mobility model for large-scale natural disasters [38]. This model features seven different node roles (including citizens and professional disaster response teams (DRTs)), various points-of-interest (base camps, etc.), and time-of-day depen- dent activities. Depending on the group, we consider that users are registered () or unregistered (). We evaluate RESCUE in the Typhoon Haiyan scenario (city of Tacloban, Philippines) with a total of 500 nodes. We show the inter- group message delivery rates for PS in Figs. 5 and 6. For space reasons, we only show the results for no attacker (Fig. 5) and for the Sybil attack with a ﬂooding rate of Rϵ = 1Rn but with different values for \|S1\| (Fig. 6). The Sybil attack is the same as in Section 8.3 with the difference that the attacker node is chosen randomly from the DRO group, since this group is the best connected one due to high contact rates. Delivery Asymmetries. We detect asymmetries in the de- livery rates between certain group pairs in Fig. 5 such as the registered DROs and the unregistered injured citizens (0.54 vs. 0.28). These asymmetries can be explained by our PS levels: registered users can utilize secure relays than unregistered ones. These asymmetries might inﬂuence the design of applications building on RESCUE. 8.5 Threats to Validity We designed RESCUE to be agnostic to the underlying mobility model. While this makes our approach applica- ble to different scenarios, we acknowledge that—as with any DTN—the nodes’ mobility behavior governs its overall performance, i. e., reliability and delay. This means that RESCUE’s performance might vary in different scenarios. To assess the extend to which mobility inﬂuences perfor- mance, we considered two distinct scenarios that cover two extremes: (i) a well-connected network with several fast- moving nodes and (ii) a poorly connected network with slowly moving nodes. We expect that performance in other scenarios with similar features will fall between these two scenarios. Therefore, we can view the results presented in this section as an operating range when assessing whether RESCUE should be deployed in a particular (other) sce- nario. Impact of Mobility Model. In comparison to our synthetic scenario, the most striking difference is that we do not achieve perfect delivery rates when no attackers are present (compare Fig. 2a and Fig. 5). These differences are caused by Message delivery rate 1.0 0.0 Injured Healthy Healthy Scientist Govt. UN USRTs DROs Receiver Injured Healthy Healthy Scientist Govt. UN USRTs DROs Sender 0.18 0.31 0.22 0.29 0.33 0.33 0.32 0.28 0.40 0.58 0.56 0.59 0.64 0.63 0.58 0.58 0.46 0.69 0.63 0.62 0.70 0.71 0.69 0.72 0.52 0.73 0.59 0.69 0.75 0.77 0.80 0.75 0.57 0.80 0.75 0.79 0.88 0.88 0.89 0.92 0.58 0.81 0.75 0.75 0.89 0.90 0.90 0.93 0.51 0.74 0.63 0.75 0.81 0.81 0.96 0.94 0.54 0.75 0.68 0.75 0.82 0.82 0.94 0.98 Figure 5: Typhoon Haiyan scenario. No attack (Rϵ = 0). Message delivery rates between roles using PS. 8.3 Sybil Attack 0 1 2 3 4 5 Time [h] 0 2 4 6 8 Copies per Message (b) Message copies for PS d (pedestrian) Figure 4: Sybil attack. Delivered messages over the delivery delay and message life Sybil attack. Delivered messages over the delivery delay and message lifetime at Rϵ = 1Rn. nodes that are not in the Sybil-secure priority set). We omit the results for FIFO and MR since they already performed poorly under a non-Sybil attack. We set the Sybil-secure priority set size \|S1\| to 10. For the pre-registered nodes, the sets are generated from a ﬁxed-degree random graph, while the sets of the post-registered nodes are determined by the \|S1\| most recently post-registered nodes. For space reasons, Fig. 4 shows only the results for the pre-registered and unregistered group. PS, the situation improves enormously. The secure copies in S1 work as intended, so that there is a small number (≤\|Sl\|) of message copies that propagate through the network. Con- versely, this experiment also shows that all \|Sl\| with l > 1 are completely vulnerable to a Sybil attack: the attacker is able to effectively replace all non-secure copies, which is con- ﬁrmed by the almost identical performance of PSo and PS (overlap in Fig. 4a). Therefore, the network already operates at its lower performance limit. Further increasing the at- tacker’s injection rate (beyond Rϵ = 1Rn) or the number of attackers should, thus, not affect the network’s performance. This behavior demonstrates the resilience aspect of RESCUE: Sybil Attacker Replaces All Non-secure Copies. In Fig. 4a, we can see that the delivery rate for SEB signiﬁcantly decreases under a Sybil attack compared to Fig. 2e. With This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing epted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing 11 even under a strong attack, RESCUE continues to operate considerably well but with reduced performance. the underlying mobility, e. g., some nodes (injured citizens) do not move at all. In addition, the synthetic scenario features fast-moving cars and a higher node density. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ 9 CONCLUSION We also leveraged the concept of secure copies to implement priority sets to offer protection against Sybil attacks. We eval- uated our solution in a synthetic scenario and have shown that, under ﬂooding attacks, RESCUE maintains a delivery rate of 100 % for all registered users, while unregistered users experience a drop of up to 20 %. In the presence of a Sybil attacker, RESCUE maintains a delivery rate close to 100 % for all registered users, while unregistered users can still deliver more than 60 % of their messages. Finally, we conﬁrmed that RESCUE performs well in a realistic natural disaster scenario and showed that the priority set size can be increased to improve the delivery success under a Sybil attack. In future work, we will integrate our buffer management strategies into an existing DTN implemen- tation to evaluate performance and energy efﬁciency on mobile devices. In addition, we want to further investigate the relation between messages copies and delivery rates to appropriately dimension the priority sets. Finally, the proposed solution should be investigated in applications of the Internet-of-Things or in Fog computing scenarios. Here, we ﬁnd similar heterogeneous device setups and often ﬂuctuating network connectivity. p pp [4] H. Yih-Chun and A. Perrig, “A survey of secure wireless ad hoc routing,” IEEE Security & Privacy Magazine, vol. 2, no. 3, pp. 28–39, 2004. [5] Q. Li, W. Gao, S. Zhu, and G. Cao, “To lie or to comply: Defend- ing against ﬂood attacks in disruption tolerant networks,” IEEE Transactions on Dependable and Secure Computing, vol. 10, 2013. [6] J. Zheng, Q. Li, G. Gu, J. Cao, D. K. Yau, and J. Wu, “Realtime DDoS Defense Using COTS SDN Switches via Adaptive Correla- tion Analysis,” IEEE TIFS, 2018. [7] F. Kohnh¨auser, M. Stute, L. Baumg¨artner, L. Almon, S. Katzen- beisser, M. Hollick, and B. Freisleben, “SEDCOS: A secure device- to-device communication system for disaster scenarios,” in IEEE Conference on Local Computer Networks (LCN), 2017. [8] T. Hossmann, P. Carta, D. Schatzmann, F. Legendre, P. Gunning- berg, and C. Rohner, “Twitter in disaster mode: security architec- ture,” in ACM Special Workshop on Internet and Disasters, 2011. [9] S. G. Weber, Y. Kalev, S. Ries, and M. M¨uhlh¨auser, “MundoMes- sage: Enabling trustworthy ubiquitous emergency communica- tion,” in ACM IMCOM, 2011. [10] Briar Project, “Website,” 2017. [Online]. Available: https://briarproject.org [11] K. Sakai, M.-T. Sun, W.-S. Ku, and J. 9 CONCLUSION Wu, “A framework for anonymous routing in delay tolerant networks,” in International Conference on Network Protocols (ICNP). IEEE, 2017, pp. 1–10. f pp [12] J. Burgess, G. D. Bissias, M. D. Corner, and B. N. Levine, “Sur- viving attacks on disruption-tolerant networks without authenti- cation,” in ACM MobiHoc, 2007. [13] T. N. D. Pham, C. K. Yeo, N. Yanai, and T. Fujiwara, “Detecting ﬂooding attack and accommodating burst trafﬁc in delay-tolerant networks,” IEEE Transactions on Vehicular Technology, vol. 67, no. 1, pp. 795–808, 2018. 9 CONCLUSION UN USRTs DROs Sender 0.01 0.04 0.03 0.06 0.07 0.07 0.03 0.03 0.04 0.19 0.17 0.18 0.15 0.15 0.07 0.09 0.11 0.40 0.36 0.33 0.40 0.41 0.26 0.30 0.17 0.39 0.29 0.43 0.50 0.50 0.54 0.47 0.20 0.48 0.45 0.54 0.72 0.71 0.69 0.72 0.21 0.48 0.43 0.50 0.73 0.74 0.71 0.76 0.18 0.44 0.36 0.53 0.62 0.64 0.87 0.85 0.19 0.46 0.41 0.52 0.65 0.66 0.85 0.89 (b) \|S1\| = 20 Injured Healthy Healthy Scientist Govt. UN USRTs DROs Receiver Injured Healthy Healthy Scientist Govt. UN USRTs DROs Sender 0.01 0.04 0.03 0.06 0.07 0.07 0.03 0.03 0.04 0.19 0.17 0.19 0.15 0.15 0.07 0.09 0.17 0.48 0.45 0.41 0.49 0.51 0.36 0.41 0.26 0.54 0.42 0.53 0.64 0.63 0.69 0.63 0.29 0.61 0.57 0.64 0.78 0.78 0.81 0.84 0.31 0.60 0.55 0.60 0.79 0.80 0.81 0.86 0.26 0.57 0.46 0.64 0.69 0.70 0.92 0.89 0.27 0.59 0.52 0.62 0.71 0.72 0.90 0.94 (c) \|S1\| = 50 Figure 6: Typhoon Haiyan scenario. Sybil Attack (Rϵ = 1Rn). Message delivery rates between roles using PS. Injured Healthy Healthy Scientist Govt. UN USRTs DROs Receiver Injured Healthy Healthy Scientist Govt. UN USRTs DROs Sender 0.01 0.04 0.03 0.06 0.07 0.07 0.03 0.03 0.04 0.19 0.17 0.18 0.15 0.15 0.07 0.09 0.11 0.40 0.36 0.33 0.40 0.41 0.26 0.30 0.17 0.39 0.29 0.43 0.50 0.50 0.54 0.47 0.20 0.48 0.45 0.54 0.72 0.71 0.69 0.72 0.21 0.48 0.43 0.50 0.73 0.74 0.71 0.76 0.18 0.44 0.36 0.53 0.62 0.64 0.87 0.85 0.19 0.46 0.41 0.52 0.65 0.66 0.85 0.89 (b) \|S1\| = 20 (c) \|S1\| = 50 Figure 6: Typhoon Haiyan scenario. Sybil Attack (Rϵ = 1Rn). Message delivery rates between roles using PS. Figure 6: Typhoon Haiyan scenario. Sybil Attack (Rϵ = 1Rn). Message delivery rates betw [3] A. Mart´ın-Campillo, J. Crowcroft, E. Yoneki, and R. Mart´ı, “Eval- uating opportunistic networks in disaster scenarios,” Journal of Network and Computer Applications, 2013. attacks even in the presence of Sybil attackers. In particular, our solution uses a minimalistic communication protocol and implements ﬂooding mitigation via source-based elastic buckets that prevent attackers from purging valid messages. ACKNOWLEDGMENTS pp [14] H. Yu, M. Kaminsky, P. B. Gibbons, and A. Flaxman, “SybilGuard: defending against sybil attacks via social networks,” in ACM SIGCOMM, vol. 36, 2006. This work has been co-funded by the LOEWE initiative (Hesse, Germany) within the emergenCITY center and by the German Federal Ministry of Education and Research and the Hessen State Ministry for Higher Education, Research and the Arts within their joint support of the National Research Center for Applied Cybersecurity ATHENE. [15] H. Yu, P. B. Gibbons, M. Kaminsky, and F. Xiao, “SybilLimit: A near-optimal social network defense against sybil attacks,” IEEE/ACM Transactions on Networking, vol. 18, no. 3, 2010. [16] D. Quercia and S. Hailes, “Sybil attacks against mobile users: Friends and foes to the rescue,” 2010. [17] S. Trifunovic, M. Kurant, K. A. Hummel, and F. Legendre, “Pre- venting spam in opportunistic networks,” 2014. [18] S. C. Nelson, M. Bakht, and R. Kravets, “Encounter-based routing in DTNs,” in IEEE INFOCOM, 2009. 9 CONCLUSION In this work, we presented RESCUE, a communication framework for resilient and secure disruption-tolerant emer- gency communication on mobile devices. RESCUE is the ﬁrst secure emergency communication solution that allows users to join the network during disasters when infras- tructure networks are unavailable by deploying mobile authorities in the ﬁeld. In addition, we are the ﬁrst to present a buffer management approach to mitigate ﬂooding Figure 5: Typhoon Haiyan scenario. No attack (Rϵ = 0). Message delivery rates between roles using PS. has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing his article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing 12 Injured Healthy Healthy Scientist Govt. UN USRTs DROs Receiver Injured Healthy Healthy Scientist Govt. UN USRTs DROs Sender 0.01 0.04 0.03 0.06 0.07 0.07 0.03 0.03 0.04 0.19 0.17 0.19 0.15 0.15 0.07 0.09 0.17 0.48 0.45 0.41 0.49 0.51 0.36 0.41 0.26 0.54 0.42 0.53 0.64 0.63 0.69 0.63 0.29 0.61 0.57 0.64 0.78 0.78 0.81 0.84 0.31 0.60 0.55 0.60 0.79 0.80 0.81 0.86 0.26 0.57 0.46 0.64 0.69 0.70 0.92 0.89 0.27 0.59 0.52 0.62 0.71 0.72 0.90 0.94 (c) \|S1\| = 50 12 12 Message delivery rate 1.0 0.0 Injured Healthy Healthy Scientist Govt. UN USRTs DROs Receiver Injured Healthy Healthy Scientist Govt. UN USRTs DROs Sender 0.01 0.04 0.03 0.06 0.07 0.07 0.03 0.03 0.04 0.19 0.17 0.19 0.15 0.15 0.07 0.09 0.07 0.31 0.27 0.25 0.29 0.29 0.17 0.20 0.11 0.29 0.21 0.35 0.39 0.38 0.38 0.33 0.16 0.38 0.35 0.43 0.68 0.66 0.59 0.63 0.16 0.38 0.34 0.40 0.67 0.70 0.61 0.68 0.12 0.33 0.28 0.44 0.59 0.59 0.84 0.81 0.13 0.36 0.32 0.42 0.61 0.62 0.82 0.86 (a) \|S1\| = 10 Injured Healthy Healthy Scientist Govt. UN USRTs DROs Receiver Injured Healthy Healthy Scientist Govt. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ REFERENCES Stefan Katzenbeisser received his Ph.D. de- gree from the Vienna University of Technology, Austria. Since 2019, he has been a professor at University of Passau, previously at TU Darm- stadt. His current research interests include em- bedded security, data privacy and cryptographic protocol design. He has authored over 200 sci- entiﬁc publications and served on the program committees of several workshops and confer- ences devoted to information security. Stefan Katzenbeisser received his Ph.D. de- gree from the Vienna University of Technology, Austria. Since 2019, he has been a professor at University of Passau, previously at TU Darm- stadt. His current research interests include em- bedded security, data privacy and cryptographic protocol design. He has authored over 200 sci- entiﬁc publications and served on the program committees of several workshops and confer- ences devoted to information security. Milan Stute received his Ph.D. degree in com- puter science from TU Darmstadt in 2020 and is now a postdoctoral researcher at the Secure Mobile Networking Lab. His research focuses on denial-of-service attacks and prevention mech- anisms for distributed wireless networks. He is recipient of several awards including outstand- ing Bachelor and Master thesis awards as well as ACM MobiCom Best Community Paper and Demo Awards. Milan Stute received his Ph.D. degree in com- puter science from TU Darmstadt in 2020 and is now a postdoctoral researcher at the Secure Mobile Networking Lab. His research focuses on denial-of-service attacks and prevention mech- anisms for distributed wireless networks. He is recipient of several awards including outstand- ing Bachelor and Master thesis awards as well as ACM MobiCom Best Community Paper and Demo Awards. Milan Stute received his Ph.D. degree in com- puter science from TU Darmstadt in 2020 and is now a postdoctoral researcher at the Secure Mobile Networking Lab. His research focuses on denial-of-service attacks and prevention mech- anisms for distributed wireless networks. He is recipient of several awards including outstand- ing Bachelor and Master thesis awards as well as ACM MobiCom Best Community Paper and Demo Awards. Bernd Freisleben received his Ph.D. degree and Habilitation from TU Darmstadt, in 1985 and 1993, respectively. He is a full professor in the Department of Mathematics and Com- puter Science, University of Marburg, Germany, and a part-time professor in the Department of Electrical Engineering and Information Tech- nology, TU Darmstadt. REFERENCES [19] F. Li, J. Wu, and A. Srinivasan, “Thwarting blackhole attacks in disruption-tolerant networks using encounter tickets,” in IEEE INFOCOM, 2009. [1] IRIN News, “Life-saving radio begins broadcasting in typhoon-hit Tacloban,” Nov. 2013. [Online]. Avail- able: http://www.irinnews.org/report/99132/life-saving-radio- begins-broadcasting-typhoon-hit-tacloban [20] ——, “Thwarting blackhole attacks in disruption-tolerant net- works using encounter tickets,” in IEEE INFOCOM, 2009. g g yp [2] Z. Lu, G. Cao, and T. La Porta, “Networking smartphones for disaster recovery,” in IEEE PerCom, 2016. [21] Q. Li and G. Cao, “Mitigating routing misbehavior in disruption tolerant networks,” IEEE TIFS, vol. 7, no. 2, 2012. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing s been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TDSC.2020.3036224, IEEE Transactions on Dependable and Secure Computing 13 [22] X. Zhang, G. Neglia, J. Kurose, and D. Towsley, “Performance modeling of epidemic routing,” Computer Networks, vol. 51, no. 10, 2007. Lars Baumg¨artner received his Ph.D. degree in computer science from the University of Mar- burg, Germany. Since 2019, he has been a re- search staff member in the Software Technology Group at the TU Darmstadt. His research inter- ests include computer/network security, resilient communication using embedded systems and mobile devices, and delay-tolerant networking for emergency scenarios in particular. [23] S. Skorobogatov, “The bumpy road towards iphone 5c nand mir- roring,” arXiv preprint arXiv:1609.04327, 2016. g p p [24] Android Developers, “Safetynet attestation api,” 2018, https://developer.android.com/training/safetynet/attestation. p p g y [25] Samsung, “White paper: An overview of Samsung KNOX,” 2013. [26] Alan Meeus, “Windows 10 Mobile security guide,” 2017, https://technet.microsoft.com/en-us/itpro/windows/keep- secure/windows-10-mobile-security-guide. y g [27] Joseph Yiu, “ARMv8-M architecture technical overview,” 2015, https://community.arm.com/docs/DOC-10896. p y [28] Trusted Computing Group, “TPM 2.0 Mo- bile Reference Architecture Speciﬁcation,” 2014, https://www.trustedcomputinggroup.org/tpm-2-0-mobile- reference-architecture-speciﬁcation/. Lars Almon received his master degrees in computer science and IT security from TU Darm- stadt in 2015. REFERENCES He started working as a research associate and doctoral candiate at the Secure Mobile Networking Lab in 2015. His research focuses on secure wireless communication pro- tocols for mobile and resource constraint de- vices. Especially the design and management of wireless testbeds. Lars Almon received his master degrees in computer science and IT security from TU Darm- stadt in 2015. He started working as a research associate and doctoral candiate at the Secure Mobile Networking Lab in 2015. His research focuses on secure wireless communication pro- tocols for mobile and resource constraint de- vices. Especially the design and management of wireless testbeds. p [29] F. Kohnh¨auser, N. B¨uscher, and S. Katzenbeisser, “A practical attestation protocol for autonomous embedded systems,” in IEEE EuroS&P, 2019, pp. 263–278. pp [30] T. Abera, R. Bahmani, F. Brasser, A. Ibrahim, A.-R. Sadeghi, and M. Schunter, “Diat: Data integrity attestation for resilient collabo- ration of autonomous systems,” in NDSS, 2019. y [31] S. Wedaj, K. Paul, and V. J. Ribeiro, “Dads: Decentralized attesta- tion for device swarms,” ACM TOPS, vol. 22, no. 3, 2019. [32] Regulation of the European parliament and of the council, “EUR- Lex - 32014R0910,” 2014. [33] A. Wiesmaier, M. Horsch, J. Braun, F. Kiefer, D. Hhnlein, F. Stren- zke, and J. Buchmann, “An efﬁcient mobile PACE implementa- tion,” in ACM ASIACCS, 2011. [34] F. C. Lee, W. Goh, and C. K. Yeo, “A queuing mechanism to alleviate ﬂooding attacks in probabilistic delay tolerant networks,” in IEEE AICT, 2010. Matthias Hollick is currently heading the Se- cure Mobile Networking Lab at TU Darmstadt. After receiving the Ph.D. degree from TU Darm- stadt in 2004, he has been researching and teaching at TU Darmstadt, Universidad Carlos III de Madrid, and the University of Illinois at Urbana–Champaign. His research focus is on resilient, secure, privacy-preserving, and QoS- aware communication for mobile and wireless systems and networks. [35] W. Galuba, P. Papadimitratos, M. Poturalski, K. Aberer, Z. Despo- tovic, and W. Kellerer, “Castor: Scalable secure routing for ad hoc networks,” in IEEE INFOCOM, 2010. [36] R. Morris and K. Thompson, “Password security: A case history,” Bell Laboratories, 1978. [Online]. Available: https://web.archive.org/web/20130821093338/http://cm.bell- labs.com/cm/cs/who/dmr/passwd.ps p p [37] A. Ker¨anen, J. Ott, and T. K¨arkk¨ainen, “The ONE simulator for DTN protocol evaluation,” in ICST SIMUTools, 2009. [38] M. Stute, M. Maass, T. Schons, and M. Hollick, “Reverse engineer- ing human mobility in large-scale natural disasters,” Nov. 2017. REFERENCES His current research in- terests include mobile computing, wireless net- works, computation- and data-intensive applica- tions, and IT security. Florian Kohnh¨auser received his Ph.D. degree in computer science from TU Darmstadt in 2019. From 2015 to 2019 he worked as a research associate and doctoral candidate at the Secu- rity Engineering Group. His research interests include system and network security for mobile and embedded devices. More particularly, his research focuses on the design, development, and analysis of cryptographic protocols to verify the integrity of embedded systems (remote at- testation). This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
https://openalex.org/W1978108239	http://www.scielo.cl/pdf/eure/v40n119/art04.pdf	Portuguese	null	O princípio de coesão territorial enquanto novo paradigma de desenvolvimento na formulação de políticas públicas: (re)construindo ideias dominantes	EURE	2,014	cc-by	8,271	vol  \| no  \| enero  \| pp. - \| artículos \| ©EURE 75 Recibido el 22 de agosto de 2011, aprobado el 1 de noviembre de 2012 E-mail: g.santinha@ua.pt O princípio de coesão territorial enquanto novo paradigma de desenvolvimento na formulação de políticas públicas: (re)construindo ideias dominantes Gonçalo Santinha6OJWFSTJEBEFEF"WFJSP "WFJSP 1PSUVHBM resumo \| A valorização da dimensão territorial na formulação de políticas públicas UFNTJEPWJTÓWFMOBTPSJFOUBÎÜFTFNBOBEBTQFMBTNBJTEJWFSTBTJOTUJUVJÎÜFTNVOEJBJT lançando recentemente para o debate político e acadêmico a Coesão Territorial en- quanto novo paradigma de desenvolvimento. Num momento de particular impor- UÉODJB FNRVFTFEJTDVUFNBTPSJFOUBÎÜFTQBSBPQFSÓPEPQØTFBBUSJCVJÎÍP dos fundos estruturais, esta questão assume um interesse adicional nos Estados- .FNCSPTEB6OJÍP&VSPQFJB&NCPSB DPNPPCKFUJWPQPMÓUJDP B$PFTÍP5FSSJUPSJBM UFOIBWJOEPBBERVJSJSVNBJNQPSUÉODJBFMFWBEB BJOEBÏOPUØSJBBBVTÐODJBEFVNB delimitação precisa do conceito, diﬁcultando a capacidade de lidar de forma ana- lítica e normativa com este princípio. É sobre esta questão que o texto se debruça, identiﬁcando-se caminhos que poderão contribuir como referencial para a formu- lação e territorialização de políticas públicas, melhorando, assim, os processos de decisão com base neste princípio. palavras-chave \| desenvolvimento territorial, distribuição espacial, integração territorial. abstract \| There has been a growing interest by several worldwide institutions in both academia and policy making arenas in the territorial dimension of public policy design, discussing the added value of Territorial Cohesion as the new development paradigm. In a stage of particular importance, in which the main guidelines for the post-2013 period and the allocation of structural funds are being discussed, this issue assumes an additional interest amongst the European Union Member-states. Albeit as a political objective Ter- ritorial Cohesion is here to stay, there is no agreement on its precise meaning, hence raising diﬃculties on how to deal with it in an analytical and normative way. This article seeks to bring critical commentary to this arena on this matter by providing a guiding reference for public policies design and territorialization, thus improving decision policy processes based on this principle. key words \| territorial development, spatial distribution, territorial integration. key words \| territorial development, spatial distribution, territorial integration. issn impreso 0250-7161 \| issn digital 0717-6236 ©EURE \| vol  \| no  \| enero  \| pp. -  Introdução O texto inicia com uma panorâmica das políticas públicas Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... \| ©EURE  OPFTQBÎPFVSPQFV DPNPJOUVJUPEFDMBSJmDBSPDBNJOIPUSBÎBEPBUÏBBEPÎÍPEB ct enquanto novo paradigma de desenvolvimento. No seguimento desta narrativa IJTUØSJDB FMFODBNTFBTSB[ÜFTRVFKVTUJmDBNBBEPÎÍPEFTUFQSJODÓQJPOPDPOUFYUP atual. Os argumentos expostos constituem o ponto de partida da terceira seção, na qual se procura apresentar uma interpretação processual do princípio de ct, para que se possam formular políticas públicas que caminhem no sentido de operaciona- MJ[ÈMPFQSPNPWÐMP1PSmN OBÞMUJNBTFÎÍP ÏBQSFTFOUBEPVNRVBESPTÓOUFTFEP exercício efetuado que poderá beneﬁciar, quer futuros debates no âmbito desta te- mática, quer os próprios processos de tomada de decisão política, num momento de QBSUJDVMBSJNQPSUÉODJBFNRVFTFEJTDVUFNBTPSJFOUBÎÜFTQBSBPQFSÓPEPQØT Introdução %F VNB GPSNB DSFTDFOUF B EJNFOTÍP UFSSJUPSJBM WFN BTTVNJOEP WJTJCJMJEBEF OB formulação de políticas públicas. A título de exemplo, uma análise dos trabalhos QSPEV[JEPTQFMB0SHBOJ[BÎÍPQBSBB$PPQFSBÎÍPF%FTFOWPMWJNFOUP&DPOÙNJDP (ocde FN JOHMÐT 0SHBOJTBUJPO GPS &DPOPNJD $PPQFSBUJPO BOE %FWFMPQNFOU PVQFMP#BODP.VOEJBM NPTUSBDPNPFTUBTJOTUJUVJÎÜFT EF referência em escala mundial, sublinham a importância de abordagens de base te- rritorial como fatores essenciais na formulação de políticas públicas. No quadro concreto das políticas da Comissão Europeia (ce FTUBRVFTUÍPÏJHVBMNFOUFWJTÓWFM tendo sido inclusive adotado, recentemente, o princípio da Coesão Territorial (ct) como o novo paradigma de desenvolvimento do território europeu. g A noção de ctBEWÏNEFVNEPTDPODFJUPTQJMBSFTEB1PMÓUJDB3FHJPOBMoPEB iDPFTÍPwoRVF BVTFOUFEP5SBUBEPEF3PNB FNFSHFGPSNBMNFOUFFN OP"DUPÁOJDP&VSPQFV GB[FOEPEB$PFTÍP&DPOÙNJDBF4PDJBMVNBOPWBDPN- QFUÐODJBEB6OJÍP&VSPQFJB ue) com o intuito de se promover o desenvolvimento IBSNPOJPTPEPUFSSJUØSJP&VSPQFV.BTÏTØDPNP5SBUBEPEF"NTUFSEÍ que surge, pela primeira vez, o conceito de ctOVNUFYUPPmDJBM4FÏDFSUPRVFFTUF surge formalmente neste Tratado, com o principal intuito de moderar os efeitos associados à (tendência de) liberalização dos Serviços de Interesse Geral (Faludi, OÍPÏNFOPTWFSEBEFRVFIÈRVBUSPFQJTØEJPTJOTUJUVDJPOBJTSFDFOUFTRVF marcam um ponto de não retorno na adoção deste princípio, do ponto de vista da política pública: a publicação da Agenda Territorial da União Europeia (primeiro, em 2007, e a versão renovada, em 2011), o lançamento do Livro Verde sobre a Coesão Territorial BJODMVTÍPEBCT no Tratado de Lisboa (2009) enquanto 3º pilar EFBUVBÎÍPBQBSEB$PFTÍP&DPOÙNJDBFEB$PFTÍP4PDJBMF mOBMNFOUF PEFTFOIP EB&TUSBUÏHJB&VSPQB BEPUBEBFN DPNPJOUVJUPEFSFMBOÎBSBFDPOPNJB europeia e na qual a ct surge como um dos principais objetivos a atingir. Procurando incutir a importância do território na agenda política, o princípio da ctTVSHF OPTFJPEFTUBTPSJFOUBÎÜFT DPNPJOUVJUPHFOFSBMJ[BEPEFBMDBOÎBSP desenvolvimento harmonioso de todos os territórios (ou seja, um desenvolvimento TPDJPFDPOÙNJDPFRVJMJCSBEPFFRVJUBUJWP WBMPSJ[BSBTVBEJWFSTJEBEFFDPNQMFNFO- taridades, e facultar a possibilidade da população tirar o melhor partido das caracte- rísticas existentes em cada território. Embora o Livro Verde e a abertura do processo de consulta pública sobre o mesmo procuraram clariﬁcar o que se entendia por ct e quais as políticas a tomar para se alcançar esse princípio, ainda se veriﬁca alguma ambiguidade inerente ao seu conceito (o quê) e à sua operacionalização (como). Esta Ï QSFDJTBNFOUF VNBQSFPDVQBÎÍPRVFWFNFNFSHJOEPOPTÞMUJNPTBOPTFNEJWFS- TPTQBÓTFT DVKBTPSJFOUBÎÜFT DMBSBNFOUFJOnVFODJBEBTQFMBBHFOEBEBce, apontam no sentido de adotar este princípio na formulação de políticas públicas. É justamente sobre esta ausência de delimitação precisa do conceito e sobre o seu valor acrescentado, enquanto princípio integrante da formulação de políticas públicas, que este texto se debruça, identiﬁcando-se alguns caminhos que poderão contribuir para a melhoria dos processos de decisão e para a própria territorialização das políticas públicas. A dimensão territorial nas políticas públicas: rumo à coesão territorial %FTEFTVBPSJHFN RVFBue tem por missão promover um desenvolvimento har- NPOJPTPEBTBUJWJEBEFTFDPOÙNJDBTQFMBSFEVÎÍPEBTEJTQBSJEBEFTFOUSFBTEJWFSTBT SFHJÜFTFEPBUSBTPEBTNFOPTGBWPSFDJEBT 5SBUBEPEF3PNB 6NBNJTTÍP RVF BERVJSJV VNB NBJPS SFMFWÉODJB QPS JOUFSNÏEJP EF +BDRVFT %FMPST FN com a adoção do Acto Único Europeu, estabelecendo uma política comunitária EFDPFTÍPFDPOÙNJDBFTPDJBMBmNEFDPOUSBCBMBOÎBSPTFGFJUPTEBSFBMJ[BÎÍPEP mercado interno nos Estados-Membros menos desenvolvidos, e de atenuar as dis- DSFQÉODJBTEFEFTFOWPMWJNFOUPFOUSFBTSFHJÜFT4FNBMUFSBSBCBTFFPTPCKFUJWPTEB sua gênese, com o passar do tempo esta política foi sendo redesenhada: se, numa fase inicial, a sua concretização na procura da diminuição das assimetrias regio- nais efetuava-se com base numa distribuição equitativa dos fundos comunitários, mais tarde, constatou-se que nem sempre isso aconteceu (Santinha, Rodrigues & Almeida, 2009). Os trabalhos desenvolvidos por Martin (2003) e pela própria ce &VSPQFBO$PNNJTTJPO TÍPFTDMBSFDFEPSFTEFTUBRVFTUÍP NPTUSBOEPRVF BP longo dos tempos, embora as desigualdades entre Estados-Membros tenham dimi- nuído (em valores de pib per capita BTEJTQBSJEBEFTFOUSFSFHJÜFTnuts ii dentro de cada Estado-Membro aumentaram, contrariando assim, os objetivos de base para os quais a Política de Coesão foi desenhada. Surgindo formalmente com o Tratado de Amsterdã (1997), a ct vem dar corpo a esta preocupação com as disparidades, procurando incutir a importância do território na agenda política, não só europeia, NBTUBNCÏNEFDBEB&TUBEP.FNCSPterritory mattersÏVNBFYQSFTTÍPRVFQBTTBB ser recorrente nas diversas políticas e relatórios de orientação comunitária. Este recrudescimento do interesse na dimensão territorial das políticas públicas BEWÏNNVJUPEPUSBCBMIPEFTFOWPMWJEPEFTEFBEÏDBEBEF OPÉNCJUPEPPSEF- namento do território, evidenciando o quão importante pode ser para se atingirem os objetivos de coesão tendo em conta a diversidade territorial no espaço europeu (Ionica, 2007). Sem competência formal na ce FNNBUÏSJBEPPSEFOBNFOUPEP território, durante vários anos os defensores do princípio de ct optaram por uma FTUSBUÏHJBEJEÈUJDB DPNVNBEVQMBBCPSEBHFN +PVFO VNBWFSUFOUFQSÈUJDB PQFSBUJWB BUSBWÏT QPSFYFNQMP EPQSPHSBNBDPNVOJUÈSJP/5&33&(oDPNVNB WJTÍPEFPSEFOBNFOUPUSBOTOBDJPOBMo FVNBWFSUFOUFNBJTDPODFJUVBM OBQSPDVSBEF VNBWJTÍPFTUSBUÏHJDBDPNVNQBSBPDPOKVOUPEPFTQBÎPFVSPQFV DBQB[EFPSJFOUBS ©EURE \| vol  \| no  \| enero  \| pp. -  as políticas públicas com impacto em todas as escalas (da europeia à local). A dimensão territorial nas políticas públicas: rumo à coesão territorial \| ©EURE Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... \| ©EURE   FYJTUFOUFTFËJOUFHSBÎÍPEFOPWPTQBÓTFT P3FMBUØSJPTPCSFB$PFTÍP&DPOÙNJDB F4PDJBM OPRVBMÏNFODJPOBEPPEFTFOWPMWJNFOUPQPMJDÐOUSJDPFTVCMJOIBEB a importância da ct para o desenvolvimento do espaço europeu. Por outro lado, ao apontar para a diﬁculdade existente na cooperação interse- UPSJBMEBTEJGFSFOUFTEJSFÎÜFTHFSBJTEBce, e para a necessidade de uma articulação entre os diferentes níveis institucionais (desde a comunitária à local), o edec Ï referenciado no Livro Branco da Governança Europeia (cec, 2001) enquanto um dos quadros orientadores para uma melhor coordenação entre as políticas setoriais da ce e dos Estados-Membros. Ainda mais relevantes sobre esta questão foram os SFTVMUBEPTPCUJEPTQPS4UFBEF8BUFSIPVU OBTFOUSFWJTUBTRVFFGFUVBSBNOBT WÈSJBTEJSFÎÜFTHFSBJTEBce sobre a inﬂuência do edec no seu modus operandi. A JEFJBHFSBMÏBEFRVFPSFGPSÎPEBJNQPSUÉODJBEBEJNFOTÍPUFSSJUPSJBMOBGPSNV- lação das políticas públicas, proporcionando consequentemente um maior número EFRVFTUÜFTFNDPNVNFVNBNFMIPSEJWJTÍPEFSFTQPOTBCJMJEBEFT GPSUBMFDFVBT MJHBÎÜFTFOUSFBTEJGFSFOUFTEJSFÎÜFTHFSBJT Paralelamente a estas consequências comunitárias do ponto de vista estra- UÏHJDP JNQPSUBSFGFSJSRVFPedec teve inﬂuência num âmbito de intervenção mais alargado (pan-europeu): a cemat (Conferência Europeia dos Ministros do Ordenamento do Território), criada em 1970, e constituída pelos representantes dos Estados-Membros do Conselho da Europa, aprova em 2000 o documento Guiding Principles for Sustainable Spatial Development of the European Continent, RVFiBEPUBQBSBUPEPPFTQBÎPFVSPQFV DPNMJHFJSBTBEBQUBÎÜFTËFTQFDJmDJEBEF dos países de leste, os princípios e a visão do edecw 'FSSÍP Q %F uma forma geral, estes princípios orientadores têm como objetivo último, a har- NPOJ[BÎÍP FOUSF BT FYQFDUBUJWBT FDPOÙNJDBT F TPDJBJT FN SFMBÎÍP BP UFSSJUØSJP contribuindo assim para um desenvolvimento territorial equilibrado. Reﬁra-se, inclusive, que o primeiro dos dez princípios orientadores que elencava, visava à promoção da ctQPSJOUFSNÏEJPEFVNEFTFOWPMWJNFOUPFDPOÙNJDPFTPDJBMNBJT FRVJMJCSBEPEBTSFHJÜFTFEFVNBNBJPSDPNQFUJUJWJEBEF Numa perspectiva mais operacional, o edec forneceu um excelente impulso para o desenvolvimento de dois instrumentos promotores da cooperação territorial, no âmbito da ue. Por um lado, veriﬁcou-se o lançamento, menos de um ano depois, da iniciativa interreg iii nas suas três vertentes (transnacional, transfronteiriça e inter-regional), indicando uma forte inﬂuência do edec ao nível dos conceitos, QSJODÓQJPTFFTUSBUÏHJBT 4UFBE8BUFSIPVU 4BMF[ 1PSPVUSPMBEP GPJ DSJBEP P &VSPQFBO 4QBUJBM 1MBOOJOH 0CTFSWBUPSZ /FUXPSL (espon), que consiste num programa de investigação que apoia a realização de estudos de base territorial, seja numa ótica de análise de disparidades territoriais, seja do ponto de vista da formulação de políticas mais adequadas ao desenvolvimento territorial da Europa. A experiência mostra que os estudos desenvolvidos no âmbito do espon RVF FNCPSB NBOUFOEP P NFTNP BDSÙOJNP DVSJPTBNFOUF BMUFSPV B TVB EFTJHOBÎÍPQBSB0CTFSWBUJPO/FUXPSLGPS5FSSJUPSJBM%FWFMPQNFOUBOE$PIFTJPO enfatizando, portanto, a questão da ct) têm sido bastante úteis. A dimensão territorial nas políticas públicas: rumo à coesão territorial Foi neste sentido que se sucederam um conjunto de estudos acadêmicos, como o Europe 2000 (1991), o Europe 2000+ (1994) e, em antecipação à entrada de 12 novos membros na ue, o EDEC: &TRVFNBEF%FTFOWPMWJNFOUPEP&TQBÎP$PNVOJUÈSJP GBUPSDIBWF QBSBiBmSNBSVNBWJTÍPFTUSBUÏHJDBEFPSEFOBNFOUPTVQSBOBDJPOBMEPUFSSJUØSJPDP- NVOJUÈSJPw 'FSSÍP Q %PDVNFOUPTBQSPWBEPT TVCMJOIFTF FNSFVOJÜFT informais por parte dos ministros do ordenamento do território e desenvolvimento regional dos Estados-Membros, pelo que quer o vocabulário adotado quer o con- UFÞEPFTUSBUÏHJDPEFmOJEPDPOTUJUVÓSBNPSFTVMUBEPEFVNDPNQSPNJTTPQPMÓUJDP Pela sua relevância para o ordenamento do território e pelo seu impacto político e cientíﬁco, importa aqui destacar o edec, quer enquanto documento (e processo) chave em torno do próprio conceito de ct, quer enquanto elemento demonstrador da importância da dimensão territorial nas políticas públicas. Amplamente reconhecido e mencionado na arena acadêmica e política, o edec representa um momento crítico OBFNFSHÐODJBEFVNOPWPEJTDVSTPFVSPQFVTPCSFPPSEFOBNFOUPEPUFSSJUØSJP6N processo de aprendizagem mútuo, desenhado por uma mistura de diferentes culturas EFPSEFOBNFOUPEPUFSSJUØSJPFRVFDPOUSJCVJVQBSBVNBBMUFSBÎÍPOPNÏUPEPFOB agenda das políticas de ordenamento do território (Albrechts, 2001), signiﬁcando um deslocamento de abordagem típico de regulação e uso do solo, para uma maior ÐOGBTFOBTQPMÓUJDBTFFTUSBUÏHJBTEFPSEFOBNFOUPEPUFSSJUØSJP #ÚINF & de fato, o edec constituiu um bom instrumento para os passos que se seguiram de forte cooperação entre diferentes atores, e enquanto encorajamento de novas formas de discussão, em que a dimensão territorial ganha importância na formulação de políticas públicas em diferentes níveis de decisão (comunitário, nacional e regional). As consequências do edec foram particularmente visíveis ao nível comunitário, quer EPQPOUPEFWJTUBFTUSBUÏHJDP RVFSOVNBQFSTQFDUJWBNBJTPQFSBDJPOBM NBTUBNCÏN em alguns Estados-Membros as inﬂuências ﬁzeram-se sentir. g "POÓWFMDPNVOJUÈSJP BPiEJTQPOJCJMJ[BSVNSFGFSFODJBMFTUSBUÏHJDPQBSBPEFT- FOWPMWJNFOUPEPFTQBÎPFVSPQFVw 'FSSÍP Q Pedec produziu impactos relevantes imediatos na elaboração de políticas comunitárias, e na aplicação dos GVOEPT FTUSVUVSBJT QBSB P QFSÓPEP %F GBUP Ï OPUØSJB TVB JOnVÐODJB OP3FMBUØSJPTPCSFB$PFTÍP&DPOÙNJDBF4PDJBM OPRVBMÏTVCMJOIBEB a existência de disparidades territoriais no espaço europeu e feita uma referência explícita à questão da ct, em consonância com o primeiro princípio diretor do edecoEFTFOWPMWJNFOUPFTQBDJBMQPMJDÐOUSJDPFFRVJMJCSBEPOBue (Ferrão, 2004). Importa aqui mencionar o papel crucial que o Comissário Michel Barnier desem- penhou neste processo, inﬂuenciando a adoção do princípio de ct em textos oﬁ- ciais, como o 2º Relatório ou, previamente, o Tratado de Amsterdã (1997). Sua BEPÎÍPTFSWJSJBJODMVTJWF DPNPiEFTDVMQBwQBSBBce ultrapassar o problema da falta EFDPNQFUÐODJBTGPSNBJTFNNBUÏSJBEFPSEFOBNFOUPEPUFSSJUØSJPF BTTJN BUSJ- buir oﬁcialmente uma importância à dimensão territorial (Faludi, 2009). Ainda no DBNQPEBJOnVÐODJBFTUSBUÏHJDBEPedec FNUFSNPTEFEPDVNFOUPTFJOTUJUVJÎÜFT BPOÓWFMDPNVOJUÈSJP ÏQPTTÓWFMFODPOUSBSSFGFSÐODJBTTVCTUBODJBJTBPTTFVTQSJO- cípios no relatório seguinte da Política de Coesão (que, por sua vez, ia ganhando DBEBWF[NBJTQSFQPOEFSÉODJB DPNPSFTQPTUBBRVFTUÜFTJOFSFOUFTËTEJTQBSJEBEFT Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... A dimensão territorial nas políticas públicas: rumo à coesão territorial Vejam-se, a título de exemplo, os estudos Territory Matters for Competitiveness and Cohesion (2006) e Scenarios on the Territorial Future of Europe (2007) que permitiram, por ©EURE \| vol  \| no  \| enero  \| pp. - ©EURE \| vol  \| no  \| enero  \| pp. -  um lado, capitalizar informação bastante rica e diversa; por outro, sublinhar a importância da dimensão territorial na formulação de políticas públicas e, por ﬁm, DIBNBSBBUFOÎÍPQBSBDPNPEFUFSNJOBEBTUFOEÐODJBT EFTBmPT oNPCJMJEBEFF QSFTTÍPNJHSBUØSJB FOWFMIFDJNFOUPEFNPHSÈmDP BMUFSBÎÜFTDMJNÈUJDBTFFmDJÐODJB FOFSHÏUJDB HMPCBMJ[BÎÍPFDPNQFUJUJWJEBEFFDPOÙNJDB UFDOPMPHJBTEFJOGPSNBÎÍP FDPNVOJDBÎÍPFBDSFTDFOUFJNQPSUÉODJBEBTDJEBEFT FORVBOUPBUPSFTHMPCBJT o embora atingindo a ue na globalidade, afetam de forma diferenciada os territórios e, por conseguinte, constituem fatores catalisadores de disparidades territoriais. A MØHJDBTVCKBDFOUFÏBEFRVFQSPCMFNBTDPNVOTSFRVFSFNTPMVÎÜFTDPNVOT FB utilização de recursos comuns de forma harmonizada, tirando partido da diversidade territorial. Finalmente, e embora com menor incidência, a inﬂuência do edec não se fez só TFOUJSOPOÓWFMEBTJOTUJUVJÎÜFTFEPDVNFOUPTDPNVOJUÈSJPT5BNCÏNUFWFDPOTFRVÐO- cias no nível nacional, apesar desta ter se veriﬁcado bastante variável e, na maior parte EBTPDBTJÜFT NVJUPMJNJUBEB GB[FOEPTFTFOUJSFTTFODJBMNFOUFFNQBÓTFTPOEFBTQP- líticas de ordenamento do território foram reformuladas pouco tempo depois da sua QVCMJDBÎÍP DPNPGPSBNPTDBTPTEB(SÏDJB EB&TQBOIBFEF1PSUVHBM espon, 2007). É É certo que a falta de competências formais no seio da ce FNNBUÏSJBEFPSEF- namento do território, torna a aplicabilidade de documentos da natureza do edec menos plausível. No entanto, o fato deste ter se revelado um processo longo e par- ticipativo (em parceria entre os diversos Estados-Membros e a ce), fez com que a sua elaboração tenha se tornado mais importante do que o conteúdo do próprio documento (Albrechts, 2001; Ferrão, 2004). Este argumento baseia-se na ideia de RVFPTBUPSFTFBTJOTUJUVJÎÜFTOÍPTÍPNFSPTSFDJQJFOUFTQBTTJWPT RVFGBDJMNFOUF JNQMFNFOUBNPRVFPTEPDVNFOUPTQSPQÜFN1FMPDPOUSÈSJP TÍPBUPSFTBUJWPTOVN QSPDFTTPFNRVFPTDPODFJUPTFPTPCKFUJWPTFTUSBUÏHJDPTEFEPDVNFOUPTEBOBUVSF[B do edec não são, à partida, uma preocupação desses mesmos agentes, pelo que pro- cessos de aprendizagem como os ocorridos durante a elaboração do edec e a ade- quação das suas propostas num âmbito mais vasto da agenda política europeia (isto Ï TFSWJOEPEFGPOUFEFJOTQJSBÎÍP QPEFNJOnVFODJBSGPSUFNFOUFPTQSPDFTTPTEF UPNBEFEFDJTÍP 4UFBE8BUFSIPVU & EFGBUP BEJOÉNJDBEFTFODBEFBEB pelo edec não só contribuiu para um maior reconhecimento do ordenamento do UFSSJUØSJP FORVBOUPQSÏSFRVJTJUPQBSBPEFTFOWPMWJNFOUPIBSNPOJPTPEB$PNVOJ- EBEF QBSBVNBNFMIPSJOUFHSBÎÍPEBTÈSFBTNBJTQFSJGÏSJDBTFDPNPNFJPOFDFTTÈ- SJPQBSBVNBNBJPSSBDJPOBMJEBEFOBBGFUBÎÍPEFSFDVSTPT "ULJOTPO DPNP UBNCÏN TFSWJVEFCBTFQBSBPQSPDFTTPEFFMBCPSBÎÍPFMBOÎBNFOUP FN EB "HFOEB5FSSJUPSJBMEB6OJÍP&VSPQFJB SFTVMUBOUFEFVNBSFVOJÍPJOUFSHPWFSOBNFO- tal informal dos diversos ministros do ordenamento do território da ue, e preparado conjuntamente com a ce. Apresentado como um quadro político orientado para a ação e promotor de DPPQFSBÎÍPFOUSFPT&TUBEPT.FNCSPTFNNBUÏSJBEFPSEFOBNFOUPEPUFSSJUØSJP F sustentado nos objetivos do edec, um dos grandes desígnios da Agenda Territorial consiste em reforçar a ct da ue e colocar o seu princípio no centro do discurso eu- SPQFV4VCMJOIB BJOEB BJNQPSUÉODJBEFBSUJDVMBSBTPSJFOUBÎÜFTEFCBTFUFSSJUPSJBM com as outras políticas da ce EFTJHOBEBNFOUFB&TUSBUÏHJBEF-JTCPBFB1PMÓUJDB Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... ©EURE \| vol  \| no  \| enero  \| pp. - \| ©EURE  EF$PFTÍP QBSBQSPNPWFSPCFNFTUBSTPDJBMFFDPOÙNJDPEPFTQBÎPFVSPQFV&TUB ideia resulta, essencialmente, do fato dos respectivos ministros terem percebido que a procura de ct incluía o reforço da competitividade e a redução das disparidades, BSHVNFOUBOEPRVFB&TUSBUÏHJBEF-JTCPBOÍPUJOIBTVmDJFOUFNFOUFFNDPOTJEF- ração o potencial da diversidade territorial europeia, pelo que abordagens integradas de desenvolvimento territorial (explorando o potencial endógeno existente), pode- riam melhorar a sua prestação e, consequentemente, aumentar a competitividade da Europa (Faludi, 2009). Com o processo de elaboração do Tratado de Lisboa em mãos, e encorajada pela Agenda Territorial, a ceEFDJEJVQVCMJDBS FN PLivro Verde sobre a Coesão Te- rritorial. Já com uma interpretação de ct que extravasa a tradicional abordagem de QSPDVSBEFDPOWFSHÐODJBEBTSFHJÜFTNBJTEFTGBWPSFDJEBT FTUFEPDVNFOUPFOGBUJ[BB necessidade de se tirar partido da diversidade e identidade territoriais, de maneira a GB[FSVNNFMIPSVTPEBTTVBTDBSBDUFSÓTUJDBTJOEJWJEVBJTFTQFDÓmDBT BUSBWÏTEFNF- canismos de cooperação. Por outro lado, no processo de reformas conduzidas pela ce, e já a pensar no pós-2013, a pedido da ce e da então Comissária para a Política 3FHJPOBM%BOVUB)VCOFS PFOUÍP%JSFUPS(FSBMEP.JOJTUÏSJPEBT'JOBOÎBTFEF Economia da Itália, Fabrizio Barca, desenvolveu o relatório Barca (2009), com o ob- KFUJWPEFBQSFTFOUBSQSPQPTUBTQBSBPGVUVSPEB1PMÓUJDBEF$PFTÍP/FTTFSFMBUØSJP Ï visível a relevância da dimensão territorial na formulação de políticas públicas, argu- NFOUBOEPTFRVFTÍPOFDFTTÈSJBTQPMÓUJDBTEFGPNFOUPFDPOÙNJDPFTPDJBMUBMIBEBT QBSBBTOFDFTTJEBEFTFTQFDÓmDBTEFSFHJÜFTNVJUPEJGFSFOUFTFOUSFTJ HBOIBOEPQBSUJ- cular relevância o princípio de subsidiariedade (aumentando, consequentemente, a importância do nível local) e a articulação entre os vários níveis de políticas e insti- UVJÎÜFT EBFTDBMBFVSPQFJBËMPDBM &NCPSBOÍPNFODJPOBOEPFTQFDJmDBNFOUFBct, ÏDMBSBBSFMBÎÍPFOUSFBabordagem de base territorial que advoga e este princípio. O objetivo consiste em reduzir a ineﬁciência causada pela subutilização dos recursos, e em combater a exclusão social. Todo este percurso conduziu a um reconhecimento mais generalizado da importância da componente territorial para o projeto europeu, colocando nas agendas política e mediática esta temática. Exemplo disso foi a inclusão, em dezembro de 2009 no Tratado de Lisboa, do princípio de ct enquanto 3º pilar de BUVBÎÍPBQBSEB$PFTÍP&DPOÙNJDBFEB$PFTÍP4PDJBM DMBSBNFOUFTVCMJOIBOEPB QSFQPOEFSÉODJBEFTUFQSJODÓQJPOBEFmOJÎÍPEFFTUSBUÏHJBTFOBQSØQSJBDPOTPMJEBÎÍP do projeto europeu (e uma visão sistêmica do espaço europeu implica que cada Estado-Membro possua ideias claras sobre o caminho mais adequado a percorrer). "UVBMNFOUF FEFNBOFJSBBBUVBMJ[BSBEJNFOTÍPFTUSBUÏHJDBEPTEPDVNFOUPT europeus de ordenamento do território, foi aprovada e lançada uma nova versão EB "HFOEB 5FSSJUPSJBM o ta (European Commission, 2011) com a dupla intenção de reforçar o papel da ct enquanto novo paradigma de desenvolvimento do território europeu, e de tornar os territórios mais resilientes a desaﬁos globais RVF FOUSFUBOUP TVSHJSBNPVBMDBOÎBSBNPVUSBTQSPQPSÎÜFT EFTJHOBEBNFOUFBDSJTF FDPOÙNJDBFmOBODFJSB PTGFOÙNFOPTEFNPHSÈmDPT FOWFMIFDJNFOUPFNJHSBÎÍP PEFTQPWPBNFOUPEFEFUFSNJOBEBTÈSFBT BTBMUFSBÎÜFTDMJNÈUJDBT BFmDJÐODJBFB TFHVSBOÎBFOFSHÏUJDB BWVMOFSBCJMJEBEFEFDFSUBTFDPOPNJBTFDPNVOJEBEFTMPDBJTF ©EURE \| vol  \| no  \| enero  \| pp. ©EURE \| vol  \| no  \| enero  \| pp. - -  o papel das cidades enquanto motores de desenvolvimento (European Commission, ²WFSEBEFRVFBWFSTÍPEFKÈFGFUVBWBVNBBOÈMJTFEFTUBTRVFTUÜFT NBT nesta nova versão a profundidade e a abrangência são maiores, procurando tirar partido de um novo quadro institucional mais favorável (decorrente do Tratado de -JTCPB FBSUJDVMBSTFDPNBOPWBFTUSBUÏHJBEurope 2020. Não sendo, portanto, uma preocupação totalmente nova, o princípio da CT TVSHFOPTFJPEFTUBTPSJFOUBÎÜFTDPNPJOUVJUPHFOFSBMJ[BEPEFBMDBOÎBSPEFTFO- volvimento harmonioso de todos os territórios, valorizar sua diversidade e com- plementaridades e facultar a possibilidade da população tirar o melhor partido das características existentes em cada território. É certo que o Livro Verde e a abertura do processo de consulta pública sobre o mesmo procuraram clariﬁcar o que se entendia por ctFRVBJTBTQPMÓUJDBTBUPNBSQBSBTFBMDBOÎBSFTTFQSJODÓQJP NBTOÍPÏNFOPT verdade que ainda se veriﬁca alguma ambiguidade inerente ao seu conceito. Sendo notória a ausência de uma delimitação precisa do conceito, mas que do ponto de vista discursivo inﬂuencia claramente as agendas políticas (e cientíﬁcas), qual então o seu valor acrescentado enquanto princípio integrante da formulação de políticas públicas? A discussão em torno desta questão será efetuada num duplo registro: QSJNFJSP FMFODBOEPBTQSJODJQBJTSB[ÜFTRVFKVTUJmDBNBBEPÎÍPEPQSJODÓQJPEF ct na formulação de políticas públicas no contexto atual e, depois, aprofundando conceitualmente este princípio para que se possam formular políticas públicas que caminhem no sentido de operacionalizá-la e promovê-la. A adoção do princípio de coesão territorial no contexto atual das políticas públicas "FNFSHÐODJBEBTPSJFOUBÎÜFTFTUSBUÏHJDBTEFOBUVSF[BQPMJUJDPOPSNBUJWBToEFT- DSJUBTOBTFÎÍPBOUFSJPSoEFTUBDBNBJNQPSUÉODJBEBEJNFOTÍPUFSSJUPSJBMFEBT especiﬁcidades locais na formulação de políticas públicas, considerando o territó- rio não só enquanto espaço físico de suporte às atividades existentes, mas, sobre- UVEP DPNPVNTJTUFNBTPDJBMFJOTUJUVDJPOBMDPNNÞMUJQMBTSFMBÎÜFT BUJWJEBEFTF USBOTBÎÜFTFDPOÙNJDBT$PNFTUFFOGPRVF TVCMJOIBTFRVFPTQSPDFTTPTTPDJBJTF FDPOÙNJDPTUPNBNGPSNBOVNDPOUFYUPFTQBDJBMFFTUÍPDPOEJDJPOBEPTQPSGBUP- SFTHFPHSÈmDPTDPNPBEJTUÉODJBQSPYJNJEBEF BTDBSBDUFSÓTUJDBTGÓTJDBT DMJNB UP- pograﬁa) e, ainda, a identidade de cada local. Contraria-se, portanto, a lógica que NVJUPTBVUPSFTEFGFOEJBNOPmOBMEPTÏDVMPQBTTBEPDPNGFOÙNFOPEBHMPCBMJ- zação e o advento e a proliferação das tecnologias de informação e comunicação, o espaço (ou a geograﬁa) deixaria de ser relevante, na medida em que haveria uma perda de importância do acesso a fatores de produção tangíveis, em prol do interesse da disponibilidade de fatores de produção intangíveis, designadamente a informação e o conhecimento. &TUBTRVFTUÜFTUSBÎBSBNVNDBNJOIPQBSBBBEPÎÍPEPQSJODÓQJPEFct na agenda política, nomeadamente da ce OBQSPDVSBEFVNEFTFOWPMWJNFOUPIBSNPOJPTP equilibrado do território, em que a dimensão territorial adquire uma maior prepon- derância. A título de exemplo, vejam-se os inúmeros estudos elaborados, seja, por exemplo, ao nível do espon, seja designadamente no âmbito da ceQPSJOUFSNÏEJP Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... \| ©EURE  EPTSFMBUØSJPTEFDPFTÍP EFNPOTUSBOEPPTFGFJUPTBTTJNÏUSJDPTRVFVNDPOKVOUP de tendências atuais têm produzido, comprovando que os contextos territoriais podem, eles mesmos, serem ampliﬁcadores de desequilíbrios. Exemplos incluem BNPCJMJEBEFFBQSFTTÍPNJHSBUØSJB PFOWFMIFDJNFOUPEFNPHSÈmDP BTBMUFSBÎÜFT DMJNÈUJDBTFBFmDJÐODJBFOFSHÏUJDB BHMPCBMJ[BÎÍPFDPNQFUJUJWJEBEFFDPOÙNJDB BT tecnologias de informação e comunicação, e a crescente importância das cidades (enquanto atores globais). Tendências que se tornam claramente um desaﬁo para as políticas públicas, afetando quer a competitividade dos territórios quer os equilí- brios territoriais (numa lógica de igualdade de oportunidades), muitas vezes vistos como objetivos de certa forma conﬂituosos (Bertrand & Peyrache-Gadeau, 2009). A adoção do princípio de ct vem, precisamente, procurar lidar com esta questão, realçando as características especíﬁcas de cada território, não só ao nível dos proble- NBT NBTUBNCÏNUFOEPFNDPOUBBTPQPSUVOJEBEFT5BMQSJODÓQJPSFWFMBVNBOPWB abordagem, que une à ideia (sempre necessária) de combater disparidades regionais FTTFODJBMNFOUFBQBSUJSEFTJUVBÎÜFTQSPCMFNÈUJDBT VNBMØHJDBEFWBMPSJ[BÎÍPUFSSJ- UPSJBM JTUPÏ EFDPODFQÎÍPEPUFSSJUØSJPFORVBOUPPQPSUVOJEBEF Por outro lado, reconhecendo-se que as políticas públicas com impacto territorial são orientadas no âmbito da ue, quer pelas próprias políticas de ordenamento do UFSSJUØSJP BUSBWÏT QPSFYFNQMP EB"HFOEB5FSSJUPSJBM RVFSQPSQPMÓUJDBTTFUPSJBJT quer ainda, especiﬁcamente, pela Política de Coesão (estabelecendo várias referências com signiﬁcativo impacto no ordenamento do território), não se veriﬁca ainda uma estreita articulação entre elas, podendo o princípio de ct servir de plataforma de articulação dessas políticas. p Acresce, no contexto atual de políticas públicas, outro motivo para a adoção do princípio de ct. A adoção do princípio de coesão territorial no contexto atual das políticas públicas Não detendo competências formais no domínio do ordenamento do território, como já salientado, e apesar do contexto institucional favorável que hoje se veriﬁca no seio da ce (em que a adoção do princípio de ct no Tratado de Lis- CPBÏFYFNQMPEJTTP ÏFTTFODJBMNFOUFBUSBWÏTEFSFVOJÜFTJOGPSNBJTFOUSFNJOJTUSPT EFDBEB&TUBEP.FNCSPRVFTÍPEFCBUJEBT DPNBMHVNBQFSJPEJDJEBEF BTRVFTUÜFT de ordenamento do território, procurando harmonizar e corresponsabilizar a im- plementação de princípios e desaﬁos comuns e potenciais respostas. Neste sentido, como forma de harmonização e de difusão, a ue procura adotar mecanismos capazes de alterar culturas e modos de atuação territoriais nos Estados-Membros. Exemplos JODMVFNBJOUFHSBÎÍPEJTDVSTJWB DPOTUSVÎÍPEFWJTÜFTFTQBDJBJTQBSUJMIBEBTFQSÈUJDBT DPMBCPSBUJWBTUSBOTOBDJPOBJT 'FSSÍP & EFGBUP DFSUBTPSJFOUBÎÜFT WBMPSFT JEFOUJEBEFT NFUPEPMPHJBT FTUSBUÏHJBT UÐNTJEPJOEV[JEBTQPSEJTDVSTPTIFHFNÙ- nicos e políticas da euPMPOHPQSPDFTTPEFUSBCBMIPUÏDOJDP DJFOUÓmDPFQPMÓUJDP conjunto que culminou com a aprovação do edec, constitui um bom exemplo desta RVFTUÍP DSJBOEPDPOEJÎÜFTQBSBBIBSNPOJ[BÎÍPEFEJTDVSTPTFWJTÜFTQBSUJMIBEBT FBHFOFSBMJ[BÎÍPEFCPBTQSÈUJDBT%BNFTNBGPSNB PTEFCBUFTRVFDPOEV[JSBN mais recentemente à adoção quer da Agenda Territorial e do seu 1º Programa de Ação, quer do Livro Verde sobre a Coesão Territorial, contribuíram para consolidar este efeito edec, mobilizando diferentes comunidades nacionais e internacionais em torno de temas decisivos para a ue em geral, e para as políticas de ordenamento do território, em particular (Ferrão, 2010). Neste âmbito, a ct tornou-se um conceito ©EURE \| vol  \| no  \| enero  \| pp. -  catalisador, em torno do qual, diversos valores e políticas (territoriais e não territo- riais) têm gerado práticas discursivas. %FGBUP PEFTFOWPMWJNFOUPEFOPWPTDPODFJUPTSFQSFTFOUBÎÜFTNFUÈGPSBTOPTFJP da ue produzem vários efeitos: i) chamam a atenção das entidades nacionais compe- UFOUFTEPPSEFOBNFOUPEPUFSSJUØSJPQBSBRVFTUÜFTUFNÈUJDBTDPNVOTJJ BKVEBNB criar uma linguagem coletiva comum; iii) entram nos meios e nos debates políticos rapidamente, e com facilidade; e iv) procuram simpliﬁcar realidades complexas, pelo que constituem mensagens simples de uma ou duas palavras. Acresce que são conceitos que surgem da necessidade que as entidades sentem de formular quadros de referência, capazes de deﬁnir princípios orientadores para as políticas públicas (em geral) e o ordenamento do território (em particular). E, por isso mesmo, o aparecimento de conceitos como a ct representam a possibilidade de discutir as atuais tendências territoriais, e devem ser considerados como referenciais que permitem es- truturar diferentes valores e culturas, e promover um diálogo intersetorial (Tatzberger, 2006; Servillo, 2010). Dimensão heterogeneidade territorial Lidar com a heterogeneidade territorial e a complexidade das dinâmicas territo- riais signiﬁca, por um lado, reconhecer e combater a problemática associada às disparidades territoriais (amplamente divulgadas nos diversos relatórios produzi- dos, quer no âmbito da ocde, quer no seio da ce) e, por outro, tirar partido da diversidade territorial. A questão das disparidades desde cedo se constituiu em uma das principais QSFPDVQBÎÜFT JOFSFOUFT ËT QPMÓUJDBT QÞCMJDBT o WFKBTF QPS FYFNQMP P DBTP EB Política de Coesão da ueFBBUSJCVJÎÍPEPTGVOEPTFTUSVUVSBJTo DVKBTJOUFSWFOÎÜFT UÐNTJEP EFDFSUBGPSNB BODPSBEBTOPEFTFOWPMWJNFOUPEFFTUSBUÏHJBTEFQSPNPÎÍP de justiça social, um princípio normativo associado à equidade ou igualdade de PQPSUVOJEBEFT 4FO "BTTPDJBÎÍPEPDPODFJUPEFKVTUJÎBTPDJBMBRVFTUÜFT UFSSJUPSJBJTQPSBVUPSFTDPNP-FGFCWSF )BSWFZ :PVOH e, mais recentemente, Soja (2010) e Fainstein (2010), veio, não só contribuir para um aprofundamento da noção de território, que não deve ser encarado simplesmente FORVBOUPPCKFUPJOBOJNBEPFQBTTJWPPOEFPDPSSFNQSPDFTTPTTPDJBJT DPNPUBNCÏN chamar a atenção para a relação que existe entre esta dimensão e a sociedade. Aliás, ÏOFTUFDPOUFYUPRVFFTUFTBVUPSFTBWBOÎBNDPNBOPÎÍPEFKVTUJÎBFTQBDJBM FN DPNVNBSHVNFOUBOEPRVFBQSPDVSBEFKVTUJÎBJNQMJDBQFSDFCFSBSFMBÎÍPEJBMÏUJDB FOUSF OÍPTØBTDPOEJÎÜFTFDPOÙNJDBTFTPDJBJTEFEJGFSFOUFTHSVQPT NBTUBNCÏN a geograﬁa da injustiça, não sendo possível, portanto, compreender a questão da equidade sem considerar de que forma os serviços, as oportunidades e o próprio poder institucional se encontram geograﬁcamente distribuídos. Reconhecendo, então, que as disparidades territoriais merecem a maior atenção possível, torna-se crucial intervir sobre os territórios se os quisermos mais equitativos, devendo, neste sentido, a distribuição de recursos, serviços e o seu acesso ser considerado um direito humano básico. E, de fato, esta foi, de algum modo, a primeira preocupação explícita por trás da noção de ct, do ponto de vista das políticas públicas da ce, na procura do que se designa por um desenvolvimento territorial homogêneo da ue e em cada um dos Estados-Membros, como ﬁcou explícito na narrativa histórica apresentada anteriormente. Neste sentido, abraçar a problemática associada às disparidades territoriais im- plica lidar, de forma diferenciada, as áreas com as características geográﬁcas (eco- MØHJDBT TPDJBJTFFDPOÙNJDBT FTQFDJBJTODVUF QPJT VNKVÓ[PEFEJTDSJNJOBÎÍP positiva, atribuindo uma dimensão de justiça espacial à política pública e apelando Ë TPMJEBSJFEBEF OÍP TØ FOUSF DJEBEÍPT NBT UBNCÏN FOUSF UFSSJUØSJPT %BWPVEJ 2009). Signiﬁca que a ação pública deverá ser diferenciada por uma questão moral e de acordo com as necessidades existentes (Boyne, Powell, & Ashworth, 2001; Harvey, 1973), tratando de forma diferente realidades que são estruturalmente dis- tintas, pelo que a solução one-size-ﬁts-allÏJOTVmDJFOUFQBSBMJEBSDPNBIFUFSPHF- neidade territorial existente. Coesão territorial: uma interpretação processual O debate acadêmico e político em torno da ct tem sido amplo e incorpora inúme- SBTDPOUSJCVJÎÜFTEPQPOUPEFWJTUBIJTUØSJDP QPSRVFFDPNPTVSHJV FDPODFJUVBM (seu propósito), seja ao nível da literatura acadêmica no domínio do ordenamento do território, da geograﬁa e da economia, seja no nível de documentos oﬁciais e PSJFOUBÎÜFTEBce. Embora, como conceito e objetivo político, este princípio tenha vindo a adquirir uma importância elevada, a consulta e análise desta extensa do- DVNFOUBÎÍPNPTUSBRVFPTFVTJHOJmDBEPBJOEBOÍPÏDPOTFOTVBM EJmDVMUBOEPB capacidade de lidar, de forma analítica e normativa, com o princípio. Mas, mais do que a procura de uma deﬁnição precisa, importa clariﬁcar o seu conteúdo efetivo e GVODJPOBM VNQFSDVSTPRVFUBNCÏNTFFODPOUSBQPSUSJMIBS "QFTBSEFUVEP VNBMFJUVSBTJTUÏNJDBEFTUFDPSQPUFØSJDPFEBTRVFTUÜFTDIBWF que levanta, permite formalizar um conjunto de dimensões analíticas subjacentes ao princípio de ct, designadamente: i) reconhecer e lidar com a heterogeneidade terri- torial e a complexidade das dinâmicas territoriais; ii) estabelecer uma organização UFSSJUPSJBMRVFQSPNPWBBJOUFOTJmDBÎÍPEBTSFMBÎÜFTJOUFSVSCBOBTFBDPNQMFNFO- UBSJEBEFEFVTPT GVOÎÜFTFDPNQFUÐODJBTFOUSFÈSFBTVSCBOBTFFTQBÎPTSVSBJTJJJ EFTFOWPMWFSJOUFSWFOÎÜFTJOUFHSBEBTBTTFOUFTFNTJTUFNBTEFHPWFSOBOÎBUFSSJUPSJBM H H 0BSHVNFOUPRVFBRVJTFBQSFTFOUBÏPEFRVFBQFSDFQÎÍPFPBQSPGVOEBNFOUP EFTUBTUSÐTEJNFOTÜFTQFSNJUFNDMBSJmDBSPTJHOJmDBEPFMBOÎBSBTCBTFTQBSBBPQF- racionalização do princípio de ct. Sua análise será efetuada tendo por base duas QSFPDVQBÎÜFTQPSVNMBEP QSPDVSBOEPFOUFOEFSPRVFDBSBDUFSJ[BDBEBEJNFOTÍPF clariﬁcando de que forma estas se relacionam com a ideia de ct; por outro, preten- dendo enquadrar e balizar a análise a efetuar, numa ótica de intervenção de políticas públicas. A análise culminará na apresentação, na próxima seção, de um quadro síntese de ideias-chave, que servirá como referencial para a formulação de políticas públicas tendo por base o princípio de ct. Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... \| ©EURE  Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... \| ©EURE  ©EURE \| vol  \| no  \| enero  \| pp. -  a esta questão, aﬁrmando: “a Europa enfrentará nos próximos anos um número de EFTBmPT DIBWF JODMVJOEP BEBQUBÎÍP Ë HMPCBMJ[BÎÍP ËT BMUFSBÎÜFT EFNPHSÈmDBT F DMJNÈUJDBT FBPEFTBmPFOFSHÏUJDP5PEBTBTSFHJÜFT&VSPQFJBTTFSÍPBGFUBEBT/P FOUBOUP DBEBVNEFTUFTEFTBmPTFYJCFVNQBESÍPEJTUJOUPw Q 3FDPOIFDFSBIFUFSPHFOFJEBEFUFSSJUPSJBMJNQMJDB BTTJN UBNCÏNUJSBSQBSUJEP da diversidade dos territórios, ou seja, considerar as potencialidades e as fragilidades territoriais diferentes, o que consequentemente se traduz na análise das capacida- des (bens, capitais, recursos…) existentes para formular políticas territorialmente EJWFSTJmDBEBT6NBQFSTQFDUJWBRVFUFNTJEPBNQMBNFOUFEFGFOEJEBOPTUFNQPT NBJTSFDFOUFTÏBRVFQSPDVSBBTTPDJBSBJEFJBEFRVFVNUFSSJUØSJPRVFUJSBQBSUJEP EBTVBEJWFSTJEBEF FSFTQFDUJWBTTJOHVMBSJEBEFT ÏVNUFSSJUØSJPNBJTDPFTPFDPN NBJPSQPUFODJBMQBSBTFUPSOBSNBJTDPNQFUJUJWPFSFTJMJFOUF )BNEPVDI %FQSFU Tanguy, 2011), ou seja, com maior capacidade de se antecipar e se adaptar, face às adversidades e às grandes tendências evolutivas. Está-se, aqui, perante o que na lite- SBUVSBTFEFTJHOBQPSWBMPSJ[BSPDBQJUBMUFSSJUPSJBM JTUPÏ TFSDBQB[EFUJSBSQBSUJEP EPTFMFNFOUPT DBSBDUFSÓTUJDBTMPDBJTQPSWF[FTÞOJDBT EFRVFEJTQÜFPUFSSJUØSJPBP nível material (na forma de recursos naturais e culturais, bens públicos e a estrutura urbana em geral), e imaterial (no que respeita à capacidade relacional, ao empreen- dedorismo, à criatividade e à liderança) (Camagni & Capello, 2010). 0SB FTUBÏBQPTUVSBRVFBQSØQSJBce tem crescentemente defendido no âmbito EBBUSJCVJÎÍPEPTGVOEPTFTUSVUVSBJTFBNQMBNFOUFEJGVOEJEP TFKBBUSBWÏTEB1P- MÓUJDBEF$PFTÍP TFKBQPSJOUFSNÏEJPEB&TUSBUÏHJBEuropa 2020. Esta mudança de orientação assenta, pois, no pressuposto de que a coesão não decorrerá somente da DPODFOUSBÎÍPEFSFDVSTPTDPNVOJUÈSJPTOBTSFHJÜFTNFOPTEFTFOWPMWJEBT QPTUVSB tradicional de uma política redistributiva, como resposta a assimetrias regionais de EFTFOWPMWJNFOUP"NFMIPSJBEFTUBTSFHJÜFTFYJHFJOWFTUJNFOUPTFMFUJWP RVBMJmDB- dor e capaz de valorizar os seus recursos endógenos (Santinha et al., 2009). Por PVUSBTQBMBWSBT BTJOUFSWFOÎÜFTQÞCMJDBTEFWFNEFJYBSEFTFSUFOEFODJBMNFOUFEFT- FOIBEBTBQBSUJSEFTJUVBÎÜFTQSPCMFNÈUJDBT QBSBQBTTBSFN UBNCÏN BDPOTJEFSBSP território enquanto recurso fundamental para aumentar o leque de oportunidades de quem aí vive ou desenvolve a sua atividade, e enquanto fonte de qualidade de vida e bem-estar para as pessoas, e de conhecimento e bom desempenho para as PSHBOJ[BÎÜFT 'FSSÍP Dimensão heterogeneidade territorial Acresce que, esta ação pública diferenciada inﬂuencia igualmente a forma como em cada território se consegue lidar com os mais diversos EFTBmPTRVFBTPDJFEBEFFNHFSBMFOGSFOUBFFOGSFOUBSÈBDVSUPFNÏEJPQSB[P0 documento Regions 2020 da ce RVFDPOTUJUVJVNQPOUPEFSFGFSÐODJBQBSB BEFmOJÎÍPEFPSJFOUBÎÜFTFNNBUÏSJBEF1PMÓUJDBEF$PFTÍP EFEJDBTFQSFDJTBNFOUF ©EURE \| vol  \| no  \| enero  \| pp. - Dimensão organização territorial A ideia subjacente a esta questão reside nas próprias palavras-chave que o Livro Verde menciona, designadamente: concentração (ultrapassar as diferenças de den- TJEBEF FWJUBOEPDPODFOUSBÎÜFTFYDFTTJWBTEFDSFTDJNFOUPFGBDJMJUBOEPPBDFTTPË QBSUJMIBEPTCFOFGÓDJPTQSPQPSDJPOBEPTQFMBTBHMPNFSBÎÜFTFYJTUFOUFTFNUPEPTPT territórios); conectividade (melhorar o acesso a Serviços de Interesse Geral de forma TPDJBMNFOUFFFTQBDJBMNFOUFFRVJUBUJWB FDPPQFSBÎÍP FMJNJOBSEJWJTÜFT OBNFEJEB em que os problemas da conectividade e da concentração só podem ser eﬁcazmente resolvidos com a estreita cooperação entre todos os níveis de intervenientes). Poderá argumentar-se que a lógica por detrás da (des)concentração, conectividade e coope- ração está associada aos conceitos de desenvolvimento policêntrico e nova relação Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... \| ©EURE  VSCBOPSVSBM que têm estado na base de grande parte das políticas de ordenamento EPUFSSJUØSJPEFTUBÞMUJNBEÏDBEBFRVFTVSHFNJOTUJUVDJPOBMJ[BEPTOPedec. A procura de uma Europa mais policêntrica e territorialmente equilibrada surge como resposta às persistentes disparidades territoriais veriﬁcadas no espaço euro- QFVFËTEJTDVTTÜFTFNUPSOPEPEFTFOWPMWJNFOUPUFSSJUPSJBM RVFFSBNEPNJOBEBT QFMPQFOTBNFOUPDPOWFODJPOBMDFOUSPQFSJGFSJB FNRVFBÈSFBNBJTEFTFOWPMWJEB economicamente era circunscrita por Londres, Paris, Milão, Munique e Hamburgo (designado no próprio edec por Pentágono) e, a restante (periferia), necessitava de auxílio urgentemente. A noção de desenvolvimento policêntrico, não só introduz no debate uma interpretação mais soﬁsticada da organização territorial da Europa, sublinhando o potencial que outros centros (fora da área dominante) possuem .FJKFST 8BUFSIPVU;POOFWFME DPNPUBNCÏNQSFTTVQÜF A convicção de que uma Europa mais forte necessita de um maior número de polos capazes de competir com sucesso nos mercados globais. Por outro lado, a ideia de que uma Europa mais equilibrada implica a existência de polos de valia europeia tanto no seu core como nas suas várias periferias. Finalmente, a defesa de que uma WJTÍPTJTUÐNJDBEPDPOKVOUPEPFTQBÎPFVSPQFVQFSNJUJSÈSFDPSSFSËTFTUSBUÏHJBT transnacionais de cooperação e complementaridade como meio ativo para valorizar de forma mais inteligente os recursos de toda a Comunidade (Ferrão, 2002, p. 33). A lógica por detrás da adoção do desenvolvimento policêntrico, enquanto obje- tivo político favorável para o território Europeu, residiu igualmente no reconhe- DJNFOUPEB&VSPQBTFOUJSPTFGFJUPTEFDPSSFOUFTEPGFOÙNFOPEFNFUSPQPMJ[BÎÍP associados, em parte, aos custos da sobre-concentração (poluição urbana, dispo- nibilidade de solo, tempo de deslocação, fragmentação social…), e à perda de JEFOUJEBEFUFSSJUPSJBM BPTRVBJTTFBEJDJPOBNQSFPDVQBÎÜFTDPNBQSPDVSBEFVN desenvolvimento mais sustentável. ©EURE \| vol  \| no  \| enero  \| pp. -  Associado ao desenvolvimento policêntrico, surge a ideia da nova parceria VSCBOPSVSBM DPNP VNB GPSNB EF JOUFHSBS BT ÈSFBT SVSBJT OVNB QPMÓUJDB EF ordenamento conjunta, procurando uma situação de win-win %F BDPSEP DPN FTUB ØUJDB B EJDPUPNJB SVSBMVSCBOP o FN RVF TF WFSJmDBN EFUFSNJOBEBT características contrastantes tradicionais como isolamento versus acessibilidade, FOWFMIFDJNFOUP WFSTVT KVWFOUVEF EFDMÓOJP WFSTVT EJOBNJTNP FDPOÙNJDP EFTFSUJmDBÎÍPWFSTVTEFOTJEBEFQPQVMBDJPOBM 'FSSÍP C oÏTVCTUJUVÓEBQPS novos relacionamentos “retratando as interdependências funcionais e espaciais e a OFDFTTJEBEFEFQSPNPWFSVNBNBJPSJOUFHSBÎÍPFDPNQMFNFOUBSJEBEFUFSSJUPSJBMw (Marques, 2003, p. 507). Está-se, portanto, perante uma nova forma de deﬁnir QPMÓUJDBTQBSBPFTQBÎPSVSBMoBRVFBocde (2006) designou por novo paradigma de desenvolvimento rural, e que visa o incremento de competitividade dessas ÈSFBT F B WBMPSJ[BÎÍP EPT SFDVSTPT FOEØHFOPT o F QBSB P FTQBÎP VSCBOP DVKB participação em lógicas reticulares sistêmicas e de proximidade (seja ao nível de nVYPTEFQFTTPBT CFOTFJOGPSNBÎÍP TFKBQPSJOUFSNÏEJPEBQSFTUBÎÍPEFTFSWJÎPT seja ainda ao nível de políticas) pode constituir um fator, não só integrador, mas UBNCÏNBMBWBODBEPSEFEFTFOWPMWJNFOUPQBSBPTEPJTFTQBÎPT Dimensão organização territorial "MDBOÎBSVNEFTFOWPMWJNFOUPNBJTQPMJDÐOUSJDPQSFTTVQÜFVNBPSHBOJ[BÎÍP UFSSJUPSJBMDBQB[EFFODPSBKBSBDPPQFSBÎÍP JTUPÏ BQUBQBSBQSPNPWFSVNBNBJPS complementaridade entre as áreas urbanas, de forma a que estas desempenhem um papel estrutural ao criar espaços equilibrados e, ao mesmo tempo, desenvolver NBTTBDSÓUJDBBPOÓWFMFDPOÙNJDPHMPCBM FNVNBQFSTQFDUJWBDPTNPQPMJUB"TTJN sendo, o desenvolvimento policêntrico baseia-se na ideia de que i) a organização do território assenta nas cidades, pelo que importa desenvolver a sua capacidade de atratividade e desenvolvimento, ii) as cidades constituem os nós com os quais se criará uma rede interurbana, cuja integração poderá ser intensiﬁcada com a cons- tituição de redes de sinergia (cooperação entre áreas urbanas com perﬁs funcionais idênticos para a obtenção de massa crítica suﬁciente e troca de experiências) e de complementaridade (cooperação entre áreas urbanas com perﬁs complementares EFFTQFDJBMJ[BÎÍPGVODJPOBMFQSPEVUJWB QSØYJNBTFOUSFTJ 'FSSÍP .FJKFST FUBM FJJJ BTJOUFSBÎÜFTTÍPGBDJMJUBEBT RVFSWJBSFEFTEFJOGSBFTUSVUVSBT físicas e virtuais (com recurso às tecnologias de informação e comunicação), quer QPSJOUFSNÏEJPEFSFMBÎÜFTJOTUJUVDJPOBJTFOUSFEJGFSFOUFTBHFOUFT WJBFTUSBUÏHJBT conjuntas de desenvolvimento. ©EURE \| vol  \| no  \| enero  \| pp. - Dimensão governança territorial O termo governança tem sido crescentemente usado nos domínios acadêmico e polí- UJDP FTQFDJBMNFOUFBQBSUJSEBEÏDBEBEF FÏBUVBMNFOUFDPOTJEFSBEPDPNPGBUPS crucial para a implementação de políticas públicas (veja-se, por exemplo, o relatório Barca, 2009). A emergência deste tema e o respetivo crescente interesse no contexto Europeu estão, de certa forma, relacionados com a discussão em torno das transfor- NBÎÜFTEPT&TUBEPTEB&VSPQB RVFSFNUFSNPTEFBSUJDVMBÎÍPUFSSJUPSJBM QSPDVSBOEP BEJWJTÍPEFDPNQFUÐODJBTWJBEFTDFOUSBMJ[BÎÍPSFHJPOBMJ[BÎÍP 4IBSQF RVFSFN termos de processos de tomada de decisão, com o envolvimento e a mobilização de diferentes atores públicos e não públicos (Rhodes, 1997), quer ainda no que respeita à FVSPQFJ[BÎÍPEFJEFJBT QSJODÓQJPT PSJFOUBÎÜFTFQSÈUJDBTOPÉNCJUPEFQPMÓUJDBTTPCJO- ﬂuência do processo de construção do projeto europeu, com impacto signiﬁcativo em diversas estruturas organizacionais e administrativas dos diversos Estados-Membros 'FSSÍP BC É extensa a literatura que se debruça sobre a temática da governança no campo das DJÐODJBTQPMÓUJDBT EBTPDJPMPHJBQPMÓUJDBFEBTDJÐODJBTFDPOÙNJDBT NBTPTFVDPODFJUPF PTDPOUFYUPTPOEFÏBQMJDBEPOÍPTÍPDPOTFOTVBJT QFMPRVFOÍPFYJTUF a priori, um me- canismo de governança superior aos demais. Apesar de tudo, há um reconhecimento generalizado de que o termo governança se refere ao desenvolvimento de diferentes for- mas de atuação, em que a fronteira entre os setores público e privado têm se atenuado 3IPEFT 4UPLFS 6NBBOÈMJTFEBFWPMVÎÍPIJTUØSJDBEPQBQFMEP&TUBEPOBGPSNVMBÎÍPEFQPMÓUJDBT públicas permite veriﬁcar que há uma relação estreita entre estes mecanismos e os prin- DÓQJPTRVFDBSBDUFSJ[BNBHPWFSOBOÎBQÞCMJDB%FGBUP EVSBOUFWÈSJBTEÏDBEBT QBSUJDV- MBSNFOUFOPQFSÓPEPRVFEFDPSSFEFTEFPQØT(VFSSB.VOEJBMBUÏBEÏDBEBEF veriﬁcou-se um consenso generalizado em torno do papel do Estado, considerado cru- cial para responder às necessidades dos cidadãos, não só para permitir a reconstrução Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... \| ©EURE  EP&TUBEPOBÎÍPEPQPOUPEFWJTUBFDPOÙNJDP NBTUBNCÏNQPSRVFIBWJBBOFDFTTJ- EBEFEFTFFOWPMWFSOBHFTUÍPEFDPOnJUPTFDPODFSUBÎÜFTFOUSFDMBTTFTFHSVQPTTPDJBJT de forma a minimizar as desigualdades existentes. Neste período de intervenção direta do Estado na sociedade (Estado Providência ou de bem-estar), as despesas públicas eram interpretadas como um investimento coletivo (Hughes, 2003). Contudo, por um con- KVOUPEFSB[ÜFTFOEØHFOBTBP&TUBEP OPNFBEBNFOUFPBVNFOUPEBEFTQFTBQÞCMJDB e exógenas (como, por exemplo, a crise global desencadeada pelo aumento dos preços do petróleo), aliadas à eleição de Margaret Tatcher para o governo britânico (1979) e Ronald Reagan para os eua oRVFBQSFTFOUBWBNVNEJTDVSTPGBDJMJUBEPSEF EJNJOVJÎÍPEFSFDFJUBT WJBSFEVÎÍPEPTJNQPTUPT FEBEFTQFTBQÞCMJDB BUSBWÏTEFNF- DBOJTNPTEFQSJWBUJ[BÎÍPPVPVUSBTGPSNBTEFDPOUSBUPT oTVSHJSBNSFGPSNBTRVF EF uma forma geral, introduziriam o movimento da Nova Gestão Pública e cujo período EFBmSNBÎÍPEFDPSSF TFOTJWFMNFOUF FOUSFFPmOBMEPTBOPT )VHIFT 4JMWFTUSF "JEFJBTVCKBDFOUFÏBEFVTBSBTBCPSEBHFOTFNQSFTBSJBJTEPTFDUPS privado no setor público e, consequentemente, tornar este mais magro e competitivo BUSBWÏT QPSFYFNQMP EPBVNFOUPEPSFDVSTPBNFDBOJTNPTUJQPNFSDBEP DPNPB contratação externa, e da privatização de empresas públicas) (Lane, 2000). ©EURE \| vol  \| no  \| enero  \| pp. -  MPDBM BRVFFTUÈBTTPDJBEPPQSJODÓQJPEFTVCTJEJBSJFEBEF RVFSFOUSFFOUJEBEFT BUPSFTEPNFTNPOÓWFM JEFJBEFDPPQFSBÎÍPDPPSEFOBÎÍPIPSJ[POUBMFOUSFEJWFSTPT BUPSFT JODMVJOEPFTGFSBTQÞCMJDB QSJWBEBFUFSDFJSPTFUPS RVFTUÜFTEFDJTJWBTQBSB BGPSNVMBÎÍPEFQPMÓUJDBTFDPOTUSVÎÍPEFFTUSBUÏHJBTDPMFUJWBTUFSSJUPSJBJT #PULB 2009; Feio & Chorincas, 2009; Janin Rivolin, 2010). J &TUBDPPQFSBÎÍPJOTUJUVDJPOBM TFKBWJBSFQBSUJÎÍPWFSUJDBMEFBUSJCVJÎÜFTFDPN- petências numa ótica de subsidiariedade, seja via cooperação entre atores baseada OVNBWJTÍPQBSUJMIBEBTPCSFVNNFTNPUFSSJUØSJP QSFTTVQÜFUBNCÏNBCPSEBHFOT BSUJDVMBEBTFJOUFHSBEBTEBTWÈSJBTEJNFOTÜFTTFUPSJBJT PVTFKB TVCFOUFOEFBQSPDVSB EFDPFSÐODJBEFQPMÓUJDBT UBNCÏNFTUBTCBTFBEBTOVNBWJTÍPQBSUJMIBEBTPCSFVN mesmo território. Esta ideia de articular, integrar e territorializar as políticas públi- DBToJODPSQPSBOEPBEJNFOTÍPUFSSJUPSJBMOBTQPMÓUJDBTTFUPSJBJTFBSUJDVMBOEPFTUBT com as políticas de base territorial, contribuindo-se assim para uma maior sinergia FOUSFEJGFSFOUFTQPMÓUJDBTFQBSBBNBYJNJ[BÎÍPEPTTFVTJNQBDUPTUFSSJUPSJBJToSF- TVMUBEPGBUPEFBTEFDJTÜFTQPMÓUJDBTTFSFNUPNBEBTFTTFODJBMNFOUFEFVNQPOUP EFWJTUBTFUPSJBMFEFGPSNBEFTBSUJDVMBEBDPNBTQPMÓUJDBTUFSSJUPSJBJT&TUBÏVNB questão mencionada por vários documentos de orientação política da ce, sendo sublinhado que a articulação, integração e territorialização das políticas públicas tornam-se cruciais, quer para uma maior eﬁciência das próprias políticas, quer para tornar os territórios mais resilientes, capazes de se anteciparem e adaptarem às gran- EFTUFOEÐODJBTFWPMVUJWBTFJNQBDUPTEFBTQFDUPFDPOÙNJDP TPDJBM BNCJFOUBMPV mesmo político. Tal implica considerar o território enquanto contexto diferencia- dor de políticas sectoriais, tornando-as consequentemente mais sensíveis ao territó- rio, e enquanto referência para a coordenação de políticas públicas, constituindo as QPMÓUJDBTEFCBTFUFSSJUPSJBMVNSFGFSFODJBMFTUSBUÏHJDPRVFEÈinputs importantes às políticas setoriais. É neste contexto que o território passa a ser considerado o lugar, por excelência, de articulação e racionalidade de políticas públicas (Feio & Chorin- cas, 2009), reforçando a visibilidade, a legitimidade e a eﬁcácia das políticas de base territorial, por um lado, e fomentando a simbiose entre as políticas setoriais sensí- veis ao território e entre estas e as políticas de base territorial, por outro, tornando- as mais capazes de sustentar processos de desenvolvimento territorial. Dimensão governança territorial Apesar de tudo, foram várias as críticas apontadas a este movimento, associadas essencialmente à sua crença nas forças do mercado enquanto panaceia de todos os problemas e en- quanto mecanismo capaz de representar o interesse social geral, pelo que se procurou VNBOPWBBHFOEBSFGPSNJTUB EFDFSUBGPSNB KÈWJTÓWFMOBUSBOTJÎÍPEPTÏDVMP"QFTBS de apresentar alguns pontos de continuidade com a Nova Gestão Pública (realçando os mecanismos de mercado e as parcerias público-privadas como fatores-chave para a reforma administrativa), neste novo período, procura-se não enfatizar a importância da competição, mas antes realçar os benefícios da colaboração, designadamente entre Estado, Mercado e Sociedade Civil (Newman, 2002). Esta diferença implica (poten- DJBMNFOUF BMUFSBÎÜFTOBMØHJDBEFUPNBEBEFEFDJTÍPFOBTSFMBÎÜFTJOUFSJOTUJUVDJPOBJT sublinhando-se a necessidade de uma maior e melhor integração horizontal (entre setor QÞCMJDP TFUPSQSJWBEPFTFDUPS FWFSUJDBM FOUSFBTBENJOJTUSBÎÜFTDFOUSBM SFHJPOBM FMPDBMFFOUSFRVFNEFmOFBTQPMÓUJDBTFRVFNÏBGFUBEPQPSTVBTEFDJTÜFT %FOIBSEU %FOIBSEU Ora, se a este contexto de mudança se associa a dimensão territorial, ﬁca-se na QSFTFOÎBEFVNRVBESPEFHPWFSOBCJMJEBEFUFSSJUPSJBMNBJTJOUFHSBEPFFTUSBUÏHJDP assente na interação e cooperação de diferentes atores e na coordenação de diferentes políticas (de base territorial e setoriais) em torno de uma agenda territorial comum (Ferrão, 2010). Autores como Feio e Chorincas (2009) argumentam que a governança territorial engloba a: Capacidade de os atores públicos e privados de um determinado território conseguirem: i) construir um ‘consenso organizacional’, envolvendo diferentes atores, de forma a deﬁnir ob- jetivos e metas comuns; ii) acordar a contribuição de cada ator para a realização dos objeti- vos previamente deﬁnidos e iii) acordar a visão comum para o futuro do território (p. 140). A noção de governança territorial está, portanto, associada à ideia de uma articu- lação mais eﬁciente, quer entre diferentes níveis de administração (ideia de gover- OBOÎBNVMUJOÓWFMPVEFDPPQFSBÎÍPDPPSEFOBÎÍPWFSUJDBM EBFTDBMBFVSPQFJBËFTDBMB ©EURE \| vol  \| no  \| enero  \| pp. - Comentário ﬁnal e apresentação de quadro síntese A incursão histórica e teórica, e a reﬂexão aqui apresentadas em torno da ct mos- USBNRVFFTUFQSJODÓQJPÏIPKFBNQMBNFOUFSFDPOIFDJEP UBOUPOBBSFOBQPMÓUJDB como no debate acadêmico, enquanto novo paradigma de desenvolvimento, tor- nando-se um veículo catalisador e atuando inclusive como referência normativa e cognitiva favorável ao processo de integração europeia. Mas, porque a adoção deste princípio vem acrescentar valor à tomada de decisão política? Em primeiro lugar, QFMPTFOUJEPEFDPFTÍP JTUPÏ QFMBQSPNPÎÍPEFMJHBÎÜFTFSFMBÎÜFTGPSUFTFOUSFJO- EJWÓEVPTPVJOTUJUVJÎÜFTQBSBBMDBOÎBSmOBMJEBEFTDPNVOT BUSBWÏT QPSVNMBEP EF VNNBJPSFRVJMÓCSJPTPDJBMFFDPOÙNJDP OBEVQMBØUJDBEFEJTUSJCVJÎÍPEFSFDVSTPT e de participação nos processos de tomada de decisão, e, por outro, de uma afetação de recursos mais eﬁciente, valorizando a diversidade e a singularidade. Em segundo Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... \| ©EURE  lugar, pela importância da dimensão territorial que completa o próprio conceito de coesão: a própria ce vem acrescentar esta componente na sua Política de Coesão BPTPVUSPTPCKFUJWPTEFDPFTÍPTPDJBMFFDPOÙNJDB PRVF OVNQFSÓPEPEFQBSUJDV- lar importância devido à atribuição de fundos comunitários para o pós-2013, vem introduzir uma racionalidade adicional aos processos de tomada de decisão. Con- tudo, a abrangência e a falta de consenso, não só em torno do seu signiﬁcado, mas, sobretudo, no que respeita ao seu conteúdo efetivo e funcional, têm diﬁcultado a capacidade de utilizar este princípio nos processos de tomada de decisão. Ora, foi este mesmo o principal propósito do presente texto: avançar com um possível caminho para tornar o princípio de ct mais compreensível e operacional, em uma ótica de intervenção de políticas públicas. "T EJNFOTÜFT BOBMÓUJDBT BRVJ BQSPGVOEBEBT o heterogeneidade territorial, organização territorial e governança territorialoQSPDVSBN FNDPOKVOUP DPOUSJCVJS para uma melhor percepção do conceito em discussão, delineando um percurso analítico e normativo para alcançar o desenvolvimento harmonioso de todos os UFSSJUØSJPT JTUPÏ EJNJOVJSQSFWFOJSBTEJTQBSJEBEFTUFSSJUPSJBJT UPSOBOEPPTNBJT IPNPHÐOFPTJOUFSOBNFOUFFDPNJOUFSSFMBÎÜFTFRVJMJCSBEBT BWBMPSJ[BÎÍPEBTVB diversidade e complementaridades e a possibilidade da população tirar o melhor partido das características existentes em cada território, aliada a uma capacidade de interação e cooperação de diferentes agentes e políticas distintas (de base territorial F TFUPSJBJT 0 2VBESP TVNBSJB FTUBT EJNFOTÜFT BOBMÓUJDBT EFDPSSFOUFT EB interpretação efetuada do princípio de ct, visando criar um referencial que permita JEFOUJmDBSFTFMFDJPOBSQPTTÓWFJTDSJUÏSJPTEFJOUFSWFOÎÍPEFQPMÓUJDBTQÞCMJDBT©EURE ©EURE \| vol  \| no  \| enero  \| pp. Comentário ﬁnal e apresentação de quadro síntese -  quadro 1 \| Dimensões analíticas do princípio de coesão territorial e respectivos componentes de intervenção componente analítica componente de intervenção (políticas públicas) dimensões subdimensões finalidade objetivos Heterogeneidade Territorial %JTQBSJEBEFT Justiça espacial Combater a marginalização de pessoas em contextos territoriais especíﬁcos Tornar os territórios mais resilientes %JWFSTJEBEF Valorização do capital territorial Criar oportunidades a partir dos ativos especíﬁcos de cada território Promover a identidade territorial como GBUPSEFVOJEBEFEJGFSFODJBÎÍP Organização Territorial 4JTUFNB6SCBOP (policentrismo) %FTFOWPMWJNFOUP urbano equilibrado e em rede %FTFOWPMWFSDBQBDJEBEFTEF atratividade e competitividade dos centros urbanos Promover a complementaridade funcional interurbana (massa crítica em rede) Estimular uma cultura de cooperação FTUSBUÏHJDBJOUFSVSCBOB 6SCBOPSVSBM (parceria urbano- rural) Integração territorial de proximidade 3FGPSÎBSBTSFMBÎÜFTGVODJPOBJTFP desenvolvimento de redes (ﬂuxos) entre os centros urbanos e as áreas rurais %FTFOWPMWFSNFDBOJTNPTEF interdependência e de complementaridade entre centros urbanos e áreas rurais Estimular uma cultura de participação em lógicas reticulares sistêmicas e de proximidade Governança Territorial Coordenação vertical Governança multinível Clariﬁcar e articular a repartição WFSUJDBMEFBUSJCVJÎÜFTFDPNQFUÐODJBT (ótica de subsidiariedade) Melhorar a capacidade de coordenação e governabilidade entre diferentes escalas de intervenção Coordenação horizontal Cooperação institucional e coerência de políticas %FTFOWPMWFSGPSNBTEFDPPQFSBÎÍP entre diversos atores baseada numa visão partilhada sobre um mesmo território %FTFOWPMWFSGPSNBTEFEJTDVTTÍPF participação ativa dos cidadãos nos processos de tomada de decisão %FTFOWPMWFSBCPSEBHFOTBSUJDVMBEBT integradas e territorializadas das várias políticas públicas fonte elaboração própria Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... \| ©EURE Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... \| ©EURE   Referências bibliográﬁcas Albrechts, L. (2001). In pursuit of new approaches to strategic spatial planning. A European perspective. International Planning Studies, 6(3), 293-310. doi: "ULJOTPO 3 ɨF FNFSHJOH AVSCBO BHFOEB BOE UIF &VSPQFBO TQBUJBM EFWFMPQNFOU perspective: towards an eu urban policy? European Planning Studies, 9 #FSUSBOE /1FZSBDIF(BEFBV 7 A$PIÏTJPOTPDJBMFFUDPIÏSFODFUFSSJUPSJBMFRVFM DBESFEFSÏnFYJPOTQPVSMBNÏOBHFNFOUFUMFEÏWFMPQQFNFOU Géographie, Économie, Société, 11 %JTQPOÓWFMem XXXDBJSOJOGPSFWVFHFPHSBQIJFFDPOPNJF TPDJFUFQBHFIUN #ÚINF , "/PSUIFSO7JFXPOesdp. North Boyne, G.; Powell, M. & Ashworth, R. (2001). Spatial equity and public services: an empirical analysis of local government ﬁnance in England. Public Management Review, 3(1), EPJ Camagni, R. & Capello, R. (2010). Macroeconomic and territorial policies for regional com- petitiveness: an eu perspective. Regional Science Policy & Practice, 2(1), 1-19. doi: KY %BWPVEJ 4 5FSSJUPSJBMDPIFTJPO &VSPQFBOTPDJBMNPEFMBOEUSBOTOBUJPOBMDPPQFSB- tion. En J. Knieling & F. Othengrafen (Eds.), Planning cultures in Europe. Decoding cultural phenomena in urban and regional planning (pp. 269-279). Aldershot: Ashgate. %FOIBSEU +7%FOIBSEU 3# The new public service: serving, not steering/FX:PSL M. E. Sharpe. Fainstein, S. (2010). The just city. Ithaca, ny$PSOFMM6OJWFSTJUZ1SFTT Faludi, A. (2009). Territorial cohesion under the looking glass: Synthesis paper about the history of the concept and policy background to territorial cohesion. Recuperado de sito Web EF&VSPQFBO$PNNJTTJPO 3FHJPOBM1PMJDZOGPHFSJPIUUQFDFVSPQBFVSFHJPOBM@ QPMJDZBSDIJWFDPOTVMUBUJPOUFSDPQEGMPPLJOHHMBTTQEG Faludi, A. (2010, julio). Territorial cohesion post-2013: to whomsoever it may concern. Artículo presentado en 24th aesop"OOVBM$POGFSFODF %FMGU6OJWFSTJUZPG5FDIOPMPHZ 0UC Research Institute for the Built Environment, Finland. %JTQPOÓWFM FO IUUQCJU MZE,S/ Feio, P. & Chorincas, J. (2009). Governação territorial e inovação das políticas públicas. Prospectiva e Planeamento, 16 %JTQPOÓWFMFNIUUQCJUMZ+N 'FSSÍP + B 3FEVSCBOB JOTUSVNFOUPEFFRVJEBE DPIFTJØOZEFTBSSPMMPEURE, 24(71), %JTQPOÓWFMFNIUUQCJUMZ3YN&P 'FSSÍP + C 3FDPOTUSVJS P JOUFSJPS EFTUSVJOEP B JOUFSJPSJEBEF QBSB VNB FTUSBUÏHJB activa de inclusão de actores. En Presidência da República (Ed.), Perspectivas de desenvolvimento do Interior. Lisboa: Presidência da República e Imprensa Nacional $BTBEB.PFEB%JTQPOÓWFMFNIUUQCJUMZUI07R 'FSSÍP + 1PMJDFOUSJTNPFDPFTÍPUFSSJUPSJBMVNOPWPQBQFMQBSBBTSFHJÜFT Revista Europa - Novas Fronteiras, 12 %JTQPOÓWFMFNIUUQCJUMZQ& ©EURE \| vol  \| no  \| enero  \| pp. -  Ferrão, J. (2003). Para uma nova relação entre território, competitividade e acção pública. A EFNPDSBDJBFDPOÙNJDBDPNPGPOUFEFDSJBÎÍPEFWBMPS&O(SVQP1BSMBNFOUBSEP Partido Socialista (Ed.), Novas Políticas para a Competitividade (pp. 123-135). Oeiras: Celta Editora. 'FSSÍP + "FNFSHÐODJBEFFTUSBUÏHJBTUSBOTOBDJPOBJTEFPSEFOBNFOUPEPUFSSJUØSJPOB 6OJÍP&VSPQFJBSFJNBHJOBSPFTQBÎPFVSPQFVQBSBDSJBSOPWBTGPSNBTEFHPWFSOBOÎB territorial? EURE, 30 %JTQPOÓWFMFNIUUQCJUMZT-ZSP Ferrão, J. (2010). As condições sociais de funcionamento da política pública de ordenamento do território. 1SPHSBNBEFOWFTUJHBÎÍP QSPWBTEFIBCJMJUBÎÍPQBSBPFYFSDÓDJPEFGVOÎÜFT EF DPPSEFOBÎÍP DJFOUÓmDB OTUJUVUP EF $JFODJBT 4PDJBMFT 6OJWFSTJEBE EF -JTCPB Portugal. Referências bibliográﬁcas )BNEPVDI "%FQSFU .5BOHVZ $ BCSJM Globalisation and territorial resilience: a suggested approach and preliminary insights from 3 case studies. Trabalho apresentado na The rsa Annual International Conference, Newcastle-upon-Tyne, uk. h )BSWFZ % Social justice and the city-POEPO&EXBSE"SOPMEBOEɨF+PIOT)PQLJOT 6OJWFSTJUZ1SFTT )BSWFZ % Justice, nature and the geography of diﬀerence. Malden, ma#MBDLXFMM Hughes, O. (2003). Public management and administration: An introduction /FX :PSL Palgrave Macmillan. POJDB % ɨF UFSSJUPSJBM EJNFOTJPO PG UIF $PIFTJPO 1PMJDZ Romanian Economic Journal 10(25), 201-220. %JTQPOÓWFMFNIUUQCJUMZB:K(+ +BOJO 3JWPMJO 6 eu territorial governance: learning from institutional progress. European Journal of Spatial Development %JTQPOÓWFMFNIUUQCJUMZC0LS(C +PVFO. Territorial cohesion: from theory to practice. Paris: Notre Europe. Lane, J. (2000). New public management. London: Routledge. Lefebvre, H. (1969). O direito à cidade4ÍP1BVMP&EJUPSB%PDVNFOUPT .BSRVFT 5 %JOÉNJDBTUFSSJUPSJBJTFBTSFMBÎÜFTVSCBOPSVSBJTRevista da Faculdade de Letras – Geograﬁa, 29 %JTQPOÓWFMFNIUUQMFSMFUSBTVQQUVQMPBET mDIFJSPTQEG Martin, P. (2003). 1VCMJDQPMJDJFTBOEFDPOPNJDHFPHSBQIZ&O#'VODL-1J[[BUJ &ET European integration, regional policy and growth (pp. 19-32). Washington dc: The OUFSOBUJPOBM#BOLGPS3FDPOTUSVDUJPOBOE%FWFMPQNFOUɨF8PSME#BOL Meijers, E.; Waterhout, B. & Zonneveld, W. (2007). Closing the gap: Territorial cohesion through polycentric development. European Journal of Spatial Development, 24, 1-24. %JTQPOÓWFMFNIUUQCJUMZ#[ZW, Newman, J. (2002). The new public management, modernization and institutional change: %JTSVQUJPOT EJTKVODUVSFTBOEEJMFNNBT&N,.D-BVHIMJO 40TCPSOF&'FSMJF (Eds.), The New public management: Current trends and future prospects (pp. 77-92). London: Routledge. Rhodes, R. (1997). Understanding Governance: Policy Networks, Governance, Reﬂexivity and Accountability. #VDLJOHIBN0QFO6OJWFSTJUZ1SFTT Rhodes, R.(1996). The New Governance: Governing without Government. Political Studies, EPJKUCY Richardson, T. & Jensen, O. (1999). The Trouble with the esdp. NORTH, 10(6), 11-16. Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... \| ©EURE   Salez, P. (2009). How Europe comes to spatial planning: from the birth of regional policy to the Green Paper on territorial cohesion, the emergence of the Community as a player over more than 20 years. &N(#BVEFMMF : +FBO . 7BOJFS . &ET L’Europe, aménager les territoires, Paris: A. Colin. 4BMMB 7 1PMJUJDBM NZUI NZUIPMPHZ BOE UIF &VSPQFBO 6OJPO Journal of Common Market Studies, 48 EPJKY Santinha, G.; Rodrigues, C. & Almeida, A. (2009). Rumo a um novo ciclo de apoio comuni- tário: que impactos na deﬁnição de políticas territoriais de desenvolvimento local? O caso do Município de Arouca. Revista Inforgeo, 24 %JTQPOÓWFMFNIUUQCJU MZCX2K4 Sen, A. (2010). A ideia de justiça$PJNCSB&EJÎÜFT"MNFEJOB Servillo, L. (2010). 5FSSJUPSJBM $PIFTJPO %JTDPVSTFT )FHFNPOJD 4USBUFHJD $PODFQUT JO European Spatial Planning. Referências bibliográﬁcas Planning Theory & Practice, 11(3), 397-416. doi: Sharpe, L. (1993). The rise of meso-government in Europe. London: Sage. Silvestre, H. (2009). Gestão pública: modelos de prestação no serviço público. Lisboa: Escolar Editora. Soja, E. (2010). Seeking spatial justice..JOOFBQPMJT6OJWFSTJUZPG.JOOFTPUB1SFTT 4UFBE % 8BUFSIPVU # -FBSOJOH GSPN UIF BQQMJDBUJPO PG UIF esdp: In- ﬂuences on european territorial governance. disP, 44(172), 21-34. doi: 4UPLFS ( (PWFSOBODFUIFPSZ'JWFQSPQPTJUJPOTInternational Social Science Journal, 50 EPJ Tatzberger, G. (2006). Spatial visions, concepts and metaphors: their essential role in Euro- QFBOTQBUJBMEFWFMPQNFOUBOEQPMJDZEJTDPVSTF&O-%PSJB 7'FEFMJ$5FEFTDP (Eds.), Rethinking European Spatial Policy as a Hologram: actions, institutions, discourses QQ "MEFSTIPU"TIHBUF :PVOH Justice and the politics of diﬀerence.1SJODFUPO /+1SJODFUPO6OJWFSTJUZ1SFTT Documentos oﬁciales Barca, F. (2009). An agenda for a reformed cohesion policy: a place-based approach to meeting the European Union challenges and expectations (Independent Report prepared at the SFRVFTUPG%BOVUB)àCOFS $PNNJTTJPOFSGPS3FHJPOBM1PMJDZ 3FDVQFSBEPEFMTJUJP XFCEF&VSPQFBO1BSMJBNFOUIUUQCJUMZ3[X #PULB . $PPSE Handbook on territorial cohesion – application of territorial approaches in developments supported by the public sector. Budapest: ğėĘĞ o ħĒĥĚ. %JTQPOÓWFMFNIUUQCJUMZ.HX Commission of the European Communities (1991). Europa 2000. Perspectivas para o desenvolvimento do território da Comunidade-VYFNCVSHP4FSWJÎPEBT1VCMJDBÎÜFT Oﬁciais das Comunidades Europeias. Commission of the European Communities (1994). Europa 2000+. Cooperação para o ordenamento do território Europeu-VYFNCVSHP4FSWJÎPEBT1VCMJDBÎÜFT0mDJBJTEBT Comunidades Europeias. Commission of the European Communities (2001). Governança Europeia - Um Livro Branco. com mOBM #SVYFMBT ©EURE \| vol  \| no  \| enero  \| pp. -  Commission of the European Communities (2001). Unidade da Europa, solidariedade dos povos, diversidade dos territórios. Segundo relatório sobre a coesão econômica e social. -VYFNCVSHP4FSWJÎPEBT1VCMJDBÎÜFT0mDJBJTEBT$PNVOJEBEFT&VSPQFJBT Commission of the European Communities (2004). Uma nova parceria para a coesão: conver- gência competitividade cooperação. Terceiro relatório sobre a coesão econômica e social. -VYFNCVSHP4FSWJÎPEBT1VCMJDBÎÜFT0mDJBJTEBT$PNVOJEBEFT&VSPQFJBT &VSPQFBO0CTFSWBUJPO/FUXPSLGPS5FSSJUPSJBM%FWFMPQNFOUBOE$PIFTJPO Territory matters for competitiveness and cohesion: facets of regional diversity and potentials in Europe. espon ﬁnal Synthesis Report. Luxemburg: espon. &VSPQFBO0CTFSWBUJPO/FUXPSLGPS5FSSJUPSJBM%FWFMPQNFOUBOE$PIFTJPO Scenarios on the Territorial Future of Europe. espon Project 3.2. Luxemburg: espon. &VSPQB TB 1PSUBMEB6OJÍP&VSPQFJBA construção europeia através dos tratados%JTQPOÓWFMFN IUUQFVSPQBFVMFHJTMBUJPO@TVNNBSJFTJOTUJUVUJPOBM@BĊBJSTUSFBUJFTJOEFY@QUIUN European Commision. (2011, mayo). Territorial Agenda 2020 – Towards an Inclusive, Smart and Sustainable Europe of Diverse Regions<TO>(ÚEÚMMƸ )VOHBSZ"VUPS%JTQPOÓWFM FNIUUQCJUMZUL,3, European Commission. (1999). European Spatial Development Perspective. Towards balanced and sustainable development of the territory of the European Union. Luxembourg: Oﬃce for 0ċDJBM 1VCMJDBUJPOT PG UIF &VSPQFBO $PNNVOJUJFT %JTQPOÓWFM FN IUUQCJU MZ5M2 European Commission. (2004). Interim Territorial Cohesion Report (Preliminary results of ESPON and EU Commission studies). Luxembourg: Oﬃce for Oﬃcial Publications of the European Communities. %JTQPOÓWFMFNIUUQCJUMZ5M) European Commission. (2007, mayo). Agenda Territorial da União Europeia. Reunión infor- NBMEF.JOJTUSPTEF%FTBSSPMMP6SCBOPZDPIFTJØOUFSSJUPSJBM 1SFTJEFODJB"MFNBOBEF 6OJØO&VSPQFB -FJQ[JH "MFNBOJB%JTQPOÓWFMFNIUUQXXXEHPUEVQUVF7''"- HFOEB5FSSJUPSJB@QEG &VSPQFBO$PNNJTTJPO B Livro Verde sobre a Coesão Territorial Europeia – Tirar Partido da Diversidade Territorial <TO>. Bruselas: Autor. %JTQPOÓWFMFNIUUQXXXEHPUEV QUVF-JWSP7FSEF5$@QUQEG European Commission. C Regions 2020 - An Assessment of Future Challenges for EU Regions<TO>#SVTFMBT"VUPS%JTQPOÓWFMFNIUUQCJUMZRYOZU European Commission. (2010). Europa 2020 – Estratégia para um crescimento inteligente, sustentável e inclusivo.<TO>.#SVTFMBT"VUPS%JTQPOÓWFMFNIUUQCJUMZ5SI- European Conference of Ministers responisble for Regional Planning (cemat). (2000). Guiding principles for sustainable spatial development of the European continent <TO>. Strasbourg: "VUPS%JTQPOÓWFMFNIUUQXXXCLBHWBU%PD7JFXBYE $PC*E Observatorio en Red de la Ordenación del Territorio Europeo (espon). (2007). 0SHBOJTBUJPOGPS&DPOPNJD$PPQFSBUJPOBOE%FWFMPQNFOU oecd). (2009). Regions matter: Economic recovery, innovation and sustainable growth. <TO>1BSJTAutor. ɨF 8PSME #BOL World development report 2009: Reshaping economic geography. Washington dc"VUPS%JTQPOÓWFMFNIUUQCJUMZ197L2OV Documentos oﬁciales Application and eﬀects of the ESDP in the Member States. ESPON Project 2.3.1<TO>-VYFNCVSH"VUPS %JTQPOÓWFMFNIUUQCJUMZR+ 0SHBOJTBUJPOGPS&DPOPNJD$PPQFSBUJPOBOE%FWFMPQNFOU oecd). (2001). Territorial outlook <TO>1BSJT"VUPS%JTQPOÓWFMFNIUUQCJUMZS.Y$ 0SHBOJTBUJPO GPS &DPOPNJD $PPQFSBUJPO BOE %FWFMPQNFOU oecd). (2006). The new rural paradigm: Policies and governance <TO> 1BSJT "VUPS %JTQPOÓWFM FN IUUQXXX VOJTJJUDJQBTSFGoecd@@3VSBM@1BSBEJHNQEG Santinha \| O princípio de coesão territorial enquanto novo paradigma de desenvolvimento... \| ©EURE  
https://openalex.org/W2892391378	https://nhm.openrepository.com/bitstream/10141/622453/1/Mondillo_et_al_2018%20AcceptedPreprintVersion.pdf	English	null	A global assessment of Zn isotope fractionation in secondary Zn minerals from sulfide and non-sulfide ore deposits and model for fractionation control	Chemical geology	2,018	cc-by	14,439	Title A global assessment of Zn isotope fractionation in secondary Zn minerals from sulfide and non-sulfide ore deposits and model for fractionation control Authors Mondillo, N; Wilkinson, JJ; Boni, M; Weiss, DJ; Mathur, R Date Submitted 2018-11-27 Title A global assessment of Zn isotope fractionation in secondary Zn minerals from sulfide and non-sulfide ore deposits and model for fractionation control Authors Mondillo, N; Wilkinson, JJ; Boni, M; Weiss, DJ; Mathur, R Date Submitted 2018-11-27 A global assessment of Zn isotope fractionation in secondary Zn minerals from sulfide and non-sulfide ore deposits and model for fractionation control Accepted Manuscript A global assessment of Zn isotope fractionation in secondary Zn minerals from sulfide and non-sulfide ore deposits and model for fractionation control Nicola Mondillo, Jamie J. Wilkinson, Maria Boni, Dominik J. Weiss, Ryan Mathur PII: S0009-2541(18)30480-7 DOI: doi:10.1016/j.chemgeo.2018.09.033 Reference: CHEMGE 18926 To appear in: Chemical Geology Received date: 25 June 2018 Revised date: 25 September 2018 Accepted date: 26 September 2018 Please cite this article as: Nicola Mondillo, Jamie J. Wilkinson, Maria Boni, Dominik J. Weiss, Ryan Mathur , A global assessment of Zn isotope fractionation in secondary Zn minerals from sulfide and non-sulfide ore deposits and model for fractionation control. Chemge (2018), doi:10.1016/j.chemgeo.2018.09.033 Accepted Manuscript Accepted Manuscript A global assessment of Zn isotope fractionation in secondary Zn minerals from sulfide and non-sulfide ore deposits and model for fractionation control Nicola Mondillo, Jamie J. Wilkinson, Maria Boni, Dominik J. Weiss, Ryan Mathur Nicola Mondillo, Jamie J. Wilkinson, Maria Boni, Dominik J. Weiss, Ryan Mathur PII: S0009-2541(18)30480-7 DOI: doi:10.1016/j.chemgeo.2018.09.033 Reference: CHEMGE 18926 To appear in: Chemical Geology Received date: 25 June 2018 Revised date: 25 September 2018 Accepted date: 26 September 2018 S0009-2541(18)30480-7 doi:10.1016/j.chemgeo.2018.09.033 CHEMGE 18926 Chemical Geology 25 June 2018 25 September 2018 26 September 2018 Please cite this article as: Nicola Mondillo, Jamie J. Wilkinson, Maria Boni, Dominik J. Weiss, Ryan Mathur , A global assessment of Zn isotope fractionation in secondary Zn minerals from sulfide and non-sulfide ore deposits and model for fractionation control. Chemge (2018), doi:10.1016/j.chemgeo.2018.09.033 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT Abstract ACCEPTED MANUSCRIPT We investigated extent and direction of Zn isotope fractionation in secondary zinc minerals formed during low temperature hydrothermal and/or supergene oxidation of primary sulfide deposits. Zinc isotope data have been obtained from non-sulfide zinc mineral separates (willemite - Zn2SiO4, smithsonite - ZnCO3, hemimorphite - Zn4(Si2O7)(OH)2·H2O, hydrozincite - Zn5(CO3)2(OH)6, and sauconite - Na0.3Zn3(Si,Al)4O10(OH)2·4H2O) collected from several Zn deposits in Ireland, Belgium, Poland, Namibia, Peru, Yemen and Zambia. The data are compared with Zn isotope compositions measured on Zn sulfides collected in the same areas and/or derived from the existing literature to establish the controls of direction and likely extent of any fractionations. We find that willemite has the greatest compositional variability, with measured 66ZnJCM-Lyon values ranging from –0.42 to 1.39‰, spanning the entire range of terrestrial variation in Zn isotopes recorded to date. Overall, significant fractionations in positive and negative directions are recorded relative to the precursor phase (primary sphalerite or an earlier secondary phase), with primary sphalerite falling in a relatively narrow range of isotopic values (approximately –0.1 to +0.4‰). Most of the data observed on willemite, hemimorphite and hydrozincite can be explained with a model of isotopic fractionation in which partial dissolution of primary sphalerite is followed by precipitation of an initial secondary phase that preferentially incorporates heavy Zn isotopes. Smithsonite, instead, preferentially incorporates light Zn isotopes. This reflects the variation in the Zn-x bond strengths of these secondary phases with respect to the original sulfides. We also observed that isotope compositions do not depend only on the difference between the fractionation factors of the involved phases but also on the amount of the secondary mineral precipitated after dissolution of primary sulfide, and that the greatest ACCEPTED MANUSCRIPT We investigated extent and direction of Zn isotope fractionation in secondary zinc minerals formed during low temperature hydrothermal and/or supergene oxidation of primary sulfide deposits. Zinc isotope data have been obtained from non-sulfide zinc mineral separates (willemite - Zn2SiO4, smithsonite - ZnCO3, hemimorphite - Zn4(Si2O7)(OH)2·H2O, hydrozincite - Zn5(CO3)2(OH)6, and sauconite - Na0.3Zn3(Si,Al)4O10(OH)2·4H2O) collected from several Zn deposits in Ireland, Belgium, Poland, Namibia, Peru, Yemen and Zambia. The data are compared with Zn isotope compositions measured on Zn sulfides collected in the same areas and/or derived from the existing literature to establish the controls of direction and likely extent of any fractionations. PT A global assessment of Zn isotope fractionation in secondary Zn minerals from sulfide and non-sulfide ore deposits and model for fractionation control ANUSC Nicola Mondillo1,2, Jamie J. Wilkinson 2,3, Maria Boni1, Dominik Weiss J.3,4, Ryan Mathur5 ED MA 1 Dipartimento di Scienze della Terra, dell'Ambiente e delle Risorse Università degli Studi di Napoli Federico II, Complesso Universitario Monte S. Angelo, Via Cintia, 80126 Napoli, Italy. Email: nicola.mondillo@unina.it; boni@unina.it corresponding author ED MA 1 Dipartimento di Scienze della Terra, dell'Ambiente e delle Risorse Università degli Studi di Napoli Federico II, Complesso Universitario Monte S. Angelo, Via Cintia, 80126 Napoli, Italy. Email: nicola.mondillo@unina.it; boni@unina.it corresponding author EPTE 2 LODE, Department of Earth Sciences, Natural History Museum, Cromwell Road, London SW7 5BD, UK. Email: j.wilkinson@nhm.ac.uk CCEP 3 Department of Earth Science and Engineering, Imperial College London, Exhibition Road, London SW7 2AZ, UK. Email: d.weiss@imperial.ac.uk AC 4 School of Earth, Energy & Environmental Sciences, Stanford University, Stanford, California 94305, United States 5 Juniata College 1700, Moore St Huntingdon PA 16652, USA. Email: mathurr@juniata.edu 1 ACCEPTED MANUSCRIPT ACC Abstract We find that willemite has the greatest compositional variability, with measured 66ZnJCM-Lyon values ranging from –0.42 to 1.39‰, spanning the entire range of terrestrial variation in Zn isotopes recorded to date. Overall, significant fractionations in positive and negative directions are recorded relative to the precursor phase (primary sphalerite or an earlier secondary phase), with primary sphalerite falling in a relatively narrow range of isotopic values (approximately –0.1 to +0.4‰). Most of the data observed on willemite, hemimorphite and hydrozincite can be explained with a model of isotopic fractionation in which partial dissolution of primary sphalerite is followed by precipitation of an initial secondary phase that preferentially incorporates heavy Zn isotopes. Smithsonite, instead, preferentially incorporates light Zn isotopes. This reflects the variation in the Zn-x bond strengths of these secondary phases with respect to the original sulfides. We also observed that isotope compositions do not depend only on the difference between the fractionation factors of the involved phases but also on the amount of the secondary mineral precipitated after dissolution of primary sulfide, and that the greatest 2 2 ACCEPTED MANUSCRIPT TED MANUSCRIPT fractionations occur when only small amounts of secondary mineral are precipitated. Progressive precipitation from migrating fluids that form phases enriched in heavy zinc isotopes would lead to a gradual decrease in the 66Zn values of such phases, and the fluids involved, in time and space. Strong negative isotopic shifts are almost only observed for late crystallizing phases, such as those in vugs. These are interpreted to reflect precipitation from residual, isotopically-light fluids that are the inevitable highly-fractionated product of the above-described process. Where a more complete replacement of primary sulfide has occurred, such as in the high-grade core of nonsulfide zinc orebodies, there is limited net isotopic fractionation because dissolved primary zinc is nearly quantitatively reprecipitated locally. In addition, in only one case (Yanque, Peru) we observed that the fringes of non-sulfide zinc deposit were characterized by isotopically fractionated compositions, with highly negative values implying extensive precipitation (earlier, or elsewhere) of isotopically heavy secondary phases. The higher-grade ore zones, where complete breakdown of primary sulfides and quantitative reprecipitation of zinc have occurred, show instead less fractionated compositions. CCEP Keywords: Zn isotopes; isotope fractionation; supergene environment; Rayleigh-type distillation; non-sulfide zinc deposits; secondary zinc minerals. 3 ACCEPTED MANUSCRIPT 1. Introduction Zn2+, solvated by water molecules (Ducher et al., 2018), as Zn complexed to organic ligands or humic acids (Jouvin et al., 2009 Mar ovi et al., 2017), and complexed with inorganic ligands such as chloride. Experimental and theoretical investigations on Zn complexation by organic ligands (Ban et al., 2002; Black et al., 2011; Jouvin et al., 2009, Fuji and Albarède, 2012; Mar ovi et al., 2017) has shown that such complexes favor in general heavy Zn isotopes controlled by the affinity constant. Other experiments on isotope fractionation associated with sorption on Fe- and Mn-oxyhydroxides (e.g. Balistrieri et al., 2008; Bryan et al., 2015; Juillot et al., 2008; Pokrovski et al., 2005) have shown that heavy Zn isotopes are preferentially adsorbed, generating a positive ∆66Zn between the adsorbing phase and the solution ACCEPTED MANUSCRIPT as meteorites (Rosman, 1972; Luck et al., 2005) and in lunar rocks (Moynier et al., 2006). In terrestrial geological samples (i.e. sediments, igneous rocks and ores), δ66Zn values are clustered around +0.5‰, ranging from ca. –0.5‰ to +2.5‰ (Moynier et al., 2017). This limited spread reflects the fact that Zn does not undergo redox transitions which often produce large isotope fractionations (Moynier et al., 2017). The main mechanisms causing Zn isotope fractionation at the Earth’s surface are equilibrium isotope distribution between dissolved aqueous species (e.g. organic complexes) and equilibrium and kinetic effects caused by interactions between solids and aqueous solutions (e.g. sorption, precipitation). In natural waters and hydrothermal solutions, Zn occurs as the free ion, i.e. Zn2+, solvated by water molecules (Ducher et al., 2018), as Zn complexed to organic ligands or humic acids (Jouvin et al., 2009 Mar ovi et al., 2017), and complexed with inorganic ligands such as chloride. Experimental and theoretical investigations on Zn complexation by organic ligands (Ban et al., 2002; Black et al., 2011; Jouvin et al., 2009, Fuji and Albarède, 2012; Mar ovi et al., 2017) has shown that such complexes favor in general heavy Zn isotopes controlled by the affinity constant. Other experiments on isotope fractionation associated with sorption on Fe- and Mn-oxyhydroxides (e.g. 1. Introduction EPTED MANUSCRIPT Zinc, an element with atomic number 30, which occurs in the natural environment with the +2 oxidation state, has five stable isotopes, 64Zn (49.2%), 66Zn (27.8%), 67Zn (4.0%), 68Zn (18.4%), and 70Zn (0.6%) and an average relative atomic mass of 65.37777(22) (Cloquet et al., 2008; Moynier et al., 2017; Rosman, 1972). The measurement of Zn isotopic ratios by Thermal- Ionization Mass-Spectrometry (TIMS) is difficult, because the element has a high first ionization potential which leads to a low Saha constant; for this reason, most of the existing data have been produced since the late 90s, after the advent of Multiple-Collector Inductively-Coupled-Plasma Mass Spectrometry (MC-ICP-MS) (Moynier et al., 2017). In the last twenty years, several reference standards have been adopted for Zn isotopic investigations. However, many studies used a Johnson Matthey Corporation (JMC) Zn standard solution, batch 3–0749L, commonly denoted as JMC-Lyon (Maréchal et al., 1999), which is no longer available (Moynier et al., 2017). Möller et al. (2012) calibrated IRMM-3702 as the new certified Zn standard, which has a δ66Zn value of +0.29 ‰ relative to the JMC-Lyon standard. Despite this, recent papers (e.g. Wanty et al., 2013) have still reported Zn isotope ratios relative to JMC-Lyon: ACCE δ66Zn (‰)= A The main processes affecting the Zn isotope composition of natural materials are evaporation– condensation processes, in which the vapor phase is depleted in the heavier isotopes relative to the solid phase, adsorption, complexation and diffusion (Cloquet et al., 2008; Moynier et al., 2017). Evaporation-condensation processes have produced the large variations in δ66Zn JCM-Lyon (–11‰ to +7‰: Moynier et al., 2011; Kato et al., 2015) observed in extraterrestrial samples such 4 ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT as meteorites (Rosman, 1972; Luck et al., 2005) and in lunar rocks (Moynier et al., 2006). In terrestrial geological samples (i.e. sediments, igneous rocks and ores), δ66Zn values are clustered around +0.5‰, ranging from ca. –0.5‰ to +2.5‰ (Moynier et al., 2017). This limited spread reflects the fact that Zn does not undergo redox transitions which often produce large isotope fractionations (Moynier et al., 2017). The main mechanisms causing Zn isotope fractionation at the Earth’s surface are equilibrium isotope distribution between dissolved aqueous species (e.g. organic complexes) and equilibrium and kinetic effects caused by interactions between solids and aqueous solutions (e.g. sorption, precipitation). In natural waters and hydrothermal solutions, Zn occurs as the free ion, i.e. 1. Introduction Balistrieri et al., 2008; Bryan et al., 2015; Juillot et al., 2008; Pokrovski et al., 2005) have shown that heavy Zn isotopes are preferentially adsorbed, generating a positive ∆66Zn between the adsorbing phase and the solution A variety of processes are considered as possible explanations for the isotopic variability of Zn AC A variety of processes are considered as possible explanations for the isotopic variability of Zn in hydrothermal ore systems, including source-rock control, temperature, and kinetic or equilibrium fractionation during precipitation (Albarède, 2004; Archer and Vance, 2002; Kelley et al., 2009; Maréchal et al., 1999, 2000; Mason et al., 2005; Pašava et al., 2014 Pichat et al., 2003; Wilkinson et al., 2005). For example, the variability in Zn isotopic data observed for 5 ACCEPTED MANUSCRIPT USCRIPT sphalerite in the Irish ore field is likely generated by kinetic fractionation and Rayleigh distillation during progressive sulfide precipitation from the hydrothermal fluids (Wilkinson et al., 2005). Low temperature water-rock interaction imparts measurable Zn isotope fractionation (Balistrieri et al., 2008; Bermin et al., 2006; Borrok et al., 2008; Fernandez and Borrok, 2009; Lv et al., 2016; Opfergelt et al., 2017). To date, the majority of the Zn isotope studies on surficial environments have focused on how the isotopes can trace metal contamination (Araújo et al., 2017; Bigalke et al., 2010, 2016; Cloquet et al., 2008; John et al., 2008; Kelley et al., 2009; Mattielli et al., 2009; Sivry et al., 2008; Sonke et al., 2008; Wanty et al., 2013; Weiss et al., 2007). ACCEPTED MANU Several experimental and natural studies have provided important insights into the controls of Zn isotope fractionation. Fernandez and Borro (2009) observed that “continuous batch leaching of sphalerite at pH=2 and pH=4 was characterized by a modest (0.2‰) initial increase in the δ66Zn of the leachate relative to the sphalerite, which quickly dissipated and the δ66Zn of the fluid began to reflect the δ66Zn of the sphalerite”. The initial fractionation was explained by the transient formation of a surface coating composed of Zn-sulfates that, changing the bonding environment for Zn, incorporated preferentially the heavier Zn isotopes. Wanty and co-workers (2013) studied the isotopic effects related to the precipitation of hydrozincite in a Zn-polluted river in Sardinia and observed that δ66Zn values of hydrozincite samples were consistently about 0.35‰ heavier than the waters with which they were in contact. 1. Introduction Thus, it was concluded that hydrozincite preferentially incorporated the heavier isotope, after a biologically-mediated mineralization process, leaving the residual waters depleted in 66Zn and leading to lower δ66Zn values in downstream water samples. Isotopic fractionation of Zn, associated with a solution isotopically lighter that the parent material, was also observed during experimental leaching of a 6 6 ACCEPTED MANUSCRIPT biotite granite with oxalic and hydrochloric acid solutions which, at the early stages, produced a solution ca. 1‰ lighter than the original sample (Weiss et al., 2014). biotite granite with oxalic and hydrochloric acid solutions which, at the early stages, produced a solution ca. 1‰ lighter than the original sample (Weiss et al., 2014). ANUSCRIPT Ducher et al. (2016) reported theoretical studies of Zn isotope fractionation in Zn-bearing minerals. They investigated the isotope fractionation properties of Zn in several minerals at equilibrium using density functional theory (DFT). Assuming that equilibrium, mass-dependent isotopic fractionation arises mainly from the vibrational motions of atoms, the equilibrium isotopic properties of the element were determined by the calculation of the reduced partition function ratio, also called the β-factor, by computing vibrational frequencies using a harmonic approach. 1. Introduction The β-factor is defined as the ratio at equilibrium of the isotope ratio of the substance of interest to the isotope ratio of a perfect gas of atoms (Bigeleisen and Mayer, 1947), which in this case is represented by monatomic Zn: D MA EPTED The β-factors are useful to our aims because there is a relationship between the equilibrium isotopic fractionation factor α (a, b, Y) between two phases a and b for an element Y (in this case Zn), and the reduced partition function ratios of a and b: EPTED The β-factors are useful to our aims because there is a relationship between the equilibrium isotopic fractionation factor α (a, b, Y) between two phases a and b for an element Y (in this case Zn), and the reduced partition function ratios of a and b: CCE α (a,b,Y) = β (a,Y) / β (b,Y) α (a,b,Y) = β (a,Y) / β (b,Y) α (a,b,Y) = β (a,Y) / β (b,Y) ACC This allows simple approximation of the β-factors with the δ values, which are normally measured on natural samples: ACC This allows simple approximation of the β-factors with the δ values, which are normally measured on natural samples: 103 ln α (a,b,Y) ≈ ∆ = δa – δb 103 ln α (a,b,Y) ≈ ∆ = δa – δb As the calculated β-factors are very close to 1, they are presented as logarithmic quantities (lnβ) expressed in per mil units (‰). Ducher and co-workers (2016) demonstrated that at ambient temperature gahnite, hemimorphite and hydrozincite are amongst the minerals having the highest logarithmic values of the reduced partition function ratio (ln β 66Zn/64Zn of 4.18‰, 3.78‰ and 7 ACCEPTED MANUSCRIPT NUSCRIPT 3.19‰ at 22 °C, respectively), whereas smithsonite and sphalerite have the lowest β-factors (ln β 66Zn/64Zn of 2.30‰ and 2.40‰ at 22 °C, respectively). This means that, at equilibrium, an alteration process that releases Zn from sphalerite and reprecipitates it into secondary minerals could produce either positive or negative fractionations: hemimorphite and hydrozincite would be isotopically heavier than parental sphalerite (at 22 °C ∆hem-sph(‰) = +1.38, ∆hydr-sph (‰) = +0.80), whereas smithsonite would be isotopically lighter (at 22 °C ∆sm-sph (‰) = –0.10). 1. Introduction Crystal- chemical parameters control these isotopic properties (Ducher et al., 2016): indeed, excellent linear correlations exist between ln β and Zn interatomic force constants, and β-factors increase when the Zn-first neighbor bond lengths decrease and charges on atoms involved in the bonding increase and vice versa. ACCEPTED MAN The present study sought to determine the controls of Zn isotope fractionation during sphalerite dissolution and formation of oxidized secondary Zn minerals in sulfide and non-sulfide ore deposits, and to establish a preliminary model for Zn isotope fractionation for this process. To this end, Zn isotope compositions were measured on 28 secondary zinc minerals (willemite - Zn2SiO4, smithsonite - ZnCO3, hemimorphite - Zn4(Si2O7)(OH)2·H2O, hydrozincite - Zn5(CO3)2(OH)6, sauconite - Na0.3Zn3(Si,Al)4O10(OH)2·4H2O) collected from several localities in Ireland, Belgium, Poland, Namibia, Peru, Yemen and Zambia (Fig. 1), and compared, where possible, with the Zn-isotopic ratios measured on primary sulfides (9 samples of sphalerite and one galena) from the same areas. Of particular interest was the behavior of the Zn silicate, willemite, which is observed in some of the deposits (Belgium, Namibia, Zambia) that may have either a supergene or hypogene genesis. 8 ACCEPTED MANUSCRIPT 2.1. Ireland TED MANUSCRIPT The zinc sulfide deposits in Ireland (carbonate-hosted SEDEX, Mississippi Valley-type (MVT) and vein-type deposits) occur mainly in the Lower Carboniferous of the Central Midlands (e.g. Wilkinson et al., 2003; Wilkinson and Hitzman, 2015). Locally, nonsulfide zinc-lead mineralization was developed above these deposits (Balassone et al., 2008), and represents an example of supergene oxidation of primary sulfide ores, with local redeposition and preservation under glacial till. The dominant composition of these oxidized ores is relatively simple: smithsonite, hemimorphite, cerussite and Cu-carbonates. Prior to the 1960s, lead and zinc were also mined from low temperature hydrothermal veins throughout much of Ireland. These veins are hosted by Lower Paleozoic rocks which, in the case of the Glendalough and Old Luganure deposits, comprise Caledonian granite. Supergene minerals in these occurrences probably relate to exhumation and exposure of the deposits between 36-10 Ma. 2.2. Belgium ACCE In the Liège Zn-Pb-Ba mining district (NE Belgium), primary zinc-lead sulfides (MVT) and secondary nonsulfides are the main typology of deposits (Dejonghe et al., 1993). Sulfide mineralogy includes sphalerite, galena, and pyrite/marcasite with collomorphic and brecciated textures. The mineralized veins cut the Palaeozoic rocks folded during the Hercynian orogeny, and the main mineralizing event dates back to the end of the Jurassic (~150 Ma). The Belgian nonsulfide ores mainly represent the supergene oxidation products of primary sulfides. Nonsulfide ores consist of an association of smithsonite, hydrozincite, hemimorphite, willemite 9 9 ACCEPTED MANUSCRIPT USCRIPT and Zn-clays, with a variable content of Pb minerals. The extensive occurrence of the Zn-silicate willemite (Fig. 2A), which appears to be the first deposited nonsulfide mineral directly replacing sphalerite, is a peculiarity of the Belgian ores and its genesis has not been fully resolved. In fact, willemite may have formed during a hydrothermal event (Brugger et al., 2003) with moderate temperatures, high silica activities and deeply reaching O-rich fluids, possibly active in the deposit area between the Jurassic and Cretaceous, as well as the first product of sulfide weathering, post Cretaceous (Coppola et al., 2008). By contrast, smithsonite and hemimorphite have all the characteristics of supergene products, and were commonly developed at the expense of primary sulfides as well as of willemite (Coppola et al., 2008). 2.3. Poland ACCEPTED MA Sulfide ores from Upper Silesia (Poland) are hosted by the so-called Ore Bearing Dolomite, a hydrothermal alteration product corresponding to the epigenetic replacement of Muschelkalk (Triassic) carbonates. Two important kinds of mineralization have been described: (a) replacement bodies and (b) in-filling of open spaces in fractures and breccias. The mineral association in the sulfide deposits includes sphalerite, galena, marcasite, pyrite and sulfoarsenides (Kucha, 2005). The age of mineralization is controversial and both early Mesozoic and Tertiary have been proposed (Heijlen et al., 2003; Leach et al., 1996). The nonsulfide ores are commonly associated with horst features delimited by Tertiary faults (Coppola et al., 2009) and were formed by weathering of primary Zn-Pb sulfides. Their prevailing mineralogical association comprises smithsonite, cerussite, hemimorphite and Fe- (hydr)oxides (Fig. 2B). 10 ACCEPTED MANUSCRIPT 2.4. Namibia TED MANUSCRIPT The Otavi Mountain Land (OML) in Namibia broadly coincides with the Neoproterozoic carbonate platform of the Damara Orogen. In the OML, several ore deposits and prospects (both sulfides and nonsulfides) are hosted in the carbonate successions of the Otavi Group. The primary sulfide ores have been subdivided into the ‘Berg Au as-type’ (Zn-Pb MVT) and ‘Tsumeb-type’ (Pb-Cu-Zn pipes) deposits (Pirajno and Joubert, 1993). A maximum age for the ‘Tsumeb-type’ sulfide ores in the OML is 530 Ma, whereas the ‘Berg Au as-type’ ores seem to be older, possibly related to fluids expelled during the Damaran orogeny. In the Namibian occurrences, willemite is in several generations (Fig. 2C). It commonly replaces sphalerite, but may also occur as a newformed phase or as a direct replacement of the carbonate host rocks (Schneider et al., 2008). The willemite concentrations in the Namibian deposits (Berg Aukas, Abenab West, Baltika etc.) have been interpreted to be mainly of hydrothermal origin (Hitzman et al., 2003; Schneider et al., 2008). Smithsonite, hemimorphite and descloizite commonly follow willemite and have been interpreted as late, weathering-derived supergene phases. 2.5. Zambia ACCE The Zambian Zn-Pb deposits occur in metasedimentary rocks of the Late Proterozoic Katangan Supergroup. The most important orebodies are located in the Kabwe area, and contain both sulfides and nonsulfides hosted in dolomite. The Zn-Pb sulfide mineralization has been interpreted as having been formed before the last Lufilian deformational phase (ca 650 - 500 Ma) (Kamona and Friedrich, 2007), and would be broadly in the same age range as the OML sulfides. Similar to the OML, several willemite generations of possibly hydrothermal origin (Fig. 2D) are 11 ACCEPTED MANUSCRIPT present at Kabwe (Mondillo et al., 2018), which are clearly followed by supergene phases, like smithsonite, cerussite, hemimorphite, descloizite and tarbuttite (Fig. 2E). present at Kabwe (Mondillo et al., 2018), which are clearly followed by supergene phases, like smithsonite, cerussite, hemimorphite, descloizite and tarbuttite (Fig. 2E). 2.6. Yemen PTED MANUSCRIPT The Jabali mixed sulfide-nonsulfide Zn mineralization is located about 110 km east of Sana'a (Yemen) (Mondillo et al., 2014a). The host rocks of both ore types are intensively dolomitized Jurassic carbonates. The primary sulfide deposit shows the general features of MVT ores, and its formation age was attributed to Early Cretaceous (Ostendorf et al., 2015). Smithsonite (Fig. 2F) is the most abundant economic mineral in the nonsulfide mineralization, where it replaces both the host dolomite and sphalerite (Fig. 2G). It is associated with minor hydrozincite, hemimorphite, acanthite and greenoc ite. The Jabali smithsonite shows δ18O values substantially lower and with a larger variability in comparison with other smithsonites considered supergene in the literature (Gilg et al., 2008), suggesting that this Zn carbonate could have been precipitated from fluids with a temperature between 30 and 65 °C, probably derived from the mixing of local groundwaters with low-temperature hydrothermal waters (Mondillo et al., 2014a). 2.7. Peru 2.7. Peru ACC The Yanque nonsulfide Zn(Pb) prospect is located in the Cuzco area (southern Peru, Mondillo et al., 2014b). The primary sulfides are considered to represent polymetallic mineralization genetically related to the emplacement of the Andahuaylas–Yauri batholith (Middle Eocene to Early Oligocene, ca. 48–32 Ma). However, these sulfides have been totally weathered, and the economic concentrations in the prospect consist prevailingly of supergene Zn(Pb) nonsulfides, mostly hosted in siliciclastic conglomerates. The mineral association contains sauconite, 12 ACCEPTED MANUSCRIPT SCRIPT hemimorphite, smithsonite and cerussite. Remnants of the primary sulfides are represented by moderate amounts of galena and pyrite, but residual sphalerite is very rare. Sauconite, the most abundant ore mineral in the Yanque deposit, formed after a process of supergene wall-rock replacement (Fig. 2H; sensu Hitzman et al., 2003), where the Zn released from the complete dissolution of sphalerite precipitated as a newformed mineral after the reaction of Zn-bearing solutions with the silicatic country rock. The process likely started in the center of the present orebody, and progressively propagated southward from surficial zones into more deeply buried strata (Mondillo et al., 2014b). SCRIPT hemimorphite, smithsonite and cerussite. Remnants of the primary sulfides are represented by moderate amounts of galena and pyrite, but residual sphalerite is very rare. Sauconite, the most abundant ore mineral in the Yanque deposit, formed after a process of supergene wall-rock replacement (Fig. 2H; sensu Hitzman et al., 2003), where the Zn released from the complete dissolution of sphalerite precipitated as a newformed mineral after the reaction of Zn-bearing solutions with the silicatic country rock. The process likely started in the center of the present orebody, and progressively propagated southward from surficial zones into more deeply buried strata (Mondillo et al., 2014b). 3. Sample locations and descriptions ACCEPTED MA The samples from the Glendalough Mine, Glendalough (BM1964R-6305) and from the Old Luganure mine, Glendasan (BM1964R-6307), originate from the Mineralogy Collection of the Natural History Museum (London), and consist of two sphalerite-smithsonite pairs. In the Glendalough sample, the sphalerite occurs as nice, idiomorphic crystals, with smithsonite in globular concretions that have overgrown and partly replaced it. In the Old Laganure sample, the smithsonite is commonly green-yellow, possibly due to the presence of Cd, and appears to have overgrown sphalerite. The samples from the Galmoy mine were collected underground from the CW orebody (Doyle et al., 1992; Balassone et al., 2008) and consist of vug-filling, late-stage, greenish sphalerite, overgrown by smithsonite. Samples from the mines of Fossey (RN2300) and La Calamine (R2B35/3925) in Belgium (Coppola et al., 2008) originate from the Royal Belgian Institute of Natural Sciences in Brussels 13 ACCEPTED MANUSCRIPT RIPT and consist of two willemite-smithsonite pairs (Table 1). In both samples, rhombohedral crystals of smithsonite have overgrown partially-altered, anhedral crystals of willemite that have replaced the original sphalerite. The sample (RN5011) from the Moresnet old mine (a subsidiary of La Calamine) consists of crystalline willemite, which has replaced collomorphic sphalerite; galena from this sample was also analyzed to provide an indication of the Zn isotopic composition of primary sulfides in the deposit. RIPT and consist of two willemite-smithsonite pairs (Table 1). In both samples, rhombohedral crystals of smithsonite have overgrown partially-altered, anhedral crystals of willemite that have replaced the original sphalerite. The sample (RN5011) from the Moresnet old mine (a subsidiary of La Calamine) consists of crystalline willemite, which has replaced collomorphic sphalerite; galena from this sample was also analyzed to provide an indication of the Zn isotopic composition of primary sulfides in the deposit. D MANUSCR The samples from the Polish mines (Coppola et al., 2009) were collected on site, from old mine workings. They are represented by a sphalerite-smithsonite pair (sample CV05-71, Chrzanòw), and by a sphalerite-hemimorphite pair (sample CV05-80, Pomorzany) (Table 1). In the Chrzanòw sample, smithsonite occurs as globular concretions around massive sphalerite (complete replacement of sphalerite by smithsonite was not observed), whereas in the Pomorzany sample, hemimorphite has overgrown collomorphic sphalerite. ACCEPTE The Namibian samples (Schneider et al., 2008) are from two different localities. 3. Sample locations and descriptions Three near-surface samples (<20 m) are YA13-SAM12, collected from the northern limb of the ore body, YA17-SAM16 from the central part of the ore body and YA20-SAM19 from its eastern side. The samples YA31-SAM24, YA34-SAM25, YA40-SAM29 and YA45-SAM32 originate from the southernmost limb of the ore body and at greater depths (120-160 m); they are separated by around 100 m and are sequentially located along a traverse to the south (i.e. YA45-SAM32 most distal from the deposit core). SCRIPT feldspar-mica siliciclastic rock. The samples are texturally similar, but originate from different parts of the deposit, which locally crops out in the north of the prospect area and plunges southward with a dip of ~30°. Three near-surface samples (<20 m) are YA13-SAM12, collected from the northern limb of the ore body, YA17-SAM16 from the central part of the ore body and YA20-SAM19 from its eastern side. The samples YA31-SAM24, YA34-SAM25, YA40-SAM29 and YA45-SAM32 originate from the southernmost limb of the ore body and at greater depths (120-160 m); they are separated by around 100 m and are sequentially located along a traverse to the south (i.e. YA45-SAM32 most distal from the deposit core). 3. Sample locations and descriptions Two generations of willemite were collected from the Berg Aukas mine (generation 1 BA2003-4: botryoidal yellow willemite that has replaced the carbonate host rock; generation 2 BA2003-13.2: white needles of late willemite in vein and cavity fillings; Terracciano, 2008). A sphalerite-willemite pair was also sampled at the small Baltika prospect near Otavi. The sample NARO427A-1 consists of black-reddish sphalerite, and sample NARO427A-2 contains an early generation of willemite that has partly replaced sphalerite. 14 ACCEPTED MANUSCRIPT RIPT The Zambian samples originate from the Kabwe mining area (Terracciano, 2008). The first sample consists of a sphalerite-willemite pair from the Kabwe 5/6 pit (ZA0502-D: black-grayish sphalerite; ZA0502-M: early, botryoidal willemite that has replaced the carbonate host rock). The other specimen contains a smithsonite-hemimorphite pair sampled from the outcropping gossan of Kabwe pit 2 (ZA0511-2: brownish smithsonite crystals in cavity; ZA0511-1: white hemimorphite needles that have overgrown the smithsonite). ACCEPTED MANUSCR The samples from the Jabali deposit originated from drill core J125, which encounters the mineralization at ~50 m depth below a thick cover of carbonate rocks, and from two outcrops located on the side of a valley which cuts the sedimentary rocks hosting mineralization. Sphalerite samples J125-5 and J125-7 represent two specimens of vein-type sphalerite, texturally associated with saddle dolomite (Mondillo et al., 2014a). Sample J125-9 contains smithsonite that has directly replaced the vein-type sphalerite, whereas the concretionary smithsonite sample J125-32 came from the deeper zones of the Jabali mineralization and is not texturally associated with any sphalerite. Smithsonite sample JS-MON-6 and hydrozincite sample JS-MON-22B were collected from the best mineralized outcrops occurring at Jabali: JS-MON-6 represents an outcropping part of the smithsonite orebody, sampled alongside the J125-9 specimen, whereas the hydrozincite is a late phase formed from the supergene alteration of the former smithsonite generations. Seven sauconite samples were selected from drill cores in the Yanque deposit (Peru). The samples consist of massive sauconite that formed from Zn precipitation from Zn-bearing groundwaters, produced by leaching of an original sulfide body, and reacting with a quartz- 15 ACCEPTED MANUSCRIPT SCRIPT feldspar-mica siliciclastic rock. The samples are texturally similar, but originate from different parts of the deposit, which locally crops out in the north of the prospect area and plunges southward with a dip of ~30°. 4. Analytical methods ACCEPTED MA Samples from Ireland, Belgium, Poland, Namibia and Zambia were analyzed at the MAGIC Laboratories, Imperial College London, UK. Samples from Yemen and Peru were analyzed at the University of Arizona, USA. The Zn minerals were handpicked from each specimen, powdered and sieved to <500 micron. At the University of Arizona, 0.02-0.04 g of sample was weighed into 15 ml Teflon beakers, and dissolved/leached at 120°C for greater than 12 hours 4 ml in ultrapure aqua regia. 0.5 ml of the resulting solution was then taken and dried. At Imperial College London, sample digestion was carried out following the procedure described in Chapman et al. (2006). In both the laboratories, the Zn in the resulting salts was then separated using two subsequent passages through an anion exchange resin as outlined in detail by Chapman et al. (2006). 16 ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT At Imperial College London, isotope measurements were made using an IsoProbe MC-ICP-MS (Thermo Instruments, U.K.), following the analytical procedures described in detail by Gioia and co workers (2008) and by Araújo et al. (2016). A Cetac Aridus desolvating nebulizer (Cetac Technologies, Omaha) was used for sample introduction. The instrument was used in the “hard extraction” mode and and signal contributions from the cone were insignificant. Instrumental background and acid matrix blank corrections were performed using on-peak acid blank measurements taken before every sample and standard (Gioia et al., 2008). Measurements are quoted relative to the Zn standard JMC 3-0749 L (JMC-Lyon). The procedural blank was around 5 ng which was insignificant compared to the amount of Zn processed through the samples. Accuracy and precision of the Zn isotope ratio measurements following our ion exchange procedure were assessed using a well-characterized synthetic element solution from our laboratory (RomilZn), certified (for concentrations) reference materials BCR-027 (blende ore) and G-2 (granite) and the in-house standard HRM-24 (granite). Possible isotope fractionation during the ion exchange separation procedure was tested comparing the isotopic compositions of the synthetic element solutions before and after the passage through the ion exchange columns. We did not find significant isotope fractionation during the passage of the ion exchange column and δ66ZnLyon values of RomilZn, G-2and BCR-027 agreed within error with published data using a variety of different analytical procedures (Möller et al., 2012). The in-house standard HRM-24 showed values for δ66ZnJMC-Lyon of +0.267 ± 0.05‰ (2SD and n=3). 4. Analytical methods These ratios fall well within the range observed for granites and andesite (Moynier et al., 2017). Repeated measurement of RomilZn over the study period (36 months) showed an external precision of ACCEPTED MANUSCRIPT At Imperial College London, isotope measurements were made using an IsoProbe MC-ICP-MS (Thermo Instruments, U.K.), following the analytical procedures described in detail by Gioia and co workers (2008) and by Araújo et al. (2016). A Cetac Aridus desolvating nebulizer (Cetac Technologies, Omaha) was used for sample introduction. The instrument was used in the “hard extraction” mode and and signal contributions from the cone were insignificant. Instrumental background and acid matrix blank corrections were performed using on-peak acid blank measurements taken before every sample and standard (Gioia et al., 2008). Measurements are quoted relative to the Zn standard JMC 3-0749 L (JMC-Lyon). The procedural blank was around 5 ng which was insignificant compared to the amount of Zn processed through the samples. Accuracy and precision of the Zn isotope ratio measurements following our ion exchange procedure were assessed using a well-characterized synthetic element solution from our laboratory (RomilZn), certified (for concentrations) reference materials BCR-027 (blende ore) and G-2 (granite) and the in-house standard HRM-24 (granite). Possible isotope fractionation during the ion exchange separation procedure was tested comparing the isotopic compositions of the synthetic element solutions before and after the passage through the ion exchange columns. We did not find significant isotope fractionation during the passage of the ion exchange column and δ66ZnLyon values of RomilZn, G-2and BCR-027 agreed within error with published data using a variety of different analytical procedures (Möller et al., 2012). The in-house standard HRM-24 showed values for δ66ZnJMC-Lyon of +0.267 ± 0.05‰ (2SD and n=3). These ratios fall ACCEP We did not find significant isotope fractionation during the passage of the ion exchange column and δ66ZnLyon values of RomilZn, G-2and BCR-027 agreed within error with published data using a variety of different analytical procedures (Möller et al., 2012). The in-house standard HRM-24 showed values for δ66ZnJMC-Lyon of +0.267 ± 0.05‰ (2SD and n=3). These ratios fall well within the range observed for granites and andesite (Moynier et al., 2017). Repeated measurement of RomilZn over the study period (36 months) showed an external precision of ±0.05‰ pamu. 17 ACCEPTED MANUSCRIPT SCRIPT At the University of Arizona, isotopic measurements were also made on an IsoProbe (MC-ICP- MS). 4. Analytical methods Solutions were diluted to approximately 260 ppb for Zn prior to measurement, which generated a 66Zn signal between 3-4 V. Sample intensities matched the standard intensities within 20%. The reported data are the average of 2 separate measurements of a block of 25 ratios. Isotope variations were determined using IRMM 3702 and then recalculated to standard delta notation relative to JMC-Lyon, summing at the IRMM 3702 values the published offset between the two standards of +0.29‰ (Cloquet et al., 2008; Ponzevera et al., 2006; Moeller et al., 2012). 5. Results ACCEPTED MA The δ66Zn values for the analyzed European sulfide and non-sulfide samples are given in Table 1 and summarized in Figure 3. The δ66ZnJCM-Lyon values for all the primary sphalerites samples fall in the relatively narrow range between −0.06 to +0.41‰, typical of hydrothermal sphalerite from low-moderate temperature ore deposits (e.g. Wilkinson et al., 2005; Moynier et al., 2017). The minerals that generally occur as the earliest overgrowth/replacement of primary sulfides (excluding the Pomorzanj hemimorphite) are willemite and smithsonite. Differently from Araújo et al. (2017), we found that willemite has the greatest compositional variability, with measured 66ZnJCM-Lyon values ranging from –0.42 to 1.39‰, spanning the entire range of terrestrial variation in Zn isotopes recorded to date. The majority of the willemite samples (with the exception of the second willemite generation from Berg Aukas, Namibia) show a distinct positive isotopic shift from the precursor sulfide (Fig. 4), with the most extreme fractionation observed in the Belgian samples (shift larger than 0.5‰). Li ewise, apart from the samples from 18 ACCEPTED MANUSCRIPT CRIPT Glendalough and Old Luganure (Ireland), all the smithsonite samples show a positive shift in δ66ZnJCM-Lyon relative to the primary sulfide of up to ~0.5‰ (Jabali, Yemen). The single case where hemimorphite was observed as the first secondary phase (Pomorzanj, Poland) also shows a positive isotopic shift of +0.23‰ (Fig. 4). Although primary sphalerites have not been measured from Yanque (Peru), based on the assumption that it is likely to have typical hydrothermal compositions in the range of −0.1 to 0.4‰, it can be inferred that the sauconite formation also involved zinc isotope fractionation with an increase in δ66ZnJCM-Lyon. TED MANUSCR For the secondary minerals that replaced the earlier non-sulfide zinc phases (Table 1, Fig. 2) the pattern is the opposite. Smithsonite formed after willemite in the case of the Belgian deposits shows a strong negative shift (up to −0.87‰) relative to the willemite that it has overgrown and partly replaced (Fig. 3). Hydrozincite after smithsonite in the Jabali deposit (Yemen) shows a shift of −0.18‰ in δ66ZnJCM-Lyon. Late hemimorphite after smithsonite from the Kabwe deposit (Namibia) shows a shift of −0.66‰ and the late acicular willemite from Berg Au as (Namibia) a shift of −0.50‰ relative to the earlier willemites. ACCEPTE cussion 6. Discussion 6. Discussion AC The analyzed sphalerites have δ66ZnJCM-Lyon values in the same range as those of hydrothermal sphalerites occurring in other deposits (Albarède, 2004; Archer and Vance, 2002; Kelley et al., 2009; Maréchal et al., 2000; Mason et al., 2005; Sonke et al., 2008; Wilkinson et al., 2005; Weiss et al., 2007; Zhou et al., 2014). The primary sphalerite displays a relatively narrow 19 ACCEPTED MANUSCRIPT compositional range (−0.06 to +0.41‰) in comparison with the replacing secondary minerals (Table 1). ACCEPTED MANUSCRIPT From the Zn isotope data on secondary Zn-minerals analyzed for this study (Table 1), it is evident that the mobilization of zinc via dissolution and reprecipitation in various oxidized minerals results in measurable zinc isotope fractionation. Whether the initial replacement is hydrothermal, as may be the case for some of the willemites (e.g. Kabwe specimens; Mondillo et al., 2018), or supergene for the other phases (e.g. Pomorzanj hemimorphite; Coppola et al., 2009), there is a common pattern of enrichment of heavy zinc isotopes in the newly-formed phase. A similar positive shift of δ66ZnJCM-Lyon compositions in secondary minerals deriving from sulfide alteration was observed in the previous studies of non-sulfide zinc minerals, which were focused on sphalerite and hydrozincite from several small deposits of the Cantabrian region in Northern Spain (Pašava et al., 2012), and on hydrozincite precipitation in the Naracauli river cross-cutting an old mining area in SW Sardinia (Italy) (Wanty et al., 2013). In the latter case, the groundwaters flowing into the river had previously leached waste dumps and hidden Zn- sulfide bodies (Wanty et al., 2013). The isotopic compositions of primary sphalerites from the above studies (δ66ZnJCM-Lyon −0.80 to 0.15‰) are similar to the compositions measured in other hydrothermal deposits and to those reported from our samples. In addition, the hydrozincite samples show a comparable shift to higher values, i.e. +0.01 to +0.30‰ (Pasava et al., 2012; Wanty et al., 2013). In regard to Sardinia, Wanty and co-workers (2013) argued that “complete sphalerite dissolution yields a solution with the same δ66ZnJCM-Lyon as the precursor sulfide”, and that it is in the following precipitation step that the fractionation occurs, with the preferential incorporation of heavy zinc isotopes in the secondary Zn minerals, producing residual fluids with progressively lighter isotopic signature. 6. Discussion However, considering that, in the upper reaches of the ACCEPTED MANUSCRIPT From the Zn isotope data on secondary Zn-minerals analyzed for this study (Table 1), it is evident that the mobilization of zinc via dissolution and reprecipitation in various oxidized minerals results in measurable zinc isotope fractionation. Whether the initial replacement is hydrothermal, as may be the case for some of the willemites (e.g. Kabwe specimens; Mondillo et al., 2018), or supergene for the other phases (e.g. Pomorzanj hemimorphite; Coppola et al., 2009), there is a common pattern of enrichment of heavy zinc isotopes in the newly-formed phase. A similar positive shift of δ66ZnJCM-Lyon compositions in secondary minerals deriving from sulfide alteration was observed in the previous studies of non-sulfide zinc minerals, which were focused on sphalerite and hydrozincite from several small deposits of the Cantabrian region in Northern Spain (Pašava et al., 2012), and on hydrozincite precipitation in the Naracauli river cross-cutting an old mining area in SW Sardinia (Italy) (Wanty et al., 2013). In the latter case, the groundwaters flowing into the river had previously leached waste dumps and hidden Zn- sulfide bodies (Wanty et al., 2013). The isotopic compositions of primary sphalerites from the above studies (δ66ZnJCM-Lyon −0.80 to 0.15‰) are similar to the compositions measured in other hydrothermal deposits and to those reported from our samples. In addition, the hydrozincite samples show a comparable shift to higher values, i.e. +0.01 to +0.30‰ (Pasava et al., 2012; Wanty et al., 2013). In regard to Sardinia, Wanty and co-workers (2013) argued that “complete sphalerite dissolution yields a solution with the same δ66ZnJCM-Lyon as the precursor sulfide”, and that it is in the following precipitation step that the fractionation occurs, with the preferential incorporation of heavy zinc isotopes in the secondary Zn minerals, producing residual fluids with progressively lighter isotopic signature. However, considering that, in the upper reaches of the 20 ACCEPTED MANUSCRIPT T stream, hydrozincite precipitates in a biologically-mediated environment, facilitated by microalgae and cyanobacteria, and that synthetic hydrozincites produced without bacteria had δ66Zn identical to the dissolved Zn, the above authors concluded that the Zn isotopic fractionation between hydrozincite and waters was a biologically-mediated mineralization process (Wanty et al., 2013). 6. Discussion T stream, hydrozincite precipitates in a biologically-mediated environment, facilitated by microalgae and cyanobacteria, and that synthetic hydrozincites produced without bacteria had δ66Zn identical to the dissolved Zn, the above authors concluded that the Zn isotopic fractionation between hydrozincite and waters was a biologically-mediated mineralization process (Wanty et al., 2013). ACCEPTED MANUSCRIPT Fernandez and Borrok (2009) conducted an experimental investigation on Zn isotopic fractionation during oxidative dissolution of pure sphalerite, and observed a modest (ca. 0.2‰) initial increase in the δ66Zn of the leachate relative to the sphalerite separate, which they explained invoking a change in the bonding environment for Zn in the altered surface layer of the sphalerite. In detail, as the surface coating was composed of Zn-sulfates, the oxygen-bonding would favor the preferential incorporation of the heavier Zn isotopes in the Zn-sulfate relatively to the solution that has the same composition of the original Zn-sulfide (Fernandez and Borrok, 2009). The observations of Fernandez and Borrok (2009) (and also of Wanty et al. [2013], if we exclude the biological mediation) are in agreement with the theoretical models of Ducher et al. (2016), which yielded positive ∆66Zn between hydrozincite and sphalerite (+0.80‰), and between the hydrous Zn-sulfate (gunningite) and sphalerite (+0.26‰), at ambient temperature (22 °C), and attribute these differences to the crystal-chemical parameters and the Zn interatomic force constants of the considered mineral phases. In detail, β-factors are higher when the Zn-first neighbor bond lengths is smaller and charges on atoms involved in the bonding are higher and vice versa (Ducher et al., 2016). Broadly, the observed ∆66Zn shifts for all the oxidized phases analyzed for this study from the precursor sulfides are in agreement with the theoretical calculations of Ducher et al. (2016), which predicted, at equilibrium, positive ∆66Zn between several oxidized minerals (e.g. 21 ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT hydrozincite and hemimorphite) and sphalerite, and negative ∆66Zn between smithsonite and the Zn-sulfide, on the basis of the “changing bonding environment of the Zn atoms from sulfide to non-sulfide minerals and the different crystal-chemical parameters and Zn interatomic force constants between the different phases” (Ducher et al., 2016). There are apparent exceptions to this general pattern in our data set. 6. Discussion For example, an isotopically lighter smithsonite than the original sulfide only observed in the Irish Glendalough and Luganure samples, whereas in all the other smithsonite samples from Zambia, Poland and Yemen this Zn-carbonate shows positive shifts in δ66Zn relative to the different precursor sphalerites (∆66Znsm-sph is +0.42‰ at Kabwe- Zambia, +0.22‰ at Chrzanòw-Poland, and +0.61‰ at Jabali-Yemen). These positive shifts can be explained (i) by a lack of direct relationship between sphalerite breakdown and smithsonite formation, for example by the involvement of an isotopically heavy intermediate phase; or (ii) by unusual zinc complexation that favored the enrichment of heavy isotopes (for example, on organic complexes; Ban et al., 2002 Blac et al., 2011 Jouvin et al., 2009, Fu i and Albar de, 2012 Mar ovi et al., 2017). The lack of a direct relationship between sphalerite breakdown and smithsonite formation can be suggested for the Belgian smithsonites, which do not directly derive from sulfide alteration but have an intermediate willemite precursor (Coppola et al., 2008). In this case, the Belgian smithsonites are characterized by δ66Zn lower than their willemite precursor (∆66Znsm-will is between −0.5‰ and −0.9‰) and therefore fit the theoretical equilibrium Zn isotope fractionation model (Ducher et al., 2016). A similar argument can also be made for the Zambian samples where willemite is intermediate between sphalerite and smithsonite (Fig. 4). The best example suggesting an effect/influence of soil-derived organic ligands in the Zn isotopic fractionation is in the Polish samples. The smithsonites from Poland are mostly characterized by isotopically light carbon, reflecting derivation of organic carbon 22 ACCEPTED MANUSCRIPT from C3-vegetated soils overlying the deposits (Gilg et al., 2008; Coppola et al., 2009). It is therefore well plausible that organic complexes were present in the waters percolating through the vegetated soils, and that these were responsible for alteration of sulfides, Zn transportation and secondary Zn mineral precipitation (McPhail et al., 2003). ACCEPTED MANUSCRIPT Assessing the other minerals, several important conclusions can be drawn. In the Namibian Berg Aukas and Baltika specimens, the first willemite generations (NAR0427A-2 and BA2003-4) formed after sphalerite alteration have compositions (+0.07‰, +0.08‰) slightly heavier than the precursor sulfide (−0.06‰). Although β-factors for willemite do not exist (Ducher et al., 2016), because Zn is tetrahedrally coordinated with oxygen in willemite like in hemimorphite, it is li ely that the β-factors for these phases would be comparable. 6. Discussion Thus, willemite formation from sphalerite might be expected to produce a positive ∆66Zn relative to the original sulfide, as is observed in the Namibian willemite 1 samples. The second willemite generation (BA2003-13.2), however, is strongly depleted relatively to willemite 1 (∆66Znw2-w1 is −0.35‰). Considering that the observed fractionation concerns specimens of the same mineral species, likely formed at similar temperatures from similar fluids (Schenider et al., 2008; Terracciano, 2008), it is not possible to explain their diverse composition by simply using equilibrium isotope fractionation mechanisms. In this case, it is possible to explain the depleted compositions of the para- genetically late willemite via a Rayleigh-type distillation model. It is envisaged that initial dissolution of sphalerite and the formation of a Zn-bearing solution with initial composition equal to the original sulfide is followed by Zn re-precipitation into a series of secondary oxide phases. As long as Zn reprecipitation is incomplete, the formation of isotopically-heavy, neo- 23 ACCEPTED MANUSCRIPT formed Zn minerals will produce a depleted fluid that can progressively evolve back towards the primary sulfide composition, or even pass it (Fig. 5A). The equation describing this process is: formed Zn minerals will produce a depleted fluid that can progressively evolve back towards the primary sulfide composition, or even pass it (Fig. 5A). The equation describing this process is: ACCEPTED MANUSCRIPT where, Rdm0 is the isotope ratio of the initial solution corresponding to the isotope ratio of the original dissolved mineral (assuming its complete dissolution), and Rpm is the isotope ratio of the precipitated mineral; f is the fraction of residual metal in solution and the fractionation factor α is given by Rpm / Rdm. In this model, a solution with starting δ66Zn bul composition of 0.1‰, resulting from the complete dissolution of a sphalerite with the same composition (δ66Znsph = 0.1‰ is the median value of all the sphalerite shown in Fig. 4), will produce precipitates with different δ66Zn compositions using a α-fractionation factors between the mineral and the fluid, and the mass percent of metal precipitated (Fig. 5). For example, an α-fractionation factor of 1.0015 (which is ca. 6. Discussion the α-fractionation factor between hemimorphite and sphalerite at T = 22 °C; Ducher et al., 2016) reproduces both the highest (which are formed from the first small amounts of precipitation), and the lowest (which form when about 75% of the metal has been deposited) δ66Zn values occurring in our dataset (that are δ66Zn = 1.39‰ La Calamine-Belgium willemite, δ66Zn = −0.42‰ Berg Au as-Namibia willemite 2). However, at any point, sudden supersaturation of the solution could dump out the total Zn with the same composition of the fluid at that point (i.e. no further fractionation occurs). This model is consistent with the nature of the isotopically-light samples from Namibia, Zambia and Yemen which all occur as paragenetically late phases (willemite, hemimorphite and hydrozincite respectively; see Figure 4). In detail, from this model, the Kabwe hemimorphite could have been precipitated once about 70% of the metal had been deposited from a fluid with an initial δ66Zn value of 0.1‰ (i.e. the composition of the dissolved precursor willemite). Assuming an initial δ66Zn value of about ACCEPTED MANUSCRIPT where, Rdm0 is the isotope ratio of the initial solution corresponding to the isotope ratio of the original dissolved mineral (assuming its complete dissolution), and Rpm is the isotope ratio of the precipitated mineral; f is the fraction of residual metal in solution and the fractionation factor α is given by Rpm / Rdm. In this model, a solution with starting δ66Zn bul composition of 0.1‰, resulting from the complete dissolution of a sphalerite with the same composition (δ66Znsph = 0.1‰ is the median value of all the sphalerite shown in Fig. 4), will produce precipitates with different δ66Zn compositions using a α-fractionation factors between the mineral and the fluid, and the mass percent of metal precipitated (Fig. 5). For example, an α-fractionation factor of 1.0015 (which is ca. the α-fractionation factor between hemimorphite and sphalerite at T = 22 °C; Ducher et al., 2016) reproduces both the highest (which are formed from the first small amounts of precipitation), and the lowest (which form when about 75% of the metal has been deposited) δ66Zn values occurring in our dataset (that are δ66Zn = 1.39‰ La Calamine-Belgium willemite, δ66Zn = −0.42‰ Berg Au as-Namibia willemite 2). However, at any point, sudden supersaturation of the solution could dump out the total Zn with the same composition of the fluid at that point (i.e. no further fractionation occurs). 6. Discussion This model is consistent with the nature of the isotopically-light samples from Namibia, Zambia and Yemen which all occur as paragenetically late phases (willemite, hemimorphite and hydrozincite respectively; see Figure 4). In detail, from this model, the Kabwe hemimorphite could have been precipitated once about 70% of the metal had been deposited from a fluid with an initial δ66Zn value of 0.1‰ (i.e. the composition of the dissolved precursor willemite). Assuming an initial δ66Zn value of about 24 ACCEPTED MANUSCRIPT 0.60‰ for the fluid sourcing the Jabali hydrozincite (corresponding to the mean value of precursor smithsonites), the model predicts a precipitation of the hydrozincite (0.25‰) after almost 75% of the metal in solution had been already deposited (Fig. 5B). 0.60‰ for the fluid sourcing the Jabali hydrozincite (corresponding to the mean value of precursor smithsonites), the model predicts a precipitation of the hydrozincite (0.25‰) after almost 75% of the metal in solution had been already deposited (Fig. 5B). ACCEPTED MANUSCRIPT The δ66ZnJCM-Lyon values (comprised between 0.14‰ to 0.61‰) of the sauconite samples from the Yanque deposit (Peru) show that significant isotopic variability also occurs in the same secondary mineral within one system. There is evidence of zonation in the orebody with the highest isotopic values in the center and a trend towards lower values towards the southern limits of the deposit at depth, as well as a decrease upwards and outwards (see Table 1 for sample positions and compositions). The genetic model of the Yanque deposit proposed by Mondillo et al. (2014b) suggests that sauconite formed after a process of supergene wall-rock replacement (sensu Hitzman et al., 2003), where Zn released from the complete dissolution of sphalerite precipitated as sauconite after the reaction of Zn-bearing solutions with the silicatic country rock. The process likely started in the center of the present orebody, and progressively propagated southward from surficial zones into more deeply buried strata (Mondillo et al., 2014b). In line with this model , the observed δ66Zn zonation in the analyzed Yanque sauconites can be explained by preferential incorporation of heavy Zn isotopes in earlier formed sauconite in the core of the system, with fluids depleted in these isotopes generating the lower values in peripheral areas (e.g. YA 45 SAM 32). 7. Conclusions MANUSCRIPT The results of this study clearly indicate that supergene and hydrothermal processes on Zn mineralizations produce more significant fractionation of zinc isotopes, than previously observed (e.g. Wanty et al., 2013). The variation in Zn isotope composition through the different minerals/sampled sites can be explained by a combination of equilibrium isotope fractionation (Ducher et al., 2016) and open system Rayleigh distillation during incomplete dissolution- reprecipitation reactions. We also observed that zones where more complete replacement of primary sulfide occurs will result in minor fractionation, as the majority of mobilized zinc is redeposited without significant removal of the light zinc isotopes. 6. Discussion Thus, this spatial variation is consistent with the Rayleigh-type distillation model proposed above that can also account for a temporal (paragenetic) evolution from initial isotopically heavy phases, toward later, isotopically light ones. ED MANUSCRIPT The δ66ZnJCM-Lyon values (comprised between 0.14‰ to 0.61‰) of the sauconite samples from the Yanque deposit (Peru) show that significant isotopic variability also occurs in the same secondary mineral within one system. There is evidence of zonation in the orebody with the highest isotopic values in the center and a trend towards lower values towards the southern limits of the deposit at depth, as well as a decrease upwards and outwards (see Table 1 for sample positions and compositions). The genetic model of the Yanque deposit proposed by Mondillo et al. (2014b) suggests that sauconite formed after a process of supergene wall-rock replacement (sensu Hitzman et al., 2003), where Zn released from the complete dissolution of sphalerite precipitated as sauconite after the reaction of Zn-bearing solutions with the silicatic country rock. ACCEPTED The process likely started in the center of the present orebody, and progressively propagated southward from surficial zones into more deeply buried strata (Mondillo et al., 2014b). In line with this model , the observed δ66Zn zonation in the analyzed Yanque sauconites can be explained by preferential incorporation of heavy Zn isotopes in earlier formed sauconite in the core of the system, with fluids depleted in these isotopes generating the lower values in peripheral areas (e.g. YA 45 SAM 32). Thus, this spatial variation is consistent with the Rayleigh-type distillation model proposed above that can also account for a temporal (paragenetic) evolution from initial isotopically heavy phases, toward later, isotopically light ones. 25 ACCEPTED MANUSCRIPT D MA Acknowledgements CCEPTE The authors wish to thank the Curators of the Natural History Museum, London, and of the Royal Belgian Institute of Natural Sciences, Brussels, for allowing sampling of the Irish and Belgian nonsulfide minerals from the respective Museum collections. ACC Funding: This work was supported by the departmental research funds granted by DiSTAR to M. Boni and by the EU Marie Curie training center ACCORD, based at the Natural History Museum, London. 26 ACCEPTED MANUSCRIPT References T Albarède, F., 2004. The stable isotope geochemistry of copper and zinc. In: Johnson, C.M., Beard, B.L., Albarède, F. (Eds.), Geochemistry of nontraditional stable isotopes. Reviews in Mineralogy, 55, 409–427. USCRIPT Araújo, D., Boaventura, G., Viers, J., Mulholland, D., Weiss, D., Araújo, D., Lima, B., Ruiz, I., Machado, W., Babinski, M., Dantas, E., 2016. Ion exchange chromatography and mass bias correction for accurate and precise Zn isotope ratio measurements in environmental reference materials by MC-ICP-MS. J. Brazil. Chem. Soc., 28, 225-235. MANU Araújo, D., Boaventura, G., Machado, W., Viers, J., Weiss, D., Patchineelam, S.R., Ruiz, I., Rodrigues, A.P.C., Babinski, M., Dantas, E., 2017. Tracing of anthropogenic zinc sources in coastal environments using stable isotope composition. Chem. Geol., 449, 226-235. CCEPTED MANU Araújo, D., Boaventura, G., Machado, W., Viers, J., Weiss, D., Patchineelam, S.R., Ruiz, I., Rodrigues, A.P.C., Babinski, M., Dantas, E., 2017. Tracing of anthropogenic zinc sources in coastal environments using stable isotope composition. Chem. Geol., 449, 226-235. Archer, C., Vance, D., 2002. Large fractionations in Fe, Cu, and Zn isotopes associated with Archean mircrobially-mediated sulfides. Geochim. Cosmochim. Acta, 66, p. A26. Balassone, G., Rossi, M., Boni, M., Stanley, G., McDermott, P., 2008. Mineralogical and geochemical characterization of nonsulfide Zn–Pb mineralization at Silvermines and Galmoy (Irish Midlands). Ore Geol. Rev., 33, 168–186. ED M Archer, C., Vance, D., 2002. Large fractionations in Fe, Cu, and Zn isotopes associated with TED Archean mircrobially-mediated sulfides. Geochim. Cosmochim. Acta, 66, p. A26. CEPTE Balassone, G., Rossi, M., Boni, M., Stanley, G., McDermott, P., 2008. Mineralogical and geochemical characterization of nonsulfide Zn–Pb mineralization at Silvermines and Galmoy (Irish Midlands). Ore Geol. Rev., 33, 168–186. ACC Balistrieri, L.S., Borrok, D.M., Wanty, R.B., Ridley, W.I., 2008. Fractionation of Cu and Zn isotopes during adsorption onto amorphous Fe(III) oxyhydroxide: Experimental mixing of acid rock drainage and ambient river water. Geochim. Cosmochim. Acta, 72, 311-328. Ban, Y., Aida, M., Nomura, M., Fujii, Y., 2002. Zinc isotope separation by ligand-exchange 27 ACCEPTED MANUSCRIPT Bermin, J., Vance, D., Archer, C., Statham, P.J., 2006. The determination of the isotopic composition of Cu and Zn in sea water. Chem. Geol., 226, 280-297. Bigalke, M., Schwab, L., Gygax, S., Mestrot, A., 2016. Soil Contamination with Trace Metals: Quantification, Speciation, and Source Identification. Chimia, 70, 899-899. RIPT Bigalke, M., Weyer, S., Kobza, J., Wilcke, W., 2010. References Stable Cu and Zn isotope ratios as tracers of sources and transport of Cu and Zn in contaminated soil. Geochim. Cosmochim. Acta, PT Bigalke, M., Weyer, S., Kobza, J., Wilcke, W., 2010. Stable Cu and Zn isotope ratios as tracers RIPT of sources and transport of Cu and Zn in contaminated soil. Geochim. Cosmochim. Acta, 74, 6801-6813. USCR Bigeleisen, J., Mayer, M.G., 1947. Calculation of equilibrium constants for isotopic exchange reactions. J. Chem. Phys., 15, 261–267. MANU Black, J. R., Kavner, A., Schauble, E.A., 2011. Calculation of equilibrium stable isotope partition function ratios for aqueous zinc complexes and metallic zinc. Geochim. Cosmochim. Acta, 75 (3), 769−783. TED M Boni, M., Mondillo, N., 2015. The “Calamines” and the “Others”: The great family of supergene nonsulfide zinc ores. Ore Geol. Rev., 67, 208–233. CCEPTE Borrok, D.M., Nimick, D.A., Wanty, R.B., Ridley, W.I., 2008. Isotopic variations of dissolved copper and zinc in stream waters affected by historical mining. Geochim Cosmochim Acta, 72, 329-344. ACC Bryan, A.L., Dong, S., Wilkes, E.B., Wasylenki, L.E., 2015. Zinc isotope fractionation during adsorption onto Mn oxyhydroxide at low and high ionic strength. Geochim. Cosmochim. Acta, 157, 182-197. Brugger, J., McPhail, D.C., Wallace, M., Waters, J., 2003. Formation of Willemite in Hydrothermal Environments. Econom. Geol., 98, 819-835. 28 ACCEPTED MANUSCRIPT Chapman, J.B., Mason, T.F.D., Weiss, D.J., Coles, B.J., Wilkinson, J.J., 2006. Chemical separation and isotopic variations of Cu and Zn from five geological reference materials. Geostandards Geoanalyt. Res., 30, 5-16. IPT Cloquet, C., Carignan, J., Lehmann, M., Vanhaecke, F., 2008. Variation in the isotopic composition of zinc in the natural environment and the use of zinc isotopes in biogeosciences: a review. Analyt. Bioanalyt. Chem., 390, 451-463. USCRIP Coppola, V., Boni, M., Gilg, H.A., Balassone, G., De onghe, L., 2008. The “calamine” nonsulfide Zn–Pb deposits of Belgium: petrographical, mineralogical and geochemical characterization. Ore Geol. Rev., 33, 187-210. ANU Coppola V., Boni, M., Gilg, H.A, Strzelska-Smakowska, B., 2009. Nonsulfide zinc deposits in the Silesia-Cracow district, Southern Poland. Min. Deposita, 44, 559-580. TED MA Doyle, E., Bowden, A.A., Jones, G.V., and Stanley, G.A., 1992. The geology of the Galmoy deposits, in Bowden, A., Earls, G., O’Connor, P., and Pyne, J., eds., The Irish Minerals Industry 1980–1990: Dublin, Irish Association for Economic Geology, p. 211–225. TED Industry 1980–1990: Dublin, Irish Association for Economic Geology, p. 211–225. CCEPTE Dejonghe, L., Ladeuze, F., Jans, D., 1993. References Atlas des gisements plombo-zincifères du Synclinorim de Verviers (Est de la Belgique). Mémoires pour servir à l'explication des Cartes Géologiques et Miniére de Belgique, 33, 1-483. ACC Ducher, M., Blanchard, M., Balan, E., 2016. Equilibrium zinc isotope fractionation in Zn bearing minerals from first-principles calculations. Chem. Geol., 443, 87–96. AC minerals from first-principles calculations. Chem. Geol., 443, 87–96. Ducher, M., Blanchard, M., Balan, E., 2018. Equilibrium isotopic fractionation between aqueous Zn and minerals from first-principles calculations. Chem. Geol., 483, 342-350. Zn and minerals from first-principles calculations. Chem. Geol., 483, 342-350. Fernandez, A., Borrok, D.M., 2009. Fractionation of Cu, Fe, and Zn isotopes during the oxidative weathering of sulfide-rich rocks. Chem. Geol., 264, 1-12. 29 ACCEPTED MANUSCRIPT Fujii, T., Albarède, F., 2012. Ab initio calculation of the Zn isotope effect in phosphates, citrates, and malates and applications to plants and soil. PLoS One, 7, (2), doi:e3072610.1371/journal.pone.0030726. IPT Gagnevin, D., Boyce, A.J., Barrie, C.D., Menuge, J.F., Blakeman, R.J., 2012. Zn, Fe and S isotope fractionation in a large hydrothermal system. Geochim. Cosmochim. Acta, 88, 183-198. USCRIP Gilg, H.A., Boni, M., Hochleitner, R., Struck, U., 2008. Stable isotope geochemistry of carbonate minerals in supergene oxidation zones of Zn–Pb deposits. Ore Geol. Rev. 33, 117–133. ANUS Gioia, S., Weiss, D., Coles, B., Arnold, T., Babinski, M., 2008. Accurate and precise zinc isotope ratio measurements in urban aerosols. Anal. Chem., 80, 9776–9780. TED MA Heijlen, W., Muchez, Ph., Banks, D.A., Schneider, J., Kucka, H., Keppens. E., 2003. Carbonate- hosted Zn-Pb deposits in Upper Silesia, Poland: origin and evolution of mineralizing fluids and constraints on genetic models. Econ. Geol., 98, 911-932. PTE Hitzman, M.H., Reynolds, N.A., Sangster, D.F., Allen, C.R., Carman, C.E., 2003. Classification, ACCEPT Hitzman, M.H., Reynolds, N.A., Sangster, D.F., Allen, C.R., Carman, C.E., 2003. Classification, genesis, and exploration guides for nonsulfide zinc deposits. Econ. Geol., 98, 685–714. John, S.G., Rouxel, O.J., Craddock, P.R., Engwall, A.M., Boyle, E.A., 2008. Zinc stable isotopes in seafloor hydrothermal vent fluids and chimneys. Earth Planet. Sci. Lett., 269, 17-28. Juillot, F.,Maréchal, C., Ponthieu, M., Cacaly, S., Morin, G., Benedetti, M., Hazemann, J.L., Proux, O., Guyot, F., 2008. Zn isotopic fractionation caused by sorption on goethite and 2-Lines ferrihydrite. Geochim. Cosmochim. Acta, 72, 4886-4900. Jouvin, D., Louvat, P., Juillot, F., Marechal, C.N., Benedetti, M.F., 2009. Zinc isotopic fractionation: why organic matters. Environ. Sci. Technol., 43, 5747-5754. EP genesis, and exploration guides for nonsulfide zinc deposits. Econ. Geol., 98, 685–714. References CCE John, S.G., Rouxel, O.J., Craddock, P.R., Engwall, A.M., Boyle, E.A., 2008. Zinc stable isotopes ACC loor hydrothermal vent fluids and chimneys. Earth Planet. Sci. Lett., 269, 17-28. Jouvin, D., Louvat, P., Juillot, F., Marechal, C.N., Benedetti, M.F., 2009. Zinc isotopic fractionation: why organic matters. Environ. Sci. Technol., 43, 5747-5754. 30 ACCEPTED MANUSCRIPT Kamona, A.F., Friedrich, G.H., 2007. Geology, mineralogy and stable isotope geochemistry of the Kabwe carbonate-hosted Pb–Zn deposit, Central Zambia. Ore Geol. Rev., 30, 217- 243. T Kato, C., Moynier, F., Valdes, M.C., Dhaliwal, J.K., Day, J.M., 2015. Extensive volatile loss during formation and differentiation of the Moon. Nat. Communication, , 6, 7617. SCRIPT Kelley, K.D., Wilkinson, J.J., Chapman, J.B., Crowther, H.L., Weiss, D.J., 2009. Zinc isotopes in sphalerite from base metal deposits in the Red Dog district, Northern Alaska. Econ. Geol., 104, 767-773. MANUSC Kucha, H., 2005. Oxysulfides, smithsonite-siderite and Fe-free smithsonite as indicators of conditions of formation of primary and supergene non-sulfide Zn-Pb deposits, Upper Silesia, Poland. ESF Workshop on Nonsulfide Zn-Pb Deposits, 21th - 23th April, 2005, Iglesias, Italy 2005, Abstract, 24-25. PTED M Leach, D.L., Viets, J.G. Powell, J.W., 1996. Textures of ores from the Silesian-Krakow zinc-lead deposits, Poland: clues to the ore-forming environment. Prace Instytutu Geologicznego, 154, 37-50. ACCEP Lv, Y., Liu, S.-A., Zhu, J.-M., Li, S., 2016. Copper and zinc isotope fractionation during deposition and weathering of hihly metalliferous black shales in China. Chem. Geol., 445, 24-35. AC Luck, J.m., Othman, D.B., Albarède, F., 2005. Zn and Cu usotopic variations in chondrites and iron meteorites: Early solar nebula reservoirs and parent-body processes. Geochim. Cosmochim. Acta, 69, 5351-5363. Maréchal, C.N., Télouk, P., Albarède, F., 1999. Precise analysis of copper and zinc isotopic compositions by plasma-source mass spectrometry. Chem. Geol., 156, 251–273. 31 ACCEPTED MANUSCRIPT IPT Maréchal, C.N., Nicolas, E., Douchet, C., Albarède, F., 2000. Abundance of zinc isotopes as a marine biogeochemical tracer. Geochem. Geophis. Geosyst., 1, 17 pp. Mar ovi , T., Manzoor, S., Humphreys-Williams, E., Kirk, G.J.D., Vilar, R., Weiss, D.J., 2017. Experimental determination of zinc isotope fractionation in complexes with the Phytosiderophore 2′-Deoxymugeneic Acid (DMA) and its structural analogues, and implications for plant uptake mechanisms. Environm. Sc. Technol., 51, 98-107. IPT Maréchal, C.N., Nicolas, E., Douchet, C., Albarède, F., 2000. Abundance of zinc isotopes as a marine biogeochemical tracer. Geochem. Geophis. Geosyst., 1, 17 pp. Mar ovi , T., Manzoor, S., Humphreys-Williams, E., Kirk, G.J.D., Vilar, R., Weiss, D.J., 2017. References Experimental determination of zinc isotope fractionation in complexes with the Phytosiderophore 2′-Deoxymugeneic Acid (DMA) and its structural analogues, and implications for plant uptake mechanisms. Environm. Sc. Technol., 51, 98-107. Maréchal, C.N., Nicolas, E., Douchet, C., Albarède, F., 2000. Abundance of zinc isotopes as a marine biogeochemical tracer. Geochem. Geophis. Geosyst., 1, 17 pp. NUSCRIP Mason, T.F.D., Weiss, D.J., Chapman, J.B., Wilkinson, J.J., Tessalina, S.G., Spiro, B., Horstwood, M.S.A., Spratt, J., Coles, B.J., 2005. Zn and Cu isotopic variability in the Alexandrinka volcanic-hosted massive sulfide (VHMS) ore deposit, Urals, Russia. Chem. Geol., 221, 170-187. D MAN Mattielli, N., Petit, J.C.J., Deboudt, K., Flament, P., Perdrix, E., Taillez, A., Rimetz-Planchon, J., Weis, D., 2009. Zn isotope study of atmospheric emissions and dry depositions within a 5 km radius of a Pb–Zn refinery. Atmospheric Environm., 43, 1265-1272. EPTED McPhail, D.C., Summerhayes, E., Welch, S., Brugger, J., 2003. The Geochemistry of Zinc in the Regolith. In: Roach, I.C. (Ed.), Advances in Regolith. CRC for Landscape Environments and Mineral Exploration, 287–291. ACCE Möller, K., Schoenberg, R., Pedersen, R-B., Weiss, D., Dong S., 2012. Calibration of the new certified reference materials ERM-AE633 and ERM-AE647 for copper and IRMM-3702 for zinc isotope amount ratio determinations. Geostand Geoanalytical Res., 36, 177–99. ACCE Möller, K., Schoenberg, R., Pedersen, R-B., Weiss, D., Dong S., 2012. Calibration of the new certified reference materials ERM-AE633 and ERM-AE647 for copper and IRMM-3702 for zinc isotope amount ratio determinations. Geostand Geoanalytical Res., 36, 177–99. Mondillo, N., Boni, M., Balassone, G., Joachimski, M., Mormone, A., 2014a. The Jabali Nonsulfide Zn-Pb-Ag Deposit, Western Yemen. Ore Geol. Rev., 61, 248–267. Mondillo, N., Boni, M., Balassone, G., Villa, I.M., 2014b. The Yanque Prospect (Peru): from polymetallic Zn-Pb mineralization to a Nonsulfide deposit. Econ. Geol., 109, 1735-1762. Mondillo, N., Boni, M., Balassone, G., Joachimski, M., Mormone, A., 2014a. The Jabali Nonsulfide Zn-Pb-Ag Deposit, Western Yemen. Ore Geol. Rev., 61, 248–267. Mondillo, N., Boni, M., Balassone, G., Villa, I.M., 2014b. The Yanque Prospect (Peru): from 32 ACCEPTED MANUSCRIPT Mondillo, N., Herrington, R., Boyce, A.J., Winkinson, C., Santoro, L., Rumsey, M., 2018. Critical elements in nonsulphide Zn deposits: A reanalysis of the Kabwe Zn-Pb ores (central Zambia). Miner. Mag., 82(S1), S89–S114. Moynier, F., Albarède, F., Herzong, G., 2006. Isotopic composition of zinc, copper, and iron in T lunar samples. Geochim. Cosmochim. Acta, 70, 6103-6117. RIPT Moynier, F., Paniello, R.C., Gounelle, M., Albarède, F., Beck, P., Podosek, F., Zanda, B., 2011. References CRIPT Moynier, F., Paniello, R.C., Gounelle, M., Albarède, F., Beck, P., Podosek, F., Zanda, B., 2011. Nature of volatile depletion and genetic relationships in enstatite chondrites and aubrites inferred from Zn isotopes Geochim Cosmochim Acta 75 297 307 CRIP Nature of volatile depletion and genetic relationships in enstatite chondrites and aubrites SCR inferred from Zn isotopes. Geochim. Cosmochim. Acta, 75, 297-307. NUSC Moynier, F., Vance, D., Fujii, T., Savage, P., 2017. The Isotope Geochemistry of Zinc and Copper. Rev. Miner. Geochem., 82, 543-600. D MAN Opfergelt, S., Cornélis, J.T., Houben, D., Givron, C., Burton, K.W., Mattielli, N., 2017. The influence of weathering and soil organic matter on Zn isotopes in soils. Chem. Geol., 466, 140-148. CCEPTED Ostendorf, J., Henjes-Kunst, F., Mondillo, N., Boni, M., Schneider, J., Gutzmer, J., 2015. Formation of Mississippi Valley–type deposits linked to hydrocarbon generation in extensional tectonic settings: Evidence from the Jabali Zn-Pb-(Ag) deposit (Yemen). Geology, 43, 1055-1058. ACC Pašava, J., Tornos, F., Chrastn , V., 2014. Zinc and sulfur isotope variation in sphalerite from carbonate-hosted zinc deposits, Cantabria, Spain. Min. Deposita, 49, 797-807. Pichat, S., Douchet, C., Albarède, F., 2003. Zinc isotope variations in deep-sea carbonates from the eastern equatorial Pacific over the last 175 a. Earth Planet. Sci. Lett., 210, 167−178. Pichat, S., Douchet, C., Albarède, F., 2003. Zinc isotope variations in deep-sea carbonates from the eastern equatorial Pacific over the last 175 a. Earth Planet. Sci. Lett., 210, 167−178. Pirajno, F., Joubert, B.D., 1993. An overview of carbonate-hosted mineral deposits in the Otavi Pirajno, F., Joubert, B.D., 1993. An overview of carbonate-hosted mineral deposits in the Otavi Mountain Land, Namibia: implications for ore genesis. J. Afr. Earth Sci., 16, 265- 272. 33 ACCEPTED MANUSCRIPT CRIPT Pokrovsky, O.S., Viers, J., Freydier, R., 2005. Zinc stable isotope fractionation during its adsorption on oxides and hydroxides. J. Colloid and Interface Sci., 291, 192–200. Ponzevera, E., Quétel, C.R., Berglund, M., Taylor, P.D.P., Evans, P., Loss, R.D., Fortunato, G., 2006. Mass Discrimination During MC-ICPMS Isotopic Ratio Measurements: Investigation by Means of Synthetic Isotopic Mixtures (IRMM-007 Series) and Application to the Calibration of Natural-Like Zinc Materials (Including IRMM-3702 and IRMM-651). J. Amer. Soc. Mass Spectr., 17, 1413-1428. Pokrovsky, O.S., Viers, J., Freydier, R., 2005. Zinc stable isotope fractionation during its adsorption on oxides and hydroxides. J. Colloid and Interface Sci., 291, 192–200. CRIPT Pokrovsky, O.S., Viers, J., Freydier, R., 2005. AC References Zinc stable isotope fractionation during its adsorption on oxides and hydroxides. J. Colloid and Interface Sci., 291, 192–200. Ponzevera, E., Quétel, C.R., Berglund, M., Taylor, P.D.P., Evans, P., Loss, R.D., Fortunato, G., 2006. Mass Discrimination During MC-ICPMS Isotopic Ratio Measurements: Investigation by Means of Synthetic Isotopic Mixtures (IRMM-007 Series) and Application to the Calibration of Natural-Like Zinc Materials (Including IRMM-3702 and IRMM-651). J. Amer. Soc. Mass Spectr., 17, 1413-1428. USCR Rosman, K.J.R., 1972. A survey of the isotopic and elemental abundances of zinc. Geochim. Cosmochim. Acta, 36, 801-819. MANU Schneider, J., Boni, M., Laukamp, C., Bechstädt, T., Petzel, V., 2008. Willemite (Zn2SiO4) as a possible Rb–Sr geochronometer for dating nonsulfide Zn–Pb mineralization: Examples from the Otavi Mountainland (Namibia). Ore Geol. Rev. 33, 152–167. ED M Sivry, Y., Riotte, J., Sonke, J.E., Audry, S., Schäfer, J., Viers, J., Blanc, G., Freydier, R., Dupré, PTED B., 2008. Zn isotopes as tracers of anthropogenic pollution from Zn-ore smelters The Riou Mort–Lot River system. Chem. Geol., 255, 295-304. ACCEP Sonke, J.E., Sivry, Y., Viers, J., Freydier, R., Dejonghe, L., André, L., Aggarwal, J.K., Fontan, F., Dupré, B., 2008. Historical variations in the isotopic composition of atmospheric zinc deposition from a zinc smelter. Chem. Geol., 252, 145-157. AC Terracciano, R., 2008. Willemite mineralisation in Namibia and Zambia. Ph.D. Dissertation, Università degli Studi di Napoli Federico II, 178 pp. Wanty, R.B., Podda, F., De Giudici, G., Cidu, R., Lattanzi, P., 2013. Zinc isotope and transition- element dynamics accompanying hydrozincite biomineralization in the Rio Naracauli, Sardinia, Italy. Chem. Geol., 337 (338), 1–10. 34 ACCEPTED MANUSCRIPT Weiss, D.J., Boye, K., Caldelas, C., Fendorf, S., 2014. Zinc isotope fractionation during early dissolution of biotite granite. Soil Sci. Soc. Am. J., 78, 171–179 T Weiss, D., Rausch, N., Mason, T., Coles, B., Wilkinson, J., Ukonmaanaho, L., Arnold, T., Nieminen, T., 2007. Atmospheric deposition and isotope biogeochemistry of zinc in ombrotrophic peat. Geochim. Cosmochim. Acta 71, 3498–3517. RIPT Wilkinson, J.J., Hitzman, M.W., 2015. The Irish Zn-Pb Orefield: The View from 2014. In: S. M. SCRIP Archibald and S. J. Piercey (Eds.), Current Perspectives on Zinc Deposits Irish Association for Economic Geology, Dublin , 59-72. USC Wilkinson, J.J., Boyce, A.J., Everett, C.E., Lee, M.J., 2003. Timing and depth of mineralisation MANUSC Wilkinson, J.J., Boyce, A.J., Everett, C.E., Lee, M.J., 2003. Timing and depth of mineralisation in the Irish Zn-Pb orefield. Figure captions Figure captions References In: Kelly, J., Andrew, C.J., Ashton, J.H., Boland, M.B., Earls, G., Fusciardi, L., Stanley, G. (Eds.), Europe’s ma or base metal deposits, Dublin, Irish Association for Economic Geology, 483–498. ANU in the Irish Zn-Pb orefield. In: Kelly, J., Andrew, C.J., Ashton, J.H., Boland, M.B., Earls, MANU G., Fusciardi, L., Stanley, G. (Eds.), Europe’s ma or base metal deposits, Dublin, Irish Association for Economic Geology, 483–498. PTED M Wilkinson, J.J., Weiss, D.J., Mason, T.F.D., Coles, B.J., 2005. Zinc isotope variation In hydrothermal systems: preliminary evidence from the Irish Midlands ore field. Econ. Geol., 100 583–590. ACCEP Zhou, J.X., Huang, Z.L., Zhou, M.F., Zhu, X.K., Muchez, P., 2014. Zinc, sulphur and lead isotopic variations in carbonate-hosted Pb–Zn sulfide deposits, southwest China. Ore Geol. Rev., 58. 41-54. 35 ACCEPTED MANUSCRIPT Figure 1 – Locations of the nonsulfide Zn deposits sampled for this study. TED MANUSCRIPT Figure 2 – Typical ore samples from the studied localities. A. Belgium. Willemite from the La Calamine deposit. B. Poland. Sphalerite concretions directly replaced by smithsonite (from Coppola et al., 2009). C. Namibia. Willemite 1 directly replacing sphalerite, and cross-cut by veins and geodes filled by willemite 2 (Berg Aukas). D. Zambia. Willemite directly replacing sphalerite (Kabwe). E. Zambia. Supergene smithsonite concretions, with hemimorphite overgrowths (Kabwe). F. Yemen. Smithsonite concretions in drillcore J125 (Jabali) (from Mondillo et al., 2014a). G. Yemen. Galena and sphalerite oxidized to cerussite and smithsonite in drillcore J125 (Jabali) (from Mondillo et al., 2014a). H. Peru. Massive sauconite in drillcore YA20 (Yanque) (from Mondillo et al., 2014b). Hm = hemimorphite, Sau = sauconite, Sm = smithsonite, Sp = sphalerite, Will = willemite. ACCEPT Figure 3 – Diagram showing the distribution of 66ZnJCM-Lyon compositions of the studied Zn minerals, compared with literature data. 1 = data from Pašava et al. (2012); 2 = data from Wanty et al. (2013); 3 = data from Araújo et al. (2017); 4 = data from Wilkinson et al. (2005) and Wilkinson (unpublished); 5 = data from Archer and Vance (2002), Albarède (2004), Kelley et al. (2009), Maréchal et al., (1999), Mason et al. (2005), Pašava et al. (2014), Zhou et al. (2014). Figure 4 – Paragenetic relationships between the studied Zn minerals in the various districts. = data from Sonke et al. (2008); # = data from Wilkinson et al. (2005) and Wilkinson 36 ACCEPTED MANUSCRIPT (unpublished): 5th, 25th, 75th, and 95th percentiles are whiskers and box, the star is the median value. Grey field is the range of primary sulfides. (unpublished): 5th, 25th, 75th, and 95th percentiles are whiskers and box, the star is the median value. Grey field is the range of primary sulfides. ACCEPTED MANUSCRIPT Figure 5 - Rayleigh-type distillation model predicting the composition of the precipitating phase, assuming progressive precipitation from an evolving solution composition. A) The starting (bulk) 66ZnJCM-Lyon composition of the solution is 0.1‰ - based on the assumption of initial dissolution of sphalerite with bulk composition (the median value of all the sphalerites shown in Figure 4) - and the curves for a range of different -fractionation factors are shown. Figure 1 – Locations of the nonsulfide Zn deposits sampled for this study. =1.0015 reproduces the heaviest (which are formed from the first small amounts of precipitation) and the lightest values (which form when about 75% of the metal has been deposited) occurring in our dataset. At any point, the supersaturation of the solution could dump out the total Zn with the same composition of the fluid at that point (i.e. no further fractionation). B) The starting (bulk) 66ZnJCM-Lyon composition of the solution is 0.6‰, representing the fluid sourcing the Jabali hydrozincite (corresponding to the mean value of the precursor smithsonites): the model predicts precipitation of a hydrozincite with a composition of 0.25‰, after almost 75% of the metal in solution had been already deposited. [1] = -fractionation factors for the pairs hemimorphite- sphalerite, hydrozincite-sphalerite, and smithsonite-sphalerite (Ducher et al., 2016) – here we assume that the precipitating fluid has the same δ66Zn composition of dissolved sphalerite. [2] = -fractionation factor for the pairs hydrozincite-smithsonite (Ducher et al., 2016) – here we assume that the precipitating fluid has the same δ66Zn composition of dissolved smithsonite. 37 ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT Table 1. Zn isotopic compositions of the analyzed samples. Sample ID Location δ66Zn/64Z n ± δ66Zn/64Z n ± δ67Zn/64Z n ± δ68Zn/64Z n ± δ68Zn/64Z n ± ‰ IRMM 3702 2 S.D. ‰ Lyon JMC 3- 0749L 2 S.D. ‰ Lyon JMC 3- 0749L 2 S.D. ‰ IRMM 3702 2 S.D. ‰ Lyon JMC 3- 0749L 2 S.D. RN5011 Galena Belgium - Moresnet mine 0.34 0.19 0.52 0.28 0.71 0.35 RN5011 Willemite Belgium - Moresnet mine 0.98 0.09 1.51 0.22 2.24 1.46 RN2300 Willemite Belgium - Fossey mine 1.31 0.13 1.83 0.33 2.39 0.32 RN2300 Smithsonite Belgium - Fossey mine 0.44 0.06 0.50 0.55 0.75 0.54 R2B35/3925 Willemite Belgium - La Calamine mine 1.39 0.07 2.02 0.03 2.70 0.03 R2B35/3925 Smithsonite Belgium - La Calamine mine 0.85 0.08 1.30 0.22 1.61 0.39 BM1964R-6305 Sphalerite Ireland - Glendalough mine, Glendalough, Co Wicklow. 0.31 0.25 0.44 0.35 0.55 0.28 BM1964R-6305 Smithsonite Ireland - Glendalough mine, Glendalough, Co Wicklow. -0.15 0.08 -0.19 0.27 -0.27 0.21 BM1964R-6307 Sphalerite Ireland - Old Luganure mine (dump), Glendasan, Co Wicklow. 0.24 0.04 0.39 0.12 0.55 0.18 BM1964R-6307 Smithsonite Ireland - Old Luganure mine (dump), Glendasan, Co Wicklow. -0.09 0.07 -0.12 0.09 -0.14 0.08 CW-1 Sphalerite Ireland - Galmoy 0.33 0.078 CW-1 Smithsonite Ireland - Galmoy 0.353 0.147 CW-AF Smithsonite Ireland - Galmoy 0.455 0.071 Table 1. Figure 1 – Locations of the nonsulfide Zn deposits sampled for this study. Zn isotopic compositions of the analyzed samples. NUSCRIPT δ66Zn/64Z n ± δ67Zn/64Z n ± δ68Zn/64Z n ± δ68Zn/64Z n ± ‰ Lyon JMC 3- 0749L 2 S.D. ‰ Lyon JMC 3- 0749L 2 S.D. ‰ IRMM 3702 2 S.D. ‰ Lyon JMC 3- 0749L 2 S.D. 0.34 0.19 0.52 0.28 0.71 0.35 0.98 0.09 1.51 0.22 2.24 1.46 1.31 0.13 1.83 0.33 2.39 0.32 0.44 0.06 0.50 0.55 0.75 0.54 1.39 0.07 2.02 0.03 2.70 0.03 0.85 0.08 1.30 0.22 1.61 0.39 0.31 0.25 0.44 0.35 0.55 0.28 -0.15 0.08 -0.19 0.27 -0.27 0.21 0.24 0.04 0.39 0.12 0.55 0.18 -0.09 0.07 -0.12 0.09 -0.14 0.08 0.33 0.078 0.353 0.147 0.455 0.071 ACCEPTED BM1964R-6305 Sphalerite Ireland - Glendalough mine, Glendalough, Co Wicklow. BM1964R-6305 Smithsonite Ireland - Glendalough mine, Glendalough, Co Wicklow. BM1964R-6307 Sphalerite Ireland - Old Luganure mine (dump), Glendasan, Co Wicklow. BM1964R-6307 Smithsonite Ireland - Old Luganure mine (dump), Glendasan, Co Wicklow. ACCEPTED M mbia - Kabwe mine Pit 5/6 mbia - Kabwe mine Pit 5/6 mbia - Kabwe mine Pit 2 mbia - Kabwe mine Pit 2 D MAN D MAN D MANU ACCEPTE mbia - Kabwe mine Pit 2 mbia - Kabwe mine Pit 2 CEPTED MANU 0 0 -0 -0 0 -0 0 D MANU -0.0 0.08 0.07 -0.4 -0.0 ACCEPTE ACCEPTE et al. (2012) ACCEPTED M Figure 1 – Locations of the nonsulfide Zn deposits sampled for this study. CW-1 Sphalerite Ireland - Galmoy CW-1 Smithsonite Ireland - Galmoy CW-AF Smithsonite Ireland - Galmoy 38 ACCEPTED MANUSCRIPT ACCEPTED MA CV05-80 Sphalerite Poland - Pomorzanj mine CV05-80 Hemimorphite Poland - Pomorzanj mine CV05-71 Sphalerite Poland - Chrzanòw CV05-71 Smithsonite Poland - Chrzanòw NAR0427A-1 Sphalerite Namibia - Baltika mine NAR0427A-2 Willemite Namibia - Baltika mine BA2003-4 Willemite I Namibia - Berg Aukas mine BA2003-13.2 Willemite II Namibia - Berg Aukas mine ZA0502-D Sphalerite Zambia - Kabwe mine Pit 5/6 ZA0502-M Willemite I Zambia - Kabwe mine Pit 5/6 ZA0511-1 Hemimorphite Zambia - Kabwe mine Pit 2 ZA0511-2 Smithsonite Zambia - Kabwe mine Pit 2 J125-5 Sphalerite Yemen - Jabali -0.05 < ACCE CV05-80 Sphalerite Poland - Pomorzanj mine CV05-80 Hemimorphite Poland - Pomorzanj mine CV05-71 Sphalerite Poland - Chrzanòw CV05-71 Smithsonite Poland - Chrzanòw NAR0427A-1 Sphalerite Namibia - Baltika mine NAR0427A-2 Willemite Namibia - Baltika mine BA2003-4 Willemite I Namibia - Berg Aukas mine BA2003-13.2 Willemite II Namibia - Berg Aukas mine ZA0502-D Sphalerite Zambia - Kabwe mine Pit 5/6 ZA0502-M Willemite I Zambia - Kabwe mine Pit 5/6 ZA0511-1 Hemimorphite Zambia - Kabwe mine Pit 2 ZA0511-2 Smithsonite Zambia - Kabwe mine Pit 2 J125-5 Sphalerite Yemen - Jabali D MANUSCRIPT -0.01 0.05 -0.13 0.57 -0.01 0.16 0.22 0.07 0.34 0.15 0.45 0.15 0.26 0.12 0.41 1.04 0.36 0.51 0.48 0.06 0.67 0.12 0.83 0.21 -0.06 0.02 -0.06 0.08 -0.19 0.12 0.08 0.12 0.16 0.14 0.09 0.26 0.07 0.05 0.07 0.01 0.18 0.09 -0.42 0.07 -0.70 0.10 -0.92 0.11 -0.02 0.01 -0.10 0.05 -0.15 0.05 0.10 0.01 0.15 0.02 0.14 0.00 -0.26 0.04 -0.35 0.09 -0.40 0.07 0.40 0.11 0.67 0.17 0.77 0.14 δ66Zn/64Z n calculate d to ‰ JMC- Lyon (+0.29‰) * -0.05 <0.17 0.24 0.02 <0.17 ANUSCRIPT -0.01 0.05 -0.13 0.57 -0.01 0.16 0.22 0.07 0.34 0.15 0.45 0.15 0.26 0.12 0.41 1.04 0.36 0.51 0.48 0.06 0.67 0.12 0.83 0.21 -0.06 0.02 -0.06 0.08 -0.19 0.12 0.08 0.12 0.16 0.14 0.09 0.26 0.07 0.05 0.07 0.01 0.18 0.09 -0.42 0.07 -0.70 0.10 -0.92 0.11 -0.02 0.01 -0.10 0.05 -0.15 0.05 0.10 0.01 0.15 0.02 0.14 0.00 -0.26 0.04 -0.35 0.09 -0.40 0.07 0.40 0.11 0.67 0.17 0.77 0.14 E ZA0511-1 Hemimorphite Zambia - Kabwe mine Pit 2 CE ZA0511-2 Smithsonite Zambia - Kabwe mine Pit 2 J125-5 Sphalerite Yemen - Jabali J125-5 Sphalerite 39 ACCEPTED MANUSCRIPT TED MANUSCRIPT J125-7 Sphalerite Yemen - Jabali 0.11 <0.17 0.40 0.16 <0.17 J125-9 Smithsonite Yemen - Jabali 0.56 <0.17 0.85 1.19 <0.17 J125-32 Smithsonite Yemen - Jabali 0.23 <0.17 0.52 0.36 <0.17 JS MON 6 Smithsonite Yemen - Jabali 0.14 <0.17 0.43 0.25 <0.17 JS MON 22B Hydrozincite Yemen - Jabali -0.04 <0.17 0.25 -0.02 <0.17 YA 13 SAM 12 Sauconite Peru - Yanque - northern limb of the orebody (surficial) -0.04 <0.17 0.25 -0.21 <0.17 YA 17 SAM 16 Sauconite Peru - Yanque - central part of the orebody (surficial) 0.24 <0.17 0.53 0.38 <0.17 YA 20 SAM 19 Sauconite Peru - Yanque - eastern side of the orebody (surficial) -0.15 <0.17 0.14 -0.36 <0.17 YA 31 SAM 24 Sauconite Peru - Yanque - southern limb of the orebody (deep) 0.32 <0.17 0.61 0.70 <0.17 YA 34 SAM 25 Sauconite Peru - Yanque - southern limb of the orebody (deep) 0.25 <0.17 0.54 0.44 <0.17 YA 40 SAM 29 Sauconite Peru - Yanque - southern limb of the orebody (deep) 0.15 <0.17 0.44 0.25 <0.17 YA 45 SAM 32 Sauconite Peru - Yanque - southern limb of the orebody (deep) 0.07 <0.17 0.36 -0.01 <0.17 * ti l f M ll t l (2012) 40 ACCEPTED MANUSCRIPT CCEPTED MAN D MAN Highlights Zinc isotope data from nonsulfide zinc mineral separates (willemite, smithsonite, Zinc isotope data from nonsulfide zinc mineral separates (willemite, smithsonite, hemimorphite, hydrozincite, sauconite) collected from seven major ore deposits worldwide IPT Significant fractionations both in positive and negative directions are recorded relative to the precursor phase SCRIP Partial dissolution of primary sphalerite is followed by precipitation of a secondary phase preferentially incorporating heavy Zn isotopes USC Negative isotopic shifts are only observed for late crystallizing phases, and are interpreted to NUSC Negative isotopic shifts are only observed for late crystallizing phases, and are interpreted to reflect precipitation from residual, isotopically light fluids NU reflect precipitation from residual, isotopically light fluids 41 Figure 1 Figure 2 Figure 2 Figure 2 Figure 3 Figure 4 Figure 5
https://openalex.org/W4205257137	https://link.springer.com/content/pdf/10.1007/s13202-022-01455-8.pdf	English	null	Shale gas well productivity potential evaluation based on data-driven methods: case study in the WY block	Journal of petroleum exploration and production technology	2,022	cc-by	8,324	Shale gas well productivity potential evaluation based on data‑driven methods: case study in the WY block Chaodong Tan1,2 · Hanwen Deng1,2 · Wenrong Song3 · Huizhao Niu3 · Chunqiu Wang1,2 Received: 12 September 2021 / Accepted: 3 January 2022 / Published online: 16 January 2022 © The Author(s) 2022 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 https://doi.org/10.1007/s13202-022-01455-8 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 https://doi.org/10.1007/s13202-022-01455-8 ORIGINAL PAPER-PRODUCTION ENGINEERING 1 State Key Laboratory of Petroleum Resource and Prospecting, China University of Petroleum, Changping, Beijing 102249, China * Hanwen Deng denghanwen1998@163.com Vol.:(0123456789) 1 3 Abstract Evaluating the productivity potential of shale gas well before fracturing reformation is imperative due to the complex frac- turing mechanism and high operation investment. However, conventional single-factor analysis method has been unable to meet the demand of productivity potential evaluation due to the numerous and intricate influencing factors. In this paper, a data-driven-based approach is proposed based on the data of 282 shale gas wells in WY block. LightGBM is used to conduct feature ranking, K-means is utilized to classify wells and evaluate gas productivity according to geological features and fracturing operating parameters, and production optimization is realized through random forest. The experimental results show that shale gas productivity potential is basically determined by geological condition for the total influence weights of geologic properties take the proportion of 0.64 and that of engineering attributes is 0.36. The difference between each category of well is more obvious when the cluster number of well is four. Meanwhile, those low production wells with good geological conditions but unreasonable fracturing schemes have the greatest optimization space. The model constructed in this paper can classify shale gas wells according to their productivity differences, help providing suggestions for engineers on productivity evaluation and the design of fracturing operating parameters of shale gas well. Keywords Shale gas · Productivity potential evaluation · LightGBM · K-means · Optimization Abbreviations r(X, Y) Linear correlation coefficient Cov(X, Y) Covariance 휎(X) and 휎(Y) Variances M Number of trees vt Splitting feature ̂I2 j Corresponding empirical improvement in squared-error a(i) Mean distance between i and all other data points in the same cluster d(i, j) Distance between data points i and j in the cluster Ci b(i) Smallest mean distance of i to all points any other cluster Bk Between group dispersion matrix Wk Within-cluster dispersion N Number of data points Cq Point set in cluster q cq Center of cluster q c Center of matrix trace E nq Number of points in cluster q b Individual bootstrap samples B Sum of trees t∗ b Individual learner ̂yi True value yi Predicted value Introduction (2019) established a prediction model of gas production in Marcellus shale gas area with 31 input parameters including well information, reservoir parameters, rock mechanics parameters, comple- tion parameters, and fracturing parameters by using neural network algorithm, and optimized parameters such as perfo- ration cluster number, fracturing stage and proppant volume through this model. Liu et al. (2021) applied deep learning method to predict estimated ultimate recovery (EUR) based on 17 parameters of the geological, engineering and produc- tion factors involved with the EUR. Liang et al. (2021) built a multi-objective random forest method to predict the shale gas production performance throughout the entire shale gas production process. However, there were no specific con- clusions about the controlling factors of gas productivity. Vikara et al. (2020) used gradient boosted regression tree (GBRT)-based predictive model to estimate ultimate recov- ery and concluded that net thickness and porosity were the vital influencing factors. However, whether geologic attrib- ute is the first principal controlling factor remains unclear. Meanwhile, it is still not explicit to know how important the engineering factors such as fracturing parameters are to the productivity. Therefore, it is necessary to rank the impor- tance of the factors that affect the productivity of shale gas wells and determine the category of parameter attributes with the largest influence weight. If the weight of geological attribute ranks first, it is of great significance for shale gas wells to develop different fracturing optimization schemes according to different geological conditions to increase the output of single well and reduce production cost. In view of the current research status, this study seeks to Shale gas is an important unconventional natural gas resource with huge exploration potentials (Yao et al. 2019; Zhu et al. 2020). The success of the shale gas revolution in North America further highlights its resource potential and development prospects (Ma et al. 2018). Shale gas reservoirs are characterized by low porosity, low perme- ability and heterogeneity, and the flow law in the matrix of shale gas reservoirs no longer follows Darcy flow com- pared with conventional reservoirs (Yan et al. 2017). Due to the ultralow porosity and permeability, it requires the hori- zontal well drilling and the multi-stage hydraulic fractur- ing technology to produce the shale gas (Jiang et al. 2020). Introduction 1 State Key Laboratory of Petroleum Resource and Prospecting, China University of Petroleum, Changping, Beijing 102249, China The research area in this paper focused on Weiyuan (WY) shale gas block. WY block, which is located in Sichuan basin, is a representative block with the achievement of industrial production in the Lower Cambrian shale gas exploration in China (Cao et al. 2020). The whole block has 2 College of Petroleum Engineering, China University of Petroleum, Changping, Beijing 102249, China 3 Beijing Yadan Petroleum Technology Development Co., LTD., Changping 102200, China Vol.:(0123456789) 1 3 Vol.:(0123456789) 1 3 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 2348 between geological parameters and gas productivity. There- fore, the evaluation of shale gas production potential should be established on the determination of the controlling factors of productivity. Furthermore, well optimization will be more reliable when based on the important factors influencing gas recovery. a monoclinic structure inclined to the southeast, and faults are not well developed in this block. The reservoir properties are generally high, with total organic carbon (TOC) value of 2.1–8.1% and average of 3.2%, porosity of 2.1–9.6% and average of 5.9%, brittle mineral content of 60–82% and aver- age of 74%, gas content of 3.3–8.5 m3/t and average of 5.5 m3/t. The shale gas play is characterized by small thickness (1–7 m and an average of 3.8 m) of high-quality reservoir, large horizontal stress difference (7.7–18.7 MPa) (Ma et al. 2020; Liu et al.2021). With the rapid development of AI, machine learning methods are widely applied in reservoir waterflooding (Ma et al.2019), heavy oil recovery (Lu et al.2021), performance evaluation of shale gas (Syed et al.2021) and other oil and gas fields. Compared with data-driven models, the mecha- nism-based reservoir simulation models for parameter sensi- tivity analysis are too time-consuming and computationally rigorous to conduct (Bettin et al. 2019; Mo et al. 2019). Selection of the principal parameters influencing gas produc- tion should consider both geologic and engineering param- eters. Currently, the commonly used data-driven method to determine the main influencing factors of gas production is to establish a prediction model between the influencing fac- tors and productivity parameters. He et al. 1 3 Introduction The productivity of shale gas wells is affected by the energy utilization rate of the fracturing fluid, the closure law of main fractures and induced fractures, etc. (Mohammadmo- radi and Kantzas, 2018). The challenge in developing shale gas lies in the inherent problem of multi-parameter decision making, especially when substantial uncertainty may exist with underground conditions and major economic factors. Additionally, since the development cost of shale gas wells is generally higher than that of conventional oil and gas, especially the fracturing cost, it is important to develop wells with high production potential and design reasonable frac- turing parameters to improve the economic benefits. i In petroleum geoscience, sweet spots are regarded as tar- get regions within the best production potential (Hauge et al. 2017). Engineers have employed traditional approaches like core analysis, well logging and seismic test to identify the locations of sweet spots (Liu et al.2019; Zhao et al.2019). However, these methods are not persuasive, because there is no obvious equivalence between the measured param- eters and the well's productivity, and the fact that there is no uniform standard for the identification of the sweet spot location from one field to another makes conventional methods less applicable. Some scholars have also adopted data-driven methods, such as neural network, fuzzy logic and support vector machine, to conduct productivity evalu- ation. The usual practice is to use logging data, seismic test data and production data to predict geological parameters such as TOC that related to the location of sweet spots or perform other analyses (Tahmasebi et al. 2017; Qian et al. 2018; Kurison et al. 2019). But the above techniques remain uncertain as they still do not fully explain the relationship In view of the current research status, this study seeks to achieve the well productivity evaluation, dominating factors analysis and production optimization through machine learn- ing methods. The model constructed in this paper can give 1 3 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 2349 fracturing productivity. Therefore, evaluating the produc- tivity potential of shale gas wells prior to fracturing will greatly reduce the probability of low production caused by the failure of fracturing parameter design, thus improving the overall benefit. engineers a more comprehensive understanding of the effect of various factors on productivity. Data description and statistical analysis The dataset obtained from the 282 fractured wells in the study area contains 12 different types of data items, includ- ing 7 geological features (vertical depth, TOC content, porosity, thickness of organic-rich gas reserves, gas satura- tion, formation pressure coefficient, brittle mineral content), 4 engineering features or fracturing operating features (aver- age fracturing interval length, amount of fracturing fluid, amount of fracturing sand, average fracturing fluid displace- ment), and first-year daily gas production rate ( Qg) is taken as the evaluation index of gas productivity. Due to the lim- ited field conditions, some parameters such as permeability cannot be obtained temporarily, but the method proposed in this paper is still feasible and can be extended to more data. Table 1 is the descriptive statistics of all wells, serial number of geological features is 1–7, and that of fractur- ing operating features is 8–11, and No.12 is the label of the dataset. Among them, geological factors are difficult to be changed artificially, while fracturing factors are more con- trollable to field engineers, and can be adjusted manually, both of which are related to Qg . The productivity of these wells varies widely, with the maximum Qg value being more than 170 times of the minimum Qg value, and the variance of Qg is large, indicating that the productivity data is scattered. From a mechanical point of view, poor geological condi- tions or unreasonable fracturing parameters will lead to low The dataset obtained from the 282 fractured wells in the study area contains 12 different types of data items, includ- ing 7 geological features (vertical depth, TOC content, porosity, thickness of organic-rich gas reserves, gas satura- tion, formation pressure coefficient, brittle mineral content), 4 engineering features or fracturing operating features (aver- age fracturing interval length, amount of fracturing fluid, amount of fracturing sand, average fracturing fluid displace- ment), and first-year daily gas production rate ( Qg) is taken as the evaluation index of gas productivity. Due to the lim- ited field conditions, some parameters such as permeability cannot be obtained temporarily, but the method proposed in this paper is still feasible and can be extended to more data. Table 1 is the descriptive statistics of all wells, serial number of geological features is 1–7, and that of fractur- ing operating features is 8–11, and No.12 is the label of the dataset. Introduction Additionally, it can help provide suggestions for engineers on the design of fracturing parameters and resource allocation of new wells, which can also improve the economic benefits. i This paper will calculate the influence weight of each feature through feature ranking, find the principal fac- tors affecting productivity, evaluate productivity potential through feature reconstruction and clustering, and finally establish a random forest predictive model to optimize the comprehensive index of fracturing operating parameters of low production wells. A preliminary data analysis was conducted before the above work. The relationship between each feature and the response variable was independently measured by the univariate feature selection method, and the linear correlation coefficient r between each parameter was calculated by Eq. (1). From Fig. 1, it can be seen that there is almost no strong linear relationship between geological and fracturing operation parameters with Qg , only vertical depth is positively related to formation pressure coefficient. The correlation coefficients of X1 and X2 with X12 are all 0.5, indicating thickness of organic-rich gas reserves and gas saturation are moderately correlated with Qg . Therefore, con- ventional single-factor analysis method has been unable to meet the demand of productivity potential evaluation due to the numerous and intricate influencing factors. Data-driven methods provide the possibility to solve the problem. And the scattered correlation matrix of these features is shown in Fig. 2. 1 3 Feature ranking by LightGBM Fig. 1 Heat diagram of linear correlation coefficient of each param- eter, X1 is vertical depth, X2 is TOC content, X3 is porosity, X4 is thickness of organic-rich gas reserves, X5 is gas saturation, X6 is for- mation pressure coefficient, X7 is brittle mineral content, X8 is aver- age fracturing fluid displacement, X9 is average fracturing interval length, X10 is amount of fracturing fluid, X11 is amount of fracturing sand, X12 is Qg The feature ranking in this paper is model-based, because some machine learning methods have a scoring mechanism for features. Therefore, LightGBM (Light gradient boost- ing machine learning) algorithm is selected to establish the prediction model between each feature and daily gas production, and its ranking mechanism is used to calculate the influence weight of geological features and fracturing parameters. p LightGBM is an improved variant of gradient boosting decision tree (GBDT) algorithm (Dj et al.2020); it com- bines the predictions of multiple decision trees to make the final prediction generalizes well and contains two novel techniques: Gradient-based One-Side Sampling and Exclu- sive Feature Bundling to deal with large number of data samples and large number of features, respectively (Ke et al. 2017). LightGBM algorithm divides the training set into multiple models, and in each iteration, the selection of the first k features is determined by local voting, and then the selection of the first 2k features is determined by global voting. The default training decision tree in Light- GBM algorithm uses histogram algorithm, which does not need to find splits on the classified feature values, but stores continuous feature values in discrete bins which are used to construct feature histograms during the training process. This reduces the number of data instances and features, thus shortening training time (Qin et al. 2021). LightGBM can be used for featuring ranking since greedy feature selection already happens in the algorithm when selecting splitting features, for example, splitting fea- tures and splitting points of regression trees are found to minimize the squared-error loss for any given partition of the data. Moreover, as a byproduct, a sorted list of rela- tive importance of features is automatically generated for each boosted trees model (Feng et al.2009).Friedman (J.H, 2001) generalizes the global importance of feature xj is measured by the average value of the importance of feature xj in a single tree: Fig. Data description and statistical analysis Among them, geological factors are difficult to be changed artificially, while fracturing factors are more con- trollable to field engineers, and can be adjusted manually, both of which are related to Qg . The productivity of these wells varies widely, with the maximum Qg value being more than 170 times of the minimum Qg value, and the variance of Qg is large, indicating that the productivity data is scattered. From a mechanical point of view, poor geological condi- tions or unreasonable fracturing parameters will lead to low r(X, Y) = Cov(X, Y) √ 휎(X) ⋅휎(Y) Table 1 Statistics of the dataset from the research block Table 1 Statistics of the dataset from the research block a.u. refers to arbitrary unit (dimensionless) No Dataset Unit Standard deviation Average value Minimum value Maximum value 1 Vertical depth m 343 3015 2200 3800 2 TOC content % 0.75 5.38 3.30 7.80 3 Porosity % 0.94 7.15 5.70 8.90 4 Thickness of organic-rich gas reserves m 1.37 5.19 2.30 7.60 5 Gas saturation % 3.17 75.79 60 83 6 Formation pressure coefficient a.u 0.18 1.74 1.35 2.05 7 Brittle mineral content % 7.26 78.26 62.50 96 8 Average hydraulic fracturing fluid displacement m3/min 1.30 11.82 6.19 15 9 Average hydraulic fracturing interval length m 11.61 68.30 43.10 145.70 10 Amount of hydraulic fracturing fluid m3/m 4.39 27.31 7.12 47.94 11 Amount of hydraulic fracturing sand t/m 0.39 1.64 0.29 3 12 First-year daily gas production rate (Qg) 104m3/d 5.49 9.50 0.19 33.30 1 3 2350 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 Feature ranking by LightGBM 1 Heat diagram of linear correlation coefficient of each param- eter, X1 is vertical depth, X2 is TOC content, X3 is porosity, X4 is thickness of organic-rich gas reserves, X5 is gas saturation, X6 is for- mation pressure coefficient, X7 is brittle mineral content, X8 is aver- age fracturing fluid displacement, X9 is average fracturing interval length, X10 is amount of fracturing fluid, X11 is amount of fracturing sand, X12 is Qg r(X, Y) is the linear correlation coefficient between vari- ables X and Y , Cov(X, Y) is the covariance between X and Y,휎(X) and 휎(Y) are the variances. The value of r ranges from − 1 to 1. The greater the absolute value of r , the stronger the correlation between two parameters, and vice versa. When r is equal to 0, it can only indicate that there is no linear correlation between the two parameters, not irrelevant. 1 3 Methodology There are wide variations in gas productivity of these wells, in order to improve shale gas production, it is of great sig- nificance to further explore the reasons for the productivity differences among wells and find out the controlling factors that affect the productivity.i (2) ̂J2 j = 1 M ∑M m=1̂J2 j Tm (2) M is the number of trees, and the relative importance of feature xj in each tree is calculated as: f Specifically, the data from 282 shale gas wells will be used to: (1) calculate the importance weight of each influ- encing factors and conduct feature ranking by LightGBM. (2) Calculate comprehensive index of geological parameters and fracturing operating parameters through feature recon- struction, and combine new features with first-year daily gas production rate to form a new dataset. (3) Classify the new dataset so as to evaluate the gas productivity potential through K-means. (4) Using random forest predictive model to optimize wells with low Qg. (3) ̂J2 j (T) = ∑L−1 t=1̂I2 j 1(vt = j)) (3) where the summation is over the internal nodes t of a L -terminal node tree T , vt is the splitting feature associated with node t , and ̂I2 j is the corresponding empirical improve- ment in squared-error as a result of the split.l Through this method, the influence weight of each fea- ture, the total influence weight of geological features and 1 3 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 2351 Fig. 2 Scattered correlation matrix for features distribution Fig. 2 Scattered correlation matrix for features distribution The commonly used method of feature dimension reduction is feature extraction. Feature extraction refers to the use of some specific feature extraction methods to transform the original data, such as PCA (Principal Com- ponent Analysis), LDA (Linear Discriminant Analysis), etc. (Martinez et al. 2001; Shlens, 2014). The dimension of the data is reduced after feature extraction. Although the new feature has statistical significance and retains the main information of the original data, the physical mean- ing is not clear. Feature construction refers to artificially fracturing operating features will be obtained. Thus, features have greater impact on gas production will be known. 1 3 Feature reconstruction Since the ground truth labels are not known, two internal evaluation methods, Silhouette coefficient (SC) and Calinski–Harabasz (CH) index, will be adopted as the evaluation indexes of clustering performance.i In this study, there are 282 samples, 7 geological fea- tures and 4 fracturing operating features, thus the compre- hensive index of geologic attribute c1 = x282×7×w1×7 T ∑7 j=1yj , and the comprehensive index of fracturing operating attrib- utec2 = x282×4×w1×4 T ∑4 j=1yj . Therefore, c1 , c2 and Qg will form a new data matrix of 282 rows and 3 columns. In this study, there are 282 samples, 7 geological fea- tures and 4 fracturing operating features, thus the compre- hensive index of geologic attribute c1 = x282×7×w1×7 T ∑7 j=1yj , and the j 1 j comprehensive index of fracturing operating attrib- utec2 = x282×4×w1×4 T ∑4 j=1yj . Therefore, c1 , c2 and Qg will form a new data matrix of 282 rows and 3 columns. SC was first proposed by Peter (1999); it is based on the comparison of its tightness and separation, and can be used to evaluate the influence of different algorithms or different operation modes on clustering results based on the same original data. SC ranges from − 1 to 1, the closer the value is to 1, the more reasonable the clustering result is. If the value is close to − 1, it means that the sample should be classified into another cluster. And the closer the value is to − 1, the more reasonable the sample clustering should be classified into another cluster。. The SC for a set of samples is given as the mean of the SC for each sample. Feature reconstruction In order to facilitate the classification of shale gas wells, it is necessary to reduce the dimensionality of the original fea- ture parameters, and feature reconstruction is used to obtain the comprehensive index of geologic features and fracturing operating features, respectively. 1 3 1 3 1 3 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 2352 constructing new features from original data, such as using mixed attributes or combined attributes to create new fea- tures, or decomposing and segmenting original features to create new features. Therefore, this paper will adopt feature construction, combined with domain knowledge to reconstruct new features to achieve feature reduction. The detailed calculation process is as follows. (1) Input N samples {x1, x2, … , xN }, xn ∈RD , and input the proposed number of clustering k (1) Input N samples {x1, x2, … , xN }, xn ∈RD , and input the proposed number of clustering k (2) Randomly initialize K vectors of D dimension, and select K different sample points as the initial cluster- ing center g (3) Repeat the following process until convergence g (3) Repeat the following process until convergence A data set with N samples, M features and F attributes can be set as XN×M . Suppose mi is the number of features belonging to the i th attribute, thus M = ∑F i=1mi , XN×M = [xN×m1, xN×m2, ⋯, xN×mF].Set the importance weight matrix as W = [w1×m1, w1×m2, … , w1×mF ]T, w1×mi = [y1, y2, ⋯, ymi ], yj is the feature weight, ∑F i=1 ∑mi j=1yj = 1 , and the new feature matrix can be represented as: CN×F=[c1, c2, … , cF ], where: For each object xn , calculate the class it should belong to, as in the following formula: ck = {n ∶k = argmin\|xn −휇k\|2} Recalculate the centroid of the class: Recalculate the centroid of the class: (6) 휇k = 1 \|\|ck\|\| ∑ n∈ck xn 휇k = 1 \|\|ck\|\| ∑ n∈ck xn (6) ci = xN×mi × w1×mi T ∑mi j yj The iterative convergence definition is that the clustering center remains unchanged or the sum of distances from each sample to the corresponding clustering center no longer has a great change. 1 3 Results and discussion LightGBM algorithm is selected as the feature selection algorithm to determine the importance weight of each fea- ture, and normalization processing was conducted to elimi- nate any differences in the order of magnitude among the various input parameters. In the training process, 80% of the data set is divided into training set and 20% is divided into test set, and the grid search method is used to adjust the hyperparameters of the LightGBM predictive model, and the hyperparameters are shown in Table 2. In this paper, Mean square error (MSE) is used as the evaluation index of the regression model, it is calculated by Eq. (15), where n represents the number of shale gas wells, ̂yi is true value, yi is predicted value, the smaller the value of MSE, the higher the accuracy of the prediction model. MSE value of training set is 0.006 and that of test set is 0.01, which is small enough to make the model feasible. (10) CH(k) = Tr (Bk ) Tr (Wk ) × N −k k −1 (11) Wk = ∑k q=1 ∑ x∈Cq (x −cq)(x −cq )T (12) Bk = ∑ q nq(cq −c)(cq −c)T (10) CH(k) = Tr (Bk ) Tr (Wk ) × N −k k −1 (10) (11) Wk = ∑k q=1 ∑ x∈Cq (x −cq)(x −cq )T (12) Bk = ∑ q nq(cq −c)(cq −c)T (11) (12) Bk = ∑ q nq(cq −c)(cq −c)T (12) Bk is between group dispersion matrix,Wk is within- cluster dispersion,N is the number of data points, Cq is the point set in cluster q , cq is the center of cluster q , c is the center of matrix trace E , and nq is the number of points in cluster q. (15) MSE = 1 n n ∑ i=1 (̂yi −yi )2 (15) Productivity potential evaluation by K‑means After constructing the new data set, the wells will be clas- sified by clustering algorithm, and the productivity poten- tial will be evaluated, so as to select those low productivity wells with good geological conditions but unreasonable fracturing parameters design to facilitate subsequent pro- ductivity optimization.i The SC of object i is: K-means algorithm is selected for well classification as it is one of the most commonly used and important clustering algorithms, the basic idea is to find the partition scheme of K clusters through iteration, and use the mean value of these K clusters to represent the corresponding sample types, so as to minimize the total error (Chu X et al. 2020). Suppose 휇1, 휇2, … , 휇k are K cluster centers, rnk ∈{0, 1} is used to indicate whether xn belongs to clus- ter K , the loss function is defined as: (7) SC(i) = b(i) −a(i) max{a(i), b(i)} SC(i) = b(i) −a(i) max{a(i), b(i)} (7) a(i) = 1 \|\|Ci\|\| ∑ j∈C,j≠i d(i, j) b(i) = min k≠i 1 \|\|Ck\|\| ∑ j∈Ck d(i, j) (5) J(휇, r) = ∑K n=1 ∑K k=1 rnkxn −휇k 2 (5) a(i) is the mean distance between i and all other data points in the same cluster,d(i, j) is the distance between data points i and j in the cluster Ci,b(i) is the smallest mean dis- tance of i to all points in any other cluster, of which i is not a member. The cost function of above formula can only be mini- mized by iterative methods. The K-means algorithm clusters samples into k clusters, where k is given by the user. The solution process of the algorithm is described as follows: The CH index, also known as the variance ratio crite- rion—can be used to evaluate the clustering model, where 1 3 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 2353 a higher CH score relates to a model with better defined clusters. Production optimization through random forest Through the previous steps, wells with good geological conditions but low productivity will be selected for opti- mization, and random forest (RF) will be used to establish a regression model between new characteristic parameters and Qg . When optimizing the gas productivity of a sin- gle well, fix the geological index of the well, change the fracturing operating index and iterate through the trained predictive model to search for the corresponding fractur- ing operating index when the output is maximum, so as to achieve productivity optimization. The influence weight of each feature is shown in Fig. 3. It can be seen that thickness of organic-rich gas reserves and amount of hydraulic fracturing sand are vital parameters affecting gas productivity. The total weight of geological fea- tures is 0.64 and that of fracturing operating features is 0.36, indicating that geological condition is more deterministic to gas productivity. The result also proves that gas production of well in poor reservoir condition is difficult to be improved by adjusting engineering parameters. New sample matrix C282×3 is constructed according to Eq. (4); the new data statistics is shown in Table 3. Through feature transformation, 11 features were compressed to 2-dimension according to 2 feature attributes, and a three- dimensional dataset, including comprehensive index of RF is an ensemble of CART (Classification and regres- sion tree) and uses a bootstrap method based on the clas- sification and regression tree analysis to predict a con- tinuous response variable (Di Fina et al.2015; Ftpa B et al.2021). Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 Fig. 3 Histogram of importance weight of features Table 3 Statistics of the reconstructed dataset Feature Unit Standard deviation Average value Minimum value Maximum value Qg 104m3/d 5.49 9.50 0.19 33.30 Comprehensive index of geologic features ( c1) a.u 31.73 310.96 234.31 384.57 Comprehensive index of fracturing operating features ( c2) a.u 2.11 24.85 19.37 38.42 Fig. 4 Change curve of SC with cluster number k Fig. 5 Change curve of with CH index with cluster number k Table 3 Statistics of the reconstructed dataset Feature Unit Standard deviation Average value Minimum value Maximum value Qg 104m3/d 5.49 9.50 0.19 33.30 Comprehensive index of geologic features ( c1) a.u 31.73 310.96 234.31 384.57 Comprehensive index of fracturing operating features ( c2) a.u 2.11 24.85 19.37 38.42 Fig. 4 Change curve of SC with cluster number k Fig. 5 Change curve of with CH index with cluster number k Fig. 4 Change curve of SC with cluster number k Fig. 5 Change curve of with CH index with cluster number k geological features ( c1 ), comprehensive index of fractur- ing operating features ( c2 ) and Qg was formed for further analysis.f acceptable. In order to evaluate the productivity potential of the well more reasonably, the cluster diagrams when k is 2, 3, and 4 are drawn, as shown in Fig. 6. When the cluster number is 4, the difference between various types of wells is more obvious, which is more conducive to subsequent analy- sis. Therefore, this article divides the 282 shale gas wells in the study block into 4 categories, detailed feature description and data statistics of each cluster are shown in Table 4. The values of Qg and c1 values in cluster 2 are very small, which demonstrate that this type of wells has the poorest geological conditions, low productivity potential, and low economic development benefits. The values of c1 in cluster 1 and clus- ter 3 is relatively large, indicating that these two types of Due to the large difference between the values of Qg , c1 and c2 , data preprocessing is required before classification. This article uses the method of MinMaxScaler to compress these values to between 0 and 1. Then, clustering will be performed by K-means after the normalization of new data- set. 1 3 Production optimization through random forest Regression in RF is generated as an ensemble estimate from a number of decision trees based on bag- ging; the final output is the average of outputs of indi- vidual trees (Prasad et al.2006; Adusumilli et al.2013) 1 3 Table 2 Hyperparameters of LightGBM Hyperparameter Type/Value boosting_type gbdt objective Regression num_leaves 20 max_depth 6 learning_rate 0.09 feature_fraction 0.8 bagging_freq 1 bagging_fraction 0.8 bagging_seed 1 lambda_l1 0.001 lambda_l2 1 verbosity − 1 n_estimators 50 1 3 Table 2 Hyperparameters of LightGBM Hyperparameter Type/Value boosting_type gbdt objective Regression num_leaves 20 max_depth 6 learning_rate 0.09 feature_fraction 0.8 bagging_freq 1 bagging_fraction 0.8 bagging_seed 1 lambda_l1 0.001 lambda_l2 1 verbosity − 1 n_estimators 50 Table 2 Hyperparameters of LightGBM (13) ̂휃B(x) = 1 B ⋅∑B b=1 t∗ b(x) (13) b is the individual bootstrap samples, B is the sum of trees, t∗ b is the individual learner, and the individual deci- sion tree: b is the individual bootstrap samples, B is the sum of trees, t∗ b is the individual learner, and the individual deci- sion tree: (14) t∗ b(x) = t (x;z∗ b1, z∗ b2, … , z∗ bK ) (14) z∗ bk(k=1,2,…,K ) is the k th training sample with pairs of values for the target variable ( y ) and covariates(x ):z∗ bi = (xk, yk). z∗ bk(k=1,2,…,K ) is the k th training sample with pairs of values for the target variable ( y ) and covariates(x ):z∗ bi = (xk, yk). 2354 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 For wells with good geological conditions, the investment can be increased as designing reasonable fracturing parameters for hydrau- lic fracturing will bring greater economic benefits, and for wells with poor geological conditions, the investment cost can be relatively reduced. By considering different reservoir characteristics, the fracturing scheme is customized to make it more specific and applicable. i After completing the above work, six wells are selected from cluster 1 for productivity optimization, because these wells have large optimization space. Taking c2 as the opti- mization variable, the objective function is to maximize Qg by changing the value of c2 , and RF algorithm is used to establish a predictive model between c1 , c2 and Qg of 282 wells. The ratio of the training set to the test set of the predictive model is 7:3. When the accuracy of the predic- tive model reaches the standard, it starts to iterate. During the optimization process of each well, only the value of c2 is adjusted while c1 remained unchanged in every itera- tion, and the value of c2 varies in the range of 0 and 1. The results are shown in Fig. Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 In order to obtain the best cluster number, sensitivity analysis of cluster number k is carried out by using two internal evaluation methods: SC and CH. From Fig. 4 and Fig. 5, it can be seen that when k is 2, the clustering effect is the best, while when k is 3 or 4, the clustering effect is also 1 3 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 2355 Fig. 6 Clustering graph of normalized Qg , comprehensive index of geologic features ( c1 ) and fracturing operation features ( c2) Fig. 6 Clustering graph of normalized Qg , comprehensive index of geologic features ( c1 ) and fracturing operation features ( c2) Fig. 6 Clustering graph of normalized Qg , comprehensive index of geologic features ( c1 ) and fracturing operation features ( c2) 1 3 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 2356 Table 4 Data statistics of 4 types of wells in the study block Cluster Number of wells Normal- ized feature Average value Minimum value Maximum value Description 1 77 Qg 0.24 0.024 0.45 Good geological condition, poor fracturing measures and low produc- tivity c1 0.79 0.64 0.95 c2 0.41 0 0.70 2 67 Qg 0.19 0.07 0.41 Poor geological condition and the productivity is low regardless of the fracturing measures c1 0.19 0 0.38 c2 0.50 0.31 0.71 3 31 Qg 0.61 0.37 1 Superior geological condition, good fracturing measures and high productivity c1 0.83 0.60 1 c2 0.61 0.29 1 4 107 Qg 0.27 0 0.58 Medium geological condition, medium productivity c1 0.49 0.30 0.64 c2 0.43 0.11 0.75 Table 4 Data statistics of 4 types of wells in the study block wells have superior geological conditions, while the Qg value of cluster 1 is low, which proves that the fracturing opera- tion parameter design of this type of wells is unreasonable and has large optimization space. If secondary fracturing is required, such kind of wells should be considered first. The fracturing parameter design of wells in cluster 3 is reason- able and has important significance. Table 5 summarizes the parameter range of wells in cluster 3, and the design of fracturing operating parameters can refer to the wells in cluster 3. Conclusively, the geological conditions of the well should be precisely evaluated to determine the productiv- ity potential of the well before fracturing. Table 5 Statistical distribution of parameters in cluster 3 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 7, the red line in the figure represents the original Qg , and the blue dotted curve is the change trend of the optimized Qg with the number of 1 3 Table 5 Statistical distribution of parameters in cluster 3 Feature Unit Mean value Minimum value Maximum value Vertical depth m 2991.94 2600.00 3600.00 TOC content % 5.41 4.85 6.40 Porosity % 7.47 6.40 8.70 Thickness of organic-rich gas reserves m 6.83 5.20 7.60 Gas saturation % 80.16 73.50 83.00 Formation pressure coefficient a.u 1.73 1.52 1.90 Brittle mineral content % 82.39 73.80 96.00 Average fracturing fluid displacement m3/min 12.37 10.60 15.00 Average fracturing interval length m 71.05 44.95 94.00 Amount of fracturing fluid m3/m 28.59 23.58 47.94 Amount of fracturing sand t/m 2.10 0.92 3.00 Qg 104m3/d 20.54 12.61 33.30 Feature Unit Mean value Minimum value Maximum value Vertical depth m 2991.94 2600.00 3600.00 TOC content % 5.41 4.85 6.40 Porosity % 7.47 6.40 8.70 Thickness of organic-rich gas reserves m 6.83 5.20 7.60 Gas saturation % 80.16 73.50 83.00 Formation pressure coefficient a.u 1.73 1.52 1.90 Brittle mineral content % 82.39 73.80 96.00 Average fracturing fluid displacement m3/min 12.37 10.60 15.00 Average fracturing interval length m 71.05 44.95 94.00 Amount of fracturing fluid m3/m 28.59 23.58 47.94 Amount of fracturing sand t/m 2.10 0.92 3.00 Qg 104m3/d 20.54 12.61 33.30 1 3 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 2357 Fig.7 Change curve of optimized Qg with iterations from six wells in cluster 1 Fig.7 Change curve of optimized Qg with iterations from six wells in cluster 1 In addition, the data obtained in this study is limited, with only 7 geological parameters and 4 engineering parameters. If more features can be obtained in the future, such as per- meability, stimulated reservoir volume, drilling and com- pletion parameters, they will also be added to the model to improve its generalization ability. Table 6 Parameter statistics of six wells in cluster 1 Well name Original value of c2 Optimal value of c2 Original value of Qg Optimal value of Qg Increase rate of Qg Well_163 0.31 0.90 0.11 0.72 5.81 Well_174 0.36 0.90 0.15 0.65 3.32 Well_175 0.37 0.93 0.15 0.69 3.57 Well_181 0.36 0.90 0.13 0.67 4.19 Well_194 0.37 0.93 0.16 0.65 3.13 Well_277 0.39 0.95 0.12 0.71 4.76 Table 6 Parameter statistics of six wells in cluster 1 References poor fracturing measures and low gas production; wells with poor geological condition and the productivity is low regardless of the fracturing measures; wells with superior geological condition, good fracturing measures and high production; wells with medium geological con- dition and moderate gas production. Adusumilli S, Bhatt D, Wang H et al (2013) A low-cost INS/GPS integration methodology based on random forest regression[J]. Expert Syst Appl 40(11):4653–4659 Bettin G, Bromhal G, Brudzinski M et al., (2019) Real-time Decision Making for the Subsurface Report. Carnegie Mellon University Wilson E. Scott Institute for Energy Innovation, Pittsburgh, Penn- sylvania.2. Methodology. 3. The model constructed in this paper can classify shale gas wells according to their productivity difference, it can select wells with good fracturing measures and good geological conditions as sample wells. Meanwhile, wells with high productivity potential but low production due to poor fracturing measures can also be selected as wells that can be refractured. Cao C, Zhang M, Li L et al (2020) Tracing the Sources and Evolution Processes of Shale Gas by Coupling Stable (C, H) and Noble Gas Isotopic Compositions: Cases from Weiyuan and Changning in Sichuan Basin, China. J Nat Gas Sci Eng 78:103304 Sichuan Basin, China. J Nat Gas Sci Eng 78:103304 Chu X, Lei J, Liu X, et al. KMEANS Algorithm Clustering for Massive AIS Data Based on the Spark Platform[C]// 2020 5th International Conference on Control, Robotics and Cybernetics (CRC). 2020. 4. Six low production wells with good geological condi- tions but unreasonable design of fracturing operating parameters have been optimized successfully, and the gas production rate of those wells increased by an aver- age of five times after optimization. Besides, this kind of wells should be highly valued as their productivity potential has not been fully exploited. Dj A, Yla B, Jq A et al (2020) SwiftIDS: Real-time intrusion detec- tion system based on LightGBM and parallel intrusion detection mechanism. Comput Security 97:101984 Feng P, Converse T, Ahn D , et al. Feature selection for ranking using boosted trees[C]// Acm Conference on Information & Knowledge Management. ACM, 2009. Di Fina, D., Karaman, S., Bagdanov, A.D., Del Bimbo, A., 2015. MORF: Multi-ObjectiveRandom Forests for Face Character- istic Estimation, 2015 12th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS). IEEE, pp. 1–6. In the future, the number of samples and features will be increased to improve the generalization ability of the model. Conclusions This paper utilized machine learning methods to evaluate the productivity potential of 282 fractured shale gas wells in WY block, and the specific content of this study includes: featuring ranking, feature reconstruction, well classification, productivity potential evaluation and optimization. Several conclusions from this research are presented below: iterations. It can be seen from the figure that each well has achieved the largest Qg value within the range of 70 iterations. The statistics of the parameters before and after optimization of the 6 wells are shown in Table 6. The experimental result shows the gas production rate of these six wells has been greatly improved, indicating that the previous design of fracturing operating parameters such as the amount of hydraulic fracturing fluid and the amount of hydraulic fracturing sand is not reasonable. These wells with greater productivity potential should be taken seri- ously as their gas production has not been fully exploited. iterations. It can be seen from the figure that each well has achieved the largest Qg value within the range of 70 iterations. The statistics of the parameters before and after optimization of the 6 wells are shown in Table 6. The experimental result shows the gas production rate of these six wells has been greatly improved, indicating that the previous design of fracturing operating parameters such as the amount of hydraulic fracturing fluid and the amount of hydraulic fracturing sand is not reasonable. These wells with greater productivity potential should be taken seri- ously as their gas production has not been fully exploited. 1. Thickness of organic-rich gas reserves and amount of hydraulic fracturing sand are vital parameters affecting gas productivity, and geological condition is more deter- ministic to gas productivity. 2. The model established in this paper can accurately divide shale gas wells into four categories according to their geological conditions and productivity potential, mainly including wells with good geological condition, 1 3 1 3 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 2358 References Friedman JH (2001) Greedy Function Approximation: A Gradient Boosting Machine. Ann Statist 29(5):1189–1232 Ftpa B, As A, Ma C, et al. 2021 Predicting the abatement rates of soil organic carbon sequestration management in Western European vineyards using random forest regression[J]. Cleaner Environmen- tal Systems,. Funding This work was supported by the Natural Science Foundation of China (51974327), and the Major national science and technology project of China (2016ZX05043-004). Hauge, V. and Hermansen, G. 2017. Machine Learning Methods for Sweet Spot Detection: A Case Study. In Geostatistics Valencia 2016, ed. J. Gómez ernández, J. Rodrigo-llarri, M. Rodrigo- Clavero, E. Cassiraga, and J. Vargas-Guzmán, Vol. 19. Cham, Switzerland: Quantitative Geology and Geostatistis Series, Springer.i Declarations J Petrol Sci Eng 203:108637 1 3 2359 Journal of Petroleum Exploration and Production Technology (2022) 12:2347–2359 Using Machine Learning Algorithms[J]. Comput Biol Chem 91:107456 Lu C, Jiang H, You C et al (2021) A novel method to determine the thief zones in heavy oil reservoirs based on convolutional neural network. J Petrol Sci Eng 201:108471 Shlens J 2014 A Tutorial on Principal Component Analysis[J]. Inter- national Journal of Remote Sensing 51(2). g Xinhua Ma, Xie J . The progress and prospects of shale gas explora- tion and development in southern Sichuan Basin, SW China[J]. Petroleum Exploration and Development, 2018.l Syed FI, Muther T, Dahaghi AK et al (2021) AI/ML assisted shale gas production performance evaluation. J Petrol Explor Prod Technol 11:3509–3519. https://doi.org/10.1007/s13202-021-01253-8 Ma H, Yu G, She Y et al. 2019 Waterflooding Optimization under Geological Uncertainties by Using Deep Reinforcement Learning Algorithms. Paper presented at the SPE Annual Technical Confer- ence and Exhibition, Calgary, Alberta, Canada, 30–2 September. SPE-196190-MS. https://doi.org/10.2118/196190-MS. Tahmasebi P, Javadpour F, Sahimi M (2017) Data mining and machine learning for identifying sweet spots in shale reservoirs. Expert Syst Appl 88(435):447 Vikara D, Remson D, Khann V (2020) Machine learning-informed ensemble framework for evaluating shale gas production poten- tial: Case study in the Marcellus Shale[J]. J Nat Gas Sci Eng 84:103679 Ma X, Li X, Liang F et al (2020) Dominating factors on well produc- tivity and development strategies optimization in Weiyuan shale gas play, Sichuan Basin, SW China[J]. Petrol Explor Develop 47(3):594–602 Yan B, Mi L, Wang Y, Tang H, An C, Killough JE (2017) Mechanistic Simulation Workflow in Shale Gas Reservoirs. Soc Petrol Eng. https://doi.org/10.2118/182623-MS Martinez AM, Kak AC (2001) PCA versus LDA. IEEE Trans Pat- tern Anal Mach Intell 23(2):228–233. https://doi.org/10.1109/34. 908974 Yao M, Chang H, Li X et al (2019) An Integrated Approach for History Matching of Multiscale-Fractured Reservoirs. SPE J 24(4):1508–1525 Mo S, Zhu Y, Zabaras N et al (2019) Deep convolutional encoder- decoder networks for uncertainty quantification of dynamic mul- tiphase flow in heterogeneous media. Water Rresources Research 55:703–728 Zhao X, Pu X, Jin F et al (2019) Geological characteristics and key exploration technologies of continental shale oil sweet spots: a case study of Member 2 of Kongdian Formation in the Cang- dong sag in Huanghua depression, Bohai Bay Basin. Petrol Res 4:97–112 Mohammadmoradi P, Kantzas A (2018) Wettability and Capillary Imbibition in Shales Analytical and Data-Driven Analysis. Soc Petrol Eng. Declarations Conflict of interest The author(s) declare(s) that there is no conflict of interest regarding the publication of this paper. He Q, Zhong Z, Alabboodi M, Wang G (2019) Artificial Intelligence Assisted Hydraulic Fracturing Design in Shale Gas Reservoir. Soc Petrol Eng. https://doi.org/10.2118/196608-MS Ethical approval On behalf of all the co-authors, the corresponding author states that there are no ethical statements contained in the manu- scripts. Jiang Z, Song Y, Tan X et al (2020) Controlling factors of marine shale gas differential enrichment in southern China. Petrol Explor Develop 47(3):661–673fi Ke GL, Meng Q, Finley et al, 2017 Lightgbm: a highly efficient gra- dient boosting decision tree. Conference on Neural Information Processing Systems. Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, 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 licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Kurison C, Sadi Kuleli H, Mubarak M (2019) Unlocking well produc- tivity drivers in Eagle Ford and Utica unconventional resources through data analytics. J Nat Gas Sci Eng 71:1–24 Liang X , Yla B , Yxc D , et al. A data-driven shale gas production forecasting method based on the multi-objective random forest regression. Journal of Petroleum Science and Engineering 196 Liu W, Zhang G, Cao J et al (2019) Combined petrophysics and 3D seismic attributes to predict shale reservoirs favorable areas. J Geophys Eng 16:974–991 Liu YY, Ma XH, Zhang XW et al (2021) A deep-learning-based predic- tion method of the estimated ultimate recovery (EUR) of shale gas wells. Petrol Sci 17:1450–1464 Liu Y, Ma X, Zhang X et al (2021) Shale Gas Well Flowback Rate Pre- diction for Weiyuan Field Based on a Deep Learning Algorithm. Declarations https://doi.org/10.2118/189806-MS Peter RJ (1999) Silhouettes: a graphical aid to the interpretation and validation of cluster analysis [J]. J Comput Appl Math 20i Zhu G, Kou J, Yao J et al (2020) A Phase-Field Moving Contact Line Model with Soluble Surfactants. J Comput Phys 405:109170. https://doi.org/10.1016/j.jcp.2019.109170 Prasad AM, Iverson LR, Liaw A (2006) Newer classification and regression tree techniques: bagging and random forests for eco- logical prediction. Ecosystems 9(2):181–199 Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Qian K, He Z, Lu X, Chen Y (2018) Intelligent prediction and integral analysis of shale oil and gas sweet spots. Petrol Sci 15:744–755 Qin X, Liu M, Zhang L et al (2021) Structural Protein Fold Recogni- tion Based on Secondary Structure and Evolutionary Information 1 3
https://openalex.org/W2007755558	https://europepmc.org/articles/pmc3823024?pdf=render	English	null	Presence of N‐glycosylated transthyretin in plasma of V30M carriers in familial amyloidotic polyneuropathy: an escape from <scp>ERAD</scp>	Journal of Cellular and Molecular Medicine	2,013	cc-by	6,202	J. Cell. Mol. Med. Vol 17, No 3, 2013 pp. 429-435 J. Cell. Mol. Med. Vol 17, No 3, 2013 pp. 429-435 Introduction Familial amyloidotic polyneuropathy (FAP) is characterized by extra- cellular amyloid deposition, in particular, in the peripheral nervous system. Deposits are mainly composed by mutated transthyretin (TTR) being V30M the most common mutation associated with FAP. More than 100 TTR mutations have been described, the great majority associated with FAP. The vast majority of FAP patients are heterozy- gous for the mutant TTR. Some forms are not neuropathic, but rather cardiomyopathic whereas other variants are not pathogenic (for a tab- ulation see: http://amyloidosismutations.com/attr.html). port of thyroxine (T4) and retinol-binding protein (RBP). Native TTR comprises four identical subunits each containing 127 amino acid resi- dues with a molecular mass of approximately 14 KD. The mechanism of extracellular ﬁbrillogenesis is not fully understood, but several studies point out that amyloidogenic TTR mutants inﬂuence conformational changes that induce tetramers dissociation into partially unfolded mono- mers which self-assemble into amyloid ﬁbrils. (for a review see [1]). Although no evidence exists for the circulation of TTR aggregates in plasma from FAP patients [2], recent biophysical studies on TTR L55P, associated with a very aggressive form of FAP, revealed signiﬁ- cative conformational changes as compared with the wild-type pro- tein or other amyloidogenic clinically less aggressive mutations [3]. Thus, the three-dimensional studies of the L55P variant indicated that the OH group of tyrosine 78 plays an important role in maintaining the tertiary structure of the AB loop. Based on these ﬁndings, a spe- ciﬁc mutation was designed to replace tyrosine for phenylalanine [4]. Biochemical characterization of Y78F showed that this variant adopts a tetrameric conformation as normal TTR and retains the ability to bind T4, indicating a functional tetrameric structure. Under acidic pH, TTR is a secreted non-glycosylated tetrameric protein of 55 KD that is mainly synthesized in liver whose major function is the plasma trans- Correspondence to: Maria J. SARAIVA, Molecular Neurobiology, Instituto de Biologia Molecular e Celular R. Campo Alegre, 823. 4150-180 Porto, Portugal. Tel.: +351 226074900 Fax: +351 226099157 E-mail: mjsaraiv@ibmc.up.pt Correspondence to: Maria J. SARAIVA, Molecular Neurobiology, Instituto de Biologia Molecular e Celular, R. Campo Alegre, 823. 4150-180 Porto, Portugal. Tel.: +351 226074900 Fax: +351 226099157 E-mail: mjsaraiv@ibmc.up.pt Correspondence to: Maria J. SARAIVA, Molecular Neurobiology, Instituto de Biologia Molecular e Celular, R. Campo Alegre, 823. 4150-180 Porto, Portugal. Tel.: +351 226074900 Fax: +351 226099157 E-mail: mjsaraiv@ibmc.up.pt doi: 10.1111/jcmm.12024 doi: 10.1111/jcmm.12024 ª 2013 The Authors. ª 2013 The Authors. Published by Foundation for Cellular and Molecular Medicine/Blackwell Publishing Ltd. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. doi: 10.1111/jcmm Abstract Familial amyloid polyneuropathy (FAP) is an autosomal dominant disease characterized by deposition of amyloid related to the presence of mutations in the transthyretin (TTR) gene. TTR is mainly synthesized in liver, choroid plexuses of brain and pancreas and secreted to plasma and cerebrospinal ﬂuid (CSF). Although it possesses a sequon for N-glycosylation N-D-S at position 98, it is not secreted as a glycoprotein. The most common FAP-associated mutation is TTR V30M. In a screening for monoclonal antibodies developed against an amyloidogenic TTR form, we detected a distinct TTR with slower electrophoretic mobility in Western of plasma from carriers of the V30M mutation, not present in normal plasma. Mass spectrometry analyses of this slower migrating TTR (SMT) identiﬁed both wild-type and mutant V30M; SMT was undetectable upon N-glycosidase F treatment. Furthermore, SMT readily disappeared in the plasma of V30M - FAP patients after liver transplantation and appeared in plasma of transplanted domino individuals that received a V30M liver. SMT was also detected in plasma, but not in CSF of trans- genic mice for the human V30M mutation. A hepatoma cell line transduced to express human V30M did not present the SMT modiﬁcation in secretion media. Glycosylated TTR was absent in ﬁbrils extracted from human kidney V30M autopsy tissue or in TTR aggregates extracted from the intestine of human TTR transgenic mice. Studies on the metabolism of this novel, glycosylated TTR secreted from FAP liver are warranted to provide new mechanisms in protein quality control and etiopathogenesis of the disease. Keywords: Transthyretin N- glycosylation ERAD Keywords: Transthyretin N- glycosylation ERAD a Molecular Neurobiology, Instituto de Biologia Molecular e Celular, IBMC, Porto, Portugal b Instituto de Ci^encias Biomedicas de Abel Salazar, ICBAS, University of Porto, Portugal a Molecular Neurobiology, Instituto de Biologia Molecular e Celular, IBMC, Porto, Portugal b Instituto de Ci^encias Biomedicas de Abel Salazar, ICBAS, University of Porto, Portugal Received: August 28, 2012; Accepted: December 20, 2012 Presence of N-glycosylated transthyretin in plasma of V30M carriers in familial amyloidotic polyneuropathy: an escape from ERAD Anabela C. Teixeira a, Maria J. Saraiva a, b, a Molecular Neurobiology, Instituto de Biologia Molecular e Celular, IBMC, Porto, Portugal b Instituto de Ci^encias Biomedicas de Abel Salazar, ICBAS, University of Porto, Portugal Introduction Published by Foundation for Cellular and Molecular Medicine/Blackwell Publishing Ltd. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. doi: 10.1111/jcm the tyr—phe substitution at position 78 is more prone to form ﬁbrils as compared with non-mutated TTR. It was hypothesized that Y78F exhibits the characteristics of an intermediate structure in the ﬁbrillogenesis pathway and might represent an early event in TTR am- yloidogenesis. Interestingly, this mutation designed in silico was found associated with peripheral neuropathy, carpal tunnel syndrome and skin amyloidosis [5]. collected upon sacriﬁce (6 months) were analysed. Mice used in the immunization protocol were 16-months old TTR null mice [9]. Mice were housed in pathogen-free conditions in a controlled-temperature room and were maintained under a 12-hrs light/dark period. Water and food were freely available. All experiments were performed in accor- dance with the European Communities Council Directive (2010/63). Monoclonal antibodies (mabs) produced in mice against highly aggressive amyloidogenic synthetic TTR mutants were shown to react with high molecular weight TTR aggregates, but do not recognize solu- ble native TTR when tested under ELISA (enzyme-linked immunoassay). It was hypothesized that these mabs recognize cryptic epitopes that are exposed in mutant TTRs resembling aggregated TTR [6]. Interestingly, these mabs, under speciﬁc conditions, reacted with TTR from plasma of FAP patients and/or asymptomatic carriers of neuropathic TTR mutants, but not with plasma from normal individuals, thus detecting subtle structural changes that occur in amyloidogenic TTR tetramers [7]. Mab puriﬁcation As described throughout the results section, plasma samples from carriers of different published TTR mutations were available in the labo- ratory and had been collected after informed consent. Isolation of the mab from cell culture supernatants was carried out on a Protein G Sepharose High Performance column (MabTrap G II protein G – GE Healthcare). The mab was eluted from the afﬁnity column with Immuno Pure Gentle Elution Buffer (Pierce). Kidney autopsy tissue from a FAP patient was available at Hospital Geral de Santo Antonio, Porto, Portugal. Human samples In a direct approach, microtitre plates (96 wells, Nunc) were coated with 1 lg of isolated plasma TTR or 100 ll of serum diluted 1:10 in coating buffer. Following blocking, plates were incubated with undiluted hybridoma culture supernatants and then with sheep antimouse immu- noglobulins G-HRP conjugated (Pierce) (1:5000 dilution in PBST). Plates were developed using 5 mM 2.2′-azinobis (3-ethybenzthiazoline-6-sul- fonic acid) (ABTS) (Sigma), and measured at 405 nm. Plasma samples from asymptomatic (n = 180), symptomatic heterozyg- otic (n = 12) and homozygotic (n = 3) V30M carriers, as demonstrated by DNA analysis and their non-carrier individuals (n = 214) were obtained either from the Center for Predictive siblings and Preventive Genetics at IBMC, or different hospitals, all with informed consent, fol- lowing the Declaration of Helsinki. Plasma samples from individuals who had undergone domino liver transplantation (DLT), hosting the liver of V30M FAP patients (n = 2) were subsequently collected at different times up to 1 year, ranging from 1 to 2, 10,20,30,45, 60, 120 and 180 days. In a ‘sandwich approach’, plates were ﬁrst incubated at 4°C with rabbit anti-TTR polyclonal antibody (DAKO) in a 1:500 dilution in coating buffer. About 5 to 100 lg or 100 ll of diluted serum (1:10) were applied and incubated with undiluted hybridoma culture supernatant, for 1 hr at RT. After washing, bound antibodies were detected as described above. Plasma from transplanted V30M patients (n = 6) was collected 24 hrs and on the followings days, as well as several months after transplantation; both procedures were approved by the Ethics Commit- tee of the Transplantation Department, University Hospitals of Coimbra, Portugal and included informed consent. Mice immunization and fusion protocol Mice (n = 3) were immunized with recombinant mutant TTR Y78F; the best responder animal exhibiting positive results for TTR recognition by direct ELISA was killed, blood collected and spleen removed to perform fusions, following standard procedures and multiple clones were then screened by sandwich ELISA for positive IgG to human plasma from carriers of the V30M mutation, but not to normal human plasma and then re-cloned by serial dilution. To identify the possible existence of altered TTR conformations/ modiﬁcations in tissues and plasma of FAP individuals, we produced several monoclonal antibodies against the Y78F mutant. In the pres- ent report, we characterize a particular mab designated by AD7 that detects a glycosylated form of TTR in the plasma of V30M carriers not present in normal plasma. Mab isotyping Microtitre plates (96 wells, Nunc) were coated with rabbit antimouse IgG – Fc speciﬁc. After blocking, plates were incubated with 100 ll/well of undiluted serum-free hybridoma culture supernatant, and then isotyped for IgG1, IgG2a, IgG2b and IgG3 classes. (AbD, Serotec). Bound antibodies were detected with streptavidin-biotinylated horseradish peroxidase (GE Health- care). ª 2013 The Authors. Published by Foundation for Cellular and Molecular Medicine/Blackwell Publishing Ltd. Screening Following fusion, clones were screened for antibody production by testing the culture supernatants in a sandwich ELISA assay that selec- tively detects antibodies recognizing native antigens. Selected clones recognized: (i) recombinant Y78F TTR; (ii) puriﬁed V30M TTR from the plasma of heterozygotic carriers, but not TTR puriﬁed from nor- mal plasma; (iii) TTR from the plasma of heterozygotic V30M carriers, but not from normal individuals. One stable hybridoma, named AD7F6, thereafter referred to as AD7, of the IgG2b isotype was positive for the three above described criteria. Immunoprecipitation of TTR Five micrograms of puriﬁed mab pre-adsorbed to 50 ll of Protein G Sepharose was mixed with 25 ll of human plasma from healthy individ- uals or V30M carriers, overnight at 4°C, centrifuged and the pellet anal- ysed by 15% acrylamide SDS-PAGE. TTR bands were visualized after immunoblotting using the mab. ELISA The discriminative properties of AD7 were further studied by sand- wich ELISA in different carriers and conditions. Discrimination of V30M versus normal plasma was conﬁrmed in heterozygotic V30M patients (n = 12), in homozygous V30M patients. (n = 3). The inten- sity of the reaction did not correlate between hetero and homozygotic patients, suggesting that under the conditions tested, AD7 is of value only in qualitative tests. We then tested plasma from asymptomatic V30M carriers (n = 200) and observed positive immunoreactivity as compared with non-carriers (n = 180). Samples were considered positive after subtraction of OD values from three negative controls tested on the same ELISA plate. Fibril extraction All data were expressed as mean SE. D’Agostino and Pearson tests evaluated normal distributions. The values of treated cells were expressed as percentage relative to controls. The differences were analysed by one-way analysis of variance (ANOVA) with Tukey’s post hoc tests. GraphPad Prism version 5.04 for Windows (GraphPad Software, San Diego California USA, www.graphpad.com) was employed. The Kaplan method [11] was applied to extract human TTR amyloid ﬁbril from FAP kidney and/or TTR aggregates from the intestine of transgenic mice for human V30 M. Tunicamycin treatment of puriﬁed plasma TTR to use in clone screening, and immunization, TTR was isolated by ion exchange chromatography and preparative electrophoresis [10]. Y78F TTR was produced in an Escherichia coli expression system, isolated, and puriﬁed as described [4]. Approximately 3 9 105 cells of a hepatoma cell line – SaHep [12] were stably transduced with a CSCW2 lentiviral vector to express the human wild-type and the V30M TTR variant. Cells cultured in MEM, supple- mented with 10% foetal bovine serum, 2 mM glutamine (all materials from Gibco, Invitrogen) and 1 mM of non-essential amino acids (Sigma) were maintained at 37°C in a 5% CO2 humidiﬁed atmosphere. Cells were washed and medium without FBS was added. One ﬂask was used as control and 1 lg/ml of tunicamycin (Sigma) added to the other ﬂask. After 16 hrs incubation, supernatants were recovered for further analysis. About 100 lg of protein from conditioned media without tu- nicamycin treatment were analysed by immunoblotting with the mab as described above. Concentration of secreted TTR was assessed by ELISA by an in house-assay. Results Proteins were analysed either in native or denaturing conditions. Native electrophoresis was carried out on 10% (w/v) acrylamide gels. Electro- phoresis under denaturing conditions was performed in SDS-PAGE gels (15% acrylamide, 0.1% (w/v) SDS), after heat treatment of samples and addition of 0.1 M b-mercaptoethanol. Proteins were transferred from gels into nitrocellulose membranes (Hybond TM-C pure, GE Health- care), using a Tris-Glycine system, for 1 hr. After blocking and washing with PBST, immunodetection was performed with the mab in pure hybridoma supernatant, for 1 hr at RT, and sheep antimouse immuno- globulins-HRP conjugated (Pierce, 1:5000 dilution). TTR was visualized using either the enhanced chemiluminescence method (ECL, GE Health- care) or 3,3′-Diaminobenzidine (DAB) substrate. AD7 discrimination of V30M plasma MALDI mass spectroscopic analysis of TTR was performed after band excision from a silver stained gel and digestion with trypsin followed by identiﬁcation of the originated peptides from the SWISSPROT data- base; N-terminal sequencing was carried out after transfer of the pro- teins of interest from a SDS gel to a PVDF membrane and Coomassie staining. ª 2013 The Authors. Published by Foundation for Cellular and Molecular Medicine/Blackwell Publishing Ltd. TTR puriﬁcation Small quantities of plasma TTR were puriﬁed by immunoafﬁnity chro- matography by cross-linking (with dimethyl pimelimidate) 8 mg of puri- ﬁed antibody to 1 ml Protein G Sepharose. To obtain larger quantities Thirty plasma samples, nine cerebrospinal ﬂuid samples - CSF (age between 6 and 18 months) and intestine (6-months old) from trans- genic mice for the human V30M mutation in a TTR null background [8] 430 J. Cell. Mol. Med. Vol 17, No 3, 2013 Deglycosylation Approximately 2 ll of whole plasma were treated with 5 units of N-gly- cosidase F by a denaturing standard protocol according to the supplier (Calbiochem) followed by immunoblotting analysis with the mab. Five microlitres of plasma of homozygotic V30M plasma were treated with 25 units of N-glycosidase F by the same non-denaturing standard proto- col and analysed by a direct ELISA procedure using the mab, as above described. We then analysed plasma from V30M patients who had under- gone liver transplantation and found that the ELISA values resembled 431 those found for the normal individuals. Moreover, the ELISA levels in the plasma of normal recipients of livers from V30M carriers after domino transplantation were similar to those found for carriers of the V30M mutation, even after 1 day post transplant. from transgenic mice for the human V30M (n = 25) (immunoblots 2 of Figs. 1A and B). A doublet SMT band was observed under denatur- ing conditions; however, it was not observed in CSF samples of these animals (n = 4 pools of 3 animals) (not shown). Furthermore, the SMT band was observed in the blots of plasma from recipients of domino FAP livers collected at different time-points, starting at day 1, as shown in blot 3 of Figure 1A. On the other hand, SMT was not visi- ble on the blots of sera of transplanted V30M FAP patients collected right after 24 hrs after transplant and on the following days as well as several months (not shown). Further testing by ELISA of plasma from carriers of non-amyloido- genic TTR mutations, such as G6S (n = 5) and T119M (n = 2), or from patients with senile systemic amyloidosis (SAA; n = 4) showed no reactivity. Finally, we conducted a limited study on the reactivity of AD7 for plasma of carriers of amyloidogenic mutations including G49A, S50R and F64L (n = 3, each) and although the two ﬁrst mutants were positive, carriers of F64L behaved as control individu- als. Plasma from carriers of other amyloidogenic TTR mutations, namely G47A and T59K were also recognized by the mab, but these results must be conﬁrmed because only one sample of each of these mutants was available for analysis. SMT was also observed in double V30M/T119M compound het- erozygotes; it was absent in heterozygotic carriers of non-amyloido- genic TTR mutations included in the study, i.e. G6S and T119M. Immunoblotting and immunoprecipitation Pl f i d i V Plasma from asymptomatic and symptomatic V30M carriers (n = 25 and 12 respectively) were separated by both native and SDS-PAGE and blotted into a nitrocellulose ﬁlter which was probed with the mab. As documented in Figure 1, besides the regular TTR band, a slow migrating TTR band (SMT, labelled with an asterisk) was readily detected in the plasma of V30 M carriers under both conditions; this band was absent in non-carrier controls (n = 25) (immunoblot 1 in Fig. 1A for native PAGE and immunoblot 1 in Fig. 1B for denaturing PAGE conditions respectively). This SMT band was also detected in immunoblot analyses (from native and denaturing gels) of plasma Deglycosylation Preliminary analyses of limited samples of carriers of non-V30M mutations (as above described in 3.2.) showed the SMT band in K70N and V30A carriers. The intensity of SMT correlated with the amount of TTR monomer with the regular migration, representing approximately 0.2% and was similar between hetero-and homozygot- ic V30M carriers either asymptomatic or symptomatic. We then searched for the SMT band in blots of TTR V30M aggregate/ﬁbrils extracts of human and mice tissue origin. The SMT band was not detected in these samples (not shown). Using a direct ELISA protocol, the mab maintained the dis- criminative properties between carriers of the V30M, G49A and S50R mutations over controls or carriers of non-amyloidogenic mutations. Carriers of the F33V, V30A and K70N mutants (n = 1 each) that showed negative reactivity by AD7 by the sandwich ELISA protocol reacted in a positive fashion by the direct proto- col, as compared with normal plasma, probably due to exposition of cryptic epitopes in the direct ELISA approach (binding to the plastic). Finally, plasma from V30M carriers and normal controls was im- munoprecipitated with the mab and the immunoprecipitates analysed by immunoblots using the mab. As seen in panel 3 of Figure 1B, SMT was immunoprecipitable only from carriers of V30M and absent from controls, corroborating the immunoblot analyses performed on direct plasma. Similarly, after immunoprecipitation of plasma from trans- genic mice for human V30M, two SMT bands were visualized. Identiﬁcation of SMT Plasma of V30M carriers was subjected to afﬁnity chromatography using the immobilized mab, the proteins in the eluate separated by SDS-PAGE and both the silver-stained SMT and normal TTR A B Fig. 1 (A) Immunoblotting analysis of plasma TTR after native PAGE electrophoresis. The SMT - TTR band () was detected in: (1) plasma from TTR V30M carriers (a–d) and absent in negative controls (e–g); (2) plasma from transgenic mice for V30M TTR (a–g); (3) plasma from domino recipients, collected at different time-points after transplantation (day 1 – lane a; days 10,20,30 and 45 – lanes b,c,d and e respectively), and 4 months (lane f). (B) Immunoblotting analysis of plasma TTR after SDS-PAGE electrophoresis. The SMT - TTR band () was detected in: (1) human plasma from TTR V30M carriers (b–g) and is absent in the negative control of lane a; (2) transgenic mice for V30M TTR (a–d); (3) immuno- precipitates of human V30M plasma (a,b and d) absent in control plasmas (c and e). Fig. 1 (A) Immunoblotting analysis of plasma TTR after native PAGE electrophoresis. The SMT - TTR band () was detected in: (1) plasma from TTR V30M carriers (a–d) and absent in negative controls (e–g); (2) plasma from transgenic mice for V30M TTR (a–g); (3) plasma from domino recipients, collected at different time-points after transplantation (day 1 – lane a; days 10,20,30 and 45 – lanes b,c,d and e respectively), and 4 months (lane f). (B) Immunoblotting analysis of plasma TTR after SDS-PAGE electrophoresis. The SMT - TTR band () was detected in: (1) human plasma from TTR V30M carriers (b–g) and is absent in the negative control of lane a; (2) transgenic mice for V30M TTR (a–d); (3) immuno- precipitates of human V30M plasma (a,b and d) absent in control plasmas (c and e). Fig. 1 (A) Immunoblotting analysis of plasma TTR after native PAGE electrophoresis. The SMT - TTR band () was detected in: (1) plasma from TTR V30M carriers (a–d) and absent in negative controls (e–g); (2) plasma from transgenic mice for V30M TTR (a–g); (3) plasma from domino recipients, collected at different time-points after transplantation (day 1 – lane a; days 10,20,30 and 45 – lanes b,c,d and e respectively), and 4 months (lane f). (B) Immunoblotting analysis of plasma TTR after SDS-PAGE electrophoresis. ª 2013 The Authors. Published by Foundation for Cellular and Molecular Medicine/Blackwell Publishing Ltd. Discussion We reasoned that the SMT recognized by the mab might represent glycosylation, considering the approximately 2000 kD increment rela- tive to the regular monomer and the heterogeneity found in plasma of V30M transgenic mice; therefore, plasma from V30M carriers and double heterozygotic V30M/T119M or from V30M transgenic mice were treated with N-glycosidase F and analysed by immunoblot using the mab, before and after treatment. As seen in Figure 2, after N-gly- cosidase F treatment, SMT was no longer detected by the mab after enzymatic removal of N-glycans. We conclude that SMT represents a N-glycosylated form of TTR, secreted from livers producing the V30M mutation both in humans and transgenic mice. SMT doublet band from transgenic mice is thus possibly related to sugar heterogeneity. The presented mab has unique properties as it is immunoreactive under native conditions of ELISA of human plasma from V30M carri- ers or recipient individuals that received a V30M liver. It is not reac- tive with plasma from normal individuals, transplanted V30M FAP patients that received a normal liver or from carriers of two tested non-amyloidogenic mutations. Other amyloidogenic mutations were also immunoreactive with the mab, but the limitation of the number of samples under test does not allow us to conclude on whether the mab properties are speciﬁc for the V30M mutant where we have solid evidence. The most signiﬁcant ﬁnding was the identiﬁcation by the mab of the SMT band in the plasma of the very same carriers of V30M tested by ELISA whether derived from genetic carriers or by recipients of TTR V30M liver and the disappearance of the SMT band after trans- plantation of a normal liver. SMT was also in plasma from transgenic V30M mice, but not in CSF. The overall data suggested the circulation of TTR conformers in the plasma of V30M carriers recognized by the mab. It should be noted that TTR is resistant to complete unfolding in SDS-PAGE, migrating as a 16,000 Kd band, above the predicted molecular mass of 14,300. The non-amyloidogenic T119M mutant did not suppress SMT in V30M/T119M allelic double compound het- erozygotes. T119M TTR has been described as a non-amyloidogenic TTR variant. In compound heterozygotic individuals, the evolution of the disease seems to be more benign than in typical V30M TTR patients, suggesting a protective effect of T119M TTR on the patho- genic effects of V30M TTR. Identiﬁcation of SMT The SMT - TTR band () was detected in: (1) human plasma from TTR V30M carriers (b–g) and is absent in the negative control of lane a; (2) transgenic mice for V30M TTR (a–d); (3) immuno- precipitates of human V30M plasma (a,b and d) absent in control plasmas (c and e). ª 2013 The Authors. Published by Foundation for Cellular and Molecular Medicine/Blackwell Publishing Ltd. 432 J. Cell. Mol. Med. Vol 17, No 3, 2013 Fig. 2 Immunoblotting after deglycosylation with N-glycosidase F. Human plasma from V30M TTR carriers before (a) and after (b) deglycosylation and from transgenic mouse before (c) and after N-glycosidase treatment (d). Human plasma from a carrier of a double mutant V30M/T119M before (e and g) and after deglycosylation (f and h). In all samples, SMT () is no longer detected after enzymatic removal of N-glycans. Fig. 2 Immunoblotting after deglycosylation with N-glycosidase F. Human plasma from V30M TTR carriers before (a) and after (b) deglycosylation and from transgenic mouse before (c) and after N-glycosidase treatment (d). Human plasma from a carrier of a double mutant V30M/T119M before (e and g) and after deglycosylation (f and h). In all samples, SMT () is no longer detected after enzymatic removal of N-glycans. monomer bands used for characterization. Mass spectrometry of SMT after enzymatic digestion revealed peptides with the expected masses for digested TTR (except for the ﬁrst 16 residues not detected on the analyses); both normal peptide entailing residues 16–35 and the mutant peptide with a mass shift corresponding to a substitution of methionine for valine were detected as well as their oxidation prod- ucts as previously reported for TTR [13]. The results obtained for the regular TTR monomer were similar, i.e. both the regular monomer and the SMT bands contained mutated and non-mutated TTR forms. From the N-terminal analyses of the SMT band, we concluded that the N-terminal was intact. of a high concentration of the regular monomer (not shown). The same conclusions were reached in parallel experiments with a HEK cell line. Under the experimental conditions, the concentration of secreted WT or V30M TTR in the media averaged 5 lg/ml/105 cells; when incubated with tunicamycin, the relative percentage levels of secreted TTR in treated versus untreated conditions was not signiﬁcantly dif- ferent between WT and V30 M producing cells (Fig. Discussion Physical–chemical studies on T119M plasma showed a higher resistance to dissociation of TTR tetramers We ﬁnally investigated whether the discriminative properties of the mab for V30M carriers versus non-carriers on the ELISA proce- dure could relate to V30M glycosylation. With this purpose, we trea- ted V30M homozygotic plasma with N-glycosidase F and compared the mab reactivity by sandwich ELISA in untreated versus untreated samples. No change in immunoreactivity was found. ª 2013 The Authors. Published by Foundation for Cellular and Molecular Medicine/Blackwell Publishing Ltd. Identiﬁcation of SMT 3), which corrob- orates the discrepancy in the observation of N-glycosylation found in vitro and in vivo observations. All together the results were suggestive of a post-translational modiﬁcation occurring in TTR in V30M liver, but not in normal liver, which is then secreted in plasma. V30M TTR secretion by hepatomas They proposed that nascent poly- peptides meet multiple OST complexes as they emerge in the ER lumen to provide proteins proper glycosylation and folding before they leave the secretory pathway. Recently, seminal extensive studies by Sato et al. [19] hypothesized STT3B-dependent post-translational N-glycosylation as part of a triage-salvage system recognizing cryptic N-glycosylation sites of secretory proteins to preserve protein homeostasis. This STT3B-dependent alternative pathway for degrada- tion is EDEM3-mediated N-glycan-dependent ERAD, distinct from the major pathway of Herp-mediated N-glycan-independent ERAD. These authors used TTR mutants in their studies to conclude that prolonged TTR unfolding in the ER occurs with highly amyloidogenic TTR mutants, as the D18G mutation and induces cryptic N-glycosylation of asparagine side chain at position 98 of TTR, a process dependent of STT3B. This conclusion emerged from genetic studies with D18G TTR and other TTR mutations (V30M included) transfected in HEK cells; in their studies V30M behaved as WT-TTR without evidence for N-glycosylation. In the present in vitro studies in human hepatomas secreting WT and V30M, we did not detect glycosylated TTR in the media, neither did tunicamycin affect secretion efﬁciency. Do the present in vivo ﬁndings underlie escape from ERAD systems and improved secretion of the TTR variant? Why liver secretes glycosylat- ed V30M into plasma, but choroid plexus does not, needs investiga- tion to dissect further unknown mechanisms in the surveillance system for protein secretion that most likely are cell speciﬁc. into monomers, in contrast to a lower resistance in V30M plasma and a similar resistance to tetrameric dissociation between T119M/V30M compound heterozygotes and normal plasma [14]. Of even more interest was the identiﬁcation in plasma of a N- glycosylated TTR fraction in carriers of the V30M mutations, unde- tectable in plasma of normal individuals. Because TTR is not a gly- cosylated protein, despite possessing a sequon for N-glycosylation at position 98 (Asn-Asp-Ser) the question arises what factor/s inﬂu- ence N-glycosylation during liver synthesis and intracellular trafﬁc in the secretory pathway, and the disposal of N-glycosylated V30M once in circulation. Because both events contribute to the pool of N-glycosylated TTR in plasma, it is difﬁcult to mechanistically explain both the total levels of N-glycosylated TTR, and selectivity of glycosylation. V30M TTR secretion by hepatomas We next searched by immunoblotting for SMT in the secretion media of a human hepatoma cell line that does not produce/secrete endoge- nous TTR [12] after infection with lentiviral vector stocks that lead to the production and secretion of WT and V30M TTR. SMT was not detected in either case, in contrast with the in vivo situation, despite 433 Fig. 3 Effect of tunicamycin (TN) on secretion of TTR from WT and V30M lentiviral infected SaHep cells. Data were expressed as relative percentage to control (untreated) of tunicamycin treated cells, from ﬁve independent experiments. plexus relate to protein chaperones and metabolite chaperone con- tent of the ER in different tissues; eventually some other factors, are likely to strongly inﬂuence tissue speciﬁcity, severity, and age of onset of TTR amyloidoses [15]; most TTR mutants, such as the V30M are secreted with the similar efﬁciencies as WT and only most highly destabilized TTR mutants, as the case of D18G, associated with leptomeningeal amyloidosis are subjected to ER-associated degradation (ERAD); however, the precise mechanisms used in qual- ity control to deal with the different amyloidogenic TTR mutants were not explored in this work, namely post-translational events, in particular N-glycosylation. N-glycosylation, a highly conserved process that is characterized by the addition of a sugar complex to the nascent polypeptide chain as soon it emerges into the ER, affects protein folding, oligomeriza- tion and stability [16]. Oligosaccharyl transferase (OST) catalyses the transfer of the oligosaccharide to the asparagine side chain of the acceptor polypeptide into a consensus sequence Asn-Xaa-Ser/Thr. Fig. 3 Effect of tunicamycin (TN) on secretion of TTR from WT and V30M lentiviral infected SaHep cells. Data were expressed as relative percentage to control (untreated) of tunicamycin treated cells, from ﬁve independent experiments. p p yp p q N-glycosylation is normally a cotranslational process, but in 2005 Bolt et al. [17] demonstrated for the ﬁrst time post-translational N- glycosylation in human coagulation factor VII, a non-modiﬁed protein in mammalian cells raising the question about how many glycopro- teins that undergo post-translational N-glycosylation can be a part of natural processing. In 2009, Ruiz-Canada et al.[18] reported that oligosaccharyl transferase isoform STT3B efﬁciently mediates post-translational N-glycosylation when a sequon is found close the C-terminus of an unfolded protein. ª 2013 The Authors. Published by Foundation for Cellular and Molecular Medicine/Blackwell Publishing Ltd. References transthyretin (TTR) mutations in familial am- yloidotic polyneuropathy (FAP). J Neurol Sci. 2000; 173: 154–9. 7. Palha JA, Moreira P, Olofsson A, et al. Antibody recognition of amyloidogenic transthyretin variants in serum of patients with familial amyloidotic polyneuropathy. J Mol Med. 2001; 78: 703–7. 1. Saraiva MJ, Magalh~aes J, Ferreira N, et al. Transthyretin deposition in familial amyloi- dotic polyneuropathy. Curr Med Chem. 2012; 19: 2304–11. 14. Alves IL, Hays MT, Saraiva MJM. Compara- tive stability and clearance of Met30 trans- thyretin and Met119 transthyretin. Eur J Biochem. 1997; 249: 662–8. 2. Saraiva MJM, Costa PP, Goodman DS. Studies on plasma transthyretin (prealbu- min) in familial amyloidotic polyneuropathy, Portuguese type. J Lab Clin Med. 1983; 102: 590–603. 8. Kohno K, Palha JA, Miyakawa K, et al. Analysis of amyloid deposition in a trans- genic mouse model of homozygous familial amyloidotic polyneuropathy. Am J Pathol. 1997; 150: 1497–508. 15. Sekijima Y, Wiseman RL, Matteson J, et al. The biological and chemical basis for tissue-selective amyloid disease. Cell. 2005; 121: 73–85. 3. Sebasti~ao MP, Saraiva MJ, Damas AM. The crystal structure of amyloidogenic Leu55 –> Pro transthyretin variant reveals a possible pathway for transthyretin polymerization into amyloid ﬁbrils. J Biol Chem. 1998; 273: 24715–22. 9. Episkopou V, Maeda S, Nishiguchi S, et al. Disruption of the transthyretin gene results in mice with depressed levels of plasma reti- nol and thyroid hormone. Proc Natl Acad Sci USA. 1993; 90: 2375–9. 16. Hammond C, Braakman I, Helenius A. Role of N-linked oligosaccharide recognition, glu- cose trimming, and calnexin in glycoprotein folding and quality control. Proc Natl Acad Sci USA. 1994; 91: 913–7. 4. Redondo C, Damas AM, Olofsson A, et al. Search for intermediate structures in trans- thyretin ﬁbrillogenesis: soluble tetrameric Tyr78Phe TTR expresses a speciﬁc epitope present only in amyloid ﬁbrils. J Mol Biol. 2000; 304: 461–70. 10. Almeida MR, Damas AM, Lans MC, et al. Thyroxine binding to transthyretin Met 119 - comparative studies of different heterozygot- ic carriers and structural analysis. Endo- crine. 1997; 6: 309–15. 17. Bolt G, Kristensen C, Steenstrup TD. Post- translational N-glycosylation takes place during the normal processing of human coagulation factor VII. Glycobiology. 2005; 15: 541–7. 11. Kaplan B, German G, Ravid M, et al. Deter- mination of amyloid type by ELISA using milligram amounts of tissue. Clin Chim Acta. 1994; 229: 171–9. 5. Magy N, Liepnieks JJ, Gil H, et al. V30M TTR secretion by hepatomas The data from the electrophoretic and mass spec- trometry analyses raise a number of questions: (i) homozygotic and heterozygotic carriers have approximately the same proportion of the SMT band relative to the total monomeric unglycosylated band both under native or SDS-PAGE conditions; is this related to steric hindrance and/or different conformation between hetero- and homotetrameric V30M with consequences in glycosylation, or to a different disposal once in circulation, which could relate to differen- tial uptake by cells or unstability ? (ii) the SMT band revealed N- glycosylation of both wt and V30M monomers; does this reﬂect N- glycosylation of the normal monomer in V30M heterotetramers ? The tunicamycin experiments in hepatomas corroborate the notion that wt homotetramers are retained in the Golgi with this treatment to the same extent as V30M homotetramers; it is difﬁcult to predict the behaviour of wt/V30M or T119M/V30M heterotetramers as the V30M mutant monomer might impact on the conformation of the normal (or T119M) monomer in tetrameric hybrid species. More than 100 different TTR mutants have been reported, although onset of the disease, tissue selectivity and severity are dif- ferent among them. Earlier in vitro studies using HEK cells on the secretion efﬁciency of different mutants, including V30M, support the hypothesis that secretion differences by liver and the choroid At any rate, the absence of glycosylated TTR species in extracts of material from tissues with TTR deposition not specialized in TTR syn- thesis might relate to degradation events associated with instability of circulating N-glycosylated species or differential uptake by cells. Future studies will determine the impact of plasma N-glycosylated 434 J. Cell. Mol. Med. Vol 17, No 3, 2013 022718 (PEst- C/SAU/LA0002/2011) through FEDER and the COMPETE pro- gramme. 022718 (PEst- C/SAU/LA0002/2011) through FEDER and the COMPETE pro- gramme. TTR in the etiopathogenesis of FAP and should identify the sugar moi- ety, and the glycosylated peptide. TTR in the etiopathogenesis of FAP and should identify the sugar moi- ety, and the glycosylated peptide. Conﬂict of Interest Statement This study was supported by Fundac~ao para a Ci^encia e Tecnologia (FCT) through grant PTDC/SAU-ORG/11313/2009 and FCOMP-01-0124-FEDER- ª 2013 The Authors. Published by Foundation for Cellular and Molecular Medicine/Blackwell Publishing Ltd. References A trans- thyretin mutation (Tyr78Phe) associated with peripheral neuropathy, carpal tunnel syndrome and skin amyloidosis. Amyloid. 2003; 10: 29–33. 18. Ruiz-Canada C, Kelleher DJ, Gilmore R. Cotranslational and posttranslational N-gly- cosylation of polypeptides by distinct mam- malian OST isoforms. Cell. 2009; 136: 272– 83. 12. Sousa MM, Saraiva MJ. Internalization of transthyretin: evidence for a novel yet unidentiﬁed receptor-associated protein (RAP) - sensitive receptor. J Biol Chem. 2001; 276: 14420–5. 19. Sato T, Sako Y, Sho M, et al. STT3B- dependent posttranslational N-glycosyla- tion as a surveillance system for secretory protein. Mol Cell. 2012; 47: 99–110. 6. Goldsteins G, Persson H, Andersson K, et al. Exposure of cryptic epitopes on trans- thyretin only in amyloid and in amyloido- genic mutants, Proc. Natl. Acad. Sci. USA. 1999; 96: 3108–13. 13. Yamashita T, Ando Y, Suhr O, et al. A new diagnostic procedure to detect unknown ª 2013 The Authors. Published by Foundation for Cellular and Molecular Medicine/Blackwell Publishing Ltd. ª 2013 The Authors. Published by Foundation for Cellular and Molecular Medicine/Blackwell Publishing Ltd. 435
https://openalex.org/W4385963564	https://zenodo.org/records/8216631/files/THE%20COMPLIANCE%20WITH%20COVIDq-19%20HEALTH%20PROTOCOLS%20AT%20ST.%20PAUL%20UNIVERSITY%20SURIGAO.pdf	English	null	The Compliance with Covid-19 Health Protocols at St. Paul University Surigao	Zenodo (CERN European Organization for Nuclear Research)	2,023	cc-by	7,824	International Journal of Research (IJR) e-ISSN: 2348-684 p-ISSN: 2348-795X Vol. 10 Issue 0 August 202 The Compliance with Covid-19 Health Protocols at St. Paul University Surigao e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 INTRODUCTION A coronavirus outbreak (COVID-19) in 2019 put everyone's health in danger. The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) that causes this Coronavirus disease 2019 (COVID-19), which is highly contagious and has killed more than 6 million people worldwide, has had a devastating impact on the world's demographics and is now the most significant global health crisis since the influenza pandemic of 1918. (Cascella et al., 2022). In addition, this virus mainly spreads through contact with an infected person's cough or sneeze. The virus can also spread when a person touches something that has the virus on it before touching their eyes, nose, or mouth. Virus COVID-19 can survive for up to 72 hours (Bakar & Rosbi, 2020). Allegrante et al. (2020) assert that in order to effectively prevent the spread of COVID-19, health education and information must be targeted at various subpopulations and accurate, understandable, convincing, and relevant information must be provided. World Health Organization (WHO) and Centers for Disease Control and Prevention (CDC) have stated that one of the main prevention measures for the entire population is the appropriate use of personal protective equipment (PPE) as well as the adoption of effective hygiene systems (Cirrincione et al., 2022). Moreover, necessary precautions have been put in place to stop the spread of the virus and lower mortality rates, such as the requirement that everyone wears masks, consistent hand washing and hand sanitizing, social isolation, avoiding crowded places, remote working, and postponing public events. So, following COVID-19 prevention guidelines could assist to lower the prevalence of other infectious diseases like influenza, pneumonia, and Mycobacterium tuberculosis (Dadras, et al., 2021). Compliance with COVID-19 health protocols is crucial in preventing the spread of the virus and protecting public health. The specific protocols may vary based on recommendations from health authorities and the prevailing situation. The Department of Health's report from 2021 states that physical separation and hygiene standards will be needed in all venues, so when neighborhood businesses restart operations, they must adhere to hygiene standards to stop the spread of the virus. To ensure the security of staff, customers, and patients, health facilities and DOH divisions have modified their standards and procedures. Alternative work arrangements, utilizing virtual platforms to their fullest potential, and ensuring that infection control protocols are carefully followed are some strategies. Jobelle S. Teves, MAN, RN (jobelle.teves@spus.edu.ph) Marcu Augustu E. Mantilla (mantillamarcu@gmail.com) Jichell Grace A. Basol (jichellbasol@gmail.com) Lucy L. Teves, PhD, RN (ORCID No. 0000-0003-0939-2824) The novel COVID-19 disease rapidly spread worldwide and was declared a pandemic by the World Health Organization (WHO, 2020). The study aimed to evaluate the compliance of the college students, faculty, and non-teaching staff with COVID-19 health protocols at St. Paul University Surigao. The study used a quantitative descriptive survey that made use of adherence to health protocol, where the calculated sample size for the participants' occupation at the St. Paul University Surigao – Main campus was 379. The statistical method applied to the collected data was Mean and Standard Deviation, Frequency of Percentage Distribution, and Analysis of Variance. Findings drawn from the study and interpretations of the data gathered, the majority of participants are students 86.54%, according to their occupation. Moreover, there is no significant degree of variance in the compliance with COVID-19 health protocols in the occupation of the participants in terms of No Facemask, No Entry and Screening Prior to Entry. However, there is a significant degree of variance in the indicators of Physical Distancing, Practice Respiratory Etiquette, Hand washing and Hand disinfection, and Information dissemination. Therefore, it is recommended in this study that the administrators of St. Paul University Surigao continue to apply no facemask, no entry, practice respiratory etiquette, hand washing and hand disinfection and screening prior to entry while improving and developing different approaches on physical distancing, practice respiratory etiquette, hand washing and hand disinfection, and information dissemination to better cater the groups to further improve compliance to the health protocols. Future researchers are encouraged to further examine the differences and group them according to their profiles, address the same research problem in a different setting, context or location. The researchers also recommend conducting research that focuses on utilizing a convergent parallel design and incorporating natural observation. This research approach will provide valuable insights into the implementation and effectiveness of health protocols in promoting a safe and healthy environment for students, staff, and the wider school community. Keywords: COVID-19, Compliance, Health Protocols Keywords: COVID-19, Compliance, Health Protocols Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 1 1 DOI: https://doi.org/10.5281/zenodo.8216631 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 INTRODUCTION As a result, following government directives and public health advice is essential in lowering transmission rates during the COVID -19 pandemic, which is a worldwide health emergency. Nevertheless, a small percentage of people reportedly disregard laws and regulations (Banai et al., 2021). Compliance with COVID-19 health protocols is crucial in preventing the spread of the virus and protecting public health. The specific protocols may vary based on recommendations from health authorities and the prevailing situation. In this study, the researchers were eager to compile and assess the safety and health practices currently used at St. Paul University Surigao as required by the government to stop the spread of the virus and also to protect its staff and students. The researchers primarily focused on Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 2 DOI: https://doi.org/10.5281/zenodo.8216631 e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 International Journal of Research (IJR) the compliance by the teachers, students, and non-teaching staff to health protocols implemented at St. Paul University Surigao. Additionally, since the start of face-to-face sessions in the academic year 2022–2023, St. Paul University Surigao has strictly adhered to the government's necessary health procedures. Conceptual Framework of the Study This study focused on the examination of COVID-19 health protocols, building upon the research conducted by Daniel et al. (2022). The research highlighted the importance of complying with health protocols during the pandemic. The study recommended various practical measures, including hand washing in public spaces, maintaining physical distance, wearing masks, avoiding crowded areas, practicing respiratory etiquette, engaging in physical exercise, taking vitamins or supplements, and adopting a balanced diet. The implementation of these health protocols is crucial in preventing the transmission of COVID-19 within the community. It is essential for society to adhere to these protocols in order to effectively control the spread of the pandemic. Additionally, the health protocols support society's efforts to remain productive while staying safe. To raise public awareness, all institutions must work together (Saputra et al, 2020; Christiarini et al, 2020; as cited in Christian, 2021). The focus of this study, on the other hand, is based on the concept of Daniel et al., (2022) which supports health protocols such as hand washing in public spaces, physical distancing, frequency of using the mask, and avoiding crowded places which were highly correlated. Hence, in this study, the researchers focused on the health protocols practiced at St. Paul University Surigao which were similar to the concept of Daniel et al., (2022) such as no facemask means no entry, physical distancing, practicing respiratory etiquette, hand washing and disinfection, information dissemination, and screening prior to entry. Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 DOI: https://doi.org/10.5281/zenodo.8216631 3 Occupation of the Participants: • Faculty • Students • Non-Teaching Staffs Proposed Recommendations The extent of compliance with the COVID-19 health protocols in terms of: • No Facemask, No Entry • Physical Distancing • Handwashing and Hand Disinfection • Practice Respiratory Etiquette • Information Dissemination • Screening Prior to Entry Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 Occupation of the Participants: • Faculty • Students • Non-Teaching Staffs The extent of compliance with the COVID-19 health protocols in terms of: Proposed Recommendations 3 3 DOI: https://doi.org/10.5281/zenodo.8216631 Participants The researchers selected the faculty, non–teaching staff, and students at St. Paul University, Surigao. The participants were randomly selected using the multistage Cluster Sampling method to form a sample. The researchers procured the lists of the faculty, non-teaching staff, and students to get the sample size of targeted participants. Sample participants were taken using Slovin’s formula, used to calculate the sample size (326) given the population size (1,759) and a margin of error (e) which is computed as n=N/(1+Ne2). METHOD Figure 1: Schematic Diagram METHOD Figure 1: Schematic Diagram This study used descriptive-survey research design. This design investigates the compliance of the participants without any of them being managed or controlled by the researcher. This method is suitable since the purpose of the research is to evaluate participants’ compliance. Moreover, information on compliance with health practices was gathered using a standardized survey questionnaire. International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 Instrument In this study, a researcher-made questionnaire based on the Health Services COVID-19 Implementing Guidelines of St. Paul University, Surigao (2020) was used as a survey questionnaire to collect data from the respondents. Consequently, the researchers have selected the following indicators: 1.) No facemask, no entry; 2.) Physical distancing; 3.) Practice respiratory etiquette; 4.) Hand washing and disinfection; 5.) information dissemination; and 6.) entry screening to evaluate compliance with minimum health standards implemented at St. Paul University, Surigao. Faculty, non-teaching staff, and college students were asked to describe themselves in reference to verbal interpretation: always, often, sometimes, and never in the qualitative description; excellent compliance, good compliance, average compliance, and poor compliance. International Journal of Research (IJR) Mean and Standard Deviation. This tool was utilized to determine the effectiveness of the participant’s compliance. p p p The following are the basis for the interpretation of data: p p p The following are the basis for the interpretation of data: ANOVA (Analysis of Variance). This tool was employed to measure the degree of variance in the extent of compliance among the three group which are: college students, faculty, and non- teaching staff. International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 RESULTS AND DISCUSSION This chapter presents, analyzes, and interprets the data gathered. The presentation follows the sequence of problems posed in Chapter 1. This chapter presents, analyzes, and interprets the data gathered. The presentation follows the sequence of problems posed in Chapter 1. Occupation of the Participants Data Analysis Data Analysis The following statistical tools were used to answer the specific problems: Frequency Count and Percentage Distribution. This tool was used to determine the occupation of the participants. Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 4 DOI: https://doi.org/10.5281/zenodo.8216631 q p p Occupation of the Participants Table 1 shows the profile of the participants on whether they are faculty, students, or non- teaching staff. teaching staff. Table 1. Occupation of the Participants. Occupation f (n=379) % Faculty 21 5.54 Non-teaching Staff 30 7.92 Student 328 86.54 The table provided information that in terms of their occupation, it can be noticed that most of the participants are students. Specifically, 328 (86%) are students, Participants from the faculty and non–teaching staff were 21 (5.54%) and 30 (7.92%). Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 DOI: https://doi.org/10.5281/zenodo.8216631 5 Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 5 DOI: https://doi.org/10.5281/zenodo.8216631 DOI: https://doi.org/10.5281/zenodo.8216631 Legend: A-Always; EC-Excellent Compliance As shown in table 2, the indicator, “I wear facemask upon entering the school premises,” got the highest mean (M=3.70, SD=68), and verbally interpreted as Always with a qualitative description of Excellent Compliance. This means that it was prominent that the people entering the school premises would wear facemasks. While the indicator, “I wear a facemask even in classroom/offices,” got the lowest mean (M=3.51, SD=0.77), it is still verbally interpreted as Always with a qualitative description of Excellent Compliance. Despite being the lowest indicator, it still yielded a highly valued description. A study from Liang et al. (2020), found that masks shield other populations against respiratory virus infections and demonstrated the general effectiveness of masks in reducing the spread of respiratory virus infections. While wearing a mask may not keep people from getting COVID-19, it can help lower the odds. If they are sick, a mask can help keep germs from infecting others. If they are healthy, a mask can help keep respiratory droplets from someone who is sick from landing in their noses and mouth. On average, the compliance with the COVID-19 health protocols in terms of “No facemask, No Entry” got the mean of M=3.64, SD=0.71 with a qualitative description of Excellent Compliance. International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 Mean and Standard Deviation on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao Table 2 presents the compliance with COVID-19 health protocols in terms of No Facemask, No Entry. Table 2. The Compliance with COVID-19 Health Protocols in terms of No Facemask, No Entry Indicators Mean SD VI QD No Facemask, No Entry I wear facemask upon entering the school premises. 3.70 0.68 A EC I wear a facemask that covers both the nose and mouth. 3.69 0.68 A EC I wear facemask even in classrooms/offices. 3.51 0.77 A EC Average: 3.64 0.71 A EC Legend: A-Always; EC-Excellent Compliance Table 2. Mean and Standard Deviation on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao Table 3 presents the compliance with COVID-19 health protocols in terms of Physical Distancing. The Compliance with COVID-19 Health Protocols in terms of Physical Distancing Indicators Mean SD VI QD Physical Distancing Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 DOI: https://doi.org/10.5281/zenodo.8216631 6 The Compliance with COVID-19 Health Protocols in terms of Physical Distancing Indicators Mean SD VI QD Physical Distancing Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 Physical Distancing 6 DOI: https://doi.org/10.5281/zenodo.8216631 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 I observe physical distancing of at least (1) meter or six feet at all times. 2.57 1.02 O GC I follow the designated entrance and exit routes to minimize congestion. 3.49 0.76 A EC I follow the floor markings and other signages that are visible in the school premises. 3.27 0.75 A EC Average: 3.11 0.84 O GC Legend: A-Always; O-Often; EC-Excellent Compliance; GC-Good Compliance International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 I observe physical distancing of at least (1) meter or six feet at all times. 2.57 1.02 O GC I follow the designated entrance and exit routes to minimize congestion. 3.49 0.76 A EC I follow the floor markings and other signages that are visible in the school premises. 3.27 0.75 A EC Average: 3.11 0.84 O GC gend: A-Always; O-Often; EC-Excellent Compliance; GC-Good Compliance International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 I observe physical distancing of at least (1) meter or six feet at all times. 2.57 1.02 As shown in Table 3, the indicator, “I follow the designated entrance and exit routes to minimize congestion,” got the highest mean (M=3.49, SD=0.76), and verbally interpreted as Always with a qualitative description of Excellent Compliance. In addition, the indicator, “I follow the floor markings and other signages that are visible in the school premises,” got a high mean as well (M=3.27, SD=0.75), verbally interpreted as Always with a qualitative description of Excellent Compliance. This points that the respondents were compliant in following signages and routes to reduce possible contractions of the COVID-19. Mean and Standard Deviation on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao However, the indicator, “I observe physical distancing of at least (1) meter or six feet at all times,” got the lowest mean (M=2.57, SD=1.02) and verbally interpreted as Often with a qualitative description of Good Compliance. Despite having acceptable results, results reveal that the respondents were slightly challenged in distancing themselves apart by one meter. As COVID – 19 vaccinations roll out, people may be less compliant in socially distancing themselves. According to a study from Teslya, A. (2022), people may perceive themselves protected from COVID – 19, relying on a reduction in transmission brought on by increased vaccine coverage, therefore complying less. This is corroborated by a study from Andersson et al., (2021), where an increased focused on vaccines may result in low compliance to public health recommendations and hasten the spread of COVID – 19. On average, the compliance with the COVID-19 health protocols in terms of “Physical distancing” got the mean of M=3.11, SD=0.84 with a qualitative description of Good Compliance. Mean and Standard Deviation on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao Mean and Standard Deviation on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao Table 4 presents the compliance with COVID-19 health protocols in terms of Practice Respiratory Etiquette. Table 4. The Compliance with COVID-19 Health Protocols in terms of Practice Respiratory Etiquette Indicators Mean SD VI QD Practice Respiratory Etiquette I cover my mouth and nose when coughing and sneezing using a tissue, handkerchief, or with my inner elbow. 3.57 0.74 A EC I avoid touching my eyes, nose and mouth. 3.18 0.79 O GC I wash my hands with soap and water after coughing and/or sneezing. 3.27 0.77 A EC Average: 3.34 0.76 A EC Legend: A-Always; O-Often; EC-Excellent Compliance; GC-Good Compliance Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 7 DOI: https://doi.org/10.5281/zenodo.8216631 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 As shown in Table 4, the indicator, “I cover my mouth and nose when coughing and sneezing using a tissue, handkerchief, or with my inner elbow,” got the highest mean (M=3.57, SD=0.74), and verbally interpreted as Always with a qualitative description of Excellent Compliance. Additionally, the indicator, “I wash my hands with soap and water after coughing and sneezing,” gained a high mean (M=3.27, SD=0.77), and verbally interpreted as Always with a qualitative description of Excellent Compliance. Conversely, the indicator, “I avoid touching my eyes, nose, and mouth,” got the lowest mean (M=3.18, SD=0.79), verbally interpreted as Often with a qualitative description of Good Compliance. The data presented above points that most of the respondents would cover themselves when they cough and sneeze and would wash their hands after thoroughly. However, they found it hard to avoid touching their eyes, noses and mouth. According to Patel, et al., (2020), A possible decline of respiratory hygiene was observed after the lockdown was lifted because people were less afraid in contracting the COVID – 19. On average, the compliance with the COVID-19 health protocols in terms of “Practice Respiratory Etiquette” got the mean of M=3.34, SD=0.76 with a qualitative description of Excellent Compliance. Mean and Standard Deviation on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao Table 5 presents the compliance with COVID-19 health protocols in terms of Hand Washing and Hand Disinfection. Table 5. The Compliance with COVID-19 Health Protocols in terms of Hand Washing and Hand Disinfection Indicators Mean SD VI QD Hand washing and Hand disinfection I use liquid hand soap and clean water that are accessible in all hand washing stations. 3.44 0.76 A EC I wash my hands for at least 20 seconds while inside the campus. 2.71 1.08 O GC I use alcohol with 70% isopropyl or ethyl which are available in every classroom/office. 3.63 0.73 A EC Average: 3.26 0.86 A EC Legend: A-Always; O-Often; EC-Excellent Compliance; GC-Good Compliance As presented in Table 5, the indicator, “I use alcohol with 70% isopropyl or ethyl which are available in every classroom/office,” got the highest mean (M=3.63, SD=0.73) and verbally interpreted as Always with a qualitative description of Excellent Compliance. However, the indicator, “I wash my hands for at least 20 seconds while inside the campus,” got the lowest mean (M=2.71, SD=1.08), and verbally interpreted as Often with a qualitative description of Good Compliance. Washing hands for at least 20 seconds while inside the campus was less likely observed and the people preferred using alcohol instead. Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 DOI: https://doi.org/10.5281/zenodo.8216631 8 Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 8 DOI: https://doi.org/10.5281/zenodo.8216631 e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 International Journal of Research (IJR) According to a study by Roy et al. (2020), disinfection using the right and recommended physical or chemical disinfectants will not only reduce the spread of the illness but will also greatly help to flatten the curve. It is equally important to wash your hands with soap and water or with products containing alcohol. On average, the compliance with the COVID-19 health protocols in terms of “Hand washing and Hand disinfection” got the mean of M=3.26, SD=0.86 with a qualitative description of Excellent Compliance. Mean and Standard Deviation on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao Table 6 presents the compliance with COVID-19 health protocols in terms of Information Dissemination. Table 6. The Compliance with COVID-19 Health Protocols in terms of Information Dissemination Indicators Mean SD VI QD Information dissemination I use the official Facebook Page of the Health Services to update myself on health concerns and reminders, especially about COVID-19 2.64 1.09 O GC I follow the posters that are placed inside the campus to remind me of the following: Hand washing, Cough and sneeze etiquette, Proper disposal of waste, Physical distancing 3.36 0.75 A EC Average: 3.00 0.92 O GC Legend: A-Always; O-Often; EC-Excellent Compliance; GC-Good Compliance Mean and Standard Deviation on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao Mean and Standard Deviation on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao Table 7 presents the compliance with COVID-19 health protocols in terms of Screening Prior to Entry. Table 7. The Compliance with COVID-19 Health Protocols in terms of Screening Prior to Entry Indicators Mean SD VI QD Screening Prior to Entry I check myself prior to going to school to ensure that I can enter the school premises. 3.57 0.75 A EC Using the non-contact thermometer, I check if my temperature is below 37.5 °C. 3.54 0.77 A EC I use the treated foot bath that is placed on the entrance gate to disinfect my footwear. 3.34 0.99 A EC Average: 3.48 0.84 A EC Legend: A-Always; EC-Excellent Compliance International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 Mean and Standard Deviation on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao Information dissemination As shown in Table 6, the indicator, “I follow the posters that are placed inside the campus to remind me of the following: Hand washing, Cough and sneeze etiquette, Proper disposal of waste, Physical distancing,” got the highest mean among the two (M=3.36, SD=0.75), and verbally interpreted as Always with a qualitative description of Excellent Compliance. Indicating that the posters placed around the campus was effective. Contrastingly, the indicator, “I use the official Facebook Page of the Health Services to update myself on health concerns and reminders, especially about COVID-19,” acquired a low mean (M=2.64, SD=1.09), and verbally interpreted as Often with a qualitative description of Good Compliance. Pointing that most of the respondents would update themselves on Facebook with regards to health updates. However, though the indicator acquired an acceptable result, it also shows that the posters placed inside the campus was effectively followed more than by checking into the Facebook page itself. On average, the compliance with the COVID-19 health protocols in terms of “Information dissemination” got the mean of M=3.00, SD=0.92 with a qualitative description of Good Compliance. Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 9 DOI: https://doi.org/10.5281/zenodo.8216631 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 Mean and Standard Deviation on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 (IJR) Vol. 10 Issue 08 August 2023 Table 8 presents the summary on the compliance with COVID-19 health protocols at St. Paul University Surigao Table 8. Summary on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao The Compliance with Covid-19 Health Protocols Mean SD VI QD No Facemask, No Entry 3.64 0.71 A EC Physical Distancing 3.11 0.84 O GC Practice Respiratory Etiquette 3.34 0.76 A EC Hand washing and Hand disinfection 3.26 0.86 A EC Information dissemination 3.00 0.92 O GC Screening Prior to Entry 3.48 0.84 A EC General Average: 3.31 0.82 A EC Legend: Scale Range Verbal Interpretation Qualitative Description 4 3.25-4.00 Always (A) Excellent Compliance (EC) 3 2.50-3.24 Often (O) Good Compliance (GC) 2 1.75-2.49 Sometimes (S) Average Compliance (AC) 1 1.00-1.74 Never (N) Poor Compliance (PC) 8 presents the summary on the compliance with COVID-19 health protocols at St y Surigao Table 8 presents the summary on the compliance with COVID-19 health protocols at St. Paul University Surigao Table 8. Summary on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao The Compliance with Covid-19 Health Protocols Mean SD VI QD No Facemask, No Entry 3.64 0.71 A EC Physical Distancing 3.11 0.84 O GC Practice Respiratory Etiquette 3.34 0.76 A EC Hand washing and Hand disinfection 3.26 0.86 A EC Information dissemination 3.00 0.92 O GC Screening Prior to Entry 3.48 0.84 A EC General Average: 3.31 0.82 A EC L d Summary on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao The Compliance with Covid-19 Health Protocols Mean SD VI QD No Facemask, No Entry 3.64 0.71 A EC Physical Distancing 3.11 0.84 O GC Practice Respiratory Etiquette 3.34 0.76 A EC Hand washing and Hand disinfection 3.26 0.86 A EC Information dissemination 3.00 0.92 O GC Screening Prior to Entry 3.48 0.84 A EC General Average: 3.31 0.82 A EC Legend: Scale Range Verbal Interpretation Qualitative Description 4 3.25-4.00 Always (A) Excellent Compliance (EC) 3 2.50-3.24 Often (O) Good Compliance (GC) 2 1.75-2.49 Sometimes (S) Average Compliance (AC) 1 1.00-1.74 Never (N) Poor Compliance (PC) Based on Table 8, the respondents gave the highest overall rating on No Facemask, No Entry (M=3.64, SD=0.71), which is verbally interpreted as Always and qualitatively described as Excellent Compliance. Legend: A-Always; EC-Excellent Compliance Legend: A-Always; EC-Excellent Compliance As shown in Table 7, indicator one, “I check myself prior to going to school to ensure that I can enter the school premises,” got the highest mean (M=3.57, SD=0.75), and verbally interpreted as Always with a qualitative description of Excellent Compliance. The indicator, “Using the non-contact thermometer, I check if my temperature is below 37.5 °C,” also got a high mean (M=3.54, SD=0.77), with a verbal interpretation of Always and a qualitative description of Excellent Compliance. While the indicator, “I use the treated foot bath that is placed on the entrance gate to disinfect my footwear,” although having the lowest mean (M=3.34, SD=0.99), is still verbally interpreted as Always with a qualitative description of Excellent Compliance. All three indicators acquired high results. This means that the respondents would thoroughly check themselves prior to entry and followed proper screening. Measures of prevention, protection, screening, isolation and distribution have been shown to be efficient in similar settings (Basile et al., 2021). On average, the compliance with the COVID-19 health protocols in terms of “Screening Prior to Entry” got the mean of M=3.31, SD=0.82 with a qualitative description of Excellent Compliance. Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 DOI: https://doi.org/10.5281/zenodo.8216631 10 Mean and Standard Deviation on The Compliance with COVID-19 Health Protocols at St. Paul University Surigao Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 10 Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 DOI: https://doi.org/10.5281/zenodo.8216631 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 Significant Difference between The Compliance with COVID-19 Health Protocols at St. Paul University Surigao and the Occupation of the Participants The table below present the significant difference between The Compliance with COVID- 19 Health Protocols at St. Paul University Surigao and the Occupation of the Participants. ab e 9. Significant Difference on The Compliance with COVID-19 Health Protocols at St Surigao and the Occupation of the Participants with respect to their Occupation Dependent Sum of Squares df Mean Square F p-value Decision No Facemask, No Entry 0.28 2 0.14 0.37 0.694 Do not reject Ho Physical Distancing 4.47 2 2.23 5.71 0.004 Reject Ho Practice Respiratory Etiquette 2.31 2 1.15 3.14 0.045 Reject Ho Hand washing and Hand disinfection 5.30 2 2.65 7.18 0.001 Reject Ho Information dissemination 9.13 2 4.57 8.63 0.000 Reject Ho Screening Prior to Entry 2.63 2 1.31 2.69 0.069 Do not reject Ho As to the significant difference between the occupation of the participants and the variables No Facemask, No Entry and Screening Prior to Entry, findings revealed that there is no significant difference between both variables (p-values=0.694 and 0.069, respectively). However, as to the significant difference between the occupation of the participants and the variables Physical Distancing, Practice Respiratory Etiquette, Hand washing and Hand disinfection, Information dissemination, findings revealed that there is significant difference between these variables (p-values=0.004, 0.045, 0.001, 0.000, respectively). International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 Consecutively, the second-highest variable as evaluated by the respondents is Screening Prior to Entry with the mean of M=3.48, SD=0.84, and is verbally interpreted as Always with a qualitative description of Excellent Compliance. Lastly, Information Dissemination got the lowest average (M=3.00, SD=0.92), verbally interpreted as Often and qualitatively described as Good Compliance. In general, The Compliance with COVID-19 Health Protocols at St. Paul University Surigao got the overall mean of M=3.31, SD=0.82 and verbally interpreted as Always with a qualitative description of Excellent Compliance. Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 11 DOI: https://doi.org/10.5281/zenodo.8216631 DOI: https://doi.org/10.5281/zenodo.8216631 DOI: https://doi.org/10.5281/zenodo.8216631 Findings Based on the analysis and interpretations done on the data gathered, the different findings in this study, based on each statement of the problems articulated, summarized as follows: 1. As to the occupation of the participants, majority of the participants are students (328 or 86.54%). 1. As to the occupation of the participants, majority of the participants are students (328 or 86.54%). 2. As to Compliance with COVID-19 Health Protocols at St. Paul University Surigao: 2. As to Compliance with COVID-19 Health Protocols at St. Paul University Surigao: Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 12 DOI: https://doi.org/10.5281/zenodo.8216631 e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 International Journal of Research (IJR) 2.1 No Facemask, No Entry is excellently complied with an average mean of 3.64, and I wear facemask upon entering the school premises is qualitatively described as Excellent Compliance (M=3.70, SD=0.68). 2.1 No Facemask, No Entry is excellently complied with an average mean of 3.64, and I wear facemask upon entering the school premises is qualitatively described as Excellent Compliance (M=3.70, SD=0.68). 2.1 No Facemask, No Entry is excellently complied with an average mean of 3.64, and I wear facemask upon entering the school premises is qualitatively described as Excellent Compliance (M=3.70, SD=0.68). p ( ) 2.2 Physical Distancing had good compliance with an average mean of 3.11, and I follow the designated entrance and exit routes to minimize congestion is qualitatively described as Excellent Compliance (M=3.49, SD=0.76). p ( ) 2.2 Physical Distancing had good compliance with an average mean of 3.11, and I follow the designated entrance and exit routes to minimize congestion is qualitatively described as Excellent Compliance (M=3.49, SD=0.76). p ( ) 2.2 Physical Distancing had good compliance with an average mean of 3.11, and I follow the designated entrance and exit routes to minimize congestion is qualitatively described as Excellent Compliance (M=3.49, SD=0.76). 2.3 Practice Respiratory Etiquette is excellently complied with an average mean of 3.34, and I cover my mouth and nose when coughing and sneezing using a tissue, handkerchief, or with my inner elbow is qualitatively described as Excellent Compliance (M=3.57, SD=0.74). 2.4 Hand Washing and Hand Disinfection is excellently complied with an average mean of 3.26, and I use alcohol with 70% isopropyl or ethyl which are available in every classroom/office is qualitatively described as Excellent Compliance (M=3.63, SD=0.73). 2.5 Information Dissemination had good compliance with an average mean of 3.00, and I follow the posters that are placed inside the campus to remind me of the following: Hand washing, Cough and sneeze etiquette, Proper disposal of waste, Physical distancing is qualitatively described as Excellent Compliance (M=3.36, SD=0.75). 2.5 Information Dissemination had good compliance with an average mean of 3.00, and I follow the posters that are placed inside the campus to remind me of the following: Hand washing, Cough and sneeze etiquette, Proper disposal of waste, Physical distancing is qualitatively described as Excellent Compliance (M=3.36, SD=0.75). International Journal of Research (IJR) 2.6 Screening Prior to Entry is excellently complied with an average mean of 3.48, and I check myself prior to going to school to ensure that I can enter the school premises is qualitatively described as Excellent Compliance (M=3.57, SD=0.75). 3. There is no significant degree of variance in the compliance with COVID-19 health protocols in the occupation of the participants in terms of No Facemask, No Entry and Screening Prior to Entry (p-values=0.694 and 0.069, respectively). However, there is a significant degree of variance in the indicators Physical Distancing, Practice Respiratory Etiquette, Hand washing and Hand disinfection, Information dissemination (p-values=0.004, 0.045, 0.001, 0.000, respectively). The conclusions that may be drawn from the study's findings are as follows: The study and interpretations of the data gathered, the majority of participants are students (86.54%), according to their occupation. Moreover, there is no significant degree of variance in the compliance with COVID-19 health protocols in the occupation of the participants in terms of No Facemask, No Entry and Screening Prior to Entry. However, there is a significant degree of variance in the indicators Physical Distancing, Practice Respiratory Etiquette, Hand washing and Hand disinfection, and Information dissemination. This suggests that health protocols are effective in maintaining clean facilities on campus, the administrator of St. Paul University Surigao may continue to use them as a health guideline. 13 Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 DOI: https://doi.org/10.5281/zenodo.8216631 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 DOI: https://doi.org/10.5281/zenodo.8216631 15 REFERENCES Allegrante, J.P., Auld, M.E., Natarajan, S. (2020). Preventing covid 19 and its sequela: "There is no magic bullet... it's just behavior" American Journal of Preventive Medicine. https://doi.org/10.1016/j.amepre.2020.05.004 Andersson, O., Campos-Mercade, P., Meier, A. N., & Wengström, E. (2021). Anticipation of COVID-19 vaccines reduces willingness to socially distance. Journal of Health Economics, 80, 102530. https://doi.org/10.1016/j.jhealeco.2021.102530 Backer, J., Mollema, L., Vos, E., Klinkenberg, D., Klis, F., Melker, H., Holf, S., Wallinga, J. (2021). Impact of physical distancing measures against COVID-19 on contacts and mixing patterns: repeated cross-sectional surveys, the Netherlands, 2016 17, April 2020 and June 2020. Eurosurveillance. https://www.eurosurveillance.org/content/10.2807/15607917.ES.2021.26.8.2000994?cra wler=true#abstract_content Bakar, N.A., Rosbi, S. (2020). Effect of coronavirus disease to tourism industry. International Journal of Advanced Engineering Research and Science. https://ijaers.com/detail/effect-of-coronavirus-disease-covid-19-to-tourism industry/ Recommendations 1. The school administration can continue to apply no facemask, no entry, practice respiratory etiquette, hand washing and hand disinfection and screening prior to entry as safety health protocols inside the facilities to be complied by the teachers, students and non – teaching staff. 1. The school administration can continue to apply no facemask, no entry, practice respiratory etiquette, hand washing and hand disinfection and screening prior to entry as safety health protocols inside the facilities to be complied by the teachers, students and non – teaching staff. 2. However, the school administration may need to improve physical distancing and information dissemination as safety health protocols to be complied by the teachers, students and non – teaching staff. 2. However, the school administration may need to improve physical distancing and information dissemination as safety health protocols to be complied by the teachers, students and non – teaching staff. 3. In terms of physical distancing, practice respiratory etiquette, hand washing and hand disinfection, and information dissemination, the school administration may develop different approaches to better cater the groups to further improve compliance to the health protocols. 4. Future researchers could further examine the differences and group them according to their profiles, address the same research problem in a different setting, context or location. Future researchers of St. Paul University Surigao are encouraged to address the problem in this event and furthermore, look into finding ways to help mitigate the effects of COVID – 19 and improving the lives of everyone affected. The researchers also recommend conducting research that focuses on utilizing a convergent parallel design and incorporating natural observation. This research approach will provide valuable insights into the implementation and effectiveness of health protocols in promoting a safe and healthy environment for students, staff, and the wider school community. Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 14 DOI: https://doi.org/10.5281/zenodo.8216631 DOI: https://doi.org/10.5281/zenodo.8216631 DOI: https://doi.org/10.5281/zenodo.8216631 REFERENCES Allegrante, J.P., Auld, M.E., Natarajan, S. (2020). Preventing covid 19 and its sequela: "There is no magic bullet... it's just behavior" American Journal of Preventive Medicine. https://doi.org/10.1016/j.amepre.2020.05.004 15 DOI: https://doi.org/10.5281/zenodo.8216631 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 Banai, I., Banai, B., Mikloušić, I. (2021). Beliefs in COVID-19 conspiracy theories, compliance with the preventive measures, and trust in government medical officials. https://link.springer.com/article/10.1007/s12144-021-01898-y#Sec21 Breidablik, H., Lysebo, D., Johannessen, L., Skare, A., Andersen, J., Kleiven, O. (2020). Effects of hand disinfection with alcohol hand rub, ozonized water, or soap and water: time for reconsideration? Science Direct. https://www.sciencedirect.com/science/article/pii/S0195670120301134 Cai, M., Lou, H., Meng, X., Cui, Y., Wang, W. (2022). Influence of information attributes on information dissemination in public health emergencies. Humanities and Social Sciences Communications. https://www.nature.com/articles/s41599-022-01278- 2#Sec29 Cascella, M., Rajnik, M., Aleem, A., Dulebohn, S.C., Napoli, R.D. (2022). Features, Evaluation, and Treatment of Coronavirus. https://www.ncbi.nlm.nih.gov/books/NBK554776/ Chea, B., Bolt, A., Chaab, M., Dincer, I. (2021). Assessment of effectiveness of optimum physical distancing phenomena for COVID-19. AIP Publishing LLC. https://aip.scitation.org/doi/full/10.1063/5.0046429 Christian, A.C.S., Sa'id, M. (2021). Public obedience to health protocols during covid-19 pandemic in Indonesia: A perspective from health belief model theory. Journal of Socia Sciences and Humanities. ojs.pnb.ac.id/index.php/SOSHUM/ Chu, D., Akl, A., Duda, S., Solo, K., Yaacoub, S., Schunemann, H. (2020). Physical International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 Cai, M., Lou, H., Meng, X., Cui, Y., Wang, W. (2022). Influence of information attributes on information dissemination in public health emergencies. Humanities and Social Sciences Communications. https://www.nature.com/articles/s41599-022-01278- https://www.ncbi.nlm.nih.gov/books/NBK554776/ 16 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 distancing, face masks, and eye protection to prevent person-to-person transmission of SARS-CoV-2 and COVID-19: a systematic review and meta- analysis. Science Direct. https://www.sciencedirect.com/science/article/pii/S0140673620311429#cesec10 Chiodni, J. (2020). Good Practice Guidance for Providing a Travel Health Service. National Library of Medicine. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7679228/ Chavez, S., Long, B., Koyfman, A., Liang, S. (2021).Coronavirus Disease (COVID-19): A primer for emergency physicians. Science Direct. https://www.sciencedirect.com/science/article/pii/S0735675720301789 Cirrincione, L., Plescia, F., Ledda, C., Rapisardo, V., Martorana, D., Lacca, G., Argo, A., Zerbo, S., Vitale, E., Vinnikov, D., Cannizzaro, E. (2022). Covid 19 pandemic: New prevention and protection measure. https://www.mdpi.com/2071 1050/14/8/4766 Claresta, A., Christian, S. (2021). Public Obedience to Health Protocols during Covid-19 Pandemic in Indonesia: A Perspective from Health Belief Model Theory. Soshum: Jurnal Sosial dan Humaniora. https://ojs.pnb.ac.id/index.php/SOSHUM/article/view/2549 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7679228/ https://www.sciencedirect.com/science/article/pii/S0735675720301789 Cirrincione, L., Plescia, F., Ledda, C., Rapisardo, V., Martorana, D., Lacca, G., Argo, A., Zerbo, S., Vitale, E., Vinnikov, D., Cannizzaro, E. (2022). Covid 19 pandemic: New prevention and protection measure. https://www.mdpi.com/2071 1050/14/8/4766 Claresta, A., Christian, S. (2021). Public Obedience to Health Protocols during Covid-19 Pandemic in Indonesia: A Perspective from Health Belief Model Theory Soshum: Jurnal Claresta, A., Christian, S. (2021). Public Obedience to Health Protocols during Covid-19 Pandemic in Indonesia: A Perspective from Health Belief Model Theory. Soshum: Jurnal Sosial dan Humaniora. https://ojs.pnb.ac.id/index.php/SOSHUM/article/view/2549 Coronavirus & COVID-19: Glossary of Terms \| UVA Health. (n.d). UVA Health. Retrieved April 22, 2022, from https://uvahealth.com/services/covid19-glossary#:%7E:text=The %20coronavirus %20that %20causes%20COVID,acute%20respiratory%20syndrome %20coronavirus %202.&text=The%20name%20of%20the%20illness,for %20%22coronavirus%20disease %202019.%E2%80%9D Dadras, O., Alinaghi, S., Karimi, A., Mohsseni, M., Barzegary, A., Vahedi, F., Pashaei, Z., Mirzapour, P., Fakhfouri, A., Zargari, G., Saeidi, S., Mojdeganlou, H., Badri, Qaderi, K., Behnezhad, Mehraeen, E. (2021). Effects of COVID-19 prevention Received: 22 July 2023 Revised: 2 August 2023 procedures on other common infections: a systematic review. European Journal of Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 17 DOI: https://doi.org/10.5281/zenodo.8216631 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 Daniel, D., Kurniawan, A., Pinawati, A., Thohira, M., Annaduzzaman, M. (2022). The COVID-19 health protocol among university students: Case studies in thress cities in Indonesia. International Journal of Environmental Research and Public Health. https://doi.org/10.3390/ijerph191710630 l, D., Kurniawan, A., Pinawati, A., Thohira, M., Annaduzzaman, M. (2022). The tment of Health. (2021). The new normal for health. Received: 22 July 2023 Revised: 2 August 2023 Fi l A t d 10 J l 2023 Deviante, R., Michael, T. (2021). Accountability of mosque administrators against violation of Covid-19 health protocols in the mosque environment. International Journ of Research in Business and Social Science. https://ssbfnet.com/ojs/index.php/ijrbs/article/view/1050 Edet, C., Harry, A., Wegbom, A., Raimi, O., Fagbamigbe, A., Kiri, V. (2020). Face Mask Utilization in the Era of COVID-19: Nigeria Experience. Research Square. https://assets.researchsquare.com/files/rs-103515/v1/61079b39-e6e0-48ed-bac5- 6fdc55311762.pdf?c=1631860114 Eikenberry, S., Mancuso, M., Iboi, E., Eikenberry, K., Kuang, Y., Kostelich, E., Gumel, A. (2020). To mask or not to mask: Modeling the potential for face mask use by the general public to curtail the COVID-19 pandemic. Science Direct. https://www.sciencedirect.com/science/article/pii/S2468042720300117 https://ssbfnet.com/ojs/index.php/ijrbs/article/view/1050 https://assets.researchsquare.com/files/rs-103515/v1/61079b39-e6e0-48ed-bac5- 6fdc55311762.pdf?c=1631860114 A. (2020). To mask or not to mask: Modeling the potential for face mask use by the general public to curtail the COVID-19 pandemic. Science Direct. Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 18 DOI: https://doi.org/10.5281/zenodo.8216631 DOI: https://doi.org/10.5281/zenodo.8216631 01404-4 Jones, N., Qureshi, Z., Temple, R., Larwood, J., Greenhalgh, T., Bourouiba, L. (2020). Two metres or one: what is the evidence for physical distancing in covid-19? BMJ Publishing Group Ltd. https://www.bmj.com/content/370/bmj.m3223 Lathifa, A., Kamalia, F., Putra, F., Nuryanti, L. (2021). Student Compliance in Doing Health Protocols during the Covid-19 Pandemic. Proceding of Inter-Islamic University Conference on Psychology. https://press.umsida.ac.id/index.php/iiucp/article/view/598 Liang, M., Gao, L., Cheng, C., Zhou, Q., Uy, J., Heiner, K., Sun, C. (2020). Efficacy of face mask in preventing respiratory virus transmission: A systematic review and meta analysis. Science Direct. International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 Fadda, M., Albanese, E., Suggs, L. (2020). When a COVID-19 vaccine is ready, will we all be ready for it? Springer Link. https://link.springer.com/article/10.1007/s00038-020- https://www.sciencedirect.com/science/article/pii/S1477893920302301#sec5 Liao, M., Liu, H., Wang, X., Hu, X., Huang, Y., Liu, X., Brenan, K., Mecha, J., Nirmalan, M., Lu, J. (2021). A technical review of face mask wearing in preventing respiratory COVID-19 transmission. Science Direct. https://www.sciencedirect.com/science/article/pii/S1359029421000017#coi0010 Ma, R., Deng, Z., Wu, M. (2020). Effects of Health Information Dissemination on User Follows and Likes during COVID-19 Outbreak in China: Data and Content Analysis. International Journal of Environmental Research and Public Health. https://www.mdpi.com/1660-4601/17/14/5081 19 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 Marano, N. (2020). Considerations for Health Screening for COVID-19 at Points of Entry. Journal of Travel Medicine. https://www.cdc.gov/coronavirus/2019- ncov/downloads/global-covid-19/Point-of-Entry-Screening-for-COVID-19.pdf Nsawotebba, A., Ibanda, I., Ssewanyana, I., Ogwok, P., Ocen, F., Okiira, C., Kagirita, A., Mujuni, D., Tugumisirize, D., Kabugo, J., Nyombi, A., Majwala, R., Bagaya, B.,..., Nabadda, S. (2021). Effectiveness of thermal screening in detection of COVID-19 among truck drivers at Mutukula Land Point of Entry, Uganda. PLOS Org. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0251150#sec018 Offei, M., Salifu, Y., Adewale, B., Bayuo, J., Puko, R., Lokko, E. (2021). A rapid review of the use of face mask in preventing the spread of COVID-19. Science Direct. https://www.sciencedirect.com/science/article/pii/S2666142X20300126 Patel, D., Desai, R., Parmar, J., Ramavat, M. (2020). A cross-sectional study regarding respiratory etiquette and stress during novel coronavirus pandemic. National Journal of Physiology, Pharmacy and Pharmacology. http://www.njppp.com/fulltext/28- 1597733221.pdf Pelatihan, B., Cikarang, K. (2020). Implementation of health protocols in PUSKESMAS offices in pandemic. MONAS Jurnal Inovasi Aparatur. https://ejournal- bpsdm.jakarta.go.id/index.php/monas/article/view/52 Roy, A., Parida, S., Bhatia, V. (2020). Role of disinfection and hand hygiene: a COVID-19 Marano, N. (2020). Considerations for Health Screening for COVID-19 at Points of Entry. Journal of Travel Medicine. https://www.cdc.gov/coronavirus/2019- ncov/downloads/global-covid-19/Point-of-Entry-Screening-for-COVID-19.pdf Nsawotebba, A., Ibanda, I., Ssewanyana, I., Ogwok, P., Ocen, F., Okiira, C., Kagirita, A., Mujuni, D., Tugumisirize, D., Kabugo, J., Nyombi, A., Majwala, R., Bagaya, B.,..., Nabadda, S. (2021). Effectiveness of thermal screening in detection of COVID-19 among truck drivers at Mutukula Land Point of Entry, Uganda. PLOS Org. 20 International Journal of Research (IJR) e-ISSN: 2348-6848 p-ISSN: 2348-795X Vol. 10 Issue 08 August 2023 https://academic.oup.com/jpubhealth/article/43/2/e305/6130168 Taylor, C. (2020). Information Dissemination. Cyberhoot. https://cyberhoot.com/cybrary/dissemination-of-information/ http://www.scielo.org.co/scielo.php?script=sci_arttext&pid=S0120-53072018000200009 Virginia Department of Health (VDN). (2021). Respiratory Hygiene and Cough Etiquette. https://www.vdh.virginia.gov/content/uploads/sites/3/2016/01/RespiratoryHygieneCough Etiquette_FAQ.pdf Received: 22 July 2023 Revised: 2 August 2023 Final Accepted 10 July 2023 Copyright Ó authors 2023 21 DOI: https://doi.org/10.5281/zenodo.8216631 DOI: https://doi.org/10.5281/zenodo.8216631
https://openalex.org/W2906775972	https://europepmc.org/articles/pmc6334357?pdf=render	English	null	Presentation of an Approach on Determination of the Natural Frequency of Human Lumbar Spine Using Dynamic Finite Element Analysis	Applied bionics and biomechanics	2,019	cc-by	5,919	Fan Ruoxun ,1 Liu Jie,1 Liu Jun,2 and Wang Weijun1 1Department of Automotive Engineering, Jilin Institute of Chemical Technology, Jilin 132022, China 2No. 2 Hospital of Jilin University, Jilin University, Changchun 130025, China Fan Ruoxun ,1 Liu Jie,1 Liu Jun,2 and Wang Weijun1 1Department of Automotive Engineering, Jilin Institute of Chemical Technology, Jilin 132022, China 2No. 2 Hospital of Jilin University, Jilin University, Changchun 130025, China Fan Ruoxun ,1 Liu Jie,1 Liu Jun,2 and Wang Weijun1 1Department of Automotive Engineering, Jilin Institute of Chemical Technology, Jilin 132022, China 2No. 2 Hospital of Jilin University, Jilin University, Changchun 130025, China Fan Ruoxun ,1 Liu Jie,1 Liu Jun,2 and Wang Weijun1 1Department of Automotive Engineering, Jilin Institute of Chemical Technology, Jilin 132022, China 2No. 2 Hospital of Jilin University, Jilin University, Changchun 130025, China Correspondence should be addressed to Fan Ruoxun; fanruoxun@126.com Received 9 August 2018; Revised 25 October 2018; Accepted 6 November 2018; Published 2 January 2019 Academic Editor: Jan Harm Koolstra Copyright © 2019 Fan Ruoxun et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Occurring resonance may negatively aﬀect the health of the human lumbar spine. Hence, vibration generated in working and living environments should be optimized to avoid resonance when identifying the natural frequency of the human lumbar spine. The range of the natural frequency of the human lumbar spine has been investigated, but its speciﬁc numerical value has not been determined yet. This study aimed at presenting an approach based on resonance for predicting the speciﬁc numerical value of the natural frequency of the human lumbar spine. The changes in the numerical ﬂuctuation amplitudes and the cycles of lumbar mechanical parameters during resonance are greater than those during nonresonant vibration. Given that the range of the natural frequency has been identiﬁed, vibrations at diﬀerent excitation frequencies within this range can be applied in a human lumbar ﬁnite element model for dynamic ﬁnite element analysis. When the excitation frequency is close to the natural frequency, resonance occurs, causing great changes in the numerical ﬂuctuation amplitudes and the cycles of lumbar mechanical parameters. Therefore, the natural frequency of the lumbar ﬁnite element model could be back-calculated. Results showed that the natural frequency of the established model was 3.5 Hz. Meanwhile, the closer the excitation frequency was to the natural frequency, the greater the changes in the numerical ﬂuctuation amplitudes and cycles in the parameters would be. Hindawi Applied Bionics and Biomechanics Volume 2019, Article ID 5473891, 8 pages https://doi.org/10.1155/2019/5473891 Hindawi Applied Bionics and Biomechanics Volume 2019, Article ID 5473891, 8 pages https://doi.org/10.1155/2019/5473891 Fan Ruoxun ,1 Liu Jie,1 Liu Jun,2 and Wang Weijun1 1Department of Automotive Engineering, Jilin Institute of Chemical Technology, Jilin 132022, China 2No. 2 Hospital of Jilin University, Jilin University, Changchun 130025, China This study presented an approach for predicting the speciﬁc numerical value of the natural frequency of the human lumbar spine. Identifying the natural frequency assists in ﬁnding preventive measures for lumbar injury caused by vibration and in designing the vibration source in working and living environments to avoid approximating to the natural frequency of the human lumbar spine. 1. Introduction should reduce vibrating time to decrease the incidence of lumbar degeneration. If a vibrational environment is inev- itable, the vibration source should be optimized to at least prevent resonance. Lumbar degeneration is a spine degenerative disorder and one of the common causes of low-back pain [1, 2]. Mean- while, vibration is one of the important factors leading to lumbar degeneration [3]. Vibrational environment has a nonnegligible inﬂuence on human lumbar tissue, and even long-term vibration in low frequency and small amplitude may cause or accelerate lumbar degeneration [4]. Resonance occurs when the excitation frequency is close to the natural frequency of the human lumbar spine. The force and deformation in the lumbar during resonance are more seri- ous than those in other vibrations [5, 6]. Therefore, we To optimize the vibration source and subsequently pre- vent resonance, we need to determine the speciﬁc numerical value of the natural frequency of the human lumbar spine ﬁrst. Previous experiments measured that the natural fre- quency of the human lumbar spine was within 1–6 Hz [7, 8]. At present, the natural frequency of the human lum- bar spine is mainly predicted by dynamic ﬁnite element (FE) analysis because of the restrictions in the law and human lumbar specimen. Several FE analyses have predicted Applied Bionics and Biomechanics 2 2.2. Assignment of Lumbar Material Parameters in the FE Model. Porous seepage-stress coupled dynamic FE analysis was performed on the established lumbar FE model, which needs to assign biphasic materials including solid and porous parameters [15, 18]. Most of the lumbar spinal structures were assigned with the biphasic materials, except for the lig- aments, facets, and annulus ﬁbrosus, which were assigned with the solid phase [19]. The solid phase in the vertebral bone was simulated with linear elastic material; in the annu- lus, ground substance and nucleus pulposus were assumed as neo-Hookean hyperelastic materials [20]. The nonlinear stress-strain relationships among each ligament were ﬁtted from a previous experiment, and the elastic moduli of the annulus ﬁbrosus were obtained from literature [21, 22]. All the porous material parameters were selected based on liter- ature [19, 23, 24]. The speciﬁc material parameters were shown in Table 1. and narrowed the range of the natural frequency to 2–5 Hz [9–11]. However, applying dynamic FE analysis to precisely predict natural frequency is challenging. The human lumbar spine is composed of vertebral bone, intervertebral disc, and ligament. 1. Introduction The bones and soft tissues cannot be simulated with the same material parameter. According to vibration theory, the mechanical properties of a structure composed of multiple materials are diﬃcult to predict by dynamic FE analysis because the elastic moduli in each material cannot be distributed in the structure [12, 13]. Therefore, the speciﬁc numerical value of the natural frequency of the human lum- bar spine is diﬃcult to precisely predict using conventional FE analysis. To overcome the limitations in conventional FE analysis, we intended to predict the natural frequency of the human lumbar spine based on resonance, that is, the changes in the numerical ﬂuctuation amplitudes and the cycles of lumbar mechanical parameters caused by resonant vibration are evi- dently greater than those caused by other nonresonant vibra- tions [6, 9, 14]. Once the range of the natural frequency of the human lumbar spine is known, the excitation frequency that induced resonance can be explored, and this excitation fre- quency should be close to the natural frequency of the human lumbar spine. Permeability was set to be changed with void ratio and can be directly implemented into the ABAQUS code by pro- viding the calculated permeability values for a range of allow- able void ratios. The variable permeability formulation was taken from literature [18]. The permeability k is dependent on void ratio e: k = k0 e 1 + e0 e0 1 + e 2 exp 8 5 1 + e 1 + e0 −1 , 1 p Accordingly, this study aimed at presenting an approach to predict the speciﬁc numerical value of the natural fre- quency of the human lumbar spine. A poroelastic FE model of human lumbar spinal segments L2–L3 was primarily established based on healthy human computed tomography (CT) images. Vibrations at the excitation frequencies of 2– 5 Hz, which was within the range of the natural frequency of the human lumbar spine, were applied to the lumbar FE model to perform time-dependent dynamic FE analysis. The excitation frequency, which may induce resonance, was investigated by observing the changing processes of the lum- bar mechanical parameters under diﬀerent vibrations. This frequency was then utilized to predict the speciﬁc numerical value of the natural frequency of the human lumbar spine. 1 where k0 is the initial permeability and e0 is the initial void ratio. 2.3. Prediction on the Natural Frequency of Human Lumbar Spine. 1. Introduction The mechanical properties in the lumbar spine are determined by the solid and porous material parameters [15, 17]. In this study, the axial eﬀective stress and maximum radial strain in the intervertebral disc were used to express the solid mechanical properties. Pore pressure due to com- pression in the intervertebral disc was used to express the porous mechanical properties. The natural frequency of the human lumbar spine is within 2–5 Hz; thus, the excitation frequency was set within 2–5 Hz [9–11]. The interval was set to 0.5 Hz; therefore, the excitation frequencies of the applied vibrations were 2, 2.5, 3, 3.5, 4, 4.5, and 5 Hz. According to the dynamic FE analysis results, the changing processes of axial eﬀective stress, the radial maximum strain, and the pore pressure in the lumbar FE model under vibrations at abovementioned excitation frequencies were observed and investigated. The excitation frequency that caused the greatest changes in the numerical ﬂuctuation amplitudes and the cycles of lumbar mechanical parameters was subsequently identiﬁed, which can be considered as the natural frequency of the established lumbar FE model. 2. Materials and Methods 2.1. Establishment of the Lumbar FE Model. A healthy young volunteer was selected to conduct lumbar spine image scan- ning. Images with a resolution of 1 mm were imported into Mimics software (Materialise, Leuven, Belgium) to reestab- lish the 2D outside surface of the lumbar spine. Then, the 2D surface was converted into a 3D lumbar FE model using tetrahedral element with ABAQUS software (Simulia, Provi- dence, USA). Most of the geometric appearances in the FE model were derived from the images. The soft tissues that cannot be developed during scanning were created and assembled using CATIA software (Simulia, Providence, USA) [15]. The intervertebral disc structure included annu- lus ground substance, nucleus pulposus, and annulus ﬁbro- sus. In particular, the annulus ﬁbrosus was modeled as ﬁber-reinforced composite, which was embedded in the annulus ground substance in eight layers [16, 17]. The estab- lished lumbar FE model was shown in Figure 1. 2.4. Boundary and Loading Conditions. The lower surface of the L3 segment was ﬁxed in all directions. To apply the vibra- tional load, we built a reference point above the lumbar FE model and coupled with the upper surface of L2 segment. The ﬁve degrees of freedom of the reference point other than the loading direction were constrained. According to litera- ture, for a weight of 70 kg, approximately 800 N of load is 3 Applied Bionics and Biomechanics 3 Annulus matrix Nucleus Annulus ﬁbrosus z y x Figure 1: The schematic diagram of the poroelastic lumbar FE model of L2–L3 segments. Annulus matrix Annulus matrix Nucleus Annulus ﬁbrosus Figure 1: The schematic diagram of the poroelastic lumbar FE model of L2–L3 segments. Table 1: The material parameters of the poroelastic lumbar FE model. Elastic formulation Poroelastic formulation References Elastic modulus (MPa) Poisson’s ratio Permeability (m4/Ns) Void ratio Cancellous bone Linear elastic 100 0.2 1e−13 0.4 [15, 19, 23] Cortical bone Linear elastic 10,000 0.3 1e−20 0.02 [15, 19, 23] Annulus ﬁbrosus Linear elastic 357–550 0.3 [22] Annulus ground substance Hyperelastic C10 = 0.315, D = 0.688 9e−16 2.33 [15, 20, 24] Nucleus pulposus Hyperelastic C10 = 0.125, D = 2.475 3e−16 4 [15, 20, 24] Ligament/facet Hyperelastic Fitting from previous experiment [21] pressure predicted by the simulation was consistent with the experimental results. 2. Materials and Methods compressed on the lumbar spinal segments L2–L3 while sit- ting on a chair with backrest, and the vibrational amplitude range is ±10% of the static pressure [25, 26]. Therefore, sinu- soidal vertical vibrations within the range of 720–880 N pres- sure at the excitation frequencies of 2, 2.5, 3, 3.5, 4, 4.5, and 5 Hz were applied at the reference point. The vibrational time was set to 1 hour, and a damping ratio of 0.08 was adopted [11]. The facet contacts were set to frictionless interaction, and other contacts among the vertebral bodies, intervertebral disc, and ligaments were all deﬁned as “TIE” interaction in the ABAQUS software. Healthy disc swelling occurs because of osmotic potential and thus was simulated by a 0.2 MPa boundary pore pressure at the outer surface of the spine [17]. Second, the FE model was validated under vibrational loading condition. The model was loaded at a sinusoidal anterior-posterior displacement of 0.6 mm with 1 Hz excita- tion frequency [28]. The predicted force-displacement curve was consistent with the curve shape of the specimens in the experiment, and the vibrational force with time was also in the range of the experimental results, as shown in Figure 2(b). 3.2. The Eﬀects of Diﬀerent Excitation Frequencies on the Axial Eﬀective Stress. Figure 3 illustrates the eﬀects of diﬀer- ent excitation frequencies on the average axial eﬀective stress in the intervertebral disc with vibrational time (average axial eﬀective stress was obtained by dividing the total value in the axial eﬀective stress of all elements by the total element number in the intervertebral disc, including annulus ground substance and nucleus pulposus). Vibrations at diﬀerent excitation frequencies led to great diﬀerences in the numeri- cal ﬂuctuation amplitude and the numerical changing cycle. When the excitation frequency increased, the numerical ﬂuc- tuation amplitude of axial eﬀective stress initially increased; when the excitation frequency increased to 3.5 Hz, the peak of stress reached the maximum; and when the excitation fre- quency continued to increase, the numerical ﬂuctuation 3. Results 3.1. Lumbar FE Model Validation. First, the FE model was validated under static loading condition. The intradiscal pressure in the lumbar spine under a compressive force of 500 N for 15 min was compared with the in vitro experiment [27]. As shown in Figure 2(a), both the experimental and simulated results showed that the intradiscal pressure increased to the highest point ﬁrst and then slightly decreased. Meanwhile, the maximum value of the intradiscal Applied Bionics and Biomechanics 4 0 0.0 0.2 0.4 0.6 300 Present model curve 600 Time (s) Intradiscal pressure (Mpa) 900 Experimental curve (a) 0.0 0 −50 −100 −150 50 100 0.5 1.0 Test specimen 1 1.5 Time (s) Force (N) 2.0 Test specimen 2 Present model (b) Figure 2: Comparison of the mechanical responses between the established lumbar FE model and the experiments. (a) Static loading condition; (b) vibrational loading condition. 0 0.0 0.2 0.4 0.6 300 600 Time (s) Intradiscal pressure (Mpa) 900 0.0 0 −50 −100 −150 50 100 0.5 1.0 1.5 Time (s) Force (N) 2.0 Force (N) (a) Figure 2: Comparison of the mechanical responses between the established lumbar FE model and the experiments. (a) Static loading condition; (b) vibrational loading condition. 0 −0.12 −0.16 −0.04 −0.08 0.00 900 1800 2 Hz 2700 Time (s) Axial effective stress (MPa) 3600 2.5 Hz 3 Hz 3.5 Hz (a) 0 −0.12 −0.16 −0.04 −0.08 0.00 900 1800 3.5 Hz 2700 Time (s) Axial effective stress (MPa) 3600 4 Hz 4.5 Hz 5 Hz (b) Figure 3: The eﬀects of vibrations at diﬀerent excitation frequencies on the average axial eﬀective stress in the intervertebral disc. 0 −0.12 −0.16 −0.04 −0.08 0.00 900 1800 2700 Time (s) Axial effective stress (MPa) 3600 0 −0.12 −0.16 −0.04 −0.08 0.00 900 1800 2700 Time (s) Axial effective stress (MPa) 3600 (b) (a) cts of vibrations at diﬀerent excitation frequencies on the average axial eﬀective stress in the intervertebral disc. Figure 3: The eﬀects of vibrations at diﬀerent excitation frequencies on the average axial eﬀective stress in t amplitude and the maximum axial eﬀective stress decreased. The 3.5 Hz vibrational curve expressed only one complete numerical changing cycle in an hour, and all the numerical changing cycles of axial eﬀective stress in the remaining curves were shorter than those in the 3.5 Hz vibration. 3. Results Figure 4: The eﬀects of vibrations at diﬀerent excitation frequencies on the average pore pressure in the of pore pressure value between 3.5 Hz and other excitation frequency vibrations. was investigated by observing the changing processes of the lumbar mechanical parameters under diﬀerent vibrations. This frequency was then used to predict the speciﬁc numerical value of the natural frequency of the human lumbar spine. 3.4. The Eﬀects of Diﬀerent Excitation Frequencies on the Radial Strain. Figure 5 shows the eﬀects of diﬀerent excita- tion frequencies on the maximum radial strain in the inter- vertebral disc with vibrational time. At the initial stage of loading, all the strain curves elevated linearly, reached a high value in a short time, and ﬁnally slowed down. When the excitation frequency increased, the numerical ﬂuctua- tion amplitude and the maximum radial strain gradually increased; when the excitation frequency increased to 3.5 Hz, the peak in the strain reached the maximum; and when the excitation frequency continued to increase, the numerical ﬂuctuation amplitude and the peak gradually decreased. The 3.5 Hz vibrational curve expressed only one complete numerical changing cycle in an hour; all the numerical changing cycles of the radial strain in the remaining curves were shorter than those in the 3.5 Hz vibration. The radial strain varied less with excitation fre- quency, and the diﬀerences in the value of strain generated by vibrations at diﬀerent excitation frequencies were smaller than the changes in the values of the axial eﬀective stress and pore pressure. The changing processes of the axial eﬀective stress, max- imum radial strain, and pore pressure in the intervertebral disc under vibrations at diﬀerent excitation frequencies were observed. As shown in Figures 3–5, although the three mechanical parameters showed diﬀerent values, the changing trends with the excitation frequency were similar. Thus, the axial eﬀective stress was used as an example to discuss the vibrational change process. When the excitation frequency increased from 2 Hz to 3.5 Hz, the numerical ﬂuctuation amplitudes in the vibrational curves increased and reached the maximum level at 3.5 Hz. When the excitation frequency increased from 3.5 Hz to 5 Hz, the numerical ﬂuctuation amplitudes decreased. Meanwhile, the peaks of axial eﬀective stress under vibrations at 2, 4.5, and 5 Hz were at the mini- mum level within 0.117–0.119 MPa. Under vibrations at 2.5, 3, and 4 Hz, the peaks were at the moderate level within 0.121–0.123 MPa. 3. Results values under 2, 4.5, and 5 Hz vibrations were at the maxi- mum level; the values under 2.5, 3, and 4 Hz vibrations were centered; and the value caused by 3.5 Hz vibration was at the minimum level. The shapes of the pore pressure curve showed that vibration at a relatively low excitation frequency produced a long numerical changing cycle, and the shortest numerical changing cycles appeared at the 5 Hz vibrational curve. The pore pressure curve of 3.5 Hz vibration was diﬀer- ent from other curves. Although the pore pressure at 3.5 Hz vibration changed periodically, the numerical changing cycle was long and only showed one complete cycle. Meanwhile, the diﬀerence between the maximum and minimum values of the pore pressure in this cycle was less than that in other curves, and the pore pressure continued to decrease with periodic change, thus possibly leading to an increasing gap 3.3. The Eﬀects of Diﬀerent Excitation Frequencies on the Pore Pressure. Figure 4 shows the eﬀects of diﬀerent excitation fre- quencies on the average pore pressure in the intervertebral disc with vibrational time (average pore pressure in the inter- vertebral disc was obtained by dividing the total value in the pore pressure of all elements by the total element number in the intervertebral disc, including the annulus ground sub- stance and nucleus pulposus). The pore pressure under dif- ferent vibrations could be divided into three levels: the 5 Applied Bionics and Biomechanics 5 0 0.0 0.1 0.3 0.2 0.4 900 2 Hz 1800 2700 Time (s) Pore pressure (MPa) 3600 3.5 Hz 3 Hz 2.5 Hz (a) 0 0.0 0.1 0.3 0.2 0.4 900 3.5 Hz 1800 2700 Time (s) Pore pressure (MPa) 3600 5 Hz 4.5 Hz 4 Hz (b) Figure 4: The eﬀects of vibrations at diﬀerent excitation frequencies on the average pore pressure in the intervertebral disc. 0 0.0 0.1 0.3 0.2 0.4 900 1800 2700 Time (s) Pore pressure (MPa) 3600 0 0.0 0.1 0.3 0.2 0.4 900 1800 2700 Time (s) Pore pressure (MPa) 3600 (a) (b) Figure 4: The eﬀects of vibrations at diﬀerent excitation frequencies on the average pore pressure in the intervertebral disc. cts of vibrations at diﬀerent excitation frequencies on the average pore pressure in the intervertebral disc. 3. Results Under vibrations at 3.5 Hz, the peak was at the maximum level at 0.129 MPa. Furthermore, the longest numerical changing cycle occurred at 3.5 Hz, and the cycles at 3 and 4 Hz were longer than those at 2.5, 4.5, and 5 Hz. Combining these numerical ﬂuctuations and the resonant phenomenon, resonance may occur under the vibration at 3.5 Hz. Thus, the natural frequency of the established lumbar FE model should be 3.5 Hz. 4. Discussion Therefore, in these cases, the axial eﬀective stress and radial strain greatly increased, and the pore pressure remarkably decreased, possibly leading to lumbar injury and degeneration. This phenomenon explained why resonant vibration seriously aﬀects the human lumbar spine from the changes in the numerical ﬂuctuation amplitudes. ﬂuctuation amplitude; when the vibrational cycle prolonged, the damping capacity and eﬀect weakened due to the exuding ﬂuid in the disc in one cycle, resulting in great stress ﬂuctua- tion amplitude [4, 9]. Given that the resonant curve expressed one complete cycle in an hour, the damping eﬀect was rather weak in the middle and late stages of loading, causing the stress to increase signiﬁcantly over time in one cycle. There- fore, the occurrence of resonance may prolong the numerical changing cycle and lead to great numerical ﬂuctuation ampli- tude in one cycle, ultimately changing the lumbar mechanical parameters to be detrimental to the lumbar spine health. This phenomenon explained why resonant vibration seriously aﬀects the human lumbar spine from the changes in the numerical ﬂuctuation cycle. This paper has several limitations in the process of dynamic FE analysis. The boundary condition of the lumbar FE model was simpliﬁed. First, a ﬁxed osmotic pore pressure was set at the outer surface of the spine. As the lumbar spine was compressed and the ﬂuid in the intervertebral disc was exuded, the density of the ﬁxed charges within the nucleus increased, causing changes in the osmotic pore pressure with interstitial ﬂuids surrounding the disc. However, Galbusera et al. concluded that the diﬀerences between the ﬁxed and altered osmotic pore pressure in the predicted results were minimal [31]. Second, the constraint condition for a person in sitting position under vibration was simpliﬁed, which may exert certain eﬀects on the predicted results. The conclu- sions in this study were obtained by comparing the mechan- ical responses of diﬀerent vibration frequencies. The eﬀects of ﬁxed osmotic pore pressure and simpliﬁed constraint condition were considered for all analyses. Therefore, the conclusions may not be signiﬁcantly inﬂuenced by the sim- pliﬁed boundary condition. Except for the boundary condi- tion, only one lumbar FE model was established, indicating that the predicted absolute values may not be representative for an average person. This study aimed at presenting an approach to predict natural frequency; thus, our lumbar FE The shapes of the resonant curves clearly diﬀered from those of the nonresonant curves. 4. Discussion The excitation frequency generated by operating transpor- tation, such as cars, can be controlled through optimization design. However, the speciﬁc numerical value of the natural frequency of the human lumbar spine must be determined rather than merely identifying the range. Therefore, this study aimed at predicting the speciﬁc numerical value of the natural frequency of the human lumbar spine according to the phenomenon of resonance. Vibrations at the excita- tion frequencies of 2–5 Hz were applied to the lumbar FE model. The excitation frequency, which induced resonance, High axial eﬀective stress increases the bearing burden of the annulus ground substance. Pore pressure can resist the external pressure in the intervertebral disc. If pore pressure dissipates quickly, most of the external pressure may be toler- ated by the solid skeleton. Excessive radial strain in the annu- lus ground substance causes serious deformation of the annulus ﬁbrosus. Under the combined inﬂuences of the above three conditions, nucleus pulposus may break through 6 Applied Bionics and Biomechanics 0 0.05 0.00 0.15 0.10 0.20 900 1800 2 Hz 2700 Time (s) Radial strain 3600 2.5 Hz 3 Hz 3.5 Hz (a) 0 0.05 0.00 0.15 0.10 0.20 900 1800 3.5 Hz 2700 Time (s) Radial strain 3600 4 Hz 4.5 Hz 5 Hz (b) Figure 5: The eﬀects of vibrations at diﬀerent excitation frequencies on the maximum radial strain in the intervertebral disc. 0 0.05 0.00 0.15 0.10 0.20 900 1800 2700 Time (s) Radial strain 3600 0 0.05 0.00 0.15 0.10 0.20 900 1800 2700 Time (s) Radial strain 3600 Radial strain (b) (a) Figure 5: The eﬀects of vibrations at diﬀerent excitation frequencies on the maximum radial strain in the intervertebral disc. ﬀects of vibrations at diﬀerent excitation frequencies on the maximum radial strain in the intervertebral disc. Figure 5: The eﬀects of vibrations at diﬀerent excitation frequencies on the maximum radial strain in the int the injured annulus ﬁbrosus, damage the intervertebral disc, and lead to lumbar degeneration in long-term eﬀects [29, 30]. Herein, the axial eﬀective stress, pore pressure, and radial strain peaked at 0.117 MPa, 0.322 MPa, and 14.892%, respectively, when the excitation frequency was far from the natural frequency, such as 2 Hz, 0.129 MPa, 0.237 MPa, and 16.094%, respectively, under resonant frequency. Mean- while, the numerical ﬂuctuation amplitudes of the lumbar mechanical parameters were great when the excitation fre- quency was closed to the natural frequency. Conflicts of Interest [13] S. Nachum and E. Altus, “Natural frequencies and mode shapes of deterministic and stochastic non-homogeneous rods and beams,” Journal of Sound and Vibration, vol. 302, no. 4-5, pp. 903–924, 2007. The authors declare that they have no conﬂicts of interest. 5. Conclusions This study presented an approach to predict the speciﬁc numerical value of the natural frequency of the human lum- bar spine through dynamic FE analysis. Results showed that the natural frequency of the established lumbar FE model was 3.5 Hz. Meanwhile, the reasons of the resonance causing remarkable changes in the lumbar mechanical parameters were explained by comparing and analyzing the diﬀerences between the resonant and nonresonant vibrational curves and data. [9] J. T. M. Cheung, M. Zhang, and D. H. K. Chow, “Biome- chanical responses of the intervertebral joints to static and vibrational loading: a ﬁnite element study,” Clinical Biome- chanics, vol. 18, no. 9, pp. 790–799, 2003. [10] W. Z. Kong and V. K. Goel, “Ability of the ﬁnite element models to predict response of the human spine to sinusoidal vertical vibration,” Spine, vol. 28, no. 17, pp. 1961–1967, 2003. [11] W. Fan and L. X. Guo, “Inﬂuence of diﬀerent frequen- cies of axial cyclic loading on time-domain vibration response of the lumbar spine: a ﬁnite element study,” Computers in Biology and Medicine, vol. 86, no. 1, pp. 75–81, 2017. Data Availability The data used to support the ﬁndings of this study are available from the corresponding author upon request. [12] L.-X. Guo, M. Zhang, Y.-M. Zhang, and E.-C. Teo, “Vibration modes of injured spine at resonant frequencies under vertical vibration,” Spine, vol. 34, no. 19, pp. E682–E688, 2009. Applied Bionics and Biomechanics Applied Bionics and Biomechanics of Occupational and Environmental Health, vol. 91, no. 6, pp. 689–694, 2018. of Occupational and Environmental Health, vol. 91, no. 6, pp. 689–694, 2018. model was used as an example to introduce this approach. Although the dynamic FE analysis contained assumptions, model validation ensured the accuracy of the FE model. Hence, accurate results and meaningful conclusions could be obtained from the investigation on the basis of precise modeling and analytical method. [4] [4] L. X. Guo, E. C. Teo, K. K. Lee, and Q. H. Zhang, “Vibration characteristics of the human spine under axial cyclic loads: eﬀect of frequency and damping,” Spine, vol. 30, no. 6, pp. 631–637, 2005. [5] L. X. Guo and E. C. Teo, “Prediction of the modal characteris- tics of the human spine at resonant frequency using ﬁnite ele- ment models,” Proceedings of the Institution of Mechanical Engineers, Part H: Journal of Engineering in Medicine, vol. 219, no. 4, pp. 277–284, 2005. The innovation of this approach was to predict the spe- ciﬁc numerical value of the natural frequency of the human lumbar spine by using the porous seepage-stress coupled FE analysis [32]. Initially, a real time-dependent vibrational condition was simulated, and the changes in lumbar mechanical parameters with time were investigated. Mean- while, the porous mechanical parameters in the lumbar spine can be predicted to analyze changes in the internal ﬂuid properties under vibration. Accordingly, the porous seepage-stress coupled FE analysis might be applicable to predict the speciﬁc numerical value of the natural frequency of a person who has CT images with either healthy or degenerated lumbar spine. [6] W. Wang, B. Bazrgari, A. Shirazi-Adl, S. Rakheja, and P.-É. Boileau, “Biodynamic response and spinal load estima- tion of seated body in vibration using ﬁnite element model- ing,” Industrial Health, vol. 48, no. 5, pp. 557–564, 2010. [7] M. H. Pope, D. G. Wilder, L. Jorneus, H. Broman, M. Svensson, and G. Andersson, “The response of the seated human to sinusoidal vibration and impact,” Journal of Biome- chanical Engineering, vol. 109, no. 4, pp. 279–284, 1987. [8] M. M. Panjabi, G. B. Andersson, L. Jorneus, E. Hult, and L. Mattsson, “In vivo measurements of spinal column vibra- tions,” Journal of Bone and Joint Surgery, vol. 68, no. 5, pp. 695–702, 1986. Acknowledgments [14] M. Malekian, D. Trieu, J. S. Owoc, S. S. Park, and C. J. Hunter, “Investigation of the intervertebral disc and fused joint dynamics through experimental modal analysis and the recep- tance coupling method,” Journal of Biomechanical Engineer- ing, vol. 132, no. 4, article 041004, 2010. This work was supported by the Education Science and Technology Project of Jilin Province (JJKH20180560KJ). 4. Discussion The three lumbar mechan- ical parameters in the resonant curves only expressed one complete numerical changing cycle in an hour. For the exci- tation frequency away from the natural frequency, such as 2, 4.5, and 5 Hz, the numerical changing cycle gradually shortened as the excitation frequency increased; when the excitation frequency was close to the natural frequency, the numerical changing cycle prolonged. Meanwhile, the changes in numerical ﬂuctuation cycle under various excita- tion frequencies led to diﬀerent numerical ﬂuctuation amplitudes in one cycle. Here, we considered the axial eﬀec- tive stress in the last cycle as an example. The stress ﬂuctua- tion amplitude was 0.022 MPa at 2 Hz vibration in the last cycle, 0.024 MPa at 2.5 Hz, 0.026 MPa at 3 Hz, 0.035 MPa at 3.5 Hz, 0.028 MPa at 4 Hz, 0.020 MPa at 4.5 Hz, and 0.019 MPa at 5 Hz. The stress ﬂuctuation amplitude in one cycle decreased as the cycle shortened. This phenomenon may be related to the continuous damping capacity and time inside the lumbar intervertebral disc. The damping eﬀect was evident in a short cycle, directly leading to low stress 7 References [15] R. Fan, H. Gong, S. Qiu, X. Zhang, J. Fang, and D. Zhu, “Eﬀects of resting modes on human lumbar spines with diﬀerent levels of degenerated intervertebral discs: a ﬁnite element investiga- tion,” BMC Musculoskeletal Disorders, vol. 16, no. 1, p. 221, 2015. [1] P. Strube, M. Pumberger, L. Sonnow et al., “Association between lumbar spinal degeneration and anatomic pelvic parameters,” Clinical Spine Surgery, vol. 31, no. 6, pp. 263– 267, 2018. [16] H. Schmidt, A. Kettler, F. Heuer, U. Simon, L. Claes, and H. J. Wilke, “Intradiscal pressure, shear strain, and ﬁber strain in the intervertebral disc under combined loading,” Spine, vol. 32, no. 7, pp. 748–755, 2007. [2] M. R. McCann, M. A. Veras, C. Yeung et al., “Whole-body vibration of mice induces progressive degeneration of inter- vertebral discs associated with increased expression of Il-1β and multiple matrix degrading enzymes,” Osteoarthritis and Cartilage, vol. 25, no. 5, pp. 779–789, 2017. [17] H. Schmidt, A. Shirazi-Adl, F. Galbusera, and H. J. Wilke, “Response analysis of the lumbar spine during regular daily activities-a ﬁnite element analysis,” Journal of Biomechanics, vol. 43, no. 10, pp. 1849–1856, 2010. [3] J. Wahlstrom, L. Burstrom, P. W. Johnson, T. Nilsson, and B. Jarvholm, “Exposure to whole-body vibration and hospital- ization due to lumbar disc herniation,” International Archives Applied Bionics and Biomechanics 8 [18] M. Argoubi and A. Shirazi-Adl, “Poroelastic creep response analysis of a lumbar motion segment in compression,” Journal of Biomechanics, vol. 29, no. 10, pp. 1331–1339, 1996. [19] H. Schmidt, M. Bashkuev, F. Galbusera, H. J. Wilke, and A. Shirazi-Adl, “Finite element study of human lumbar disc nucleus replacements,” Computer Methods in Biomechanics and Biomedical Engineering, vol. 17, no. 16, pp. 1762–1776, 2014. [20] H. Schmidt, M. Bashkuev, M. Dreischarf et al., “Computa- tional biomechanics of a lumbar motion segment in pure and combined shear loads,” Journal of Biomechanics, vol. 46, no. 14, pp. 2513–2521, 2013. [21] M. Sharma, N. A. Langrana, and J. Rodriguez, “Role of liga- ments and facets in lumbar spinal stability,” Spine, vol. 20, no. 8, pp. 887–900, 1995. [22] H. Xu, H. Tang, X. Guan et al., “Biomechanical comparison of posterior lumbar interbody fusion and transforaminal lumbar interbody fusion by ﬁnite element analysis,” Neurosurgery, vol. 72, no. 1, pp. ons 21–ons 26, 2013. [23] S. Zanjani-Pour, C. P. Winlove, C. W. Smith, and J. R. References Meakin, “Image driven subject–speciﬁc ﬁnite element models of spinal biomechanics,” Journal of Biomechanics, vol. 49, no. 6, pp. 919–925, 2016. [24] L. X. Guo, R. Li, and M. Zhang, “Biomechanical and ﬂuid ﬂow- ing characteristics of intervertebral disc of lumbar spine pre- dicted by poroelastic ﬁnite element method,” Acta of Bioengineering and Biomechanics, vol. 18, no. 2, pp. 19–29, 2016. [25] M. M. Verver, J. van Hoof, C. W. J. Oomens, N. van de Wouw, and J. S. H. M. Wismans, “Estimation of spinal loading in vertical vibrations by numerical simulation,” Clinical biome- chanics, vol. 18, no. 9, pp. 800–811, 2003. [26] L.-X. Guo and W. Fan, “Dynamic response of the lumbar spine to whole-body vibration under a compressive follower pre- load,” Spine, vol. 43, no. 3, pp. E143–E153, 2018. [27] F. Heuer, H. Schmitt, H. Schmidt, L. Claes, and H. J. Wilke, “Creep associated changes in intervertebral disc bulging obtained with a laser scanning device,” Clinical Biomechanics, vol. 22, no. 7, pp. 737–744, 2007. [28] I. A. Stokes, M. Gardner-Morse, D. Churchill, and J. P. Laible, “Measurement of a spinal motion segment stiﬀness matrix,” Journal of Biomechanics, vol. 35, no. 4, pp. 517–521, 2002. [29] F. Galbusera, H. Schmidt, C. Neidlinger-Wilke, A. Gottschalk, and H. J. Wilke, “The mechanical response of the lumbar spine to diﬀerent combinations of disc degenerative changes investi- gated using randomized poroelastic ﬁnite element models,” European Spine Journal, vol. 20, no. 4, pp. 563–571, 2011. [30] M. A. Adams and P. J. Roughley, “What is intervertebral disc degeneration, and what causes it?,” Spine, vol. 31, no. 18, pp. 2151–2161, 2006. [31] F. Galbusera, H. Schmidt, J. Noailly et al., “Comparison of four methods to simulate swelling in poroelastic ﬁnite element models of intervertebral discs,” Journal of the Mechanical Behavior of Biomedical Materials, vol. 4, no. 7, pp. 1234– 1241, 2011. [32] J. Zheng, L. Tang, and J. Hu, “A numerical investigation of risk factors aﬀecting lumbar spine injuries using a detailed lumbar model,” Applied Bionics and Biomechanics, vol. 2018, no. 8626102, 8 pages, 2018.
W1496441967.txt	https://newprairiepress.org/cgi/viewcontent.cgi?article=1947&context=kaesrr	en		Economics of sealing horizontal silos	Kansas Agricultural Experiment Station research reports	1,997	cc-by	1,473	Kansas Agricultural Experiment Station Research Reports Volume 0 Issue 1 Cattleman's Day (1993-2014) Article 544 1997 Economics of sealing horizontal silos G.L. Huck J.E. Turner M.K. Siefers See next page for additional authors Follow this and additional works at: https://newprairiepress.org/kaesrr Part of the Other Animal Sciences Commons Recommended Citation Huck, G.L.; Turner, J.E.; Siefers, M.K.; Brent, B.E.; Bolsen, K.K.; Young, Matthew A.; and Pope, Ronald V. (1997) "Economics of sealing horizontal silos," Kansas Agricultural Experiment Station Research Reports: Vol. 0: Iss. 1. https://doi.org/10.4148/2378-5977.1947 This report is brought to you for free and open access by New Prairie Press. It has been accepted for inclusion in Kansas Agricultural Experiment Station Research Reports by an authorized administrator of New Prairie Press. Copyright 1997 the Author(s). Contents of this publication may be freely reproduced for educational purposes. All other rights reserved. Brand names appearing in this publication are for product identification purposes only. No endorsement is intended, nor is criticism implied of similar products not mentioned. K-State Research and Extension is an equal opportunity provider and employer. Economics of sealing horizontal silos Authors G.L. Huck, J.E. Turner, M.K. Siefers, B.E. Brent, K.K. Bolsen, Matthew A. Young, and Ronald V. Pope This research report is available in Kansas Agricultural Experiment Station Research Reports: https://newprairiepress.org/kaesrr/vol0/iss1/544 Cattlemen's Day 1997 ECONOMICS OF SEALING HORIZONTAL SILOS G. L. Huck, J. E. Turner, M. K. Siefers, M. A. Young, R. V. Pope, B. E. Brent, and K. K. Bolsen Summary losses, it is so awkward, cumb ersome, and labor intensive that many producers feel the silage saved is not worth their time and effort. Determinin g the value of silage saved by effectively sealing a hori zontal silo requires only a few simple calculations, bu tit is still a concept that is often overlooked by many livestock producers who store large amounts of silage in that manner. Kansas produces about 3.0 million tons of silage annually, primarily from corn and sorghum. A majority of thi ssilage is made and stored in either bunker, trench, or "drive-over" pile silos . Only 20 to 30% of these silos are sealed after filling. Producers who do not seal need to take a second look at the economics of this highly troublesome "technology" before they rej ect it as unnecessary and uneconomical. The loss from a 100 × 250 ft silo filled with corn silage can exceed $10,000. Top-spoilag e research ha s been conducted at Kansas State University since 1989, and the results d ocument the magnitude of the DM and nutrient losses in the original top 3 ft of the ensiled crop. However, these losses can not be seen until the silo is opened. Even then, the spoilag e might be apparen t only in the top 6 to 12 inches of silage, obscuring the fact that this area of spoiled silage represents substantially more silage as originally stored. (Key Words: Silage, Top Sp oilage, Silo, Bunker Silo, Trench Silo, Pile Silo.) We provide here a few simple equations, that can be hand-calc ulated or incorporated into a computer spreadsheet. They allow producers to estim ate the value of silage saved by sealing, based on their crop value, silo dimensions, cost of the sealing material, and labor to cover their silage. Introduction Calculations and Examples Three economically attractive methods in Kansas for storing large amounts of ensiled forage are the horizontal silos (i.e., bunker, trench, or pile), but because so much of the surface of the ensiled material is exposed, dry matter (DM) and nutrient losses can be extensive . If left unprotected, losses in the top 2 to 4 ft can exceed 50%. This is particularly disturbing when one consid ers that in the typical horizontal silo, over 20% of the silage might be within the top 4 feet. Calculating the value of silage saved by sealing is based on four economic inputs and two silo/silage inputs. The four economic inputs are: These losses can be minimized by sealing (covering ) the ensiled mass with polyethylene sheets, which usually are weighted with tires or soil. Although this method minimizes Ten hours per 4,000 ft 2 of polyethylene sheet were used to calculate the labor cost. 1) Value of the silage ($/ton) 2) Cost of the polyethylene sheet (cents/ft 2 × number of f t2) 3) Cost of the weighting material (zero was used in the examples) 4) labor cost ($/hr × number of hrs). 84 In order to accoun t for overlapping from sheet to sheet and along the side walls or base, we assumed a covering efficiency of 80%. Loss, unsealed: 270 tons × $25/ton × 50% = $3,375 Loss, sealed: 270 tons × $25/ton × 20% = $1,350 Cost of sealin g = $ 420 Net, seale d = $1,770 Net return to sealing: $3,375 – $1,770 = $1,605 The first of the two silo/silage inputs determine s the amount of silage within the origina l top 3 ft of the silo after filling is complete . It is determined by multiplying the silo width(ft) by length(ft) by depth of interest (3 ft) by the silage density (lb/f 3t) and dividing the product by 2,000 (lb/ton). The concepts sh own above are presented in a user-friendly spreadsheet format in Table 1. The first nine lines are economic inputs determined by the producer, and the next six lines are results that are based on formulas utilizing the producer's inputs. They can be programme d easily into the spreadsheet using the row letters as guides. The second silo/silage input estimates the amount of silage within the original top 3 ft of the silo that is lost as spoilage. These values (50% of sealed , 20% if unsealed) are based on researc h conducted at Kansas State University and published in KAES Reports of Progress 623, p. 70; 651, p. 127; and 727, p. 59 and 63. The most important single facto rinfluencing preservation efficiency of ensiled forages is the degree of anaerobic fermentation achieved durin g ensiling. When silage is not sealed or when the seal is inadequate, air and moisture enter the mass and affect both the ensiling process and silage quality durin gthe storage and feedo ut phases. Based on the examples in Table 1, sealing a 40 ft × 100 ft silo could save approximatel y $1,600 worth of silage. Using the same concept, covering a 10 0ft × 400 ft silo could save the producer over $16,000. The following example estimates the net return from sealin g a horizontal silo 40 ft wide 2 by 100 ft long (4,000 f t). Economic assumptions : 1) Corn silage price: $25/ton 2) Polyethylen e film: $.055 per ft 2 of surface covered. $.055 × 4,000 ft 2 = $220 3) Weighting m aterial: zero cost assumed 4) Labor cost: 10 hr/4,000 ft 2 sheet × $20/hr = $200 Sealing cost = $220 + $200 = $420 Although future technolo gy might introduce a more environmentally and user-friendly product , polyethylene (6 mm) is the most effectiv e sealing material available today . The most common sealing method is to place the polyethylen e sheet over the ensiled forage and weight it down with rubber tires (20 to 25 tires per 100 sq ft). Silo/silage assumptions : 1) Assuming a silage density of 45 lb/ft 3 (4000 ft2 surface × 3 ft deep × 45 lb/ft3)/2000 = 270 t ons of silage within the original top 3 ft (total capacity of the silo is about 1,080 tons) Research-base d calculations confirm that the financial loss incurred b ynot sealing silage is substantial and reinforces our recommendation that sealing the exposed surface of a horiz ontal silo is one of the most important management decisions in any silage program. 2) Assum e 20% loss in the top 3 feet if sealed, 50% loss if unsealed. 85 Table 1. Value of Silage Saved by Sealing Three Horizontal Silos Differing in Size Economic inputs Silage crop Corn Corn Corn 25 25 25 A 45 45 45 B Silo width, ft 40 100 100 C Silo length, ft 100 250 400 D Cost of 40 ft × 100 ft poly sheet, $ 175 175 175 E Efficiency of sheet, % 80 80 80 F Silage lost if unsealed, % 50 50 50 G Silage lost if sealed, % 20 20 20 H Labor cost, $/hr 20 20 20 I 270 1,688 2,700 J Silage value lost if unsealed, $ 3,375 21,094 33,750 K J×(G/100)×A Silage value lost if sealed, $ 1,350 8,438 13,500 L J×(H/100)×A Silage value, $/ton Silage density, lb/ft 3 Spreadsheet Formulas Results Silage in the top 3 ft, tons Cost per ft 2 of poly sheet, ¢ Sealing cost, $ Value of silage saved, $ 5.5 5.5 5.5 419 2,617 4,188 1,606 10,039 16,063 86 (C×D×3×B)/2000 M ([E/(F/100)]/4000)×100 N [(C×D×M)/100)]+ [(I×C×D×10)/4000] P K–(L+N)
https://openalex.org/W4308494612	https://link.springer.com/content/pdf/10.1007/s10896-022-00451-0.pdf	English	null	Childhood Predictors of Adult Intimate Partner Violence Perpetration and Victimization	Journal of family violence	2,022	cc-by	12,813	Ashlee Curtis1 · Travis Harries1 · Bengianni Pizzirani1 · Shannon Hyder1 · Ryan Baldwin1 · Richelle Mayshak1 · Arlene Walker1 · John W Toumbourou1 · Peter Miller1 Accepted: 2 October 2022 / Published online: 3 November 2022 © The Author(s) 2022 Journal of Family Violence (2023) 38:1591–1606 https://doi.org/10.1007/s10896-022-00451-0 Journal of Family Violence (2023) 38:1591–1606 https://doi.org/10.1007/s10896-022-00451-0 REVIEW ARTICLE Richelle.mayshak@deakin.edu.au Arlene Walker arlene.walker@deakin.edu.au John W Toumbourou john.toumbourou@deakin.edu.au Peter Miller peter.miller@deakin.edu.au 1 School of Psychology, Deakin University, Geelong, VIC, Australia Childhood Predictors of Adult Intimate Partner Violence Perpetration and Victimization Ashlee Curtis1 · Travis Harries1 · Bengianni Pizzirani1 · Shannon Hyder1 · Ryan Baldwin1 · Richelle Mayshak1 · Arlene Walker1 · John W Toumbourou1 · Peter Miller1 Eligibility criteria Couched within a developmental perspective, this review focused on risk factors identified in childhood and adoles cence that contribute to subsequent IPV experienced in the context of adult romantic relationships. To be included for review, studies were required to report IPV within current or recent romantic relationships in adulthood. Thus, studies that assessed the occurrence of at least one type of adult IPV (e.g., sexual, emotional, psychological, or physical), via a single item or scale score (self-reported or partner-reported) were included for review. Specifically, in line with Costa et al., articles were included in the review if they met the fol lowing inclusion criteria: (a) were published in English from January 2015 to November 2020; (b) were peer reviewed; (c) described a longitudinal study (i.e., minimum two waves of data collection; where the first data collection occurred prior to age 18, and at least one follow-up assessment was made in adulthood); (d) empirically tested the strength of one or more childhood/adolescent predictors of IPV perpe tration and/or victimization in adulthood; and (e) the full text was accessible. No limits were placed on sample size or recruitment methodology. Understanding the key developmental factors and influ ences on experiences of IPV in adulthood is critical to ensure effective prevention efforts (for a summary of this litera ture, see Costa et al., 2015). In 2015, Costa et al. identified 25 studies within five domains in their systematic review focusing on longitudinal predictors of IPV perpetration and victimization. The first domain, experiencing abuse in child hood and adolescence, was associated with an increased likelihood of perpetrating physical abuse and psychologi cal aggression and experiencing victimization in an adult romantic relationship. The second domain, experiences of adversity in the family of origin (e.g., poor parent-child rela tionship; witnessing parental violence), increased the likeli hood of both IPV perpetration and victimization as an adult. Similarly, the third domain, behavioral risks (e.g., aggres sive behavior; alcohol and drug use) increased the likeli hood of IPV perpetration and victimization in adulthood. The fourth domain, adolescent peer risks (e.g., conflict with peers), predicted IPV perpetration and victimization as an adult. The last domain, sociodemographic risks, including family of origin and low socioeconomic status, were also found to be consistent predictors of both physical and psy chological IPV perpetration and victimization. Abstract Purpose This systematic review is a 5-year update of a previously conducted review on the longitudinal predictors of domestic violence perpetration and victimization. This review adopted the term ‘Intimate Partner Violence (IPV)’ to align with current literature and addressed two aims: to identify any novel longitudinal risk factors since the previous review, and to determine if a distinction could be drawn between risk factors for perpetration and victimization (a limitation identified by the previous review). Methods Twelve studies met the inclusion criteria of prospectively investigating childhood/adolescent predictors (prior to age 18) for adulthood IPV perpetration and victimization. Peer-reviewed papers were identified via the following databases in November 2020: MEDLINE, APA PsycINFO, SocINDEX, EMBASE, and Scopus. Study quality was assessed using the Cambridge Quality Checklists. Results Consistent with the previous review, child and adolescent abuse, family of origin risks, child and adolescent behav ioral problems, and adolescent peer risks were identified as significant predictors of IPV perpetration and victimization. The current review, however, adds nuance to these findings, identifying potential moderating and/or mediating factors and additional risk factors, including mental health and cultural and attitudinal risks. Conclusion This review re-emphasizes the importance of developmental risk factors for adulthood IPV perpetration and victimization, and their role in prevention and intervention efforts. Keywords Intimate partner violence · longitudinal predictors · systematic review · intervention · prevention Intimate Partner Violence (IPV) impacts approximately 16% of individuals in population-based studies (Langhin richsen-Rohling et al., 2012) and is associated with a range of negative physical health outcomes (e.g., injury, chronic pain, and gastrointestinal problems), as well as negative mental health outcomes (e.g., substance use and depressive symptoms) (Campbell, 2002; Coker et al., 2002). Despite these well-established adverse outcomes, there remains a lack of evidence for effective prevention and intervention options (Graham et al., 2021; Karakurt et al., 2019). Abstract This Ashlee Curtis ashlee.curtis@deakin.edu.au Travis Harries travis.harries@deakin.edu.au Bengianni Pizzirani bengianni.pizzirani@deakin.edu.au Shannon Hyder Shannon.hyder@deakin.edu.au Ryan Baldwin ryan.baldwin@deakin.edu.au Richelle Mayshak Ashlee Curtis ashlee.curtis@deakin.edu.au Travis Harries travis.harries@deakin.edu.au Bengianni Pizzirani bengianni.pizzirani@deakin.edu.au Shannon Hyder Shannon.hyder@deakin.edu.au Ryan Baldwin ryan.baldwin@deakin.edu.au Richelle Mayshak Richelle.mayshak@deakin.edu.au Arlene Walker arlene.walker@deakin.edu.au John W Toumbourou john.toumbourou@deakin.edu.au Peter Miller peter.miller@deakin.edu.au 1 School of Psychology, Deakin University, Geelong, VIC, Australia Richelle.mayshak@deakin.edu.au Arlene Walker arlene.walker@deakin.edu.au John W Toumbourou john.toumbourou@deakin.edu.au Peter Miller peter.miller@deakin.edu.au 1 School of Psychology, Deakin University, Geelong, VIC, Australia Richelle Mayshak 1 3 1592 Journal of Family Violence (2023) 38:1591–1606 relative paucity of evidence likely results from both the complexity of, and interplay between, the multitude of risk and protective factors that have been identified in both IPV perpetration and victimization, as well as a need to better understand the etiology of IPV (Capaldi & Langhinrich sen-Rohling, 2012). Prospective longitudinal studies can identify key developmental patterns, experiences, and influ ences which inform early prevention efforts, or intervention efforts, throughout the life-course.l Information sources Peer-reviewed papers were identified by searching elec tronic databases: MEDLINE, APA PsycINFO, SocINDEX with full text, EMBASE, and Scopus. Costa et al., (2015) highlighted the consistent overlap between IPV victimization and perpetration in adulthood indicating it might be difficult to identify unique predic tors for each, suggesting that this interrelationship may be a result of common risk factors and via reciprocal IPV. As such, the current systematic review provides an update on, and extension of the Costa et al. (2015) systematic review, highlighting recent evidentiary advancements for the role of longitudinal influences on IPV perpetration and victimiza tion. Specifically, this systematic review aimed to identify if, since the Costa et al. review, there were newly identified developmental predictors of IPV perpetration and victim ization that could be used to inform prevention and inter vention efforts. While a significant proportion of IPV is bidirectional (as much as 68% in some population studies), IPV is also often unidirectional (Langhinrichsen-Rohling et al., 2012). Thus, we also aimed to explore whether risk factors for IPV perpetration and victimization can be distin guished from one another. Method A systematic literature review was conducted using the Pre ferred Reporting Items for Systematic Reviews and Meta- Analyses (PRISMA) Checklist (Page et al., 2021). Results Eight studies utilized a version of the Conflict Tactics Scale (Abajobir et al., 2017; Goodnight et al., 2017; Grest, Amaro, et al., 2018; Grest, Lee, et al., 2018; Herrenkohl & Jung, 2016; Narayan et al., 2017; Theobald et al., 2016; Zamir et al., 2018). Two studies utilized the 14-item Conflict Tactics Scale – Short Form (CTS-SF) (Grest, Amaro, et al., 2018; Selection process All relevant titles and abstracts identified through the searches were exported into Covidence (Veritas Health Innovation, 2014), a systematic review support program for screening. The title and abstract of all articles were screened by one author (BP) to assess suitability for inclusion. Author SH independently screened 25% of the titles and abstracts. The inter-rater agreement between the two researchers was 96% and any differences were resolved by Author AC. Author BP conducted screening of full-text articles for inclusion. Author RB independently screened 10% of the full-text articles with no disagreement found. See Fig. 1 for PRISMA flowchart. Table 2 shows the characteristics of the 12 studies included in the current review, including study cohort, pre dictor variables, outcome measures, and predictors. Five studies focused on both perpetration of IPV and victimiza tion by IPV in adulthood (Grest, Amaro, et al., 2018; Grest, Lee, et al., 2018; Handley et al., 2019; Herrenkohl & Jung, 2016; Narayan et al., 2017), four focused on perpetration only (Goodnight et al., 2017; Milaniak & Widom, 2015; Neppl et al., 2019; Theobald et al., 2016), and three on vic timization only (Abajobir et al., 2017; Vézina et al., 2015; Zamir et al., 2018). Nine of the 12 studies were conducted in the USA (Goodnight et al., 2017; Grest, Amaro, et al., 2018; Grest, Lee, et al., 2018; Handley et al., 2019; Herrenkohl & Jung, 2016; Milaniak & Widom, 2015; Narayan et al., 2017; Neppl et al., 2019; Zamir et al., 2018), and one each in the UK (Theobald et al., 2016), Australia (Abajobir et al., 2017), and Canada (Vézina et al., 2015). Study characteristics Table 1 shows the citation, cohort, country, participants, and length of follow up for each study listed by cohort. Nine of the studies collected data from both males and females (Abajobir et al., 2017; Goodnight et al., 2017; Grest, Amaro, et al., 2018; Grest, Lee, et al., 2018; Handley et al., 2019; Herrenkohl & Jung, 2016; Milaniak & Widom, 2015; Nara yan et al., 2017; Neppl et al., 2019), two from only females (Vézina et al., 2015; Zamir et al., 2018), and one from only males (Theobald et al., 2016). The sample sizes within the cohorts ranged from 80 (Zamir et al., 2018) to 3322 (Aba jobir et al., 2017), with an average sample size of 752. The studies were from 10 longitudinal cohorts (see Table 1). Three studies from two cohorts (Narayan et al., 2017; Theo bald et al., 2016; Zamir et al., 2018) were based on different waves of data from cohorts that were included in the Costa et al., (2015) review., and which were published within the timeframe of the current review. Data collection process Data was collected from articles by Author TH and cross- checked by Author AC for accuracy. Data was populated into a summary table (see Table 1). Quality assessment The quality of the studies included in the current review were assessed using the Cambridge Quality Checklists (Jol liffe et al., 2012; Murray et al., 2009). Points were allocated to each study based on three domains: Correlate (5 points; sample/sampling method, response rate, reliability of mea sures), Risk Factor (3 points; study design), and Causal (7 points; attribution of causality; see Supplementary Table 1 for detail). Search strategy All available records were searched using the following combination of search concepts (and their related terms; see Supplementary material 1 for full search syntax) in the title or abstract of the article: ‘intimate partner violence’, ‘interpersonal relationships’ and ‘longitudinal predic tors.’ Each search concept was created by first developing a list of search terms relevant to each concept (e.g., for the ‘intimate partner violence’ concept search terms included ‘interpersonal violence’, ‘domestic abuse’, ‘maltreatment’, etc.), with this process continuing until all keywords for each concept were identified. To create each concept, the search terms were searched collectively using the operator ‘or’. Each concept was then combined into the final search by using the operator ‘and’. Search terms included indexed 1 1 3 Journal of Family Violence (2023) 38:1591–1606 1593 terms unique to each database, as well as terms used in pre vious reviews (e.g., Costa et al., 2015). terms unique to each database, as well as terms used in pre vious reviews (e.g., Costa et al., 2015). met the inclusion criteria. During this process, 3,510 papers did not meet the inclusion criteria and were removed (Fig. 1). The remaining 109 papers were read in full whilst applying the inclusion criteria, at which point a further 97 papers were excluded. The remaining 12 papers were identi fied as relevant to the research aim. To minimize the impact of indexing errors, several jour nals were identified for hand searching based on these jour nals publishing a high volume of IPV studies. These journals included Journal of Family Violence and Journal of Inter personal Violence. Our hand searches did not reveal any further relevant articles for inclusion beyond our database searches, and as such we opted not to conduct any further hand searches in specific journals. Reference lists of eligible studies and review articles were also searched for additional studies that were not captured in the database searches. Study selection After removal of duplicates a total of 3,619 articles were retained, and the abstracts were read to assess if the article 1 1 3 1594 Journal of Family Violence (2023) 38:1591–1606 Fig 1 PRISMA al., 2019), and one study combined two existing IPV mea sures (including the CTS 2) (Vézina et al., 2015). Studies were mixed in terms of the type of IPV they assessed in adulthood. Ten of the 12 studies specifically measured physical IPV (Abajobir et al., 2017; Goodnight et al., 2017; Grest, Amaro, et al., 2018; Grest, Lee, et al., 2018; Herrenkohl & Jung, 2016; Milaniak & Widom, 2015; Narayan et al., 2017; Theobald et al., 2016; Vézina et al., 2015; Zamir et al., 2018), six measured psychological IPV Grest, Lee, et al., 2018), one study used the full CTS (Theo bald et al., 2016), four utilized a modified version of the full CTS, including a 7-item (Abajobir et al., 2017), 8-item (Zamir et al., 2018), 10-item (Narayan et al., 2017), and a 33-item version (Herrenkohl & Jung, 2016), and one study utilized a shortened version of the CTS 2 (15-item (Good night et al., 2017)). Two studies used a one-item measure of IPV in adulthood (Milaniak & Widom, 2015), one study used a 4-item partner and observer report scale (Handley et Fig. 1 PRISMA Fig. 1 PRISMA al., 2019), and one study combined two existing IPV mea sures (including the CTS 2) (Vézina et al., 2015). al., 2019), and one study combined two existing IPV mea sures (including the CTS 2) (Vézina et al., 2015). Studies were mixed in terms of the type of IPV they assessed in adulthood. Ten of the 12 studies specifically measured physical IPV (Abajobir et al., 2017; Goodnight et al., 2017; Grest, Amaro, et al., 2018; Grest, Lee, et al., 2018; Herrenkohl & Jung, 2016; Milaniak & Widom, 2015; Narayan et al., 2017; Theobald et al., 2016; Vézina et al., 2015; Zamir et al., 2018), six measured psychological IPV Grest, Lee, et al., 2018), one study used the full CTS (Theo bald et al., 2016), four utilized a modified version of the full CTS, including a 7-item (Abajobir et al., 2017), 8-item (Zamir et al., 2018), 10-item (Narayan et al., 2017), and a 33-item version (Herrenkohl & Jung, 2016), and one study utilized a shortened version of the CTS 2 (15-item (Good night et al., 2017)). Study selection Two studies used a one-item measure of IPV in adulthood (Milaniak & Widom, 2015), one study used a 4-item partner and observer report scale (Handley et Grest, Lee, et al., 2018), one study used the full CTS (Theo bald et al., 2016), four utilized a modified version of the full CTS, including a 7-item (Abajobir et al., 2017), 8-item (Zamir et al., 2018), 10-item (Narayan et al., 2017), and a 33-item version (Herrenkohl & Jung, 2016), and one study utilized a shortened version of the CTS 2 (15-item (Good night et al., 2017)). Two studies used a one-item measure of IPV in adulthood (Milaniak & Widom, 2015), one study used a 4-item partner and observer report scale (Handley et Studies were mixed in terms of the type of IPV they assessed in adulthood. Ten of the 12 studies specifically measured physical IPV (Abajobir et al., 2017; Goodnight et al., 2017; Grest, Amaro, et al., 2018; Grest, Lee, et al., 2018; Herrenkohl & Jung, 2016; Milaniak & Widom, 2015; Narayan et al., 2017; Theobald et al., 2016; Vézina et al., 2015; Zamir et al., 2018), six measured psychological IPV Studies were mixed in terms of the type of IPV they assessed in adulthood. Ten of the 12 studies specifically measured physical IPV (Abajobir et al., 2017; Goodnight et al., 2017; Grest, Amaro, et al., 2018; Grest, Lee, et al., 2018; Herrenkohl & Jung, 2016; Milaniak & Widom, 2015; Narayan et al., 2017; Theobald et al., 2016; Vézina et al., 2015; Zamir et al., 2018), six measured psychological IPV 1 3 Journal of Family Violence (2023) 38:1591–1606 1595 Table 1 Summary of reviewed studies Study/cohort (country) Participants Length of follow up (number of waves) Waves uin current study Cohort in Costa et al? 1. Cambridge Study in Delinquent Development (CSDD; UK) Theobald et al., (2016) 411 males aged 8 at T1 and fol lowed up to age 48 (n = 365). 40 years (9 waves) 5 waves (age 8–10, age 18, age 32, age 48) Yes 2. Minnesota Longitudinal Study of Risk and Adaptation (MLSRA; USA) Narayan et al., (2017) 93 males and 86 females followed from birth up to 32 years old. 36 years (24 waves) 11 waves (all from 12–54 months and at least 1 of 3 adulthood waves from ages 23, 26, 32) Yes Zamir et al., (2018) 80 females followed from birth to age 32. Study selection 36 years (24 waves) 15 waves (all waves between birth-17.5 years, waves from ages 19–32) Yes 3. Child development project (USA) Goodnight et al. (2018) 226 males and 240 females aged 5 at T1 and followed to age 23. 23 years (23 waves) 18 waves (yearly from age 5–23) No 4. Family Transitions Project (FTP; USA) Neppl et al., (2019) 93 males and 100 females; aver age age 12.7 at T1, followed to age 29. 17 years (17 waves – 2nd generation only) 9 waves (ages 13–16,18,19,21,23,29) No 5. Project RED (USA) Grest et al., (2018) 345 males and 478 females aged 13.9 to 16.9 at T1 and followed to age 21.8–24.2. 6 years (6 waves) Wave 2 and Wave 6 No Grest et al., (2018) 418 males and 642 females aged 13.9–17.7 at T1 and followed to age 21.6–24.2. 6 years (6 waves) Wave 2, Wave 3 Wave 4, and Wave 6, No 6. Lehigh Longitudinal Study (USA) Herrenkohl & Jung (2016) 248 males and 209 females aged 18 months to 6 at T1, followed for 34 years. 34 years (4 waves) 3 waves (preschool, adolescent, adult) No 7. Mater Hospital-University of Queensland Study of Pregnancy (MUSP; Australia) Abajobir et al. (2017) 1495 males and 1827 females fol lowed from birth to age 21. 30 years (7 waves) 5 waves (birth – 21 years) No No Cohort Listed 8. Vezina et al. (2015) 443 females aged 6 at T1 and followed to age 21 who had been dating during at least 2 weeks in the past year. 15 years (3 waves) All No 9. Milaniak & Widom (2015) 613 males and 582 females fol lowed from birth to age 19-40.7 22 years (2 waves) All No 10. Handley, Russotti, Rogosch and Ciccetti (2019) 392 males and females aged 11 at T1 and followed to age 20. IPV perpetration (physi cal/threat of physical) • Shortened CTS2–15 items at 18, 22, 23 years of age 2. MLSRA (USA) 2. MLSRA (USA) Narayan et al., (2017) Developmental timing of exposure to interparental violence (EIPV) within early childhood; abuse/neglect Control variables: Child sex (female); socioeconomic status; maternal age IPV perpetration & IPV victimization • 10-item CTS (2 verbal, 8 physical) administered at ages 23, 26, 32. P t ( p b C 1 Zamir et al., (2018) Childhood sexual or physical abuse, timing of abuse Control variables: child hood socioeconomic status; occupational prestige; educa tional attainment; race IPV victimization • 8-item CTS at age 23, 26, 32 (physical items only) D ( I ( v 3. Child development project (USA) Goodnight et al. (2018) Individual factors: tempera mental resistance of control; psychopathic traits; antisocial behavior Social and environmental factors: maternal warmth; maternal monitoring; parent- adolescent problem solving; friend antisocial behavior; family stress Control variables: Sex; race/ ethnicity; socioeconomic status IPV perpetration (physi cal/threat of physical) • Shortened CTS2–15 items at 18, 22, 23 years of age M r a n I r p ( o p 4. FTP (USA) Neppl et al., (2019) Parental psychological vio lence in early adolescence IPV perpetration (psychological) • Partner and observer report of behavior to partner. 4 items covering psychological violence. ↑ p → h 5. ↑ Parental psychological violence in early adolescence (ß=0.29, p < .001) → ↑ psychological violence towards partner in adolescence → ↑ perpetration in emerging adulthood (ß=0.26, p < .001) and adult hood (ß=0.37, p < .001) IPV perpetration • CTS – interviewed female partner nomi nated by the man at age 48 IPV perpetration & IPV victimization • 10-item CTS (2 verbal, 8 physical) administered at ages 23, 26, 32. IPV perpetration & IPV victimization • 10-item CTS (2 verbal, 8 physical) administered at ages 23, 26, 32. IPV victimization • 8-item CTS at age 23, 26, 32 (physical items only) 1. CSDD (UK) IPV perpetration • CTS – interviewed female partner nomi nated by the man at age 48 ↓ harsh erratic parenting (OR = 0.27, 95% CI[0.10, 0.73]); ↑ impul sivity (OR = 2.82, 95% CI[1.32, 6.00]) 2. MLSRA (USA) Narayan et al., (2017) Developmental timing of exposure to interparental violence (EIPV) within early childhood; abuse/neglect Control variables: Child sex (female); socioeconomic status; maternal age Zamir et al., (2018) Childhood sexual or physical abuse, timing of abuse Control variables: child hood socioeconomic status; occupational prestige; educa tional attainment; race 3. Child development project (USA) Goodnight et al. (2018) Individual factors: tempera mental resistance of control; psychopathic traits; antisocial behavior Social and environmental factors: maternal warmth; maternal monitoring; parent- adolescent problem solving; friend antisocial behavior; family stress Control variables: Sex; race/ ethnicity; socioeconomic status 4. FTP (USA) Neppl et al., (2019) Parental psychological vio lence in early adolescence 5. Project RED (USA) Study selection 10 years (2 waves) All No (Abajobir et al., 2017; Grest, Amaro, et al., 2018; Grest, Lee, et al., 2018; Herrenkohl & Jung, 2016; Neppl et al., 2019; Vézina et al., 2015), four measured sexual IPV (Grest, Amaro, et al., 2018; Grest, Lee, et al., 2018; Her renkohl & Jung, 2016; Vézina et al., 2015), three measured injury from IPV (Grest, Amaro, et al., 2018; Grest, Lee, et al., 2018; Herrenkohl & Jung, 2016), one study measured verbal IPV (Narayan et al., 2017), and one study measured any experience of, or perpetration of IPV in the past 12 months (Handley et al., 2019). any experience of, or perpetration of IPV in the past 12 months (Handley et al., 2019). The proportion of IPV experienced within the studies varied, with perpetration of physical IPV ranging from 4.9% (Grest, Amaro, et al., 2018) to 38.4% (Milaniak & Widom, 2015). One study reported psychological IPV was perpe trated by 78.6% of the sample, and 75.2% of the sample had been victimized by psychological IPV (Herrenkohl & Jung, 2016). The proportion of IPV experienced within the studies varied, with perpetration of physical IPV ranging from 4.9% (Grest, Amaro, et al., 2018) to 38.4% (Milaniak & Widom, 2015). One study reported psychological IPV was perpe trated by 78.6% of the sample, and 75.2% of the sample had been victimized by psychological IPV (Herrenkohl & Jung, 2016). 1 3 1596 Journal of Family Violence (2023) 38:1591–1606 Table 2 Summary of predictor variables, outcome measures, and findings Predictors/pathways from childhood and adolescence IPV perpetration & IPV victimization • 10-item CTS (2 verbal, 8 physical) administered at ages 23, 26, 32. Perpetration Sexual: ↑ child maltreatment (b = 1.21, p < .05) Injury: ↑ child maltreatment (b = 2.56, p < .05) 7. MUSP (Australia) Abajobir et al. Interaction effects (victimization) Interaction effects (victimization) Psychological: ↑ alcohol use for females only (AOR = 1.57, 95% CI[1.26, 1.96]) Grest et al., (2018) Acculturation Control variables: child hood abuse; witnessing domestic violence before 18 6. Lehigh Longitudinal Study (USA) Herrenkohl & Jung (2016) Child abuse/maltreat ment (parent-report and substantiated) Control variables: age; gen der; socioeconomic status Bidirectional IPV (Psychological, sexual, multiform) • CTS-SF – 14 items Sexual: ↑ childhood abuse (OR = 2.85, 95% CI[1.82, 4.45]) Multiform: ↑ childhood abuse (OR = 5.39, 95% CI[1.58, 18.41]) Sexual: ↑ childhood abuse (OR = 2.85, 95% CI[1.82, 4.45]) Multiform: ↑ childhood abuse (OR = 5.39, 95% CI[1.58, 18.41]) 2. MLSRA (USA) Project RED (USA) Perpetration and victimization: ↑ EIPV in toddlerhood (perpetra tion: ß=0.23, p < .01; victimization: ß=0.25, p < .01), but not infancy (Perpetration: ß=0.00, p > .05; victimization: ß=-0.07, p > .05), predicts new IPV cases at age 23 and a pattern of increasing IPV between ages 26 and 32, compared to desisting IPV (OR = 1.37, 95% CI[1.07, 1.76]) and non-violent patterns (OR = 1.41, 95% CI[1.15, 1.72])f IPV perpetration & IPV victimization IPV victimization • 8-item CTS at age 23, 26, 32 (physical items only) Direct effects: ↑ physical or sexual abuse in childhood/adolescence (ß=0.27, p = .007)f Direct effects: ↑ physical or sexual abuse in childhood/adolescence (ß=0.27, p = .007)f Indirect effects: ↑ physical or sexual abuse in childhood/adolescence (ß=0.26, p = .020) → ↑ dissociation at age 19 (ß=0.30, p = .003) → victimization Main effects: ↑ psychopathic traits (b = 0.28, p = .001[intercept]); ↑ resistance to control temperament (b = 0.02, p = .031[slope]); ↑ friend antisocial behavior (b = 0.05, p = .011[slope]); ↓ higher socioeco nomic status (b = − 0.15, p = .041[intercept]) Interaction effects: ↓ parent-teen relationship only at levels of resistance to control temperament 2.2 SD above average (b=-0.11, p = .028[slope]); ↑ friend antisocial behavior for males, not females (b=-0.06, p = .011[slope]); ↑ friend antisocial behavior when levels of resistance to control temperament 1 SD above average (b = 0.03, p = .025[intercept]), but not 1 SD below average IPV perpetration (psychological) • Partner and observer report of behavior to partner. 4 items covering psychological violence. IPV perpetration (psychological) • Partner and observer report of behavior to partner. 4 items covering psychological violence. 5. 7. MUSP (Australia) Abajobir et al. (2017) Childhood maltreatment/ abuse (sexual, physical, emotional, neglect) 2. MLSRA (USA) Project RED (USA) 1 3 1 3 1 3 1597 Journal of Family Violence (2023) 38:1591–1606 Journal of Family Violence (2023) 38:1591–1606 Study/cohort (country) Predictor variables measured in childhood and adolescence Outcome measures Predictors/pathways from childhood and adolescence Grest et al., (2018) Acculturation; traditional gender role attitudes; substance use; depressive symptoms Control variables: child hood abuse; witnessing domestic violence before 18 IPV victimization (physical, psychologi cal, injury, sexual) & IPV perpetration (physi cal, psychological, injury, sexual) • CTS-SF – 14 items Main effects (perpetration) Psychological: ↑ alcohol use (AOR = 1.31, 95% CI[1.12, 1.54]); ↑ witnessed domestic violence (AOR = 2.25, 95% CI[1.29, 3.93]) Physical: ↑ alcohol use (AOR = 1.37, 95% CI[1.12, 1.68]); ↓ Hispanic-oriented acculturation (AOR = 0.70, 95% CI[0.50, 0.99]) Injury: ↑ traditional gender roles (AOR = 1.21, 95% CI[1.03, 1.41]); ↓ Hispanic-oriented acculturation (AOR = 0.37, 95% CI[0.18, 0.75]) Sexual: ↑ traditional gender roles (AOR = 1.07, 95% CI[1.01, 1.13]); Table 2 (continued) Predictors/pathways from childhood and adolescence Main effects (victimization) Psychological: ↑ traditional gender roles (AOR = 1.08, 95% CI[1.02, 1.14]); ↑ alcohol use (AOR = 1.20, 95% CI[1.03, 1.39]) Physical: ↑ traditional gender roles (AOR = 1.13, 95% CI[1.05, 1.22]) Injury: ↑ traditional gender roles (AOR = 1.18, 95% CI[1.03, 1.35]); ↓ Hispanic-oriented acculturation (AOR = 0.49, 95% CI[0.27, 0.88]); ↑ depressive symptoms (AOR = 1.05, 95% CI[1.00, 1.10]) Sexual: ↑ history of child abuse (AOR = 2.77, 95% CI[1.69, 4.54]) Interaction effects (perpetration) Psychological: ↑ traditional gender roles for males only (AOR = 1.16, 95% CI[1.03, 1.29]); ↑ alcohol use for females only (AOR = 1.67, 95% CI[1.34, 2.09]) Physical: ↑ traditional gender roles for males only (AOR = 1.28, 95% CI[1.09, 1.49]); ↑ US-oriented acculturation for males only (AOR = 3.15, 95% CI[1.13, 8.76])f Bidirectional IPV (Psychological, sexual, multiform) • CTS-SF – 14 items Main effects (perpetration) Main effects (perpetration) Psychological: ↑ alcohol use (AOR = 1.31, 95% CI[1.12, 1.54]); ↑ witnessed domestic violence (AOR = 2.25, 95% CI[1.29, 3.93]) Physical: ↑ alcohol use (AOR = 1.37, 95% CI[1.12, 1.68]); ↓ Hispanic-oriented acculturation (AOR = 0.70, 95% CI[0.50, 0.99]) Injury: ↑ traditional gender roles (AOR = 1.21, 95% CI[1.03, 1.41]); ↓ Hispanic-oriented acculturation (AOR = 0.37, 95% CI[0.18, 0.75]) Sexual: ↑ traditional gender roles (AOR = 1.07, 95% CI[1.01, 1.13]); ↑ history of child abuse (AOR = 3.27, 95% CI[1.82, 5.88]) Main effects (victimization) Perpetration (2017) Childhood maltreatment/ abuse (sexual, physical, emotional, neglect) Emotional IPV: ↑ any abuse (AOR = 1.84, 95% CI[1.31, 2.57]); ↑ physical abuse (AOR = 1.84, 95% CI[1.11, 3.03]); ↑ emotional abuse (AOR = 3.19, 95% CI[1.99, 5.14]); ↑ neglect (AOR = 2.64, 95% CI[1.58, 4.42]) Physical IPV: ↑ any abuse (AOR = 2.14, 95% CI[1.51, 2.99]); ↑ sexual abuse (AOR = 2.31, 95% CI[1.27, 4.18]); ↑ physical abuse (AOR = 1.76, 95% CI[1.06, 2.92]); ↑ emotional abuse (AOR = 2.76, 95% CI[1.72, 4.43]); ↑ neglect (AOR = 2.74, 95% CI[1.62, 4.63]) Harassment: ↑ emotional abuse (AOR = 1.63, 95% CI[1.02, 2.59]) Combined/severe IPV: ↑ any abuse (AOR = 2.12, 95% CI[1.28, 3.51]); ↑ emotional abuse (AOR = 3.97, 95% CI[2.74, 7.04]); ↑ neglect (AOR = 4.62, 95% CI[2.51, 8.52]) Emotional IPV: ↑ any abuse (AOR = 1.84, 95% CI[1.31, 2.57]); ↑ physical abuse (AOR = 1.84, 95% CI[1.11, 3.03]); ↑ emotional abuse (AOR = 3.19, 95% CI[1.99, 5.14]); ↑ neglect (AOR = 2.64, 95% CI[1.58, 4.42]) Physical IPV: ↑ any abuse (AOR = 2.14, 95% CI[1.51, 2.99]); ↑ sexual abuse (AOR = 2.31, 95% CI[1.27, 4.18]); ↑ physical abuse (AOR = 1.76, 95% CI[1.06, 2.92]); ↑ emotional abuse (AOR = 2.76, 95% CI[1.72, 4.43]); ↑ neglect (AOR = 2.74, 95% CI[1.62, 4.63]) Harassment: ↑ emotional abuse (AOR = 1.63, 95% CI[1.02, 2.59]) Combined/severe IPV: ↑ any abuse (AOR = 2.12, 95% CI[1.28, 3.51]); ↑ emotional abuse (AOR = 3.97, 95% CI[2.74, 7.04]); ↑ neglect (AOR = 4.62, 95% CI[2.51, 8.52]) Emotional IPV: ↑ any abuse (AOR = 1.84, 95% CI[1.31, 2.57]); ↑ physical abuse (AOR = 1.84, 95% CI[1.11, 3.03]); ↑ emotional abuse (AOR = 3.19, 95% CI[1.99, 5.14]); ↑ neglect (AOR = 2.64, 95% CI[1.58, 4.42]) IPV victimization (emotional, physical, harassment, combined/ severe) • Modified CTS – 7 items Physical IPV: ↑ any abuse (AOR = 2.14, 95% CI[1.51, 2.99]); ↑ sexual abuse (AOR = 2.31, 95% CI[1.27, 4.18]); ↑ physical abuse (AOR = 1.76, 95% CI[1.06, 2.92]); ↑ emotional abuse (AOR = 2.76, 95% CI[1.72, 4.43]); ↑ neglect (AOR = 2.74, 95% CI[1.62, 4.63]) Harassment: ↑ emotional abuse (AOR = 1.63, 95% CI[1.02, 2.59]) Combined/severe IPV: ↑ any abuse (AOR = 2.12, 95% CI[1.28, 3.51]); ↑ emotional abuse (AOR = 3.97, 95% CI[2.74, 7.04]); ↑ neglect (AOR = 4.62, 95% CI[2.51, 8.52]) No Cohort Listed 1 3 1 3 Journal of Family Violence (2023) 38:1591–1606 1598 Study/cohort (country) Predictor variables measured in childhood and adolescence Outcome measures Predictors/pathways from childhood and adolescence 8. Perpetration Vezina et al. (2015) Parental monitoring; affiliation with deviant peers; childhood behavior prob lems; high-risk behaviors IPV victimization (physical/sexual, psychological) • Combination of Violence faite aux Filles dans les Fréquentations à l’Adolescence and CTS2 (2 items from CTS2, edited) Psychological (victimized in both early adulthood and ado lescence – i.e., revictimized): ↑ childhood behavior problems (OR = 1.10, 95% CI[1.03, 1.18]), ↑ high-risk behaviors (OR = 1.75, 95% CI[0.99, 3.11]) Psychological (victimized in early adulthood only): ↑ childhood behavior problems (OR = 1.09, 95% CI[1.01, 1.17]) 9. Milaniak & Widom (2015) Childhood physical abuse, sexual abuse, and neglect IPV Perpetration • One item from Antisocial Personal ity Disorder module of DIS-III-R “have you ever hit or thrown things at your partner?” ↑ Abuse/neglect (AOR = 1.54, 95% CI[1.10, 1.97]) 10. Handley, Russotti, Rogosch and Ciccetti (2019) Childhood maltreatment; childhood antisocial behavior; childhood relational aggression T2: IPV involvement • One item assessing whether participants had experienced or perpe trated domestic violence in the past 12 months. ↑ Childhood maltreatment (t = − 3.60, p < .05) ↓ negative association with outcome, ↑ positive association with outcome, → interaction pathway, odds ratio (OR), adjusted odds ratio (AOR), confidence interval (CI), standard deviation (SD) Table 2 (continued) Psychological (victimized in both early adulthood and ado lescence – i.e., revictimized): ↑ childhood behavior problems (OR = 1.10, 95% CI[1.03, 1.18]), ↑ high-risk behaviors (OR = 1.75, 95% CI[0.99, 3.11]) Psychological (victimized in early adulthood only): ↑ childhood behavior problems (OR = 1.09, 95% CI[1.01, 1.17]) ↑ Abuse/neglect (AOR = 1.54, 95% CI[1.10, 1.97]) ↑ Childhood maltreatment (t = − 3.60, p < .05) ↓ negative association with outcome, ↑ positive association with outcome, → interaction pathway, odds ratio (OR), adjusted odds ratio (AOR), confidence interval (CI), standard deviation (SD) psychological childhood abuse specifically (Neppl et al., 2019), suggesting homogeneity of violence type continuity from childhood to adulthood. Indirect exposure to violence (i.e., witnessing violence in the home) was also shown to increase the likelihood of both psychological (Grest, Amaro, et al., 2018), verbal and physical IPV (Narayan et al., 2017) in adulthood. 1 3 Predictors Zamir et al., (2018) demonstrated that physical or sexual abuse led to increases in symptoms of dissociation in young adulthood, which increased the risk for physical IPV victimization in adulthood. Behavioral and Personality Risks There was mixed evi dence regarding the impact of antisocial behaviors in childhood on IPV in adulthood. Although Handley et al., (2019) demonstrated no significant relationship between adolescent antisocial behavior and IPV involvement, more nuanced evidence suggests that antisocial behavior may predict victimization of IPV (specifically psychological forms; Vézina et al., 2015), but not IPV perpetration (Good night et al., 2017; Vézina et al., 2015) found that young girls who displayed problematic behaviors (e.g., disruptiveness) at age six, or high-risk behaviors at age 15 (e.g., risky sexual activity), were at increased risk of experiencing a pattern of psychological IPV victimization in adulthood (as well as adolescence). While Goodnight et al., (2017) found no pro spective association between antisocial behaviors at age 16 and patterns of IPV perpetration in adulthood, their model also incorporated co-occurring psychopathic traits, which was found to be positively predictive. It may be that psy chopathic traits are a mechanism for both antisocial behav iors in adolescence as well as IPV in later life. Regarding other personality factors, impulsivity was also demonstrated to be influential by one study. Theobald et al., (2016) found that higher impulsivity in late childhood were nearly three times more likely to report perpetrating physical family vio lence in adulthood. This association was unique to family violence and did not carry through to the perpetration of general violence. Family of Origin Risks Three studies explored the prospec tive effect of factors related to the family system during childhood and adolescence, however many of these fac tors were not significantly predictive of IPV in adulthood. Parent-factors, such as criminal background, nervousness, as well as disagreement between parents do not impact the likelihood of IPV in adulthood. Instead, dyadic parent-child factors are more impactful, although not with the antici pated effect; Theobald et al., (2016) found that harsh and erratic parenting reduced the likelihood of IPV perpetration. However, this effect was unique to those who perpetrated family violence only, with the absence of any general vio lent offending. Predictors anticipated direction, whereby increases in harsh and erratic discipline increased the likelihood of violence in adulthood. Other dyadic factors, including warmth and monitoring/ supervision had no effect on adult IPV, although a positive parent-child relationship (which included aspects of moni toring) reduced the risk of IPV but only in young people with an impulsive and unmanageable temperament (Good night et al., 2017; Vézina et al., 2015). Further, Vézina et al., (2015) speculate that parental monitoring could have an indirect effect on IPV victimization in girls through asso ciation with deviant peers and behavioral problems, which were positively associated with monitoring and IPV. As with perpetration, there are inconsistencies in the rep licability of these effects. Indeed, Grest, Amaro, et al. (2018) found that a history of child abuse only predicted sexual IPV victimization, with no links to physical or psychologi cal IPV victimization in adulthood. Further, Herrenkohl & Jung (2016) found no significant associations between child abuse or substantiated maltreatment and IPV victim ization. Herrenkohl & Jung (2016) used a robust process to distinguish serious physical and emotional abuse, which would have isolated only the most serious abuse cases, yet compared to Abajobir et al. (2017) - who reported several significant associations - Herrenkohl & Jung (2016) had far lower rates of IPV victimization in their cohort (e.g., physi cal victimization: 39.4% vs. 16.4%, respectively). Given the sample in Abajobir et al. (2017) reported twice the rate of victimization compared with the sample in Herrenkohl & Jung (2016), it may be that those in the Abajobir et al. (2017) study represent a highly traumatized sample that may not be obtainable or replicable in other longitudinal studies of a similar design. Whilst this may limit the generalizability of the Abajobir et al. findings, they remain an important con tribution to the development of a wholistic understanding of IPV. Similarly, Narayan et al., (2017) found no signifi cant associations while looking at patterns of IPV over time in adulthood and using a robust measure of abuse/neglect. Indeed, abuse and neglect frequently co-occur with expo sure to inter-parental violence in the home, which Narayan et al., (2017) demonstrated to be a more prominent predictor of increasing patterns of IPV victimization. Alternatively, the prospective effect on victimization may be partially mediated by the individual’s response to the trauma of the event(s). Predictors The majority of predictors identified in the current review were consistent with the five risk domains identified in the Costa et al., (2015) review: child and adolescent experi ences of violence and abuse (8 studies, 67%), family of ori gin risks (3 studies, 25%), behavioral and personality risks (5 studies, 42%), adolescent peer risks (2 studies, 17%), and sociodemographic risks (5 studies, 42%). The current review identified one additional domain: cultural and atti tudinal risks (2 studies, 17%). Consistent with Costa et al., risk factors for both IPV perpetration and victimization for both males and females will be discussed within each of these domains. However, Grest, Amaro, et al. (2018) did not replicate the association with physical IPV. This was despite the two studies possessing the same methodological quality and using a similar measure of IPV. Further, there appears to be no association with bidirectional IPV (Grest, Lee, et al., 2018). As such, the timing of this exposure may be impor tant; Narayan et al., (2017) found that exposure to violence in the home did not impact adult IPV when it occurred dur ing infancy (birth-24 months), only that which occurred during toddlerhood (25–64 months of age) increased the risk of IPV. Child and Adolescent Exposure to Violence and Abuse Expe riences Experiences of abuse, neglect, and maltreatment in childhood and adolescence were consistently reported as increasing the risk of involvement in IPV in adulthood. Regarding perpetration, the experience of abuse prospec tively predicts an increase in the likelihood of reporting physical, sexual, and psychological IPV in adulthood. Abuse also predicts bidirectional sexual and bidirectional multiform IPV (i.e., involving more than one type of vio lence). The association with psychological IPV perpetration only appears to hold when the individual has experienced Regarding victimization in adulthood, results from this review suggest there may be differential associations between abuse/neglect and IPV depending on the form of violence experienced. Physical, sexual, and emotional abuse in childhood/adolescence is associated with an increased risk of experiencing physical and sexual IPV. Emotional IPV victimization does not appear to be related to childhood sexual abuse, but associations with physical and emotional abuse remain. However, despite these associations, Abajo bir et al. (2017) reports that the most consistent predictor 1 1 3 Journal of Family Violence (2023) 38:1591–1606 1599 across forms of IPV victimization, when controlling for other abuse, is the experience of childhood neglect. Predictors For those who reported both general violent offending as well as family violence, the effect was in the One study has explored the prospective association of alcohol use and depressive symptoms in adolescence and IPV in adulthood. Grest, Amaro, et al. (2018) found that alcohol use in adolescence increased the risk of perpetrating physical and psychological IPV in adulthood. However, the association with psychological IPV perpetration was only found in females. Further, for victimization, alcohol use increased the likelihood of psychological IPV, but not other forms (and again, only for females). Neither marijuana, nor tobacco consumption, had any effect on victimization or perpetration. Regarding mental health, symptoms of 1 3 3 Journal of Family Violence (2023) 38:1591–1606 1600 Table 3 Cambridge Quality Checklists Ratings Study Correlate Score Risk Factor Score Causal Score Theobald et al., (2016) 4 3 5 Narayan et al., (2017) 3 3 5 Zamir et al., (2018) 3 3 5 Milaniak and Widow (2015) 3 3 5 Handley, Russotti, Rogosch and Ciccetti (2019) 1 3 5 Goodnight et al. (2018) 5 3 5 Neppl et al., (2019) 3 3 5 Abajobir et al. (2017) 3 3 5 Grest et al., (2018) 3 3 5 Grest et al., (2018) 3 3 5 Herrenkohl & Jung (2016) 2 3 5 Vezina et al. (2015) 1 3 5 able 3 Cambridge Quality Checklists Ratings depression increased the risk of injury from IPV victimiza tion in adulthood (Grest, Amaro, et al., 2018). depression increased the risk of injury from IPV victimiza tion in adulthood (Grest, Amaro, et al., 2018). Adolescent Peer Risks Two studies have investigated the influence of peer relationships in adolescence on the like lihood of perpetrating IPV in adulthood. Negative peer influence appears to increase the likelihood of physical IPV perpetration (Goodnight et al., 2017), but not IPV victimiza tion in girls (Vézina et al., 2015). Indeed, these effects may be gender specific; Goodnight et al., (2017) found that lev els of peer antisocial behavior in adolescence led to physical IPV, but only in males. This effect was also stronger in those who displayed impulsive and unmanageable behaviors in early childhood. Cultural and Attitudinal Risks The degree to which a young person held gender normative beliefs (i.e., adherence to tra ditional conceptualizations of male and female roles) was a more consistent predictor of victimization than perpetra tion, across violence forms. Using a Latino sample, Grest, Amaro, et al. Predictors (2018) identified a 7% increase in the risk of sexual IPV perpetration, and a 21% increase in the likeli hood of causing injury for the overall sample. However, when the results were stratified by gender these findings were no longer significant, and instead, being male and holding gender normative beliefs was associated with a 16% increase in the risk of perpetrating psychological IPV and a 28% increase in the risk of perpetrating physical IPV. Further, these attitudes predicted victimization from psy chological and physical violence, as well as the likelihood of injury, but not sexual IPV (Grest, Amaro, et al., 2018). studies, were not measuring concurrent abuse or maltreat ment, which may therefore be a more important predictor than SES, as evidenced by the findings in the other four studies that included SES. Quality of studies Table 3 shows the overall ratings on the Cambridge Quality Checklists (Murray et al., 2009) independently completed by Authors AC and TH, with a 96% agreement rate. All dis agreements were resolved in consultation with Author RB. Only one study reached the highest score for the Cor relate domain (Goodnight et al., 2017), whereas all studies reached the highest rating (3) for the risk factor domain. All studies included in this review received a 5 out of 7 for the causal score, reflective of a lack of clear comparison groups or randomized control designs in this area. Table 3 shows the overall ratings on the Cambridge Quality Checklists (Murray et al., 2009) independently completed by Authors AC and TH, with a 96% agreement rate. All dis agreements were resolved in consultation with Author RB. Only one study reached the highest score for the Cor relate domain (Goodnight et al., 2017), whereas all studies reached the highest rating (3) for the risk factor domain. All studies included in this review received a 5 out of 7 for the causal score, reflective of a lack of clear comparison groups or randomized control designs in this area. In two related studies using Latino samples, Grest, Amaro, et al. (2018) and Grest et al., (2018) also explored the effect of developmental acculturation on IPV in adult hood. Although acculturation did not predict any bidirec tional forms of IPV (Grest, Lee, et al., 2018), the prospective associations for victimization and perpetration indepen dently were more nuanced. Growing up in a Hispanic-ori ented culture reduced the risk of perpetrating physical IPV and reduced the likelihood of injury from victimization. In contrast, a United States acculturation during childhood increased the risk of perpetrating physical IPV in adulthood, but as with gender normative beliefs, only for males. Discussion The current review updated the previous longitudinal review conducted by Costa et al. in 2015. In doing so, the review intended to address two key aims: (1) whether there were any newly discovered developmental predictors of IPV perpetration and victimization that could be used to inform prevention and intervention efforts, and (2) whether risk factors for IPV perpetration and victimization could be distinguished from one another. Sociodemographic Risks Socio-economic status (SES) dur ing childhood was not a reliable predictor of IPV perpetra tion or victimization in adulthood. Four studies found no significant association with IPV outcomes (Herrenkohl & Jung, 2016; Narayan et al., 2017; Theobald et al., 2016). While Goodnight et al., (2017) found that SES reduced the risk of physical violence perpetration, they, unlike other Predictors of IPV The current review found that experiences of abuse during development had a consistent impact on both IPV victimiza tion and perpetration. Broadly, this finding is consistent with 1 1 3 1601 Journal of Family Violence (2023) 38:1591–1606 Costa et al., (2015) however the current review adds much needed specificity regarding the forms of abuse experienced and the forms of IPV predicted (noted by Costa et al. as being underexplored at the time). Regarding perpetration, there appears to be homogeneity between the type of abuse experienced, and the type of IPV perpetrated (e.g., those who experience psychological abuse as a child are more likely to perpetrate psychological IPV). For victimization, the review identified mixed relationships with abuse forms which depend on the type of IPV experienced in adulthood. The most reliable predictor across all adulthood IPV forms being childhood neglect. There was some evidence to sug gest, however, that variation (i.e., individual differences) in how a person responds to trauma (i.e., levels of dissocia tion) may mediate the relationship between childhood pre dictors and victimization. this same study found that alcohol use in adolescence was only related to psychological IPV perpetration in females (in contrast to Costa et al.). Further, the association between alcohol use and victimization was only evident for females and psychological victimization, again conflicting with the Costa et al. review which found consistent relationships between perpetration and victimization of IPV, and sub stance use. Importantly, one study in the current review identified the role of mental health, noting that depressive symptoms increased the likelihood of experiencing injury from IPV victimization in adulthood. This is a new risk fac tor identified by the current review as no studies included in the Costa et al. review discussed or identified the role of mental health of the participant.i The current review identified that adolescent peer risks were associated with IPV, but only male perpetration, dif fering from the Costa et al., (2015) review which found that poor quality adolescent peer networks were associated with both perpetration and victimization for males and females. Further to this, the current review identified that this asso ciation was stronger for males who exhibited impulsive and unmanageable behaviors in early childhood, suggesting that a difficult temperament may predispose susceptibility to negative peer influence in adolescence, which then predicts the perpetration of physical IPV in adulthood. Predictors of IPV Similarly, the pathway from poor parental supervision may also be medi ated through negative peer influence. Regarding family of origin risk factors, parent-child rela tionships and discipline were important factors in predict ing IPV in adulthood. While Costa et al., (2015) found that harsh parenting was not associated with IPV in adulthood for males, the current review provides more clarity here; for males, harsh parenting may reduce the likelihood of adult IPV offences in isolation, but increase the likelihood of IPV offences which form a part of a more general pattern of offending. Furthermore, Costa et al. reported that weak parent-child attachment and negative interactions were a strong and consistent predictor of both IPV perpetration and victimization across genders. We add that this effect may interact with the temperament of the young person; a positive parent-child relationship protects against IPV only when the young person has a difficult temperament (Good night et al., 2017). A new domain of longitudinal risk factors for IPV was identified in the current review: cultural and attitudinal risks. The Costa et al., (2015) review did not identify any cultural or attitudinal risks, however these have impor tant implications for prevention and intervention for IPV. Specifically, one study utilizing a culturally and socioeco nomically specific sample (Latino high-school students in Southern California) (Grest, Amaro, et al., 2018) identified that holding gender normative beliefs as a young person (male or female) was related to a small increase in the risk of perpetrating sexual IPV and in the likelihood of caus ing injury to others. Importantly, Grest, Amaro, et al. (2018) also identified comparable rates of IPV perpetration and vic timization for males and females in their sample, consistent with findings identifying the bidirectional nature of intimate partner violence (Caetano et al., 2004; O’Leary et al., 2008; Renner & Whitney, 2012; Ulloa & Hammett, 2016). The bidirectional nature of IPV aligns with two types of IPV as discussed by Johnson (2006a);situational couple violence and mutual violence. Situational couple violence describes IPV in which the violence occurs in a dyadic context, how ever neither partner is violent and controlling (i.e., violence occurs as a result of the ‘situation’), whereas mutual vio lence refers to IPV in which both partners are both violent and controlling. Distinguishing between victimization and perpetration For males specifically, holding gender normative beliefs was associated with an increase in perpetrating physical and psychological IPV. These associations may reflect the increased severity of IPV which commonly characterizes another of Johnson’s (2006a) typologies -- intimate terror ism, in which only one member of the relationship is vio lent and controlling (usually the male), and suggests that the attitudes underpinning intimate terrorism may originate early in life. Further, such attitudes also predicted psycho logical and physical victimization, which may suggest that some intimate terrorism could be bi-directional, consistent with the findings of Johnson et al., (2014). However, to fur ther understand the relationship between gender normative beliefs and victimization, specifically, person-based analy sis is required. Given the smaller volume of studies in the current review (compared to Costa et al., 2015), it was difficult to isolate factors which may differentiate IPV victimization from IPV perpetration. Many risk factors were only explored for either victimization or perpetration, rendering the compari son somewhat biased; there was significant heterogeneity between studies in the instruments used to assess IPV. Fur ther, no studies reported estimating the unique effect of risk factors on either perpetration or victimization, while con trolling for the opposing experience. Given the substantial bidirectionality of IPV, this could reduce the reliability of the effects reported. Despite this, where comparisons can be made, there were some preliminary trends to consider. Anti-social behavior in childhood/adolescence emerged as a stronger predictor of victimization for females (Vézina et al., 2015), while association with anti-social peers may be more unique to perpetration for males (Goodnight et al., 2017). However, when considering the interaction between gender and the distinction between victimization- and perpe tration-specific risk factors, it is important to note that males are often under-represented in victimization literature, and females underrepresented in perpetration literature (Las key et al., 2019; Mackay et al., 2018). Nonetheless, gender normative attitudes were related to more forms of IPV vic timization than perpetration. Exposure to family violence and victimization, alcohol use, as well as acculturation all appear consistent across victimization and perpetration. In terms of acculturation, the current review highlighted a reduced risk of perpetrating physical IPV and injury from IPV victimization if a young person was raised in a His panic-oriented culture specifically. However, being raised as a male with a United States acculturation increased the risk of perpetrating physical IPV in adulthood. Predictors of IPV In contrast to Costa et al., (2015), a history of antiso cial behavior in childhood did not emerge as a strong or consistent predictor of IPV. Antisocial behavior in adoles cence was only demonstrated to predict a greater likelihood of psychological IPV victimization, not perpetration, and only in females (Vézina et al., 2015). Instead, the findings of the current review suggest that it may be more impor tant to consider the co-occurring psychopathic traits and early temperament than the delinquency likely facilitated by these factors, when predicting IPV perpetration in adult hood. Young people with psychopathic traits likely belong to a population whose use of violence does not cease in ado lescence (i.e., adolescent-limited) and instead continues into adulthood (i.e., life course persistent aggressors) (Moffitt et al., 2002), putting them at particular risk for perpetrating adult partner violence. In terms of substance use, the only study to assess this found that alcohol use as an adolescent increased the likeli hood of perpetrating both physical and psychological IPV as an adult, consistent with Costa et al., (2015). However, 1 3 1602 Journal of Family Violence (2023) 38:1591–1606 Distinguishing between victimization and perpetration It is important to note that this finding is from two studies within the same longitudinal cohort which specifically followed a Latino sample, and as such investigating acculturation in future longitudinal studies will be crucial to identifying the role that these risk factors play in IPV perpetration and victim ization, and in determining which specific cultural factors may be contributing. It is possible that young Latinos who retain a Hispanic-oriented culture experience protective fac tors common within Latino families, such as a strong family network (Dupont-Reyes et al., 2015), which reduces the risk of violent relationships in adulthood. Bidirectional IPV is frequently over-looked, despite being an important understanding for risk assessment and intervention (Bates, 2016), and the most prevalent manifes tation of IPV (Langhinrichsen-Rohling et al., 2012). Our review identified only one prospective study which consid ered bidirectional IPV. As such, we were not able to provide a meaningful interpretation of how this form may differ from unidirectional perpetration or victimisation. Similarly, Costa et al., (2015) included only two prospective studies on bidi rectional IPV, where depressive symptoms in adolescence was identified as a risk factor. Bidirectional IPV should be a unique consideration in prospective designs moving for ward, and also controlled for when considering the unique risk factor profiles of victimisation and perpetration. In terms of sociodemographic risk factors, the current review found that socio-economic status (SES) during childhood was not associated with either IPV perpetration or victimization, in contrast with Costa et al., (2015) who found that low SES was a significant predictor of domestic violence in two studies. It appears that the studies included in the current review found that other predictors included in their predictive models were stronger predictors of IPV perpetration and victimization than SES, which is consistent with the findings from Fergusson et al., (2008) in the Costa et al. review. Further, the measures used to assess SES in the studies identified within the current review focused on a combination of income and education, whereas the two studies in Costa et al., that reported SES as a significant predictor of IPV experiences utilized either income (Man chikanti Gómez, 2010), or only occupation (Magdol et al., 1998), with the operationalization of SES potentially result ing in differing findings across the studies. Methodological considerations The major methodological limitation of the studies included in the current review was the lack of control or compari son groups, and randomized control trial designs. Without methodologically strong designs, there is a risk that the find ings of the current review are biased, potentially limiting 1 3 Journal of Family Violence (2023) 38:1591–1606 1603 due to the limited ability of the CTS to consider this type of context, without modification (e.g., Babcock et al., 2019). As noted by Costa et al., and reiterated in a recent meta- analysis (Love et al., 2018), samples drawn from clinical settings (e.g., participants of interventions, hospital settings, domestic violence shelters, or from police records) are likely to have risk factors that are strongly associated with intimate terrorism, in contrast to situational couple violence which is more likely to be identified in non-clinical samples (i.e., general community samples). However, as these differ ent types of violence were not assessed in any of the studies included in the review, we are not able to draw conclusions about whether the identified developmental predictors are related to situational violence or intimate terrorism. This has important implications as these types of violence present dif ferently, and as such likely require different prevention and intervention approaches. Importantly, many existing inter vention options target intimate terrorism, but are unlikely to be effective for situational violence as such programs do not address the relationship factors known to be present in couples who experience situational violence (Love et al., 2018). Future research focused on adulthood experiences of IPV would benefit from inclusion of an assessment of type of IPV experienced (as a perpetrator and/or victim) to assist in better understanding the contribution of developmental predictors to adulthood IPV, and to better inform prevention and intervention efforts. its generalizability. Largely, the key methodological issue identified with the studies in the current review was a lack of an adequate sampling method (i.e., total population sam pling or random sampling), with many studies employing convenience or case control samples. The current review, however, only included prospective studies that all attained the highest rating for the risk factor category on the Cam bridge Quality Checklists (Murray et al., 2009). As such, it is possible to infer temporal precedence. Further, all studies were classified as non-randomized non-experimental stud ies in which there is not adequate control of covariates or within-individual change. Methodological considerations Whilst this study design is on the higher end of the Cambridge Quality Rating scale, it falls short of the gold standard (i.e., randomized control trials tar geting a specific risk factor). Improving the quality of stud ies through stronger methodology would go some way in improving the overall quality of the literature investigating developmental predictors of IPV. However, for such stud ies to adopt a randomized control trial design feasibly and ethically, it is likely they would need to take an interven tion approach whereby a specific risk factor is targeted and the long-term impact on IPV perpetration, victimization, or both is assessed. Regarding measurement, the majority of studies employed the Conflict Tactics Scale (CTS), or a variation thereof, to measure IPV. The CTS has received criticism for not replicating the gender asymmetry found with other IPV measures, as well as its lack of clinical utility when applied without consideration of contextual factors (Jones et al., 2017). The CTS was also developed for heterosexual couples, and while it appears valid for use with same-sex relationships (Regan et al., 2002), it may lead to under- reporting of IPV experiences, compared with same-sex relationship-specific measures (e.g., Stephenson & Finneran 2013). Indeed, IPV between same-sex partners may actually be more prevalent than in heterosexual couples (Messinger, 2011). Such measurement bias may have impacted studies in the current review, though, no studies specifically consid ered gender-diverse populations, or same-sex couples; the vast majority did not report the prevalence of these popula tions within their sample (apart from two, where over 90% of couples were heterosexual; Grest, Amaro, et al., 2018; Grest, Lee, et al., 2018), while one study noted the exclusion of same-sex couples (n = 20; Theobald et al., 2016). Implications availability of parenting skill building programs and pro grams focused on building younger children’s cognitive and social skills both of which have been shown to have posi tive impacts on antisocial and disruptive behavior (Forgatch et al., 2009; Menting et al., 2013; Reynolds et al., 2007), or trauma informed cognitive behavioral therapy for those children who have witnessed or experienced abuse (Cohen et al., 2011; Scheeringa et al., 2011; Smith et al., 2007). In addition, children or adolescents who are identified as using alcohol or who have mental health concerns would benefit from intervention to address these concerns early in an attempt to prevent these resulting in increased risk of IPV perpetration or victimization in adulthood. Impor tantly, intervention programs which incorporate a focus on a range of key risk factors for IPV (including experience of trauma and substance use) are more likely to yield positive outcomes for IPV perpetration (Karakurt et al., 2019). As one example, Communities That Care supports coalitions to design and implement community prevention strategies, which have evidence for reducing risk factors identified in the current review (Toumbourou et al., 2019). Registration and Protocol The current review was registered with Prospero, ID: CRD42020207038. Registration and Protocol The current review was registered with Prospero, ID: CRD42020207038. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, 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 licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/. References The current review was largely consistent with findings from Costa et al., (2015), however key differences within larger developmental predictor categories emerged. These related to the relationship with gender, whether the devel opmental predictors were associated with IPV perpetration or victimization as an adult, and the inclusion of possible moderating or mediating effects of other variables on IPV perpetration or victimization. Further, a new developmen tal risk domain emerged: cultural and attitudinal risks. This new information has important implications for prevention and intervention options, including consideration of tem peramental aspects of the person experiencing (as the per petrator or victim) IPV, such as impulsivity or psychopathy, the mental health of the young person (e.g., experience of depressive symptoms), and attitudinal and cultural factors that may impact on experiences of IPV. *Abajobir, A. A., Kisely, S., Williams, G. M., Clavarino, A. M., & Naj man, J. M. (2017). Substantiated Childhood Maltreatment and Intimate Partner Violence Victimization in Young Adulthood: A Birth Cohort Study. Journal Of Youth And Adolescence, 46(1), 165–179. https://doi.org/10.1007/s10964-016-0558-3 Babcock, J. C., Snead, A. L., Bennett, V. E., & Armenti, N. A. (2019). Distinguishing Subtypes of Mutual Violence in the Context of Self-Defense: Classifying Types of Partner Violent Couples Using a Modified Conflict Tactics Scale. Journal of family violence, 34(7), 687–696. https://doi.org/10.1007/s10896-018-0012-2 Bates, E. A. (2016). Current controversies within intimate partner violence: Overlooking bidirectional violence. Journal of family violence, 31(8), 937–940 Caetano, R., Ramisetty-Mikler, S., & McGrath, C. (2004). Accultura tion, drinking, and intimate partner violence among Hispanic couples in the United States: A longitudinal study. Hispanic Jour nal of Behavioral Sciences, 26, 60–79 Campbell, J. C. (2002). Health consequences of intimate part ner violence. The Lancet, 359(9314), 1331–1336. https://doi. org/10.1016/S0140-6736(02)08336-8 Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s10896- 022-00451-0. Capaldi, D. M., & Langhinrichsen-Rohling, J. (2012). Informing Intimate Partner Violence Prevention Efforts: Dyadic, Develop mental, and Contextual Considerations. Prevention Science, 13, 323–328 Funding Open Access funding enabled and organized by CAUL and its Member Institutions Cohen, J. A., Mannarino, A. P., & Iyengar, S. (2011). Community Treatment of Posttraumatic Stress Disorder for Children Exposed to Intimate Partner Violence: A Randomized Controlled Trial. Archives of Pediatrics & Adolescent Medicine, 165(1), 16–21. https://doi.org/10.1001/archpediatrics.2010.247 Implications The current review highlights the importance of addressing key risk factors for IPV perpetration and victimization in childhood and adolescence from a primary and secondary prevention perspective, before these risk factors escalate into IPV experiences. In particular, many of the risk fac tors identified in this review, and within Costa et al., (2015) are factors that may not be obvious or identified early. This emphasizes the importance of policy change which supports population-wide strategies that may reduce the likelihood of these risk factors occurring or preventing them from progressing from risk factors into experiences of violent behavior. Firstly, as gender normative attitudes were identified as playing a role in later experiences of IPV, primary prevention strategies focused on this (i.e., mass education campaigns about gender-related attitudes) are important. Further, early intervention efforts for those considered at-risk (secondary prevention), such as those children identified as engaging in disruptive or impulsive behavior, or indeed who have been identified as having experienced abuse, may reduce the like lihood of IPV perpetration and victimization as those chil dren progress to adulthood. Such strategies may include the Consistent with Costa et al., (2015) this review was unable to separate adulthood experiences of IPV into those which were situational in nature, whereby violence results from a mutual escalation of conflict between partners (John son & Leone, 2005), and those which would be considered intimate terrorism, described as the need for one partner to exert control over the other, typically through the use of dominance and violence (Johnson, 2006b). This is primarily 1 3 1604 Journal of Family Violence (2023) 38:1591–1606 ment Research Fund, Foundation for Alcohol Research and Education, Cancer Council Victoria, Central Australian Aboriginal Congress, Northern Territory government, Australian Rechabites Founda tion, Northern Territory Primary Health Network, Lives Lived Well, Queensland government and Australian Drug Foundation, travel and related costs from Queensland Police Service, Queensland Office of Liquor Gaming and Racing and the Australasian Drug Strategy Con ference. He has acted as a paid expert witness on behalf of a licensed venue and a security firm. John Toumbourou is a volunteer Director of the not-for-profit company Communities That Care Ltd, which pro vides prevention training to municipal coalitions. Declarations Coker, A. L., Davis, K. E., Arias, I., Desai, S., Sanderson, M., Brandt, H. M., & Smith, P. H. (2002). Physical and mental health effects Conflict of interest Peter Miller receives funding from Australian Re search Council and Australian National Health and Medical Research Council, grants from NSW Government, National Drug Law Enforce 1 3 Journal of Family Violence (2023) 38:1591–1606 1605 of intimate partner violence for men and women. American jour nal of preventive medicine, 23(4), 260–268 Jones, R. T., Browne, K., & Chou, S. (2017). A critique of the revised Conflict Tactics Scales-2 (CTS-2). Aggression and Violent Behav ior, 37, 83–90. https://doi.org/10.1016/j.avb.2017.08.005 Costa, B. M., Kaestle, C. E., Walker, A., Curtis, A., Day, A., Toum bourou, J. W., & Miller, P. (2015). Longitudinal predictors of domestic violence perpetration and victimization: A systematic review. Aggression and Violent Behavior, 24, 261–272 Karakurt, G., Koç, E., Çetinsaya, E. E., Ayluçtarhan, Z., & Bolen, S. (2019). Meta-analysis and systematic review for the treatment of perpetrators of intimate partner violence. Neuroscience and biobehavioral reviews, 105, 220–230. https://doi.org/10.1016/j. neubiorev.2019.08.006 Dupont-Reyes, M., Fry, D., Rickert, V., & Davidson, L. L. (2015). Adolescent Relationship Violence and Acculturation Among NYC Latinos. Maternal and Child Health Journal, 19(7), 1543– 1552. https://doi.org/10.1007/s10995-014-1659-9 Langhinrichsen-Rohling, J., Misra, T. A., Selwyn, C., & Rohling, M. L. (2012). Rates of bidirectional versus unidirectional intimate partner violence across samples, sexual orientations, and race/eth nicities: A comprehensive review. Partner Abuse, 3(2), 199–230 Fergusson, D., Bowden, J., & Horwood, L. J. (2008). Developmen tal antecedents of interpartner violence in a New Zealand birth cohort. Journal of Family Violence, 28(8), 737–753. https://doi. org/10.1037/0012-1649.35.1.258 Laskey, P., Bates, E. A., & Taylor, J. C. (2019). A systematic litera ture review of intimate partner violence victimisation: An inclu sive review across gender and sexuality. Aggression and Violent Behavior, 47, 1–11. https://doi.org/10.1016/j.avb.2019.02.014 Forgatch, M. S., Patterson, G. R., Degarmo, D. S., & Beldavs, Z. G. (2009). Testing the Oregon delinquency model with 9-year follow-up of the Oregon Divorce Study. Development And Psychopathology, 21(2), 637–660. https://doi.org/10.1017/ s0954579409000340 Love, H. A., Spencer, C. M., May, S. A., Mendez, M., & Stith, S. M. (2018). Perpetrator Risk Markers for Intimate Ter rorism and Situational Couple Violence: A Meta-Analy sis. Trauma Violence & Abuse, 21(5), 922–931. https://doi. org/10.1177/1524838018801331 Goodnight, J. A., Bates, J. E., Holtzworth-Munroe, A., Pettit, G. S., Ballard, R. H., Iskander, J. M., Swanson, A., Dodge, K. A., & Lansford, J. E. (2017). Declarations Dispositional, demographic, and social predictors of trajectories of intimate partner aggression in early adulthood [doi:https://doi.org/10.1037/ccp0000226].950–965. https://doi.org/10.1037/ccp0000226 Mackay, J., Bowen, E., Walker, K., & O’Doherty, L. (2018). Risk factors for female perpetrators of intimate partner violence within criminal justice settings: A systematic review. Aggression and Violent Behavior, 41, 128–146. https://doi.org/10.1016/j. avb.2018.06.004fi Graham, L. M., Embry, V., Young, B. R., Macy, R. J., Moracco, K. E., Reyes, H. L. M., & Martin, S. L. (2021). Evaluations of Prevention Programs for Sexual, Dating, and Intimate Partner Violence for Boys and Men: A Systematic Review. Trauma Violence & Abuse, 22(3), 439–465. https://doi.org/10.1177/1524838019851158 Magdol, L., Moffitt, T. E., Caspi, A., & Silva, P. A. (1998). Develop mental antecedents of partner abuse: a prospective-longitudinal study. Journal of abnormal psychology, 107(3), 375–389. https:// doi.org/10.1037//0021-843x.107.3.375 Grest, C. V., Amaro, H., & Unger, J. (2018). Longitudinal Predictors of Intimate Partner Violence Perpetration and Victimization in Latino Emerging Adults. Journal Of Youth And Adolescence, 47(3), 560–574. https://doi.org/10.1007/s10964-017-0663-y Manchikanti Gómez, A. (2010). Testing the Cycle of Violence Hypoth esis: Child Abuse and Adolescent Dating Violence as Predictors of Intimate Partner Violence in Young Adulthood. Youth & Soci ety, 43(1), 171–192. https://doi.org/10.1177/0044118X09358313f Grest, C. V., Lee, J. O., Gilreath, T., & Unger, J. B. (2018). Latent Class Analysis of Intimate Partner Violence Perpetration and Victimization among Latino Emerging Adults. Journal Of Youth And Adolescence, 47(3), 575–585. https://doi.org/10.1007/ s10964-017-0807-0 Menting, A. T., de Castro, O., B., & Matthys, W. (2013). Effective ness of the Incredible Years parent training to modify disruptive and prosocial child behavior: a meta-analytic review. Clinical Psychology Review, 33(8), 901–913. https://doi.org/10.1016/j. cpr.2013.07.006 p Messinger, A. M. (2011). Invisible Victims: Same-Sex IPV in the National Violence Against Women Survey. Journal of Interpersonal Violence, 26(11), 2228–2243. https://doi. org/10.1177/0886260510383023 Handley, E. D., Russotti, J., Rogosch, F. A., & Cicchetti, D. (2019). Developmental cascades from child maltreatment to negative friend and romantic interactions in emerging adulthood. Devel opment and Psychopathology, 31(5), 1649–1659. https://doi. org/10.1017/S095457941900124Xf Milaniak, I., & Widom, C. S. (2015). Does child abuse and neglect increase risk for perpetration of violence inside and outside the home? Psychology of violence, 5(3), 246fi Herrenkohl, T. I., & Jung, H. (2016). Effects of child abuse, adolescent violence, peer approval and pro-violence attitudes on intimate partner violence in adulthood. Criminal Behaviour and Mental Health, 26(4), 304–314. https://doi.org/10.1002/cbm.2014l Moffitt, T. E., Caspi, A., Harrington, H., & Milne, B. J. (2002). Declarations Males on the life-course-persistent and adolescence-limited antisocial pathways: Follow-up at age 26 years. Development and Psychopathology, 14(1), 179–207. https://doi.org/10.1017/ S0954579402001104 Johnson, M. P. (2006a). Conflict and control: gender symmetry and asymmetry in domestic violence. Violence Against Women, 12(11), 1003–1018. https://doi.org/10.1177/1077801206293328 Murray, J., Farrington, D. P., & Eisner, M. P. (2009). Drawing conclu sions about causes from systematic reviews of risk factors: The Cambridge Quality Checklists. J Exp Criminol, 5, 1–23 Johnson, M. P. (2006b). Gendered communication and intimate part ner violence. In B. J. Dow, & J. T. Wood (Eds.), The SAGE hand book of gender and communication (pp. 71–87). Sageff Narayan, A. J., Labella, M. H., Englund, M. M., Carlson, E. A., & Egeland, B. (2017). The legacy of early childhood violence expo sure to adulthood intimate partner violence: Variable- and person- oriented evidence. Journal of family psychology: JFP : journal of the Division of Family Psychology of the American Psychological Association (Division, 43(7), 833–843. https://doi.org/10.1037/ fam0000327. 31 Johnson, M. P., & Leone, J. M. (2005). The differential effects of inti mate terrorism and situational couple violence: Findings from the National Violence Against Women Survey. Journal of Family Issues, 26, 322–349 Johnson, M. P., Leone, J. M., & Xu, Y. (2014). Intimate Terrorism and Situational Couple Violence in General Surveys: Ex-Spouses Required. Violence Against Women, 20(2), 186–207. https://doi. org/10.1177/1077801214521324f Neppl, T. K., Lohman, B. J., Senia, J. M., Kavanaugh, S. A., & Cui, M. (2019). Intergenerational continuity of psychological violence: Intimate partner relationships and harsh parenting. Psychology of violence, 9(3), 298–307. https://doi.org/10.1037/vio0000129 Jolliffe, D., Murray, J., Farrington, D., & Vannick, C. (2012). Testing the Cambridge Quality Checklists on a review of disrupted fami lies and crime. 22(5),303–314. https://doi.org/10.1002/cbm.1837 1 1 3 1606 Journal of Family Violence (2023) 38:1591–1606 O’Leary, K. D., Slep, A. M. S., Avery-Leaf, S., & Cascardi, M. (2008). Gender differences in dating aggression among multiethnic high school students. Journal of Adolescent Health, 42, 473–479f Stephenson, R., & Finneran, C. (2013). The IPV-GBM Scale: A New Scale to Measure Intimate Partner Violence among Gay and Bisexual Men. Plos One, 8(6), e62592. https://doi.org/10.1371/ journal.pone.0062592 Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Bren nan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDon ald, S., & Moher, D. Declarations (2021). The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Systematic Reviews, 10(1), 89. https://doi.org/10.1186/s13643-021-01626-4 Theobald, D., Farrington, D. P., Coid, J. W., & Piquero, A. R. (2016). A longitudinal analysis of the criminal careers of intimate part ner violence offender subtypes: Results from a prospective sur vey of males. Violence And Victims, 31(6), 999–1020. https://doi. org/10.1891/0886-6708.VV-D-14-00194 Toumbourou, J. W., Rowland, B., Williams, J., Smith, R., & Patton, G. C. (2019). Community intervention to prevent adolescent health behavior problems: Evaluation of Communities that Care in Australia. Health Psychology, 38(6), 536–544. https://doi. org/10.1037/hea0000735f Regan, K. V., Bartholomew, K., Oram, D., & Landolt, M. A. (2002). Measuring Physical Violence in Male Same-Sex Relationships. Journal of Interpersonal Violence, 17(3), 235–252. https://doi. org/10.1177/0886260502017003001 Renner, L. M., & Whitney, S. D. (2012). Risk factors for unidirectional and bidirectional intimate partner violence among young adults. Child Abuse & Neglect, 36, 40–52 Ulloa, E. C., & Hammett, J. F. (2016). The effect of gender and perpe trator–victim role on mental health outcomes and risk behaviors associated with intimate partner violence. Journal of Interper sonal Violence, 31, 1184–1207 Reynolds, A. J., Temple, J. A., Ou, S. R., Robertson, D. L., Mersky, J. P., Topitzes, J. W., & Niles, M. D. (2007). Effects of a school- based, early childhood intervention on adult health and well- being: a 19-year follow-up of low-income families. Archives of Pediatrics and Adolescent Medicine, 161(8), 730–739 Veritas Health Innovation. (2014). Covidence systematic review soft ware. Veritas Health Innovation. www.covidence.org Vézina, J., Hébert, M., Poulin, F., Lavoie, F., Vitaro, F., & Tremblay, R. E. (2015). History of family violence, childhood behavior prob lems, and adolescent high-risk behaviors as predictors of girls’ repeated patterns of dating victimization in two developmental periods. Violence Against Women, 21(4), 435–459. https://doi. org/10.1177/1077801215570481 Scheeringa, M. S., Weems, C. F., Cohen, J. A., Amaya-Jackson, L., & Guthrie, D. (2011). Trauma-focused cognitive-behavioral therapy for posttraumatic stress disorder in three-through six year-old children: a randomized clinical trial. Journal of Child Psychology and Psychiatry, 52(8), 853–860. https://doi. org/10.1111/j.1469-7610.2010.02354.x Zamir, O., Szepsenwol, O., Englund, M. M., & Simpson, J. A. (2018). J. C. a., & neglect. The role of dissociation in revictimization across the lifespan: A 32-year prospective study. 79, 144–153 Smith, P., Yule, W., Perrin, S., Tranah, T., Dalgleish, T. I. M., & Clark, D. M. (2007). Cognitive-Behavioral Therapy for PTSD in Children and Adolescents: A Preliminary Randomized Con trolled Trial. Publisher’s Note Springer Nature remains neutral with regard to juris dictional claims in published maps and institutional affiliations. Declarations Journal of the American Academy of Child & Ado lescent Psychiatry, 46(8), 1051–1061. https://doi.org/10.1097/ CHI.0b013e318067e288 Publisher’s Note Springer Nature remains neutral with regard to juris dictional claims in published maps and institutional affiliations. Publisher’s Note Springer Nature remains neutral with regard to juris dictional claims in published maps and institutional affiliations. 1 3 1 3
https://openalex.org/W2049339234	https://europepmc.org/articles/pmc2099484?pdf=render	English	null	Computational analysis of biological functions and pathways collectively targeted by co-expressed microRNAs in cancer	BMC bioinformatics	2,007	cc-by	10,923	BioMed Central BMC Bioinformatics Open Access Proceedings Computational analysis of biological functions and pathways collectively targeted by co-expressed microRNAs in cancer Yuriy Gusev1, Thomas D Schmittgen2, Megan Lerner1,3, Russell Postier1 and Daniel Brackett1,3 Address: 1Department of Surgery, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma, USA, 2Division of Pharmaceutics, Ohio State University, Columbus, Ohio, USA and 3Veterans Affairs Medical Hospital, Oklahoma City, Oklahoma, USA Email: Yuriy Gusev - yuriy-gusev@ouhsc.edu; Thomas D Schmittgen - schmittgen@pharmacy.ohio-state.edu; Megan Lerner - megan- lerner@ouhsc.edu; Russell Postier - russell-postier@ouhsc.edu; Daniel Brackett - daniel-brackett@ouhsc.edu * Corresponding author BioMed Central BMC Bioinformatics Open Access Proceedings Computational analysis of biological functions and pathways collectively targeted by co-expressed microRNAs in cancer Yuriy Gusev1, Thomas D Schmittgen2, Megan Lerner1,3, Russell Postier1 and Daniel Brackett1,3 Address: 1Department of Surgery, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma, USA, 2Division of Pharmaceutics, Ohio State University, Columbus, Ohio, USA and 3Veterans Affairs Medical Hospital, Oklahoma City, Oklahoma, USA Email: Yuriy Gusev - yuriy-gusev@ouhsc.edu; Thomas D Schmittgen - schmittgen@pharmacy.ohio-state.edu; Megan Lerner - megan- lerner@ouhsc.edu; Russell Postier - russell-postier@ouhsc.edu; Daniel Brackett - daniel-brackett@ouhsc.edu * Corresponding author BioMed Central Open Acce Proceedings Computational analysis of biological functions and pathways collectively targeted by co-expressed microRNAs in cancer Yuriy Gusev1, Thomas D Schmittgen2, Megan Lerner1,3, Russell Postier1 and Daniel Brackett1,3 Open Access Address: 1Department of Surgery, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma, USA, 2Division of Pharmaceutics, Ohio State University, Columbus, Ohio, USA and 3Veterans Affairs Medical Hospital, Oklahoma City, Oklahoma, USA Address: 1Department of Surgery, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma, USA, 2Division of Pharmaceutics, Ohio State University, Columbus, Ohio, USA and 3Veterans Affairs Medical Hospital, Oklahoma City, Oklahoma, USA Email: Yuriy Gusev - yuriy-gusev@ouhsc.edu; Thomas D Schmittgen - schmittgen@pharmacy.ohio-state.edu; Megan Lerner - megan- lerner@ouhsc.edu; Russell Postier - russell-postier@ouhsc.edu; Daniel Brackett - daniel-brackett@ouhsc.edu * Corresponding author from Fourth Annual MCBIOS Conference. Computational Frontiers in Biomedicine New Orleans, LA, USA. 1–3 February 2007 Published: 1 November 2007 Published: 1 November 2007 BMC Bioinformatics 2007, 8(Suppl 7):S16 doi:10.1186/1471-2105-8-S7-S16 This article is available from: http://www.biomedcentral.com/1471-2105/8/S7/S16 © 2007 Gusev et al; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background While bioinformatics methods continue to improve spe- cificity and sensitivity of target predictions, the unresolved challenge still remains to utilize even the most accurate predictions for biologically meaningful interpretation of miRNA profiling data. This problem is due to the fact that the majority of known miRNAs are predicted to target a very large number of transcripts. Each miRNA might have up to several hundred targets. In addition, many tran- scripts from protein-coding genes are targeted by more than one miRNA and some transcripts might have over a hundred target sites for different miRNAs. It has been esti- mated that nearly 50% of all human gene transcripts are regulated by relatively small number of 474 miRNAs that are known to date with average of ~200 targets per miRNA [13,14]. g microRNAs (miRNAs) are single-stranded, non-coding RNAs approximately 22 nucleotides in length that post- transcriptionally regulate gene expression in a sequence specific manner by binding with imperfect complimenta- rity at multiple sites of 3'-UTR regions of mRNA, thereby facilitating mRNA degradation or inhibiting translation initiation (reviewed in [1,2]). First discovered in C. ele- gans by Victor Ambros [3], miRNAs are shown to be important negative feedback regulators of many biologi- cal processes such as development, differentiation, and proliferation in vertebrates and plants. Aberrant expres- sion of miRNAs in human cancer was first reported in 2002 for leukemia [4]. Since then multiple studies have found aberrant expres- sion profiles of miRNAome in all major human cancers (reviewed in [5]). While several target genes were experi- mentally identified for some miRNAs in various tumors [6], the global pattern of cellular functions and pathways that are affected by miRNAs in cancer remains elusive. This conundrum of target multiplicity has become espe- cially evident with the discovery of a large number of dif- ferentially expressed miRNAs in many human cancers [4,9]. A significant number of miRNAs in a range from ten to one-hundred were found to be aberrantly expressed in breast cancer, colon cancer, lung cancer, and other major cancers with a predicted total number of targets ranging from several hundred to as many as several thousand. This again presents a problem for global analysis and the bio- logical interpretation of the regulatory impact of miRNAs in cancer cells. Background There is a clear need for data reduction methods which would allow reducing the list of targets and determining cellular processes that are most signifi- cantly affected by miRNAs in cancer. In our recent studies we analyzed expression profiles of more than 200 miRNAs in human cancer cells lines [7] and samples of pancreatic tumors [8]. Using high- throughput real-time PCR we found over one-hundred miRNAs to be differentially expressed in pancreatic cancer in comparison to normal pancreatic tissue. A cluster of forty-seven most significantly over-expressed miRNAs from this study was included in the current analysis along with 4 other datasets published by others (breast, colon and lung cancers [9], and lymphomas [10,11]). The Gene Ontology (GO) enrichment analysis is one of the data reduction techniques that could be used to reduce the number of targets of a large group of co-expressed miRNAs and to find biological functions potentially affected by multiple miRNAs. miRNA expression signatures are shown to be specific and allow classification of tumor type as well as different stages in tumor progression and in some cases predict out- come of a disease [12]. A number of studies have shown that expression of some of the proteins affected in cancer was negatively correlated with the expression of specific miRNAs. Based on these findings, several groups have hypothesized that miRNAs may play important role in tumorigenesis and tumor progression and could function as oncogenes or tumor suppressor genes [4,9]. However, such global interpretation of miRNA expression profiling data is impaired by the lack of high throughput target val- idation methods and mostly relies upon computational analysis of potential mRNA targets. The concept of combinatorial target regulation by miR- NAs has been discussed in the literature [2,13,14] and was incorporated into several current prediction algorithms such as PicTar [13,14] and miRgate (Actigenics/Cepheid) [16,17]. It is based on experimental evidence that some co-expressed miRNAs may all target the same genes or genes from the same functional categories. Several studies have reported results of computational analysis of func- tional annotation of genes targeted by single miRNAs [18], all known miRNAs [19,20], or small groups of miR- NAs that were selected based on high similarity of "seed" sequences in the 5' region and/or large overlap of pre- dicted target sets [21]. Abstract Background: Multiple recent studies have found aberrant expression profiles of microRNAome in human cancers. While several target genes have been experimentally identified for some microRNAs in various tumors, the global pattern of cellular functions and pathways affected by co- expressed microRNAs in cancer remains elusive. The goal of this study was to develop a computational approach to global analysis of the major biological processes and signaling pathways that are most likely to be affected collectively by co-expressed microRNAs in cancer cells. Results: We report results of computational analysis of five datasets of aberrantly expressed microRNAs in five human cancers published by the authors (pancreatic cancer) and others (breast cancer, colon cancer, lung cancer and lymphoma). Using the combinatorial target prediction algorithm miRgate and a two-step data reduction procedure we have determined Gene Ontology categories as well as biological functions, disease categories, toxicological categories and signaling pathways that are: targeted by multiple microRNAs; statistically significantly enriched with target genes; and known to be affected in specific cancers. Conclusion: Our global analysis of predicted miRNA targets suggests that co-expressed miRNAs collectively provide systemic compensatory response to the abnormal phenotypic changes in cancer cells by targeting a broad range of functional categories and signaling pathways known to be affected in a particular cancer. Such systems biology based approach provides new avenues for biological interpretation of miRNA profiling data and generation of experimentally testable hypotheses regarding collective regulatory functions of miRNA in cancer. Page 1 of 17 (page number not for citation purposes) http://www.biomedcentral.com/1471-2105/8/S7/S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 BMC Bioinformatics 2007, 8(Suppl 7):S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 Page 2 of 17 (page number not for citation purposes) Background However, in case of experimentally obtained miRNA profiling data these approaches are not very practical when the task is to determine common bio- logical functions and regulatory pathways that are tar- geted by experimentally detected groups of co-expressed miRNAs. Specifically in cancer such groups of miRNAs are Computational algorithms have played a central role in discovery of the majority of miRNAs known to date, as well as in prediction of their targets (reviewed in [13-15]). However, virtually all existing programs generate rela- tively high levels of false positive predictions (up to a twenty percent). Experimental evidence also suggests that these programs generate some false negative predictions. Page 2 of 17 (page number not for citation purposes) http://www.biomedcentral.com/1471-2105/8/S7/S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 BMC Bioinformatics 2007, 8(Suppl 7):S16 BMC Bioinformatics 2007, 8(Suppl 7):S16 BMC Bioinformatics 2007, 8(Suppl 7):S16 often found to have fewer common targeted genes and not to share similar "seed" sequences. often found to have fewer common targeted genes and not to share similar "seed" sequences. sent the whole spectrum of group sizes of co-expressed miRNAs that are observed in cancers: from a small set of several co-expressed miRNAs (7 miRNAs, cistron miR-17- 92) to a large set (47 miRNAs, pancreatic cancer) to avoid possible bias of sample size. In this study we address this problem of biological inter- pretation of miRNA profiling data using systems biology analysis of major biological proccesses, disease categories and signalling pathways that are targeted collectively by co-expressed miRNAs in cancer cells. We assumed that fil- tering GO categories on the total number of hits by miR- NAs targeting the same category would reduce the number of false positive target predictions and at the same time would allow narrowing down the large target lists and determining those biological functions and pathways that are most likely to be affected by co-expressed miRNAs. Results Data sets Five groups of co-expressed miRNAs were selected from the literature for this study: 3 groups that were reported by Calin et al. [9] as being overexpressed in breast cancer (14 miRNAs), colon cancer (20 miRNAs) and lung cancer (33 miRNAs). We have also included a set of miRNAs that we found to be significantly overexpressed in pancreatic can- cer (47 miRNAs) [8]. An additional group of seven miR- NAs was reported as being overexpressed in lymphomas [10]. This group of miRNAs is encoded by a single gene (cistron miR-17-92) and expressed as a single primary transcript. Overexpression of cistron miR-17-92 was found in B-lymphomas [10] and also was shown to have strong correlation with T-lymphoma development in an animal model [11]. These datasets were selected to repre- Combinatorial analysis of miRNA targets Computational analysis of predicted targets for clusters of overexpressed miRNAs was performed using novel combi- natorial target prediction algorithms (miRgate 2.1 suite, Actigenics/Cepheid [16,17]). Similar to other known algorithms, a list of potential target sites (conserved between human and mouse) on 3'-UTRs of human gene- coding transcripts was determined based on a search for complimentary binding sites for the sequence of "seed" regions of 5'ends of mature miRNA positioned at 2–8 nucleotide region. These target lists were then analyzed by several statistical methods. A flow diagram of data analy- sis is shown on Figure 1. Flow Diagram of Data Analysis Figure 1 Flow Diagram of Data Analysis. Gene Ontology enrichment analysis As a first step, the Gene Ontology (GO) [22] enrichment analysis [23] of biological processes targeted by each of five groups of miRNAs was performed using the miRgate GO profiling algorithm that is specifically designed to take in account information about the number of miRNAs that are targeting the same genes and GO categories i.e. number of miRNA hits per GO category (miRgate 2.1 suite [16,17]). The GO categories were determined for all pre- dicted targets of miRNAs from each of the five groups. This set of GO categories was then filtered based on signif- Flow Diagram of Data Analysis Figure 1 Flow Diagram of Data Analysis. miRgate Target Prediction miRgate GO Enrichment Analysis p 0.05 Ingenuity Pathway Enrichment Analysis: known cancer genes vs predicted microRNA targets p 0.05 Heatmaps UH Clustering microRNA Group Or P fil miRgate GO Enrichment Analysis p 0.05 microRNA Group Or P fil miRgate Target Prediction Heatmaps UH Clustering Or P fil Ingenuity Pathway Enrichment Analysis: known cancer genes vs predicted microRNA targets p 0.05 Page 3 of 17 (page number not for citation purposes) http://www.biomedcentral.com/1471-2105/8/S7/S16 BMC Bioinformatics 2007, 8(Suppl 7):S16 eatmap of unsupervised hierarchical clustering of all Gene ntology (GO) categories obtained by enrichment analysis r 5 datasets of miRNAs gure 2 eatmap of unsupervised hierarchical clustering of Gene Ontology (GO) categories obtained by richment analysis for 5 datasets of miRNAs. Color adient represents number of miRNAs targeting each cate ry # microRNAs Gene Ontology BP Categories icance of overrepresentation using a selected threshold for p-values of hypergeometric distribution (p ≤ 0.05) [17]. An additional filter could then be applied to select only those overrepresented GO categories which are targeted by at least several miRNAs. For our study we have gener- ated lists of all overrepresented GO categories for each data set of over-expressed miRNAs from five types of can- cer. The number of miRNAs targeting the same GO cate- gory was included in the analysis and was used as the parameter (Nm) under investigation. Unsupervised hierarchical clustering of GO categories The resulting matrices of enriched GO categories that were affected by groups of co-expressed miRNA were further analyzed by unsupervised hierarchical clustering using information about the number of miRNAs targeted each category to find similarities and differences in patterns of affected biological functions for five types of cancer (Fig. 2). Using uncentered Pearson correlation as a distance the clusters of GO categories that were specific for each cancer were identified with a surprisingly low level of overlap. Gene Ontology BP Categories Subsets of GO categories were generated by trimming the original matrix on the minimal number of miRNAs target- ing the same category (Nm ≥ threshold) and subjected to hierarchical clustering to find categories targeted by the number of miRNAs above the threshold. Resulting clus- ters were compared among data sets from different can- cers to determine how this trimming affects the clustering results. We observed that trimming on Nm actually improved the separation of clusters of GO categories that are specific for each cancer. As an example a heat map with clustering results for Nm ≥ 6 (i.e. at least 6 miRNAs targeting each GO category) is included on Figure 3. A more detailed view of the same clusters is presented as a bar graph (Figure 4). Each bar represents a single GO category for each of five types of cancer and showing the number of miRNAs that target the same category. These clusters of GO Biological Processes contain many categories known to be affected in cancer such as general categories of proliferation, regulation of cell cycle, and transcription, as well as more specific cate- gories such as ras-oncogene signaling pathway, chromo- some segregation, and others. Systems Biology Analysis of predicted miRNA targets 0 10 20 30 40 50 # microRNAs protein kinase cascade regulation of cell migration transcription metabolic process regulation of RNA export from nucleus potassium ion transport N-glycan processing mitosis amino acid biosynthetic process metal ion transport positive regulation of protein biosynthetic process G-protein signaling, adenylate cyclase inhibiting pathway apical junction assembly glycerol-3-phosphate metabolic process sodium ion transport pinocytosis extracellular matrix organization and biogenesis activation of NF-kappaB transcription factor negative regulation of follicle-stimulating hormone secretion positive regulation of fatty acid oxidation protein ubiquitination during ubiquitin-dependent protein catabolic process feeding behavior G1 phase of mitotic cell cycle protein complex assembly regulation of smooth muscle contraction sulfate transport positive regulation of follicle-stimulating hormone secretion negative regulation of keratinocyte differentiation transcription from RNA polymerase II promoter induction of apoptosis macrophage activation response to radiation t bli h t f t i l li ti Cis17_92 Breast Cancer Colon Cancer Lung Cancer Pancreatic Cancer Cis17_92 Breast Cancer Colon Cancer Lung Cancer Pancreatic Cancer Bar graph of the same clusters of GO categories for 5 types of cancer that are presented on the Figure 3 Figure 4 Bar graph of the same clusters of GO categories for 5 types of cancer that are presented on the Figure 3. graph of the same clusters of GO categories for 5 types of cancer that are presented on the Figure 3 re 4 graph of the same clusters of GO categories for 5 types of cancer that are presented on the Figure 3. 2. Targets that belong to GO categories targeted by at least 50% of the co-expressed miRNAs; 2. Targets that belong to GO categories targeted by at least 50% of the co-expressed miRNAs; ries, canonical signaling pathways, and drugs associated with each data set. 3. Targets that belong to GO categories targeted by 100% of the co-expressed miRNAs. 3. Targets that belong to GO categories targeted by 100% of the co-expressed miRNAs. To select most important functional categories and path- ways for analysis and to understand relevance of affected categories to the specific cancer, we generated reference sets of genes for each of five cancers by keyword search of the Ingenuity Knowledge Base. This search provided us with very conservative lists of genes that were reported to be affected in a specific cancer by multiple research groups and were then manually curated by a group of expert biol- ogists. Systems Biology Analysis of predicted miRNA targets Systems Biology Analysis of predicted miRNA targets For each set of co-expressed miRNAs we have generated 3 sets of predicted gene targets using GO ontology enrich- ment analysis as a statistical filter: Heatmap of unsupervised hierarchical clustering of all Gene Ontology (GO) categories obtained by enrichment analysis for 5 datasets of miRNAs Figure 2 Heatmap of unsupervised hierarchical clustering of all Gene Ontology (GO) categories obtained by enrichment analysis for 5 datasets of miRNAs. Color gradient represents number of miRNAs targeting each cate- gory. 1. Targets that belong to GO categories targeted by at least one of the co-expressed miRNAs; Page 4 of 17 (page number not for citation purposes) BMC Bioinformatics 2007, 8(Suppl 7):S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 map of unsupervised hierarchical clustering of a subset of GO categories that are targeted by al least 6 miRNAs eac re 3 map of unsupervised hierarchical clustering of a subset of GO categories that are targeted by al least 6 miRNAs eac #microRNAs Gene Ontology BP Categories Gene Ontology BP Categories Gene Ontology BP Categories Heatmap of unsupervised hierarchical clustering of a subset of GO categories that are targeted by al least 6 miRNAs each Figure 3 Heatmap of unsupervised hierarchical clustering of a subset of GO categories that are targeted by al least 6 miRNAs each. Heatmap of unsupervised hierarchical clustering of a subset of GO categories that are targeted by al least 6 miRNAs each Figure 3 Heatmap of unsupervised hierarchical clustering of a subset of GO categories that are targeted by al least 6 miRNAs each. Page 5 of 17 (page number not for citation purposes) BMC Bioinformatics 2007, 8(Suppl 7):S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 Bar graph of the same clusters of GO categories for 5 types of cancer that are presented on the Figure 3 Figure 4 Bar graph of the same clusters of GO categories for 5 types of cancer that are presented on the Figure 3. Comparative analysis of toxicology categories Comparative analysis of toxicology categories Using IPA 5.0 we have also analyzed top ranked toxicol- ogy related gene lists for each of five reference gene lists and compared them with toxicology categories found in our miRNA target lists (Additional File 4). We found that 8 top ranked toxicology gene lists for each cancer were statistically significantly overrepresented among miRNA targets (Additional File 4). We found it particularly interesting that several categories related to oxidative stress and hypoxia were among top ranked over- represented categories for miRNA targeted genes (Addi- tional File 4). These findings are in agreements with recent experimental data reporting over expression of multiple miRNAs in response to oxidative stress or hypoxia [26] and showing a functional link between hypoxia, a well- known tumor microenvironment factor, and microRNA expression. Systems Biology Analysis of predicted miRNA targets These 5 reference sets of genes known to be affected in lymphoma, breast cancer, colon cancer, lung cancer, and pancreatic cancer were used in the Ingenuity Pathway Analysis system to generate sets of biological functions, disease categories, and pathways known to be most affected in each of these cancers. We started with raw lists of all targets predicted for each of the miRNAs in a range from 2175 genes (cistron mir-17- 92) to 5356 genes (pancreatic cancer). Filtering raw sets of predicted targets by miRgate GO enrichment analysis algorithm provided significant reduction of target lists in a range from 2.5 fold to over 4 fold (Table 1). Three sets of genes were generated for each of 5 groups of miRNAs and were then analyzed by Ingenuity Pathway Analysis tools to determine major biological functions, disease categories, toxicological and pathological catego- Page 6 of 17 (page number not for citation purposes) BMC Bioinformatics 2007, 8(Suppl 7):S16 Table 1: Number of predicted targets for 5 datasets miRNA Data Set Number of co- expressed miRNAs Total Number of Targets Total number of targets from enrichment analysis (GO BP), p < 0.05 Number of targets from GO BP categories targeted by >50% of miRNAs Number of targets from GO BP categories targeted by 100% of miRNAs Lymphoma (cistron miR-17-92 only) 7 2175 962 918 851 Breast Cancer 14 3462 805 720 546 Colon Cancer 20 3439 828 711 578 Lung Cancer 33 4942 700 560 192 Pancreatic Cancer 47 5356 996 841 702 Table 1: Number of predicted targets for 5 datasets Comparative analysis of biological functions and disease categories represented functional categories for colon cancer included cell death, cell growth and proliferation, cell cycle, cell movement, and DNA replication and repair (Additional File 1) and we found all of these categories to be significantly overrepresented in our sets of predicted miRNA targets. Similar results were obtained for pancre- atic cancer (Additional File 2). To evaluate the specific functional categories of genes from broad GO categories that are targeted by miRNAs, we performed more detailed analysis using IPA 5.0 (Inge- nuity Systems, Redwood, CA). We compared gene sets determined by GO enrichment algorithm against groups of genes known to be affected in specific cancer to deter- mine which top ranked categories would be statistically enriched with miRNA targets. The results indicate that many top ranked biological functions and disease catego- ries as well as toxicological categories that were tissue spe- cific for each specific cancer were also statistically significantly overrepresented in our target lists. Top bio- logical functions and disease related categories were com- pared among 5 groups of data using gene lists generated by trimming miRNA collectively targeted genes at the 50% level (Fig. 5). The top ranked disease category for all 5 datasets was Cancer (Fig. 5B) with highly significant enrichment (p ~ 10-10 ÷ 10-20). Within this top category we identified a large number of genes known as tissue specific biomarkers of each of 5 cancers. For example a list of miRNA targets for colon cancer included APC gene (ade- nomatosis polyposis coli) (Table 2) among other well known oncogenes. For pancreatic cancer a list of miRNA targets included both kras and p53 genes (Table 3) that are well known biomarkers of pancreatic tumors [24]. Importantly, several Ras oncogenes were experimentally validated as targets of multiple miRNAs from let-7 family [25]. Overall, in our analysis we identified 25 known can- cer related genes that have been already experimentally validated as targets of miRNAs. Comparative pathway analysis To further evaluate the specific functions of genes from the broad GO categories that are targeted by miRNAs, we performed additional, more detailed pathway analysis (IPA 5.0, Ingenuity Systems). We compared gene sets determined by GO enrichment algorithm against known signaling pathways to determine which pathways would be statistically enriched with miRNA targets. We were also interested in determining which pathways were affected the most by multiple miRNAs from the same co-expressed group in each specific cancer. Other top ranked categories included tissue specific dis- eases such as gastrointestinal systems diseases for colon cancer; reproductive systems diseases for breast cancer etc. (Fig. 5B). The detailed information and statistics of IPA enrichment analysis for all datasets are available in Additional Files (Additional Files 1, 2, and 3). Comparative pathway analysis For example, the top 5 over- To select pathways for analysis and to understand rele- vance of affected pathways to the specific cancer, we used Page 7 of 17 (page number not for citation purposes) Page 7 of 17 (page number not for citation purposes) BMC Bioinformatics 2007, 8(Suppl 7):S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 Table 2: List of predicted targets that are known genes affected in Colon Cancer Name Description Location Type Drugs AKT3 v-akt murine thymoma viral oncogene homolog 3 (protein kinase B, gamma) Cytoplasm kinase enzastaurin APC adenomatosis polyposis coli Nucleus enzyme BAX BCL2-associated X protein Cytoplasm other BCL2L11 BCL2-like 11 (apoptosis facilitator) Cytoplasm other BIRC5 baculoviral IAP repeat-containing 5 (survivin) Cytoplasm other BRCA1 breast cancer 1, early onset Nucleus transcription regulator CCND1 cyclin D1 Nucleus other DIABLO diablo homolog (Drosophila) Cytoplasm other EP300 E1A binding protein p300 Nucleus transcription regulator ERBB3 v-erb-b2 erythroblastic leukemia viral oncogene homolog 3 (avian) Plasma Membrane kinase FASLG Fas ligand (TNF superfamily, member 6) Extracellular Space cytokine FLT1 fms-related tyrosine kinase 1 (vascular endothelial growth factor/ vascular permeability factor receptor) Plasma Membrane kinase sunitinib, axitinib FST follistatin Extracellular Space other GSK3A glycogen synthase kinase 3 alpha Nucleus kinase ITGB1 integrin, beta 1 (fibronectin receptor, beta polypeptide, antigen CD29 includes MDF2, MSK12) Plasma Membrane transmembrane receptor MAP2K1 mitogen-activated protein kinase kinase 1 Cytoplasm kinase PD 0325901 MAPK9 mitogen-activated protein kinase 9 Cytoplasm kinase MAPK14 mitogen-activated protein kinase 14 Cytoplasm kinase SCIO-469 NFAT5 nuclear factor of activated T-cells 5, tonicity-responsive Nucleus transcription regulator NUAK1 NUAK family, SNF1-like kinase, 1 Unknown kinase PARP1 poly (ADP-ribose) polymerase family, member 1 Nucleus enzyme INO-1001 PDCD4 programmed cell death 4 (neoplastic transformation inhibitor) Nucleus other PLK1 polo-like kinase 1 (Drosophila) Nucleus kinase PPP2R1B protein phosphatase 2 (formerly 2A), regulatory subunit A (PR 65), beta isoform Unknown phosphatase PRKCA protein kinase C, alpha Cytoplasm kinase safingol PRKG1 protein kinase, cGMP-dependent, type I Cytoplasm kinase SRC v-src sarcoma (Schmidt-Ruppin A-2) viral oncogene homolog (avian) Cytoplasm kinase dasatinib TGFBR2 transforming growth factor, beta receptor II (70/80kDa) Plasma Membrane kinase YES1 v-yes-1 Yamaguchi sarcoma viral oncogene homolog 1 Cytoplasm kinase dasatinib enriched GO categories and two sets trimmed by exclud- ing genes from those GO categories that are targeted by less than 50 percent or less than 100 percent of miRNAs in a group. Comparative pathway analysis the same reference sets of genes for each of five cancers that were obtained by keyword search of the Ingenuity Knowledge Base. These 5 reference sets of genes known to be affected in lymphomas, breast cancer, colon cancer, lung cancer, and pancreatic cancer were used in the Inge- nuity Pathway Analysis system to analyze sets of pathways known to be affected in each of these cancers. For example, Figure 6 demonstrates the results of compar- ison of top 12 pathways known to be affected by colon cancer and targeted by miRNAs that are over-expressed in colon cancer. Bar color represents 4 data sets: dark blue – reference gene list for colon cancer generated by Ingenuity Knowledgebase; medium blue, light blue and black – 3 gene lists that were obtained from GO enrichment analy- sis; a complete list (all targets) and two sets trimmed by excluding genes from GO categories targeted by less than 50% or less than 100% of miRNAs in the group. The same type of comparison of top 12 pathways for pancreatic can- cer is shown on Figure 7. For each of the five datasets we performed analysis of all canonical signaling pathways known to be affected by each cancer to reveal pathways that are targeted by miR- NAs from our datasets for each particular cancer. We used Fisher's exact test to select pathways that were statistically significantly enriched with miRNA targets (p ≤ 0.05). The total number of targeted genes in each pathway was compared with the number of reference genes from the same pathway that are known to be affected by cancer. For each type of cancer a reference set of genes was compared with three target sets: a complete list of genes from Results of the pathway analysis were similar for all data- sets. Comparative pathway analysis We found that a large fraction of pathways known to Page 8 of 17 (page number not for citation purposes) BMC Bioinformatics 2007, 8(Suppl 7):S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 Table 3: List of predicted targets that are known genes affected in Pancreatic Cancer Name Description Location Type Drugs AKT1 v-akt murine thymoma viral oncogene homolog 1 Cytoplasm kinase enzastaurin BCL2 B-cell CLL/lymphoma 2 Cytoplasm other oblimersen CCKBR cholecystokinin B receptor Plasma Membrane G-protein coupled receptor CR 2945 CDK6 cyclin-dependent kinase 6 Nucleus kinase flavopiridol CTSB cathepsin B Cytoplasm peptidase CTSL2 cathepsin L2 Cytoplasm peptidase E2F1 E2F transcription factor 1 Nucleus transcription regulator FAS Fas (TNF receptor superfamily, member 6) Plasma Membrane transmembrane receptor HGF hepatocyte growth factor (hepapoietin A; scatter factor) Extracellular Space growth factor HMGA1 high mobility group AT-hook 1 Nucleus transcription regulator IL6 interleukin 6 (interferon, beta 2) Extracellular Space cytokine KRAS v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog Cytoplasm enzyme PLAU plasminogen activator, urokinase Extracellular Space peptidase RHOB ras homolog gene family, member B Cytoplasm enzyme THOC1 THO complex 1 Nucleus other TOP1 topoisomerase (DNA) I Nucleus enzyme elsamitrucin, T 0128, CT-2106, BN 80927, tafluposide, TAS-103, irinotecan, topotecan, 9-amino-20- camptothecin, rubitecan, gimatecan, karenitecin TP53 tumor protein p53 (Li-Fraumeni syndrome) Nucleus transcription regulator Table 3: List of predicted targets that are known genes affected in Pancreatic Cancer elsamitrucin, T 0128, CT-2106, BN 80927, tafluposide, TAS-103, irinotecan, topotecan, 9-amino-20- camptothecin, rubitecan, gimatecan, karenitecin tumor protein p53 (Li-Fraumeni syndrome) Nucleus TP53 be affected in a particular cancer was also collectively tar- geted by co-expressed miRNAs that were experimentally detected in that cancer. However the sets of most affected pathways were specific for each type of cancer (Additional File 5). be affected in a particular cancer was also collectively tar- geted by co-expressed miRNAs that were experimentally detected in that cancer. However the sets of most affected pathways were specific for each type of cancer (Additional File 5). known as kinase inhibitors (specific examples are pro- vided in the next section). Since all miRNAs in this study were over-expressed in cancer, their overall effect would be to down-regulate many or, sometimes, all of the abnor- mally activated alternative signal transduction cascades in most of the pathways known to be affected by a particular cancer. Such effect would be comparable with the effect of several kinase inhibitor drugs combined. known as kinase inhibitors (specific examples are pro- vided in the next section). Comparative pathway analysis Since all miRNAs in this study were over-expressed in cancer, their overall effect would be to down-regulate many or, sometimes, all of the abnor- mally activated alternative signal transduction cascades in most of the pathways known to be affected by a particular cancer. Such effect would be comparable with the effect of several kinase inhibitor drugs combined. The detailed inspection of overlay diagrams revealed an interesting pattern of genes targeted by miRNAs. In the majority of inspected pathways the miRNA targets were often different from genes known to be affected by cancer and were found among genes that are directly down- stream and/or upstream of the cancer related genes in the same branches of signaling cascades (Figures 8, 9, 10, 11, Additional file 6). Compariso of over ex Figure 5 Comparison of top 10 biological function categories and disease categories enriched with miRNA target genes from 5 datasets of over expressed miRNAs Figure 5 Comparison of top 10 biological function categories and disease categories enriched with miRNA target genes from 5 datasets of over expressed miRNAs. (caspase inhibitor) which is indicated for hepatitis C. Both of these genes are also predicted targets of miRNAs that are over-expressed in pancreatic cancer (Fig. 11B). all branches of this signal transduction pathway (Figure 9A). Interestingly, three of these genes are also predicted targets of miRNAs that are co-expressed in colon cancer (Figure 9B). The p38 MAP kinase pathway plays an impor- tant function in the cellular response after infection by pathogens or inflammatory stimulation and has been also implicated in breast, colon and other types of cancer [27,28]. all branches of this signal transduction pathway (Figure 9A). Interestingly, three of these genes are also predicted targets of miRNAs that are co-expressed in colon cancer (Figure 9B). The p38 MAP kinase pathway plays an impor- tant function in the cellular response after infection by pathogens or inflammatory stimulation and has been also implicated in breast, colon and other types of cancer [27,28]. Bcl-2 encodes an integral mitochondrial outer membrane protein that blocks the apoptotic death of some cells such as lymphocytes. Constitutive expression of Bcl-2 is thought to be the cause of some types of cancer [29]. Therefore down regulation of Bcl-2 by over expressed miRNAs would have pro-apoptotic anticancer effect simi- lar to the effect of oblimesen. We found similar examples of other pathways with kinase genes that are targeted by both the anticancer drugs and miRNAs. Importantly, Bcl-2 is one of the few microRNA targets that were confirmed experimentally to be targeted collectively by at least two miRNAs. It has been recently shown by Cimmino et al. [29] that miR-15a and miR-16-1 expres- sion is inversely correlated with Bcl-2 expression in CLL and that both miRNAs negatively regulate Bcl-2. It has also been shown that repression of Bcl-2 by these miRNAs induces apoptosis in leukemia cell lines. The authors of this study have proposed that miR-15 and miR-16 are nat- ural antisense Bcl-2 regulators that could be used for can- cer therapy of some tumors. We have also found other types of genes that are targeted by anticancer drugs and miRNAs within the same path- ways. Anti-cancer drugs and microRNA targets Using IPA knowledge base (Ingenuity Systems, Redwood, CA) we have analyzed known anticancer drugs and found that several drugs are targeting the same cancer related genes as the miRNAs (Tables 2, 3 and Additional Files 7, 8) including several experimentally validated microRNA targets. All overlay diagrams show target genes (highlighted in gray color) corresponding to a set of target genes from sta- tistically enriched GO categories targeted by at least 50% of miRNAs in a group. This type of analysis was performed on all five datasets. One of the most common groups of drugs sharing the same targets with miRNAs was a relatively new class of kinase inhibitors that are designed to inhibit abnormally activated kinases in signal transduction pathways in can- cer cells. Analysis of pathways diagrams has shown that many of these kinase inhibitors target the same kinases as do miRNAs. Two examples are shown for colon cancer (Figures 8, 9) and two – for pancreatic cancer (Figures 10, 11), with additional data presented for lymphoma in Additional File 6. For example, we found that multiple genes from p38 sig- naling pathway are affected in colon cancer and that kinase inhibitor drug SCIO targets all 4 izoforms of the p38 MAP kinase super-family effectively down regulating In many pathways we found that miRNAs target multiple kinases that are important mediators of signal transduc- tion pathways and are often targeted by anticancer drugs Page 9 of 17 (page number not for citation purposes) Page 9 of 17 (page number not for citation purposes) http://www.biomedcentral.com/1471-2105/8/S7/S16 BMC Bioinformatics 2007, 8(Suppl 7):S16 Comparison of top 10 biological function categories and disease categories enriched with miRNA target genes from 5 datasets of over expressed miRNAs Figure 5 Comparison of top 10 biological function categories and disease categories enriched with miRNA target genes from 5 datasets of over expressed miRNAs. B. Disease Categories A. Biological Functions A. Biological Functions A. Biological Functions B. Disease Categories Compariso of over ex Figure 5 Compariso cancer Figure 6 Co pa so o pat ways ow to be a ecte co o ca ce a ta gete by s t at a e ove e p esse co o ca ce gu e 6 Comparison of 12 pathways known to be affected in colon cancer and targeted by miRNAs that are over- expressed in colon cancer. Color of the bars represents 4 data sets. Detailed description could be found in the Results sec- tion. A. Ratio of affected genes to total number of genes in the pathway. Reference gene set and three target sets are included: a complete list (all targets) and two sets trimmed by excluding genes from GO categories targeted by less than 50% or less than 100% of miRNAs in the group. B. Significance of enrichment. Threshold p < 0.05 is shown as yellow line. Bars that are above the line indicate significant enrichment of a pathway. Paradoxically, down regulation of caspase-3 by miRNAs should have anti-apoptotic effect similar to the effect of IDN-6556 and would be beneficial for cancer cell survival. This counterintuitive mode of miRNA regulation has been recently discussed in the literature in a context of relevant experimental observations showing that miR-34a has tumor suppressor activity when ectopically expressed in NB cell lines through induction of caspase 3/7 apoptotic pathway [30]. miR-34a may have a pro-apoptotic effect, in part, through targeting the E2F3 transcription factor. Recently, it has been shown that miR-17-5p and miR-20a also act as tumor suppressors by targeting and reducing E2F1 levels [31]. Similar to miR-34a, the chromosome region with the miR-17 cluster is deleted in some human tumors. This same region, however, is amplified in diffuse large B-cell lymphoma samples [29]. Thus miRNAs may have a tumor suppressor or oncogenic effect depending upon the cell type in which they are expressed. It is also important to keep in mind that when multiple miRNAs are co-expressed within the same tumor cells a systemic regulatory impact should be considered in a context of many regulatory targets that are affected simultaneously. In case of pancreatic cancer we have found that E2F1 is a predicted microRNA target as well as caspase3. E2F1 has been also validated experimentally as a target of multiple miRNAs [30]. Compariso cancer Figure 6 In this regard, it is interesting to note that similar complex regulatory responses were previously reported for the members of E2F family of transcription factors that can also have cell proliferative or pro-apop- totic effects in different cellular and regulatory contexts [[32], 33]. Paradoxically, down regulation of caspase-3 by miRNAs should have anti-apoptotic effect similar to the effect of IDN-6556 and would be beneficial for cancer cell survival. This counterintuitive mode of miRNA regulation has been recently discussed in the literature in a context of relevant experimental observations showing that miR-34a has tumor suppressor activity when ectopically expressed in NB cell lines through induction of caspase 3/7 apoptotic pathway [30]. miR-34a may have a pro-apoptotic effect, in part, through targeting the E2F3 transcription factor. Recently, it has been shown that miR-17-5p and miR-20a also act as tumor suppressors by targeting and reducing E2F1 levels [31]. Similar to miR-34a, the chromosome region with the miR-17 cluster is deleted in some human tumors. This same region, however, is amplified in diffuse large B-cell lymphoma samples [29]. Thus miRNAs may have a tumor suppressor or oncogenic effect depending upon the cell type in which they are expressed. It is also Compariso of over ex Figure 5 For instance we found that apoptosis signaling pathway has many genes affected in pancreatic cancer including two critical regulators of apoptosis: Bcl-2 and Caspase-3 (Figure 11A). Bcl-2 is targeted by recently devel- oped drug oblimesen which is an antisense synthetic oli- gonucleotyde-based anticancer (pro-apoptotic) drug effectively silencing Bcl-2 transcripts (Figure 11A). Capase-3 is targeted by IDN-6556, an anti-apoptotic drug Page 10 of 17 (page number not for citation purposes) BMC Bioinformatics 2007, 8(Suppl 7):S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 Comparison of 12 pathways known to be affected in colon cancer and targeted by miRNAs that are over-expressed in colon cancer Figure 6 Comparison of 12 pathways known to be affected in colon cancer and targeted by miRNAs that are over- expressed in colon cancer. Color of the bars represents 4 data sets. Detailed description could be found in the Results sec- tion. A. Ratio of affected genes to total number of genes in the pathway. Reference gene set and three target sets are included: a complete list (all targets) and two sets trimmed by excluding genes from GO categories targeted by less than 50% or less than 100% of miRNAs in the group. B. Significance of enrichment. Threshold p < 0.05 is shown as yellow line. Bars that are above the line indicate significant enrichment of a pathway. B. A. A. A. B. B. Discussion In this study we have addressed the problem of identify- ing major biological processes and signaling pathways that are targeted collectively by co-expressed miRNAs in Page 11 of 17 (page number not for citation purposes) Page 11 of 17 (page number not for citation purposes) BMC Bioinformatics 2007, 8(Suppl 7):S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 ( pp ) p Comparison of 12 pathways known to be affected in Pancreatic Cancer and targeted by miRNAs that are over-expressed in Pancreatic Cancer Figure 7 Comparison of 12 pathways known to be affected in Pancreatic Cancer and targeted by miRNAs that are over- expressed in Pancreatic Cancer. Color of the bars represents 4 data sets. Detailed description could be found in the Results section. A. Ratio of affected genes to total number of genes in the pathway. Reference gene set and three target sets are included: a complete list (all targets) and two sets trimmed by excluding genes from GO categories targeted by less than 50% or less than 100% of miRNAs in the group. B. Significance of enrichment. Threshold p < 0.05 is shown as yellow line. Bars that are above the line indicate significant enrichment of a pathway. B. A. A. A. B. B. Compariso Pancreatic Figure 7 Comparison of 12 pathways known to be affected in Pancreatic Cancer and targeted by miRNAs that are over expressed in Pancreatic Cancer Figure 7 Comparison of 12 pathways known to be affected in Pancreatic Cancer and targeted by miRNAs that are over- expressed in Pancreatic Cancer. Color of the bars represents 4 data sets. Detailed description could be found in the Results section. A. Ratio of affected genes to total number of genes in the pathway. Reference gene set and three target sets are included: a complete list (all targets) and two sets trimmed by excluding genes from GO categories targeted by less than 50% or less than 100% of miRNAs in the group. B. Significance of enrichment. Threshold p < 0.05 is shown as yellow line. Bars that are above the line indicate significant enrichment of a pathway. very large raw list of predicted target genes to a smaller set of target genes from significantly enriched GO categories. cancer cells. Recently, multiple studies have reported aber- rant expression profiles of miRNAome in human cancers. Discussion Several target genes were experimentally identified for some miRNAs in various tumors; however the global pat- tern of cellular functions and pathways affected by miR- NAs in cancer remains elusive. We proposed a two step data reduction procedure and have tested it on experi- mentally obtained data sets of aberrantly expressed miR- NAs in five human cancers. Second, we have tested the idea that additional trimming of the over-represented GO categories on the total number of hits by multiple miRNAs would allow determi- nation of those biological functions that were affected the most by miRNAs and that were more specific for each can- cer. It has been shown in simulation studies [13] that the level of random noise in miRNA target predictions declined sharply even with trimming by condition of at least 3 miRNAs targeting each of the same genes. Even though we have already filtered the target list on signifi- cance of GO enrichment, we assume that using additional criteria of multiple miRNAs targeting the same categories would reduce the random predictions noise even further. First, we performed a statistical enrichment analysis of GO categories to find categories that are enriched with targets of co-expressed miRNAs. The miRgate GO profiling algo- rithm is specifically designed to take in account informa- tion about the number of miRNAs that are targeting the same genes and information about the number of miR- NAs hits per GO category [17]. It allowed us to reduce a Page 12 of 17 (page number not for citation purposes) http://www.biomedcentral.com/1471-2105/8/S7/S16 BMC Bioinformatics 2007, 8(Suppl 7):S16 Diagram of PTEN Signaling Pathway Figure 8 Diagram of PTEN Signaling Pathway. A. Genes known to be affected in Colon Cancer are shown in gray. B. Genes from significantly enriched GO categories that are targeted by at least 50% of over-expressed miRNAs from Colon Cancer dataset are shown in gray A. B. A. A. B. g g g y g Diagram of PTEN Signaling Pathway. A. Genes known to be affected in Colon Cancer are shown in gray. B. Genes from significantly enriched GO categories that are targeted by at least 50% of over-expressed miRNAs from Colon Cancer dataset are shown in gray and to determine how trimming on the number of target- ing miRNAs would affect these clusters. Our results show that trimming actually improved the separation of clus- ters. Discussion and to determine how trimming on the number of target- ing miRNAs would affect these clusters. Our results show that trimming actually improved the separation of clus- ters. We have tested this assumption by performing unsuper- vised clustering analysis of GO categories using the number of miRNAs that target each category to calculate correlation-based distance and to find groups of catego- ries that were most specifically targeted for each cancer Diagram of p38 signaling pathway Figure 9 Diagram of p38 signaling pathway. A. Genes known to be affected in Colon Cancer are shown in gray. B. Genes from sig- nificantly enriched GO categories that are targeted by at least 50% of over-expressed miRNAs from Colon Cancer dataset are shown in gray. Drug that are known to target this signaling pathway are outlined in dark blue. Drug targets are outlined in light blue. A. B. A. B. B. Diagram o Figure 9 Diagram of p38 signaling pathway Figure 9 Diagram of p38 signaling pathway. A. Genes known to be affected in Colon Cancer are shown in gray. B. Genes from sig- nificantly enriched GO categories that are targeted by at least 50% of over-expressed miRNAs from Colon Cancer dataset are shown in gray. Drug that are known to target this signaling pathway are outlined in dark blue. Drug targets are outlined in light blue. g p g g p y g Diagram of p38 signaling pathway. A. Genes known to be affected in Colon Cancer are shown in gray. B. Genes from sig- nificantly enriched GO categories that are targeted by at least 50% of over-expressed miRNAs from Colon Cancer dataset are shown in gray. Drug that are known to target this signaling pathway are outlined in dark blue. Drug targets are outlined in light blue Page 13 of 17 (page number not for citation purposes) Page 13 of 17 (page number not for citation purposes) BMC Bioinformatics 2007, 8(Suppl 7):S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 Diagram of Integrin Signaling Pathway Figure 10 Diagram of Integrin Signaling Pathway. A. Genes known to be affected in Pancreatic Cancer are shown in gray. B. Genes from significantly enriched GO categories that are targeted by at least 50% of over-expressed miRNAs from Pancreatic Cancer dataset are shown in gray B. A. A. Diagram of Integrin Signaling Pathway Figure 10 Diagram of Integrin Signaling Pathway. A. Genes known to be affected in Pancreatic Cancer are shown in gray. B. Authors' contributions Our pathway analysis suggests that co-expressed miRNAs seem to collectively targeting a broad range of down- stream signaling cascades and down regulating expression of genes in abnormally activated pathways. Such compu- tationally inspired hypothesis could be tested experimen- tally by comparing microRNA expression data with mRNA expression data and protein expression data obtained from the same samples of the tumor tissue. YG designed and performed this study, YG and DB wrote the manuscript. TS, MG and RP provided guidance with biological interpretation, comments on the study and on the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Discussion Diagram of Apoptosis Signaling Pathway Figure 11 Diagram of Apoptosis Signaling Pathway. A. Genes known to be affected in Pancreatic Cancer are shown in gray. B. Genes from significantly enriched GO categories that are targeted by at least 50% of over-expressed miRNAs from Pancreatic Cancer dataset are shown in gray. Drug that are known to target genes from this signaling pathway are outlined in dark blue. Drug targets are outlined in light blue. Diagram of Apoptosis Signaling Pathway Figure 11 Diagram of Apoptosis Signaling Pathway. A. Genes known to be affected in Pancreatic Cancer are shown in gray. B. Genes from significantly enriched GO categories that are targeted by at least 50% of over-expressed miRNAs from Pancreatic Cancer dataset are shown in gray. Drug that are known to target genes from this signaling pathway are outlined in dark blue. Drug targets are outlined in light blue. biology based approach could provide new avenues for biological interpretation of miRNA profiling data as well as generation of experimentally testable hypotheses regarding collective regulatory functions of miRNA in can- cer. to be affected in specific cancers. Importantly, several well known cancer related genes such as ras, Bcl-2 and E2F1 that we identified in our analysis have been already vali- dated in wet lab experiments and were reported by others as targets of multiple miRNAs [25,29,31]. Overall, in our analysis we identified 25 known cancer related genes that have been already experimentally validated as targets of miRNAs. to be affected in specific cancers. Importantly, several well known cancer related genes such as ras, Bcl-2 and E2F1 that we identified in our analysis have been already vali- dated in wet lab experiments and were reported by others as targets of multiple miRNAs [25,29,31]. Overall, in our analysis we identified 25 known cancer related genes that have been already experimentally validated as targets of miRNAs. Discussion Genes from significantly enriched GO categories that are targeted by at least 50% of over-expressed miRNAs from Pancreatic Cancer dataset are shown in gray B. B. Diagram of Integrin Signaling Pathway Figure 10 Diagram of Integrin Signaling Pathway. A. Genes known to be affected in Pancreatic Cancer are shown in gray. B. Genes from significantly enriched GO categories that are targeted by at least 50% of over-expressed miRNAs from Pancreatic Cancer dataset are shown in gray We have further investigated the effect of trimming in pathway analysis where only significantly overrepresented canonical pathways were selected and compared against reference gene sets known to be affected in particular can- cer types. We found that for the majority of pathways the trimming on the number of miRNAs targeting the same GO category produced little or no change in the number of affected genes and at times even provided additional improvement on the significance of enrichment (Figs. 6B, 7B), most probably by eliminating unrelated genes that were diluting the target gene set. Clustering analysis of enriched GO categories and path- ways analysis of major signaling pathways have both shown that using additional criteria of multiple miRNA targeting of the same category allows narrowing down the sets of predicted targets without losing the specificity of cancer related gene lists. Clustering analysis of enriched GO categories and path- ways analysis of major signaling pathways have both shown that using additional criteria of multiple miRNA targeting of the same category allows narrowing down the sets of predicted targets without losing the specificity of cancer related gene lists. Using a combinatorial target prediction algorithm we have found GO categories as well as biological functions, disease categories, toxicological categories, and regulatory pathways that are: targeted by multiple miRNAs; statisti- cally significantly enriched with target genes; and known Page 14 of 17 (page number not for citation purposes) BMC Bioinformatics 2007, 8(Suppl 7):S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 Diagram of Apoptosis Signaling Pathway Figure 11 Diagram of Apoptosis Signaling Pathway. A. Genes known to be affected in Pancreatic Cancer are shown in gray. B. Genes from significantly enriched GO categories that are targeted by at least 50% of over-expressed miRNAs from Pancreatic Cancer dataset are shown in gray. Drug that are known to target genes from this signaling pathway are outlined in dark blue. Drug targets are outlined in light blue. A. B. B. A. B. A. Additional file 1 Supplemental Figure 2 – Top ranked biological functions and disease categories targeted by miRNAs and known to be affected in Colon Cancer. A. Top 5 Biological Functions. B. Top 5 Disease Categories Click here for file [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S1.doc] Additional file 4 Supplemental Figure 1 – Comparison of Toxicology categories and canonical pathways targeted by miRNA for 5 types of cancer. A. Toxi- cology related gene lists from top ranked Toxicology categories. B. Top ranked Canonical Pathways p y 4. Calin GA, Dumitru CD, Shimizu M, Bichi R, Zupo S, Noch E, Aldler H, Rattan S, Keating M, Rai K, Rassenti L, Kipps T, Negrini M, Bullrich F, Croce CM: Frequent deletions and down-regulation of micro-RNA genes miR15 and miR16 at 13q14 in chronic lym- phocytic leukemia. Proc Natl Acad Sci USA 2002, 99:15524-15529. [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S4.doc] p 5. Calin GA, Croce CM: MiRNA signatures in human cancers. Nat Rev Cancer 2006, 6(11):857-66. ( ) 6. Adams BD, Furneaux H, White B: The micro-RNA miR-206 Tar- gets the Human Estrogen Receptor-{alpha}, and Represses ER {alpha} mRNA and Protein Expression in Breast Cancer Cell Lines. Mol Endocrinol in press. 2007, Mar 13 Additional file 5 Supplemental Figure 5 – Comparison of top ranked signaling path- ways targeted by microRNAs in 5 types of cancer. A. Ratio of the number of genes targeted by miRNAs to the total number of genes in each pathway. B. Significance of overrepresentation of microRNA targets in each pathway p 7. Jiang J, Lee EJ, Gusev Y, Schmittgen TD: Real-time expression pro- filing of miRNA precursors in human cancer cell lines. Nucleic Acids Res 2005, 33(17):5394-403. ( ) 8. Lee EJ, Gusev Y, Jiang J, Nuovo GJ, Lerner MR, Frankel WL, Morgan DL, Postier RG, Brackett DJ, Schmittgen TD: Expression profiling identifies miRNA signature in pancreatic cancer. Int J Cancer 2007, 120(5):1046-54. [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S5.doc] ( ) 9. Volinia S, Calin GA, Liu CG, Ambs S, Cimmino A, Petrocca F, Visone R, Iorio M, Roldo C, Ferracin M, Prueitt RL, Yanaihara N, Lanza G, Scarpa A, Vecchione A, Negrini M, Harris CC, Croce CM: A miRNA expression signature of human solid tumors defines cancer gene targets. Proc Natl Acad Sci USA 2006, 103:2257-2261. Additional file 7 Supplemental Table 1 – Number of predicted miRNA target genes known to be associated with specific or any cancer and targeted by anti-cancer drugs. 14. Krek A, Grun D, Poy MN, Wolf R, Rosenberg L, Epstein EJ, MacMe- namin P, da Piedade I, Gunsalus KC, Stoffel M, Rajewsky N: Combi- natorial miRNA target predictions. Nature Genetics 2005, 37:495-500. 15. John B, Sander C, Marks DS: Prediction of human miRNA tar- gets. Methods Mol Biol 2006, 342:101-13. g 16. miRgate 2.1 suite, Actigenics/Cepheid [http://www.actigen ics.com] Additional file 6 Supplemental Figure 4 – TGF-b Signaling pathway. A. Genes known to be affected in Lymphoma (shown in gray) and known anticancer drugs. B. Targets of microRNA family miR-17-92 (shown in gray) and targets of known anticancer drugs. Drug that are known to target genes from this signaling pathway are outlined in dark blue. Drug targets are outlined in light blue. 10. He1 L, Thomson JM, Hemann MT, Hernando-Monge E, Mu D, Good- son S, Powers S, Cordon-Cardo C, Lowe SW, Hannon GJ, Scott M, Hammond SM: A miRNA polycistron as a potential human oncogene. Nature 2005, 435:828-833. g 11. Wang CL, Wang BB, Bartha G, Li L, Channa N, Klinger M, Killeen N, Wabl M: Activation of an oncogenic miRNA cistron by provi- rus integration. Proc Natl Acad Sci USA 2006, 103:18680-18684. g 12. Chen Y, Stallings RL: Differential patterns of miRNA expression in neuroblastoma are correlated with prognosis, differentia- tion, and apoptosis. Cancer Res 2007, 67(3):976-83. [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S6.doc] [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S6.doc] 13. Chen K, Rajewsky N: Natural selection on human miRNA bind- ing sites inferred from SNP data. Nature Genetics 2006, 38:1452-1456. Acknowledgements Additional file 2 Supplemental Figure 3 – Top ranked biological functions and disease categories targeted by miRNAs and known to be affected in Pancreatic Cancer. A. Top 5 Biological Functions. B. Top 5 Disease Categories Click here for file [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S2.doc] Additional file 2 Supplemental Figure 3 – Top ranked biological functions and disease categories targeted by miRNAs and known to be affected in Pancreatic Cancer. A. Top 5 Biological Functions. B. Top 5 Disease Categories Click here for file [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S2.doc] We would like to thank Olivier Delfour for technical assistance and on-line support in using miRgate software suite. This article has been published as part of BMC Bioinformatics Volume 8 Sup- plement 7, 2007: Proceedings of the Fourth Annual MCBIOS Conference. Computational Frontiers in Biomedicine. The full contents of the supple- ment are available online at http://www.biomedcentral.com/1471-2105/ 8?issue=S7. Conclusion Our global analysis of predicted miRNA targets in 5 human cancers demonstrates that co-expressed miRNAs might collectively provide systemic compensatory response to the abnormal functional and phenotypic changes in cancer cells by targeting a broad range of func- tional categories and abnormally activated pathways known to be affected in a particular cancer. Such systems Page 15 of 17 (page number not for citation purposes) Page 15 of 17 (page number not for citation purposes) BMC Bioinformatics 2007, 8(Suppl 7):S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 Additional file 2 Supplemental Figure 3 – Top ranked biological functions and disease categories targeted by miRNAs and known to be affected in Pancreatic Cancer. A. Top 5 Biological Functions. B. Top 5 Disease Categories Click here for file [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S2.doc] Additional file 3 Top 5 functions, disease categories and toxicology lists for 5 cancers. Click here for file [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S3.xls] Additional file 4 Supplemental Figure 1 – Comparison of Toxicology categories and canonical pathways targeted by miRNA for 5 types of cancer. A. Toxi- cology related gene lists from top ranked Toxicology categories. B. Top ranked Canonical Pathways Click here for file [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S4.doc] Additional file 5 Supplemental Figure 5 – Comparison of top ranked signaling path- ways targeted by microRNAs in 5 types of cancer. A. Ratio of the number of genes targeted by miRNAs to the total number of genes in each pathway. B. Significance of overrepresentation of microRNA targets in each pathway Click here for file [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S5.doc] Additional file 6 Supplemental Figure 4 – TGF-b Signaling pathway. A. Genes known to be affected in Lymphoma (shown in gray) and known anticancer drugs. B. Targets of microRNA family miR-17-92 (shown in gray) and targets of known anticancer drugs. Drug that are known to target genes from this signaling pathway are outlined in dark blue. Drug targets are outlined in light blue. Click here for file [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S6.doc] Additional file 7 Supplemental Table 1 – Number of predicted miRNA target genes known to be associated with specific or any cancer and targeted by anti-cancer drugs. Click here for file [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S7.doc] Additional file 8 Supplemental Tables: genes known to be affected in specific cancer and targeted by miRNAs (lymphoma, breast cancer, lung cancer). Cli k h f fil References Top 5 functions, disease categories and toxicology lists for 5 cancers. Click here for file [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S3.xls] Top 5 functions, disease categories and toxicology lists for 5 cancers. Click here for file [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S3.xls] 1. Ambros V: The functions of animal miRNAs. Nature 2004, 431:350-355. 2. Bartel DP: MiRNAs: genomics, biogenesis, mechanism, and function. Cell 2004, 116:281-297. 3. Lee RC, Rhonda L, Feinbaum RL, Ambros V: The C. elegans hete- rochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell 1993, 75:843-854. Page 16 of 17 (page number not for citation purposes) BMC Bioinformatics 2007, 8(Suppl 7):S16 BMC Bioinformatics 2007, 8(Suppl 7):S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 http://www.biomedcentral.com/1471-2105/8/S7/S16 20. Gaidatzis D, van Nimwegen E, Hausser J, Zavolan M: Inference of miRNA targets using evolutionary conservation and path- way analysis. BMC Bioinformatics 2007, 8:69. y y 21. Yoon S, De Micheli G: Prediction of regulatory modules com- prising microRNAs and target genes. Bioinformatics 2005, 21(Suppl 2):ii93-ii100. ( pp ) 22. Gene Ontology Consortium: The Gene Ontology (GO) project in 2006. Nucleic Acids Res 2006, 34(Database):D322-D326. ( ) 23. Beissbarth T, Speed TP: GOstat: find statistically overrepre- sented Gene Ontologies within a group of genes. Bioinformatics 2004, 20:1464-1465. 24. Dergham ST, Dugan MC, Kucway R, Du W, Kamarauskiene DS, Vaitkevicius VK, Crissman JD, Sarkar FH: Prevalence and clinical significance of combined K-ras mutation and p53 aberration in pancreatic adenocarcinoma. Int J Pancreatol 1997, 21(2):127-43. ( ) 25. Johnson SM, Grosshans H, Shingara J, Byrom M, Jarvis R, Cheng A, Labourier E, Reinert KL, Brown D, Slack FJ: RAS is regulated by the let-7 microRNA family. Cell 2005, 120(5):635-47. 26. Kulshreshtha R, Ferracin M, Wojcik SE, Garzon R, Alder H, Agosto- Perez FJ, Davuluri R, Liu CG, Croce CM, Negrini M, Calin GA, Ivan M: A microRNA signature of hypoxia. Mol Cell Biol 2007, 27(5):1859-67. ( ) 27. Zarubin T, Han J: Activation and signaling of the p38 MAP kinase pathway. Cell Res 2005, 15:11-18. 28. Ostrander JH, Daniel AR, Lofgren K, Kleer CG, Lange CA: Breast Tumor Kinase (Protein Tyrosine Kinase 6) Regulates Hereg- ulin-Induced Activation of ERK5 and p38 MAP Kinases in Breast Cancer Cells. Cancer Res 2007, 67(9):4199-209. Breast Cancer Cells. Cancer Res 2007, 67(9):4199-209. 29. Cimmino A, Calin GA, Fabbri M, Iorio MV, Ferracin M, Shimizu M, Wojcik SE, Aqeilan RI, Zupo S, Dono M, Rassenti L, Alder H, Volinia S, Liu CG, Kipps TJ, Negrini M, Croce CM: miR-15 and miR-16 induce apoptosis by targeting BCL2. Proc Natl Acad Sci USA 2005, 102(39):13944-9. ( ) 30. Welch C, Chen Y, Stallings RL: MicroRNA-34a functions as a potential tumor suppressor by inducing apoptosis in neurob- lastoma cells. Oncogene . advance online publication 12 February 2007; doi: 10.1038/sj.onc.1210293 31. O'Donnell KA, Wentzel EA, Zeller KI, Dang CV, Mendell JT: c-Myc- regulated microRNAs modulate E2F1 expression. Nature 2005, 435:839-843. 32. Johnson DG, Degregori J: Putting the Oncogenic and Tumor Suppressive Activities of E2F into Context. Curr Mol Med 2006, 6:731-738. Additional file 8 17. Delfour O, Vilanova D, Atzorn V, Michot B: The passionate race for miRNA detection and function deciphering. In miRNA: Biol- ogy, Function and Expression Edited by: Clarke NJ, Sanseau P. DNA Press; 2007:335-362. Supplemental Tables: genes known to be affected in specific cancer and targeted by miRNAs (lymphoma, breast cancer, lung cancer). Click here for file [http://www.biomedcentral.com/content/supplementary/1471- 2105-8-S7-S16-S8.xls] 18. Wang X, Wang X: Systematic identification of microRNA func- tions by combining target prediction and expression profil- ing. Nucleic Acids Research 2006, 34(5):1646-1652. g ( ) 19. Cui Q, Yu Z, Purisima EO, Wang E: Principles of microRNA reg- ulation of a human cellular signaling network. Mol Syst Biol 2006, 2:46. Page 16 of 17 (page number not for citation purposes) Page 16 of 17 (page number not for citation purposes) BMC Bioinformatics 2007, 8(Suppl 7):S16 Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp BioMedcentral Page 17 of 17 (page number not for citation purposes) Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp BioMedcentral Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp BioMedcentral Publish with BioMed Central and every scientist can read your work free of charge

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.