id
stringlengths 3
9
| source
stringclasses 1
value | version
stringclasses 1
value | text
stringlengths 1.54k
298k
| added
stringdate 1993-11-25 05:05:38
2024-09-20 15:30:25
| created
stringdate 1-01-01 00:00:00
2024-07-31 00:00:00
| metadata
dict |
|---|---|---|---|---|---|---|
253397322
|
pes2o/s2orc
|
v3-fos-license
|
WDHD1 is over‐expressed in nasopharyngeal carcinoma and may control the expression of ITGAV
Nasopharyngeal carcinoma (NPC) is a highly metastatic and invasive malignant tumor that originates in the nasopharynx. The DNA‐binding protein WD repeat and HMG‐box DNA‐binding protein 1 (WDHD1) are highly expressed in a variety of tumours, but its expression and mechanism of action in NPC have not been reported to date. To investigate the involvement of WDHD1 in NPC, we first mined databases for the gene expression profile of NPC. Immunohistochemistry (IHC) was performed on 338 cases of NPC and 112 non‐NPC samples to verify the results. We report that the expression of WDHD1 is significantly elevated in NPC. ChIP‐seq was used to show that integrin alpha V (ITGAV) and WDHD1 exhibit a significant binding peak in the promoter region of the ITGAV gene. The expression levels of ITGAV and WDHD1 exhibit a significant positive correlation, and IHC was performed to show that ITGAV is highly expressed in NPC. Expression of ITGAV increased after overexpression of WDHD1, suggesting that ITGAV may be a potential target gene of WDHD1. Pathway analysis showed that both genes were closely related to the cell cycle, and flow cytometry was used to further confirm that decreased expression of WDHD1 significantly increased the number of apoptotic cells. In conclusion, our results suggest that expression of WDHD1 is increased in NPC and is likely to be associated with the NPC cell cycle; thus, we propose that WDHD1 may have the potential as a target gene for primary screening and treatment of NPC.
Nasopharyngeal carcinoma (NPC) is a highly metastatic and invasive malignant tumor that originates in the nasopharynx. The DNA-binding protein WD repeat and HMG-box DNA-binding protein 1 (WDHD1) are highly expressed in a variety of tumours, but its expression and mechanism of action in NPC have not been reported to date. To investigate the involvement of WDHD1 in NPC, we first mined databases for the gene expression profile of NPC. Immunohistochemistry (IHC) was performed on 338 cases of NPC and 112 non-NPC samples to verify the results. We report that the expression of WDHD1 is significantly elevated in NPC. ChIP-seq was used to show that integrin alpha V (ITGAV) and WDHD1 exhibit a significant binding peak in the promoter region of the ITGAV gene. The expression levels of ITGAV and WDHD1 exhibit a significant positive correlation, and IHC was performed to show that ITGAV is highly expressed in NPC. Expression of ITGAV increased after overexpression of WDHD1, suggesting that ITGAV may be a potential target gene of WDHD1. Pathway analysis showed that both genes were closely related to the cell cycle, and flow cytometry was used to further confirm that decreased expression of WDHD1 significantly increased the number of apoptotic cells. In conclusion, our results suggest that expression of WDHD1 is increased in NPC and is likely to be associated with the NPC cell cycle; thus, we propose that WDHD1 may have the potential as a target gene for primary screening and treatment of NPC.
Nasopharyngeal carcinoma (NPC) is a highly metastatic and invasive malignant tumour that originates in the nasopharynx. Available statistics show that 133 354 new cases and 80 008 deaths from NPC occurred worldwide in 2020 [1], indicating the serious danger of this cancer to human health. The incidence of NPC has clear regional characteristics, with 70% of new cases occurring in Southeast and East Asia, especially in Guangdong and Guangxi in China [2]. In the recent years, the age of onset has become younger and Abbreviations AUC, the area under the curve; Cistrome DB, Cistrome Data Browser; DEG, differentially expressed gene; EMT, epithelial-mesenchymal transformation; GEO, Gene Expression Omnibus; GO, Gene Ontology; HCC, hepatocellular carcinoma; IGV, Integrative Genomics Viewer; ITGAV, integrin alpha V; KEGG, Kyoto Encyclopedia of Genes and Genomes; NPC, nasopharyngeal carcinoma; NSCLC, non-small cell lung cancer cells; SAC, spindle assembly checkpoint; sROC, summary receiver operating characteristic curve; TCGA, The Cancer Genome Atlas; WDHD1, WD repeat and HMG-box DNA-binding protein 1. younger, imposing a heavy economic burden on patients' families and society. Radiotherapy is still the preferred treatment for NPC due to its hidden location and sensitivity to radiation. However, in areas with a high incidence of NPC, about 70% of patients are already in an advanced stage when diagnosed [3], and clinical intervention is late. For these reasons, the overall prognosis is poor and the recurrence rate is high. Clarifying the specific pathogenesis of NPC and finding new diagnostic and therapeutic targets are goals that require further efforts.
WD repeat and HMG-box DNA-binding protein 1 (WDHD1) is a DNA-binding protein with a total molecular weight of 125 kDa, composed of 1127 amino acids [4]. Extensive research has shown that WDHD1 is associated with the cell cycle. Its possible mechanisms of action may involve: (a) regulation of entry into the S phase, as WDHD1 depletion leads to G1 stagnation and delays progression to the S phase [5,6]; (b) regulation of the assembly and activation of the pre-replication complex (pre-RC), two key steps in the initiation of DNA replication [7]; and (c) control of the repair process to correct base mismatch and insertion-deletion mismatch during DNA replication [8]. For these reasons, WDHD1 is implicated in the development of many cancers. TCGA analysis and genome-wide siRNA screening performed by Ertay et al. [9] showed a higher expression of WDHD1 mRNA in triple-negative breast cancer than in normal breast tissue, with expression related to tumour size, proliferation, and poor prognosis [9]. Other studies have shown overexpression of WDHD1 in the esophageal, lung, bile duct, and cervical cancers [10][11][12], as well as an association with poor prognosis and malignant progression. However, WDHD1 expression and its effects have not yet been reported for NPC.
In our research, we searched The Cancer Genome Atlas (TCGA) (https://cancergenome.nih.gov/), Gene Expression Omnibus (GEO) (http://www.ncbi.nlm.nih. gov/geo/), ArrayExpress database (http://www.ebi.ac. uk/arrayexpress/), and related literature, followed by combined bioinformatic and immunohistochemical analysis to obtain a comprehensive picture of WDHD1 expression in NPC. We also predicted the downstream target gene of WDHD1, and explored the expression of the predicted target gene in NPC by immunohistochemistry (IHC). PCR was used to analyze the relationship between their expressions. In addition, possible molecular functions and signaling pathways were forecasted by pathway analysis, and the results were preliminarily verified by flow cytometry apoptosis assay (Fig. 1A). Our study provides a basis for developing new diagnostic markers and therapeutic targets for NPC.
Bioinformatics data mining and processing
Nasopharyngeal carcinoma-related datasets were downloaded from the GEO databases (Fig. 1B). We searched for "(nasopharyngeal OR nasopharynx) AND (neoplasm OR cancer OR carcinoma OR malignancy)" and included datasets according to the following criteria: (a) human NPC tissues were studied, (b) each dataset contained WDHD1 mRNA expression profile data, and (c) the experimental group contained no < 3 samples. The expressed data were processed in the R 4.0.2 software environment (Lucent Technologies, Jasmine Hill, NJ, USA), and fragments per kilobase of exon model per million mapped fragments (FPKM) from RNA-seq experiments were converted to tethered particle motion (TPM) to eliminate the effects of gene length and sequencing depth on the expression level. The unstandardized data were converted by the formula log 2 (x + 0.001), followed by the use of the normalize Between Arrays function of the limma package to eliminate the inter-sample batch effect in the same dataset. We also combined the same platform datasets and normalized the batch effect using the combat function of the SVA package.
Statistical analysis
After obtaining the expression data for WDHD1, we conducted an analysis of the expression data using the independent-samples t-test of IBM SPSS STATISTICS version 25.0 (IBM Corp., Armonk, NY, USA) to confirm the significance of WDHD1 expression differences between NPCs and adjacent non-NPCs (mean was used to describe the average level of this group of data; SD was used to describe the degree of dispersion). The data were also analyzed using STATA 15.0 (Stata Corp LP, College Station, TX, USA) to increase the accuracy of the results and to evaluate the ability of WDHD1 expression to distinguish between NPCs and non-NPCs. The I-square (I 2 ) and chi-square tests were used to determine the level of heterogeneity between different datasets when drawing forest maps. A value of P < 0.05 or I 2 > 50% indicated significant heterogeneity, and a random effect model was chosen for analysis. We used both Begg's and Egger's methods for funnel plots to evaluate publication bias, with a value of P > 0.05 suggesting that publication bias was not significant. A summary receiver operating characteristic curve (sROC) was drawn, and the area under the curve (AUC) was used to measure the ability of WDHD1 to differentiate non-cancerous tissue from NPC.
Immunohistochemistry
Tissue microarrays were prepared by Pantomics, Inc. (Richmond, CA, USA) from 338 NPC tissues and 112 nasopharyngeal mucosa chronic inflammation tissues (NPC TMA1, NPC TMA2, NPC131, NPC241 and NPC482). Paraffin-embedded sections were prepared and deparaffinized using conventional procedures. After hydration, tissue sections were placed in ethylene diamine tetraacetic acid (EDTA) antigen retrieval solution for antigen repair, followed by 3% H 2 O 2 blocking endogenous peroxidase. The primary antibody (rabbit polyclonal antibody of WDHD1, ab224221, Abcam (Cambridge, MA, USA); rabbit monoclonal antibody of ITGAV, ab179475, Abcam) was diluted (Zhong Shan Golden Bridge Biotechnology, Beijing, China) at a ratio of 1 : 100 and then incubated at 37°C for 70 min. The rest of the steps were performed according to a previously reported procedure [13]. Staining was assessed under a light microscope, and we used two different methods to assess the IHC results of WDHD1 and ITGAV. For WDHD1, we randomly selected five nonoverlapping high-magnification fields, and calculated the number of positive cells per 100 cells in each field. The final result was the mean number of the five fields. For ITGAV, the final total score was the product of the percentage of positive cells and the staining intensity. The percentage of positive cells was evaluated as follows: 0 (< 5%); 1 (6-25%); 2 (26-50%); 3 (51-75%); and 4 (> 76%). The staining intensity score criteria were as follows: 0 (colourless); 1 (light yellow, mild staining); 2 (tan, medium staining); and 3 (brown, strong staining). We obtained a more intuitive display of the differential expression of WDHD1 and ITGAV in control tissues and experimental tissues by drawing a violin diagram. We then used STATA 15.0 to draw the ROC curve to evaluate the ability of WDHD1 and ITGAV to distinguish between NPC and non-NPC tissues. In addition, the clinicopathological parameters of the patients were collated, and the statistical differences among the groups were estimated using the Student's t-test.
Potential transcriptional mechanisms of WDHD1 in NPC
To obtain a better understanding of the mechanism of WDHD1 in NPC, we gathered the following three gene sets (WDHD1 target genes, co-expressed genes of WDHD1 and NPC differentially expressed genes) to predict the potential target gene of WDHD1 with their intersection. We first downloaded all target genes of WDHD1 from the Cistrome Data Browser (Cistrome DB), and selected putative target genes with scores ≥ 1 for further analysis. We then used R 4.0.2 and the Pearson correlation algorithm to screen the co-expressed genes of WDHD1 (CEGs), and collected genes with correlation coefficients ≥ 0.3, P-values < 0.05, and presence in more than two datasets. We then screened differentially expressed genes (DEGs) with the limma package, and we collected genes with log 2 FoldChange > 1, P < 0.05 and presence in more than two datasets. Some up-regulated and positively correlated putative target genes were obtained using the intersection of the three gene screens. We then selected a gene that had significant binding peaks with WDHD1, downloaded the bigwig files, and imported them into the INTEGRATIVE GENOMICS VIEWER (IGV) (Cambridge, MA, USA) to draw a peak graph. We further explored the relationship between the predicted target gene and WDHD1 and its expression in NPC tissues by plotting the sROC curve and correlation scatter plots based on GEO databases and IHC, and calculating the SMD value.
PCR was used to detect the expression of ITGAV after overexpressed the expression of WDHD1 in C666-1 and IHC was used to further verify the expression of the predicted target in NPC tissues (NPC TMA1, including 96 NPC cases and 30 non-NPC cases). We used the CLUSTERPROFILER R Package to obtain a deeper understanding of the biological function and mechanism of WDHD1 in NPC by conducting Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses on the intersection of the genes collected above and ITGAV-related DEGs (the intersection of ITGAV positive-related genes and NPC DEGs; The screening method of ITGAV positively related genes was consistent with that of WDHD1 positively related genes, and the Pearson correlation algorithm was used to screen them in R 4.0.2. The screening criteria were correlation coefficients ≥ 0.3, P-values < 0.05, and presence in more than four datasets; DEGs appeared in no less than three datasets. We then used the intersection of the two gene screens for further analysis).
Lentivirus infection
The concentrated virus solution was purchased from Shanghai Jikai Gene Chemical Technology Co., Ltd (Shanghai, China), and stored at −80°C (the component sequence of the WDHD1 overexpression group was Ubi-MCS-3FLAG-CBh-gcGFP-IRES-puromycin, GV492; the negative control was constructed with empty vectors, and nonsense sequences were not inserted, CON335). C666-1 cells with good-growing conditions in the logarithmic growth phase were taken and inoculated in 6-well plates with 2 × 10 5 cells per well. The cells were infected after adherence on the second day. The corresponding amount of virus was added according to MOI = 75, and 40 μL of transfection reagent HitransG A was added. The complete medium was supplemented to 1 mL per well, and the culture was continued for 15 h in a 37°C, 5% CO 2 incubator. The transfection was observed under an inverted fluorescence microscope 72 h after infection. If the cell growth condition is good and the degree of fusion is more than 70%, the medium containing puromycin can be used to screen the infected cells (the initial screening concentration is 3 μgÁmL −1 , which is changed to 1 μgÁmL −1 for maintenance culture for about a week). The fluorescence was observed under an inverted fluorescence microscope, and when the fluorescence efficiency was almost 100%, the cells were expanded and collected for qPCR. siRNA transfection CNE-2 cells in the logarithmic growth phase and a good growth state were inoculated in 6-well plates at 2 × 10 5 cells per well and transfected after cell adhesion (the target sequence of WDHD1 siRNA: 5 0 -GCUGUGAAUUUA GCCAUUATT-3 0 ; and the sequence of negative control was confidential, which did not target any gene product. Both were purchased from Huzhou Hippo Biotechnology Co., Ltd [Huzhou, Zhejiang, China]). At 2 h before transfection, the culture medium was washed and then replaced with serum-free 1640 medium, and transfection was carried out according to the instructions. After 6 h, the medium was exchanged with 2 mL of fresh complete medium, and the cells were incubated for 48 h in a constant-temperature incubator (37°C, 5% CO 2 ).
RNA extraction, reverse transcription and realtime fluorescence quantitative PCR
We extracted the total RNA with TRIzol reagent (Shanghai Pufei Biotechnology Co., LTD, Shanghai, China) according to the instructions, and then reverse-transcribed the RNA into cDNA using a Promega M-MLV kit (Promega (Beijing) Biotech Co., Ltd, Beijing, China). The product was evaluated with light absorption at 260 and 280 nm using a spectrophotometer. Primer sequences were designed, and all the primers were synthesized by Sangon Biotech (Shanghai) Co., Ltd (Shanghai, China). The primer sequences were as follows: GAPDH (F): was then used for amplification on an ABI 7500HT system (Thermo Fisher Scientific, Waltham, MA, USA). SYBR-Green analysis was used for quantitative detection of gene expression, and 2 ÀΔΔCT was used for calculation (ΔC T = C T value of target gene − C T value of reference gene GAPDH, −ΔΔC T = ΔC T of the control group − ΔC T of experimental group). Finally, we used an independent-samples t-test to determine whether the difference between the negative control group and the experimental group was statistically significant.
Cell apoptosis detection by flow cytometry
After 48 h of culture, 0.25% trypsin without EDTA was added to digest the transfected CNE-2 cells. The cells were washed twice with phosphate buffered saline, centrifuged at 1500 g for 3 min, and the supernatant was removed. An Annexin V-FITC/PI apoptosis detection kit (Tianjin Sungene Biotech Co., Ltd, Tianjin, China) was used as follows: 500 μL binding buffer was added to resuspend the cells, followed by 5 μL Annexin V-FITC and 5 μL PI. The tube was mixed well and then incubated and protected from light at room temperature for 5-15 min. Cell apoptosis was detected using flow cytometry and statistically analyzed by an independent-samples t-test (the upper right quadrant should be the late apoptotic cell, and the lower right quadrant should be the early apoptotic cell). All the experiments were repeated three times.
Results
The association between WDHD1 expression and NPC was analyzed based on online database information After screening, a total of 14 datasets (nine arrays: GSE13597, GSE40290, GSE53819, GSE61218, GSE126683, GSE39826, GSE64634, GSE34573 and GSE12452; and five RNA-seq datasets: GSE118719, GSE68799, GSE63381, GSE102349 and GSE134886) were collected; datasets from the same platform were combined, and batch effects were removed using the combat function of the SVA package (GSE68799, GSE63381 and GSE102349 were combined into GPL11154, while GSE64634, GSE34573 and GSE12452 were combined into GPL570). The WDHD1 expression data in each dataset were then extracted for comprehensive analysis. The values of mean AE SD revealed a significant expression of WDHD1 in NPC in most datasets (P < 0.05) ( Table 1). The forest map showed high heterogeneity (I 2 = 79.3%, P < 0.001), so a random effect model was selected for analysis. The expression of WDHD1 mRNA was significantly higher in the NPC group than in the control group (SMD = 1.22, 95% CI = 0.40-2.04) ( Fig. 2A). The Begg's test did not detect significant publication bias (P = 0.592, > 0.05) (Fig. 2B), nor did the Egger's test (P = 0.332, > 0.05) (Fig. 2C) (Fig. 2D), indicating that WDHD1 had a strong ability to distinguish between NPC and non-NPC tissues.
IHC analysis of the expression of the WDHD1 protein in NPC
We also used IHC to explore the expression of the WDHD1 protein in NPC. Pathology sections of 338 NPC tissues and 112 nasopharyngeal mucosa with chronic inflammation were prepared. Expression of WDHD1 protein in NPC tissues showed nuclear localization ( Fig. 3A-F) and was higher in the NPC tissues than in nasopharyngeal mucosa with chronic inflammation (P < 0.001) (Fig. 3G), and the ROC curve showed AUC = 0.957 (P < 0.001), indicating that WDHD1 has a strong ability to distinguish NPC from non-NPC tissues (Fig. 3H), in agreement with the results of the bioinformatic analysis. A further statistical analysis of the clinicopathological information of the patients revealed that WDHD1 expression was not significantly correlated with age or gender ( Table 2). Further investigation is needed regarding other pathological features.
Analysis of the potential mechanism of WDHD1 in NPC
We also explored the possible role of WDHD1 in NPC by screening NPC DEGs, WDHD1 positively related genes and target genes. According to the screening criteria, a total of 1115 up-regulated DEGs, 6137 WDHD1 positively related genes, and 4112 target genes were screened. The intersection of the three gene sets was 181 ( Fig. 4A; Table S1). For ITGAV-related DEGs, there were 523 up-regulated DEGs, and 1712 ITGAV positively related genes (Fig. 4B). We found that integrin alpha V (ITGAV) and WDHD1 had significant binding peaks upstream of the transcription start site (Fig. 5A) and that their expressions were significantly positively correlated ( Fig. 5B-L). Extraction of the ITGAV expression data to clarify its mRNA expression level in NPCs revealed a higher expression level in NPC than in non-NPC tissues (SMD = 2.01, 95% CI = 1.33-2.69); I 2 = 65.4% and P = 0.002, indicating significant heterogeneity, and a random effect model was chosen for analysis (Fig. 6A). The Egger test did not detect significant publication bias (P = 0.484, > 0.05) (Fig. 6B), and the sROC curve showed AUC = 0.99 [0.98-1.00], indicating a strong ability of ITGAV to distinguish NPC from non-NPC tissue (Fig. 6C). Subsequent IHC confirmed that ITGAV was also highly expressed at the protein level ( Fig. 7A-F), and the violin diagram shows this result more intuitively (Fig. 7G). The ROC curve suggested that ITGAV had a strong ability to distinguish between NPC and non-NPC tissue (Fig. 7H). The results of PCR indicated that we had successfully constructed the WDHD1 overexpressed stable strains in C666-1, the cell transfection efficiency rate of nearly 100% (Fig. 8A-D) and the mRNA expression level of WDHD1 was higher in the overexpressed group than in the control group (P < 0.05) (Fig. 8E). More importantly, we found that the expression of ITGAV increased after the overexpression of WDHD1 in C666-1 (Fig. 8F). These results revealed that WDHD1 may affect the progression of NPC through the transcriptional regulation of ITGAV expression. We further explored the cellular functions and biological processes of these two genes by conducting enrichment analyses. The first 10 records of the results are shown in Fig. 9. The GO analysis indicated that WDHD1-related target genes were mainly enriched in functions related to the cell cycle, such as organelle fission, nuclear division, and DNA replication (P < 0.05). The KEGG enrichment analysis showed that these genes were closely connected with the cell cycle (P < 0.05). The GO functional items of the ITGAV-related DEGs were enriched in organelle fission, nuclear division, DNA replication, and positive regulation of the cell cycle process (P < 0.05) (Fig. 10A), while the KEGG analysis revealed that these genes mainly participated in the cell cycle and PI3K-Akt signaling pathway (Fig. 10B). These enrichment results suggest that WDHD1 and ITGAV are closely related to the regulation of the cell cycle in NPC.
WDHD1 is closely related to cell apoptosis
At 48 h after siRNA transfection, RT-QPCR was used to detect the relative expression level of WDHD1 mRNA in CNE-2 cells, and it was found that the expression level of WDHD1 mRNA in the interference group was lower than that in the control group (P < 0.05) (Fig. 11). Subsequently, we conducted a more in-depth study of the mechanism of WDHD1 in NPC by flow cytometry of CNE-2 cells that were successfully transfected with WDHD1 siRNA. The numbers of early and late apoptotic cells increased significantly after knockdown of WDHD1 expression with siRNA (NC: 2.93 AE 0.88%; WDHD1 siRNA: 8.31 AE 0.54%) (Fig. 12).
Discussion
The WDHD1 is a DNA-binding protein containing the WD40 domain and an "HMG" box, which is closely related to the post-transcriptional steps of the centromeric silencing pathway, homologous recombination repair, and the assembly and activation of the pre-replication complex (pre-RC), two key steps in the initiation of DNA replication [7,14,15]. It plays a significant role in regulating protein signal transduction, transcription, cell proliferation, and apoptosis [16]. A number of existing studies have shown that WDHD1 is closely related to tumor progression. For example, Liu et al. [11] found that miR-494 may play a role in epithelial-mesenchymal transformation (EMT), tumor growth and metastasis of cholangiocarcinoma by negatively regulating WDHD1, and Zhou and Chen [17] demonstrated that STAT3 may affect DNA replication through transcriptional regulation of WDHD1. He et al. [18] analyzed the expression of WDHD1 in hepatocellular carcinoma (HCC) by combining IHC and bioassay, and found that WDHD1 was highly expressed in HCC and may be associated with its poor prognosis. Our research group was surprised to find that WDHD1 was also highly expressed in laryngeal squamous cell carcinoma [19]. In a recent study, Gou et al. knocked down the expression of WDHD1, tested the radiotherapy sensitivity of non- small cell lung cancer cells (NSCLC) by clonal formation assay, flow cytometry, comet assay, and immunofluorescence and verified it in a vivo xenograft tumor model. The results showed that knocking down the expression of WDHD1 significantly increased the radiotherapy sensitivity of NSCLC [20]. Despite these studies, its role in NPC has not been reported. In our study, we first combined high-throughput data with immunohistochemistry, followed by prediction of the biological function of WDHD1 and its target genes with GO and KEGG enrichment analysis, then preliminary verification using flow apoptosis assays, and finally an evaluation of WDHD1 in NPC to determine its possible molecular mechanism. To the best of our knowledge, this is the first study to verify the expression level of WDHD1 in NPC by combining multiple methods. It is also innovative in its use of pathway analysis, PCR and flow cytometry detection of apoptosis to explore the molecular mechanism of WDHD1 as a transcription factor in the malignant progression of NPC.
The bioprediction and immunohistochemistry results presented here indicate that WDHD1 is significantly overexpressed in NPC in most datasets. The forest map showed SMD = 1.22 (95% CI = 0.40-2.04) and I 2 = 79.3%, indicating significant heterogeneity. This may reflect the fact that the data samples come from different countries or that the microarray platforms differ. The reason for the significantly higher expression of WDHD1 in the control group than in the experimental group in GPL11154 (platform merged datasets) might be that some of the included datasets had no control group or that the sample size difference between the two groups was too large. Enrichment analysis suggested that WDHD1 was closely related to cell cycle functions, which was further verified by the results of the flow cytometry apoptosis assay.
We also evaluated the mechanism by combining the WDHD1 target genes for analysis. The promoter of the ITGAV gene is a DNA sequence that RNA polymerase identifies, binds to, and uses to initiate transcription. ITGAV showed an obvious peak with WDHD1 and was significantly positively correlated with WDHD1 expression. PCR indicated that the expression of ITGAV increased with the increase of WDHD1 in C666-1. Therefore, we speculated that WDHD1 regulates the cell cycle by transcriptional regulation of ITGAV expression. Integrin is a kind of transmembrane glycoprotein receptor that binds cells and the matrix. It is composed of an α subunit, a β subunit and bidirectional signaling molecules, and participates in cell adhesion and differentiation. ITGAV is located on chromosome 2q31-q32, and as a proteincoding gene, it is closely related to tumor angiogenesis, invasion and metastasis. Existing studies have shown that ITGAV overexpression is related to the progression of some tumors, such as glioblastoma and laryngeal, breast, esophageal, and colorectal cancers [21][22][23].
With regard to the relationship between ITGAV and NPC, Ding et al. [24] demonstrated experimentally that ITGAV is the target gene of miR-9-3p, which can promote the proliferation and metastasis of NPC cells by promoting the EMT process. In our research, ITGAV was significantly overexpressed in NPC tissues and was closely related to the p53 and PI3K/Akt sig- Fig. 11. Detection of knockdown efficiency of WDHD1 gene by RT-qPCR. Error bars show means AE standard deviation (SD; n = 3); ***, P < 0.001; P, P value of Student's t-test. naling pathways. The p53 gene is the most commonly mutated gene in human cancers and was initially considered to be an oncogene; however, subsequent studies have revealed that the wild-type p53 gene can inhibit tumor growth and cell canceration [25,26]. At present, many studies have found that the p53 signaling pathway plays an important role in tumor proliferation [27,28]. The PI3K/Akt signaling pathway is recognized as closely related to cell apoptosis, tumor angiogenesis, and the cell cycle. Many studies have shown that this pathway is abnormally activated in a variety of malignant tumours, such as liver, gastrointestinal, and pancreatic cancers, and regulates cell proliferation, survival, transformation and other functions [29][30][31]. Therefore, we speculate that WDHD1 plays a role in the malignant progression of NPC through transcriptional regulation of ITGAV expression and subsequent regulation of the p53 or PI3K/Akt signaling pathways. However, this possibility requires further experimental verification, and our research group is currently conducting relevant studies.
In the present study, a role for WDHD1 in the malignant progression of NPC was evident, although the study had some limitations. One was that the heterogeneity among the included studies was large, which we speculate was due to a number of reasons: (a) some of the data samples came from different countries or the microarray platforms differed; (b) in some studies, there was a large difference in the number of samples between the control group and the experimental group, and some had no control group at all; and (c) the sample sources were different. For the experimental group, the sample sources included NPC biopsy, primary tumor, and cell lines. For the control tissues, the sources included normal nasopharyngeal mucosa, inflamed nasopharyngeal mucosa, and primary tonsillar epithelial cells. Another limitation is that the number of included specimens was small, so further expansion of the sample size is needed to reduce sample size errors. A third limitation is that the results presented here are derived mainly from bioprediction, so they lack full in vivo and in vitro verification. Our future goal is to conduct in vitro and in vivo experiments to further improve this study.
Overall, we found that WDHD1 is highly expressed in NPC and may influence the NPC cell cycle by regulating ITGAV expression. Bioinformatic analysis showed that WDHD1 has potential value in differentiating an NPC group from a non-cancer group and may be useful as a new NPC screening molecule. Therefore, this research may serve as the basis for developing new diagnostic and therapeutic targets for the treatment of NPC.
Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Table S1. A total of 181 WDHD1 positively related differentially expressed target genes are obtained.
|
2022-11-09T06:16:57.255Z
|
2022-11-07T00:00:00.000
|
{
"year": 2022,
"sha1": "ebd6ad205527316753e148cf65a45795d4a51e7d",
"oa_license": null,
"oa_url": null,
"oa_status": "CLOSED",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "3f7067bb889f0e8b17aac1dc9a67ab1aff6dedef",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
5550409
|
pes2o/s2orc
|
v3-fos-license
|
Coordination in reconnaissance-attack parsing
A proposal for recognizing coordinate structures using the 'reconnaissance-attack' model is presented. The approach concentrates on distinguishing predicate coordination from other types of coordination and suggests that low-level structural cues (such as the number of predicates, coordinators, and subordinators occurring in the input string) can be exploited at little cost during the early phase of the parse, with dramatic results. The method is tested on a text of 16,000 words.
a particular phase if more than one answer is possible in principle given the information available to that point. (If, at the end of the process, unresolved indeterminacies remain, ambiguity is predicted.) Intuitively, the difference between Reconnaissance and Attack is that Reconnaissance constitutes the gathering of information while Attack constitutes anything which involves decision-making. More formally, Reconnaissance can be viewed as a series of parameter-setting operations each of which is done independently of any of the others while Attack requires simultaneous access to all parameters.
It is worth noting that there does not appear to be any reason to exclude in principle the possibility of hybrid models in which principles of the sort we shall develop below are invoked prior to the application of a parser along the lines of those described in e.g. Dahl andMcCord 1983 or Fong andBerwick 1985. Our principal contention is that whatever choices are made about how to go about 'parsing proper' (that is, actually building a syntactic representation for an input sentence), there is an advantage to having certain global structural information already available rather than starting 'blind '. Following Kac 1978 and1985, we subsume under a single rubric of 'predicate coordination' the coordination of verbs, VP's, and S's on the rationale that common to all three types is that they have the effect of rendering predicates 'equiordinate' (that is, so related that neither is sub-or superordinate to the other). In e.g.
(2) I believe that John likes Mary and Han'y admires Sue. the verbs likes and admires are both subordinate to believe but neither is subordinate to the other. Similarly, in a sentence like (ib) above, hits and attacks are both 'topmost' in the ordination scheme. (For a more detailed development of the theory of ordination relations, see Rindflesch forthcoming.) In this approach a distinction is made between STRICT and LOOSE coordination (two coordinate expressions are strictly so if separated by at most a conjunction, loosely coordinate otherwise, as in e.g. John and Mary. ran vs. John ran, and Mary (too)) and also between PRIMARY and SECONDARY coordination. The primary coordinates in a coordinate structure are the largest coordinate expressions (e.g. the S's in sentential coordination), while the secondary coordinates are smaller expressions contained in the primary ones taken (by the theory) to be coordinate by virtue of the coordination of the containing expressions; for example, the predicates of coordinate sentences (both VP's and V's) are secondary coordinates in a sentential coordination.
For purposes of parsing, we assume that the first task is to coordinate WORDS rather than the larger expressions containing them; that is, secondary coordinates are sought first, and the primary coordinates in which they appear are identified later. This is consistent with the overall theoretical approach, described in more detail in Rindflesch op. cit., which is much more akin to dependency syntax than to phrase structure analysis. (See also Kac and Manaster-Ramer 1986.)
A Sketch of the Parsing Strategy
In this paper, our focus will be on determining, from a minimally analyzed string, whether or not a given instance of and or or enters into a predicate coordination as defined above. (A longer paper giving full details of the approach is in preparation.) In the earliest stages of parsing a given sentence containing a coordinating conjunction, each conjunction is identified as either (a) definitely involved in a predicate coordination, (b) as definitely not involved in such a coordination, by virtue of falling certain necessary conditions for being so involved, or (c) as of indeterminate status which must be resolved (if possible) in a later phase of the parse. The following principles are invoked for this purpose: Applied early in Attack: (3) LIMITS CONSTRAINT (Rindflesch forthcoming) The number of predicate-coordinating conjunctions in a sentence must be smaller than the number of verbs.
(4) POSITION CONSTRAINT (Kac 1978(Kac , 1985 If a coordinating conjunction conjoins expressions X and Y, it lies somewhere between X and Y.
Applied late in Attack: There is at least one predicate in every sentence which is not subordinate to any other predicate in that sentence.
(6) EQUIORDINATION CONSTRAINT If two predicates are coordinate then they are also equiordinate.
28g
The principles (3-6) are all rather straightforward, even common-~sensical; it is nonetheless not entirely uninteresting to learn that they Ibrm the basis for an extremely effective parsing strategy.
Reconnaissance involves a single pass through the currant string, the first steps being lexical lookup and counting and indexing all categories. The information gained from this counting a0d indexing is then used to eliminate impossible structures, via a check for compatibility with the principles (3-6) above.
In order to deal with coordination two ancillary lists, As will be discussed below, in the vast majority of cases in at least one domain the type of coordination occurring in a sentence cart be determined solely on the basis of these straighb forward principles. In these eases, the structure encountered is similar to that seen in (1 a). In order to determine whether predicates are being coordinated in structures like those seen in (lb) 2B7 and (lc) it is necessary to have somewhat more information about the input string.
The additional information required to deal with strings such as (lb) and (lc), only one of which involves predicate coordination despite the fact that the two are nearly identical, concerns the relationships which obtain between predicates in a complex sentence. These relationships are enforced by constraints (5-6) above, in conjunction with Here we will concentrate on subordinators. Each subordinator in a sentence r0ust be associated with a verb in that sentence, and this association causes that verb to be necessarily subordinate to some other predicate. The fact which is of value in parsing coordinate structures is that this can be known even before the superordinate partner of the subordinate predicate has been identified. For example in (lc) even before anything else is known about the structure of the sentence, it can be determined that the subordinator when is associated with hits and that therefore hits will have to be subordinate to some other predicate in that sentence.
As noted above, the parsing principles applied during Attack remove words from the PCL's. In the parse of (lb), while there are nouns and verbs in both PCL's at the beginning of Attack, all the nouns are removed, as Attack proceeds, from both PCL's, leaving only the verbs to be coordinated. The way in which Attack accomplishes this is as follows.
There is more than one predicate in (lb) and thus the predicates have to be in an ordination relation in order to satisfy the Multipredicate Constraint. This relation cannot be subordination, since no subordinating ORS is present; assuming coordination to be the only other possibility, and given that there is a coordinating conjunction between the two predicates, we conclude that the predicates are in fact coordinate. In order to satisfy all of the constraints Attack must therefore remove John and Fred from PCL-L leaving hits as the sole member of that list. It must also remove guys and him from PCL-R leaving attack as the only word in that list. The configuration of these lists thus indicates that the only possible coordinates in (lb) are hits and attack.
These same principles determine that predicate coordination cannot obtain in (lc). As Attack begins, PCL-L for the conjunction in this string contains John, hits', and Fred. PCL-R contains guys, attack, and him. Since there is more than one predicate in this string, the predicates will have to be in an ordination relationship, but it will have to be a relationship of subordination rather than coordination. Hits will have to be subordinate to some predicate in this sentence by vil'tue of the fact that it is associated with the subordinator when. ( (1 c) is one in which hits is subordinate to attack. Therefore, hits must be removed from the PCL-L and attack must be removed from the PCL-R. From this it can at least be determined that (lc) does not involve predicate coordination.
Empirical Support for the Approach
To test the effectiveness of the strategy described above, we subjected to analysis a corpus of nearly 16,000 words (15,985 to be exact). The texts used were specifications and design requirements (5 in all) applying to hardware manufactured by Control Data Corporation, supplied to us in machine-readable form. Each text was run through a concordance program which identified all tokens of and and or; and for each token of each conjunction, tile containing sentence was then analyzed (by hand). A total of 431 tokens of the two conjunctions occurred in the corpus, 362 of them in complete sentences (as opposed to section heads or fragments, which were ignored). As noted earlier, we did not, in undertaking the analysis, take into account the fact that there is widespread category-label ambiguity ('CLA') in English; this represents a significant idealization of the data, but it is nol a cheat. The problem with regard to coordination with which we m'e concerned is that even in cases where no CLA occurs, problems of the sort exemplified by (1) arise. That the overall problem is even worse than we make it out to be does not invalidate our claims, though it meaus --and we are fully aware of this --th,~t tile account is incomplete.
Of the conjunctions occurring in complete sentences, the type of coo,-dination in which each was involved was correctly ascertainable via application of the five constraints in 91% of the
Examples
We conclude with an analysis of some sentences from tile corpus, to illustrate the approach in more detail. The discussion here concentrates on our stated goal of determining for any conjunction what kinds of expressions are being coordinated. A large number of the sentences in the corpus, with respect to coordination, have a structure resembling (9) A single sector single port buffer will provide speed matching between the host interface and the controller.
In this sentence, there is only one predicate (will provide) and fi~rthermore there is no predicate to the right of the conjunction.
Either the Limits Constraint or the Position Constraint can therefore determine solely on the basis of information determined during Reconnaissance that there is no predicate coordination in (9).
The somewhat more complex structure of (10) can also be handled without difficulty.
Although there are two predicates in (10) (are and to support), The Position Constraint correctly predicts that they cannot be coordinate since they are not separated by the conjunction in this sentence.
Sentences containing more than one conjunction submit to the principles we propose in this paper, as illustrated by (11) The primary slructures and relationships of these memory blocks are illustrated in Figure 11 and are defined more precisely in later sections.
The first conjunction in (11) does not effect predicate coordination, while the second does. 'file Position Constraint assures the correct analysis for the first conjunction: PCL-L for the first conjunction will not contain a verb since there are no verbs to the left of this conjunction; consequently, no verb will be put in the con'esponding PCL-R, thus precluding predicate coordination for the first conjunction in (1 I). When the PCL's are filled for the second conjunction in (11), they will both contain nouns as well as predicates; hence either could potentially be coordinated.
However, since there are two predicates in (11) (are illustrated and are defined) and since there are no subordinating ORS's in the sentence, the predicates in fact must be coordinate in order to satisfy tile Multipredicatc Constraint.
Although the PCL's for the conjunction or in (I2) will initially contain both nouns and verbs, the conect analysis of this sentence does not involve predicate coordination.
(12) When switch position 1 is set to the "off' position, a 2 byte or a 16 bit word will be available on the data bus bits 0-F.
The analysis of (12) is similar to the analysis of (lc). There arc two predicates in the string (is set and will be available), one of which (is set) is necessarily non-main due to its association with the subordinating conjunction when. Were these predicates to be coordinated they would both be non-main by the Equiordination Constraint. Therefore, the only way the Multipredicate Constraint and the Main Predicate Constraint can be satisfied is to consider there to be no predicate coordination in this sentence.
(I0) The pfimary purposes of the special functions arc tt~ support diagnostic analysis, data recovery, and download capabilities.
|
2014-07-01T00:00:00.000Z
|
1988-08-22T00:00:00.000
|
{
"year": 1988,
"sha1": "ada7582851085efedd43dde1606694df1053da3b",
"oa_license": null,
"oa_url": "http://dl.acm.org/ft_gateway.cfm?id=991693&type=pdf",
"oa_status": "BRONZE",
"pdf_src": "ACL",
"pdf_hash": "ada7582851085efedd43dde1606694df1053da3b",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science"
]
}
|
266561013
|
pes2o/s2orc
|
v3-fos-license
|
Body mass index and weight gain in pregnancy and cardiovascular health in middle age: A cohort study
To examine associations between body mass index (BMI) in early pregnancy and gestational weight gain (GWG) with cardiovascular health in middle age using the ‘Life's Essential 8’ (LE8) concept of the American Heart Association (AHA).
Funding information
Swedish Heart-Lung Foundation; Knut and Alice Wallenberg Foundation; Swedish Research Council; VINNOVA; University of Gothenburg and Sahlgrenska University Hospital; Karolinska Institutet and Stockholm County council; Linköping University and University Hospital; Lund University and Skåne University Hospital; Umeå University and University Hospital; Uppsala University and University Hospital; Autonomous University of Madrid
| I N TRODUC T ION
Pregnancy is commonly a period where substantial weight is gained to support maternal requirements and ensure fetal growth. 1 However, having overweight before pregnancy and excessive gestational weight gain (GWG) have been associated with long-term maternal adiposity, [2][3][4] and the issue seems to be on the rise; it has been reported that almost half of women worldwide exceed the guidelines for appropriate weight gain during pregnancy. 5Previous studies suggest that excessive GWG is associated with perinatal complications, such as macrosomia and preterm delivery, 6 and along with a high prepregnancy body mass index (BMI) have been shown to increase the risk of cardiometabolic conditions in both the mother and the offspring. 7,8The negative effects of excessive body weight and GWG thus present a major public health concern, making pregnancy a potential target for obesity prevention and the promotion of long-term cardiovascular health.For this reason, an effort to evaluate the impact of pregnancy weight patterns on future maternal cardiovascular risk is important, to counteract these alarming trends.Although studies have confirmed the unfavourable effect of obesity on long-term cardiometabolic status, [9][10][11][12] the association between weight status in pregnancy and cardiovascular health has not been sufficiently explored, particularly from a longitudinal perspective.One way to study the impact of maternal weight status and GWG on later cardiovascular health is to examine the association between gestational weight and the concept of 'Life's Essential 8' (LE8) developed by the American Heart Association (AHA) to assess and improve cardiovascular health in a given population. 13It is based on a risk score system consisting of eight metrics of known risk components, corresponding to health behaviours and health factors.Importantly, LE8 has been strongly related to later all-cause and cardiovascular mortality, 14,15 making it a suitable proxy for general cardiovascular health.However, little is known regarding the risk factors for poor cardiovascular health as measured by LE8.4][15] The aim of this study was therefore to investigate the association between BMI in early pregnancy and GWG with cardiovascular health (measured as LE8) in middle-aged women.
| Study population
This cohort study linked information on cardiovascular health in middle age, utilising data from SCAPIS, to BMI in early pregnancy and GWG, obtained from the Swedish Medical Birth Register. 16The study participants in SCAPIS (n = 30 154, 51.4% women), aged 50-64 years, were randomly selected from the Swedish population register between 2013 and 2018.Six university hospitals in Sweden (Gothenburg, Linköping, Malmö/Lund, Stockholm, Umeå and Uppsala) collaborated in performing the recruitment and examinations for SCAPIS. 17e linked data from women in SCAPIS (n = 15 508) with data from the Swedish Medical Birth Register using the unique personal IDs that every citizen in Sweden has.
| Derivation of the analytical sample
To minimise the impact of consecutive pregnancies, information from the participant's most recent pregnancy was collected.If data were missing from the most recent pregnancy, information from the second latest pregnancy was used, and so forth.A flow chart illustrating the recruitment and inclusion criteria of the study participants is presented in Figure S1.Women from SCAPIS who had delivered at least one child in Sweden and who had data available on BMI in early pregnancy and the LE8 score were included in the analyses (n = 8871).Thus, women with missing data on any of the covariates (age at SCAPIS examination, age at index pregnancy, educational level, parity and study site) were excluded.We also excluded individuals whose index pregnancy was a twin pregnancy (n = 117), as they have different recommendations on GWG compared with women carrying singleton fetuses. 18In total, 3693 of the women with data from the Swedish Medical Birth Register were excluded, and thus 8871 women were included in the analysis with BMI.Of these women, 6264 women had a weight measured at delivery and were included in the analysis of GWG.
| Primary exposures
Pregnancy weight data were collected from the Swedish Medical Birth Register, to which the SCAPIS study was linked. 17BMI (kg/m 2 ) in early pregnancy was measured at the first prenatal visit (usually occurring before 12 weeks of gestation) or estimated as the difference between weight at delivery and self-reported weight gain if the pregnancy occurred between 1982 and 1989.It was categorised according to current classifications: underweight (<18.5 kg/ m 2 ), normal weight (18.5-24.9kg/m 2 ), overweight (25.0-29.9kg/m 2 ) and obesity (≥30 kg/m 2 ).GWG was calculated as the difference between the weight recorded at the first prenatal visit and the weight before delivery.The derived weight difference was categorised as having inadequate, adequate or excessive GWG, as defined by the BMI-specific guidelines from the Institute of Medicine (IOM). 18For additional analyses independent of BMI, classifications of GWG were also made based only on kilograms gained during pregnancy (<5.0, 5.0-9.9,10.0-14.9,15.0-19.9 and ≥20.0 kg).
| Life's Essential 8
Information regarding cardiovascular health status in middle age was collected from the SCAPIS project.The available data, including information of all components in LE8, were described previously in detail. 19As for the behavioural components, the diet score was calculated based on the criteria for the Mediterranean Eating Patterns for Americans, 20 assessed through the web-based questionnaire MiniMeal-Q. 21,22Physical activity levels were estimated using triaxial accelerometers (GT3X, wGT3X+ and wGT3X-BT; ActiGraph LLC, Pensacola, FL, USA). 23icotine exposure and sleep health were measured using self-reported questionnaires.The LE8 health factors (BMI, non-HDL (high-density lipoprotein) cholesterol, blood glucose and blood pressure) were calculated using standardised laboratory and clinical procedures.Based on these assessments, each metric was given a score ranging from 0 to 100 using the AHA's definitions, where 100 is seen as an ideal value.By summing the scores of each component and dividing it by the number of metrics present, an unweighted average cardiovascular score of 0-100 was generated.In this study, LE8 was calculated when data were available for seven or eight of the components.The classifications and arithmetic score are all categorised in accordance with the AHA's definition of the concept. 13LE8 scores below 60 points were classified as poor, whereas scores of 60 points or higher were coded as 'not poor', as we have previously shown that an LE8 score below 60 is strongly related to atherosclerosis. 19The usefulness of LE8 as a proxy of cardiovascular health has been confirmed in several studies. 19,24
| Covariates
Information on all covariates except age at index pregnancy was collected at the SCAPIS examination.Educational status was categorised as 'did not finish primary education', 'primary education level', 'high school level' or 'higher educational level' (university or equivalent).Parity was defined as the number of times the woman had given birth and was classified as 1, 2, 3, 4 or ≥5 births.Birth region was categorised as: Central Europe, Eastern Europe and Central Asia; high-income countries excluding Sweden; Latin America and The Caribbean; North Africa and the Middle East; South Asia; Southeast and East Asia; Sub-Saharan Africa; or Sweden.Current menopausal status during the outcome measurements was classified as a dichotomised answer of 'yes' or 'no'.
| Statistical analysis
Descriptive statistics are presented as means and standard deviations (SDs) for continuous variables and frequencies and percentages for categorical variables.The associations between BMI in early pregnancy and GWG with later cardiovascular health were analysed using binary logistic regression models.We present an unadjusted model and a main adjusted model, which includes age at SCAPIS examination, educational level, age at index pregnancy, parity and study site as covariates.Results are presented as adjusted odds ratios (aORs) with 95% confidence intervals (95% CIs).When analysing GWG, adjustments were also made for BMI in early pregnancy.Analyses of BMI in early pregnancy and GWG were performed separately.Statistical significance was set as P < 0.05 in all analyses, and the analyses were conducted in SPSS 28.0 (IBM, Armonk, NY, USA).To further study a potential nonlinear relationship between LE8 and the exposure variables, we used restricted cubic splines with four knots located at the 5th, 35th, 65th and 95th percentiles in Stata 17 (StataCorp LLC, College Station, TX, USA). 25 We performed all analyses as 'complete cases', and in Table S3 we present descriptive statistics on women who were excluded from the study.
We conducted a series of sensitivity analyses to examine whether the observed associations were robust.First, an analysis with additional adjustments for menopausal status and maternal birth region was performed.Second, to exclude any influence of previous pregnancies, we analysed women that gave birth to only one child.Third, analyses excluding women whose index pregnancy was a preterm birth were executed.Fourth, to ensure that consecutive pregnancies after measurements did not affect the outcome, another sensitivity analysis of only women whose index pregnancy was their last was made.Analyses where women with preeclampsia were excluded were also performed.Moreover, a new LE8 score was calculated including a stricter cut-off level for physical activity, where the highest score was achieved at 300 minutes (instead of 150 minutes) of moderate (or greater) intensity activity per week.Each minute of moderate activity was counted as 1 minute and each minute of vigorous activity was counted as 2 minutes. 23We also examined the influence of two other LE8 cut-off levels for poor cardiovascular health (50 and 70 points).We performed an analysis excluding the participants from the years where the pregnancy BMI (and consequently GWG) were estimated.Furthermore, additional adjustments for a history of gestational diabetes and maternal hypertension, gestational age, polycystic ovarian syndrome (PCOS), smoking at pregnancy admission and hormone replacement therapy (HRT) were performed separately.To address potential selection bias, we conducted logistic models that incorporated inverse probability weighting to account for missing data in exposures, outcomes and the covariates used in the analysis. 26Finally, we also calculated the goodness of fit for our main models.
| Descriptive characteristics
Table 1 presents the descriptive characteristics of the study participants.The mean age at the index pregnancy was 32.3 (SD 4.8) years and the age at SCAPIS was on average 56.8 (SD 4.1) years, resulting in a mean follow-up time of 24.5 (SD 6.1) years.In total, 398 (4.5%) women were classed as underweight, 6689 (75.4%) women were classed as normal weight, 1460 (16.5%) women were classed as having overweight and 324 (3.7%) women were classed as having obesity.Based on the IOM categories of GWG, 18 1676 (26.8%) women were classed as having inadequate weight gain in their index pregnancy, 2687 (42.9%) women were classed as having adequate weight gain in their index pregnancy and 1901 (30.3%) women were classed as having excessive weight gain in their index pregnancy.At the outcome assessments in SCAPIS, the mean LE8 score was 73.6 (SD 11.5) points out of 100.Of all participants, 1107 (12.5%) had a score under 60, which was set as the cut-off point for having poor cardiovascular health.Table S1 depicts the distribution of GWG in relation to BMI categories and Table S2 shows the gestational age at delivery by GWG and BMI categories.Compared with those excluded from the study, the women included were slightly younger at the SCAPIS assessment, had a lower BMI and a higher mean LE8 score, but the differences between the groups were small (Table S3).
| Associations between BMI in early pregnancy and LE8 in middle age
Poor cardiovascular health in middle age was strongly associated with BMI in early pregnancy (Figure 1A).As presented in Figure 1B, when using BMI categories, the aOR of poor cardiovascular health in middle age was significantly greater in women who had overweight (aOR 3.30, 95% CI 2.82-3.88)or obesity (aOR 7.63, 95% CI 5.86-9.94),compared with women classified as being of normal weight.The odds of having poor cardiovascular health were lower in the underweight group as compared with individuals of normal weight in the adjusted model (aOR 0.67, 95% CI 0.46-0.96).The estimates for both the unadjusted and the adjusted models are presented in Table S4.When BMI was analysed as a continuous variable with linear regression, the odds of having poor cardiovascular health increased by 21% for each increase in BMI unit (aOR 1.21, 95% CI 1.19-1.24).When analysing the LE8 health factors and behaviours separately, BMI in early pregnancy was more strongly associated with LE8 factors as compared with behaviours in middle age (Figure 2).Associations of BMI with each individual LE8 component are enclosed in Table S5.
| Associations between GWG and LE8 in middle age
A U-shaped association illustrates that the odds of having poor cardiovascular health were greater when gaining both more or less than the recommended weight in pregnancy (Figure 3A).In Figure 3B, aORs are shown both for categories according to weight gain in kilograms as well as the IOM classifications that also include the prepregnancy BMI category.For those who gained <5.0 kg during their pregnancy, the odds of having poor cardiovascular health were 2.6 times higher compared with those who gained between 10.0 and 14.9 kg, which was set as the reference group.For those who gained 5.0-9.9kg, the odds of having poor cardiovascular health were 1.2 times higher, for those who gained 15.0-19.9kg, the odds were 1.3 times higher, and for those who gained 20 kg or more during pregnancy, the odds were 1.8 times higher than for the reference group.When categorising according to the IOM classifications, the women who gained excessively during pregnancy had 31% higher odds of having poor cardiovascular health (aOR 1.31, 95% CI 1.09-1.57),compared with those who gained weight within the recommendations, whereas the result for those who gained inadequate weight was not statistically significant after adjustments (aOR 1.12, 95% CI 0.92-1.37).The estimates for both the unadjusted and the adjusted models are presented in Table S6.Finally, the U-shape in the association between GWG and LE8 was evident for both factors and behaviours (Figure S2).Associations of GWG with each individual LE8 component are enclosed in Table S7.
| Sensitivity and complementary analysis
In general, the results in all sensitivity analyses were comparable with our main analyses and the conclusions were similar (Tables S4, S6, S8 and S9).However, when excluding births with estimated GWG, associations with excessive GWG were attenuated.Nevertheless, conclusions regarding the associations GWG in kilograms (e.g.<5.0, 5.0-9.9,10.0-14.9,15.0-19.9 and ≥20.0 kg) remained similar to our main results.Regarding selection bias, the analysis using inverse probability weighting to account for missing data rendered similar results compared with the main analysis (Tables S8 and S9).We also calculated the goodness of fit for our main models (Table S10), the results showed that the addition of BMI, not GWG, to the models increased the explained variance.
| DISCUS SION
In this large population-based cohort study, BMI in early pregnancy and GWG were associated with future cardiovascular health, measured through the AHA's LE8 construct.Women classified as having overweight obese at the beginning of pregnancy had significantly greater odds of a poor LE8 score (<60 points) later in life, whereas those who were underweight had lower odds of a poor LE8 score than the women classified in the normal weight category, after adjustments.The relationship between a higher BMI in early pregnancy and an unfavourable score proposes a cumulative negative effect on later cardiovascular health when exceeding the BMI recommendations.For GWG, higher odds of poor cardiovascular health were seen in the group who gained excessive weight during pregnancy, but these odds were not of the same magnitude as the odds for a high BMI in early pregnancy.In contrast to low BMI in early pregnancy, both low and high GWG were associated with higher odds of having poor cardiovascular health, indicating a U-shaped association.The relationship between inadequate GWG and high Associations between BMI in early pregnancy and LE8 in middle age.The associations of BMI in early pregnancy with aORs for a poor LE8 score in middle age were examined using cubic splines (A) and binary logistic regression according to BMI categories (B).Adjustments were made for age at index pregnancy, age at SCAPIS assessment, educational level, parity and study site.aOR, adjusted odds ratio; BMI, body mass index; CI, confidence interval; LE8, Life's Essential 8.
F I G U R E 2
Associations between BMI in early pregnancy and LE8 factors and behaviours.The associations of BMI in early pregnancy with aORs for a poor LE8 factor score (A) and LE8 behaviour score (B) were examined using cubic splines.Adjustments were made for age at index pregnancy, age at SCAPIS assessment, educational level, parity and study site.aOR, adjusted odds ratio; BMI, body mass index; CI, confidence interval; LE8, Life's Essential 8.
odds of a LE8 score may reflect the general health consequences of gaining too little weight during pregnancy. 18revious research has clearly shown that prepregnancy BMI carries a significant influence for future cardiovascular health.Several have found strong associations between prepregnancy weight and a higher risk of several adverse cardiometabolic outcomes, such as diabetes, hypertension and cardiovascular events in midlife. 3,8,27A cumulative impact on cardiometabolic factors, such as diabetes, hypertension and vascular diseases, with each pregnancy in which excessive weight was retained has also been observed. 28owever, associations between gestational BMI and LE8 have not been investigated.Thus, our study provides novel evidence for the role of BMI in early pregnancy for future cardiovascular health as measured by LE8 in middle age.
Although our associations between overweight or obesity in early pregnancy and a high risk for cardiovascular disease may be reconciled with previous research, tendencies towards an increased risk when being underweight have also been observed, contrary to our results.One study found an increased risk of major cardiovascular events later in life for both the underweight and obese BMI category during pregnancy. 29In our study, being underweight first indicated lower odds for poor cardiovascular health, but after examining the individual components (Table S5) it was clear that this association was mainly due to a better LE8 score in the BMI component.Thus, as having a low BMI can be detrimental to many aspects of a women's health, maintaining a normal and stable BMI over the years should be emphasised. 30Furthermore, our analyses demonstrate that obesity is more strongly linked to individual LE8 factors, compared with individual LE8 behaviours (Table S5), which explains the stronger associations observed for LE8 factors.
While previous studies have confirmed associations between prepregnancy BMI and later cardiovascular risk, the importance of GWG is more controversial.In one study, no associations were found between excessive GWG and higher lipid levels, blood pressure, glucose, insulin or metabolic syndrome at 8 years postpartum. 2 In contrast, another study found that excessive GWG was followed by a small but significant increase in atherosclerotic cardiovascular disease risk, compared with those who did not gain weight excessively during pregnancy. 31Similarly, when examining the risk of later cardiovascular disease and hypertension related to perinatal weight change, another study found that an increase of one BMI unit or more increased the risk of hypertension for all weight categories, as well as the cardiovascular risk among those who were of normal weight or underweight before pregnancy. 32As obesity is known to be associated with a greater cardiovascular risk, 33 the association between GWG and poor cardiovascular health is presumably mediated by the increased risk of obesity that GWG brings.In our study, a GWG of <5 kg was associated with a poor LE8 score, whereas inadequate GWG defined by the IOM guidelines did not exhibit any association.Discrepancies between these results could be because a GWG of <5 kg is far more uncommon than having an inadequate GWG.Interestingly, when analysing the associations of GWG (in kg) with the LE8 components (Table S7), it was evident that both low (<5 kg) and high (≥20 kg) GWG were inversely associated with several individual components.These results may explain the observed U-shaped associations of GWG with LE8 total, factor and behaviour scores.Although several pregnancy-related disorders, such preterm birth, 34 gestational hypertension and PCOS, are linked to a greater risk for cardiovascular disease, screening is currently only recommended for women with pre-eclampsia. 35studies may further explore the independent and combined effects of different adverse pregnancy outcomes that are linked to BMI and cardiovascular health, such as pre-eclampsia and preterm birth, although our sensitivity analysis excluding pregnancies with preeclampsia and preterm birth provided very similar conclusions as our main analyses.Furthermore, it is important to be aware of the potential presence of omitted variable bias that this study carries arising from its long-term perspective.Unaccounted confounding variables may have influenced cardiovascular outcomes later in life.This does not dismiss the findings but highlights the possibility of pregnancy as an important window of opportunity to identify individuals for health screening and supportive interventions.Potential strategies for optimising weight gain in pregnancy could, for example, include counselling and interventions on behavioural components such as diet and exercise. 36Thus, in conjunction with the increased prevalence of being overweight or obese, 37 it is essential to focus on pregnancy weight optimisation to reduce the future risk of cardiovascular disease.
| Strengths and limitations
This study consisted of a large population-based sample and involved a rich source of information on participant characteristics, which allowed us to account for a variety of possible confounding factors.Additionally, the results are further strengthened by the fact that all LE8 components that the AHA recommends were included in the analysis.Another strength is the longitudinal nature of the study, allowing us to examine associations over an average follow-up time of 24 years.
Limitations in this study included its observational design, which limits causal inferences.Some of the participants that reported BMI in early pregnancy did not have data on delivery weight, which represents another limitation.Another limitation was that body weight was estimated using self-reported weight gain and delivery weight in the early years of the Swedish Medical Birth Register, although excluding these women in a sensitivity analysis did not change our main findings.Furthermore, as postpartum weight trajectories were not available in this study, we are unaware of any longitudinal weight changes after pregnancy that could have affected the outcome.Finally, the moderate level of missing data in our study and the subtle variations in the characteristics of those included and excluded from the study indicate that missingness could potentially impact our estimates.Nevertheless, we conducted a series of sensitivity analyses including inverse probability weighting to account for selection bias, and the main conclusions remained.
| CONCLUSION
Our study demonstrates that women with a high BMI in early pregnancy and women with excessive GWG have a greater risk of a poor cardiovascular health in middle age than women who were of normal weight and who gained weight within the recommendations.Although further research is needed to assess the impact of body weight during pregnancy on later cardiovascular health, these findings support the notion that obesity and excessive GWG are important risk factors for poor cardiovascular health decades later.Additional pregnancy weight optimisation initiatives should be considered to promote future cardiovascular health.
AU T HOR C ON T R I BU T ION S
LW, AH-A, VHA, SH-F, DB, CJÖ, KR and PH: contributed to the conception and design of the study.JS, CJÖ and KR: contributed to data acquisition.LW, AH-A, VHA and PH: conducted the statistical analysis.SH-F, DB, PW, BD, CL, JS, CJÖ and KR: contributed to data analysis and interpretation.LW, AH-A and PH: drafted the article, which was reviewed and revised by SH-F, VHA, DB, PW, BD, CL, JS, CJÖ and KR.All authors contributed to the article and approved the final version for publication.
AC K NO W L E D GE M E N T S
We thank the participants and staff of the SCAPIS project for their valuable contributions.
F U N DI NG I N FOR M AT ION
The main funding body of the Swedish CArdioPulmonary bioImage Study (SCAPIS) is the Swedish Heart-Lung Foundation.The study is also funded by the Knut and Alice Wallenberg Foundation, the Swedish Research Council and VINNOVA (Sweden's Innovation Agency), the University of Gothenburg and Sahlgrenska University Hospital, Karolinska Institutet and Stockholm County Council, Linköping University and University Hospital, Lund University and Skåne University Hospital, Umeå University and University Hospital, Uppsala University and University Hospital.SH-F is supported by a Margarita Salas grant from the Autonomous University of Madrid.
C ON F L IC T OF I N T E R E S T S TAT E M E N T
JS reports stock ownership in Anagram Kommunikation AB and Symptoms Europe AB.The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflicts of interest.
DATA AVA I L A BI L I T Y S TAT E M E N T
The data used in this article cannot be shared publicly, to protect the privacy of the participants in this study as well as for legal reasons.However, by contacting the SCAPIS (www.scapis.org) or the corresponding author, information will be provided regarding the procedures for accessing data following Swedish legislation.Requests to access data can be directed to pontus.henriksson@liu.se.
E T H IC S PPROVA L
The studies involving human participants were reviewed and approved by the Swedish Ethical Review Authority (ref. 2021-01).The participants provided their written informed consent to participate in this study.
F I G U R E 3
Associations between GWG and LE8 in middle age.(A) The associations of GWG in early pregnancy with aORs for a poor LE8 score in middle age were examined using cubic splines.(B) Binary logistic regressions were made for estimating aORs according to GWG categories classified after IOM recommendations (lower panel) or kilograms (upper panel).Adjustments were made for age at index pregnancy, age at SCAPIS assessment, educational level, parity, BMI at index pregnancy and study site.aOR, adjusted odds ratio; CI, confidence interval; GWG, gestational weight gain; IOM, Institute of Medicine; LE8, Life's Essential 8.
Descriptive characteristics of the study sample.Results are presented as mean ± SD for continuous variables or numbers and percentages for categorical variables.
|
2023-12-28T06:16:50.595Z
|
2023-12-27T00:00:00.000
|
{
"year": 2024,
"sha1": "f6259b1258b261f5a8e2527a1e8309e3dbf14b2c",
"oa_license": "CCBY",
"oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/1471-0528.17740",
"oa_status": "HYBRID",
"pdf_src": "Wiley",
"pdf_hash": "f20f613bd73c6cad3b726a636c85da278ddd8fc2",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
117906571
|
pes2o/s2orc
|
v3-fos-license
|
Metrizable uniform spaces
Three themes of general topology: quotient spaces; absolute retracts; and inverse limits - are reapproached here in the setting of metrizable uniform spaces, with an eye to applications in geometric and algebraic topology. The results include: 1) If f: A ->Y is a uniformly continuous map, where X and Y are metric spaces and A is a closed subset of X, we show that the adjunction space X\cup_f Y with the quotient uniformity (hence also with the topology thereof) is metrizable, by an explicit metric. This yields natural constructions of cone, join and mapping cylinder in the category of metrizable uniform spaces, which we show to coincide with those based on subspace (of a normed linear space); on product (with a cone); and on the isotropy of the l_2 metric. 2) We revisit Isbell's theory of uniform ANRs, as refined by Garg and Nhu in the metrizable case. The iterated loop spaces \Omega^n P of a pointed compact polyhedron P are shown to be uniform ANRs. Four characterizations of uniform ANRs among metrizable uniform spaces X are given: (i) the completion of X is a uniform ANR, and the remainder is uniformly a Z-set in the completion; (ii) X is uniformly locally contractible and satisfies the Hahn approximation property; (iii) X is uniformly \epsilon-homotopy dominated by a uniform ANR for each \epsilon>0; (iv) X is an inverse limit of uniform ANRs with"nearly splitting"bonding maps. Several chapters are devoted primarily to exposition: (I) an introduction to uniform spaces, with a focus on the metrizable case; (V) the space of measurable functions; (VI) the space of probability measures and other measure spaces.
Introduction
Although topological and uniform approaches to foundations of what was then known as Analysis Situs originated in the same works by M. Fréchet and F. Riesz (cf. Remark 2.4 below), uniform spaces hopelessly lagged behind in development ever since, and were never taken seriously in algebraic and geometric topology, due in part to the lack of a coherent theory of quotient spaces, and of a reasonable notion of a polyhedron in the uniform sense. Yet the opposite side of the coin features painful side effects of the usual topological foundations, including (i) the non-metrizability of the cone over R; (ii) the non-metrizability of RP ∞ with the topology of a CW-complex (or the geometric realization of a simplicial set); (iii) the uncountability of a family of polyhedra (Mardešić's resolution [59]) needed to capture geometrically relevant homotopy properties (the shape) of a non-compact space. This results, in particular, in the awkward situation that the shape invariant homology of a space as simple as N + × N (where N denotes the infinite countable discrete space, and + denotes the one-point compactification) cannot be calculated in ZFC, as its value depends on additional axioms [60].
1.A. Quotient spaces
One of the goals of the present paper is to show that (the topology of) quotient uniformity is, after all, far nicer than quotient topology in the context of metrizable spaces. In particular, we show that finite homotopy colimits (=homotopy direct limits) of metrizable uniform spaces and uniformly continuous maps are metrizable when done uniformly (Corollary 3.24); they are certainly not metrizable with the quotient topology, as we will now discuss. Which topology on the cone is the 'right one' ? It turns out that in many situations where some actual work is being done, the quotient topology does not do its job, and has to be replaced by something else: (i) In his classifying space construction BG = (G * G * . . . )/G, Milnor had to use a weak (initial) topology on his joins (including finite joins) rather than the strong (final) topology of the quotient [64] (with his reversed terminology he called his topology a "strong topology").
(ii) In showing that the usual homotopy category is a closed model category in the sense of Quillen ("The homotopy category is a homotopy category" [83]), A. Strøm had to modify the quotient topology of the mapping cylinder in order to show that if f : E → B is a (Hurewicz) fibration, then so is the projection MC(f ) → B. His modified mapping cylinder can be identified with a subspace of Milnor's modified join.
There are other examples of this sort in the literature (including e.g. teardrop neighborhoods); J. Strom called attention to them in his questions on Mathoverflow.net.
It turns out that the trouble disappears entirely in the setting of uniform spaces, which is the subject of §3 below. If X ⊃ A f − → Y is a uniformly continuous partial map of metric spaces, where A is closed in X, we show that the adjunction space X ∪ f Y with the quotient uniformity (hence also with the topology thereof) is metrizable, by an explicit metric (Theorem 3.8). This yields natural constructions of cone, join and mapping cylinder in the category of metrizable uniform spaces, which we show to coincide with various other natural constructions (see §3.B). In particular, we show equivalence, up to uniform homeomorphism, of a number of definitions of join of metric spaces: one based on quotient uniformity; another based on embedding in a normed linear space; and those based on the amalgamated union X × CY ∪ CX × Y , where the cones are defined using either any of the previous methods or the approach of geometric group theory, based on the isotropy of the l 2 metric (i.e., the Law of Cosines).
1.2. Topological fundamental group. Among other serious troubles with quotient topology is that the (usual) fundamental group of a compact metric space can fail to be a topological group as the quotient of the space of maps [30]. The basic reason is that the product of quotient maps is not always a quotient map.
Again, there is no such problem with (the topology of) the quotient uniformity. The product of two quotient maps is a quotient map in the category of uniform spaces and uniformly continuous maps (cf. [50; Exercise III.8(c)]; less trivially this is so also for infinite products [47]). Using this it is easy to show that for a compact metric space X, the fundamental group of X is a topological group with the topology of the quotient uniformity, as a quotient of the space of maps (S 1 , pt) → (X, pt) endowed with the metric d(f, g) = sup x∈S 1 (f (x), g(x)). Remark 1.3. We note that this topology on the fundamental group of X differs from the more intriguing topology considered in [62]; but it is noteworthy that the topology of the still more intriguing Steenrod fundamental group of X considered in [62] coincides with the topology of the quotient uniformity, as a quotient of the space of base-ray preserving maps from the 2-disk into the one-point compactification of the mapping telescope.
1.B. Absolute retracts
In working with polyhedra it is convenient to separate combinatorial issues (such as simplicial approximation and pseudo-radial projection) from topological ones, which are well captured by the notion of an ANR. For instance, finite homotopy limits (=homotopy inverse limits) of PL maps between compact polyhedra are still ANRs, but no longer polyhedra in general.
The second major goal of the present paper is to prepare for the treatment of uniform polyhedra in the sequel to this paper by advancing a theory of uniform ANRs roughly to the level of the classical theory of ANRs as presented in the books by Borsuk [15] and Hu [44]. In particular, we show that finite homotopy limits and colimits of uniform ANRs are still uniform ANRs (Theorem 4.38) -although we will see in the sequel that those of uniform polyhedra and "uniformly PL" maps are no longer uniform polyhedra in general.
Uniform ANRs.
Our notion of a uniform ANR is not entirely standard. Two best-known analogues of ANRs in the uniform world are the semi-uniform ANRs studied by Michael and Torunczyk (see Remarks 4.30(b) and 4.31) and the ANRUs of Isbell, which we revisit in §4.A. While semi-uniform ANRs are more manageable in some respects, they are at best a useful but technical tool, involving a mix of topological and uniform notions. On the other hand, as long as metrizable spaces are concerned, it was realized independently by Garg and Nhu that Isbell's ANRUs are only a part of the story -namely the complete part. This understanding is, however, scarcely known, and is not well established in the literature: Garg mentioned what we now call uniform ANRs only in passing (so did not even give them any name); whereas Nhu's metric uniform ANRs (see Remark 4.30(a)), although do coincide with our uniform ANRs, but somewhat accidentally -for his metric uniform ARs differ from our uniform ARs. Above all, there seems to have been no good intuition and no readily available technique for dealing with non-complete uniform ANRs, as compared with complete ones (i.e. ANRUs).
This is now entirely changed, for we show that a metrizable uniform space is a uniform ANR if and only if its completion is an ANRU, and the remainder can be instantaneously taken off itself by a uniform self-homotopy of the completion (Theorem 4.10). Moreover, in many ways uniform ANRs turn out to be easier, and not harder than ANRUs. In particular, we show (Theorem 4.21) that a metrizable uniform space is a uniform ANR if and only if it is uniformly locally contractible (in the sense of Isbell) and satisfies the Hahn approximation property (in the sense of Isbell). This result, proved by an infinite process, improves on the metrizable case of Isbell's characterization of ANRUs as those uniform spaces that are uniformly locally contractible and satisfy the Hahn property and the uniform homotopy extension property. (Isbell's uniform homotopy extension property is for possibly non-closed subsets; it follows from completeness along with the uniform homotopy extension property for closed subsets.) The above two characterizations of uniform ANRs are at the heart of a toolkit that enables one to deal with uniform ANRs just as easily as with compact ANRs. Indeed we establish uniform analogs of what appears to be the core results of the usual theory of retracts, including Hanner's ε-domination criterion (Theorem 4.22(a)) and J. H. C. Whitehead's theorem on adjunction spaces (Theorem 4.32). We also show (see Theorem 4.35 or Corollary 4.44) that the space of uniformly continuous maps from a metrizable uniform space to a uniform ANR is a uniform ANR, and extend this to maps of pairs, which is the nontrivial part. In particular, this shows that the loop space of a compact polyhedron is a uniform ANR (Corollary 4.37(b)).
1.C. Inverse limits
The final section elaborates on a uniform theory of sequential inverse limits. The role of Mardešić resolutions is played by convergent inverse sequences (see Lemma 5.12), thereby reducing much of the hassle to the simple condition of convergence. (It can be viewed as a weakening of the surjectivity of all bonding maps, and specializes to the Mittag-Leffler condition in the case of inverse sequences of discrete spaces.) This enables us to generalize to metrizable uniform spaces virtually all known theory of inverse sequences of compacta. In particular, we establish the analogue of Milnor's lemma on extension of a map between inverse limits to the infinite mapping telescopes (Theorem 5.14 and Corollary 5.16), which amounts to a foundation of strong shape theory (see [62]). Another noteworthy result, whose compact case the author has not seen in the literature, is a characterization of inverse limits that are uniform ANRs in terms of properties of the bonding maps (Theorem 5.18).
A review of uniform spaces
This section is intended to serve as an introduction to uniform spaces for the reader who has little to no previous acquaintance with the subject. It appears to be rather different in viewpoint and in style from the existing introductions to uniform spaces in the literature, but mathematically it does not contain anything new. The section attempts to be self-contained modulo straightforwardly verified facts, for which it gives references.
The shorter subsection on metrizable uniform spaces and the denser subsection on finiteness conditions are intended to be read carefully by the reader not familiar with these matters; whereas the more leisurely subsections on general uniform spaces are largely intended to serve as a reference.
2.1. References on uniform spaces. Unsurpassed basic references for uniform spaces are still those to the founders of the subject: Isbell's book [50], which is well complemented by Chapters II, IX and X of Bourbaki's General Topology [17] (recall that the original Bourbaki group included A. Weil and J. Dieudonné). See also the historic survey [12]; further surveys exist [41], [55]. Other specialized sources include books by A. Weil (1937; in French), J. W. Tukey (1940), and I. M. James (1990) on uniform spaces.
Much of modern development of the theory of uniform spaces seems to occur not within topology but, in particular, in Geometric Nonlinear Functional Analysis (see [13], [54]).
2.A. Metrizable uniform spaces
2.2. Uniform continuity. We recall that a map f : M → N between metric spaces is uniformly continuous iff any two sequences x i , y i ∈ M with d(x i , y i ) → 0 as i → ∞ satisfy d(f (x i ), f (y i )) → 0 as i → ∞. (This sequential formulation of the familiar ε-δ definition will often be more convenient for our purposes.) Furthermore, it is not hard to see that f is uniformly continuous iff d 2.3. Metrizable uniformity. A uniform homeomorphism between metric spaces is a bijection that is uniformly continuous in both directions. Two metrics d and d ′ on a set S are uniformly equivalent if id S is a uniform homeomorphism between (S, d) and (S, d ′ ). In particular, every metric d is uniformly equivalent to the bounded metrics d ′ (x, y) = min(d(x, y), 1) and d ′′ (x, y) = d(x,y) 1+d(x,y) . A metrizable uniformity (or uniform structure) u on S is a uniform equivalence class of metrics on S; each of these metrics induces u; and a metrizable uniform space is a set endowed with a uniformity. Clearly, the topology induced by a metric d is determined by the uniformity induced by d. A uniformity u on a set S is finer than u ′ if id S : (S, u) → (S, u ′ ) is uniformly continuous.
Remark 2.4. Historically, the idea of a metrizable uniform space emerged together with those of a metric space and a metrizable topological space. Fréchet's thesis (1906, based on a series of 1904-05 papers), which introduced metric spaces as well as a variant of topological spaces based on limit points of sequences, also studied an axiomatic structure midway between metric and metrizable uniform spaces (see [12; §1.1]). Sets of axioms satisfied by the relation d(A, B) = 0 between the subsets A, B of a metric space have been considered by F. Riesz in the same ICM talk (1908, based on a 1906 paper), where he suggested a modification of Fréchet's approach based on limit points of sets as opposed to countable sequences (see [12; §1.5] and [21]).
Completeness.
We recall that a sequence of points x n of a metric space M is called a Cauchy sequence if for every ε > 0 there exists a k such that for every j > k, the ε-neighborhood of x j in X contains x k . Clearly, this notion depends only on the underlying uniform structure of M. Thus a metrizable uniform space is called complete if every its Cauchy sequence converges. Every metrizable uniform space is a dense subset of a unique complete one, which is called its completion; every uniformly continuous map into a complete space uniquely extends over the completion of the domain (see [50]). Every compactum, i.e. a compact metrizable space, admits a unique uniform structure, which is complete. A metrizable uniform space is called precompact if its completion is compact. Thus a subspace of a complete metrizable uniform space is precompact iff its closure is compact.
2.6. Covers. We recall that a cover (or a covering) of a set S is a collection of subsets of S whose union is the whole of S. A cover C of S is said to refine a cover D of S if every U ∈ C is a subset of some V ∈ D. If C is a cover of S, and f : T → S is a map, we have the covers f (C) := {f (U) | U ∈ C} and f −1 (C) := {f −1 (U) | U ∈ C} of T ; in the case where T ⊂ S and f is the inclusion map, we denote f −1 (C) by C ∩ T and call it the trace of C on T . If C and D are covers of S, then C ∧ D := {U ∩ V | U ∈ C, V ∈ D} is a cover of S refining both C and D. Similarly one defines the meet C λ of a finite family of covers C λ ; note that if the family is empty, its meet is the singleton cover {X} of X.
If T ⊂ S is a subset, the star st(T, C) of T in C is the union of all elements of C that intersect T . We say that C star-refines D if the cover {st({x}, C) | x ∈ S} refines D. Next, C is said to strongly star-refine D if {st(U, C) | U ∈ C} refines D. It is easy to see that every strong star-refinement is a star-refinement, and every star-refinement of a star-refinement is a strong star-refinement.
Uniform covers.
A cover C of a metric space M is said to be uniform if there exists a positive number λ such that every subset of M of diameter < λ is contained in some U ∈ C; such a λ is called a Lebesgue number of C. Note that if a cover C of M is refined by the cover C ε by all balls of radius ε, then C is uniform (with Lebesgue number ε); and conversely, every uniform cover of M with Lebesgue number λ is refined by C λ/2 .
It is easy to see that a map f : M → N between metric spaces is uniformly continuous iff for every uniform cover D of N, the cover f −1 (D) of M is uniform (equivalently, there exists a uniform cover C of M such that f (C) refines D ∩ f (M)). It follows that the property of being uniform for a cover of M depends only on the underlying uniform structure of M. Clearly, a cover of a compactum is uniform iff it can be refined by an open cover. A metrizable uniform space X is precompact iff every uniform cover of X has a finite uniform refinement (see [50; II.28]).
A family of disjoint subsets X α ⊂ M is called uniformly discrete if it constitutes a uniform cover of its union. (In other words, if there exists an ε > 0 such that d(X α , X β ) > ε whenever α = β, where d(X, Y ) = sup{d(x, y) | x ∈ X, y ∈ Y }.) The space M itself is called uniformly discrete if the collection of its singletons is uniformly discrete. A neighborhood U of a subset S of a metrizable uniform space X is called uniform if it contains the star of S in some uniform cover of X; or equivalently if S and X \ U constitute a uniformly discrete collection. The space M is called uniformly connected if contains no subset that is its own uniform neighborhood.
2.8. Basis of uniformity. Uniform covers can be used to axiomatize the notion of a uniform structure. Let us call a sequence of covers C 1 , C 2 , . . . of a set S fundamental if it satisfies (1) each C n+1 star-refines C n ; (2) for any distinct points x, y ∈ S there exists an n such that no element of C n contains both x and y. Two fundamental sequences of covers C n and D n are equivalent if for each n there exists an m such that C m refines D n and D m refines C n . A basis for a metrizable uniformity u on S is a fundamental sequence of covers C n such that each C n is uniform with respect to u, and every uniform cover C of (S, u) is refined by some C n . Clearly, every two bases of u are equivalent; and every fundamental sequence of covers of S that is equivalent to a basis of u is itself a basis of u. On the other hand, if d is a metric on S inducing u, then the covers C n of S by the balls of radius 3 −n about all points of (S, d) form a standard basis of u. Theorem 2.9 (Alexandroff-Urysohn [3]). There exists a bijection between metrizable uniformities on S and equivalence classes of fundamental sequences of covers of S, which assigns to a uniformity the equivalence class of any its standard basis.
Of course, the statement in the 1923 paper [3] is in different terms, even though the proof is essentially the same as that given below. It is likely, however, that the authors must have been at least partially aware of this interpretation of their result, since according to Fréchet (1928;cf. [12; p. 585]), they have been thinking of avoiding the use of metric in defining the notion of uniform continuity.
Proof. It remains to show that every fundamental sequence of covers C n of S is a basis of some metrizable uniformity. To this end consider an auxiliary 'pre-distance' function f (x, y) = inf{2 −n | x, y ∈ U for some U ∈ C 2n }, and define d(x, y) to be the infimum of the sums f (x 0 , x 1 ) + · · · + f (x n−1 , x n ) over all finite chains x = x 0 , . . . , x n = y of points of S. Clearly, d is a pseudo-metric, i.e. it is symmetric, satisfies the triangle axiom and is such that d(x, x) = 0 for every x ∈ S. Let D 2n−1 be the set of all subsets of S of d-diameter at most 2 −n . Since d(x, y) ≤ f (x, y), each U ∈ C 2n also belongs to D 2n−1 , thus C 2n refines D 2n−1 . To prove that d is a metric and that {C 2n } is a basis for the uniformity induced by d is suffices to show that each D 2n+1 refines C 2n . The latter, in turn, would follow if we prove that f (x, y) ≤ 2d(x, y). Let us show that f (x, y) ≤ 2[f (x 0 , x 1 ) + · · · + f (x n−1 , x n )] for every finite chain x = x 0 , . . . , x n = y of points of S.
2.B. General nonsense I: Objects
We shall work with concrete categories over the category of sets, that is "constructs" in the terminology of The Joy of Cats [1]. As a bridge between sets (which we have to start from anyway) and abstract categories (whose powerful machinery we do need), they enable a unified treatment of constructions from Isbell [50] and Bourbaki [17].
2.10. Uniform spaces. In some auxiliary constructions (such as quotient uniformities and semi-uniform products of metrizable uniformities) we will have to deal with general (possibly non-metrizable) uniform spaces. These will be the subject of the remainder of this section.
We call a nonempty family of covers {C α } of a set S fundamental if (0) each C α and C β are refined by some C γ ; (1) each C α is star-refined by some C β ; and (2) for any distinct points x, y ∈ S there exists an α such that no element of C α contains both x and y.
A uniformity u on S is the set of all covers that are refined by some element of a fundamental family of covers {C α }. These covers are called uniform, the family {C α } is called a basis of u, and the pair uS = (S, u) is called a uniform space. A map between two uniform spaces f : X → Y is called uniformly continuous if for every uniform cover C of Y , the cover f −1 (C) of X is uniform.
As far as single (countable) coverings are concerned, all uniform spaces are like (separable) metrizable uniform spaces (to rephrase an expression of Isbell), in the sense that every (countable) uniform cover C of a uniform space X is refined by f −1 (D) for some (countable) uniform cover D of some (separable) metrizable Y and some uniformly continuous f : X → Y [50; I.14], [36; 3.1].
If we drop condition (2), we obtain definitions of a pre-fundamental family of covers and a pre-uniformity (some authors call this a "uniformity", and refer to one satisfying (2) as a "separated uniformity"). In other words, a pre-uniformity on S is a family of covers of S that forms a filter with respect to star-refinement (cf. [50; I.6]). Every pseudo-metric induces a pre-uniformity. Similarly to Theorem 2.9, a pre-uniformity is pseudo-metrizable iff it has a basis that is a pre-fundamental sequence of covers C i (i.e. satisfies condition 2.8(1)).
Also similarly to Theorem 2.9, every pre-uniformity has a basis of covers C iα , where each C iα consists of all balls of radius 2 −i with respect to some pseudo-metric d α (cf. [50; proof of I.14], [29; proof of 8.1.10]). This yields a bijective correspondence between pre-uniformities on S and uniform equivalence classes of collections D of pseudo-metrics on S such that (i) for any d, d ′ ∈ D there exists a d ′′ ∈ D with d ′′ ≥ max(d, d ′ ); uniformities correspond to the equivalence classes of collections D such that (ii) for each pair of distinct points x, y ∈ S there exists a d ∈ D such that d(x, y) > 0 (cf. [29; 8.1.18]). Two such collections D and D ′ are uniformly equivalent if id : (X, D) → (X, D ′ ) is uniformly continuous in both directions. A definition of uniform continuity in relevant terms is, a map f : (X, D) → (Y, E) is uniformly continuous iff for each ε > 0 and e ∈ E there exists a δ > 0 and a d ∈ D such that d(x, y) ≤ δ implies e(f (x), f (y)) ≤ ε. A related criterion is, a function between pre-uniform spaces f : X → Y is uniformly continuous iff for each uniformly continuous pseudo-metric e on Y , the pseudo-metric d(x, y) = e(f (x), f (y)) is uniformly continuous (cf. [29; 8.1.22]).
If X is a uniform space and S is a subset of X, the uniform structure of subspace on S is given by the covers C ∩ S, where C runs over all uniform covers of X.
If X is a pre-uniform space, its induced topology is defined by declaring a subset S ⊂ X open iff for each x ∈ S there exists a uniform cover C of X such that st(x, C) ⊂ S. In other words, a base of neighborhoods of x is given by the stars of x in basic uniform covers of X. Every uniform cover of X is refined by an open cover, since this is obviously so in the metrizable case (cf. [50; I.19]). Thus every pre-uniformity has a basis consisting of open covers.
We refer to [50] for the definition and properties of complete uniform spaces.
Let U (resp.Ū ) denote the category of (pre-)uniform spaces and uniformly continuous maps, and T the category of topological spaces and continuous maps -all viewed as concrete categories over the category of sets.
2.11. Initial uniformity. Given a set X and a family f of maps f λ : X → Y λ into pre-uniform spaces, all finite meets of the form f −1 , where each C i is a uniform cover of Y λ i , clearly form a basis of a pre-uniformity u f . Clearly u f is the coarsest pre-uniformity on X making all the f λ uniformly continuous (cf. [50; I.8]). Moreover, it is easy to see that u f is initial inŪ with respect to f ; that is, a map g : Z → X, where Z is a pre-uniform space, is uniformly continuous if (and, obviously, where the non-trivial part of the argument is redundant). Conversely, if a pre-uniformity on X is initial with respect to f , then it has to be the coarsest pre-uniformity making all the f λ uniformly continuous (cf. [1; 10.43]). Thus we may call u f the initial pre-uniformity inŪ with respect to f . The induced topology of u f is initial in T with respect to f [50; I.16]. Corresponding to the empty family ∅ of maps on X we have the anti-discrete pre-uniformity u ∅ = {{X}}, which is not a uniformity (cf. [1; 8.3]).
The pre-uniformity u f is a uniformity (and is initial in U with respect to f ), provided that each Y λ is a uniform space, and f is point-separating, i.e. for every pair of distinct points x, y ∈ X there exists a λ such that f λ (x) = f λ (y) (cf. [50; I.8 and I.17]).
2.12. Finest uniformity. (Pre-)uniformities on a set X are ordered by inclusion, as subsets of the set of all covers of X. Given a family u of pre-uniformities u λ on a set X, the initial pre-uniformity u f corresponding to the family f of maps f λ : X → (X, u λ ) is the least upper bound sup u of the family u (cf. [17; §II. 1.5]). If at least one u λ is a uniformity, then so is sup u. By the above, A cover C 1 of a set X is called normal with respect to a family F of covers of X, if it can be included in an infinite sequence C 1 ; C 2 , C 3 , . . . of covers of X such that each C i+1 star-refines C i , and each C i is refined by some element of F . If F is nonempty and every two elements of F have a common refinement in F , then it is easy to see that the family of covers of X, normal with respect to F , constitutes a pre-uniformity u F on X. Clearly, u F is the finest among those pre-uniformities u λ that have a basis contained in F ; that is, u F = sup u and u F ∈ u, where u is the family of all the u λ (cf. [50; I.10]).
2.13. Fine uniformity. The induced topology of every uniformity is Tychonoff (=completely regular Hausdorff=T 3 1 2 ) [50; I.11], and every pre-uniformity whose induced topology is T 1 is clearly a uniformity. Given a Tychonoff topological space X, its topology is initial with respect to the family f of all continuous maps f λ : X → R, and therefore is induced by the uniformity u f , with respect to the usual uniformity on R (cf. [50; I.15]). Once the set of uniformities inducing the given topology on X is non-empty, there exists a finest such uniformity u X , consisting of all covers of X that are normal with respect to the family of all open covers of X (cf. [50; I.20]). This is the fine uniformity of the Tychonoff topological space X. A map from (X, u X ) into a pre-uniform space is uniformly continuous iff it is continuous, and u X is characterized by this property [17; Exer. IX. 1.5]; at the same time, a map (X, u f ) → R is uniformly continuous iff it is continuous. By [17; Exer. IX. 1.5], u X corresponds to the family of all pseudo-metrics on X that are uniformly continuous as functions X × X → R; whereas by [17; Example at the end of §IX.1.2], u f corresponds to the family of all pseudo-metrics It is well-known that every open cover of a paracompact topological space has an open star-refinement (cf. [29; 5.1.12]) and that every metrizable topological space is paracompact (see [14] or [29]). It follows that the fine uniformity of a metrizable topological space X consists of all covers that can be refined by open covers. This uniformity is itself almost never metrizable -specifically, it is metrizable if and only if the set K of non-isolated points of X is compact, and the complement to any uniform neighborhood of K is uniformly discrete (see [4], [56], [75]).
2.14. Final pre-uniformity. Given a set X and a family f of maps f λ : Y λ → X from pre-uniform spaces, the collection u F of all covers of X, normal with respect to the family F of all covers C of X such that the cover f −1 λ (C) of Y λ is uniform for each λ, is a pre-uniformity u f on X. Clearly u f is the finest pre-uniformity on X making all the f λ uniformly continuous (cf. [50; Exer. I.7]). Moreover, it is easy to see that u f is final inŪ with respect to f ; that is, a map g : (X, u f ) → Z, where Z is a preuniform space, is uniformly continuous if (and, obviously, only if) each composition is uniformly continuous. (To see this, note that if D ′ star-refines D, then g −1 (D ′ ) star-refines g −1 (D).) Conversely, if a pre-uniformity on X is final with respect to f , then it is the finest pre-uniformity making each f λ uniformly continuous (this is the dualization of [1; 10.43]). Thus we may call u f the final pre-uniformity in U with respect to f . Corresponding to the empty family ∅ of maps into X we have the discrete pre-uniformity u ∅ , which is a uniformity (cf. [1; 8.1]). We note that the fine uniformity of a topological space (S, t) is nothing but the final uniformity corresponding to the family of inclusions in S of all compacta in (S, t) considered as uniform spaces (with their unique uniformity).
In general, the question when u f is a uniformity is not easy (see [50; Exer. I.7], [36; Theorem 2.2] for partial results). Instead, one has the following construction.
2.15. Uniform space associated to a pre-uniform space. If X is a pre-uniform space, let ∼ be the separating equivalence relation on X defined by x ∼ y iff every uniform cover of X contains an element U such that x, y ∈ U. Equivalently, x ∼ y iff d(x, y) = 0 for each uniformly continuous pseudo-metric on X. Let f : X → X/ ∼ assign to each point its separating equivalence class. If C is a uniform cover of X, then f (C) is normal with respect to f ; indeed, f −1 (f (U)) ⊂ st(U, C ′ ) for every U ∈ C and every uniform cover C ′ of X (in particular, this holds with C ′ = C), and therefore f −1 (f (C)) refines D whenever C strongly star-refines D. It follows that the covers f (C), where C runs over all uniform covers of X, form the pre-uniformity u f on X/ ∼ ; by construction it is a uniformity.
2.16. Coarsest pre-uniformity. Given a family u of pre-uniformities u λ on a set X, the final pre-uniformity u f corresponding to the family f of maps f λ : (X, u λ ) → X is the greatest lower bound inf u of the family u. Alternatively, inf u = sup u * , where the set u * of lower bounds of u among all pre-uniformities on X is non-empty as it contains the anti-discrete pre-uniformity u ∅ (cf. [17; §II.1.5]). Similarly sup u = inf u * , where the set u * of upper bounds of u among all pre-uniformities on X is non-empty as it contains the discrete uniformity u ∅ . By the above, inf u consists of all covers, normal with respect to the family λ u λ .
If F is a family of covers of a set X such that every C ∈ F is normal with respect to F , then it is easy to see that the family of covers C 1 ∧ · · · ∧ C k , where k ∈ N and each C i ∈ F , is a base of a pre-uniformity u F on X. Clearly, u F is the coarsest among those pre-uniformities u λ that contain F ; that is, u F = inf u and u F ∈ u, where u is the family of all the u λ (cf. [50; I.9]).
Coarse uniformity.
A Tychonoff space X admits a coarsest uniformity u X inducing its topology iff X is locally compact; when X is locally compact, u X coincides with the uniformity of the subspace of the one-point compactification of X, as well as with the initial uniformity with respect to the family of all continuous maps X → R that vanish on the complement to a compact set [ 2.18. Product. The product X λ of uniform spaces X λ is their set-theoretic product X endowed with the initial uniformity with respect to the family of projections π λ : X → X λ (cf. [1; 10.53]). Thus the induced topology of X λ is the product topology, and a cover of X λ is uniform iff it is refined by π −1 λ 1 (C 1 ) ∧ · · · ∧ π −1 λ k (C k ) for some uniform covers C 1 , . . . , C k of some finite subcollection X λ 1 , . . . , X λ k . It is easy to check that (X, π λ ) is also the product of X i 's in U in the sense of abstract category theory (see [50; p. 14]).
Now let X 1 , X 2 , . . . be metrizable uniform spaces and for each i let C (i) n , n = 1, 2, . . . be a basis of uniform covers of X i . Then a basis of uniform covers of X i is given by n−1 ) ∧ · · · ∧ π −1 n (C (n) 1 ), which is indeed a fundamental sequence. It follows that if d i is a metric on X i such that X i is of diameter at most 1, the uniformity on X i is induced by the l 1 metric d(( 2.19. Disjoint union. The disjoint union X λ of uniform spaces X λ is their settheoretic disjoint union X endowed with the final pre-uniformity with respect to the injections ι λ : X λ → X (cf. [1; 10.67(2)]). This pre-uniformity is obviously a uniformity (cf. [50; Exer. I.7(i)]). A cover C of X λ is uniform iff ι −1 λ (C) is uniform for each λ; indeed, every cover C satisfying the latter condition is star-refined by another such cover ι λ (C λ ), where each C λ star-refines ι −1 λ (C). It is easy to check that (X, ι λ ) is the coproduct of X i 's in U in the sense of abstract category theory (see [50; p. 14]), and that its underlying topology is the topology of disjoint union (see [50; II.8]).
We note that infinite disjoint unions of metrizable uniform spaces normally fail to be metrizable. Finite disjoint unions preserve metrizability. Indeed, if X and Y are metric spaces of diameter ≤ 1, we can extend their metrics to a metric on the set-theoretic disjoint union of X and Y by d(x, y) = 1 whenever x ∈ X, y ∈ Y ; clearly, it induces the uniformity of the disjoint union.
A metrizable replacement of the countable disjoint union X i is the inverse limit of the finite disjoint unions Y i := (X 1 ⊔ · · · ⊔ X i ) ⊔ N and the maps f i : More generally, a family f of uniformly continuous maps f λ : X → Y λ is a monosource inŪ iff it is point-separating; the former means that two maps g, h : Z → X coincide whenever the compositions Z Dually, a family f of uniformly continuous maps f λ : Y λ → X is an epi-sink inŪ iff it is jointly surjective (i.e. λ f (Y λ ) = X); the former means that two maps (a) Given a family of uniformly continuous maps f λ : X → Y λ between pre-uniform spaces, there exist a uniformly continuous surjection h : X → Z and a point-separating family g of uniformly continuous maps g λ : Z → Y λ , where Z has initial uniformity with respect to g, such that each f λ = g λ h. Such a factorization is unique up to uniform homeomorphism.
(b) Given a family of uniformly continuous maps f λ : Y λ → X between pre-uniform spaces, there exist a uniformly continuous injection h : Z → Y and a jointly surjective family g of uniformly continuous maps g λ : Y λ → Z, where Z has final uniformity with respect to g, such that each f λ = hg λ . Such a factorization is unique up to uniform homeomorphism.
Embedding.
A uniformly continuous map f : A → X of uniform spaces is called an embedding if it is injective, and the uniformity on A is initial with respect to f (cf. [1; 8.6]). Thus if f : A → X is an embedding, a basis of the uniformity of A is given by the covers f −1 (C), where C runs over all uniform covers of X. In fact, all uniform covers of A are of this form; for if D is refined by f −1 (C), then the cover E : to the injectivity of f . Thus an injective map between uniform spaces is an embedding iff it is a uniform homeomorphism onto its image with the subspace uniformity.
Composition of embeddings is an embedding; and if the composition X f − → Y → Z is an embedding, then so is f [1; 8.9]. By Proposition 2.21(a), every uniformly continuous map between pre-uniform spaces is a composition of a uniformly continuous surjection and an embedding; this factorization is unique up to uniform homeomorphism (cf. [50; II.5]).
Extremal and regular monomorphisms.
Embeddings coincide with extremal monomorphisms inŪ and also with regular monomorphisms inŪ [1; 21.13(4) and 21.8(1)]. AŪ-monomorphism (i.e. a uniformly continuous injection) f : A → X is extremal inŪ if, once f factors inŪ through aŪ-epimorphism (i.e. a uniformly continuous surjection) g : A → B, this g must be a uniform homeomorphism. A uniformly continuous map f : A → X is a regular monomorphism inŪ if there existŪ-morphisms g, h : X → Y such that f is their equalizer; that is, gf = hf , and anyŪ-morphism It is easy to see that extremal monomorphisms in U coincide with embeddings onto closed subspaces. To see that regular monomorphisms in U coincide with embeddings onto closed subspaces, note that if A is a closed subspace of a uniform space X, then for each x ∈ X \ A there exists a uniformly continuous map g x : X → [0, ∞) such that g x (A) = {0} and g x (x) = 0 [50; I.13]; consequently, the uniformly continuous map It follows that the pullback of an embedding is an embedding [1; 11.18].
Quotient.
A uniformly continuous map f : X → Q of pre-uniform spaces is called a quotient map if it is surjective, and the pre-uniformity on Q is final with respect to f (cf. [1; 8.10]). This quotient pre-uniformity therefore consists of all covers C 1 of Y such that C 1 admits a star-refinement C 2 , which in turn admits a star-refinement C 3 , etc., so that f −1 (C i ) is uniform for each i. Note that the uniform space associated to a pre-uniform space (see 2.15 above) is its quotient. On the other hand, if X is a uniform space, then the quotient pre-uniformity does not need to be a uniformity in general. For instance, if f has a non-closed point-inverse, then no uniformity on Q can make f uniformly continuous. See [36; Theorem 2.2] for a characterization of quotients of uniform spaces whose pre-uniformity is a uniformity. Given a uniform space X and an equivalence relation R on the underlying set of X, the quotient X/R is the set of equivalence classes of R endowed with the quotient pre-uniformity. Composition of quotient maps is a quotient map; and if a composition X → Y f − → Z is a quotient map, then so is f [1; 8.13]. Every uniformly continuous retraction is a quotient map [1; 8.12 (2)]. By Proposition 2.21(b), every uniformly continuous map between pre-uniform spaces is a composition of a quotient map and a uniformly continuous injection; by the above, the latter is in turn a composition of a uniformly continuous bijection and an embedding; this factorization into three maps is unique up to uniform homeomorphism (cf. [50; II.5]).
Extremal epimorphisms in U coincide again with quotient maps, for they coincide with extremal epimorphisms inŪ as long as they are surjective -which they have be due to the second condition in the definition of an extremal epimorphism. To see that regular epimorphisms in U coincide with quotient maps, note that a quotient map q : X → Q is the coequalizer of the projections of the subspace {(x, y) | q(x) = q(y)} of X × X onto the factors; and the coequalizer of a pair of maps from a pre-uniform space Y into X equals the coequalizer of the resulting maps from the uniform space associated to X.
It follows that the pushout of a quotient map is a quotient map (dually to [1; 11.18]).
2.26. Extremal mono-sources and epi-sinks. The uniqueness part of Proposition 2.21 implies the following. A family f ofŪ-morphisms f λ : X → Y λ is an extremal mono-source inŪ iff it is point-separating, and X has the initial uniformity with respect to f . The extremality means that once f factors inŪ through aŪ-epimorphism (i.e. a uniformly continuous surjection) g : X → Z, this g must be a uniform homeomorphism.
Dually, a family f ofŪ-morphisms f λ : Y λ → X is an extremal epi-sink inŪ iff it is jointly surjective, and X has the final uniformity with respect to f . Here the extremality means that once f factors inŪ through aŪ-monomorphism (i.e. an uniformly continuous injection) p : R → Q, this p must be a uniform homeomorphism.
Combining the above with [1; 10.26 (2)] and its dual, we obtain and only if f is point-separating and X has the initial uniformity with respect to f .
(b) Let f be a family ofŪ-morphisms f λ : Y λ → X. Then f : λ Y λ → X is a quotient map if and only if f is jointly surjective and X has the final uniformity with respect to f .
(a) If each f λ is an embedding, then so is From [1; 27.15] (see also [1; 28.14]) and its dual we deduce Proposition 2.29. (b) If g : A → X is an embedding and Y is a uniform space, then g ⊔id Y : A⊔Y → X ⊔Y is an embedding. In particular, ι 2 : Y → X ⊔Y is an embedding.
In fact, a product of quotient maps is a quotient map [47; Theorem 1] (for the case of a finite product see also [50; Exer. III.8(c)]). A product of two sequential direct limits is the sequential direct limit of the products [8] (which is not the case for topological spaces).
2.30.
Graph. If f : X → Y is a possibly discontinuous map between uniform spaces, inclusion and the projection, whose inverse is given by X In particular, every continuous map f : X → Y between uniform spaces is the composition of the homeomorphism X → Γ f and the uniformly continuous map Γ f → X ×Y → X. Similarly, every uniformly continuous map between metric spaces is a composition of a uniform homeomorphism and a 1-Lipschitz map.
2.D. Function spaces
If X and Y are uniform spaces, and C is a uniform cover of Y , let U(X, C) be the cover of the set U(X, Y ) of uniformly continuous maps X → Y by the sets ∈ U} for all f ∈ U(X, Y ). One can check that the covers U(X, C) form a base of a uniform structure on U(X, Y ), which therefore becomes a uniform space.
Absolute extensors.
We call a uniform space Y an A[N]EU if for every (not necessarily metrizable) uniform space X and every (not necessarily closed) subspace A ⊂ X, every uniformly continuous map A → X extends to a uniformly continuous AEUs are also known as "injective spaces" [50] since they are the injective objects of the category of uniform spaces and uniformly continuous maps.
In particular, U(D, I) is an AEU.
Proof. If Z a uniform space, it is easy to see (cf. [50; III.13]) that a map f : We note that U(N, I) is inseparable, where N denotes the infinite countable discrete space and I = [0, 1]. On the other hand, q 0 := U((N + , ∞), (I, 0)) is separable, where + stands for the one-point compactification. (We recall that the functional space U((N + , ∞), (R, 0)) is known as c 0 .) Clearly, r is a uniformly continuous retraction. Since U(N, I) is an AEU, we infer that so is q 0 .
2.34. Semi-uniform product. A subset S ⊂ U(X, Y ) is called uniformly equicontinuous (or equiuniformly continuous) if for each uniform cover D of Y there exists a uniform cover C of X such that for each f ∈ S, the cover f −1 (D) is refined by C (equivalently, f (C) refines D ∩ f (X)).
Given a map f : • f is uniformly continuous and its image is uniformly equicontinuous; • F is uniformly continuous; • F (z, * ) : X → Y is a uniformly equicontinuous family and F ( * , x) : Z → Y is a uniformly equicontinuous family.
The semi-uniform product Z ⋉ X of a uniform space Z and a metrizable uniform space X is the set Z × X endowed with the uniformity with basis consisting of covers Given a map f : • f is uniformly continuous; a uniformly equicontinuous family.
Remark 2.35. Note that the uniform continuity of Φ(z, * ) : X → Y for each z ∈ Z and of Φ( * , x) : Z → Y for each x ∈ X is equivalent to the uniform continuity of f : Z → U pw (X, Y ), where U pw is the space of uniformly continuous maps with the uniformity of pointwise convergence; the corresponding uniformity X ⊲⊳ Y on the product X × Y need not be pre-compact even if both X and Y are compact [50; Exer. III.7]. Another symmetric version sup(u XY , u Y X ) of the uniformity u XY of X ⋉ Y is studied in [10].
We note that the proof essentially involves non-metrizable spaces when Y is metrizable. A rather technical proof, not involving non-metrizable spaces, of the case Y = I is found in [50; III.18].
Proof. If X is an AEU, every uniformly continuous map
The ANEU case is similar, using additionally that every uniform neighborhood of C ⋉ X in Z ⋉ X contains U ⋉ X for some uniform neighborhood U of C in Z. [11]).
2.E. Finiteness properties of coverings
It is not hard to see that every uniform cover of a separable uniform space has a countable uniform refinement (cf. [36; 2.3]).
Dimension.
A cover C of a set S is said to be of multiplicity ≤ m if every x ∈ X belongs to at most m elements of C. We say that a uniform space X is of dimension ≤ d if every uniform cover of X is refined by a uniform cover of multiplicity ≤ d + 1. (Isbell calls this the "large dimension" of X; however, when finite, it coincides with what he calls the "uniform dimension" [50; Theorem V.5].) We call X residually finite-dimensional if every uniform cover of X is refined by a uniform cover of finite multiplicity (in the literature these are also known as "finitistic" or "distal" uniform spaces). Note that every compactum is residually finite-dimensional.
Lemma 2.39. The completion of a residually finite dimensional space is residually finite dimensional.
Proof. This is similar to [50; IV.23] but for the reader's convenience we include a direct proof. Let C be a uniform cover of X with a Lebesgue number 3λ. Then every subset of X of diameter λ has its λ-neighborhood contained in some U ∈ C, and therefore is itself contained in The coverD ofX by the closures of the elements of D is uniform [50; II.9], and obviously its multiplicity does not exceed that of C.
2.40. Point-finite, star-finite and Noetherian spaces. We recall that a cover {U α } of a set X is called point-finite if each x ∈ X belongs to only finitely many U α 's. Next, The following implications are straightforward.
It is well-known that every metrizable space is paracompact so in particular weakly paracompact, i.e. every its open cover admits an open point-finite refinement; and that every separable metrizable space is strongly paracompact, i.e. every its open cover admits an open star-finite refinement (see [14] or [29]). It is asserted in [42; §7] that every open cover of every paracompact space admits an open Noetherian refinement.
A uniform space X is called point-finite (star-finite; Noetherian) if every uniform cover of X has a point-finite (star-finite; Noetherian) uniform refinement. We caution the reader that in the literature, "uniform paracompactness" and its variations refer to a completely different menagerie of properties, involving arbitrary open coverings besides uniform coverings.
Proposition 2.41. A uniform cover of every d-dimensional (resp. of every residually finite-dimensional) star-finite uniform space has a star-finite uniform refinement of multiplicity ≤ d + 1 (resp. of finite multiplicity). Proof. Let C be the given uniform cover of the given space X. Then C has a starfinite uniform refinement D, which in turn has a uniform refinement E of multiplicity [42]); only recently it has been shown that U(N, I) fails to be point-finite [42].
In contrast, the separable space Proof. Let D be the given uniform cover of the given space X, and let C = {U i } be a countable strong star-refinement of a strong star-refinement of D. Let us consider Theorem 2.42 also follows from Aharoni's theorem: every separable metric space admits a Lipschitz (hence uniform) embedding in c 0 (see [70; Theorem 3.1]). We find it easier, however, to prove 2.42 directly and then use it to give a short proof (following [70; 2.1]) of the relevant part of Aharoni's theorem.
Proof. Let M be the given space with some fixed metric. Let C n be the covering of M by balls of radius 2 −n . By 2.42 it has a countable uniform point-finite refinement D n . Let λ n be a Lebesgue number of D n , and suppose that If d(x, y) > 2 −n+1 , then no element of D n contains both x and y. Let V ni be an element of D n containing the 1 2 λ n -neighborhood of x. Then f (x, n, i) > 1 2 λ n and f (y, n, i) = 0.
Thus F is injective and F −1 is uniformly continuous. It is not hard to see that the separable space CN, the cone over the countable uniformly discrete space, is not star-finite. (See §3.B for the general definition of a cone; for the time being, we may define CN as the subspace of U(N, I) consisting of all functions with support in at most one point.) Every subspace of the countable product R ∞ of lines is star-finite: a fundamental sequence of star-finite covers of R ∞ is given by f −1 n (C n ), where f n : R ∞ → R n is the projection and C n is the set of the open (cubical) stars of vertices of the standard cubulation of R n by cubes with edge length 2 −n .
Proof. Let C 1 , C 2 , . . . be a basis for a (completely) metrizable uniformity uX on the given topological space X. Without loss of generality, each C i is an open cover. Set D 1 = C 1 , and suppose that an open cover D i of X has been constructed. Since X is paracompact, there exists an open star-refinement D ′ i of D i (see [29]). Then is an open cover of X refining C i+1 and star-refining has an open refinement of multiplicity ≤ d + 1, and so itself will be assumed without loss of generality to be of multiplicity ≤ d + 1.] Since X is strongly paracompact, there exists an open star-finite refinement D i+1 of D ′′ i (see [14] or [29]). Thus we obtain a sequence of star-finite open covers D 1 , D 2 , . . . such that each D i+1 refines C i+1 and [refines a cover of multiplicity ≤ d + 1 that in turn] star-refines D i . The former implies that the sequence D 1 , D 2 , . . . separates points of X, and thus is a basis for some [d-dimensional] star-finite metrizable uniformity u ′ X on the underlying set of X.
If a subset T ⊂ S is open in the induced topology of u ′ X (that is, for every x ∈ T there exists an n such that st(x, D n ) ⊂ T ), then T is open in X due to the openness of the covers D i . Since each D i refines C i , the identity map u ′ X → uX is uniformly continuous. Hence u ′ X induces the original topology of X. It remains to observe that a uniformity finer than a complete one and inducing the same topology is complete [17]; for (in the metrizable case) every Cauchy sequence in u ′ X is Cauchy in uX and hence convergent in X.
2.46. Uniform local compacta. By a local compactum we mean a locally compact separable metrizable space 1 ; or equivalently a metrizable topological space that is a countable union of compacta X i such that each X i ⊂ Int X i+1 ; or equivalently the complement to a point in a compactum; or equivalently the complement to a compactum in a compactum (see e.g. [29; 3.3.2, 3.8.C, 3.5.11]).
By a uniform local compactum we mean a metrizable uniform space that has a countable uniform cover by compacta. A map f from a uniform local compactum X into a metric space is uniformly continuous iff it is continuous and every two proper maps ϕ, ψ : N → X such that d(ϕ(n), ψ(n)) → 0 as n → ∞ satisfy d(f ϕ(n), f ψ(n)) → 0 as n → ∞. Every closed subset of a finite-dimensional Euclidean space is a uniform local compactum.
Clearly, every uniform local compactum is complete and its underlying topological space is a local compactum. The converse to the latter is false: N × N with the metric d((m, n), (m ′ , n ′ )) = 1 if m = m ′ and d((m, n), (m, n ′ )) = 1 m is not a uniform local compactum, although it is topologically discrete. However, an ANRU (see §4.A) whose underlying topological space is a local compactum is a uniform local compactum Proof. Let d be some metric on the one-point compactification of the given local compactum X.
| is a complete metric on X, inducing the same topology as d, moreover every ball (of finite radius) in (X, d ′ ) is compact.
The constructed uniform structure is not canonical: it depends in general on the choice of d in its uniform equivalence class. Proof. Let Q = {q 1 , q 2 , . . . } be a countable dense subset of X, and suppose that closed 3ε-balls in X are compact. Define Z ⊂ Q by q 1 ∈ Z, and q i+1 ∈ Z unless q i+1 is contained in the union of the open ε-balls about the points of {q 1 , . . . , q i } ∩ Z. On the other hand, let {V α } be the cover of X by the open ε/2-balls about all x ∈ X. If K is the closed 3ε-ball about some z 0 ∈ Z, its cover {V α ∩ K} has a finite subcover {W j }. Since Z contains no pair of points at distance < ε from each other, each W j contains at most one z ∈ Z. So K ∩ Z is finite. Then the cover of X by the 3 2 ε-balls about the points of Z is star-finite; clearly, it is also uniform.
2.49. Bornological conditions. Let X be a uniform space.
We recall that X is precompact if and only if for each uniform cover C of X there exists a finite set F ⊂ X such that st(F, C) = X. Every separable metrizable space admits a uniformly continuous homeomorphism onto a (non-canonical) precompact metrizable space, cf. [40].
A subset S of X is called R-bounded if for each uniform cover C of X there exists a finite set F ⊂ X and a positive integer n such that S ⊂ st n (F, C); here st 0 (F, C) = F , and st n+1 (F, C) = st(st n (F, C), C) (see [45], [17; Exer. II. 4.7]; an anticipatory definition is found in [46] and in an older edition of [17]). It is not hard to see that R-bounded subsets of a uniformly locally precompact space are precompact (cf. [38; 1.18]). On the other hand, every uniformly contractible uniform space (for instance, U(X, I) for every uniform space X) is R-bounded (as a subset of itself).
It turns out that S ⊂ X is R-bounded if and only if every uniformly continuous map of X into the real line R with the Euclidean uniformity is bounded on S [38; 1.14] (see also [4; Theorem 2], [5; Theorem 7]). In particular, R-boundedness is preserved under uniformly continuous surjections. In fact, S is R-bounded if and only if it has finite diameter with respect to every uniformly continuously pseudo-metric on X; when X is metrizable, S is R-bounded if and only if it has finite diameter with respect to every metric on X [38; 1.12, 1.13]. Further references on R-bounded spaces include [39], [33] and those therein.
A uniform space X is called fine-bounded if every uniformly continuous map from X to a fine uniform space F has precompact image. Equivalently (see [85]), if every uniformly continuous map from X to R with the fine uniformity is bounded.
A uniform space X is called N-bounded if it admits no countable cover by uniformly disjoint sets (see [85]). Equivalently, if every uniformly continuous map of X into the countable uniformly discrete space N is bounded. Clearly, The fine uniformity on R is N-bounded but not fine-bounded, and the Euclidean uniformity on R is fine-bounded but not R-bounded. However, metrizable N-bounded spaces are fine-bounded [84] (see also [77; proof of Theorem 3]).
Metrization of quotient spaces
The following proposition does not seem to be particularly helpful, but may be relevant for motivation of subsequent considerations.
Proposition 3.1. Let X be a metrizable uniform space with a basis C 1 , C 2 , . . . of uniformity, and let Y be a pre-uniform space and f : X → Y a quotient map. Consider An analogue of (a) in the language of pseudo-metrics is established in [71].
Proof.(a). Each D ϕ is star-refined by D ϕσ , where σ(k) = k + 1. Each D ϕ and D ψ are refined by D χ , where χ(n) = max(ϕ(n), ψ(n)). Thus the D ϕ for increasing ϕ form a pre-fundamental family of covers of Y . Finally, suppose that E 0 , E 1 , . . . is a sequence of covers of Y such that each E i+1 star-refines E i and each (b). Suppose that Y is metrizable, and let E 1 , E 2 , . . . be a basis of its uniformity. Then by (a) each E k is refined by D ϕ k for some increasing ϕ k : N → N, and each D ϕ is refined by some E n .
Conversely, suppose that there exist ϕ i as specified. Then by (a), every uniform cover of Y is refined by some D ϕ , and by the hypothesis, D ϕ is refined by D ϕn for some n. Also for each m there exists an n such that D ϕmσ is refined by D ϕn , where σ(k) = k + 1 so D ϕm is star-refined by D ϕmσ . By a renumbering we may assume that n = m + 1 without loss of generality. Then D ϕ 1 , D ϕ 2 , . . . is a basis of the pre-uniformity of Y .
3.A. Quotient maps of finite type
We say that a map q : X → Y between two uniform spaces is a quotient map of finite type if it is uniformly continuous and surjective, and there exists an n such that for every uniform cover C of X, the cover f n (C) = f (st(f −1 (f n−1 (C)), C) is uniform, where f 0 (C) is the cover of Y by singletons. Specifically we will say that q is of type n. Most important are type 1 quotient maps, which were introduced in [90].
Lemma 3.2. Let X be a metrizable uniform space, Q a set, q : X → Q a surjection, and u q the quotient pre-uniformity on Q. If d is a metric on X, let d n (x, y) = inf x=x 0 ,...,xn=y d(q −1 (x i ), q −1 (x i+1 )) and d ∞ (x, y) = inf n∈N d n (x, y). Clearly, d ∞ is a pseudo-metric on Q.
(a) [61] If u q is metrizable, then it is induced by d ∞ for some metric d on X.
(a ′ ) If u q is metrizable, and (Q, d ∞ ) id − → (Q, d n ) is uniformly continuous for some n (regardless of whether d n is a pseudo-metric), then u q is induced by d ∞ .
(b) If q : X → (Q, u) is a quotient map of finite type for some uniformity u on Q, then it is a quotient map (i.e. u = u q ), and u is pseudo-metrizable.
(c) If d 2n is a metric, then q : is uniformly continuous (regardless of whether d 2n is a pseudo-metric), then q : is uniformly continuous for some n (regardless of whether d n is a pseudo-metric), then u q is induced by d ∞ .
(e) The composition of a type m quotient map and a type n quotient map is a type mn quotient map.
To this lemma one should add that if X is a compactum and u q is a uniformity, then q : X → (Q, u q ) is a type 1 quotient map [90].
The proof of (a) is easy, and we repeat it as it is used in the proof of (a ′ ). Our proof of (d) is much simpler than the original one.
Part (e) is surprising, because the composition of two quotient maps between uniform spaces for which d m and (d m ) n are metrics is obviously a quotient map for which d mn is a metric, and easy examples show that this estimate is sharp. Thus a composition of k of quotient maps for which d 2 is a metric is still a type 1 quotient map, but with only d 2 k being a metric! Proof. (a). Suppose that u q is induced by a metric d Q . Then d ′ (x, y) := d Q (q(x), q(y)) is a uniformly continuous pseudo-metric on X. The pseudo-metric d ′ ∞ (x, y) on Q is defined as above in the case of a metric. Since d Q satisfies the triangle axiom, d ′ ∞ = d Q . If d is a metric on X, then d ′′ := d ′ + d is a metric, which is uniformly equivalent to d, and d ′′ ∞ ≥ d ′ ∞ = d Q . Since d Q is a metric (rather than just a pseudo-metric), so is d ′′ ∞ . Finally, d ′′ ∞ is uniformly continuous on (Q, d Q ) = (Q, u q ), and therefore is uniformly equivalent to d Q .
(a ′ ). In the notation of the proof of (a), we note that d ′′ ∞ ≥ d ∞ . Then to prove that d ∞ is uniformly equivalent to d ′′ ∞ (and hence induces u q ), it suffices to show that d ′′ ∞ is uniformly continuous on (Q, d ∞ ). Now D(x, y) := d ′′ ∞ (q(x), q(y)) is a uniformly continuous pseudo-metric on X. So for each ε > 0 there exists a δ > 0 such that d(x, y) < δ implies D(x, y) < 1 n ε. By the hypothesis there exists a γ > 0 such that d ∞ (v, w) < γ implies d n (v, w) < 1 2 δ. Then there exist x 0 , y 1 , x 1 , . . . , y n ∈ X such that q(x 0 ) = v, q(x i ) = q(y i ) for i = 1, . . . , n, q(y n ) = w and d(x 0 , y 1 ) + · · · + d(x n−1 , y n ) < δ. In particular, each d( (b). To prove that q is a quotient map, we need to show that u includes every sequence of covers D i of (Q, u) such that f −1 (D i ) is uniform and D i+1 strongly star-refines D i for each i.
, and so f n (C) refines D 1 .
To prove that u is pseudo-metrizable, we use Theorem 2.9. If E 1 , E 2 , . . . is a fundamental sequence of covers for X, and D 1 is a uniform cover of (Q, u), let D i+1 strongly star-refine D i , and let C be an E i that refines f −1 (D n ). Then f n (C) refines D 1 , so u has a countable basis.
(c). Suppose that d 2n is a metric. From its definition, q : X → (Q, d 2n ) is uniformly continuous. Let C ε be the cover of X by closed balls of radius ε. It suffices to show that f n (C ε ) is uniform for every ε > 0. If x ∈ X, the element f n (x) of f n (C ε ) consists of all y ∈ Q such that there exists a chain f (x) = x 0 , . . . , x n = y satisfying d(q −1 (x i ), q −1 (x i+1 )) < ε for each i. Then every pair of points y, y ′ ∈ Q at d 2n -distance < ε is contained in some element of f n (C ε ).
(c ′ ). q : X → (Q, d 2n ) is uniformly continuous by the definition of d ∞ . By the hypothesis, for each ε > 0 there exists a δ > 0 such that every pair of points y, y ′ ∈ Q at d ∞ -distance < δ satisfies d 2n (y, y ′ ) < ε. Now the proof of (c) applies. [61]). Let X be a metrizable uniform space and S α a uniformly discrete family of closed subsets of X. Then the quotient Q of X by the equivalence relation whose only non-singleton equivalence classes are the S α is metrizable.
Moreover, if d is a metric on X then d 2 is a metric on Q.
Proof. It is easy to see that d ∞ = d 2 (cf. [61]; see also [90]) and that d 2 (x, y) = 0 whenever x = y. . If X is a metrizable uniform space and A ⊂ X is a closed subset, then every bounded metric on A extends to a bounded metric on X.
The proof given below depends on 3.3, in contrast to Nhu's paper, which the author became aware of long after writing up the proof below. The similar assertion for pseudometrics is well-known [48; Lemma 1.4] (a direct proof), [50; III.16]. An extension theorem for a (possibly unbounded) metric on a closed subset of a metrizable topological space was proved by F. Hausdorff (1930).
We note that the use of functional spaces in the following proof may be replaced by a reference to [48; Lemma 1.4], whose proof is elementary.
Proof. We may assume that the diameter of A does not exceed 1. Then there exists an isometric embedding of A into U(dA, I), where dA is A endowed with discrete uniformity (see Remark 2.44(i)), which in turn extends to a uniformly continuous map f : X → U(dA, I) (see Theorem 2.32). 2 Pick some metric on X/A, which exists by Corollary 7.4, and consider the l ∞ product metric (or the l 1 product metric) on U(dA, I)×(X/A). Then g = f × q : X → U(dA, I) × (X/A) is injective, uniformly continuous, and isometric on A. It suffices to prove that its inverse is uniformly continuous. Suppose that x i , y i ∈ X are such that d(g(x i ), g(y i )) → 0 as i → ∞ but d(x i , y i ) is bounded below by some ε > 0. By passing to a subsequence and interchanging x i with y i , we may assume that either In the second case, d(q(x i ), q(y i )) is bounded below by δ/2 for sufficiently large i, hence d(g(x i ), g(y i )) is bounded below. In the third case, let Z be the complement in X to the open δ-neighborhood of A. Then q| Z is a uniform embedding, so d(q(x i ), q(y i )) is bounded below since d(x i , y i ) is. In the first case, there exist ) is bounded below. By the triangle axiom, d(f (x i ), f (y i )) is bounded below. Thus d(g(x i ), g(y i )) is bounded below.
3.5. Amalgamated union. Let f : A → X and g : A → Y be embeddings between uniform spaces with closed images. The pushout of the diagram X f ← − A g − → Y is called the amalgamated union of X and Y along the copies of A, and is denoted X ∪ A Y ; a more detailed notation is X ∪ h Y , where h = gf −1 is the uniform homeomorphism between f (A) and g(A). Thus X ∪ A Y is the quotient of X ⊔ Y by the equivalence relation x ∼ y if {x, y} = {f (a), g(a)} for some a ∈ A. It is not hard to see that X and Y (and consequently also A) can be identified with subspaces of X ∪ A Y . Corollary 3.6. Let X and Y be metrizable uniform spaces and h is a uniform homeomorphism between closed subsets A ⊂ Y and B ⊂ Z, then the amalgamated union X ∪ h Y is metrizable.
Moreover, there exists a metric on X such that h is an isometry; and if d is any such metric, then d 2 is a metric on X ∪ h Y .
On the other hand, in Corollary 4.41 below we give an alternative proof of the metrizability of the amalgamated union, which is more direct (based on Theorem 2.9) but does not produce any nice explicit metric. It is, however, the explicit d 2 metric that will be crucial for applications of Corollary 3.6.
Proof. The existence of a metric such that h is an isometry follows from Lemma 3.4. If h is an isometry, then it is easy to see that d ∞ = d 2 and that d 2 (x, y) = 0 whenever x = y (cf. [18; I.5.24]).
Adjunction space.
Let X and Y be uniform spaces, A ⊂ X and f : A → X a uniformly continuous map. The adjunction space X ∪ f Y is the pushout (in the category We are now ready to prove the main result of this subsection. Theorem 3.8. Let X and Y be metrizable (complete) uniform spaces and A a closed subset of X. If f : A → Y is a uniformly continuous map, then X ∪ f Y is metrizable (and complete). Moreover, there exists a metric d on X ⊔ Y such that f is 1-Lipschitz; for any such metric, d 3 is a metric on X ∪ f Y .
We note that this includes 3.3 and 3.6 as special cases, apart from the explicit metrics. It might be possible to prove the metrizability of the adjunction space, without producing any nice explicit metric, by a more direct method, akin to the proof of Corollary 4.41 below. In fact, G. L. Garg gave a (rather complicated) construction of a metrizable uniform space with properties resembling those of the adjunction space [35]; in his Zentralblatt review of Garg's paper, J. R. Isbell claims, without giving any justification, that "the author constructs the pushout of a closed embedding and a surjective morphism in the category of metrizable uniform spaces". We note that due to the nature of Garg's construction, a proof of Isbell's claim would be unlikely to produce any nice explicit metric on the adjunction space. It is, however, the explicit d 3 metric that will be crucial for applications of Theorem 3.8.
Proof. Let d X and d Y be some bounded metrics on X and Y . Let D(a, b) f (b)) for all a, b ∈ A. Then D is a bounded metric on A; clearly it is uniformly equivalent to the restriction of d X over A; and d Y (f (a), f (b)) ≤ D(a, b) for all a, b ∈ A.
(Note that D is nothing but the restriction of the l 1 product metric to the graph of f .) By Lemma 3.4 D extends to a metric, also denoted D, on the uniform space X. Define a metric d on X ⊔ Y by d(x, x ′ ) = D(x, x ′ ) if p, q ∈ X, by d(y, y ′ ) = d Y (y, y ′ ) if y, y ′ ∈ Y and by d(x, y) = 1 whenever x ∈ X and y ∈ Y . Then d(f (a), f (b)) ≤ d(a, b) for a, b ∈ A. Since A is closed, d(x, A) > 0 for every x ∈ X \ A. We have for all y, z ∈ f (X), and it follows that for all y, z ∈ Y . Therefore Y is identified with a subspace of X ∪ f Y , and it is easy to see that for all x ∈ X \ A and y ∈ Y and for any pair of distinct x, x ′ ∈ X \ A. Thus d 3 is a metric (see the first assertion of Lemma 3.2(d)). Hence by the second assertion of Lemma 3.2(d), X ∪ f Y is metrizable. Now suppose that X and Y are complete. Let q 1 , q 2 , . . . be a Cauchy sequence in Q. Suppose that d 3 (q n , Y ) is bounded below. Then each q n = {x n }, where x n ∈ X \ A and d(x n , A) is bounded below. Then x n is a Cauchy sequence in X and so converges to an x ∈ X \ A. Then q n converges to {x} in Q. If d 3 (q n , Y ) is not bounded below, then by passing to a subsequence we may assume that d 3 (q n , Y ) → 0 as n → ∞. Then there exist y 1 , y 2 , · · · ∈ Y such that d 3 (q n , [y n ]) → 0 as n → ∞. Then [y n ] is a Cauchy sequence in Q, whence y n is a Cauchy sequence in Y and so has a limit y ∈ Y . Then [y n ] converges to [y] in Q. Since q 1 , [y 1 ], q 2 , [y 2 ], . . . is a Cauchy sequence and it has a subsequence converging to [y], it converges to [y] itself.
3.B. Cone, join and mapping cylinder
3.9. Uniform join. Let X and Y be uniform spaces. We define their join to be the quotient X × Y × [−1, 1]/ ∼, where (x, y, t) ∼ (x ′ , y ′ , t ′ ) if either t = 1 and x = x ′ or t = −1 and y = y ′ . Thus we have the pushout diagram where π(x, y, −1) = x and π(x, y, 1) = y. By Lemma 2.29(b) and Lemma 2.28(b), π is a quotient map. So if X and Y are metrizable or complete, then by Theorem 3.8 so is X * Y . More specifically, given metrics, denoted d, on X and Y , then a metric on X * Y is given by (In the case t, t ′ ∈ (−1, 1), the four options correspond respectively to the chains other cases involve subchains of these chains.) The formula for d 3 implies that if X and Y are metrizable uniform spaces, and A ⊂ X and B ⊂ Y are their subspaces, then A * B is a subspace of X * Y . In particular, X * ∅ and ∅ * Y are subspaces of X * Y . Since they are uniformly disjoint, their union may be identified with X ⊔ Y .
3.10. Uniform cone. The uniform join X * pt is denoted CX and called the cone over the uniform space X. Thus CX may be viewed as the quotient space X × [0, 1]/X × {1}. Given a metric d on X, by Lemma 3.3, a metric on CX is given by If the diameter of X is ≤ 2, then (X, d) is isometric to the subset X × {0} of (CX, d 2 ).
Lemma 3.11. Let X and Y be metrizable uniform spaces. Then X * Y is uniformly homeomorphic to CX × Y ∪ X×Y X × CY .
Proof. Let us fix some metrics on X and Y such that X and Y have diameters ≤ 2.
Viewing CX as X ×[0, 1]/X ×{1} and CY as Y ×[−1, 0]/Y ×{−1}, a metric on CX ×Y is given by and a metric on X × CY is given by Since the diameters of X and Y are ≤ 2, the metrics on CX × Y and X × CY both induce the same metric d(x, x ′ ) + d(y, y ′ ) on X × Y . Then it is easy to see that the metric on CX × Y ∪ X×Y X × CY given by 3.6 coincides (exactly, not just up to uniform equivalence) with the above metric d 3 of X * Y . , which is bounded below by 1 2 td(x, y) and above by td(x, y). This is somewhat "smoother" than with our standard cone metric: Lemma 3.13. If X is a metrizable uniform space, the Euclidean cone over X is uniformly homeomorphic to CX keeping X fixed.
Let x i , y i ∈ CX be two sequences. By passing to subsequences and interchanging x i with y i , we may assume that either Indeed, both hold in the case (i); none holds in the case (ii); and to do with the case (iii) in suffices to show that the restrictions of d 2 and d e over X × [0, 1 − ε] are uniformly equivalent for each ε > 0.
It is clear that the restriction of d 2 over X × [0, 1 − ε] is uniformly equivalent to the l 1 product metric for each ε > 0. The restriction of d e over X × [0, 1 − ε] is uniformly equivalent to the l 2 product metric for each ε > 0, using that where ts ∈ [ε 2 , 1], and z 2 ≤ sin z ≤ z as long as z ∈ [0, π 2 ].
The proof is based on Lemma 3.13.
Proof. Let us consider the l ∞ norm ||(v, t)|| = max{||v||, t} on V × R. By scaling the norm of V we may assume that X lies in the unit ball Q of V . Without loss of generality X = Q. Let f : Q×I → cQ send Q×{0} onto the cone vertex {(0, 1)}, and each cylinder generator {x} × I linearly onto the cone generator c{x}.
We consider two cases. When |s − t| ≤ 1 2 max(t, s), we get ts ≥ max(t, s) 2 − max(t, s)|s − t| ≥ 1 2 max(t, s) 2 and therefore In either case, E ≥ 1 4 max(t, s)D and E 2 ≥ (s − t) 2 . Hence 4E + E ≥ max(t, s)D + |s − t| ≥ S. Lemma 3.17. If X, Y and Z are uniform spaces, x ∈ X and y ∈ Y , then the subspace This can be seen as a purely category-theoretic fact.
Proof. The amalgamated union maps onto the subspace in the obvious way. Given maps f : X × {y} × Z → W and g : {x} × Y × Z → W agreeing on {x} × {y} × Z, one defines f ∪ g using the projections of X × Y × Z onto X × Z and onto Y × Z.
Lemma 3.18. Let V and W be normed vector spaces, and let X ⊂ V and Y ⊂ W be bounded subsets. Then the subspace cX × Y ∪ X × cY of cX × cY is uniformly homeomorphic to the amalgam cX × Y ∪ X×Y X × cY .
Proof. Lemma 3.16 implies that a uniform neighborhood of X in cX is uniformly homeomorphic to X × I. Denote this uniform neighborhood by U X . Then the uniform neighborhood U X ×U Y of X ×Y in cX ×cY is uniformly homeomorphic to X ×Y ×I ×I. Hence by Lemma 3.17, the subspace The definition of the amalgam as a pushout yields a uniformly continuous bijection f : cX ×Y ∪ X×Y X ×cY → cX ×Y ∪X ×cY such that f −1 is uniformly continuous on cX × Y and on X × cY . By the above, f −1 is also uniformly continuous on U X × Y ∪ X × U Y . Since these three subsets of cX × Y ∪ X × cY form its uniform cover, f −1 is uniformly continuous.
Theorem 3.19. Let V and W be normed vector spaces, and let X ⊂ V and Y ⊂ W be bounded subsets. Then the independent rectilinear join XY is uniformly homeomorphic to X * Y .
Proof. By Lemma 3.15, XY is uniformly homeomorphic to the subspace cX ×Y ∪X ×cY of cX × cY . By Lemma 3.18, the latter is uniformly homeomorphic to the amalgam cX × Y ∪ X×Y X × cY . By Lemma 3.16 the latter is in turn uniformly homeomorphic to the amalgam CX × Y ∪ X×Y X × CY . Finally, by Lemma 3.11, the latter is uniformly homeomorphic to X * Y . Corollary 3.20. Let V and W be normed vector spaces, X and Y metrizable uniform spaces, and f : X ֒→ V and g : Y ֒→ W embeddings onto bounded subsets. Then the uniform homeomorphism type of the independent rectilinear join f (X)g(Y ) does not depend on the choices of V , W , f and g.
Corollary 3.21. Given metrizable uniform spaces X and Y , there exists a uniform homeomorphism between C(X * Y ) and CX × CY taking X * Y onto CX × Y ∪ X × CY .
Proof. Pick bounded metrics on X and Y . Then there exists an isometric embedding of X onto a bounded subset of the normed vector space U(dX, R), where dX denotes X endowed with discrete uniformity (see Remark 2.44(i)), and similarly for Y . (
Mapping cylinder. The mapping cylinder MC(f ) is the adjunction space
Lemma 3.23. Let f : X → Y be a uniformly continuous map between metrizable uniform spaces. ( Proof. (a). Given bounded metrics d X on X and d on Y , define a new metric d on X by d(x, ). Clearly, d is bounded and uniformly equivalent to d X , and d(f (x), f (y)) ≤ d(x, y) for all x, y ∈ X. Using d we define the l 1 product metric on X × I, which is bounded; and hence the metric of disjoint union, also denoted d, on X × I ⊔ Y . Then d(f ′ (p), f ′ (q)) ≤ d(p, q) for all p, q ∈ X × {1}, and thus from Theorem 3.8 we have the following d 3 metric on MC(f ): for all y, y ′ ∈ Y , x, x ′ ∈ X and t, t ′ ∈ I. Clearly, this coincides with the similar d 3 metric on MC(f | A ).
(b). The uniform homeomorphism between X and Γ f yields a uniform homeomorphism between MC(f ) and MC(π| Γ f ), where π : X × Y → Y is the projection. On the other hand, by the proof of Lemma 3.11, CX × Y is uniformly homeomorphic to MC(π). Hence the assertion follows from (a).
Corollary 3.24. If ∆ is a finite diagram of metrizable uniform spaces and uniformly continuous maps, its homotopy colimit is a metrizable uniform space.
4.A. Absolute retracts and extensors
4.1. Absolute retracts and extensors. As long as we have finite coproducts, as well as embeddings and quotient maps in a concrete category C over the category of sets, we also have in C the notions of map extension and neighborhood.
where f is a C-morphism and a is a C-embedding is called a partial C-morphism, and is said extend to a C-morphismf : X → Y if the composition A a ֒→ Xf − → Y equals f . Given a C-embedding g : X ֒→ Y , its decomposition into a pair of C-embeddings An AE(C) is any C-object Y such that every partial C-morphism X a ←֓ A f − → Y extends to a C-morphism X → Y . If Y satisfies this for the special case of f = id Y , then Y is called an AR(C). Absolute extensors are also known as "injective objects", cf. [50; p. 39], [1; 9.1; see also 9.6] An ANE(C) is any C-object Y such that for every partial C-morphism X If Y satisfies the above for the special case of f = id Y , then Y is called an ANR(C). Note the obvious implications ; and also with A[N]Rs in the notation of Borsuk [15], which are the same as compact A[N]Rs in the notation of Hu [44] (c) A Banach space V is an ANRU if it has uniformly normal structure [13; 1.26], that is, if there exists a γ < 1 such that every convex subset of E of diameter 1 is contained in a ball of radius γ. The Banach spaces L p (µ) with 1 < p < ∞ and every measure µ, including the sequence spaces l p with 1 < p < ∞, have uniformly normal structure (see [13; A.9] and [58; II.1.f.1, "in addition" part]). In particular, they are ANRUs, though this has more direct proofs [57; Theorem 8] (correcting a mistake in [49; 3.1, proof of (c)]), [13; 1.29]. In particular, the unit balls of these spaces are ARUs by (a).
A[N]Rs and
The unit ball of L 1 (µ) for every measure µ is uniformly homeomorphic to those of L p (µ) for 1 < p < ∞ (see [13; Remark to 1.29]), hence is also an ARU. In particular, the unit ball of l 1 is an ARU; in fact, according to [50; Remarks after the proof of 3.2], l 1 itself is an ANRU. A necessary condition for a Banach space V to have unit ball uniformly homeomorphic to that of an L p (µ), 1 ≤ p < ∞, is that l n ∞ (i.e. the n-dimensional vector space with the l ∞ -norm) do not uniformly embed in V for each n [13; 9.21]. This condition is not sufficient [13; 9.23], but can be made sufficient by further hypotheses [13; 9.4, 9.7].
(d) Given a normed vector space V identified with a subspace of the countably dimensional vector space ∞ i=1 R and given normed vector spaces V 1 , V 2 , . . . , let us write The Banach space (l 1 p ⊕ l 2 p ⊕ . . . ) c 0 , where 1 ≤ p ≤ ∞ and l n p is the n-dimensional vector space with the l p -norm, is an ANRU [57; proof of Theorem 13].
On the other hand, none of the following Banach spaces, nor even their unit balls, is an ANRU: (l p 1 ⊕ l p 2 ⊕ . . . ) c 0 and (l p 1 ⊕ l p 2 ⊕ . . . ) lp , as well as (l n 1 Indeed, if V is the limit of an inverse sequence . . . Banach spaces V i and continuous linear maps, then each f ∞ i : V → V i is continuous, so its kernel K i is closed. Without loss of generality 0 is an interior point of B. Then B i := B + K i is a closed convex body in V that is absorbing (i.e. for every v ∈ V there exists an r such that v/r ∈ B i ), and then similarly to (a), B i is a uniform retract of V , and in particular of B i−1 . Since B is the inverse limit of the B i , it follows (see the proof of Theorem 5.18 below) that B is a uniform retract of B [0,∞] , and therefore (see the proof of Theorem 4.21 below) also of V . The remainder of the proof is similar to that in (a).
(h) If V is an injective object in the category of Frechét spaces, then every closed bounded convex body B in V is an ARU. Indeed, V is injective as a locally convex space (see [26; Lemma 0]), and B is a uniform retract of V by the argument in (g). Now the assertion follows from [66; Theorem 1.6].
4.B. Uniform ANRs
Isbell's proof that ANRUs are complete uses non-metrizable test spaces such as the ordinal ω 1 in the order topology in an essential way, even when the given ANRU is itself metrizable. Relevance of this argument for the purposes of geometric topology is questionable; one would not be comfortable using this construction as a basis for geometrically substantial results. An alternative, more transparent approach (implicit in the papers of Garg [35] and Nhu [66] Remark 4.7. (i) The above proof is much easier than the original one in [35]; a proof that is arguably still easier (as it does not use metrizability of the adjunction space) is presented in [91; Appendix] modelled on an argument in [66].
(ii) Proposition 4.6 holds, with the same proof, if "metrizable" is replaced by "separable metrizable" throughout, including the definitions of uniform A[N]R and A[N]E.
Homotopy completeness.
We say that a uniform space X is homotopy complete if there exists a uniform homotopy H :X × I →X, whereX is the completion of X, such that H(x, 0) = x and H(X × (0, 1]) ⊂ X. Note that if X is homotopy complete, then it is uniformly ε-homotopy equivalent to its completion, for each ε > 0. Remark 4.9. A subspace A of a separable metrizable topological space Y such that there exists a homotopy h t : Y → Y satisfying h 0 = id and h t (Y ) ∩ A = ∅ for t > 0 is called a 'Z-set' by some authors including S. Ferry [31] and a 'homotopy negligible set' by some authors including T. Banakh (see [7] A few month after having written up the proof of Theorem 4.10, the author learned that its analogue for semi-uniform ANRs (see Remark 4.30(b)) had been known [79].
Proof. For the "if" direction, we consider the case of ARUs; the case of ANRUs is similar. Conversely, suppose X is a uniform A[N]R. Let Z t denote the subspace X × {0} ∪ X × (0, t] ofX × [0, 1]. Then X × {0} is a closed subset of Z 1 . Since X is a uniform ANR, id X extends to a uniformly continuous map Z ε → X for some ε > 0. The latter has a unique extension over the completions, which yields the required homotopy.
It This follows from Lemma 4.3(a) and the proof of Theorem 4.10, which also works in the separable case. 4.13. Uniform local contractibility. We call a metric space X uniformly locally contractible if for each ε > 0 there exists a δ > 0 such that for every uniform space Y , every two δ-close uniformly continuous maps f, g : Y → X are uniformly ε-homotopic (that is, are joined by a uniformly continuous homotopy Y × I → X that is ε-close to Obviously uniform local contractibility does not depend on the choice of the metric on X in its uniform equivalence class. Remark 4.14. If X is uniformly locally contractible, then for each ε > 0 there exists a δ > 0 such that δ-balls in X contract within their concentric ε-balls, by a uniformly equicontinuous family of null-homotopies (see [49; 4.2]).
Remark 4.15. If X is uniformly locally contractible, then for each ε > 0 there exists a δ > 0 such that for each n and every uniform space Y , every (n−1)-sphere F : Y ×S n−1 → X of maps Y → X that are within δ of each other bounds an n-ballF : Y × B n → X of maps Y → X that are within ε of each other. Indeed, the definition of uniform local contractibility yields an ε-homotopy h t between F and the composition of the projection Y × ∂I n → Y and F | Y ×pt : Y → X; thenF can be defined byF (y, [(r, ϕ) Lemma 4.16. Let X be a (separable; compact) metric space. Then X is uniformly locally contractible if either (a) X is a uniform ANR, or (b) for each (separable; compact) metrizable uniform space Y and every ε > 0 there exists a δ > 0 such that every two δ-close uniformly continuous maps f, g : Y → X are uniformly ε-homotopic.
We note that the hypothesis of (b) is a weakening of the condition of uniform local contractibility, with restrictions imposed on Y and with δ allowed to depend on Y . It follows from (b) that X is uniformly locally contractible if and only if it is LCU in the sense of Isbell (whose δ is allowed to depend on Y ) [49]. That ANRUs are LCUs was known to Isbell [49;4.2]; his proof of this fact is rather different (via functional spaces).
Proof. Let U i be the 1 i -neighborhood of the diagonal in X × X in the l ∞ product metric on X × X, and let f i , g i : U i ⊂ X × X → X be the two projections. We shall show that, under the hypothesis of either (a) or (b), for each ε > 0 there exists an i such that f i and g i are uniformly ε-homotopic.
Since X is a uniform ANR, f has a uniformly continuous extensionf over a uniform neighborhood U of A 1 in Y 1 . This U contains the 1 n -neighborhood of A for some n, in the l ∞ product metric on Y ⊂ X × X × N × [0, 1], which in turn contains Y n . Moreover, for each ε > 0 there exists an m ≥ n such thatf takes 1 m -close points into ε-close points. Thenf restricted to U m × [0, 1 m ] is a uniform ε-homotopy between f m and g m .
(b) Consider F = i∈N f i and G n := i∈[n] f i ⊔ i∈N\[n] g i both mapping Y := U i into X. Note that d(F, G n ) → 0 as n → ∞. Then by the hypothesis of (b), for each ε > 0 there exists an n such that F and G n are uniformly ε-homotopic. Then also f n+1 and g n+1 are uniformly ε-homotopic.
We have thus shown, in both (a) and (b), that for each ε > 0 there exists a uniform ε-homotopy h t : U i → X between f i and g i for some i. Now given 1 i -close maps f, g : Y → X, the image of f × g : Y → X × X lies in U i . Then the composite homotopy Y f ×g − − → U n ht − → X is a uniform ε-homotopy between f and g.
Remark 4.17. Let us call a metric space X uniformly locally equiconnected if for each ε > 0 there exists a δ > 0 such that for every uniform space Y , and a subset A of Y , every two δ-close uniformly continuous maps f, g : Y → X that agree on A are uniformly ε-homotopic keeping A fixed. Then Lemma 4.16 has a relative version: (a) uniform ANRs are uniformly locally equiconnected; (b) in the definition of uniform local equiconnectedness, Y may be assumed to be metrizable and A to be closed; and if X is separable (compact), Y may be additionally assumed to be separable (compact). The proof is similar, and shows that, under the hypothesis of either (a) or (b), the uniform ε-homotopy between f i and g i can be chosen so as to keep the diagonal ∆ X ⊂ U i fixed.
Uniformly locally equiconnected compacta coincide with locally equiconnected compacta [28; Theorem 2.5]. Whether locally equiconnected compacta are ANRs is a long standing open problem going back to Fox [32]. Its special case is the so-called "compact AR problem" of whether a compact convex subset of a metrizable topological vector space is an AR.
Hahn property. Following Isbell [49]
, we say that a metric space satisfies the Hahn property if for each ε > 0 there exists a δ > 0 such that for every uniform space Y and every δ-continuous map Ξ : Y → X (that is, a possibly discontinuous map such that there exists a uniform cover C of Y such that Ξ sends every element of C into a set of diameter at most δ), there exists a uniformly continuous map f : Y → X that is ε-close to Ξ. Obviously the satisfaction of the Hahn property does not depend on the choice of the metric on X in its uniform equivalence class.
Lemma 4.19. Let X be a (separable; compact) metric space. Then X satisfies the Hahn property if either (a) X is a uniform ANR, or (b) for each (separable; compact) metrizable uniform space Y and every ε > 0 there exists a δ > 0 such that every δ-continuous map Y → X is ε-close to a uniformly continuous map Y → X.
We note that the hypothesis of (b) is a weakening of the Hahn condition, with restrictions imposed on Y and with δ allowed to depend on Y . That ANRUs satisfy the Hahn property was shown by Isbell [49;4.2]; this along with Theorem 4.10 implies (a).
Proof of (b). Let F be a (separable; compact) metric ARU containing X (see Theorem 2.43 and Remark 2.44). Let U i be the 1 i -neighborhood of X in F , and let U = U i .
Pick a 1 3i -continuous retraction ξ i : U i → X (cf. the proof of Theorem 4.21), and define Ξ n : U → X via ξ i on each U i for i ≥ n, and by a constant map on U 1 ⊔ · · · ⊔ U n−1 . Thus Ξ n is 1 3n -continuous. By the hypothesis, for each ε > 0 there exists an n ≥ 1 ε such that Ξ n is ε 2 -close to a uniformly continuous map U → X. Then ξ n is ε 2 -close to a uniformly continuous map r n : U n → X.
By (a), F satisfies the Hahn property. So for each n there exists a δ > 0 such that every δ-continuous map Ξ : Y → X ⊂ F is 1 n -close to a uniformly continuous map Theorem 4.20. The following are equivalent for a compactum X: (i) X satisfies the Hahn property; (ii) for each ε > 0 there exists a compact polyhedron P and continuous maps X f − → P g − → X whose composition is ε-close to id X ; (iii) for each ε > 0 there exists a compactum Y satisfying the Hahn property and This means that compacta satisfying the Hahn property coincide with the "approximate ANRs" of Noguchi and Clapp [23], which are also known as "NE-sets" after Borsuk [16] and as compacta that are "approximate polyhedra" in the sense of Mardešić [59].
We shall generalize Corollary 4.20 to all separable metrizable uniform spaces in the sequel to this paper which deals with uniform polyhedra.
Proof. (i) implies (ii) using that X is the inverse limit of an inverse sequence of compact polyhedra (see also Lemma A.12(a) below). Since compact polyhedra are ANRs, hence uniform ANRs, and in particular satisfy the Hahn property, (ii) implies (iii). The implication (iii)⇒(i) is an easy exercise. A few months after having written up the proof of Theorem 4.21, the author discovered that a rather similar characterization of non-uniform ANRs is stated without proof in [2; Theorem 3.7] (some relevant ideas can be found also in Theorems 3.2 and 4.8 in [2]).
Proof. If X is a uniform ANR, it is uniformly locally contractible by Lemma 4.16(a) and satisfies the Hahn property by Lemma 4.19(a).
Conversely, suppose Y is a metric space, A is a closed subset of Y , and f : A → X is a uniformly continuous map. Fix some metric on X. Let γ i = δ LCU (2 −n ) be given by the uniform local contractibility of X with ε = 2 −n ; let β i = δ Hahn (γ i ) be given by the Hahn property of X with ε = γ i ; and let α i = δ f (β i ) be such that f sends 3α i -close points into β i -close points. We may assume that α i+1 ≤ α i , γ i+1 ≤ γ i and β i ≤ γ i for each i. Let U i be the α i -neighborhood of A in Y ; thus A = U i .
(When Y is separable, it is easy to find such an x ′ i using only the countable axiom of choice.) Then Ξ i sends α i -close points x, y into β i -close points f (x ′ i ), f (y ′ i ) (using that x ′ i is 3α i -close to y ′ i ). By the Hahn property, there exists a uniformly continuous map f i : Then by the uniform local contractibility, f i | U i+1 is uniformly 2 −i -homotopic to f i+1 . These homotopies combine into a uniformly continuous map H on the extended mapping telescope U [1,∞] Pick a uniformly continuous function ϕ : [1,∞] and restricts to the identity on A. Hence the composition (b) Suppose that a metric space X is uniformly embedded in a uniform ANR Z. Then X is a uniform ANR if and only if for each ε > 0 there exists a uniformly continuous map ϕ of a uniform neighborhood of X in Z into X such that ϕ| X is uniformly ε-homotopic to id X with values in X.
To prove the "if" assertion, fix an ε > 0 and let δ be such that that the uniform neighborhood U provided by the hypothesis contains the δ-neighborhood of X in Z, and the map ϕ provided by the hypothesis is (δ, ε)-continuous.
By Lemma 4.16(a), there exists a γ > 0 such that for every uniform space Y , every two γ-close uniformly continuous maps f, g : Y → X are uniformly δ-homotopic with values in Z. Then the homotopy actually has values in U, and therefore ϕf and ϕg are uniformly ε-homotopic with values in X. On the other hand, by the hypothesis they are uniformly ε-homotopic to f and g, respectively, with values in X. Thus X is uniformly locally contractible.
By Lemma 4.19(a), there exists a β > 0 such that for every uniform space Y , every β-continuous map Ξ : Y → X is δ-close to a uniformly continuous map h : Y → Z. Then h actually has values in U, and then ϕh : Y → X is ε-close to ϕΞ. By the hypothesis, the latter is in turn ε-close to Ξ. Thus X satisfies the Hahn property.
So we infer from Theorem 4.21 that X is a uniform ANR.
(a). Let X be the given space. We may assume that it is embedded in a uniform ANR (for instance, in some ARU) Z. Given an ε > 0, the hypothesis provides a uniform ANR Y that uniformly ε-dominates X. Thus we are given uniformly continuous maps u : X → Y and d : Y → X such that the composition X u − → Y d − → X is uniformly εhomotopic to the identity. Let δ be such that d is (δ, ε)-continuous. Since Y is a uniform ANR, and X × {0} is closed in X × {0} ∪ Z × (0, 1], u is uniformly δ-homotopic to the restriction of a uniformly continuous mapū : U → Y for some uniform neighborhood U of X in Z. Then du is uniformly ε-homotopic to dū| X . On the other hand, du is uniformly ε-homotopic to the identity. So we infer from (b) that X is a uniform ANR.
Another consequence of Theorem 4.21 is the following Proof. Let X be a uniformly locally contractible n-dimensional space. Then its comple-tionX is still n-dimensional (see Lemma 2.39). By [50; Theorem V.34],X is the limit of an inverse sequence of n-dimensional uniform polyhedra P i . Since X is uniformly locally contractible, every collection of small spheroids S k → X bounds a uniformly equicontinuous family of small singular disks D k+1 → X. Using this, it is easy to construct maps r i : P i → X by induction on skeleta, so that for each ε > 0 there is an i such that the composition X ⊂X [48; 1.9], each P i is a uniform ANR and so (see Lemma 4.19(a)) satisfies the Hahn property. Since each p ∞ i :X → P i is uniformly continuous, we infer that X also satisfies the Hahn property. By Theorem 4.21, X is a uniform ANR.
Remark 4.24. Borsuk's example of a locally contractible compactum that is not an ANR (see [15]) shows that uniform local contractibility of a residually finite dimensional metrizable uniform space X does not follow from uniform local connectedness of each of the functional spaces F n = U(S n × N, X). By the proof of Theorem 4.23, it does follow when X is finite-dimensional. This is parallel to [48; 1.11], apart from using the metrizability of the cone.
Proof. If Y is a uniform AR, by definition it is a uniform ANR. The cone CY is metrizable, so id Y extends to a uniformly continuous map CY → Y ; the composition Y × I → CY → Y is then a uniform null-homotopy of id Y .
Conversely, let A be a closed subset of a metrizable uniform space X. If Y is uniformly contractible, every uniformly continuous map f : A → Y is homotopic to a constant map, which extends over X. If additionally Y is a uniform ANR, then by Lemma 4.26 the homotopy extends over X × I; in particular, f extends over X. Lemma 4.28. Let X be a metrizable uniform space and A a closed subset of X.
(a) Suppose that A is a uniform ANR. Then A is a uniform deformation retract of X if and only if the inclusion A ֒→ X is a uniform homotopy equivalence.
(b) Suppose that X is a uniform ANR. Then A is a uniform deformation retract of X if and only if A is a uniform strong deformation retract of X. (b) Michael calls a map f : X → Y of metric spaces uniformly continuous at a closed subset A ⊂ X if for each ε > 0 there exists a δ > 0 such f is (δ, ε)-continuous on the δ-neighborhood of A [63]. Michael [63] and Torunczyk [86] [86]. Sakai established the analogue of Theorem 4.10 for semi-uniform ANRs [79].
(c) Lipschitz, and 1-Lipschitz ANRs and ARs have been studied. According to [52; p. 65], 1-dimensional topologically complete ARs are metrizable as 1-Lipschitz ARs (see [72] for a proof in the compact case). On the other hand, Isbell showed that 2dimensional non-collapsible compact polyhedra are not metrizable as 1-Lipschitz ARs [52]. It appears to be unknown whether every ANR is homeomorphic to a Lipschitz ANR (cf. [43]). However, Hohti showed that every LC n compactum can be remetrized so as to be Lipschitz n-LC for all n > 0 [43].
Remark 4.31. We mention some facts relating to semi-uniform ANRs.
(a) Similarly to the proof of Theorem 4.21, a metrizable uniform space is a semiuniform ANR if and only if it is semi-uniformly locally contractible and satisfies the weak Hahn property. We call a metric space M semi-uniformly locally contractible if for each ε > 0 there exists a δ > 0 such that every two δ-close continuous maps of a metrizable topological space into M are ε-homotopic. (That semi-uniform ANRs satisfy this property but not conversely was observed by Michael [63].) We say that M satisfies the weak Hahn property if for each ε > 0 there exists a δ > 0 such that for every γ > 0, every (γ, δ)-continuous map f of a metric space N into M is ε-close to a continuous map. (Note that the metric on N is irrelevant, and the hypothesis on f is equivalent to saying that for every convergent sequence (x n ) in N, the set of limit points of the sequence (f (x n )) has diameter < δ.) (b) Similarly to the proof of Theorem 4.23, the following are equivalent for a finite dimensional metrizable uniform space X: (i) X is a semi-uniform ANR; (ii) X is semiuniformly locally contractible; (iii) for each ε > 0 there exists a δ > 0 such that every continuous map S n → X with image of diameter ≤ δ bounds a continuous map B n+1 → X with image of diameter ≤ ε. Lemma 4.33. Assume the hypothesis of Theorem 4.32 and fix a metric on X ∪ f Y . Then for each ε > 0 there exists a uniform ε-homotopy H t : X ∪ f Y → X ∪ f Y keeping Y fixed and such that H 0 is the identity, and H 1 retracts a uniform neighborhood of
4.C. Homotopy limits and colimits of uniform ANRs
Moreover, H t lifts to a uniform homotopy h t : X → X keeping A fixed and such that h 0 is the identity and h 1 retracts a uniform neighborhood of A in X onto A.
Proof. Since A is a uniform ANR, there is a uniform retraction r of a uniform neighborhood U of A in X onto A. The projection X × {0} ∪ A × I → X combines with r into a uniformly continuous map R : X × {0} ∪ A × I ∪ U × {1} → X. Since X is a uniform ANR, the latter extends to a uniformly continuous mapR on a uniform neighborhood W of X × {0} ∪ A × I ∪ U × {1} in X × I. This W contains the region Φ below the graph of a function ϕ : (X, V ) → ([δ, 1], {1}), where V is a uniform neighborhood of A in U and δ > 0. Then h t : X → X, defined by h t (x) =R(x, tϕ(x)), is a uniform homotopy, keeping A fixed, between id X and an extensionr of r over X. Since Φ may be assumed to be contained in any uniform neighborhood of X × {0} ∪ A × I given in advance, h t may be assumed to be an ε-homotopy.
Consider the self-homotopy H t of X ∪ f Y defined by H t (y) = y for each y ∈ Y and all t ∈ I and by H t ([x]) = [h t (x)] for x ∈ X and all t ∈ I. It is well-defined since h t fixes A. To prove that H t is uniform, we consider the d 3 metric as in Theorem 3.8. Given Proof of Theorem 4.32. We only consider the case of ANRs; the case of ARs is similar, and alternatively it can be deduced from the case of ANRs using Theorem 4. 27.
Suppose X ∪ f Y is a closed subset of a metrizable uniform space Z. We are going to apply Theorem 4.22(b); to this end, fix an ε > 0, and feed it into the preceding lemma. Let U Y be the uniform neighborhood of A in X provided by the preceding lemma, and let h t and H t be the homotopies provided by the preceding lemma. Definef : Since A is a uniform ANR, h 1f −1 | V A extends to a uniformly continuous map ϕ A : Since Z X and Z Y are uniformly disjoint, so are W X and W Y , and therefore the map Modulo Corollary 3.6, this also follows from the results of Nhu [66], [67], who used the d 2 metric on the underlying set of X ∪ A Y but did not identify it as a metric of the quotient uniformity. Proof. We consider the case of ANRUs; the case of ARUs follows using Theorem 4.27.
By Theorem 4.10, the completionX of X is an ANRU. Hence by Theorem 2.36, U(Y,X) is an ANRU. Again by Theorem 4.10, there exists a uniform homotopy h t ofX such that h 0 = id and h t (X) ⊂ X for t > 0. Now H t : f → h t f is a uniform homotopy of U(Y,X) such that H 0 = id and H t (U(Y,X)) ⊂ U(Y, X) for t > 0. Hence by Theorem 4.10, U(Y, X) is a uniform ANR.
Since A is a uniform ANR, by Theorem 4.22(b) for each ε > 0 there exists uniformly continuous map g : Y, B)) is a uniform ANR. It follows, for instance, that the subspace of self-homotopy equivalences in U((X, A), (X, A)) is a uniform ANR.
4.D. Relative functional spaces via non-metrizable amalgam
This subsection is devoted to an alternative proof of Theorem 4.35 (with a similar result for non-metrizable ARUs and ANRUs as a byproduct), avoiding the infinite construction in the proof of Theorem 4.21 but involving a study of amalgamated unions of nonmetrizable uniform spaces.
Lemma 4.40. A cover of X ∪ A Y is uniform if and only if it is refined by a cover of the form C + D := {st(z, C ∪ D) | z ∈ X ∪ A Y }, where C is a uniform cover of X and D a uniform cover of Y .
Proof. Let E be a uniform cover of X ∪ A Y . Then E is star-refined by a cover E * such that C := E * ∩ X is a uniform cover of X and D := E * ∩ Y is a uniform cover of Y . Then C ∪ D refines E * , hence C + D refines E.
Conversely, let C be a uniform cover of X and D a uniform cover of Y . We need to construct a sequence of covers E 0 , E 1 , . . . of X ∪ Y such that E 0 = C + D, each E i+1 star-refines E i , and each E i ∩ X is a uniform cover of X and each E i ∩ Y is a uniform cover of Y . First note that (C + D) ∩ X itself is uniform, for it is refined by {st(x, C) | x ∈ X}, which in turn is refined by C. Similarly (C + D) ∩ Y is uniform.
Let C * be a uniform cover of X star-refining C and let D * be a uniform cover of Y star-refining D. Then C A := C * ∩ A and D A := D * ∩ A are uniform covers of A, hence so is F * := C * ∧D * = C A ∧D A . Since C A is a uniform cover of A, it is of the form C Y ∩A for some uniform cover C Y of Y . Similarly D A is of the form D X ∩ A for some uniform cover D X of X. Then C ′ * := C * ∧D X is a uniform cover of X star-refining C and D ′ * := D * ∧C Y is a uniform cover of Y star-refining D. In addition, C ′ * ∩ A = F * = D ′ * ∩ A. Next, let C * * be a uniform cover of X star-refining C ′ * and let D * * be a uniform cover of Y star-refining D ′ * . Let F * * = C * * ∧ D * * , and define C ′ * * and D ′ * * similarly to the above. Then C ′ * * is a uniform cover of X star-refining C ′ * and D ′ * * is a uniform cover of Y star-refining D ′ * ; in addition, C ′ * * ∩ A = F * * = D ′ * * ∩ A. We claim that C ′ * * + D ′ * * star-refines C + D; iterating the construction of C ′ * * and D ′ * * would then yield the required sequence E 1 , E 2 , . . . (with E 1 = C ′ * * + D ′ * * ). Given a z ∈ X ∪ A Y , we will show that st(z, C ′ * * + D ′ * * ) lies in st(z ′ , C) ∪ st(z ′ , D) = st(z ′ , C ∪ D) for some z ′ ∈ X ∪ A Y . By symmetry we may assume that z ∈ X. Let U be an element of C ′ * * + D ′ * * containing z. Then U = st(w, C ′ * * ) ∪ st(w, D ′ * * ) for some w ∈ X ∪ A Y . We consider two cases.
I. First suppose that z / ∈ st(A, C ′ * * ). Then w ∈ X \ A, hence U = st(w, C ′ * * ). Then U is contained in an element of C ′ * , which is in turn contained in an element of C. Thus U ⊂ st(z, C) and we may set z ′ = z.
1. If z ∈ st(w, D ′ * * ) then z ∈ Y , whence z ∈ A. Since z and z ′ are contained in one element of C ′ * * and also in A, they are contained in one element of F * * , hence in one element of D ′ * * . Thus z ∈ st(z ′ , D ′ * * ) and w ∈ st(z, D ′ * * ), whence w ∈ st(z ′ , D ′ * ). In particular, w ∈ Y . We consider two cases. a) If w / ∈ A, then U = st(w, D ′ * * ). Then U ⊂ st(z ′ , D) and we are done. b) If w ∈ A, then w ∈ st(z ′ , F * ), and therefore w ∈ st(z ′ , C * ). Then U = st(w, C ′ * * ) ∪ st(w, D ′ * * ) is contained in st(z ′ , C) ∪ st(z ′ , D) and we are done.
We note that Lemma 4.40 yields an alternative proof of Corollary 4.41, apart from the explicit metric: Corollary 4.41. If X and Y are metrizable uniform spaces, every amalgamated union X ∪ A Y is metrizable.
Proof. Let C 1 , C 2 , . . . be a basis of the uniformity of X and D 1 , D 2 , . . . be a basis of the uniformity of Y . If C is a uniform cover of X and D is a uniform cover of Y , then there exists an i such that C i refines C and D i refines D. Then C i + D i refines C + D. Hence every uniform cover of X ∪ A Y is refined by one of C 1 + D 1 , C 2 + D 2 , . . . . By Theorem 2.9, X ∪ A Y is metrizable.
Corollary 4.42. Let X and Y be uniform spaces and A ⊂ X and B ⊂ Y closed subspaces.
(a) The subspace Part (a), whose proof is similar to (b) (but easier), will not be used below; the metrizable case of (a) can also be deduced from Corollary 3.6.
The first two assertions of (b) are proved in [50].
Proof of (b The case A = ∅ is known (see Theorem 2.36); the proof of the general case is based on the same idea but additionally employs Corollary 4.42(b).
Proof. We consider the ARU case; the ANRU case is similar (cf. the proof of Theorem 2.36). Since B is an ARU, it is complete, and therefore closed in Y . Then without loss of generality A is closed in X (else it can be replaced by its closure). Pick a pair (Z, C) of uniform spaces with C closed in Z and a uniformly continuous f : C → U((X, A), (Y, B)). We now apply Corollary 4.42(b). Since C ⋉ A → C ⋉ X is a uniform embedding, f determines a uniformly continuous map Φ : (C ⋉ X, C ⋉ A) → (Y, B). Since C ⋉ A → Z ⋉ A is a uniform embedding and B is an ARU, the restriction ψ : C ⋉ A → B of Φ extends to a uniformly continuousψ : For an alternative proof, see Theorem 4.35.
5.A. Convergence and stability
5.1. Inverse limits. Let X 1 , X 2 , . . . be metrizable uniform spaces. Given uniformly continuous maps f i : X i+1 → X i for each i, the inverse limit L := lim ← − ( is defined to be the subset of X i consisting of threads, i.e. sequences (x 1 , x 2 , . . . ) such that each f i (x i+1 ) = x i . The map f ∞ i : L → X i is defined by restricting the projection π i : X j → X i . The bonding maps f i have compositions X j Since every two uniform covers of each X i+1 can be refined by a single uniform cover, we conclude that a cover of L is uniform iff it can be refined by the preimage of a single uniform cover of some X i . It is easy to check that (L, f ∞ i ) is the categorytheoretic inverse limit, i.e. every family of uniformly continuous maps ϕ i : L ′ → X i commuting with the bonding maps f i factors through a unique map ϕ : L ′ → L (so that each ϕ i = f ∞ i ϕ). It is easy to check that if each X i is complete; separable; point-finite; star-finite; or Noetherian, then so is L. And if each X i is a uniform local compactum and each f i is proper (i.e. the preimage of every compactum is a compactum), then L is a uniform local compactum.
5.2. ε-Separating maps. Let f : X → Y be a map between metric spaces. We recall that f is uniformly continuous iff for each ε > 0 there exists a δ > 0 such that f is (δ, ε)continuous, that is, sends δ-close (=at most δ-close) points into ε-close points. Note that the Hahn property (see §4.B) involved ε-continuous maps, i.e. maps that are (δ, ε)continuous for some δ > 0. Dually, we say that f is (ε, δ)-separating if δ-close points have ε-close point-inverses; and ε-separating if it is (ε, δ)-separating for some δ > 0. Note that when X is compact, the latter is equivalent to the more familiar notion of an "ε-map", which is that f −1 (x) is of diameter < ε for every x ∈ X. ε-Separating maps were known to Isbell [51], who called them simply "ε-mappings". Lemma 5.3. Given an inverse sequence of metric spaces X i and uniformly continuous maps p i , for each ε > 0 there exists an i such that p ∞ i : lim ← − X j → X i is ε-separating.
Proof. Let C be the cover of lim ← − X j by all sets of diameter ε. Since C is uniform, it is refined by (p ∞ i ) −1 (C i ) for some uniform cover C i of X i . If λ is the Lebesgue number of C i , we obtain that p ∞ i is (ε, λ)-separating.
5.4.
Freudenthal's space and mapping telescope. Consider an inverse sequence X = (. . . Note that X ∞ is identified with lim ← − X, and its complement X N is homeomorphic to the topological space i∈N X i . If each X i is complete, X N∪∞ is the completion of X N . We may further define X [0,∞] to be the inverse limit of the finite mapping telescopes Proof. Parts (a) and (b) have been proved above. Parts (c) and (d) follow using that each p ∞ i : X {i,i+1,...,∞} → X i (which is the restriction of p ∞ i : X N∪∞ → X 1 ⊔ · · · ⊔ X i ) is uniformly continuous and ε i -separating, where ε i → 0 as i → ∞, by Lemma 5.3. Parts (e) and (f) follow from (c); part (g) follows from (c) and (d).
From (b) and (d) we immediately obtain (compare [62])
Corollary 5.7 (Bourbaki's Mittag-Leffler Theorem). Let L be the limit of an inverse sequence X = (. . . The substantial result of this subsection is that inverse limits are stable under sufficiently small perturbations of inverse sequences. and q i+1 i f i+1 are α i -close for each i, where (ii) α i > 0 is such that q i j is (α i , 2 j−i β j )-continuous for each j ≤ i; (b) is unique, if in addition to (i) and (ii) the following holds: (iii) each β i > 0 is such that q ∞ i is (δ i , 9β i )-separating, where δ i > 0 is such that q ∞ i is δ i -separating and δ i → 0 as i → ∞; (c) is a uniform homeomorphism onto its image, if (iv) each f i is (γ i , 5β i )-separating, where (v) each γ i > 0 is such that p ∞ i is (ε i , γ i )-separating, where ε i > 0 is such that p ∞ i is ε i -separating and ε i → 0 as i → ∞; (d) is surjective, if in addition to (i)-(v) the following holds: (vi) every y ′ i ∈ Y i is α i -close to some y i ∈ f i (X i ); (vii) X is complete and . . .
Admittedly the statement of Proposition 5.9 is rather cumbersome. For some purposes it becomes more revealing if simplified in one or both of the following ways. (e) if each X i has diameter ≤ 1, and condition (v) is replaced by (v ′ ) each γ i > 0 is such that p i j is (γ i , 2 j−i )-continuous for all j ≤ i; (f) if each Y i has diameter ≤ 1, and condition (iii) is replaced by (iii ′ ) each β i > 0 is such that q i j is (9β i , 2 j−i )-continuous for j ≤ i; (g) when (e) and (f ) are combined.
Proof. If each X i has diameter ≤ 1, we may endow X with the metric d((x i ), (x ′ i )) = sup{2 −i d(x i , x ′ i ) | i ∈ N} and take ε i = 2 −i . Then (v) follows from (v ′ ). Similarly, if each Y i has diameter ≤ 1, we may endow Y with the metric d((y i ), (y ′ i )) = sup{2 −i d(y i , y ′ i ) | i ∈ N} and take δ i = 2 −i . Then (iii) follows from (iii ′ ). In the compact case, a version of Corollary 5.10(g) was obtained by Rogers [78]. His proof is by a different method, reducing the general case to the case where p i j and q i j are embeddings, and involving what appears to be a substantial use of compactness.
We note the following regarding the proof of Proposition 5.9. The proof of (d) will not be simplified if the f i are assumed to be surjective. If the f i are only assumed to be continuous, rather than uniformly continuous, then the hypotheses of (a) and (c) still imply that f is a homeomorphism onto its image and f −1 is uniformly continuous. The constant 9 in (iii) and (iii ′ ) is relevant for (d), but can be replaced by 5 for the purposes of (b).
Proof. (a). Let us consider the compositions F
and F (i) j are 2 j−i β j -close for every i ≥ j. Since Y j is complete, F (j+k) j uniformly converge to a map F j : X → Y j . Since 2 0 + 2 −1 + · · · = 2, it is 2β j -close to F (j) j = f j p ∞ j . Each f i is uniformly continuous, hence so is each F (j+k) j and consequently their uniform limit F j . Since each F (i) j = q j+1 j q j k (y i j ) is 2β k -close to y i k , and q j k (y (j) j ) is 2β k -close to y (j) k := f k p ∞ k (x (j) ). Summing up, y (j) k is 9 2 β k -close to y i k . By the above, the latter is in turn α k -close, hence by (ii) β k -close to y (k) k . Thus y (j) k is 11 2 β k -close to y (k) k . Since by (iv) and (v), f k p ∞ k is (ε k , 5β k )-separating, we conclude that x (j) is ε k -close to x (k) . Consequently x (1) , x (2) , . . . is a Cauchy sequence, and since X is complete, it converges to some x ∈ X.
By (b) we have q ∞ j f = F j in the notation in the proof of (a), where F j (x) is the limit of F j . Without loss of generality, l > j; then by the above, y (l) j is 9 2 β j -close to y m j , where m = ϕ(l). From (i) and (ii), y m j is 2β j -close to q m j (y m ). By our choice of y m , this y m is α m -close to q ∞ m (y), hence by (ii), q m j (y m ) is 1 2 β j -close to q ∞ j (y) (using that m ≥ l < j + 1). To summarize, F (i) j (x (l) ) = q i j (y (l) i ) is 9β j -close to q ∞ j (y); in fact, they are even (9β j − ι)-close for some ι > 0. Hence F (i) j (x) is (9β j − ι 2 )-close to q ∞ j (y) for each i. Then F j (x) is 9β j -close to q ∞ j (y). Since F j = q ∞ j f , by (iii), f (x) is δ j -close to y for each j. Since δ j → 0 as j → ∞, we obtain that f (x) = y.
The following is a direct consequence of Corollary 5.10(a,b,f); the compact case (apart from the uniqueness) is due to Mioduszewski [65]. − → X 0 be an inverse sequence of uniformly continuous maps between complete metric spaces, and let X be its inverse limit. Then there exists a sequence of β * i > 0 such that for each sequence of β i ∈ (0, β * i ] there exists a sequence of α i > 0 such that the following holds. Suppose . . .
− → Y 0 is an inverse sequence of uniformly continuous maps between metrizable uniform spaces, and Y is its inverse limit. If n i is a non-decreasing unbounded sequence of natural numbers, and f i : X n i → Y i are uniformly continuous maps such that the diagram − → Y 0 be an inverse sequence of uniformly continuous maps between uniform ANRs, where Y 0 is a uniform AR, and let Y be its inverse limit. Suppose f : X → Y is a uniformly continuous map, where X is the limit of a convergent inverse sequence . . . − → X 0 be a convergent inverse sequence of uniformly continuous maps between uniform ANRs, where X 0 is a uniform AR [resp. no condition on X 0 ], and let X be its inverse limit.
Then The "only if" direction follows by the definition of a convergent inverse sequence. − → X 0 be a convergent inverse sequence of uniformly continuous maps between uniform ANRs, and let X be its inverse limit. Then there exists a sequence of α * respectively α m i -and α n i−1 -commute, and the compositions Y m i+1 some zero-convergent sequence of ε i > 0. Thus f g is ε i -close to the identity for each i, whence it is the identity. Similarly gf is the identity. Theorem 5.20 immediately implies the following well-known result, whose compact case is due to M. Brown [19].
(i) X and Y are uniformly homeomorphic;
|
2018-08-30T11:05:21.000Z
|
2011-06-16T00:00:00.000
|
{
"year": 2011,
"sha1": "6d08f2d208996aeae207435f04384237bbb1219f",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "7fae04f80a96785f1c60963820da5055ccde68a6",
"s2fieldsofstudy": [
"Mathematics"
],
"extfieldsofstudy": [
"Mathematics"
]
}
|
250669370
|
pes2o/s2orc
|
v3-fos-license
|
75As and 139La NMR/NQR investigations of iron-based superconductor LaFeAs(O0.89F0.11)
We report 75As and 139La NMR/NQR results in LaFeAs(O0.89F0.11). In the normal state, 1/T1T decreases with lowered temperature, which is reminiscent of the pseudogap behavior in the high-Tc cuprates. In the superconducting (SC) state, 1/T1 decreases suddenly below Tc without a Hebel-Slichter coherence peak, followed by a T3 dependence, which is characteristics of unconventional superconductors with lines of nodes. However, the residual density of states in the low temperatures, which is usually observed in unconventional superconductors with crystal imperfections and/or impurity phases, was not observed. Knight shift measurements show that spin susceptibility decreases in the SC state.
Introduction
The recent discovery of iron-based superconductor LaFeAs(O 1−x F x ) with T c up to 43 K under pressure has renewed the record of T c among non-cuprate superconductors [1,2]. One of the notable features is the highly two-dimensional (2D) crystal structure. Fe atoms form a 2D square lattice, and the cylindrical Fermi surfaces deriving from Fe-3d orbitals are reported [3]. Besides, superconductivity occurs when carriers are doped into the parent compound LaFeAsO that undergoes a SDW transition at T N ∼ 140 K accompanied with a structural phase transition at ∼ 160 K [4,5,6,7]. These results imply interplay between magnetism and superconductivity in LaFsAs(O 1−x F x ). In order to reveal magnetic fluctuations in the normal state and the SC properties in LaFeAs(O 0.89 F 0.11 ), we performed 75 As and 139 La NMR/NQR experiments.
Experimental
Polycrystalline samples of LaFeAs(O 0.89 F 0.11 ) synthesized by solid-state reactions [1] are ground into powder for NMR/NQR measurements. The crystal structure, phase purity, and lattice constants of the sample were examined by powder X-ray diffraction [1]. Electrical resistivity measurements via a conventional four-probe method show that resistivity ρ follows ρ 0 + AT 2 below 200 K down to T c , which yields a large residual resistivity ratio RRR=ρ(300 K)/ρ 0 22.5 with ρ 0 ≡ ρ(T → 0) = 0.153 mΩcm (Fig. 1). T c determined from the zero-resistivity temperature is 22.5 K in zero magnetic field. This value of T c is in good agreement with AC susceptibility measurements using an in situ NMR coil at ∼ 12 MHz. A standard spinecho technique was used for NMR/NQR measurements. The nuclear spin-lattice relaxation rate 1/T 1 was measured using a saturation recovery method. The recoveries of the nuclear magnetization M (t) of 75 As (I = 3/2) and 139 La (I = 7/2) after a saturation pulse were fitted with m(t) = (M (∞) − M (t))/M (∞) = 0.1e (−t/T 1 ) + 0.9e (−6t/T 1 ) and m(t) = 0.0119e (−t/T 1 ) + 0.0682e (−6t/T 1 ) + 0.2061e (−15t/T 1 ) + 0.7137e (−28t/T 1 ) respectively, which were measured at the intense peak of the central transition. For Knight shift measurements, the powder sample was mixed with epoxy (Stycast 1266), and then was aligned in a magnetic field of 5.5 T to obtain a narrow NMR linewidth as described in Ref. 8 [8]. Figure 2 shows 75 As and 139 La NMR spectra of the powder sample at 30 K obtained by sweeping magnetic field at 72.1 MHz. Note that this powder sample mostly aligned along the ab-plane, which was inferred from the NMR line shape of the aligned sample. From a splitting between the 1st satellites of the 75 As NMR spectrum, we deduced the NQR frequency to be about 11 MHz, which is in good agreement with the previous report [8]. The 75 As NQR signal arising from the ±1/2 ↔ ±3/2 transitions was observed at 11.2 MHz in zero magnetic field. Figure 3 shows T -dependence of 1/T 1 T of 75 As in 9.82 T and 139 La in 2.55 T. 1/T 1 T both of 75 As and 139 La decreases with lowered temperature, which is reminiscent of the pseudogap behavior in the underdoped regime of high-T c cuprates [9,10], and approaches a nearly constant value in a narrow T -region just above T c . The dashed line in Fig. 3 is a fit to 1
Results and Discussions
.04 (sK) −1 , B = 0.17 ± 0.01 (sK) −1 , and ∆ PG = 172 ± 17 K for 1/T 1 T of 75 As. We point out a difference between the pseudogap behavior of LaFeAs(O 0.89 F 0.11 ) and the underdoped cuprate. In the cuprate, 1/T 1 T decreases from temperatures far above T c and no clear anomaly is observed at T c . In contrast, in LaFeAs(O 0.89 F 0.11 ), the Korringa behavior (T 1 T = const.) is observed in the narrow T -region from 30 K to T c , which is related to the T 2 behavior of the resistivity, and a clear anomaly of 1/T 1 T is observed at T c .
We now turn to 1/T 1 in the SC state. Fig. 4 shows T -dependence of 1/T 1 of 75 As in 9.82 T. 1/T 1 decreases suddenly below T c (H)( 20 K) and exhibits a T 3 dependence down to the However, we point out that a residual density of states (DOS), which is usually inferred from the 1/T 1 ∝ T behavior at the lowest temperatures, is not observed in LaFeAs(O 0.89 F 0.11 ). In unconventional superconductors with line-node SC gap, 1/T 1 follows a T 3 dependence below T c , then turns into linear T behavior at the lowest temperatures due to the residual DOS originating from lattice imperfections and/or impurities. The absence of the residual DOS appears to be contrary to the existence of line-nodes, but be favor for the extended s-wave scenario in which the SC gap vanishes along lines without sign change of the SC gap function. Further NMR measurements using high quality samples and impurity-doped samples are important in order to fully determine the SC gap function.
The spin symmetry of the Cooper pair is inferred from Knight shift measurements in the SC state. Fig. 5 shows the 75 As NMR spectra of the aligned sample obtained by sweeping external field at a fixed frequency of 40.5 MHz. The spectral peak (corresponding to the magnetic field along the ab-plane direction) shifts toward higher fields with lowered temperature, indicating a decrease of the Knight shift in the SC state (see the inset). By measuring a splitting between the 1st satellites of 75 As NMR spectrum at a series of different temperatures, we found that electric quadrupole interaction is T -independent in this temperature range. We also found that the entire shift cannot be ascribed to a SC diamagnetic effect, which was investigated from measuring 139 La NMR spectrum. Therefore, the shift originates from the decrease of the spin susceptibility in the SC state, which is consistent with previous reports [8,11,12]. Detailed analysis of the Knight shift will be published elsewhere.
Conclusion
From 75 As and 139 La NMR/NQR measurements in LaFeAs(O 0.89 F 0.11 ), we observed pseudogap behavior in 1/T 1 T both at the 75 As and 139 La sites, , which is reminiscent of the high-T c cuprates. In the SC state, 1/T 1 decreases suddenly below T c without a Hebel-Slichter coherence T (K) T c Figure 5. 75 As NMR spectra of the central line of the aligned powder corresponding to the magnetic field along the ab-plane obtained by sweeping external field at 40.5 MHz.
peak followed by a T 3 dependence, suggesting that LaFeAs(O 0.89 F 0.11 ) is an unconventional superconductor in which the SC gap vanishes along lines or possesses line-nodes in the SC gap. Knight shift measurements show that spin susceptibility decreases in the SC state. However, the absence of residual DOS at low T is different from other unconventional superconductors with lines of nodes, further NMR studies using high-quality sample and impurity-doped sample are needed to identify the SC gap function.
|
2022-06-28T01:29:31.687Z
|
2009-01-01T00:00:00.000
|
{
"year": 2009,
"sha1": "f98eab5c16b60a55f43e38b96dc10fefe1da807c",
"oa_license": null,
"oa_url": "https://doi.org/10.1088/1742-6596/150/5/052179",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "f98eab5c16b60a55f43e38b96dc10fefe1da807c",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
}
|
250709352
|
pes2o/s2orc
|
v3-fos-license
|
Association between SEMA3A signaling pathway genes and BMD/OP risk: An epidemiological and experimental study
Objective This study aimed to explore the associations of genetic variants in the semaphorin 3A (SEMA3A) signaling pathway genes, including SEMA3A, NRP1, PLXNA1, PLXNA2 and PLXNA3 with osteoporosis (OP) risk and bone mineral density (BMD) in a Chinese Han older adult population. Study design and method A two-stage design was adopted. Total of 47.8kb regions in the 5 genes were sequenced using targeted next-generation sequencing (NGS) technology in the discovery stage, and the discovered OP-related single nucleotide polymorphisms (SNPs) were further genotyped using improved multiple linkage detection reaction technique in the validation stage. Methods of ALP/TRAP staining, real-time fluorescent quantitative PCR, and cell proliferation and apoptosis assays were performed with MC3T3-E1 and RAW 264.7 cell lines to clarify biological effects of observed functional variants in cell lines responsible for bone mass remodeling. Results Total of 400 postmenopausal women (211 OP cases) were involved in the discovery stage, where 6 common and 4 rare genetic variants were found to be associated with OP risk. In the validation stage among another 859 participants (417 women, 270 OP cases), the PLXNA2 rs2274446 T allele was associated with reduced OP risk and increased femoral neck (FN) BMD compared to the C allele. Moreover, significant associations of NRP1 rs2070296 with FN BMD/OP risk and of NRP1 rs180868035 with lumbar spine and FN BMDs were also observed in the combination dataset analysis. Compared to the osteoblasts/osteoclasts transfected with the wild-type NRP1 rs180868035, those transfected with the mutant-type had reduced mRNA expression of osteoblastic genes (i.e., ALP, RUNX2, SP7 and OCN), while elevated mRNA expression of osteoclastic genes (i.e., TRAP, NFATc1 and CTSK). Furthermore, mutant NRP1 rs180868035 transfection inhibited osteoblast proliferation and osteoclast apoptosis, while promoted osteoclast proliferation and osteoblast apoptosis in corresponding cell lines. Conclusion Genetic variants located in NRP1 and PLXNA2 genes were associated with OP risk and BMD. The NRP1 rs180868035 affects bone metabolism by influencing osteoblasts and osteoclasts differentiation, proliferation and apoptosis.
Introduction
Osteoporosis (OP) is a systemic skeletal disease characterized by low bone mineral density (BMD) and deteriorative bone microstructure. Fragility fracture is the most common and harmful complication of OP (1). Dual-energy Xray absorptiometry (DXA)-derived BMD measurements at skeletal sites of total hip, lumbar spine (LS), and femoral neck (FN) were recommended for the diagnosis of OP (2). A systematic review and meta-analysis study demonstrated that the prevalence of OP was 18.3% worldwide, with a significantly higher prevalence in women (23.1%) than in men (11.7%) (3). It was reported the number of fragility fractures was estimated at 3.5 million in 2010, and it would increase 28%, from 3.5 million to 4.5 million, by 2025 in Europe (4). In a study involving 3157 American participants aged ≥50 years, 30% men and 49% women were diagnosed with osteopenia, and 2% men and 10% women were diagnosed with OP based on the BMD measurements at the FN (5). A recent nationwide epidemiological study in China estimated the age-standardized prevalence of OP was 6.46% and 29.13% for men and women aged ≥50 years, corresponding to 10.9 million men and 49.3 million women (6). Above-mentioned evidence suggests that the prevalence of OP stands at a high level, and OP not only causes high disability and mortality, but brings a heavy burden to society, becoming a serious public health issue.
Bone remodeling is a dynamically balanced process regulated by both osteoblasts and osteoclasts. OP occurs when osteoclast-induced bone mass loss exceeds osteoblast-induced bone mass formation. Semaphorin 3A (SEMA3A) is one of the semaphorin family members that are involved in various physiological and pathological activities in human body, e.g., embryonic and nervous system development, immune regulation, promotion of inflammation reduction and oncogenesis (7)(8)(9)(10). Recently, SEMA3A was also found to possess dual role of promoting osteoblastogenesis and inhibiting osteoclastogenesis, suggesting that it may be a potential therapeutic target for OP (11). Animal studies found that the SEMA3A supplementation via transcutaneous injection into the center of the distraction zone elevated BMD, tissue mineral density (TMD) and vascular density of operative area (of tibia) in adult male mice underwent tibia osteotomy surgery (12). Hayashi et al. observed obvious OP phenotype in SEMA3A gene knocked out mice (13). Liu et al. found that fasting serum SEMA3A was positively correlated with bone formation marker -osteocalcin (OCN), while no significant association was found of fasting serum SEMA3A with BMD/osteoporotic fractures (14).
Neuropilin-1 (NRP1) is a type I transmembrane protein which possesses high affinity to SEMA3A, and it plays an indispensable role in the chemorepulsion mediated by SMEA3A (15). PlexinA, one class of the plexin proteins, is the major receptors for SEMA3A. SEMA3A binds to NRP1 to assemble and to active NRP-plexinA holoreceptor complexes (16). It was found that the NRP1-mutant mice exhibited a same osteoporotic phenotype as the SEMA3A-deficient mice, suggesting that the SEMA3A function through NRP1 (13). SEMA3A enhanced the competitive binding of NRP1 (against TREM2) with plexinA1 (PLXNA1), thereby blocking the PLXNA1-TREM2-DAP12 complex formation to inhibit osteoclast differentiation (13). Otherwise, SEMA3A can also bind to the NRP1-PLXNA1 receptor complex to promote osteoblast formation (17). In vitro, plexinA2 (PLXNA2) were found to influence the expression of RUNX2 and SP7-two major osteogenic transcription factors, further regulating osteogenic differentiation and mineralization (18). PlexinA3 (PLXNA3) may participate in the process of SEMA3A regulating bone innervation (19).
Taking all evidence above, the SEMA3A signaling pathway may be involved in the OP pathology, and relevant genes may be associated with the disease susceptibility. However, epidemiological and experimental evidence is still scarce. Herein, we conducted a two-stage study in 1259 Chinese Han elderly to explore OP-related genetic variants in SEMA3A signaling pathway genes, including SEMA3A, NRP1, PLXNA1, PLXNA2 and PLXNA3, using next generation sequencing (NGS) and improved multiple linkage detection reaction (iMLDR) technique. And in vitro experiments were also conducted to explore potential functions and biological mechanisms of revealed OP-related genetic loci.
Study design and participants
The work of this study consisted of four parts. First, we conducted NGS of 5 candidate genes (SEMA3A, NRP1, PLXNA1, PLXNA2, and PLXNA3) using Illumina Hiseq 2500 sequencing platform based on microarray capture sequencing method to discover OP-related genetic variants in 400 Chinese postmenopausal women (211 OP cases). Second, the discovered OP-related single nucleotide polymorphisms (SNPs) were further genotyped using iMLDR technique in another 859 Chinese participants (417 women, 270 OP cases) in the validation stage. Third, we conducted a combination dataset analysis of genotyping data in all participants in the discovery and validation stages to improve the test performance. After the three steps for OP-related genetic variant identification, bioinformatic tools were adopted for function annotations to the identified variants. Fourth, in vitro experiments were conducted to clarify the roles of diseaserelated and functional genetic variants in bone mass remodeling. Considering that postmenopausal women is a high-risk group for OP, we included postmenopausal women exclusively in the discovery stage to identify as many OP-related genetic variants as possible. Detailed illustration to the study design is shown in Figure 1.
Participants were recruited from two communities in Wuhan and physical examination and rehabilitation center of Wuhan Union hospital during 2017-2018. The trained staff conducted an interview by phone to determine whether these participants were eligible for this study. We included Chinese Han participants aged ≥60 years old. We excluded participants: (i) with previous hysterectomy, ovarian and adnexal resection; (ii) with other prevalent endocrine diseases that may potentially influenced bone metabolism (i.e., hyperthyroidism and hypothyroidism); (iii) with hereditary bone disease, including osteogenesis imperfecta, osteochondrosis and multiple myeloma; (iv) who ever took hormonal medicines for a long time (>3 months) or in the past 6 months; and (v) with unavailable data on candidate genes and BMD. Finally, a total of 1259 participants were involved in this study. OP diagnosis is defined as the T value ≤-2.5 (BMD values is ≤2.5 standard deviation below the BMD values in normal adult with same sex and race) based on the World Health Organization OP diagnostic criteria (2). Thus, those passed the inclusion and exclusion criteria and met OP diagnostic criteria were assigned FIGURE 1 Framework of the study design.
Data collection and BMD measurement
Questionnaire survey was conducted to collect information on general characteristics (name, gender and age), medical history (fracture, uterus and ovaries surgery and thyroidrelated disease), medication history (hormones), lifestyle (smoking and drinking), and menopause status for women only (menarche age and menopause age) from each participant. Body mass index (BMI) was obtained from physical examination, and it was calculated as weight (kg) divided by the square of height (m). BMD at the skeletal sites of the LS and FN was measured by Dual-energy X-ray absorptiometry (Lunar Prodigy, GE, USA).
Genotyping and function annotation
Before the physical examination, 3 mL venous blood was drawn from each participant for genotyping. The Relax Gene Blood DNA System Kit (Tiangen, Beijing, China) was adopted to extract genomic DNA. The NGS was used to screen for genetic variants that were potentially associated with OP in the discovery stage. IMLDR technique was used for genotyping the discovered OP-related SNPs in the validation stage. During this process, we set negative controls for each plate, and 5% of double-blind samples were randomly selected for repeat typing verification. The consistency of typing results was 100%, indicating the reliability of iMLDR technique.
In vitro experiments 2.4.1 Plasmid construction and lentivirus packaging
Sequence of NRP1 gene was obtained from the NCBI website. The cDNA sequence of wild-type NRP1 (allele T at rs180868035, I140) and mutant NRP1 (allele G at rs180868035, L140) were inserted into the vector plasmid, named pLVX-NRP1-Puro and pLVX-NRP1-Mut-Puro, respectively. The correctness of the inserted target fragment was verified by sequencing. Both plasmid construction and validation were done by Wuhan Vapol Biological company. The pLVX-NRP1-Puro and pLVX-NRP1-Mut-Puro were introduced into 293 T cells by lentiviral packaging to produce high titer lentivirus containing the target fragment.
Measurement of osteoblast differentiation ability
Alkaline phosphatase (ALP) staining test was used to measure the activity of ALP, which was reckoned as the hallmark enzyme of mature osteoblast. The HiScript II Q Select RT SuperMix for qPCR (+ gDNA wiper) kit was used to reverse transcribe RNA into DNA, and real-time fluorescent quantitative PCR was performed to detect the mRNA expression levels of osteoblast marker genes, including ALP, RUNX2, SP7 and osteocalcin (OCN). In the process, the cDNA synthesized by reverse transcription, as a template, was amplified on an ABI QuantStudio 6 real-rime fluorescence quantitative PCR instrument using SYBR Green Master Mix reagent. The GAPDH gene was used as an internal reference for the relative quantification of osteoblast marker genes mRNA levels, and the relative quantification was calculated using the 2 -DDCt method.
Measurement of osteoclast differentiation ability
Tartrate-resistant acid phosphatase (TRAP) staining test was used to measure the activity of TRAP, which was reckoned as the hallmark enzyme of mature osteoclast. The HiScript II Q Select RT SuperMix for qPCR (+ gDNA wiper) kit was used to reverse transcribe RNA into DNA, and real-time fluorescent quantitative PCR was performed to detect the mRNA expression levels of osteoclast marker genes, including TRAP, nuclear factor of activated T cells c1 (NFATc1) and cathepsin K (CTSK). Other processes were consistent with above.
Cell proliferation assay
The proliferation ability of osteoblast and osteoclast was detected by CCK8 cell proliferation and cytotoxicity assay kit. MC3T3-E1 and RAW 264.7 cells in good logarithmic growth phase were inoculated into 96-well plates at a concentration of 5×10 3 and 2×10 3 cells, respectively, per well, and three replicate wells were set up for each group. Then they were incubated at 37°C in a 5% CO 2 incubator for 48 hours. And 10 mL CCK8 solution was added to each well and incubated for 4 hours at 37°C , then the optical density (OD) value at 450 nm of each well was measured by enzyme-labeled instrument.
Cell apoptosis assay
The AnnexinV-APC/7-AAD apoptosis detection kit was adopted for apoptosis assay. The kit uses fluorescein APClabeled Annexin V as a probe to reflect early apoptosis by binding phosphatidylserine exposed on the outside of the cell to the cytosol of early apoptosis cells. The 7-AAD, provided in this kit, is a nucleic acid dye that cannot penetrate the intact cell membrane of normal or early-stage apoptotic cells, but it can penetrate and bind to DNA within late-stage apoptotic or necrotic cells. Thus, it can be used to distinguish surviving early-stage cells from necrotic or late-stage apoptotic cells. When Annexin V-APC was used in combination with 7-AAD, 7-AAD was excluded from live cells and early apoptotic cells, while late apoptotic cells and necrotic cells were stained, presenting double positive status.
Statistical analyses
Continuous variables were presented as mean ± standard deviation (SD) or median (interquartile range, IQR) where appropriate, and Student's t test or Mann-Whitney U test was used to exam between-group difference. Categorical variables were presented as number (percentage), and Pearson c 2 test was adopted to exam difference between two groups. The c 2 Goodness-of -Fit test was applied to assess whether the genotype distribution of genetic variants in the control group conformed to the Hardy-Weinberg Equilibrium (HWE). The associations between common variants and BMD/OP risk were analyzed using four kinds of multiple linear regression models or multivariate-adjusted logistic regression models, including codominant, additive, dominant and recessive models. For rare variants, considering the small sample size of control group, the East Asian population in the GnomAD database was also used as the control. Fisher's exact test was adopted to compare the allele frequency and genotype distribution of rare variants between OP and non-OP groups, and linear regression was used to analyze the association of rare variants with BMD. The discovery stage was designed to investigate potentially OP-related genetic variant therefore the results of this stage were not corrected for false discovery rate (FDR). The other results were corrected for FDR to control the false positive rate caused by multiple comparisons. Statistical analyses were performed using IBM SPSS v22.0 and R v4.0.3 (R Core Team, Vienna, Austria), and picture production was achieved by GraphPad Prism v5.01. All statistical analyses were two-sided, and a P value < 0.05 indicated statistical significance.
General characteristics
General characteristics of all participants were presented in Table 1. In the discovery stage involving 400 postmenopausal women (211 OP cases), the mean (SD) age of OP cases was 67.67 (5.87) years, comparatively larger than the non-OP individuals aged 66.26 years (4.63) (P=0.008). The mean (SD) BMI of OP cases was 23.22 (3.14) kg/m 2 , significantly lower than the counterpart values (25.24 kg/m 2 (SD: 3.05)) of non-OP group (P<0.001). The proportions of smoker and drinker in both OP and non-OP groups were no more than 3.3% and 2.4%, and both two proportions were not significantly different between two groups. The mean (SD) BMD measurement and T value at the skeletal sites of LS were 0.80 (0.08) g/cm 2 and -3.12 (0.63), and were 0.66 (0.08) g/cm 2 and -2.66 (0.67) of FN in OP group, significantly lower than the counterpart values, sequentially were 1.08 (0.15) g/cm 2 , -0.80 (1.09), 0.83 (0.09) g/cm 2 and -1.25 (0.77), in non-OP group (all P<0.001).
Results in the discovery stage
A total of 794 genetic variants were genotyped in the 5 SEMA3A signaling pathway genes. After excluding genetic variants (i) with low quality and possible repetitive sequences, (ii) with average sequencing depth ≤30×, (iii) with detection rate ≤90%, and (iv) not met HWE equilibrium, 586 variants were retained for subsequent analysis. Among them, 119 variants were new, the rs numbers of whom were unavailable based on databases of 1000G, ExAC, ESP6500 and GnomAD. Among the newly found variants, 12 were nonsynonymous mutations locating in the SEMA3A (n=2), NRP1 (n=1), PLXNA1 (n=3), PLXNA2 (n=3) and PLXNA3 (n=3) genes, respectively. It was predicted that two of them were deleterious to protein function by at least two of three software i.e., the SIFT, POLYphen2 and MutationTaster (Table S1).
Results of associations between the sequenced common variants and OP risk are presented in Table S2. After adjusting for age, BMI and menopause age, 6 common variants were associated with OP risk, including NRP1 rs2070296, PLXNA1 rs4679323 and rs73861745, and PLXNA2 rs1664227, rs2274446 and rs3748735. Compared to the wild genotype carriers (i.e., the rs4679323 CC, rs1664227 GG, rs2274446 CC and rs3748735 CC), carriers of rs4679323 CA (OR=0.53, 95% CI: 0.31, 0.90), rs2274446 CT (OR=0.55, 95% CI: 0.34, 0.88) and rs3748735 CT (OR=0.55, 95% CI: 0.32, 0.93) genotypes were less likely to have prevalent OP, while the rs1664227 CC genotype carriers were more likely to have prevalent OP (OR=2.04, 95% CI: 1.07, 3.86). Results of dominant, recessive and additive models were also presented ( Table S2). Table S3, the MAF distribution of four rare variants, i.e., the rs180868035 and rs767142032 in NRP1, and rs369477952 and rs146550621 in PLXNA1 were found to be significantly different between OP group and the East Asian population in the GnomAD database, with P values of 0.019, 0.022, 0.005 and 0.022 sequentially (Table S3).
Results in the validation stage
Above-mentioned 10 OP-related genetic variants (6 common and 4 rare ones) were further genotyped in another 859 subjects in the validation stage. Characteristics of the 10 variants are shown in Table S4. Compared to the NRP1 rs2070296 CC/CT genotype, the TT genotype was associated with an increased OP risk (OR=1.46, 95%CI: 1.01, 2.11), while the association did not achieve an FDR-corrected significance ( Table 2). The PLXNA2 rs2274446 CT/TT genotype carriers had a 37% reduction in OP risk (OR=0.63, 95%CI: 0.46, 0.87, P FDR =0.027) compared to the CC genotype carriers; and OP risk decreased by about 32% on average in response to a T to C mutant in the PLXNA2 rs2274446 (OR=0.68, 95%CI: 0.52, 0.89, P FDR =0.029).
Results of associations of the 6 common variants and BMD measurements at skeletal sites of LS and FN are presented in Table 3. We found the PLXNA2 rs2274446 CT/TT genotype was As to the rare variant-BMD association, it was found that the BMD decreased in response to per rs180868035 G allele mutation at skeletal sites of LS (b=-0.282 g/cm 2 , 95%CI: -0.507, -0.057) and FN (b=-0.208 g/cm 2 , 95%CI: -0.375, -0.035), respectively.
Results in combination dataset analysis
In general, the significant associations found in the combination dataset analysis were consistent with those found in the validation stage. Notably, we also found the NRP1 rs2070296 TT genotype was associated with higher OP risk (OR=1.63, 95%CI: 1.21, 2.19, P FDR =0.007) and lower FN BMD (b=-0.030 g/cm 2 , 95% CI: -0.050, -0.010, P FDR =0.004) compared to the CC/CT genotype (Tables 4-6).
Effect of NRP1 rs180868035 mutation on osteoblast and osteoclast differentiation
Considering our findings from above work and the bioinformatics analysis of function annotations, the NRP1 rs180868035 was preferentially selected for function research via in vitro experiments. The NRP1 rs180868035 is a nonsynonymous mutation located in the exonic region. It was predicted by the SIFT and Mutation Taster software that the NRP1 rs180868035 T>G mutation can cause the coded amino acid (p.I140L) change (Table S4).
The ALP staining was positive in osteoblasts transfected with an empty vector, and vectors with either the wild-type (allele T at rs180868035, I140) or mutant (allele G at rs180868035, L140) NRP1 gene sequence. Compared to the osteoblasts transfected with empty vector, wild-type transfected osteoblasts apparently enhanced the depth of ALP staining ( Figure S1), as well as mRNA expression of the osteoblast marker genes, i.e., ALP, RUNX2, SP7 and OCN ( Figure S2). Moreover, compared to the wild-type transfected cells, the depth of ALP staining was apparently reduced (Figure S1), and osteoblast marker genes all were lower in the mutant transfected cells ( Figure S2).
The TRAP staining was positive in osteoclasts transfected with an empty vector, and vectors with either the wild-type or mutant NRP1 gene sequence. Compared to the osteoclasts transfected with empty vector (Figure S3), the depth of TRAP staining and osteoclast marker genes expression level, i.e., TRAP, NFATc1 and CTSK, all were reduced in the wild-type transfected osteoclasts ( Figure S4). Besides, compared to the wild-type transfected cells, mutant transfection significantly enhanced the depth of TRAP staining ( Figure S3), and mRNA expression level of osteoclast marker genes ( Figure S4).
Effect of NRP1 rs180868035 mutation on osteoblast and osteoclast proliferation
The average proliferation rates of osteoblasts transfected an empty vector, and vectors with either the wild-type and mutant NRP1 were 97.3%, 136.8% and 103.5%, respectively. As shown in Figure 2A, compared to those transfected with the empty vector, the proliferation rate was significantly promoted in osteoblasts transfected with the wild-type NRP1; and compared to those transfected with the wild-type, the proliferation rate was significantly inhibited in osteoblasts transfected with the mutant NRP1. The average proliferation rates of osteoclasts transfected an empty vector, and vectors with either the wild-type and mutant NRP1 were 92.8%, 61.5% and 86.1%, respectively. As shown in Figure 2B, compared to those transfected with the empty vector, the proliferation rate was significantly inhibited in osteoclasts transfected with the wild-type NRP1; and compared to those transfected with the wild-type, the proliferation rate was significantly promoted in osteoclasts transfected with the mutant NRP1.
Effect of NRP1 rs180868035 mutation on osteoblast and osteoclast apoptosis
The average apoptosis rates of osteoblasts transfected with an empty vector, and vectors with either the wild-type and mutant NRP1 were 6.2%, 3.8% and 5.8%, respectively. Compared to those transfected with the empty vector, the apoptosis rate was significantly reduced in osteoblasts transfected with the wildtype NRP1; and compared to those transfected with the wildtype, the apoptosis rate was significantly increased in osteoblasts transfected with the mutant NRP1 ( Figure 3). The average apoptosis rates of osteoclasts transfected with an empty vector, and vectors with either the wild-type and mutant NRP1 were 8.1%, 27.0% and 10.2%, separately. Compared to those transfected with the empty vector, the apoptosis rate was significantly elevated in osteoclasts transfected with the wildtype NRP1; and compared to those transfected with the wildtype, the apoptosis rate was significantly decreased in osteoclasts transfected with the mutant NRP1 (Figure 4).
Discussion
In the present study, we conducted a two-stage designed epidemiological research in combination with bioinformatics analysis to identify functional genetic variants associated with OP risk/BMD in 5 SEMA3A pathway genes (SEMA3A, NRP1, PLXNA1, PLXNA2 and PLXNA3). In vitro experiments were further performed to investigate roles of identified variants in bone mass remodeling. The NRP1 rs2070296 and rs180868035, and PLXNA2 rs2274446 were revealed to be associated with OP risk/BMD. Based on function annotations to those variants, in vitro experiments were conducted and showed that the NRP1 rs180868035 mutation affect differentiation, proliferation and apoptosis of osteoblasts and osteoclasts.
Our findings showed that the NRP1 rs2070296 TT genotype was associated with increased OP risk and decreased FN BMD, According to the rSNPBase database (http://rsnp3.psych.ac.cn/ index.do), the rs2070296 is either a regulatory SNP or an expression quantitative trait locus, indicating it may be involved in the regulation of NRP1 gene expression. We found little evidence on the association of the rs2070296 variant with specific outcomes other than one literature observed a cumulative effect of SNP rs2070296 on visual acuity underwent ranibizumab treatment (20). The rs180868035 T>G (I140L) is a nonsynonymous mutation at physical position of chr10:33559615 located in exon 3 of the NRP1 gene, corresponding to the change of Ile amino acid to Leu amino acid (p.I140L). It was predicted by the SIFT algorithm to be deleterious to protein function (21). No study on the association between genetic variant of NRP1 rs180868035 and bone metabolism was reported before ours. We further performed in vitro experiments to explore the influence of NRP1 mutation on bone metabolism. Our findings revealed that overexpression of a rs180868035 T>G mutant decreased the expression levels of osteoblast marker genes, i.e., ALP (22), RUNX2 (23), SP7 (24) and OCN (25), inhibited osteoblast proliferation and promoted osteoblast apoptosis. In the meanwhile, the overexpression of the mutant was found to enhance the expression of osteoclast marker genes, i.e., TRAP The N-terminal of SEMA3A contains a seven-bladed b-propeller structure, which was reckoned as the signature 'Sema' domain, determining its unique ability to bind to NRP1, and the NRP1 rs180868035 genetic variant is located in the functional domain in extracellular region of NRP1 that binds to the SEMA3A (28,29). The Npn1-mutant mice (NRP1 Semamice), lacking the Sema-binding site, exhibited a same osteoporotic phenotype as SEMA3Adeficient mice, indicating that the NRP1 functions through binding to SEMA3A (13). In addition, the binding of SEMA3A to NRP1 could inhibit osteoclast differentiation induced by RANKL signaling and enhance osteoblast differentiation through classic WNT/b catenin pathway (9,13). We therefore speculated that the NRP1 mutation caused amino acid sequence alteration that resulted in the inability of NRP1 to bind to SEMA3A, further disrupting the differentiation, proliferation and apoptosis of osteoblasts and osteoclasts. The NRP1 served as a co-receptor and interacted with the discoidin domain receptor 2 (DDR2)-a receptor tyrosine kinase possessing a role of promoting bone-depositing cells and osteoblasts differentiation. Animal experiments suggested that NRP1 enhanced the bone formation induced by DDR2, and knockdown of NRP1 or ectopic expression of NRP1 can influence osteoclastogenesis in cells with different DDR2 expression levels (30). Thus, NRP1 mutation might inhibit the interaction between NRP1 and DDR2. It is an intriguing idea for the future studies. Besides, NRP1 expression was tightly regulated by growth factors, transcription factors, and injury, e.g., tumor necrosis factor-alpha, Prox-1, and ischemia (31). But evidence on how these modulators regulate NRP1 in the process of OP development is still scarce. More studies are needed to investigate potential mechanisms. We found genetic variant of PLXNA2 rs2274446 (C > T) was associated with reduced OP risk and increased FN BMD. PLXNA2 is one of transmembrane receptors and can mediate SEMA3A signals through binding to NRP1 as a co-receptor for SMEA3A (32). Oh et al. found that the PLXNA2 expression could be upregulated by the bone morphogenic protein 2, subsequently activating the expression of two major osteogenic transcription factors, Runx2 and SP7, and reinforcing osteogenic mineralization and differentiation, indicating that the PLXNA2 possesses potential ability to affect bone remodeling and is a noteworthy target for OP treatment (18). PLXNA2 is a receptor of SEMA6A in osteoclasts. The SEMA6A-PLXNA2 axis participated in osteoclastogenesis induced by RANKL, and this function is dependent on the activation of NFATc1 induced by PLCg (33). It offers potential therapeutic targets in the intervention of OP. In a study including 560 postmenopausal women of Korean ethnicity, ten SNPs in the PLXNA2 gene were selected for genotyping and the rs3748735 variant was found to be significantly associated with FN BMD (32). Genetic variant of PLXNA2 rs3748735 was found to be associated with FN BMD at the discovery stage in our study, while the significant association did not maintain at the further validation stages. This may be partly due to different sample sizes and study populations.
As far as we known, our study is one of few studies to explore the association of genetic variants in the SEMA3A signaling pathway genes with OP risk. Our work consisted of populationbased epidemiological study and in vitro experiments, concluding genetic variant of NRP1 rs180868035 may affect BMD by increasing bone resorption and reducing bone formation. Results from these two parts corroborate and complement each other. Thus, our study can provide stable and valid evidence for OP susceptibility loci identification. Meanwhile, several limitations should be mentioned. First, though we found genetic variant of NRP1 rs180868035 could affect osteoblast and osteoclast proliferation and apoptosis, the specific biological mechanism through which NRP1 rs180868035 influence bone metabolism needs to be elucidated in the future. Second, the extensibility of our findings will be limited because we only included Chinese Han residents living in Wuhan, China. Other population-based studies are needed to confirm our findings. Third, in the in vitro experiment part, the early stage of osteoblasts and osteoclasts differentiation was missed to be measured in this study. At last, several potential confounders, e.g., physical activity (34), serum calcium and vitamin D levels (35), were not adjusted in the present study due to insufficient data, which would bias the conclusions.
Conclusion
This study firstly and systematically analyzed the associations between genetic variants in the SEMA3A signaling pathway genes and OP risk/BMD in a Chinese population. We concluded genetic variants of NRP1 rs2070296 and rs180868035, and PLXNA2 rs2274446 were associated with OP risk and BMDs. The NRP1 rs180868035 influences bone metabolism by regulating osteoblast and osteoclast differentiation, proliferation and apoptosis. More researches are needed to prove our findings.
Data availability statement
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 below: www.ncbi.nlm.nih. gov/bioproject/, PRJNA884817.
Ethics statement
The studies involving human participants were reviewed and approved by The Ethics Committee of Tongji Medical College of Huazhong University of Science and Technology. The patients/ participants provided their written informed consent to participate in this study.
|
2022-07-21T15:21:06.861Z
|
2022-11-08T00:00:00.000
|
{
"year": 2022,
"sha1": "0948e1cfb9d0423fc4f0ce45fb04b2b62b9f2652",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Frontier",
"pdf_hash": "052e7d7168796e6cdf9b8f2de0fe2ca18bbb46ee",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
12292612
|
pes2o/s2orc
|
v3-fos-license
|
Current Status of Early Blight Resistance in Tomato: An Update
Early blight (EB) is one of the dreadful diseases of tomato caused by several species of Alternaria including Alternaria linariae (which includes A. solani and A. tomatophila), as well as A. alternata. In some instances, annual economic yield losses due to EB have been estimated at 79%. Alternaria are known only to reproduce asexually, but a highly-virulent isolate has the potential to overcome existing resistance genes. Currently, cultural practices and fungicide applications are employed for the management of EB due to the lack of strong resistant cultivars. Resistance sources have been identified in wild species of tomato; some breeding lines and cultivars with moderate resistance have been developed through conventional breeding methods. Polygenic inheritance of EB resistance, insufficient resistance in cultivated species and the association of EB resistance with undesirable horticultural traits have thwarted the effective breeding of EB resistance in tomato. Several quantitative trait loci (QTL) conferring EB resistance have been detected in the populations derived from different wild species including Solanum habrochaites, Solanum arcanum and S. pimpinellifolium, but none of them could be used in EB resistance breeding due to low individual QTL effects. Pyramiding of those QTLs would provide strong resistance. More research is needed to identify additional sources of useful resistance, to incorporate resistant QTLs into breeding lines through marker-assisted selection (MAS) and to develop resistant cultivars with desirable horticultural traits including high yielding potential and early maturity. This paper will review the current understanding of causal agents of EB of tomato, resistance genetics and breeding, problems associated with breeding and future prospects.
Introduction
Tomato (Solanum lycopersicum Linnaeus), native to the Andean region of South America, is one of the most common horticultural crops and cultivated throughout the world. It can be grown in a wide range of climates from tropical to temperate; it also can be cultivated under cover conditions when outdoor temperatures are not favorable. Tomato is the world's second most consumed vegetable after potato [1]. The total world production of tomato is 161.7 million metric tons with a value of ∼$59 billion. USA tomato production contributes 13.2 million metric tons with a value of $5 billion to the total world production [2]. The USA ranks in third position in the total world production of tomato after China and India [2]. Tomatoes are consumed in several ways: fresh, mixed in other food items or processed and canned as sauce, ketchup, juice, salsa, paste, soup and pickled. Tomato is the richest source of vitamin A and C and supplies a sufficient amount of the antioxidant lycopene pigment that helps to protect the body against cancer and heart disease [3,4]. Because of its wide use according to the host plant of origin. One major cluster contained 62% of isolates from potato, whereas the other major cluster contained 87% from tomato.
Similarly Gannibal, Orina, Mironenko and Levitin [10] analyzed sixty three large-spored Alternaria isolates from different hosts, mainly potato and tomato and a few from wild potato species in Russia, and compared them with representatives of A. solani and A. tomatophila collected in the USA [15]. Sequence analysis of Alternaria major allergen (Alt a1), glyceraldehyde 3-phosphate dehydrogenase (gpd) and calmodulin revealed six, four and four single-nucleotide polymorphism (SNP) markers, respectively. Haplotype analysis showed two major groups representing A. solani and A. tomatophila [10]. Furthermore, the majority of A. solani from the potato cluster had mating type locus 1 (MAT1) idiomorph MAT1-1, while 88% of A. tomatophila isolates had the MAT1-2 idiomorph.
Secondary metabolite profiling has been used to distinguish large-spored Alternaria species, A. dauci, A. porri, A. solani and A. tomatophila [15]. Metabolite cluster analysis of 56 isolates confirmed A. dauci, A. solani and A. tomatophila as three distinctive species with their own metabolite profiles. Furthermore, this study revealed that altersolanol A, altertoxin I and macrosporin are produced both in A. solani and A. tomatophila, but not in A. dauci. Another chemotaxonomic approach of proteins, using matrix-assisted laser desorption ionization time of flight mass spectrometry (MALDI-TOF MS) of 37 isolates showed three major clusters corresponding mainly to A. dauci, A. porri and A. solani/A. tomatophila, but was able to distinguish A. solani from A. tomatophila [16].
Recently, the genome sequence of A solani, A. tomatophila and several other Alternaria species has been made publicly available [17]. Future comparative genome analyses will likely shed additional light on the molecular basis for host specialization between these closely-related pathogens. In the following sections, we will primarily focus on A. solani, recognizing that several others closely-related Alternaria species can cause EB of tomato.
Disease Cycle
A. solani reproduces asexually; a sexual stage of this fungus is unknown. The fungus overwinters in soil, plant debris, seed and alternate hosts in the form of either conidia or mycelia, which may serve as primary sources of inoculum ( Figure 1). The thick cell wall of conidia enables the fungus to adapt to adverse climatic conditions [18]. Infection occurs during warm and humid conditions. Conidia germinates at temperature of 8-32 • C in cool and humid conditions in the presence of moisture to form germ tubes [19,20]. Germ tubes penetrate host tissue directly or enter through stomata or wounds, thereby causing infection. Lesions appear after 2-3 days of infection depending on environmental conditions, leaf age and cultivar susceptibility, and spores are produced 3-5 days after the appearance of lesions [19][20][21]. Generally, a long period of wetness is needed for spore production, but spores are also produced during alternate wet and dry conditions. First, conidiophores are developed during wet nights, which then produce spores or conidia in another wet night after the period of day light and dryness. In the next step, conidia are rapidly dispersed through wind and rain splash and continue the disease cycle in other healthy parts of the same plant or different plants. EB has the potential of causing polycyclic infection because of its short disease cycle [19,20].
Variability among Isolates
Despite having only asexual reproduction, isolates of A. solani highly differ in their morphology, toxin production and genetic composition. Such variation might have arisen from heterokaryosis and natural mutation [23]. Morphological variations among A. solani isolates are observed in terms of types of conidia (beaked or non-beaked), length of conidiophores, dimension of the spore body and vegetative compatibility. Morphological variations are often affected by environmental factors such as substrate, light intensity and temperature. Therefore, other variations should be taken into account while differentiating among species in addition to morphological variation. High variations among isolates have been detected at the molecular level using several molecular markers. In addition, A. solani also possesses extensive variation in terms of its toxin production. Several Alternaria species, which were difficult to distinguish by their morphological characteristics, have been classified based on their toxin production. Nonetheless, differentiation of A. solani based on toxin production is still not unanimous among researchers and scientists [18].
Because of high genetic variation, A. solani can easily adapt to the changing environment and develop resistance to fungicides [24]. Furthermore, high genetic variation also has a higher risk of overcoming existing genetic resistance of the host, thus challenging the control of EB and the development of a completely EB-resistant cultivar [25]. However, there are no physiological races of A. solani known or confirmed so far, suggesting host non-specific recognition of the pathogen. This makes the studies of the pathology and genetics of EB more complex [18].
Molecular Mechanism of Infection
Fungal pathogens produce a number of enzymes and secondary metabolites that help to infect the host plant. The secondary metabolites may kill or activate the host cells and penetrate into the cells before invasion. A. alternata has been reported to produce AAL toxin and causes Alternaria stem canker in tomato [26]. A polyketide synthase gene (PKS) ALT1 was involved in producing the AAL
Variability among Isolates
Despite having only asexual reproduction, isolates of A. solani highly differ in their morphology, toxin production and genetic composition. Such variation might have arisen from heterokaryosis and natural mutation [23]. Morphological variations among A. solani isolates are observed in terms of types of conidia (beaked or non-beaked), length of conidiophores, dimension of the spore body and vegetative compatibility. Morphological variations are often affected by environmental factors such as substrate, light intensity and temperature. Therefore, other variations should be taken into account while differentiating among species in addition to morphological variation. High variations among isolates have been detected at the molecular level using several molecular markers. In addition, A. solani also possesses extensive variation in terms of its toxin production. Several Alternaria species, which were difficult to distinguish by their morphological characteristics, have been classified based on their toxin production. Nonetheless, differentiation of A. solani based on toxin production is still not unanimous among researchers and scientists [18].
Because of high genetic variation, A. solani can easily adapt to the changing environment and develop resistance to fungicides [24]. Furthermore, high genetic variation also has a higher risk of overcoming existing genetic resistance of the host, thus challenging the control of EB and the development of a completely EB-resistant cultivar [25]. However, there are no physiological races of A. solani known or confirmed so far, suggesting host non-specific recognition of the pathogen. This makes the studies of the pathology and genetics of EB more complex [18].
Molecular Mechanism of Infection
Fungal pathogens produce a number of enzymes and secondary metabolites that help to infect the host plant. The secondary metabolites may kill or activate the host cells and penetrate into the cells before invasion. A. alternata has been reported to produce AAL toxin and causes Alternaria stem canker in tomato [26]. A polyketide synthase gene (PKS) ALT1 was involved in producing the AAL toxin biosynthesis. Additional toxins associated with Alternaria sp. infection have been reported such as Brefeldin, tentoxin, ACT toxin, AF-toxin, AK-toxin, ACTG-toxin, ACR-toxin, AM-toxin, AS-toxin and maculosin destruxin B [27]. Enzymes involved in the pathogen establishment on the host tissue surface include cutinases, lipase, endo-glucanases and exo-glucanases [27].
To date, limited studies have been conducted to determine the molecular basis of infection and host responses to A. solani. Many plant pathogenic fungi including Alternaria species produce phytotoxins. Based on selectivity, phytotoxins are classified as either host-specific toxins (HSTs) or non-host-specific toxins (NST). Alternaria species produce a variety of HSTs, and most are low molecular weight secondary metabolites. The HTSs are structurally diverse and some are involved in pathogenicity or virulence during host-plant interaction, as has been shown in A. alternata [27,28].
A. solani produces various types of phytotoxic metabolites, including alternariol, altersolanol A, altertoxin, macrosporin and solanapyrone [15,29]. Secondary metabolite profiling of large-spored Alternaria species revealed that A. solani, as well as A. dauci, A. porri and A. tomatophila strains produce common secondary metabolites including altersolanol A, altertoxin I and macrosporin, as well as unknown A. solani-specific secondary metabolites [15]. The role they play in disease has generally not been determined, with a few exceptions.
Several forms of solanapyrone A, B and C are produced in A. solani. Solanapyrone is derived from a biosynthetic gene cluster consisting of six sol genes encoding a polyketide synthase, an O-methyltransferase, a dehydrogenase, a transcription factor, a flavin-dependent oxidase and cytochrome P450 [29]. Sol5 encodes Diels-Alderase, which is involved in the last step of solanapyrone biosynthesis. Sol5 gene knock out mutants were unable to synthesize solanapyrone, but showed no change in virulence compared to the wild type. Thus, despite the strong phytotoxicity of solanapyrone, it is not required for pathogenicity by the fungus [30].
Extracellular proteases are key enzymes produced by fungi during the host-pathogen interaction and are involved in growth, development, survival and pathogenicity [31]. The presence of an extracellular serine protease and metalloprotease was recently found to be secreted by A. solani and proposed to be involved in phytopathogenicity [32,33]. However, additional molecular studies are needed to demonstrate that they play a role in virulence.
Chitin is the major component of most fungal cell walls. Plants express chitinases during fungal infection and are recognized as pathogenicity-related (PR) proteins as part of the defense response against fungal attack [34]. Against A. solani, plant chitinases produce strong antifungal effects, and transgenic expression of an endo-chitinase in potato provided disease resistance [35]. Similarly, transgenic expression of a rice chitinase gene in tomato enhanced disease resistance to early blight [34].
Microarray expression profiling has been used to investigate differential gene expression in susceptible and resistance tomato varieties in response to A. solani [36]. A number of genes including members of the ethylene responsive transcription factor (ERF) family, ribosomal protein transcription factors and zinc finger transcription factors were responsive in resistance tomato. Such genes have been proposed to play a role in plant defense against A. solani possibly by activating PR protein synthesis [36]. However, direct evidence for such a role remains to be determined.
A. solani is sensitive to quinone outside inhibitor, strobilurin and fungicides. However, resistance is being encountered. An amino acid substitution at position 129 in A. solani cytochrome b (so-called F129L, where phenylalanine is changed to leucine) is responsible for the decreased sensitivity [37]. A single-nucleotide polymorphism responsible for sensitivity has been shown to be increasing rapidly in populations in recent years [38].
Recently, a whole genome sequence for A. solani, as well as several other key Alternaria species has been published [17]. Based on Illumina whole genome sequencing technology, A. solani (BMP 0185) was predicted to contain 11,726 genes in a 32.9-Mb genome. In the coming years, comparative and functional genome analysis among Alternaria spp. will help shed new light on the molecular mechanism of disease caused by A. solani.
Disease Symptoms
The symptoms of A. solani appear in leaf (leaf blight or early blight), stem (collar rot or stem lesions) and fruit (fruit rot) of tomato plant [21,39]. Leaf blight is the most important phase of the disease. It is characterized by dark, small, necrotic, coalescing and concentric lesions giving a target-like appearance on the leaf surface. The lesions are surrounded by yellow rings [21]. The disease first appears on lower older leaves and moves upward as the plant becomes mature [6]. Older leaves are more susceptible than younger leaves. As disease progresses, heavy defoliation occurs, raising the respiration rate and lowering the photosynthetic rate. Defoliation also exposes fruits to the Sun, causing sunscald. This leads to poor fruit quality and significant yield loss [18]. Collar rot is typified by initial dark and sunken lesions on the stem, which later expand to form lens-shaped lesions with concentric rings similar to those on leaves. Lesions at the ground level girdle the stem in young seedlings, destroy the vascular system and form "collars" [18]. Collar rot also serves as a source of inoculum.
The infection on fruit causes dark, sunken, leathery and purple lesions on the stem-end. These lesions expand to a significant size and extend deep into the flesh of the fruit. Infected fruits mostly drop prematurely, and those reaching to maturity also become unmarketable [23].
Control Measures
EB can be controlled by three measures: cultural practices, fungicide treatment and the use of resistant varieties. Cultural practices and fungicidal treatment are more common practices [18]. Cultural practices include maintenance of a healthy field and crop vigor, sanitation, removing infected vines and fruits, plant debris and volunteer weeds from the vicinity of the field, crop rotation and reducing the leaf wetness by soil-directed irrigation systems [18,23]. However, cultural practice alone is not sufficient for the control of EB.
Several types of fungicides have been developed for the control of EB, but fungicide treatment is not economically feasible, nor environmentally sound. Fungicides are first applied 1-2 days after transplantation and then require routine application at the interval of 7-10 days for effective control, thereby increasing production cost and environment pollution [18,20]. In North Carolina, a total of 15 applications of fungicides is suggested per growing season including pre-harvest and during the harvest period [40]. Fungicides often fail to work under high disease pressure [18]. The frequent use of fungicide also results in the emergence of new fungicide-resistant isolates due to high selection pressure.
The use of biological controls has also been documented in several research papers as an alternative method for the management of EB. The four species of Pseudomonas-P. fluorescens, P. aeruginosa, P. putida and P. cepacia-have resulted in less disease severity and higher fruit weights compared to control in tomato [41]. Similarly, P. gladioli B25 reduced EB disease by 60.2% by inducing systemic resistance in the host due to an enhanced level of phenol accumulation, peroxidase activity and PALase activity [42]. Extracts of several plants such as Cinnamomum zeylanicum, Syzygium aromaticum, Ferula foetida, Inula racemosa, Hemidesmus indicus, Rubia cordifolia, Glycyrrhiza glabra and Saussurea lappa also showed antifungal activity against EB pathogen ranging from a medium to a high level [43]. In addition, the role of fungi such as Trichoderma viride [44] and T. harzianum [45] in the management of EB has been reported. The plant growth-promoting rhizobacteria Bacillus subtilis also increased systemic resistance in tomato by inducing growth hormones and defense-related enzymes such as peroxidase, polyphenol oxidase and superoxide dismutase [45]. Other antagonistic bacillus strains (3F-II, 3F-VII, 13F-III, 15F-III) decreased EB severity on tomato, which could be due to an antibiosis-like mechanism or ecological competition [46].
The association of the microbial community including mycorrhiza with plants' growth and development has been known for a long time. These microbes produce different types of hormones including auxins and cytokinins and impact plant growth [47][48][49]. A new approach at the holobiont level is one in which microbial communities are also considered in the plant selection process [50].
This will ultimately help to improve the overall performance of plants under field conditions. While this is a relatively new approach, targeting the microbial community for the improvement of plant performance, it may be a novel approach and can be exploited for the overall performance of plants to address biotic and abiotic stresses in plants [51,52].
Use of resistant varieties is the most provident, effective and sustainable control measure of EB of tomato. The genetic integration of effective EB resistance into cultivated tomato would reduce dependence on chemical inputs and have significant environmental, health and financial benefits.
Screening Methods
Disease screening plays a key role in the identification of resistant lines from diverse germplasm, which then can be used as a source of resistance in breeding programs. Proper disease screening methods are required in disease resistance breeding to get reliable and accurate data. Disease symptoms can be evaluated in the field, greenhouse and growth chamber.
Inoculum Preparation
Inoculum of A. solani can be prepared in various culture media such as potato-dextrose agar (PDA) [53], V-8 juice agar [54], lima bean agar [55] and sporulation medium [56]. Sporulation is enhanced by mycelial wounding or transferring culture onto minimal media and incubating culture at 21-23 • C under a cool-white fluorescent diurnal light with a 12 h photoperiod [56,57]. The conidia are collected after 10-14 days by flooding the plates with ddH 2 O (containing 0.01% Tween 20) and brushing the agar surface with a paintbrush [58]. A spore concentration of 10,000-40,000 conidia/mL is maintained prior to inoculation using a hemocytometer. The inoculation methods may vary depending on the screening methods [59].
Field Screening
Field screening is the most common method of screening. It allows screening of a large number of populations at the same time under natural conditions and keeping track of disease severity in every phase of the plant life cycle [18]. EB severity in the field is usually evaluated by recording the final percent of defoliation and calculating the area under the disease progress curve (AUDPC) using the data observed over time [60]. AUDPC combines the effects of host, pathogen and environment [23].
However, field screening has some disadvantages. It requires a large amount of inoculum, disease spreader rows and intensive labor for the evaluation of large populations. It is also difficult to maintain treatment uniformity in the field. The high disease pressure and confounding effects of other foliar diseases reduce the effectiveness of field screening [18]. To avoid such confounding effects, other foliar diseases should be eliminated using fungicides before the application of inoculum [61]. Besides, the environment may not be favorable for the disease incidence in every season in every location, and even if every condition is optimal, field screening allows only one cycle of screening per year, which decelerates the pace of breeding for EB resistance [18].
Greenhouse Screening
Greenhouse screening is useful in regions where field screening is difficult due to an unfavorable environment and confounding effects of other complex diseases. As the environment can be controlled in the greenhouse, it provides a conducive, uniform and repeatable environment. Unlike field evaluation, greenhouse screening allows several cycles of screening in a year [18]. During greenhouse experiments, inoculated seedlings are first kept under high relative humidity (RH) (100% RH) for 24 h. This is then reduced to 14-16 h only at night for 5-7 days in a mist chamber to allow disease development. The severity of the disease is measured in terms of the percentage of necrotic area on leaves or the number of lesions when there are few necrotic spots usually after seven days of inoculation [57].
Several studies have been conducted in greenhouses with different results. Greenhouse studies involving moderately-resistant and susceptible lines did not show a significant difference in their symptoms [57,62], whereas those involving highly-resistant and susceptible lines showed significant differences in their symptoms [63]. This indicates that greenhouse screening may only be able to distinguish two extreme groups of genotypes. However, in the study conducted by Foolad, Ntahimpera, Christ and Lin [58], the greenhouse results were comparable with those of the field test, suggesting its use for initial screening to expedite the selection process. Greenhouse screening is not very common for genetic analysis studies [25].
Greenhouse screening is affected by several factors including plant age, inoculum quality and quantity, inoculation technique and pre-and post-inoculation environmental conditions [18]. As younger leaves are more resistant to EB than older and mature leaves, it is difficult to distinguish the true EB-resistant genotype seedling stage. Therefore, the greenhouse may be more suitable for the selection of collar rot-resistant lines [64]. However, plants at their late seedling and early flowering stage (7-8 weeks) can be successfully evaluated in the greenhouse [58]. In a study by Vloutoglou and Kalogerakis [65], with the increase in inoculum quantity from 6.2 to 11.5 conidia/mL, the defoliation percentage increased linearly. Likewise, the disease was developed after 4 h of leaf wetness, and the severity increased with the increase in the leaf wetness period up to 24 h, but did not increase further after 24 h [65].
Laboratory Assays (Detached Leaflet Assay)
The detached leaflet assay is performed in a growth chamber. Unlike field and greenhouse evaluation, it can only measure foliar resistance. A droplet of spore suspension is deposited on the upper surface of either punctured [53] or non-punctured [58], young, fully-expanded leaflets obtained from the greenhouse. After inoculation, leaflets are incubated in the dark for 24 h at 22 • C, followed by incubation under cool-white fluorescent diurnal light with a 12-h photoperiod [58]. In the study of Foolad, Ntahimpera, Christ and Lin [58], the detached leaflet assays did not correlate with either field or greenhouse results, but field and greenhouse results had a good correlation, suggesting the requirement of whole plant for the evaluation of EB resistance. On the other hand, Locke [53] obtained successful results with this method. In general, this method alone is not reliable for EB resistance evaluation, but can be used for comparing the virulence of different strains of A. solani and the evaluation of EB resistance along with other methods.
Resistant sources have been identified in wild species S. habrochaites, S. pimpinellifolium and S. peruvianum, and utilized to develop resistant lines in a tomato breeding program [18]. S. habrochaites is the richest source of resistant genes among wild species of tomato, but these species are late maturing, indeterminate and low yielding. Some highly-resistant S. habrochaites accessions are: PI127827, PI126445, PI390513, PI390514, PI390516, PI390658, PI390660, PI390662, PI390663, LA2100 and LA2650 [72,73]. These resistant accessions have been used to develop other EB-resistant lines. S. habrochaites accession PI390662 was used to develop EB-resistant USDA Line 87B187 [57] and PI126445 to develop NCEBR-1 [66]. Similarly, B6013 was used to develop H-7, H-22 and H-25 EB-resistant lines [74]. The F 1 hybrids of resistant accessions of S. habrochaites and the cultivated tomato showed a moderate to high degree of resistance, yet they also possessed other undesirable characteristics of S. habrochaites such as late maturity and small fruit size.
S. pimpinellifolium is a close relative of cultivated tomato, hence, it can be easily crossed with cultivated tomato. It also possesses fewer undesirable traits compared to other wild species of tomato. Therefore, S. pimpinellifolium has been frequently used in tomato breeding programs. Several highlyand moderately-resistant S. pimpinellifolium accessions have been obtained in a field, greenhouse and growth chamber evaluation [18,75]. S. pimpinellifolium Accession LA1921 and its BC4F4 progeny exhibited high resistance towards EB [74].
Although some resistant accessions PI129152, PI127829, PE33, LA1292, LA1365, LA1910, LA1983 and PI270435, have been identified in S. peruvianum, they have not been fully utilized in tomato breeding for EB resistance because of high genetic variation within each accession and the cross-incompatible nature of S. peruvianum with cultivated tomato [57,73] All the identified genetic sources of resistance for EB disease are summarized in Table 1.
Characterization of EB Resistance
EB resistance is a complex trait that is quantitatively inherited and controlled by the additive or non-additive interaction of multiple genes and their interaction with the environment [39,61]. No studies have detected a single qualitative gene for EB resistance. Tomato lines such as C1943, 71B2, IHR1939 and IHR1816 exhibited a polygenic recessive inheritance pattern [39,76,77]. Other lines such as NCEBR-1 and 83602029 showed dominant genes involving additive, dominance and epistatic effects, whereas NCEBR-2, 87B187 and NC39E had polygenic partial dominance [23]. The study of EB resistance is mainly focused on the leaf blight phase. Only a few studies have been conducted on the collar rot and fruit rot phase, where the inheritance of collar rot and fruit rot was independent of leaf blight inheritance [75,78]. The few studies on collar rot and fruit rot might be due to the requirement of a longer time compared to leaf blight, as the disease first appears on leaves.
A low to moderate heritability (h 2 ) of 0.26-0.38 [61], 0.65-0.75 [59,79] and 0.43-0.72 [80] was detected for EB resistance. Therefore, breeding for EB resistance totally based on phenotypic selection will take a longer time. The marker-assisted breeding methods may accelerate the transfer of resistant genes from wild species to cultivated species, but no such markers for the selection of EB resistance are currently available. Some studies have investigated the physiological and molecular defense response of resistant tomato lines against A. solani compared to susceptible lines. EB resistance was associated with enhanced expression of TPK1b (a defense-related gene), increased accumulation of H 2 O 2 , phenolic compounds and superoxide anion, a higher percentage of cross-linking and polyphenol activity, decreased DNA damage and less reduction in cell viability [81]. These data might be useful in understanding the basis of EB resistance, hence to develop resistant tomato lines against EB [81].
Breeding for EB Resistance
Tomato breeding for various desirable traits began more than 100 years ago in the United States [82]. Several breeding lines with EB resistance have been developed utilizing resistance from wild species [39,57,[66][67][68]70,74,83] (Table 1). Improving EB resistance through traditional breeding is difficult and challenging mainly because of two reasons: (i) the polygenic nature of resistance and (ii) the association of resistance with several undesirable characteristics such as physiological maturity, fruit load and plant type. Generally, late-maturing, low fruit-yielding and indeterminate tomato lines are more EB resistant, which has impeded the development of high EB-resistant cultivars with desirable traits through phenotypic selection [62,84,85]. In a cross between S. habrochaites and S. lycopersicum, resistant lines were obtained, but with late maturity, indeterminate growth and low yield, which were inherited from S. habrochaites accessions [23]. However, some recombinant inbred lines (RILs) that were derived from the cross of S. pimpinellifolium accession LA2093 and S. lycopersicum tomato breeding line NCEBR-1 were reported with EB resistance and other desirable characteristics including high yield, earliness in maturity and small plant size [86]. Although EB resistance in those RILs was partially affected by growth habit, plant size, maturity and fruit yield, no known genetic relationship between EB resistance and other horticultural traits could be detected in this study.
Younger leaves are more resistant to EB than older leaves because of high sugar content and glycoalkaloids (solanine, chaconine and solanidine) [6,87]. Therefore, a late maturing crop may appear EB resistant without even having resistant genes because of the high sugar content and glycoalkaloids [23,88]. Such a correlation of negative horticultural traits and EB resistance has created difficulty in phenotypic selection during tomato breeding. - [75] The development of molecular markers opens the possibility of solving the problems seen in traditional breeding of EB resistance. Molecular markers, such as restriction fragment length polymorphisms (RFLPs), amplified fragment length polymorphisms (AFLPs), simple sequence repeats (SSRs), cleaved amplified polymorphic sequences (CAPS), resistance gene analogues (RGAs), expressed sequenced tags (ESTs) and conserved orthologs sets (COS), have already been mapped onto all 12 tomato chromosomes and used for locating QTLs governing important agricultural traits [1] ( Table 2). The emergence of next generation sequencing (NGS) and high throughput genotyping platforms has enhanced the discovery of a large number of single nucleotide polymorphisms (SNPs). As SNPs can detect minor variation among populations and can exploit intraspecific variation, they are now the "marker of choice" in tomato breeding [91]. Such markers would be very useful in the dissection of complex quantitative EB resistance into individual genes, selection of an EB-resistant genotype at the early seedling stage and understanding the genetic basis of correlation between EB resistance and negative horticultural traits [80]. Therefore, quantitative trait loci (QTL) mapping efforts are being conducted to identify markers linked to QTL controlling the EB resistance. The location of QTLs governing different horticultural traits and EB resistance can then be compared to determine if the correlation is due to a pleiotropic effect or linkage drag [92]. Unfortunately, no such markers have been developed yet for the selection of EB resistance because of the lack of QTL strong enough to be used in breeding programs. The QTLs associated with EB resistance that have been identified so far are summarized in Table 2. The first QTL mapping effort for EB resistance was performed on backcross populations (BC 1 and BC 1 S 1 ) resulting from a cross between susceptible S. lycopersicum (NC84173) and highly-resistant S. habrochaites (PI126445) [90]. In this study, the self-incompatible, late-maturing, indeterminate and low-yielding plants were removed from the BC 1 population before EB evaluations to avoid confounding effects of those factors on EB resistance. Nine significant QTLs on chromosomes 1, 2, 5, 9, 10, 11 and 12 and two minor QTLs on chromosomes 8 and 11 were detected in the BC 1 population for EB resistance using the simple interval mapping method. The individual effects ranged from 8.4% (chromosome 5) to 21.9% (chromosome 1). Simple interval mapping analysis also detected thirteen significant QTLs in the BC 1 S 1 population on chromosomes 1, 2, 3, 5, 8, 9, 10, 11 and 12 for EB resistance, with individual effects varying from 7.9% (chromosome 5) to 25.9% (chromosome 9). The six QTLs out of 11 in the BC 1 population on chromosomes 1, 5, 9 (two QTLs), 10 and 11 and, again, the six QTLs out of 13 located on chromosomes 1, 5, 8 (two QTLs), 9 and 10 in the BC 1 S 1 population were confirmed as independent QTLs using the composite interval mapping method [90].
Zhang, Lin, Nino-Liu and Foolad [93] conducted a mapping study on the same BC 1 population using a selective genotyping approach to detect and confirm QTLs governing EB resistance in PI 126445. They identified seven QTLs on chromosomes 3, 4, 5, 6, 8, 10 and 11; four of which, on chromosome 5, 8, 10 and 11, were also detected in the previous study of Foolad, Zhang, Khan, Nino-Liu and Lin [90], suggesting the potential utility of these QTLs in molecular breeding of tomato for EB resistance. All of the QTLs were inherited from a resistant parent except on chromosome 3.
Another study identified six QTLs governing EB resistance on chromosomes 1, 2, 5, 6, 7 and 9 that explained 7-16% of total phenotypic variance, using different environmental conditions and disease scoring methods in F 2 and F 3 . F 2 and F 3 populations were obtained from a cross between EB-resistant S. arcanum (LA2157) and susceptible S. lycopersicum (Solentos) [25]. QTLs on chromosomes 2 and 7 were inherited from a susceptible parent. Three EB QTLs on chromosomes 2, 5 and 9 were detected in both field and glasshouse tests and also coincided with stem lesions, whereas two QTLs on chromosomes 1 and 6 were detected only in a field test, and one QTL on chromosome 7 was detected only in a glasshouse test. Only QTLs on chromosomes 2 and 9 were effective in both environments. QTLs on chromosome 9 also contributed 35% of total phenotypic variance of stem lesions; hence, they could be useful in breeding. The detected QTLs were also similar to QTLs detected in Foolad, Zhang, Khan, Nino-Liu and Lin [90] and Zhang, Lin, Nino-Liu and Foolad [93].
QTL mapping for EB resistance has also been conducted in a population derived from S. pimpinellifolium. After extensive screening of 300 accessions of S. pimpinellifolium, Accession LA2093 (previously designated as PSLP125) was selected for QTL mapping. LA2093 has good EB resistance plus other desirable characteristics [18,80,86]. First, the F 2 , F 3 and F 4 populations derived from the cross of LA2093 and tomato breeding line NCEBR-1 were used for QTL mapping, which detected 10 QTLs conferring EB resistance on chromosomes 2, 3, 4, 5, 6, 7, 9 and 12 with an individual effect of 7.6-13.4% and a combined effect of 44% of total phenotypic variance [18]. Out of ten QTLs, three were from NCEBR-1. This result was further verified using the F 2 population of the same cross for QTL mapping through a selective genotyping approach, where nine EB QTLs were detected on chromosomes 1, 2, 3, 4, 5, 6 and 11 [18]. Therefore, these QTLs can be further utilized in EB resistance breeding.
Again, five major QTLs for EB resistance were identified on chromosomes 2 (two QTLs), 5, 6 and 9, using recombinant inbred lines (RILs), i.e., the F 7 , F 8 , F 9 and F 10 generations of the same cross (LA2093 x NCEBR-1) [80]. QTLs on chromosome 2 and 6 were from LA2093, whereas QTLs on chromosomes 5 and 9 were from NCEBR-1. QTLs on chromosomes 5 and 6 were most stable across generations with the largest effects that explained 11-17% and 10-16% of the phenotypic variation, respectively. Therefore, these two QTLs are important in EB resistance breeding. The detected QTLs were also co-localized with other resistant genes and candidate ESTs, suggesting the possibility of the detection of candidate genes for EB resistance after more research and investigation [80].
Several breeding efforts are ongoing to develop early blight-resistant varieties, but no commercial early blight-resistant variety has been reported yet [94]. The lack of qualitative genes and markers and the association of undesirable traits have limited the development of a resistant cultivar. Researches are also directed towards the transgenic tomato to increase the EB resistance. Arshad et al. [95] reported enhanced early blight tolerance with high fruit and nutritional quality in tomato plants transformed with the rolB gene of Agrobacterium rhizogenes. Babu, Jogaiah, Ito, Nagaraj and Tran [94] reported that the plant growth-promoting rhizobacteria (PGPR) are involved in the suppression of A. solani. Tomato plants pre-treated with PGPR showed improved resistance against early blight, which was associated with increased synthesis of antioxidant peroxidase (POX) and polyphenol oxidase (PPO) enzymes. This suggests the potential utility of POX and PPO genes to develop transgenic tomato lines with improved resistance against early blight [94].
Conclusions and Future Prospects
To conclude, EB resistance is a complex trait, governed by multiple genes, and yet, no pathogen races have been reported. There is no single gene-for-gene model for EB resistance and no major resistance gene identified. Furthermore, EB resistance is affected by different factors including physiological maturity, plant type, growth habit and fruit load. More importantly, the existing EB resistance in cultivars is not enough, and still, growers depend on cultural practices and extensive use of fungicides for the control of EB [86]. Although EB resistance has been identified in wild species, only a moderate level of resistance has been obtained in breeding lines and cultivars, and often, they are associated with undesirable horticultural traits like low yield and late maturity. Some desirable epistatic interactions between the QTLs and other segregating genes from wild germplasm may be lost while transferring alleles from wild to cultivated species, leading to the change in QTL effects in cultivated species [96]. Because of this, it is difficult to obtain full resistance as that of the original wild sources.
Breeding programs require stable QTLs having large additive effects and independent of epistatic interactions to be used in MAS. In the above studies, QTLs for EB resistance have been detected in almost all tomato chromosomes in different interspecific crosses. Among them, QTLs on chromosome 5 and 6 were reported in every study and have also been verified in a study by Foolad, Merk and Ashrafi [18]. QTLs on chromosome 5 and 6 were also the most stable QTL in a study of Ashrafi and Foolad [80]. However, the detected QTLs had small effects, large marker intervals, i.e., >10 cM, that may contain multiple closely-linked genes for the single trait, and only a few of the detected QTLs have been validated. It is also difficult to determine whether the detected QTL controls other traits or not. Further study of these QTLs and fine mapping might lead to important outputs in EB resistance breeding programs. Fine mapping allows precise estimation of the effects and chromosomal positions of the detected QTLs, which can be conducted by establishing introgression lines (ILs), near isogenic lines (NILs) and sub-NILs containing the resistance portion from wild species so as to incorporate resistant QTLs into cultivated tomato. Such lines also help to break down the possible negative correlation of the resistance QTLs with other undesirable traits prior to introgression in elite cultivars and to identify the physiological basis of QTL effects, pleotropic effects and multiple genetic factors controlling the trait [96].
Additionally, narrow genetic diversity among cultivated tomatoes, limited polymorphic markers for intraspecific populations, extreme labor for markers like AFLP and RFLP and the unavailability of reliable PCR-based markers are major problems in molecular breeding of tomato. Most genetic maps were based on interspecific crosses of cultivated tomato and wild species because of the greater number of polymorphic markers, but most breeding populations are based on intraspecific crosses of cultivated tomato or with that of the closely-related species S. pimpinellifolium, where there are limited marker polymorphisms [91]. The polymorphic markers selected from the wide crosses were not polymorphic within cultivated species [97].
With the development of next generation sequencing technologies and the decreased cost of sequencing, highly polymorphic markers such as SNPs can now be incorporated into the tomato breeding program, which can detect the polymorphisms among closely-related individuals within S. lycopersicum or between S. lycopersicum and closely-related species. The work in [98] identified one SNP per 8500 bases with 101 candidate SNPs within cultivated germplasm for 44 genes. Labate and Baldo [99] identified several SNPs among more than 15 tomato lines by EST mining and resequencing. Hamilton, Sim, Stoffel, Van Deynze, Buell and Francis [97] identified 62,576 SNPs through NGS of the transcriptome for six tomato accessions: three fresh market types, two processing, one cherry and one S. pimpinellifolium. Identification of SNPs and their utilization to uncover genetic variation among elite cultivars provides new insights into tomato breeding for EB resistance.
In general, breeding for EB resistance in tomato without compromising other desirable horticultural traits is a demanding task. More research and investigations are needed to identify additional resistant sources, to validate, utilize and introduce detected QTLs in cultivated tomato species for the release of EB-resistant breeding lines and cultivars with high yield and early to mid-season maturity.
QTLs have been identified in S. habrochaites, S. pimpinellifolium and S. arcanum. QTLs on chromosome 5 and 6 were stably observed across generations and environments with the effect of 10-16% [80] and also reported in every study, whereas the QTL on chromosome 9 explained the largest percentage of phenotypic variance (25.6%) for EB resistance in the S. habrochaites PI 126445 accession [90,93]. The QTL on chromosome 9 of the S. arcanum LA2157 accession was also stable and showed a major effect on the stem lesion resistance, i.e., 35% of total phenotypic variance [25]. However, none of these QTLs have been fine mapped. Therefore, research can focus on these chromosomes with the largest QTL effects in the future. Fine mapping of these QTLs will allow the determination of the precise location of these QTLs and the prevention of the introgression of other undesirable regions of the donor genome. Since the detected QTLs have less individual effects, QTLs from different species can be combined after fine mapping and QTL validation in a single cultivar via gene pyramiding to provide strong resistance. Besides, there is a need for additional sources of genetic resistance and its characterization, which can be obtained through germplasm screening.
The new approach of crop improvement may be exploited to investigate if there are any microbial species that may suppress the growth of Alternaria solani or A. alternate. Alternatively, there may be an enhanced level of resistance in tomato in the presence of certain microbes, which needs to be investigated. This is a completely new area of investigation.
|
2017-10-03T13:42:50.896Z
|
2017-09-21T00:00:00.000
|
{
"year": 2017,
"sha1": "850f343d38d391922f2a603d7a91219ae48c839b",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/1422-0067/18/10/2019/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "850f343d38d391922f2a603d7a91219ae48c839b",
"s2fieldsofstudy": [
"Agricultural And Food Sciences"
],
"extfieldsofstudy": [
"Biology",
"Medicine"
]
}
|
270816795
|
pes2o/s2orc
|
v3-fos-license
|
Showcasing the Saudi e-referral system experience: the epidemiology and pattern of referrals utilising nationwide secondary data
Introduction Referrals are an integral part of any healthcare system. In the Kingdom of Saudi Arabia (KSA) an electronic referral (e-referral) system known as the Saudi Medical Appointments and Referrals Centre (SMARC) began formally functioning in 2019. This study aims to showcase the Saudi experience of the e-referral system and explore the epidemiology of referrals nationally. Methods This retrospective descriptive study utilised secondary collected data between 2020 and 2021 from the SMARC system. Cross tabulations with significance testing and colour-coded maps were used to highlight the patterns across all regions. Results The study analysed over 600,000 referral requests. The mean age of patients was 40.70 ± 24.66 years. Males had a higher number of referrals (55.43%). Referrals in 2021 were higher than those in 2020 (56.21%). Both the Autumn and Winter seasons had the highest number of referrals (27.09% and 27.43%, respectively). The Surgical specialty followed by Medicine had the highest referrals (26.07% and 22.27%, respectively). Life-saving referrals in the Central region were more than double those in other regions (14.56%). Emergency referrals were also highest in the Southern regions (44.06%). The Central and Eastern regions had higher referrals due to unavailable sub-speciality (68.86% and 67.93%, respectively). The Southern regions had higher referrals due to both unavailable machine and unavailable beds (18.44% and 6.24%, respectively). Conclusion This study shows a unique system in which referrals are between secondary, tertiary, and specialised care. It also highlights areas of improvement for equitable resource allocation and specialised care in slightly problematic areas as well as the use of population density in future planning.
Introduction
Digital health is transforming healthcare into real-time, individualised care, enhancing diagnosis, treatment, and patient empowerment (1).It provides opportunities beyond conventional healthcare for prevention, early illness detection, and chronic disease management (2).However, literature shows mixed results of this digital transformation across different countries (3).In new medicine, digital technologies can reinforce best practices like electronic referrals (4).
E-referrals are critical for providing quality healthcare.Efficient referral systems promote collaboration across all levels of care (5).
Referral system success depends on many factors including patient barriers, resources, technology, and patient behaviour (6).
Saudi Arabia has recently undergone significant healthcare reforms and system changes as part of the National Transformation Programme launched in 2015 under Saudi Vision 2030.This aims to provide equitable, high-quality healthcare for all through innovations such as a robust digital health infrastructure (7)(8)(9).One key component of the digital health transformation is the establishment of a national electronic referral system known as the Saudi Medical Appointments and Referrals Centre (SMARC).SMARC facilitates referrals between healthcare facilities across all levels of care in the Kingdom.It utilises a Unified System of Medical Referrals (USMR) to receive and coordinate referral requests nationally through a centralised platform (10).
Whilst Saudi Arabia has made significant progress in implementing digital health, few studies have evaluated the impacts and effectiveness of these efforts.One study found preparedness amongst Saudi facilities for adapting to Vision 2030 changes was varied (11).Understanding patterns and utilisation of the new e-referral system across regions can provide insights into its performance and areas needing improvement.
The e-referral system within the Kingdom of Saudi Arabia (KSA), previously known as Ehalati, faced challenges when initially launched in 2012 including fragmented systems across hospitals, lack of integration between public and private facilities, and inadequate expertise in digital health solutions.The information technology platform at that time lacked features like artificial intelligence, robust data analytics, and interoperability, making centralised data management difficult.With many hospitals relying on their own individual platforms, doctors often depended on informal referral networks to coordinate care.However, aligned with Vision 2030, the centralised electronic referral system was revamped and fully reimplemented in 2019 as the SMARC (12).
Since 2019, substantial improvements have been made with Ministry of Health (MoH) support to transition to a unified, national e-referral platform.Targeted training programmes were implemented to build digital health capabilities across facilities.The SMARC system leverages advanced health information technologies like artificial intelligence and predictive analytics for improved care coordination.By standardising the e-referral platform and workflows across all public and private hospitals, SMARC addressed fragmentation and seamlessly integrates referrals digitally.With all governmental and majority of private healthcare facilities now connected to the centralised system, SMARC facilitates efficient nation-wide referral management and represents a major milestone in the digital transformation of Saudi Arabia's health sector.
Other countries have also implemented effective digital health systems, such as Catalonia's electronic health information exchange which has been a European leader since 2009.This system enabled critical health data sharing during the COVID-19 pandemic (13).Additionally, the European Health Data Space (EHDS) promotes individuals' electronic health data access and use for research and public benefit (14).Global digital health initiatives like Catalonia's and the EHDS exemplify how digital systems can improve health outcomes and research.Lessons from these efforts can inform Saudi Arabia's digital health advancements under Vision 2030.
Saudi Arabia currently serves a population of almost 34 million through a combination of public and private facilities across 13 administrative regions.As part of Vision 2030 reforms, the healthcare system is being upgraded to boost quality, efficiency and value through integrating public and private sectors.This includes establishing five new business units to manage the 13 healthcare regions alongside national insurance companies, overseen by the MoH and new insurance centres (15)(16)(17).
This study is the first to showcase the KSA's nationwide referral patterns using routine data from the new SMARC e-referral system.Examining referral epidemiology and trends will provide insights into the system's effectiveness and inform future optimisations to enhance its impact as a key digital health initiative under Vision 2030.
Setting and data source
Under the new healthcare transformation adopted by the MoH, the 13 administrative areas will be pooled into five BUs as follows; Asir, Jazan, and Najran in the Southern BU; Aljouf, Hail, Northern Border and Tabuk in the Northern BU; Riyadh and Alqassim in the Central BU; Makkah, Medina, and Albaha in the Western BU, and the Eastern administrative area in the Eastern BU (15).
All hospitals have a designated coordination department usually known as the Office of Coordination and Eligibility for Treatment (OCET), which has access to the USMR.The OCET receives a referral request from the treating physician which is then uploaded to the USMR.Depending on a patient's medical condition, the referral request is uploaded as either lifesaving, emergency, or routine.These three types of referrals are categorised by SMARC to facilitate the referral process, and to timely secure acceptance to patients who are in most need.
For emergency and routine referrals, the OCET has the privilege to choose up to three hospitals that can potentially offer the needed service at the same region, or alternatively the USMR will automatically choose three appropriate hospitals.The SMARC system has built-in timeframes for referral requests to be accepted, depending on the urgency.For emergency referrals, hospitals have 72 h to accept the request, whilst routine referrals have a 14-day timeframe.If the initially chosen hospitals reject the request within the allotted timeframe, the request is sent to additional hospitals for consideration.If no hospital has accepted the referral request once the timeframe elapses, the case is escalated to the SMARC medical referral management team.They will find an appropriate alternative from the pool of public and private hospitals, searching both within the same region and in other regions if needed.Importantly, whilst awaiting referral acceptance, patients continue receiving necessary healthcare management at the sending hospital to ensure stability until the transfer is arranged.
To expedite the referral request for life-threatening cases, SMARC offers a 24-h lifesaving hotline (1937) in which any treating physician can call directly.The call is answered by a SMARC lifesaving agent and directed to an on-call medical consultant for review and acceptance.If the request is accepted as a lifesaving by the on-call consultant, the treating physician through the OCET, will upload the request to the USMR along with the acceptance code and the name of the receiving hospital.Treating physicians who requested emergency referrals can also call and use this service when patients' health conditions deteriorate whilst waiting the emergency referral acceptance.
Additionally, SMARC oversees referrals for Saudi patients seeking to return to the KSA for treatment.For these cases, Saudi Embassies abroad have access to the USMR and may initiate a referral request.Figure 1 describes the process of the referral requests acceptance.
This study utilised routinely collected secondary data extracted from the SMARC e-referral system database between 2020 and 2021.Permission was obtained to access and analyse this de-identified dataset for research purposes, which was provided by the SMARC team after obtaining necessary approvals.The informed consent was waived given the retrospective nature of this study which relied solely on anonymized secondary data.The dataset was checked for completeness and consistency, and any incomplete or inconsistent records were removed prior to analysis.No personally identifiable information was included to maintain patient confidentiality.The variables included in the dataset are described in the Measurements section.
Study design
This retrospective study utilised secondary routinely collected data provided by the SMARC e-referral system.The dataset includes all referral requests submitted through the SMARC system nationally in 2020 and 2021, with no exclusion criteria applied.
The process of referral acceptance across Saudi regions.
Measurements
The dataset includes variables on sex, age, date of referral (month and year), type of referral (e.g., lifesaving, routine), bed type (e.g., ward bed or burn bed), reason for referral (e.g., unavailable speciality or unavailability of a specialised physician), medical speciality requesting the referral (e.g., medicine or surgery), region of referral request according to the five business units of the New Model of Care as well as according to the entire 13 administrative regions of the country.
Statistical analysis
To answer the objectives of the study, cross tabulations of explanatory variables according to the five BUs were performed, and tests of significance through Chi-squared tests and ANOVA tests were computed where appropriate.All analyses were run using the Stata statistical software version 16 (18).To further study the distribution of referral requests across the 13 administrative areas, colour-coded maps were drawn in ArcGIS (GIS software) version 10.0 (19), according to the percentage of referrals of each area.
Sociodemographic characteristics of patients
Table 1 presents the sociodemographic characteristics of all patients.The total number of patients was 671,672 with an average age of 40.70 ± 24.66 years.Over 55% of referrals were for males.Non-Saudi's made up 15.11% of the total referrals.Most referral requests originated from the Western BU (34.99%), and the least originated from the Eastern BU (11.02%).Referrals were higher in 2021 compared to 2020 (56.21% and 43.79%, respectively).
Pattern of referrals across months
Upon examining the overall monthly pattern of referrals in Figure 2, both years of 2020 and 2021 have commenced with a high percentage of referrals, dipping to their lowest levels in April 2020 and May 2021.In 2020, the chart displays an initial decline in medical case referrals between February and March.From the beginning of April onwards, there is a significant graduate increase in referrals through to the year end.In contrast, 2021 displays a less consistent trend with more fluctuations and a significant increase in referrals in March, June, August, and December.Comparatively, referral rates from both years meet by the end of the year, indicating an expected new standard for medical referrals has developed.
Patterns across medical specialties
Figure 3 shows the pattern of medical specialties requesting referrals.Patients with referrals pertaining to internal medicine were the most common, reaching over a quarter of all requests (27.74%).Followed by general surgery and cardiac surgery (25.23% and 9.63%).The least common referral requests were for anaesthesia (0.03%).
Sociodemographic variables and region of referral request
Associations between sociodemographic variables and the region of referral request are presented in Table 2. Referrals originating from the Western and Southern regions were for patients who were relatively older than those from other regions (average age 42.38 and 41.63 years).Males dominated referrals from the Southern regions (59.18%), whereas for females, referrals were similarly high for both the Central and the Eastern regions.Requests for non-Saudis was highest in the Western region and lowest in the Eastern region (19.46% and 9.06%, respectively).All associations were significant at the 0.05 level.
Referral characteristics and region of referral request
The Central region had the highest number of referrals due to life saving events (14.56%), whereas the Northern region had the lowest at only 2.32%.For routine outpatient referrals, those originating from the Northern region were the highest reaching 57.39%.Emergency related referrals were most common in the Southern region and least common in the Eastern region (44.06% and 25.94%, respectively).As for dialysis, no referrals were registered from the Eastern region, whereas the Western region had 93 requests.Also, 44.85% of referrals for ward beds were found to be in the Southern region.The Western region had the highest requests for isolation beds (5.63%).Requests for ICU beds was highest in the Central region (9.36%), whereas for CCU beds it was highest in the Western region (3.26%).As for PICU and NICU beds, they were highest in the Northern and Western regions, respectively.
As for reasons for referral, unavailable subspeciality was the most common reason and was highest in the Central region followed by the Eastern region (68.86% and 67.93%, respectively).The unavailability of a specialised physician was mostly reported in the Northern region (24.03%).The Southern region mostly reported the unavailability of a machine and the unavailability of a bed compared to all other regions (18.44% and 6.24%, respectively).Referrals due to social reasons were most commonly reported in the Western region, whilst there were 213 referrals due to a royal order from the Eastern region.Referrals due to injuries were only reported in the Southern region, whereas referrals due to health crises were highly reported in the Western region.All associations were significant at the 0.05 level (Table 3).
Total referral requests and referrals received by administrative areas
Both the Eastern and Makkah administrative areas were in the highest quintile with requests beyond 10.49% for both.However, Hail, The e-referral requests by medical speciality for 2020 and 2021 across the Kingdom of Saudi Arabia.Tabuk, and Najran administrative areas were within the lowest quintiles (Figure 4).As for receiver areas, both Riyadh and Makkah were within the highest quintile both reaching above 12.80% of the total requests received.Whereas, Hail, Najran and the Northern areas were amongst the lowest (Figure 5).
Discussion
This study is the first to present the current status of the Saudi e-referral system.It also explored the patterns of e-referrals across the country utilising routinely collected data stored by the SMARC system.Patterns of referrals have been enormously studied worldwide (20)(21)(22)(23).However, making clear comparisons are likely to be difficult due to differences between countries in, for example, local contexts and health care systems (24).Also, patterns of referrals in the current literature were mostly limited to primary healthcare referrals (20)(21)(22)(23).This contrasts with the SMARC system in the KSA, which is concerned with secondary, tertiary, and specialised levels of care only.Also, this analysis of the Saudi e-referral system provides the first empirical evidence of inequalities across the different BUs.Previous studies in the KSA have shown that there are discrepancies in the quality of treatment provided to COVID-19 patients amongst the five different BUs at the outbreak's onset (15).Other several noteworthy observations can be drawn from the findings of this study.
Sex variations in e-referrals suggest the presence of disparities in healthcare-seeking patterns.Higher referrals amongst males are reflected in the higher proportion of males compared to females as shown in the 2022 census (25).However, sex variations were observed in referrals in other countries including America and Canada (23,26,27).
The observed discrepancy in the ratio of Saudi/non-Saudi patients may be ascribed to the inherent characteristics of the healthcare system and expatriates' situation in the KSA.Expatriates are primarily in the country for work purposes and are obligated to be medically fit in order to have a work visa.Also, since employers are required to provide health insurances for their foreign employees, most of them attain health services from private hospitals.This is despite the fact that free healthcare is provided to all citizens regardless of nationality in MoH facilities especially during the COVID-19 pandemic (28).
Regional variations in referral patterns likely stem from differences in healthcare resources and infrastructure.The uneven distribution of healthcare services across and within regions is well-established (29)(30)(31)(32)(33). Disparities between the five new business units in Saudi Arabia have been noted in prior studies on quality indicators for COVID-19 patients (15,34,35).Our analysis provides further evidence of disproportionality amongst the BUs regarding referral initiation and receipt.
The Eastern region and Makkah initiated the highest total number of referral requests.Contributing factors could include their greater population density (25) and regional health system capacities.
However, when examining referral request rates per 10,000 population, the Northern and Albaha regions were actually highest (Supplementary Table 1).This suggests medical resource limitations, also reflected when grouped into their respective Northern and Western BUs.Conversely, the Eastern region had the second lowest rate per capita despite having the most total referrals, highlighting the need to consider population density in resource allocation.
Riyadh and Makkah had the most referral requests.As the country's major healthcare hubs with advanced facilities and specialties (36), these regions likely attract more referrals due to advanced medical capabilities.Similar regional differences have been observed elsewhere globally (26).Further research into the distribution of health system resources, such as workforce and facilities, is needed to fully explain the variations in medical referrals across Saudi Arabia's regions.
Internal medicine emerged as the most commonly referring speciality highlighting the prevalence of chronic diseases and cardiovascular conditions in the population (37).Surgical related specialities followed which may be due to shortage of surgical staff, particularly in surgical sub-specialities.Patients and referring physicians often prefer and trust specialised centres, further driving the demand for surgical services (38).Comparatively, in Canada, dermatology was one of the top referred specialties, whereas in the KSA, dermatology related referrals were low (39).
One notable finding is the high number of referrals due to unavailable subspecialties, which is particularly high in the Central and Eastern regions.This may be an indication of a shortage of certain sub-specialities, where despite the fact that these areas are home to two of the main cities of the country namely Riyadh and Dammam, which both include excellent healthcare services and high quality of care, these cities are surrounded by smaller towns with hospitals equipped with lesser specialised staff referring to those main cities.It may also indicate that healthcare staff within those areas pursue a high-quality of care (40).Also, the Northern region stands out with a significant proportion of referrals attributed to the unavailability of a specialised physician, indicating a potential need for improved access to specialised care in that region.In contrast, the Southern region reports a higher frequency of referrals due to the unavailability of a machine and bed, indicating infrastructure-related challenges.Additionally, it is worth noting that referrals due to injuries were exclusively reported in the Southern region.This suggests that the Southern region may have a higher prevalence of injury-related incidents or a greater need for specialised care for injuries compared to other regions.The Western region shows a higher occurrence of referrals due to social reasons, potentially reflecting the influence of social and cultural factors on healthcare-seeking behaviour.
Discrepancies in bed types across regions indicates potential differences in healthcare needs and allocation of resources.The Southern region exhibits a relatively higher number of referrals for ward beds, this may be attributed to the concentration of general hospitals or specialised facilities within that particular geographic location.The assignment of distinct bed categorisations, such as burn beds and isolation beds, may be a result of various factors, such as the prevalence of illnesses in a specific geographical region, the demand for specialised medical treatments, and the demographic attributes of the populace.
Lastly, referral rates are influenced by national and international incidents.In 2020, COVID-19 pandemic and its consequences including the lockdown could explain the low referral rate in 2020 compared to 2021.The drop in referrals during the pandemic has also been seen in different settings such as emergency departments, and heart diseases in countries including Italy and the United Kingdom (41,42).However, this is the first study to observe the influence on a national level.With the implementation of the new regional healthcare transformations under the 2030 Vision and the merging of the existing 13 regions into five BUs; this research promises to encourage greater Current findings show a momentary view of national referral patterns during a two-year time frame.These results provide an opportunity for improvement in terms of equity in resource allocation as well as enhancement of specialised care especially in problematic areas highlighted here.Furthermore, the use of this nationwide secondary data enabled us to explore the patterns of e-referrals across the country.However, certain limitations should be addressed.The reliance on secondary data obtained from the e-referral system limits the scope of variables examined.Also, the absence of similar studies in the wider literature makes direct comparisons challenging.Additionally, the study's focus on referral patterns may overlook other important aspects of healthcare, such as primary care utilisation or patient outcomes.Future research should address these limitations to provide a more comprehensive understanding of healthcare utilisation and effectiveness.
Additionally, several key insights for healthcare systems worldwide can be drawn from the evolution of SMARC.First, a unified e-health platform not only enhances service quality and efficiency but also improves access, conserves resources and eliminates service redundancies.Second, centralised tracking allows effective monitoring of health outcomes and resource utilisation, which aids in the identification of strengths and weaknesses within the system.Finally, this integrated approach increases strategic resource distribution, informs health policy and advances academic research, leading to greater optimization of healthcare delivery (43,44).Investments in e-health and digital health provide economic benefits as well through streamlining operations and reducing administrative burdens, both of which are achieved by automating processes and minimising the need for in-person consultations.Digital health technologies improve diagnostic and treatment accuracy, improving patient outcomes and reducing medical errors; and extend service reach, particularly in underserved areas, maximising resource utilisation.These benefits contribute to an overall improvement in the efficiency of the healthcare system, ultimately leading to lower costs over the longterm (43,44).
Conclusion
This study examined the underlying mechanism of an important telehealth tool, namely, e-referrals.Certain patterns were observed which included higher referrals for males, as well as in internal medicine and surgical related specialities, and unavailable subspeciality being the most commonly reported reason for referrals.We recommend the use of population density in the future planning of resource allocation and specialised care.
FIGURE 2
FIGURE 2Pattern of monthly e-referrals for the years 2020 and 2021 across the Kingdom of Saudi Arabia.
FIGURE 4
FIGURE 4Total e-referral requests sent according to the 13 administrative areas of the Kingdom of Saudi Arabia.
FIGURE 5
FIGURE 5Total e-referral requests received according to the 13 administrative areas of the Kingdom of Saudi Arabia.
Both the MoH and Imam Abdulrahman Bin Faisal University institutional review boards have approved this study (23-77-E) and (IRB-2023-01-305).Standard precautions were taken to protect the confidentiality and privacy of patients' data involved.
TABLE 1
Sociodemographic characteristics of patients with referral requests.
TABLE 2
Associations between sociodemographic variables and region of referral request.
|
2024-06-29T15:17:53.611Z
|
2024-06-27T00:00:00.000
|
{
"year": 2024,
"sha1": "0d2cd63b826ebab5ecce30a59107b2fc6cde4e1a",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.3389/fmed.2024.1348442",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "c4bfe3949eb1dd3f33d35cc6fd1a5c2b3448de07",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
246361235
|
pes2o/s2orc
|
v3-fos-license
|
Poor Respiratory Health Following Relapsing SARS-CoV-2 Infection in Children with Cystic Fibrosis
Children with cystic fibrosis (CF) constitute a high-risk group for COVID-19 with underlying chronic lung disease. COVID-19 severity varying from mild infection to need of intensive care has been described in children with CF. Two children with significant underlying pulmonary morbidity are described here, who developed severe disease following SARS-CoV-2 infection. Case 1 (a 9-y-old boy) had pneumonia with respiratory failure requiring noninvasive ventilation support. He had delayed clearance of SARS-CoV-2, with recurrence of symptomatic disease with short asymptomatic period in between. He was also diagnosed with CF-related diabetes and allergic bronchopulmonary aspergillosis during the second episode. Case 2 (an 18-mo-old boy) had two episodes of SARS-CoV-2–related severe lower respiratory infection within a period of 2 mo, requiring high-flow nasal oxygen support. Both children had 3rd pulmonary exacerbation but SARS-CoV-2 was not detected in respiratory secretions. To conclude, children with CF with underlying pulmonary morbidity, can develop severe COVID-19 and prolonged SARS-CoV-2 shedding. Supplementary Information The online version contains supplementary material available at 10.1007/s12098-021-04057-0.
Introduction
Children with cystic fibrosis (CF)-related chronic lung disease are predisposed to severe pulmonary infections with severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2). Though initial reports demonstrated asymptomatic and mild SARS-CoV-2 infection in children with CF, larger series have described all degrees of severity including intensive care admission especially in post-transplant patients [1][2][3]. Children with CF from developing countries frequently have advanced lung disease [4]. The authors' unit runs a weekly CF clinic with nearly 200 children in regular follow-up cared by clinical specialists, physiotherapist, nutritionist, and specialist nurse. Presentation of two children with CF with COVID-19 are described here.
Case 1
A 9-y-old boy (homozygous F508del, Staphylococcus aureus colonization, BMI -2.18 z-score, CF-related liver disease, on regular azithromycin) had no pulmonary exacerbation in the last 11 mo. He developed fever, increased cough, and fast breathing for 3 d. He received intravenous piperacillin-tazobactam and amikacin for 14 d, but did not improve. At presentation to the authors' hospital, respiratory specimen was positive for SARS-CoV-2 reverse-transcriptase polymerase chain reaction (RT-PCR). He required noninvasive ventilation for 2 d, followed by high-flow nasal canula (HFNC) support. Inflammatory markers showed no significant abnormalities (Supplementary Table S1). Chest radiograph showed hyperinflation, mucous plugging, and interstitial shadows (Fig. 1a). As per the institutional protocol, he received remdesivir (5 d), hydrocortisone, and cefoperazone-sulbactam and vancomycin for 14 d. Child became afebrile on day 8 of admission. He was discharged after 14 d while still SARS-CoV-2 RT-PCR positive.
Within 4 d of discharge, he again developed intermittent fever spikes and increased cough. But he was brought to the authors'center with breathing difficulty and persistent fever after 20 d. His SARS-CoV-2 RT-PCR was still positive. He received oxygen by facemask and intravenous antibiotics. Sputum culture grew E. coli. He was also diagnosed with CF-related diabetes (CFRD) and allergic bronchopulmonary aspergillosis during this admission (Supplementary Table S1). SARS-CoV-2 RT-PCR turned negative by day 10 of hospital stay.
He had another pulmonary exacerbation at 3 mo, but this time, SARS-CoV-2 RT-PCR was negative. Sputum cultures grew E. coli, Klebsiella pneumoniae,and methicillin-resistant Staphylococcus aureus. He was discharged after 20 d of treatment.
Case 2
An 18-mo-old boy (diagnosed by elevated sweat chloride, heterozygous c.1333delA, BMI -0.90 z score, on regular azithromycin) had last pulmonary exacerbation 7 mo back. He had increased cough for 3 d, and fever, fast breathing, and desaturations for 1 d. He had hypoxia with bilateral crepitations and wheeze. His SARS-CoV-2 RT-PCR was positive. Inflammatory markers were not elevated (Supplementary Table S1). Chest radiograph showed interstitial infiltrates (Fig. 1b). He received HFNC support for 4 d and oxygen was gradually tapered. He received meropenem, vancomycin, and prednisolone (1 mg/kg/d). He was discharged after 14 d with negative SARS-CoV-2 RT-PCR.
Thirty-four days following discharge, the child received antibiotics for 10 d for pulmonary exacerbation locally.While on treatment, he again developed fever, worsening cough, fast breathing, and wheezing for 3 d. He was admitted to the authors'center; respiratory sample was positive for SARS-CoV-2. He was treated with intravenous antibiotics, HFNC (20 L/min), and prednisolone. He was discharged after 14 d with negative SARS-CoV-2 RT-PCR. There was no fresh SARS-CoV-2 infection in the house. IgG against SARS-CoV-2 was significantly elevated. After 3 mo from the first episode, he again had increased cough, fast breathing, and wheezing, which was managed in the outpatient setting with oral antibiotics and a short course of prednisolone.
Discussion
Two children with CF with severe SARS-CoV-2 infection have been discussed. While the first case had persistent infection, the second might have had recurrent infection due to SARS-CoV-2. Both had frequent pulmonary exacerbations following initial COVID-19. In published literature, a significant proportion of CF patients require hospitalization (29%) when they develop COVID-19 as compared to overall pediatric age group (0.1%-2.2%) [1,5]. Oxygen requirement was reported in 6% and 29% in two studies [1,2]. Both the cases had severe disease.
Prolonged shedding of SARS-CoV-2 up to 120 d has been described in immunocompromised adults [6]. The first case described here had SARS-CoV-2 RT-PCR positive till 55 d of onset of symptoms. The second case had one negative SARS-CoV-2 RT-PCR in between, so it is difficult to conclude if the second episode was re-infection or persistent infection. CF-related innate-immune dysfunction, use of steroids, and new-onset diabetes could cause prolonged viral shedding [7,8].
Viral infection in children with CF can cause pulmonary exacerbations due to direct infection, as well as secondary bacterial infection [9]. Both of the above-mentioned cases had poor respiratory health with pulmonary exacerbations after SARS-CoV-2 infection, despite having wellcontrolled disease in the previous year. Case 1 also had new onset CFRD. New-onset type 1 diabetes has been seen with COVID-19; the presence of SARS-CoV-2 receptors on pancreatic exocrine cells, use of steroids, and infection-related stress response have been the postulated mechanisms [10].
Conclusion
Children with CF having significant lung morbidity can develop severe infection and repeated pulmonary exacerbations. Strict preventive measures should be taken to protect children with CF from SARS-CoV-2.
|
2022-01-29T06:17:20.021Z
|
2022-01-28T00:00:00.000
|
{
"year": 2022,
"sha1": "33d879c32390c67b1938865dbc40b96fe5a45481",
"oa_license": null,
"oa_url": "https://link.springer.com/content/pdf/10.1007/s12098-021-04057-0.pdf",
"oa_status": "BRONZE",
"pdf_src": "PubMedCentral",
"pdf_hash": "a59668bbe0cb5a0729897abee0f18dab8958c0f1",
"s2fieldsofstudy": [
"Environmental Science",
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
264915665
|
pes2o/s2orc
|
v3-fos-license
|
Intra-cluster correlation coefficients in adults with diabetes in primary care practices: the Vermont Diabetes Information System field survey
Background Proper estimation of sample size requirements for cluster-based studies requires estimates of the intra-cluster correlation coefficient (ICC) for the variables of interest. Methods We calculated the ICC for 112 variables measured as part of the Vermont Diabetes Information System, a cluster-randomized study of adults with diabetes from 73 primary care practices (the clusters) in Vermont and surrounding areas. Results ICCs varied widely around a median value of 0.0185 (Inter-quartile range: 0.006, 0.037). Some characteristics (such as the proportion having a recent creatinine measurement) were highly associated with the practice (ICC = 0.288), while others (prevalence of some comorbidities and complications and certain aspects of quality of life) varied much more across patients with only small correlation within practices (ICC<0.001). Conclusion The ICC values reported here may be useful in designing future studies that use clustered sampling from primary care practices.
Background
Multi-level or clustered sampling designs are increasingly deployed in medical and health care surveys. In these designs, clusters are identified (e.g. medical practices) and then subjects (e.g. patients) are sampled from each cluster. The analysis and sample size estimation for such designs must take the clustering into account or the resultant significance tests (P values) and confidence intervals will be in error [1]. Generally, failure to account for clustering leads to nominal confidence intervals that are too narrow and to P values that are too small. To the extent that patient characteristics are independent of cluster, the effective sample size will be close to the number of individual subjects studied. If the subject characteristics are highly associated within clusters, the effective sample size approaches the number of clusters. In the extreme case, if all the subjects within a cluster are identical, there is no advantage to measuring more than one subject per cluster.
To estimate statistical power or required sample size in a study based on simple random sampling or allocation, one requires an estimate of the minimal important effect and (for continuous measures) the standard deviation of the outcome in the population studied. For clustered designs, however, one must also inflate the sample size to account for the clustering effect. The design effect, sometimes referred to as the variance inflation factor, is a function of the extent of correlation within clusters, the intraclass (or intra-cluster) correlation coefficient (ICC). Unfortunately, pre-study estimates of ICC are difficult to come by and obtaining them constitutes "the main difficulty in calculating sample size for cluster randomized studies" [2].
Several groups have published estimates of ICCs for various patient characteristics observed in large surveys of patients clustered within primary care or general practices from around the world [3][4][5][6]. Here we expand their estimates to include those derived from a survey of adults with diabetes clustered within primary care practices in the northeast United States.
Methods
This study was part of a larger project, the Vermont Diabetes Information System (VDIS), a cluster-randomized trial of a laboratory-based diabetes decision support system in a region-wide sample of 8808 adults with diabetes from 73 Primary Care practices in Vermont and nearby parts of the United States [7]. Primary care in these predominantly rural practices is provided by General Internists, Family Physicians, Physician Assistants, and Nurse Practitioners who provide the bulk of long-term care for these and other patients. There are few diabetes specialists in the region and most diabetes care is provided in the practices. All 119 eligible primary care practices near the thirteen participating hospitals were invited to participate [7]. The participating practices range in size from one provider (in 41 practices) to two practices with six providers each.
A field survey targeted at a sub-sample of subjects was designed to provide a better understanding of the nonlaboratory features of the patients before intervention. Field survey subjects were selected at random from the patients participating in the VDIS and invited by telephone to participate in an in-home interview. Patient names were randomly sorted and patients contacted until a sample of approximately 15% of the patients from each practice agreed to an interview. We attempted to contact 4,209 patients and reached 1,576 (37%). Of these, 1,006 (64%) agreed to be interviewed.
Subjects who agreed were mailed a questionnaire and were scheduled for an interview by a trained field interviewer. During the visit, the interviewer reviewed any missing or ambiguous questionnaire items. If necessary, the interviewer read the questions aloud for subjects and recorded their responses for them. Then the interviewer measured the subject as described below and administered a few more instruments that were not included in the questionnaire. The interviews took place during the baseline phase of the study before any interventions were in place. All subjects provided written informed consent.
The protocol was approved by the institutional review board of the University of Vermont.
Demographic, social and economic characteristics
Income was recorded in seven ordered self-reported categories from less than US$15,000 per year to US$100,000 per year or more. Education was also recorded as the highest level completed in seven categories from "Less than 9th Grade" to "Graduate or Professional Degree." We collapsed self-reported race and ethnicity into two categories: Non-Hispanic white and all others. Marital status was collapsed into two categories: Married or living as married vs. all others (single, widowed, divorced or separated). We recorded the presence or absence of four types of health insurance: private (commercial indemnity or health maintenance organization benefits often supplied by an employer), Medicare (government health coverage for the elderly and disabled), Medicaid (government health coverage for low income patients), and military (including active duty or veteran's benefits). Subjects may have more than one insurance type.
The shortest driving distance from the patients' homes to their site of care was calculated in kilometers using ArcView 3.3 by Environmental Systems Research Institute, Inc., and a geographic data set purchased from TeleAtlas, Inc. Driving distance was defined as the shortest distance along roads and highways [8].
Physical characteristics
Height was measured using a portable stadiometer (SECA, Inc.), weight with a portable scale (LB Dial Scale HAP200KD-41, Healthometer, Inc.), and blood pressure with an automated sphygmomanometer (Omron Model HEM-711). Blood pressure was obtained in the seated position in the left arm (unless contraindicated), using the cuff size recommended by the manufacturer. Three readings were obtained at 5-minute intervals and were averaged for the final result. Body mass index was calculated as weight in kilograms divided by height in meters squared.
Laboratory results
Glycosolated hemoglobin A1C was measured at 13 clinical laboratories in the patients' home communities. All laboratories used the same high-pressure liquid chromatography method with identical reference ranges. Serum creatinine, urine microalbumin-to-creatinine ratio, total cholesterol, high density lipoprotein cholesterol, and triglycerides were likewise measured by the laboratories. Low density lipoprotein cholesterol (LDL) was calculated using the Friedwald formula (LDL = Total cholesterolhigh density lipoprotein cholesterol -triglycerides/5) [9] from fasting specimens. Each patient was classified as being above or below certain laboratory value thresholds recommended by the American Diabetes Association (A1C >8%; A1C <7%; microalbumin-to-creatinine ratio <30 mg/mmol) [10]. If the LDL was 100 mg/dl or greater, or if it could not be calculated because the triglycerides were above 400 mg/dl, we categorized lipids as above goal. Tests were ordered by the primary care provider when clinically indicated. We report the most recent laboratory assays done before the home visit.
Quality of care
Where possible, we classified each subject as meeting or not meeting recommendations for care made by the American Diabetes Association [10] and the Vermont Program for Quality in Health Care [11].
Health habits
Alcohol consumption was measured by asking: "How many drinks of the following alcoholic beverages do you have in a typical week (including weekends)?
Bottles or cans of beer: _________ Glasses of wine or wine coolers: _________ Mixed drinks or shots of liquor: _________" Subjects who indicated that they do not currently drink alcohol were assigned zero to each of the three beverage categories. A summary variable representing total consumption was constructed as the sum of the three beverage-specific responses. Subjects were also asked the four CAGE screening questions [12].
Tobacco use was assessed by asking: "Have you smoked a cigarette -even one puff -during the past seven days?" Those responding "yes" were asked "How many cigarettes do you smoke on an average day?"
Self care
We assessed self-care behavior with the Summary of Diabetes Self Care Activities Measure [13]. This instrument asks the subject to record how many days in the last week they performed recommended self-care activities such as following a healthful eating plan, or participating in at least 30 minutes of physical activity. Eleven items are used to generate 5 summary scores representing the fraction of days the subject performs recommended activities related to general diet, diabetes-specific diet, exercise, blood glucose self-monitoring, and foot care. Each score ranges from 0 to 100.
Comorbidity
The Self-Administered Comorbidity Questionnaire is a modification of the widely used Charlson Index. It uses patient interview or questionnaire responses rather than chart abstraction for assessment of comorbidity and has excellent agreement with the chart-based Charlson Index [16,17]. We calculated the rate of endorsement of each of 18 specific conditions as well the number of conditions endorsed. We also calculated a score with one point if the condition is endorsed and additional points if the subject reports currently receiving treatment for it, or if it limits activities. Each condition may, therefore, contribute 0 to 3 points for a possible maximum of 54 points. One of the conditions, "eye, nerve, or kidney damage due to diabetes" may be considered a complication of diabetes rather than strictly a comorbidity.
Functional status and depression
The Medical Outcomes Trust SF-12 Health Survey is a widely used, validated instrument for assessment of general (rather than disease-specific) functional status [18]. Two summary scales are calculated: the Physical Component Summary and the Mental Component Summary. The Patient Health Questionnaire-9 is a brief self-report instrument that quantifies the presence and degree of mental depression [19].
Complications
We assessed the presence of diabetes complications by asking six questions. The responses were "Yes," "No," and "Don't know." 1. Have you ever had an ulcer or sore on your leg or foot that took more than 4 weeks to heal?
Has your doctor or health care provider ever told you that you have these problems: 2. Problems with vision or retinopathy related to your diabetes?
3. Pain, burning, or numbness in the feet or legs related to your diabetes?
4. Problems with stomach emptying related to your diabetes?
5. Problems with sexual function?" 6. Problems with your kidneys related to your diabetes?
Medications
The subjects were asked to produce "all medications you have used in the past month including prescriptions, overthe-counter products, vitamins, and herbs." The field assistant recorded the name, strength, dose, route, and frequency of each preparation.
Quality of life
The Audit of Diabetes-Dependant Quality of Life is an 18item questionnaire regarding the impact of diabetes on specific aspects of a person's life with patient weighting of the impact of each domain [20,21]. We employed 17 of the 18 domains of this instrument. The scores for each domain can range from -9 (maximum negative impact of diabetes on that domain) to +9 (maximum positive impact).
Resource utilization
The survey included items asking the subjects to record whether they had used various services in the last year: Endocrinologist, Dietician, Podiatrist, Diabetes Educator, Ophthalmologist, and Diabetes Class. Those answering "Don't Know" were assigned a value of "No." It also prompted subjects to report the number of Emergency Room visits they had in the last year and "In the past month, how many times have you been to a doctor or health care professional?"
Statistical analyses
In the random effects model, the ICC is the proportion of the total variance that is between clusters (practices).
where is the between-cluster component of variance while is the within-cluster component. If a measurement varies across patients without regard to which practice they are in, the ICC will be close to zero. If the value of the variable is largely a function of which practice they are in, the ICC will be close to 1.0 [2]. We used the analysis of variance estimator [22][23][24] provided by the "loneway" command in STATA 8.2 (Stata Corp., College Station, Texas). This estimator uses the F statistic to calculate the ICC for N total subjects in k groups of size N o : where Further, is the asymptomatic standard error of the ICC, and the 100(1-α)% confidence interval is: For each characteristic, we recorded the sample size (N), the sample size per cluster (N o ) the mean (or proportion for dichotomous variables), the standard deviation (SD) for continuous variables, the standard error of the mean (or percentage) adjusted for clustering within practices (SE), the ICC, and the 95% confidence interval of the ICC.
We assessed the association between the value (reported proportion) of binary variables and the ICC [25] with Spearman's non-parametric correlation coefficient. For proportions greater than 0.5, we used the complement of the proportion so that all proportions for this analysis were less than 0.5. To compare groups of ICCs, we used the two-sample Wilcoxon rank-sum (Mann-Whitney) test.
Results
The results appear in The ICCs for the 62 binary variables were significantly associated with their proportions (Spearman's correlation coefficient = 0.53; P < 0.0001) to a degree sometimes classified as "Large" [26].
Discussion
These data provide estimates of intra-cluster correlations for 112 patient characteristics relevant to the analysis of = Number per cluster; SD = standard deviation of mean; SE = standard error adjusted for clustering; ICC = Intra-cluster (intra-practice) correlation coefficient. *In 39 cases in which LDL could not be estimated because triglycerides were over 400 mg/dl, LDL was assumed to be over 100 mg/dl. † Because only a few subjects have this condition, the ICC may not be estimated consistently. §The method used truncates the estimate of ICC (and its CI) at 0 for some values that may be negative. adults with diabetes receiving primary care in Vermont and nearby regions of the United States. They may usefully be applied to the design and sample size estimation of future surveys that are clustered on primary care practices.
In the design of clustered-based studies, the ICC may be used to calculate the design effect which is the degree to which the sample size must be inflated above that of a simple random sample to account for the loss of information inherent in the clustered design. If the average number of subjects sampled per cluster is m, the design effect is given by: If m or ICC is large, the total number of individual subjects needed may be substantially greater than suggested by a sample size calculation that is not adjusted for clustering. Alternatively, if both m and ICC are small, the design effect may be very close to 1.0 indicating that the clustered design does not inflate the sample size.
In the VDIS, the cost of enrolling subjects for laboratory data collection within a cluster, once the practice was enrolled, was relatively low. Therefore, large values of m (120.7 subjects per practice on average) were not problematic. However, the cost per patient of obtaining interview data was relatively high. Therefore, we reduced the mean sample size per cluster to 14.5 by random sampling within practice. The design effects experienced in VDIS are not representative of those faced in other designs unless they happen to have the same mean cluster sample size as VDIS (which is extremely unlikely). Therefore, unlike previous publications, we have elected not to report design effects. Study designers should use the ICCs and their own estimates of m to understand their own design effects.
Campbell et al [27] suggest that ICCs are higher for process measures than for outcomes. We see evidence for this in that the eight quality of care variables (process measures, see Table 1) have a median ICC of 0.088 (IQR 0.049, 0.181) while the ten laboratory outcomes measures have a median ICC of 0.038 (IQR 0.029, 0.055). This difference is significant by Wilcoxon rank-sum test with P = 0.013. We note that the nine physical characteristics of the subjects, presumably under less control of the provider than either laboratory results or even process measures, have a median ICC of 0.017 (IQR 0.011, 0.028) and are significantly different than the laboratory measures (P = 0.034).
Within practice correlation was most prominent for process measures associated with quality of care. The likelihood of receiving a creatinine measurement on time had the highest ICC (0.288) with other quality of care meas-ures also among the most highly correlated. This may represent that process measures are heavily influenced by the practice style of the practitioners and any office-based procedures (reminders, registries, flow sheets, etc.) that only some practices employ. In a similar vein, physiologic control of some aspects of diabetes (especially achievement of tight control of A1C and LDL) appears to vary importantly across practices with ICCs of 0.046 for A1C below 7% and 0.029 for LDL below 100 mg/dl. Some demographic aspects of the population were correlated within practices: age, sex, and especially travel burden. As patients tend to stay with their primary provider as they age, some practices accumulate older patients (ICC = 0.077). Some patients express a preference for same-sex providers, with more women visiting practices that have female providers. This may account for the relatively high ICC for sex (0.038). Travel distance may be related to the geographic location of the practice office. Practices in more densely populated areas may tend to have lower typical travel burdens.
With the possible exception of blood pressure (which may be under the control of the providers to some degree), the physical characteristics and health habits of patients vary little across practices.
Although apparently under the control of the providers, the utilization of health care services had generally small ICCs. The exception was consultation by an Endocrinologist with an ICC of 0.089. This may reflect the geographic proximity of an endocrinologist to some of the practices.
The aspects of diabetes that are most directly experienced by patients (complications, quality of life, functional status, comorbidity, and self-care) vary little across practices. It does not appear that some primary care practices tend to accumulate more complicated or difficult patients than others. Likewise, low health literacy is a substantial, and perhaps unrecognized, problem for all practices, with little clustering within practices.
These data demonstrate a large correlation between the proportion of the 62 binary variables and their ICCs. This finding has been noted by others [25] and may be useful in estimating an ICC for sample size calculations.
For many of these variables, the impact of within-cluster correlation on sample size requirements appears to be relatively small. Thirty-three variables had an ICC <0.010 with a design effect less than 1.14 indicating that the VDIS clustered design required an increase in sample size of 14% compared to a non-clustered study. However, depending on the number of clusters and the number of subjects per cluster, even small ICCs can result in the need for costly increases in overall sample size.
These estimates of ICC are not useful for studies that cluster on factors other than practice (such as community, hospital, individual provider, family, classroom, etc.). The VDIS study population was drawn from Vermont and nearby New York and New Hampshire and is, therefore, predominantly white and rural. All the subjects were under care for diabetes. Although some have Type 1 diabetes, this older, overweight population is largely comprised of patients with Type 2 diabetes. The practices in the VDIS are small with a median of 2 providers per practice. For these reasons, generalization to other populations and settings may be problematic.
Several recent reports provide some comparisons from other settings ( Table 2). A study of British patients aged 75 and older reported intra-practice correlations from 106 general practices [4]. A study from Rhode Island and nearby Massachusetts enrolled 15 primary care practices [5]. Several surveys of general practices from Australia and New Zealand provided a few ICCs comparable to VDIS [3,6]. The ICCs for most of these variables vary substantially across the studies. For instance, the ICCs for weight and body mass index varied between 0.011 and 0.081. Differences in the practice structures, referral patterns, social and geographic factors and practice patterns may explain these differences. Only recently have determinants of ICC begun to be studied [26][27][28]. We suggest that more catalogues of ICCs, drawn from a variety of settings, will be useful both to investigators designing new clustered studies and to researchers investigating the role of setting on patient characteristics.
Conclusion
Intra-practice correlation coefficients in this survey of adults receiving care for diabetes varied widely. Some characteristics (such as the likelihood of having a recent creatinine measurement) were highly associated with the practice (ICC = 0.288), while others (prevalence of some comorbidities and complications and certain aspects of quality of life) varied much more across patients with virtually no correlation within practices (ICC<0.001). The values reported here may be useful in designing future clustered studies.
|
2014-10-01T00:00:00.000Z
|
2006-01-01T00:00:00.000
|
{
"year": 2006,
"sha1": "df710e5912d7d79e183e16f31009da1d5a5e0aed",
"oa_license": "CCBY",
"oa_url": "https://bmcmedresmethodol.biomedcentral.com/counter/pdf/10.1186/1471-2288-6-20",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "df710e5912d7d79e183e16f31009da1d5a5e0aed",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": []
}
|
118659554
|
pes2o/s2orc
|
v3-fos-license
|
A compact entanglement distillery
Large-scale quantum-correlated networks could transform technologies ranging from communications and cryptography to computation, metrology, and simulation of novel materials. Critical to achieving such quantum enhancements is distributing high-quality entanglement between distant nodes. This is made possible in the unavoidable presence of decoherence by entanglement distillation. However, current versions of this protocol are prohibitively costly in terms of resources. We introduce a new scheme for continuous-variable entanglement distillation that requires only linear temporal and constant physical or spatial resources, both of which are exponential improvements over existing protocols. Our scheme uses a fixed module - an entanglement distillery - comprising only four quantum memories of at most 50 % storage efficiency and allowing a feasible experimental implementation. Tangible quantum advantages are obtained by using non-ideal quantum memories outside their conventional role of storage. By creating, storing and processing information in the same physical space, the scheme establishes a potentially valuable technique for designing stable, scalable protocols across different quantum technologies.
Using the same physical space for storing and processing information is an attractive paradigm for material and information science [1]. It will make devices smaller, and increase robustness and speed by dispensing with the necessity of moving information around. Such a development is particularly promising for quantum information science, where stored quantum states awaiting processing are extremely vulnerable to noise [2,3], and shuttling information-bearing quantum states around a quantum device presents formidable challenges. Although enormous progress is being made in manipulating quantum information in a variety of systems and scenarios [4][5][6][7], theoretical proposals are necessary that minimize the demands on such manipulation. This is vital if the resources and advantages offered by quantum physics, in applications ranging from communications to metrology [8,9], are to be harnessed. Preparing distributed quantum states with high quality quantum entanglement lies at the heart of this endeavour. Unfortunately, noise and decoherence inevitably deteriorate the quality of entanglement in states that are shared between distant parties [10]. The situation can be remedied by entanglement distillation, in which multiple copies of a less entangled state are transformed by local operations and classical communications (LOCC) into fewer copies of a more entangled state [11].
We introduce the notion of a quantum entanglement distillery. For continuous variable (CV) entanglement, it provides improved use of fixed resources to achieve the same levels of improvement in entanglement as earlier schemes. In fact, our distillery has doubly-exponential temporal and exponential spatial advantage over existing * animesh.datta@physics.ox.ac.uk distillation schemes [12,13]. It also surpasses crucial limitations of finite-dimensional entanglement pumping schemes [14,15]. In particular, failed local operations merely reduce the entanglement of the states involved as opposed to a finite-dimensional instance where failed attempts lead to completely unentangled states. Ours is a repeat-until-success scheme using a linear number of initially poorly entangled states to obtain a final state with higher entanglement. The distillery is a fixed module, consisting of only four quantum memories where the final amount of entanglement is governed by the initial states and the number of iterations. Memories [16] allow us to store results from previous iterations while the subsequent ones succeed, providing an exponential advantage in time. The additional exponential advantage in space and time is provided by our entanglement distillation protocol which we describe later.
The quantum memories not only store quantum information, but also process it concurrently in the same physical space. They allow us to repeatedly perform probabilistic operations on the same copy of the quantum state, further saving time and enhancing resilience against decoherence. This is vital as all local schemes for distilling entanglement must be probabilistic, since entanglement cannot, on average, increase under LOCC. Furthermore, distillation of CV entanglement is not possible if all states and operations involved are Gaussian [17][18][19]. This can be circumvented by performing probabilistic non-Gaussian operations on initially Gaussian states [12,13,[17][18][19][20]. A major advantage over finitedimensional schemes is that failed local operations do not require starting the whole process anew. Finally, our scheme is also event-ready, in that the protocol's success is reported by fixed detector outcomes.
There are two key features the quantum memories must possess to be suitable for our distillery. Firstly, their FIG. 1. Mashing. (a) A linear-optics schematic of the iterative mashing protocol, in which an entangled resource state |ψ 0 , distributed between two parties Alice and Bob, is interfered with the shared entangled state |ψ i on 50:50 beamsplitters. Detection of vacuum by Alice and Bob heralds the success of the protocol, which produces a more entangled state |ψ i+1 . (b) Implementation of mashing using four quantum memories. The resource state is generated between memories A2 and B2, while the state |ψ i is shared between memories A1 and B1. The grey panels show the control pulses required to drive the memory interactions: full retrieval (red) and 50:50 beamsplitter (green). Mashing is achieved by retrieving the resource state from A2, B2, and sending it through memories A1, B1 while driving a 50:50 beamsplitter interaction, as described in the main text. Vacuum detections herald the successful production of |ψ i+1 between A1 and B1 (not shown).
time-bandwidth product, which determines the number of iterations that can be executed within its coherence lifetime, must be sufficiently large. Secondly, they need a transparent failure mode, i.e. they transmit any unstored excitation, allowing them to be used as a beamsplitter, and enabling in-situ generation of the initial two-mode squeezed state. Most importantly, however, at no point do we require a perfect memory.
We begin by describing our protocol, consisting of two major steps we call malting and mashing. We first describe mashing, which provides iterative improvement of a weak entanglement resource, and then the design of our quantum entanglement distillery that implements our scheme. Then we describe malting, the generation of a weak entangled resource, and analyze its success probability. Finally, we show how realistic memories allow us sufficiently many iterations to get very close to the limiting state.
FIG. 2. Malting using quantum memories. The grey panels show the pulses required to drive the memory interactions: squeezing (blue), 50:50 beamsplitter (green), weak beamsplitter (purple). (a) A squeezing interaction is driven in both memories A and B, which each emit photons entangled with spinwave excitations. The propagating photons from A are directed into B and vice versa, while a 50:50 beamsplitter interaction is driven, causing the photons and spin-wave excitations to interfere. Vacuum detections herald the generation of a two-mode squeezed state shared by A and B. (b) De-Gaussification is achieved by photon subtraction at A and B. Because the states are stored in the memories, a repeat-until-success strategy can be employed. Weak control pulses drive a series of beamsplitter interactions with small effective reflectivities. When a retrieved photon is detected at both A and B, the malting process is over and the non-Gaussian entangled resources state |ψ 0 has been successfully generated.
MASHING
The mashing step of entanglement distillation begins with a non-Gaussian resource state |ψ 0 . This state is produced in a process called malting, which we describe later. Any pure, bipartite state is locally equivalent to a state in Schmidt form as where |n A1 (|n B1 ) denotes an n-photon Fock state in Alice's (Bob's) mode. This is the resource at the end of malting stage from which we will distill our final state using an iterative protocol. In the first step of the iteration, Alice and Bob combine two copies of the state |ψ 0 on two 50/50 beamsplitters. In the case that each party detects vacuum on one of the emerging modes from each beamsplitter, the resultant state in the other two modes is |ψ 1 . Next, |ψ 1 is interfered with a fresh copy of |ψ 0 to produce |ψ 2 upon vacuum detection, and so on. At stage i of the protocol, we combine |ψ i with |ψ 0 mode-wise on the beamsplitters, and detect vacuum, as in Fig. 1 (a), to produce the state where U ab represents a 50/50 beamsplitter across modes a and b. If we denote we obtain, from Eq. (2), an iterative relation of the form where M depends on |ψ 0 . The convergence of this map is proven in [21]. Its behavior when different outcomes are detected in the mashing step, and its performance under decoherence is also discussed in [21]. A key innovation of our work is that we implement both mashing and malting steps in the same hardware. We call this hardware an entanglement distillery, and present a design for it which uses only four quantum memories.
AN ENTANGLEMENT DISTILLERY
Our distillery consists of four quantum memories to store entangled states during the protocol, and act as nonlinear and linear elements for generating and processing them. In fact, four photodetectors are the only other elements required to implement the distillery.
A quantum memory typically involves three modes, the input, the control, and a localised storage mode. The writing operation uses the control pulse to write the input into the storage mode, and the reading operation uses the control pulse to do the reverse. The storage mode b is generally a matter degree of freedom, while the other two are optical. The simplest interaction for transferring a single excitation amongst three modes is of the form The beamsplitter required for the mashing step in Fig. 1 (b) can be readily achieved by setting the field a to be classical. Note that one of the modes involved is optical, while the other is material. This allows us to exploit the best of both worlds: the optical modes for transferring information across the distillery, and the material modes for processing it. Once copies of |ψ 0 are malted between memories A 1 and B 1 , and between A 2 and B 2 , the matter modes in A 2 and B 2 are converted entirely into optical modes using strong control pulses. On Alice's side, the photons retrieved from A 2 are directed into A 1 , and interfered with the matter mode via a 50/50 beamsplitter interaction, as in Fig. 1 (b). Correspondingly on Bob's side, the photons retrieved from B 2 are interfered with the matter mode in B 1 . In the case that no photons are detected emerging from the ensembles, the state shared by Alice and Bob is projected into |ψ 1 . The second iteration proceeds by malting another copy of |ψ 0 between A 2 and B 2 , and interfering this with the matter modes in A 1 and B 1 as described above, which produces |ψ 2 provided both Alice and Bob detect vacuum again. Subsequent iterations are performed in the same way, by malting successive copies of |ψ 0 and mashing them into the state held in A 1 B 1 . We note that this stage of the protocol requires a beamsplitter interaction with T = 0 -that is, 100% retrieval efficiency from the memories A 2 , B 2 . The Raman scheme studied later makes this technically feasible [21].
Setting the mode c in Eq. (5) to be classical leads to a two-mode squeezing Hamiltonian, once again between an optical and material mode. This forms the first step of the malting process.
MALTING
Malting is a two-step process, depicted in Figs. 2 (a) and (b). The first step is a two-mode squeezing interaction in memories A and B, held by Alice and Bob, denoted by the blue pulse in Fig. 2 (a). These generate a pair of two-mode squeezed states of the form |Φ = √ 1 − λ 2 n λ n |n |n , where λ is the squeezing parameter, and the matter mode of each memory is now entangled with its corresponding optical mode. The emitted photons are then directed over the channel connecting Alice and Bob, so that Alice receives Bob's photons, and Bob receives Alice's, as in Fig. 2 (a). Each party now uses a control pulse (green pulse in Fig. 2 (a)) to drive the same 50/50 beamsplitter interaction as used in the mashing step, so that Alice's photons are interfered with Bob's matter modes and vice versa. Finally, a photon counting detector placed behind each memory measures the optical mode emerging from the beamsplitter interaction. In the case that no photons are detected, the joint state of the two memories is again a two-mode squeezed state, now between the matter modes of Alice's and Bob's memories [ Fig. 2 (a)], In order to prepare a suitable non-Gaussian resource state |ψ 0 , some non-Gaussian operation is now required.
That is the aim of the second part of the malting process. We can perform any non-Gaussian operation on one or both of the modes of the two-mode squeezed state, and the memory provides a lot of versatility, but for the rest of this paper we will concentrate on photon subtraction, which has been studied in the context of entanglement distillation previously [12,13,[22][23][24][25].
Typically, photons are subtracted from optical modes using low-reflectivity beamsplitters and photon counters. This is a probabilistic process. In the same way, phonons can be subtracted from the matter modes by sending in weak control pulses and detecting the emission of a photon at the output ( Fig. 2 (b)). The advantage of using the matter modes to process information now becomes apparent, as the subtraction process can be tried repeatedly on the same copy of the initial state. By contrast, an optical implementation requires fresh preparation of the initial state if the subtraction fails. If after several weak control pulses (green pulses in Fig. 2 (b)), a photon has been detected by both Alice and Bob, a successful subtraction on matter modes in both the memories has been heralded and our non-Gaussian resource state |ψ 0 is now ready. This completes the malting process. An example pulse sequence, as applied by Alice and Bob, is depicted in Fig. 2 (b), where they require three and five attempts, respectively, to successfully implement subtraction. This is a fundamental advantage of a memory-based CV distillery, without any counterpart in free-space, finite-dimensional distillation schemes.
Each failed detection, however, alters the state. Since the initial state in the memory is a two-mode squeezed state, the quantum state after f vacuum detections (over both arms) is still a two-mode squeezed state of the form of Eq. (6), but with a squeezing parameter of x = λT f , where T is the effective transmissivity of the beamsplitter interaction. It is related to the reflectivity R by R 2 +T 2 = 1. The larger the value of T the better. Rather conveniently, T can be made arbitrarily close to 1 simply by reducing the energy of the subtracting control pulses. If we succeed in detecting photons at the photon counters in Fig. 2 (b) after f trials, our resource state takes the (unnormalized) form with µ = xT 2 = λT f +2 . Equating this with the entanglement of the initial two-mode squeezed state allows us to find the maximum number of tries f c within which we must succeed if we are to have a net gain in entanglement. If T = 1 − η, for small η, where R is the real root of the equation . For typical parameters such as λ = 0.2 and T = 0.99, f c = 60. This means we have 60 attempts at subtracting excitations before there is no net advantage over the initial two-mode squeezed state in terms of entanglement. The dependence of f c on λ is weak, but a lower readout efficiency allows us more attempts at detecting photons.
Probability of successful malting: The probability of success for the first step ( Fig. 1 (b)), in which vacuum detections at each party herald the creation of |Φ AB , is 1 − λ 2 , assuming perfect detectors. The probability of success of a memory-based subtraction at trial f +1 after f vacuum detections is and the probability of succeeding by trial f is P f = f j=0 P j . For the initial state |ψ 0 , the results of the distillation scheme are presented in Fig. (3). As can be seen, the performance of our scheme is almost indistinguishable from that of the exponential protocol [12]. Additionally, the state converges to the limiting state after a small number of iterations.
In principle, our distillery could perform an arbitrary number of iterations, but in practice the number of iterations is limited by the finite storage lifetime t mem of the memories. The useful lifetime of a memory is captured by its time-bandwidth product B = t mem /τ , which is the number of clock cycles, as defined by the duration τ of the control pulses, that fit within the lifetime of the memory. If p s ∞ is the probability of success of mashing two copies of the limiting state, the maximum number of iterations i m satisfies [21] For λ = 0.15, T = 0.75, and B = 20000, i m = 54. Fig. (4) shows the number of iterations allowed for a broad range of parameters. For the instances we are interested in, a small number of iterations -three, for example -suffices to get close to the limiting case as shown in Fig. (3). Note that a memory with a timebandwidth product on the order of B ∼ 10 3 was recently implemented [27], while B 10 5 is feasible with modest technical modifications, such as improved magnetic shielding. We next describe a quantum memory which can be used to implement a Hamiltonian of the form Eq. .
RAMAN MEMORIES
Off-resonant Raman interactions in atomic ensembles provide a method for controllably coupling photons to non-propagating matter modes of atomic coherences known as spin-wave excitations. The archetypal system consists of a vapour of three-level atoms with a Λ-type level configuration, in which a ground state is coupled to a long-lived storage state via an excited state. Two types of interactions can be implemented [28], as shown in Fig. (5). The first is a two-mode squeezing interaction, in which the emission of a Stokes photon is accompanied by one of the atoms flipping from the ground to the storage state, producing a distributed excitation across the atomic ensemble, a spin wave. Stokes photons and spin-wave excitations are therefore produced in correlated pairs, and the interaction Hamiltonian has the form of a two-mode squeezer [29,30] where C 2 S ≈ τ dγΩ 2 /∆ 2 S is the coupling strength, with d the resonant optical depth of the ensemble and γ the homogeneous linewidth of the excited state. Here a and b are annihilation operators for the Stokes and spin wave modes respectively.
The second kind of interaction has the form of a beamsplitter between optical and material modes. If there is a spin-wave excitation, it is converted into an anti-Stokes photon. On the other hand, an incident photon in the anti-Stokes mode will be absorbed and mapped into the spin wave mode. This allows for the storage and on-demand retrieval of optical pulses [31]. More generally, if photons and spin-wave excitations are simultaneously present, they will interfere, precisely as two optical (11) in which a strong, off-resonant pump pulse with Rabi frequency Ω, detuning ∆S from resonance and duration τ drives spontaneous Stokes scattering from the ground state. (c) Here, the pump pulse is replaced by an off-resonant control pulse, with detuning ∆BS, that couples the storage state to the excited state. This results in an interaction as in Eq. (12). Note that we have associated the same operator a to both the anti-Stokes mode for the beamsplitter and the Stokes mode for the squeezer. In order to match the modes for the two types of interaction, we arrange the detunings such that ∆BS = ∆S + δ, where δ is the Stokes splitting between the ground and storage states. modes would at a beamsplitter; the energy in the control pulse determines the effective reflectivity of the beamsplitter interaction. The Hamiltonian for this interaction takes the form [32] H BS = C BS ab † + h.c., (12) where the coupling strength C BS is the same as C S , except that the detuning ∆ S is replaced with ∆ BS . The beamsplitter ratio is T ≈ 1 − C 2 BS .
DISCUSSION
The Herculean task of maintaining coherence in one part of a quantum device while another decoheres is the biggest roadblock to scalable quantum technologies. Quantum computation, communication, simulation, and metrology should all benefit immensely from the act of storing and processing quantum information is the same physical space. We have shown that protocols can be designed that cater to this requirement, and that systems exist to implement them. Quantum memories can be used not only to store, but process quantum information. They would also allow us to perform other local gaussian operations, such as squeezing [33], as well as nongaussian operations. Our quantum distillery can produce high-quality entangled states between distant parties using just four quantum memories with imperfect storage efficiencies. We also showed that convergence to the limiting state can be achieved using realistic memories with limited storage times.
We have presented a protocol for CV entanglement distillation that is exponentially more efficient in spatial and temporal resources compared with previous schemes. The most attractive feature of our protocol is the ability to continue after failed probabilistic events. This is partly due to the use of continuous variables, which are more amenable to economic use than finitedimensional systems like qubits. Critically, however, it is the localised nature of the continuous modes that really allows us to recycle the outcomes of failed attempts.
Preparation of entangled states is the starting point of most quantum information protocols. Designing protocols which minimise the experimental resource requirements will be a key step in taking quantum information science into the realm of practical realisation. Our work shows the potential for achieving this by making best creative use of existing components such as quantum memories. We hope that this will encourage a stronger focus on the development of resource-efficient quantum protocols.
I. APPENDIX
A. Convergence of the mashing step of distillation protocol The mashing step, depicted in Fig. 1 (a), is written as Denoting |ψ i = n α i n |n A1 |n B1 , for i = 0, 1, 2..., the mashing step can be expressed as Each iteration, shown in Eq. (2) and Fig. 2 (a), maps the set of coefficients {α i n } ∞ n=0 into {α i+1 n } ∞ n=0 . Hence, all the properties of the distillation scheme are encapsulated where Θ(x) is the Heaviside step function. Expressed in matrix form It is easy to see that the fixed points of this map is given by α 0 = {λ n } ∞ n=0 . Note that the iteration in Eq. (13) leads to α 1 0 = (α 1 0 ) 2 , whereby α 1 0 = 1. Also, α 1 i is fixed, say to λ all the remaining terms in the sequence are fixed. This proves that the fixed point is unique as well.
As we are ignoring normalization here, we can choose, α 0 0 = 1. Then the map has eigenvalues of the form 1/2 k , k = 0, 1, 2, · · · , independent of the actual form of the α 0 n . Let us denote the (right) eigenvectors of M by |m i . Decomposing the initial state in terms of these eigenvectors as After i iterations, we get Note that the right eigenvectors do not form an orthonormal basis but still span the space. Thus, This implies that the limiting state after an infinite number of iterations is the eigenvector corresponding to the eigenvalue 1. Following Eq. (3), let us designate We note that α i+1 From Eq. (13), α i+1 1 = 1 2 (α i 1 + α 0 1 ). Iterating this recursion leads to Also α i+1 2 = 1 4 (α i 2 + 2(α 0 1 ) 2 + α 0 2 ), which leads to and so, Now we note a couple of features of this linear distillation scheme in general, after which we will specialize to specific input states. Our distillation protocol leaves the first two coefficients of the state unchanged, and only enhances the entanglement by modulating the higher Fock layers of the initial state. Also, in this linear scheme, the limiting state is determined by not just the α 0 1 term of the initial state, but by all the terms in the initial sequence. This is the primary reason that makes the analytic derivation of the limiting state hard.
We now consider the a non-gaussian state as a resource, the candidate for our |ψ 0 . The one we consider is a two-mode squeezed state with a photon subtracted from each arm. Such a state a given by Eq. (7). For this particular input state, α 1 n = µ n (n 2 + 3n + 4)/4, (23) α 2 n = µ n (n 3 + 7n 2 + 24n + 32)/32.
It can be seen that the for this input state, α n = a n µ n , with 0 ≤ a n ≤ 1. We can then rewrite the recursion relation in terms of the a n as a n = 1 2 n n t=0 n t (t + 1)a n−t .
Its is interesting to note that outcomes different from that discussed above also lead to states with enhanced entanglement. Indeed, one could consider a mashing step as detecting a i and b i photons on Alice's and Bob's sides respectively leading to the iteration The case we have considered in the simplest to analyze, where a i = b i = 0 for all i. More general cases, involving a i = b i = 0, ∀i, as well as those involving different detection events in each iteration can also be analyzed, and adaptive schemes involving classical communication between the parties designed that lead to highly entangled distilled states. A more complete analysis of such distillation strategies is beyond the scope of this work, and will be presented elsewhere.
B. Mashing in the presence of dephasing
We discuss in brief the role of decoherence in our protocol. We envisage a scenario, where the distilled state at the end iteration i has to wait in the memories while the source state |ψ 0 is malted. The state |ψ i will suffer dephasing in that time, and thereby loose its purity. To proceed further, we need to express the mashing step for mixed states. Let Then the mashing step corresponds to This can be expressed as an iteration on individual elements of the density matrix by We consider dephasing on the state ρ phenomenologically asρ pq;rs = ρ pq;rs e − (p+q−r−s) 2 v denotes the amount of dephasing suffered by a state, and in the limit of v → ∞, the resulting state is completely diagonal, and consequently classical. In Fig. 6, we show the effect of depashing on 3 iterations of our protocol. The largest value of v that makes the process of distillation break even in the presence of noise is around v = 2. Note that in the limit of complete dephasing of one of the states in an iteration, the state after the iteration still has as much entanglement as ρ 0 . This is another advantage of the linear protocol, which is much more susceptible to decoherence due to the exponential number of states involved. In a practical scenario, this would translate to requirements on the length of time one can wait between iterations, and other physical and material parameters.
C. Maximum number of subtraction attempts
The entanglement, as quantified by the logarithmic negativity [26], in the phonon-subtracted state in Eq. (7) is where µ = λT f +2 . The logarithmic negativity of twomode squeezed state with squeezing parameter λ is Equating the two above equations allows for a solution to f = f c as where R is the real root of the equation
D. Probability of success of the malting step
As is to be expected, P f ≥ P 0 , ∀f. Denoting the logarithmic negativity after i iterations by N i , a more reasonable figure of merit is the averaged entanglement gain given by P c N fc /P 0 N 0 , and we have found this ratio to be larger than unity for a broad range of parameters, typically for T > 0.7, independent of λ. As noted earlier, a larger value of T allows for a larger value of f c . To get an idea of how a finite time-bandwidth product affects the number of iterations possible in our quantum distillery, we need to calculate the probability of success of an arbitrary iteration.
We begin by denoting the probability of success of the i th iteration as p s i . The combined probability of an entire sequence of i successful iterations is then P c i j=1 p s j , where the P c is the probability of successfully preparing |ψ 0 within f c trials as described above. Note that we are accounting for the worst-case scenario, whereby the initial state needs f c attempts to be realized. As already mentioned, our memory-based subtraction scheme increases the probability of photon subtraction. Also, i+1 copies of this state are required for i iterations of the protocol. Then the total number of memory operations is given by (i + 1)f c /P c i j=1 p s j , which must be less than B.
where for the first inequality we have used p s i ≥ p s ∞ , the probability of the step defined in Eq. (2) succeeding in the limit of an infinite number of iterations. Since the limiting state is, by definition, invariant under Eq. (2), Inequality (33) can be solved numerically.
E. Efficient readout
After malting a resource state |ψ 0 between the ensembles A 2 , B 2 , this must be mashed into the current entangled state |ψ i by retrieving it and interfering it with |ψ i in the ensembles A 1 , B 1 . This therefore assumes that we are able to retrieve a state with 100% efficiency from the ensembles A 2 , B 2 , which is to say that we can implement a beamsplitter interaction with T = 0. Fortunately, the Raman interaction allows for near perfect retrieval using multi-pulse readout. That is, a train of several control pulses is directed into the ensembles. Each pulse may only achieve partial readout, but the combined effect of all the pulses enables, asymptotically, the complete extraction of the stored excitation. The retrieved state is now distributed over several temporal modes, but it remains coherent. Interfering this state with |ψ i now requires a single T = 1/2 interaction in A 1 , B 1 . Since the incident field is delocalized over several time bins, the appropriate control field to couple it to the stored spin wave should have the form of a pulse train [32], but this is precisely what was used to drive the retrieval from A 2 , B 2 . Therefore the same train of control pulses can be re-used to mediate the interference. In fact, since only T = 1/2 is required for interference, perfect modematching is not needed, and some attenuation of the control intensity can also be accommodated. This establishes the technical feasibility of each stage of our distillation protocol. Perfect storage is never required, and where near-perfect retrieval is desirable, it can be implemented easily with a train of several pulses -the ability to interfere the resulting temporally delocalized state is also retained with this scheme.
|
2011-07-20T21:33:44.000Z
|
2011-07-20T00:00:00.000
|
{
"year": 2011,
"sha1": "970bc0b77a3ec7f8d5d09bdfc971c64049719fad",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/1107.4122",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "970bc0b77a3ec7f8d5d09bdfc971c64049719fad",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
}
|
12947970
|
pes2o/s2orc
|
v3-fos-license
|
Models for the Polarized Parton Distributions of the Nucleon
Polarized deep inelastic scattering (DIS) data are analyzed in leading and next-to-leading order of QCD within the common `standard' scenario of polarized parton distributions with a flavor-symmetric light sea (antiquark) distribution $\delta\bar{q}$, and a completely SU(3)$_f$ broken `valence' scenario with totally flavor-asymmetric light sea densities $(\delta\bar{u}\neq\delta\bar{d}\neq\delta\bar{s})$. The latter flavor-broken light sea distributions are modelled with the help of a Pauli-blocking ansatz at the low radiative/dynamical input scales of $\mu_{\rm LO(NLO)}^2=0.26$ (0.40) GeV$^2$ which complies with predictions of the chiral quark-soliton model and expectations based on the statistical parton model as well as with the corresponding, well established, flavor-broken unpolarized sea ($\bar{d}>\bar{u}$). Present semi-inclusive DIS data cannot yet uniquely discriminate between those two flavor-symmetric and flavor-broken polarized light sea scenarios.
Introduction
The polarized parton distributions of the nucleon have been intensively studied in recent years [1 -14]. The conclusion has been that the experimental data dictate a negatively polarized antiquark component, and show a tendency toward a positive polarization of gluons. Presently we possess a lot of precise data [15 -24] on the polarized structure functions of the nucleon, some of them very recent [23,24], which justify a renewed investigation of the aforementioned issue. This alone, however, does not provide the main motivation for this project, rather the improved understanding in recent years of the situation in the unpolarized parton sector [25 -28] provides important insights for the corresponding polarized parton densities. In particular, one notes that the unpolarized sea (antiquark) distributions are flavor-asymmetric (d >ū), which can be understood in terms of flavor mass asymmetries and Pauli-blocking effects [29,30]. The main objective of the present paper is to transcribe these insights into the polarized parton sector as will be described in Section 3. In Section 2 we shall, for completeness, present an analysis within the framework of the simplified SU(3) f symmetric 'standard' scenario in which the flavor-asymmetries in the polarized antiquark sector are neglected. This is done in view of the fact that in many situations these flavor asymmetries are unobservable as is the case for (most) presently available data which cannot provide any reliable information concerning this issue.
We emphasize that, as in our original analysis [1], we compute both g 1 and F 1 entirely in leading-twist QCD. In particular, in order to obtain F 1 , we use the parton densities of GRV98 [25] along with LO (note that R = 0 at leading order) or NLO coefficient functions for F 2 and R in (1.1). An alternative, frequently adopted [3,4,5,9,10,13] approach is to take F 2 (x, Q 2 ) and R(x, Q 2 ) from experimental measurements, which is motivated by the fact that leading-twist calculations of F 2 (x, Q 2 ) and R(x, Q 2 ) do not agree very well with experimental determinations in the region of low Q 2 and W 2 = Q 2 (1 − x)/x. These regions are affected by power-suppressed contributions and are therefore excluded from all unpolarized DIS analyses. The presently available data in the polarized case, however, do not allow to impose similarly 'safe' cuts (Q 2 ≥ 4 GeV 2 , W 2 ≥ 10 GeV 2 ) without losing too much information. On the other hand, the Q 2 range accessed so far in polarized DIS does not allow for extracting the magnitude and shape of power-suppressed contributions reliably. To study the issue further, we performed fits to the A 1 data in both possible ways, i.e. with leading-twist calculations of F 2 and R as well as with their experimental results, admitting at the same time an 'effective higher-twist' contribution to A 1 in terms of a factor (1 + A(x)/Q 2 ) with A(x) to be determined by the data. The outcome of this analysis was that A(x) is consistent with zero if we use leading-twist QCD for F 2 and R, but that it is sizeable and important in the fit if F 2 and R are taken from experiment. We take this as an indication that our preferred approach is more consistent and less liable to modifications by higher-twist terms. This view is also corroborated by the fact that the DIS A 1 data show only a very mild Q 2 -dependence, even toward low values of Q 2 . The consistency of the polarized parton densities as extracted from DIS data at comparatively low values of Q 2 with measurements of other hard processes at higher scales can be studied soon at RHIC and perhaps in the future at a polarized ep collider.
In NLO, g N 1 (x, Q 2 ) is related to the polarized (anti)quark and gluon distributions δf (x, Q 2 ) ≡ f + − f − in the following way: with the convolutions δC f * f being defined in the usual way. The MS coefficient functions can be also found in [1], where all necessary ingredients for the Q 2 -evolution have been formulated as well. A similar expression holds for the unpolarized structure function F N 1 (x, Q 2 ) with its spin-averaged parton distributions f (x, Q 2 ) ≡ f + + f − and the unpolarized Wilson coefficients can be found, for example, in [31]. The LO and NLO(MS) input scales, running coupling constants and parton distributions, employed in the positivity constraints |δf | ≤ f , will be adopted from GRV98 [25]. Furthermore we shall, as always, use the notation δq p ≡ δq and q p ≡ q, and neglect the marginal charm contribution to g N 1 stemming from the subprocess γ * g → cc [32]. The charm contribution to F N 1 is also small in the kinematic range covered by present polarization experiments.
The total helicity of a specific parton f = u,ū, d,d, s,s, g is given by the first (n = 1) Thus, according to (1.2), since ∆C q = −3C F /2 = −2 and ∆C g = 0. Therefore we have in general with the flavor-nonsinglet components being conserved, i.e. Q 2 -independent, and the flavor-singlet component is given by These quantities will be subject to various constraints (derived from hyperon β-decays) depending on the specific model scenarios under consideration to which we shall turn in the next two Sections.
Finally, the fundamental helicity sum rule reads where L q+g refers to the total orbital contribution of all (anti)quarks and gluons to the spin of the proton.
SU(3) f symmetric 'standard' (unbroken-sea) scenario
As stated in the Introduction, present data do not provide sufficient information concerning the flavor-asymmetries of the polarized sea distributions. Thus present day studies must, as a first approximation, neglect this issue unless one is willing to adopt some models for the flavor-asymmetries as will be done in Section 3. Here we follow the procedure presented in GRSV95 [1]. The searched for polarized NLO (as well as LO) parton distributions δf (x, Q 2 ), compatible with present data [15 -24] on A N 1 (x, Q 2 ), are constrained by the 'standard' sum rules ∆q 3 = F + D = g A = 1.2670 (35) (2.1) where the updated values for F and D have been taken into account [33] and the error estimate in Eq. (2.2) is due to [34]. Thus, Eq. (1.5) becomes i.e. one needs here a finite sizeable strange sea polarization ∆s(Q 2 ) < 0 in order to achieve the experimentally required reduction of the Ellis-Jaffe LO expectation [35] Γ p 1,EJ = Furthermore, in the 'standard' scenario one assumes an unbroken SU(3) f symmetric sea, For the determination of the NLO (LO) polarized parton distributions δf (x, Q 2 ) we follow our original analysis [1] by relating the polarized input densities to the unpolarized ones, using some intuitive theoretical arguments [36] as guidelines. We employ the following ansatz for the LO and NLO(MS) polarized parton distributions at an input scale Q 2 = µ 2 [37]: with the LO and NLO unpolarized input densities referring to the ones of GRV98 [25] andq ≡ (ū +d)/2 should be considered as the reference light sea distribution for the 'standard' unbroken-sea scenario in (2.5). The parameters of our optimal LO densities at µ 2 LO = 0.26 GeV 2 and the ones of the NLO(MS) densities at µ 2 NLO = 0.40 GeV 2 are given in Table I. These optimal LO and NLO(MS) fits correspond to a χ 2 per degree of freedom (χ 2 DF ) of χ 2 DF, LO = 0.84 and to χ 2 DF, NLO = 0.81. The polarized gluon density in (2.6) is, as usual, rather weakly constrained by present data. Although a fully saturated (via the positivity constraint) gluon input δg(x, µ 2 ) = ±g(x, µ 2 ) is disfavored, a less saturated δg(x, µ 2 ) = ±xg(x, µ 2 ) input or even a vanishing (purely dynamical) input δg(x, µ 2 ) = 0 are fully compatible with present data. The latter choice, however, seems to be unlikely in view of δq(x, µ 2 ) = 0.
In Fig. 1 our NLO results are compared with the data on A N 1 (x, Q 2 ) as well as with our old original NLO(MS) fit [1]. The differences between these two results are small, except perhaps for A n 1 in the large-x region. Our new LO fit is similar to the NLO one shown in Fig. 1 by the solid curves. The Q 2 -dependence of our LO and NLO 'standard' scenario fits at various fixed values of x is shown in Fig. 2 and compared with all recent data on g p 1 (x, Q 2 ), including the most recent E155 proton data [24]. The main reason for our LO results being larger than the NLO ones in the small-x region is due to the vanishing Fig. 3 where the expected extrapolations into the yet unmeasured small-x region down to x = 10 −3 are shown as well. The solid curves refer to our optimal NLO fit (with the input given in Table I) and allowing our optimal total χ 2 in Table I to vary by one unit, δχ 2 = ±1, gives rise to the shaded areas due to different choices of the polarized gluon input at Q 2 = µ 2 NLO in (2.6) such as δg = ±xg, etc. In particular a vanishing polarized gluon input δg(x, µ 2 NLO ) = 0, is for the time being entirely compatible with all present data as shown by the dashed curves. On the other hand, fully saturated (via the positivity constraint) gluon inputs δg(x, µ 2 NLO ) = ±g appear to be disfavored by present data as shown by the dotted curves in Fig. 3. It should be furthermore noted that the shaded bands in Fig. 3 contain polarized gluon densities which correspond to first moments ∆g(Q 2 = 5 GeV 2 ) between −0.81 and 1.73, according to input moments between ∆g(µ 2 NLO ) = −0.45 and 0.7, respectively, i.e. even negative total gluon polarizations are compatible with present data. The same results hold of course also in LO. Future dedicated polarized small-x measurements and upcoming determinations of δg(x, Q 2 ) should be useful in removing such extrapolation ambiguities caused by our present poor knowledge of δg(x, Q 2 ).
Our corresponding LO and NLO parton distributions at the respective input scales LO, NLO in Eq. (2.6) with the 'standard' scenario fit parameters given in Table I are shown in Fig. 4. The main differences between our new input densities and our old GRSV95 ones [1] are somewhat harder δd (due to the new neutron data) and δg distributions although, as discussed above, the polarized gluon distribution in Fig. 4 is only slightly preferred by our 'optimal' fit to presently available data. The polarized input densities in Figs. 4(a) and 4(b) are compared with our reference unpolarized valence-like LO and NLO dynamical input densities of [25] which satisfy of course the positivity requirement |δf | ≤ f as is obvious from Eq. (2.6). It should be nevertheless emphasized that the parameters, resulting from our rather general LO and NLO fits, are always such that these positivity conditions are automatically satisfied, i.e. there is practically no need to impose them separately. The distributions at Q 2 = 5 GeV 2 , as obtained from these LO and NLO inputs at Q 2 = µ 2 , are shown in Fig. 5 where they are also compared with our old NLO GRSV95 [1] results. It should be noted that the substantially harder input polarized gluon density xδg(x, µ 2 ), in particular in LO, in Fig. 4(a) as compared to our old GRSV95 fit, causes the sea density xδq(x, Q 2 ) to oscillate (slightly) in the large-x region at Q 2 > µ 2 as shown in Fig which result in ∆Σ = 0.259 and Our NLO results are summarized in Table II at some typical values of Q 2 . Both our LO and NLO results for Γ p,n 1 (Q 2 ) are in satisfactory agreement with recent experimental determinations [17 -24]. Furthermore, due to the constraint (2.1), the Bjorken sum rule [39] holds manifestly in LO and, according to (1.5), the NLO α s -corrected sum rule reads It is also interesting to observe that at our low input scales Q 2 = µ 2 LO, NLO = 0.26, 0.40 GeV 2 the nucleon's spin carried by the total helicities of quarks and gluons amounts only which implies for the helicity sum rule (1.9) already a sizeable orbital contribution L q+g (µ 2 LO, NLO ) ≃ 0.18, 0.15 at the low input scales. Although this is in contrast to our somewhat more intuitive previous GRSV95 result [1], L q+g (µ 2 LO, NLO ) ≃ 0, it should be kept in mind that, for the time being, ∆g(Q 2 ) is rather weakly constrained by present data as was discussed above.
Finally, for completeness we have also performed a NLO analysis in a different factorization scheme, the so-called chirally invariant (CI) or JET scheme [8,40,41], but any other choice would do as well for studying the scheme dependence of our MS fit results.
Here, among other things [8,40,41], the total helicity of quarks, ∆Σ CI , is conserved, i.e. Q 2 -independent, and is related to our ∆Σ in the MS scheme via Similarly agreeable fits as the ones in Figs. 1 -3 can be obtained in this scheme, e.g. by choosing a large positive gluon density with a total (input) helicity ∆g(µ 2 NLO ) CI ≃ 0.6 − 0.7 being about three times larger than in Table II and the sea density δq CI (∆q CI ) turns out to be roughly 50% smaller than the one of our best fit in the MS 'standard' scenario; here the total quark helicity increases to ∆Σ CI ≃ 0.4.
SU(3) f broken 'valence' (broken-sea) scenario
The assumption of the flavor symmetric 'standard' scenario with its unbroken sea density in (2.5) is expected to be unrealistic, following our experience in the unpolarized case where a suppression of the strange sea component is required, as accomplished by the vanishing input s(x, µ 2 ) =s(x, µ 2 ) = 0 in GRV98 [25], in order to comply with experimental indications [42,43] of an SU(3) f broken sea, and the positivity constraint δs ≤ s.
Thus, in GRSV95 [1] we also considered a 'valence' scenario where, in contrast to (2.5), Furthermore the full SU(3) f flavor symmetry, giving rise to the constraints (2.1) and (2.2), is broken in the 'valence' scenario to the extent [44] that the flavor-changing hyperon βdecay data fix only the total helicity of valence quarks ∆q v ≡ ∆q − ∆q: i.e. ∆q 3 = ∆u v −∆d v and ∆q 8 = 3F −D+4∆q at Q 2 = µ 2 [1] according to (3.1). Therefore a light polarized sea ∆q < 0 suffices here to account for the reduction of the Ellis-Jaffe estimate (2.4). This is the reason for our simplifying assumption of a maximally broken SU(3) f strange sea input in (3.1) in order to reduce the number of input distributions to be fitted to the rather scarce available polarization data, which are now sufficient for fixing these input distributions. (Future high-statistics data should allow, at least in principle, to extract the total strange sea polarization without employing any simplifying assumption, as for example from (1.8), ∆(s +s) = (∆Σ − ∆q 8 )/3 .) The quality of the fits obtained is comparable to that of the 'standard' flavor-symmetric scenario discussed and presented in the previous Section, cf. Fig. 1. We refrain, however, from presenting these results here explicitly because the assumed remaining flavor-isospin symmetry of the light sea components in (3.1) appears to be somewhat artificial and unnatural in view of the flavor-asymmetric unpolarized light sea distributionsd(x, Q 2 ) >ū(x, Q 2 ) [25 -28].
Turning now to the presumably more realistic scenario where also the flavor-isospin symmetry of the polarized sea is broken, we note, as already pointed out in the Introduction, that some model assumptions are needed for the corresponding input distributions.
The analysis of the unpolarized structure functions yields which holds to a rather good accuracy for the GRV98 distributions [25]. This proportionality relation is expected to hold approximately at least for 0.01 x 0.3 where the breaking of the light sead >ū is directly tested experimentally via Drell-Yan dilepton production in pp and pd collisions [45] and semi-inclusive π ± production in ep and ed reactions [46]. The relation (3.4) may be considered [30] as a manifestation of the Pauli-blocking effect [47] which should be relevant also in the polarized parton sector.
Our resulting input distributions can be parametrized as in (2.6) where now, instead of the unbroken δq sea, we have a similar parametrization for δū(x, µ 2 ) and δd(x, µ 2 ) which are constrained by (3.5) [51], together with the respective flavor-broken unpolarized input densitiesū(x, µ 2 ) undd(x, µ 2 ) taken from [25]. The parameters of our optimal LO densities at µ 2 LO = 0.26 GeV 2 and the ones of the NLO(MS) densities at µ 2 NLO = 0.40 GeV 2 are given in Table III. These optimal fits correspond to χ 2 DF, LO = 0.823 and χ 2 DF, NLO = 0.816 similarly to the 'standard' scenario in Sec. 2. The quality of these LO and NLO fits to A N 1 (x, Q 2 ) in the broken 'valence' scenario is practically identical to our new fit in the 'standard' scenario shown in Fig. 1 by the solid curves. The same holds true also for g p,n,d Table III are shown in Fig. 6, which are also compared with our reference unpolarized valence-like dynamical input densities of [25] which satisfy the positivity constraint |δf | ≤ f . The polarized gluon densities turn out to be somewhat larger here, in particular in NLO, than the ones in the 'standard' scenario shown in Fig. 4.
It should be furthermore emphasized that we always expect for the broken light-sea input densities to have a positive δū and a negative δd with |δd| > δū, i.e. δū − δd > 0 and δū + δd < 0.
In Fig. 7 we present the flavor asymmetry x(δū−δd)(x, µ 2 ) separately, as obtained from Fig. 6, which compares favorably with predictions of the relativistic field theoretical chiral quark-soliton model [52,50]. Similar results have been obtained by a recent analysis [53] based on the statistical parton model which are supposed to hold at a somewhat larger input scale Q 2 0 = M 2 p ≃ 0.9 GeV 2 . We note furthermore that the prediction δg(x, M 2 p ) = 0 of this latter model is consistent with the results of the present analysis which do not exclude this possibility, cf. Fig. 3. The resulting NLO distributions at Q 2 = 5 GeV 2 are shown in Fig. 8 where they are also compared with our new NLO results obtained in the 'standard' scenario as shown by the solid curves in Our NLO results are summarized in Table IV (3.12) which is larger than the 'standard' scenario results (2.10), and thus a very small orbital contribution L q+g (µ 2 LO, NLO ) ≃ 0.08, 0.02 is required at the low input scales in order to comply with the sum rule (1.9). This is somewhat similar to our previous results [1], but again ∆g(Q 2 ) is not strongly constrained by present data. Nevertheless it is intuitively appealing that this nonperturbative orbital (angular momentum) contribution to the helicity sum rule (1.9) vanishes at our low input scales, L q+g (µ 2 ) ≃ 0. This is in contrast to larger scales Q 2 > µ 2 where hard radiative effects give rise to sizeable orbital components due to the increasing k T of the partons, which eventually have to compensate in (1.9) the strongly increasing gluon polarization ∆g(Q 2 ) ∼ α −1 s (Q 2 ) : in both scenarios we obtain, for example, ∆g(10 GeV 2 ) ≃ 1.
Finally let us conclude with a few remarks concerning the flavor-symmetry breaking which was implemented in our broken 'valence' scenario via the entirely empirical relation (3.5). On rather general grounds one expects the product to be a universal flavor-independent function P (x), since the effect of Pauli-blocking is only related to the spin (helicities) of quarks and antiquarks irrespective of their flavor degree of freedom. This implies δu(x, µ 2 ) δū(x, µ 2 ) = δd(x, µ 2 ) δd(x, µ 2 ), i.e. Eq. (3.5).
Furthermore, we have seen that the data select, within our 'valence' scenario with its totally flavor-broken polarized light sea densities in (3.5) and (3.6), the solution of Eq.
(3.5) which satisfies P (x) > 0 in (3.13) for q = u, d as can be seen in Fig. 6. This can be understood [30] as a consequence of the expected predominant pseudoscalar configuration [29,54] of the quark-antiquark pairs in the nucleon sea. In fact, the two relations uū ≃ dd and δu δū = δd δd at the input scale Q 2 = µ 2 can be rewritten as with P = P p − P a in (3.13) and the common helicity densities being given by q )/2 where for brevity we have dropped the x-dependence. A predominant pseudoscalar configuration of (qq)-pairs in the nucleon sea implies, via Pauli-blocking, that the aligned quark-quark configurations q + (q +q− ) and q − (q −q+ ) are suppressed relatively to the antialigned q + (q −q+ ) and q − (q +q− ) 'cloud' configurations, i.e. P p (x) > P a (x) which implies P (x) > 0 in (3.13). The result for P p /P a , corresponding to our optimal fit, is shown in Fig. 9: clearly, this ratio will be maximal where xq(x, µ 2 ) and xδq(x, µ 2 ) are maximal at x ≃ 0.2 − 0.4, cf. Fig. 6, i.e. where the Pauli-blocking, Eq. (3.13), is most effective which is nicely exhibited in Fig. 9 in LO and NLO.
It is interesting to mention that some of these expectations, which derive mainly from our light-sea flavor breaking relation (3.5), have been already confirmed by a recent entirely independent simultaneous analysis [14] of polarized DIS and semi-inclusive deep inelastic scattering (SIDIS) asymmetry-data. In particular the more recent high precision SIDIS HERA-HERMES data [55] on h + production (h = π,K dominantly) off a proton target, e p → eh + X, seem to play a decisive role in favoring flavor-broken light sea densities δū(x, Q 2 ) = δd(x, Q 2 ), despite the fact that these asymmetry data on A h + 1p refer to rather small scales Q 2 > ∼ 1 GeV 2 . The reason for this discriminative power is, when combined with the data from inclusive DIS, due to the fact that A h + 1p is proportional [14], besides to the dominant valence contribution, also to δd − 4δū, multiplied by a 'favored' fragmentation function, which is significantly more sensitive to δū than to δd. A clear preference for a positive δū has been observed [14], which is very similar to our NLO δū shown in Fig. 6(b), and a flavor symmetric 'standard' light sea scenario seems to be strongly disfavored.
We have calculated at NLO the spin asymmetries for semi-inclusive DIS using the well known theoretical SIDIS framework [56,14] together with our results for the polarized parton distributions of the 'standard' and 'valence' scenario with their flavor-symmetric and flavor-broken light sea densities, respectively, employing the fragmentation functions of [57]. (We did not use the alternative set of recent fragmentation functions suggested in [58], since they refer to scales Q 2 larger than 2 GeV 2 .) The results for the relevant SIDIS asymmetry A h + 1p are shown in Fig. 10. Although the high precision HERMES data [55] seem to favor slightly the 'valence' scenario with its flavor-broken light sea, the results of the 'standard' scenario with its flavor-symmetric light sea cannot yet be ruled out. Both scenarios in Fig. 10 give rise to a comparable χ 2 /(9 data points) of 7.6 and 8.5 for the 'valence' and 'standard' scenario, respectively.
Summary and conclusions
All recent polarized DIS data, including the most recent SLAC-E155 proton data [24], have been analyzed and studied within the 'standard' and 'valence' scenario in LO and NLO of QCD. The 'standard' scenario, characterized by (2.1) and (2.2), refers to the common simplified, but probably unrealistic, assumption of an SU(3) f flavor-symmetric polarized light sea. The original 'valence' scenario [1], characterized by (3.2) and (3.3), is now modified by employing a totally SU(3) f asymmetric polarized light sea δū = δd = δs which leads to the modified constraints (3.7) and (3.8). Since inclusive polarized DIS data cannot fix the flavor-broken sea densities, we have modelled the flavor-asymmetric light sea densities δū = δd using a Pauli-blocking ansatz [30] in (3.5), because similar 'blocking' effects can also explain the flavor-asymmetry of unpolarized sea densities (d >ū). All our resulting polarized parton distributions respect the fundamental positivity constraints down to the low resolution scales Q 2 = µ 2 LO = 0.26 GeV 2 and µ 2 NLO = 0.40 GeV 2 . The polarized gluon distribution δg(x, Q 2 ) is weakly constrained by present data in both scenarios. In particular, a vanishing gluon input δg(x, µ 2 ) = 0 is equally compatible with all present measurements of A N 1 (x, Q 2 ) or g N 1 (x, Q 2 ). Only a fully saturated (via the positivity constraint) gluon input δg(x, µ 2 ) = ±g(x, µ 2 ) appears to be disfavored by present data.
The presumably more realistic 'valence' scenario with its flavor-broken light sea quark distributions δū = δd ( = δs = δs ≃ 0) leads to a positive δū(x, Q 2 ) density and a sizeably larger negative δd(x, Q 2 ). These results are supported by a recent combined analysis [14] of polarized DIS and semi-inclusive DIS data and agree with predictions of the relativistic field theoretical chiral quark-soliton model [52,50] and of the statistical parton model [53]. Present high statistics HERA-HERMES data [55] on semi-inclusive asymmetries 1N for h ± production off nucleon targets cannot, however, yet uniquely discriminate between our 'valence' scenario with flavor-broken polarized light sea densities and the common 'standard' scenario with a flavor-symmetric light sea-quark distribution.
A FORTRAN package containing our optimally fitted 'standard' and fully flavorbroken 'valence' NLO(MS) as well as LO distributions can be obtained by electronic mail.
δg(x, Q 2 )/g(x, Q 2 ) = 0.41 ± 0.18 at x ≃ 0.17 and Q 2 ≡ p 2 T = 2.1 GeV 2 , which implies xδg(x, Q 2 ) ≃ 0.35±0.15, actually compatible with our results in Fig. 5, does not include systematic theoretical errors. These could be sizeable in the low transverse momentum region considered. Table II. First moments (total polarizations) ∆f of polarized NLO parton densities δf (x, Q 2 ) and g p,n 1 (x, Q 2 ), defined in (1.3) and (1.4), as obtained in the 'standard' scenario. The marginal differences between ∆ū and ∆d at Q 2 > µ 2 , generated dynamically by the NLO evolution, are not displayed. Table I] with present data [16 -24]. The Q 2 values adopted here correspond to the different values quoted in [16 -24] for each data point starting at Q 2 ≥ 1 GeV 2 at the lowest available x bin. Our old NLO GRSV95 fit [1] is shown for comparison as well (dashed curves). Our present LO fit is very similar to the NLO one shown by the solid curves. 3. The x-dependence of g N 1 at Q 2 = 5 GeV 2 in the NLO 'standard' scenario. Different choices of the gluon input δg(x, µ 2 NLO ) in (2.6) are shown by the dashed and dotted curves as indicated. Allowing our optimal total χ 2 to change by one unit, δχ 2 = ±1, results in the shaded areas shown. The data are taken from [17, 18, 20, 22 -24]. Table I at µ 2 LO = 0.26 GeV 2 with our previous old GRSV95 fit [1] and with the unpolarized dynamical GRV98 input densities of [25]. (b) The same as in (a) but for the NLO input densities at µ 2 NLO = 0.40 GeV 2 . δū(x, µ 2 ) and δd(x, µ 2 ) in the 'valence' scenario as obtained from Fig. 6. The prediction of the chiral quark-soliton model is taken from K. Goeke et al. [52]. 9. The ratio of P p (x) and P a (x), defined in (3.14), at the LO and NLO input scales µ 2 LO, NLO which demonstrates the Pauli-blocking of the disfavored antiparallel q ±q∓ configurations relative to the favored parallel q ±q± configurations as discussed in the text. Fig. 10. NLO predictions for the semi-inclusive DIS asymmetry A h + 1p for h + production off a proton target within the 'valence' and 'standard' scenario. The HERMES data are taken from [55].
|
2014-10-01T00:00:00.000Z
|
2000-11-17T00:00:00.000
|
{
"year": 2000,
"sha1": "5bfadebaaee1988ae854c8a9d6612c4c9aafb160",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/hep-ph/0011215",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "ce51536faffc0f019182916695d12537b0c37259",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
}
|
256445233
|
pes2o/s2orc
|
v3-fos-license
|
Ovarian Morphometric and Histologic Characteristics and Correlation with Clinical Factors: A Cross-Sectional Study
Reproductive lifespan is determined by the reserve of ovarian follicles; their quality and quality determine the fertility potential at a given point in time for a particular individual. Inter-individual variations related to morphometry, laterality, medical history, demographic characteristics and ethnicity may impact ovarian histology, which however, has not been extensively studied or documented. The present cross-sectional study aims to investigate the potential association of clinical factors (age, medical and obstetric history) with ovarian morphometry and histology in females of reproductive age in the local population. The sample included 31 specimens of whole human ovaries, obtained from surgical/autopsy procedures in reproductive-aged women, processed at the Pathology Department. Morphometric characteristics were assessed, including shape, color, length, width, thickness and gross ovarian pathology. Random samples of specific dimensions were histologically examined to determine follicular counts. The results were analyzed statistically in correlation to morphometric characteristics and medical history. The majority of the patients had oval-shaped ovaries (77.8% right; 92.3% left; p = 0.368) of whitish color (38.9% right; 46.2% left; p > 0.999). Right ovaries had significantly greater length, width and volume (p-values 0.018, 0.040 and 0.050, respectively). Thickness was equivalent, as well as follicular distribution of all classes. Age correlated inversely with ovarian volume and primordial/primary follicular count on histology. Women with a caesarian-section history yielded significantly lower primordial/primary follicular counts. As estimated by ovarian histology, macroscopic and clinical factors may be significantly associated with actual ovarian reserve.
Introduction
Human ovaries lie within the lesser pelvis, laterally to the uterus and inside a shallow depression of the lateral pelvic wall termed as the "ovarium fossa". They serve as reserves for ovarian follicles and reproductive hormones, playing a crucial role in female fertility and endocrine function [1]. The dimensions of the ovaries vary greatly throughout the lifespan of a female and exhibit considerable inter-individual differences, approximately ranging between 3-5 cm × 1.5-3 cm × 0.6-1.5 cm (length × width × thickness), and weighing 5-8 g [2]. They are smooth and pinkish during the early reproductive life. However, subsequent cycles of ovulation-scarring-corpus luteum and corpus alba formation distort their initially smooth surface [2,3].
The exact duration of the female reproductive lifespan has received accelerated interest in the recent years [4,5]. Reproductive aging is a matter of significant concern, as the mean age of childbearing has had an upward trend during recent decades [6]. Maternal age is firmly associated with the remaining ovarian reserve and has a substantial impact on the specific Assisted Reproductive Technology (ART) protocols and outcomes [7]. ART has a broad application in the field of infertility, assembling 2.0% of live births in USA in 2018 [8].
Ovarian reserve is defined by oocyte quantitative and qualitative characteristics and depicts the fertility potential of a woman at a specific time [4,9]. The need for quantification of the ovarian reserve has emerged during this process, mostly in terms of defining the prognosis and optimizing the protocols for ART [7]. Ovarian reserve is assessed by several clinical tools including transvaginal ultrasound measuring antral follicle count and ovarian volume as well as laboratory tests, including the follicle stimulating hormone (FSH), estradiol E2, and anti-Müllerian hormone (AMH) [10,11].
It has been documented that ovarian volume is associated with actual ovarian reserve, counted as the number of non-growing follicles present in histopathology [12]. Previous studies have demonstrated that an ovarian volume of less than 3 cm 3 is associated with an increased risk of cycle cancellation in in-vitro fertilization (IVF) despite aggressive stimulation protocols. It is, also, associated with a requirement for higher doses of gonadotropins, and lower oocyte yields [3,13], suggesting a potential association of macroscopic characteristics such as volume with the actual number of recruitable follicles. Nonetheless, it has been suggested by previous findings that laterality (right versus left side) may significantly affect the ovarian volume and the antral follicle count (AFC) [14].
Concerning ovarian volume, anatomical descriptions for human ovaries are mainly retrieved from classic anatomy and histology textbooks [15,16]. There are several studies in the literature, concerning the morphologic and histopathologic features of human ovaries, but they are highly heterogeneous in terms of methodology, studied parameters, population, ethnicity and other potential variables that may affect the results. Nevertheless, the largest and most comprehensive studies regarding ovarian volume were conducted by means of ultrasound scanning [17,18]. There are reasonable concerns about the potentially limiting role of ultrasound technology, as many studies date back several years. The accuracy of ultrasound measurements may not reflect current practice, as ultrasound technology has greatly advanced over the years. However, the initial findings are enhanced by more recent publications [19].
Taking into consideration the growing interest in reproductive medicine both in terms of assisted reproduction and fertility preservation techniques, it is extremely useful to investigate the potential association of ovarian morphology and morphometry in relation to histopathology, past medical history, demographic characteristics and laterality in specific populations. Within our scope is to comment on the morphometric and morphologic characteristics of human ovaries and their association with histopathologic and clinical features in our study population, using a method which is cost effective and easily reproducible.
This study is motivated by the fact that variations related to ethnicity and environmental factors may have an impact on ovarian morphology and histopathology, which has not been extensively studied or proven. The findings aim to build a database for future studies in this particular population.
Materials and Methods
The sample of the present study consists of human ovaries prospectively collected from surgical or cadaveric specimens during the time period from November 2016 to November 2021 at (i) the Department of Pathology of Democritus University of Thrace, (ii) the Laboratory of Forensic Sciences and Toxicology of Democritus University of Thrace in Alexandroupolis, Evros and (iii) the Department of Pathology of "Thriasio" General Hospital of Elefsina, Attica in Greece. The samples were embedded in formaldehyde solution 10% and analyzed at the Department of Pathology in Democritus University of Thrace (Alexandroupolis, Greece). The study is in compliance with the declaration of Helsinki for medical research and has been approved by the Ethics and Research Committees of the University Hospital of Alexandroupolis and the General Hospital of Elefsina "Thriasio" (under documentation codes ES1/08-02-2018 and 185/22-06-2020, respectively).
All of the specimens that were collected during this time interval in females of reproductive age were included in the present study. Since the existence of differences in the different sides (right and left) is controversial in the literature [20,21], we have included both right-sided and left-sided ovaries, when applicable, in order to (i) avoid bias and (ii) run a comparison between the different sides.
A total of 19 patients and 31 ovaries were analyzed. Morphological characteristics such as dimensions (length, width, thickness), shape and color were measured. Subsequently, the specimens underwent histopathologic analysis, and the number of follicles was evaluated. The samples were subjected to the following process.
Macroscopic Observation
Macroscopic observation, including the shape and the color, was followed by measurement of the dimensions (length, width, thickness) for each ovary with electronic calipers. The presence of any gross pathology, such as cysts, was also addressed. The ovarian volume was estimated according to the following formula: where V comprises the measured volume in cm 3 or ml, and a, b, c comprise each of the three radii of the ellipsoid in cm. The value of π is equal to 3.14. As the actual diameters were measured, the formula was modified as follows: where l: length, w: width, and t: thickness ( Figure 1). All of the specimens that were collected during this time interval in females of reproductive age were included in the present study. Since the existence of differences in the different sides (right and left) is controversial in the literature [20,21], we have included both right-sided and left-sided ovaries, when applicable, in order to (i) avoid bias and (ii) run a comparison between the different sides.
A total of 19 patients and 31 ovaries were analyzed. Morphological characteristics such as dimensions (length, width, thickness), shape and color were measured. Subsequently, the specimens underwent histopathologic analysis, and the number of follicles was evaluated. The samples were subjected to the following process.
Macroscopic Observation
Macroscopic observation, including the shape and the color, was followed by measurement of the dimensions (length, width, thickness) for each ovary with electronic calipers. The presence of any gross pathology, such as cysts, was also addressed. The ovarian volume was estimated according to the following formula: where V comprises the measured volume in cm 3 or ml, and a, b, c comprise each of the three radii of the ellipsoid in cm. The value of π is equal to 3.14. As the actual diameters were measured, the formula was modified as follows: where l: length, w: width, and t: thickness ( Figure 1).
Figure 1.
Graphical demonstration of the measured parameters. Macroscopic appearance of human ovary and demonstration of measured parameters; with red axis for length; black axis for width; orange axis for thickness.
Histopathologic Evaluation
Each ovary was further processed to obtain a random ovarian tissue specimen of fixed dimensions of 3 mm × 3 mm × 5 mm, containing an area of 3 mm × 3 mm of ovarian cortex. The tissue specimens, after being processed with the appropriate solution of alcohol and xylol, were embedded into paraffin wax. Slices of 5 μm every 50 μm of tissue were obtained from a microtome until the tissue specimens were exhausted. Only histopathology slices for which the whole surface was intact were included in the final analysis. As a result, a total of 25 slices were obtained for each ovary. Previous studies have reported measurements of variable width for cortical tissue such as 5 μm or 6 μm [22], 7 μm [23], 6 μm, 10 μm [22] or even 50 μm [24,25]. Taking into consideration that the mean diameter
Histopathologic Evaluation
Each ovary was further processed to obtain a random ovarian tissue specimen of fixed dimensions of 3 mm × 3 mm × 5 mm, containing an area of 3 mm × 3 mm of ovarian cortex. The tissue specimens, after being processed with the appropriate solution of alcohol and xylol, were embedded into paraffin wax. Slices of 5 µm every 50 µm of tissue were obtained from a microtome until the tissue specimens were exhausted. Only histopathology slices for which the whole surface was intact were included in the final analysis. As a result, a total of 25 slices were obtained for each ovary. Previous studies have reported measurements of variable width for cortical tissue such as 5 µm or 6 µm [22], 7 µm [23], 6 µm, 10 µm [22] or even 50 µm [24,25]. Taking into consideration that the mean diameter of a primordial follicle is estimated at approximately 39.5 µm (±7.6 µm) with a maximum up to 49 µm [24], the histopathology slice width was predetermined at 5 µm to retain a high quality of tissue and facilitate microscopic evaluation. To avoid double measurements one single 5µm slice was extracted from each 50 µm of cortical width. Based on previous studies, the random comparison and microscopic evaluation of ovarian tissue is sufficiently accurate to draw safe conclusions, in so far as the comparison and not the actual absolute number of follicles is the investigated outcome [26,27]. As noted, a total of 25 slices available for microscopic evaluation were obtained for each ovary included in this study. The tissue was dyed with conventional hematoxylin-eosin stain and observed with an optical microscope Nikon Eclipse 50i with magnification × 10 HPF (high-power field) (Nikon Group Companies, Tokyo, Japan). Follicular counts were obtained in a total of 6 samples from each ovary, randomly selected as the 1st, 5th, 10th, 15th, 20th and 25th of each series. Subsequently, they were classified based on their characteristics, as demonstrated by Gougeon et al. [28], to primordial, primary, secondary, and tertiary follicles ( Figure 2). The slides were observed by two independent researchers and the results were cross-checked. The findings were statistically analyzed to detect any significant association in follicular number with age, BMI, laterality, ovarian volume, height, width, thickness, previous gynecological and obstetric history. of a primordial follicle is estimated at approximately 39.5 μm (±7.6 μm) with a maximum up to 49 μm [24], the histopathology slice width was predetermined at 5 μm to retain a high quality of tissue and facilitate microscopic evaluation. To avoid double measurements one single 5μm slice was extracted from each 50 μm of cortical width. Based on previous studies, the random comparison and microscopic evaluation of ovarian tissue is sufficiently accurate to draw safe conclusions, in so far as the comparison and not the actual absolute number of follicles is the investigated outcome [26,27]. As noted, a total of 25 slices available for microscopic evaluation were obtained for each ovary included in this study. The tissue was dyed with conventional hematoxylin-eosin stain and observed with an optical microscope Nikon Eclipse 50i with magnification × 10 HPF (high-power field) (Nikon Group Companies, Tokyo, Japan). Follicular counts were obtained in a total of 6 samples from each ovary, randomly selected as the 1st, 5th, 10th, 15th, 20th and 25th of each series. Subsequently, they were classified based on their characteristics, as demonstrated by Gougeon et al. [28], to primordial, primary, secondary, and tertiary follicles ( Figure 2). The slides were observed by two independent researchers and the results were cross-checked. The findings were statistically analyzed to detect any significant association in follicular number with age, BMI, laterality, ovarian volume, height, width, thickness, previous gynecological and obstetric history.
Statistical Analysis
Quantitative variables were expressed as mean values (Standard Deviation) and as median (interquartile range), while qualitative variables were expressed as absolute and relative frequencies. The Mann-Whitney test was used for the comparison of continuous variables between two groups. Spearman's correlation coefficient (rho) was used to explore the association of two continuous variables. Multiple linear regression analysis was used with the volume and the number of primordial/primary follicles as the dependent variables. The regression equation included terms for all patients' demographic and clinical characteristics. In cases where volume was the dependent variable, the number of primary cells was also included in the model as an independent variable. Linear regression analysis was conducted after the logarithmic transformation of the dependent variables. All reported p values are two-tailed. Statistical significance was set at p < 0.05 and analyses were conducted using SPSS statistical software (version 22.0).
Statistical Analysis
Quantitative variables were expressed as mean values (Standard Deviation) and as median (interquartile range), while qualitative variables were expressed as absolute and relative frequencies. The Mann-Whitney test was used for the comparison of continuous variables between two groups. Spearman's correlation coefficient (rho) was used to explore the association of two continuous variables. Multiple linear regression analysis was used with the volume and the number of primordial/primary follicles as the dependent variables. The regression equation included terms for all patients' demographic and clinical characteristics. In cases where volume was the dependent variable, the number of primary cells was also included in the model as an independent variable. Linear regression analysis was conducted after the logarithmic transformation of the dependent variables. All reported p values are two-tailed. Statistical significance was set at p < 0.05 and analyses were conducted using SPSS statistical software (version 22.0).
Results
The sample consisted of 31 ovaries from reproductive-aged women with a mean age of 43.3 years (SD = 7.5 years) and regular menstrual cycles. Their characteristics are presented in Table 1. Mean BMI was 26.0 kg/m 2 (±3.7) and 63.2% were overweight. Smokers comprised 36.8% of the sample. A total of 77.8% had an obstetrical history of at least one vaginal labor and 27.8% of at least one caesarian section. Only one woman had a history of malignancy (breast cancer) and 15.8% had undergone a previous adnexal surgery. The majority of the patients had oval shaped ovaries (77.8% right ovaries and 92.3% left ovaries, p = 0.368). As far as the ovarian color is concerned, it was similar in left and right ovaries (p > 0.999) and whitish in the majority of the cases (38.9% for the right and 46.2% for the left ovaries). Right ovaries had significantly greater length, width and volume, compared to left ovaries (Table 2). Thickness was similar in left and right ovaries, as well as the distribution of all classes of follicles. Spearman's correlation coefficients of ovarian volume with follicular counts are presented in Table 3. Ovarian volume was significantly positively associated with primordial, primary and secondary follicles. Greater age was significantly associated with lower volume (Table 4). Women over 45 years old had significantly lower volume compared to younger women. Moreover, significantly greater ovarian volume was associated with benign adnexal pathology as a primary indication for surgery, with type of surgery confined to adnexectomy or where there was a macroscopic appearance of a cyst. Greater age was significantly associated with less primordial/primary follicles (p = 0.004). Women over 45 years old had significantly less primordial/primary follicles compared to younger women(p = 0.019). Furthermore, significantly less primordial/primary follicles were found in ovaries from women who had undergone a caesarian section in the past (p = 0.033).
With multiple linear regression, greater ovarian volume was associated with higher primordial/primary follicular density. In contrast, greater age and having had a caesarian section in the past were significantly associated with lower primary follicular density. Greater age was associated with significantly reduced ovarian volume (Table 5).
Discussion
The present study aimed to investigate the macroscopic and histopathological characteristics of human ovaries, regarding follicular counts, as well as their association with clinical factors, by using a method that is easily applicable and reproducible in common clinical practice.
The dominant shape was oval and the dominant color whitish, with no differences concerning laterality. Our findings suggest that right ovaries had significantly greater length, width and volume, compared to left ovaries. However, thickness was similar in both sides.
In the present study, ovarian volume was negatively affected by increasing age and primordial and primary follicles were negatively affected by both age and previous caesarian deliveries. Ovarian volume was independent from BMI, smoking habits and past obstetrical history (vaginal, caesarian deliveries, abortions). Clinical factors such as BMI, smoking habits, indication and type of surgery, previous adnexal surgeries, presence of a macroscopically evident cyst and previous vaginal deliveries do not seem to significantly affect the total follicular counts. Regarding the correlation of morphometry to histopathologic findings, increased ovarian volume was significantly correlated with increased follicular counts on ovarian histopathology.
Our findings demonstrate variability compared to previous studies. According to the findings of Rani et al., the predominant shape of human ovaries is almond in women aged 16 to 55 years [29]. This is inconsistent with our findings, where the predominant shape is oval at both sides. The authors, also, reported a gradual increase in all ovarian dimensions in older women. However, comparison between different sides (right or left) was not conducted, neither were the findings correlated with histopathology.
Perven at al. conducted a post-mortem study of human ovaries in the specific ethnic group of Bangladeshi women and found that the weight of right ovaries was significantly greater, compared to the left side [30]. As expected, the ovarian weight was significantly lower in the subgroup of women aged over 45 years. Of note, the findings were not correlated with histopathology [30]. In a previous study conducted in the same population, the authors found that right ovaries were associated with significantly greater length, breadth and thickness, compared with the left side and the findings were consistent within different age groups [21].
Ahmed at al. investigated the morphometric characteristics of human ovaries again in Bangladeshi women, including the variables of length, width and thickness [31]. Concerning the different age subgroups, the findings suggested that ovarian length was significantly greater in the older age groups, in both sides. The average width and thickness were significantly higher in the older age groups, but only on the right side. Breadth and thickness were significantly smaller in females aged under 13 years, but no significant differences in women over or under 45 years of age were observed.
Pavlik et al. determined the association of ovarian volume measured by ultrasound with age, height and weight in a large cohort of 58,673 women. Notably, ovarian volume negatively correlated with age, but also with small height [17].
The findings described above suggest high variability of the measured parameters, concerning laterality, age, and probably ethnicity, as specific ethnic subgroups are separately investigated. One would expect that as ovaries atrophy with ageing their morphometric characteristics would decrease in older age groups; however, this finding is not consistent among different studies.
The bottom-line in any case is the estimation of ovarian reserve. Several clinical and laboratory tools have been used in clinical practice to provide an indirect estimate of ovarian reserve, such as AFC or AMH [32][33][34]. Korsholm et al. conducted a cross-sectional study in a large sample consisting of 1423 reproductive-aged women in order to investigate antral follicular count and ovarian volume, measured by transvaginal sonography [14]. The study included both healthy and infertile women. The key findings suggested that right ovaries are larger and included more antral follicles than the left. To adjust for significant confounders such as age and AMH, patients were subcategorized to quartiles based on their chronological or biological (based on AMH) age. The difference in volume was consistent in all subgroups, except for AFC, which was significantly higher in the right ovary, except for the lowest AMH and the highest age quartile.
Alserri et al. conducted a retrospective observational study using a large sample of 6617 ultrasound scans, to assess a potential association of AFC with laterality [35]. The findings supported that antral follicular counts are significantly higher in the right side, compared to the left. This association was prevalent both in patients with polycystic ovarian morphology (PCO) and controls, when ovarian reserve was considered normal based on AFC counts [35]. Interestingly, the significant difference was eliminated in the subgroup of women with low ovarian reserve, as classified by an AFC count between 1 and 9 [35]. This is in line with findings of Korsholm et al. [14].
Jokubkiene et al. investigated ovarian volume and the antral follicular counts by using 3D-transvaginal ultrasound in 213 women receiving combined oral contraceptives (COCs). Based on their findings, laterality had significant impact on measured parameters. The right ovary was associated with greater volume, more antral follicles and larger predominant follicle, compared to the contralateral side, in the age group of 20-29 years. In women aged 30-39 years, the differences, although present, did not reach statistical significance [36]. A tendency towards reduced ovarian volume and less antral follicles was observed within older subjects (30-39 years of age) under COCs. Ovarian volume and vascularization, as measured by Doppler indices, seemed to decrease with advancing age, but the difference did not reach statistical significance. In this particular study, if only one ovary included a follicle > 10 mm, this was considered as "dominant" and a comparison of the measured variables in dominant and non-dominant ovaries was carried out, with no distinctions concerning laterality. The number of such cases was limited (24 cases) to allow for sound conclusions. Ovaries containing follicles >10 mm (dominant) did not present significant differences in ovarian volume and follicular number between the different age groups. On the contrary, the "non-dominant" ovaries were associated with diminished ovarian volume and antral follicles in the older age group.
Jokubkiene et al. also conducted a cross-sectional study with prospectively collected data in reproductive-aged women between 20 to 39 years to evaluate the normal 3Dtransvaginal ultrasound findings in terms of ovarian volume, antral follicular count, total follicular volume and vascular Doppler indices [37]. The current use of COCs was an exclusion criterion. The analysis included a total of 303 women and the results were analyzed separately when all follicles measured <10 mm or when only one follicle in one ovary exceeded 10 mm. The findings suggested that right ovaries had significantly higher volume, more follicles and increased vascularization, compared with the left side, in the first subgroup. Ovarian volume and total follicular counts seemed to substantially diminish with age in both sides. Notably, AFC was significantly associated with total ovarian volume with a positive correlation.
In the subgroup of women with an ovary containing a follicle > 10 mm, the dominant ovary was associated with significantly higher volume but no difference in total follicular counts were observed between the two sides [37]. The total number of follicles were negatively associated with advancing age, a finding consistent with the group with no follicles exceeding 10 mm. Interestingly, ovarian volume of the dominant ovary was not affected by age, in contrast to the nondominant ovary; however, no distinction according to laterality was made.
Several theories for this laterization have been proposed in the current literature. Theoretically, a per se increased pool of primordial follicles in the right ovary established during early fetal development cannot be excluded [14]. Other potential causes may demonstrate their effect through accelerated follicular loss and atresia in the left-sided ovaries [14]. The anatomical peculiarities in blood drainage between different sides have been, also, implicated, as causal factors [35,36]. The drainage of the left ovarian vein is directed to the left renal vein, in contrast with the right draining directly to the inferior vena cava [38][39][40]. However, significant anatomical variations may exist, especially in venous drainage [39]. There are specific peculiarities concerning laterality, reflected also on the anatomy of the testicular veins that are involved in the etiopathogenesis of the varicocele in males [38]. Firstly, the inferior vena cava has a greater diameter. Secondly, the left renal vein develops a higher intravascular pressure, mainly due to its "entrapment" between the aorta and its branch, the superior mesenteric artery [41]. A wide spectrum of clinical manifestations arising from this "nutcracker phenomenon" may be presented in both genders, especially during the second and third decades of life [42,43]. Thirdly, the insertion of the left ovarian vein to the left renal vein forms a 90-degree angle [38]. These factors predispose to elevated intravascular pressure in the left gonadal vein, dilation with subsequent valve malfunction. The sequelae of blood stasis are increased local temperature and relative hypoxia, contributing to testicular dysfunction and impaired spermatogenesis in males with varicoceles [44]. A potential similar effect on females expressed as decreased ovarian volume and follicular number in the left side remains hypothetical [35].
An additional concern is that ultrasound evaluation of the left ovaries may be more challenging due to their anatomic location in close proximity with the sigmoid, potentially contributing to differences in measured parameters in favor of the right side [36]. However, these technical parameters are not considered to be a great contributor to these background differences, between the right and left side. The findings of our study, also, support the existence of a potential true biologic variation between the different sides.
Accurate knowledge of factors that significantly affect ovarian histopathology as defined per counted follicular density has significant clinical implications in assisted reproduction and fertility preservation. Our study contributes to the current knowledge by investigating the macroscopic characteristics of human ovaries and relating them to histopathology, demographics and clinical parameters, associated with the previous medical and obstetrical history.
Conclusions
Macroscopic and clinical factors may be significantly associated with actual ovarian reserve, as estimated by ovarian histopathology. Age was significantly associated with decreased ovarian volume and decreased counted primordial and primary follicles on histopathology. Furthermore, significantly decreased primordial and primary follicles were found among women with a history of a caesarian section.
Strengths and Limitations
The strength of our study is that macroscopic characteristics were measured in surgical or autopsy specimens and not indirectly by means of ultrasound. However, the volume still represents an estimate, as it was calculated by a mathematical formula and the ovaries are not perfect ellipsoids [45]. A degree of error is, therefore, expected, but it should interfere with measurements on both sides. Additionally, the follicular counts were estimated by histopathology through a standard procedure for sample handling and inspection. The results were double-checked by two independent investigators. Additionally, all patients included in the analysis had a history of at least one delivery and proven previous fertility, comprising a healthy and fertile sample. Concerning the limitations, the number of patients is limited due to the content of our protocol. Despite the statistical significance of the findings, due to the limited patients' number and the small absolute differences in follicular counts, the extent of clinical significance is unknown and cannot be determined by the present study [35]. An additional limitation is that the specimens-in some cases-derived from human cadavers. Cadaveric ovarian tissue histopathology may potentially deviate from surgical specimen histopathology. Although this difference has been investigated for other tissue types [46], cadaveric ovarian tissue histopathology has not been previously validated or standardized, and thus, a degree of error may be attributed to the source of the specimens. Our sample, however, predominantly consisted of surgical specimens, and therefore, the contribution of cadaveric specimens to the final outcome is proportionally low. This study does not aim to establish causality, but to highlight potential associations and areas of interest for subsequent studies.
|
2023-02-01T16:08:06.668Z
|
2023-01-28T00:00:00.000
|
{
"year": 2023,
"sha1": "018ae0c3f6a8953f68cd6ea57c77318f627907a7",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2075-4426/13/2/232/pdf?version=1674879863",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "70a88724e99b8ab8ed3c1851f8bded483b6f17ca",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
233619193
|
pes2o/s2orc
|
v3-fos-license
|
Effect of L-citrulline, L-leucine, and multicomponent exercises on body compositions, physical 1 activity, and amino acid concentrations in older Japanese women with low body mass index: A 2 randomized double-blind placebo-controlled study 3
Background: The intake of citrulline (CIT) and leucine (LEU) can stimulate protein synthesis. However, the efficacy of the combined intervention of CIT and LEU intake with exercise on body 18 composition and physical activity (PA) remains unclear. This study aimed to investigate the combined effect of CIT and LEU intake and weight-bearing exercises (WBE) and square stepping exercise (SSE) 20 for 20 weeks on body composition, PA, and amino acid concentration in older women with low body 21 mass index (BMI) (16 to 21 kg/m 2 ). Methods: A total of 23 participants practiced WBE and SSE once a week for 75 minutes and were 23 administered supplement (Ex + CIT·LEU group: CIT 0.8 g and LEU 1.6 g; Ex + Placebo group: 3.5 g 24 carbohydrate) twice a day for 20 weeks. Body composition was measured using dual-energy X-ray 25 absorptiometry. PA, including leisure-time, household, and occupational PA, was assessed using the 26 Physical Activity Scale for the Elderly. Amino acid concentrations in the blood were analyzed by high-
Aging strongly affects the body's composition, particularly body cells, and bone and muscle mass 44 [1]. Muscle mass begins to decrease at age 25, further accelerating at age 50, and by 80 years when it 45 has decreased by an average of 40% compared with age 20 [2]. Asian women (-1.3 kg/m 2 ) and men (-46 1.4 kg/m 2 ) have a lower body mass index (BMI) than Europeans, which may lead to several 47 consequences [3]. A previous study reported that low BMI may lead to poor physical performance in 48 Japanese older adults [4]. In this population, abnormal BMI, either high or low, is a major cause of 49 mortality, and a previous study found a higher mortality rate in older adults with low BMI than in 50 those with normal BMI [5]. The low BMI of older individuals is strongly associated with sarcopenia, 51 and sarcopenia is paired with phenotypes such as decrease in muscle mass [6]. Sarcopenia has been 52 associated with disabilities in the performance of instrumental activities of daily living, mobility 53 limitation [7], poor physical functioning [8], high risk of falls or fractures [9], and mortality [10]. To 54 prevent the loss of body composition, the importance of exercise and a well-balanced diet have been 55 proven for Japanese adults with low BMI. The combined therapy of exercise and amino acid intake 56 has been proven to increase muscle protein synthesis [11] and leg muscle mass [12] in older adults. 57 Therefore, exercise programs such as weight-bearing exercises (WBE), which can be easily performed 58 without equipment at any time, are necessary for older adults. 59 Previous studies have shown that different exercise programs offer different benefits for older adults. 60 The WBE program has been proven to improve bone mineral density [13] and promote muscle 61 activation [14]. By contrast, the square stepping exercise (SSE), an aerobic exercise, has been reported 62 to improve both lower-extremity [15] and cognitive functioning [16] and prevent falls [17] in older 63 adults. Furthermore, the combined intervention of exercise and nutrition has shown more significance 64 in the improvement of the body composition of underweight older adults compared with the single use 65 of either one. 66 Leucine (LEU) is a well-known essential amino acid for protein metabolism and is composed of 67 branched-chain amino acids (BCAA), including isoleucine and valine. Additionally, a unique 68 characteristic of LEU is the stimulation of protein synthesis among BCAA. Through the activation of 69 the mammalian target of rapamycin (mTOR), LEU regulates the signaling pathway of insulin PI 3-70 kinase and stimulates the translational control of protein synthesis [18]. In addition, citrulline (CIT), 71 a non-protein amino acid, is a potent endogenous precursor of arginine. Direct intake of arginine can 72 cause gastrointestinal distress; however, intake of CIT can reduce gastrointestinal distress and promote 73 arginine production [19]. CIT is converted to arginine in the kidneys after ingestion, which is converted 74 into nitric oxide in the ornithine circuit [20]. Arginine increases nitric oxide causing vasodilatation, 75 which helps the circulation of both central and peripheral blood vessels, and increases growth hormone 76 secretion, which promotes protein anabolism and wound healing [21-23]. A previous animal study 77 showed that CIT intake increased protein synthesis and content in muscle [24]. Another study reported 78 that the intake of both CIT and LEU could stimulate muscle protein synthesis, which resulted from the 79 activation of the mTOR complex 1 (mTORC1) signaling pathway [25]. Therefore, the intake of both 80 CIT and LEU, combined with WBE and SSE, can be expected to prevent the decrease in body 81 composition of older adults with low BMI. 82 To our knowledge, no study has examined the effect of the combined intake of CIT and LEU with 83 exercise practice on older women with low BMI. We hypothesized that the simultaneous 84 administration of CIT and LEU with exercise practice may increase amino acid concentration in the 85 blood and improve the body composition and physical activity (PA) of older women. This study aimed 86 to examine the effect of multicomponent exercises (WBE and SSE) and amino acid intake (CIT and 87 LEU) on body compositions, PA, and amino acid concentrations of older Japanese women with low 88 BMI were recruited through regional information magazines (Joyo Living Co., Ltd., Tsukuba, Japan). A 98 screening survey was conducted through telephone interview using self-reported, general health 99 questionnaires. The inclusion criteria were as follows: 1) age ranging from 65 to 80 years, 2) BMI 100 ranging from 16 to 21 kg/m 2 [3,26], 3) no exercise prohibition from doctors, and 4) independent 101 mobility and active participation in the exercise classes of the study. As regards the exclusion criteria, 102 participants were ineligible if they 1) take neurological medicine; 2) have a medical history of 103 comorbid diseases such as diabetes, brain, liver, kidney, heart, and peripheral vascular diseases; 3) 104 qualify to at least one question of the physical strength section (questions 6 to 10) of the "Kihon 105 Checklist" [27], which is a measure used to identify frailty; 4) have excessive intake of alcohol (>60 106 g/day) [28], 5) smoking, 6) have allergies to the supplements administered in this study; 7) have 107 undergone blood collection of more than 400 mL within 10 weeks, or 200 mL within 4 weeks, or have 108 donated blood in the past 2 weeks; and 8) have participated in other clinical studies in the past 10 109 weeks. A total of 43 older women applied for this study, but 15 applicants were excluded according to 110 the criteria; further, two applicants dropped out because of conflicting schedules. Eligible participants 111 were fully informed face-to-face about the study objectives, design, criteria of inclusion and exclusion, 112 intervention of exercise program and supplements, assessments, insurance compensation for injury, 113 withdrawal of consent and privacy protection. Finally, written consent to participation and data 114 publishing was obtained from 26 participants. 115
Study design 116
We commissioned THF Co., Ltd. to divide 26 participants into two groups by the simple 117 randomization method: 13 female participants were assigned to the Ex + CIT·LEU group (exercises 118 and L-CIT and L-LEU) and 13 to the Ex + Placebo group (exercises and placebo). First, random 119 numbers were generated by a computer and distributed to the participants. Random numbers less than 120 0.5 were determined as A (Ex + CIT·LEU group), and numbers greater than 0.5 were determined as B 121 (Ex + Placebo group). No difference in age and BMI was found between the two groups. However, 122 three participants in the Ex + CIT·LEU group withdrew their consent to participate because of personal 123 reasons. Finally, 23 participants were analyzed for this study (Fig. 1). To ensure the efficacy of the 124 double-blinding method, all data (class attendance records, distribution and collection of food, diary, 125 and measurement results) were managed by THF Co., Ltd. Members of each group remained 126 completely anonymous to both participants and researchers until the key codes were revealed after 127 completion of the trials at 20 weeks. Measurements were obtained at baseline (pre-intervention 128 measurement: Pre), after 10 weeks (mid-intervention measurement: Mid), and after 20 weeks (post-129 intervention measurement: Post). 130 131 Within the 20-week intervention period, the two groups equally performed WBE and SSE with the 132 guidance of a professional instructor once a week. Exercise sessions lasted for 75 min (10-min warm-133 up, 25-min SSE, 30-min WBE, and 10-min cool down). WBE consisted of two training patterns 134 (pattern A, five exercises on the chair including knee extension, knee raise, squat, heel raise, and 135 abdominal roll-up; pattern B, five exercises on the mat including pressing towel between palms, supine 136 bridge, side-lying leg raise, pressing towel between knees, and sit-up), and only one set of WBE was 137 performed with each exercise repeated 10 times. In addition, to reduce possible burden of this exercise 138 intervention, which lasts for 6 months, participants were required to practice pattern A or pattern B 139 alternately on a daily basis at home. Then, their performance was evaluated at four levels, i.e., 1, 140 certainly; 2, moderately; 3, slightly; 4, could not. Following this, their daily records were checked and 141 feedback provided once every 2 weeks. SSE, detailed description of which has been introduced in 142 previous studies [15], was practiced on a thin mat measuring 250 cm × 100 cm with 40 squares (25 × 143 25 cm). SSE contained step patterns of forward, backward, lateral, and oblique movements; in addition, 144 step patterns increasingly became more complex and were categorized into six levels: junior, basic, 145 semiregular, regular, senior, and master. Participants were required to memorize the patterns and then 146 step forward continuously without treading on the front and side lines of squares. 147 All participants were required to take 4.1 g of supplements (KYOWA HAKKO BIO Co., Ltd., Tokyo, 148 Japan) with 100 mL of water twice a day (8.2 g) for the duration of the trial. The 16.0 kcal total calorie 149 supplementation in the Ex + CIT·LEU group consisted of 0.8 g CIT, 1.6 g LEU, 0.3 g valine, 0.3 g 150 isoleucine, and 1.1 g carbohydrate. The 16.1 kcal total calorie supplementation in the Ex + Placebo 151 group consisted only of 0.3 g valine, 0.3 g isoleucine, and 3.5 g carbohydrate. These supplements were 152 distributed to all participants once every 2 weeks; their supplement intake was confirmed using self-153 reported diaries. 154 We calculated the participation rate of exercise classes, practice rate of WBE at home, and intake 155 rate of amino acid supplement. 156
Body compositions 157
We measured systolic and diastolic blood pressure plus heart rates (OMRON HEM-7111, Kyoto, 158 Japan). Participants were required to roll up their sleeves, and the sensor of the blood pressure monitor 159 was placed on the area of their left arm where the brachial artery was located; thereafter, systolic blood 160 pressure, diastolic blood pressure, and heart rate were measured. The unit of blood pressure was 161 millimeters of mercury (mmHg) and that of the heart rate was beats per minute (bpm). To measure 162 height, participants were required to stand on the height scale with bare feet and look straight. 163 Centimeter was used as the unit of height, and the value was specified to only one decimal place. As 164 for the body weight, a weighing scale (MC-980A, TANITA, Tokyo, Japan) was used and weight was 165 measured in kilograms. Participants were required to wear light clothes and take off their shoes before 166 measurement. Considering the weight of the clothes, 0.5 kg was subtracted from the obtained body 167 weight and specified to only one decimal place. BMI (kg/m 2 ) is calculated as body weight divided by 168 height squared. Using a dual-energy X-ray absorptiometry (DXA, QDR Discovery Wi, Hologic, Tokyo, 169 Japan) performed by a clinical technician, we obtained data regarding body fat, bone area, bone 170 mineral density, lean mass, fat mass, bone mass, and body mass (the sum of the lean mass, fat mass, 171 and bone mass) at baseline and after intervention. Participants were required to wear light clothes, take 172 off their shoes, and lie in a supine position on the DXA machine for 7 min. 173 Physical activity 174 PA was assessed using the Japanese version of the Physical Activity Scale for the Elderly (PASE), 175 the validity and reliability of which have been analyzed by previous studies [29]. PASE is a 12-item 176 questionnaire that measures the average time spent on daily physical activities during the past 7 days. 177 The 12 items comprised sections on leisure time PA (such as walking, recreational activities, and 178 strength training), household PA (such as home repair and garden maintenance), and occupational PA 179 (such as job and volunteer work). These items are weighted based on the intensity of each activity, and 180 the PASE (total PA) score is the sum of the 12 weighted items [29]. 181
Amino acid concentrations 182
We entrusted the analysis of blood samples to a microbiological institute (Kotobiken Medical 183 Laboratories Inc., Tsukuba, Japan). Participants were instructed to fast for 10 h in preparation for 184 venipuncture of the brachial veins. Blood was collected using the EDTA-2NA containing tubes by 185 nurses. To separate the plasma, 2 mL of blood was collected and centrifuged at 3000 rpm for 10 min 186 at 4°C, and 0.5 ml of plasma was stored at -80°C until analysis. Amino acid concentrations were 187 analyzed using high-performance liquid chromatography (HPLC) with an Acquity Ultra Performance 188 Liquid Chromatography with both bone mineral density (r = 0.66) and CIT change amount (r = 0.72). In the Ex + Placebo 255 group, significantly negative correlations were shown in the intake rate of supplementation with valine 256 (r = -0.59), and practice rate of WBE at home with both LEU (r = -0.70) and isoleucine (r = -0.64). 257 In addition, the participation rate of exercise classes showed significantly positive correlations with 258 arginine (r = 0.59), asparagine (r = 0.73), glutamine (r = 0.81), and histidine (r = 0.57), but 259 significantly negative correlations with height (r = -0.60), leisure time PA (r = -0.72), and LEU (r = 260 -0.57
Ex + CIT·LEU ( n = 10 ) Ex + Placebo ( n = 13 )
Note: Spearman rank correlation, † P < 0.05, ‡ P < 0.001 Change rate of all parameters were calculated using the formula (post intervention value -baseline value) / baseline value) *100 ) # Body fat and citrulline change amount were calculated using the formula (post intervention value -baseline value) the combination of exercise and nutrition intervention is more essential for older women with low 298 BMI to prevent sarcopenia. The efficacy of the combined intervention of exercise and nutrition is 299 controversial. Combined intervention of exercise and LEU-rich essential amino acid for 3 months was 300 reported to increase leg muscle mass in Japanese women with sarcopenia [12]. They defined 301 sarcopenia as BMI ≤22 kg/m 2 and found that exercise improved the muscle mass of older women with 302 low BMI, which is consistent with our study. In another study, supplement intake (6 g CIT maleate, 5 303 g creatine, 3 g LEU, 3 g isoleucine, 1.5 g valine, etc.) and resistance exercise practice for 4 weeks 304 significantly increased body mass and lean body mass compared with the placebo group (digestion-305 resistant maltodextrin) [40]. The participants of the previous study were recreationally active men, 306 whereas our study focused on older women with low BMI, but the effects of CIT and LEU intake and 307 exercise practice were consistent. However, other studies have reported conflicting results. A previous 308 study reported that a high dose of whey protein (45 g) intake, rather than combined intake of low dose 309 whey protein (15 g) and CIT (10 g), promoted postprandial muscle protein synthesis after resistance 310 exercise in older men [41]. Another study reported that CIT intake (0.18 g/kg/day) increased plasma 311 CIT and promoted arginine availability, but exerted no influence on LEU oxidation and whole-body 312 protein synthesis [42]. This finding contradicted a previous study reporting that CIT intake increased 313 arginine availability in the urea cycle, which promotes protein anabolism, indicating that CIT may 314 affect protein synthesis [24,25]. 315 These conflicting findings may result from various doses and durations of CIT and LEU intake. 316 According to a review study, an effective dose for daily CIT supplementation was recommended at a 317 minimum of 3 g to a maximum of 10 g [43]. The dose of CIT administered in our study was less than 318 that of the review study. In addition, the WHO recommended doses for daily intake of LEU (0.039 319 g/kg), isoleucine (0.020 g/kg), and valine ( that CIT and LEU supplements were well tolerated by the participants, and a long-term intake of CIT 331 and LEU effectively increased their plasma levels. In addition, the concentration of tyrosine 332 significantly increased in both groups, but the effect size was larger in the Ex + CIT·LEU group (d = 333 1.12) than in the Ex + Placebo group (d = 0.51). The concentration of phenylalanine increased 334 significantly only in the Ex + CIT·LEU group. Phenylalanine is an essential amino acid converted to 335 tyrosine during protein metabolism, which plays a vital role in transmitting signals between the brain 336 and nerve cells. After a mixed diet (49% carbohydrate, 22% protein, 29% lipid), leucine and 337 phenylalanine-tyrosine tracers are produced in the plasma or intracellularly increased whole-body 338 protein synthesis [45]. With reference to previous studies, LEU intake may have increased 339 phenylalanine and tyrosine concentrations in our study; however, this requires further studies. 340 Our study has several limitations. First, there was sampling bias and a relatively small number of 341 participants. Because our participants were Japanese older women with low BMI, the results may not 342 be generalizable to the population of older women with low BMI. Furthermore, 23 of 35 participants 343 were included in this study based on the exclusion criteria and personal reasons. Therefore, a study 344 with a larger population that adheres to the same intervention design is warranted. Second, a nutritional 345 survey was not conducted among the participants. Results of the main effect of time showed significant 346 differences in most variables of amino acid concentrations, indicating that in both groups, amino acid 347 levels in the plasma increased after the intervention (Pre < Post). However, this may have resulted 348 from the leniency toward the participant's lifestyle during the intervention. Diet, sleep, and other types 349 of exercises were not limited to verify the effect of the combined intervention without interfering with 350 the patient's usual lifestyle. Therefore, rather than diet limitation, it is necessary to conduct a 351 nutritional survey to investigate the increase in amino acid concentrations of the placebo group. Third, 352 the amount of PA was measured using the PASE questionnaire, which is a reliable tool to evaluate 353 different types of PA in the elderly [29]. However, reporting bias may exist, and other methods to 354 obtain objective data, such as using an accelerometer, should be considered in future research. 355
Conclusion 357
This study demonstrated that body weight, BMI, lean mass, bone mineral density, bone mass, 358 household and total PA, tyrosine, and phenylalanine were increased significantly after 20 weeks in the 359 Ex + CIT·LEU group. Our findings suggest that the combination practice of multicomponent exercises 360 (WBE + SSE) and intake of L-citrulline and L-leucine can improve body weight, muscle mass, and 361 PA in older women with low BMI, contributing to the prevention of sarcopenia and frailty. 362 363 List of Abbreviations 364
|
2021-05-05T00:09:29.083Z
|
2021-03-13T00:00:00.000
|
{
"year": 2021,
"sha1": "47106ff87815362f412185f2caebbd3cc1e355b8",
"oa_license": "CCBY",
"oa_url": "https://www.researchsquare.com/article/rs-311313/v1.pdf?c=1631893310000",
"oa_status": "GREEN",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "3450850b7e15f4e9b1af5103ea958fe936d070f6",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Chemistry"
]
}
|
115903324
|
pes2o/s2orc
|
v3-fos-license
|
Development of high strength self-compacting fibre reinforced concrete for prefabricated concrete industry
Prefabricated construction is an emerging industry in Australia and considered a key mechanism to boost productivity in the construction industry. The use of fibre reinforced concrete has a huge potential in the prefabricated industry as the concrete can be delivered straight to the precast mould, eliminating in many cases the steel reinforcement, thus increasing production quotas and cost savings. Such results can be further improved by utilising self-compacting concrete reinforced with fibres. Although the use of steel fibres as reinforcement is now well established, in the precast industry thin walls and shape of the moulds can be a limitation to steel fibre as well as work health and safety concerns for handling. Under such conditions, the use of polymeric fibres can be extremely beneficial, reducing labour hours and placement time as well as improving safety. This paper reports the development of high strength self-compacting fibre reinforced concrete for application in prefabricated concrete industry, exploring the effect of Forta-Ferro and ReoShore fibres on concrete fresh and mechanical properties.
Introduction
Recent developments in precast concrete manufacturing has created a high demand for new concrete types with enhanced properties, both at early and late ages, and fibre reinforced concrete plays an important role in this scenario. Macro high modulus fibres, especially steel fibres, are already used for replacing rebars in nonstructural applications and have a huge potential in the precast industry eliminating in several cases the need of steel reinforcement and improving the production efficiency, reducing the time and labour cost related to installation of steel rebar, and thereby enabling the production of thin precast elements. Steel fibres have various beneficial effects on the mechanical performance of concrete and the use of it has spread out in different applications such as industrial pavements, precast structural elements, tunnel lining, etc. [1]. In addition, guidelines, specifications and test methods for steel-fibre reinforced concrete are well established [2][3][4][5]. In Australia, the national standard for the design of concrete bridges was reviewed and the new release AS 5100.5 [6] now includes design procedures for steel fibre reinforced concrete in a comprehensive way. However, the limitations of steel fibres in terms of unfavourable effect on concrete rheology [7] and reduced workability, difficulty in finishing, hazards in handling, and corrosion of fibres cannot be disregarded. In addition, the stiff steel fibres may be problematic to be used in thin precast elements of different shapes and therefore there is a crescent demand for alternative fibres that can effectively be used in fibre reinforced concrete. However, there are some limitations in terms of the unfavourable effect of steel fibres on concrete rheology [7] and reduced workability, difficulty in finishing, hazards in handling, and corrosion of fibres cannot be disregarded. In addition, the stiff steel fibres may be problematic to be used in thin precast elements of different shapes and therefore there is a crescent demand for alternative fibres that can effectively be used in fibre reinforced concrete. Micro synthetic (polymer) fibres are mainly used to reduce the cracking due to shrinkage and are not as effective in improving the structural properties and postcracking behaviour of concrete with a small volume addition. Macro-synthetic fibres have been recently made available and can be suitable alternatives to steel fibres in some applications, although fibres of one type cannot offer all the improved properties due to the limitations of each type. For example, micro-fibres can be beneficial to arrest the micro-cracks and delay but are not effective in bridging macro-cracks whereas macrofibres can bridge the larger cracks and thereby improve the post-cracking performance of concrete. Optimised combinations (hybridisation) of two or more types of fibres in concrete can develop better engineering properties than using only one type of fibre and can be achieved combining fibres with different shapes, dimensions, tensile strength, Young's modulus, ductility, and bond properties to cementitious matrices [8]. Forta-Ferro fibres are a recently-developed blend of synthetic fibres, designed as a cheaper, non-corrosive and easier-to-finish replacement for steel fibres. The product is a blend of two types of fibres: (i) fibrillated polypropylene fibres to reduce and control shrinkage and temperature cracking, and (ii) very heavy-duty twistedbundle monofilament fibre made of a strong synthetic copolymer, to increase load-transfer and post-crack performance. ReoShore is a high-performance monofilament structural fibre made of 100% virgin polypropylene. The fibre is 45 mm, which is shorter than Forta-Ferro, and has higher stiffness when compared to Forta-Ferro, which may improve concrete toughness and crack control. The fibres used together are expected to complement each other by combining the beneficial effect of stronger and stiffer fibres in the first phase of crack opening to improve the flexural strength whereas the softer fibres in the later phase to improve the residual flexural performance. Self-consolidating concrete (SCC) has fresh concrete properties with high flowing ability without segregation and thereby compacted by the self-weight without any vibration or added compacting effort [9]. These properties of SCC are highly beneficial in terms of reduction of labour cost to compact the concrete with vibration. This is especially true for the use of SCC in the casting of structural elements with congested reinforcement and in manufacturing of thin precast elements. The addition of fibres can further improve the mechanical performance of SCC. Forta-Ferro is already used in a range of commercial precast products produced by Aus Pits in Australia. Aus Pits is an experienced pre-cast concrete manufacturing company, specialising in the supply of custom designed products including pits for storm water, drainage and connections for electrical, communications and data networks. Deakin University was engaged by Aus Pits for a R&D project to extend the existing range of fibre reinforced concrete products beyond the current limits by improving concrete properties. This paper presents an experimental investigation conducted to develop high strength self-compacting fibre reinforced concrete for application in prefabricated concrete industry, using Forta-Ferro and ReoShore 45 fibres as reinforcement. This is an ongoing research project with the industry partner and the paper only highlights the material testing phase of the project.
Experimental program
Four batches of self-consolidating concrete were used in the research, where three different dosages of Forta-Ferro fibres ( Fig. 1-a) were incorporated in three batches and a hybrid combination of Forta-Ferro and ReoShore 45 ( Fig. 1-b) fibres (hybrid fibres) was added to the fourth batch of concrete. All mixes were produced in in a ready mix concrete plant (Local Mix, Geelong, Australia), using a standard concrete mixer truck. To mimic a real scale production, for each batch of concrete 4 m 3 of concrete was produced. Detailed properties of the two fibres are presented in Table 1 and the concrete batch identifications with the corresponding fibre dosages are given in Table 2
Fresh properties
All mixes were designed to flow on the test apparatus in the range of 510 to 590 mm (diameter), with ideal flow of 560 mm in 4 seconds. Figures 2 shows a fibre reinforced self-compacting concrete produced in this work during the execution of the flow test. Visual observations during the test also offer additional information about bleeding capacity and segregation potential. All mixes accepted as self-compacting concrete flowing beyond the minimum flow range, except for the the mix with the highest dosage of Forta-Ferro fibres, FF-HD. Although the mix presented good fibre dispersion and enough paste, the higher dosage of fibres reduced the fluidity of the mix and required longer mixing time to obtain a reasonable mix workability. However the desirable fresh properties were not achieved in that particular mix.
Compressive strength, tensile strength and modulus of elasticity
The 28-day test results for compressive strength, split tensile strength and modulus of elasticity are summarised in Table 3 and the comparison of the results for four concrete batches are plotted in Fig. 3
Flexural strength and equivalent flexural strength
The flexural strength determined from the notched beam testing under three-point bending and Re,3 values for all four concrete batches are summarised in Fig. 5 and 6. The crack mouth opening was measured with CMOD transducer conforming to BS EN 14651 [2]. These results were calculated from the load-CMOD plots (Fig. 7). The increase in fibre dosage from low dosage in FF-LD to high dosage in FF-HD did not improve the flexural strength of concrete, however, residual flexural strength was increased significantly from 36% to 79%, which is clearly Regarding the flexural performance, the technical report 'TR63' by Concrete Society [3] states that the flexural strength and ductility of a concrete beam is only predicted to increase if Re,3 is greater than 0.5. The equivalent flexural strength ratio Re,3 of FF-MD and FF-LD/RS-LD were 0.46 and 0.55, respectively, which are much higher than that of FF-LD and justify much better residual flexural performance of these two batches. However, the values are significantly lower than batch FF-HD (Re,3 = 0.76) due to comparatively lower dosage of fibres used in these two batches. Considering this criterion, it is expected that both FF-HD and FF-LD/RS-LD mixes would improve the flexural performance of the fibre-reinforced pits (the target application of the mixes) eliminating the conventional steel reinforcement. However, much lower compressive strength and unacceptable fresh properties of FF-HD hinder its suitability for target application.
Conclusions
This paper investigated fresh and mechanical properties of four different fibre-reinforced high strength selfcompacting fiber reinforced concrete batches fabricated using three different dosages of Forta-Ferro fibres and a hybrid combination of Forta-Ferro and ReoShore fibres, in order to identify suitable mix(es) for the production of pre-cast components, as a part of a collaborative research project between Deakin University and Aus Pits. Considering overall performance, the hybrid mix FF-LD/RS-LD (low dosage of Forta Ferro + low dosage of RheoShore fibres) seems to be the most suitable mix design out of the four for the intended application. The hybrid mix not only met the criteria for compressive, tensile and flexural strength but also showed significantly improved post-cracking flexural behaviour representing improved ductility. This feature is highly desirable in the design of concrete structures and can be beneficial in the construction of precast concrete products of thin sections by replacing the conventional rebar and thereby reducing the construction time and labour cost.
|
2019-04-16T13:29:40.904Z
|
2019-01-01T00:00:00.000
|
{
"year": 2019,
"sha1": "4928643a1ce622aba3998eb841cadb5b0ea282a9",
"oa_license": "CCBY",
"oa_url": "https://www.matec-conferences.org/articles/matecconf/pdf/2019/24/matecconf_acem2019_02011.pdf",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "43105187f37f655626f8a4d0b3ce5a0d49ec9211",
"s2fieldsofstudy": [
"Engineering",
"Materials Science"
],
"extfieldsofstudy": [
"Engineering"
]
}
|
119266143
|
pes2o/s2orc
|
v3-fos-license
|
Ordering in ternary nitride semiconducting alloys
We present a thorough theoretical study of ordering phenomena in nitride ternary alloys GaInN, AlInN, and AlGaN. Using the Monte Carlo approach and energetics based on the Keating model we analyze the influence of various factors on ordering in bulk crystals and epitaxial layers. We characterize the degree of both short range order (SRO) and long ranger order (LRO) for different compositions, temperatures and for substrates associated with different epitaxial strain. For the description of the SRO the Warren-Cowley parameters related to the first four coordination shells are used. The LRO is detected by means of the introduced sim-LRO parameter, based on the Bragg-Williams approach. The description of the observed long-range ordering patterns and conditions for their occurrence follows.
I. INTRODUCTION
Ordering in semiconducting nitride alloys is recently an intensively investigated issue. Both experimental as well as theoretical activities deal with mixtures of AlN, GaN, and InN, which are important from the application perspectives such as optoelectronics, high-power and high frequency electronics, and sensors. There are three essential issues that trigger such high research interest.
First, there is much experimental controversy over which ordering patterns can be obtained in nitrides and under what conditions do they occur. Actually, various modes of ordering are reported to be observed in the nitride samples. These include clustering and precipitation, 1,2 compositional modulation, 3,4 or uniform random alloy possibly with some degree of short-range ordering. [5][6][7] However, many of the aspects here still remain unclear. For example, recently Galtrey and coworkers, 5 on the basis of three dimensional atomic probe (3DAP) measurements, concluded that there is no evidence of clustering in Ga x In 1−x N except for the natural spatial fluctuations of composition. Based on a previous report by Smeeton,8 they suggested that observed "clusters" could be artifacts related to the radiative damage of the sample during high electron resolution transmission microscopy (HRTEM) observation. However, this argumentation met with serious critique from TEM experimental groups, 9,10 pointing out that the radiative damage impact can be highly minimized throughout the experiment and that during such careful measurements clustering of In atoms still remains visible. In the presence of such ambiguities, modeling could provide a valuable insight into the nature of ordering occurring in nitrides.
The second reason is related to the fact that ordering influences the electronic structure and the optical proper-ties of alloys. These properties, including band gap, carrier localization, mobility, etc., are in turn crucial from the applications viewpoint, which include laser diodes (LDs), light-emitting diodes (LEDs), high-electron mobility transistors (HEMTs), sensors, etc. It turns out that even local short-range ordering significantly influences optical properties of nitrides as well as other semiconducting systems, see e.g. the work of Bellaiche et al . 11 Of course, global long-range ordering has also a pronounced impact on optical properties in nitride alloys, as it was demonstrated, e.g., in Ref. 12. Recently, also Gorczyca and coworkers published a series of papers comparing electronic structure of small supercells calculated assuming either clustered or uniform distribution of cations. [13][14][15] Their studies show significant differences for these extreme distribution cases, for materials including ternary nitride alloys Ga x In 1−x N, Al x In 1−x N, and Al x In 1−x N as well as selected compositions of quaternaries Al x Ga y In 1−x−y N. Yet, another important example of ordering significance for electronic properties is related to the luminescence intensity of In containing nitride samples. There are indications that excitonic recombination occurring at In microclusters could significantly contribute to the luminescence signal. 16,17 Therefore, the presence or absence of In clusters should have considerable impact on efficiency of light-emitting devices.
The third issue has methodological origin. The type and degree of homogeneity in the considered alloy influences the range of methods that can be applied to modeling its properties. Many theoretical approaches assume the perfectly random uncorrelated alloy, characterized by both short-and long-range order parameters equal to zero. Examples of such methods include virtual crystal approximation (VCA), coherent potential approximation (CPA), special quasirandom structures approach (SQS), etc. The question about the presence of correlations and ordering within nitride alloys is, therefore, also very vital in terms of accuracy and applicability of various modeling schemes. Moreover, the detailed knowledge about the ordering patterns could provide more realistic atomistic configuration for input to semiempirical electronic structure computation methods such as tight-binding or empirical pseudopotential schemes. Thus, the overall reliability of nitride alloys modeling can be seriously improved, once more detailed knowledge about ordering is available.
Keeping in mind the above reasons, we try to shed some light on the ordering phenomena occurring in nitrides. We carry out the detailed studies of ordering in different nitride ternary alloys including Ga x In 1−x N, Al x In 1−x N, and Al x Ga 1−x N. We verify how various factors could influence the ordering in these systems. The crucial variables include composition, temperature, and epitaxial strain. Our simulations are performed under assumption of the thermodynamics equilibrium. Therefore, we do not directly address the effects specific to the growth method. The growth of nitride layers is often performed using methods operating in non-equilibrium conditions such as molecular beam epitaxy (MBE). Moreover, usually this epitaxial growth takes place on the surface. Therefore, the features related to surface ordering and details of growth process could in principle remain "quenched" in the sample, forming a metastable state. Phenomena of this kind are not directly included in our modeling. The main computational tool of our investigation is the static Monte Carlo method. This is not the only option. There are also kinetic studies available in the literature for nitrides. 18 The energetics in our calculations is computed on the basis of the Keating valence force field model and, for the sake of simplicity, we focus on the zinc-blende structures. The studies are carried out in the lattice-coherent thermodynamics regime, i.e., we assume that the alloy constituents form common lattice and can not decay into separate components, each with its own independent lattice parameter. This assumption was often not included and there were numerous studies examining the behavior in the so called lattice-incoherent case. These approaches rely mainly on the free-energy difference between alloy and binary constituents relaxed to their own lattice constants. Examples of this type of calculation for ternary [19][20][21][22][23][24][25][26][27][28] and for quaternary nitrides 29,30 are present in the literature. These studies provided qualitatively similar phase diagrams with unimodal curves separating the uniform region in the composition-temperature parameter space. Moreover, they predicted very high critical temperatures T M , above which the alloy components are fully miscible. For the most analyzed case of the ternary Ga x In 1−x N, typical reported values of T M are in the range of T M ≈ 1500-2000 K (see the work of Liu and Zunger 31 for a thorough review of different model findings). These high T M values yield very low estimates for the maximum In composition, which can be incorporated in Ga x In 1−x N in typical growth temperatures, be-fore phase separation occurs. On the other hand, there are experimental findings showing that it is possible to grow Ga x In 1−x N samples containing up to 20-30% of In without triggering significant clustering, which is in disagreement with aforementioned predictions of the latticeincoherent case. Therefore, recently also a different approach was introduced, 31-33 where the lattice coherence is assumed. It turned out that in this case T M decreases to values below typical growth temperatures. Instead of the phase separation, the random alloy phase with some degree of short-range order is predicted. Therefore, in our study, we focus on lattice-coherent case, which seems to correspond better to the alloys obtained during epitaxial growth. Moreover, we also analyze the influence of the stress related to different epitaxial substrates employed, as this can be an important factor suppressing or triggering ordering. 22,23,28,31,34 In order to characterize the degree of ordering we examine equilibrium structures obtained from our Monte Carlo simulations. To quantify the degree of short-range order in the generated samples we use the Warren-Cowley family of parameters. For description and detection of long-range ordering we employ the approach based on the Bragg-Williams long range order measure. We present a complete and systematic study of these parameters in the whole composition range for all ternary combinations Ga x In 1−x N, Al x In 1−x N, and Al x Ga 1−x N. We also examine different temperatures, beginning from growth temperature range to higher values which could correspond to the sample annealing conditions. Moreover, we examine how the above ordering metrics are influenced by biaxial strain related to application of a different substrate for growth process.
The paper is organized as follows. We start with a short overview of the short-and long-range ordering concepts in Sec. II. The commonly employed Warren-Cowley and Bragg-Williams parameters are also briefly discussed there. Next, in Sec. III, the details of the employed computational approach are explained. Section IV describes how the order parameters are computed from the Monte Carlo data. In particular, it introduces the sim-LRO parameter, constructed on the basis of the Bragg-Williams characteristic. The sim-LRO metric proposed in this work is a handy indicator of the long-range ordering emergence during Monte Carlo simulations. The results for the ternary bulk alloys Ga x In 1−x N, Al x In 1−x N, and Al x Ga 1−x N are described in Sec. V. The ordering in biaxially strained epitaxial Ga x In 1−x N layers is studied in Sec. VI. Finally the paper is concluded in Sec. VII.
II. QUANTITATIVE DESCRIPTION OF ORDERING
The concepts of order and disorder are intuitively easily understood. However, there exist many ways of converting this intuition to numbers describing the degree of ordering in crystals. Therefore, to provide a neces-sary background and fix the notation, a brief description of the concepts used throughout the forthcoming analysis of ordering in nitrides is given in this section. For more thorough insight into the ordering phenomena, see classical monographs of Ziman, 35 Ducastelle, 36 or a very informative introductory paper by Klein. 37 A. Long-range order vs short-range order When analyzing the ordering in alloys, it is important to distinguish between two different aspects of this phenomena, namely, the short-range order (SRO) and the long-range order (LRO). The term short-range order is used to describe the preference of certain types of atoms to reside near each other. This effect manifests itself in the form of statistical correlations between occurrences of atomic types on their respective coordination shells. Usually, alloys exhibit non-zero degree of short-range order, as a result of energetic preference towards particular atomic arrangements. The long-range order is related to the development of a global pattern spread throughout the whole crystal. It is worth underlining that in the temperatures above 0 K the observed pattern is never perfect and contains some misplaced atoms. Both types of ordering can be described in a quantitative manner. In the following we recall the definitions of the commonly used Warren-Cowley SRO parameter and the Bragg-Williams LRO characteristic.
B. Warren-Cowley short-range order parameter
One of the most commonly used measures of SRO is the Warren-Cowley family of parameters Γ. 38 In multicomponent alloys, these SRO parameters between components A and B are defined on the basis of conditional probability of finding an A-type atom in the specified coordination shell of a B-type atom: where c A denotes the concentration of A-type atoms in the system. It is straightforward to note that the relation Γ (i) AB < 0 indicates the preference to AB neighborhood on the ith coordination shell, whereas Γ and that for each of the components P the following sum rule holds: Therefore, for the alloy A x B 1−x , we can in principle define four different SRO measures Γ AA , Γ AB , Γ BA , and Γ BB . However, because of one symmetry relation [Eq. (2)] and two sum rules [Eq. (3)], actually only one SRO parameter is really independent. In this paper, we use standard convention and focus on Γ AB . It is worth underlining that, even though technically we deal with ternary alloys A x B 1−x N, they can be described identically as in the binary case. This is because nitrogen lattice remains mostly unaffected, as the probability of substitution in the nitrogen site is much lower than exchanges in cationic lattice because of the energetic reasons. Therefore, we focus only on the cationic fcc sublattice and investigate in a great detail order parameters related to the first four coordination shells within this lattice.
C. Bragg-Williams long-range order parameter
The long-range order must be specified with respect to a certain pattern. To calculate it, e.g., for A x B 1−x binary alloy we have to know which sites should be occupied by atoms of type A, type B, etc. Once we know the spatial pattern (PAT), we can define the Bragg-Williams longrange order parameter as For the SRO parameter, a key role was played by the conditional probability, whereas for the LRO measure the most important variable is f . It denotes the fraction of A sites from pattern PAT, which are actually occupied by A type atoms in the structure under consideration. Symbol c A stands here for concentration of A atoms in the sample. If S (PAT) A ≈ 1, it indicates that the location of A atoms in the structure is similar to pattern PAT. If S (PAT) A = 0, it means that no significant similarity to PAT in terms of A-atom locations was detected. From S (PAT) A < 0, it can be deduced that A atoms in the structure avoid A sites from PAT. For the binary al- , so we have only one independent Bragg-Williams LRO parameter. For the technical reasons, the direct use of the Bragg-Williams parameter for the quantification of the LRO in Monte Carlo simulation data is difficult. Therefore, to detect the presence of the LRO in our results, we develop a derived quantity, the sim-LRO parameter. It is described in greater detail in Sec. IV.
III. COMPUTATIONAL DETAILS
All calculations, were performed for the zinc blende 6 × 6 × 6 supercell containing 1728 atoms. As resulted from our tests, this size of the supercell ensured good convergence of the examined order parameters. See When it comes to the applied simulation techniques, the standard static Metropolis algorithm within the NVT ensemble was applied. The concentration of species within alloy was held constant during the simulation, atomic shifts and cationic exchanges were employed as trial moves. Because of the employment of atomic shifts the lattice vibrational effects are directly included in the performed calculations. The magnitude of atomic shifts was adjusted so that exchange probability was equal to 0.5, as recommended for the most efficient phase space exploration. 39,40 The exchange trials were performed five times more often than atomic shifts, since in the majority of the cases their probability was much lower than for shifts. The total length of simulation was 8.3 million Monte Carlo sweeps per concentration. For each run, we allowed 0.4 million Monte Carlo sweeps for equilibration, before gathering the simulation statistics. The energy calculations were carried out using Keating model 41 and its parametrization for nitrides described in the earlier work of the authors. 42 Periodic boundary conditions were used throughout all presented simulations. For every type of ternary alloy, A x B 1−x N a series of simulations with 17 different concentrations x = 0.056, 0.111, 0.167, . . . , 0.944 was performed. For the initial conditions, random and uncorrelated structures were generated with lattice constant optimized to their energy minimum. For the pseudorandom number generator, the RanLux method was used as implemented in the GSL library. 43 Also tests with different algorithms were performed, showing that the simulation results are insensitive to pseudorandom number generation technique employed. Figure 2 displays the difference between the Γ to the formula The observed values of DΓ (i) GaIn are small. As it can be seen from Fig. 2, they do not exceed a few percent.
IV. DETERMINATIONS OF ORDER PARAMETERS FROM SIMULATION DATA
The computation of Warren-Cowley parameters Γ (i) AB from structures obtained in Monte Carlo simulations is relatively easy. It is enough to estimate the conditional probability from formula (1) by average percentage of type A atoms present in the ith shell of the B atoms in the set of structures generated during each Monte Carlo run. However, as opposed to the short-range order, the situation for the long-range ordering is not so straightforward. To be able to estimate the Bragg-Williams measure S (PAT) A , the pattern PAT has to be known in advance. Its correct selection is crucial for the detection of the long-range order. When calculating S (PAT) A for a highly ordered structure, however, very different from assumed PAT, one may easily obtain values close to zero, indicating the absence of LRO. Moreover, even if the exact pattern is known, one should check for all its symmetry-equivalent variations, since the value of S (PAT) A clearly depends on assumed orientation of PAT. Therefore, in principle, all variants of PAT produced by symmetry operations should be checked, as the variant developed during simulation is usually a result of random fluctuations and can not be predicted a priori. However, it would be very useful to have a quick way of checking whether long-range order of some type develops during Monte Carlo simulation or not, preferably without the a priori knowledge about ordering pattern.
It turns out that, in order to detect the presence of long-range ordering, quantities derived from S (PAT) A are more useful. Instead of a priori selecting ordered configuration PAT, it is helpful to calculate the Bragg-Williams parameter with one of the structures obtained in the developed stage of simulation used as pattern PAT. It is important that this reference structure PAT is selected far after the thermalization phase. The quantity calculated with this method we will call the simulation LRO parameter (sim-LRO) and denote it as S A . The typical behavior patterns of S A during Monte Carlo simulation run are displayed in Fig. 3. They are taken from calculations for Ga x In 1−x N on InN substrate, which are discussed in greater detail in Sec. VI. It turns out that three types of behavior for S A can be distinguished. First, if no LRO is present then S A simply fluctuates around 0 value, corresponding to unordered alloy. This is illustrated in Fig. 3 (a). Second, if certain LRO pattern was developed during simulation, the successive structures fluctuate around it generating values of S A that are considerably different from 0. This is depicted in Fig. 3 (b). Third, sometimes different arrangements of atoms could repeat during simulations, e.g., related to different orientation or shift of LRO pattern. It will correspond to series of peaks visible in S A values during simulation time, see Fig. 3 (c). These peaks correspond to moments when the structure is similar to the one selected for comparison. Even though, the average values of S A throughout the whole simulation could be in this case close to 0, this behavior also indicates the development of LRO or precipitation. Therefore, to detect this type of ordering, except the average value of S A we also analyze the spread of observed S A values. Convenient measure here is the difference of percentiles q related to 5 and 95 percent, estimated on the basis of S A population generated during Monte Carlo (MC) sampling This quantity provides the interval width which includes 90% of the observed data. The occurrence of high values for either average sim-LRO parameter S A or its spread ∆S A , enables for convenient detection of long range ordering in MC simulation results. The average value of S A together with the percentiles forming ∆S A are marked with dashed lines in Figs. 3 (a)-(c).
V. ORDERING IN THE TERNARY BULK ALLOYS
In this section, the ordering in the bulk nitride ternary materials Ga x In 1−x N, Al x In 1−x N, and Al x Ga 1−x N is examined in a great detail. In the analyzed cases, no longrange ordering was observed. The sim-LRO parameter for all presented cases fluctuated around zero, similarly to the situation displayed in Fig. 3 (a). Therefore, we focus here on the short-range ordering. The Warren-Cowley SRO parameters for the first four coordination shells in the cationic lattice are examined in the whole concentration range. Moreover, this study covers a few different temperatures, beginning roughly from the lower end of growth temperatures range, i.e., 873 K, and including higher values up to 1673 K. This upper limit could correspond to annealing during the post-processing phase. In the following the results for each of the ternary combinations, namely, Ga x In 1−x N, Al x In 1−x N, and Al x Ga 1−x N, are described.
GaIn , Γ GaIn , and Γ GaIn on composition for different temperatures are presented in Fig. 4. The most interesting composition range, corresponding to violet-blue-green wavelength 400-570 nm is gray shaded on all graphs. It is easy to observe that Γ (1) GaIn and Γ (4) GaIn are negative in the whole concentration range and for all examined temperatures. This indicates the preference toward Ga-In neighboring on the first and fourth coordination shells.
Conversely, Γ (2) GaIn and Γ (3) GaIn remain positive, which indicates that Ga atom is more likely to have another Ga atom on its second and third coordination shell, than it would result from concentration x. This sign pattern is in qualitative agreement with recent results of Chan and coworkers 33 obtained on the basis of the cluster expansion model. This −/ + / + /− sequence observed for the Γ (i) parameters is usually interpreted as manifestation of stability for the chalcopyrite structure, which exhibits the same SRO signs pattern. 33,45 For each dependence of the short-range order parameter on gallium concentration x, the third order polynomial was fitted. This type of curve is selected as the simplest model capable of describing observed shapes with asymmetric maximum/minimum. These fits are presented as continuous lines in Fig. 4. It turns out that all short-range order parameters dependencies have single extremum, corresponding to maximum absolute value of Γ parameter. On the basis of obtained polynomial fits, concentrations x Table I. For the nearest neighbors and next nearest neighbors, the x (i) ext is located at lower gallium concentration roughly around 35%, or in other words at high indium content 65%. These compositions are outside of the most interesting violet-blue-green range. For the third and fourth coordination shells, these extrema move toward approximately 50% compositions. Also the overall curve shape gets more symmetric around extremum for the case of Γ (3) GaIn and Γ (4) GaIn than it is in the case of the first and second shell. Nevertheless, these higher-order parameters should have weaker influence, e.g., on electronic properties of Ga x In 1−x N than Γ (1) GaIn and Γ (2) GaIn . Therefore, we conclude that the region, where effects related to ordering are the most pronounced is located around x ≈ 0.35. It is also worth noting that the position of x (i) ext is virtually independent on temperature, and the difference for all coordination shells i = 1, . . . 4 does not exceed 0.04. We also calculate the difference of extremum values for each Γ (i) This quantity indicates that the ordering decreases roughly by half when the temperature changes from T = 873 to T = 1673 K. All the above findings are summarized in Table I. As far as modeling of electronic structure for Ga x In 1−x N is concerned, the above results indicate that it would be the most interesting to focus on structures with Γ GaIn = 0 such as, e.g., special quasirandom structures (SQS) or virtual crystal approximation (VCA) could lead to systematic inaccuracy, caused by neglecting of the SRO. The range, where these inaccuracies should be the most pronounced, is determined by extrema for the lowest coordination shells SRO parameter, namely around indium concentration of 65% or so. On the other hand, when considering alloys corresponding to the violet end of spectra, say around 10% of indium, the influence of ordering effects should be much less pronounced. The comparison of the above results with experiment is not straightforward. There is a heated debate about the range of concentrations and, more generally, conditions that lead to In clustering in Ga x In 1−x N samples. This issue so far is by no means clarified, even on the phenomenological level. Our computations, to a certain extent, support these findings that report no clustering or long-range ordering in the samples even with high In concentration. [5][6][7] In our simulations we show that under the lattice-coherent assumption, thermodynamics does not prohibit obtaining mixtures without precipitates or LRO in very broad indium concentration range and in temperatures higher or in the region of the typical growth conditions. However, one has to stress the fact, that this state could be difficult to obtain experimentally, mostly due to artifacts related to epitaxial growth methods. These methods, such as MBE or metalorganic vapor phase epitaxy (MOVPE), take place on the surface, in conditions at least partially corresponding to non-equilibrium. Therefore, some features specific to growth process could form metastable state and remain "quenched" in the samples. To fully describe these phenomena, the detailed growth models should be developed. This is, however, far more difficult than already not-easy computation of equilibrium properties. Yet another issue not addressed on this stage of modeling is the presence of strain, both local and global. By local strain we mean deformations related to defects, in particular dislocations, whereas global strain occurs in thin lay- ers and is related to lattice mismatch between substrate and alloy material. We reference it as global, since it is present in the whole volume of the considered layer. The influence of dislocations and other sources of local strain is not included in our modeling, since we assume perfect crystalline lattice. Taking into account such extended defects would require calculations with much larger supercells with complicated geometry, prohibitive for the detailed composition and temperature scans provided in this study. Nevertheless, we can gain certain insight into the influence of strain on ordering by modeling global epitaxial deformation. It was already indicated in the literature that this epitaxial strain can have significant impact on ordering phenomena. 22,23,28,31,34 Therefore, its role is analyzed in detail in Sec. VI.
B. Ordering in AlxIn1−xN
When speaking of ternary nitride alloys with application prospects, Ga x In 1−x N is not the only option. Recently, also Al x In 1−x N attracts considerable attention. It is particularly interesting, because by adjusting the In composition one can lattice match it to GaN substrate, producing in principle strain-free interfaces and heterojunctions. The absence of strain has many advantages such as lack of electric field component related to piezoelectric effect, lower density of defects on the interface, etc. Therefore, such Al x In 1−x N layers are promising candidates for applications in optoelectronic devices, e.g., for high quality factor microcavities or distributed Bragg reflectors. Moreover, the AlInN/GaN junction could be an important building block for high electron mobility transistors (HEMTs). For more detailed information about properties and applications of Al x In 1−x N, see a recent review paper by Butté and co-workers. 46 In our study of ordering in the ternary Al x In 1−x N, we observe qualitatively similar behavior as in the case of Ga x In 1−x N (compare Figs. 4 and 5). Again, the same alternating pattern of signs for order parameters is observed, i.e., Γ AlIn and Γ (4) AlIn are negative, whereas Γ Table II and the analogous data for Ga x In 1−x N gathered in Table I. Important difference in the case of Al x In 1−x N is that the composition dependencies of Γ (2) AlIn , Γ AlIn , and Γ (4) AlIn exhibit features of bimodal shape, particularly in lower temperatures. Therefore, in these cases the third-order polynomial fits carried out in order to locate extrema, were performed in the narrowed range: x ∈ [0.0, 0.6] for Γ (2) AlIn , Γ AlIn and x ∈ [0.0, 0.7] for Γ (4) AlIn . The bimodal AlIn at the lowest examined temperature T = 873 K are provided. Since in AlxIn1−xN, for Γ (2) AlIn and Γ (3) AlIn two extrema are present, both of these are quoted in the table. The influence of temperature is illustrated using the difference ∆x AlIn , Γ AlIn in the lowest temperature T = 873 K. In these cases, the second maxima are clearly visible and could be easily calculated from polynomial fits. The detailed information about all extrema is provided in Table II. Apart from the bimodal character, other features observed for the SRO parameters in Al x In 1−x N are similar to Ga x In 1−x N. The shape of extremum for Γ (1) AlIn and Γ (2) AlIn is asymmetric and located in the high indium concentration range around 70%, a little bit higher than for the Ga x In 1−x N case. For the highest coordination shells, the extremum gets more symmetric and shifts toward 50% composition. Again for this alloy, the location of extremum concentration weakly depends on temperature and changes maximally around 3%, as presented in Table II.
Since the properties of Al x In 1−x N lattice matched to GaN are of high practical importance, we gather the information about the short-range order parameters of this material in Table III. In the context of wurtzite Al x In 1−x N, usually the aluminum concentration x ≈ 0.82 is quoted as corresponding to lattice matching. 46 In the case of cubic materials and lattice parameters employed in this study, the Vegard's law leads to the concentration x ≈ 0.79. For simplicity, we neglect here the small unphysical lattice bowing effect which is a feature of the Keating model. 42 The composition of the Al x In 1−x N lattice matching GaN, is also marked for convenience in Fig. 5. It turns out that the magnitude of SRO parameters in this case does not exceed 0.05. Interestingly, it is the lowest for the first coordination shell, which should have the greatest impact on the alloy properties. When it comes to modeling implications, the situation here is analogous to Ga x In 1−x N. Our calculations indicate that for electronic structure modeling the configuration with SRO sign pattern −/ + / + /− should be the most interesting. The deviation from random uncorrelated alloy is the biggest for high indium concentration around 70%, therefore, in this region models assuming SRO=0 should have the largest systematic inaccuracy. Finally, the last possible ternary combination for the examined family of nitrides is Al x Ga 1−x N. This material is very well suited for high electron mobility transistors, operating in high-power range as well as ultraviolet light emitters and detectors. It is also a promising building block for biosensors. Since the lattice mismatch and the differences in force field parameters are the smallest in this case, also the ordering effects here are the weakest. For this material, we carried out the studies only for two temperatures 873 and 1673 K. None of the examined Γ (i) AlGa values exceeded 0.01. Therefore, our conclusion is that Al x Ga 1−x N follows quite closely the picture of of uncorrelated random alloy, i.e., with both SRO=0 and LRO=0.
VI. ORDERING IN TERNARY ALLOYS ON THE SUBSTRATE
In the previous section we have examined ordering phenomena in bulk crystals. However, for applications in optoelectronic devices or sensors usually thin epitaxial layers grown on substrate are employed. In this part, we describe, how the presence of substrate associated with biaxial strain influences the ordering phenomena. We focus on the technologically most important case of Ga x In 1−x N and examine its behavior on a variety of substrates. However, before moving on to the discussion of ordering, the basic facts about thin epitaxial layers are summarized below.
The main parameter for epitaxial layer grown on the substrate is misfit strain. For a cubic material, it is defined as where a substrate and a layer denote substrate and layer alloy lattice constants respectively. If we consider the simplest model, the misfit strain is compensated by purely elastic deformation of the thin film material. The thick substrate is assumed to stay undeformed in this approach. See Fig. 6 for illustration of such a situation. Within this model, the deformation of layer unit cell on the interface (in order to match the substrate lattice) is compensated by relaxation in the perpendicular direction. One can approximate the relaxation strain relax associated with accommodation of misfit strain using linear elasticity theory of lattice mismatch between the substrate and epitaxial, material there exists a critical thickness h, beyond which layer starts to relax the strain in the form of misfit dislocations and other defects, instead of purely elastic deformation. This means that because of the high defects density, epitaxial layers beyond h are virtually useless from the point of view of light-emitting heterostructures. However, estimation of the critical thickness is a difficult problem. Various models were proposed for this purpose.
In this paper, in order to get a rough estimate of the critical thickness, we employ a simple approach developed by Matthews and Blakeslee 47 and later on employed to cubic nitrides. 48 It predicts that h is given by the transcenden- where ν is the Poisson ratio of the layer material ν = c 12 /(c 11 + c 12 ). Note that since both misfit strain and Poisson ratio depend on alloy composition, the resulting critical thickness also depends on the proportion of compounds in alloy epitaxial layer. Employed values of lattice constants are given in Table IV, the composition dependencies of the elastic parameters in Eq. (11) were taken from our previous work 42 and are given by For certain compositions related to large misfit strains, the real-valued solutions of the Eq. (11) do not exist, which indicates that fabricating even very thin layers of high crystalline quality is impossible for certain alloy/substrate combinations. However, one has to bear in mind, that accurate modeling of the critical thickness is complicated task in itself. The Matthews and Blakeslee approach is simple, but crude model. Therefore, it gives more qualitative than quantitative picture of the h dependence on composition and limits of the elastic strain accommodation regime.
In the present section we examine the Ga x In 1−x N ternary alloy grown on a large variety of possible substrates, both employed experimentally as well as possibly suitable in the future, including 3C-SiC, zb-ZnO, CaO, AlN, GaN, and InN. Currently, the most promising well lattice matched substrates for the zinc blende nitrides seem to be 3C-SiC or GaN obtained either in the form of thick buffer layer on a different substrate or recently fabricated zb-GaN free-standing crystals. 49-51 However, with technological progress, other options can gain significance in the future. The information about misfits bounds for each of these materials is gathered in Table IV. To simplify the discussion, we separate our results into two groups of substrates related to moderate and large misfit strains, since the observed phenomena in both cases are quite different. The first group consists of GaN, zb-ZnO, and CaO, whereas the second comprises 3C-SiC, AlN, and InN. All the results have been obtained in T = 873 K, corresponding to a typical growth temperature range.
A. Moderate misfits regime
For substrates related to moderate misfit strains (i.e., GaN, zb-ZnO, and CaO), no long-range ordering is observed in Ga x In 1−x N, similarly to the bulk case. Since the intensity of ordering effects in materials is quantified by the absolute value of short-range order parameter GaIn < 0, the grown thin layer behaves closer to the uncorrelated random alloy than the bulk case.
As an example of obtained results, the case of GaN substrate is presented in Fig. 7. The findings for CaO and ZnO are qualitatively similar. The graphs illustrate the comparison of the short-range order parameters Γ (1) GaIn , Γ GaIn , Γ GaIn , and Γ (4) GaIn between strained and unstrained cases. Order parameters are accompanied by the corresponding ∆Γ (i) GaIn values. Also, the misfit strain misfit and the relaxation strain relax dependence on composition are provided for completeness. Finally, the information about the critical thickness h calculated from the Matthews-Blakeslee model given by Eq. (11) is included. This provides insight, as to range of composition is of practical importance.
The relax presented in Fig. 7 (c) has been calculated using two methods: first, within the theory of elasticity (TOE) and Eq. (10), second by relaxing the unit cell to the shape corresponding to minimum energy employing the Keating valence force field (VFF). For GaN and other examined substrates, both methods give similar results. The small differences emerged for compositions related to larger misfit strains. This is intuitively well understood, since the Keating model was constructed to recover the predictions of the TOE for small strains. 41 We also observed that for negative misfit strains (compression in the substrate plane), the theory of elasticity slightly overestimates the relax , whereas for positive misfits (in the case of zb-ZnO and CaO substrate), the relax predicted by the Keating force field is a little bit larger than calculated from elasticity. When it comes to the analysis of the substrate influence on order parameters, it can be clearly seen from Figs. 7 (a) and (b), that the most sensitive to the presence of substrate is Γ (2) GaIn , as the magnitude of ∆Γ (2) GaIn is the largest. Similar behavior is observed not only for the GaN, but also for zb-ZnO, and CaO.
Interestingly, for both negative and positive misfits (present in the case of zb-ZnO and CaO), the ∆Γ (2) GaIn is positive, indicating that the presence of the substraterelated strain increases the ordering within this shell. On the other hand, the quantities ∆Γ (1) GaIn and ∆Γ (4) GaIn are mostly negative, showing the decreasing degree of ordering compared to the bulk case for these coordination shells. The magnitude of this effect is, however, lower than in the case of ∆Γ (2) GaIn . Finally, the observed behavior of ∆Γ (3) GaIn is dependent on the sign of misfit strain. For the negative misfit , the decreased ordering is observed (∆Γ GaIn > 0). Let us also mention that the previous calculations by Liu and Zunger using epitaxial cluster expansion 31,32 predict that already the influence of the GaN substrate triggers the phase separation and long-range ordering in T = 873 K. They obtained the full miscibility in temperatures above T M = 1080 K. Our model predicts similar phenomena, however, in order to observe them substrates related to higher strains are necessary, as discussed in the next section.
B. Large misfits regime
In this section, we describe the behavior of Ga x In 1−x N on substrates that are related to larger strains, i.e., 3C-SiC, AlN, and InN. It turns out that in these cases the long-range ordering can be triggered. This is detected by analyzing the sim-LRO parameter S Ga and its spread ∆S Ga , as described in Sec. IV. For conditions where either S Ga or ∆S Ga are considerably larger than in the bulk case, various ordered structures can be observed in the configurations generated during the Monte Carlo run. The details of formed patterns depend on composition as well as on the magnitude and sign of strain induced by the substrate. For the InN base layer, the Ga x In 1−x N alloy undergoes tensile strain, whereas for the 3C-SiC and AlN the strain is compressive. Generally, both 3C-SiC and AlN correspond to similar misfits as displayed in Table IV; therefore, only the 3C-SiC case will be presented in detail. The effect of the AlN substrate is analogous.
Let us begin with the analysis of the InN substrate. The complete set of characteristic quantities for this case is presented in Fig. 8. Graphs include the dependencies on composition of the long-range order parameter S Ga and its spread ∆S Ga , the short-range order parameters Γ critical thickness. One can easily see that the sim-LRO parameter indicates the presence of the long-range ordering in the range 0.40 < x < 0.90, since either the S Ga or ∆S Ga reaches high values, compared to the bulk case. The observed dependencies of the LRO parameters on Monte Carlo time are in these cases similar to the sample series presented in Fig. 3 (b) and (c). The presence of the LRO in the range 0.40 < x < 0.90 is also indicated by much higher values of the Γ (i) GaIn compared to the bulk case. It is worth noting that the whole violet-blue-green composition region lies completely in the discussed LRO regime for the considered InN substrate. The visual inspection of the structures obtained during simulations reveals that for concentrations around x = 0.5 the so called chalcopyrite (CH) pattern develops in the examined crystal. The ideal chalcopyrite structure is presented in Fig. 9 (a) and compared with one of the structures obtained during simulations in Fig. 9 (b). Note that even though both the deformation due to epitaxial strain and the effect of thermal vibrations are visible in the Monte Carlo structure, its similarity to the perfect chalcopyrite is clear. The concentrations corresponding to the chalcopyrite ordering pattern are marked in Fig. 8 with the CH label. It is well known from the literature that this type of ordering is energetically very favorable in the case of strained semiconductor alloys, not only nitrides. [52][53][54] Therefore, its appearance in our results is consistent with the previous findings. Moreover, the chalcopyrite pattern developed in our simulations is oriented perpendicularly to the substrate plane (CH ⊥ ) as predicted in the recent work of Liu and coworkers for large misfits. 54 The study of the remaining part of the concentration range reveals that the CH pattern is not the only ordering option. It turns out that different atomic arrangement occurs for the higher Ga concentration x > 0.7 . It consists of the In planes perpendicular to the substrate. The concentrations corresponding to this behavior are marked by label PE in Fig. 8. Sample structure with this type of ordering is presented in Fig. 11 (a). Such a behavior can be viewed as a certain mode of phase separation. This is also indicated by the fact that Γ GaIn , Γ GaIn , and Γ (4) GaIn shift toward high positive values, indicating the preference toward Ga-Ga and In-In neighboring. The structures without the CH/PE symbols correspond to mixed ordering carrying certain features of both arrangements. The rest of the concentration range, namely for x < 0.4 and for x > 0.9, manifests no long range ordering, similarly to the case of GaN, zb-ZnO, and CaO substrates. It is also worth noticing that the CH ordering is observed close to the limit of elastic accommodation of strain regime, marked by the existence of critical thickness solutions in Eq. (11). The PE ordering in turn is observed well beyond this regime. Therefore, in reality the PE case might be difficult to observe due to the very poor crystalline layer quality in this misfit region.
The behavior of Ga x In 1−x N on substrates corresponding to large compressive strains also leads to the formation of the long-range ordering. The detailed dependencies on concentration for the long-range order characteristics S Ga and ∆S Ga , the short-range order parameters Γ (i) GaIn , the relaxation and misfit strains accompanied by the critical thickness for SiC substrate are presented in Fig. 10. The analogous results for AlN are not included, since they were almost identical. From our simulations it emerges that for the high gallium concentrations in the range 0.60 < x < 0.85 the long-range ordering can be observed. Even though S Ga remains close to zero, the dependence of ∆S Ga on concentration experiences abrupt change in the ordering region. This change indicates that the S Ga dependence on MC time is similar to the sample presented in Fig. 3 (c). This is also confirmed by the larger values of the short-range order parameters in this regime. It is worth noting that all Γ (i) GaIn become positive, indicating Ga-Ga and In-In neighboring preference. Visual inspection of configuration from the region with LRO reveals that indeed atoms tend to cluster in the planes parallel to the substrate. This is depicted in Fig. 10 as the PA pattern. Such a behavior can be inter-preted as a certain mode of phase separation. The sample structure corresponding to this arrangement is displayed in Fig. 11 (b). For the 3C-SiC substrate, this ordering occurs partially within the region of elastic strain accommodation marked by the existence of solutions for the critical thickness model from Eq. (11), so it should be possible to access it experimentally. Moreover, the whole discussed area of the LRO existence lies within the technologically important violet-blue-green composition region. For the Ga concentrations below the x = 0.6, which correspond to the very high compressive strains, no significant indications of the long-range order are observed. Interestingly, the parameters Γ (1) GaIn , Γ GaIn , and Γ (4) GaIn in this concentration region have absolute values lower than in the unstrained bulk case. The Γ (2) GaIn , in turn, similarly to the moderate strain case, deviates mostly from the bulk ordering and its absolute value increases. These phenomena, however, occur well outside the elastic accommodation of strain regime, where in reality extremely poor crystal quality would be obtained.
VII. SUMMARY
In this paper, the ordering phenomena in ternary nitride alloys have been examined in a great detail. We have investigated how temperature, biaxial strain related to the presence of substrate, and the change of composition in the whole available range could influence type and degree of ordering. Our approach included vibrational contribution, which is not so common in the literature, where restricting to configurational degrees of freedom only is a frequent approximation. As a starting point bulk Ga x In 1−x N, Al x In 1−x N, and Al x Ga 1−x N have been investigated. The Al x Ga 1−x N has been found to follow closely uncorrelated random alloy picture, without both SRO and LRO. For the mixtures containing indium, no signs of long-range ordering or precipitation have been found in the bulk case; however, a considerable degree of short-range ordering has been observed. The short-range order parameters corresponding to the first four coordination shells followed the −/ + / + /− sign pattern, which agrees with the behavior observed in simulations for Ga x In 1−x N on the basis of cluster expansion model. 33 Both materials deviate the strongest from the random uncorrelated alloy at high indium concentrations around 65%-70%. This means that theoretical methods neglecting SRO should yield the largest systematic error in this concentration range. When it comes to resolving the In clustering controversy reported in the literature, 5,7,9,10 our findings could provide a certain support for the experimental results showing uniform distribution of In atoms in Ga x In 1−x N samples. However, the model of bulk alloy might not be fully appropriate here, since examined nitride alloys are usually employed in (opto)electronic devices in the form of strained epitaxial layers. Therefore, the next phase of our studies has been devoted to the influence of the substrate related biaxial strain on Ga x In 1−x N layers. A number of different substrates has been examined and classified into two groups. The first group, associated with moderate strains, contained GaN, CaO, and zb-ZnO. In this case, in the examined temperature T = 873 K, still only SRO has occurred. The second and more interesting group comprised substrates yielding larger misfits, namely AlN, 3C-SiC, and InN. For these substrates, the appearance of the LRO has been detected by introduced sim-LRO parameter. For compressive misfit strain (AlN, 3C-SiC), the parallel planes (PA) pattern has developed for the In content 15%-40%. For tensile misfit strain (InN) the LRO regime has covered the region of 10%-60% of In content. The chalcopyrite ordering (CH) has emerged, when the indium concentration was around 50%, whereas for lower values down to 10%, the perpendicular planes (PE) pattern has been found. The PA and the PE cases represent certain types of precipitation. The above results shed more light on the issue of ordering in nitrides occurring in thermodynamical equilibrium. We believe that the knowledge presented here can facilitate further modeling, e.g., of electronic structure in nitride alloys. It can be also be helpful in the interpretation of experimental findings.
|
2019-04-13T00:07:50.971Z
|
2011-08-11T00:00:00.000
|
{
"year": 2011,
"sha1": "daee2e69b18ccd472f48ad0bdfc27b95d39d4e30",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/1108.2447",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "1d0aa0fb45c15bf37a58e934ee6d4d6df0344f4c",
"s2fieldsofstudy": [
"Materials Science"
],
"extfieldsofstudy": [
"Materials Science",
"Physics"
]
}
|
69925828
|
pes2o/s2orc
|
v3-fos-license
|
Implementation Method CUSUM To Determine The Accident Prone Areas in Web Based
At the Unit Laka Lantas Polres Lhokseumawe determining accident-prone areas on roads in Lhokseumawe still using manual systems. Determination of accident-prone areas are less effective because the police Unit Laka Lantas Polres Lhokseumawe just using the data of the last year and had to repeatedly calculate manually accident-prone areas based on the number of human casualties. In this study the system implementation CUSUM method (Cummulative Summary) to determine the accident-prone areas designed using the web-based programming language PHP. In this system there are three processes to determine areas prone to accidents, the accident rate calculations based on the weighted severity, blacksite calculations using the Z-Score method for determining areas prone to accident blackspots and calculations using the CUSUM method to determine the critical points of an accident. The data used are secondary data obtained from the Police Unit Laka Lantas Polres Lhokseumawe years 2009-2013.
Introduction
Along with the increase of population in the city of Lhokseumawe each year causes the transport demand is also increasing, will indirectly increase the risk of growing traffic problems, such as congestion and accidents, which will have an impact on the decline in the performance of the service road. Traffic accidents, according to Law No. 22 in 2009 was an event on the highway unexpected and unintentional involve with or without other road users, which resulted in human casualties and property loss. Traffic accidents generally occur due to various factors such as: violations or acts not careful users (drivers and pedestrians), road conditions, vehicle conditions, the weather and an unobstructed view. Traffic violation is high enough and private vehicle ownership is increasingly rising, it will indirectly lead to traffic accidents [1].
Based on data from Lhokseumawe City Police, during the period 2009 -2013, there were approximately 160 people who died due to traffic accidents, and 3,154 others were injured (heavy and light). Under these conditions, the city of Lhokseumawe government must immediately address this, because many people assume the crash was not a factor because the fate of the other factors. But on the other hand according to the viewpoint of transportation causes of traffic accidents is in addition to the human factor that does not obey traffic rules are also influenced by the geometric factor of road and traffic arrangements. From these data it is necessary to attempt to reduce the number of accidents. As an initial step necessary to determine accident-prone areas (black site) using the Z-Score and accident-prone spots (black spots) using Cusum (cummulative Summary) [2].
Cusum (Cumulative Summary) is a method that can be used to identify black spots. Blackspot is a point on accident-prone roads (black site). Cusum is a standard statistical procedures as a quality control to detect change from mean values. Cusum of calculation can be known a point where the road is an accident-prone spot or black spot [3]. So far the determination of accident-prone areas by Laka Police Unit Lhokseumawe only based data on the number of accidents that occur based on last year's data. According Austroads (1992), determination of accident-prone areas should ideally take into account historical data. When referring to the Austroads (1992), is not currently available data is systematically accident-prone areas (black site) and accident-prone points (blackspot) in Lhokseumawe City. In addition, people can not know which areas become accident-prone areas in the absence of a web-based system to determine the accident-prone areas in the city of Lhokseumawe [4]. Chen have proposed the Gamma CUSUM for online customer churn prediction [5]. In this implementation, according to Abdullah et al. must consider about the safety and efficiency and will be know with benchmarking process [6] and also the resources involved in the process [7] and for answering the question we must implement an inovative sollution [8]. The performance of CUSUM is significantly influenced by the sensitivity to the presence of seasonal trends in the data considered [9]. Talking about the seasonal trend in the data, then the process of classification of data is an important thing to do. It should be noted that data classification is a process for retrieving minority data which is an interesting pattern to analyze [10]. Thats why we must consider about the class imbalance problem [11] and the data sensitivity [12].
Related Works
Ospina et al. have used CUSUM in basic anaesthesia procedures and obtain satisfactory results [13]. Sibanda and Sibanda used the CUSUM chart method as a tool for continuous monitoring of clinical outcomes using routinely collected data [14]. The research of Chetouani get a satisfactory results in implementing CUSUM Test for Detecting Abrupt Changes in the Process Dynamics [15]. The research of Schuh et al. used CUSUM Control Chart in realt-time occupational safety monitoring. The frequency of accidents within a subset of historical accident data is identified. The performance of the Poisson CUSUM tables and the exponential CUSUM (time between events) is compared in the illustrative example to show that shorter aggregation periods and timing between events lead to earlier indications of increased accident frequency [16] [17] [18].
Research Methodology
In determining the accident-prone areas first step is to find the value of the Z-Score to determine vulnerable road traffic accidents (black site). Z-score is used to standardize the data is the number of accidents each road segment to ensure comparability between the accident on a road section of one over another road. Z-Score calculation using numerical computation result of traffic accidents. After the calculated Z-Score for growth in the number of accidents in recent years, namely in 2013, will now be calculated Cusum method.
Literature study, conducted by collecting and read and understand the relevant reference regarding Accident Prone Area also collects theories from multiple sources such as library books, articles from the internet, journals, and reference of final project regarding the CUSUM method. As for crash data as the data source of the accident victim, the data of roads, the number of accidents and others derived from Lhokseumawe Police Unit Laka. Interview, Retrieving data via interviewing or orally directly with data sources, either through a face to face or by telephone. Of respondents answers summarized by researcher. Analysis system, Authors analyze problems found process Regional determination Troubled Accidents applied previously and perform comprehension particular.
CUSUM Formula
The method of calculation results can be known a point where the road which is the point of hazardous or black spot. Cusum value can be searched by the formula: 1. Finding the mean value (W) Calculation to find the mean value of secondary data, as follows:
Results
In the process of computer-based system design, problem analysis plays an important role in making the details of the application to be developed, the analysis of the problem is a step in understanding the issue before taking action or decision final settlement. During this determination by the accident-prone areas Laka Police Unit Lhokseumawe is still done manually based on data from the past year. The system is currently running felt still less effective and efficient because it must calculate it manually every year and will slow down the performance of Laka Police Unit of Lhokseumawe. With the support of this system, Unit Laka Police Lhokseumawe can manage and update online accident data and can identify accident-prone areas (black site) and points prone to accidents (black spots) and can display it online to the public both inside and outside the city Lhokseumawe.
Public Form is a form that can be accessed by the public or society. Society does not need to log into the system. The public can see information on the number of human victims on the roads in the city of Lhokseumawe. It also can view information areas prone to accidents and accident-prone points on the road section of Lhokseumawe and can see a map of the area roads Lhokseumawe city that never happened a traffic accident. The Public Form can be seen in Figure 1.
Conclusion
From the results of the design and implementation methods cusum (cumulative summary) to determine the webbased accident-prone areas in the city of Lhokseumawe, the author can draw conclusions: 1. This system is used to determine the accident-prone areas based on web in the city of Lhokseumawe. In addition, this system can also be used to determine the critical points of accidents on roads in the city of Lhokseumawe. In the process of decision making using Z-Score to determine Blacksite (accident-prone areas) and methods Cusum (cummulative Summary) to determine Blackspot (accident-prone points). 2. With this system it is part Laka Police Unit Lhokseumawe easily be able to determine areas prone to accidents on the roads in the city of Lhokseumawe. This system can also be accessed by the public so add information to the public about the accident-prone areas on roads Kota Lhokseumawe. 3. Based on the calculation method of the Z-Score and methods Cusum acquired areas that have the highest accident rate of vulnerability is Jl. Medan Aceh -Banda Panggoi Muara Dua Village in Sta 266 -Sta 267 (km 266.00 -267.00 km). While the region has the lowest accident rate of vulnerability is Jl. Desa Bukit Indah Padang Muara Sakti One in Sta 0 -Sta 1 (km 00,00 -00,01 km).
|
2019-02-19T14:08:22.328Z
|
2018-11-01T00:00:00.000
|
{
"year": 2018,
"sha1": "227eddb59c1482e0ed23540f4480140a19017137",
"oa_license": null,
"oa_url": "https://doi.org/10.1088/1742-6596/1114/1/012093",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "7851d5415df9870e89ff7fc5e464264a217c5406",
"s2fieldsofstudy": [
"Engineering",
"Environmental Science",
"Computer Science"
],
"extfieldsofstudy": [
"Physics",
"Computer Science"
]
}
|
13720228
|
pes2o/s2orc
|
v3-fos-license
|
Potential Clinical Benefits of D-ribose in Ischemic Cardiovascular Disease
Cardiovascular disease still remains the leading cause of deaths worldwide. Atherosclerosis, the most common type of cardiovascular disease, has continued to progress due to many factors, genetics, and lifestyles. All cells require adequate adenosine triphosphate (ATP) levels to maintain their integrity and function. Myocardial ischemia commonly found in atherosclerosis can produce lower levels of ATP, which affects not only cellular energy, but also alters normal function. D-ribose, a naturally occurring pentose carbohydrate, has been shown to increase cellular energy levels and improve function following ischemia in pre-clinical studies and have demonstrated potential benefits in clinical evaluations. This review paper presents an overview of ischemic cardiovascular disease and the potential role that D-ribose could play in improving myocardial energy levels and function in the area of ischemic cardiovascular diseases.
Introduction And Background
Cardiovascular disease still remains the leading cause of deaths worldwide for both men and women. There are approximately 61 million individuals afflicted with heart disease in the United States according to the Centers for Disease Control and Prevention. The World Health Organization has reported that approximately 29% of all deaths worldwide are related to this disease state [1]. Atherosclerosis or arterial intimal plaque formation has been shown to progress over time due to factors such as genetics and lifestyles. Obviously, trying to modify genetic factors is not practical; however, changes in lifestyle are plausible. Major lifestyle factors can include consuming a non-healthy diet, smoking, hypertension, lack or inadequate amount of exercise, and potentially continuous stress. Constant efforts have continued to adequately address and to stress to patients that making lifestyle changes could play a positive role in the lower progression of this disease and thereby potentially lower their risk of a cardiovascular event. Precisely, the medical profession has instructed their patients to alter their lifestyle by changing to a healthier diet, regular exercise, cessation of smoking, control of blood pressure, and seek efforts to lower stress. Increasing plaque formation can ultimately limit the delivery of blood and thereby oxygen to viable tissues. Atherosclerosis is not confined to a sole anatomic region, for it is a systemic disease that can involve various arterial regions, such as the heart, cerebrovascular, and peripheral circulation. Clinically significant coronary arterial atherosclerotic disease can produce symptoms of angina or chest pain, commonly during and following stressful situations; however, some patients can even have symptoms at rest. Furthermore, with the progression of coronary artery disease with increasing atherosclerosis plaque formation, these patients are susceptible to sustaining an acute myocardial infarction or live with a chronic, debilitating condition, which over a prolonged period of time, can potentially advance into the development of congestive heart failure. Coupled with coronary artery ischemia and with longstanding hypertension, many patients develop myocardial diastolic dysfunction, an abnormality in ventricular relaxation. Ventricular dysfunction has continued to be a challenging clinical problem because, as of today, there is not yet an approved, solely directed therapy, both pharmaceutical and device-related, targeted at effectively treating this myocardial dysfunctional state [1].
Medical investigations have found that left ventricular diastolic dysfunction is more prevalent than expected. Redfield et al. stated that "systolic dysfunction is frequently present in individuals without cognized congestive heart failure. Furthermore, diastolic dysfunction is common, often not accompanied by recognized congestive heart failure, and is associated with marked increases in all-cause mortality." In 2003, Redfield et al. reported in randomly selected adults over the age of 45 that 21% had mild and 7% had at least moderate diastolic dysfunction. Six percent had moderate or severe diastolic dysfunction with a normal ejection fraction [2]. Bursi et al. found that "more than half of patients with heart failure have preserved ejection fraction, and isolated diastolic dysfunction is present in more than 40% of cases" [3]. Fischer et al. found diastolic dysfunction in 2.8% of individuals between 25-35 years of age with an increased incidence (15.8%) in individuals older than 65 years of age [4]. Compared to women, men had a higher rate of diastolic abnormalities (13.8% vs. 8.6%). Associated factors responsible for the development of diastolic abnormalities have included longstanding hypertension, evidence of left ventricular hypertrophy, and coronary artery disease. Additionally, diastolic dysfunction has been related to a high body mass index, high body fat, and individuals with diabetes mellitus.
Myocardial ischemia alters oxidative cellular metabolism, reflected in lower levels of adenosine triphosphate (ATP). Different states of ischemia, transient versus chronic, can reflect varying degrees of cellular reductions in ATP levels, as well as functional abnormalities and altered cellular homeostasis, and if severe enough, cellular viability. Myocardial ischemia is a common underlying cause in patients with congestive heart failure, and many of these patients also experience a degree of diastolic dysfunction. Clinically, at least 50% of patients diagnosed with congestive heart failure will have various degrees of diastolic dysfunction solely or potentially combined with systolic dysfunction. Within the last decade, Ingwall and Weiss proposed that the failing heart is energy starved [5]. This theory had been previously proposed but was not actively pursued because it was unclear whether ATP levels, in fact, were really decreased; if these levels did decrease in the failing heart, many physicians and scientists thought that the remaining pool of high energy phosphates should be sufficient enough to satisfy myocardial energy-requiring reactions. The advancement in biophysical tools, such as nuclear magnetic resonance spectroscopy and positron emission tomography scans, have aided in providing a better understanding of this myocardial energy/functional relationship.
Review Cellular bioenergetics and function
Every cell requires adequate levels of high-energy phosphates to maintain their integrity and function. This cellular supply of energy (precisely, ATP) usually meets demands. Adenosine triphosphate is produced by intracellular pathways, such as glycolysis and the tricarboxylic acid cycle, with glucose as the starting substrate. However, some cells can also rely on alternative pathways, such as the pentose phosphate pathway, for the production of ATP [6]. There is a direct relationship between adequate myocardial ATP levels and the development of ventricular diastolic dysfunction. Calcium plays a major role in this interaction. Adenosine triphosphate provides the energy for this interaction between cytosolic calcium and the sarcoplasmic reticulum. Depleted ATP levels can result in calcium remaining fixed to troponin longer in diastole, producing a state of diastolic dysfunction or altered ventricular compliance. Following ischemia, a return in diastolic function may be limited by the availability of the highenergy phosphate, ATP [7]. Theoretically, efforts to maximize the recovery of myocardial ATP levels during and following ischemia could aid in minimizing the functional untoward effects of ischemia.
When cells are subjected to ischemia or hypoxia, the normal production of energy compounds is reduced. Myocardial ischemia depletes ATP levels, which can affect intracellular reactions and the cells' function. There are degrees of ischemia, and if severe enough, the viability of the cell is jeopardized. Researchers have also demonstrated that this depletion in myocardial ATP levels following ischemia can last for a considerable time period due to a slow adenine nucleotide synthesis, which is reflected in an alteration in function [8]. Zimmer reported that an extended time period was required for the recovery of depressed myocardial energy levels, as well as an improvement in the alteration in mechanical function, following ischemia [9]. Many of these myocardial ischemic studies involved "isolated hearts", which have limitations due to the nature of the isolated, excised heart preparations. Therefore, an intact acute animal model has also been developed to assess and confirm the findings observed in the "isolated heart" models. Short-term investigations involving an intact animal model found similar findings in myocardial adenine nucleotide levels following ischemia. Reibel and Rovetto reported in isolated perfused rat hearts that a moderate ischemic insult resulted in a 50 -70% decrease in myocardial ATP levels [10]. Similarly, Reimer et al. reported that 12 to 30 minutes of myocardial ischemia produced a substantial drop in ATP levels with a significant decrease in total adenine nucleotides, and days were required for total recovery [11]. However, to better appreciate this important energy-functional relationship following ischemia, a chronic animal model was developed to provide the means for long-term assessment of myocardial energy levels and function during and following ischemia [12][13], (Ward HB, Kriett J St.Cyr J, et al.: Relationship between recovery of myocardial ATP levels and cardiac function following ischemia. J Am Coll Cardiol. 1984, 3:544 abstr.). In this chronic canine model, 20 minutes of normothermic, global myocardial ischemia produced a significant 50% decrease in ATP levels, which was accompanied by a state of left ventricular diastolic dysfunction. Furthermore, the data from this chronic animal study confirmed that myocardial ischemia has a long-term effect. These researchers reported that there was a substantial delay, over one week, in myocardial energy levels and functional recovery following this moderate, reversible ischemic insult [12][13]
Adenine nucleotide metabolism and D-ribose
Ward et al. reported that myocardial precursor availability is an important limiting factor in the recovery of myocardial ATP molecules following ischemia [13]. Adenine nucleotide catabolism eventually produces precursors, which can be lost from the cell, and therefore, inhibits recovery of ATP (Figure 1). The ultimate repletion of myocyte ATP levels following ischemia depends on the availability of the precursor phosphoribosyl-pyrophosphate (PRPP), and D-ribose has shown to accelerate this nucleotide synthesis by directly increasing PRPP levels [14]. Various adenine nucleotides have been investigated for their potential to replenish post-ischemic lower levels of myocardial ATP, such as adenosine, 5-amino-4-imidazole carboxamide riboside, inosine, adenine, and D-ribose. Not only precursors have been evaluated, but also the use of adenine degradative enzymatic inhibitors for the maintenance or accelerated replenishment in energy levels following ischemia. The majority of these substrates and inhibitors have, at best, shown minimal improvement in both ATP recovery and function. However, D-ribose supplementation has been shown to offer great benefits. D-ribose, a naturally occurring pentose carbohydrate, has repeatedly demonstrated in numerous animal studies its significant benefit in enhancing the return in ATP levels and improving diastolic function following myocardial ischemia.
Metabolism of D-ribose
D-ribose can produce nucleotides and metabolic intermediates, such as ribose-5-phosphate. Ribose-5-phosphate is an important intermediate of the pentose phosphate pathway, also known as the hexose monophosphate shunt or phosphogluconate pathway (Figure 2).
FIGURE 2: Pentose phosphate pathway
F-6-P: fructose-6-phosphate; F-1,6-diP: fructose-1,6-diphosphate; GAP: GTPase activating protein; G-6-PDH: glucose-6-phosphate dehydrogenase; G-6-PGDH: glucose-6phosphogluconate dehydrogenase; NADP+: nicotinamide adenine dinucleotide phosphate; NADPH + H: Nicotinamide adenine dinucleotide phosphate (reduced form) plus hydrogen; NADPH: nicotinamide adenine dinucleotide phosphate; PRPP: phosphoribosyl-pyrophosphate; R-5-P: ribose-5-phosphate; Ru-5-P: ribulose-5-phosphate; 6-P-G: 6-phosphogluconate Ribose enters the pentose phosphate pathway by being phosphorylated to ribose-5-phosphate by ribokinase. Ribose-5-phosphate can be utilized in different ways: for the synthesis of glucose, for glycolysis, and in the synthesis of pyrimidine nucleotides or purine nucleotides. Ribose is the substrate for the formation of PRPP, a precursor for de novo synthesis of ATP ( Figure 3). In the de novo pathway, nucleotide bases are assembled from simpler compounds and then attached to a ribose molecule, whereas in the salvage pathway, nucleotides are synthesized by recycling intermediates in the degradative pathway for nucleotides, as well as bases, with the addition of a ribose unit (Figure 4). There exists an interrelationship between the de novo and salvage pathways with PRPP playing a significant role. Ribose plays a vital role in myocardial metabolism, largely through its participation in the formation of PRPP leading to the synthesis of adenine nucleotides, mainly ATP. Supplemental ribose is unique in that it bypasses the rate-limiting enzymatic step in the pentose phosphate pathway, glucose-6-phosphate dehydrogenase, thereby directly increasing PRPP levels, leading to an enhanced production of myocardial adenine nucleotide biosynthesis. This process generates ATP molecules at an enhanced rate, which offers benefits in not only the cell's bioenergetics status but also in the functional aspect of the cell.
D-ribose investigations
Zimmer and Gerlach found in isolated adult rat heart preparations that D-ribose supplementation produced an increased rate of adenine nucleotide synthesis following ischemia [15]. Pasque et al. also reported similar results in isolated, perfused working rat hearts subjected to 15 minutes of ischemia. D-ribose supplementation improved the recovery in myocardial ATP levels, as well as an improvement in functional recovery following ischemia [16]. Using a chronic, canine model, St.Cyr et al. reported the benefits of D-ribose supplementation when used with adenine supplementation in producing an 85% return in ATP levels by 24 hours as compared to no ATP recovery in controls following 20 minutes of global myocardial ischemia [17]. In a separate study using a chronic in vivo canine model, Schneider Significant coronary arterial ischemia has the potential to cause an acute myocardial infarction, which can result in functional compromise and over time can progress to the development of heart failure. The remote myocardium, adjacent but not involved in the infarcted tissue, is subjected to an increased workload, further taxing its energy supply. Zimmer et al. found a progressive decline in left ventricular systolic pressure, a decline in left ventricular dP/dTmax, an elevation in left ventricular end diastolic pressure, and lower cardiac outputs and stroke volume indices post-infarction. D-ribose supplementation post-infarction resulted in an improvement in the above measured hemodynamic parameters with stimulation in adenine nucleotide synthesis [18]. Similar findings were observed when Befera et al. The positive findings of D-ribose supplementation in ischemic cardiovascular disease in the initial pre-clinical animal studies generated interest in investigating its potential role in patients afflicted with ischemic cardiovascular diseases, both as a potential diagnostic and also as a therapeutic agent. Its diagnostic potential in identifying hibernating myocardium, found in ischemic coronary artery disease, as well as its potential benefits in patients with ischemic congestive heart failure, and during and following cardiovascular surgery have also been observed. As of today, oral D-ribose is referred to as a supplement, not requiring a prescription. Clinical patient study applications have commonly used approximately 5 gm/dose of D-ribose supplementation, multiple times/day.
Hibernating myocardium represents regional areas of myocardial dysfunction due to the prolonged hypoperfusion or ischemia. This state has the potential to be reversed upon restoration of oxygenated blood flow to these regions. Theoretically, Thallium-201 scans, dobutamine stress echocardiography, positron emission tomography, and magnetic resonance imagery can aid in the identification of these areas in the myocardium, which can provide helpful information in management strategies for revascularization. Hibernating myocardium is likely to be associated with lower levels of ATP; therefore, supplementation with an adenine nucleotide agent, such as D-ribose, might aid in increasing identification of these viable regions. Pre-clinical animal studies employing D-ribose supplementation demonstrated this benefit in enhancing the identification of these areas of hibernation. Subsequent clinical studies by Perlmutter et al. reported that D-ribose identified more reversible defects using Thallium-201 scan technology [20]. Furthermore, Hegewald et al. found that D-ribose supplementation also increased the detection of viable ischemic myocardial regions using SPECT thallium imaging [21]. More recently, Sawada et al. reported that D-ribose supplementation provided anti-ischemic effects with improving the identification of wall motion dysfunctional abnormalities found during dobutamine stress echocardiography [22].
Previously reported positive pre-clinical studies demonstrated the enhancement of D-ribose supplementation in regenerating depressed ATP levels and improved function following myocardial ischemia [1]. Pliml et al. argued that since there is significant impairment in ATP levels in the ischemic myocardium, D-ribose supplementation might offer a potential role in enhancing deficient ATP levels in patients with ischemic coronary artery disease. They investigated D-ribose supplementation in patients with stable coronary arterial disease by using serial treadmill exercise testing. D-ribose supplementation demonstrated significant benefits in increasing treadmill exercise time before the onset of angina and/or the development of ischemic electrocardiographic changes during exercise [23].
Long-term myocardial ischemia does play a factor in the development of congestive heart failure and further reports have proposed that the failing heart is energy starved [5]. Since preclinical animal studies have demonstrated that D-ribose supplementation enhanced the recovery of myocardial ATP levels and improved diastolic dysfunction following ischemia, Omran et al. investigated the role of D-ribose supplementation in Class II-III ischemic, congestive heart failure patients. They found an improvement in diastolic dysfunction using echocardiographic parameters. They observed an improvement in atrial contribution to left ventricular filling, a smaller left atrial chamber size, and a shortened E wave deceleration. Subjectively, these patients also experienced an improved quality of life and physical function [24]. Bayram et al. more recently demonstrated an improvement in tissue Doppler velocity, which was maintained at nine weeks in 64% of Class II-IV congestive heart failure (CHF) study patients when taking D-ribose supplementation. Approximately 45% also showed an improvement in their early diastolic filling velocity (E) to early annulus relaxation velocity (E') [25]. ). More recently, oral D-ribose supplementation was added to a metabolic designed protocol for patients undergoing "off pump" coronary artery bypass revascularization. Perkowski et al. observed that the addition of D-ribose resulted in lower mortality and morbidities, along with a significant early postoperative improvement in cardiac index following revascularization [27].
Conclusions
Cardiovascular disease still ranks as the leading cause of deaths worldwide, even with continued advances in cardiovascular therapeutic technologies. Myocardial ischemia, commonly found in cardiovascular disease, plays a major role a majority of these deaths. There continues to be strong efforts to educate the population in addressing atherosclerotic risk factors, such as refinements in a more healthy diet, regular exercise, smoking cessation, blood pressure control, and stress reduction.
All cells require adequate levels of high-energy phosphates to maintain their integrity and function. Numerous studies have demonstrated that myocardial ischemia lowers ATP levels, which can be reflected in alterations in diastolic function. Myocardial ischemia, which lowers ATP levels and causes a dysfunctional state, is a main category in cardiovascular disease. Myocardial adenine nucleotide precursor availability is an important limiting factor in the recovery of myocardial ATP molecules following ischemia, for which various substrates have been investigated. D-ribose, a naturally occurring pentose carbohydrate, has been shown in both pre-clinical and in pilot clinical studies to enhance the recovery of ATP levels and also to aid in improving left ventricular diastolic dysfunction following ischemia. These studies have shown its potential clinical benefits in acute and chronic myocardial ischemic conditions, enhancing the identification of hibernating myocardium, aiding congestive heart failure patients, and its perioperative use in cardiovascular surgery. Future, larger cohort, multicenter clinical investigations are mandatory to further substantiate these initially observed benefits in patients afflicted with ischemic cardiovascular disease.
|
2018-05-21T20:27:36.721Z
|
2018-03-01T00:00:00.000
|
{
"year": 2018,
"sha1": "2d4001bcb3fa0397288160d30b22cc98f5dec27e",
"oa_license": "CCBY",
"oa_url": "https://assets.cureus.com/uploads/review_article/pdf/10915/1566922984-20190827-4135-1dehgl4.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "8f48379e2555f76076408ae0a359d5bc963e18e6",
"s2fieldsofstudy": [
"Biology",
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
250497831
|
pes2o/s2orc
|
v3-fos-license
|
Identification of Effective Diagnostic Biomarkers and Immune Cell Infiltration in Atopic Dermatitis by Comprehensive Bioinformatics Analysis
Background: Atopic dermatitis (AD) is a dermatological disorder characterized by symptoms such as chronically inflamed skin and frequently intolerable itching. The mechanism underlying AD development is still unclear. Our study aims to identify the diagnostic and therapeutic biomarkers for AD and provide insight into immune mechanisms at the molecular level through bioinformatics analysis. Methods: The GSE6012, GSE32924, and GSE36842 gene expression profiles were obtained for analysis from the Gene Expression Omnibus database. Differentially expressed genes (DEGs) were segregated using the “Batch correction” and “RobustRankAggreg” methods. Weighted gene co-expression network analysis (WGCNA) was performed to screen for module genes with AD traits. Then, common DEGs (co-DEGs) were screened out via combined differential expression analysis and WGCNA. Functional enrichment analysis was performed for these co-DEGs using Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG), followed by protein-protein interaction network analysis. Candidate hub genes were identified using the “cytoHubba” plugin in Cytoscape, and their value for AD diagnosis was validated using receiver operating characteristic curve analysis in the external database GSE120721. Immunohistochemical staining was performed for further validation. The CIBERSORT algorithm was used to evaluate skin samples obtained from healthy controls (HCs) and lesions of AD patients, to determine the extent of immune cell infiltration. The association between the identified hub genes and significant differential immune cells was analyzed using Pearson correlation analysis. Results: A total of 259 DEGs were acquired from the intersection of DEGs obtained by the two independent procedures, and 331 AD-trait module genes were separated out from the blue module via WGCNA analysis. Then, 169 co-DEGs arising from the intersection of the 259 DEGs and the 331 AD-trait module genes were obtained. We found that co-DEGs were significantly enhanced in the type I interferon and IL-17 signal transduction pathways. Thirteen potential hub genes were identified using Cytoscape. Five hub genes (CCR7, CXCL10, IRF7, MMP1, and RRM2) were identified after screening via external dataset validation and immunohistochemical analysis. We also identified four significant differential immune cells, i.e., activated dendritic cells, plasma cells, resting mast cells, and CD4+ naïve T cells, between AD patients and HCs. Moreover, the relationship between the identified hub genes and significant differential immune cells was analyzed. The results showed that the CCR7 expression level was positively correlated with the number of CD4+ naïve T cells (R = 0.42, p = 0.011). Conclusion: CCR7, CXCL10, IRF7, MMP1, and RRM2 could be potential diagnostic and therapeutic biomarkers for AD. CCR7 expression level was positively correlated with the number of CD4+ naïve T cells in AD. These findings need to be corroborated in future studies.
INTRODUCTION
Atopic dermatitis (AD) is a dermatological disorder that presents as chronically inflamed skin and often intolerable itching (Langan et al., 2020). Globally, it is one of the most prevalent skin conditions, affecting approximately 2.1-4.9% of adults and 20% of children (Nutten, 2015;Barbarot et al., 2018). AD causes severe psychological and social hardships and is associated with a high risk of depression, anxiety, work absenteeism, and suicidal tendencies (Eckert et al., 2017;Patel et al., 2019;Sandhu et al., 2019;Schonmann et al., 2020).
The occurrence of AD could be attributed to various factors, including environmental factors, hereditary tendencies, epidermal dysfunction, skin microbiome abnormalities, and immune dysregulation. Immune dysregulation plays a substantial role in the progression of AD . The main immunological mechanisms underlying AD have been attributed to an imbalance of T helper (Th)1/Th2 differentiation. A Th2 cell-mediated response triggers the acute phase, while the chronic phase is triggered by a shift to a Th1 cell-mediated response (Salvati et al., 2021). Damage to the epidermal barrier results in penetration of the skin by antigens, causing the release of alarmins such as interleukin (IL) -25, IL-33, and thymic stromal lymphopoietin (Salimi et al., 2013). These events result in the stimulation of innate type 2 lymphoid cells and inflammatory epidermal dendritic cells. Naïve CD4 + T cells also gravitate towards Th2 cells, following the generation of IL-4, IL-5, and IL-13 (Cosmi et al., 2019). In addition to triggering type 2 inflammation, immune cells including Th1, Th17, and mast cells are involved in mixed inflammatory responses (Ho and Kupper, 2019;Abreu and Kim, 2021). However, the immune mechanisms underlying AD have not been described adequately. Investigation of the role played by immune cells and the crucial genes related to associated immune responses needs to be conducted in a methodical manner, and was the focus of the current study.
Several novel genes and biomarkers have been identified for various diseases following the widespread use of bioinformatics analysis techniques and the development of various algorithms.
Weighted gene co-expression network analysis (WGCNA) is a sophisticated data processing tool that is widely used to detect clusters or modules of highly correlated genes. It can identify a cluster of genes with similar biological functions, and it can be used to investigate connections between clinical characteristics and gene expression (Langfelder and Horvath, 2008). Besides, an algorithm known as cell-type identification by estimating relative subsets of RNA transcripts (CIBERSORT) has been developed for estimating numbers of immune cells (Chen et al., 2018). This tool has been utilized to measure the extent of immune cell infiltration in various immune-mediated skin disorders, including psoriasis and acne vulgaris Zhang et al., 2021).
Here, we aimed to identify diagnostic and therapeutic biomarkers for AD via the analysis of gene expression omnibus (GEO) database using R packages and online bioinformatics tools. We also aimed to determine the pathways related to AD pathogenesis and mechanisms of immune cell infiltration. We believe this is the first study to apply both the WGCNA and CIBERSORT methods for integration of datasets from various platforms to understand the molecular mechanisms involved in AD, and to verify the results via immunohistochemical method. We are confident that this study would provide a new direction to research on AD management.
Data Acquisition and Workflow
The GEO database (http://www.ncbi.nlm.nih.gov/geo) was used to segregate gene expression profiles. Gene expression profiles were included if 1) high-throughput sequencing or array-based mRNA expression profiling had been performed; 2) it was possible to obtain lesion skin tissues and healthy skin tissues from AD patients and healthy controls (HCs), respectively; and 3) a minimum of six specimens were included in the dataset. Datasets with drug, placebo, vehicle administration, or other treatments were excluded. Finally, the datasets GSE6012, GSE32924, GSE36842, and GSE120721 were selected for use in our study. The data are summarized in Table 1 and the workflow is depicted in Figure 1.
Data Pre-processing
Series matrix files and platform annotation information were downloaded for three microarray datasets, namely, GSE6012, GSE32924, and GSE36842. The three-probe expression matrixes were converted into gene expression matrixes using Perl script. The average value was derived if there was more than one probe corresponding to a gene. Batch effects were then eliminated using the R package "sva." Thus, a merged, normalized gene expression matrix was obtained for further NR, not reported. Frontiers in Molecular Biosciences | www.frontiersin.org July 2022 | Volume 9 | Article 917077 3 analysis. The gene expression matrix for the external validation dataset GSE120721 was obtained in a similar manner.
Construction of WGCNA
The "WGCNA" R package was used to derive modules identified with clinical characteristics, and to identify hub genes. The merged gene matrix was checked to eliminate anomalous specimens which might have escaped from clustering of samples. The topological overlap matrix (TOM) was derived from the adjacency matrix. Genes were divided into modules based on their dissimilarity in TOM. The cut height, minimal module size, and soft-thresholding power were set at 0.25, 30, and 4 (scale-free R 2 = 0.85), respectively. The absolute value of the correlation coefficient between traits and genes was used to determine the relationship between clinical traits and module eigengenes. A gene was considered to be significant based on the absolute correlation coefficient between traits and genes. A gene was considered to be a member of a module based on its association with certain expression profiles. Genes with a gene significance (GS) > 0.2 and module membership (MM) > 0.8 in the module most relevant to clinical traits were considered as module genes with AD traits (Hu et al., 2020).
Screening for Differentially Expressed Genes
Two independent methods were used to obtain the most robust differentially expressed genes (DEGs) from several microarray cohorts. In the first approach ("Batch correction"), the three downloaded raw datasets were combined into an expression matrix after the elimination of batch effects and normalization. Then, DEGs were analyzed using the R package "limma". In the second approach ("RobustRankAggreg, RRA"), the R package "limma" was used to analyze the DEGs of the downloaded gene matrixes. The DEGs from each such derived dataset were then combined using the R package "RRA" . The final DEGs identified by the two methods were determined from their intersection using a Venn diagram. A |log fold change| > 1 was set along with an adjusted p value <0.05, as the threshold points for DEGs.
Functional Annotation and Pathway Enrichment Analysis
The gene from the module of interest and the final DEGs obtained by the two methods were intersected to identify the common DEGs (co-DEGs) using a Venn diagram. Then, the co-DEGs were subjected to functional enrichment analysis. Gene names to gene ID conversion of the co-DEGs was achieved using the R package "org.Hs.eg.db." Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis for co-DEGs was conducted using the R package "clusterProfiler". Metascape (http://metascape.org/gp/index.html) was then used for comprehensive data analysis. Significance was set at a p value <0.05 and q value <0.05.
Identification and Validation of Hub Genes
The protein-protein interaction (PPI) network for the co-DEGs was constructed using the STRING tool (https://string-db.org/). The interaction file was downloaded and each node gene was scored by 12 algorithms (Maximal Clique Centrality (MCC), Density of Maximum Neighborhood Component (DMNC), Maximum Neighborhood Component (MNC), Degree, Edge Percolated Component (EPC), BottleNeck, EcCentricity, Closeness, Radiality, Betweenness, Stress, and Clustering Coefficient), using the "cytoHubba" plugin of Cytoscape (v 3.7.2). Candidate hub genes were identified from the genes at the junction of these 12 algorithms, then visualized using the R package "UpSet". The preciseness of these genes was evaluated in the gene expression profile GSE120721 by ROC curve analysis using the R package "pROC". Hub genes were validated with an area under the curve (AUC) criterion of >0.8 and a differential expression level significance of p < 0.05 using the Wilcoxon test (Sun et al., 2021).
Evaluation of Immune Cell Infiltration and Correlation Analysis Between Hub Genes and Immune Cells
The immune cell proportions were determined using the CIBERSORT R package (http://cibersort.stanford.edu/ download.php), with data from the normalized gene expression matrix of AD patients and HCs and the data from expressed reference signature genes (LM22, http://cibersort. stanford.edu/download.php) . A violin plot was used to visualize the distribution of these immune cells using the "vioplot" package in R software. A significance criterion of p < 0.05 was used for this calculation. The correlation between the identified hub genes and the infiltrating immune cells was conducted by using Pearson correlation analysis. The resulting relationships were studied using the "ggplot2" and "ggpubr" packages.
Patient Recruitment
The study proposal was reviewed and approved by the Ethics Committee of Shengjing Hospital of China Medical University. AD patients and HCs were included in the study after obtaining written informed consent. AD was diagnosed as per the Hanifin-Rajka criteria (Hanifin and Rajka, 1980). Full thickness skin samples were collected from 12 individuals. These included six AD patients and six HCs matched for age, sex, and race. Details of the study subjects are shown in Supplementary Table S1.
Immunohistochemical Verification
Extracted skin tissues were fixed with 4% formaldehyde buffer, and 4-μm-thick sections were obtained from paraffinized specimens. Tissue sections were incubated at 60°C for 2 h before the dewaxing process. For antigen retrieval, the sections were autoclaved in a citric acid buffer (pH 6.0) at 115°C for 3 min and quenched in 0.3% H2O2 for 15 min for endogenous peroxidase activity. Then, sections were treated with goat serum, which was used as a blocking solution, for 45 min, and incubated overnight at 4°C with primary antibodies against CCR7 . These sections were treated with the goat anti-rabbit secondary antibody for 30 min at room temperature. Then, 3,3′diaminobenzidine was used to visualize protein expression and a Nikon Eclipse 80i microscope (Nikon Corporation) was used to capture images. The public domain program "Image-Pro Plus 6.0" was used to measure the immunohistochemical integral optical density (IOD). Sum IODs and areas of each photo were measured, then average optical densities in different groups were compared.
Statistical Analysis
R studio (version 4.0.3) was used for the analysis of data and GraphPad Prism 7 software (version 7.0, San Diego, CA, United States) was used for the generation of images.
Immunohistochemical data were assessed using the Student's t-test. p < 0.05 was considered statistically significant.
Identification of DEGs
DEGs were identified by two methods using three microarray datasets, including 39 AD and 33 HC samples. A total of 303 DEGs were acquired by "RRA." These consisted of 139 upregulated and 164 downregulated genes, some of which are shown in Figure 2A. A total of 860 DEGs were obtained by "Batch correction." These consisted of 478 upregulated genes and 382 downregulated genes, which were visualized via a heatmap ( Figure 2B) and a volcano map ( Figure 2C). A total of 259 DEGs, including 130 upregulated genes and 129 downregulated genes, were acquired from the intersection of the DEGs obtained by the two procedures ( Figure 2D and Supplementary Table S2). Frontiers in Molecular Biosciences | www.frontiersin.org July 2022 | Volume 9 | Article 917077 5
WGCNA and Module Analysis
In WGCNA analysis, the merged gene matrix from "Batch correction" was clustered to exclude outliers and abnormal samples which might have escaped during sample clustering ( Figure 3A). The soft-thresholding power β was set to 4 to attain a scale-free network evaluation coefficient R 2 of 0.85 ( Figure 3B). Similar modules from the co-expression network were combined using a cut height of 0.25 to obtain a total of eight modules ( Figure 3C). Correlations between clinical traits and module eigengenes are shown in Figure 3D. The genes of blue module were determined to be the most relevant to AD, based on the module-trait relationship heatmap (correlation coefficient 0.51, p < 0.001). The MM vs. GS for AD in blue module were calculated and presented with a scatter diagram ( Figure 3E). Finally, 331 module genes with AD traits from the blue module were identified for further studies (Supplementary Table S3).
Identification of Co-DEGs and Functional Enrichment Analysis
A total of 169 co-DEGs were identified from the intersection of the 331 AD traits module genes and 259 DEGs ( Figure 4A). GO and KEGG pathway analyses were used to further investigate the biological properties of these co-DEGs. We identified several biological properties, including cellular components (CCs), biological processes (BPs), and molecular functions (MFs), using GO annotation enrichment terms. With regard to BPs, the co-DEGs were substantially enhanced in the cellular response to type I Frontiers in Molecular Biosciences | www.frontiersin.org July 2022 | Volume 9 | Article 917077 interferon (IFN) and the type I IFN signaling pathway. As for CC and MF, co-DEGs were substantially enhanced in the collagen-containing extracellular matrix and peptidase regulator activation, respectively ( Figure 4B). We identified critical signaling pathways using KEGG enrichment analysis. The co-DEGs were substantially enhanced in the IL-17 signaling pathway ( Figure 4C). Metascape analysis indicated that co-DEGs were substantially enhanced in the cytokine-mediated signaling pathway ( Figures 4D-F).
Immunohistochemical Analysis
We found that the IOD/area values of CCR7, CXCL10, IRF7, MMP1, and RRM2 in AD tissue were significantly higher than those in HC tissue (p < 0.05) (Figure 8). However, the IOD/area values of CCNA2, ISG15, MKI67, and NCAPG were comparable between the AD and HC tissues (Supplementary Figure S1).
Ananalysis of Immune Cell Infiltration and the Relationship Between Identified Hub Genes and Differential Immune Cells in AD
A total of 23 AD and seven HC tissues met the criteria for analysis with CIBERSORT (p < 0.05). The constitutions of 22 types of immune cells in each specimen are shown in Figure 9A.
Compared with HC tissues, AD tissues exhibited a higher number of activated dendritic cells (DCs), CD4 + naïve T cells, and plasma cells, whereas the number of resting mast cells was relatively lower ( Figure 9B). Relationships between five identified hub genes (CCR7, CXCL10, IRF7, MMP1, and RRM2) and four significant differential immune cells (naive CD4 + T cells, plasma cells, activated DCs, and resting mast cells) in AD tissue were analyzed ( Figure 9C). Significantly related hub genes and immune cells were segregated using thresholds of R > 0.4 and p < 0.05. We found that the CCR7 expression level was positively correlated with the number of CD4 + naïve T cells (R = 0.42, p = 0.011) ( Figure 9D).
DISCUSSION
Though research into the molecular processes involved in AD has made rapid advances, the mechanisms of pathogenesis and progression of this disease are still poorly understood, which In this study, we methodically segregated potential new biomarkers for AD and evaluated the extent of infiltration of the skin tissue by immune cells in AD patients. We screened DEGs using two independent methods to improve the precision of the results. We also screened 169 final co-DEGs in combination with WGCNA and differential expression analysis. The results of GO and KEGG analyses showed that co-DEGs were substantially enhanced in the type I IFN and IL-17 signaling pathways. Frontiers in Molecular Biosciences | www.frontiersin.org July 2022 | Volume 9 | Article 917077 10 Multiple IFNα isotypes and IFNβ of the type I family have been extensively studied. Type I IFNs can activate crucial processes of innate and adaptive immune systems, including the presentation of antigens and the production of cytokines responsible for triggering the activation of B cells, T cells, and natural killer cells. These contribute substantially to the pathogenesis of AD (Kader et al., 2021;Li et al., 2021). IL-17 is consistently upregulated in AD in both the acute and chronic phases (Gittler et al., 2012). It may promote abnormalities in immune regulation in AD by upregulating S100A7/8/9 in combination with IL-22 (Nograles et al., 2008). The S100A proteins are highly upregulated in AD, and can operate as both promoters of inflammation and as antimicrobials (Mansouri and Guttman-Yassky, 2015). It has been reported that IL-17 downregulated filaggrin and promotes abnormalities in the skin barrier. It has also been shown to affect the expression of genes in keratinocytes related to cell adherence, thereby increasing the risk of AD (Gutowska-Owsiak et al., 2012).
In the present study, 13 prospective hub genes were screened using STRING and the cytoHubba plugin in Cytoscape. We found nine hub genes (CCR7, CCNA2, CXCL10, IRF7, ISG15, MKI67, MMP1, NCAPG, and RRM2) with high diagnostic sensitivity and specificity by external dataset GSE120721 validation and ROC curve analysis. For experimental validation, we performed immunohistochemistry analysis for 12 individuals, including 6 AD patients and 6 age-matched, race-matched and sex-matched healthy individuals, to detect expression levels of these nine identified hub genes. The expression levels of CCR7, CXCL10, IRF7, MMP1, and RRM2 were found to be upregulated in AD patients, as compared to the levels observed in HCs. This was in accordance with the results obtained using bioinformatics analysis. Expression levels of CCNA2, ISG15, MKI67, and NCAPG did not differ significantly in AD patients and HCs, though AD patients had higher expression levels of all four genes. We hypothesize that the non-significance of these differences could be attributed to the small sample size. This was because a greater number of skin samples could not be obtained from AD patients due to ethical reasons. Moreover, we only performed immunohistochemistry analysis. We could not simultaneously investigate protein expression levels via western blotting because of the small amount of skin tissue collected.
In this study, we identified CCR7, CXCL10, IRF7, MMP1, and RRM2 to be potential diagnostic biomarkers for AD. The chemokine receptor CCR7 was reportedly present on DCs as well as naïve, regulatory, and memory T cells (Förster et al., 2008). The movement of skin DCs to lymphoid tissue has been shown to be conditionally dependent upon the activation of the chemokine receptor CCR7 under both stable and inflammatory conditions (Ohl et al., 2004). After CCR7 −/− mouse bone marrow-derived DCs were subcutaneously injected into wild-type mice, these cells were unable to travel to the lymph nodes and access lymphatic drainage, which indicates that DCs also need to express CCR7 to facilitate their homing to lymph nodes (Hintzen et al., 2006). Moreover, it has been reported that the movement of neutrophils from skin to skin-draining lymph nodes via lymphatic vessels was also mediated by CCR7 (Özcan et al., 2022). In the present study, we found that the CCR7 expression level was upregulated in AD patients. As AD is a complex disease involving various immune cells such as DCs and neutrophils (Novak, 2012;Dhingra et al., 2013), it could be speculated that CCR7 might mediate the movement of DCs and neutrophils to skin-draining lymph nodes in AD and induce subsequent immune responses.
CXCL10, a 10 kDa protein classified as a Th1-chemokine, binds to CXCR3 receptors. CXCL10 exhibits chemotactic activity toward activated T lymphocytes and monocytes in the peripheral blood and is produced by activated T cells, monocytes, endothelial cells, and keratinocytes (Qi et al., 2009). In AD patients, the serum levels of CXCL10 and the expression levels of CXCL10 in skin lesions were significantly increased in comparison to the levels observed in HCs (Esaki et al., 2016;Brunner et al., 2017;Lang et al., 2021). This is consistent with our findings. CXCL10 can recruit activated T cells expressing CXCR3 (especially Th1 cells). These recruited Th1 cells secrete IFN-γ, stimulate local CXCL10 production, and further facilitate Th1 cell recruitment (Colvin et al., 2004). As T lymphocytes play vital roles in AD pathogenesis (Sroka-Tomaszewska and Trzeciak, 2021), lymphocyte chemotaxis regulated by CXCL10 may be involved in the AD immune response.
IFN regulatory factor 7 (IRF7) belongs to the IFN regulatory transcription factor family, which plays a vital role in the control of several biological processes, including inflammation, apoptosis, and immune response generation. He et al. found that the expression of IRF7 is increased in type 2 lymphoid cells during allergic inflammation, whereas an IRF7 deficiency leads to its remission (He et al., 2019). Cohen et al. reported that the phenotypic transformation from pro-inflammatory macrophages to anti-inflammatory macrophages (M1-to-M2) is modulated by IRF7, which is down-regulated by the transforming growth factor-beta 1 pathway (Cohen et al., 2014).
MMP-1, an interstitial collagenase, cleaves type I and II collagen, which are major constituents of the dermis (Parks et al., 2004). It has been shown that the serum levels of MMP-1 were significantly higher in the AD group than in HCs, and may correlate with the degree of damage to the epidermal barrier, which is represented by the extent of trans-epidermal water loss (Basałygo et al., 2021). Harper et al. reported that an increase in MMP-1 levels was associated with a corresponding increase in the severity of AD (Harper et al., 2010). Another report demonstrated that the severity of AD was proportional to the severity of damage to the epidermal barrier (Weidinger and Novak, 2016). MMP-1 also triggers other MMPs, such as MMP-9, which is highly specific for substrates such as dermal elastin and fibrillin (Tsoureli-Nikita et al., 2006). The cleavage of the components in the basement membrane allows T cells to cross the basement membrane and enter the epidermal compartment during skin inflammation (Wölfle et al., 2015).
RRM2, a small subunit of ribonucleotide reductase, is overexpressed by tumors, and plays a role in resistance to chemotherapy (Zhan et al., 2021). Tang et al. reported that there was a positive correlation between the expression level of RRM2 and the extent of infiltration by neutrophils and macrophages. They also reported that RRM2 inhibition effectively suppressed macrophage infiltration, and affected the balance of macrophage polarization, which promoted M1 phenotype polarization and suppressed the M2 phenotype in vitro and in vivo (Tang et al., 2021). However, there have been no reports about the relationship between RRM2 and AD until now. We found that RRM2 was upregulated in AD skin tissue. This may contribute to the chronic inflammation involved in AD, since RRM2 is known to promote macrophage infiltration and polarization. However, this hypothesis needs to be validated by further studies.
Immune cell infiltration is a hallmark of AD. Therefore, we studied infiltration of immune cells, and relationships between the identified hub genes (CCR7, CXCL10, IRF7, MMP1, and RRM2) and significant differential immune cells (naive CD4 + T cells, plasma cells, activated DCs, and resting mast cells) in AD. We found that CCR7 expression level was positively correlated with the number of CD4 + naïve T cells (R = 0.42, p = 0.011) in AD patients. It has been shown that CCR7 mediates the homing of naive T cells. Studies in CCR7-deficient (CCR7 −/− ) mice have demonstrated that on examination of lymph nodes and Peyer's patches, naive T cell numbers are reduced. When the T cells of CCR7 −/− mice were adoptively transferred to wild-type mice, they were unable to home to Peyer's patches and draining lymph nodes (Förster et al., 1999). We hypothesized that CCR7 might also mediate the homing of naive T cells in AD. However, more in vitro and in vivo studies need to be conducted to validate this hypothesis.
Several limitations are associated with our study. First, the human sample size was limited. Additional samples are required to validate the present findings. Second, we used only immunohistochemistry techniques to validate hub genes in vitro. Additional experiments with more quantitative methodologies will be needed for further confirmation of our results. Finally, the precise role of identified hub genes in AD requires additional elucidation both in vitro and in vivo. Further research is essential to determine whether CCR7, CXCL10, IRF7, MMP1, and RRM2 could be used as predictive biomarkers for the diagnosis and treatment of AD.
CONCLUSION
To summarize, our study uses both WGCNA and CIBERSORT methods to understand the molecular mechanism of AD, and to verify the result by immunohistochemical technology. Our study has not only identified effective biomarkers for AD but has increased our understanding of how immune cells related to AD. CCR7, CXCL10, IRF7, MMP1, and RRM2 could be used as biomarkers in AD for both diagnostic and therapeutic purposes. Additionally, we have shown that CCR7 expression level was positively correlated with the number of CD4 + naïve T cells. We therefore have confidence that our study supports a new trajectory for investigation into the management of AD and provides new knowledge on the immune, cellular, and molecular mechanisms for its pathogenesis.
ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Shengjing Hospital of the China Medical University. The patients/participants provided their written informed consent to participate in this study.
AUTHOR CONTRIBUTIONS
CL downloaded and analyzed the data, prepared charts, and wrote the draft of the paper. YL initiated and planned the experiments. XH initiated and planned the experiments and reviewed and revised the article. We declare that all authors have read and approved the final manuscript.
|
2022-07-14T13:35:57.956Z
|
2022-07-14T00:00:00.000
|
{
"year": 2022,
"sha1": "b79bf6a7eea792ea82d544955a42cd8906fe0a77",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Frontier",
"pdf_hash": "b79bf6a7eea792ea82d544955a42cd8906fe0a77",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
27484401
|
pes2o/s2orc
|
v3-fos-license
|
A set of loop-1 and -3 structures in the novel vascular endothelial growth factor (VEGF) family member, VEGF-ENZ-7, is essential for the activation of VEGFR-2 signaling.
The vascular endothelial growth factor (VEGF) family plays important roles in angiogenesis and vascular permeability. Novel members of the VEGF family encoded in the Orf virus genome, VEGF-E, function as potent angiogenic factors by specifically binding and activating VEGFR-2 (KDR). VEGF-E is about 45% homologous to VEGF-A at amino acid levels, however, the amino acid residues in VEGF-A crucial for the VEGFR-2-binding are not conserved in VEGF-E. To understand the molecular basis of the biological activity of VEGF-E, we have functionally mapped residues important for interaction of VEGF-E with VEGFR-2 by exchanging the domains between VEGF-E(NZ-7) and PlGF, which binds only to VEGFR-1 (Flt-1). Exchange on the amino- and carboxyl-terminal regions had no suppressive effect on biological activity. However, exchange on either the loop-1 or -3 region of VEGF-E(NZ-7) significantly reduced activities. On the other hand, introduction of the loop-1 and -3 of VEGF-E(NZ-7) to placenta growth factor rescued the biological activities. The chimera between VEGF-A and VEGF-E(NZ-7) gave essentially the same results. These findings strongly suggest that a common rule exists for VEGFR-2 ligands (VEGF-E(NZ-7) and VEGF-A) that they build up the binding structure for VEGFR-2 through the appropriate interaction between loop-1 and -3 regions.
The vascular endothelial growth factor (VEGF) family plays important roles in angiogenesis and vascular permeability. Novel members of the VEGF family encoded in the Orf virus genome, VEGF-E, function as potent angiogenic factors by specifically binding and activating VEGFR-2 (KDR). VEGF-E is about 45% homologous to VEGF-A at amino acid levels, however, the amino acid residues in VEGF-A crucial for the VEGFR-2-binding are not conserved in VEGF-E. To understand the molecular basis of the biological activity of VEGF-E, we have functionally mapped residues important for interaction of VEGF-E with VEGFR-2 by exchanging the domains between VEGF-E NZ-7 and PlGF, which binds only to VEGFR-1 (Flt-1). Exchange on the amino-and carboxylterminal regions had no suppressive effect on biological activity. However, exchange on either the loop-1 or -3 region of VEGF-E NZ-7 significantly reduced activities.
On the other hand, introduction of the loop-1 and -3 of VEGF-E NZ-7 to placenta growth factor rescued the biological activities. The chimera between VEGF-A and VEGF-E NZ-7 gave essentially the same results. These findings strongly suggest that a common rule exists for VEGFR-2 ligands (VEGF-E NZ-7 and VEGF-A) that they build up the binding structure for VEGFR-2 through the appropriate interaction between loop-1 and -3 regions.
Vascular endothelial growth factor-A (VEGF-A) 1 plays a pivotal role in vasculogenesis, angiogenesis, and differentiation of hemangioblasts to hematopoietic precursor cells in embryogenesis. VEGF-A is also known to be closely involved in a variety of pathological conditions such as tumor angiogenesis and diabetic retinopathy (1)(2)(3).
VEGF-A is a member of the PDGF superfamily because of their structural similarities. VEGF is found to be a dimeric glycoprotein of M r 34,000 -42,000, and have conserved eight cysteine residues in each monomer. These cysteine residues construct a particular folding consisting of two intermolecular and three intramolecular disulfide bonds that generate three loop-like structures, loop-1, -2, and -3. VEGF-A exerts its biological activity by interacting with receptor-type tyrosine kinases, VEGFR-1 (Flt-1) and VEGFR-2 (KDR/Flk-1) (4 -7). Homozygous loss of the VEGFR-1 or VEGFR-2 genes resulted in embryonic lethality between days 8.5 and 9.5, indicating that these VEGF receptors play important roles in vasculogenesis and angiogenesis (8,9). The different phenotypes of these VEGFR-mice suggest that VEGFR-2 is the major positive signal transducer, whereas VEGFR-1 has a negative regulatory role in angiogenesis at early embryogenesis.
The VEGF family in vertebrate genomes includes VEGF-A, PlGF (placenta growth factor), VEGF-B, -C, and -D. PlGF and VEGF-B specifically bind to VEGFR-1 (10 -13), whereas VEGF-C and -D bind and activate VEGFR-3 (Flt-4), regulating lymphangiogenesis as well as angiogenesis in the middle stage of embryogenesis (14 -18). Amino acids in VEGF-A essential for binding with VEGFR-2 have been studied by the alanine scanning method. Keyt et al. (19) have reported that Arg-82, Lys-84, and His-86 are indispensable for the interaction between VEGF-A and VEGFR-2.
Recently, we have shown that the VEGF-E NZ-7 protein, a novel member of the VEGF family could bind specifically to VEGFR-2, activate the receptor, and promote the growth of endothelial cells in vitro and in vivo at a transient condition (20). VEGF-E genes were originally found as an open reading frame in the genome of the NZ-7, NZ-2, and D1701 strains of parapoxvirus, Orf virus (21).
These three genes were structurally very similar to each other compared with VEGF family proteins, and were designated as VEGF-E NZ-7 , ORFV2-VEGF/VEGF-E NZ-7 , and VEGF-E D1701 (20,22,23). Orf virus causes contagious pustular dermatitis in sheep, goats and, sometimes, humans. Histologically, the lesions are highly vascularized and edematous with proliferation of endothelial cells and inflammatory cells. In addition to VEGF-E NZ-7 , other two VEGF-E members were also shown to specifically bind to VEGFR-2 but not to other receptors.
VEGF-E NZ-7 has a high affinity to VEGFR-2 at similar levels as VEGF-A, and efficiently competes to VEGF-A (20). This indicates that the binding pocket on VEGFR-2 for VEGF-A and VEGF-E NZ-7 is significantly overlapped to each other. Interestingly, however, three basic amino acids on VEGF-A essential for the VEGFR-2-binding are not conserved in VEGF-E NZ-7 , and these basic amino acids were changed to hydrophobic or non-charged ones, Val, Gly, and Ser, respectively. These results indicate that the local structure built up by these basic amino acids in VEGF-A are not always required for the ligands that bind to VEGFR-2.
To elucidate a novel rule for the ligand-binding to VEGFR-2, we carried out a series of domain-exchange analysis between VEGF-E and PlGF, or VEGF-E and VEGF-A. Our results clearly indicate that an intimate relationship between the loop-1 and -3 of VEGF-E NZ-7 as well as VEGF-A is crucial for the formation of the three-dimensional structure important for the high-affinity binding to VEGFR-2.
Polyclonal and Monoclonal Antibodies-Polyclonal antisera against VEGF-E NZ-7 were raised in rabbits using a 20-amino acid sequence of the carboxyl terminus as antigen (20). Anti-human VEGFR-2 antiserum (B2) was prepared previously (24). A monoclonal antibody specific to phosphotyrosine (PY20) was obtained from ICN Biochemicals (Costa Mesa, CA). Secondary antibodies conjugated to horseradish peroxidase were purchased from Amersham Biosciences. Monoclonal neutralizing antibody to VEGF-A 165 was purchased from R&D Systems (Minneapolis, MN).
Construction of VEGF-E NZ-7 Chimeric Mutant-The synthetic cDNA encoding VEGF-E NZ-7 was cloned to the BamHI and EcoRI restriction sites of pUC18 (20), and PlGF cDNA was cloned to pUC18 (13). At first, sequences encoding six histidines were introduced by using a double stranded oligonucleotide encoding amino acid residues Cys-130 to Arg-148 of VEGF-E NZ-7 and six histidines followed by stop codon. This oligomer also had the cohesive end for DraIII at the NH 2 terminus and for EcoRI at the COOH terminus, and it was ligated to the 3-kb DraIII-EcoRI fragment of the plasmid that contains cDNA of VEGF-E NZ-7 cloned to pUC18 (pUCE-his). With the same manner, histidine residues FIG. 1. Alignment of the amino acid sequence for the VEGF family and three-dimensional structure of the dimeric molecule of VEGF-A. A, the amino acid sequences of VEGF-E (VEGF-E NZ-7 , VEGF-E NZ-2 /VEGF-OR-FV NZ-2 , VEGF-E D1701 ) and human VEGF-A, -B, -C, and -D are shown. The amino acid residues of VEGF-A critical for interacting with VEGFR-2 are shaded in orange. The numbers in red and blue boxes represent the amino acid residues of VEGF-A and VEGF-E NZ-7 , respectively. The numbers and thin lines in black represent the regions of VEGF-E NZ-7 that were replaced with corresponding amino acid residues of human PlGF and/or VEGF-A. The double black line with number 27 indicates a unique amino acid stretch only found in VEGF-E NZ-7 . Red character, the eight cysteine residues conserved in all VEGF family members. Blue character, the amino acid residues identical to VEGF-E NZ-7 . Black line, the core region that is conserved in all VEGF family. Red bar, loop regions 1, 2, and 3 of VEGF-A. Green and blue bars, the ␣-helix regions and -strands of VEGF-A, respectively. Signal peptide sequences at the NH 2 -terminal region of the VEGF family were omitted in this figure. Ϫ, no amino acids. B, dimer structure of VEGF-A. One monomer of the VEGF-A dimer is shown in red and another in green. The VEGF-A molecule forms three intramolecular and two intermolecular disulfide bonds. The disulfide bonds and cysteine residues of a homodimer are shown in yellow in one monomer and in red in another. The numbers in black indicate the loop structure. The amino acid residues of VEGF-A important for interacting VEGFR-2 are shaded in red. This figure was adapted from Muller et al. (30,31).
were introduced into the COOH terminus of PlGF-1 cDNA. The double stranded synthetic oligonucleotide encoding Arg-131 to Arg-149 of PlGF-1 was followed by six histidine residues and a stop codon. This oligomer had cohesive ends for the BsmI site at the NH 2 terminus and EcoRI site at the COOH terminus. This oligomer was cloned to a 3-kb BsmI-EcoRI fragment of the plasmid (pUCP-his) with full-length cDNA of PlGF-I cloned to BamHI and the EcoRI site of pUC18.
The constructs of chimera proteins were produced by exchanging variable regions among VEGF-E NZ-7 , VEGF-A, and PlGF. They were achieved by ligating a series of digested fragments of pUCP-his and/or pUCE-his with synthetic double stranded oligonucleotides and/or the fragments produced by the PCR technique. The details of each plasmid construction are available on request.
Each construct contains the following amino acid sequence: A, a variety of constructs for the replacement of amino-and carboxyl-terminal regions in VEGF-E NZ-7 (orange) with the corresponding regions of PlGF (green). a, VEGFR-2 (KDR) autophosphorylation was measured by using NIH3T3-KDR cells. The cells were stimulated with chimera mutant proteins (4 -120 ng/ml), lysed, and then subjected to SDS-PAGE for Western blotting with an anti-phosphotyrosine antibody (␣-PY) or with an anti-VEGFR-2 antibody (␣-KDR IK-5). The activity was indicated by ϩ. ϩϩϩ, equal to wild type VEGF-E NZ-7 ; ϩϩ, 2-3-fold weaker than wild type VEGF-E NZ-7 ; ϩ, 3-10-fold weaker than wild type VEGF-E NZ-7 ; Ϫ, phosphorylation was not detected in this concentration range. b, proliferation of HUVEC. Quiescent HUVEC were stimulated with 1, 10, and 100 ng/ml chimera mutant proteins. After 4 days, cells were stained and the cell number was determined by averaging the counting of five spots to wild type VEGF-E NZ-7 . ϩϩ, 2-fold weaker than wild type VEGF-E NZ-7 at 100 ng/ml; ϩ, 4-fold weaker than wild type VEGF-E NZ-7 ; Ϫ, HUVEC proliferation was not detected. B, binding activities of the chimeric proteins between VEGF-E NZ-7 and PlGF to VEGFR-2. The activities were measured by competition experiments using 125 I-VEGF-A and VEGFR-2 binding systems (see "Materials and Methods"). C, the activities of chimeric proteins for stimulation of HUVEC growth.
Expression of Mutant Proteins and Purification-The full-length coding regions of chimera mutants were subcloned into the BamHI-EcoRI site on the multicloning region of pVL1393. These transfer vector DNAs were used for co-transfection into Sf9 cells along with the linearized baculovirus DNA "BaculoGold" by liposome transfection. The recombinant viruses were amplified at 3-day intervals. Sf9 cells grown in serum-free medium EX-CELL 400 were used to produce chimera proteins. For a large scale preparation of proteins, Sf9 cells were infected with viruses at a multiplicity of infection of about 10. Three days after infection, the culture media were harvested and centrifuged to remove debris. The supernatants were resolved by SDS-PAGE on a 15% gel followed by Western blotting.
One-hundred ml of the supernatant of cells infected with chimera recombinant virus was collected, concentrated, and dialyzed to 22.5 mM sodium phosphate buffer, pH 8.0, containing 375 mM NaCl and 62.5 mM imidazole. After dialysis and filtration, glycerol was added for a final concentration at 20%, giving rise to the final buffer with 20 mM sodium phosphate, pH 8.0, 300 mM NaCl, and 50 mM imidazole. The nickelnitrilotriacetic acid beads (Qiagen, Germany) were applied to the dialyzed sample and rotated for 3 h at 4°C. The nickel-nitrilotriacetic acid containing sample was loaded onto a 5-ml column, and washed with washing buffer (50 mM imidazole, 20 mM sodium phosphate buffer, 300 mM NaCl, and 20% glycerol). The chimera proteins were eluted with elution buffer (250 mM imidazole, 20 mM sodium phosphate buffer, pH 8.0, 300 mM NaCl, and 20% glycerol). The positive fractions were collected and further concentrated by Microcon (Millipore). For detection of VEGF-E protein, aliquots of fractions were analyzed by Western blotting using anti-His antibody (Qiagen, Germany) and Coomassie staining. The purity of chimera proteins was above 80%.
Binding Competition Assays-Binding competition assay was performed as previously described (25,26) using receptor protein immobilized to solid phase. Preparation of the receptor protein consisting of the extracellular region of VEGFR-2 tagged with the Fc portion of IgG was described previously (25). Aliquots (50 l) of the Fc-tagged receptor (0.3 g/ml) in phosphate-buffered saline were attached to 96-well plates, Immunon 2 (Dynex Technologies, Inc., Chantilly, VA), overnight at 4°C. The plates were washed twice and blocked with binding buffer (1% bovine serum albumin in phosphate-buffered saline) for 30 min at 25°C. Competition activity was examined by incubating the KDRcoated plates with a fixed concentration of 125 I-VEGF-A and increasing concentrations of non-radiolabeled chimera mutant proteins for 3 h at 25°C. The wells were washed three times with binding buffer, then the bound 125 I-labeled proteins were quantified in a ␥-counter. All experiments were performed in duplicate.
Human Endothelial Cell Proliferation Assay-Human umbilical vein endothelial cells (HUVEC) (Morinaga, Tokyo) were grown in HuMedia-EG2 (Kurabo, Tokyo) and used for endothelial cell growth assay. HU-VEC were seeded at 4,000 cells/well on 24-well collagen-coated plates (CELLTIGHT, Sumitomo Bakelite Co., Ltd., Tokyo) in medium with free-growth factor and low concentration serum (0.2%). Four h after plating, mutant proteins containing medium were added. Two days after, medium and mutant proteins were replaced with fresh sample. At day 4, cells were stained with the crystal violet. The cell numbers were determined by averaging the counting of five spots randomly chosen at each well using a Coulter counter.
VEGFR-2 Autophosphorylation Assay-For in vivo phosphorylation, NIH3T3-KDR (VEGFR-2) cells were grown to semiconfluence and stimulated with a variety of ligands at 37°C for 5 min. The cells were washed in ice-cold phosphate-buffered saline with 0.1 mM Na 3 VO 4 twice and lysed in 1% Triton X-100 lysis buffer (50 mM HEPES, pH 7.4, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl 2 , 2% aprotinin, 1 mM phenylmethylsulfonyl fluoride, 50 mM sodium fluoride, 10 mM Na 4 P 2 O 7 , and 2 mM Na 3 VO 4 ). The lysates were clarified by centrifugation (15,000 rpm for 10 min). Protein concentrations were measured using a Bio-Rad protein assay and the same amounts of protein of each sample were used for analysis. For immunoblotting, the cell lysates were subjected to 7.5% SDS-PAGE and transferred to a nitrocellulose sheet. The blots were incubated with a blocking solution (5% bovine serum albumin containing washing buffer (20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.3% Tween 20)) and probed with the primary antibody diluted in blocking solution. The signal was visualized using peroxidase-conjugated secondary antibodies and exchanged chemiluminescence (ECL, Amersham) according to the manufacturer's instructions.
Tubular Formation Assay-The angiogenesis kit was purchased from Kurabo (Tokyo) and the tubular formation assay was demonstrated as per the manufacturer's instructions. In this system, human endothelial cells and fibroblasts were co-cultured in 24-well plates at 37°C, 5% CO 2 . At days 1, 4, 7, and 9, medium was exchanged with fresh medium, which contains chimera proteins and the excess amount of neutralizing antibody for VEGF-A. At day 11, incubation was terminated, fixed with 70% ethanol, and immunostained for CD31/PECAM-1. For immunostaining, mouse anti-human CD31, alkaline phosphataseconjugated goat anti-mouse IgG and 5-bromo-4-chloro-3-indolyl phosphate were purchased from Kurabo (Tokyo). The cells were incubated with mouse anti-human CD31 for 60 min at 37°C in blocking buffer (1% bovine serum albumin in phosphate-buffered saline) and washed with blocking buffer. Secondary, alkaline phosphatase-conjugated goat antimouse IgG was added to the cells followed with incubation for 60 min at 37°C. After washing, 5-bromo-4-chloro-3-indolyl phosphate was used to develop a color to visualize endothelial cells. Tubules were analyzed under a bright-field microscope and total length of branching in a fixed area (mm/mm 2 ) of the randomly chosen 5 spots per well.
Production and Dimer Formation of VEGF-E NZ-7 Mutants-
Alanine scanning analysis of VEGF-A has shown that basic amino acids at 82, 84, and 86 as well as other amino acids in loop-3 are essential for VEGFR-2 binding, however, the VEGF-E family that binds to VEGFR-2 at a similar affinity as VEGF-A does not conserve these amino acids (Fig. 1). To examine which region(s) in VEGF-E is crucial for the binding and activation of VEGFR-2, we constructed a series of chimeric mutants by exchanging the variable region of VEGF-E NZ-7 with that of PlGF and/or VEGF-A. The carboxyl terminus of all chimeric mutants had an insertion of six histidines as a tag sequence for the convenience of protein purification. This His tag did not suppress the activity of VEGF-E NZ-7 .
VEGF-E NZ-7 was exchanged with PlGF from the NH 2 -terminal, and the exchanging region was extended toward the COOH-terminal, serially (Figs. 2 and 3). Secondary, exchanging was performed from the COOH terminus toward serially. Loop-wise exchanging on loop-1, -2, and -3 was produced by both PlGF and VEGF-A amino acid sequences (Fig. 4A). The site-directed mutagenesis was also performed by exchanging only 4 amino acid residues (Fig. 4B). Finally, loop-1 and -3 regions of VEGF-E NZ-7 were introduced to PlGF as the gain of function mutants (Fig. 5) (see "Materials and Methods").
The molecular size of mutant proteins in non-reducing conditions were in the range of M r 40,000 to 45,000, and 20,000 to 25,000 in reducing conditions, consistent with wild type VEGF-E NZ-7 and PlGF (Fig. 3, data not shown). These results indicate that all mutants did not show any disruption in dimerization. affinity of each mutant to receptor, a binding assay was carried out by using the extracellular domain of VEGFR-2 tagged with the Fc portion of IgG (see "Materials and Methods"). Because VEGF-E NZ-7 could compete with VEGF-A in binding to VEGFR-2, chimeric proteins were tested for their ability to compete with 125 I-labeled VEGF-A 165 for interaction to VEGFR-2-Fc immobilized onto 96-well plates.
When VEGF-E NZ-7 was exchanged with PlGF from the NH 2terminal toward COOH-terminal serially ( Fig. 2A), 34 amino acids of the NH 2 -terminal region in VEGF-E NZ-7 were replaceable without any reduction of affinity to VEGFR-2 (Fig. 2, chimera 1). An extension of the PlGF region to the first conserved cysteine residue resulted in a minor reduction of the affinity (Fig. 2B, 2). However, as shown by chimeras 3, 4, and 5, the PlGF amino acid sequence close to and over loop-1 resulted in the loss of affinity similar to intact PlGF (Fig. 2B).
Secondary exchanging was performed from the COOH terminus toward the NH 2 terminus serially ( Fig. 2A). Among the mutants 6, 7, and 8, chimera 6 with 18 amino acids of the COOH-terminal exchange from the 8th conserved cysteine to the COOH-terminal end remained its affinity. However, the replacement over loop-3 such as chimeras 7 and 8 had lost their binding ability. Taken together, NH 2 and COOH termini were not critical for VEGF-E NZ-7 to interact with VEGFR-2.
To confirm this result, a chimera with simultaneous exchanging of both NH 2 and COOH termini were evaluated ( Fig. 2A, chimera 9), and showed no significant suppression in its affinity. These results may imply that at least loop-1 and -3 are very critical for VEGF-E NZ-7 , whereas the importance of loop-2 remained unclear. (27) reported that several amino acids in loop-3 are essential for the binding of VEGF-A to VEGFR-2 (27). This result suggests that loop-3 is a direct binding site for VEGFR-2 in VEGF-A. To analyze the importance of each loop region of VEGF-E NZ-7 , a variety of chimera mutants with loop region-specific replacement were tested (Figs. 4 and 5). These exchanged regions do not contain -strands that flank the loop site. Therefore, the basic architecture of these chimeric proteins would not be disrupted.
We found that the loop-2 region in VEGF-E NZ-7 is exchangeable to that of PlGF or VEGF-A (Fig. 4, chimera 12 and 13), whereas loop-1 and -3 specific exchanges to PlGF caused a significant defect (Fig. 4, chimera 11, 15, and 17). These results indicate that the important regions for VEGF-E NZ-7 to interact to VEGFR-2 are not only loop-3 but also loop-1.
Along with these PlGF replacements, the VEGF-A region was also introduced into loop regions of VEGF-E NZ-7 as a reference. Surprisingly, independent introduction of loops of VEGF-A, which is a strong ligand for VEGFR-2, also resulted in significant reduction in biochemical and biological activities (Fig. 4, chimeras 10, 14, and 16). Therefore, both loop-1 and -3 regions are equally important in VEGF-A as well as in VEGF-E NZ-7 . This result implies that VEGF-E NZ-7 and VEGF-A may have a common mechanism in the interaction to VEGFR-2, by having critical regions in loop-1 and -3.
Loop-1 but Not the Adjacent Short Sequences Are Important for Biological Function-More detailed mutations were introduced to identify important amino acid residues. The region for loop-1 exchange was divided into 3 parts, which were composed of YLGE, ESTN, and LQYN. These residues were changed to PlGF counterparts, DVVS, SEVE, and HMFS (Figs. 1A and 4B). Among these mutants, a chimera with ESTN exchanged with SEVE severely lost the affinity, and exchange of YLGE to DVVS had minor loss in affinity, whereas no effect in another exchange mutant (chimeras 20, 19, and 21, respectively). These results suggest that the affinity reduction conferred by the exchange of the loop-1 region in chimera 11 was mainly caused by the replacement of ESTN on VEGF-E NZ-7 to SEVE in PlGF. In the loop-1 of VEGF-E NZ-7 , accumulation of negatively charged residues might have disrupted the binding action of VEGF-E NZ-7 . In addition, regions upstream from loop-1 (chimera 22 and 23) and a region downstream of loop-3 (number 26) were replaceable to the corresponding regions of PlGF, however, short stretches upstream from loop-3 were not (numbers 24 and 25).
VEGF-E NZ-7 have its specific insertion-like stretch with 8 amino acid residues in the loop-3 region (Fig. 1A). This stretch FIG. 5. Functional analysis of VEGF-E NZ-7 /PlGF chimera mutant proteins, 28 -33. A, the construct map of chimera mutants and their activities for VEGFR-2 autophosphorylation and HUVEC growth (see Fig. 2, legend). B, binding activities of the chimeric proteins to VEGFR-2. C, the activities of chimeric proteins for stimulation of HUVEC growth.
was not found in other VEGF-E members or in the VEGF family. Interestingly, the flanking region of this stretch showed similarities in amino acid sequence to other VEGF-E genes. This stretch was deleted to test its significance for the VEGF-E NZ-7 protein. The deletion resulted in a complete loss of biological activities, indicating that this short stretch is indispensable for VEGF-E NZ-7 (Fig. 4B, chimera 27).
Cooperation between a Proper Set of Large Loop-1 and Large Loop-3 Is Crucial for Restoration of Biological Activity in
VEGF-E NZ-7 -To examine the regions in VEGF-E NZ-7 necessary and sufficient for the activation of KDR/VEGFR-2, we further constructed chimeric molecules between VEGF-E NZ-7 and PlGF. Under the background of the PlGF sequence, the loop-1-containing sequence and the loop-3-containing sequence in VEGF-E NZ-7 were introduced to the corresponding regions (Fig. 5A, 31-33). For a control experiment, under the background of VEGF-E NZ-7 , the loop-1 and -3 regions in VEGF-A were replaced to the corresponding regions (Fig. 5A, 28 -30). As shown in chimera 33, both loop-1 and -3-containing regions were found to be required for the activation of VEGFR-2 kinase.
Although VEGF-E NZ-7 and VEGF-A are able to stimulate VEGFR-2, to our surprise, a single replacement of either loop-1 or -3 in VEGF-A to the corresponding region in VEGF-E NZ-7 strongly suppressed the biological activity (Fig. 5, chimeras 28 and 29). Both loop-1 and -3 of VEGF-A are required for a partial recovery of the activation of VEGFR-2 (chimera 30). Taken together, these results strongly suggest that an appropriate pair of loop-1 and -3 is essential for the construction of threedimensional structure for the binding and activation of VEGFR-2.
The Binding Ability of Mutants to VEGFR-2 Correlates Well with the Activity of Receptor Autophosphorylation Assay and HUVEC Proliferation Assay-These mutant proteins were tested for their ability to induce autophosphorylation of VEGFR-2 using a cell line, NIH3T3-KDR (see "Materials and Methods") ( Fig. 6, data not shown). The ability was highly correlated with the affinity of mutant proteins to the receptor (Figs. 2, 4, and 5). The mutants with high affinity, as the wild type VEGF-E NZ-7 protein to receptor, demonstrated a strong activity in induction of autophosphorylation of KDR/VEGFR-2, whereas the mutants with no affinity did not in a concentration range up to 150 ng/ml in the final medium. These results indicate that the critical regions of VEGF-E NZ-7 for binding to VEGFR-2 are also important for inducing autophosphorylation of VEGFR-2.
VEGF-A-induced signal transduction for the proliferation of endothelial cells is mainly mediated by VEGFR-2. To evaluate the relationship of the abilities to induce autophosphorylation of VEGFR-2 and to induce a proliferation of endothelial cells, mutant proteins were tested for their activity to induce proliferation of HUVEC. As expected, mutant chimeras 1, 2, 6, 9, 12, 13, 19, 21-23, 26, 30 (weak), and 33 with high affinity for VEGFR-2 could lead the proliferation of HUVEC, whereas mutants with low receptor affinity facilitated weak mitogenesis (Figs. 2, 4, and 5). Those mutants with no affinity failed to show such activity, with one exception, number 18. Mutant number 18 did not show detectable affinity to VEGFR-2, but it partially induced mitogenesis of endothelial cells. These results suggest that the affinity strength of mutant proteins mostly correlates with their activity to induce mitogenesis of HUVEC.
Evaluation of Mutants in Tubular Formation Assay-VEGF-A is known to stimulate the endothelial cells to form a tubule-like structure in vitro and in vivo. The chimeric mutants of VEGF-E NZ-7 were tested for their activity for tubular formation in a recently developed co-culture system between HUVEC and human diploid fibroblasts (see "Materials and Methods") ( Fig. 7). To decrease a background tube formation in this system detectable without any exogenous angiogenic factors, we added anti-VEGF-A neutralizing antibody into culture medium.
At first, we demonstrated that VEGF-E NZ-7 could induce the tubular formation in vitro (Fig. 7A). Next we examined tubular formation by chimera mutants. Representative results and quantitative analysis are shown in Fig. 7, B and C. The strength of chimera proteins for tubular-forming activity was correlated with the affinity to the receptor, and with the abilities to autophosphorylate the receptor and induce the proliferation of endothelial cells. Interestingly, the tubules induced by VEGF-E NZ-7 were morphologically distinguished from that of VEGF-A, where VEGF-E NZ-7 -induced tubules were slightly thicker. Chimera number 18, which showed only weak HUVEC proliferation without the significant affinity to receptor and the receptor autophosphorylation ability in vitro could induce the tubular formation as effectively as VEGF-A and VEGF-E NZ-7 . The morphology induced by number 18 was closely related to that of tubules induced rather by VEGF-A than VEGF-E NZ-7 . This result suggests that this tubular formation system is more sensitive than other assays such as the binding assay, receptor autophosphorylation assay, and HUVEC proliferation assay, so that the activity of chimera 18 was detected. In conclusion, both loop-1 and -3-containing regions in VEGF-E NZ-7 were found to be required for the binding and activation of VEGFR-2 in a cooperative manner (Fig. 8). DISCUSSION In this study, the important regions of VEGF-E NZ-7 for binding to VEGFR-2 were analyzed by using domain exchanging with PlGF and site-directed mutagenesis. PlGF does not bind to VEGFR-2 but only to VEGFR-1 (12,13,28). VEGF-E NZ-7 and PlGF have a similar amino acid composition, and they also completely conserve the critical cysteine-knot motif that was composed of eight cysteine residues (see Fig. 1B). Recently, the crystal structure study for PlGF (29) revealed that PlGF actually conserves the tertially structure with VEGF (30,31) and PDGF (32). Thus, the VEGF-E NZ-7 /PlGF chimera mutants were assumed to fold appropriately in a similar manner and build up tertiary structure. We confirmed that these chimeric mutants conserve the dimerization property in post-translational modification (Fig. 3).
In previous reports, the receptor interacting amino acids for PDGF-B and VEGF-A were identified. For PDGF, site-directed mutagenesis revealed that Arg-27 and Ile-30 were critical amino acid residues in interacting to PDGF receptor (33). These amino acid residues were found in the loop-1 region of PDGF-B. The alanine scanning mutagenesis of VEGF-A demonstrated important amino acid residues for binding to receptors. For VEGFR-1 binding, negatively charged residues Asp-63, Glu-64, and Glu-67 were critical, and they were found in the loop-2 region (27). For VEGFR-2 binding, mutation in positively charged residues, Arg-82, Lys-84, and His-86 resulted in severe reduction in receptor binding, and these basic amino acid residues are located in the loop-3 region (27). Because these three basic amino acids are not conserved and changed to hydrophobic ones in the VEGF-E family, we suggest that these basic amino acids in VEGF-A are not directly interacting with VEGFR-2 but are important for keeping the proper tertially structure within the VEGF-A molecule.
In another report (31), supporting the previous result, critical amino acids for binding VEGFR-2 were found clustered at ␣-helix, loop-1, -2, and -3 regions. They were designated "hot spots." The majority of these hot spot amino acid residues are located in loop-1 and -3 regions. In addition, VEGF-A neutralizing antibody and peptides, which inhibit the interaction of VEGF-A to VEGFR-2, have been reported. Interestingly, the epitopes for these inhibitors are located at loop-1 and -3, thus, they might modify the structure of loop-1 and/or loop-3 regions (34,35).
We found that PlGF could replace the amino and carboxyl terminus of VEGF-E NZ-7 without any changes in biochemical and biological activities. This result indicates that the amino and carboxyl termini in VEGF-E NZ-7 are not involved directly in binding with VEGFR-2. As shown in mutants 3-5, 7, and 8 ( Fig. 2), extending the exchange region into the core region and over loop-1 and -3 regions resulted in a significant defect in biological and biochemical activities. These results suggest that the regions containing loop-1 and -3 are critical for VEGF-E NZ-7 to be functionally intact. As expected, exchanging the narrowed regions of VEGF-E NZ-7 only containing loops-1 and -3 (mutant FIG. 7. Analysis of tubular formation by chimera mutant proteins. VEGF-E chimera proteins were tested for stimulatory activity of endothelial tubular formation using the HUVEC-fibroblast co-culture system (see "Materials and Methods"). At days 1, 4, 7, and 9, medium was replaced with fresh medium that contains angiogenic factor and neutralizing antibody for VEGF-A (500 ng/ml). At day 11, incubation was terminated, fixed, and immunostained with ␣-CD31/PECAM-1 antibody for endothelial cells (arrows). The co-cultured fibroblasts were not stained but slightly giving rise to the yellow background. A, VEGF-E NZ-7 (50 ng/ml) was tested for tubular formation activity. a, the supplied medium induced tubular formation. b, tubular formation was completely inhibited by an excess amount of neutralizing antibody (500 ng/ml) for human VEGF-A. c and d shows tubular formation by VEGF-E NZ-7 without and with an excess amount of neutralizing antibody for human VEGF-A, respectively. It indicates that VEGFR-2 activation only by VEGF-E NZ-7 is sufficient for tubular formation. B, chimera mutants (50 ng/ml) were tested for tubular formation. The number in each panel indicates the serial number of chimera mutants. 9, 12, and 18 show the tubular formation activity as effective as wild type VEGF-E NZ-7 . C, quantitative analysis of tubular formation with mutant VEGF-E NZ-7 (see "Materials and Methods"). Control, VEGF-A (50 ng/ ml); Ab, anti-VEGF-A antibody (500 ng/ml); E, VEGF-E NZ-7 (50 ng/ml); number, each chimera mutant VEGF-E NZ-7 (50 ng/ml).
FIG. 8. Schematic representation of exchangeable regions and non-exchangeable regions for VEGF-E NZ-7 to interact with VEGFR-2. Exchangeable regions are shown in red, unexchangeable regions obtained from the chimera study in blue, and the regions conserved in VEGF-E NZ-7 and PlGF in green. Regions exchangeable with PlGF in a detailed chimeric analysis (Fig. 4B) are shown in weak red color. Exchangeable regions include the NH 2 terminus, COOH terminus, loop-2, ␣-helix, and part of loop-1 regions. Non-exchangeable regions include the middle part of loop-1 and the broad region associated with loop-3. By introducing these "non-exchangeable regions" of PlGF to VEGF-E NZ-7 significantly reduces activities for receptor interaction and endothelial proliferation. Note that the conserved regions were not mutated. These results suggest that loop-1 and -3 play the important roles in providing the binding determinant for VEGFR-2.
10 and 11 for loop-1, 14 -17 for loop-3) also produced functionally defective proteins, whereas the loop-2 exchange did not reduce activity (mutants 12 and 13). Mutant proteins with low affinity to VEGFR-2 could not induce autophosphorylation of VEGFR-2, proliferation of HUVEC, or tubular formation by endothelial cells except for chimera 18. These results suggest that the affinity strength to the receptor basically reflect their effectiveness in biological response.
In addition to the replacement by PlGF, VEGF-A regions were also introduced to VEGF-E NZ-7 in a loop-specific manner, as a reference (Figs. 4A and 5A). Surprisingly, these mutants with VEGF-A loops also had significant reduction in biochemical and biological activities. It could suggest that a single loop of VEGF-A is not enough to be biologically functional. A pair of loop-1 and -3 of VEGF-E NZ-7 or VEGF-A may be required to build up the receptor-binding determinant for VEGFR-2. Mutants 10 and 14 did not bind to VEGFR-2 and lost biological activity. Mutant 18, which is a combination of numbers 10 and 14 with regions of loops-1 and -3 of VEGF-A onto VEGF-E NZ-7 , had a tubular formation activity, whereas it did not show activities in other VEGFR-2-associated analysis at detectable levels. Chimera 18 molecule might interact with VEGFR-2 better in a two-dimensional co-culture system that seems closer to the in vivo situation, compared with the one-dimensional culture of HUVEC.
As reported previously, regional exchange of the VEGF-A loop-3 region with that of PlGF resulted with significant reduction in VEGFR-2 binding and proliferation of endothelial cells (36,37). However, its activity in inducing vascular permeability was still functional (37). It was proposed that the determinant of VEGF-A to facilitate vascular permeability could be different from that to bind to VEGFR-2 and induce endothelial cell proliferation. It is also possible that the Miles assay, an assay for activity to facilitate vascular permeability could be very sensitive compared with other assays, so that activity of that mutant was detectable. The construction pattern of two chimera proteins in this study, numbers 15 and 17, are similar to such loop-3 exchanged mutants of VEGF-A. Interestingly, these VEGF-E NZ-7 chimeric proteins lost the activity to facilitate vascular permeability, 2 indicating that the loop-3 region of VEGF-E NZ-7 was critical in this particular biological activity. This result suggests that some of the biochemical properties of VEGF-A and VEGF-E could be different, especially in the aspect of the mechanisms for facilitating vascular permeability.
The chimera proteins that could be clinically applied as angiogenic stimulators include numbers 1, 6, 9, and 33. In keeping with the strong activities as wild type VEGF-E NZ-7 , they have human PlGF amino acid residues at the amino and/or carboxyl termini instead of viral amino acid residues of VEGF-E NZ-7 . Amino-or carboxyl-terminal peptide sequences are generally highly immunogenic. Therefore, these "humanized VEGF-E family" are expected to be less immunogenic compared with the wild type VEGF-E NZ-7 and clinically more useful.
The important region of VEGF-E NZ-7 for interacting with VEGFR-2 was identified as loop-1 and -3. Loop-1 and -3 may be playing an important role in presenting the binding determinant of VEGF-E NZ-7 for VEGFR-2. According to these results, a heterodimer in which one molecule binds only VEGFR-2 but the other does not bind either VEGFR-1 or VEGFR-2 could be designed and synthesized. Such peptides may be very useful as a VEGFR-2 antagonist for clinical application to inhibit the pathological angiogenesis in which VEGFR-2 plays the major role as a signal transducer.
|
2018-04-03T00:56:53.218Z
|
2003-04-11T00:00:00.000
|
{
"year": 2003,
"sha1": "619e3e02bb4b45283ac74f62a788ec583a270807",
"oa_license": "CCBY",
"oa_url": "http://www.jbc.org/content/278/15/13453.full.pdf",
"oa_status": "HYBRID",
"pdf_src": "Adhoc",
"pdf_hash": "7bc0c99c4491a77aa10770b4117a8823906e34ca",
"s2fieldsofstudy": [
"Biology",
"Chemistry"
],
"extfieldsofstudy": [
"Biology",
"Medicine"
]
}
|
119460808
|
pes2o/s2orc
|
v3-fos-license
|
General Backgrounds for higher spin massive particles
We consider the propagation of totally symmetric bosonic fields on generic background spacetimes. The mutual compatibility of the dynamical equations and constraints severely constrains the set of geometries where consistent propagation is possible. To enlarge this set in this article we allow several background fields to be turned on. We were able to show that massive fields of spin s greater than or equal to three may consistently propagate in a large set of non-trivial spacetimes, such as asymptotically de-Sitter, flat and anti-de-Sitter black holes geometries, as long as certain conditions between the various background fields are met. For the special case of massive spin-2 fields the set of allowed spacetimes is larger and includes domain-wall-type geometries, such as the Freedman-Robertson-Walker metric. We comment on the assumptions underlying our study and on possible applications of our results.
Introduction
Consistent theories of interacting higher spin fields are difficult to construct. Higher spin gauge invariance poses strong constraints even in flat space, leading to several no-go theorems [1,2,3,4,5,6,7]. However, a non-linear, interacting theory has been succesfully constructed in AdS space [8,9]. Moving away from AdS space still remains challenging.
In fact, even the free propagation of massive higher spin fields in non-trivial backgrounds is a challenging task. As first noted by Fierz and Pauli [10], the system of equations of motions and constraints governing the propagation of higher spin particles generically ceases to remain mutually compatible in the presence of nontrivial background fields. One may resort to a Lagrangian formulation, which is free from these kind of difficulties, but generically suffers from the Velo-Zwanziger problem, i.e., leads to superluminal propagation [11,12,13,14,15,16].
For a special class of spacetimes, those of constant curvature, a consistent Lagrangian desciption of free, massive higher spin fields is available as long as non-minimal terms are introduced [17,18,19,20,21,22,23,24,25]. It is natural to wonder whether a consistent system of equations of motions and constraints can be found without the help of a Lagrangian formulation. Indeed this is possible even for backgrounds of non-constant curvature [26,27]. The method employed in [26] is rooted in a systematic deformation of the dynamical equations and constraints away from the flat case using the formalism developed in [28].
In this article we will focus on totally symmetric, bosonic fields, of mass m denoted by rank-s symmetric and traceless Lorentz tensors, i.e., ϕ µ1···µs . The Fierz-Pauli equations describing the free propagation of these fields in flat space are given by: (∂ 2 − m 2 )ϕ µ1···µs = 0, ∂ · ϕ µ1···µs−1 = 0, ϕ ′ µ1···µs−2 ≡ ϕ µ1···µ in D spactime dimensions. The commuting nature of the ordinary derivatives renders the Fierz-Pauli system mutually compatible. This is no longer true in the presence of nontrivial background fields, because covariant derivatives do not commute with each other. Yet one may restore consistency by systematically including non-minimal terms to the covariantised version of (1) [28]. In practice, one requires the closure of the algebra generated by the deformed d'Alembertian, the divergence and the trace operators. The resulting set of equations of motion and constraints become mutually compatible, but there is a price to pay: the background fields are constrained to satisfy identities derived from the closure of the algebra.
Here Y µνρ , Z µνρ , X σλ µνρ are irreducible Lorentz tensors constructed out of the Ricci scalar, the Ricci curvature and the Weyl tensor of the background spacetime (explicit expressions can be found in eq. (21). When s = 1 there are no restrictions on the spacetime and when s = 2, (4) reduces to Y µνρ = Z µνρ = 0, leaving X σλ µνρ free to take any value. Here we attempt to find a consistent set of dynamical equations and constraints for massive higher spins in more general spacetimes. We will use the same methodology, but we will allow additional background fields to be turned on. To simplify the analysis, we will focus on traceless, symmetric fields and consider deformations only to the dynamical equations and divergence constraints.
The organisation of this article is as follows: First we review the methodology and results of [28,26,27]. In section 3, we introduce further deformations to the involutive system by allowing more background fields to be turned on. By insisting on the vanishing of commutator between the d'Alembertian and the divergence, we determine necessary conditions the varioius background fields must meet. Our results are presented in section 4, where we also discuss background geometries of special interest. In section 5 we detail the conditions necessary for consistent propagation to all orders in the curvature and we conclude in section 6, with a discussion of the methodology and its implicit assumptions together with possible applications of our results. Most of the technical details can be found in the appendices.
Conventions & Notations:
We use a mostly positive metric convention. The notation a(n) denotes the collection of indices: a 1 · · · a n . When the indices are placed in parenthesis, (a 1 · · · a n ), the expression is totally symmetrised in all the indices a 1 , · · · a n with a normalisation factor 1 n! . The tensor symbol g a1···an,µ1···µn , with g µν the inverse metric of the backrgound spacetime, denotes the following expression: g a1···an,µ1···µn ≡ g a1b1 · · · g anbn δ µ1···µn b1···bn (5) with δ µ1···µn b1···bn = δ µ1 b1 · · · δ µn bn . The gravitational covariant derivative is denoted by ∇ µ and the commutator obeys: The notation A a(n) · B b(m) implies a signle index contraction as follows: 2 Review of the formalism Consider the Fierz-Pauli system acting on symmetric and traceless higher spin fields: This is an involutive system of partial differential equations of second order. The term involutive implies that the system comprises of all the differential consequences of order ≤ p from any order-p subsystem.
From the differential consequences one can derive the so called "gauge identities", which should not be confused here with any gauge symmetries. It is easy to see that the Fierz-Pauli system (8) satisfies the following "gauge identity" The nature of the gauge identity can be made manifest if we express it as follows: it is clear that it corresponds to the vanishing of the commutator between the d'Alembertian operator and the divergence. Had we included the trace condition in the system, we would have had three gauge identities, in other words, three commuting operators defining an abelian algebra: the d'Alembertian, the divergence and the trace [29].
In what follows it will be useful to introduce the gauge identity generators: L a1···as−1 L a1···as−1,µ1···µs = η a1···as−1,(µ1···µs−1 ∂ µs) , L a1···as−1,µs···µs−1 = −η a1···as−1,µ1···µs−1 (∂ 2 − m 2 ) . (11) and express the gauge identity using the compact notation: where the index µ collectively denotes the indices of each subsystem of the involutive system and the sum runs over all subsystems. Now we consider coupling the Fierz-Pauli system to gravity and analyze the deformations order by order in the curvature. Starting from the minimal coupling ansatz, we promote all the derivatives into covariant ones: The resulting system is consistent to zeroth-order in the curvature. At linear order the covariantised gauge identity (7) fails to be satisfied due to the non-commutativity of the covariant derivatives. This is precisely the anomaly noticed by Fierz and Pauli [10]: To push the failure of the gauge identities to O(R 2 ) we introduce first order deformations to the dynamical equations, T 1 , and gauge identity generators, L 1 : This implies that the first order deformations should satisfy In practice, one tries to bring the expression for the anomaly term (14) in a form which will allow for the identification of the apporpriate deformations L 1 and T 1 in accordance with (16). The proposed identification may sometimes give non-local solutions for the first order deformations. If we insist on preserving locality and the number of degrees of freedom, we may have to impose restrictions on the external background. Constraints on the higher spin field are not allowed since they will ruin the involutive form of the system and/or the number of degrees of freedom count. As shown in [26,27] the anomaly tensor for the covariantised Fierz-Pauli system can be expressed as follows: where R µνρσ , R µν , R respectively denote the Riemann curvature tensor, the Ricci tensor and the Ricci scalar of the external spacetime, and α is defined in (3). Comparing eq. (16) with the first order terms of the gauge identity (17) we identify the first order deformation of the dynamical equationsfrom the first line of (17): and the first order deformation of the gauge identity generators from the second line of (17): The remaining terms denoted collectively as B a(s−1) must be set to zero. B a(s−1) contains gradients of the Riemann and Ricci tensors which can be split into irreducible Lorentz tensors resulting in [26,27]: where with C µναβ the Weyl tensor. Summarising, as long as B a(s−1) = 0 or equivalently, Y µνρ = Z µνρ = X a1a2 µνρ = 0 the involutive system (2) is consistent to first order in the curvature. It is however easy to see that given the above restrictions on the background spacetime, (2) is consistent to all orders in the curvature. Firstly, the covariantised anomaly tensor (14) does not contain O(R 2 ) terms and secondly, all second and higher order terms in (15), i.e., L 1 ⊲ T 1 , vanish by virtue of (18) and (19).
3 Deformations including more general backgrounds.
As reviewed above, deformations T 1 of the covariantised Fierz-Pauli system T 0 are allowed in background spacetimes fulfilling B a(s−1) = 0 .
This condition is quite strong and consistent propagation is achieved for a very limited set of nontrivial backgrounds, such as constant curvature spaces and their products, including a particular set of domainwall geometries [26,27]. Here we would like to investigate the possibility of enlarging the allowed set of spacetimes upon which the deformed involutive system of Fierz and Pauli is mutually compatible. In particular, we are interested in exploring this scenario by turning on more background fields. In practice, we would like to ask the question: are there any further modifications to T = T 0 + T 1 allowing for more general backgrounds where higher spin particles can propagate consistently? In other words, is it possible to find deformations which would respect the gauge identity? Let us consider deformations of the system T = T 0 + T 1 in the form tensors should be completely symmetric and traceless and composed of background fields other than the curvature. In writing (23) we are implicitly making the assumption that consistent propagation is always possible in constant curvature spaces (implicit in using T 1 as defined in (18)) 1 . On the other hand, we remain agnostic as to the specific background field content of the tensors Q 1 ; we simply demand that there exist some other background fields such that Q are of the same order as the curvature. For reasons of simplification, we also preserve the structure of the covariant derivative. This is a very strong assumption; we know for example that in string theory, higher spin fields are charged under several background fields other than the metric. We hope to lift this assumption in future work.
The deformations (23) will in principle be accompanied by deformations of the gauge identity generators: with the L 1 given in (19) and M 1 to be specified by consistency requirements. Demanding that the gauge identity is satisfied to first order in the curvature leads to: Recall that T 1 , L 1 are defined so that: with B a(s−1) given in eq. (20). Hence, the vanishing of the gauge identity to first order in the curvature yields: Our task will be to specify Q (26) is true. Note that the terms involving M 1 should cancel certain O(R) terms coming from the first line of (26). To this order, they do not have an impact on the consistency conditions of the background fields. However, they generically produce new, non-trivial O(R 2 ) terms on the right-hand side of equation (25). Requiring then that (25) is true to all orders in the deformation, will result in additional conditions for the background fields which are explored in section 5.
Let us now discuss in detail the contributions to the deformed Fierz-Pauli equations coming from the background fields. Generic tensors Q µ(s) and Q µ(s−1) with the required symmetry properties are constructed from all the possible couplings between irreducible tensors of mixed symmetry and the higher spin fields: and where (a 1 , · · · a 7 ) and (b 1 , · · · b 3 ) are arbitrary coefficients. All the information on the external background fields is hidden in the irreducible tensors of (27) and (28) which correspond to the following Young diagrams : Tensors corresponding to the diagrams do not couple to the higher spind fields due to symmetry properties. Similarly, the deformation of the constraint equation involves coupling the fields to the tensors: Tensors in the following representations do not couple to the higher spin fields. Finally, higher rank background tensors are not expected to contribute in the anomaly cancellation.
In order to compute the gauge identity we need to perform basically the computations involved in the (L 0 ⊲ Q 1 ) a(s−1) terms of eq. (26). In practice, we need to bring the sum of the two terms in (29) in the form −M 1 ⊲ T 0 + rest by pushing the derivatives all the way to the higher spin fields. The computations are tedious but straightforward. Here we will give a few examples, for more details the interested reader should consult appendix B. The first term in Q µ(s) 1 is given by a 1 U ϕ µ(s) . We need to compute: The second term in (30) is exactly of the type we are after, defining a part of M 1 . The first term must be cancelled against terms in B µ(s−1) coming from the orginal anomaly. Let us do one more example, which is slightly less trivial: where we used the definition of indices symmetrisation to go from the first to second line and from the second to the third. Once more there are terms which cannot be written in the form (M 1 ⊲ T 0 ) a(s−1) and should be cancelled by terms in the anomaly. With similar techniques we can work out all the relavant terms produced by hitting with a derivative on Q µ(s) 1 of eq. (27). Likewise we treat terms of the form −(∇ 2 − m 2 )Q 1 µ(s−1) . Computations in this case are slightly simpler. As an example consider: Proceeding this way leads to: where C a(s−1) is a complicated expression involving the extra background fields and their derivatives, parts of which we saw in the examples above. The explicit form of M 1 can be found in Appendix D.
Preservation of the gauge identity to first order in the curvature demands that: This condition corresponds to a system of differential equations the background fields should satisfy for higher spin fields to porpagate consistently on them. Explicit results are presented in the following section. The reader interested in the intermediate steps of the computation is encouraged to consult Appendices B and C.
Results
In this section we focus on the conditions the backgrounds fields should satisfy. For fields of generic spin s ≥ 3, eq (34) allows for a very small set of non-vanishing background tensors, namely: In other words, without loss of generality one may set 2 : The remaining non-trivial background fields (35) must satsify the following set of equations and The latter conditions, (38) follow from the tracelessness and the other symmetry properties of the fields involved in (37). Assuming the above equations are satisfied, propagation of higher spin fields is described by the following involutive system 3 : The case of spin-2 fields is special. As expected, there is more freedom in the choice of background fields. In particular, X µ1µ2 νρσ is not constrained and the same is true for several of the additional background fields introduced. For a minimal solution we set: The remaining background fields should satisfy: 2 With the help of the identities in Appendix B, one can show that all the allowed terms in (27) are proportional to (a 4 − a 5 ) and thus vanish. 3 Notice that background scalar function U is allowed as long as it satisfies : a 1 ∇ν U = 0. So U is a constant. This reflects the ambiguity in the definition of the mass parameter in arbitrary spacetimes (recall that U couples to the higher spin field exactly like the mass). The simplest solution is to set: a 1 = 0 and the involutive system reduces to .
Let us end this section with a final comment. Eqs (37,38) and (42) should be viewed as (part of) the set of differential equations imposed upon the background fields by an underlying consistent, but otherwise unkown theory, where massive higher spin fields can propagate.
General comments on the solutions
Fields of spin s ≥ 3.
• The first observation we can make is that spacetimes with non-vanishing Y µνρ are not allowed even in the presence of other background fields. This is a very strong constraint and excludes all conformally flat spacetimes except for the ones studied in [27].
• Given a background geometry, one can try to solve the conditions (37) to obtain explicit expressions for the other background fields in terms of the geometric data. In general, this is a difficult problem. We can simplify it by noticing the following consequences of (37) and (38): and Notice that (45) and (46) define an algebraic system of n equations for n unkowns; where n are the independent components of the unkown background fields S ρ , S λ|ρσ 4 . Thus, given a certain background spacetime, S ρ andŜ µ|ρλ can be completely determined from (45) and (46). Recall that knowledge of (S ρŜµ|ρλ ) is sufficient to completely determine the involutive system since: , and Hence finding an appropriate solution has been reduced to a simple algebraic problem with the use of (45) and (46). Once a solution is determined, it should be checked that (37) and (38) are identically satisfied. If this is not the case, then a solution does not exist.
Notable cases of background geometries which may be interesting to study are those where besides Y µνρ = 0, X στ µνρ or Z µνρ also identically vanish.
-When Y µνρ = X µ1µ2 µνρ = 0, the background spacetime should satisfy: 4 In the former case, n = D whereas in the latter n = D(D−1) 2 It would be interesting to have examples of such spacetimes, if they exist at all. If they do, the simplest consistent solution of (37,38) in such a geometry would correspond to setting a 7 = b 3 = 0. We would then be left to specify the background field S ρ (since U ρλ is completely determined from S ρ and the ambient spacetime).
-When on the other hand, Y µνρ = Z µνρ = 0, then according to (21) In this case, a simple solution corresponds to setting a 2 = b 1 = 0, together with where Λ µνρσ is an arbitrary tensor with the symmetry properties of the Weyl tensor and additionally satisfies: For instance, Λ µνρσ can be equal to the Weyl tensor of any symmetric space.
One should then determineŜ µ|ρσ by combining (46) together with (38), ensuring in particular that: A large number of spacetimes belongs to this class. Recall that the divergence of the Weyl tensor can be expressed in terms of the Cotton tensor: which is in turn defined in terms of the Schoutten tensor through Fields of spin s = 2.
• The simplest possible solution to the set of conditions (42) corresponds to setting a 1 = b 2 = a 2 = 0, and reducing the system to: .
Note that in this case only the constraint equation is modified and the relevant background fields satisfy a rather simple equation.
• It is clear that Y µνρ does not need to vanish and as a result it is possible to use the involutive system (43) or (55) to describe the propagation of spin-2 fields in conformally flat spacetimes. Conformally flat spacetimes are particularly simple to treat because Z µνρ = X µ1µ2 µνρ = 0.
• Particular cases of interest are the asymptotically AdS domain-wall geometries (relevant for holographic applications) and the Friedmann-Robertson-Walker metrics (relevant for cosmology). Consider the following metric: where D denotes the dimensionality of the spacetime and Ω D−3 the unit hypersphere. k takes the values (−1, 0, +1) which correspond to (D − 1)-dimensional AdS, flat and dS slicings respectively. The metric is conformally flat, so Z µνρ = X µ1µ2 µνρ = 0 automatically. On the other hand, Y µνρ does not vanish but takes a rather simple form with non-zero components: In order to find a solution of (56) we set b 3 = 0 and make the following ansatz: Substituting the ansatz into (56) leads to a differential equation s(y) must satisfy for consistency: Thus, given a conformally flat metric (57) with known f (y), the involutive system (55) describing the propagation of spin-2 fields in this background metric is consistent as long as another background field, S µνρ , which satisfies eqs (59, 60) is turned on.
For completeness, we present also the case of the four-dimensional Friedmann-Robertson-Walker metric: It is easy to see that as long as h(t) solves the following differential equation:
Conditions for an exact solution
In the previous sections we restricted our attention to the cancellation of the anomaly tensor to first order in the curvature. To obtain the cancelation without imposing restrictions on the ambient spacetime, we introduced new background fields. With the help of these new fields, our goal was reached. However, the new fields are bound to modify the anomaly tensor to second order in the curvature. This definitely limits the applications of our results.
For an all order's result, we must require the cancellation of the anomaly to second order in the curvature (higher order cancellation will be automatically achieved). To this end, we compute here the O(R 2 ) and demand that they vanish, i.e.
Fields of spin s ≥ 3. Explicit computation yields: where we took into account that a 3 = a 6 = b 2 = a 4 − a 5 = 0. Eq. (66) leads to the following necessary conditions the background fields should satisfy: The first line of (67) leads to the following simple solutions: where we used (45) to set S λ R λσ = 0 and (38) to express U ρσ in terms of S σ . The other two equations in (67) do not yield a simple solution but for specific curved spacetimes they can be combined with (46) and (38) to check whether an all orders solution exists.
Fields of spin s = 2. The conditions for spin-two fields can be worked out in the same manner as the general case. Focusing on the simple solution where the only non-vanishing parameters are (b 2 , b 3 ) we obtain the following quadratic equations: A comment is in order. Eqs (69) are sufficient but not necessary conditions for the existence of an all orders solution. The reason is that for spin-2 fields, various background fields can be turned on without any effect on the anomaly to linear order in the curvature. Such background fields will in general contribute to the anomaly only to second order in the curvature. One may thus fully determine them from the requirement of an all-orders cancellation of the anomaly, whenever (69) is not satified.
Discussion & Outlook
In this article we used the involutive method to derive equations describing the consistent propagation of higher spin fields in non-trivial spacetimes. To expand the set of possible background geometries where consistent propagation is possible, we allowed for additional background fields to be turned on. We remained agnostic as to the nature of these background fields. For simplicity, we limited the discussion to deformations of the dynamical equations and divergence constraints, leaving the trace constraint undeformed. We found that massive fields of spin s ≥ 3 cannot propagate consistently on geometries characterised by Y µνρ = 0. These unfortunately include all the non-trivial domain-wall type metrics. We were however able to show that consistent propagation may be possible in a large set of non-trivial spacetimes, such as asymptotically flat or dS/AdS black hole geometries, as long as certain conditions between the various background fields are satisfied. On the other hand, massive spin-2 fields are found to propagate consistently on dS, flat or AdS-domain wall spacetimes. Section 4 contains the main results of this work.
Several open questions remain. Most of them are related to the explicit or implicit assumptions under which this work was carried out. For reasons of completeness we list them here: • Consistency of free propagation of massive higher spin fields in arbitrary backgrounds is related to the existence of an abelian algera defined by the d'Alembertian, the divergence and the trace operators.
This assumption is plausible, being a natural generalisation of the flat space case. Nonetheless, other possibilities may also exist. Indeed, the authors of [29] argued that a bigger non-abelian algebra, which includes the algebra discussed here, may be necessary for consistency. It would be interesting to include other background fields in the analysis of [29] and examine the implications.
• The trace constraint remains undeformed.
In other words, the trace vanishes identically instead of as an on-shell condtion. The assumption was made for the technical simplification it provides. It is clear that it may have non-trivial consequences in a possible Lagrangian formulation of the system. On the other hand, as discussed in [27], one may view the original system (1) as the zero-trace gauge fixing of a system of symmetric rank-s fieds with Weyl symmetry: δϕ a1···as = g (a1a2 λ a3···as) . Then the freedom of the rank-(s − 2) parameter λ a1···as−2 allows one to choose the trace to vanish at the interaction level.
One may wonder whether, had we allowed for the deformation of the trace constraint, our results would be modified. In particular, would backgrounds with Y µνρ = 0 be allowed? We leave this as future work.
• At the linearised level, higher spin fields decouple from each other as well as from background fluctuations.
In practice we assumed that a basis exists such that diagonalisation is possible. This is a natural assumption when considering the free propagation of massive higher spin fields in curved spacetime, like in [26,27]. However, one expects mixing to occur at least with graviton fluctuations, when other background fields are present. For the sake of simplicity, we did not consider this case here but is definitely worth of further investigation.
• Inhomogeneous terms included in the dynamical equations and constraints need not be considered.
This assumption is justified from the standard analysis of linearised fluctuations in Lagrangian systems. At the same time, inhomogeneous terms would change the nature of the algebra formed by the d'Alembertian, the divergence and the trace operators. As a result, they would appear to spoil the consistency of the system. It is likely however, that with a careful choice of such terms, we could preserve the algebraic structure but promote it to a non-abelian one. In [29], a bigger non-Abelian algebra was considered as necessary for consistency. In this spirit, the requirement of an abelian algebra may be too strong. This point certainly deserves further study.
• Fields of mixed symmetry need not be included.
The absence of mixed-symmetry fields may be too strong of an assumption. It appears that in string theory neglecting mixed-symmetry fields corresponds to considering only the first Regge trajectory of string excitations, thereby excluding all the subleading ones [29]. Put differently, consistency in string theory requires the inclusion of mixed-symmetry fields. In a positive scenario, our simplified approach may lead to possibilities which are not realised in string theory. In a negative one, it may strongly restrict the set of backgrounds where consistency of propagation can be achieved.
• Higher-derivative, including various other quadratic terms, need not be included.
By preserving the derivative structure of the system, we ensure hyperbolicity and avoid superluminality. However, it is indeed expected that higher-derivative terms may be necessary for the contstruction of a consistent theory. At the linear level it may not be so crucial, nevertheless it would be interesting to see if more general spacetimes are allowed if this assumption is removed.
We would also like to draw our attention to the treatment of background fields. We did not specify the precise nature of the various tensors appearing in (27) and (28). They may correspond to higher spin fields taking on background values or they may be composed from expectation values of low spin fields and the metric tensor, similarly to the spacetime tensors X σλ µνρ , Y µνρ , Z µνρ . However, we assumed that whichever background fields are turned on, they do not alter the structure of the covariant derivative. This is a very strong assumption which may significantly restrict the applicability of our results. Moreover, we did not search for a theory which could derive the equations of section 4, involving various background fields and gravity. It would be interesting to see if such a theory can be constructed and/or if it is related to string theory in any way. One possibility, would be to consider string solutions supported on a non-trivial dilaton and a closed two-from B µν .
It is a curious fact that, for spacetimes whose Ricci tensor is covarianty constant, the involutive system obtained in [26] coincides for spin s = 2 fields with the system of equations derived to linear order in (α ′ ) from the string sigma model [34]. It would be interesting to see if the matching persists when additional string theory background fields are considered. Note however that in the presence of supersymmetry our results are likely to be modified.
Finally, we would like to mention posible applications of our results. Immediate applications concern the study of spin s = 2 fields in AdS-domain wall geometries both at zero and finite temperature. Similar studies in FRW cosmologies would also be of interest. In practice, one would like to study the equations semiclassically and determine whether bound or quasi-bound states would be possible. A stability analysis should also be performed.
Recently, there has been renewed interest in the study of massive spin s = 2 fields, motivated by the contruction of a consistent massive (bi)gravity theory [35,36]. Here we did not attempt to construct a full-fledged theory of higher spin massive particles, but focused only on their propagation in external backgrounds. A hasty calculation shows that the linearised equations of [35,36] differ not only from the ones obtained herein but also from [26,27] (see for example, [37,38] or [39] and references therein). It would be interesting to confirm this and study its phenomenological implications.
For higher spin fields, the simplest and most interesting case involves black hole geometries. In particular, investigations in the context of the AdS/CFT correspondence may be relevant for the description of different phases in strogly coupled quantum field theories.
where we used the symmetry properties ofV µ|νρσ together with the identity (A.4) to write: Similarly, And from the constraint: Gathering the above terms with appropriate coefficients according to (27) and (28) -except for those which can be related to a deformation of the constraint and/or the dynamical equations -defines what is denoted as C a(s−1) in section 3. Demanding that C a(s−1) cancels out the remaining anomaly terms B a(s−1) leads to the following equations between the background fields 5 : C Simplifying the equations for the background fields.
We deal here with fields of generic spin s ≥ 3. The spin s = 2 case is much simpler and can be similarly which implies that all the terms proportional to a 4 and a 5 in ( 27) are actually proportional to the linear combination (a 4 − a 5 ), and thus vanish for a 4 = a 5 . The set of equations necessary for the vanishing of the anomaly term, can be then expressed simply as follows: a 2 s U νρ + 2b 1 ∇ ν S ρ = 0, ⇒ ∇ (ν S ρ)=0 , ∇ ρ S ρ = 0, ∇ 2 S ρ = 0 a 1 ∇ ν U = 0 . (C. 5) Notice, that there is an ambiguity in the value of U , which reflects the ambiguity in the definition of the mass parameter in arbitrary spacetimes.
D First order deformations of the gauge identity generators.
Here we write down explicit expressions for the correction terms to the gauge identity generators. They play an important role for deformations to second (and higher) order in the curvature. It is straightfowrard to write down M a(s−1) 1 µ(s−1) and M a(s−1) 1 µ(s) using eqs (B.1-B.8) while bearing in mind the relevant deformation of the dynamical equations and constraints, e.g., (27) and (28). In particular, we obtain: where we took into account that a 3 = a 4 − a 5 = a 6 = b 2 = 0 as required by the anomaly cancellation to linear order in the curvature for massive fields of spin s ≥ 3.
In the case of spin s = 2 massive fields, and for the minimal solution leading to (55), we have that:
|
2019-04-13T19:58:32.398Z
|
2017-11-30T00:00:00.000
|
{
"year": 2017,
"sha1": "169e14610a77686e1640beb506cf7cee7f11eb0f",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "1d7abdf05ad2a769792e36e2a692bc6162d12587",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
}
|
270585129
|
pes2o/s2orc
|
v3-fos-license
|
A competence set for sustainable urban development: framing a research agenda
The primary purpose of this paper is to call for more research on the competencies needed to promote sustainable urban development. Our premise is that sustainable urban development does not occur without transcending differing interests and identifying novel ways to collaborate across organisational and institutional boundaries. Relatedly, sustainable urban development calls for mobilisation and pooling of scattered assets. Therefore, sustainable urban development calls for enhanced competencies that significantly differ from the traditional capabilities valued in public administration. This premise leads us to determine what competencies are needed to support sustainable urban development and to ask how fragmented capabilities can be pooled to serve the common purpose. Sustainable urban development necessitates transformative system change, dependent on diverse stakeholders, relying on many actors’ knowledge and capabilities instead of a few selected actors’ expertise. To achieve a systemic perspective, we need to be able to group the capabilities and competencies; otherwise. We propose a conceptual framework drawing on insights from public administration, management studies, organisation studies, and the Intergovernmental Panel on Climate Change (IPCC) that might allow us to move from individual capabilities to shared competencies and collective learning processes, thus adding analytical leverage to our efforts to strengthen the competencies embedded in systems but held by individuals. For analytical purposes, we adopt the concept of the competence set. It is geared toward identifying how different capabilities of many actors could be integrated at a systemic level so that a set would serve both the entire urban system and the actors embedded in it.
INTRODUCTION
The primary purpose of this paper is to call for more research on the competencies needed to promote and implement sustainable urban development (SUD).
Cities are not only centres of economic activity but also major sites in addressing the climate crisis, as they represent approximately 75% of global final energy use (Ranalder et al., 2021).Consequently, making cities sustainable, inclusive, safe and resilient is one of the United Nations Sustainable Development Goals (Transforming Our World, 2015), and SUD is the guiding theme in the New Urban Agenda (Habitat III, 2017) and several policy documents published at the European scale (Territorial Agenda 2030, 2020).As cities are confronted with ecological, social and economic challenges (Millennium Ecosystem Assessment, 2005;OECD, 2010), the essential question is posed as follows: what competencies are needed to effect such a profound change that would allow cities to simultaneously meet ecological requirements and social and economic needs?
The difficulty in SUD is that social and environmental goals are often regarded as obstacles to economic development, and vice versa.SUD is difficult to achieve in practice because of a short-term perspective dominating politics and the many conflicts of interest.SUD is characterised by multi-actor and multi-value processes involving overlapping networks of public, private, community and academic actors, in which no organisation has a primacy in governance (Padt, 2007).Moreover, the vague nature of the SUD concept enables it to be understood and utilised in numerous ways and affects how SUD is understood and how actors perceive it.Thus, the practical and political implications of various interpretations attached to it in practice remain unknown (Griggs et al., 2017).
Consequently, in this short paper, we suggest that the urban and regional development community pay greater attention to what they need to learn to successfully meet the new demands, as the new challenges are not possible to be addressed with yesterday's competencies.While we know that actors working for urban development have recognised their key positions in promoting SUD, why cities need to strive for it (Habitat III, 2017) and what action should be taken (Alvarez-Risco et al., 2020), we do not yet have in-depth knowledge of how all necessary measures could be carried out and which actor capabilities and competencies are required.
Our premise is that (a) SUD does not occur without transcending differing interests and identifying novel ways to collaborate across organisational and institutional boundaries.Relatedly, (b) SUD calls for mobilisation and pooling of scattered assets.Therefore, (c) SUD calls for enhanced competencies that significantly differ from the traditional capabilities valued in public administration (PA).This premise leads us to determine what competencies are needed to support SUD and to ask how fragmented capabilities can be pooled to serve the common purpose.In the following sections, we first define the concepts of capability and competence and finally argue for the need to pool competencies into competence sets.We follow Borrás et al. (2023) and believe that existing insights can be mobilised to understand the formation of capabilities and competencies for SUD at multiple levels -from the individual level to complex networks and systems (Capasso et al., 2019).
As the literature on capabilities in urban and regional development is scarce, not to mention SUD, we believe it is necessary to make use of the research work done in other disciplines.At the same time, regional development studies are uniquely placed to contribute to such debates, given previous work on such topics as spatial policies during crises (Martin et al., 2022), place leadership (Sotarauta & Beer, 2021), and the geography of sustainability transitions (Binz et al., 2020;Hansen & Coenen, 2015).We acknowledge that the many concepts around competency-related issues have been interchangeably applied.In this paper, we propose a conceptual framework drawing on insights from public administration, management studies, organisation studies and also the Intergovernmental Panel on Climate Change (IPCC) that might allow us to move from individual capabilities to shared competencies and collective learning processes, thus adding analytical leverage to our efforts to strengthen the competencies embedded in systems but held by individuals.
BASIC TENETS OF COMPETENCE THINKING
The concepts of capability and competence are often synonymously employed.However, for research purposes, we need to make an analytical distinction between capabilities, competencies and competence sets (Pralahad & Hamel, 1990;Teece et al., 1997).This line of thinking has mainly, but not only been applied in the business world.Those studies that have focused on capabilities in urban and regional development have mostly followed business scholars and applied their conceptualisations.For example, Best (1999) used Penrose's (1965) theory, highlighting dynamics between capabilities and market opportunities.Also, Laasonen and Kolehmainen (2017) followed the resource-based view (Penrose, 1965;Wernerfelt, 1984) in their study on regional economic actors' and networks' capabilities (see also, for example, Laasonen, 2023;Lawson, 1999;Pihkala et al., 2007;Sotarauta, 2005).These studies show it would be possible to enhance SUD by systematically identifying the complementing capabilities and consciously constructing them.This task is challenging as capabilities must be pooled at a system level, at the intersection of multiple -often conflicting -intentions, interests and aims.A more nuanced conceptual framework may prove helpful when analysing capabilities in systems.Javidan (1998, p. 62) used the concept of competence to refer to the combining and coordinating of capabilities, cutting across functions.Teece et al. (1997) defined capabilities as the firm's ability to integrate, build and reconfigure internal and external skills and resources to address rapidly changing environments.Supporting SUD requires capabilities ranging from communication to marketing, from leading to regulation, from finances to technologies and so forth.We approach competencies as interlinked capabilities that are scattered throughout a network or a system but that may also be embedded mainly in a single organisation having a central position in a system (Sotarauta & Heinonen, 2016).In summary, in the context of SUD, competence is a nested concept that cuts across specific capabilities of several organisations.We assume collective competencies, drawing on individual actors' specific capabilities, to determine the quality of sustainable development strategies and their implementation.
According to Pralahad and Hamel (1990), identifying and developing core competencies is an arena for coordinating and integrating dispersed functions and capabilities.The authors approach, accordingly, core competencies as collective learning.For Pralahad and Hamel (1990), core competencies comprise a collection of capabilities shared in a multi-industry and multi-site organisation.In the context of SUD, identifying and enhancing core competencies could be a way to close scattered actors' capability gaps, facilitating the identification of various organisations' roles in the system.
INDIVIDUAL LEVEL CAPABILITIES
According to Borrás et al. (2023), few studies address the competencies needed at the individual level for transitions towards sustainability, even if it is widely acknowledged that the needed competencies are rapidly changing.As competence is a nested concept in the SUD context that cuts across the capabilities of several organisations, we need to understand individual capabilities from a relational perspective, highlighting competencies as a nexus of internal and external assets, powers and capabilities.Consequently, we need to contrast the needed competencies regarding the traditional PA philosophy of government with the new public management (NPM) and new public governance (NPG).In PA, the responsibility of civil servants is to support the construction of rational policies and to ensure that they are properly implemented.Civil servants are supposed to make decisions by relying on their expertise in relevant fields.The longheld assumption is that the challenges civil servants have to overcome in policy development and implementation are mainly technical.Consequently, key capabilities include understanding juridical procedures, having administrative skills, and foreseeing and avoiding future risks (Kruyen & Van Genugten, 2020).
The central idea of NPM is the implementation of management principles from the private sector in government; consequently, the capabilities required from civil servants are very different.However, briefly summarised, many of the needed capabilities are very similar to the capabilities required by private-sector employees.Capabilities that ensure that the public sector delivers 'value for money' are emphasised.These capabilities include good skills in preparing, signing and surveying contracts with private sector actors.Thus, having negotiating and controlling skills is very important in NPM.Additionally, being customer-oriented is essential as citizens are essentially considered customers who buy certain services (Kruyen & Van Genugten, 2020).
The issue then is how the complexities involved in SUD influence the needed capabilities.This issue is related to the increasing emphasis on 'New Public Governance' (Osborne, 2006), which relies more on interactive modes of governance than the more straightforward NPM.Consequently, it is crucial to determine what capabilities public sector employees need when the policy is aimed at stimulating transformative change through, by and with diverse groups of actors.Kattel and Mazzucato (2018) focus on the needed capabilities in the public sector to successfully develop mission-oriented innovation policies but remain at a generic level (all quotes on p. 797): 'capabilities for leadership and engagement' -'capabilities to encourage bottom-up engagement' -'the ability to identify coherent policy mixes (instruments and funding) and capabilities for coordination' -'experimentation capabilities' -'evaluation capabilities'.
OECD (2017) and Nesta (2019) provided us with more specific and detailed conceptualisations of the needed capabilities.The former emphasises six skills: iteration, data literacy, usercentricity, curiosity, storytelling and insurgency. 1 In particular, the latter very clearly illustrates changes in comparison to the needed capabilities from a traditional PA perspective, in which being disciplined and following the rules of bureaucracy are underlined.Nesta (2019) provided a toolkit that is explicitly developed to highlight the specific capabilities in the public sector that are needed to address wicked problems such as SUD.The toolkit makes a distinction between attitudes and skills.Attitudes are here understood as (p.9) 'settled ways of thinking or feeling about something'.Examples include being imaginative, outcomes-focused and courageous.In addition to attitudes, several specific capabilities are identified as important.These capabilities relate to three main areas: collaboration, leading change and accelerating change.Compared with OECD (2017), there is a considerable overlap: for example, storytelling, data literacy and curiosity are highlighted in both documents.
As an illustration of the changing needs in capabilities at the level of the specific positions related to SUD, we consider two different positions: the public procurement officer and cluster facilitator.A public procurement officer has traditionally worked to ensure low costs and set quality requirements for the procured goods and services.Traditionally, a mix of capabilities relating to traditional PA and NPM were needed: juridical knowledge, attention to detail, negotiation skills and commercial skills.However, with an emphasis on SUD, procurement practices undergo fundamental changes, requiring, for example, that procurement officers procure using functional requirements.Consequently, procurement officers must learn to translate long-term visions into specific procurement requirements and to communicate to a broader audience how procurement assists in addressing grand challenges (Edler & Georghiou, 2007;Edquist & Zabala-Iturriagagoitia, 2012).
A cluster facilitator, or a network manager, has traditionally worked to connect different actors to stimulate innovation.To successfully perform this task, one needs project management, facilitation and communication skills.However, with the transformation towards sustainability as the central goal, the objective is to facilitate specific types of innovations and to set or follow a particular direction for change for a cluster (Sjøtun & Njøs, 2019).This approach requires system thinking capabilities to understand the localised consequences of the activities of the cluster.In addition, 'selective facilitation skills' are called for; i.e., the capability to understand and select the actors that need to be connected and supported to achieve a certain direction of change are valued (Grillitsch et al., 2019).Selective facilitation is crucial, as public actors are expected, on their part, to launch SUD-supporting processes and mobilise key stakeholders.Other actors are needed to contribute to the identification of the opportunities and to the finding a way to achieve them.
To make the individual capabilities serve SUD, we need to identify a way to study and develop them at a systemic level.Next, we discuss the concept of competence set as an approach to understanding how to pool individual capabilities and translate them into systemic competencies.
COMPETENCE SET TO SUPPORT SYSTEMS THINKING
Sustainable urban development is not only about boosting technological advancements.It necessitates transformative system change, dependent on diverse stakeholders (Coenen et al., 2015), relying on many actors' knowledge and capabilities instead of a few selected actors' expertise.To achieve a systemic perspective, we need to be able to group the capabilities and competencies; otherwise we would lose ourselves in detail.Therefore, for analytical purposes, we adopt the concept of the competence set.It is geared toward identifying how different capabilities of many actors could be integrated at a systemic level so that a set would serve both the entire urban system and the actors embedded in it (Sotarauta & Heinonen, 2016).
As Markard et al. (2012) pointed out, the literature on socio-technical transitions has provided us with a dynamic conceptual framework to explain and understand how systems change.Meadows defines a system as 'an interconnected set of elements that is coherently organised in a way that achieves something' (Meadows, 2008, p. 11).She posited elements, interconnections and a purpose as the three main constituents of a system.Drawing on their literature review on systems, Arnold and Wade (2015, p. 675) defined systems thinking as follows: 'Systems thinking is a set of synergistic analytic skills used to improve the capability of identifying and understanding systems, predicting their behaviours, and devising modifications to them in order to produce desired effects.These skills work together as a system'.(Arnold & Wade, 2015, p. 675) Renewing or transforming any system's elements, interconnections or purpose is daunting, calling for well-established capabilities to think and understand systems and a variety of stakeholders.Representing a globally significant body that assesses the science related to climate change and translates scientific observations into policy, the United Nations' Intergovernmental Panel on Climate Change (IPCC) provides a broad perspective on what is needed in policy and practice.Even though the IPCC has strongly emphasised the need for system-level changes, it has concluded that the capabilities (from individual to organisational to societal level) required to push for such a change are unclear due to a shortage of empirical evidence and successful practical examples (IPCC 2022a).The IPCC also acknowledged -with high confidence -that mitigation of climate change is place-specific, and therefore 'the enabling conditions for shifting development pathways towards increased sustainability will therefore also differ, giving rise to different needs' (IPCC 2022b, p. 43).In its reports, the IPCC mentions numerous capabilities, ranging from testing to asset mobilisation to expectations management.To understand the wide range of needed capabilities, drawing on an earlier extensive empirical study in Finland (Sotarauta et al., 2007) and our reasoning based on the conceptual analysis, we group capabilities into three main competence categories, which are substance, institutional and process competencies.(Figure 1) In our framework, substance competencies include knowledge and expertise in specific technologies, services, industries, sciences, etc. Substance competencies involve an endless array of 'know-why' and 'know-what' related capabilities.'Know-why' represents 'an understanding of the principles underlying the construction of each component and interactions between them' (Garud, 1997, p. 84).Thus, 'know-why' refers, for example, to obtaining appropriate technological solutions or incentive structures for SUD strategies and relates measures and tools.'Know-why' also refers to the understanding and internalising of the requirements set by SUD-policies for practical work.More specifically, 'know-why' related capabilities may revolve around waste management, circular economy, digitalisation of services, green public procurement, etc. Basically, anything with a promise to reduce emissions or otherwise enhance SUD is included.Importantly, as Garud (1997) reminds us, knowing what to do, and why to do it, resides at the nexus of substance competencies.
Institutional competencies are a collection of capabilities to improve the playground and rules of the game for SUD (applying North, 1991), creating distinct and flexible frameworks for the substantial SUD work.Institutional competencies also include all the capabilities needed to improve regulative, normative and cultural-cognitive institutions in a specific SUD context (applying Scott, 2001).The foremost thing is to improve the institutional environment and arrangements to enhance human agency and to identify ways to align actors' expertise and capabilities.For example, the public sector not only provides various actors with indirect and diret subsidies, but it also encourages the efficient mobilisation of private sector finances by creating a supportive institutional framework by correcting failures, removing barriers, providing information and sharing risk (IPCC, 2022b, p. 135).The IPCC (2022a; 2022b) also accentuates the importance of cities' institutional capacity to develop, coordinate and integrate sectoral strategies within their jurisdictions, acknowledging the significant differences between cities globally.In an ideal world, the SUD development network forms a tightly knit entity where the fundamental issues of SUD are widely shared and internalised and different organisations recognise their role as part of the whole.Of course, the world is far from being ideal.
Moreover, as Mazutis and Abolina (2019) remind us, we should never overlook the complex act of implementing all the strategies and concrete ideas to support SUD.Therefore, we include not only technical tasks such as ensuring that a system or a network is aligned with policies and formal alliances but also tasks such as securing funding and lobbying the government into the broad category of institutional competencies (Mazutis & Abolina, 2019).Moreover, all the new strategies and ideas need to be institutionalised, requiring well-established and shared institutional and process competencies.In summary, institutional competencies are needed to improve institutions, efficiently act on them and institutionalise innovations and new practices.Process competencies refer to the capabilities for leading and managing interactive processes, creating conditions for commitment, and seeking future possibilities, but also to personal, energising charisma.'Know-whom' and 'know-how' capabilities are typical building blocks for integrated process competencies (Defillippi & Arthur, 1994;Garud, 1997).'Know-how' related capabilities enable one to understand and guide complex SUD-related networks -how to get new ideas through planning and decision-making processes, how to persuade different actors to collaborate, how to prevent conflicts and identify common interests among the goals of other actors, etc. (Hinchlifffe, 2002;Lundvall & Johnson, 1994).'Know-whom' capabilities are social capital skills -who are the key individuals, how do they think, what do they support and/or oppose, what are their goals, etc.
The IPCC (2022b) showed how institutional competencies are dependent on well-developed process competencies.According to the IPCC, clear climate policy signals support mobilising privare funding and guides investment decisions.If governments and the international community were capable at credibly signalling about the investments opportunities and their future prospects, it would reduce uncertainty for financial decision-makers and thus mitigate anticipated risk.We see process competencies supporting what the IPCC is after, and what it deemed as missing in many cities (IPCC, 2022a(IPCC, , 2022b)).Process competencies include the capability to involve people and empower them to act as a network; the capability to make people work to reach joint and separate goals and renew them in an ongoing process, the capability to promote interactive processes serving as an intermediary in interaction between actors and steering activities towards seeking goals and enabling co-operation, and the capability to connect various actors to the knowledge pool from their own starting points (Sotarauta, 2005).These competencies also includes social capabilities referring to producing shared and often tacit knowledge potentially leading to the social integration of actors, reaching far beyond the integrative force of institutions.Absorbing new knowledge, people and money into SUD work is essential.Therefore, interpretive capabilities surface, i.e., capabilities to value, assimilate and apply new knowledge and to transfer visions and strategies into action are crucial in managing processes.
We should never take for granted that all the actors needed to work for SUD are keen on contributing to or participating in collective efforts.Thus, motivational capabilities belong to process competencies and need to be included in the competence set.One needs to understand many 'silent messages' or various actors conveying the compatibility or incompatibility of beliefs, values and identities.Thus, process competencies also need non-technical know-why capabilities.Here, process competencies also deal with 'existential' issues -why do we do all this, what are we aiming for, why do we implement these strategies, what is the nature and identity of different agents contributing to or opposing SUD strategies, etc.These capabilities are crucial, as the future only exists in the present as some form of anticipation (Miller, 2018).Therefore, to make sense of the many overlapping development trajectories, we should not only believe in a predetermined future and work for the desirable vision but also work tirelessly to assess the continuously evolving SUD-paths.Such capabilities that assist us in identifying conflicting views about the future and the actor groups behind them are needed, which may be crucial in mitigating conflicts and working towards collective action.Capabilities to identify individual and conflicting visions and to detect and create common dominators -to search for the 'third way' -may be a prerequisite for collective action.
CONCLUSIONS
In this article, we underline the importance of systematically pooling scattered capabilities into competence sets.We propose a tentative and still-porous conceptual framework to explore the existing competence sets, especially gaps in them, thus supporting a more integrated SUD work.
Overall, it is easy to conclude that without sufficient substantive expertise in sciences and engineering (know-why), we will never succeed in the green transformation.We also need specialised actors to lead and manage SUD processes, knowing how to work on a system and change it step by step.
In an ideal world, actors could compensate for each other's competence deficiencies with smooth collaboration.However, as the world is not perfect, there is a need to boost core competencies (learning processes) to integrate different capabilities into competence sets.Enhancing processual and institutional competencies may be crucial, as the NPG call for actors to produce results in an ever more complex environment and, consequently, the mismatch between the grand ambitions guiding SUD and the knowledge of delivering change is not diminishing.Therefore, it is essential to acknowledge that various actors may value needed competencies differently, causing difficulties and hindering SUD efforts.
SUD is governed across networks of public, private, community and academic actors, necessitating consideration of the variation or lack of it in the competencies needed among various cities.However, studies that systematically connect competence thinking to SUD are rare.We suggest that the competence set approach may prove promising in terms of sustainable transitions and as a tool for city leadership.In addition to a general need for more attention to urban competence sets and individual capabilities in the core of them, we propose three specific research topics.First, as indicated above, the needed competencies vary according to the functions and roles of the public, private and higher education sector actors.To date, we have insufficient knowledge about this variation.Second, we need to understand better competence requirements across different types of organisations and their roles in the system (Sotarauta et al., 2021).Third, while individual capabilities can be considered the micro-foundations for core competencies and competence sets at higher levels of aggregation, we have minimal knowledge about the occurrence of the aggregation.Places characterised by similarities in individual capabilities may substantially vary in competence sets, highlighting the importance of understanding differences in aggregation processes.
We still lack sufficient knowledge about how the aggregation of individual capabilities contributes to SUD across cities. Perhaps, the concept of the competence set has been the missing link.
DATA AVAILABILITY STATEMENT
As the paper is conceptual, we do not have data to share.
DISCLOSURE STATEMENT
No potential conflict of interest was reported by the author(s).
Figure 1 .
Figure 1.The competence set model in a nutshell.
FUNDING
This research was funded by the Nordic Green Growth Research and Innovation Programme in cooperation with NordForsk, Nordic Innovation and Nordic Energy Research [grant number 83130].
|
2024-06-19T15:02:07.448Z
|
2024-06-17T00:00:00.000
|
{
"year": 2024,
"sha1": "8a2621ed7fceb091cdb405df061ddb3c9c2109db",
"oa_license": "CCBY",
"oa_url": "https://www.tandfonline.com/doi/pdf/10.1080/21681376.2024.2359910?needAccess=true",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "67859817cfbe8f19c88142fa49e884f86a8e4233",
"s2fieldsofstudy": [
"Environmental Science",
"Sociology"
],
"extfieldsofstudy": []
}
|
239691207
|
pes2o/s2orc
|
v3-fos-license
|
Robust Single Molecule Magnet Monolayers on Graphene and Graphite with Magnetic Hysteresis up to 28 K
The chemical functionalization of fullerene single molecule magnet Tb2@C80(CH2Ph) enables the facile preparation of robust monolayers on graphene and highly oriented pyrolytic graphite from solution without impairing their magnetic properties. Monolayers of endohedral fullerene functionalized with pyrene exhibit magnetic bistability up to a temperature of 28 K. The use of pyrene terminated linker molecules opens the way to devise integration of spin carrying units encapsulated by fullerene cages on graphitic substrates, be it single‐molecule magnets or qubit candidates.
time of magnetic states, should be sustained on contact. On a more practical side, the magnetic units should possess a certain thermal and chemical stability to facilitate their processing. Concepts of several devices based on single-molecule magnets (SMMs) were already successfully realized, [1] but at this moment their function has only been demonstrated at sub-Kelvin temperatures. In the meantime, continuous development of lanthanide SMMs boosted their temperature range up to 80 K, [2] but the transfer of these properties from the bulk to nanoscale assemblies is lagging behind, and the magnetic bistability of SMM submonolayers is limited to 10 K at this moment. [3] Here, we describe a breakthrough in the field of surfacesupported SMMs through the use of chemically functionalized magnetic metallofullerenes. We show that the functionalization of Tb 2 @C 80 (CH 2 Ph) with a pyrene-terminated linker molecule facilitates the formation of self-assembled monolayers (SAMs) on substrates like graphene and highly oriented pyrolytic graphite (HOPG). Layers of these molecules exhibit magnetic hysteresis up to 28 K, showing no negative impact of the functionalization or deposition process on the magnetic properties. The exceptionally simple method of obtaining 2D-fullerene-graphene hybrid materials described here opens the way for device integration of spincarrying endohedral fullerenes, be it SMMs or qubit candidates.
Tb 2 @C 80 (CH 2 Ph) ({Tb 2 } hereafter) is a member of the lanthanide dimetallofullerene series with unconventional electronic structure and excellent SMM performance. [4] The endohedral metal dimer in {Tb 2 } features a single-electron covalent TbTb bond and can be formally described as a [Tb 3+ -e-Tb 3+ ] unit. This bond is responsible for the unique magnetic structure of the molecule. Two Tb ions have moderately large uniaxial magnetic anisotropy with a collinear arrangement of their magnetic moments. Each of the local Tb-4f moments is coupled ferromagnetically to the spin of the unpaired electron residing on the TbTb bonding molecular orbital with spd-hybrid character. As a result, {Tb 2 } has a large magnetic moment of 19 μ B and demonstrates SMM behavior with a high blocking temperature of 29 K surpassed only by some metallocene-based SMMs. [2,5] Besides, {Tb 2 } shows exceptionally broad magnetic hysteresis with a coercive field of 8 Tesla at 10 K. Despite the unconventional lanthanide oxidation state of +2.5, {Tb 2 } is an air-stable compound and does not require special handling conditions. This combination of outstanding SMM properties and
Introduction
The realization of single-spin devices for magnetic memory or quantum information processing relies on the possibility to create arrays of discrete magnetic units on suitable substrates, which can function as device electrodes. Useful properties of these units, such as the magnetic bistability and the long life-stability makes {Tb 2 } a prime candidate for the seamless integration into spintronic devices.
Fullerene Derivatization
The benzyl group on {Tb 2 }, which is needed for stabilization of the otherwise open-shell electronic structure, precludes the clean evaporation of the molecule onto surfaces. A straightforward way to overcome this limitation and facilitate the formation of SAMs is the functionalization of the fullerene with a linker moiety, terminated by a functional group tailored to the desired substrate. Here, we concentrated on the deposition of {Tb 2 } onto graphene and graphite (HOPG). Pyrene, known to show high affinity to graphitic substrates, was used to functionalize SMMs for their deposition on graphene and nanotubes in the past. [1b,6] To introduce pyrene groups to {Tb 2 }, we utilized the approach developed by Zaragoza-Galán et al. [7] for functionalization of C 60 with pyrene groups and then used the derivative for subsequent SAM growth. In short, 3,5-bis(4-(pyren-1-yl)butoxy) benzaldehyde was synthesized and then reacted with {Tb 2 } and N-methylglycine (sarcosine) in o-dichlorobenzene at 140 °C (Figure 1a). The crude product of this dipolar cycloaddition reaction contained the targeted fulleropyrrolidine monoadduct bearing two pyrene groups, {Tb 2 }-pyr 2 , along with unreacted educts and several other fulleropyrrolidine adducts. The chromatographic isolation and characterization (b) and toluene (c) (room temperature, peak force tapping mode at 150 pN). d) STM topography of {Tb 2 }-pyr 2 submonolayer on HOPG grown from toluene after 24 h degassing in UHV conditions (room temperature, bias 1.7 V, current 150 pA). e) Height profiles along the dashed lines in (c, AFM) and (d, STM) and the height profiles for a stripe of 12-molecules from molecular dynamics simulations (MD). f,g) Laying-down and standing-up configurations of {Tb 2 }-pyr 2 molecule on graphene from MD simulations, corresponding to mol 1 and mol 2 profiles in (e) at the time moment 120 ps.
Self-Assembled Monolayers
{Tb 2 }-pyr 2 is soluble in toluene and, to a lesser extent, dimethylformamide (DMF). Solutions of {Tb 2 }-pyr 2 in these solvents were used for self-assembly of the fullerene derivative on two types of graphitic substrates, freshly cleaved HOPG and chemical vapor deposition graphene on SiO 2 /Si. The binding of {Tb 2 }-pyr 2 to either substrate is very easily achieved by depositing a droplet of the solution for some time, before washing said droplet and excess fullerene molecules from the substrate with clean solvent. In the case of DMF, a drop of ≈3 mm diameter is stable for at least 30 min under a glass lid. Toluene is more volatile and tends to evaporate in under 10 min, leaving precipitated fullerene molecules on the substrate, which can be washed off with clean toluene. The whole procedure is described in full detail in the Supporting Information. Atomically flat graphene layers in HOPG substrate give a clear contrast with the fullerene islands and thus allow detailed analysis of the layer coverage and thickness. Thus, in the discussion below we focus on {Tb 2 }-pyr 2 SAMs grown on HOPG. Characterization of the films grown on graphene/SiO 2 /Si by atomic force microscopy (AFM) and Raman spectroscopy is provided in the Supporting Information.
Characterization of the resulting {Tb 2 }-pyr 2 films on HOPG by atomic force microscopy (AFM) reveals, that the deposition from DMF leads to the formation of a layer with uniform thickness in the whole area exposed to the solution (Figure 1b, see also Figure S16a, Supporting Information, for different magnification). The layer has orifices with the lateral size of 20-40 nm, and the height profile measurements along these orifices gives the layer thickness in the order of 1.8-2 nm (Figure S16b, Supporting Information), indicating that the layer has a singlemolecule thickness. Excluding the area of the orifices allows estimation of the surface coverage in DMF-deposited fullerene layer as 80-85%.
The deposition from toluene solution leads to the formation of smaller islands with gaps of clean substrate in between (Figure 2c and Figure S17a, Supporting Information). This is likely owed to the difference in solubility, favoring the binding to the substrate in case of DMF and the removal of some of the deposited {Tb 2 }-pyr 2 when rinsing with toluene. The coverage of the substrate with fullerene estimated from AFM topographies is about 35% (Figure S17b, Supporting Information). and during cooling down in magnetic field (FC); magnetic field 0.2 T, temperature sweep rate 5 K min −1 , the curves for two compounds are shown with a vertical off-set, vertical dotted lines mark blocking temperature of magnetization T B . c) Magnetic hysteresis curves of {Tb 2 }-pyr 2 measured between 1.8 and 28 K; magnetic field sweep rate 2.9 mT s −1 . d) Magnetization relaxation times of {Tb 2 }-pyr 2 determined at different temperatures; straight lines are contributions of the QTM and Orbach relaxation mechanisms, the orange curve is a total fit including QTM, Raman, and Orbach processes: . The inset magnifies the temperature range 21-29 K and shows the definition of 100-s blocking temperature T B100 .
www.afm-journal.de www.advancedsciencenews.com The characterization of {Tb 2 }-pyr 2 deposited on HOPG from a toluene solution by scanning tunneling microscopy (STM) yields similar topography to the AFM technique with the coverage near 30% (Figure 1d). The measurement was complicated by the tendency of deposited molecules to attach to the STM tip, and the image in Figure 1d was acquired by keeping the tip relatively far away from the sample. The possibility to perform both AFM and STM topographic characterization at room temperature indicates that the molecules are anchored to the substrate even if the layer is not closely packed.
The well-defined borders between fullerene islands and bare substrate in toluene-grown SAMs are instrumental for a more reliable determination of the film thickness than for the SAM deposited from DMF. The height profiles of {Tb 2 }-pyr 2 films extracted from Figures 1b (AFM) and 1c (STM) prove the formation of layers with uniform height (Figure 1d). However, noticeable distinction of height differences between unfunctionalized graphite and the fullerene islands were detected by AFM (2 nm) and STM (0.8 nm). Molecular dynamics (MD) simulations showed that a single {Tb 2 }-pyr 2 molecule on graphene would adopt a laying-down conformation with the apparent height of 1.1 nm, corresponding to the size of the fullerene cage ( Figure 1e). However, when surrounded by other molecules, either of solvent (Supporting Information) or of {Tb 2 }-pyr 2 neighbors, the fullerene core tends to elevate above the substrate, giving the height of 1.5-1.6 nm for a still rather sparse arrangement of molecules in the island (Figure 1f, see Supporting Information for more details on molecular dynamics simulations). Further increase of the molecular height may naturally occur in a densely packed layer. Similar conclusions were obtained in the study of functionalized C 60 SAMs on graphene. [8] We thus conclude that the results of AFM and STM measurements point to the single-molecule thickness of the {Tb 2 }-pyr 2 SAMs, and that the apparent height difference between the two techniques can likely be attributed to the different sample-tip interactions of the two methods. In particular, different local density of states in graphite and fullerene can strongly affect the tunneling current, giving different apparent layer heights depending on bias and current settings.
Magnetic Properties of {M 2 }-pyr 2
Facile preparation of SAMs indicates that they might be well suited for design of molecular spintronic devices based on integration with graphene electronics, unless magnetic properties of {Tb 2 } are negatively affected by derivatization and deposition. To evaluate the effect of derivatization, we characterized spin properties of {Y 2 }-pyr 2 and {Tb 2 }-pyr 2 by electron paramagnetic resonance (EPR) spectroscopy and SQUID magnetometry, respectively.
{Y 2 } exhibits a very characteristic hyperfine structure in its EPR spectrum caused by the interaction of the unpaired electron spin residing on the YY bonding orbital with nuclear spins of 89 Y (I Y = 1/2). The isotropic hyperfine coupling constant in {Y 2 } is a iso ( 89 Y) = 224 MHz and the g-factor is 1.9733. [4c] {Y 2 }-pyr 2 has a similar hyperfine structure in toluene solution (Figure 2a), but the rotational averaging appears incomplete even at 350 K, leading to a distorted intensity ratio (see Supporting Information for further analysis). The measurement in frozen solution at 150 K revealed a well-resolved powder-like pattern with slightly distorted axial g-tensor (1.9591, 1.9616, and 1.9997) and two inequivalent Y atoms with axial A( 89 Y)-tensors and principal A z /A x,y values of 246/219 and 258/207 MHz (see further discussion in Supporting Information). Two Y atoms in {Y 2 }-pyr 2 remain inequivalent at 350 K. Thus, EPR measurements prove that the cycloaddition leaves the single-electron MM bond intact and hinders rotation of the M 2 dimer inside a fullerene cage.
The effect of the cycloaddition on the SMM behavior of the endohedral [Tb 3+ -e-Tb 3+ ] unit was studied by SQUID magnetometry of a powder {Tb 2 }-pyr 2 sample. Blocking of magnetization, as determined from the comparison of a field-cooled and a zero-field-cooled temperature sweep (5 K min −1 ) is observed at 30.3 K (Figure 2b), which is 1.4 K higher than in the pristine {Tb 2 }. [4a] Another SMM figure of merit, the 100-s blocking temperature T B100 , is found to be 26.6 K in {Tb 2 }-pyr 2 versus 25.2 K in {Tb 2 }. When measured with a moderate sweep rate of 2.9 mT s −1 , magnetic hysteresis in {Tb 2 }-pyr 2 is still open at 28 K (Figure 2c). The hysteresis of {Tb 2 }-pyr 2 is narrower than in {Tb 2 }, but still exhibits a very large coercive field of 5.0-5.4 T below 20 K.
To elucidate the relaxation mechanism, magnetization relaxation times were measured at different temperatures between 2 and 28 K by magnetizing the sample at 7 T, quickly removing the external field and measuring the magnetization decay of the sample. The decay curves and their fitting with a stretched exponential function are presented in Supporting Information. Figure 2d plots the obtained relaxation times as a function of reciprocal temperature. In zero magnetic field, {Tb 2 }-pyr 2 shows two distinct relaxation regimes. Below 10 K, the relaxation time of 17.1(3) h is temperature-independent, which is a characteristic sign of the quantum tunneling of magnetization (QTM). Above 22 K, the relaxation is dominated by the Orbach mechanism as evidenced by the Arrhenius behavior, , with the effective barrier U eff of 725(6) K and τ 0 of 1.6(4) 10 −10 s. For comparison, the pristine {Tb 2 } exhibits the QTM mechanism with τ QTM of 18.0(3) h below 15 K and thermally activated relaxation with the barrier of 799(2) K above 20 K. [4a] Cycloaddition thus reduced the U eff barrier by 10%. The thermal barrier U eff corresponds to the exchange excitation, when one Tb magnetic moment attains opposite orientation versus the magnetic moment of the other Tb ion, which remains ferromagnetically coupled to the unpaired electron spin. The somewhat smaller U eff value obtained for {Tb 2 }-pyr 2 when compared to {Tb 2 } shows that the exchange interactions within the Tb 2 dimer are affected when the fullerene π-system is changed in the course of exohedral derivatization, but the effect is not very pronounced and the overall properties of the magnetic core remain similar. Evidently, the chemical functionalization even improves some SMM parameters, as it was found in another fullerene-SMM, DySc 2 N@C 80 . [9]
Surface Magnetism of {Tb 2 }-pyr 2 Monolayers
Surface magnetism of {Tb 2 }-pyr 2 monolayers on HOPG and graphene was studied by X-ray magnetic circular dichroism www.afm-journal.de www.advancedsciencenews.com (XMCD) with synchrotron radiation at the Swiss Light Source, Paul Scherrer Institute (X-Treme beamline [10] ) and ALBA (Boreas beamline [11] ). In agreement with the AFM and STM results discussed above, the ratio of X-ray absorption intensity at the Tb-M 4,5 edge for the {Tb 2 }-pyr 2 SAMs deposited from DMF and toluene is close to 2:1 ( Figure S29, Supporting Information), proving that the SAMs deposited from DMF do not exceed a monolayer coverage. Since SAMs deposited from and rinsed with DMF showed higher and more uniform coverage of the substrate (Figure 1b,c), we discuss XMCD measurements performed for such monolayers here. Results obtained for toluene-deposited SAMs are, however, similar (see Supporting Information).
X-ray absorption spectra of a {Tb 2 }-pyr 2 monolayer at the Tb-M 4,5 edge measured with clockwise and counter-clockwise polarizations of incoming X-ray beam develop a dichroism in the presence of a net magnetization along the beam direction (Figure 3a). A magnetic field collinear with the X-ray beam was employed in the measurements, and therefore the size of the XMCD signal is proportional to the magnetization in the direction of the X-ray beam. The angular dependence of the XMCD was investigated by varying the incidence angle of the magnetic field and X-ray beam with respect to the sample surface. The lack of a strong angular dependence as seen in Figure 3b shows that Tb 2 dimers attain different orientations in the {Tb 2 }-pyr 2 monolayer on HOPG with a slight preference for the 70° orientation versus the substrate.
The maximum of the XMCD signal at 1235 eV was used to monitor the changes of the SAM magnetization during magnetic field sweeps. At 2 K, {Tb 2 }-pyr 2 monolayers exhibit open magnetic hysteresis on both HOPG and graphene substrates (Figure 3c,e). Measurements with different X-ray fluxes revealed that the increase of the X-ray intensity leads to a narrowing of the hysteresis ( Figure S32, Supporting Information), which indicates that the hysteresis width is limited by the demagnetization effect of the X-rays. [3f,12] Thus, the intrinsic hysteresis of the {Tb 2 }-pyr 2 monolayer is probably broader, in line with the bulk magnetization measurement (Figure 2c). Interestingly, on graphene the hysteresis opening is narrower than on HOPG, which may indicate that the electronic properties of the substrate influence the relaxation of magnetization, [6d] although the charging of the silicon substrate may have a similar effect and thus cannot be excluded either. Spectacularly, the hysteresis of {Tb 2 }-pyr 2 monolayers remains open up to 28 K and is closed at 30 K (Figure 3d,f and Figure S31, Supporting Information, for intermediate temperatures and 30 K). Thus, {Tb 2 }-pyr 2 retains slow relaxation of magnetization in monolayer form on conducting graphitic substrates and exhibits measurable magnetic bistability in the same temperature range as in the bulk powder form. Figure 3. a) Circularly polarized X-ray absorption spectra of {Tb 2 }-pyr 2 monolayer on HOPG: I + and I − are clockwise and counter-clockwise circular polarizations, their sum gives non-polarized spectrum, whereas their difference is XMCD; T = 30 K, μ 0 H = 6 T, normal X-ray incidence. b) Angular dependence of the XMCD on HOPG, T = 2 K, θ is the angle between X-ray beam and the surface (90° corresponds to the normal incidence). c,d) Magnetic hysteresis curves of {Tb 2 }-pyr 2 monolayer on HOPG measured by XMCD at 2 and 28 K. e,f) Magnetic hysteresis curves of {Tb 2 }-pyr 2 monolayer on graphene measured by XMCD at 2 and 28 K (see Figure S30, Supporting Information, for measurements at other temperatures). Normal X-ray and magnetic field incidence, sweep rate 1.5 T min −1 .
Conclusions
Over a decade after the first observation of magnetic hysteresis in an Fe 4 monolayer, [13] the temperature range in which this phenomenon could be observed for SMM monolayers expanded from sub-K regime to 10 K. [3b,c,f ] With this robust fullerene-based SMM, we make a major step forward and push the temperature range of the phenomenon to 28 K. This is comparable to Ho atoms on a thin layer of MgO|Ag(100), which demonstrate broad magnetic hysteresis up to 30 K [14] and magnetic bistability up to 45 K. [15] But atomic submonolayers are stable only in UHV condition and require cryogenic temperature deposition to avoid cluster formation and to maintain a stable adsorption site on the surface since Ho atoms start to move on the surface already at 50 K. The solution-based preparation procedure of the {Tb 2 }-pyr 2 SAMs on graphitic substrates is exceptionally simple and fast, consumes tiny amounts of material, can be carried out in air, and does not require UHV conditions or special temperature regimes. The method should lend itself well to larger scale manufacturing techniques, like ink-jet printing and the use of nanospotters, and is compatible with graphene electronics and spintronics, [16] opening the way toward patterned devices with hybrid graphene-SMM interfaces. Furthermore, while {Tb 2 } is an excellent SMM, its non-4f analogs {Y 2 } and especially {Sc 2 } have recently been shown to be promising qubit candidates, [17] and the functionalization and deposition methods presented here might be a way of addressing them on a single molecule level. Finally, the linker molecule presented in this work is but the first option tested, and other approaches to hybrids of fullerenes with graphene and 2D materials can be applied to create assemblies with required properties. [18] For instance, increasing the number of surface binding pyrene groups in the linker molecule has been shown to enhance the stability of the resulting SAMs, [19] and examples with fullerenes can already be found in literature. [8]
Supporting Information
Supporting Information is available from the Wiley Online Library or from the author. www.advancedsciencenews.com
|
2021-10-22T15:52:53.038Z
|
2021-09-01T00:00:00.000
|
{
"year": 2021,
"sha1": "f9a459873fa9c0ee9132f2dde6fecf65220c98c6",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.1002/adfm.202105516",
"oa_status": "HYBRID",
"pdf_src": "Wiley",
"pdf_hash": "892023d7c5c90247d002e342f060d6a7c8dd61f8",
"s2fieldsofstudy": [
"Chemistry",
"Physics"
],
"extfieldsofstudy": [
"Materials Science"
]
}
|
234075322
|
pes2o/s2orc
|
v3-fos-license
|
Visual level measurement system application in the technological process automated control system of the recuperative furnace bath
The paper considers the melted material level determining problem in the recuperative furnace bath melter in technological process automated control system for the basalt mineral fiber production. To solve this problem, it is proposed to use a visual level measurement system based on the analysis and processing of the melt boundary image. A block diagram of the visual level measurement system developed by the authors is presented, and an algorithm for determining the melt level using a modified Sobel edge detection method is proposed.
1.Introduction
Currently, in the construction and industry heat and sound insulation materials based on mineral fiber (mineral wool) are widely used. The popularity of these materials is due to their high thermal-insulation properties, the ability to work at high temperatures (up to 700°C) and relatively low cost. Blast furnace slags, natural rocks, as well as industrial waste (brickbat, keramzite and cement production emitted dust) can serve as raw materials for the mineral fiber production. At the same time, mineral wool obtained from basalt group rocks has the highest quality. The mineral fiber production process includes raw material mixture (furnace charge) preparation operations, its melting, obtaining fiber from the melt, mineral wool subsidence, mineral wool carpet formation with subsequent processing, cutting into products of specified sizes and packaging.
Problem statement
For melting the raw material mixture in modern production bath regenerative and recuperative furnaces are mainly used. Recuperative bath furnaces contain a recuperator in which the heat from the hot exhaust gases is used to heat the air coming to the burner. Due to this recuperative furnaces are more efficient than regenerative furnaces. The bath furnace consists of two parts: cooking and working (feeder). The preprepared charge is fed to the cooking pool using loaders through special loading windows in the furnace side walls. There are burner devices in the cooking part that provide the furnace charge melting, after which the resulting melt enters the feeder. There is a threshold between the cooking part and the feeder preventing unmelted raw material from entering the feeder. The melt from the feeder is issued for the subsequent mineral fiber forming operation. One of the most important technological process parameters during the bath recuperative furnace operation is the melt level in the cooking part, which, in accordance with the regulations, must be maintained with an accuracy of ±20 mm. The melt level is maintained by the automatic control system by changing the loaders mode time. The melt level direct control is a complex problem, since the known contact and non-contact methods for its determination [1,2] are hardly suitable due to the constant presence of foam on the melt surface ( Fig. 1a) and clusters of unmelted raw materials (Fig. 1b). For this reason, the melt level automatic control is usually based on indirect data representing the difference between the amount of raw material loaded and the melt output. It is also possible to work with the periodic actuation of loaders on time without any feedback. These two control methods have a significant drawback: the accurate data lack on the melt level necessarily leads to errors in the automatic loader control system operation, which leads to either an overflow of the furnace with melt, or a decrease in the melt level below the minimum allowable value. Therefore, the melt level direct continuous measurement problem in the furnace cooking part is relevant. a) b) Figure 1. The internal cavity of the recuperative furnace
Theory
Based on the analysis of the recuperative melting furnace loading nodes studies results [3], the authors proposed a method for measuring the melt level drawing on visual observation. The method is based on the fact that the furnace wall and melt light intensity almost always have different values. The successful using of video camera for temperature measuring [4] and mineral wool quality control [5] is known. The melt height is determined by the difference in the furnace wall light intensity and the melt light intensity. In the course of research, it was found that a grayscale black-and-white image (256 grayscale) is more informative than a color image. Also, during the experiments it was found that hot air and gas flows inside the furnace make it difficult to determine the melt level in addition to foam and unmelted raw materials. They cause image instability, resulting in geometric distortions and blurring. In the course of performing field tests it was found that to improve the melt level determining quality pre-processing of the obtained video image is required. Image pre-processing consists of several stages. At the first stage, the effect of geometric distortions and image blurring is reduced. To do this, the image is averaged frame-by-frame over time. The averaging time is selected depending on the process and can be set from 1 second to 5 minutes. At the second stage, the image contrast increases. Increasing the contrast made it possible to use the image brightness scale more fully. Obtained video images histograms analysis showed a low contrast of the recuperative furnace internal cavity image. Figure 2 shows a histogram of the recuperative furnace internal cavity image. As one can see from the presented histogram, the most informative brightness image values are in the range from 80 to 160, so it was decided to use the brightness range from 80 to 160, which was stretched to the full scale (from 0 to 255). This made it possible to display the melt boundary more clearly on the screen. The third stage is to increase the image transitions sharpness, which makes it even more possible highlight the melt level against the furnace wall background. Since the melt level is always parallel to the horizon, it is possible to use the horizontal transitions sharpness, the so-called edge enhancement, to display it more clearly. In this case, the edge is a sharp change in the image brightness. The edge brightness is proportional to the change in brightness surrounding the edge in the image. There are several methods of edge enhancement: convolution, Laplace, shift and difference, and others [6,7]. By analyzing video images obtained during field tests, it was found that the best result is obtained when the edge is strengthened by the convolution method. In this case, the convolution core has the form shown in figure 3: The only area with the melt level is preprocessed instead of entire image to minimize the image preprocessing time. The melt level is determined using a modified Sobel edge detection method. The essence of this method is as follows [6]. Image area selected in size of 3x3 pixel is indicated in Figure 5: Four lines can be drawn through the center point "e" of this area: Line 1: a-e-i, Line 2: b-e-h, Line 3: c-eg, Line 4: d-e-f. Each of these lines divides the space adjacent to the point "e" into two three-point areas, and the absolute value of the difference between these two subdomains average values is calculated. The image element under consideration is assigned the largest value of the four absolute differences. This new element value is then compared to a specific threshold value. A modification of the Sobel algorithm is that the melt boundary is always horizontal, so it makes sense to determine the boundary only along the d-e-f line, and not consider the other lines. This allows to speed up the melt level calculation. In addition, it was proposed to use a 10x10 matrix to determine the melt boundary more correctly. Accordingly for each point located in the melt level vicinity a new value is calculated using the formula: where I x,y+k is a brightness over the pixel x,y; I x,y-k is a brightness under the pixel x,y; k is a whole number. The obtained value of the point I x,y is compared with a threshold value that is set in the configuration program, depending on furnace and technological process. If the point I x,y new value is greater than the threshold, then the corresponding element of the one-dimensional array H is increased, which stores information about how many points I x,y of the corresponding row has a value greater than the threshold. After processing all points in the melt level vicinity, the largest element is searched in the array H, which determines the melt boundary position. The melt level is calculated based on the array H maximum element index.
Experimental results
The block diagram of the visual level measurement system developed by authors is shown in Fig. 6. In addition, this system also allows visually controlling the processes occurring in the furnace internal cavity. Two specially designed high-resolution IP video cameras are used to record the melt level in the system. The presence of two video cameras allows to monitor the melt level in the presence of foam and unmelted raw materials on its surface more accurately. Video streams from IP cameras are sent via an Ethernet commutator to an industrial computer with specialized software developed by the authors. The software processes the received data, calculates the melt level and transmits this information to the automatic control system of the furnace charge loaders in the furnace via the RS485 interface. Figure 6. The visual level measurement system block diagram The system software consists of three modules: • the administration module; • the melt height calculation module; • module for working with archived data.
The administration module allows to enter new or delete existing system users and change their access level. When starting the administration module, it is necessary to enter the user name and password in the window that appears. If the user has "Administrator" access, the main module window is loaded. In this window there is the possibility to add a new user, delete a user, or change the user's access level. In total, the system has four access levels: • Guest; • Operator; • Engineer; • Administrator; The "Guest" access level is set by default when the program starts and only allows viewing information about the melt height. When the software is idle, it also automatically switches to this access level from any other level. The "Operator" access level allows not only viewing information about the melt height, but also correcting the melt height in case of a failure in the melt level determining. To do this it is necessary to select the "Melt level" menu item and specify the current level on the image. After that, the program begins to determine the melt level based on the newly received information. The "Engineer" access level allows to configure the system to determine the melt level in addition to the above actions. The "Administrator" access level allows to perform all the listed actions, as well as enter new or delete existing users of the system and change their access level. Figure 7 shows the main window of the administration module. The melt height calculation module is the main software module. This module performs the following functions: • calculating the melt height; • basic software settings; • data archivation. The main window of the melt height calculation module, software for the visual level measurement system, is shown in figure 8. The system configuration is made in the window "Setting", the input to which is carried out via the menu item "Setting". The "Settings" menu is active only if you log in to the program with the "Engineer" or "Administrator" access. The window "Setting" is shown in figure 9. The "Melt level" tab sets various fixed points or areas on the visualized image of the furnace internal cavity. They allow to determine the melt level more accurately. The "Configuration" tab allows to perform general system configuration. For example, it is possible to specify the path to save data, or select the port to connect to the loader management system. The "Add. Settings" tab sets parameters for image preprocessing and melt level determining. These parameters include the contrast threshold, minimum brightness, maximum brightness and filter order. Figure 9. The "Setting" window The module for working with archived data displays the trend of changes in the melt level and the furnace internal cavity image the period under present observation. Archived data can be accessed over the network. Figure 10 shows the archive module window.
Conclusions
The visual level measurement system developed by the authors makes it possible to determine the melt level with an accuracy of ±2 mm. The level measurement system has been tested and implemented as part of the automatic control system for the recuperative furnace bath loaders at TECHNONICOL Corporation (Chelyabinsk) -one of the enterprises producing mineral wool products.
|
2021-05-10T00:03:42.573Z
|
2021-02-01T00:00:00.000
|
{
"year": 2021,
"sha1": "6706eebdd424df7cb82d370aea7b84a58a06aa6d",
"oa_license": null,
"oa_url": "https://doi.org/10.1088/1742-6596/1791/1/012111",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "05b84b8e0c8fbe0a60f25fec4df2f7096f2e6eb1",
"s2fieldsofstudy": [
"Engineering",
"Physics"
],
"extfieldsofstudy": [
"Physics",
"Computer Science"
]
}
|
245134251
|
pes2o/s2orc
|
v3-fos-license
|
Association of pigment epithelium derived factor expression with cancer progression and prognosis: a meta-analysis study
Background Pigment epithelium derived factor (PEDF) is a secreted protein that strongly suppresses angiogenesis and directly inhibits cancer cells proliferation. The differential expression of PEDF has been observed in multiple types of human tumors. However, it is unclear as to how PEDF expression is associated with cancer progression and if PEDF could serve as a prognostic marker for cancer patients. Methods We performed a comprehensive search for the studies on PEDF expression in 14 top-ranked types of solid tumor cancer with the highest incidence. A systemic approach was used to screen for qualified studies and to extract data. Meta-analysis was performed to investigate if PEDF expression is associated with the TNM staging, tumor size, lymph node invasion, distal metastasis and pathological grade of tumor in a pan-cancer manner. A Kaplan–Meier curve was plotted with the digitally-reconstituted patient survival data to study the effect of PEDF expression on the prognosis of cancer patients. Results A total of nine studies were selected, reviewed and analyzed. Meta-analysis suggested that decreased PEDF protein expression was associated with higher TNM staging (OR = 2.13, 95% CI: 1.61–2.81), larger tumor size (OR = 1.42, 95% CI: 1.1–1.84), larger possibility of lymph node invasion (OR = 1.68, 95% CI: 1.26–2.22) and higher pathological grade (OR = 1.6, 95% CI: 1.2–2.13). No correlation was found between PEDF expression and tumor distal metastasis, gender or age. In addition, low PEDF protein level in tumor tissue is correlated with shorter overall survival (P < 0.05). Conclusions Low PEDF protein expression in cancer is significantly associated with more advanced cancer progression and significantly poorer survival. The differential clinical outcome among patients with various PEDF expression suggests its prognostic value. Supplementary Information The online version contains supplementary material available at 10.1007/s12672-021-00457-y.
Introduction
Pigment epithelial-derived factor (PEDF) was first identified as a protein factor secreted by human retinal pigment epithelium cells [1]. It is also called serine protease inhibitor F1 (SERPINF1) and belongs to the serine protease inhibitor (serpin) superfamily. However, it has no serine protease inhibitor activity [2]. Numerous previous studies suggested that PEDF is a protein with multifaceted anti-tumor activities. Its most recognized function is the ability to inhibit angiogenesis, a process that drives tumor growth and metastasis [3,4]. In in vitro assays, PEDF's anti-angiogenic activity was found to be more potent than other endogenous anti-angiogenic factors such as endostatin, angiostatin and thrombospondin-1 [5]. In a variety of in vivo models, PEDF reduced microvascular density in tumor tissues which is one of the mechanisms leading to tumor inhibition [6][7][8][9][10][11]. PEDF's receptor that mediates its antiangiogenic activity has been identified as a cellsurface transmembrane protein Plexin domain-containing protein 1 (PLXDC1) which is also known as tumor endothelial marker 7 (TEM7) [12]. PLXDC1 and its homologue PLXDC2 are the only two proteins that have been demonstrated to confer cell surface binding to extracellular PEDF and to transduce PEDF signal into the cell. PLXDC1/TEM7 was highly enriched in the blood vessels of tumor tissues but not in the blood vessels of adjacent normal tissue. PLXDC1's tumor blood vessel expression was found in a wide range of cancer types, including liver cancer, breast cancer, ovarian cancer, pancreatic cancer, colorectal cancer, lung cancer, neuroblastoma and sarcomas [13][14][15][16][17]. This high specificity of PEDF receptor's expression in tumor blood vessels matches the specificity of PEDF's tumor inhibitory effect. Other than inhibiting angiogenesis, PEDF also directly inhibits certain tumor cells and promotes cell differentiation [18]. Previous studies showed that PEDF can suppress the growth of lung cancer cells [19] and promote apoptotic cell death in melanoma cells [20,21]. In addition, PEDF was found to induce the differentiation of tumor cells, such as eliciting neuron-like morphology in neuroblastoma cells [22] and promoting neuroendocrine function of prostate cancer cells [23]. Interestingly, another PEDF receptor, PLXDC2 was found to be expressed in a variety of cancers, including colon cancer, hepatocellular carcinoma, laryngeal cancer, testicular seminoma, and vulva squamous cell carcinoma [24][25][26][27][28]. PLXDC2-mediated signaling could be responsible for the direct effect of PEDF on cancer cells. The collective anti-cancer activity (anti-angiogenesis and direct anti-tumor cell effect) of PEDF was observed in in vivo studies. PEDF treatment by gene therapy or administration of recombinant protein was shown to inhibit the growth of pancreatic cancer [8], 1 3 hepatoblastoma [6], prostate cancer [7], retinoblastoma [10], ocular melanoma [11], lung cancer and colon cancer [9]. It also significantly reduced thoracic metastasis of colon cancer [9].
PEDF is widely expressed in most human organs and tissues such as eye, liver, heart, brain, bone and lung [29]. A significant decrease of PEDF level was found in age-related macular degeneration (AMD) and diabetic retinopathy (DR), two pathological processes dependent on angiogenesis. PEDF expression in cancer was also studied in a wide range of cancer types. Although the association between PEDF expression and tumor development has been reported by many studies, each result only applies to individual cancer and there is no broad view on PEDF's effect on diverse types of cancer. Given the multifaceted anti-tumor effects of PEDF and the presence of PEDF signaling machinery in various types of cancer, it would be interesting to know if PEDF protein expression level is associated with cancer progression and the prognosis of patients in general. To answer this question, we use systemic review and meta-analysis to inclusively identify research on this topic, unbiasedly select qualified studies and reach a weighted average with pooled data from the selected studies. Specifically, we reviewed the studies on PEDF expression in 14 types of top-ranked solid tumor malignancies with the highest incidence worldwide. We investigated if PEDF protein expression in cancer tissue is associated with the clinical development (TNM staging) and pathological feature (cancer grade) of tumors in a pan-cancer pattern by meta-analysis. We also explored the prognostic value of PEDF protein and discussed its biological importance and other potential values in cancer management.
Search strategy
PubMed and EMBASE were searched for studies on PEDF expression in cancer through December 2020. The search included the 15 most prevalent cancer types worldwide in 2020 (the 13th-ranked cancer is leukemia and thus 14 types of solid tumor cancers were included in the study) [30]. No language restrictions were imposed. The search strategy was built using the key words for PEDF and the key words for each type of cancer. The 14 malignancies and the search strategies are summarized in Table S1.
Study selection
Studies were included for meta-analysis if they meet the following criteria: (1) had a case-control design; (2) reported PEDF expression levels in human samples; (3) presented categorical data as case number. The search results and the literature were reviewed by two authors (G.C. and C.S.) independently.
Data extraction and quality assessment
Data extraction form was designed to record data from selected studies by two reviewers (G.C. and C.S.) independently. The following information was extracted: the first author's last name, year of publication, study design, age, gender, total number of cases, PEDF detection method, number of cases categorized based on age, gender, tumor TNM stage, tumor size, lymph-node invasion, distal metastasis, and tumor histopathological grade. Quality assessment was conducted using Newcastle-Ottawa Scale (NOS) and studies with an NOS score ≥ 7 were considered high quality [31].
Reconstructing individual patient data of survival
Individual patient data (IPD) of survival were reconstructed from the published Kaplan-Meier curve (KMC) by the following steps: (1) KMC of overall survival (OS) was copied from the original study by Microsoft Windows "snipping" tool.
(2) Published KMC was uploaded to the online data extraction tool "WebPlotDigitizer" [32] to extract survival probability (S ti ) and time point (t i ) from the curve. Data were extracted at the time point of every 10 months. (3) Due to the missing number at risk (N ti ) in the published studies, we did approximate calculation of N ti with the formula "N ti = (S ti-10 / S ti ) × N ti-10 -C ti ". The number of censored patients (C ti ) were simplified as zero as five out of seven studies did not reveal this number. (4) S ti , t i and N ti were input to the algorithm in R developed by Guyot et al. [33] for the generation of IPD. IPD from all 7 studies were pooled for pan-cancer survival analysis.
Retrieving individual patient data of survival from The Cancer Genome Atlas (TCGA)
We used an online platform, Oncolnc, which incorporates TCGA database and provides users with the access to IPD on gene expression (RNA level) and survivorship for 21 common malignancies [34]. IPD were sorted into two groups on Oncolnc: low gene expression (lower 50 percentile mRNA level) and high gene expression (higher 50 percentile mRNA level). IPD of 21 malignancies which composed of PEDF mRNA level and survival information of 7970 cases were downloaded and pooled for Kaplan-Meier curve plotting.
Statistical analysis
Meta-analysis was performed if PEDF expression and clinical feature were reported in more than two studies. Pooled odds ratio (OR) and 95% confidence intervals were calculated to estimate the strength and credibility of the association. Either fixed-effect (I 2 ≤ 50%, P ≥ 0.05) or random-effect (I 2 ≥ 50%, P < 0.05) model was chosen based on the heterogeneity test [35]. The I 2 test was used to assess the heterogeneity among studies. The I 2 value was explained as of no (0-25%), low (25-50%), moderate (50-75%) or high heterogeneity (75-100%) [36]. Publication bias was evaluated with funnel plots. If publication bias was detected, it would be corrected by trim-and-fill method. Sensitivity test was performed by omitting one study at a time to assess the robustness of the pooled results (leaving-one-out method). Statistical analysis were conducted using the software Rstudio (version 1.3.1093). Package "meta" and "ggplot" were used for data analysis and graph generation.
Kaplan-Meier curve was created with "survfit" and "ggsurvplot" in Rstudio (version 1.3.1093). Pooled, reconstructed IPD were imported and the curve was plotted with a 10-month interval for 100 months. IPD downloaded from Oncolnc were pooled and plotted into Kaplan-Meier curve with a 20-month interval for 360 months.
Systematic review and summary of selected studies
The search of PEDF expression in cancer generated a total of 770 articles. After removing those that are duplicated or not published in English, 391 literature remained. We further excluded 107 articles with irrelevant topics, 90 studies using in vitro or animal models, 63 conference reports and 84 review articles. Forty-seven articles are left that reported PEDF expression in human samples acquired clinically. Among them, 9 studies met the criteria described above for meta-analysis. The search and screening processes are shown in Fig. 1. As summarized in Table 1, the 9 studies examined PEDF protein expression in cancer tissues by immunostaining semi-quantitatively. With a case-control design, they compared the clinical features, such as age, gender, tumor TNM stage and pathological grade, between PEDF high-expression group and low-PEDF expression group. Although we reviewed literatures on 14 types of primary solid tumor malignancies, the selected studies reported PEDF protein expression in 8 cancer types. The quality of the selected studies was evaluated by NOS manual as shown in Table S2. All the studies were identified as low risk for bias (NOS score ≥ 7).
Association of PEDF protein expression and TNM staging
The first analysis we performed is to examine if PEDF protein expression is associated with cancer progression clinically manifested by TNM staging. TNM stage II and below were classified as "low stage" while stage III and above were classified as "high stage". The pooled effect from those studies were shown in Fig. 2a. Low PEDF-expressing tumors had increased odds to be high TNM stage compared to high PEDF-expressing tumors (OR = 2.30, 95% CI: 1.70-3.12, fixed-effect model, I 2 = 0%, P = 0.56). Publication bias was found and 2 artificial studies were added by the "meta" package to correct the pooled OR (Fig. 2b). The adjusted odds ratio remained significant (OR = 2.13, 95% CI: 1.61-2.81, I 2 = 0%, P = 0.53). Sensitivity analysis (Fig. 2c) suggested that none of the individual study significantly affected the pooled effect.
Association of PEDF protein expression and tumor size
We further investigated if PEDF protein expression is correlated with tumor size, lymph node invasion, and distal metastasis, each of which could contribute to tumor progression. Although different studies adopted different cut-off values to classify "small-" and "large-" tumors as shown in Table 1, each threshold is consistent with the feature of a specific type of cancer and is conventionally used to classify tumors of that type based on size. Therefore, we did not use any absolute value to define tumor size; instead, we directly extracted and pooled data based on the classification in each original study. Meta-analysis showed that low PEDF expression is associated with larger tumor size compared to high PEDF expression, presenting an increased odds (OR = 1.42, 95% CI: 1.1-1.84, fixed-effect model, I 2 = 0%, P = 0.55) (Fig. 3a). No publication bias was detected (Fig. 3b). In the sensitivity test, omitting any studies did not change the conclusion (Fig. 3c).
Association of PEDF protein expression with local or distal invasion
Next we examined if PEDF protein expression is also associated with lymph node invasion. Low PEDF-expressing tumors showed an increased odds to metastasize to lymph node compared to high PEDF-expressing tumors (OR = 1.68, 95% CI: 1.26-2.22, fixed-effect model, I 2 = 39%, P = 0.13) (Fig. 4a). Publication bias was detected and one artificial study was added (Fig. 4b). After correcting the bias, the OR remained significant (OR = 1.7, 95% CI:1.28-2.27; I 2 = 38%, P = 0.12). Omitting any study in the sensitivity test did not change the result (Fig. 4c). Only two studies reported the observation on PEDF expression and distal metastasis. Meta-analysis on these studies did not find a significant association between PEDF expression and distal metastasis (OR = 1.42, 95% CI: 0.75-2.67; I 2 = 67%, P = 0.08) (Fig. S1).
Association of PEDF protein expression and histological grade of cancer
Other than TNM stating, we are also interested in the association of PEDF protein expression with tumor histological grades. While reviewing the original studies, we noticed that tumors had been categorized into "low-grade" and pT primary tumor, LN lymph node, OS overall survival "*" well-differentiated and moderately-differentiated groups from the original study were combined as "low grade" while poorly-differentiated group was categorized as "high grade" Investigators Zhou [40] Zhang [42] Li [43] Hou [37] Yi [38] Uehara [41] Jang [55] Lv [56] Jiang [39] Publication year "high-grade" groups based on the conventional cut-off criteria of each type of cancer (Table 1). Specifically, Hou et al. [37], Yi et al. [38], and Jiang et al. [39] grouped tumors with grade G1-2 to be "low grade" and tumors with grade G3-4 to be "high grade". Zhou et al. [40] and Uehara et al. [41] defined G1 tumors as "low grade" and G2-3 tumors as "high grade"; while Jang et al. [39] did not provide grading information for the "low grade" and "high grade" groups in their study. Zhang et al. [42] and Li et al. [43] stratified tumor grades into "well-", "moderately-" and "poorly-differentiated". According to the categorization by the original studies, we extracted and pooled the data into the "low-grade" and "high-grade" groups. We also combined well-differentiated and moderately-differentiated tumors in Zhang et al. [42] and Li et al. [43]'s studies into the "low-grade" group while pooled the poorly-differentiated cases into the "high-grade" group. Meta-analysis with the pooled data revealed that cancers with low PEDF expression have an increased odds to be high grade than those with high PEDF expression (OR = 1.6, 95% CI: 1.2-2.13, fixed-effect model, I 2 = 33%, P = 0.17) (Fig. 5a). Funnel plot (Fig. 5b) showed the publication bias and 3 studies were added to correct the publication bias. The adjusted meta-analysis still suggested an increased odds for low PEDF-expressing tumor to be high grade (OR = 1.94, 95% CI: 1.26-2.99, randomized model, I 2 = 55%, P = 0.01). Excluding any one of the studies did not change the pooled OR in sensitivity test (Fig. 5c).
Prognostic value of PEDF
Lastly, we investigated the correlation between PEDF protein expression level and the prognosis of cancer patients. As the conclusions from previous studies are not consistent (Table 1), we aim to pool the patients' data from the original studies and to analyze the effect of PEDF protein expression on the overall survival of cancer patients. We used a novel method to reconstruct individual patient data from the published Kaplan-Meier curve following the protocol developed by Guyot, et al. [33]. After combining the data of a total of 906 individual patients, we plotted a Kaplan-Meier curve to compare the overall survival of patients with high PEDF protein expression and that of patients with low PEDF protein expression. Patients with high-PEDF protein expression in the cancer tissue have a significant longer overall survival than those with low-PEDF protein expression (P = 0.00035) (Fig. 6).
We also studied if PEDF mRNA level is correlated with patients' survival on a pan-cancer basis with the data from The Cancer Genomic Atlas (TCGA), a cancer genomic program established by the National Cancer Institute (https:// www. cancer. gov/ tcga). We used Oncolnc, an online platform that incorporates TCGA database and provides gene expressionsurvival data [34]. A total of 7970 individual patient data of 21 malignancies were downloaded and pooled. The analysis of the correlation between PEDF mRNA and overall survival was performed. To our surprise, low PEDF expression at mRNA level is associated with longer survival time (with a marginal statistical significance P = 0.045, Fig. S2). The finding is the opposite of the observation on the PEDF protein-survival correlation (Fig. 6). The discrepancy could likely be explained by the fact that gene expression at the mRNA level and gene expression at the protein level are not consistent in many scenarios [44]. Many layers of regulation such as translation efficiency, post-translational modification and protein degradation have impact on protein concentration independent of the change of mRNA quantity [44,45]. In addition, mRNA level is much more susceptible to fluctuation than protein upon the transition of cell status. It was found that the global mRNA number in a cell shrinks substantially when the cell switches from a proliferating state to quiescence while the global protein number only drops by ~ 9.5% [46]. Therefore, we make the following hypothesis to explain the contradictory data: the aggressive tumors associated with shorter overall survival likely have more proliferating cells and thus higher global mRNA level including PEDF mRNA; meanwhile, PEDF protein expression in those tumors is largely suppressed due to the translation and/or post-translation regulation. On the other hand, tumors associated with longer survival have relatively more cells in quiescence and significantly lower global mRNA including PEDF mRNA; however, PEDF protein translation is very efficient in those tumors which results in higher PEDF protein expression.
Association of PEDF protein expression and age or gender
According to the NOS manual, it is not known if patients with different gender or age are evenly distributed between high PEDF-expression group and low PEDF-expression group. Thus, gender or age could be confounding factors in the analysis of the association between PEDF expression and cancer progression or overall survival. For example, if the average age in low-PEDF expression group is higher, this group would likely show a shorter overall survival compared to high-PEDF expression group. To assess this possibility, we performed meta-analysis on the association of PEDF expression
Discussion
This is the first systemic review and meta-analysis of PEDF expression in major types of human cancer since the identification of PEDF three decades ago. This study shows that the decreased PEDF expression at protein level is associated with more adverse clinical outcomes in cancer patients, manifested by higher TNM staging, higher tumor grading, and a significantly shorter overall survival. Using meta-analysis, this study provides a quantified strength of the association between PEDF and the clinical features of malignancy in a pan-cancer manner. Thus this study reveals the value of PEDF as a general biomarker to assess cancer progression and prognosis in cancer management. In addition, the result of this study is consistent with the findings in preclinical models that PEDF has multifaceted anti-tumor activities and suggests the broad anti-tumor effect of PEDF in a wide range of human malignancies. Fig. 2 Cancer with low PEDF expression has higher odds to have higher TNM staging. a Forest plot showing the association of PEDF expression and cancer TNM staging. Squares indicate study-specific odds ratios (ORs). The size of the box is proportional to the weight of the study. Horizontal lines indicate 95% confidence intervals (CI). A diamond indicates the summary OR with its corresponding 95% CI. b Funnel plot on the publication bias of studies on PEDF expression and TNM staging. Each filled circle represents one study. Open circles represent added artificial studies to correct publication bias. c Forest plot of sensitivity test. Squares indicate the summary OR after omitting one specific study. Horizontal lines indicate 95% CI. A diamond indicates the summary OR of all the studies with its corresponding 95% CI. IDC intra-ductal carcinoma, NSCLC non-small cell lung cancer, HCC hepatocellular carcinoma, CRC colorectal carcinoma, PDAC pancreatic ductal adenocarcinoma, PTC papillary thyroid carcinoma, RCC renal cell carcinoma Despite the anti-angiogenesis and anti-tumor cell activity of PEDF, we did not find the association between PEDF expression and distal metastasis. This could be due to the lack of detection power as only two studies were available for meta-analysis. Biologically, the local PEDF expression in the primary tumor may not influence the tumor implantation at a distal tissue/organ, where a tumor-favorable microenvironment is necessary for successful metastasis [47]. In fact, it would be interesting to study if systemic PEDF protein level, such as PEDF serum concentration, is correlated with local invasion of primary tumors and distal metastasis.
The relation of PEDF mRNA-survivorship does not agree with the finding that low PEDF expression at protein level is correlated with poorer overall survival. Although the central dogma of biology described the tight link from DNA to RNA and to protein, transcript level by themselves are not sufficient to predict protein levels and to explain genotype-phenotype relationships in many scenarios [44]. Previous research found that the levels of mRNAs and their corresponding proteins showed limited correlation (R 2 = 0.55 in cycling cells) globally [46,48] and only 40% of the variance of protein levels could be explained by mRNA levels [49]. Besides, protein level is more stable relative to mRNA level under unsteady state. As protein is the molecule that realizes the gene function and shows higher stability in quantity, the prognostic value of protein is probably higher than that of transcripts. To gain more knowledge on how PEDF expression correlates with clinical prognosis, more studies on PEDF at both transcriptomic and proteomic levels are needed.
Other than the predictive value of PEDF protein in caner progression and prognosis, measuring PEDF protein expression in cancer tissue might guide cancer treatment, especially the usage of anti-angiogenic therapy. The anti-angiogenic therapies approved by Food and Drug Administration (FDA) are primarily targeting the vascular endothelial growth Squares indicate study-specific odds ratios (ORs). The size of the box is proportional to the weight of the study. Horizontal lines indicate 95% confidence interval (CI). A diamond indicates the summary OR with its corresponding 95% CI. b Funnel plot showing no publication bias of studies on PEDF expression and tumor size. Each filled circle represents one study. c Forest plot of sensitivity test. Squares indicate the summary OR after omitting one specific study. Horizontal lines indicate 95% CI. A diamond indicates the summary OR of all the studies with its corresponding 95% CI. IDC intra-ductal carcinoma, NSCLC non-small cell lung cancer, HCC hepatocellular carcinoma, CRC colorectal carcinoma, PDAC pancreatic ductal adenocarcinoma, PTC papillary thyroid carcinoma, RCC renal cell carcinoma factor (VEGF) pathway, such as the monoclonal antibody against VEGF [50] or chemical inhibitors of tyrosine kinase receptors stimulated by VEGF [51,52]. However, the use of anti-angiogenic therapy is not individualized to minimize the therapy-related adverse events. Like endocrine therapy, which is indicated specifically by the expression of hormonal receptors in cancer tissue [53,54], it would be interesting to know if any pathological feature could specifically suggest the sensitivity of tumors to anti-angiogenic reagents. Given that PEDF is the strongest endogenous anti-angiogenic factor that counteracts multiple pro-angiogenic factors including VEGF, the down-regulation of PEDF in cancer tissue could suggests a tilted balance toward over-angiogenic. Therefore, it is worth researching if PEDF expression in cancer tissue could be a useful indication for anti-angiogenic therapy. In addition, the inverse correlation between PEDF expression and tumor progression in multiple human tumor types also suggests the potential clinical value of developing cancer treatment by targeting this pathway.
In conclusion, we reviewed studies on PEDF expression in 14 solid tumor malignancies with the highest incidence worldwide. Nine case-control studies on eight types of cancers were selected for meta-analysis. The pooled result suggested that low PEDF protein expression in cancer tissue is significantly associated with cancer progression. Low PEDF expression in cancer tissue is also correlated with a poorer survival. This study indicates a potential clinical value of PEDF as a prognostic marker and as an indicator for anti-angiogenesis therapy. Fig. 4 Cancer with low PEDF protein expression has higher odds to invade lymph nodes. a Forest plot showing the association of PEDF expression and lymph node invasion. Squares indicate study-specific odds ratios (ORs). The size of the box is proportional to the weight of the study. Horizontal lines indicate 95% confidence interval (CI). A diamond indicates the summary OR with its corresponding 95% CI. b Funnel plot on the publication bias of studies on PEDF expression and lymph node invasion. Each filled circle represents one study. Open circles represent filled studies to correct publication bias. c Forest plot of sensitivity test. Squares indicate the summary OR after omitting one specific study. Horizontal lines indicate 95% CI. A diamond indicates the summary OR of all the studies with its corresponding 95% CI. +LN positive lymph node, IDC intra-ductal carcinoma, NSCLC non-small cell lung cancer, HCC hepatocellular carcinoma, CRC colorectal carcinoma, PDAC pancreatic ductal adenocarcinoma, PTC papillary thyroid carcinoma, RCC renal cell carcinoma 1 3
Fig. 5
Cancer with low PEDF expression has increased odds to have higher pathological grade. a Forest plot showing the association of PEDF expression and tumor pathological grade. Squares indicate study-specific odds ratios (ORs). The size of the box is proportional to the weight of the study. Horizontal lines indicate 95% confidence interval (CI). A diamond indicates the summary OR with its corresponding 95% CI. b Funnel plot on the publication bias of studies on PEDF expression and tumor pathological grade. Each filled circle represents one study. Open circles represent filled studies to correct publication bias. c Forest plot of sensitivity test. Squares indicate the summary OR after omitting one specific study. Horizontal lines indicate 95% CI. A diamond indicates the summary OR of all the studies with its corresponding 95% CI. IDC intra-ductal carcinoma, NSCLC non-small cell lung cancer, HCC hepatocellular carcinoma, CRC colorectal carcinoma, PDAC pancreatic ductal adenocarcinoma, TCC (bladder) transitional cell carcinoma, RCC renal cell carcinoma
|
2021-12-15T14:31:33.395Z
|
2021-12-01T00:00:00.000
|
{
"year": 2021,
"sha1": "693cbdcdd1d9582d79c4a4bb2ea1d51cbf46fe3d",
"oa_license": "CCBY",
"oa_url": "https://link.springer.com/content/pdf/10.1007/s12672-021-00457-y.pdf",
"oa_status": "GOLD",
"pdf_src": "Springer",
"pdf_hash": "693cbdcdd1d9582d79c4a4bb2ea1d51cbf46fe3d",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
261049422
|
pes2o/s2orc
|
v3-fos-license
|
Inductive-bias Learning: Generating Code Models with Large Language Model
Large Language Models(LLMs) have been attracting attention due to a ability called in-context learning(ICL). ICL, without updating the parameters of a LLM, it is possible to achieve highly accurate inference based on rules ``in the context'' by merely inputting a training data into the prompt. Although ICL is a developing field with many unanswered questions, LLMs themselves serves as a inference model, seemingly realizing inference without explicitly indicate ``inductive bias''. On the other hand, a code generation is also a highlighted application of LLMs. The accuracy of code generation has dramatically improved, enabling even non-engineers to generate code to perform the desired tasks by crafting appropriate prompts. In this paper, we propose a novel ``learning'' method called an ``Inductive-Bias Learning (IBL)'', which combines the techniques of ICL and code generation. An idea of IBL is straightforward. Like ICL, IBL inputs a training data into the prompt and outputs a code with a necessary structure for inference (we referred to as ``Code Model'') from a ``contextual understanding''. Despite being a seemingly simple approach, IBL encompasses both a ``property of inference without explicit inductive bias'' inherent in ICL and a ``readability and explainability'' of the code generation. Surprisingly, generated Code Models have been found to achieve predictive accuracy comparable to, and in some cases surpassing, ICL and representative machine learning models. Our IBL code is open source: https://github.com/fuyu-quant/IBLM
Inroduction
Scaled up Language models "understand" data relationships "in the context" by inputting some a training data(e.g., input-output pairs) directly to a prompt without fine-tuning. Moreover, it has been shown to perform inference with high accuracy and it is expected applicable to a wide range of tasks by characteristics of LLMs [2]. This is a property of a large language models(LLM) called In-Context Learning(ICL)(eq.1, Figure 1): LLM.input prompt [D [(x 1 , y 1 ), · · · , (x n , y n )] , x input ] → y pred (1) Figure 1: In-Context Learning Where, the left side of the figure is the prompt that expresses the input for the training data and the inference, and the right side of the figure is the output that expresses the inference result of the LLM. ICL is a developing field, we give briefly review its development: First, as studies on its structures, ( [3] [13]) describe necessary conditions of training data for LLMs to acquire ICL ability. Moreover, in [1][8] [10] a relationship between an inference process of LLMs by ICL and a gradient descent is studied. There are still many unclear points about structures of ICL, and further researches are required.
Furthermore, as researches on its performance, [6] shows that it is possible to learn various functions by ICL such as sparse linear functions, 2-layer neural networks, decision trees, etc. by pre-training Transformers. Therefore, an LLM that satisfies the appropriate conditions appears to "understand" an unknown data relationships by giving some examples of the inputs and the outputs according to an ability of ICL and can predict the value of the output value for the input. Now we encounter a natural question related to ICL: "Isn't it possible to directly the output the resulting "model" that LLMs learned from the input and the output examples?" If the above issues can be resolved, we can explicitly show what data relationships LLM has found from the input and output examples. Furthermore, we can eliminate the "traditional machine learning with inductive bias by human selection" (Figure 2) and make direct inferences like ICL. This also means that LLM can be thought of as "learning how to learn" from a data, and can handle a perspective one step higher than ICL. This can be interpreted as bringing a "Meta-learning" perspective to ICL.
To solve this question, we formulate the problem concretely as follows: "Can LLM take a training data (here, a pair of explanatory and objective variables) as input and output a model for inferring the objective variable from the explanatory variables?" Our answer to this proposition is "YES". We named this method an Inductive-bias Learning (IBL) (eq. 2, Figure 3). The name the "Inductive-bias" Learning comes from outputting models with various structures for each data without explicitly assuming the inductive bias. In this paper, we performed the following verification: We set binary classification problems as a task, and the GPT-4 as an LLM. Moreover we used multiple datasets and seed values to sample the data, using IBL and representative machine learning algorithms and a comparison of ICL accuracy were performed. As results, we find that IBL can provide accuracy comparable to or even better than typical machine learning algorithms and ICL in some cases.
As given in the proposition, IBL outputs the inference structure itself from the training data. Moreover, this can be done for a variety of data. This suggests the possibility that LLM learns how to learn the relationship between input and output when given data, and LLM also has an aspect as a kind of Meta-Learning. This paper is organized as follows: Section 2 summarizes related work on ICL and the Meta-Learning. Section 3 gives an architecture and training method of IBL using a LLM. Next, Section 4 presents two results of verifying the performance of IBL. One is accuracy verification with conventional major machine learning models, and the other is accuracy verification with ICL. Finally, Section 5, 6 summarizes the verification of IBL and the efforts and applicability required for future IBL.
Related works
There are two major areas of research related to IBL: ICL, and the Meta-Learning.
In-context Learning
In the research by [6], ICL was demonstrated that by pre-training Transformer models, linear functions could be learned with accuracy comparable to the least squares estimator. Furthermore, this approach has been shown to be applicable not only to linear functions but also to sparse linear functions, decision trees, and two-layer neural networks. In IBL of this study, high accuracy in prediction was achieved in several datasets, but the relationship with the learning ability of ICL, and whether IBL can learn linear functions, decision trees, and two-layer neural networks is unknown.
In [12], the semantic prior distribution in ICL was removed, and accuracy testing was conducted using methods like Flipped-Label ICL (inverting the correct labels) or SUL-ICL (providing semantically unrelated labels). As a result, it became clear that if sufficient datas are provided, the semantic prior distribution can be overwritten. During the execution of IBL in this inverstigate, there is also a possibility that the Code Model is generated based on the provided training data by overwriting the semantic prior distribution if a sufficient number of data are given, but more detailed verification is needed.
Meta-Learning
IBL might suggest that a certain type of the Meta Learning, represented by approaches like MAML [5], can be realized with LLMs. In IBL, structures for predicting unknown data are inferred without explicitly setting the inductive bias from the training data alone. This can be seen as LLMs being able to learn how to infer a logic for predicting from the data alone during its pre-training stage. This might, in a sense, suggest that LLMs are performing the Meta Learning within its learning process.
For instance, in the research of Meta ticket [4], it is possible to find a neural network structure that achieves high predictive accuracy in few-shot learning from data. In contrast, in this research, GPT-4 is used to find the structure of a model with high predictive accuracy only from data. While the aim of finding the optimal model structure is the same, the approaches and the data handled are different.
What happens internally when LLMs execute Inductive-bias Learning, and how LLM acquires the ability of IBL (Meta-Learning-like properties) during a learning process, is an intriguing theme. The inference by ICL being realized through gradient descent is suggested by researches such as [6], [1], [10], but how IBL is being realized is still not understood.
Inductive-bias Learning
The Inductive-bias learning (IBL) can be divided into two phases, learning and inference, similar to the usual supervised machine learning algorithms. The learning phase takes the input explanatory variable and objective variable pairs as inputs and determines the structures of the model for predicting the objective variable from the explanatory variables. In inference, the created model can be used to make predictions about unknown data. In this study, we focused on binary classification. Details are explained below (Figure 4).
Learning
A learning phase of IBL uses a given training data to generate a code what is represented as Python functions to output an objective variable from explanatory variables. If necessary, we also create a Python file in which the code is written. We call this generated Python code a "Code Model".
Moreover to a training data, which are input-output pairs, the prompts for an LLM include instructions to output a probability of label 1 for accuracy evaluation as a binary classification task, and to avoid including code that uses the general machine learning models in the generated Python code. In this verification, the use of machine learning models was prohibited because we were interested in whether GPT-4 could output "logic" to make predictions from the data.
For an output, we generate a code for Pyhton function that takes the explanatory variable as input and the probability value what the label of the objective variable is 1 as output value. The Figure 4: Above: Model generation by IBL: As in normal supervised learning, input training data consisting of multiple pairs of explanatory and objective variables. A code is created based on the input data. If necessary, a Python file is also created. Below: Prediction of a model generated by IBL: As with ordinary machine learning models, input the explanatory variables for the data you wish to predict. Make predictions using the generated code. Alternatively, you can execute the generated Python file to make predictions. generated Python code is generated as conditional branches by certain features or linear expressions involving multiple features. Various output results are also obtained depending on a dataset and a number of data used for a learning.
Inference
The inference phase of IBL is performed using the Python code generated by the Learning. The Python code that takes explanatory variables as inputs and probability values as outputs is executed.
The Python code generated by IBL learning is very fast because it is generated for simple processes such as conditional branching by certain features or linear expressions involving multiple features. On the other hand, despite such simple processing, the generated Code Models have been verified to have very high prediction accuracy.
Furthermore, the Code Models generated by IBL are outputs as Python codes. This allows the user to see what features the model uses and how much they contribute to a forecast, making the model very easy to interpret.
Experiments and Results
In this section, we introduce two experiment results to test the performance of IBL. First, we compare the predictive ability of IBL and typical machine learning models on several datasets. Second, we compare the predictive ability of IBL and ICL on the "Titanic dataset".
Datasets and Task
We compared accuracy of Code Models generated by IBL with well-known machine learning algorithms and ICL, focusing on the binary classification tasks. The datasets used in this evaluation are the well-known Titanic dataset from Kaggle, and pseudo and moon datasets available through scikit-learn.
For the Titanic dataset, we performed imputation using the mean for missing values and applied One-hot encoding for categorical data to align the inputs across all machine learning models. Additionally, we created training data by sampling with three different seed values to validate with several input datasets. Datas that was not used for training in each seed value was considered as test data. For the pseudo and moon data from scikit-learn, we generated synthetic data using three different seed values and split it into training and test data. Since having too many decimal places would reduce the number of data points that could be inputted into IBL or ICL, all generated values were rounded to three decimal places. To eliminate the imbalance in the training data, we made the number of positive and negative examples the same.
Number of IBL executions
One of the issues with IBL is the instability of outputs which are generated Code Models; the Code Models generated by IBL do not always produce high prediction accuracy in a stable manner, and the accuracy varies. For this reason, we generated 30 Code Models for each dataset, training data size, and seed value, and compared the one with a highest AUC.
How to execute ICL
To predict test data based on training data through ICL, the following method is utilized. First, all the training datas are converted into a string and included in the prompt. Additionally, instructions are written into a prompt to output the explanatory variables of the data that you want to predict, along with the target variable for that data. This prompt is input into GPT-4, and the target variable is produced as the output. These processes are carried out for all the test datas, and the resulting predictions are considered the output of ICL. The intention behind this method is to include as much training data as possible in the prompt.
In this paper, gpt-4-0613 was used for IBL and ICL. No additional training such as fine tuning was performed, and only prompt input was used.
IBL vs Machine Learning Models
In the first validation, we compared the AUC of the Code Models generated by IBL with typical machine learning models using the Titanic dataset, the pseudo dataset, and the moon dataset, with various training data size. The Logistic Regression, the K-Nearest Neighbors(K-NN), and the Linear Kernel SVM were used as machine learning models for comparison. In IBL, code models were generated 30 times for each seed value, and the models with the highest AUC were selected for the results. If there are NO Code Models that can be successfully generated, the results are not plotted in the graph. The following are the validation results for each dataset( Figure 5, 6, 7). For all seed values and all training data size, IBL achieves higher AUC than the K-NN and the SVM. IBL also achieves higher AUC than all machine learning models in some conditions.
Pseudo dataset
(a) seed=3655 (b) seed=3656 (c) seed=3657 Figure 6: CComparison of AUC for each training data size on the pseudo dataset for IBL and each machine learning algorithm In some cases, the pseudo dataset outputs AUC comparable to other machine learning models, but the overall accuracy is often low. Furthermore, depending on the dataset size, there may be cases where viable models are not generated. The AUC in the moon dataset often exceed those of the SVM, although they are inferior to those of the K-NN and the logistic regression. IBL achieves very high AUC and accuracy on the Titanic dataset. This may be due to the fact that the gpt-4-0613 model includes codes that performed EDA on the Kaggle dataset in the training data. It is also possible that the model predicts Survived, the objective variable of the Titanic dataset, based on the name of each explanatory variable alone. Conversely, it is impossible to predict the value of the objective variable from the names of the explanatory variables for the pseudo dataset and the moon dataset, and it is also considered very difficult to identify that the data are generated using libraries such as sklearn in a very small dataset size. It is therefore considered that the LLM outputs logic for inference based only on the features from the data.
Moon dataset
Additionally, the high accuracy with the Titanic dataset might be attributed to the fact that it contains more categorical variables compared to the pseudo datasets or the moon dataset. This could have made it easier for the LLM to generate the logic for prediction when creating Code Models. To determine whether continuous or discrete explanatory variables are more easily handled by IBL, further verification with different datasets will be necessary.
Some of the generated Code Models are attached in the Appendices.
IBL vs ICL
The second validation compares ICL with IBL. The data for the comparison was Kaggle's Titanic dataset. These training data were sampled with three different seed values from the Titanic dataset, while the other data was used as test data. Below are the validation results( Figure 8). Figure 8 shows that IBL can be predicted with an AUC higher than ICL for all seed values. Furthermore, IBL can predict with an AUC that is more than 10 points higher than ICL, depending on the seed value. These results suggest that ICL may be making predictions based on training data input to the prompt, but when dataset size and features are large, it is difficult to find the characteristics of the entire dataset, and some data may be used to make predictions similar to some kind of over-trained model. On the other hand, IBL uses all of the training data for code generation, so it may be able to achieve higher accuracy than ICL due to its higher generalization performance.
Moreover, the contents of the Titanic dataset and the code used to analyze it are likely to have been used in the training of gpt-4-0613, which may have affected the code generation in IBL and the prediction in ICL. It is an interesting topic to see how much IBL and ICL are affected by the LLM training data on which they are based.
In this verification, we were unable to run the same test data multiple times on the same training data, or use other datasets or seed values, because the cost of the OpenAI API is very high when running ICL. We would like to work on these validations in the future.
Future works
IBL has the potential to be a paradigm shift for building new "machine learning algorithm". The following are summaries of validation items that are expected to become important in the future.
Improved accuracy of IBL
The Code Model with accuracy that exceeds that of ICL, when compared to machine learning models, accuracy is often inferior as the dataset size increases. This is because machine learning models generally improve in accuracy as the dataset size increases, but we have not observed such a phenomenon with IBL. This may be due to the fact that the training data input to the LLM are not sufficiently analyzed and the necessary patterns for prediction are not output, as in CoT(Chain-of-Thought) [11] , where the first instruction is to output text-based elements for prediction from the training data, and the second instruction is to output unknown data based on these texts. Then, a Code Model is created to predict unknown data based on the text, which is expected to improve the prediction accuracy.
Additionally, it is expected that more accurate models can be generated by incorporating statistical learning theory techniques like Bagging and Boosting, re-editing the created Code Model, and optimizing the prompts [9] that are input together with the training data.
We are also considering the possibility of IBL replacing socially implemented machine learning algorithms, such as the LightGBM [7]. Current Code Models generated by IBL have problems such as inferior accuracy and unstable output of Code Models. Conversely, if these problems can be solved to some extent, the performance of the Code Model will be useful enough for social implementation in terms of model inference speed and model interpretability.
Understanding of IBL capabilities
It remains to be analyzed how IBL infers a logic for inference from the training data to create Code Models and how good the inference logic of the created Code models is. For the latter, the inverstigate by [6] suggests that ICL can approximate linear functions with accuracy comparable to that of least squares estimators. This may allow us to verify how accurately LLMs can obtain regression coefficients from data using IBL.
Analysis of the generated Code model
More detailed analysis of the generated Code Models are needed, as they are still unclear what kind of logics are used inside the Code models and what kind of relationships exists between the number of data prompts (IBL training data size) and an internal logic. The former question is about the types of Code Models that can be generated, and while logistic regression-like structures and conditional branching with if statements were generated in most cases in this verification, it would be very interesting to see what other types of structures can be generated (e.g., nonlinear relationships involving exponential functions). It would be very interesting to see what other structures can be generated (e.g., nonlinear relationships involving exponential functions, etc.). The relationship between the latter and how the Code Model changes as the number of input data is increased is not yet known.
Conclusion
In this paper, we proposed a new learning method called "Inductive-bias Learning "(IBL), which uses LLM to create inference models. We found that the inference models generated by this learning method, Code Models, sometimes outperform conventional machine learning models. We also show that the accuracy of IBL exceeds that of ICL on the Titanic datasets. We believe that the predictive models generated by IBL (Code Model) have the potential to replace existing machine learning models in terms of interpretability and inference speed as their accuracy improves in the future. Furthermore, we believe that IBL may lead to a better understanding for ICL, which still has many unknowns.
A Appendices
The following is an example of the execution of the Code Model generated in this experiment.
A.1 Successful execution Code Model
Some examples of successfully created Code Models that can predict with high AUC for each data set are shown below. # Please describe the process required to make the prediction below . 8 9 # The logic here is to give more weightage to the features that are more likely to result in survival . 10 # For example , passengers in first class ( pclass =1) , female passengers ( sex_female =1) , passengers who embarked at Cherbourg ( embarked_C =1) , 11 # passengers who are alone ( alone_True =1) , passengers who are not adult males ( adult_male_False =1) , passengers who are women ( who_woman =1) , 12 # passengers in Deck B , C , D , E ( deck_B =1 , deck_C =1 , deck_D =1 , deck_E =1) and passengers who embarked at Cherbourg ( e m b ark_ town_ Cher bourg =1) 13 # are more likely to survive . Therefore , these features are given more weightage in the prediction . # Please describe the process required to make the prediction below . 9 10 # The logic here is to give higher weightage to the features that are more likely to result in survival ( target =1) 11 # and lower weightage to the features that are more likely to result in non -survival ( target =0) .
12
# The weights are determined based on the data provided . # Please describe the process required to make the prediction below . # Please describe the process required to make the prediction below . 10 11 # Here we are using a simple linear regression model to predict the target . 12 # The coefficients are chosen based on the observation of the data . 13 # The intercept is set to 0.5 to make sure the output is between 0 and 1. # Please describe the process required to make the prediction below . 10 11 # Here we are using a simple linear regression model for prediction . 12 # The coefficients are calculated based on the given data . 13 # The intercept is assumed to be 0 for simplicity . # The coefficients for Feature_1 and Feature_2 are 0.4 and -0.6 respectively .
15
# These coefficients are chosen such that the predicted value is close to 1 for target =1 and close to 0 for target =0.
A.2 Failed execution Code Model
Code Models that fail to execute could be roughly divided into the following three types of patterns: 1. Failure to import generated Code Model 2. If the import succeeds in executing but the predict method fails 3. The predict method is successfully executed, but the predicted probability value is incorrect Below are examples of Code Models that failed to execute in the verification of pseudo datasets.
First is the first pattern. In the following, the Python code itself is generated, but the entire code block is enclosed in backquotes, or a natural sentence describing the generated Python code is included at the beginning of the sentence, causing a syntax error and resulting in import failure.
Listing 7: Code Model of pattern 1 that failed to execute 1 Here is a simple Python code that uses a basic logistic regression model to predict the probability that the " target " of the unknown data is 1. This code does not use any existing machine learning model , but rather implements the logistic regression model from scratch . In this code , the ' sigmoid ' function is used to map any real -valued number into the range [0 , 1] , which can be interpreted as probabilities . The ' predict ' function applies the logistic regression model to each row of the input DataFrame 'x ' to predict the probability that the " target " of the unknown data is 1. The coefficients of the logistic regression model (0.1 , 0.2 , 0.3 , 0.4) are chosen arbitrarily and should be adjusted based on the actual data to improve the prediction accuracy .
The second pattern is an implementation that is fine as Python code, but causes a behavior error when the data is actually given and processed.
In the following the Code Models process, there are a part where predictions are converted to probability values. In the denominator of this part, although y is a scalar value, min and max are calculated and the difference is taken, resulting in a division by zero and all the values become nan. # Please describe the process required to make the prediction below . The third pattern is where the predictions are not properly normalized and exceed 1, or the predictions are negative.
In the following example, a min/max calculation should be performed for y in the normalization calculation section, but a min/max calculation is performed for df.
|
2023-08-22T06:43:01.561Z
|
2023-08-19T00:00:00.000
|
{
"year": 2023,
"sha1": "3bab38ba17a9fea40b0862befa0d1ac247d7ea03",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "3bab38ba17a9fea40b0862befa0d1ac247d7ea03",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science"
]
}
|
17062127
|
pes2o/s2orc
|
v3-fos-license
|
Spatial positioning of CFTR’s pore-lining residues affirms an asymmetrical contribution of transmembrane segments to the anion permeation pathway
CFTR is a chloride channel and a member of the ABC transporter superfamily; however, its structure is unknown. By making a series of cysteine mutants, Gao and Hwang show that CFTR lacks the twofold pseudo-symmetry seen in the permeation pathway of bone fide ABC transporters.
I N T R O D U C T I O N
Regulated ion movement across the cell membrane by transport proteins partakes in a broad array of physiological functions in the human body; hence, it is not surprising that disruptions of this process could lead to many disease states (Cooper and Jan, 1999). Among them, cystic fibrosis (CF) ails the suffered from their birth because of a defective chloride transport across the epithelial cells in the airway, the gastrointestinal tract, and the reproductive system (Riordan et al., 1989). Although CFTR-the culprit behind CF-is a bona fide anion channel, topological analysis of its amino acid sequence (Dean and Annilo, 2005) places CFTR in the superfamily of the ABC transporter (Riordan et al., 1989). Like most exporter members of this superfamily, CFTR possesses the prototypical architecture of two transmembrane domains (TMDs; TMD1 and TMD2), each followed by a cytoplasmic nucleotide-binding domain (NBD; NBD1 and NBD2). Each TMD consists of 6 transmembrane segments (TMs), and some of the overall 12 TMs have been shown to craft an anion-selective pore (Akabas et al., 1994;Alex-ander et al., 2009;Bai et al., 2010Bai et al., , 2011El Hiani and Linsdell, 2010;Qian et al., 2011;Wang et al., 2011Wang et al., , 2014Norimatsu et al., 2012;Gao et al., 2013;Zhang and Hwang, 2015). In contrast, the two NBDs play a key role in controlling opening/closing of the gate in TMDs. Specifically, after ATP binding to the two binding pockets formed by NBDs, dimerization of the two NBDs facilitates gate opening, whereas gate closure is greatly accelerated by ATP hydrolysis-induced disruption of the NBD dimer (Csanády and Gadsby, 1999;Gadsby and Nairn, 1999b;Jih and Hwang, 2012). In addition to these characteristic domains shared by other ABC proteins, CFTR distinguishes itself from its peers in this superfamily by harboring a unique intracellular regulatory domain (R domain) where several conserved serine/ threonine residues must be phosphorylated in order for ATP to gate the channel (Gadsby and Nairn, 1999a).
Along with the aforementioned topological similarities between CFTR and ABC exporters, functional resemblance that NBD dimerization triggers the gates in the TMDs to open has led to the proposition that CFTR The structural composition of CFTR's anion permeation pathway has been proposed to consist of a short narrow region, flanked by two wide inner and outer vestibules, based on systematic cysteine scanning studies using thiol-reactive probes of various sizes. Although these studies identified several of the transmembrane segments (TMs) as pore lining, the exact spatial relationship between pore-lining elements remains under debate. Here, we introduce cysteine pairs in several key pore-lining positions in TM1, 6, and 12 and use Cd 2+ as a probe to gauge the spatial relationship of these residues within the pore. We find that inhibition of single cysteine CFTR mutants, such as 102C in TM1 or 341C in TM6, by intracellular Cd 2+ is readily reversible upon removal of the metal ion. However, the inhibitory effect of Cd 2+ on the double mutant 102C/341C requires the chelating agent dithiothreitol (DTT) for rapid reversal, indicating that 102C and 341C are close enough to the internal edge of the narrow region to coordinate one Cd 2+ ion between them. We observe similar effects of extracellular Cd 2+ on TM1/TM6 cysteine pairs 106C/337C, 107C/337C, and 107C/338C, corroborating the idea that these paired residues are physically close to each other at the external edge of the narrow region. Although these data paint a picture of relatively symmetrical contributions to CFTR's pore by TM1 and TM6, introducing cysteine pairs between TM6 and TM12 (348C/1141C, 348C/1144C, and 348C/1145C) or between TM1 and TM12 (95C/1141C) yields results that contest the long-held principle of twofold pseudo-symmetry in the assembly of ABC transporters' TMs. Collectively, these findings not only advance our current understanding of the architecture of CFTR's pore, but could serve as a guide for refining computational models of CFTR by imposing physical constraints among pore-lining residues.
Spatial positioning of CFTR's pore-lining residues affirms an asymmetrical contribution of transmembrane segments to the anion permeation pathway may evolve from a primordial ABC exporter by degrading its cytoplasmic gate (Gadsby et al., 2006;Chen and Hwang, 2008). So far, high-resolution structures of eight ABC exporters have been solved (Dawson and Locher, 2006;Ward et al., 2007;Aller et al., 2009;Hohl et al., 2012;Jin et al., 2012;Shintre et al., 2013;Choudhury et al., 2014;Kodan et al., 2014). Regardless of whether the whole protein is composed of two identical halves or not, a twofold (pseudo) symmetry of the substrate translocation pathway is found. Experiments on CFTR's pore-lining TM6 and TM12 also appear to support such a symmetrical contribution of TM6 in TMD1 and its counterpart TM12 in TMD2 to the internal vestibule of the pore by demonstrating that cysteines introduced into the cytoplasmic portion of these two pore-lining TMs could be accessed by intracellular MTS reagents in both the open and the closed states (Bai et al., 2011). However, recent studies provided more details in the structural and functional roles of CFTR's TMDs. For example, although several studies assure a pore-lining role for TM1 in CFTR (Akabas et al., 1994;Wang et al., 2011;Gao et al., 2013), overwhelming evidence argues against such a role for TM7 (Wang et al., 2014;Zhang and Hwang, 2015), the counterpart of TM1, in pore construction, countering the long-held twofold pseudo-symmetry in the TMDs of ABC exporters. Furthermore, the extracellular end of ABC exporters' substrate translocation pathway closes to form a gate for the substrates (Kodan et al., 2014), but the external entryway of CFTR's pore is accessible to extracellular anions in a closed channel (Gao and Hwang, 2015). Specifically, our latest studies suggest that CF-TR's gate (between positions 337 and 344 in TM6) coincides with the previously identified narrow region of the pore that spans a short segment in TM6 (positions 338 to 341; Alexander et al., 2009;Bai et al., 2010). Interestingly, our cysteine scanning experiments on TM1 also identified a narrow region confined in between L102 to I106, which prevents bulky MTS reagents from passing through the channel (Gao et al., 2013). Therefore, the narrowest section in CFTR's pore could be as short as just one α-helical turn in length if both TM1 and TM6 indeed, as proposed, assume a secondary structure of an α-helix (Alexander et al., 2009;Bai et al., 2010;El Hiani and Linsdell, 2010;Gao et al., 2013). In addition to this narrow segment, the pore of CFTR comprises two flanking compartments known as internal and external vestibules (Liu et al., 2003;Linsdell, 2014;Gao and Hwang, 2015), which are relatively wide based on the observations that many bulky reagents, like MTS reagents or channel blockers (e.g., glibenclamide), can be accommodated in these regions (Sheppard and Robinson, 1997;Zhou et al., 2002;Bai et al., 2010Bai et al., , 2011Gao et al., 2013).
As described above, our molecular understanding of CFTR's pore and gate has been fostered immensely by a plethora of functional studies, but the relative positions of the pore-lining residues as well as the alignment between individual TMs remain unclear, lacking a high-resolution structure of CFTR's TMDs. Some recent experiments using the cross-linking strategy by engineering cysteine pairs do provide some insights into the positional relationship between pore-lining residues in the spacious internal vestibule (El Hiani and Linsdell, 2014). However, when mapped to homology models of CFTR (e.g., those in Corradi et al. [2015]), those supposedly cross-linkable residues are located too far away from each other than expected from the cross-linking data shown in El Hiani and Linsdell (2014). (See Fig. 1 legend for more details.) What is also surprising is a lack of studies tackling the regions close to the narrow region, which could be ideal for cross-linking experiments because of its physical constraints. Of note, although our previous study has proposed a revised topology for TM1 that places L102 residue at the internal edge of the narrow region, and hence close to S341 the corresponding position in TM6 (Gao et al., 2013), a quite different picture was depicted by another study (Wang et al., 2011). To fill these knowledge gaps, we used the paired cysteine strategy and performed more extensive experiments on TMs 1, 6, and 12, the three most extensively investigated pore-lining TMs, with a special emphasis on the residues just internal or external to the narrowest segment of the pore.
We introduced cysteine pairs to several positions in those three TMs and used Cd 2+ as our probe to test the formation of any "tight" Cd 2+ -binding sites by the introduced paired thiol groups. We found that although Cd 2+ -induced inhibition on all mutants carrying one single cysteine (except for 344C) could be reversed rapidly upon simple removal of Cd 2+ from the perfusate, this inhibitory effect on mutants with cysteine pairs 102C/341C, 106C/337C, 107C/337C, and 107C/338C becomes nearly irreversible. The observation that dithiothreitol (DTT), a strong Cd 2+ chelator (Krężel et al., 2001) that may have better access to Cd 2+ -thiol metal bridges than EGTA because of its smaller size, can rapidly reverse the effects brought by Cd 2+ implicates the formation of a metal bridge between these paired cysteines. These results are consistent with our previous cysteine scanning studies (Bai et al., 2010;Gao et al., 2013) and hence support the notion that TM1 and TM6 contribute to all three components of the pore. When we used the same approach to examine the relation between TM6 in TMD1 and its topological counterpart in TMD2, TM12, surprisingly, our data suggested that the previously defined middle section of TM12 (e.g., position 1141) does not align with the topologically equivalent position in TM6 (i.e., S341). Instead, 1141C forms a tight Cd 2+ -binding site with the engineered cysteine at position 348 in TM6, which is two helical turns internal to the S341 residue. A similar Cd 2+ -binding site was im-plicated between 1141C and 95C, the latter of which again is two helical turns cytoplasmic to the inner edge of the narrow region in TM1 (i.e., position 102). Collectively, the current study not only assigns specific amino acid residues that confine the functionally important region of the pore, but also provides experimental evidence echoing the idea of an asymmetrical construction of CFTR's pore by its two TMDs.
Mutagenesis and channel expression
The cysless background construct used in the current study was first made where 16 out of the overall 18 endogenous cysteines in CFTR were mutated to serines with the rest of 590C and 592C mutated to leucines (Mense et al., 2006). As described in Wang et al. (2007), an additional V510A mutation was introduced to all constructs to boost channel expression. Subsequently, site-directed mutagenesis was performed under this cysless background to engineer cysteines at positions of interest. All mutations were confirmed by DNA sequencing (DNA core, University of Missouri).
Chinese hamster ovary (CHO) cells grown in Dulbecco's modified Eagle's medium with 10% fetal bovine serum were used to express CFTR channels. Together with pEGFP-C3, which encodes the green fluorescence protein, all constructs were transfected into CHO cells using PolyFect reagent (QIA GEN), according to the instructions from the manufacturer. Electrophysiological data were collected 3-7 d after transfection.
Electrophysiology
Inside-out mode. Pipettes were made by pulling the borosilicate capillary glass using a vertical puller (Narishige). After fire polishing with a homemade micro-forge, the resistance of pipettes was between 1.5 MΩ and 2.5 MΩ when filled with the inside-out pipette solution, which contained (mM) 140 NMDG-Cl, 2 MgCl 2 , 5 CaCl 2 , and 10 HEP ES, with pH adjusted to 7.4 using NMDG. Before every experiment, cells expressing certain mutants were transferred into a chamber that was initially filled with a bath solution containing (mM) 145 NaCl, 5 KCl, 2 MgCl 2 , 1 CaCl 2 , 5 glucose, 5 HEP ES, and 20 sucrose, with pH adjusted to 7.4 using NaOH. After a tight GΩ seal was made, yet before our excising the patch into the inside-out mode, the background flow was switched to the standard inside-out perfusate containing (mM) 150 NMDG-Cl, 8 Tris, 10 EGTA, 2 MgCl 2 , and 10 HEP ES, with pH adjusted to 7.4 using NMDG. All inside-out perfusion solutions contained EGTA except for the one containing Cd 2+ . In addition, to ensure most of the cysteines in the patch were in their reduced form, DTT was added into the PKA/ ATP cocktail used routinely to activate the channels. All experiments with Cd 2+ were performed after CFTR currents had been activated to a steady-state with PKA and ATP. Because this phase of channel activation could take tens of minutes to complete, it was not included in our data presentations. Solution changes were achieved with a fast solution change system (SF-77B; Warner Instruments). The membrane potential was held at −50 mV for all inside-out experiments, and downward deflection indicates CFTR current. All of the currents were recorded with a patch-clamp amplifier (EPC 10; HEKA), filtered with an 8-pole lowpass Bessel filter (Warner Instruments) at 50 Hz, and digitized at 500 Hz using Pulse (V8.80; HEKA) into a PC hard drive.
Whole-cell mode. Pipettes used for whole-cell experiments were the same as those used for inside-out exper- Figure 1. Relative alignments among TM1, TM6, and TM12 depicted in a CFTR homology model. Based on the crystal structure of an asymmetrical ABC transporter, TM287/288, Corradi et al. (2015) derived a CFTR model in which the alignment between TM1 and TM6 (left) is fairly consistent with data represented in the current work. However, the distance between several positions where introduced cysteines were shown to be cross-linkable (El Hiani and Linsdell, 2014) is way beyond the physical constrain for the formation of an effective Cd 2+ metal bridge. For example, the distance between M348 and S1141 (15.5 Å, middle) and that between K95 and S1141 (17.6 Å, right) both well exceed the working distance of Cd 2+ , the cross-linker used in El Hiani and Linsdell (2014). Note that although the distances measured above are between α-carbons of the two residues involved, the minimum sulfur to sulfur distance between two engineered cysteines, calculated by subtracting 5.6 Å (twice the distance of a single cysteine side chain) from the above values (i.e., 9.9 Å for 348C and 1141C; 12 Å for 95C and 1141C), is still >6.5 Å, the distance for coordination of two thiol groups with Cd 2+ .
iments except that they were not polished for easier breaking in. The pipette resistance in whole-cell experiments was between 1.5 MΩ and 2.5 MΩ when filled with the whole-cell pipette solution containing (mM) 10 EGTA, 10 HEP ES, 20 TEACl, 10 MgATP, 2 MgCl 2 , 85 aspartate, 16 pyruvate, and 5.8 glucose, with pH adjusted to 7.4 using CsOH. The bath solution was the same as that used in inside-out experiments. After breaking in, the cell at the pipette tip was lifted from the bottom of the chamber to the outlets of the three-barrel glass where the reagents could be immediately perfused to the whole cell. The membrane potential was held at 0 mV throughout the whole recording, and upward deflections represent whole-cell CFTR currents. The whole-cell signals were filtered at 50 Hz and digitized at 2 kHz. All electrophysiological experiments were performed at the room temperature.
Reagents
PKA, ATP, and cadmium chloride were purchased from Sigma-Aldrich. Cadmium chloride was first prepared as 10 mM stock solution. Upon use, it was diluted into different concentrations using bath or the standard inside-out perfusate subject to mutants under test. CFTR inhibitor-172 (Inh-172) was provided by R. Bridges (Department of Physiology and Biophysics, Rosalind Franklin University, Chicago, IL) with support from Cystic Fibrosis Foundation Therapeutics.
Data analysis
Figure preparations and curve fittings to generate the time constant, τ, were performed with Igor Pro 6 (WaveMetrics). The cartoon depicting TM1, 6, and 12 in Fig. 1 was prepared with PyMOL (Schrödinger). The second order reaction rates were calculated as 1/(τ * [Cd 2+ ]). Student's paired or unpaired t tests were performed with Excel (Microsoft), with P < 0.05 considered as significant. All data are presented as mean ± SEM. n denotes the number of patches for each experiment.
Online supplemental material
Fig. S1 shows the effects of 5 µM Cd 2+ on 102C and 341C single mutant currents. Fig. S2 shows the effects of 5 µM Cd 2+ on 345C and 102C/345C currents. Fig. S3 shows a whole-cell recording of 106C/337C currents in response to forskolin and DTT. Fig. S4 shows reversible inhibition of 107C and 338C single mutant whole-cell currents by 1 mM extracellular Cd 2+ . Fig. S5 presents data suggesting that 341C and 1141C do not coordinate Cd 2+ together. Fig. S6 shows the effects of 50 µM Cd 2+ on 1141C current. Fig. S7 shows reversible inhibition of 1144C and 1145C currents by 5 µM Cd 2+ . Fig. S8 shows data suggesting that 95C and 1140C are not close enough to bind one Cd 2+ ion together. Fig. S9 shows the measurement of the reaction rate between Cd 2+ and 348C/1141C in the absence of ATP. Online supplemental material is available at http ://www .jgp .org /cgi /content /full /jgp .201511557 /DC1.
R E S U LT S
Although all of our experiments were performed under the cysless background where all 18 endogenous cysteines in CFTR were mutated to serine or leucine (see Materials and methods), as Cd 2+ can interact with amino acids other than cysteine (Rulíšek and Vondrášek, 1998;Choi et al., 2013), we first tested its effect on the WT cysless CFTR channels. As shown in Fig. 2 A, after phosphorylation-activated WT/cysless CFTR channels were opened by 2 mM ATP in an excised inside-out patch, 5 µM Cd 2+ applied intracellularly reduced the macroscopic current by 22.8 ± 2.8% (n = 5; Fig. 2 C), which was readily reversed upon removal of the metal ion from the solution. Even at a concentration as high as 1 mM, this inhibitory effect of Cd 2+ vanished within seconds once Cd 2+ was washed out from the perfusion solution, an observation similar to that in Fig. 2 A except for a slightly higher inhibition of the CFTR current (32.2 ± 9.6%; n = 5; Fig. 2, B and C). Although we did not know how and where Cd 2+ binds to exert this effect, its speedy reversibility within seconds of solution changes indicates a loose binding and allows inferences of the formation of a tight Cd 2+ -binding site by our engineered cysteine pairs, should we observe any Cd 2+ -dependent inhibition that reverses much more slowly. Similar control experiments were performed in whole-cell experiments with WT/cysless channels: in the presence of forskolin, 1 mM Cd 2+ applied from the extracellular side only induced negligible inhibition of the CFTR current, but the specific CFTR Inh-172 (Ma et al., 2002) abolished nearly all of the current (Fig. 2 D). It is because of such high tolerance of CFTR to external Cd 2+ that all our following whole-cell experiments adopted this concentration, which helped shorten the recording time so as to maintain a good patch quality.
Contributions of TM1 and TM6 to the pore's narrow region As our previous substituted cysteine accessibility method (SCAM) experiments using bulky MTS reagents placed L102 of TM1 and S341 of TM6 at the intracellular edge of the narrow region in the pore (Gao et al., 2013), we first tested whether introducing cysteines at these two positions can craft a tight Cd 2+ -binding site. As a control, the effectiveness of Cd 2+ on 102C and 341C single mutants was quantified. As shown in Fig. S1, application of 5 µM Cd 2+ in the continuous presence of ATP decreased 102C and 341C currents by 42.8 ± 2.7% (n = 5) and 38.5 ± 3.2% (n = 3), respectively. However, subsequent removal of Cd 2+ in the solution readily recovered the current to the control level in both cases, indicating a relatively weak interaction between Cd 2+ and either of these two cyst-eines. In stark contrast, when a double mutant with both 102C and 341C was put to the test, 5 µM Cd 2+ abolished virtually all of the current (95.6 ± 3%; n = 3) in the presence of ATP, and very little current recovered even after tens of seconds of washout with a Cd 2+ -free solution (Fig. 3). We interpreted this result as a tight coordination of one Cd 2+ ion by the two thiol groups in 102C and 341C, respectively, an idea also supported by the observation of a fast and full recovery of the current by DTT (Fig. 3 A). Because, in order for Cd 2+ to be coordinated concurrently by two cysteines, the distance between two thiols needs to be <6.5 Å (Jalilehvand et al., 2009), these results place 102C and 341C in close vicinity of each other at the intracellular edge of the narrow region in the pore (see Discussion for more detailed elaborations).
Next we tested whether 102C can form a tight binding site with a cysteine introduced at position 344, a pore-lining residue located one helical turn cytoplasmic to S341 in TM6. Although 5 µM Cd 2+ only abolished 38.1 ± 5.2% (n = 6) of 344C currents, the time course of current recovery upon washout of Cd 2+ was unusually slow. Specifically, current recovery of other single cysteine mutants upon washout of Cd 2+ takes place in seconds, whereas it takes tens to hundreds of seconds for 344C currents to recover, suggesting a slow off rate of Cd 2+ for 344C. However, in light of such a slow off rate, the low inhibition ratio seen with 5 µM Cd 2+ also indirectly infers a slow apparent on rate of Cd 2+ for 344C. To better assess the current recovery rate, we increased the concentration of Cd 2+ to 50 µM, which inhibited 86.4 ± 1.7% (n = 14) of 344C currents (Fig. 4 A). Indeed, the current recovery time course after washout of 50 µM Cd 2+ can be well fitted with a single exponential function yielding a mean time constant of 198 ± 21 s (n = 8). Although the reason for the slow on rate and off rate of Cd 2+ to 344C channels is unclear, 50 µM Cd 2+ was used for all constructs containing 344C. Interestingly, as seen in Fig. 4 A, DTT was able to expedite the recovery process with a significantly (P < 0.005) shorter time constant of 58 ± 11 s (n = 5). Because a single cysteine is not expected to afford such a slow off rate for Cd 2+ (Choi et al., 2013; also see Discussion), these results suggest that 344C may coordinate Cd 2+ with some endogenous moi-eties yet to be identified. Nonetheless, if 102C could contribute to this coordination, we would expect an even tighter binding of Cd 2+ for the double mutant 102C/344C. However, after a nearly full abolition (95.9 ± 1.8%; n = 7) of 102C/344C currents by 50 µM Cd 2+ (Fig. 4 B), upon removal of Cd 2+ , currents recovered with a time constant of 168 ± 24 s (n = 5), which is not significantly different from the recovery time constant for the 344C single mutant (P = 0.57; Fig. 4 E). In addition, DTT also significantly (P < 0.001) shortened the time course of current recovery with an averaged time constant of 43 ± 8 s (n = 4; Fig. 4, C and E), which is again not different from that measured with 344C sin- gle mutant (P = 0.36). Because the inhibition (77.6 ± 5.5%; n = 4) on 102C current by 50 µM Cd 2+ is readily reversible within seconds ( Fig. 4 D), the little difference in the effect of Cd 2+ between 344C and 102C/344C supports the notion that it is 344C with its unknown partner, rather than 102C, that may largely account for the slow recovery seen in 102C/344C double mutants. Thus, despite a confounding effect from a yet to be identified binding partner for Cd 2+ binding to 344C, these data suggest little contribution of the free energy of binding from 102C to the interaction between Cd 2+ and the double mutant 102C/344C.
Compared with the complicated interaction between Cd 2+ and 344C, V345, a neighboring pore-lining residue in TM6, offers another opportunity to address the same issue: whether cysteines placed one helical turn intracellular to S341 are close to 102C in TM1. As shown in Fig. S2, although Cd 2+ -induced inhibition on the double mutant is somewhat higher than that seen with single mutants 102C and 345C (Fig. S2 C), which is perhaps not surprising because the sheer presence of two thiol groups increases the probability of occupancy of the double mutant by Cd 2+ to disturb the chloride flow in the pore, the fast recovery of the 102C/345C current upon Cd 2+ washout (Fig. S2 B) led to a tentative conclusion that these two cysteines may not be close enough to coordinate one Cd 2+ ion (see Discussion for details). Meanwhile, the idea that L102 and S341 are in registry to each other at the intracellular edge of the narrow region is further supported by the findings made on another double mutant, 98C/341C. As shown in Fig. 5, inhibitions by 5 µM Cd 2+ on 98C (50.4 ± 8.2%; n = 4) and 98C/341C (46.4 ± 5.6%; n = 4) currents are very similar (P = 0.7), and both inhibitions are readily reversed upon removal of the metal ion, indicating that 98C, which is one helical turn cytoplasmic to position 102, is not able to bind a Cd 2+ ion together with 341C. Now that two residues at the intracellular edge of the narrow region-position 102 in TM1 and position 341 in TM6-are shown to be aligned closely to each other, under the premise that TM1 and TM6 share a similar secondary structure (Alexander et al., 2009;Bai et al., 2010;El Hiani and Linsdell, 2010;Wang et al., 2011;Gao et al., 2013), residues at the extracellular edge of the narrow region in TM1 could also be close to their counterparts in TM6 as the narrowest section in the pore spans only 3-4 amino acids along both TM1 and TM6 (Gao et al., 2013). To test this idea, we introduced cysteines to positions at the extracellular edge of the narrow region in these two segments and assessed their reactivity to Cd 2+ . We set out first with 106C/338C as these two individual positions are extracellularly closest to the narrow region. But unfortunately, this mutant did not yield any measurable current for our whole-cell experiments (n > 10), although each single mutant was functional (see below). Therefore, we moved the cysteine in TM6 from position 338 to position 337 to construct a double mutant 106C/337C. As controls, we first tested the effect of extracellular Cd 2+ on the two single mutants, 106C and 337C. As shown in Fig. 6 A, after whole-cell 106C current was elicited by 10 µM forskolin to a steady-state in the presence of DTT, external application of 1 mM Cd 2+ nearly abolished all of the 106C current (96.2 ± 0.4%; n = 4; Fig. 6 D), which was immediately recovered upon exclusion of Cd 2+ . Subsequent inhibition of the recovered current by Inh-172 (Ma et al., 2002) confirmed that the restored current was indeed from CFTR. As for 337C, Cd 2+ induced a lower (33.2 ± 2.6%; n = 7) but still reversible inhibition (Fig. 6, B and D), and the effect of Inh-172 on this mutant also seemed reduced when compared with that in other whole-cell experiments. These results suggest Cd 2+ does not bind tightly to either 106C or 337C. However, very interesting experimental results were obtained with the double mutant 106C/337C. First, as shown in Fig. 6 C, in the presence of forskolin, Cd 2+ applied right after DTT caused an inhibition (90.7 ± 4%; n = 4) that was not reversed to an appreciable extent even after washout of the metal ion for tens of seconds, but DTT could readily reverse this inhibition, suggesting that Cd 2+ is tightly coordinated by 106C and 337C. Second, although forskolin alone could elicit 106C/337C CFTR current, DTT further enhanced this current. Moreover, once DTT was removed, the current started to decay by itself as shown in the red rectangle in Fig. 6 C (also see Fig. S3). One simple explanation for the spontaneous current decay in the absence of the reducing DTT is that these two cysteines form a disulfide bond in a relatively oxidizing extracellular environment, especially in view of the observation that 106C and 337C can be close enough to coordinate a Cd 2+ ion. Of note, susceptibility of these two cysteines to possible contaminating metals could also diminish the double mutant current (see Liu et al. [2006] for more detailed discussions).
As the A107 residue that neighbors I106 in TM1 also lines the pore, we next tested whether 107C could duplicate what we observed with 106C/337C. When 1 mM Cd 2+ was applied together with forskolin to cells expressing 107C or 338C, no matter right after forskolin plus DTT or forskolin alone, the whole-cell current was greatly diminished with an averaged inhibition ratio of 93.5 ± 6.6% (n = 3) for 107C and 77.1 ± 2.7% (n = 7) for 338C ( Fig. 7 C and Fig. S4). Because this current inhibition was only seen in the continuous presence of Cd 2+ , we conclude that Cd 2+ binds loosely to both 107C and 338C. However, the reversal of Cd 2+ -induced inhibition of the double mutant 107C/337C or 107C/338C was only dramatically expedited by DTT (Fig. 7, A and B), indicating a formation of the cysteine-metal-cysteine complex in these two constructs. Therefore, the three testable cysteine pairs-106C/337C, 107C/337C, and 107C/338C-can all craft a tight binding site for Cd 2+ , corroborating the idea that these residues from TM1 and TM6 nestle in a close proximity at the extracellular edge of the narrow region. Together with data on 102C/341C (Fig. 3), we conclude that the narrow region of the pore is flanked internally by L102 and S341 and externally by I106, A107, F337, and T338, a picture consistent with our previous results (Gao et al., 2013; Gao and Hwang, 2015; but cf. Wang et al., 2011).
Asymmetrical contributions to CFTR's pore by TM6 and TM12 As described in the Introduction, twofold (pseudo) symmetry of the substrate translocation pathway is a hallmark for ABC exporters. For CFTR, all homology models based on ABC exporters' crystal structures also portray a pore so constructed that TM6 and TM12 play a symmetrical structural role (Serohijos et al., 2008;Dalton et al., 2012;Norimatsu et al., 2012;Rahman et al., 2013;Corradi et al., 2015;Mornon et al., 2015). However, such a picture is contested by the following observations. First, unlike that of TM6, the extracellular half of TM12 has been questioned as pore lining (Norimatsu et al., 2012). Second, when carrying out cysteine scanning experiments on TM6 and TM12, Bai et al. (2011) found that cysteines introduced in the topologically de-fined middle section of TM12 (e.g., 1141C) could be modified by intracellular MTS reagents much faster than those placed in the middle of TM6 (e.g., 341C), suggesting that amino acids in the "middle section" of TM12 may reside in a relatively wider space. We therefore extended our paired cysteine approaches to examine the relative position between TM6 and TM12.
We first tested the cysteine pair 341C/1141C, as previous SCAM data suggest these two positions define the Reversible inhibition of whole-cell 106C (A) and 337C (B) currents by 1 mM of extracellular Cd 2+ . Note that the effect of CFTR inhibitor 172 on 337C is much weaker than that seen with 106C (A). (C) Inhibition of whole-cell 106C/337C current by external Cd 2+ and DTT-dependent recovery. The red rectangle encloses a part of the recording where current decays by itself when DTT is not present, suggesting possible spontaneous formation of a disulfide bond between 106C and 337C (see Results for details). (D) Averaged current inhibition ratios by external Cd 2+ for 106C, 337C, and 106C/337C. Mean ± SEM is shown.
internal access limits in TM6 and TM12, respectively (Alexander et al., 2009;Bai et al., 2010Bai et al., , 2011. However, inhibition of the double mutant 341C/1141C (54.9 ± 4.5%; n = 4) by internal application of 5 µM Cd 2+ was only slightly larger than those on single mutants 341C (38.5 ± 3.2%; n = 3) or 1141C (44.4 ± 3.5%; n = 5; Fig. S5). More importantly, washout of Cd 2+ resulted in an immediate recovery of the current for either single or double mutant. Next, we moved the cysteine in TM6 to positions one helical turn cytoplasmic to construct two double mutants-344C/1141C and 345C/1141C. For 344C/1141C, although, upon application of 50 µM Cd 2+ , the current was almost fully inhibited (99.1 ± 0.3%; n = 5); simple washing out of Cd 2+ with ATP recovered the double mutant current with a mean time constant of 219 ± 23 s (n = 6). Meanwhile, DTT could greatly speed up this process with a time constant of 42 ± 7 s (n = 5; Fig. 8 A). These time constants are not statistically different from those of 344C single mutant (Fig. 4 E). In view of the ready recovery of 1141C current after removal of 50 µM Cd 2+ (Fig. S6), we again proposed that 344C and its "invisible partner" could account for the slow recovery of 344C/1141C current. As for 345C/1141C, 5 µM Cd 2+ evoked an inhibition (48.4 ± 7.1%; n = 4) that is not much different from that of 1141C (Fig. S5 C). More importantly, the immediate recovery of the double mutant current upon exclusion of Cd 2+ from the solution prompted us to conclude that 345C and 1141C are not able to provide a bi-dentate Cd 2+ -binding site (Fig. 8 C).
These results suggest that position 1141 in the middle of TM12 does not align with position 341, 344, or 345 of TM6. In contrast, the data shown in Fig. 9 lead to the conclusion that position 1141 is in registry with position 348, which is two helical turns cytoplasmic to position 341 in TM6. Specifically, the recovery rate of the Cd 2+ -inhibited current upon simple removal of Cd 2+ was too slow to be measureable; however, DTT greatly accelerated the current recovery ( Fig. 9 B). In addition, 5 µM Cd 2+ caused an inhibition of 88.6 ± 2.7% (n = 5) of the 348C/1141C current, which is much higher than that seen with either corresponding single mutant (Fig. 9 C). Although the aforementioned data put M348 in the close vicinity of S1141, a position close to the middle of TM12, examination of 348C/1144C and 348C/1145C with Cd 2+ provided further insights into the relative alignment between TM6 and TM12. As shown in Fig. S7, both 1144C and 1145C currents could be readily inhibited by 5 µM Cd 2+ , but a fast relief of the inhibition was also seen upon washout of Cd 2+ in both cases. However, this is not the case for 348C/1144C (Fig. 10 A). Besides a higher inhibition ratio of the mean current induced by Cd 2+ (Fig. 10 D), the recovery from the inhibited state takes a long time (hundreds of seconds) to reach its steady-state. Fitting the recovery time course with a single exponential function yielded a time constant of 182 ± 30 s (n = 5), which is significantly shortened by DTT (40 ± 7 s; n = 4; Fig. 10, A and B), suggesting that 348C and 1144C can coordinate a single Cd 2+ ion, but the 348C-Cd 2+ -1144C complex is not as strong as the 348C-Cd 2+ -1141C metal bridge (Fig. 9). In contrast, results with 348C/1145C double mutant were similar to those of 348C/1141C (Fig. 10 C). Collectively, these data suggest that the topologically defined middle segment of TM12 is not aligned with that of TM6 (i.e., position 341), but rather with a position that is two helical turns cytoplasmic to position 341 (i.e., position 348). More interestingly, such an alignment between TM6 and TM12 is further supported by the observations made on another double mutant formed by 95C (TM1) and 1141C. Specifically, although Cd 2+ -induced inhibition on 95C (34.6 ± 4.2%; n = 7) could be readily released within seconds upon washing Cd 2+ out with a solution with In an inside-out recording, 5 µM Cd 2+ induced reversible inhibition on 345C/1141C, suggesting 345C and 1141C are not able to construct a bi-dentate Cd 2+ -binding site. EGTA ( Fig. 11 A), reversal of the much higher inhibition (92.9 ± 2.1%; n = 7) seen with the 95C/1141C double mutant took a much longer time, as indicated by the current recovery time constants (84 ± 15 s; n = 8). However, such a time course could be greatly shortened by the metal chelator DTT (6 ± 1 s; n = 5; Fig. 11, B and C), suggesting the formation of a Cd 2+ bridge between 95C and 1141C. Thus, position 1141 in TM12 registers with both position 95 in TM1 and position 348 in TM6; they are all located two helical turns cytoplasmic to the internal edge of the narrow region. Collectively, these data support a picture that the nominal counterparts-TM6 and TM12-do not contribute to CFTR's pore symmetrically. Figure 9. 348C, but not 341C, of TM6 can tightly coordinate Cd 2+ with 1141C of TM12. (A) An inside-out recording shows that the inhibition of 341C/1141C currents was subject to the continuous presence of Cd 2+ in the solution. (B) Contrary to that shown in A, for the double mutant 348C/1141C, the current inhibition brought by Cd 2+ in an inside-out patch was reversed extremely slowly by simple washout, but DTT rapidly restored the current. The current decay phase was fitted with a single exponential function (see Fig. S9 for more details). (C) Summary of the percent inhibition of 348C, 1141C, and 348C/1141C currents by 5 µM Cd 2+ . Mean ± SEM is shown. Figure 10. 348C of TM6 forms a tight Cd 2+ -binding site with either 1144C or 1145C of TM12. (A) Acceleration of current recovery from Cd 2+ -induced inhibition of 348C/1144C currents by DTT in an inside-out recording. Although 348C/1144C currents could recover from the Cd 2+ -inhibited state irrespective of the presence of DTT, comparison of the time constants generated by fitting both current relaxations with a single exponential function (red and blue solid curves) indicates DTT can accelerate such a process. (B) Comparison of 348C/1144C current relaxation time constants indicates there is significant difference when DTT is in the solution. **, P < 0.01. (C) DTT-dependent recovery of 348C/1145C inside-out current after Cd 2+ application implicates a tight Cd 2+ -binding site constructed by these two cysteines. (D) Summary of effect of Cd 2+ on each mutant as indicated in the panel. Mean ± SEM is shown.
D I S C U S S I O N
Site-directed mutagenesis offers investigators a unique tool to study the spots of interest in an ion channel protein by allowing them to replace original amino acids with others and examine post hoc the functional consequences of these mutation-induced perturbations with single-molecule patch-clamp techniques. SCAM has supplemented this already powerful combination as the result of a rapid reaction of cysteine's thiol group with swaths of thiol-specific modification reagents (Wilson and Karlin, 1998;Serrano et al., 2006). In the past two decades, this method has been implemented successfully in the studies of structure/function of CFTR. Especially because of the known molecular sizes of these bulky reagents, one can also, to some extent, use them to gauge the physical dimension of the pore. Consequently, a rough picture of an hourglass-shaped pore for CFTR has subsequently emerged (see Introduction). In contrast, reducing the size of the probe to approximate that of a permeating anion (e.g., [Au(CN) 2 ] − ) can afford "chemical scanning" along the whole ion permeation pathway . Such an effective application of the permeant thiol-specific reagents enabled our latest localization of a gate in CFTR (Gao and Hwang, 2015).
Moreover, by engineering cysteine pairs, one cannot only take advantage of the potential for disulfide bond formation between two thiol groups in a close proximity, but also use a broad array of soft metals (e.g., Cd 2+ ) as a tool to cross-link two cysteines that may not be close enough to form a disulfide bond. Indeed, a series of elegant studies by Yellen's group on potassium channels and hyperpolarization-activated cation (HCN) channels showed that Cd 2+ could be tightly coordinated by multiple physically close cysteines or histidines to irreversibly alter the channel's gating or permeation properties (Yellen et al., 1994;Liu et al., 1997;Holmgren et al., 1998;del Camino and Yellen, 2001;Rothberg et al., 2002Rothberg et al., , 2003Webster et al., 2004;Ryu and Yellen, 2012). Similar strategies of engineering cysteine pairs have also been adopted by several groups in the CFTR field with fruitful results. For example, by engineering cysteine pairs in the head, central, and tail regions of CFTR's two NBDs, Mense et al. (2006) demonstrated that the two NBDs in CFTR can form a head to tail "heterodimeric" complex upon ATP binding, which likely resembles those observed in other ABC proteins. Guided by a homology model built using the crystal structure of Sav 1866 as the template, Riordan's group introduced cysteine pairs into the cytoplasmic-membrane domain interface and successfully identified several potential contact sites between NBDs and TMDs (He et al., 2008;Serohijos et al., 2008). In addition, a recent study, by engineering cysteine pairs in the extracellular loop, provided evi-dence for electrostatic interactions at the outer mouth of the pore (Cui et al., 2014).
However, we were surprised that very few studies actually implement this strategy to tackle CFTR's TMDs, especially because a low sequence similarity between CFTR's TMDs and those of other ABC proteins should Figure 11. Positive contribution of 95C in TM1 to Cd 2+ binding to 1141C in TM12. (A) Reversible inhibition of 95C inside-out currents by Cd 2+ . (B) After Cd 2+ -induced inhibition, 95C/1141C currents can recover at a discernable rate back to the preinhibition level by simple Cd 2+ removal; in the same inside-out patch, current recovery is dramatically accelerated by DTT. Time constants were generated from fitting both current relaxations with a single exponential function (red and blue curves). (C) Summary of the relaxation time constants for current recovery in the presence or absence of DTT. Mean ± SEM is shown. **, P < 0.01. Here, we need to point out that, although another study drew the same conclusion that position 95 is close to position 1141 in the pore (El Hiani and Linsdell, 2014), observations made between the current study and that study are somewhat different. Specifically, in El Hiani and Linsdell (2014), both 95C and 95C/1141C currents were increased upon Cd 2+ exposure, opposite to what we observed. The reason behind this discrepancy is unclear, but numerous differences in SCAM results have been noted before in the literature (Bai et al., 2010(Bai et al., , 2011El Hiani and Linsdell, 2010;Qian et al., 2011;Wang et al., 2011;Gao et al., 2013). Regardless, because the interaction between Cd 2+ and the binding site constructed by 95C and 1141C is relatively weak judged from a relatively faster relaxation time constant, when compared with other positive pairs in the current study (e.g., 102C/341C), we wondered whether moving the cysteine at position 1141 to its neighboring positions (i.e., 1140C or 1142C) could potentially identify a tighter binding site for Cd 2+ with the 95C background. Unfortunately, although the 95C/1142C double mutant failed to yield any current in our experiments (n > 10), 95C/1140C seems less likely to compose such a tight binding site we sought (see Fig. S8). pose a tremendous challenge to rely solely on computational approaches to gain a molecular understanding of the structure/function relation for CFTR's TMDs. We therefore adopted the paired cysteine method to investigate TM-TM interactions with a focus on the three most extensively studied pore-lining TMs in our laboratory-TM1, 6, and 12 (Bai et al., 2010(Bai et al., , 2011Gao et al., 2013). Of note, a recent study has provided evidence for positive interactions among these three TMs (El Hiani and Linsdell, 2014). Although their primarily interested candidates are residues lining the inner vestibule of CFTR's pore, we made a deliberate decision to place an emphasis on residues around the narrow region for the following reason. It is generally accepted that the inner and outer vestibules in CFTR's pore are spacious, but the narrow region in between should bear a width that is <5.3 Å based on the permeability ratio measured among different anions (Linsdell et al., 1997), making Cd 2+ an ideal cross-linker at the two edges of this restrictive region.
To our delight, 102C in TM1 and 341C in TM6 indeed form a tight binding site for Cd 2+ applied from the intracellular side of the membrane (Fig. 3). The inferred physical proximity between these two positions not only corroborates our previous conclusion (Gao et al., 2013) that L102, originally assigned to define the external end of CFTR's TM1 based on hydropathy analysis (Riordan et al., 1989), should instead be placed at the internal edge of the narrow region similar to S341 in TM6 (Bai et al., 2010), but also demonstrates that both TM1 and TM6 contribute to the internal end of the narrow region. Moreover, the observation that cysteine pairs (including 106C/337C, 107C/337C, and 107C/338C) introduced three to five residues external to L102 and S341, respectively, can tightly coordinate Cd 2+ applied from the extracellular side of the membrane supports the notion that residues from TM1 and TM6 also flank the external edge of the narrow region (Alexander et al., 2009;Bai et al., 2010;Gao et al., 2013). Thus, given that both TM1 and TM6 also harbor additional residues that build the internal and external vestibules, we posit that both TM1 and TM6 span the whole length of CFTR's anion permeation pathway (also see Gao and Hwang [2015]).
However, the almost irreversible binding of Cd 2+ with the aforementioned cysteine pairs is puzzling as such tight binding of a Cd 2+ ion is reserved for a preferred tetrahedral coordination seen in numerous metalloproteins (Rulíšek and Vondrášek, 1998). Indeed, the formation of a Cd 2+ -binding site of covalent bond strength with four cysteines has been shown in several ion channel studies (Yellen et al., 1994;Webster et al., 2004). Noticeably, a recent study by Choi et al. (2013) reported an off rate constant (k off ) of 3.5 × 10 −2 s −1 for two cysteines that are cross-linkable in the staphylococcal α-hemolysin (αHL) transmembrane pore, which corresponds to a dissociation time constant of ∼29 s. Thus, although the pore structure of CFTR is different from that of αHL, it is likely that the positive pairs we identified (e.g., 102C/341C) borrow some additional help from some yet to be identified elements in the pore so that they can together construct a binding site that has a much higher affinity for Cd 2+ than one constructed by two lone cysteines. Moreover, the idea that other chemical moieties in CFTR's pore may participate in coordinating Cd 2+ may also explain the observation that recovery of the single cysteine mutant 344C currents from Cd 2+ -induced inhibition takes hundreds of seconds to complete (Fig. 4 A), which is in stark contrast to that shown in Choi et al. (2013) (a time constant of <0.08 s for single cysteine in αHL). Notwithstanding, introduction of an additional cysteine-102C-did not seem to alter coordination of Cd 2+ by 344C and its yet to be identified partners, suggesting that, unlike positions 102 and 341, 102C and 344C may not be close enough to coordinate one Cd 2+ ion. Interestingly, if we accept the premises that the diameter of the narrow region is ∼5.3 Å, as indicated above, and that both TM1 and TM6 assume the same α-helical structure with a pitch of 5.4 Å, then the diagonal distance (>7.5 Å assuming a vertical triangle formed by residues 102, 341, and 344) between 102C and 344C should exceed that required for a multi-liganded Cd 2+ -binding site (Fig. 4). The same geometric predictions can be made for cysteine pairs 102C/345C and 98C/341C. Indeed, our data ( Fig. 5 and Fig. S2) show no evidence for a cooperative binding of Cd 2+ by any of these two paired cysteines. In contrast, we found experimental evidence for a close proximity among four positions at the external end of the narrow region along TM1 and TM6 (i.e., 106C/337C, 107C/337C, and 107C/338C). Thus, the very region defining the narrowest section of the pore is made partly by TM1 and TM6 over a distance of one helical pitch or 5.4 Å.
Although the well-known coordination chemistry of Cd 2+ with the cysteine thiols affords a more exquisite structural interpretation for a relatively compact environment where the space between each party is constrained, this favorable attribute of physical constraint is lost once we move our discussion focus to a more spacious compartment of CFTR's pore-the inner vestibule. Nonetheless, we were surprised by the extremely slow off rates of Cd 2+ in several of the engineered cysteine pairs (e.g., 348C/1141C and 348C/1145C), suggesting the presence of binding partners other than the introduced thiols in the inner vestibule. Although more extensive future studies on the mechanism of this apparent tight Cd 2+ binding are needed, the current data do reveal an unexpected role played by TM12 in the pore. Specifically, our data in Fig. 9 place 1141C of TM12 in close vicinity to 348C, contradicting the proposition that position 1141 in TM12, like its counterpart S341 in TM6, may be close to the internal brink of the narrow region (Bai et al., 2010(Bai et al., , 2011. Then, results in Fig. 10 further suggest 348C may straddle between 1141C and 1144C (or 1145C), which may line on the same face of two adjacent helical turns. This geometrical constrain subsequently predicts that S1141 should be located one to two helical turns away from the narrow region of the pore. Interestingly, 1141C indeed contributes to Cd 2+ binding to 95C (Fig. 11), a position two helical turns internal to L102 of TM1 that defines the internal edge of the narrow region. Collectively, these data lead to a surprising picture for the structural role of CFTR's TM12: the middle section of TM12 (i.e., S1141) is located two helical turns cytoplasmically from the proposed inner brim of the narrow segment of the pore.
Before we wrap up our discussion, a few subjects should be revisited and/or inspected in a more critical manner so that the readers are fully aware of the potential caveats in our approach. First, although the marker used in the current study to judge the formation of a tight Cd 2+ -binding site is the sluggishness of current recovery after washout of Cd 2+ , a fast recovery from the Cd 2+ -inhibited state for several double mutant currents (e.g., 102C/345C) may not necessarily indicate a physical alienation between the two cysteines involved, especially when a higher inhibition for double mutant was observed, as in the case of 102C/345C. One possible scenario is that even though the two cysteines in the double mutant are close to each other, a lack of support from the aforementioned unidentified partner(s) could render the Cd 2+ -binding site as weak as undetectable in our experiments.
Second, we need to point out that our measurements of Cd 2+ -induced current reductions (summarized in Table 1 for the positive pairs) and current recovery afterward with or without DTT do not tell whether the effect of Cd 2+ is on gating or ion conduction, or both. Because all of the residues examined in the current study have been shown to line the pore, we prefer a simple pore-plugging effect. Of note, this distinction becomes murkier when dealing with positions close to the narrow region of the pore because of a possible dual function for this segment (Gao and Hwang, 2015). However, even if the effect of Cd 2+ in some of our mutants is attributed to gating perturbation, the identification of positive pairs still attests to a physical proximity for the introduced cysteine pair during certain states of the gating cycle. On the other hand, because Cd 2+ was always applied in the presence of ATP for an extended period of time during which the channel can virtually visit all possible gating states, negative results obtained in our experiments suggest that the two residues under study are not close at any given point in a gating cycle, thus excluding certain structural configurations for the channel. Moreover, as many of our double mutant constructs exhibit extremely slow apparent off rates of Cd 2+ , we need to be concerned by the possibility that this strong force may actually allow the formation of the inferred metal bridge in conformations in fact rarely visited during normal gating transitions. Nevertheless, this worse scenario is probably avoided as most of the inhibition occurred within seconds of Cd 2+ applications and the gating cycle of CFTR is ∼1 s −1 .
Third, except for the 348C/1141C pair, most of the double mutants were not able to generate large enough currents in inside-out patches for us to test state-dependent effects of Cd 2+ . However, we do realize that the state-dependent data could shed more light as channel gating is a dynamic process. Therefore, whenever possible, we also directed our efforts to provide our audience with a sliver of this kind of structural insights. For example, the second order rate constant for Cd 2+ reactivity with 348C/1141C mutant in the absence of ATP is approximately threefold higher than that with ATP ( Fig. 9 and Fig. S9), suggesting these two cysteines may be either in a more accessible location or better orientated to coordinate Cd 2+ in the closed state. Besides, we also observed, in our whole-cell recordings of 106C/337C channels, a biphasic current response to forskolin application: current increase followed by a decrease, which could suggest that these two cysteines are close to each other only in the open state of CFTR (see Fig. S3 legend for more details).
Bearing nearly all of the topological hallmarks of an ABC exporter, CFTR is a bona fide member of this superfamily, as previous phylogenetic study has suggested (Dean and Annilo, 2005). Many functional data on CF-TR's NBDs and TMDs also support the hypothesis that CFTR evolved from an ABC exporter by degenerating its intracellular gate (Bai et al., 2011). We noticed that, because of CFTR's kinship with ABC transporters, nearly all homology models of CFTR were built using the crystal structures of ABC exporters as templates (Serohijos et al., 2008;Dalton et al., 2012;Norimatsu et al., 2012;Rahman et al., 2013;Mornon et al., 2015). Unsur- prisingly, although many models are able to account for those identified pore-lining components in individual TMs, few could achieve a satisfaction from all examining eyes, partly because of their confinement to the twofold pseudo-symmetry inherited in ABC transporters (e.g., the one in Fig. 1). Therefore, by providing some physical constraints between well-defined, paired positions in different TMs, data reported here could provide an effective guide for future computational modeling.
|
2018-04-03T00:47:11.388Z
|
2016-05-01T00:00:00.000
|
{
"year": 2016,
"sha1": "20972940c80447436c532027c011ada4a1ea7780",
"oa_license": "CCBYNCSA",
"oa_url": "https://rupress.org/jgp/article-pdf/147/5/407/1232585/jgp_201511557.pdf",
"oa_status": "BRONZE",
"pdf_src": "PubMedCentral",
"pdf_hash": "20972940c80447436c532027c011ada4a1ea7780",
"s2fieldsofstudy": [
"Biology",
"Chemistry"
],
"extfieldsofstudy": [
"Chemistry",
"Medicine"
]
}
|
259673374
|
pes2o/s2orc
|
v3-fos-license
|
Wounds Care on Patients During The Covid-19 Pandemic Era
Abstract
Introduction
One of the threats of COVID-19 in the world is the disruption of wound care practices, especially chronic wounds. In March 2020, the surgeon general urged all healthcare facilities across the US to stop elective procedures and suspend all non-essential treatment. This is a way to prevent transmission of the corona virus and prevent the number of people in the hospital. Furthermore, patients with various medical comorbidities require disciplinary consultation from health workers, such as diabetes control, which is not as frequent as during the COVID-19 pandemic. Of course, given the difficulty of carefully observing and actually treating wounds during the COVID-19 pandemic, wound care specialists quickly devised new treatment models to ensure the continued success of managing vulnerable outpatients, as well as managing patients with pressure ulcers (1). Chronic wounds are generally described as barrier deficiencies that have not healed the wound for up to three months (2).
Chronic wounds are wounds in which the normal healing process has been restricted at one or more points in the hemostatic, inflammatory, proliferative, and remodeling phases. In this type of wound, there is generally an underlying pathology, which results in slow healing (3). Chronic wounds include diabetic foot ulcers, venous leg ulcers, and pressure ulcers. Chronic wounds are a challenge for wound care professionals and consume a lot of health resources around the world (4). It affects millions of patients worldwide, placing a great deal of responsibility on health resources in performing wound care (5). In the United States, chronic wounds affect more than 6 million people, with more being anticipated due to an aging population and the high prevalence of patients with diabetes mellitus (6).
Chronic complex wounds are a primary health problem, with serious repercussions for the people affected, the care environment and the health care system. Approaching them requires an integrated system of care where the complete course of the discus counts for greater effectiveness (7). Effective chronic wound management is complex, and to maximize outcomes for patients, it is recommended that those involved in care and treatment should have the appropriate knowledge and skills (8).
The spread of the coronavirus has severely impacted medical practice behavior, resulting in a reduction in required medical care, including specialization in wound healing. The COVID-19 pandemic has limited the ability of wounds to heal under normal conditions. Patients with vascular foot ulcers are a highly vulnerable population, with a poor quality of life caused by pain that is directly related to ulcer duration and ulcer size. If ischemic wounds and veins are not treated or managed, the consequences can be drastic, such as infection, sepsis, amputation, or even death (9). World wound care is experiencing greater disruption causing the need for wound evaluation (1).
In this systematic review, an overview will be given of nursing management in carrying out wound care during the corona virus disease. The aim of this systematic review was to evaluate 14 nursing management procedures for treating wounds in patients with chronic wounds during COVID-19.
Search Strategy
Searching for articles in this systematic review uses an international electronic database consisting of EBSCO, ProQuest, and PubMed with years of publication from 2019-2021, and articles in English and full text. The search strategy carried out in this systematic review uses several keywords used in the search in the database used. Keywords refer to Medical Subject Heading (MeSH). The search terms used were: nursing care, nursing management, wound, COVID-19, Corona Virus Disease 2019. The following search strategies were used: nursing care OR nursing management AND wounds OR injuries AND Covid-19 OR coronavirus disease 2019. These are the keywords then entered in the search box in the electronic database and then filtered according to the following criteria: full text, issues ranging from 2019 -2021, and using English.
Selection Criteria
The article reviews systematic analysis of standards using the PICO model consisting of population, interventions, comparisons, and results in determining inclusion and exclusion criteria. Articles that will be included in the study if these specific inclusion criteria are: (1) the study focuses on nurses who practice chronic wound care in patients during Covid-19, (2) nursing management, (3) case studies, (4) years publication: 2019-2021, (5) full text, (6) articles in English. Exclusion criteria: (1) acute wound care (2) not discussing nursing management, (3) articles published before 2019, (4) full text (5) articles not in English.
Quality Assessment
The assessment of research methodology for the quality of systematic review articles uses 18 JBI Critical Appraisal Checklist guidelines. The instrument used is the JBI Critical Appraisal Checklist for case series which consists of 10 questions (10). The JBI Critical Assessment Checklist was used for the purpose of assessing the quality of research methodology and to determine the extent to which research has addressed possible bias in its design, intervention, and analysis (10).
Study Selection
Researcher uses a method in the form of a systematic review with descriptive narrative analysis of several main finding articles that discuss nursing management in treating wounds to patients during COVID-19. Guidelines for systematic review using Preferred Reporting Items for Systematic Review and Meta-Analyzes (PRISMA) and as a standard in reviewing and selecting articles. Specific information was extracted such as author, year of publication, journal name, study design, type of intervention, and results.
Result and Discussion
Based on the article search flow chart or PRISMA diagram, identification stage, 242 articles were obtained from several databases used with the following details from EBSCO 2 articles, from ProQuest 142 articles, from PubMed 98 articles. At the screening stage there was a reduction in the number of articles due to the fact that there were 12 of the same articles, so that at the screening stage there were 230 articles that would be selected based on the title and abstract of the article. After being selected, 195 articles released at this stage did not meet the inclusion criteria. At the feasibility stage, there are 40 articles that will be selected full text. At this stage, 35 articles were published because they were not related to nursing management, the research did not use English, the research was published under 2019. In the final stage, 5 articles were obtained based on selection carried out according to inclusion criteria.
The World Health Organization (WHO) describes telemedicine (TM) as "one of the services in providing health services to patients, where distance is a factor of resistance, by every health care professional using information and communication technology to exchange valid information for diagnosis, treatment and prevention." disease and injury, research and evaluation, and continuing education of health care providers, all to promote individual and community health (16).
Telehealth has the prospect of addressing many of the key challenges in healthcare delivery during the COVID-19 outbreak. Telehealth has prospects for addressing many of the key challenges in healthcare delivery during the COVID-19 outbreak. In addition, telehealth can help us stay away from direct physical contact and reduce the risk of COVID transmission and ultimately provide sustainable care to the community. Health care and patients are forced to put forward telehealth tools as a problem solution to prevent and control COVID-19 infection (17). Telehealth is used as the default consultation method to prevent patients from reducing the risk of spreading disease. The current model implemented as part of the emergency response relies on the patient's ability to discuss the injury and its symptoms in order to reduce close contact and actions outside the home. The use of telehealth has a positive impact on wound healing, including a shorter healing time and positive consumer feedback. All of the systems studied used telehealth communication between the wound specialist and the nurse present with the patient (18). The current standard model of chronic wound telecare such as diabetic foot ulcers includes in-hospital specialists who perform remote clinical examinations and decision-making, working closely with visiting nurses and patients. Wound care is a visual specialty. Numbers are the gold standard for diagnosis and treatment. Chronic wounds that condition long-term care require frequent monitoring and frequent follow-up visits. This involves traveling costs and long waiting times. This situation makes the wound care specialization an ideal specialty for telemedicine applications for wound care (16). Telehealth and telemedicine services can provide opportunities to improve nursing care for patients during COVID-19.
Conclusion
The findings of this systematic review study are that telehealth and telemedicine services can provide opportunities to improve wound care for patients.
|
2023-07-12T07:53:05.194Z
|
2023-04-22T00:00:00.000
|
{
"year": 2023,
"sha1": "4e945b80bd6f403c88681af0b89a382a553d733b",
"oa_license": "CCBYSA",
"oa_url": "https://ejournal.publine.org/index.php/mohr/article/download/9/10",
"oa_status": "HYBRID",
"pdf_src": "Anansi",
"pdf_hash": "e9b43ac6f7c9b0bf2faed16efe98c765c8d1d494",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": []
}
|
235454122
|
pes2o/s2orc
|
v3-fos-license
|
Complements or substitutes? Associations between volumes of care provided in the community and hospitals
Policymakers often suggest that expansion of care in community settings may ease increasing pressures on hospital services. Substitution may lower overall health system costs, but complementarity due to previously unidentified needs might raise them. We used new national data on community and primary medical care services in England to undertake system-level analyses of whether activity in the community acts as a complement or a substitute for activity provided in hospitals. We used two-way fixed effects regression to relate monthly counts of community care and primary medical care contacts to emergency department attendances, outpatient visits and admissions for 242 hospitals between November 2017 and September 2019. We then used national unit costs to estimate the effects of increasing community activity on overall system expenditure. The findings show community care contacts to be weak substitutes with all types of hospital activity and primary care contacts are weak substitutes for emergency hospital attendances and admissions. Our estimates ranged from 28 [95% CI 21, 45] to 517 [95% CI 291, 7265] community care contacts and from 34 [95% CI 17, 1283] to 1655 [95% CI − 1995, 70,145] GP appointments to reduce one hospital service visit. Primary care and planned hospital services are complements. Increases in community services and primary care activity are both associated with increased overall system expenditure of £34 [95% CI £156, £54] per visit for community care and £41 [95% CI £78, £74] per appointment in general practice. Expansion of community-based services may not generate reductions in hospital activity and expenditure. Supplementary Information The online version contains supplementary material available at 10.1007/s10198-021-01329-6.
Introduction
In many countries, substantial proportions of health care services are provided in the community rather than at hospitals. These services may be the first point of contact for individuals to the healthcare system and include a wide variety of roles such as "primary care physicians,…, nurses, pharmacists, auxiliaries and community health workers" [1]. Policymakers often assume that primary care strengthening and the improvement and expansion of care in community settings will ease increasing pressures on hospital services. For example, the recent national strategy for the health service in England described a reorganisation of how care services were to be delivered by introducing new care models [2]. These were intended to increase the integration between health care providers, ultimately providing better quality care for patients and moving away from specialist hospital care by providing more care in community settings.
It is believed that with the current ageing demographics and increased disease burden, the reliance on hospital care will be financially unsustainable in the long term [3] because it is more costly for individuals to be treated in hospitals than in the community. Therefore, the belief that health care 1 3 services delivered in the community can substitute for hospital services is attractive to the health system, as lowering hospital service use is expected to reduce overall health care costs.
The overall relationship between community and hospital activity is complex and will depend on the health condition and the type of care required. There is a straightforward complementary relationship between first primary care physician (GP) appointments and elective hospital care in healthcare systems where GPs act as gatekeepers for nonemergency hospital services. There is also a straightforward substitution relationship where community and hospital services offer the same intervention and patients can choose between them based on the differences in availability. This is likely to apply, for example, where patients need urgent advice or minor treatments for relatively simple acute health problems.
There are some health conditions requiring specialised equipment and care by specialists that will require treatment in hospitals. However, more generally, some community service activities could be either complements or substitutes for non-emergency hospital services if they help manage conditions outside of the hospital setting. Such activities would include the management of wound care or patients with multi-morbidities. These activities could avoid patients experiencing worsening health conditions and therefore mitigate needs for non-emergency hospital services. However, improvements in community-based care may have complementary relationships with hospital services by uncovering previously unidentified needs. This would mean that increases in the provision of services in the community may lead to increases in hospital service use. These system-level relationships are critical to the economic case for expanding services in the community and understanding the impact on population health and wellbeing.
The mechanisms by which community-based services influence hospital services will differ between planned and unplanned reasons for care requirements. The previous literature on the relationship between primary care and planned secondary care provides mixed results. A systematic review that studied the substitution of outpatient hospital care with specialist care in a primary care setting found that most (11 out of 14) studies report a substitution effect between specialist care within a primary care setting and planned outpatient hospital care [4]. We found studies outside of the scope of this systematic review that also provide mixed findings. One study from the United States of America (USA) found a substitute effect between primary care and speciality medical encounters [5]. Another USA study found no effect of increasing primary care fees, which should reduce use, on hospital admissions and outpatient visits [6]. A study from the Netherlands found no effect of expanding services performed in primary care on GP referral behaviour [7]. A study from Norway found a weak substitution effect between longterm primary care and hospital use among older patients. Still, the reduction did not significantly relieve pressures on the hospital system [8]. Another Norway study found a complementary relationship between GP consultations and outpatient visits rates [9]. These conflicting findings from the literature may be driven by the differences in the identification and type of both primary and secondary care and differences in the institutional settings for each study.
Several studies have examined elements of these relationships between hospital-and community-based care. We summarise their main findings below. However, we are not aware of studies that have examined the effects at system level, particularly the relationships between volumes of contacts with community health care workers and the use of hospital services in any country.
This paper aims to estimate the overall effect of changes in the provision of services in community settings on hospital care in England. The availability of two new datasets on care provided in community settings within England enables us to assess empirically whether community activity acts as a complement or a substitute for hospital services. We use these newly-available national datasets to determine the system-level association between the volume of services delivered in the community (including community health services and appointments in General Practice) and volumes of hospital service activity in England. We then use the data on unit costs to estimate the effect of increasing community activity on total system expenditure.
Methods
Before describing the datasets and how we combined and analysed them, we describe the basic institutional arrangements for health services in England and review previous studies of elements of the relationship between hospital-and community-based care.
Institutional background
Publicly funded health care through the National Health Service (NHS) in England is generally free at the point of use and funded through general taxation. Health care services are rationed based on the health needs.
NHS England allocates a proportion of its health care services budgets to one of 135 (as of March 2021) Clinical Commissioning Groups (CCGs) who negotiate contracts for NHS health care services at a smaller regional level This allocation is based on the health care needs of the population. The population's health care needs will differ by region due to different population characteristics such as average age, deprivation and rurality.
CCGs negotiate contracts with NHS Trusts to provide hospital and or community services and General Practices for primary care services. Public Health England allocates public health budgets through 152 Local Authorities in England. These Local Authorities negotiate contracts with service providers for public health services, including some community services.
For planned care, England's health care system is structured around a gatekeeper system in which people register with a family doctor, called General Practitioner. General Practitioners are trained to identify and meet most health care needs and are responsible for, but not limited to, the management of long-term health conditions and referring patients to other health care services in line with their care needs. General Practitioners are usually the first point of contact for patients, where the care they provide is under the umbrella term of primary care. Access to specialist care delivered by hospital specialists, secondary care, is generally controlled by General Practitioners. If a General Practitioner determines that a patient will benefit from receiving specialist care delivered in a hospital, a patient can be referred for an outpatient consultation. First, outpatient consultations are made with hospital specialists responsible for hospital treatments and do not necessarily result in treatments. If treatment for a patient's health condition requires hospital admission, hospital specialists will recommend that admission.
Definition of community services
Community care services are diverse and challenging to define as the precise nature of services differs between health systems, between regions and between population groups. A recent definition states "[t]he main types of services delivered in the community include, but are not limited to: adult community services (e.g. district nursing, intermediate care, end of life care), specialist long term condition nursing (e.g. heart failure, diabetes, cancer) planned community services (e.g. podiatry, speech and language therapy, physiotherapy) children's 0-19 services (e.g. health visitors and school nursing) health and wellbeing services (e.g. sexual health, smoking cessation, weight management) and inpatient community services (e.g. inpatient services)" [3]. Broader definitions of services provided in the community also include care delivered within General Practice [10].
Literature review focusing on the English health care system.
For unplanned care, patients can directly access hospital care through Accident and Emergency (A&E) departments without prior access to a General Practitioner. We would expect that patients attending accident and emergency services would present with severe health conditions. However, some patients may attend accident and emergency departments due to perceived or actual lack of access to a General Practitioner. A systematic review conducted by Flores-Mateo et al. [11] found that increasing primary care supply is associated with reductions in A&E attendances. Studies by Lippi Bruni, Mammi and Ugolini [12] and Whittaker et al. [13] have examined the impact of extending access to general practices on the use of emergency hospital services. Both studies found that increases in access led to reductions in emergency hospital service use. However, more recently, Parkinson et al. [14] examined the association of avoidable A&E attendances to primary care quality in England. The authors found that increases in primary care quality (and availability) are associated with reductions in avoidable A&E attendances. In addition, one study in the USA setting found that increasing fees at primary care services was not associated with emergency hospital service use [6].
Harrison et al. [15] examined the effect of a financial incentive scheme to improve the quality of primary care in England on avoidable emergency hospital admissions. They found that avoidable hospital admissions reduced more for conditions targeted under the financial incentive scheme by more than incentivised conditions. A systematic review study by Bickerdike et al. [16] reviewed studies on interventions linking primary care and care in the community through social prescribing. One element of the review examined the impact of the intervention on secondary care and found that two out of three studies found social prescribing reduced secondary care activity. However, the studies had weak designs or small samples. A subsequent study by Munford et al. [17] found reductions in secondary care utilisation (outpatient visits, ambulance call outs and A&E attendance) between individuals who participated in community assets than those who did not; these findings were not subject to regression analyses. Their follow-up [18] also found that individuals participating in community assets have lower health care costs than individuals who do not.
A systematic review by Baxter et al. [19] reviewed the literature of integrated care within and outside of the UK, where one of the three focuses was on the usage of health care resources. The authors found that five UK studies found that integrated care is associated with reductions in outpatient appointments, whereas no international studies found a relationship. The authors found the relationship between integrated care and unscheduled care, the number of hospital admissions and A&E visits-to be mixed. The majority of studies reviewed reporting either reductions in hospital services or no effects. Morciano et al. [20] examined new integrated care models in England on emergency hospital admissions using a quasi-experimental design for policy evaluation, difference-in-differences. The authors found that the new integrated care models were associated with reductions in emergency admissions, driven by individuals within residing care homes two years post-intervention.
We have identified no evidence specific to community care services, but the literature on social care services suggests community care services could be a weak substitute for hospital services. A systematic review of the association between social care and hospital care conducted by Spiers et al. [21] found no evidence on the number of social care users and A&E attendances in the UK [22] and weak substitute relationships between users of care homes and hospital admissions [23,24] and A&E attendances in the USA [23]. However, a study by Forder [25] assessed the relationship between long term care and hospital utilisation for older patients, finding them to be substitutes, with a £1 spend on social care leading to a £0.35 reduction in hospital expenditure. More recently, Crawford, Stoye and Zaranko [26] found that reductions in social care spending for people aged 65 and above have led to an increase in A&E utilisation. There is no direct evidence on community care services, but the literature on social care services suggests community care services could be a weak substitute for hospital services.
Data
We used three primary datasets in our analysis, covering the 23 months between November 2017 and September 2019: The publicly available monthly series of hospital activity from provider-level HES captures counts of all publiclyfunded hospital activity; hospital admissions, unplanned hospital admissions, outpatient visits and emergency department attendances. Providers include Trusts, Foundation Trusts and independent service providers. In total, the data contains 242 different hospital service providers.
The Appointments in General Practice data series captures all scheduled appointments from General Practice that use one of four computer systems (EMIS, TPP, MICRO-TEST and VISION). A study in 2018 by Kontopantelis et al. [30] found that of 7526 General Practices in 2016, 7477 (99%) of practices used these computer systems. Counts of appointments in General Practice are stratified by health care professional type (GP, other practice staff and unknown), mode of appointment (telephone, face to face, video, online, home visit and unknown), attendance status (attended, not attended and unknown) and the length of time between the booking date and the appointment date.
NHS Digital reports the number of General Practices who did not report appointments data each month in each CCG and the registered population sizes of those practices. We inflate the count of appointments to account for the expected amount of un-reported activity using the registered population sizes. We used the list of 191 CCGs that existed as of April 2019. We combined the earlier data from the CCGs that merged using information regarding CCG mergers from the Organisation Data Service [31].
The Community Services Data Set is collected on all publicly funded community services in England. Providers of publicly-funded community services are legally mandated to supply this information. Providers include Foundation Trusts, Acute Trusts, Mental Health Trusts, Community Health Care Trusts, Care Trusts, social enterprises, Integrated Care Organisations, independent sector providers and local authorities [28]. A total of 124 community service providers submitted community care contacts data to the CSDS within the study period. We use data on care contacts from the CSDS.
In each dataset, we calculated a relative activity index for each provider. We divided the count of activity reported by each provider in each month by the average activity reported by that provider across all months. This normalises the activity level for each provider to one. We include the definition and the data required to generate the relative activity indexes in Appendix Table A1. The relative activity indexes are calculated by dividing activity levels in each provider month by the provider's mean activity level across all months and all these are shown in detail in figures.
Combining the datasets
Each of the three activity datasets is available at a different level of aggregation. Hospital activity is published at the hospital Trust level, including independent sector providers that provide publicly funded care. Appointments in General Practice is published at the CCG level. Community health services activity is published for each community health care provider.
We adopted the hospital Trust as the unit of analysis as we sought to explain variations in hospital activity. There is no direct mapping of primary care providers and community care providers with hospital care providers. Therefore, we used two methods to assign General Practice and community care providers activities to each hospital provider.
Micro-data are available from HES on the hospitals used by populations registered with the General Practice in each CCG. We used the most recent micro-data available (April 2017 to March 2018) to create the weights for the primary care relative activity levels. We calculated the shares of each hospital's activity that came from patients registered with practices in each CCG and used these shares to weight the relative activity levels for each CCG using Eq. 1.
where app ht denotes the attributed relative index of appointments in General Practice for hospital provider h and month t, H denotes a historical attribution of the proportion of patients at hospital provider h from General Practice p and A denotes the number of appointments at a General Practice.
For example, if 75% of a hospital Trust's patients were registered with practices belonging to a CCG whose appointments activity was elevated by 10% and the remaining 25% of patients were registered with practices belonging to a CCG which reported its average level of appointments activity, the appointments activity index for the hospital Trust would be 1.075 = (75% × 1.1) + (25% × 1.0). The indexed value of 1.075 can then be interpreted as 7.5% more than the average number of GP appointments. There is no national micro-data to apply the same approach for community services activity levels. Owing to the diverse nature of community care services, the commissioning and geographical footprints for each care provider are unclear and overlapping. From a survey of 71 acute trusts providing community services, on average, community trusts were commissioned by over five different organisations, with Clinical Commissioning Groups (CCGs) accounting for 77%, NHS England 5% and local authorities 17% of the overall community services budget [3,32]. A single community service provider will be responsible for providing care to a geographical area. However, some geographical locations will have one provider providing most community care activity and multiple smaller providers providing more specific community services [10]. CCGs hold, on average, 50 different contracts for community health services [32]. We created weights based on the distances between providers and their relative sizes [33]. The distance-size weight takes values between zero and one for each possible combination of community service provider and hospital service provider.
We use the following series of Eqs. (2)(3)(4)(5) to generate a distance-size weight between each community and hospital service provider and Eq. 6 uses the distance-size weights to create a hospital-level community activity index: in which, DW ph is a set of distance weights based on the distance between community provider p and hospital provider h ( d ph ). SW is a set of size weights calculated using the relative size of a community provider compared to all community service providers based on C , the average monthly volume of community care contacts provided by community provider p. GW are the distance-size weights, generated by taking the product of the distance and size weights. GW * are the normalised distance-size weights that ensure the weights assigned to each hospital service provider sum to one. The index of relative community services activity attached to hospital h in month t is then the weighted average of the indices of relative community services activities for each community provider in month t. We present a hypothetical example of how the distance-size weights are calculated in the appendix Table A2.
Some organisations will appear in both p and h because they provide both hospital and community services. The importance of distance to the weights depends on the power of the decay function in Eq. (2). We use a distance power of two and conduct sensitivity analysis adjusting for different powers of distance. Further details on how the weights were calculated are provided in the appendix. We used all pairwise combinations of hospitals and community services providers when calculating the weights to be used in analysing influences on the volume of emergency department attendances and non-elective hospital admission. This is because patients could use emergency departments anywhere in the country. This situation is much less likely to happen for outpatient visits and elective hospital admissions. Therefore, we calculated weights for planned care using the five community service providers closest to each hospital Trust. All provider location information was obtained through the NHS Digital Organisation Data Service [31] Control variables, including figures on population size, age distribution and the proportion of the population that is male from monthly series of registered patients at each General Practice, were obtained from NHS Digital [34]. Data were aggregated to the CCG level. As the population size measure, a relative population size index was calculated by dividing the monthly figures for each CCG by the CCG average. Then these covariates were attached to hospital service providers using the same weights as for the indices of relative primary care activity levels.
Estimation strategy
We set out to estimate the following linear equation: where hosp ht denotes the index of the relative volume of activity at hospital service provider h in month t , com ht is the index of the relative volume of community services attached to hospital h in month t, app ht is the index of the relative volume of appointments in General Practice linked to hospital h in month t, X is a vector of explanatory variables (index of relative population size, proportion of the population that is male, proportion of the population aged 0-14 years, proportion of the population aged 15-64 years and proportion of the population aged 65 years and over), are a set of monthly time dummy variables and is the error term with a mean of zero. Positive values of 1 and 2 would indicate complementary relationships between activity in the community and hospitals, whereas negative values would indicate substitute relationships. All are estimated with robust standard errors.
We allow for time-varying unobservable factors common to all hospitals and unobservable factors specific to each hospital fixed over time. However, there may be time-varying unobservable factors that vary across hospitals and may bias the estimated relationships. One such variable that was initially explored is the income deprivation of individuals in (7) hosp ht = 0 + 1 com ht + 2 app ht + X ht + t + ht , the local area. We did not include this measure because it showed almost no changes throughout the study period and was therefore not suitable for inclusion in the fixed effects estimation.
We use Oster bounds to assess the likely importance of these unobservable factors [35]. This analysis adjusts the estimated coefficients from the model accounting for the R-squared and potential direction of bias due to not including all possible confounding variables.
The adjustment is provided by following formula: where β * is the bias-corrected coefficient, ⌣ is the coefficient from the full model, ̇ is the coefficient from the model without covariates, ⌣ R is the R-squared statistic for the full model, Ṙ is the R-squared statistic from the model without covariates, R max is the maximum obtainable R-squared statistic and represents the importance of the unexplained confounders compared to the explained confounders.
We make three assumptions when adjusting the coefficients using this method. First, we assume that the maximum obtainable R-squared is equal to 1.3 times the R-squared value from the full model. Second, we assume that maximum explanatory power using all unobserved confounders is equal to that with the observed confounders. Third, we assume that the direction of bias when including observed confounders and unobserved confounders is the same.
Costing
Healthcare supplied by the English NHS is funded through general taxation. This publically funded budget will be used to fund both activities in the community and activity provided in hospitals. With finite budgets, policymakers in England intend to move more health care services to be delivered in the community, which are considered cheaper than hospital services [2]. Net unit costs for services provided in the community is determined by the relationship between services provided in the community and hospital services. We would expect that net unit cost is lower than the unit costs of the community's service if there is a substitution effect with hospital services. The direct costs of community services will be (partially) offset by the indirect cost savings from hospital activity reductions.
We estimated the net unit costs associated with each care contact provided within the community net of any expected reductions (or increases) in hospital activity using the estimated regression coefficients, mean volumes of activity and unit costs from NHS England and Improvement for hospital and community services [36] and appointments in General Practice [37]. Unit costs for hospital and community services are the aggregated costs from the National Cost Collection for 2019, formally known as NHS Reference Costs. Unit cost for an appointment in General Practice was obtained from an NHS England article published in 2019, which divided the total cost to of General Practice to the NHS by the number of General Practice appointments. We estimated the net unit cost of activity provided in the community using the following equation.
where ΔNet Cost i denotes the net unit cost for each community service i (community care contact or appointment in General Practice), x denotes the mean number of contacts in the community per month, c is the unit cost of activity, hosp j denotes the mean hospital activity per hospital provider per month for hospital activity j and ̂ is an estimated coefficient from Eq. (7).
Additional analyses
We performed a set of additional analysis to test whether the associations between care provided in the community and hospital services were sensitive to different assumptions and situations. Firstly, community services delivery-and therefore, associations with hospital services-may differ in rural and urban settings. This may be due to the increased scope of community services within rural areas compared to urban areas. We stratified our sample by the size and proportion of each hospital's treated population who lived in rural areas based on the historical hospital activity.
Secondly, we limited the sample to hospital service providers that also provide community care services. Mainly because the attribution of community activity to the hospital is more specific, but we may expect the relationship between community care and hospital services to be more robust when provided by the same care organisation.
Thirdly, we limited the analysis to include NHS Trusts and Foundation Trusts only. All hospital services' providers that deliver publicly-funded activity are included in the data. Independent hospital service providers generally provide much lower publicly-funded hospital activity levels than NHS Trust and Foundation Trusts. As we use relative activity levels, smaller providers will have large fluctuations in the relative activity index following small actual volume changes.
Fourthly, as community care data are a relatively new dataset, there were initial data quality issues where providers did not submit data through all months included in our study period. In our core analysis, we did not include providers with missing data. As a sensitivity analysis, if a community provider did not report data in a particular month, we linearly interpolated the missing counts using the counts in the preceding and proceeding month. We did not extrapolate data before a provider first reports or after a provider last reports in the dataset. Fifthly, we tested the effect of changing the distance decay function when attributing community services to hospital service providers. We used a linear function of distance and then changed the power on the decay function to examine whether our results were sensitive to the distancesize power choices. The effect of increasing the distance decay function meant more weight was applied to the closest community care providers to the hospital service provider. Thus, changing the decay function altered the amount of community care services attributable to each hospital service provider.
Finally, we estimate the relationship between hospital activity and community services activity taking natural logarithms of the levels of activity in hospitals and in the community. This is to mitigate the potential effects of large outliers and the positive skew in the relative activity indices. Table 1 provides descriptive statistics on the measures of health care activity and the covariates. On average, emergency departments have 10,500 attendances each month. Hospitals, on average, have 6500 hospital admissions (of which 3300 are non-elective) and 8100 first outpatient visits each month. On average, community care providers have 47,600 care contacts each month. On average, a CCG has 135,000 appointments in General Practice each month.
Descriptive statistics
A&E providers also have the lowest between provider standard deviations in relation to the mean of all other service providers. With the largest standard deviation to mean ratio (and between provider standard deviation to mean ratio), first outpatient providers are more varied than CCGs and hospital service providers. Appointments in General Practice series are the most stable relative to each provider's mean with the lowest within provider standard deviation to mean ratio. A&E providers and Community care providers exhibit the largest within provider variations given their respective means.
Table 2 Regression results
The results from linear regression models. Models also include provider fixed effects, monthly time effects. Indexes of relative activity are calculated by dividing monthly volume by the average volume reported by the provider over all months. All regressions are estimated with robust standard errors. Estimated effect sizes are the number of contacts or appointments in the community that would change the use of hospital service by one unit. We used the coefficients and mean activity levels to generate the estimated effect sizes for the predicted association between hospital activity and activity in the community Table 3 shows the net unit cost for community services is £34.16, which is lower than a unit cost of a community care contact of £64 [36]. We estimated that each additional community care contact lowers the expected cost of hospital services on the NHS by around £30; this is mainly driven by the expected reduction for inpatient hospital admission (£17). The net unit cost of an additional appointment in General Practice is £41. Due to the complementary nature of appointment in General practice and hospital services (hospital admission and first outpatient attendances), each additional appointment in General Practice is expected to increase the cost of hospital services to the NHS by around £11. However, because the estimated relationship between hospital services and appointments in General Practice is estimated with considerable uncertainty, the estimated net system cost has a wide 95% confidence range of between £8 and £74.
Bias adjusted coefficients
For all of the relationships estimated to be substitutes, the effect of adjusting for a greater set of confounders would strengthen the substitution effect (Table 4). This suggests that the substitution effects shown by the analysis are the lower bounds. The two exceptions are the substitution effects for A&E and inpatient attendance with community care; however, the estimated coefficients' changes are small.
The estimated relationship between appointments in General Practice and total inpatient admissions and outpatient visits are not stable. As the importance of unexplained variation increases, we find that the complementary relationship becomes a substitute; however, the main findings are weak with statistical significance at the 5% level. Table 5 shows results from four additional analyses focusing on rurality, same community and hospital provider, NHS providers and interpolating data. The complementary and substitute relationships between services provided in the community and hospital services are generally larger for hospital providers serving relatively rural populations than urban populations, except for A&E activities. When comparing the results from the main analysis with the results from using only providers that provide both hospital and community care, we find that the substitution effect is larger between care provided in the community and A&E activity. When imputing missing community provider data for months where providers did not submit data, estimates for GP appointments are relatively constant, but associations with community care and hospital services weaken. The results using NHS hospital providers only are consistent and similar to the main analysis. Table 6 shows the sensitivity analysis results on exploring the use of different distance-size powers on the distance decay function DW ph . We find that the results are largely robust to changing the distance function. Community service contacts are all estimated to have a substitution effect with all hospital services, the substitution being strongest between community contacts and A&E attendances activity and weakest between community activity and outpatient visits. Generally, increases in the decay function's power lead to reductions in the association between community and hospital services.
Additional analyses
The results from models using logged activity measures are included in Appendix Table A4. We find that community care contact and hospital services are substitutes, although only the effect of A&E activity is statistically significant at the 95% level. Similar to the main results, GP appointments are substitutes for emergency hospital activities and complements for elective hospital admissions. GP appointments are also complements for first outpatient visits, a relationship that is statistically significant at the 95% level.
Conclusion
Utilising new national data, we examined the frequently suggested notion that community activity acts as a complement or a substitute for hospital services. We find that there is a weak substitution effect between community care contacts and all hospital services, with an additional community care contact (direct cost of £64) associated with a decrease in hospital costs to the NHS of £30 [95% CI £10, £48]. This weak substitution means that each additional community care contact is expected to increase net system costs by £34 [95% CI £6, £54]. We find the substitution effects are weak in magnitude, with 28 [95% CI 21, 45] community care contacts associated with a reduction of one A&E attendance and 517 [95% CI 291, 7265] community care contacts associated with a reduction of one non-elective admission. However, we are unable to estimate whether these initial increase in costs are due to meeting unmet needs or result in lower longterm health costs by potentially identifying the health needs of the population within the community. All our findings are robust to changes in the distance decay function when attributing community service to hospitals.
Appointments in General Practice and A&E attendances (34 [95% CI 17, 1283] GP appointments are associated with a reduction of one A&E attendance) and non-elective hospital admissions (1655 [95% CI − 1995, 70145] GP appointments are associated with a reduction of one non-elective admission) have a substitution relationship. Appointments in General Practice and total hospital admission and first outpatient visits are weak complements (the largest relationship Table 2) that have been adjusted in line the Oster bias adjustment. Movements in coefficients and r-squared values are used to predict the direction and size of bias of not adjusting for all possible confounders. We calculated how our estimated coefficients will change when increasing the importance of the unexplained variation compared to explained variation from 0.1 (one tenth) to 1 (equal)
Importance of unexplained variation
Index of relative GP appointments activity Index of relative community care activity The substitutability between community care and hospital services is stronger for providers that serve higher levels of rural populations compared to urban populations except for A&E. This may be due to the access to services available to the population and therefore the relative importance of such health care services in the community for the population in rural settings. With A&E services not being determined by a gatekeeper, the lack of availability of A&E in rural settings will render the service a weaker substitute.
These findings imply that increases in community services could relieve pressures on the volume of care delivery for hospitals. Suppose the assumption that increases in community care delivery could identify greater unmet need levels are correct. In that case, these findings suggest that the increases in community care provision are associated with lowering hospital activity despite providing care to patients with newly identified needs. The substitution between services provided in the community and A&E services are shown to have the largest substitution effects indicating that access to care services is more of an issue for people attending A&E services.
We would expect appointments in General Practice and hospital admissions and first outpatient attendances in England to be complementary. The GP being a gatekeeper for access to secondary care means that most outpatient and hospital admissions would have to be first initiated through a GP referral. So higher appointments could lead to higher levels of hospital admissions and outpatient visits. However, this result is not consistent with the large body of literature [4] where 11 out of 14 studies found a substitution between outpatient visits and specialist GP care, where 10 of the 11 studies were carried out in the UK. However, studies in the systematic review [4] assessed the impact of GPs acting as specialists and therefore as direct substitutes. In contrast, we consider total volumes of GP activity which may stimulate more demand for hospital care. On the other hand, we find that our results are weak complements where the inclusion of a greater set of confounders may likely diminish our findings of complementarity.
Our finding that community care and first outpatient appointment are substitutes is consistent with the literature on integrated care within the UK. However, finding of a substitution effect between community care and hospital admissions adds to a literature base in which there is no current consensus [19]. A possible explanation for this is that health care services delivered in the community may not be directly substitutable for all types of care performed within the hospital. Non-surgical first outpatient visits may be more responsive to the supply of care in the community as management of conditions may not require specialist attention. In comparison, elective surgery for a hip replacement is care that is non-substitutable. Urwin et al. [38] found that informal care can substitute for home help in the community, but not GP visits or inpatient hospital activity where the latter two services require treatment from more trained individuals.
Strengths
To our knowledge, this is the first published analysis to use population-wide activity levels from the care provided in the community and hospital activity. The availability of two new datasets means that relationships between care delivered in the community and care delivered from hospitals can now be empirically tested. This is the first study to take population-wide community services data and GP appointments data to look at the association of care provided in the community and hospital care.
Furthermore, adopting distance-size weights allows for datasets available in different aggregation units to be linked. We compared GP attendances after applying historical and distance-size weighting to test the distance-size weight's performance; we found that the two weighting Table 6 Sensitivity analysis on exploring different powers on the distance decay function for the Index of relative community contacts activity The results from linear regression models. Models also include provider fixed effects, monthly time effects, proportion of population aged 0-14, 65 and over, proportion male, income and relative index of CCG population. Indexes of relative activity are calculated by dividing monthly volume by the average volume reported by the provider over all months *Distance power of 2 is used for the main analysis methods resulted in correlations of 0.99 for A&E attendances, hospital admission and outpatient visits when the distance decay function was set to the power of 2. We also conduct sensitivity analysis to explore whether the main results change based on the increasing and decreasing the distance decay function, which changes the importance of the proximity between hospital and community services providers. Ideally, a population-wide individual-level dataset containing all healthcare services activity would be used for such analysis. However, this dataset does not currently exist in England.
Limitations
Counts of community care contacts from the CSDS are not stratified by the type of care contact; we are unable to identify which services are provided from which provider. This means that we cannot refine the distance-size models based on the type of service used. Therefore, we are unable to perform analysis on subsets of patients where we would expect changes in community care provision to more directly affect the use of hospital services. Furthermore, online statistics report the number of care contacts that are not attended; this information is held on individual patient records, and as such, not available. Data on appointments within General Practice captures only planned appointments and do not include appointments which are emergencies or appointments which were arrange outside of core practice opening hours [39]. This means that not all appointments within General Practice are captured in this study. The data on the appointments within General Practice for all planned, unplanned and outside of core opening hours was not available for all GP Practices in England during this study.
Using monthly aggregated hospital activity data means that the data does not allow for the stratification by types of hospital services. We would expect types of care from A&E, outpatients and hospital admission to be unavoidable. However, outcomes such as avoidable admissions [15] would be more sensitive to changes in care delivered in the community. Furthermore, we have data only on first outpatient visits; we have not tested the relationship between care in the community and an overall number of outpatients visits. We are also limited within our analysis regarding the severity of each of the hospital services being utilised. The data are not available on the volumes of activity for specific services or population groups that may have a stronger relationship with services provided in the community, such as wound care for older age groups. This should be a priority for future research when these data become available.
Owing to the aggregated nature of available data, we do not have data on service users' location. Instead, we use postcodes of service providers. This is not ideal for providers with many sites that could be widely dispersed over a large geographical area. Using postcodes for community service providers, although not ideal, can be used as the location of service users as community service providers have geographical boundaries for where they provide care. However, as geographical boundaries for community care from each care provider are diverse and likely to be changeable over time due to short commissioner-provider contracts [32], using geographical footprints may still not be informative when linking across providers of different units.
Our models do not include any population-level health characteristics. The absence of these health variables results in a positive bias towards the estimated relationship between hospital and community services. One reason why population health variables are not included is that these variables will need to be time-varying. Owing to using fixed-effects regressions, we do control for the average level of health, but we do not account for changes in health over time. We estimate coefficient stability to test how the bias would affect coefficients and show that substitution effects estimated are assumed to be the lower bound. However, we are not as confident with the complementary findings as controlling for unobserved characteristics may result in substitutability. One assumption when using coefficient stability by Oster (2019) is that the direction of bias when including known confounders addresses the same direction of bias for the unknown confounders.
Policy implications
Expansion of care in the community may not result in a significant immediate fall in hospital service use. Despite estimating a substitute effect between hospital services and community care services, the estimated net unit cost of each additional unit of community activity results in higher overall costs to the healthcare system, as cost-saving within secondary care are not outweighed by the cost of care delivered in the community. There may be long term or lagged benefits, including better population health in the long term, but, in the short term, increasing care delivered in the community will not lead to reductions in health system expenditure.
With a fixed level of hospital beds, increasing health care services delivered in the community may not reduce overall hospital bed utilisation, as demand for hospital services outweighs the supply indicated by waiting times, which can be seen as a form of rationing within the English NHS. However, freed up hospital resources through an expansion of community care activity enables other individuals in need of health care to use hospital services, which may result in a reduction in competing demands for the same hospital bed. This could lead to a potential spill over effect of reduction
|
2021-06-17T13:39:45.730Z
|
2021-06-17T00:00:00.000
|
{
"year": 2021,
"sha1": "40ee092117b355ecfd248f2b1f916db5254dd6c6",
"oa_license": "CCBY",
"oa_url": "https://link.springer.com/content/pdf/10.1007/s10198-021-01329-6.pdf",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "40ee092117b355ecfd248f2b1f916db5254dd6c6",
"s2fieldsofstudy": [
"Political Science",
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
119693828
|
pes2o/s2orc
|
v3-fos-license
|
Pleijel's theorem for Schr\"odinger operators with radial potentials
In 1956 $\AA$. Pleijel gave his celebrated theorem showing that the inequality in Courant's theorem on the number of nodal domains is strict for large eigenvalues of the Laplacian. This was a consequence of a stronger result giving an asymptotic upper bound for the number of nodal domains of the eigenfunctions as the eigenvalues tend to $+\infty$. A similar question occurs naturally for Schr"\odinger operators. The first significant result has been obtained recently by the first author for the harmonic oscillator. The purpose of this paper is to consider more general potentials which are radial. We will analyze the case when the potential tends to $+\infty$ and the case when the potential is negative and tends to zero, where the considered eigenfucntion are associated to the eigenvalues below the essential spectrum.
case when the potential tends to zero, the considered eigenfunctions being associated with the eigenvalues below the essential spectrum.
Introduction
The goal of this paper is to extend Pleijel's theorem for the Dirichlet Laplacian H(Ω) = −∆ in a bounded domain Ω to the case of the Schrödinger operator H V = −∆ + V in R d . We are interested in counting the number of nodal domains of an eigenfunction and to relate this number with the labelling of the corresponding eigenvalue. Throughout this paper, for any function f defined over a domain D ⊂ R d , µ(f ) will denote the number of nodal domains of f , namely the number of connected components of D \ f −1 (0). The starting point of the analysis is Courant's Theorem (1923) [8].
Theorem 1.1 (Courant) If φ n is an eigenfunction associated with the nth eigenvalue λ n of H(Ω) (ordered in non decreasing order and labelled with multiplicity), then µ(φ n ) ≤ n . Let us now give the strong form of Pleijel's theorem.
2)
where j ν denotes the first zero of the Bessel function J ν .
The theorem was proved by Pleijel [26] for d = 2 and then extended by Peetre [25] and Bérard-Meyer [3]. We recall from [3,Lemma 9] that Pleijel's constant equals where • w d is the Weyl constant
4)
• ω d := |B d | , (1.5) where B d is the unit ball in R d and |D| denotes for an open set D ⊂ R d its volume; • λ(B d ) is the Dirichlet ground state energy of the Laplacian in B d .
As γ(d) < 1, one recovers as a corollary that the inequality in Courant's theorem is strict for n large. The second point to notice is that the constant is independent of the open set. Complementary properties of γ(d) have been obtained by B. Helffer and M. Persson-Sundqvist [16]. In particular d → γ(d) is decreasing exponentially to 0. Finally note that this constant is not optimal (see [4], [33] and the discussion in [15]).
The original proof of Pleijel's theorem is based on a combination between the Weyl formula and the Faber-Krahn inequality. Weyl's theorem reads, as λ → +∞, N(λ) = |Ω| w d λ There are a lot of Weyl's formulas available in the context of the Schrödinger operator H V := −∆ + V . The use of the Faber-Krahn inequality is more problematic, except of course for the case of bounded domains with bounded potential which can be treated like the membrane case. In 1989 Leydold [21] obtained in his diploma thesis a weak Pleijel theorem for the isotropic harmonic oscillator (see also [2]). Two years ago Charron [6] in his master thesis proved Pleijel's strong theorem also for the harmonic oscillator: Theorem 1.5 (Charron's theorem) Let (φ n ) n∈N be an orthogonal basis of eigenfunctions in L 2 (R d ) of the harmonic oscillator associated with (λ n ) n∈N . Then The theorem is also proven in the case of the non-isotropic harmonic oscillator [7]. The interesting fact is that the potential V (x) = i a i x 2 i (with a i > 0) does not appear on the right hand side of the upper bound. Note also that when there are no eigenvalue degeneracies a much stronger result is available in [7].
A natural question to ask is whether the theorem can be extended to more general Schrödinger operators. We will answer positively this question under the additional assumption that the potential is radial. More precisely, we assume d ≥ 2 and we consider on R d a Schrödinger operator with |x| = r. We will assume that v ∈ C 1 (0, +∞) , (1.11) and that there exists In order to allow some singularity at the origin, we assume either v ∈ C 0 ([0, +∞)) , (1.13) or that there exists s ∈ (0, 2) such that, as (1.14) Here, we say that a ≈ b is a/b and b/a are bounded, i.e. if there exists C > 0 such that We will study two cases according to the behavior of v at +∞ . Case A: v tends to +∞ as r → +∞ .
(1. 18) In the two cases there is a natural selfadjoint extension starting from C ∞ 0 (R d ) (see Section 3). In Case A, the spectrum is discrete and consists of a non decreasing sequence of eigenvalues λ n tending to +∞ . In Case B the spectrum is divided in two parts, the essential spectrum: [0, +∞) and the discrete spectrum, which consists of an infinite sequence of negative eigenvalues (λ n ) n∈N tending to 0 (see for example Reed-Simon [28], Vol. IV, Theorem XIII.6). Associated with this sequence (λ n ) n∈N , we can consider an orthonormal sequence of eigenfunctions φ n , where in Case A φ n is an Hilbertian basis of L 2 (R d ) and in Case B of the negative eigenspace.
Our analysis will contain two well-known potentials: the quantum harmonic oscillator (Case A) and the Coulomb potential (Case B). In both cases, we know the eigenvalues and an explicit basis of eigenfunctions but in the proof this property will not be used. Our aim is to prove the following result: Theorem 1.6 (Pleijel's theorem for Schrödinger) In Cases A or B, if (φ n ) n≥1 is an orthogonal sequence of eigenfunctions of H V associated with the above defined sequence λ n , then The paper is organized as follows. In Section 2, we discuss the general strategy and the methods used by Pleijel first and then by P. Charron. In Section 3, we review the general properties of the Schrödinger operator. In Section 4 we collect those Weyl-type results we need for the proof of Theorem 1.6.
In Section 5, we give the proof of our Pleijel's theorem in the two situations.
Acknowledgements.
Thanks to the ESI where the paper was initiated (B. H. and T. H.-O.). The authors would also thank I. Polterovich for helpful discussions at various stages of this work.
About the methods
As recalled in the introduction, the original proof ofÅ. Pleijel was based on a tricky combination of Weyl's formula with the Faber-Krahn inequality. When considering the case of the Schrödinger operator in R d , Weyl's formula still exists but the use of Faber-Krahn is not easy: nodal domains could be unbounded and the variation of the potential inside a nodal domain could be very high. One has consequently to find an idea for proving that these two bad situations do not occur very often.
In the case of the harmonic oscillator Charron's proof relies on specific properties of the eigenfunctions and the potential. Namely, it used the fact that every eigenfunction is a linear combination of an exponential multiplied by polynomials whose degree can be controlled by a function of the labeling of the eigenvalue, that the hypersurfaces with constant potential are hyperspheres and the fact that the counting function N(λ) behaves nicely as λ → +∞ (Weyl's law).
In addition, it also used that, every nodal domain of an eigenfunction of a Schrödinger operator intersects the classically allowed region associated with the eigenvalue λ, i.e This property is quite general and elementary.
The key was then to divide the classically allowed region in a finite number of annuli of the form V (−1) (a, b). Every nodal domain can either be contained in a single annulus or intersect more than one. To give an upper bound on the number of nodal domains not contained in one annulus, Charron uses properties of algebraic surfaces, as well as results arising in Morse theory adapted from Milnor [23].
Then, he used the Faber-Krahn inequality to give a lower bound on the volume of any nodal domain contained in a single annulus. Dividing the volume of each annulus by the volume of each nodal domain gave an upper bound on the number of nodal domains contained in that annulus.
The last step was to find an appropriate number of annuli to balance out both estimates.
To extend the methods of Charron's proof to more general potentials, we need to find Schrödinger operators such that: (i) There are good lower bounds for the number of eigenvalues below any λ.
(ii) We can count the number of nodal domains that intersect a given energy hypersurface V (−1) (b).
(iii) We can give an upper bound on the number of nodal domains that are not contained in the classically allowed region.
In the case of the harmonic oscillator, the eigenvalues are known explicitly. However, for many potentials V , Weyl's law can be extended to the Schrödinger operator H V for estimating the number of eigenvalues. Hence, we need to find a suitable class of potentials where this law holds.
So far, the only known method to give a suitable upper bound on the number of nodal domains that intersect an energy hypersurface are based on Milnor's theorem (see Subsection 3.7). Hence we need this hypersurface to be algebraic and the property that for any eigenvalue λ and any energy hypersurface (or at least a suitable family) the restriction of any associated eigenfunction u λ equals the restriction of a polynomial to this hypersurface. This is why we focus in this paper on the study of radial potentials. In this case, the energy hypersurfaces are hyperspheres {r = ρ} for some ρ > 0 and it can be shown that the restriction of an eigenfunction to a hypersphere is always a linear combination of hyperspherical harmonics, each one being the restriction to the hypersphere of a homogeneous harmonic polynomial. We will also have to control the degree of such polynomials by a function of λ or of its labelling. This last property will allow us to bound the number of nodal domains that are not contained in V (−1) (−∞, λ).
Another problem might arise when estimating the number of nodal domains contained in one annulus. In the case of the harmonic oscillator, summing over all annuli gives us an expression which can be compared directly with an integral. The error term that arises becomes negligible as λ → +∞ . It remains to show under which conditions on V the same method can be applied.
Finally, in the specific case of Coulomb-like potentials at the origin, we need to look at the behavior of the number of nodal domains near the origin.
3 On the spectral theory of the Schrödinger operators with radial potential
General theory
We first verify that our Schrödinger operator H V = −∆ + V is well defined by a Friedrichs procedure starting from its sesquilinear form defined on Note that the left term has a meaning as soon as V ∈ L 1 loc (R d \ {0}) . In our case, this is a consequence of Assumption 1.11. Our operator, will be defined through a Friedrichs extension. This works as soon as q(u) := a(u, u) is bounded from below by −C ||u|| 2 L 2 . It is consequently enough to control the integral V <0 V (x) u(x) 2 dx from below. When d ≥ 3, we use Hardy's inequality with r = |x| . This inequality extends to H 1 (R d ) by density. For d = 2, we can use the modified Hardy inequality in a disk D(0,Ř) which reads (see for example [10]) which also extends to H 1 0 (D(0,Ř)) . Using these inequalities and a partition of unity, the semi-boundedness on Let us now describe the form domain resulting from the Friedrichs extension procedure. We have: We do not need to characterize the domain of the corresponding self-adjoint operator.
Nodal domains intersect the classically allowed region
We use a similar argument as in [21] and [7] . We assume that we are either in Case A or in Case B, but the result is much more general.
Proposition 3.1 Let λ be an eigenvalue below the essential spectrum, u λ be an eigenfunction of H V associated with eigenvalue λ and Ω be a nodal domain of u λ . Then
The radial Schrödinger operator
Although the exposition there is limited to the case d = 3, one can refer to Reed-Simon [28], Vol. IV p. 90-91. The Laplace operator −∆ can be written as where r = |x| is the radial variable and ∆ S d−1 is the Laplace-Beltrami operator, acting in L 2 (S d−1 ). The following proposition is standard (see for example [31], Theorem 22.1 and Corollary 22.1).
The multiplicity of the eigenvalue ℓ(ℓ + d − 2) is given by which coincides with the dimension of the space of homogeneous, harmonic polynomials of degree ℓ.
We denote by S d−1 ∋ ω → Y ℓ,m (ω) an orthonormal basis of the Λ ℓ,d -dimensional eigenspace associated with ℓ(ℓ + d − 2). We recall that each Y ℓ,m is the restriction to S d−1 of a harmonic homogeneous polynomial of degree ℓ. We now consider the Schrödinger operator H V and assume In this case, one can determine the spectrum by using polar coordinates. In the spherical coordinates, we can determine the spectrum by considering the (closure of the) union of the spectra of the family (indexed by ℓ ∈ N) of Sturm Liouville operators L ℓ defined by acting in L 2 ((0, +∞), r d−1 dr) , with a suitable Dirichlet like condition at 0 (see Reed-Simon [28], p. 91, Proof of Lemma 1). Note that the "Dirichlet like" condition is expressed after the unitary transform u → r d−1 2 u sending L 2 ((0, +∞); r d−1 dr) onto L 2 ((0, +∞); dr) and becomes the standard Dirichlet condition for ℓ = 0. When ℓ > 0, no condition is given. The new operator is then: where λ n,ℓ is the n-th eigenvalue of L ℓ . A corresponding basis of eigenfunctions has the form where Y ℓ,m (ω) denotes an orthonormal basis of (hyper)spherical harmonics.
We recall that these functions form a basis of L 2 (R d ) in Case A (see [28]) or a basis of the negative eigenspace in Case B.
Courant's nodal theorem and nodal behavior of eigenfunctions.
For the analysis of potentials with singularities it is worth to ask under which condition one can prove Courant's theorem or describe the local nodal structure of an eigenfunction. Under our assumptions the only point is the control at the origin. Outside the origin, (1.11) implies that the potential is C 1 and the structure of the nodal set is well known.
Looking at the proof of Courant's theorem, the only thing we need is the unique continuation theorem, i.e. we need to show that if an eigenfunction u λ is identically 0 in a non empty open set ω then it is zero in R d . The argument clearly works if there is only a singularity at 0, because ω \ {0} is an open set where u λ vanishes identically. See [17] for more properties. We will show in the next subsection that no nodal domain is contained in a sufficiently small ball around the origin. Hence the counting of nodal domains can start outside this ball.
No nodal domains in a small ball
In this subsection, we show that under our assumptions the nodal domain cannot be contained in a small neighborhood of the origin. We will start with Case B which is easier.
Case B
We have the following statement: If d ≥ 2 and in Case B there exists r d > 0 such that, if λ < 0 is an eigenvalue and u λ is a corresponding eigenfunction, there is no nodal domain ω of u λ contained in B(0, r d ) .
We now use the identity When d ≥ 3 , we use Hardy's inequality (3.1) (u λ is extended in R d by 0 outside ω and this extension is in H 1 (R d )) and get that This contradicts (3.9) if ω ⊂ B(0, r d ) .
Case A
In Case A, with singularities, there is some difficulty because we consider λ large. When d ≥ 3 , the previous proposition will be true in a ball whose radius is r d (λ) ≈ λ − 1 2 . For d = 2 , r 2 (λ) could be taken as r 2 (λ) ≈ λ − 1 2 −ǫ for some ǫ > 0 . More precisely, we have Proof By (1.14), there exists C > 0 and r 0 > 0 , such that V > −Cr −s for 0 < r ≤ r 0 . As s ∈ (0, 2), there exists λ 0 such that for λ ≥ λ 0 , there exists This implies The proof is achieved by taking 0 < c < d−2 2 in the statement of the proposiiton and using the Hardy inequality as in the second part of the proof of Proposition 3.4. For the case d = 2 , a not optimal r 2 (λ) = λ − 1 2 −ǫ (for some ǫ > 0) together with the modified Hardy inequality do the job for λ ǫ large enough.
Upper bound for the degree of the polynomials associated with the spherical harmonics
In this subsection, we prove the existence of a rather optimal upper bound on the degree ℓ of the polynomials associated with the spherical harmonics Y ℓm appearing in the decomposition of an eigenfunction u λ .
Proposition 3.6
In Cases A or B, if λ is an eigenvalue of H V such that λ < lim inf r→+∞ v then there exists p λ such that, for any associated eigenfunction u λ and for any τ satisfying inf v < τ ≤ λ, we can find a polynomial P τ,λ of d variables of degree at most p λ such that on V (−1) (τ ) in R d the restriction of u λ is equal to the restriction of P τ,λ . Moreover, p λ satisfies Proof. For given λ, u λ has the form (see (3.8)) u λ = λ=λ n,ℓ c n,ℓ,m u n,ℓ,m .
If we restrict u λ to the hypersphere of radius r τ = v (−1) (τ ), we get Considering the property of Y ℓ,m , we can choose for the proof of the proposition P τ,λ = d λ,τ,ℓ,m P ℓ,m , where P ℓ,m is the homogeneous harmonic polynomial of degree ℓ such that It remains, in order to prove (3.11), to determine the highest ℓ ≥ 1 such that λ n,ℓ = λ. By the minimax principle, we have We have indeed The behavior of m ℓ should be analyzed but note that our assumptions imply that m ℓ > −∞ . Furthermore ℓ → m ℓ is strictly increasing, so we can set p λ := [p λ ] + 1 , wherep λ is the solution of λ = mp λ and [x] means the integer part of x.
Application: Determination of an upper bound of p λ .
Case A We can assume that ℓ ≥ 1. This simply implies later a choice of p λ ≥ 1 If we consider v(r) = c r m as a model case for m > 1 and c > 0 , the infimum is obtained when
This gives usp
For m = 2 , we recover what we got for the harmonic oscillator by direct computation.
To treat the general case, we use the lower bound: v(r) ≥ c r m − C r s . (3.14) We have to estimate We observe that: But, there exists (see below the computation in (3.17) with m = −s) a constant C 0 > 0, such that, for all ℓ ≥ 1, and we can use the lower bound of the model case above to get: Hence we obtain like for the model case: Corollary 3.7 In Case A, as λ → +∞ , Case B Let us now compute an example corresponding to Case B. If we take v(r) = −r m for m ∈ (−2, 0), the infimum is obtained when i.e. . (3.16) (3.17) This gives us for λ → 0 , λ > 0 , In the Coulomb case m = −1 and d = 3, we geť 20) to compare with the direct computation which can be done for the Coulomb case.
We will use twice the analysis of the model, the first time with m replaced by −s. Let us first consider We observe (see (3.16) with m = −s) that for ℓ large enough the map For the second case, we can use inf r∈(R,+∞) and what we obtained for the homogeneous model.
Nodal domains on hyperspheres
Since the considered potentials V are radial, the energy hypersurfaces {V = αλ} are hyperspheres centered at 0. Also, the restriction of any eigenfunction u λ of H V to a hypersphere equals the restriction of some harmonic polynomial. We can use the following result proven in [7], which is based on [23]: Let P be a polynomial of degree k with d variables. Then its restriction to the hypersphere S d−1 admits at most 2 2d−1 k d−1 nodal domains.
We will combine this with the previous estimates obtained in Corollaries 3.7 and 3.8 to obtain an upper bound on the number of nodal domains on any hypersphere. 4 Weyl's formula 4
.1 Preliminaries
For Schrödinger operators, Weyl's formula takes (under of course suitable assumptions to be discussed below) the form (4.1) After integration in the ξ variable, we get with w d defined in (1.4).
This formula makes sense in case A (as λ → +∞) and in case B (as λ → 0 with λ < 0). Let us just compute the right hand side for the two toy models: the harmonic oscillator and the Schrödinger operator with Coulomb potential. For the harmonic oscillator, we simply get More generally, if v(r) = r m for m > 0, we obtain, as λ → +∞, In the Coulomb case, we get, with λ < 0 More generally, if v(r) = −r m , for m ∈ (−2, 0), we obtain as λ → 0 (λ < 0), Observing that N(λ n ) = n − 1 if λ n−1 < λ n , and assuming that the Weyl formula is proven (see below for the proof), we get conversly in the case of the harmonic oscillator and in the case of the Coulomb case.
More generally we have the proposition: where the asymptotics is as λ → +∞ in Case A and as λ → 0 (λ < 0) in Case B.
Proof
Outside a ball we can use for estimating the integral defining W (λ) the comparison of v(r) with r m and then use the previous computations for the models. The control of the integral in a ball will be done in Subsection 4.3. ✷
Weyl's formula under weak assumptions
There is vast literature on this subject: Reed-Simon [28] and references therein (for the historics), D. Robert [29], H. Tamura [34], Tulovski-Shubin [32], R. Beals [1], L. Hörmander [18,19], A. Mohamed [24]. In the recent contributions the goal is to control the remainder but this is not important in the applications considered here. Here, we prefer to work under weaker asssumptions and can use Theorem XIII.81 in Reed-Simon (Vol. 4) [28] for the case V → +∞ with a condition d ≥ 2 and m > 1, and for the case V → 0 as |x| → +∞, Theorem XIII.82. The treatment of the singularity is also explained (without detail) (see the discussion p. 277, lines -7 to -1, sending to Problem 132 therein). The idea there is to first prove a statement with V continuous and then to show that the addition of a potential W with compact support or in L Theorem XIII.81 in [28] reads: and For Case B, Theorem XIII.82 in [28] reads: and
Treatment of the singularity
To cover the question of the treatment of the singularity at the origin we could think of using (Problem 132 in [28]) to treat the singularity as a perturbation. Due to the use of the Cwickel-Lieb-Rozenblum inequality [11,22,30] in the argument, this approach works only under the condition d ≥ 3 . If we remember that we only need a lower bound for N(λ) one can proceed for d ≥ 2 in the following way. We can introduce a small ball B = B(0, ǫ) around the singularity and look at the Dirichlet problem in R d \ B . We denote by N B (λ) the corresponding counting function. Because the eigenvalues are greater than the initial problem by monotonicity of the domain, the estimate of the N(λ) of the new problem will give the lower bound: The theorem in Reed-Simon [28] can be applied in R d \ B (proof unchanged) and we get by Weyl's formula It remains to compare W B (λ) and W (λ) in our two cases.
Case B Here λ < 0 . It is enough to show that, for some ǫ > 0, We have the finiteness resulting from the assumption s < 2 .
Case A. Here λ ≥ λ 0 > 0. We will show that B(0,ǫ) (λ − V ) d 2 + dx is relatively small in comparison with N(λ). In Case A, we have seen that We have This gives, as λ → +∞ , Hence in the two cases, we have shown that W A (λ) ∼ W (λ) . In conclusion, we have obtained the following theorem. where the remainder o(1) is as λ → +∞ in Case A and as λ → 0 (λ < 0) in Case B .
Preliminaries
We construct a radial partition of {V < λ} of cardinality ν(λ) with ν(λ) to be defined later. When v is increasing on (0, +∞), r λ := v −1 (λ) is well defined in (inf v, +∞) in Case A, and for any λ in (−∞, 0) in Case B. But we will only be interested λ → +∞ in Case A, and in λ → 0 (λ < 0) in Case B. Under the weaker Assumption (1.12), we can define in the two cases r λ by and obtain the asymptotics Note that in both cases r λ tends to +∞ .
A first partition of the classical region.
We recall that in this case, we assume λ ≥ λ 0 > 0 .
The determination of λ 0 "large enough" will be given during the proof. If u λ denotes some eigenfunction, we denote by D(u λ ) the set of the nodal domains of u λ . We now introduce in D(u λ ) the following subsets.
From Subsection 3.2, we know that every nodal domain is contained in at least one of these sets.
Remark 5.3
This partition will be refined by introducing, in the case of a singularity at the origin, a further partition of A 1 (u λ ).
Counting the nodal domains contained in one annulus of the partition
We first count in each of the annuli D i (λ) for i ≥ 2 . Except if there are no singularity, the treatment of A 1 (u λ ) will be done separately. Hence we first prove the
If Ω is a bounded nodal domain of u λ , the Faber-Krahn inequality (see Theorem 1.4) gives: where we recall that |Ω| denotes the volume of Ω . We know that for a given bounded nodal domain Ω, we have For all Ω ∈ A i (u λ ), we obtain that Because i ≥ 2 , we can, under Condition (5.3), combine (5.5) and (5.7) and obtain Observing that we obtain that Summing up for i = 2, . . . , ν(λ), we get (5.10) We recognize on the right hand side a Riemann sum for the function D 1 (λ)). More precisely, we can write, using the monotonicity of v , Using the asymptotic behavior of v at +∞ given in (1.17), the computation of γ(d) in (1.3), the asymptotic behavior of W (λ) given in (4.11), and (5.1), we achieve the proof of Proposition 5.4.
Counting the nodal sets meeting the boundary of the annuli
Let us now turn to the study of the sets B i (u λ ). We have shown in Corollary 3.7 that p λ ≈ λ m+2 2m . Using Proposition 3.9, we obtain that the number of nodal domains in a given B i (u λ ) satisfies .
Counting in D 1 (λ)
We still have to consider when there is a singularity around 0. We treat first the case d ≥ 3 . We know from Proposition 3.5 that we can replace D 1 (λ) by the annulus D 1 (λ, C) defined by for a sufficiently large C.
As in the proof of (5.8) we obtain: Summing over the Ω's contained in D 11 (λ) and observing that the volume of D 11 (λ) is bounded, we get the existence of a constantĈ d such that where A 11 (u λ ) := {Ω ∈ D(u λ ) | Ω ⊂ D 11 (λ)} .
In particular we get The case when d = 2 does not lead to new difficulties. (1)) . (5.22) Observing that N(λ n ) ≤ n − 1 , we obtain Theorem 1.6.
(5.26)
The proof is the same as in case A. For W (λ), we can use the asymptotics (4.11).
For the cardinalities of the sets B i (u λ ), (5.13) holds in case B under condition (5.25) and we can use Corollary 3.8 together with (4.11).
The treatment of the singularity is slightly easier in this case. We use Proposition 3.4 to make a partition of D 1 (λ) in two annuli: Again, we choose C such that v is strictly increasing for |x| > C. Since λ < 0, there exists M > 0 such that #A 11 (u λ ) < M . To give an upper bound on #A 12 (u λ ), we follow the same steps as in case A.
|
2016-04-28T10:59:49.000Z
|
2016-04-28T00:00:00.000
|
{
"year": 2016,
"sha1": "7df1a404f5d65eda6d2bcbc9594caa5bfec86f66",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/1604.08372",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "7df1a404f5d65eda6d2bcbc9594caa5bfec86f66",
"s2fieldsofstudy": [],
"extfieldsofstudy": [
"Mathematics",
"Physics"
]
}
|
132533660
|
pes2o/s2orc
|
v3-fos-license
|
Path calculation of 7-axes synchronous quasi-tangential laser manufacturing
Quasi-tangential laser processing, also called laser turning, is increasingly applied for various applications. Specifically, its ability to generate complex geometries with small feature sizes at high precision and surface quality in hard, brittle, and electrically non-conductive materials is a key benefit. Due to the geometric flexibility, the process is well suited for prototyping in hard-to-machine materials such as ceramics, carbides, and super-abrasives. However, the lack of advanced software solutions for this novel process hitherto limited the exploitation of the potential. Here, we discuss a unique computer-aided manufacturing approach for synchronous 7-axes laser manufacturing with quasi-tangential strategies. This gives the peerless possibility to process arbitrary geometries, which cannot be manufactured with conventional techniques. A detailed description of the path calculation with derivation and procedures is given. The generated machine code is tested on a laser manufacturing setup consisting of five mechanical and two optical axes. Following, a processed cylindrical ceramic specimen with a continuously varying profile along a helical path is presented. The profile is constituted by a rectangular over half-spherical to a triangular groove with defined pitch on the helix. This demonstrator provides the validation of the presented CAM solution. Measurements of the produced specimen show high adherence with the target geometry and an average deviation below 10 μm.
Introduction
In the scope of rapid prototyping and flexible production routines, laser manufacturing is a versatile tool.Various technologies have been developed in the past two decades and some are extensively used in industry.Laser welding, cutting, drilling, engraving, marking [1] and all-over additive manufacturing [2] are prominent examples.For all these technologies, the choice of beam paths, describing the relative motion of the laser and the workpiece, is crucial [3].In general, the calculation of beam paths for complex workpiece geometries is demanding and a computer-aided manufacturing (CAM) approach is necessary.Most laser processes are carried out in a layer-by-layer manner on flat surfaces with orthogonal incidence of the laser beam.This is termed 2.5D layered manufacturing and the lateral feature size is only limited by the minimal focal diameter of the laser beam [4].In case of laser-ablation processes, this leads to an accumulation of small deviations in each layer and therefore geometric errors [5].For high precision manufacturing on rotational workpieces, a quasi-tangential process is favorable to hold tight tolerances with superb surface quality [6].The principal idea of grazing-incidence laser ablation makes a self-limiting laser ablation process possible [7].This enables smoothing of a prior rough surface under oblique irradiation with θ i >75 • .
The quasi-tangential condition for laser and matter interaction leads to a small impact in terms of a heat-affected zone (HAZ) and introduced dislocations [8].Moreover, state-of-the-art ultra-short pulsed laser systems enable a competitive production time for heat sensitive samples and small geometries [9].The reduced HAZ and forceless manufacturing makes micrometer-scale production possible [10].Accordingly, machining of sensitive ultra-hard materials such as diamond and Borazon are possible without phase transitions altering the mechanical properties.Recently, ultra-short pulsed laser systems are becoming more affordable and increasingly applied by industry [11].An example is the production of cutting tools, like end mills and drills with diamond cutting edges [12,13].The kinematics of quasi-tangential laser processing is similar to the wire electric-discharge machining (EDM) process and the developed CAM models [14] can be facilitated partially.In effect, a laser can be approximated by a cone considering only the portion within the Rayleigh length.However, the curvature of the specimen has to be higher compared to the Gaussian beam caustic in order to avoid screening of the laser.A sketch in figure 1a shows the laser beam irradiating a specimen quasi-tangentially with one scan in blue.The incidence is tangential to the target, but the beam moved about one focal radius inside the material and therefore termed quasi-tangential.Hence, the relative movement of work piece and laser is assembled by a superposition of a spiral motion of the workpiece and the laser scanning with highly dynamic motion by a galvo scanner [13].Subsequently, the laser hatches are computed for each point on the helical path.Figure 1b depicts this helical path with the necessary tilting of the specimen to adjust quasi-tangential irradiation conditions.
The simpler case of 2.5D radial laser ablation has been extensively studied and the path planning is state of the art [15,16].Techniques like skywriting for constant velocity during the ablation process, beam compensation strategies and different hatching approaches are available [17].The slicing algorithms, similar to the procedures reported for milling and layered manufacturing [18,19], have been advanced and adapted to laser ablation processes.However, no comprehensive CAM solutions for quasi-tangential laser processing have been presented to date.The capability of modern numerical controls, to enable synchronous motion of multiple kinematic and optical axes, demands a fundamentally new approach to this problem.A novel path-calculation procedure is presented for a universal CAM and quasi-tangential laser processing with up to five kinematic and two optical synchronously controlled axes.This enables the efficient programming and manufacturing of arbitrary rotational geometries.Figure 2 shows a possible axis configuration incorporated into a laser ablation setup.This is constituted by a laser, beam guidance, a 2-axes scanhead (U,V) and a mechanical axis system consisting of a cross table (X,Y), a focusing axis (Z) parallel to the optical axes of the laser beam, a tilting axis (B) and a rotational axis (C).Due to the strongly varying dynamic specifications of the galvo scanner and machine axes, the mechanical system is applied for the motion of the workpiece, while the scanhead is applied for the fast laser motion.Following, a varying groove profile on a helical pitch within the physical limitations of laser processing can be produced.After a detailed discussion of the necessary functionality of the CAM solution the experimental result is presented as validation of this unique concept.
CAM Modules
There are general approaches on CAM programming for various manufacturing techniques like milling [20], drilling and grinding [21].However, laser manufacturing has to be treated differently, taking into account the non-binary behavior with the workpiece [22].The interaction of the laser beam with the material depends on several parameters with non-linear behavior.The most influencing dependencies concern the laser wavelength λ, polarization P , pulse duration τ P , fluence F , angle of incidence θ i , the materials susceptibility χ and following reflectance R. Consequently, the process parameters have to be determined experimentally to reach the desired ablation rate and surface quality.
It is necessary to follow a detailed parameter study, taking the applied material and configuration into account.If the ablation conditions for a specific set of parameters is known, the CAM tool can compute the adjacent paths and the NC code.At some point, if enough data is acquired a material database could serve as a lookup table.
In the next sections, the computation procedure of the proposed CAM tool chain is described.The steps introduced herein are realized in Matlab, however, are in principle platform independent.The process path-calculation is not limited to ablation processes, and is applicable with modifications for selective laser melting and additive path planning purposes in general [23].
Import of the Geometry
The geometry to be imported and processed by the CAM is the part to be removed by the laser process.Commercial computer-aided design (CAD) systems generate this geometry by subtracting the target geometry from the blank geometry, depicted in figure 3.One of the advantages concerning quasitangential, compared to orthogonal laser processing, is the negligible influence of the blank geometry.Unknown dimensions, irregularities and surface defects can be compensated by assuming a larger blank.Areas where the blank is defined larger than necessary, increase the processing time, but do not affect the processing result.To be flexible in terms of the used CAD system, the import is performed using the stereo lithography (STL) format.This file format was introduced in 1988 [24], but is still one of the standard formats in CAD file exchange.A closed network of triangles and the adjacent normal vectors represent the surface geometry.The proposed CAM tool encompasses a detection algorithm to control the imported geometry and calculate the adjacent normal vector of the triangular [25].
Coordinate Transformation
Generating the layers of a rigid body in Cartesian coordinates for manufacturing and rapid prototyping has been studied for years and different strategies are applicable [26].However, positioning by the mentioned kinematic setup requires the angular information given by cylindrical coordinates.For this reason, the coordinates of the STL vertices are represented in a cylindrical coordinate system with the rotational axis coinciding with the C-axis.Furthermore, the target geometry is unrolled to use the powerful approaches available for layer separation and path calculation.The transformation is formulated as follows: The sample geometry is unrolled over the entire circumference and the geometry exceeding 2π is stitched together in a sequential matter.If one transforms an arbitrary geometry with a constant helix pitch in this way, the helical path can be approximated by a line, depicted in figure 4a.After the coordinate transformation, the geometry parts are separated in pieces, where each piece represents one revolution of the rotational axis.This leads to closed surfaces and the paths can be computed in a more general way.Identical neighboring elements are connected, taking into account that there are no jumps allowed in the geometry.Thus, for example if a helix is present on the surface area of a cylinder, the transformation leads to higher angles for longer cylinders in xdirection.The connection faces for every revolution are sketched in figure 4a and illustrated dashed red.From now on the transformed coordinates are facilitated and denoted with y and z for simplicity of the formulas.
Path Determination
Considering the unrolled geometry in figure 4a an approximation of the geometry is calculated.The surface normal vectors, denoted in blue, are discriminated by the projected positive and negative values in z.This allows distinguishing between top and bottom of the groove.Following, the coordinates of the edge have to be computed.The iterative computation is carried out between the range x min and x max .Following, the approximated path segment, see figure 4b, can be calculated from the actual x-values with adjacent coefficients a k , b k and c k .For higher geometric flexibility, a second order polynomial is introduced and for the n-th y n and z n transformed coordinate follows: The adjacent angle β n for each segment follows from the slope with the coordinate transformed y n and the radius z n to The discretized points with constant distance l b have to be translated to the projected axis movement in x, y and z, which can vary dependent on the slope.The angle β n for each segment gives the necessary tilting position to hold quasitangential incidence during laser ablation.Subsequently, the step size in x is given by ∆x = l b cos β n Hence, the next iteration step for the adjacent Cartesian points y, z can be calculated with the adjacent x n+1 = x n + ∆x.
These equations give the approximated path for the laser manufacturing process of helical geometries.If the geometry is a cylinder or cone, having no helical groove or other non-rotationally-symmetric features on the lateral area, the path calculation is simplified.The roll off is fully defined between the angles 0 and 2π and the tilting axis set to zero having the optical axis normal to the workpiece axis.Following, the average value of x min and x max can be used.In order to have closed paths, the first point has to be identical to the last one after 2π.Therefore, the path distance l b has to be modified and the adjacent path point y n calculated for the given cylinder diameter with z to The adjacent next point is following equidistant on the outer contour with the modified discrete path distance l b .
Shadow Calculation
The laser hatch is determined by shading the target geometry onto the xyplane.A orientation vector v R is generated, which allows to distinguish the direction of adjacent edges relative to the projection.Subsequently, this orientation vector is computed for adjacent normal vectors of the STL triangle 130 for each segment.Therefore, a rotation of e x around y with β n denoted with Subsequently, the shadow edge can be determined, which leads to the projected shadow in the XY-plane.To calculate the laser scans the edges of the bottom and top of the groove slot are handled separately, compare figure 4a.A rotation of the edge along the orientation vector is introduced and these edges are oriented opponent to the z-direction.Hence, the two dimensional projection in the XY-plane are computed and a minimal path has to be determined.Exactly, this projection and, only using parts facing the orientation vector, makes a fast computation of the shadow contour possible.
Contour Assignment
The described projection of the shadow edges leads to an unstructured point cloud in the XY-plane.For computation the inner contour has to be determined.Exemplary, some intersection points of the projected edges lead to a minimal closed contour as sketched in figure 6.In reality, the path is more complicated and several cases have to be taken into account.Figure 7a shows the principle of the evaluation for one specific intersection point I.For illustration, considering one intersection, and the calculation the vector v n is defined between the starting S and adjacent end point E. The slope between these points is therefore given by: and the intersection point calculated from the adjacent starting point S n = I n denoted with: Following, the vector v n is evaluated considering the successor n+1 to determine the minimal inner contour.If the slope k n is in the interval [0, 1], the intersection is on the edge described by v n .Following I n+1 has to be on the same edge and the possible connection v n+1 can be found.In general, more than one potential linkage edge appears at the intersection point I n and this has to be distinguished.Figure 7 depicts several cases to be discussed and evaluated to find the inner contour sketched in 6.The biggest angle ϕ n between two edges is taken to guarantee the tracing of the inner contour from the shadow geometry.Therefore the angle between potential connection vectors is evaluated with: In the specific case shown in figure 7a, the intersection point I is the starting vectors and v 1 leads to ϕ 4 as biggest angle and therefore vector v 4 has to be the successor.Each segment of the contour is evaluated sequentially in similar manner.In some cases multiple intersection points can exist, shown in 7b, and the shortest distance to the predecessor is taken.
Contour Offset
Having the shadow contour of the geometry, the hatches have to be determined.The total volume to be ablated is shown in figure 3 and for each step of the helical path the scan of every layer can be computed.The material ablation and hatching starts from the pristine surface to the calculated inner shadow contour.In contrast, the computation starts inversely to ascertain a hatch equidistant to the inner contour.A first offset distance l e is taken into account followed by a defined offset distance l k applied multiple times until the raw geometry, top groove in figure 4a, is reached.The contour offsets are calculated along the half angle of two adjacent vectors.This ultimately leads to challenges as depicted in figure 8a, where the hatch lines cross after two layers.An event, where a certain edge disappears from geometric constraints, has to be taken into account.Concerning the 2.5D case, similar events can occur [27,28] and computational solutions were proposed.Recently, an enhancement for complex 3D shapes was introduced [29] and parts of the proposed features implemented in the offset calculation.In principle, the algorithm is based on Voronoi path diagrams and reduces two-dimensional shapes to one dimension.Linear offset paths are described, where edges are moving inwards in a straight manner and meet at a vertex, depicted in figure 8b.A straight skeleton algorithm for the edge events is implemented to create the contour offsets and vertex merge events.Initially, a normalized half-angle vector n n is calculated between v n and v n+1 , which is in case of parallel vectors v n normalized and rotated by ±π/2. Figure 9a depicts the situation in more detail for one segment causing a merge event and creating an edge event point (EEP).The black solid line shows the inner shadow contour and the gray solid line illustrates the offset contour.Exemplary, two corner points E0 and E1 have to be merged and a new EEP computed.This new point has to be calculated independent of the targeted hatch offset distances l k and l e , as it may be positioned in-between two hatch lines.The adjacent edge point is shifted in direction n n by This leads to the new merged EEP: Subsequently, the distance of the EEP has to be calculated for each segment to ensure equidistant hatches.A normal distance A is considered to the base between the involved corner points and the EEP.Therefore, the distance follows with calculating the according distance for the EEP in sequential manner to: A has to be stored for each edge point of the geometry.Every time an edge event occurs, the geometric path is shifted outward, starting with the first EEP at minimal distance A to the shadow contour from figure 6.In principle, the direction is given by the half distance and leads to EEPs inward of the shadow contour, shown in 9b.These EEPs are virtual and do not lead to merging events and therefore are neglected with the condition q n < 0 from equation 14.
Hatch Computation
Figure 10a schematically shows the calculation of a laser scan in direction of the new edge point according to equation 19.Following, several scans within the distance A are calculated with the layer thickness l k before an edge merge occurs.Exemplary, several lines are depicted in figure 10b with the shadow contour in black.Blue lines show scans without edge-merge events and gray lines with adjacent points illustrate merged events.The computation of the offset uses the angle α n , compare figure 10a, to project the layer thickness on the directional shift of the edge point this angle is calculated geometrically: From the projection of the edge point, the new edge point follows to In case of the first scan the separate line distance l e is facilitated, which is useful if the last scan should have a different offset distance for better surface quality or compensating the focal beam radius.Moreover, the calculated hatches are modified to incorporate a skywrite feature, which minimizes acceleration and deceleration effects of the scanhead on the ablation.A length l v is defined to elongate every calculated scan over the designed edge, keeping this distance constant.Taking into account the focal spot diameter and axes precision of a laser machine tool, a minimal distance l h between two edge points of the adjacent scan is defined.This reduces the number of segments and therefore axis movements tremendously.The shadow calculation and even more the contour offset procedure leads to edge points in near proximity which do not affect the precision of the design geometry.Therefore, the computational cost and the NC code size for the laser machine can be reduced.Following, the scan strategy and jump strategy is defined.Five computed scan lines are shown in figure 11a for subsequent path points with distance l b .The ablation lines are shown in royal blue and connected via the jump lines in pale blue.Depending on the intended ablation rate and surface quality the scan lines can be collected.In case of a roughening step, figure 11b, lines starting from 3 are collected and higher ablation rates with adjacent parameters applicable.Clearly, this will influence the surface quality and precision of the workpiece.Figure 11c sketches the finishing strategy, where the scan follows the shadow contour.In general, the subsequent laser scans can be oriented in the same direction with longer jump paths in between or the scan direction changed for each line.The latter leads to shorter jump lengths and therefore faster processes.The total process time can be estimated by considering the hatch and jumping paths.Depending on the process and the laser machine tool the scan speed v F and jump speed v J are defined.Taking into account all segmented scan line to line distances, the skywrite length and jump lines gives an estimation of the total process time: The galvo scanner has little inertia, resulting in fast acceleration and maximum speed compared to mechanical axes.The first laser scan, starting from the unprocessed geometry, exhibits the highest rotational speed.An estimation of the required surface speed points to: Consequently, equation 21 must be fulfilled to get the ablation paths without running into limitations of the axes.Depending on the workpiece radius, the necessary rotational speed can be calculated, or vice-versa the galvo feedrate adapted.
Parameters and Interface
The set of parameters necessary for the computation of the laser paths are given in table 1.Hence, the CAM calculation procedure realized solely depends on seven parameters and three system-dependent constants.The parameters have to be determined by a detailed experimental parameter study and set for the specific process.Depending on the utilized laser machine setup, used laser source and required surface quality these parameters can vary widely.To increase usability of the CAM solution, a graphical user interface was created.
After importing the geometry to be ablated the parameters can be set and the calculated beam paths are shown, depicted in the screenshot of figure 12.The NC code is generated with machine specific settings, considering the axes configuration and adjacent machine commands.Hence, the presented CAM tool can be used on several systems with differing industrial controllers.An opensource repository hosts the Matlab source files and a compiled version of the program licensed under the GPL-3.0[25].
Experimental Validation
A 7-axis configuration consisting of five kinematic axes (X,Y,Z,B,C) and two optical axes (U,V), as shown in figure 2, was applied for all experiments.
The axes were controlled synchronously via the Aerotech A3200 motion-control software with drives from the same supplier.Therefore, the system is highly flexible in terms of motion control and embedded in a temperature-controlled environment to reach highest precision.An ultra-short pulsed laser system with an average power of 8 W, a wavelength of 520 nm, a pulse duration of 400 fs and a repetition rate of 200 kHz was applied.
The study presented was carried out on an alumina toughened zirconia (ATZ) specimen.This technical ceramic can show heat-induced phase transitions from a tetragonal phase to a monoclinic one.Following, the validation aims to demonstrate the potential of this tool set on a temperature-sensitive specimen.After a detailed parameter study, five settings have been evaluated in a control experiment.Figure 13a shows the outcome on a cylindrical specimen, keeping the ablated volume constant with varying contour offset.Clearly, the results differ in surface quality and for example the roughness changes or oxygen vacancies lead to a change in color.Phase transitions were ruled out via a Raman spectroscopy study.
Considering the quasi-tangential laser process a smoothing of a previously roughed surface is possible.Following, a portion of the material can be ablated with fast parameters and the surface quality improved with a finishing step.Table 2 gives a summary of two parameter sets, which were used on the ATZ ceramic for roughening and finishing with the facilitated ablation setup.In or- der to demonstrate the unique possibilities of this CAM approach, a specific geometry not producible with conventional manufacturing technologies was designed.Therein, the laser hatch has to vary continuously from a triangle over a half sphere to a rectangle on a constant slope of the helix.After importing the designed CAD into the CAM tool, the laser-path calculation took some minutes and the NC code was generated.Figure 13b shows the resulting specimen with the varying geometry of the groove.The process has not been optimized and this is the first trial of the specific geometry with solely the roughening step serving as proof-of-concept.The contour was measured at a defined rotation angle with a Zoller Venturion.Figure 14 shows the attained contour of this demonstrator part with the deviations magnified one order of magnitude.The groove with varying cross-section could be successfully produced directly with the generated code from the CAM without further compensations.The average deviation is below 10 µm, however, a larger maximal deviations of about 60 µm is measured at the steep flanks of the rectangular section.This reveals the limits 0.5mm of laser processing, which is not capable of producing surfaces parallel to the optical axis of the laser beam without compensation.Steep flanks are known to be a challenge in laser ablation and taper angle compensation techniques have been proposed [30].However, these strategies are not yet implemented in the presented CAM tool.Moreover, the developed CAM system was tested with different laser machine configurations and geometries pointing to a novel flexible and fast approach for laser path calculation.
Conclusion and Outlook
A complete set of CAM routines for the path calculation of synchronous 7axis laser manufacturing of complex geometries under the quasi-tangential condition was developed.The resulting program is applicable for rapid-prototyping and successfully validated for subtractive laser manufacturing.Arbitrary geometries can be produced out-of-CAM if an appropriate laser setup is applied and certain design rules, like flank-angle limitations, are followed.Limitations with regards to the achievable geometries are mostly related to very steep beamincidence conditions on the workpiece surface.This leads to an expansion of the irradiated area and, hence, altering the ablation characteristics.The tool set is open source published and applicable for diverse path-calculation procedures on varying experimental systems.
Further investigations on design limitation during quasi-tangential laser processing and suitable simulations could enhance the presented CAM solution in future.An integrated simulation and modeling environment for the interaction of the laser with different material classes would strongly improve the applicability for rapid prototyping.However, the intrinsic material properties, like reflectance, thermal conductivity and, thus, the angle-dependent ablation behavior must be known in detail.The fast growing additive-manufacturing community is intensively discussing new formats to replace the STL format.Following, a change of the interfaces for this CAM solution will become necessary.Further developments will involve modification of the code package for parallel computation on graphical processing units.The long time goal is a ready-to-use CAM solution for zero-failure high-precision laser manufacturing of arbitrary shaped specimen.
Figure 1 :
Figure 1: Principle of the quasi-tangential laser process with a helical groove.(a) The laser scans are shown in blue and the swivel axis tilted by the helix pitch.(b) The detected helical path line and the shadow projection is shown.
Figure 2 :
Figure 2: Experimental laser ablation setup constituted of five kinematic axes (X,Y,Z,B,C) and two optical galvo axes (U,V) aligned parallel to the XY-plane.
Figure 3 : 4 :
Figure3: The target workpiece geometry is substracted from the raw specimen.This leads to the material to be removed from the blank acting is input geometry.
FigureFigure 5 :
Figure 5a sketches the situation of the geometry at a certain path point and the orientation vector to be determined.The adjacent top edges of the groove are depicted in pale blue and the bottom in royal blue.Only faces turned towards this orientation vector are respected and the averted are neglected, shown in figure 5b.The direction of these faces is evaluated taking the scalar product of the calculated orientation vector with the surface normal.If n t • v R < 0, the surface is oriented towards and taken into account for further calculation.The faces on the averted side of the geometry with n a • v R > 0 are suppressed and not considered.This generates shadow edges, illustrated by the pale blue line in 5c, for a certain angle of rotation and path point l b .a
Figure 6 :
Figure 6: Construction of the minimal inner shadow contour from a point cloud in the xyplane.The groove top is pale blue and the bottom royal blue.
155 point for the vectors v 2 to v 4 .Figure 7 :
Figure 7: (a) Intersection of vectors in the xy-plane to find the minimal inner contour and (b) the determination of the angle between crossed vectors is shown.
Figure 8 :Figure 9 :
Figure 8: Contour offset calculations for ablation paths with and without edge events (a) and merge events in (b).
Figure 10 :
Figure 10: (a) Edge event merger and generation of a new corner Enew.(b) Several hatches with offset in blue and EEP merges highlighted in gray.The first hatch distance le, following layers l k and the elongated hatch distancelv are illustrated.
Figure 11 :
Figure 11: Exemplary hatch strategies and merging laser scan lines with the discretization of l b on the helical path.Adjacent lines with layer thickness l k are summed up.In (a) the hatch is depicted and (b) merges three scan lines for a roughing process.Finishing is shown in (c) following the shadow contour.
Figure 12 :
Figure 12: GUI of the compiled CAM toolbox showing the scans in blue at the geometric transition from triangle to half sphere at a certain path point.
Figure 13 :
Figure 13: (a) Parameter study for quasi-tangential ablation on the ATZ ceramic.(b) Continuously varying groove of the demonstrator after roughening.
Figure 14 :
Figure 14: Contour measurement of the ceramic specimen; brown: target geometry; blue: measured geometry, violet/red: magnified deviations for visibility.
Table 1 :
Parameters for the path calculation acting as input to the CAM system.
Table 2 :
Parameters for quasi-tangential laser ablation showing the fluence F , scanspeed v F , angular step per radius per scan length ∆w, path distance l b , radial contour offset l k , ablation rate Q and adjacent surface quality Ra.
|
2019-04-26T14:07:07.038Z
|
2018-07-20T00:00:00.000
|
{
"year": 2019,
"sha1": "3fc373ee8f22874539fdb5064ce4eb2531fffa0a",
"oa_license": "CCBY",
"oa_url": "https://engrxiv.org/preprint/download/257/666",
"oa_status": "GREEN",
"pdf_src": "Anansi",
"pdf_hash": "2790d0c5bd43d9941c74a09c166523b9fa5676cc",
"s2fieldsofstudy": [
"Materials Science"
],
"extfieldsofstudy": [
"Materials Science"
]
}
|
56216959
|
pes2o/s2orc
|
v3-fos-license
|
The Role of the Magnetic Resonance Imaging for the Accurate Management of Focal Therapy with High-Intensity Focused Ultrasound for the Localized Prostate Cancer The Role of the Magnetic Resonance Imaging for the Accurate Management of Focal Therapy with High-Intensity Focused Ultrasound for the Localized Prostate Cancer
the chapter in “ MR Imaging, ” we want to present the role of MRI in the accurate manage- mentoffocaltherapywith HIFUforthelocalizedprostatecancer.
Introduction
High-intensity focused ultrasound (HIFU) produces ultrasound waves generated by a spherical transducer, delivering ultrasonic energy to pinpoint foci millimeters in diameter [1]. The thermal and mechanical effects of HIFU cause destruction within prostate tissue [2,3]. Specifically, coagulative degeneration quickly develops and is a primary mechanism for decreasing blood flow [3,4]. Cavitation, a mechanical effect of HIFU, occurs because of the rapid cycling from compression to refraction by the ultrasound waves and results in the formation of microbubbles in tissue. When these bubbles reach the size of resonance, they collapse and produce highpressure shock waves, destroying adjacent tissue [5]. Clinically, HIFU energy is delivered in a pulsed mode, and pulses are defined by treatment cycles and energy intensity in watts. Pulse characteristics define the size of the ablation area, including necessary margins to ensure full tumor coverage. The resulting tissue effects are a function of the frequency (wavelength) and intensity (I, in W/cm 2 ) of the applied energy. Intensity is a function of the excitation (voltage) and duration of the energy pulse. The exact size of the ablated area depends on the type of device used, shape of the piezoceramic working element of the transducer, ultrasound frequency and duration of pulsed energy applied, degree of sonication absorption by the tumor tissue, and focal intensity achieved.
Because the area involved in a single treatment using HIFU is extremely small, only minor temperature changes are observed outside of the focal zone [3], making it an attractive therapeutic modality for focal treatment of the prostate, with the aim of curing the cancer while preserving continence and erectile function. Clinical trials of focal ablation of prostate cancers with HIFU have been reported [6][7][8][9][10][11]. In the management of the focal therapy, MRI has the role in the localization of the targeted lesion, the evaluation of the treatment effectiveness, and the local recurrence after the treatment.
The technology of the HIFU in the treatment of the prostate cancer
After the early studies for the treatment of organ-confined prostate cancer, the Sonablate ® device and Ablatherm ® device have been further enhanced to include treatment planning capabilities and improved therapeutic technologies.
In the treatment planning, the Sonablate ® transrectal HIFU probes use proprietary transducer technology with low-energy ultrasound (4 MHz) for imaging of the prostate and high-energy ablative pulses (site intensity: 1300-2200 W/cm 2 ) for treatment delivery. For the Ablatherm ® device, the transrectal HIFU probes use 7.5 MHz pulses for imaging of the prostate and 3 MHz for the delivery of high-energy ablative pulses (site intensity, 1300-2200 W/cm 2 ). In recent applications of HIFU, the treatment range was planned using a MRI-TRUS fusion image. Additionally, reconstructed three-dimensional (3D) planning modes have been integrated into the SB 500 V4 and AB integrated devices, facilitating accurate planning of treatment range. Because of the inherent technology of HIFU treatments, there is no need to puncture the prostate during treatment; however, this makes accurate treatment planning essential for successful focal HIFU therapy.
In the treatment, intraoperative images are available during the procedure. The "popcorn" phenomena formed by the cavitations of tissue may act as an indicator of treatment efficacy. The SB500 tissue change monitoring system (TCM) predicts treatment efficacy by calculation of radio frequency signal from an area of interest and classifies it based on a 3-color system; green: insufficient change to cure, yellow: moderate change, and orange: large change. Based on the color indicated on the treated area, intraoperative retreatment for an insufficiently treated zone (green area) would then be performed. A recent development in the field of adaptive treatment planning is the Focal One ® . This device has a dynamic focusing transducer made of 16 isocentric rings and allows the user to electronically steer the ultrasound beam and move the focal point of the transducer to a maximum of eight different points (32-67 mm from the transducer).
3. The role of MRI in the management of focal therapy 3.1. The localization of the clinically significant cancer in the prostate with MRI-TRUS fusion image-guided prostate biopsy Multiparametric MRI (mpMRI) improves the imaging of the clinically significant prostate cancer in the prostate [12,13], and the information of mpMRI has been used for the precise diagnosis and localization of the clinically significant prostate cancer. MRI-TRUS fusion image-guided biopsy achieved accurate prostate biopsy based on MRI, combining the superior sensitivity of MRI for targeting cancer-suspicious lesions with the practicality and familiarity of TRUS. With the MRI-TRUS fusion devices, the stored MRI and real-time TRUS are superimposed using computer software to enable targeted biopsy of cancer-suspicious lesions [14]. MRI-TRUS fusion biopsy device "BioJet ® " was approved by FDA after the evaluation of the accuracy with phantoms. We reported the BioJet ® experience of the usefulness of the precise diagnosis of the clinically significant cancer [15]. In the 30 patients from whom whole-mount specimens were taken, we found 43 clinically significant cancers, of which 41 (95%) had been detected by both the targeted and systematic biopsies. The median major diameter of significant cancers in the wholegland specimens was 12 (range: 5-28) mm; lesions ≤10 mm represented 59% of the significant cancers (n = 24). Of the significant cancers ≤10 mm (n = 24) in whole-mount specimens, 96% (n = 23) were correctly diagnosed by biopsy. Gleason scores and locations of biopsy-proven significant cancers corresponded to histopathological findings for the whole-mount specimens, since the localization of the clinically significant cancer in the prostate is important for the precise focal therapy. Since February 2016, the biopsy with BioJet was approved as the advance medical technique by Japanese Ministry of Health, Labor, and Welfare.
The evaluation of the effectiveness and the local recurrence of the prostate cancer after the focal therapy with HIFU
The treated areas with HIFU in the prostate appear as nonenhanced areas on contrastenhanced T1-weighted MRI, and subsequent transrectal biopsies have been able to remove homogenous necrotic tissue sections from the nonenhanced area [16]. Therefore, contrastenhanced MRI has been considered to be the evaluation method of choice to demonstrate the effectiveness of HIFU for localized prostate cancer. We reported the time-dependent changes in blood flow within the prostate treated with HIFU on contrast-enhanced MRI between postoperative days 1 and 14 [16]. In addition, pathological analysis showed vessel damage with coagulative degeneration and detachment of vascular endothelial cells in HIFU-treated prostate tissue [16]. Previous studies [5] have shown that these time-dependent changes in blood flow within the prostate are likely due to the primary thermal effects induced on the tissue and vessel damage by coagulative degeneration in the prostate and that cavitation rapidly induces the detachment of vascular endothelial cells, gradually decreasing blood flow secondary to vessel obstruction. Future large-scale studies should investigate the most appropriate timing of contrast-enhanced MRI for precise feedback regarding the effectiveness of HIFU. Using the contrast-enhanced MRI, the evaluation of the effectiveness and the local recurrence of the prostate cancer needs to be evaluated after the focal therapy with HIFU.
4. The original protocol and the early experience of the focal therapy for the prostate cancer in our institution CTCAE Grade 2 was found in one patient (6.7%). In conclusion, the anatomical region selected focal therapy with HIFU would have potential to provide promising results with accurate treatment for the significant cancer and low morbidity.
Conclusion
MRI has the role of the management of the focal therapy with HIFU. Clinical trials have shown potential for effective focal treatment with HIFU-localized prostate cancer. Further oncological and functional outcomes in the patients treated with focal therapy with HIFU would be expected.
|
2018-12-18T16:54:30.055Z
|
2018-01-01T00:00:00.000
|
{
"year": 2018,
"sha1": "84415c5e025526e5adac607ba41fa1d1bd45f407",
"oa_license": "CCBY",
"oa_url": "https://www.intechopen.com/citation-pdf-url/57855",
"oa_status": "HYBRID",
"pdf_src": "Adhoc",
"pdf_hash": "e9d1b5b4b96d65a0318ecb93e9f75afb50916b22",
"s2fieldsofstudy": [
"Medicine",
"Physics"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
189817917
|
pes2o/s2orc
|
v3-fos-license
|
Topology on a new facet of bismuth
Significance We uncover the presence of a new topological crystalline insulator (TCI) state in bismuth, which is protected by a twofold rotational symmetry. In contrast to the recently discovered higher-order topological phase in bismuth, the present TCI phase hosts unpinned Dirac cone surface states that could be accessed directly through photoemission experiments. Our study provides a comprehensive understanding of the rich topological electronic structure of bismuth.
B ismuth is well known for its peculiar physical properties. It was long considered to be the stable element with the highest atomic mass, but relatively recent experiments have shown that bismuth is in fact weakly radioactive (1). It is a semimetal with a vanishingly small carrier density (10 17 cm −3 ) but an exceptionally high electron mobility (10 6 cm 2 ·s −1 · V −1 ) (2-5). As a result, the ultraquantum regime is reached in bismuth at a magnetic field as small as 9 T, beyond which a number of correlated electron states have been observed (4)(5)(6)(7). Because of bismuth's large spin-orbit coupling, bismuth-based materials have also played a fundamental role in topological physics (8,9). A bismuth-antimony alloy (Bi1−x Sbx ) was the first experimental realization of a 3D topological insulator (TI) (10). The Bi2X3 family supports the prototypical TI state with a single surface Dirac cone (11)(12)(13)(14). Doped Bi2X3 gives rise to the quantum anomalous Hall effect (15) and unconventional nematic superconductivity (16,17). Na3Bi is a 3D Dirac semimetal (18,19). Despite being a crucially important chemical component in many topological materials, pure bismuth has long been thought to be topologically trivial. However, distinct from an atomic band insulator, bismuth does have an even number of band inversions (10,20). The existence of these band inversions implies that bismuth cannot be smoothly connected to an atomic band insulator without undergoing closing of its band gap. This suggests that the band topology of bismuth must be far from being trivial (8,9,(20)(21)(22).
Theoretical advances on topological crystalline insulators (TCIs) (21) have greatly expanded the topological classification of band insulators beyond the Z2 TIs (8,9). After the success of theoretical prediction and experimental realization of the mirror-symmetry-protected TCIs (20-35), two novel TCI phases, rotational-symmetry-protected TCIs (36) and higherorder TIs (37)(38)(39)(40)(41)(42)(43), have been theoretically proposed recently. Rotational-symmetry-protected TCIs are predicted to harbor "unpinned" Dirac surface states in that the surface normal to an N -fold rotational axis (N = 2, 4, 6) can host N Dirac cones whose Dirac points appear at generic k points in the surface Brillouin zone. In contrast, higher-order TIs are generated through the consideration of higher-order bulk-boundary correspondence. For instance, a 3D second-order TI supports 1D topological hinge states in a rotational symmetry-preserving rod. It is important to note that the first-order and higher-order topologies are not mutually exclusive. In particular, a TCI could support both topological 2D surface states and 1D hinge states. Thus, strictly speaking, a pure higher-order TI should refer to the TCIs that do not show the presence of topological surface states. Beyond these new TCI phases, another important theoretical advance is the development of methods to systematically diagnose topological invariants in terms of the symmetry eigenvalues of the electronic states (44)(45)(46)(47)(48)(49)(50)(51). In this connection, Song et al. (47) and Khalaf et al. (48) found that, when certain additional symmetry Y is present, topological invariants of TCIs protected by symmetry X can be inferred from the Y -related symmetry eigenvalues of the energy bands. Such proposals of symmetry indicators and topological quantum chemistry have facilitated first-principles studies of new topological materials (52)(53)(54)(55)(56)(57)(58).
Building upon these theoretical advances, a recent work (41) showed that pure bismuth hosts a second-order band topology that is protected by threefold rotational and inversion symmetries. As a result, it supports 1D topological hinge
Significance
We uncover the presence of a new topological crystalline insulator (TCI) state in bismuth, which is protected by a twofold rotational symmetry. In contrast to the recently discovered higher-order topological phase in bismuth, the present TCI phase hosts unpinned Dirac cone surface states that could be accessed directly through photoemission experiments. Our study provides a comprehensive understanding of the rich topological electronic structure of bismuth. states in a crystal whose shape preserves the threefold rotational and inversion symmetries. Here we show that the band topology of bismuth is even richer in that it also hides a first-order TCI state, which is protected by its twofold rotational symmetries, resulting in a pair of unpinned topological Dirac surface states on its (110) surface. We also reveal the presence of a distinct, previously uncharacterized set of 1D topological hinge states protected by twofold rotational symmetry.
Results
First-principles calculations were carried out using the VASP package (59,60). An ultrasoft pseudopotential and the generalized-gradient approximation were applied in the selfconsistency process. The experimental crystal structure of Bi was used (61) in the calculations. A Wannier-basis-based tightbinding model, where the s and p orbitals of Bi were included, was obtained via the Wannier90 package (62).
Bismuth crystalizes in a rhombohedral structure (61). Its space group and point group are R-3m (#166) and D 3d , respectively. Each layer in this structure forms a buckled honeycomb lattice. The three principle lattice vectors ( ai=1,2,3) are tilted with respect to the Cartesian directions ( points, and with the exception of the Γ point, the lowest three valence bands at all of the TRIM points were found to have one state with positive parity and two states with negative parity.
In contrast, at Γ, all three valence bands have positive parity eigenvalues, so that bismuth has two band inversions at the Γ point.
We turn next to discuss the symmetry-based indicators. As shown in refs. 47 and 48, crystals in the space group #166 are characterized by four symmetry indicators, three Z2 invariants and one Z4 invariant (Z2,2,2,4). By enumerating the symmetry eigenvalues of the electronic states at high-symmetry points, we obtain Z2,2,2,4 = {0, 0, 0, 2}. Table 1 shows the corresponding topological states and the associated topological invariants. Importantly, the specific symmetry indicators (Z2,2,2,4 = {0, 0, 0, 2}) point to two possible topological states, one being a purely rotational-symmetry-protected TCI with n2 [110] = 1, while the other is a purely mirror-symmetry-protected TCI with nM [110] = 2. To uniquely determine the topological state of bismuth, we have further calculated the mirror Chern number nM [110] and found nM [110] = 0 for bismuth; see SI Appendix, section 1 for details. Our analysis thus reveals that bismuth is a purely rotational-symmetry-protected TCI with n2 [110] = 1. Note that there are three twofold rotational axes (2 [110] , 2 [101] , and 2 [011] ) involved here, which are related by the out-ofplane threefold rotation 3 [111] . We elaborate on the nontrivial topology protected by 2 [110] . The other two twofold rotational axes are associated with the same nontrivial topology, so that the associated surfaces can also be expected to host a pair of unpinned Dirac surface states. Moreover, the symmetry indicators point to a nontrivial topological invariant associated with space-inversion symmetry (nI = 1). This guarantees 1D topological hinge states protected by inversion symmetry, which is consistent with previous findings based on the additional C 3[111] constraint (41). We emphasize that the high-order TI predicted in ref. 41 does not support first-order surface states (36,47,48) because there are no topological invariants associated with C 3[111] alone.
To obtain a deeper understanding of Z4 = 2, we provide an intuitive physical picture that connects the band inversion property to the rotational-symmetry-protected topology along the lines of the physics discussed in ref. 41. Recall that a single band inversion at Γ can lead to a strong TI state. Therefore, the double band inversions in bismuth can be viewed as representing two sets of strong TI band structures, where each strong TI contributes a surface Dirac cone atΓ (Fig. 2C). The nontrivial topology protected by the twofold rotation 2 [110] then yields the double Dirac cones, which remain gapless on the (110) surface. To further connect the above intuitive picture to the symmetry indicators, we provide a more detailed analysis of the representations at the TRIM points. For this purpose, it is helpful to rearrange the Fu-Kane formula (63) (Eq. 1) as where ν0 is the strong TI (STI) index, and n − k is the number of occupied bands with negative parity eigenvalue. . Thus, the occupied bands can be separated into two subspaces according to their respective C2 eigenvalues (±), which are C − 2+ and C − 2− (Eq. 2). This rearrangement may lead to new topology beyond the Z2 TI state, but it can be made only when the system is Z2 trivial (ν0 = 0). From Eq. 3, we see that ν (+)/(−) plays a role similar to that of ν0 within each of the C2 subspaces. ν (±) can be directly obtained from the symmetry indicator Z4 (37,(46)(47)(48), using Eq. 4 shows that Z4 = 2 corresponds to ν (+) = ν (−) = 1. In other words, the band structure in each C2 subspace can be viewed as a strong TI. The hybridization between the two strong TI surface states in the presence of the 2 [110] rotational symmetry thus leads to the pair of unpinned Dirac surface states on the (110) surface.
With the preceding discussion in mind, we computed the surface band structure throughout the (110) surface BZ (Fig. 2). As expected, we found two sets of surface Dirac cones, whose Dirac points are located at (±0.01 π, ∓0.269 π) of (ky , kz ) axes indicated in Fig. 1B. These are generic k points, which are related only by the twofold rotational symmetry 2 [110] . The calculated surface states along the k paths passing through the Dirac points (DPs) (Fig. 2 E and F) directly show the presence of gapless Dirac surface states. Interestingly, these unpinned Dirac fermions are of type II (64) in that the velocities of the two surface bands that cross possess the same sign. By contrast, along the pathT −Γ −L1 that does not go through the DP, the surface states are gapped (Fig. 2D). To further confirm the nontrivial nature of these states, we studied the Wannier (Fig. 2H). These results unambiguously demonstrate the presence of unpinned Dirac surface states associated with the twofold rotational-symmetryprotected topology.
We now demonstrate the 1D topological hinge states protected by 2 [110] . For this purpose, we constructed two types of bismuth rods (Fig. 3 A and B), which are periodic along the [110] axis but finite sized within the plane that is perpendicular to [110]. In the first rod, the side surfaces are the (111) and (112) surfaces and the cross-section is rectangular (Fig. 3A). In the second rod, the side surfaces are (001) and (111) and the cross-section is a parallelogram (Fig. 3B). Both rods are invariant under the twofold rotation 2 [110] . The calculated band structures for rods 1 and 2 are shown in Fig. 3 C and D, from which we can clearly identify the existence of 1D helical edge states lying inside the bulk band gap. We also investigated the real-space distribution of the wave function of these 1D helical edge states. Fig. 3 E and F shows that these helical states are localized on the edges shared by adjacent side surfaces, further confirming that these are topological hinge states.
Finally, we show that the 2 [110] rotational-symmetry-protected topology sheds light on the large Rashba surface states observed on the (111) and (001) surfaces (2,3). The double bulk band inversion at Γ also leads to a pair of surface Dirac cones on the (111) and (001) surfaces. However, as (111) and (001) surfaces do not respect 2 [110] , these Dirac cones are gapped, which naturally gives rise to the Rashba surface bands (Fig. 4). Note that since the Rashba surface states are nontopological, they can be created or removed by adjusting the surface potential (69). Our analysis gives insight into the origin of Rashba bands on bismuth's (111) and (001) surfaces, which have been observed experimentally (2,3,69), in terms of the underlying symmetries and the double band inversions in the band topology of bismuth.
In summary, we have investigated topological properties of the bulk and surface band structures of bismuth. We show that bismuth is a TCI with multiple nontrivial topological invariants. In the first order, bismuth features unpinned Dirac surface states on the (110), (011), and (101) surfaces that are protected by their twofold rotational symmetries. These unpinned Dirac surface states would be amenable to direct imaging via photoemission spectroscopy. In the second order, bismuth features two sets of independent topological helical hinge states. The first set is protected by the twofold rotational symmetries whereas the second set is protected by space inversion symmetry. Our study thus provides a comprehensive picture of the rich band topology of bismuth, which is arguably the most important element involved in the field of topological materials.
Note Added in Proof. After we completed our study, we became aware of refs. 54 and 55, which also propose bismuth as a potential TCI. Our analysis, however, uniquely identifies bismuth as a rotational-symmetry-protected TCI.
|
2019-06-15T13:07:32.171Z
|
2019-06-13T00:00:00.000
|
{
"year": 2019,
"sha1": "20176979277c2786300d963947835d24b879d7de",
"oa_license": "CCBYNCND",
"oa_url": "https://www.pnas.org/content/pnas/116/27/13255.full.pdf",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "20176979277c2786300d963947835d24b879d7de",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Medicine",
"Materials Science",
"Physics"
]
}
|
85498687
|
pes2o/s2orc
|
v3-fos-license
|
On fusion rules and intertwining operators for the Weyl vertex algebra
In vertex algebra theory, fusion rules are described as the dimension of the vector space of intertwining operators between three irreducible modules. We describe fusion rules in the category of weight modules for the Weyl vertex algebra. This way we confirm the conjecture on fusion rules based on the Verlinde algebra. We explicitly construct intertwining operators appearing in the formula for fusion rules. We present a result which relates irreducible weight modules for the Weyl vertex algebra to the irreducible modules for the affine Lie superalgebra $\widehat{gl(1 \vert 1)}$.
Introduction
In the theory of vertex algebras and conformal field theory, determination of fusion rules is one of the most important problems. By a result by Y. Z. Huang [21] for a rational vertex algebra, fusion rules can be determined by using the Verlinde formula. However, although there are certain versions of Verlinde formula for a broad class of non-rational vertex algebras, so far there is no proof that fusion rules for such algebras can be determined by using the Verlinde formula. One important example is the singlet vertex algebra for (1, p)-models whose irreducible representations were classified in [2]. Verlinde formula for fusion rules was also presented by T. Creutzig and A. Milas in [12], but so far the proof was only given for the case p = 2 in [7]. We should also mention that the fusion rules and intertwining operators for some affine and superconformal vertex algebras were studied in [1], [4], [11] and [24].
In this paper we study the case of the Weyl vertex algebra, which we denote by M , also called the βγ system in the physics litarature. Its Verlinde type conjecture for fusion rules was presented by S. Wood and D. Ridout in [27]. Here, we present a short proof of Verlinde conjecture in this case. We prove the following fusion rules result:
Acknowledgment.
This work was done in part during the authors stay at the Research Institute for Mathematical Sciences (RIMS), Kyoto in July 2018. The results of this paper were also reported by V. P. at the Southeastern Lie Theory Workshop X University of Georgia, Athens, GA, June 10-12, 2018.
The authors are partially supported by the QuantiXLie Centre of Excellence, a project coffinanced by the Croatian Government and European Union through the European Regional Development Fund -the Competitiveness and Cohesion Operational Programme (KK.01.1.1.01.0004).
Fusion rules and intertwining operators
In this section we recall the definition of intertwining operators and fusion rules. More details can be found in [20], [19], [14], [10]. We also prove an important result on the action of certain automorphisms on intertwining operators. This result will enable us to produce new intertwining operators from the existing one.
Let V be a conformal vertex algebra with the conformal vector ω and let Y (ω, z) = n∈Z L(n)z −n−2 . We assume that the derivation in the vertex As a vector space, There is a unique vertex algebra (M, Y, 1) (cf. [17], [23], [16]) where the vertex operator map is In particular we have: Then β is a Heisenberg vector in M of level −1. This means that the components of the field β(z) satisfy the commutation relations Also, we have the following formula [β(n), a(m)] = −a(n + m), [β(n), a * (m)] = a * (n + m).
The vertex algebra M admits a family of Virasoro vectors This means that the components of the field L(z) satisfy the relations For µ = 0, we write ω = ω µ , L µ (n) = L(n). Then c µ = 2. Clearly Since (β(−2)1) 0 = (Dβ) 0 , we have that L µ (−1) = L(−1), for every µ ∈ C. In particular, Proof. Since W is Z ≥0 with top component W (0), the operators a(n), a * (n) must act trivially on W (0) for all n ∈ Z >0 . Since we have The Lemma holds. • For every w ∈ W , there is N ∈ Z ≥0 such that a(n)w = a * (n)w = 0, for n ≥ N.
For f, g ∈ Aut( A), we have For every s ∈ Z the Weyl algebra A admits the following automorphism ρ s (a(n)) = a(n + s), ρ s (a * (n)) = a * (n − s). Then ρ s is an automorphism of A which can be lifted to an automorphism of the vertex algebra M . Automorphism ρ s is called spectral flow automorphism.
Assume that U is any restricted module for A. Then ρ s (U ) is also a restricted module for A and ρ s (U ) is a module for the vertex algebra M .
Let K, be the category of weight M -modules such that the operators β(n), n ≥ 1, act locally nilpotent on each module N in K. Applying the automorphism ρ s to the vertex algebra M , we get M -module ρ s (M ), which is a simple current in the category K. The proof is essentially given by H. Li in [25,Theorem 2.15] in a slightly different setting.
Proposition 3.1. [25] Assume that N is an irreducible weight M -module in the category K. Then the following fusion rules hold: Proof. First we notice that if N is an irreducible M -module in K, we have the following fusion rules Using [25], one can prove that ρ s (M ) is constructed from M as: Assume that N i , i = 1, 2, 3 are irreducible modules in K. By [25, Proposition 2.4] from an intertwining operator I(·, z) of type N 3 N 1 N 2 , one can construct intertwining operator I s 2 (·, z) of type Now, the fusion rules (3.9) follows easily from the above construction using (3.10).
Consider the following automorphism of the Weyl vertex algebra Assume that U is any M -module. Then U g = U • g is generated by the following fields a g (z) = −a * (z), a * g (z) = a(z). As an A-module, U g is obtained from U by applying the following automorphism g: a(n) → −a * (n + 1), a * (n) → a(n − 1). (3.11) This implies that where σ is the automorphism of A determined by a(n) → −a * (n), a * (n) → a(n).
The automorphism g is then a vertex algebra automorphism of order 4. Denote by σ 0 the restriction of σ on A 1 . Using (3.11)-(3.13) we get the following result: Proof. We have that as an A-module: Since ρ −1 • σ = σ • ρ 1 , by applying (3.7) we get: The proof follows.
3.3.
Weight modules for the Weyl vertex algebra.
Clearly, vertex algebra M is a weight module. We will now construct a family of weight modules.
• Recall that the first Weyl algebra A 1 is generated by x, ∂ x with the commutation relation • For every λ ∈ C, has the structure of an A 1 -module.
• Note that a(0), a * (0) generate a subalgebra of the Weyl algebra, which is isomorphic to the first Weyl algebra A 1 . Therefore U (λ) can be treated as an A 1 -module by letting a(0) = ∂ x , a * (0) = x. • By applying the automorphism σ 0 on U (λ) we get • Define the following subalgebras of A: • Then we have the induced module for the Weyl algebra: as a vector space. Proof. The proof follows from Lemma 3.1 and the fact that U (λ) is a Z ≥0graded M -module whose top component is an irreducible module for A 1 .
Applying Lemma 3.2 we get: 3.4. More general weight modules. A classification of irreducible weight modules for the Weyl algebra A is given in [18]. Let us describe here a family of weight modules having infinite-dimensional weight spaces. Take λ, µ ∈ C \ Z. Let such that a(n), a * (m) act trivially for n ≥ 2, m ≥ 1. Then by [18], is an irreducible A-module. As a vector space: Since U (λ, µ) is a restricted A-module we get: is an irreducible weight module for the Weyl vertex algebra M .
One can see that the weight spaces of the module U (λ, µ) are all infinitedimensional with respect to (β(0), L(0)). In particular, vectors are linearly independent and they belong to the same weight space.
Remark 3.2. Note that modules U (λ, µ) are not in the category K, and therefore Proposition 3.1 can not be applied in this case.
The vertex algebra Π(0) and the the construction of intertwining operators
In this section we present a bosonic realization of the weight modules for the Weyl vertex algebra. We also construct intertwining operators using this bosonic realization.
There is an injective vertex algebra homomorphism f : We identify a, a * with their image in Π(0). We have (cf. [16]) Note also that Since is a vertex subalgebra of V L , for every λ ∈ C and r ∈ Z, is an irreducible Π(0)-module.
Construction of intertwining operators.
Proposition 4.2. For every ℓ 1 , ℓ 2 ∈ Z and λ, µ ∈ C there exist non-trivial intertwining operators of types in the category of weight M -modules.
In the category of M -modules, we have non-trivial intertwining operators Note that the intertwining operators (4.15) satisfy the L(−1)-derivative property. Intertwining operators (4.16) satisfy this property by using the lattice realization as before, and intertwining operators (4.17) satisfy it by Proposition 2.1, using the facts that g(ω) = ω 1 and L µ (−1) = L(−1), for µ = 1. Moreover, using relation (3.6) we see that the L µ (−1)-derivation property holds for every µ ∈ C, for all intertwining operators constructed above.
Let g = gl(1|1) be the complex Lie superalgebra generated by two even elements E, N and two odd elements Ψ ± with the following (super)commutation relations: Other (super)commutators are trivial. Let (·, ·) be the invariant supersymmetric invariant bilinear form such that All other products are zero.
Let g = gl(1|1) = g ⊗ C[t, t −1 ] + CK be the associated affine Lie superalgebra with the commutation relations where K is central and for x ∈ g we set x(n) = x ⊗ t n . Let V k (g) be the associated simple affine vertex algebra of level k.
Let V r,s be the Verma module for the Lie superalgebra g generated by the vector v r,s such that N v r,s = rv r,s , Ev r,s = sv r,s . This module is a 2-dimensional module and it is irreducible iff s = 0. If s = 0, V r,s has a 1-dimensional irreducible quotient, which we denote by A r .
Let V r,s denote the Verma module of level 1 induced from the irreducible gl(1|1)-module V r,s . If s / ∈ Z, then V r,s is an irreducible V 1 (gl(1|1))-module. If s ∈ Z, V r,s is reducible and its structure is described in [13,Section 5.3].
The Clifford vertex algebra F is generated by the fields As a vector space, Let V Zγ be the lattice vertex algebra associated to the lattice Zγ ∼ = Z, γ, γ = 1. By using the boson-fermion correspondence, we have that F ∼ = V Zγ , and we can identify the generators of the Clifford vertex algebra as follows (cf. [22]): Now we define the following vertex superalgebra: and its irreducible modules SΠ r (λ) = Π r (λ) ⊗ F = SΠ(0).e rβ+λ(α+β) .
Let U = M ⊗ F . Using [22, Section 5.8] we define the vectors Then the components of the fields satisfy the commutation relations for the affine Lie algebra g = gl(1|1), so that M ⊗ F is a g-module of level 1. (See also [5] for a realization of gl(1|1) at the critical level).
The Sugawara conformal vector is Construction of irreducible V 1 (g)-modules from irreducible Mmodules. Let V 1 (g) be the simple affine vertex algebra of level 1 associated to g.
We have the following gradation: We will present an alternative proof of the following result: We have: Proof. Let V 1 (g) be the vertex subalgebra of U 0 generated by g. Assume that V 1 (g) = U 0 . Then there is a subsingular vector v r,s / ∈ C1 for g of weight (r, s) such that for n > 0: In other words, v r,s is a singular vector in the quotient U 0 = U 0 / V 1 (g). Since E(0) acts trivially on U 0 , we conclude that s = 0. Recalling the expression for the Virasoro conformal vector (5.22), we get that in U 0 : This implies that v r,0 has the conformal weight 0 and hence must be proportional to 1. A contradiction. Therefore, U 0 = V 1 (g). Since U 0 is a simple vertex algebra, we have that V 1 (g) = V 1 (g).
We can extend this irreducibility result to a wide class of weight modules. The proof is similar to the one given in [3, Theorem 6.2].
Theorem 5.1. Assume that N is an irreducible weight module for the Weyl vertex algebra M , such that β(0) acts semisimply on N : Then N ⊗ F is a completely reducible V 1 (g)-module: Proof. Clearly L s (N ) is a U 0 (= V 1 (g))-module. It suffices to prove that each vector w ∈ L s (N ) is cyclic. Since N ⊗ F is a simple U -module, we have that U .w = N ⊗ F . On the other hand, N ⊗ F is Z-graded U -module so that U r · L s (N ) ⊂ L r+s (N ), (r, s ∈ Z).
This implies that U r .w ⊂ L r+s (N ) for each r ∈ Z. Theferore U 0 .w = L r (N ). The proof follows.
As a consequence we get a family of irreducible V 1 (g)-modules: Corollary 5.1. Assume that λ, µ ∈ C \ Z. Then for each s ∈ Z we have: We will prove in the next section that L s ( U (λ)) are irreducible highest weight modules. But one can see that L s ( U (λ, µ)) have infinite-dimensional weight spaces. A detailed analysis of the structure of these modules will appear in our forthcoming papers (cf. [9]).
The calculation of fusion rules
In this section we will finish the calculation of fusion rules for the Weyl vertex algebra M .
We will first identify certain irreducible highest weight g-modules.
By using the following natural isomorphism of the spaces of intertwining operators (cf. [6, Section 2]): Theorem 6.2 implies the fusion rules result in the category of modules for the Weyl vertex algebra M (see also [27,Corollary 6.7], for a derivation of the same fusion rules using Verlinde formula).
|
2019-03-25T11:31:54.000Z
|
2019-03-25T00:00:00.000
|
{
"year": 2019,
"sha1": "ad392ddd5360edcbaf9ca257a3b0f3885a8d49cc",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/1903.10248",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "ad392ddd5360edcbaf9ca257a3b0f3885a8d49cc",
"s2fieldsofstudy": [
"Mathematics",
"Physics"
],
"extfieldsofstudy": [
"Mathematics",
"Physics"
]
}
|
256319957
|
pes2o/s2orc
|
v3-fos-license
|
No Apparent Immediate Reproductive Costs of Overlapping Breeding and Moult in a Mediterranean Great Tit Population
Simple Summary The annual breeding and moulting calendar in birds is affected by global warming and is leading to an increased overlap between these two activities in some populations. Allocation of resources for two energy-demanding activities could negatively affect performance in one or both of them. In this paper, we examined whether the overlapping of breeding and moulting has negative effects on the reproductive performance of a population of great tit (Parus major) in eastern Spain. We found that, in pairs where both parents overlapped breeding and moulting, clutch size was smaller, fewer eggs hatched, and fewer fledglings in poorer body condition left the nest. However, these differences disappeared when the seasonal trend in breeding performance was taken into account, i.e., the poorer reproductive performance of pairs that overlap moulting and breeding was mainly due to the fact that they breed later, and reproductive performance declines as the season progresses. Thus, we conclude that the overlap of breeding and moulting does not impose additional immediate reproductive costs in this population. Abstract Some phenological events in birds, such as breeding and moulting, are being affected by rising temperatures due to global warming, and many species have undergone temporary changes in these energetically demanding phases that are often separated in time. This has led to an increased overlap between breeding and moulting in some populations. This overlap causes conflicts in resource allocation and may impose fitness costs that could affect immediate reproductive performance. We tested whether this occurs in a great tit (Parus major) population in eastern Spain. In 71% of 390 pairs, in which both parents were captured during the period of overlap between moulting and breeding, at least one parent was moulting when feeding the chicks of its second brood. Later breeders were more likely to overlap breeding and moulting, and when both parents overlapped, clutch size was smaller, fewer eggs hatched and fewer fledglings in poorer body condition were produced. Some results were intermediate when only one parent moulted. However, all these differences between moulting and non-moulting pairs disappeared when the seasonal trend in reproductive parameters was taken into account, as moulting birds bred later and reproductive performance decreased seasonally. Therefore, the overlap of breeding and moulting does not impose additional reproductive costs in this population.
Introduction
The increase in temperatures during the last decades due to global warming is affecting phenological events in many species of plants and animals [1][2][3][4][5][6][7]. Reproduction, moult and migration are some examples of phenological events that have been affected by global warming, and birds undergo through these important life stages in a specific sequence. Each stage is energetically costly, and they tend to avoid overlapping them [8][9][10][11]. In many bird species, the timing of reproduction, moult and/or migration is altered [12][13][14][15][16][17][18].
Moreover, the relative temporal changes in the duration of these stages differ in different rates, so that the periods elapsed between them are either lengthened or shortened [17]. In some extreme cases, intervals between activities can even disappear, thus two of them (as reproduction and moult) could overlap [17,18].
Overlapping breeding and moulting leads to a conflict in resource allocation between these two energy-demanding processes. This can lead to physiological trade-offs that can include a reduction in the breeding success, survival rate and/or the quality of the feathers [10,17]. However, probably because of its rare occurrence in normal conditions, few studies have explored the reproductive consequences of moult-breeding overlap. Moreover, these few studies have been mostly performed on migratory species, where the pressure to finish moulting before migration is higher. Most of these studies adopt different experimental approaches [10,[19][20][21][22][23][24], while only few of them were based on observation of non-manipulated individuals [17,[25][26][27]. Experimental studies are good for showing whether forcing the birds to overlap when they would not have done so caused overlapping birds to experience negative fitness effects. However, it is important to know whether this also happens in real (i.e., non-manipulated) conditions. It might well be that overlapping birds are in better body condition, had better territories with more available food, and/or are supporting lower reproductive burdens (i.e., lower brood size); and so they might efficiently perform both activities without apparent costs. Therefore, observational studies are important to shed light on the consequences of overlapping for non-manipulated individuals.
All studies including observations in unmanipulated conditions have been performed with the migratory pied flycatcher (Fidecula hypoleuca), each one with a different approach and measuring different variables in different ways. In an initial study, Hemborg (1999) [25] included five years of observations and 96 complete pairs (i.e., both male and female captured). He therefore compared some reproductive parameters (laying date, clutch size, number of fledglings, and fledging success) of pairs in which none; only the male; only the female; or both were moulting. Then, Hemborg et al., (2001) [26] built over the previous study, by adding data of three more European flycatcher populations. In this case, they only included in the analyses laying date, clutch size and breeding success, and males and females were treated independently (i.e., not considering pairs as a whole). Morales et al. (2007) [27] reported on the effects of overlap on hatching success and on hatchlings and fledglings produced by females. Finally, Tomotani et al., (2017) [17] performed a study with a big dataset of 35 years to study temporal changes in the timing of breeding and moulting, and the impact of their overlap on clutch size, proportion of chicks fledged and adult survival. Overall, these studies detected negative consequences of overlapping on breeding performance, but there were large differences between them in which traits were affected.
The aim of this study was to investigate the reproductive consequences of moultbreeding overlap in a Mediterranean great tit (Parus major) population, where the proportion of adults overlapping both activities have increased throughout the years [18]. This has provided the opportunity to reach a reasonable sample size of naturally overlapping individuals. Thus, this is the first observational study dealing with the potential cost of reproduction of moult-breeding overlap in a non-migrant bird. We also worked with a notably high sample size in a single population (390 complete pairs), so the consequences of different combinations (only the female, only the male, both, or neither overlapping) could be compared. Based on previous studies, and on theoretical background [28,29], we would expect that pairs in which at least one member of the pair is overlapping would show smaller clutches, lower number of hatchlings and fledglings, and produce fledglings with poorer body condition.
Materials and Methods
This study was performed between 1995 and 2019 on a resident wild population of great tits breeding in nest boxes placed in an extensive orange (Citrus aurantium) mono- . The size of the study area has increased over the years, from approximately 150 ha to about 450 ha in recent years. The density of nest boxes has remained constant at 1 nest box per ha. The type and structure of the habitat have not changed along these years. However, the increasing length of spring means ambient temperatures throughout the study period caused and advancement of the breeding season and changes in the seasonal distribution of clutches [18], which is probably the cause of the increased degree of moult-breeding overlap [30].
Each nest box was inspected weekly throughout the breeding season (March to June), and daily by the time of expected clutch completion, hatching and fledgling dates, to record basic breeding parameters [18,31], such as laying date, clutch size and number of hatchlings and fledglings. We considered the laying date as the day the first egg was laid, assuming that one egg was laid per day. Dates are always presented as "April dates" (i.e., 1 April = Day 1). Laying was assumed to have finished when no more eggs were laid during two consecutive days and the female started full incubation. As parents remove dead small nestlings, those disappearing shortly after hatching were considered dead. Nestlings were individually ringed, their tarsus length measured (digital caliper to the nearest 0.1 mm), and their body mass recorded (digital balance to the nearest 0.1 g) when they were 15 days old. Then, the body condition of each nestling was computed as the body mass to tarsus length ratio [31,32]. Mean tarsus length, body mass and body condition of the nestlings of each nest were computed and used in the analyses. Nests were visited after the expected fledging date to look for dead nestlings and determine the number of fledglings. Parents are not able to remove large dead nestlings [33], so those dead by starvation remained in the nest. On the other hand, predators active in the study area kill and usually eat all the nestlings present. Mammals, such as weasels (Mustela nivalis) [34], garden dormouse (Eliomys quercinus) [35], or black rats (Rattus rattus) [36] leave different remains, usually including legs with the metal rings. The Montpellier snake (Malpolon monspessulanus) sometimes visits nests, swallowing nestlings whole and leaving no remains in the nest [37]. However, it usually acts on smaller (i.e., before ringing) nestlings and, in any case, the nest is left very flat; a somewhat different state as when nestlings fledge. In almost 30 years ringing nestlings in the study area, we have never recovered a bird alive after it was classified as "dead in the nest" due to its disappearance after an act of predation.
Parents were captured with door traps at the nestbox when nestlings were 10-12 days old. They were ringed with individually numbered metal rings if not already ringed. Sex and age (yearling or older) were recorded [38] and it was noted whether the bird was moulting its primary feathers or not [27]. Adult great tits have a complete postnuptial moult and they rarely moult feathers other than their primaries while breeding [39]. In our case, the most advanced birds were moulting their fourth primary while feeding 10-12-day-old nestlings. All the birds included in this study were raising their second (after a successful clutch) or replacement (after a failure of the first one) clutch; none of the individuals raising their first brood were found moulting [18].
Our aim in this study was to determine the reproductive consequences of overlapping breeding and moult, so we have only used data from years when there was at least one adult moulting while breeding. For each year, we considered that moult started the first day we caught the first adult moulting at least one of its primary wing feathers. No bird captured while raising its first brood was found moulting in any of the study years, so all the birds included here were actually attempting a second clutch, whether the first one had been successful or not. Birds attempting only one clutch could have started moult before our estimated "date of moult start" [39,40], but this does not affect our analyses since none of these birds overlapped breeding and moult. Thus, the date of laying of the first egg of the second clutch of the pair in which one of the components was first found moulting each year was taken as a benchmark, and all the pairs that started laying at this date or later were considered potential overlapping pairs for that year. These will be named "late-breeding pairs" hereafter.
Considering the above restriction, our dataset was split into four groups: nests where none of the parents were overlapping breeding and moult, nests where only the male was overlapping, nests where only the female was overlapping, and nests where both parents were overlapping. Only pairs where both parents were captured were included. To analyze the effect of overlapping on each variable, we used all the pairs for which the value of this particular variable was known; therefore, sample size might differ between analyses. As we will detail in the Results section, we used 390 complete pairs, in which male and female were individually identified, for our analyses. Of these pairs, only 33 pairs were considered in two (31 pairs) or three years (2 pairs), while 322 (83%) were considered only once. Additionally, 4 males were considered twice, and 8 were considered three times paired with different females; while 5 females were considered twice, and 7 were considered three times paired with different males. As these are relatively small numbers, the potential bias due to pseudo-replication would be very low, so we avoided performing more complex analyses (as mixed models).
In order to remove the effect of the year in our analyses, actual values for all the variables were transformed into z-scores by subtracting the mean and dividing by the standard deviation of that variable for that year. To have robust estimates of these statistics, we eliminated years in which fewer than 16 pairs could have overlapped breeding and moulting. Differences in reproductive parameters between the four types of pairs were tested using ANOVAs, followed by a posteriori Tukey's HSD when significant. In addition, as reproductive performance used to decrease through the season, all the analyses were repeated as ANCOVAs, using the z-values of laying dates as a covariate to distinguish the potential effects of overlapping moult and breeding from that of seasonal variation.
All analyses were conducted with the IBM SPSS Statistics 25 software
Incidence of Moult-Breeding Overlap
Given their breeding and moult start dates for each specific year, from 1995 to 2019, individuals from a total of 564 pairs could have overlapped breeding and moulting. From these, 390 complete pairs were captured while feeding nestlings. In 29.0 % of these pairs, none of the birds were moulting; both parents were moulting in 20.8 % of the cases; only the male in 45.6 %; and only the female in 4.6 % of the pairs. In other words, at least one member of the pair overlapped moult and reproduction in 71% of late-breeding pairs.
Considering only years in which at least 16 pairs were captured during the period of potential moult-breeding overlap, those pairs in which both members overlapped moult and breeding started laying later than the rest of the groups (ANOVA: F 1, 333 = 28.378; p < 0.001, and Tukey post-hoc pairwise comparisons; Figure 1). Animals 2023, 13, x 5 of 10
Reproductive Consequences of Moult-breeding Overlap
Clutch size differed between groups, being smaller when both parents overlapped than when only the male or neither parent did, while those cases in which only the female overlapped were intermediate, not differing from any or the other groups (ANOVA: F1,331 = 9.042; p < 0.001, and Tukey tests). However, these differences disappeared when laying date was added as a covariate into the analysis (ANCOVA, moult group: F3,331 = 2.537, p = 0.057; laying date: F1,331 = 41.130, p < 0.001; Figure 2).
Reproductive Consequences of Moult-Breeding Overlap
Clutch size differed between groups, being smaller when both parents overlapped than when only the male or neither parent did, while those cases in which only the female overlapped were intermediate, not differing from any or the other groups (ANOVA: F 1,331 = 9.042; p < 0.001, and Tukey tests). However, these differences disappeared when laying date was added as a covariate into the analysis (ANCOVA, moult group: F 3,331 = 2.537, p = 0.057; laying date: F 1,331 = 41.130, p < 0.001; Figure 2).
Reproductive Consequences of Moult-breeding Overlap
Clutch size differed between groups, being smaller when both parents overlapped than when only the male or neither parent did, while those cases in which only the female overlapped were intermediate, not differing from any or the other groups (ANOVA: F1,331 = 9.042; p < 0.001, and Tukey tests). However, these differences disappeared when laying date was added as a covariate into the analysis (ANCOVA, moult group: F3,331 = 2.537, p = 0.057; laying date: F1,331 = 41.130, p < 0.001; Figure 2). Results for clutch size were exactly the same for the number of hatchlings both when comparing directly the four types of pairs (ANOVA: F 1,332 = 5.208; p = 0.002, and Tukey tests) and when laying date was added as a covariate (ANCOVA, moult group: F 3,331 = 1.339, p = 0.262; laying date: F 1,332 = 25.631, p < 0.001).
The number of fledglings produced was clearly lower for pairs where both parents overlapped than for those where none of them did it, being intermediate for those in which only one of the parents (either male or female) overlapped (ANOVA: F 1,333 = 3.217; p = 0.023, and Tukey tests) but, again, the differences disappeared when laying date was included as a covariate (ANCOVA, moult group: F 3,331 = 0.505, p = 0.679; laying date: F 1,332 = 35.570, p < 0.001; Figure 3). Results for clutch size were exactly the same for the number of hatchlings both when comparing directly the four types of pairs (ANOVA: F1,332 = 5.208; p = 0.002, and Tukey tests) and when laying date was added as a covariate (ANCOVA, moult group: F3,331 = 1.339, p = 0.262; laying date: F1,332 = 25.631, p < 0.001).
The number of fledglings produced was clearly lower for pairs where both parents overlapped than for those where none of them did it, being intermediate for those in which only one of the parents (either male or female) overlapped (ANOVA: F1,333 = 3.217; p = 0.023, and Tukey tests) but, again, the differences disappeared when laying date was included as a covariate (ANCOVA, moult group: F3,331 = 0.505, p = 0.679; laying date: F1,332 = 35.570, p < 0.001; Figure 3).
Discussion
It is well-known that males tend to start moulting earlier than females, and thus they have been found to overlap these two activities more frequently [18,25,39,41,42]. The usual explanation is that females are more energy-limited during the first stages of breeding, as they spend much more time incubating or brooding the hatchlings [11]. In our sample of 390 pairs, 259 (66%) males were moulting, while only 99 (25%) females overlapped breeding and moult. This should be considered as a conservative figure, since parents were
Discussion
It is well-known that males tend to start moulting earlier than females, and thus they have been found to overlap these two activities more frequently [18,25,39,41,42]. The usual explanation is that females are more energy-limited during the first stages of breeding, as they spend much more time incubating or brooding the hatchlings [11]. In our sample of 390 pairs, 259 (66%) males were moulting, while only 99 (25%) females overlapped breeding and moult. This should be considered as a conservative figure, since parents were trapped when nestlings were 10-12 days old, so individuals starting moulting during the second half of the nestling period remained undetected.
An important difference of our study with previous observational ones is that great tits are resident, i.e., do not have the pressure to moult early or quickly in order to migrate. We have previously shown that, in the studied population, the increase of pairs laying two clutches, probably due to recent ambient temperature increase, has also increased the number of pairs overlapping breeding and moult [18]. As presented above, results from both observational and experimental studies led us to expect a reproductive cost of such overlap of two energy-demanding activities. Our main question in this study was to test this expectation. We actually found that overlapping pairs showed a poorer breeding performance than non-overlapping ones, with generally intermediate scores for pairs in which only one of the members overlapped. However, this reduced performance is mostly explained by the seasonal decreasing trend. As overlapping breeding and moult was more frequent as the season progressed, pairs breeding late overlapped more frequently and showed reduced breeding performance, but there was not an observable additional cost of the moult-breeding overlap. We can only conclude from the available data that individuals overlap breeding and moulting if they can afford to do so, without imposing additional costs on their current breeding. We will discuss in turn the studied breeding parameters.
The first one that could be negatively affected by allocating resources to moult is clutch size. Although we declared that, based on previous studies, our expectation was that clutch size would be smaller in pairs where overlap occurred, the details of these studies merit a closer look. There are, to the best of our knowledge, two studies dealing with the effect of overlapping on clutch size, both in pied flycatchers. The first study [25] gave a clear result for a Swedish population: pairs where both parents were moulting showed smaller clutches than those in which only the male, or none of the parents, were moulting; those in which only the female was moulting showed and intermediate value, suggesting that clutch size was more affected if females were overlapping. These differences were maintained when the effects of the seasonal decrease in clutch size were removed. The second study [26] is more difficult to interpret, since it did not consider pairs but individuals, separately analyzing males and females. In none of the four European populations considered (including the Swedish one mentioned above) were differences in clutch size found between overlapping and non-overlapping females. The results with males were more striking, since in two populations clutch size was smaller in nests attended by moulting males, while in the other two the opposite was true. However, these differences disappeared when laying date was considered. We should observe that in both studies [25,26], the population of Abisko (Sweden) was included and the same data used, with totally opposite results (negative effects of overlapping in the first, no effect in the second). This difference could be only attributed to the way the data were analysed. Our results with great tits generally agree with those of Hemborg et al. (2001) [26], in that clutch size was not directly affected by moult status. In our case, if both male and female were overlapping, clutch size was lower, but this reduction of clutch size could be attributed to the later breeding dates of these pairs.
We found that the number of hatchlings produced by the different groups of pairs mirrored those of clutch size, indicating that there was no effect on incubation efficiency. The only study allowing a sensible comparison is that of Hemborg (1998) [20], an experimental study where breeding time was either advanced or delayed by exchanging clutches during incubation. In this study, clutch size could obviously not be affected, and hatching success was similar between overlapping and non-overlapping pairs. Results of this study were, therefore, in agreement with those presented here for Spanish great tits, suggesting that overlapping incubation and moult does not affect female incubation efficiency. This reinforces the idea that the effect of overlap on clutch size would be low, as the incubation period is very costly for the female. If there were negative effects on egg production, one would also expect to find costs during incubation when, if moulting has already started, moulting would be more resource intensive.
As it was stressed above, parents were captured and their moulting state assessed when their nestlings were 10-12 days old. Considering 7 days for laying eggs, 13 days for incubation and 10 days of the nestling period until capture of the parents, this means that the moult state was assessed about one month after the clutch initiation date. As it is the case that the most advanced birds were moulting their fourth primary when captured, and many of them were still moulting their first or second ones, the moulting process of most of them could not have been started, or was in its very early phase, when the clutch was produced. Thus, the lack of negative effects attributable to the overlap could be simply due to a lack of an extensive overlap at this time. Contrasting with this, parents found moulting when captured clearly overlapped this activity with the feeding of their young, so negative consequences on the number and condition of fledglings would be more probable to occur. We do not have any direct comparison to perform, since the other two observational studies available did not directly consider the number of fledglings produced. Hemborg et al., (2001) [26] found that breeding success (proportion of eggs producing fledglings) did not differ between moulting and non-moulting females in any of four studied European pied flycatcher populations, while it was lower for moulting males. The experimental studies, on the other hand, do not provide a clear guide, since two of them [20,23] found a decrease in the number of fledglings, while the other two failed to detect this effect [19,24]. Each study had a different experimental protocol, for instance, delaying breeding time, plucking feathers to simulate moult, providing extra food, altering brood size, etc., at different moments of the breeding cycle. Therefore, the effects of overlapping could be misleaded with the effects of manipulation. For example, Hemborg (1998) [20] found that delayed pairs (supposedly overlapping more frequently) produced fewer fledglings, but the moult status of the birds did not affect fledgling number. Our results are in full agreement with this conclusion: the lower number of fledglings produced by overlapping Spanish great tits is mainly due to the fact that they breed later in the season, not because they are moulting.
Some experimental studies have shown that overlapping birds produced lighter fledglings or fledglings in poorer body condition [10,21,22]. Similar to our results on other breeding parameters, nestling body condition did not differ between moulting and non-moulting pairs once the laying date was taken into account.
Conclusions
In summary, we have observed that the overlap of breeding and moulting does not impose immediate reproductive costs on birds that moult "voluntarily" (i.e., are not induced to moult) while breeding. Obviously, these costs could manifest themselves later, as differential survival of fledglings and/or as differential survival and/or future reproductive costs for adults. This has been shown in some studies, both experimental [10,24] and observational [17,25,27], but not in others [19][20][21][22][23]. Funding: This study was partially supported by the project CGL2013-48001-C2-1-P and PID2021-122171NB-I00/ AEI/10.13039/501100011033/ FEDER, UE.
Institutional Review Board Statement: Specific review and approval were waived for this study since it was implicit in the project approval (CGL2013-48001-C2-1-P and PID2021-122171NB-I00; Spanish Ministry of Science and Innovation). All the authors had the necessary permits from national and regional authorities to manipulate and ring wild birds.
Informed Consent Statement: Not applicable.
Data Availability Statement: The data presented in this study are openly available in Zenodo at DOI, 10.5281/zenodo.7398652.
|
2023-01-28T16:09:10.131Z
|
2023-01-26T00:00:00.000
|
{
"year": 2023,
"sha1": "8e5d2c16df6b0a37b9b5789329f038e9b113efe6",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2076-2615/13/3/409/pdf?version=1674712459",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "61e263e8f54851c05fef2b651a7a98b9b0f5fcae",
"s2fieldsofstudy": [
"Biology",
"Environmental Science"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
18493345
|
pes2o/s2orc
|
v3-fos-license
|
On the number of limit cycles which appear by perturbation of Hamiltonian two-saddle cycles of planar vector fields
We find an upper bound to the maximal number of limit cycles, which bifurcate from a hamiltonian two-saddle loop of an analytic vector field, under an analytic deformation.
Introduction
Consider a N -parameter analytic family of analytic plane vector elds X λ , λ ∈ (R N , 0), such that X 0 has a k-saddle cycle (a hyperbolic k-graphic) Γ k , as on g.1. The cyclicity Cycl(Γ k , X λ ) of Γ k is, roughly speaking, the maximal number of limit cycles of X λ which tend to Γ k as λ → 0. The rst results on the cyclicity of one-saddle connection (also called homoclinic connection) go back to Andronov and Leontovich in 1937 (but they were published only in 1959 [1]). The cyclicity Cycl(Γ 1 ) has been studied later in full generality by Roussarie [17,18], see also [19,12] for an extensive list of references.
The main technical tool of the Roussarie's method is an asymptotic expansion of the Dulac map (transport map near the saddle point) in terms of x k , x k ω(x, ε) where ω(x, ε) is the so called Ecalle-Roussarie compensator ω(x, ε) = x −ε − 1 ε , ω(x, 0) = − ln x, x ∈ (R + , 0) and ε = ε(λ) is the trace of the vector eld X λ at the saddle point. Let P λ be the Poincaré rst return map, associated to Γ 1 and X λ . The most delicate case to be studied is when Γ 1 is of innite co-dimension (P 0 = id).
For λ ∼ 0 the map P λ is composed by a Dulac map (near the saddle point) and an analytic map (the transport map along the homoclinic orbit). The usual derivation-division algorithm then provides an upper bound for the cyclicity in terms of the number of the coecients of the asymptotic series of the displacement map P λ − id, which vanish as λ = 0. The same method was applied more recently to one-parameter deformations of Hamiltonian twosaddle loops (called also heteroclinic Hamiltonian connections), under the non-generic assumption that one of the separatrices of Γ 2 remains unbroken [4,3]. Recall that a k-saddle cycle Γ k is said to be Hamiltonian, provided that there is a neighborhood of Γ k in which X 0 allows an analytic rst integral with only Morse critical points. The purpose of the present paper is to extend these results to the case of an arbitrary analytic perturbation of a Hamiltonian two-loop Γ 2 , having two hyperbolic equilibrium points.
Our approach is dierent, as we do not use the asymptotic series of the corresponding Dulac maps d 1 λ and d 2 λ , shown on g. 4). Recall that in the one-parameter case λ = ε ∈ (R, 0), the displacement function d 1 ε − d 2 ε can be approximated by an appropriate Abelian integral I(.) (or more generally, an iterated path integral) depending on a parameter t as follows d 1 ε (t) − d 2 (t) ε = ε d I(t) + . . . , ε ∼ 0. Therefore to count the zeros of d 1 ε − d 2 ε (corresponding to limit cycles) it is enough to count the zeros of I(.), which can be done by making use of the so called "Petrov trick" (based on the argument principle), see [15] and the references given there.
The above considerations hold true at least far from the singular points of the vector eld X λ . As discovered in [4,3], however, not all limit cycles in a neighborhood of a two-saddle loop can be approximated in such a way. The missing "alien" limit cycles are moreover non-avoidable in generic Nparameter deformations X λ with N ≥ 4. For this reason, we apply the argument principle directly to the displacement map d 1 ε (.) − d 2 (.) ε in an appropriate complex domain, in order to obtain an estimate to the number of its complex zeros (corresponding to complex limit cycles).
The main technical result of the paper is Lemma 2, which claims that the zero locus of the imaginary part of the Dulac map is a real analytic curve of Figure 1: Hamiltonian k-saddle cycles R 2 = C at the origin. This makes possible to investigate the number of the zeros of the imaginary part of d 1 λ − d 2 λ along the zero locus of the imaginary part of d 1 λ . Indeed, the intersection numbers of two analytic curves is easily computed. The proof of our main result, Theorem 1, is then completed by making use of the Petrov trick.
In the course of the proof of Theorem 1 we assume, for the sake of simplicity, that our deformations depend on a single small parameter ε. General multi-parameter deformations λ → X λ of X 0 are then studied along the same lines, as it follows from the Hironaka's desingularization theorem. We explain this in Appendix B.2, see Theorem 4.
Deformations of an arbitrary (possibly non-Hamiltonian) two-loop of innite co-dimension (P λ = id) can be studied in a similar way, and will be considered in another paper. 2 Description of the result Let X 0 be a real plane vector eld. Recall that a polycycle of X 0 is a topological polygon composed of separatrices and singular points. A k-saddle cycle of X 0 (or a hyperbolic k-graphic) denoted Γ k , is a polycycle composed of k distinct saddle-type singular points p 1 , p 2 , . . . , p k , p k+1 = p 1 and separatrices (heteroclinic orbits) connecting p i to p i+1 . Let σ be a segment transversal to the polycycle. The k-saddle cycle is said to be Hamiltonian, provided that X 0 has an analytic rst integral f having Morse critical points at p i . It follows that Γ k bounds an annulus of periodic orbits {(x, y) : f (x, y) = t} t of X 0 . Thus, a Hamiltonian 0-saddle cycle is simply a center, a Hamiltonian 1-saddle cycle is a homoclinic loop bounding a period annulus, a Hamiltonian 2-saddle cycle is a double heteroclinic loop bounding a period annulus etc., see g.1.
One can nd a "tubular neighborhood" such that Ū is compact smooth manifold with a (real three-dimensional) border.
f is analytic in some neighborhood ofŪ the border ∂Ū is transversal to the complex curves {(x, y) ∈ C 2 : f (x, y) = t}, provided that |t| is suciently small.
It follows that f : U → C denes a locally trivial bration over a punctured neighborhood of the origin in C, and each ber is homeomorphic to a genus one surface with k punctures. A one-parameter analytic deformation of X 0 is a a family X ε of realanalytic plane vector elds, depending analytically on a real parameter ε ∈ (R, 0), and dened in a suitable neighborhood of the k-saddle cycle Γ k . The corresponding foliation F ε has an extension in a complex domain denoted by the same letter, and dened by where ω ε = P (x, y, ε)dx + Q(x, y, ε)dy is a one-form, and P, Q are realanalytic in x, y, ε in a neighborhood of Γ k . Parameterize the segment σ by the "synchronized" local variable t = f | σ and let γ(t) ⊂ F t , t > 0 be the continuous family of periodic orbits of X 0 which tend to the polycycle Γ k ⊂ F 0 as t tends to 0. To the family {γ(t)} we associate the trivial rst return map P 0 : σ → σ, P 0 = id which allows an analytic continuation for ε = 0 to a rst return map The dots above mean a function in t, ε which, for every xed t such that P ε (t), ε ∼ 0 is dened, is of the type 0(ε d+1 ). The so called Poincaré-Pontryagin function function M d may be explicitly computed, see [16,5,6]. More generally, let γ(t) ⊂ F t be any continuous family of closed loops intersecting the cross-section σ. For |ε| suciently small we dene in a similar way the holonomy map h ε γ : σ → σ related to the family of loops {γ(t)} t and the deformed foliation F ε . By analogy to the Poincaré return map we have where the meaning of the dots is as before, and the number d depends on {γ(t)} t and F ε . The holonomy map h ε γ depends on the choice of σ. In contrast to this, the Poincaré-Pontryagin function M d does not depend on the cross-section σ, it depends on the free homotopy class of the loop γ(t) only. Further, it can be expressed in terms of iterated path integrals of length at most d, along suitable meromorphic dierential one-forms. It satises therefore a linear dierential equation which has a Fuchs type singularity at t = 0, see [7,9]. Thus, the leading term of M d has the form In the Hamiltonian case the number ν(P ε ) is rational, because the corresponding monodromy operator is quasi-unipotent [10].
To formulate the main result of the paper consider, more specically, the case k = 2 ( a double heteroclinic loop). As it follows from [9], the function M d is in fact an Abelian integral and can be written as for suitable analytic one-formω.
Let δ 1 (t), δ 2 (t) be two continuous families of closed loops vanishing at the saddle points p 1 and p 2 . We suppose that orientations of the loops "agree" in the sense that the intersection indices of the homotopy classes of δ 1 (t), δ 2 (t) with the homotopy class of the periodic orbit γ(t) is one and the same.
The cyclicity Cycl(Γ 2 , F ε ) of the 2-saddle cycle Γ 2 with respect to the deformed foliation F ε is the maximal number of limit cycles which bifurcate from Γ 2 near ε = 0, see [19] for a precise denition. An upper bound for the cyclicity Cycl(Γ 2 , F ε ) is given in terms of the characteristic numbers of the holonomies associated to Γ 2 as follows Theorem 1 It is tempting to conjecture that in general the cyclicity Cycl(Γ 2 , F ε ) is bounded by a similar expression in terms of the characteristic numbers of the holonomies associated to Γ k . Indeed, in the homoclinic case, k = 1, by repeating the proof of Theorem1 one obtains We have, typically for some analytic functions f 1 , f 2 , then By the Roussarie's theorem [17, Theorem C] the exact upper bound in a real domain in this case is 2p if p < q, and 2q − 1 if p ≥ q. This suggests that the bound of Theorem 1 can be improved. In fact, the bound (5) holds true for the number of complex limit cycles accumulating on Γ 2 in a suitable neighborhood of it.
Example. Suppose that d = 1 in (3), that is to say where the functions f 1 , f 2 , f 3 are analytic in a neighborhood of t = 0, Suppose further that Theorem 1 implies that the cyclicity of Γ 2 is bounded by In the case p 1 = p 2 = q = p for instance, this gives In this situation, and under the strong hypothesis that one of the connections of Γ 2 remains unbroken, it has been proved in [4,Theorem 8] that
The Dulac map
Let F ε be a real analytic foliation dened by (2) in a neighborhood of a hyperbolic Morse critical point of the function f . For all suciently small Figure 2: The Dulac map |ε| the foliation F ε has a singular point of saddle type, to which we associate a Dulac map (or the transition map), as on g.2 (i). More precisely, for all suciently small |ε| the foliation F ε has two separatrix solutions, which are transversal analytic curves, depending analytically on ε. We may suppose that they are the axes {x = 0} and {y = 0} as on g.2 (ii). Let σ, τ be two complex cross-sections (complex discs) to the two separatrices, parameterized by z = f | σ and z = f | τ . In these coordinates the Dulac map is the germ of analytic map d ε : (R + * , 0) → (R + * , 0) dened as follows: if z ∈ σ ∩R + * then d ε (z) ∈ τ ∩R + * is the intersection with τ of the orbit γ ε (z) of (2), passing through z ∈ σ. This geometric denition of d ε allows to control to a certain extent its analytic continuation in a complex domain.
Analytic continuation
The Dulac map is analytic and hence allows an analytic continuation on some open subset of the universal covering σ • of σ \ {0}. The domain of the continuation depends on ε, and obviously d 0 (z) ≡ z. Let us parameterize the universal covering σ • by polar coordinates ρ > 0, ϕ ∈ R, z = ρ exp ϕ.
Theorem 2 There exists ε 0 > 0 and a continuous function such that the Dulac map allows an analytic continuation in the domain The proof of the above Theorem in the 0-parameter case is well known, and in the multi-parameter case it is the same. For convenience of the reader it will be given in Appendix A. This proof shows even more: the analytic continuation of the Dulac map in the domain (6) can be accomplished in a geometric way as follows.
is the real orbit of df = 0 contained in the rst quadrant x ≥ 0, y ≥ 0. We note that, although the family {γ 0 (z)} z is not unique, the relative homotopy class of each loop γ 0 (z) is uniquely dened for all z ∈ σ • . It follows from the proof of Theorem 2 that {γ 0 (z)} z allows a deformation to a family of paths {γ ε (z)} z , connecting σ • to τ • , tangent to the leaves of F ε , and dened for all ε, z in the domain (6).
3.2
The Poincaré- Pontryagin integral In what follows a crucial role will be played by the integral γ 0 ω 0 , and its generalizations. Namely, let be a compact set, where ρ(ϕ) is as in Theorem 2. As γε(z) ω ε is continuous in z, ε and ω ε then the following Lemma holds uniformly in z ∈ K.
The function γ 0 (z) ω 0 is the so called Poincaré-Pontryagin integral associated to the deformed foliation F ε . It follows from the argument principle that, that if |ε| is suciently small, the number of the zeros of d ε (z) − z in the compact K is bounded by the number of the zeros of the Poincaré-Pontryagin integral γ 0 (z) ω 0 in K (counted with multiplicity). It might happen, however, that the Poincaré-Pontryagin integral vanishes identically. In all cases there is an integer d ≥ 1 and an analytic function uniformly in z ∈ K, provided that the Dulac map is not the identity map. M d is the so called higher order Poincaré-Pontryagin function and its zeros in K bound as before the number of the zeros of d ε (z)−z. As we already mentioned in section 2, there is an integral representation for M d as an iterated integral of length at most d along γ 0 (z).
Our aim is to obtain a bound for the zeros of d ε (z) − z in a domain K which is open and connected. Even if the estimate (7), (8) do allow an extension to such a domain K, the argument principle can not be directly used. For this purpose we consider rather the imaginary part of the Dulac map.
3.3
The zero locus of the imaginary part of the Dulac map We shall describe the zero locus of the imaginary part of the Dulac map d ε in an appropriate sector The surprising fact about H ε is that it is a smooth real-analytic plane curve in D ⊂ R 2 = C. Even better, the curve H ε can be conveniently approximated in terms of higher order Poincaré-Pontryagin functions. The foliation F 0 has a rst integral dening a bration with bers F t , see (1). Let δ(t) ⊂ F t be a continuous family of closed loops δ(t) ⊂ F t vanishing at the saddle point when t tends to 0. The orientation of δ(t) is chosen as follows. Let γ 0 (t) be the family of loops dened in the Appendix A. For real positive t they coincide with real orbits of F 0 connecting σ + to τ + . Then, the homotopy classes of γ 0 , δ satisfy Therefore, the exact orientation of δ(t) can be computed by the Picard-Lefschetz formula (but we do not need this).
Let τ be, as before, a cross section to the ber F 0 , see g. 2. Consider the holonomy map h ε δ associated to the family {δ(t)} t and to the deformed The anti-holomorphic involution induces, for t ∈ R an anti-holomorphic involution where R(u, ε) is an analytic function.
The above Lemma is the main technical result of the present paper. The analyticity of the zero locus H ε is responsible for the algebraic-like behavior of the Dulac map.
Proof of Lemma 2. Consider the cross-sections (complex discs transversal to the separatrices) σ ± , τ ± as shown on g.3, simultaneously parameterized as before by the restriction z of the rst integral f (x, y) on them. The crosssections σ, τ shown on g.2 are denoted, from now on, by σ + , τ + . Denote where α ε (z) is a path connecting z ∈ σ + to a point on σ − <0 and β ε (z) is a path connecting the latter point to a point on τ + <0 , where β ε (z) ⊂ R 2 . It follows that H ε is the image of σ − <0 under the holonomy map This already proves that the closure of H ε ⊂ R 2 is a smooth analytic curve. Once having said this, it is clear that H ε can be conveniently parameterized, which we do next. To prove (12), let us note that for t ∈ R the two ends of the loop α 0 (t) are real and hence this also holds true for the complex-conjugate loop α 0 (t). The loop α 0 (t) • α −1 0 (t) is therefore closed and is homotopic to δ(t) ⊂ F t dened above.
More explicitely
This, together with (11) and where, by abuse of notation, O(ε) means a function analytic in t, ε, which vanishes identically for ε = 0. This proves the identity (9).
Cyclicity of two-saddle cycles
In this section we prove Theorem 1. Using the notations introduced in section 2, we suppose that the vector eld X 0 has a two-saddle loop Γ 2 and an analytic rst integral f in a neighborhood of it. We suppose that f has Morse critical points at the two saddle points p 1 , p 2 of X 0 . Consider the Dulac maps d 1 ε , d 2 ε associated to the corresponding foliation, see g.4. Each map d i ε is a composition of a "local" Dulac map (as in section 3) and two holomorphic holonomy maps. From this it follows that Lemma 2 applies to d i ε , i = 1, 2, too. We parameterize each cross-section by the restriction z of f on it. The function d i ε , i = 1, 2, is multivalued and has a critical point at s i (ε) ∈ R, s i (0) = 0. The functions s i are real analytic. We consider rst the case ε > 0 and we may suppose that s i (0) = 0, s 1 (ε) < s 2 (ε) for all suciently small ε, (the case ε < 0 is studied in the same way). Our aim is to bound the number of those zeros of the displacement map d 1 ε − d 2 ε which are real, bigger than s 2 (ε) and tend to 0 as ε tends to 0. Note that these zeros correspond to the xed points of the Poincaré rst return map P ε . Indeed, We shall count the zeros of the displacement map in the larger complex domain D ε of the universal covering of C \ {s 1 (ε)} dened as follows. It is bounded by the circle by the interval [s 1 (ε), s 2 (ε)], and by the zero locus of the imaginary part of the Dulac map d 1 ε for (z) < s 1 (ε), as it is shown on g.5. The numbers ε, R are subject to certain conditions explained bellow. The zeros of an analytic function in a complex domain equal the increase of the argument of the function along the border of the domain, divided by 2π (the argument principle). To bound the increase of the argument we shall count the number of the zeros of the imaginary part of the function, along the border of the domain.
Choose rst the real numbers ε 0 , R > 0 as follows. Let and let z ν (log z) µ be the leading term of M d (z). Then, by (13), ν = ν(P ε ) is the characteristic number of the Poincaré map P ε associated to Γ 2 . We choose R > 0 so small, that the increase of the argument of M d (z) along the circle S R is suciently close to the increase of the argument of z ν (log z)µ SR s1(ε) s2(ε) Figure 5: The domain D ε along S R . We x R and choose ε 0 > 0 so small with respect to R, that for all ε, |ε| < ε 0 , the increase of the argument of d 1 ε (z) − d 2 ε (z) along the circle S R is suciently close to the increase of the argument of M d (z) along S R . This is indeed possible, according to Lemma 7. The conditions that we impose on ε 0 , ε and R will be denoted (by abuse of notations) as follows To evaluate the increase of the argument of d 1 ε (z)−d 2 ε (z) along the interval [s 1 (ε), s 2 (ε)], we bound the zeros of its imaginary part which equals (along the interval [s 1 (ε), s 2 (ε)]) to the imaginary part of −d 2 ε (z). In other words, we need to estimate the number of intersection points (counted with multiplicity) between the zero locus of the imaginary part of d 2 ε (z) and [s 1 (ε), s 2 (ε)]. According to Lemma 2 this number of intersection points is bounded by the multiplicity of the zero at the origin of the Poincaré-Pontryagin function of the holonomy map h ε δ 2 . This multiplicity equals ν(h ε δ 2 ). Finally, we arrive at the most delicate point in the proof of Theorem 1 : evaluate the increase of the argument of d 1 ε (z) − d 2 ε (z) along the zero locus of the imaginary part of the Dulac map d 1 ε for (z) < s 1 (ε). For this purpose we bound the zeros of the imaginary part Im (d 1 ε (z) − d 2 ε (z)), along the zero locus of Im d 2 ε (z). Thus, we need to estimate the number of intersection points (counted with multiplicity) between the zero locus of the imaginary part of d 2 ε (z), and the zero locus of the imaginary part of d 1 ε (z). Recall that to a Dulac map d i ε we associated a family of vanishing loops δ i (z) with orientation prescribed by (10). With this convention, the orientation of the loops δ 1 and δ 2 do not agree : if γ(t) is the family of periodic orbits of F 0 for t > 0, then the intersection indices of the homotopy classes of δ 1 (t), δ 2 (t) with γ(t) have opposite signs. In order to have the same convention as in the formulation of Theorem 1 we reverse the orientation of δ 2 . With this convention if then, by Lemma 2, the zero locus of the imaginary part of the holonomy maps h ε δ 1 , h ε δ 2 is given by respectively, where R 1 , R 2 are appropriate analytic functions. We conclude that the number of intersection points of the above analytic curves coincides with the multiplicity of the zero at the origin of either M d 1 . The number of intersection points equals therefore to the characteristic number ν(h ε δ 1 • h ε δ 2 ). Summing up the above information we get that the increase of the argument of d 1 ε (z) − d 2 ε (z) along the boundary of the complex domain D ε is not bigger than ν(P ε ) + ν(h ε δ 2 ) + ν(h ε δ 1 • h ε δ 2 ) + 2. The above estimate can be slightly improved, by taking into consideration the fact that the imaginary part of d 1 ε (z) − d 2 ε (z) vanishes at s 2 (ε). Theorem 1 is proved.
A
Proof of Theorem 2 In this appendix we will prove Theorem 2 in the slightly more general context of multi-parameter analytic deformations. This will be used in Appendix B.
Note that the zero-parameter case is well known [13,19]. Consider a N -parameter analytic family of analytic plane vector elds X λ , λ ∈ (R N , 0), such that X 0 has a hyperbolic singular point at the origin. It is known, since Briot and Bouquet, that X 0 has two transversal invariant analytic curves which can be supposed to coincide with the axes x = 0 and y = 0, that is to say The proof is as follows : a formal change of the variables rst removes some (but not all) non-resonant terms, and then one veries the convergency of the transformation, see [14, Appendice II]. Exactly the same proof applies, however, to the family X λ . One can show in this way that X λ is analytically orbitally equivalent to the following (slightly improved) normal form where r = r(λ), a = a(x, y, λ) are appropriate analytic functions in their arguments, r(0) < 0, see [18, Appendice 1]. We shall suppose, without loss of generality, that there exists a constant c > 0 such that r, a are analytic in the complex domain D c = {x, y, λ) : |x| < 2, |y| < 2, |λ| < c}.
After a further linear re-scaling of x, y, we may suppose that |a(x, y, λ)| is so small in D c , that r(λ) + xy.a(x, y, λ) = 0 .
After this preparation, choose the cross-sections and consider the corresponding Dulac map To prove Theorem 2 we have to show that the constant c > 0 can be chosen in such a way, that for every ϕ 0 > 0 there exists 0 < z 0 < 1, such that the Dulac map allows an analytic continuation in the sector The proof is similar to the proof of [19,Theorem 7], the only dierence being the presence of the parameter λ. We shall construct a continuous family of paths γ λ (z) contained in the leaves of the foliation F λ dened by the vector eld X λ in C 2 . Each path γ λ (z) starts at the point (x = z, y = 1) and ends at the point (x = 1, y = d λ (z). The path γ λ (z) is constructed by lifting the path l(z) contained in the x-plane {y = 0} and shown on g.6, with respect to the projection π : C 2 → C (x, y) → x Indeed, the foliation F λ is transversal to the projection π except along the leaf x = 0, provided that r + xy.a = 0. The resulting path path γ λ (z) is shown on g. 7.
To prove the existence of γ λ (z), consider the solution y = y(x) associated to the vector eld X λ , with initial condition y(z) = 1. We have to show that the solution y = y(x) exists when x is restricted to the path l(z). The path l(z) is composed by an arc and a segment. We consider them separately Along the arc x = |z|e iϕ , y = 0, 0 < ϕ < arg(z) Therefore, if |z| < z 0 is suciently small, then |y(|z|e iϕ )| < 2 when 0 < ϕ < arg(z) we have similarly xdy = −y(r + xy.a)dx xd|y| = −|y|(r + (xy.a))dx.
B
Multi-parameter deformations of Hamiltonian two-saddle loops Consider, as in the preceding Appendix, a N -parameter analytic family of analytic plane vector elds X λ , λ ∈ (R N , 0). We suppose that X 0 has a Hamiltonian two-saddle loop Γ 2 bounding a period annulus. This case is easily reduced to the one-parameter case studied in the present paper by making use of a standard procedure based on the Hironaka desingularization theorem. In this Appendix we indicate the main steps.
B.1 Principalization of the Bautin ideal
In this section we follow [8,20]. Let z 0 ∈ σ, z 0 ∈ Γ 2 and consider the Poincaré map The Bautin ideal, associated to P λ , is the ideal I =< a i >, generated by the germs of the analytic coecients a i (.). It is Noetherian, so generated by a nite number of coecients, and moreover does not depend on the choice of z 0 [19]. More generally, let {γ(z)} be any continuous family of closed loops in the bers of the foliation F 0 , intersecting the cross-section σ. For all suciently small λ the (germ of) holonomy map h λ γ : σ → σ is dened. In the same way, we associate to h λ γ (z) a Bautin ideal, generated by the coecients a i (λ) of the expansion of h λ γ (z) with respect to z − z 0 . As before it is Noetherian, and does not depend on the choice of z 0 (with the same proof).
We may assume, without any loss of generality, that the Bautin ideal is principal. For this we use a variant of Hironaka's desingularization theorem as follows.
Let ϕ 0 , ϕ 1 , . . . , ϕ p be non-zero analytic functions on a smooth complex or real analytic variety X. The indeterminacy points of the rational map ϕ : X P p can be eliminated as follows [11,2] Theorem 3 (Hironaka desingularization) There exists a smooth analytic varietyX and a proper analytic map π :X → X such that the induced mapφ = ϕ • π is analytic.X π φ @ @ @ @ @ @ @ @ Let O X be the sheaf of analytic functions on X and consider the ideal sheaf I ⊂ O X generated by ϕ 0 , ϕ 1 , . . . , ϕ p . The inverse image ideal sheaf of I under the map π :X → X will be denoted π * I. This is the ideal sheaf generated by the pull-backs of local sections of I. We note that π * I may dier from the usual sheaf-theoretic pull-back, also commonly denoted by π * I. A simple consequence of Theorem 3 is the following Corollary 1 The inverse image ideal sheaf π * I is principal. This is called the principalization of I. Indeed, as the induced mapφ is analytic, then for everyλ ∈X there exists j, such that the functionsφ i /φ j , i = 1, 2, . . . , p, are analytic in a neighborhood ofλ. Therefore there is a neighborhoodŨ ofλ such thatφ j |Ũ dividesφ i |Ũ in the ring of sections OŨ of the sheaf OX, that is to say IŨ is generated byφ j |Ũ .
In our context X = σ is the cross-section to the family of periodic orbits {γ(z)} and ϕ 0 , ϕ 1 , . . . , ϕ p are the germs of analytic functions which generate the Bautin ideal associated to the holonomy map h λ γ . To apply Theorem 3 we assume that σ is a polydisc on which ϕ i are analytic, and the divisors (ϕ i ) intersect transversally the boundary of σ. After applying the Hironaka's theorem, the origin 0 of σ is replaced by a compact divisor π −1 (0), along which the inverse image ideal sheaf π * I is principal.
Suppose now that we have a holonomy map h λ δ associated to another family of periodic orbits δ(t). Let ψ 0 , ψ 1 , . . . , ψ q be generators of the corresponding Bautin ideal J. As before we assume that ψ i are analytic on σ, with divisors transversal to its boundary. Applying twice the Hironaka's theorem we get a new analytic variety smooth analytic varietyX and a proper analytic map π :X → X such that the induced mapsφ = ϕ • π andψ = ψ • π are analytic, see diagram (19).
The inverse image ideal sheaf π * I and π * J are both principal along the compact divisor π −1 (0).
B.2
Multi-parameter version of Theorem 1 We begin by formulating the multi-parameter version of Lemma 2, let {δ(t)} be the family of vanishing loops dened in section 3.3. In agreement with the preceding section, let us suppose that the Bautin ideal of the holonomy map h λ δ is principal. We have, therefore (compare to (11) whereR(., .) is analytic,R(z, 0) = 0, andM (.) is the highest order Poincaré-Pontryagin function.
Lemma 3 The zero locus H ε of the imaginary part of the Dulac map is a smooth real-analytic curve of R 2 = C of the form (M (u) + R(u, λ)), u < 0} ∩ D where R(u, λ) is an analytic function, R(u, 0) = 0.
The proof of the above Lemma is completely analogous to that of Lemma 2 and is therefore omitted.
In the proof of Theorem 1 we used four Bautin ideals associated to the holonomy maps where {γ(z)} is the family of periodic orbits associated to the annulus, {δ 1 (z)}, {δ 2 (z)} are the vanishing families of loops associated to the saddle points. After an appropriate blow up π we may suppose that the inverse image ideal sheafs of the corresponding four Bautin ideals are principal along the compact divisor π −1 (0) ⊂X, see section B.1. Letλ be a local variable on the smooth varietyX. The cyclicity Cycl(Γ 2 , (Fλ, Fλ 0 )) is the maximal number of limit cycles which Fλ can have in an arbitrarily small neighborhood of Γ 2 , whenλ tends toλ 0 . Denote also Cycl(Γ 2 , F λ ) = Cycl(Γ 2 , (F λ , F 0 ).
|
2009-09-29T19:43:40.000Z
|
2009-06-23T00:00:00.000
|
{
"year": 2009,
"sha1": "73731678e47437e69d7ab7c6e59dc967dec2dd15",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/0906.4363",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "73731678e47437e69d7ab7c6e59dc967dec2dd15",
"s2fieldsofstudy": [
"Mathematics"
],
"extfieldsofstudy": [
"Mathematics"
]
}
|
219638006
|
pes2o/s2orc
|
v3-fos-license
|
Loss of taste and smell
fractures (5%) and other pathologies. In 58% of cases an extraction was performed, in 6% a scale and root planing of the pathological area and in 10% of cases only pharmacological treatment was indicated. Other procedures performed were first stage of root canal treatment, root canal treatments of monoradicular teeth, cementing fixed dentures and removal or repair of orthodontic appliances. Only three patients were referred to emergency hospital services. An estimated 80% of treatments provided a permanent resolution of the pathology and in the remaining cases it was possible to offer a partial or temporary solution. In conclusion, the urgent dental care provided was undoubtedly strenuous, but it was also a source of great satisfaction for the dentist, both personally and professionally. Moreover, the patients greatly appreciated his degree of availability and the importance of the work achieved. After this period the dentist was tested for SARS-COV-2 antibodies and a negative result was received. L. Varela-Montes, R. Gómez de Diego, M. Varela Morales, Madrid, Spain
symptoms through media broadcasts, leaflets, and public notices could significantly reduce the disease spread and burden. Finally, if dysgeusia and anosmia were reliable and valid premonitory symptoms of the disease, then dental, medical and para-medical services may in future include a question on the acute loss of taste and smell in all pre-treatment patient history questionnaires so as to diagnose potential, or otherwise asymptomatic, COVID-19 patients. L
Skill mix alive and well
Sir, your correspondent E. Gordon (The end of skill mix; BDJ 2020; 228: 655) derides the term 'skill mix' as 'an egalitarian fantasy supported only by salaried academic elites' . The existence of Gateway Dental Practice in Abergavenny, which became an Employee Owned Trust (similar to the John Lewis Partnership) in October 2016, contradicts this view. This innovative dental practice ownership model is doing its part in building a motivated and sustainable dental workforce. In addition to promoting skill mix, the model also promotes social sustainability in its ethos of collaborative working, alongside the encouragement of entrepreneurial activity, training and development for the employees. It reflects the fact that dental practices are assets to the communities they serve, such as in providing employment opportunities to local residents, as well as providing vital healthcare services. I encourage practice owners considering disposal of their practices to read the recent paper by Allen and Karki. 1 I would agree, however, that there is a need for regulatory reform, both to bolster
Loss of taste and smell
Sir, on 26 April 2020, the US Centers for Disease Control and Prevention included 'New loss of taste (dysgeusia/ageusia) and smell (anosmia/hyposmia)' in its list of symptoms of COVID-19 disease. 1 In the absence of any comprehensive analysis of the subject, we reviewed the published literature on COVID-19 associated early dysgeusia and anosmia, finding a total of five studies from the European community, China, Italy, USA, and Iran. 2,3,4,5,6 These yielded a total of 10,847 COVID-19 patients; 8,816 (81.27%), and 8,119 (74.85%) presented with/developed dysgeusia and/or anosmia, respectively indicating these symptoms in almost three-quarters of COVID-19 patients.
However, there are knowledge gaps. The simultaneous presence of both symptoms in the prodromal or presenting stages of COVID-19 is unclear as is the temporal association of these with other critical symptoms. Some described anosmia prior to hospitalisation followed by symptoms of dysgeusia afterwards, and others the reverse. Also, the question of how long before the definitive early symptoms of COVID-19 such as fever, sore throat, etc does dysgeusia and/or anosmia appear, particularly in otherwise asymptomatic ambulatory patients, is unresolved.
If these two symptoms were relatively reliable harbingers of COVID-19, then there are multiple clinical, community interventional strategy and disease spread implications. Both are simple for self-awareness and without medical consultation could enormously expedite self-or tele-diagnosis of COVID-19. This would be particularly pertinent in overcrowded and resource-meagre communities in the developing world, and in refugee camps. In the event, community education of these
Virus-spreading procedures
Sir, the current pandemic highlights the need to consider alternative evidence-based treatments that involve no virus-spreading procedures to prevent transmission. For example, non-cavitated dental caries lesions involving up to outer third of dentine can be managed by preventive measures including fissure sealant and proximal sealing, with minimal risk of aerosol production. 1 Cavitated shallow to deep carious lesions that do not involve pulp can be managed using selective caries removal and atraumatic restorative treatment (ART), without the use of rotary instruments. 2,3 Heavily broken-down teeth can be temporarily restored by using stainless steel crowns with no preparation required. 4 Minimally invasive endodontic treatments BRITISH DENTAL JOURNAL | VOLUME 228 NO. 11 | JUNE 12 2020 813 UPFRONT
|
2020-06-15T14:23:31.988Z
|
2020-06-01T00:00:00.000
|
{
"year": 2020,
"sha1": "93f673183d5f5558683464e63355c8a9a5a6517a",
"oa_license": null,
"oa_url": "https://www.nature.com/articles/s41415-020-1732-2.pdf",
"oa_status": "BRONZE",
"pdf_src": "PubMedCentral",
"pdf_hash": "93f673183d5f5558683464e63355c8a9a5a6517a",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
71714
|
pes2o/s2orc
|
v3-fos-license
|
Unusual morphology of the superior belly of omohyoid muscle
Though anomalies of the superior belly of the omohyoid have been described in medical literature, absence of superior belly of omohyoid is rarely reported. Herein, we report a rare case of unilateral absence of muscular part of superior belly of omohyoid. During laboratory dissections for medical undergraduate students, unusual morphology of the superior belly of the omohyoid muscle has been observed in formalin embalmed male cadaver of South Indian origin. The muscular part of the superior belly of the omohyoid was completely absent. The inferior belly originated normally from the upper border of scapula, and continued with a fibrous tendon which ran vertically lateral to sternohyoid muscle and finally attached to the lower border of the body of hyoid bone. The fibrous tendon was about 1 mm thick and received a nerve supply form the superior root of the ansa cervicalis. As omohyoid mucle is used to achieve the reconstruction of the laryngeal muscles and bowed vocal folds, the knowledge of the possible anomalies of the omohyoid muscle is important during neck surgeries.
Introduction
Omohyoid (OH) muscle is one of the infrahyoid muscles that consist of superior and inferior bellies. The two bellies are continuous to each other at the intermediate tendon. Inferior belly takes origin form the superior border of the scapula, close to the suprascapular notch. It runs across the lower part of the posterior triangle, and then continues with the intermediate tendon behind the sternocleidomastoid muscle. Superior belly begins from the intermediate tendon, and then runs vertically to across the anterior triangle to get inserted to the lower border of the hyoid bone, close to the insertion of the strnohyoid muscle [1]. Unusual origin, insertion, multiple
Discussion
Previously authors have described the anomalies of the superior belly of OH. A case of the unusual attachment of superior belly of OH to the transverse process of C6, close to the scalenus medius has been reported [3]. Absence of inferior belly of OH associated with the unusual attachment of the superior belly to clavicle has been observed [4]. Miura et al. [5] have observed a duplicated superior belly of the OH. Anderson [6] has observed unilateral duplicated superior and inferior bellies. Wood [7] has reported the triplication of superior belly and insertion of superior belly to the thyroid cartilage. Tamega et al. [8] have reported the absence of muscular part of superior belly of OH. Inferior belly continued with about 2.5 mm broad tendon, which was inserted to the hyoid bone. Our case is similar to previous finding, but the observed tendon was thin and received a nerve twig from the ansa cervicalis.
Sukekawa and Itoh [2] have studied the intermediate morphologies between normal and anomalous morphologies of the superior belly of OH. They classified the intermediate morphologies into four types; type 1 with unclear anterior margin of the superior belly due to the poor myofiber development; in type 2, the superior belly was composed of a posterior large belly and an anterior small belly; in type 3, superior belly composed of three to five bellies and the bellies were arranged in a roof tile-like morphology; in type 4, the superior belly was found to consist of two bellies arranged parallel to each other in anterior-posterior direction [2]. in the present case, we observed a rare morphology of the superior belly of OH where its muscular part is replaced by a thin fibrous tendon, on the left side of the neck.
All the infrahyoid muscles develop from the muscle primordium in the anterior cervical area [9]. Initially, the muscle primordium divides into a shallow layer and a deep layer. The deep layer forms the sternothyroid and thyrohyoid muscles. The shallow layer of the primordium forms the splenius which spread in the cervical region. Then, the splenius divides into the internal and external muscles. The internal muscle develops into sternohyoid muscle whereas the lower part of the external muscle becomes the OH muscle [9]. The absence of the muscular part of the superior belly could be related to the complex primitive morphology of the splenius.
OH muscle flap is used to repair the laryngeal muscles [10] and for treatment of the bowed vocal folds [11]. It is also used for the vocal cord abduction restoration [12]. Krishnan et al. [13] have described that OH is the reliable landmark for the endoscopic exploration of the brachial plexus. The knowledge of the possible anomalies of the OH muscle is important to head and neck surgeons during the surgical procedures involving the anterior cervical region. Occurrence of absence of the muscular part of the superior belly with presence of a fibrous tendon lying in its position is seldom reported in the literature. Nevertheless the fibrous tendon receiving the nerve supply from the ansa cervicalis has never been documented.
|
2017-09-23T07:42:52.765Z
|
2014-12-01T00:00:00.000
|
{
"year": 2014,
"sha1": "b1fedde68b72946ded3475ac39f97f8395d30762",
"oa_license": "CCBYNC",
"oa_url": "https://europepmc.org/articles/pmc4276902?pdf=render",
"oa_status": "GREEN",
"pdf_src": "PubMedCentral",
"pdf_hash": "b1fedde68b72946ded3475ac39f97f8395d30762",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
4644718
|
pes2o/s2orc
|
v3-fos-license
|
Changes in Male Labor Supply and Wages
In the 1980s, both wages and labor supply of poorly educated men fell substantially relative to those of educated men. Some observers have interpreted this positive association between changes in wages and labor supply as reflecting movement along stable labor supply curves. The author casts doubt on this interpretation by showing that the wage elasticity necessary to account, by itself, for the observed labor supply decline would greatly exceed elasticity levels typically found in prior studies. Analysis of Census data shows little relationship between changes in relative wages at the state level and changes in male labor supply. Also, panel data analysis shows no strong correlation between long-run changes in individual hours and wages. The small implied labor supply elasticities suggest that very little of the labor supply changes of men during the 1980s can be related to changes in relative wages.
Recent
decades have seen a large decrease in the labor market participation of less-educated men in the United States. During the same period, the wages of less educated men have declined relative to the wages of women and other men (Levy and Murnane 1992; Bound and Johnson 1992). The contemporaneous correlation of changes in relative wages and relative participation has led researchers to speculate that the two phenomena are related. The hypothesis expounded by Juhn (1992) and Welch (1997) is that the declines in participation experienced by less educated and low-wage men are due to the wage declines experienced by these groups: as different socioeconomic groups have moved along stable labor supply curves, the groups with large wage declines have *Paul Devereux is Assistant Professor of Economics at the University of California, Los Angeles. responded to those declines by substantially reducing their participation. However, changes in relative wages are unlikely to be the only factor influencing changes in labor supply. Other possible explanations for declining labor market participation of men are declining marriage rates, increases in the wages and labor supply of women, more generous welfare packages, greater payoff from crime, and an increase in the disutility received from work. Changes in these factors are unlikely to be uniform and may have a greater impact on less skilled men. Thus, the contemporaneous declines in wages and participa-INDUSTRIAL AND LABOR RELATIONS REVIEW tion for less educated men may be related to any of these forces. This paper uses both repeated cross-sections from Census data and individual-level panel data to estimate the labor supply response of white men to long-term changes in wages. These data allow me to relax some of the assumptions made in previous literature and to use within-group identification of labor supply responses. Previous papers have studied this issue by fitting cross-sectional labor supply functions to the changes over time. Because cross-sectional approaches have many problems, I take two alternative approaches. First, if increases in wage inequality have led to a decline in the labor supply of unskilled men relative to the more skilled, then states with larger increases in inequality should also have experienced greater relative declines in labor supply of unskilled men. I use Census data to test for the presence of such a pattern. Second, one implication of the hypothesized strong link between wages and labor force participation is that individuals with large relative increases in wages will also have large relative increases in labor supply. I test that prediction using panel data from the Panel Study of Income Dynamics (PSID). In the analysis, I focus on changes during the 1980s, because changes in relative wages were much larger in that decade than in the 1970s or 1990s.
Understanding the labor supply responses to changes in relative wages is important for at least two reasons. First, to understand the effects of changes in tax laws on labor supply, knowledge of labor supply elasticities is necessary. Estimates of wage elasticities for men in the literature have tended to be very small and even negative. The implied elasticities if relative labor supply changes during the 1980s arose solely from movement along labor supply curves are much larger than those typical estimates, and they imply large effects of tax changes on male labor supply. Second, from the perspective of labor supply estimation, these issues are important because sources of exogenous wage variation are invaluable in estimating labor supply elasticities. Thus, if changes in relative wages can be treated as exogenous to labor supply, they provide a promising source of identification in future research. Indeed, they have been used for this purpose already (see, for example, Blundell, Duncan, and Meghir 1998;Pencavel 1997). The analysis in this paper sheds light on whether this assumption is likely to be satisfied and, hence, on whether one can confidently use long-term changes in relative wages to identify labor supply responses.
Using Cross-Sectional Variation
The first approach is to use Census data to examine whether states that have had the biggest changes in relative wages have also had the biggest changes in relative hours per worker. It is well known that there are substantial differences across Census regions (see, for example, Karoly and Klerman 1994;Topel 1994) and across metropolitan areas (Borjas and Ramey 1995) in the level and trend of measures of wage inequality.1 This is also true at the state level. The approach I use is to see how this cross-state variation in relative wage changes across groups corresponds to the cross-state variation in relative labor supply changes. The advantage of this approach is that one can allow the labor supply of different groups of workers to have been affected in different ways by unobservables over time. Consider the labor supply of group g at times t and t + 10 in states defined by j: ( 1Topel (1994) found that changes in relative wages in the West between 1972 and 1990 were about 50% greater than changes in the Northeast and three times greater than changes in the South. Borjas and Ramey (1995) showed that changes in relative wages were variable across metropolitan areas between 1976 and 1990, with, for example, the return to education increasing sizably in Los Angeles but only very slightly in Pittsburgh.
Now, by differencing:
(2) (hg t+10hg]t) = (( t+lwt) + ,gt+10 -f) + (jt+10 -%jt) + (t+10 gt) Thus, by including fixed group effects and state effects in the differenced regression, one can get a consistent estimate of 3 in the case where the group-specific effect is not fixed and can differ arbitrarily across groups and over time. Note that if there were no cross-state variation, one could not allow the group-specific effect to change over time. One must make the assumption that any state-group effect is constant across time. However, it is not necessary to assume that the state-group effect is uncorrelated with wages, as it is differenced out of the regression. Thus, this methodology is robust with respect to mismeasurement of cost of living changes, to changes in federal tax rates, and to changes in nonwage benefits across groups. The state effects capture any state-specific changesthe state business cycle, for example, or state-specific changes in the cost of livingthat may affect labor supply changes. They also capture any fixed differences in group composition across states.
Researchers in the labor supply literature have drawn a distinction between labor supply responses to shifts in wage profiles and labor supply responses to movements along a given wage profile.2 Longterm trends in relative wages represent shifts in the career wage profiles faced by different kinds of individuals. The identification scheme here is appropriate because the labor supply responses are identified using variation in shifts in wage profiles across states. Further detail about the Census data and the estimation methodology is provided later in the paper. 2Blundell and MaCurdy (1999) and Pencavel (2002) provide thorough accounts of how the interpretation of the labor supply elasticity depends on the estimation specification.
Using Panel Data from the Panel Study of Income Dynamics
This approach involves using panel data to examine whether, within groups, the people who have had the biggest increases in wages have also had the biggest increases in hours. The wage change measure is the change in the wage over a ten-year period. The emphasis on long-term changes is important because it is likely that the behavioral responses to short-and long-term wage changes are quite different. The changes in the labor supply measures are over the same ten-year period. Aswith equation (2), one can allow the labor supply of different groups of workers to have been affected in different ways by unobservables over time. Consider the labor supply of person i who is a member of group g at times t and t + 10: Thus, by including fixed group effects in the differenced regression, it is possible to get a consistent estimate of P in the case where the group-specific effect is not fixed and can differ arbitrarily across groups and over time. A necessary assumption is that the individual effect is constant across time. This methodology is robust with respect to mismeasurement of cost of living changes but may be sensitive to changes in federal or state taxes that differ among individuals within groups. Given that the control variables include a quartic in actual labor market experience, the identification here is from a mixture of wage profile shifts and movement along a given profile. That is, some of the individual-level wage changes will reflect permanent changes in the individual's earning capacity, and other changes will reflect movement along the individual's career wage profile. Since the labor supply elasticity from the latter should exceed that from the former, the estimates from equation (4) will tend to overstate the labor supply effects of changes in relative wages.3 Further details about the PSID data and the estimation methodology are presented later in the paper.
Wages and Participation in the Census Data
The Census data I use come from the 5% samples of the 1980 and 1990 Censuses. I restrict the sample to individuals who are aged between 20 and 60. Persons in school or the military during the survey week are omitted, as are persons who are living in group quarters. I exclude persons who have self-employment income. The wage measure used is average hourly earnings, where earnings include wage and salary earnings only.4 Earnings are topcoded at $75,000 in 1980 and $140,000 in 1990. Half of 1% of observations are topcoded in 1980, and 0.8% are topcoded in 1990. I replace the topcoded value by the topcode times 1.33 in both years. I restrict the sample to white men.5 A generic problem in labor supply estimation arises because the market wage is not observed for non-participants. A large proportion of the relative changes in hours and weeks worked across different groups arises because of changes at the extensive rather than the intensive margin. My approach to this problem is similar to that of Juhn (1992). I impute wages of nonworkers by using the wage distribution of workers who work 1-13 weeks. First, individuals are placed in one of 20 age-education 3Browning, Deaton, and Irish (1985) showed that the intertemporal labor supply elasticity is greater than the Marshallian wage elasticity that corresponds to shifts in wage profiles.
4All monetary amounts are deflated to 1979 dollars using the Personal Consumption Expenditures price index for Gross Domestic Product. 5The sample is restricted to white men because there are inadequate numbers of non-whites in ageeducation groups to obtain unbiased estimates. Prior research has shown that grouping estimators can perform poorly when group sizes are very small (Devereux 2002). groups. These are defined by the interactions of 5 education groups (high school dropout, high school graduate, some college, college graduate, more than college) and four age groups (21-30, 31-40, 41-50, 51-60). I impute the wage for non-workers as being the average wage of individuals who worked 1-13 weeks in that group in that state. For 0.2% of non-workers, there is no person who works 1-13 weeks in that group in that state. For these individuals, I impute wages as being the average wage of persons who work 1-13 weeks in that group in that Census division. Juhn (1992) demonstrated that in the CPS this approach to imputation appears to work well.
The labor supply measures refer to the hours worked in the previous calendar year and the wage measure is the log of average hourly earnings in the previous calendar year. I include four different labor supply measures. The first, HOURS, is annual hours worked in the previous calendar year. The second, WEEKS, is the number of weeks worked in the previous calendar year. The third, Full Time/Full Year (FTFY), is the proportion in the group who worked at least 50 weeks and usually worked 35 or more hours per week. The fourth, ANNUAL PARTICIPATION, is the proportion of people in the group who worked for even one hour in the year. The value of using several participation measures is that each captures a different facet of labor supply and each has its own strengths and weaknesses. Rather than rely on one measure, I prefer to draw inferences from the effects of wages on all four. Like other researchers in this literature, I make no distinction between periods of non-employment that are classified as unemployment and periods spent out of the labor force.
There were some substantial differences in changes in labor supply across states during this period. Wyoming and West Virginia had large decreases in labor supply, with average annual hours
Changes over Time at the National Level
In this section, I describe the relationship between labor supply and wages at the national level and I discuss some of the previous literature on this topic. Some descriptive statistics on wages and participation by education group for the United States as a whole are presented in Table 1. These statistics show clearly some of the well-known characteristics of relative wages and relative participation. Hours worked are higher for persons with more education. Similarly, wages increase with education. Juhn (1992) and Welch (1997) brought cross-sectional labor supply relationships to bear on the time series changes in relative wages using the CPS. Juhn (1992) estimated the labor supply curve across the cross-section of the wage distribution for the years 1970-72 prior to the start of her period of interest. Welch (1997) estimated the labor supply equation using the entire sample period between 1967 and 1992. Juhn's and Welch's results are valid to the extent that the cross-sectional estimates are valid. Both authors acknowledged that their cross-sectional estimates may overstate how much individual participation responds to changes in individual wages. For example, highly motivated individuals are likely to have both high wages and high participation rates. Given that we cannot control fully for "motivation" in the labor supply equation, we may observe a positive relationship between wages and labor supply in the cross-section, even if wages have no causal effect on labor supply. This problem will also be serious when wages are instrumented by age and education, as it is likely that unobservables like mo-tivation differ systematically by age and education group.
All the empirical work in this paper is carried out using cell means.6 In the absence of other endogeneity problems, measurement error in reports of hours and earnings leads to bias in regressions using microdata. Mean-zero measurement error can be averaged away by taking means over cells with sufficient numbers of observations. I have estimated cross-sectional regressions of labor supply on wages using the 1980 and 1990 Census. In each year, I take the mean of each variable for each of the 20 age-education groups. In the regressions, the mean labor supply for each group is regressed on the mean of the log wage for the group. The means from each group are weighted by the number of underlying Census observations in the group. The wage elasticity is calculated as the coefficient on the log wage divided by the mean of the dependent variable. The elasticities are in Table 2a. For the 1980 sample, I find wage elasticities of 0.182 (0.056) for hours, 0.139 (0.051) for weeks, and 0.405 (0.079) for FTFYin the national sample. The elasticities from 1990 are also reported in Table 2a and are somewhat similar in magnitude. These elasticities are substantial given that this is a sample of white men.
Between 1979 and 1989, hours worked fell for the less-educated groups and rose for the groups with college degrees. As can be seen in Table 1, these changes in labor supply measures closely parallel the changes in relative wages over this time period. In Figure 1, I plot the changes in the average log wage of each group against the change in annual hours worked. There is clearly a strong positive correlation between a group's wage change and its hours change. The relationships are similar for the other labor supply measures.
The statistical analysis in the third row of Table 2a quantifies this effect more pre-6Note that the wage measure is the mean of the log wages within each cell rather than the log of the mean wage within the cell. cisely. Each column presents weighted least squares estimates from the regression of the change in labor supply for each of the 20 groups on the change in the log wage. The means from each group are weighted by the number of underlying Census observations in the group. The elasticities are quite substantial, with an estimated wage elasticity of 0.32 in the hours equation, 0.18 in the weeks equation, and 0.24 in the FTFY equation. One implication of these numbers is that in order for changes in labor supply to be explained by movement along a labor supply curve, the own wage elasticity for men must be about this magnitude. The estimates in Table 2a are consistent with the previous literature. Consider first the results for estimates that use weeks worked as the labor supply measure. The Table 2a estimates show that the cross-sectional elasticity for weeks worked is about 80% the size of the differenced elasticity, implying that if the cross-sectional relationships were fit to the time-series changes, wage changes would almost fully explain changes in weeks worked over time by different groups. This is whatJuhn (1992), who focused on weeks worked, found for white men. Turning to the results for FTFY, the cross-sectional estimates in Table 2a for FTFY imply that changes in wages should have led to larger changes in FTFY than actually occurred. Welch (1997), who concentrated on FTFYas the labor supply measure, obtained the same result. Table 2b includes results for men who worked at least one hour in the calendar year.
The advantage of restricting the sample to participators is that it obviates imputing wages; the disadvantages are that changes at the extensive margin are ignored and the sample composition changes over time as participation rates change. The cross-sectional results for participators are generally similar to those for the full sample. However, the differenced elasticities are smaller for participators, reflecting the fact that changes in labor supply across groups are not as large when one conditions on participation (because of the large changes at the extensive margin).
It is noteworthy that the strongly positive relationships between wages and labor supply in the studies mentioned above contrast with the generally small or negative wage elasticities for men in the previous litera- There are 20 observations in each regression. All elasticities are evaluated at the mean of the dependent variable. The elasticities are calculated as the coefficient on the log wage divided by the mean of the dependent variable. The means of the labor supply variables are given in the table. Robust standard errors in parentheses allow for arbitrary forms of heteroskedasticity. ture (Killingsworth 1983;Blundell and MaCurdy 1999). Also, the responsiveness is somewhat inconsistent with the fact that male hours have steadily fallen over the century as wages have steadily risen. Given the strong assumptions implicit in the crosssectional approaches, the alternative approaches implemented in this paper provide an important check on the robustness of these results.
Using Cross-State Variation
As discussed in the methodology section, cross-state variation in relative wages and relative hours can be used to identify labor supply responses in the presence of unobservable factors that have affected the labor supply of particular groups. To carry out this analysis, I calculate the mean value of the variables for the 20 groups in each of 48 states (dropped are Alaska and Hawaii) in each year. I then implement an estimator that corrects the OLS regression of the group-state cell means for small sample bias. The estimator and the logic behind it are described below.
The Estimator
On average, there are 2,048 observations per group-state cell. Even though these cell sizes are relatively large, prior research suggests that OLS estimates from regressions of the cell means may suffer from small sample biases with cell sizes of this magnitude (Devereux 2002).7 In the following discussion, I distinguish between population means and sample means. The sample mean of hours for a particular cell is the mean of the hours of all members of that cell included in the sample. The popu-7The results using national groups presented above are weighted least squares results. The national groups are so large that the corrected least squares results are basically identical to the least squares results. ( I This measurement error may have two components. First, reporting error in individual reports of earnings and hours implies that the means will be incorrect in finite samples. The individual-level reporting error issue is particularly important here, because in addition to the usual attenuation bias that arises from classical reporting error, there is non-classical reporting error induced by the correlation between reporting error in hours and in earnings divided by hours. In cells with few observations, such reporting error will induce a spurious correlation between changes in the wage and changes in hours. This correlation would bias down the own wage elasticity in an OLS labor supply regression. Second, even if individual observations were correctly reported, there will be sampling error in finite samples. As the cell size gets larger, laws of large numbers imply that the sample mean will become arbitrarily close to the population mean. Deaton (1985) showed that one can con- I report estimates both using weighted least squares and using the corrected least squares estimator. In both cases the variance-covariance matrix is calculated allowing for arbitrary forms of heteroskedasticity and non-zero covariance between observations that are in the same group in the same Census division.
Results
The estimation results are reported in Tables 3 and 4. It is standard in labor supply estimation to include controls for other income to capture negative income effects on hours worked. Thus, in Table 4, I add asset income as an additional control.8 In all specifications the observations are weighted by the number of individuals within each state-group. In the specifications in Table 3a, all white men are included and wages are imputed for nonworkers as described in the data section. I will discuss the corrected least squares estimates-the weighted least squares estimates tend to be slightly smaller.
The first two rows of Table 3a contain cross-sectional regressions of labor supply on the log wage, and a set of state indicators. As expected, the elasticities are very similar to the cross-sectional elasticities from national data. In the third row, the change in labor supply is regressed on a constant, the change in the log wage, and a full set of state indicators. There are no controls for group indicators, so the source of variation used to estimate the wage elasticity is changes in relative wages across the groups over time. Given that this is the same type of variation used in the national estimation, one would expect the results to be similar to those from the differenced regressions in Table 2a.
The estimated elasticities are somewhat smaller in magnitude than in Table 2a, but 8Asset income is defined as interest, dividend, and net rental income. Table 2). The elasticities are calculated as the coefficient on the log wage divided by the mean of the dependent variable. Robust standard errors, in parentheses, allow for arbitrary forms of heteroskedasticity and for observations in the same group and Census division to be correlated. they are much more precisely estimated. In the hours regression, the wage elasticity is 0.26. The wage elasticities for the other labor supply measures are somewhat smaller but still substantial considering that this is a sample of white men. The addition of group fixed effects dramatically reduces the size of the wage effect on all four labor Table 3a.) These 9The reported results ignore selection. I have also estimated models that use a selection correction in the spirit of Heckman (1979). I find no evidence of selection bias. My ability to test for selection bias is limited, however, by the unavailability of variables that can be included in the participation equation and omitted from the hours equation. modest elasticities for participators suggest a weak relationship between changes in wages and changes in labor supply for participators. Since they come from a specification that does not allow for differential supply shifts across groups, they imply that the observed relationship between changes in labor supply and wages at the intensive margin is consistent with modest labor supply elasticities, even in the absence of supply shifts. Table 4 contains estimates in which the specifications in Table 3 have been augmented by the addition of asset income.10 The uncompensated wage elasticities from this specification are very similar to the wage elasticities from the specification in Table 3. The marginal propensity to earn (calculated as the average wage times the coefficient on other income) is typically small and statistically insignificant. Thus, for all the measures, the compensated wage elasticity is not very different from the uncompensated elasticity. As in Table 3, the addition of group fixed effects radically reduces the size of the wage effect on all four labor supply measures. The results imply that while labor supply changes were positively correlated with wage changes over the 1980s, differences in the pattern of relative wage changes across states had little predictive power for labor supply changes. Thus, the estimates suggest the occurrence of labor supply curve shifts across groups that implied changes in labor supply that were not strongly related to changes in wages.
There are 960 observations in each regression. Each regression includes a full set of state indicators. All elasticities are evaluated at the mean of the dependent variable (listed in
One objection to this interpretation is that differences in relative wage changes across states represent sampling error rather than actual differences in wage changes. This is unlikely to be an issue, because the number of observations in each group-state cell is large (an average of 2,048 observa-1?To conserve space, I include only the weighted least squares estimates. As with the specifications in Table 3, the corrected least squares estimates are very close to, but always slightly higher than, the weighted least squares estimates. There are 960 observations in each regression. Each regression includes a full set of state indicators. All elasticities are evaluated at the mean of the dependent variable (listed in Table 2). The elasticities are calculated as the coefficient on the log wage divided by the mean of the dependent variable. The marginal propensity to earn is calculated as the coefficient on asset income multiplied by the mean wage. The mean of the wage is $9.368 for the full sample and $9.382 for the sample of participators. The mean of other income is $787.74 for the full sample and $749.62 for the sample of participators. Robust standard errors, in parentheses, allow for arbitrary forms of heteroskedasticity and for observations in the same group and Census division to be correlated. tions per group-state cell) and the estimation method corrects for any remaining variance resulting from sampling error. A second objection is that there is not sufficient variation across states in wage change differences across groups. However, as indicated by the small standard errors in the specifications with group controls, there is substantial variation in changes in relative wages across states.11 Figures 2-4 provide some graphical evidence of the relationship between wage 11A table detailing the extent of variation is available from the author on request.
Figure 2. Relationship between State-Level Wage and Hours Changes (High School
Dropouts, Aged 31-40).
Grouping by Cohort
In the analysis so far, grouping has been by age. An alternative is to follow cohorts over time. In this section, 15 groups are defined by the interactions of the 5 education groups (high school dropout, high school graduate, some college, college graduate, more than college) with three birth cohorts (men aged 21-30, 31-40, and 41-50 in 1980). Thus, individuals in 1990 are aged 31-40, 41-50, and 51-60. As be-fore, the means from each group-state cell are weighted by the number of underlying Census observations in the cell.
The results are in Table 5. For brevity, I only include the differenced estimates and I exclude specifications with asset income. As with the age-based groups, the wage elasticities are not much affected by the inclusion of asset income. The elasticities are generally slightly higher than in the age-based groups. For the full sample, the corrected least squares elasticities range from 0.02 and 0.1 in the specifications with group indicators. In the sample of participators the equivalent elasticities range from 0.1 to 0.17.
The slightly higher wage elasticities estimated from the cohort-based groups may result from the fact that they are measuring some combination of the intertemporal wage elasticity and the standard uncompensated wage elasticity. If changes in relative wages are foreseen by agents, then they do not constitute a change in wealth, and hence the marginal utility of wealth does On the other hand, when we compare the labor supply of agents of the same age in 1980 and 1990, we are comparing two cohorts with different permanent incomes resulting from the wage changes. Thus, to the extent that the changes in relative wages are permanent, the elasticities are capturing hours changes that result from shifts in the wage profile. The elasticities from the age-based groups are therefore more likely to be capturing the standard uncompensated wage elasticity. Given this consideration, it is not surprising that they are smaller. Pencavel (2002) provided an excellent discussion of how estimated elastici-ties may be interpreted as uncompensated elasticities or intertemporal elasticities depending on the exact specification. His conclusion that the uncompensated wage elasticity is very small for men is consistent with the findings of this paper.
Migration
If the composition of a group changed between 1980 and 1990, this may bias results. Since groups are defined at the state level, migration between states or immigration from abroad is a potential source of concern. Ideally, in order for the estimates to be interpreted as labor supply elasticities, there should be no migration across states. One advantage of the cohort-based strategy described in the previous section is that it allows a natural specification check for the effects of migration. The Census asks respondents about their state of residence five years previously. Thus, I can exclude from the 1990 sample all cases in which the individual resided in a different state in 1985. When I do this, the estimates are robust with respect to the exclusion of all individuals who did not reside in the same state in 1985 (the wage elasticities are typically slightly smaller when the movers are omitted, but the differences are not large).12 I do not report results, because the estimates are very similar to those in Table 5. Thus, it is unlikely that migration across states between 1980 and 1990 seriously biases my results.
120ne would expect migration to bias upward the labor supply elasticities. If the wage of a particular group rises in a certain state, individuals in that group will tend to move to that state. The individuals who move to the state will tend to be individuals with strong preferences for work. Thus average labor supply in the group will rise both because individuals in the group work more and because new entrants to the group in that state are hard-working types.
Wages and Labor Supply in the Panel Study of Income Dynamics
The Panel Study of Income Dynamics (PSID) is a longitudinal data set that has followed approximately 5,000 families from 1968 to the present. In the analysis using the PSID, I use both the random sample and the poverty sample.
The results reported are not weighted by the weights in the PSID, but the weighted results are almost identical. The sample selection criteria are as follows.
Respondents must be present for at least one year during the 1979-81 period and must also be in the sample ten years later, during the 1989-91 period. The wage measure used is annual earnings divided by annual hours.
To create a sample that is comparable to the Census sample, I restrict the sample to white men who are no more than 50 years of age in the first year. The minimum age of people in the sample is 20, and stu-423 1) CA .9 There are 720 observations in each regression. All regressions include a full set of state indicators. All elasticities are evaluated at the mean of the dependent variable. The elasticities are calculated as the coefficient on the log wage divided by the mean of the dependent variable. The means of the labor supply variables are given in the table. Robust standard errors, in parentheses, allow for arbitrary forms of heteroskedasticity and for observations in the same group and Census division to be correlated. dents are omitted. Furthermore, I only include individuals with positive actual labor market experience. I delete cases in which the reported annual hours are greater than 4,000, as these are likely to be seriously mismeasured. I also delete cases with wages of less than $1 per hour in 1979 dollars. Finally, I delete a few cases with inconsistent information on hours, weeks worked, and earnings, such as persons reporting zero weeks and positive hours and persons with positive hours and zero earnings. I also delete cases with missing education information. I treat topcoded wages in the same fashion as with the Census data, that is, I multiply the topcoded value by 1.33.13
Exploiting Individual Variation in Wage Changes
The estimating equation relates changes in labor supply for individual i between t and t+ 10 to the change in the individual's log wage over that period, and to a set of control variables x (in this case, education indicator variables and a quartic in actual [not potential] labor market experience).
13There is only one such case, and I get similar results if I delete it. The wage measure used is annual earnings divided by annual hours. Thus, by construction, the wage measure has a spurious negative correlation with hours resulting from measurement error in hours. The strategy of taking means over large numbers of observations used with the Census data cannot be applied to these individual-level data. Therefore, I take advantage of the panel structure of the data. I use the average of the log wage at t -1 and the log wage at t + 1 as an instrument for the log wage at t. Similarly, I use the average of the log wage at t + 9 and the log wage at t + 11 as an instrument for the log wage at t + 10. Thus, the instrument for the change in the log wage between t and t + 10 is the difference between these two average wages. An advantage of this instrument is that it preserves the individual-specificity of the wage changes used in the analysis.
Several caveats should be borne in mind. First, measurement error in earnings and hours may be serially correlated, and this would bias the estimates. Second, the fixed effects assumption may be violated, and changes in wages of individuals may be related to changes in skills, motivation, or other factors that have direct effects on the choice of hours.
Third, in a model of uncertainty, forecast errors may be correlated with wage changes.
There is no obvious way to impute wages for non-workers, so I restrict the sample to men who participate in both periods. The analysis is carried out using two stage least squares. Table 6 contains the wage elasticities.14 The estimates are similar for all three 10-year periods. The wage elasticities are all close to zero and are frequently negative. The estimated elasticities are a bit smaller than those from the Census analysis, but given the standard errors, one should not make too much of the differences. While both identification strategies use within-group identification, it is still meaningful that two very different sources of identification on two different data sets produce similarly small estimates. As such, these results strengthen the earlier conclusion that the wage effects on labor supply are too small to fully explain the changes in men's relative labor supply during the 1980s.
14The PSID has an alternative wage measure for a subset of workers who are paid hourly or are salaried. I have tried using this wage measure as an instrument for the wage measure used in the analysis. The results are very similar to the reported results.
Why Might Labor Supply Curves Have Shifted in for Less-Skilled Men?
Given the conclusions of this paper that the uncompensated wage elasticity for white men is close to zero and thus wage changes alone cannot explain the changes in relative labor supply, what factors can explain the implied shifts in labor supply? Possible candidates are declining marriage rates, changes in female wages and labor supply, shifts in earnings opportunities from crime, and changes in transfer income. I briefly consider these possible explanations below.
Crime rates stabilized during the 1980s as incarceration rates increased. Freeman (1996) estimated that the propensity for criminal activity by non-institutionalized men increased by between 80% and 163% between 1977 and 1992. There are no reliable estimates of earnings from crime. However, Freeman reports that two Boston surveys showed that the proportion of youths who said that they had "chances to make illegal income several times a day" roughly doubled between 1980 and 1989. The expansion of the drug market may have increased earnings from criminal activity over this period. Over the same period, the probability of incarceration increased, and this should have tended to make crime less attractive. Thus, given the incompleteness of our knowledge about crime, it is not possible to tell whether changes in criminal opportunities for different groups shifted labor supply curves over this period.
Another possible cause of shifting labor supply curves is declining marriage rates. Married men tend to work more than single men. This could be a causal effect of being married or could reflect the sorting of men into marriage. During the 1980s, marriage rates in the Census data fell for all education groups except the group with postcollege education. The proportion married fell from 0.77 to 0.66 for high school dropouts, from 0.74 to 0.68 for high school graduates, from 0.72 to 0.70 for the group with "some college," and from 0.71 to 0.70 for the college graduates, and it rose from 0.77 to 0.79 for the group with post-college education. The disproportionate decline in marriage rates for the least educated could imply a shifting inward of labor supply curves for this group. This can only be suggestive, as marriage rates are likely to be affected by wages and labor supply, and establishing any causal relationships between these outcomes might prove difficult.
The wages and labor supply of women changed enormously over this period. Among married women, labor force participation increased by about 10% for the wives of all five male education groups. Earnings of wives also increased for all five groups, with the biggest increases (in absolute and percentage terms) being experienced by the highly educated groups. In particular, the wages of wives of collegeeducated men increased about 15% while the wages of wives of less-educated men stagnated. The increasing earnings of married women would have tended to shift labor supply curves of men through income effects. However, if wives' incomes are to be invoked as an explanation for why labor supply curves have shifted in for poorly educated men only, it must be shown that the effects of wife's income differ across groups.
I have examined this possibility using the sample of married men in the Census. I take the differenced state-level hours regression without group indicators and augment it with the change in wife's earnings. I allow the effect of the change in wife's earnings to differ by husband's education. The coefficient on wife's earnings is significantly negative for high school dropouts, approximately zero for high school graduates, and significantly positive for the other education groups. Adding wife's earnings to the specification reduces the estimated wage elasticity for married men from 0.278 (0.014) to 0.083 (0.016). Given that wife's earnings are surely endogenous, one should not take these estimates too literally. However, they do indicate that changes in the wages and labor supply of women may have led to shifts in the labor supply curves of men.
One source of non-labor income available to working-age men is disability benefits. Over this period, these benefits did not become more generous, and so it is unlikely that they are a major reason for shifts in labor supply curves (Juhn 1992). In the Census, the proportion who reported having a disability that inhibited working rose from 0.17 to 0.19 among high school dropouts and remained fairly flat for the other groups. This increase in reported disabilities among the least-educated could reflect changes in attitude toward work. However, once again, reported disability may be endogenous to labor supply. I conclude that although there are many possible reasons for shifts in labor supply curves over this period, pinning down which were most important constitutes a challenging research agenda.
Conclusions
Labor supply estimation using two different data sources and identification strategies provides one consistent result: the relationship between wage changes and labor supply changes over the 1980s was very small. This relationship is much weaker than repeated cross-sectional estimates in the literature suggest. In particular, the relationship is not strong enough to suggest that labor supply changes during the 1980s reflect movement along a stable labor supply curve. The results suggest such movements along supply curves were accompanied by inward shifts in labor supply among low-skill men. Researchers should be careful when using changes in relative wages over time as instrumental variables in labor supply estimation.
Recent research has suggested that the declines in labor supply by less educated men during the 1980s represented a labor supply response to changes in relative wages. In this paper, I have shown that the labor supply elasticities required to be consistent with this mechanism are large, particularly to explain changes at the extensive margin. These large elasticities are out of line with the vast majority of previous estimates in the literature. Using two alternative identifying assumptions, I have estimated elasticities that are close to zero. These results suggest that tax changes have a negligible effect on the labor supply of men below retirement age. Policy-makers should be cautious when drawing conclusions about labor supply elasticities that are obtained by correlating trends in wages and participation rates over relatively short periods of time.
|
2018-04-06T17:46:33.510Z
|
2003-04-01T00:00:00.000
|
{
"year": 2003,
"sha1": "4f6aa628cf500a406c98b3477b1bf29c692a7914",
"oa_license": "CCBYNCSA",
"oa_url": "https://researchrepository.ucd.ie/bitstream/10197/324/4/devereuxp_article_pub_001.pdf",
"oa_status": "GREEN",
"pdf_src": "Sage",
"pdf_hash": "1223f34561aff81dcdab3286573d68759fb8bfd1",
"s2fieldsofstudy": [
"Economics"
],
"extfieldsofstudy": [
"Economics"
]
}
|
259209390
|
pes2o/s2orc
|
v3-fos-license
|
Nanomaterials for photocatalysis and applications in environmental remediation and renewable energy
Global warming and climate change are increasing global issues. In the last ten years, the intensity of these issues has drawn significant attention from many countries worldwide. One of the factors that cause climate change are industrial processes which need high power to run, and most countries have used fossil fuels for these processes [1]. The use of fossil fuels generates harmful emissions to the environment, such as carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O), nitric oxide and nitrogen dioxide (together termed NOx), and fluorinated gases (e.g., hydrofluorocarbons, perfluorocarbons, and sulfur hexafluoride) which are currently considered primary sources of environmental [2]. A Global Warming Potential (GWP) measurement was used to compare the global warming effects of different gases. It has been calculated to reflect how long gases remain in the atmosphere, on average, and how strongly it absorbs energy [3]. Besides, the discharge of persistent organic pollutants (POPs) also contributes to water pollution, increasing global environmental pollution. Recently, the reduction and conversion of CO2 into fuel as valuable hydrocarbon products has been drawing attention from scientists in materials science, chemical engineering, nanotechnology, and related fields [4]. To reduce contaminants (e.g., air pollution (CO2, NOx, SO2), POPs) there are many routes (e.g., physicochemical approaches, biological fixation, advanced oxidation process, and photocatalysis [5-8]). Among the aforementioned methods, the photocatalysis route is appropriate for treating pollutants, even in atmospheric conditions [9-11]. Moreover, the photocatalysis method is also a potential solution for environmental remediation, carbon emission reduction, and renewable energy production [12-14].
Global warming and climate change are increasing global issues. In the last ten years, the intensity of these issues has drawn significant attention from many countries worldwide. One of the factors that cause climate change are industrial processes which need high power to run, and most countries have used fossil fuels for these processes [1]. The use of fossil fuels generates harmful emissions to the environment, such as carbon dioxide (CO 2 ), methane (CH 4 ), nitrous oxide (N 2 O), nitric oxide and nitrogen dioxide (together termed NO x ), and fluorinated gases (e.g., hydrofluorocarbons, perfluorocarbons, and sulfur hexafluoride) which are currently considered primary sources of environmental [2]. A Global Warming Potential (GWP) measurement was used to compare the global warming effects of different gases. It has been calculated to reflect how long gases remain in the atmosphere, on average, and how strongly it absorbs energy [3]. Besides, the discharge of persistent organic pollutants (POPs) also contributes to water pollution, increasing global environmental pollution. Recently, the reduction and conversion of CO 2 into fuel as valuable hydrocarbon products has been drawing attention from scientists in materials science, chemical engineering, nanotechnology, and related fields [4]. To reduce contaminants (e.g., air pollution (CO 2 , NO x , SO 2 ), POPs) there are many routes (e.g., physicochemical approaches, biological fixation, advanced oxidation process, and photocatalysis [5][6][7][8]). Among the aforementioned methods, the photocatalysis route is appropriate for treating pollutants, even in atmospheric conditions [9][10][11]. Moreover, the photocatalysis method is also a potential solution for environmental remediation, carbon emission reduction, and renewable energy production [12][13][14].
Combining photocatalysts and sunlight irradiation is a potential strategy for water treatment via the effectively infinite energy from the sun and the photocatalysts. Photocatalysis based on nanostructured semiconductors can significantly contribute to tackling several environmental pollution problems, sustainable synthesis, and energy production [2,15,16]. Semiconducting photocatalyst nanomaterials, such as SnO 2 , TiO 2 , MoS 2 , g-C 3 N 4 , and Bi-nanostructures have been proven efficient for a range of applications, including organic pollutant removal, NO x degradation, renewable energy production, and waste-to-energy conversion [15,17,18]. Figure 1 shows a general photocatalysis mechanism outlining several possible targets (i.e., NO x degradation, water splitting, degradation of organic pollutants, and enhancement of electron generation in a solar-cell application).
This Thematic Issue highlights recent experimental and theoretical developments in using light harvesting by semiconductor materials for sustainable applications; for instance, dye solar cells, solar-driven water splitting, NO x removal, and contaminant degradation. The synthesis of semiconductor nanomaterials published on this thematic issue indicates a wide range of synthetic routes. The as-prepared nanomaterials with various morphologies demonstrated many preeminent features in the above applications. In detail, the MoS2 with a honeycomb-like structure was first synthesized by an electrochemical route and applied in dye-sensitized solar cells [19], which expressed a higher applicability than that of other studies [20][21][22]. Besides, Nhu et al. [23] used rosin as a green chemical approach to fabricate ZnO nanoparticles, exhibiting a high photocatalytic activity for both methylene blue (100%) and methyl orange (82.78%) decomposition after 210 min under UV radiation. Moreover, the advantages in the development of advanced materials based on semiconductors (i.e., carbon-modified hexagonal boron nitride (MBN), MgO@g-C3N4, and TiO2@MWCNTs) have indicated a highly efficient photocatalytic performance for phenol removal using a low-power visible LED light source. For NO degradation, a visible light source was used whereas for water splitting natural sunlight was used [24][25][26]. These results are mentioned as scaling up photocatalytic systems to reach net zero emission goals and the next technology to produce green hydrogen energy [14].
Up-to-date trending topics on photocatalysts based on semiconducting nanomaterials, perovskites, or Bi-based nanomaterials are presented to incentivize fine-tuning of current studies and research works on photocatalytic efficiency of nanomaterials [27]. In addition, this Thematic Issue will undoubtedly provide the reader with novel ideas for developing nanomaterials for environmental remediation and sustainable applications; for instance, dye solar cells, solar-driven water splitting, NO x removal, and contaminant degradation. This Thematic Issue will make a good reference material and be of great use for scientists in nanomaterials fields.
Viet Van Pham and Wee-Jun Ong
Ho Chi Minh City and Sepang, June 2023.
|
2023-06-22T05:07:29.535Z
|
2023-06-13T00:00:00.000
|
{
"year": 2023,
"sha1": "acf873c9cbc8567c8241b43ec3178e7d5f5fc24f",
"oa_license": "CCBY",
"oa_url": null,
"oa_status": null,
"pdf_src": "PubMedCentral",
"pdf_hash": "acf873c9cbc8567c8241b43ec3178e7d5f5fc24f",
"s2fieldsofstudy": [
"Environmental Science",
"Materials Science",
"Chemistry"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
43996730
|
pes2o/s2orc
|
v3-fos-license
|
Prognostic Models in Cirrhosis: An Anesthetist Perspective
The understanding of the pathophysiology of liver diseases has grown over the years and so has the therapeutic options available for their management. Progress has been made from shunt surgeries to Transjugular intrahepatic portosystemic shunts (TIPSS) and liver transplant. Improved survival in these patients has resulted in an increase in the number of patients suffering from liver disease who present for surgery/noninvasive procedures in diverse clinical scenarios [1] Improved care in the critical care setting has also enabled many patients with decompensated liver disease to undergo liver transplantation successfully. Risk estimation and prognostication, therefore, becomes very important for the anaesthetist who will encounter such patients in different clinical settings.
Introduction
The understanding of the pathophysiology of liver diseases has grown over the years and so has the therapeutic options available for their management. Progress has been made from shunt surgeries to Transjugular intrahepatic portosystemic shunts (TIPSS) and liver transplant. Improved survival in these patients has resulted in an increase in the number of patients suffering from liver disease who present for surgery/noninvasive procedures in diverse clinical scenarios [1] Improved care in the critical care setting has also enabled many patients with decompensated liver disease to undergo liver transplantation successfully. Risk estimation and prognostication, therefore, becomes very important for the anaesthetist who will encounter such patients in different clinical settings.
End stage liver disease is associated with significant periprocedural morbidity and mortality. Risks in such patients include further deterioration of liver function, worsening of hepatic encephalopathy, renal dysfunction, bleeding due to presence of coagulopathy, unmasking of cirrhotic cardiomyopathy and deterioration of hepatopulmonary syndrome. In order to simplify the process of risk assessment in these patients, a preoperative liver assessment (POLA) check list has been proposed by Im et al. [2].
CTP (Child Turcotte Pugh) and MELD (Model for End stage Liver Disease) scores are being commonly used for peri procedural prognostication of these patients by anesthesiologists. Many of the risks described above i.e. worsening of encephalopathy, coagulopathy, worsening liver function, and kidney dysfunction are accounted for by CTP and MELD scores. Various modifications of these scores have been proposed to predict prognosis in different clinical settings. Apart from estimation of the life expectancy, these models also tell us about the ability of these patients to withstand a particular procedure or whether the therapeutic option offers an acceptable chance of survival. The aim of this review is to help the anaesthesiologist in using the appropriate scoring system in commonly encountered clinical settings. Accordingly the background, merits and demerits of these scoring systems have been discussed.
Score derivation
The Child-Turcotte classification has been used to assess liver dysfunction and predict surgical morbidity and mortality. Developed in 1964 by Child and Turcotte, it was an empirically derived formula [3,4].
It was used for predicting the outcome after surgery (portocaval shunting and trans-section of the esophagus) in patients with cirrhosis and portal hypertension.
Score variables and range
The Child-Turcotte score included two continuous variables (bilirubin and albumin) and three discrete variables (ascites, encephalopathy, and nutritional status) [5].
Merits
The Child Turcotte score was an easy bedside assessment, not needing difficult algorithmic equation for calculation and prognostication.
The use of this score has been abandoned after its modification to the Child Turcotte Pugh Score.
Score derivation
The Child score was modified in 1972 by Pugh et al. and was termed the Child Turcotte Pugh score ( Table 1). The most subjective component of the Child-Turcotte score i.e. nutritional status was replaced by prothrombin time [5].
Score variables and range
Thus the score includes variables of bilirubin, prothrombin time, albumin, ascites and hepatic encephalopathy. The score ranges from 5-15, indicating severity as score increases. It has been used to define three classes of liver disease i.e. A, B, and C (Table 1).
In cirrhotics undergoing nontransplant surgery, CTP classes of A, B and C have been historically associated with mortality of 10%, 30%, and 76-82% respectively [6]. Other post operative complications like liver failure, worsening encephalopathy, bleeding, infection, renal failure, hypoxia and intractable ascites have also been correlated with CTP class [1]. Even in patients with CTP class A, the risk of perioperative morbidity is increased when there is associated portal hypertension. It can be reduced by preoperative placement of a transjugular intrahepatic portosystemic shunt (TIPS) in such patients [7,8]. Emergency surgery is associated with a higher mortality rate than non-emergent surgery: 22% versus 10% for patients in Child class A; 38% versus 30% for those in Child class B; and 100% versus 82% for those in Child class C [9].
Risk and morbidity varies with type of surgery and state of decompensation of liver ( Table 2). For patients with CTP class C cirrhosis, attempts should be made to improve the patients liver function to near class B before surgery. Measures to improve the hepatic function include hepatic function protection, control of ascites, nutritional support, correction of coagulopathy, and reduction of portal vein pressure.
Merits
The major advantage of the CTP score is that it is easily calculated at bedside and does not require complicated mathematical algorithm.
1.
Variables like ascites and hepatic encephalopathy are influenced by subjective interpretation.
2.
The five variables of the CTP score are given the same weight.
3. The conventional CTP system has a ceiling effect at the highest score of 15 points. For instance, patient whose serum bilirubin level is 4 mg/dL has the same CTP score as those whose bilirubin level is 20 mg/dL or higher.
4.
The variables included in CTP score are not specific markers of the synthesis (albumin and prothrombin) and elimination (bilirubin) functions of the liver. Changes in serum albumin may be also related to increased vascular permeability, especially in cases of sepsis, and large-volume ascites [10,11]. Similarly, bilirubin can be increased as a consequence of impaired renal function, hemolysis, or sepsis [12]. Prolonged prothrombin time can be a consequence of an intravascular activation of coagulation during sepsis [13].
Score derivation
The MELD score was derived from a population of 231 patients with cirrhosis who underwent elective TIPS (Transjugular intrahepatic portosystemic shunt) placement. The model was subsequently validated in an independent cohort of patients from the Netherlands undergoing TIPS placement [14]. It was found to be a good predictor of three month mortality after TIPS.
Score variables and range
The original MELD contained four variables which included etiology of liver disease. It included INR, serum creatinine, serum bilirubin level and a disease etiology factor for alcoholic liver disease and cholestatic liver disease. The etiology factor was removed as it was not observed to affect mortality prognosis. This modified MELD score was found to be a good predictor of early mortality (3 month) after placement on waiting list for liver transplant [15]. Excluding the cause of cirrhosis had minimal impact on the model accuracy. According to this modified score, patients with bilirubin and creatinine values below 1 mg/dL (17 and In patients with MELD ≥9, other treatment modalities to be considered. CTP and ASA scores also independently predict short term (30 day) mortality.
[ [70][71][72] MELD has been demonstrated as an excellent predictor of survival in patients who have end stage liver disease [15,16]. Currently MELD score is in popular use for predicting postoperative mortality for cirrhotics undergoing non transplant surgery. The other major use is presently to prioritize organ allocation for liver transplant because it is a good predictor of short term mortality on waiting list [4].
Preoperative MELD scores have been found to be related to development of acute renal failure post liver transplant [17], but poor predictors of post-transplant mortality [18].
Post operative MELD scores within first week after orthotopic liver transplant have been found to predict very early death [19].
Apart from organ allocation and assessment of severity of liver disease, MELD has been positively correlated with other organ dysfunction associated with liver disease. It has been found that higher MELD scores are associated with a higher incidence of features of cirrhotic cardiomyopathy. Some parameters which have shown a positive correlation with a higher MELD score are enlarged left atrial diameter, increased intervenricular septum thickness, increased QTc interval and cardiac output. It was also found that QTc prolongation is more common in patients with alcoholic cirrhosis (50%) as compared to the viral etiology (39%). A higher frequency of diastolic dysfunction is found in patients with MELD ≥ 20. Diastolic dysfunction has also been proposed a predictor of slow clearance of ascites [20].
In a retrospective study, it was found that if the MELD score is less than 11, the post operative mortality is low and risk of surgery is acceptable. The mortality at 30, 90 days and 1 year was 10%, 17% and 28%, respectively [21]. However, it is advisable to conduct surgery in this patient group at an institution with a centre for liver transplantation. With a MELD score of 16-20, the risk of 30 day, 90 day and 1 year mortality is 44%, 55% and 70% respectively. This increases with a rise in MELD score. Therefore, elective procedure should be postponed with score>20. For scores between 12-19, transplant evaluation should be completed before surgery so that they can proceed with urgent transplant, if required. The final score at which elective surgery should be postponed until after liver transplant may also vary with the surgical expertise and organ availability. Any surgery in a decompensated cirrhotic should be done in a tertiary care institute with intensive care support and if possible, liver transplant facilities.
Merits
The MELD score has several distinct advantages over the Child classification [4].
1. The variables which constitute MELD are selected by statistical analysis rather than clinical judgement.
2. The variables are objective and calculated from easily available laboratory parameters.
3. It does not rely on arbitrary cut off values.
4.
Appropriate weight is given to each variable according to its influence on prognosis.
5.
The MELD score is a continuous variable from 6 to 40. This helps in a better assessment of a larger population.
1.
Calculation of MELD score is difficult, not user friendly and needs a difficult algorithmic computation.
Absence of clearly defined cut-off values for categorizing cirrhotic patients
3. Absence of validation in certain clinical scenarios [4].
4.
Complications of cirrhosis which could have a significant impact on the prognosis like portopulmonary hypertension, hepatopulmonary syndrome, hepatocellular carcinoma, hyponatremia, female gender and complications of portal hypertension like ascites, variceal bleed are not considered.
5.
Patients with refractory ascites, normal creatinine, and preserved hepatic function could be under-scored with MELD.
6. Even objective parameters like creatinine are subject to changes at different laboratories.
Addition of serum creatinine
Giannini et al., [22] prospectively derived the CTP creatinine score from 145 patients and compared it with the the CTP and MELD scores to evaluate 3 month survival in patients with cirrhosis. Patients with serum creatinine<1.1 were assigned a score of 1, serum creatinine between 1.2-1.8 was assigned a score of 2 and those with serum creatinine>1.8 were assigned a score of 3. It was observed that though the creatinine modified score had better prognostic accuracy than the CTP score, it was not better than the MELD score.
CTP D score
To overcome the drawback of the ceiling effect, attempts have been made to modify the CTP score by adding another dimension i.e. CTP D class. An additional 1 point was given for patient whose serum bilirubin level and PT prolongation were more than 8 mg/ dL and 11 seconds, respectively and there was a decrease in serum albumin level below 2.3 g/dL. A modified CTP score of 16-18 indicates severely decompensated cirrhosis, was proposed as CTP class D [23]. It was prospectively compared with the original CTP score and MELD score in 436 cirrhotic patients to asses 3 and 6 month mortality.
The predictive ability of the modified CTP was significantly better than original CTP system and was similar to the MELD system. It was able to differentiate disease severity and improve its performance by partially offsetting the ceiling effect. Majority of the patients had chronic Hepatitis B infection in which this modified CTP score was evaluated therefore may not be readily applicable where alcoholism and Hepatitis C are common etiologies. In India, the major etiology of end stage liver disease is Hepatitis C [24]. Therefore, further validation this new class is necessary across different clinical scenarios.
It can be considered a good tool for assessment of severity in centres with non availability of computerised systems for calculation of MELD scores.
Apart From differentiating disease severity i.e. CTP C and CTP D, there is very little role of this classification for the anesthetist.
Modifications of MELD Score
Updated MELD score Score derivation: Liver transplant candidates with mild hepatic synthetic dysfunction and marked renal insufficiency may have a higher MELD score than candidates with severe liver disease and normal renal function [25]. Since the adoption of MELD, the number of kidney and liver transplants has increased from 2.6% in 2001 to 5.2% in 2005 [26]. This demonstrates that creatinine is heavily weighed in the existing MELD.
It is assumed that mortality is constant for a creatinine less than 1 mg/dl in the original MELD. For a hypothetical increase in serum creatinine from 0.3 mg/dl to 0.6 mg/dl, it reflects a 50% reduction in glomerular filtration rate (GFR). In view of the poor nutritional status, a relatively large numbers of patients are likely to have a serum creatinine of <1 mg/dl at the time of listing.
To overcome this, the updated MELD was derived from 38,899 retrospective patient's waitlisted for liver transplant. To preserve the non negative property of each component, and yet to retain the lower limits, the updated MELD was scored by adding 1 to the value of the individual parameters. Hence, the actual value of the individual parameters can be used instead of the values assigned as lower or upper limit in the original MELD score (Table 3).
Score variables and range: It has been found that candidates with higher serum creatinine (and, by definition, lower bilirubin and/or INR to result in the same MELD score) had significantly lower mortality than candidates with lower serum creatinine (and therefore higher bilirubin and/or INR). In contrast, patients at the same MELD score with higher bilirubin had significantly higher mortality. Hence the Updated MELD assigns lower weight to creatinine and international normalized ratio and higher weight to bilirubin. Since the score is using the actual values of parameters for calculation, no range or capping of the score is done.
Score derivation
The MELD score was originally developed based on data from patients who underwent TIPSS. The refit MELD proposed has been prepared from data of patients who are on waiting list of liver transplant [27].
Wait-list data from adult primary liver transplantation candidates from the Organ Procurement and Transplantation Network were divided into a model derivation set (number of patients=14,214) and validation set (number of patients=13,945).
Optimized MELD score implemented new upper and lower bounds for creatinine (0.8 and 3.0 mg/dL, respectively) and international normalized ratio (1 and 3, respectively). Patients receiving renal replacement therapy were automatically assigned the upper bound for creatinine (3 mg/dL) ( Table 3).
Score variables and range
The importance of INR has been reduced in the new formula because it was found that the risk of death was less beyond an INR of three [3]. The serum creatinine demonstrated a triphasic pattern with risk of death, which was linear between 0.8 mg/dL and 3.0 mg/dL. It has been argued that the original upper and lower limits set for the three variables in the United Network for Organ Sharing (UNOS) MELD were based entirely on the clinical intuition of the policy-making body. The new upper limit boundary for INR addresses recent concerns that that the INR might not be an ideal marker to gauge coagulopathy associated with liver dysfunction [28].
It is well known that serum creatinine is influenced by muscle mass, which is frequently decreased in patients with end stage liver disease [29]. The new lower limit of 0.8 mg/dL makes intuitive sense because in patients with end-stage liver disease, normal creatinine does not necessarily mean normal renal function [30]. Lowering the upper limit from 4.0 to 3.0 mg/dL because there is too much emphasis on renal function in the MELD score and that patients with a component of intrinsic renal function are disproportionately advantaged under the current scheme. Score range for refit MELD have not been prescribed yet.
MELD sodium (Na) score
MELD underscores the patients with normal creatinine, preserved hepatic function cand refractory ascites. Patients with persistent ascites with a low serum sodium and a MELD score below 21 are at high risk of early death [31]. The role of hyponatremia as a predictor of mortality has been established for patients on LT waiting list leading to several attempts to incorporate serum sodium (S Na) into the MELD score [32,33]. A modified score including serum sodium, termed MELD-Na, has been proposed as an alternative to MELD score [34] (Table 3).
The accuracy of MELD-Na was shown to be slightly superior to that of MELD in candidates for transplantation [33][34][35]. A MELD Na level more than 10 was found to be an independent risk factor for
Demerits
Scores incorporating serum sodium should be interpreted with caution. Many of these patients are on diuretics for ascites, renal dysfunction requiring dialysis, on hypotonic fluids like dextrose. All these conditions can cause alterations of serum sodium. In such patients, alternate scores should be considered.
Integrated (i) MELD Score
The i MELD score incorporates age and serum sodium to increase the prognostic capability. It has been found to be more accurate than the original MELD, in predicting the mortality at 3, 6 and 12-months in an independent cohort of patients with cirrhosis listed for liver transplantation [37] (Table 3).
In a retrospective study of 190 patients with cirrhosis undergoing elective surgery, MELD and 4 MELD based indices were compared with CTP. i MELD was found to have the highest prognostic capacity for predicting mortality after elective surgery. For an i MELD score of less than 35, 35 to 45, and more than 45, the probability of death was 4, 16 and 50.1% respectively [38].
MESO Index
MESO index was retrospectively developed from 213 cirrhotic patients. A value of more than 1.6 independently predicted a higher mortality rate [39] (Table 3).
Refit MELD Sodium
Authors who have proposed the MELD score have also proposed the Refit MELD Na score in the same study [25]. They found that the 90-days wait-list mortality increased as the Na decreased between 140 mEq/L and 125 mEq/L. There was a significant interaction between sodium and bilirubin. The impact of Na on mortality became smaller as the serum bilirubin increased. This interaction was most pronounced when serum bilirubin was between 1 and 20 mg/dL (Table 3).
MELD Gender
The serum creatinine is poorly reflective of renal dysfunction in cirrhotic patients [31]. This issue may be magnified in females because for a given level of creatinine, on an average, women have a lower GFR than men due to their reduced muscle mass [40]. In fact, this sexrelated difference in creatinine concentrations may partially account for gender disparities in outcomes on the waiting list in the MELD era. In an analysis of United Network of Organ Sharing (UNOS) data, women were more likely than men to die or become too sick for transplantation and less likely to receive a transplant [41]. Therefore, it has been proposed that a correction factor for gender should be introduced or a more accurate serum marker of renal function could be used, such as cystatin-C to be substituted in prognostic scores [42].
D MELD (Donor MELD)
MELD has been found to be a good predictor of wait list mortality since its introduction in 2002. However, it is a poor predictor of post transplant mortality. The reason for this may be that numerous donor and recipient risk factors interact to influence the probability of survival after liver transplantation. The mortality risk of different donor/recipient combinations is less well defined [43].
Avoidance of D-MELD scores above 1600 has resulted in improved results for subgroups of high-risk patients with donor age ≥ 60 and those with preoperative MELD ≥ 30. D-MELD ≥ 1600 accurately predicted worse outcome in recipients with and without hepatitis C.
Demerits
D MELD score has limited utility in regions where deceased organ availability is limited and majority of the transplants are from live related donors.
Comparative Evaluation of Prognostic Scoring Systems
The refit MELD and updated MELD have been compared with MELD, Meso index, MELD Na and Refit MELD Na by Magdee et al. in 27473 patients [44]. This study was based on the number of lives that would have been saved had additional donor livers been available. Therefore they compared the models with respect to lives saved on transplant list. With respect to number of lives saved there was no significant difference among the models. But the MELD score performed the poorest and the refit MELD performed the best. The degree to which each score predicted death in a month from best to worst were MELD Na, refit MELD Na, MESO, refit MELD and updated MELD.
A Korean study compared the refit MELD, refit MELD Na with MELD, MELD Na and CTP score to predict three month mortality in 882 patients with cirrhosis [45]. The most common etiology of cirrhosis in this study was alcohol. The refit MELD Na was found to be a poor predictor as compared to MELD, MELD NA, and refit MELD. The MELD Na was the best performing score.
The same authors have compared the refit MELD and refit MELD Na with CTP score in patients with cirrhosis and ascites to asses three month mortality [46]. Refit MELD and refit MELD Na showed good predictability for 3 month mortality. But refit MELD Na was not found to be better than refit MELD, inspite of the known relationship between hyponatremia and mortality in cirrhotic patients with ascites.
The above studies suggest that the refit MELD appears to be the most promising modified MELD score. But it has not been evaluated in perioperative/periprocedural settings. Comparison with MELD, MELD Na and CTP scores in such settings is needed.
In a recent retrospective study on 490 cirrhotics who underwent surgery under general anaesthesia, CTP, MELD and MELD Na were compared with respect to the postoperative mortality at 90 days. It was found that the CTP and MELD Na were superior to MELD score in predicting mortality at 90 days [36]. In non-transplant setting also, Cholongitas et al. reviewed literature and stated that MELD does not perform better than CTP score [47].
Cirrhotic patients with Oesophageal variceal Bleed, a MELD of 18 or more, platelet count less than 100,000 and requiring transfusion of 2 or more units of PRBC were at an increased risk of in hospital mortality [48]. In fact, Kumar et al. have suggested that adding the variceal status to CTP score improves its performance in predicting early mortality in cirrhosis [49].
In trauma patients with liver dysfunction addition of specific scores like MELD or CTP to Injury severity score (ISS) also enhances the ability of the latter to predict mortality [50].
However, in a very recent prospective, observational study of 216 cases of hospitalised patients with decompensated cirrhosis, CTP and MELD scores were calculated and followed till discharge or death. The authors concluded that MELD is superior to CTP score in predicting survival at the time of discharge in decompensated cirrhotics. Addition of renal failure carries a poor prognosis and has a good prognostic value, even better than CTP/MELD [51].
In patients with cirrhosis undergoing major surgical procedures, the risk of mortality within 7 days of surgery is best assessed by American Society of Anaesthesiologists classification of physical status of the patient, whereas mortality after 7 days is best determined by MELD score [21]. Teh et al. have added the ASA classification to the original version of MELD scale as developed by investigators at Mayo Clinic. This modified prognostic scoring system can be used to calculate 7-day, 30-day, 90-day, 1-year, and 5-year surgical mortality risk based on a patient's age, ASA class, INR, and serum bilirubin and creatinine levels (the last 3 items constitute the MELD score) [21].
Other Prognostic Indicators Sarcopenia
Muscle depletion (sarcopenia) has found to be an independent predictor of wait list mortality in patient with liver disease [52]. This is diagnosed by the measurement of L3 cross-sectional area on CT scan. Sarcopenia is present if the value is less than 52.4 and 38.5 cm 2 /m 2 in males and females, respectively. It was found that the outcomes of patients with low MELD scores and sarcopenia were similar to the outcomes of patients with high MELD scores with or without sarcopenia. A diagnosis of sarcopenia can identify those patients who may benefit from more intensive nutritional supplementation and exercise therapy, both of which have been shown to improve outcomes for patients with cirrhosis. Subjective nutritional assesment tools like body mass index and subjective global assessment have proven to be inadequate in predicting mortality in this group of patients.
Demerits
Sarcopenia is objective but is time consuming due to the need of cross sectional muscle imaging.
Von Willebrand Factor Levels
Von Willebrand factor antigen (vWF-Ag) is elevated in patients with liver cirrhosis. This may be due to endothelial activation because of portal hypertension or induction of the synthesis of vWF-Ag in the cirrhotic liver. Reduced activity of ADAMTS13 (vWF-Ag cleaving protease) also increases the levels of vWF. Recently, Ferlitsch et al. have established the clinical significance of vWF levels. They found that a level>315% identified cirrhotic patients with a higher mortality and added prognostic value to the MELD score. In compensated patients with a vWF-Ag value<315%, median time to decompensation or death was 59 months, compared to 32 months in patients with vWF-Ag levels>315% [53].
Demerits
The limitation of using vWF levels for prognostication is that it can be fallaciously high or low in certain clinical scenarios. Infections, malignancies, interferon therapy and physical therapy can elevate vWF levels, whereas active bleeding and hereditary deficiency could reduce them and lead to false prognostication [53].
Prognosis in Setting of Critical Care
Cirrhotic patients admitted to an Intensive Care Unit (ICU) have a poor prognosis. Aim of prognostic models in intensive care settings is to identify patients who will benefit from aggressive treatment. A focussed approach in this situation can either help in the recovery of hepatic function or act as a bridge to ''rescue'' transplantation.
Prognositic scores in critically ill cirrhotic patients can be classified in three main categories Patients with liver disease admitted to ICU usually present with multiorgan dysfunction. Therefore scoring systems like CTP score and MELD which determine severity of liver disease have not found to be good predictors of ICU mortality. CTP score does not include any marker of other organ function and MELD score lacks any indicator of portal hypertension, the complications of which are a frequent cause of admission to ICU.
APACHE score
The original APACHE score was developed in 1981 to classify groups of patients according to severity of illness and was divided into two sections: a physiology score to assess the degree of acute illness; and a preadmission evaluation to determine the chronic health status of the patient [54]. APACHE II, now the world's most widely used severity of illness score. In APACHE II, there are just 12 physiological variables. The worst value recorded during the first 24 hours of a patient's admission to the ICU is used for each physiological variable. The principal diagnosis leading to ICU admission is added as a category weight so that the predicted mortality is computed based on the patient's APACHE II score and their principal diagnosis at admission. Subsequently APACHE III and APACHE IV have also been developed.
Simplified acute physiology score (SAPS)
SAPS was developed and validated in France in 1984. It used 13 weighted physiological variables and age to predict risk of death in ICU patients. SAPS is calculated from the worst values obtained during the first 24 hours of ICU admission [54]. In 1993, Le Gall et al. developed SAPS II, which includes 17 variables: 12 physiological variables, age, type of admission, and 3 variables related to underlying disease. SAPS III has also been developed in 2005.
SOFA score
The SOFA score defines organ failure by a score of three or four for each of the six respective organ systems (respiratory, cardiovascular, hepatic, renal, coagulation and neurologic) [54].
The development of three or more organ failures carries an extreme risk of death, which is higher in cirrhotic patients (an average of 79%) when compared with general ICU patients (55%). The mortality rate of cirrhotic patients with septic shock is higher than in noncirrhotic patients [55,56].
Organ failure scores like Sequential organ failure assessment (SOFA) have been found to perform better [55]: SOFA>SAPS II>MELD>Child-Pugh [57]. Mortality is best correlated with a SOFA score above nine.
The APACHE II and SAPS II score are the most commonly studied scores along with SOFA score to predict mortality in critically ill patients with liver disease [55]. In all these studies SOFA score has emerged the clear winner in the ability to predict mortality. Accuracy of organ failure scores increase when they are reassed 2 days after admission To ICU. Reassesment at 48 hours therefore may be a useful guide to the degree of intensification of efforts.
The European Association for the Study of the Liver-chronic liver failure (EASL-CLIF) Consortium recently defined the CLIF SOFA score with cut off values specifically identified in cirrhotic patients [58] ( Table 4).
Like the original score, the CLIF-SOFA score assessed six organ systems (liver, kidneys, brain, coagulation, circulation, and lungs), but it also took into account some specificities of cirrhosis. The CLIF SOFA score was developed based on clinical experience of the authors. Based on the score the they identifed four groups of patients with varying number of organ failures. They found that patients with two organ failures had a 28 day mortality rate of 32%, while those with three or more organ failures had a 28 day mortality of 76%.
Sixty percent of the patients they studied had alcoholic liver disease and twenty percent had hepatitis C related liver disease. Therefore the authors have suggested evaluation of this score where other etiologies of liver disease may be predominat e.g Hepatitis B. Further validation of this score is therefore recommended.
The cause of ICU admission is also associated with the prognosis of patients. Patients admitted in ICU for acute variceal bleeding or hepatic encephalopathies have a markedly improved ICU survival of 76.5% vs. 36.2% for patients admitted for infection [59].
Karvellas et al. retrospectively assessed the outcome of 198 critically ill cirrhotic patients who received a liver transplant (LT) while in ICU in five transplant centres in Canada [60]. Eighty eight percent were on vasopressors, 56% received renal replacement therapy and 87% were mechanically ventilated prior to LT. The SOFA score was 12.5 ± 4 on ICU admission, 13 ± 5 at 48 hours and 14 ± 4 on the day of LT. Mortality after LT was 16% at 90 days, 26% at 1 year and 38% at 3 years. A SOFA score ≥ 10 in cirrhotic patients usually predicts mortality in >90% in a median time of 8 days without a liver transplant. The authors found that SOFA on admission, 48 hours after ICU admission and on the day of LT was not associated with increased risk of 90-day mortality. The only independent risk factor of death identified was the age. They concluded that SOFA at 48 hours is currently the best score to predict mortality in cirrhotic patients admitted to ICU. It is associated with a higher risk of death waiting for LT and is not associated with a worse outcome after LT while in ICU. These results appear to be promising for further prospective evaluation in regions with successful deceased donor transplant programs.
The persistence of three or more organ failures and the need for three or more organ supports (i.e. inotropic support, mechanical ventilation and continuous renal replacement therapy) may lead to consider a limitation in life sustaining treatments, as a fatal outcome is almost constant [55]. A multidisciplinary approach between hepatologists, intensivists and transplant surgeons is mandatory.
Conclusion
Inspite of the availability of various prognostic models for risk stratification and prediction of morbidity and mortality in patients with cirrhosis, the score most popularly used is the CTP score. It allows rapid bedside prognostication and is fairly reliable. It is still a good tool for anaesthesiologists for prognostication of patients with liver disease who undergo non transplant surgery. But the MELD score has recently challenged the flagship bearer status of the CTP score [2].
Prognostic scoring systems, especially the MELD score is constantly undergoing changes. In view of worldwide differences of liver transplantation with respect to indication and method (deceased donor v/s live donor), prognostication should be suited to the particular region. Few countries like Canada and United Kingdom have developed their own models of CAN wait and UKELD which are working well for them [61,62]. It is time, we developed our own Asian or Indian model for prognostication. ICU scores like the SOFA scores are more reliable in the critically ill cirrhotic patient. Modified scores like CLIF SOFA scores need further validation. Newer indicators like assessment of sarcopenia seem to be promising, but search for simpler cheaper and safer techniques of assessment is needed.
|
2019-03-12T13:05:17.024Z
|
2014-09-28T00:00:00.000
|
{
"year": 2015,
"sha1": "9c061751ef75a7d0d892b35bcc6dc6c5e5c70a5f",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.4172/2167-0870.1000186",
"oa_status": "HYBRID",
"pdf_src": "MergedPDFExtraction",
"pdf_hash": "e47456788c3077f8c49c3dde72bdab2ef2bff103",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
62906792
|
pes2o/s2orc
|
v3-fos-license
|
A Flexible Network Architecture for 5G Systems
,
Introduction
Since the early research phase of the fifth generation (5G) starting in 2012 [1][2][3][4], the development of concepts for the 5G system (5GS) has progressed at a rapid pace. Within the 5GS, 2 Wireless Communications and Mobile Computing operators unique opportunities to offer new business models to consumers, enterprises, verticals, and third-party tenants and address such various requirements. To this end, both research projects [5][6][7][8] and standardisation efforts [9,10] have described the main elements of the 5G architecture. Third generation partnership project (3GPP) has already completed the early-drop "non-standalone" release of 5G by December 2017 [11], the main-drop "standalone" release of 5G by June 2018 [12], and the late-drop release of 5G with specification of remaining architecture options by the end of 2018 [13].
Although all these aforementioned efforts have provided a solid baseline architecture, in our view there is still room for 5GS enhancements to better fulfill the 5G vision of supporting diverse service requirements while enabling new business sectors often referred to as vertical industries. This paper aims to define a flexible, adaptable, and programmable architecture for 5G mobile networks taking into account current gaps in the literature.
The rest of the paper is organised as follows. In Section 2, we perform a thorough 5GS gap analysis, in order to identify enhancements that can be included in the future refinements of the 5G architecture. In Section 3, we detail some key enabling innovations to address identified gaps. In particular, we present the design guidelines for (i) flexible cloudification of protocol stack, (ii) adaptive inter-slice control, and (iii) leveraging experiment-and implementation-driven modelling and optimisation. Section 4 provides the details on proposed architecture reference model with the envisioned four layers: (i) network layer, (ii) controller layer, (iii) management and orchestration (M&O) layer, and (iv) service layer. Section 5 presents evaluation studies and analyses on some selected identified innovation concepts in Sections 3 and 4. Finally, Section 6 concludes this paper drawing also the plans ahead.
The key novelties of the architecture and approaches proposed in this paper include the following: (i) When designing the proposed architecture, we revisit both 3GPP and the ETSI Network Function Virtualisation (NFV) network management and orchestration functions. We extend the reference architectures proposed by 3GPP and ETSI NFV by building on these architectures while addressing several gaps identified within the corresponding baseline models.
(ii) Within our proposed architecture, there are several network functionality that are not specified elsewhere and need to be designed. We present the design guidelines of some of the key modules within the architecture, corresponding to innovative elements.
(iii) One of the key enabling technologies within our architecture is network slicing. To apply this technology to specific use cases, we need new network functions that are instantiated with the network slice orchestrated by the architecture, satisfying the specific requirements of the use cases. This is addressed here by leveraging proposed enabling innovations.
Current Gaps and 5G System Enhancements
We have done a gap analysis on the consolidated view coming from the literature, the work of the relevant fora, consortia, SDOs, and 5G-PPP Phase 1 projects along with 5G-PPP working groups (WGs). A summary of the gap analysis is outlined as follows (for a thorough gap analysis, the interested readers are referred to [16]).
(#1) Interdependencies between Network Functions (NFs) Colocated in the Same Node. Traditional protocol stacks have been designed under the assumption that certain NFs residing in the same node, i.e., fixed location and NF placement; while they work close to optimality as long as such NFs are colocated in the same node, they do not account for the possibility of placing these NFs in different nodes. The logical and temporal dependencies between NFs should be relaxed and (as much as possible) removed to provide a higher flexibility in their placement. An example of such relaxation is to loosen such strict timing dependency as described in [17]. In particular, [17] proposes to opportunistically send ACKs based on the estimated channel quality instead of performing the complete decoding of the frame and then sending the corresponding N/ACKs. By decoupling the HARQ from the complete decoding, the latter can be executed in a more centralised manner in cloud data centres. This, in turn, translates into higher multiplexing gains and fewer constraints (in terms of minimum bandwidth and maximum latency) imposed by the links that connect those functions.
(#2) Orchestration-Driven Elasticity Not Supported (Lack of Slice-Aware Resource Elasticity). It is necessary for the architecture to flexibly shift NFs to nodes that better fit the specific requirements of each covered service; when doing so, we need to take elasticity considerations into account.
In the 5G systems, where each slice is composed of multiple virtual NFs (VNFs), the elastic allocation available resources (either radio resources or computational resources) to different network slices based on their demands, requirements, and Service Level Agreements (SLAs) are essential. The architecture of 5G networks should provide the required elements and flexibility to implement elastic slice-aware resource elasticity while preserving the isolation of each network slice [18].
(#3) Fixed Functional Operation of Small Cells. In the current networks [19], the functional operation of small cells does not change relative to service requirements or the location of the small cell, which can be, e.g., unplanned and dynamic. That is, the functional operation and the associated operation mode of the small cells based on the predetermined functional operation remain fixed. For instance, a fixed relay can be typically deployed as radio frequency (RF) amplify-andforward /repeater or layer 3 (L3) decode-and-forward (DF) node. This can also incur higher operational expenditure (OPEX), when the network is planned for the highest or peak service requirements [20]. However, slice awareness and 5G tight KPIs can necessitate on-demand flexible small cell operation.
(#4) Need for Support for Computational Offloading. Current architectures do not fully support delegating costly NFs beyond the network edge towards RAN (e.g., for cases like group mobility in D2D context). Addressing this gap can result in saving energy consumption, signalling overhead, or offload resource demanding tasks when needed. Some further enhancements to architecturally support such offloading scenarios started in Release 15 specifications [21] to improve remote UE reachability and to support efficient traffic differentiation, signalling, and service continuity at a controlled level of device complexity and power consumption on UEs.
(#5) Need for Support for Telco-Grade Performance (e.g., Low Latency, High Performance, and Scalability). Most of management and orchestration technologies are inherited from IT world. Adopting such technologies in the telco domain without key performance degradation is a great challenge as the added functionality in the control and M&O layer, as well as the more modular NFs, should still offer the same telco-grade performance, without degradation [22].
(#6) E2E Cross-Slice Optimisation Not Fully Supported. Architecture should allow for the simultaneous operation of multiple network slices with tailored core/access functions and functional placements to meet their target KPIs [23].
(#7) Lack of Experiment-Based E2E Resource Management for VNFs. Current 5G systems are missing E2E resource management of VNFs that takes advantage of E2E software implementations on commodity hardware in a dynamic manner. To design resource management algorithms that perform well in reality, we need to rely on more elaborate models that build insights (e.g., a quantification of the resource consumption profile per VNF) gained from E2E experiments. In this direction, the work in [24] investigates the computational consumption of state-of-the-art open source software solutions for the RAN stack. Analogously, the authors of [25] measure the computational requirements of a video server.
(#8) Lack of a Refined 5G Security Architecture Design. There are various critical gaps in the literature and architectural deployments related to management and orchestration, accountability, compliance, and liability, as well as performance and resilience. For example, there is no established security architecture for network slice deployment models which include 3rd-party-owned network infrastructure to implement highly secure mobile communication services across public and private infrastructure domains [26].
(#9) Lack of a Self-Adaptive and Slice-Aware Model for Security. E2E network slicing demands a revaluation and research on various aspects of traditional security (e.g., privacy, integrity, zoning, monitoring, and risk mitigation) [26].
(#10) Need for Enhanced and Inherent Support for RAN Reliability. RAN reliability should be a built-in solution/element of the architecture, through the application of mechanisms such as multiconnectivity and network coding, e.g., as proposed in [27].
(#11) Indirect and Rudimentary Support of Telco Cloud
Resilience Mainly through Management and Control Mechanisms. The architecture should address resilience in a structured way taking into account different aspects (e.g., individual network elements (NEs)/NFs, telco cloud components, fault management, and failsafe mechanisms) [28].
(#12) Need for (Radio and Computational) Resource Sharing Strategy for Network Slices. While basic mechanisms for multi-slice resource management have been studied in 5G-PPP Phase 1 projects, elastic mechanisms need to be devised which improve the utilisation efficiency of the computational and radio resources by taking advantage of statistical multiplexing gains across different network slices [29,30]. Furthermore, inter-slice radio resource sharing has been investigated in literature [31] where slice-aware RAN clustering, scheduler dimensioning, and adaptive resource coordination is discussed are a first attempt towards filling this gap. Furthermore, for self-backhauling RAN scenarios, inter-slice resource sharing solutions [32] can be incorporated in order to allocate backhaul/access resources optimally among slices. Table 1 provides a summary of gap analysis with respect to ongoing 5G system architecture design efforts in the industry and academia.
Enabling Innovations for 5G Services
In this section we detail some key identified enabling innovations for 5G services to address several gaps identified before. In particular, we present the design guidelines for (i) flexible cloudification of protocol stack, (ii) we present adaptive network slicing, inter-slice control and management, and (iii) we explore leveraging experiment-and implementationdriven modelling and optimisation to refine models on computational behaviour and derive the corresponding algorithms.
Cloud-Enabled Protocol Stack.
In a cloud-enabled architecture, network nodes become general-purpose processors capable of running any network functions (NF), and NFs are virtualised, decomposed, and flexibly placed in different locations. This flexibility can be beneficial for many different services. For instance, a network slice providing an eMBB service may have most of its VNFs at a centralised location, while an URLLC one may want to exploit resources at the edge.
Existing concepts for the flexible allocation of NFs use a protocol stack that is not necessarily optimised for this purpose (especially for the RAN part). Indeed, "traditional" protocol stacks have been designed under the assumption that certain functions reside in the same (fixed) location and work close to optimality as long as such NFs are colocated in the same node. As a result of this, in the traditional protocol stack we typically have interdependencies between the NFs This may compromise the overall gains obtained from the flexible function allocation. To overcome this problem, one of the key innovations needed to fully exploit the proposed orchestration-driven architecture is the redesign of the protocol stack with the goal of leveraging the benefits of the flexible function decomposition and allocation, which we call the orchestration-enabled protocol stack. The aim is to relax and (as much as possible) remove the logical and temporal dependencies between NFs, with the goal of providing a higher flexibility in their placement. This results in a fundamental piece of innovation termed as network elasticity.
Elasticity of NFs involves the ability to scale the complexity of the NFs based on the available resources (i.e., processing power, memory, and storage): in case of a resource outage, NFs should adjust their operation to the available resources while minimising the resulting impact on network performance, thus providing a graceful degradation. As a result, we have NFs that are robust against "computational outages" if there are insufficient resources to perform the required tasks within the given time. This represents a new paradigm in the design of NFs: while traditional NFs have been designed to only fight the well-known "channel outages" concept, here we advocate for designing NFs that are also robust to the fluctuation of computational resources needed to accommodate the varying load of a network slice. In this way, the performance may degrade gracefully under computational outages, in contrast to traditional schemes where the performance degrades drastically in case of insufficient computational resources to decode all received data frames.
With a graceful degradation, several KPIs can be updated. Graceful degradation of VNF supports accommodating transient huge spikes of traffic demand without disrupting service while more resources are activated. It also prevents abrupt degradation of services when resources are missing due to temporary high demands. For capacity-limited deployments, cloud-aware network design improves the efficiency of network deployments (more users can be served using the same HW).
Next step in the same research path is considering the network elasticity in a multi-tenant environment. While offering isolation to network slices, it is possible to improve resource utilisation efficiency. However, serving multiple slices with different services and quality of service (QoS) requirements is a non-trivial task. To ensure achieving acceptable network performance, scaling of NFs for each network slices has to be done based on available computational resources, the slices' SLA, and demands in addition to allocated radio resources [23].
Inter-Slice Control and Management.
The network slice awareness in 5G will strongly affect several key design requirements in different domains.
RAN Domain Aspects.
One important area is the RAN design and particularly the control plane (CP) design, where multiple slices, with different optimisation targets, will require tailored access functions and functional placements to meet their target KPIs.
In slice-aware RAN, the CP can be categorised in the following groups of functionality based on the RAN Configuration Modes (RCM) framework.
Intra-RCM RRM: for slice-specific resource management and isolation among slices, utilising the same RAN is an open topic which is currently investigated. The conventional management of dedicated resources can be seen as intra-slice RRM, which can be tailored and optimised based on slicespecific KPIs.
Inter-RCM RRM/RRC: on top of Intra-RCM RRM, Inter-RCM RRM/RRC can be defined as the set of RRM policies that allow for sharing/isolation of radio resources among slices or slice types to optimise the resource efficiency and utilisation, by flexibly orthogonalising them in coarse time scales. Inter-RCM RRM can be defined as an "umbrella" functional block which dictates the RAN sharing and level of isolation/prioritisation among network slices or slice types. In this direction, an Inter-RCM RRM mechanism is proposed in [31], where slice-aware RAN clustering, scheduler dimensioning, and adaptive placement of intra-slice RRM functions are discussed in order to optimise performance in a dense heterogeneous RAN. Given the requirement of new access functions which can be tailored for different network slices, the distribution of RRM functionality in different nodes will be a key RAN design driver which can allow for multi-objective optimisation in a multi-layer dense RAN. The adaptive allocation of such functions is also envisioned as a key feature to cope with the dynamic changes in traffic load, slice requirements, and the availability of backhaul/access resources. To this end, one further inter-slice/RCM RRM functionality is proposed in [34] which performs traffic forecasting of different slices and allocates resources to slices in a proactive manner. Figure 1 shows the possible placement of Inter-RCM and Intra-RCM RRM and Radio Resource Control (RRC) functionality. Depending on the placement the interface requirements might be different due to the time/resource granularity of the CP functionality and their possible interconnections.
Core Domain Aspects.
Another key area to realise flexible endto-end network slicing is core network design. The design paradigm within the core network has gradually shifted towards a functional service-based architecture (SBA) in line with industry consensus [10] where CP NFs are interconnected via a common bus to each other, termed as servicebased interface (SBI). SBA is expected to have the advantage of short roll-out time for new network features, extensibility, modularity, reusability, and openness [35]. SBA allows the definition and instantiation of flexible E2E networks, which can be customised by network operators' or vertical industries' requirements, in terms of performance, capabilities, isolation, etc. In other words, this allows the support of network slices, i.e., independent logical networks, either sharing partly/entirely the infrastructure they are instantiated on or isolated and deployed over separate infrastructures. 5G devices will be able to access the network core and utilise supported services from a number of network slices.
One key function, in this direction, is the network slice selection function (NSSF) as NF dedicated to selecting a proper network slice instance (NSI) for the 5G devices. There are several other NFs envisioned to customise network slices capabilities. For example, the Session Management Function (SMF) may allow the support of different UP protocol models, such as IPv4/IPv6 or Ethernet. The Policy Control Function (PCF) allows customising the policy framework on network slice basis. Finally, the Unified Data Management (UDM) function enables different authorisation, authentication, and subscription management mechanisms upon network slice tenant needs.
It should also be noted that, thanks to SBI, 3rd parties can also influence the network behaviour and extend and customise network slices capabilities via the inclusion in the system of proprietary non-standard Application Functions (AFs) or via exposing theirs services to other NFs, e.g., via Network Exposure Function (NEF).
Context-Awareness, Sharing, and Optimisation. A separate distinguishing feature of next generation networks is supporting network and user context as a means to further optimise network and cross-slice operations via supporting network analytics capability embedded in the general framework. Context-awareness is required in order to have flexible and dynamic function deployment as well as unified resource allocation and optimisation decisions between slices and users. This can be achieved by having shareable data storage mechanisms such as databases deployed as VNFs along with mechanisms to publish/subscribe to the shared context information.
Such functionality has currently been envisioned within the core via the Network Data Analytics Function (NWDAF). NWDAF provides the network core with the ability to collect and analyse per slice aggregated data and to aid network optimisation via interaction with other NFs (e.g., NSSF or PCF).
From RAN and terminal perspective, user devices are natural data collection points to gather above analytics within the network. As users can simultaneously connect to or switch across different slices (e.g., in case of mobility), they can have more prominent role for data preparation for the network to cleanse/normalise the information and to identify earlier anomalies compared to the past intra-slice and/or cross-slice information they have gathered. This opens the horizon for another level of context-awareness within the next generation networks.
Experiment-Driven Optimisation.
In a fully softwarised network architecture, the optimisation of the orchestration of the whole network needs to consider the computational behaviour of nodes and NFs. Orchestrating NFs imposes some constraints, especially at the edge of the network where it is likely that resources are limited or very expensive. To perform the placement of NFs based on computational resources, traditional approaches assume that NFs (and cloud locations) consume (and offer) a fixed amount of resources (i.e., processing power, memory, and storage) [36,37]. However, this model is very coarse and clearly insufficient to understand the performance of a real environment in which the computational load fluctuates significantly over time. Much more accurate results can be achieved if accurate models of the computational behaviour (of both VNFs and NFVI) exist.
Designing algorithms to carefully allocate VNFs to the nodes in the network involves the following challenges: (i) characterising the VNF's temporal behaviour by modelling the occurrence of peaks of resource consumption and periods of lower load, (ii) evaluating the (non-negligible) overhead incurred by computational resources used to run system management software, and (iii) assessing the impact of the communication environment as well as the logical dependencies between NFs, which introduce statistical dependencies in the computational demands of such functions.
It is worth noting that the implementation techniques (i.e., programming approaches) used in realisation of VNFs influence the behaviour VNFs. Hence, these challenges necessarily need to rely on experiments that evaluate the real computational behaviour of the different components. This can be accomplished following the methodology depicted in Figure 2. First, measurements need to be performed in real deployments of VNFs in fully operational networks. Then, their behaviour can be modelled to clearly understand both the VNF resource consumption and the infrastructure capabilities. Finally, this input can be used to design enhanced orchestration algorithms.
Via having different functionality virtualised, the cloud infrastructure providers have to develop an experimental procedure to meet the QoS requirements of each VNF optimally. Scaling and elasticity decisions (either vertical or horizontal) cannot be made without having a practical experimental optimisation approach. Experiment-driven optimisation is enabled through measurement campaigns (i.e., a monitoring process). The measurements from these campaigns feed a modelling procedure, which models the VNF behaviour regarding their computational, storage, and networking resource demands. The resulting models may facilitate the overall resource management of the cloud infrastructure. In [38], the experimental modelling of physical layer is presented.
Algorithms and functions that apply upon the 5G protocol stack can improve their performance by exploiting experiment-driven insights and, thus, taking more intelligent decisions. In this context, the experiment-driven modelling and optimisation is a key innovation enabler filling the current gap on experiment-based E2E resource management for VNFs. This also brings a new paradigm in network management and orchestration by feeding with experimentbased inputs.
Flexible Architecture Design
Following the gaps analysis in the literature and key enabling innovations to realise the 5G services, we have designed a flexible architecture enabling dynamic network slicing as part of a 5GPPP Phase 2 project, 5G-MoNArch [14], to meet 5G systems objectives. The reference architecture model proposed here extends the reference architectures proposed by 3GPP, 5GPPP Phase 1 projects, and ETSI by building on these architectures while addressing several gaps as identified in Section 2.
Overall Design Principles.
The proposed overall functional network architecture consists of four different layers, identified as network layer, controller layer, management and orchestration layer, and service layer as shown in Figure 3.
A key contribution of this paper is the definition of the role of each layer, the relationship between layers, and the identification of the required network functions within each of the layers.
The network layer comprises the VNFs and physical NFs (PNFs) of both control plane (i.e., cVNF, cPNF) and user plane (i.e., uVNF, uPNF). NFs can include, for example, 3GPP Rel. 15 Figure 3: Proposed overall functional architecture [14]. security. Generally, the network layer can comprise different CP/UP architectures; i.e., also a 4G mobile network with EUTRAN and EPC functions could constitute an instance of the network layer. Interfaces towards the M&O layer are provided via the Itf-X reference point. It is an evolution of the 3GPP Itf-S interface between element manager (EM) and network element (NE), e.g., eNB, and facilitates domainspecific fault, configuration, accounting, performance, and security (FCAPS) management as well as domain-agnostic lifecycle management (LCM) procedures. For associating a UE to the correct NSI, the network layer uses the Single Network Slice Selection Assistance Information (S-NSSAI), which is provided by the UE. Moreover, the CN part of the CP in the network layer is realised as SBA [10]. Further details of the network layer are depicted in Section 4.2.
The controller layer realises the software-defined networking concepts [39], extends them to mobile networks, and therefore accommodates two controller types: (1) the coss-slice controller (XSC), e.g., a RAN controller for the control of cross-slice network functions (XNFs) that are shared by multiple network slices, (2) the intra-slice controller (ISC), e.g., a CN controller for intra-slice network function (INFs) within a dedicated CN network slice subnet instance (CN-NSSI).
These controllers expose a northbound interface (NBI) towards control applications and a southbound interface (SoBI) towards VNFs and PNFs in the network layer.
Interfaces towards the M&O layer are provided via the MOLI reference point. The controller layer facilitates the concept of mobile network programmability. Generally, software-defined networking (SDN) splits between logic and agent for any functionality in the network. This means that the NFs are split into the decision logic hosted in a control application and the actual NF in the network layer (usually a uPNF or uVNF) that executes the decision. In other words, for the given uVNF or uPNF, the according cPNF or cVNF would disappear. The controller resides "between" application and NF and abstracts from specific technologies and implementations realised by the NF, thus decoupling the control application from the controlled NF (cf. Figure 3). If no such split between control logic and agent is applied, i.e., the cPNFs and cVNFs incorporate both, the controller layer disappears. In this sense, it is an optional layer of our proposed architecture. Further details of the controller layer are described in Section 4.3.
The management and orchestration layer is composed of the M&O functions from different network, technology, and administration domains: 3GPP public mobile network management, ETSI NFV management and orchestration (MANO) [40], ETSI multi-access edge computing functions [41], management functions of transport networks (TNs), and private enterprise networks. Further, the M&O layer comprises the end-to-end M&O sublayer hosting the Network Slice Management Function (NSMF) and Communication Service Management Function (CSMF) that manage network slices and communications services, respectively, across 8 Wireless Communications and Mobile Computing multiple management and orchestration domains in a seamless manner. In the so-called virtualisation MANO domain, the ETSI NFV MANO architecture for LCM of virtual machines (VMs) is extended towards LCM of virtualisation containers (e.g., Docker). Therefore, it comprises, besides the ETSI NFV components, corresponding functions for LCM of containers. Therefore, the Virtualised Network Function Manager (VNFM) has according components for virtual machine infrastructure (VMI) and container infrastructure (CI). Similarly, the Virtualised Infrastructure Manager (VIM) contains a VMI management function (VMIMF) and a CI management function (CIMF). NFV orchestrator (NFVO) provides the dispatching functionality. Further, the layer accommodates 3GPP network management function, such as element and domain managers (EM and DM) and network management (NM) functions. Such functions would also implement ETSI NFV MANO reference points to the VNFM and the NFVO. The CSMF transforms consumer-facing service descriptions into resource-facing service descriptions (and vice versa) and therefore works as an intermediary function between the service layer and the NSMF. The NSMF splits service requirements as received from CSMF and coordinates (negotiates) with multiple management domains for E2E network slice deployment and operation. As a major architecture novelty, NSMF further incorporates a cross-slice M&O function for inter-slice management (e.g., common context between different slices/tenants, inter-slice resource brokering for cross-slice resource allocation, particularly in the case of shared NFs, etc.). In contrast, the cross-domain M&O function works on strictly intra-slice level, but across multiple network and technology domains. The M&O layer performs the management tasks on network slice instances (NSI), which are uniquely identified by an NSI identifier. An NSI may be further associated with one or more Network Slice Subnet Instances (NSSI). Further details of the M&O layer are described in Section 4.4.
The service layer comprises Business Support Systems (BSS), business-level policy, and decision functions and further applications and services operated by a tenant or other external entities. These functions of the service layer interact with the M&O layer via the CSMF, as described above. Figure 3 implicitly illustrates three fundamental design aspects that we have followed in our proposed architecture: (1) Support for E2E network slicing: the architecture allows for combining different options of slicing support across M&O and network layers for each slice instance. The first supported option includes slicespecific functions; i.e., each slice may incorporate dedicated and possibly customised functions that are not shared with others. The second option includes the possibility of operating functions (or function instances) that are shared by multiple slices and have the capability to address requirements from multiple slices in parallel. Figure 3 depicts this split into common or so-called inter-slice functions and dedicated (intra-slice) functions. This split can be maintained in the M&O layer, the network layer, and the optional controller layer; i.e., dedicated NFs may be controlled and managed by the tenant's own instance of ISC and M&O layer functions. Shared functions are usually operated by the mobile network operator (MNO) or the communication service provider. The MNO (together with potential thirdparty infrastructure providers) is also in charge of managing the infrastructure. The policies regarding the utilisation of shared functions, particularly the resource allocation to active slices, are determined by the cross-slice M&O function, and communicated towards the respective network layer functions for further enforcement. Finally, the third option is to not only have slice-dedicated NFs but also additionally assign the associated infrastructure hardware resources (HW), including spectrum, exclusively to a single slice. The slice-specific functions and shared functions in one logical slice are bound together by the network slice identifier at the network layer.
(2) SBA: the service-based interaction between core network CP NFs provides a set of features and associated advantages. Among others, NFs can be realised in a stateless manner since such state-related data (e.g., session data) are shared via a message bus, sometimes referred to as data bus. SBA facilitates the design of modularised NFs, uniform interaction procedures between NFs (e.g., NFs can offer their functionality as a service to other NFs), unified authentication framework between NFs, and concurrent access to services.
(3) Split of control and user plane: we apply a consistent split of control plane and user plane throughout all network domains, including RAN, CN, and TN. Among others, this allows for hosting associated CP and UP NFs in different locations and also facilitates aggregating CP and UP NFs differently. The split further allows independent scalability and evolution of NFs.
Network Layer.
In this section, the highlights of the CN, also referred to as 5G core network (5GC) and RAN, are outlined. Compared to the traditional reference-point-based network architecture design, SBA has the advantage of short rollout time for new network features, extensibility, modularity, reusability, and openness [35]. Our proposed architecture, which is based on SBA principles, allows the definition and instantiation of flexible E2E networks, which can be customised by network operators' or vertical industries' requirements, in terms of performance, capabilities, isolation, etc. In other words, 5GC reference architecture allows the support of network slices, i.e., independent logical networks, either sharing partly/entirely the infrastructure they are instantiated on, or isolated and deployed over separate infrastructures. 5G devices will be able to access 5GC and requiring services from a number of supported network slices. It should also be noted that, thanks to SBI, the reference architecture also provides 3rd parties with the possibility of influencing the network behaviour and extending and customising network slices capabilities via the inclusion in the system of proprietary non-standard AFs.
Although the foundations for 5GC have been successfully established, the general framework still appears to be not entirely mature, and it seems to be still susceptible to significant technical and conceptual enhancements. Some key examples of issues still offering a large number of design options and room for further improvements are as follows: (i) The instantiation and selection of NFs for different slices in the infrastructure (ii) The specific functional customisation of NFs to address requirements of specific use cases (iii) The functional interaction among different network slices On this basis, the 5GC architecture along with the envisioned extensions and impacted NFs is shown in Figure 4. These modifications are listed in the following.
(i) To enable inter-slice context-aware optimisation: enhancements of NWDAF to collect and provide per slice/cross-slice feedback information to the NFs and to collect information from M&O layer and maybe also provide feedback to M&O layer per slice/crossslice and enhancements of NWDAF, NFs, and M&O layer to coordinate the execution of changes in the 5G system based on the feedback provided by NWDAF in case of the CP / M&O layer joint optimisation cases.
(ii) To improve slice selection and control: enhancement of NWDAF and/or NSSF to collect/process terminaldriven analytics.
(iii) To enable inter-slice coordination: introduction of inter-slice coordination function (ISCF) within AMF, which provides per service traffic flow binding and distributes the service traffic flow binding information to other NFs.
(iv) Enhancement of PCF to treat per service correlated QoS profiles considering above-mentioned inter-slice coordination.
Radio Access Network.
From the specification perspective, 3GPP Release 15 for next generation-RAN (NG-RAN) is frozen by the time of the submission of this paper (a so-called late drop of Release 15, which includes further architecture options, is planned to be frozen by the end of 2018). This specification comprises slicing awareness in RAN via NSSAI including one or more S-NSSAIs, which allow for uniquely identifying a network slice [9]. While the fundamental slicing support is achieved by Release 15, e.g., granularity of slice awareness and network slice selection, various enhancements and optimisation can be considered for future releases. Such enhancements may imply, for example, specificationrelevant signalling changes and implementation-dependent algorithms, e.g., related to resource management (RM) between slices. Accordingly, the foreseen enhancements are mapped to the specified baseline Release 15 architecture, as shown in Figure 5. It is worth noting that the proposed controller layer is covered in Section 4.3.
In principle, network slicing offers additional degree of flexibility, where NFs can be tailored according to the requirements of slice tenants. To this end, it can be expected that different tenants can have vastly changing needs which can be categorised into three levels, as illustrated in Figure 6 [15]. On one end, some of the slice tenants may only require a performance differentiation, e.g., in terms of QoS requirements, such as latency and data rate, which can be extended by further SLA requirements, such as number of connections for a given time and location. On the other end, slice tenants can require a self-operation of the network services (NSs), such as NF deployment, monitoring, and fault management with dedicated network deployment. In addition, differentiation can be partially on a functional level, where customised NFs can be introduced by the slice tenants, such as customised security and isolation.
Accordingly, slice tenant requirements can be supported by different network slicing realisation variants as depicted in Figure 7 be performed with QoS enforcement. In particular, in line with the latest 5G specification, for an NSI one or more Protocol Data Unit (PDU) sessions can be established, where a PDU session belongs to one and only one specific NSI [10]. Further, RAN maps packets belonging to different PDU sessions to different data radio bearers (DRBs), where within a PDU session there can be one or more QoS flows [9]. On this basis, the RAN treatment of different network slices can be in terms of RRM schemes performed based on the QoS profiles of QoS flows mapped onto the respective DRBs, where QoS profiles can include performance characteristics, e.g., packet delay budget and packet error rate, and allocation and retention priority (ARP). (v) The last two variants do not only share the RAN among the various slices but also the transport network (TN) in L4 and both the TN and CN in L5.
The choice towards the slicing realisation variants described above (involving the design of the slice-tailored NFs at different levels) depends on the needs of the slice tenants and how these needs can be realised on the RAN side. Thus, it is expected that different realisation variants or combinations thereof (e.g., partly shared core NFs and partly slice-specific core NFs) can coexist. Yet, it seems that the variants L0, L1, and L4 may be realised first in 5G deployments. In case high isolation is required (variants L0 and L1 in Figure 7), all RAN protocol stack functions can be tailored according to the slice requirements. In such cases, for instance, each slice can run its tailored dynamic scheduler as an intra-slice control function. In realisation variants, e.g., L4 in Figure 7, where the whole RAN protocol stack is shared by different network slices, the control functions are of the cross-slice form. Under the light of the above discussion, part of the cross-slice control functions, e.g., slice-aware RRM, can be implemented as intra-slice control functions when a high-isolation realisation variant is considered. It is worth reemphasising that even though the whole RAN protocol stack is shared by different network slices, slice-specific performance requirements can be fulfilled with appropriate QoS enforcement, as discussed under the L4 variant above.
Controller Layer.
Following the network softwarisation trends, due to the extensive application of the softwaredefined networking concept to the entire NFs in a mobile network, the control and data layer will necessarily be split by the controllers. Controllers split the functionality between the application logic (i.e., the intelligence that runs in the applications) and the agents running in the NFs. Therefore, our proposed control and data layer architecture makes use of the controller. Control plane applications are the centralised controller layer that comprises an ecosystem of applications controlling the underlying NFs (dedicated or shared), exploiting the advantages of the SDN approach.
The separation of control and execution parts of a network function implies that both parts are connected through an appropriate interface that is able to carry (i) commands from the control part to the execution part, (ii) acknowledgments to these commands back from the execution part to the control part, (iii) indications, measurements, and status reports from the execution part to the control part.
This kind of interface is shown as Southbound Interfaces (SoBIs) in Figures 3 and 5.
As the network function specifics change substantially from one another, it is near at hand that their SoBIs will require substantially different capabilities. Alternatively, all these interfaces could be bundled in a single southbound interface for the ISC/XSC. However, this SoBIs might become very feature-rich and complex.
Some network functions exist for which possibilities to split control and execution parts have been discussed already in literature and for which suitable interfaces have been described, such as the separation between RRC from lower layer RAN protocols through the NBAP protocol, OpenFlow powered mobility management protocols, and, in general, the management of the transport network.
Having a unified controller point for the full network (at least from the logical perspective) will be a fundamental building block for 5G Networks; otherwise, the management and orchestration of a diverse and extensive ecosystem of network slices, such as the one currently envisioned, will be too complex to manage.
Since SDN controllers initially developed in the scope of transport networks, the implementation of such a solution in mobile networks (e.g., RAN) requires further optimisation of their functionality and performance using modern software engineering methodologies. There are some studies focusing on the scalability of wired SDN [42,43], but it is difficult in the wireless scenarios. As the size of a network enlarges, more packets are sent to the controller. There is no doubt that the controller can hardly handle all these incoming requests. Simply improving the performance of a sole centralised controller, without altering the design from the architectural perspective, cannot scale well with the wideranging dense network dimensions. Therefore, controllers for mobile networks might consist of multiple controllers physically distributed in the system. These controllers will not conflict with the "logically centralised" principle as they communicate and cooperate with each other efficiently. This is one of the important considerations in our proposed architecture to have a controller framework that is distributed, scalable, and resilient to satisfy the stringent latency requirements of mobile networks.
Management and Orchestration
Layer. E2E management and orchestration is performed at different levels in a coordinated manner. These levels are service, network configuration, virtualisation, and transport. Our proposed management and orchestration layer takes care of this job, interworking with control layer and network layer, to deploy the required network functions and to configure the appropriate interconnections according to the service and network requirements.
The deployment and management of a network slice is performed by the management and orchestration layer to fill the request of a customer for a communication service. Two models are foreseen in this framework: network slices as network operator internals and network slice as a Service (NSaaS). In the first case, network slices are used by the network operator to fulfill the request to provide a communication service and are optimised and operated accordingly. In the second case, the network slice can be offered by a network operator to a communication service provider with a set of management functions. The communication service provider uses and manages the network slice building up his communication service on top of it.
Management and orchestration layer has to be aligned with 3GPP specifications that foresee a management system that coordinates network and slice management and orchestration. Our current architecture explicitly takes into account the interaction with the 3GPP Management Entities dedicated to network management and configuration (see Figure 8) and enhances the 3GPP 5G Management System manages NSI using two new functions: (i) The NSMF is responsible for management and orchestration of NSI and derives network slice subnet related requirements from network slice related requirements. It communicates with the Network Slice Subnet Management Function (NSSMF) and Communication Service Management Function.
(ii) The NSSMF is responsible for management and orchestration of NSSI. It communicates with the NSMF.
Lifecycle of a Network Slice Instance
De-activation Figure 9: Management aspects of network slice instance [33].
The Network Slice Management and Orchestration is also based on ETSI NFV MANO orchestration framework [44], which introduces the concept of network service (NS) as a set of NFs connected according to one or more forwarding graphs. This framework uses the Network Service Descriptor (NSD) as the templates to manage the lifecycle of a NS and to describe the requirements of VNF, respectively [45]. In addition, in ETSI NFV MANO, the VIM is responsible for control and management of NFVI compute, storage, and network resources; the VNF Manager uses the VNF Descriptor in charge of the LCM of VNF instances, and the NFVO is responsible for the orchestration of NFVI resources and lifecycle management of NS.
In the E2E service management and orchestration layer, the requirements of a new service are translated into network requirements by the CSMF. The obtained network requirements are forwarded to the NSMF, which addresses the management of each slice (Cross-Domain M&O) and the orchestration of different slices (Cross-slice M&O).
To fill the identified gaps in the current architecture (see Section 2), the management and orchestration layer has to identify the VNFs/PNFs to satisfy the service requirements and their relationships (i.e., the forwarding graph of the service); define the configuration and policies (e.g., for elasticity) to fulfill the required service and SLA and then setup the most appropriate network slice template (for network management) and Network Service Descriptor (for VNF deployment); and identify the network KPI associated with the requested SLA and then finalise the deployment and activation of the Network Slice Instance associated with the required service. Therefore, the management and orchestration layer activates the Performance Management (PM) and Fault Management (FM) functions, which continuously monitors the system KPI and triggers necessary orchestration function to fulfill service changes requests or to meet the SLA. Finally, the management and orchestration layer is in charge of exposing PM and FM data to the customer (if requested).
According to [46] the lifecycle of a network slice is comprised of the four following phases as also shown in Figure 9: (i) preparation; (ii) commissioning; (iii) operation; (iv) decommissioning.
The preparation phase includes the design of the network slice template, the on-boarding, the evaluation of the network slice requirements, the preparing the network environment, and other necessary function necessary before the creation of an NSI. From an NFV perspective the role of NFVO in the preparation phase is to ensure the resource requirements for a Network Slice Template (NST). NFVO contains the NSDs that have been previously on-boarded and that can be used to create new NSTs that are created and verified in the preparation phase. The NSDs can be updated and created from the beginning if required if a new NST is necessary.
The NSI is created during the the commissioning phase involving the allocation and configuration of network slice resources required to satisfy the network slice requirements and then the creation of the NSI. During the commissioning phase the NFV MANO functions are only involved in the network slice configuration if parameters related to virtualisation are required for any VNF instance and can be called in the network slice activation step. During the activation the NSMF or the NSSMF functions can activate VNF by means of Update NS sent towards NFVO. This operation could include adding, removing, or modifying VNF instances in the NS instance.
The operation phase comprises the activation, supervision, performance reporting, modification, and deactivation of an NSI. The activation makes the NSI active and ready to support communication services. The modification can include the creation or modification of NSI constituents of characteristics (e.g., topology). The NSI modification can be triggered by new requirements or as a result of supervision/reporting function. The deactivation makes the NSI inactive and stops providing communication services. During the operation phase NFV MANO is responsible for performance management, fault management that could affect a VNF's functioning, and lifecycle of virtualised resources. This could include for example scaling of NS.
Finally, in the decommissioning phase the non-shared constituents are depleted and the NSI specific configuration are removed from the shared constituents. After the decommissioning phase, the NSI is terminated. Table 2 summarises how captured enabling innovations (as covered by Section 3) address the identified gaps (in Section 2) and what are the architectural impacts (as covered in Section 4) per innovation concept.
Evaluation Studies and Analyses on Selected Key Innovation Concepts
In this section, we present some key findings on innovation concepts enabling our proposed flexible, adaptable and programmable network architecture, reflecting how some identified gaps are addressed.
Telco Cloud Resilience.
High resilience of the 5G mobile network (cf. Gap #11) is tightly coupled with high resilience of the telco cloud as one of its integral components. One approach for improving the resilience is applying redundancy in telco cloud deployment. Increased redundancy allows shorter failure recovery time and thus improves overall network availability. However, the increased redundancy also comes with increased costs and operational complexity. In general, a number N of components is backed up with a certain number M of additional components, forming the N+M redundancy approach. There are different modes in which N components are interacting with M redundant components [47,48]. Active-standby scheme assumes that one telco cloud instance (e.g., VM, container) is processing the load, i.e., being active instance, whereas an additional instance is prepared to take over the processing from active instance, once it fails. The procedure of taking over the processing load may incur a considerable delay. Such delay depends mainly on the level of readiness of the standby instance to take over the processing load Load sharing scheme allows sharing of processing load among instances. This scheme follows the N+M redundancy approach, where only N instances would be needed to handle the peak processing load, but additional M instances are used in parallel; thus the processing load is distributed among N+M instances. This redundancy scheme is mainly suitable for processing the tasks without major interdependencies.
Full redundancy (active-active mode) is an approach where the active and redundant instance simultaneously perform the same processing task, where the final results of processing are compared, and faults may be identified. This approach is suitable mainly for cases with extreme availability requirements.
The resulting availability of the telco cloud depends on the availability of the single instance, type of redundancy scheme, and the level of redundancy applied. In order to select the most appropriate scheme for particular context in which the telco cloud is applied, it is beneficial to take the following inputs into account, among them (a) information regarding the required availability level of the telco cloud, given the required end-to-end availability of the service/slice, e.g., 99.99 % or 99.999 %, (b) type of network functions (NFs) deployed on the telco cloud with respect to processing state, i.e., statefull or stateless NF, and (c) dependencies among NFs and their processing tasks for validating the applicability of parallel processing.
Due to its cost efficiency the load sharing redundancy scheme usually is the most reasonable approach for the cases where the processing tasks can be executed largely in parallel. Furthermore, the resulting availability that can be achieved by load sharing scheme depends on the current load in the network. For example, in an N+M load sharing scheme, in the case of lower load, i.e., only N-P instances are needed to handle the total load, the resulting redundancy scheme would be (N-P, M+P) which significantly increases the overall availability of the network. Figure 10, which is based on [48], illustrates the results of analytical analysis performed for the load sharing redundancy scheme where the following inputs are used: N=5, M=2, P=2, for different assumptions on availability of a single component. Additionally, Figure 10 shows the comparison between the load sharing approach and generic N+M redundancy scheme without load sharing. Such generic redundancy scheme provides better results in terms of overall availability, at the cost of more resource usage and no flexibility with respect to the traffic load.
Inter-Slice Control and Management.
Here, we introduce the notion of network slicing in 5G TDD networks, considering a multi-service environment with asymmetric traffic conditions. Network slices are formed on-demand with the allocated resources being dynamically adjusted with the objective to enhance the resource utilisation efficiency. Each network slice is customised to accommodate distinct service types by allowing each tenant to adopt a different TDD frame enabling a distinct UL/DL ratio, which can be reconfigured independently reducing the loss of multiplexing gain. One such TDD oriented network slicing framework is analysed in [49] considering an SDN-based architecture that enables multi-service and multi-tenancy support. However, the allocated slices have a fixed resource size for the entire duration of the service request, occupying only specific isolated subcarriers. This study builds on top of this slicing framework considering more dynamic slice allocations for dynamic radio topologies, addressing identified Gap #6 (E2E cross-slice optimisation), where slice resources can be adjusted during the time of a session request, introducing the following planned contributions.
A graph-based solutions framework is adopted as in [50] to optimise slice performance while keeping the signalling overhead and complexity. Initially, we provide a constraintbased greedy algorithm, whereas we solve the second subproblem by a novel bipartite graph-colouring based solution, which aims to perform adaptive frequency partitioning per time slots in a way that interference due to resource conflicts is avoided and at the same time resource utilisation efficiency remains in high level. Initially, a bipartite graph is translated to a line colouring graph, where each node is a combination of link and transmission time interval (TTI) (edge of the bipartite graph). The edge between two nodes in the line colouring graph appears only if a conflict exists at the receiving end of the bipartite graph, which is equivalent of having two or more links being assigned to the same TTI. The graph-colouring algorithm assigns a different colour to a node only in case of a conflict, which means that different subbands will be scheduled to avoid interference. Based on this algorithm, the output is a timetable where each link is assigned to different subbands, within distinct TTIs to ensure interference-free transmission/reception. Monte Carlo system level simulations are provided for a 5G Ultra Dense Networks (UDNs) where resources can be shared by multiple slices with diverse KPIs (example for throughput, reliability). In this evaluation study we focus on an outdoor small cell deployment of four small cell Access Points (s-APs) covering a hotspot area, using the 3GPP as baseline for our simulations (24 users uniformly distributed, 3GPP UMi channel, ideal backhaul). In each s-AP the corresponding users (6 users per cell) are randomly distributed. We run Matlab Monte Carlo simulations and random user drops for 500 snapshots. We assume four slices, whereas each slice has different TDD pattern as slice requirement (Slice 1: 80/20, Slice 2: 70/30, Slice 3: 60 /40, and Slice 4: 50/50). At each snapshot, we randomly select 6 users out of 4 cells to be connected to each slice and we also have a random traffic demand (1-10Mbps per user for both UL and DL). More details on the simulation setup and further results can be found in [50].
For our simulation comparison we consider as benchmark the service-oriented TDD slicing [49], where slices are assigned a constant amount of resources (25% of resource blocks in our case) and different TDD patterns are used independently for each slice. This solution provides a high spectral efficiency due to the interference isolation where the proposed adaptive allocation of resources in different slices outperforms the benchmark. As shown in Figure 11, significant gains are observed in both median and 90% of the CDF curves, representing the average and cell-edge UEs accordingly, due to adaptive allocation of resource blocks to users of different slices in a way that intercell inter-slice interference is avoided.
Experiment-Driven Optimisation.
To take advantage of the experiment-driven modelling and optimisation in a cloud-enabled network, new challenges arise (cf. Gap #7).
A key requirement is the conduction of exhaustive measurement campaigns per VNF and per network slice that will focus on consumption of computational, storage, and networking resources and considering cost-effectiveness and the special characteristics and peculiarities due to the use of commodity hardware (a key choice for the cloud-enabled networking).
To provide an example study towards this direction, we focus on RAN functionality and more specifically on functionality carried by protocols above the MAC layer at the gNB and UE side, more precisely, the PDCP and RLC protocols. Key target is the quantification of the computational and memory resources (CPU/RAM load) that are consumed by the higher layers of RAN protocol stack as well as investigating the impact that a function split at the RLC level can provide in terms of delay to a provided service.
In this context, the PDCP/RLC functionality was implemented in a stateless way using Python on top of the following SW/HW platform: Measurements were extracted in two different scenarios. First, with 4K video streaming, to assess the CPU and RAM consumption while a demanding application is running, and second with increasing traffic using Iperf to depict the relation of load and CPU consumption. The results are depicted in Figures 12 and 13, respectively. Some takeaways from this measurement campaign and observed findings in the above setup are listed below: (i) The type of implementation affects the performance, meaning that optimisation of the code can provide elasticity to the functions. For instance, in the current implementation the allocated CPU supports up to 65Mbps (after that the CPU is fully loaded and unstable).
(ii) The split of higher layers from the MAC/PHY adds CPU load due to the interfacing between the two network nodes. It was noticed that, in the current setup, the load for interfacing takes approx. 90-95% of the CPU load and adds a similar percentage to the total latency.
(iii) The CPU load can be used as a trigger for applying resource elasticity. However, it is not an indication for the service performance. Practically, the impact of overloading the CPU on the service performance, as revealed from the tests, is not visible to the application layer till the point that the packets cannot be served. This is due to the notion of the investigated functions, which perform a type of "forwarding", compared to more sophisticated network functions (e.g., the decoding in PHY).
Conclusions
In this paper, we followed the baseline requirements and related KPIs of 5G services and identified gaps from the literature to propose a flexible, adaptable, and programmable architecture for 5G mobile networks. Our main contribution here is to design an architecture that (i) takes the current state-of-the-art on 5G architectures, from previous projects as well as standard efforts, (ii) addresses the gaps identified within those architectures, and (iii) provides a complete architecture design, comprising state-of-the-art modules and the new ones that address the target KPIs.
The proposed architecture consists of four different layers identified as network layer, controller layer, M&O layer, and service layer. A key contribution of this paper is the definition of the role of each layer, the relationship between layers, and the identification of the required internal modules within each of the layers. Within the proposed architecture, multiple management domains for E2E network slice deployment and operation have been explored from both 3GPP and ETSI perspectives. In particular, the proposed architecture extends the reference architectures proposed by 3GPP and ETSI by building on these architectures while addressing several gaps identified within the corresponding baseline models.
Besides the overall architecture design, another key contribution of this paper is the design of some of the key modules within the architecture. To this end, we have identified enabling innovations and network functionality to address several gaps identified from flexible cloudification of protocol stack to adaptive network slicing via inter-slice control and management via leveraging experiment-and implementation-driven modelling and optimisation. Furthermore, we presented evaluation studies and analyses on a selected set of key identified innovation concepts, indicating novel approaches for the design of these modules.
The ultimate goal of the proposed architecture is to allow for the instantiation of slices that can satisfy specific requirements. The proposed architecture accommodates potential NFs and solutions to achieve slice resiliency, security, and elasticity. These functions can be thus instantiated by our architecture when deploying slices that need to provide the corresponding services.
Data Availability
Part of the findings of this paper are readily available in [14] 5G-MoNArch Deliverable D2.2, "Initial overall architecture and concepts for enabling innovations," June 2018.
|
2019-02-17T04:46:11.646Z
|
2019-02-11T00:00:00.000
|
{
"year": 2019,
"sha1": "042ccb5db6595fc248449e43d7a559a357846feb",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.1155/2019/5264012",
"oa_status": "GOLD",
"pdf_src": "MergedPDFExtraction",
"pdf_hash": "042ccb5db6595fc248449e43d7a559a357846feb",
"s2fieldsofstudy": [
"Computer Science",
"Engineering"
],
"extfieldsofstudy": [
"Computer Science"
]
}
|
232173225
|
pes2o/s2orc
|
v3-fos-license
|
Serum NGAL, BNP, PTH, and albumin do not improve glomerular filtration rate estimating formulas in children
Glomerular filtration rate (GFR) is difficult to measure, and estimating formulas are notorious for lacking precision. This study aims to assess if the inclusion of additional biomarkers improves the performance of eGFR formulas. A hundred and sixteen children with renal diseases were enrolled. Data for age, weight, height, inulin clearance (iGFR), serum creatinine, cystatin C, neutrophil gelatinase-associated lipocalin (NGAL), parathyroid hormone (PTH), albumin, and brain natriuretic peptide (BNP) were collected. These variables were added to the revised and combined (serum creatinine and cystatin C) Schwartz formulas, and the quadratic and combined quadratic formulas. We calculated the adjusted r-square (r2) in relation to iGFR and tested the improvement in variance explained by means of the likelihood ratio test. The combined Schwartz and the combined quadratic formulas yielded best results with an r2 of 0.676 and 0.730, respectively. The addition of BNP and PTH to the combined Schwartz and quadratic formulas improved the variance slightly. NGAL and albumin failed to improve the prediction of GFR further. These study results also confirm that the addition of cystatin C improves the performance of estimating GFR formulas, in particular the Schwartz formula. Conclusion: The addition of serum NGAL, BNP, PTH, and albumin to the combined Schwartz and quadratic formulas for estimating GFR did not improve GFR prediction in our population. What is Known: • Estimating glomerular filtration rate (GFR) formulas include serum creatinine and/or cystatin C but lack precision when compared to measured GFR. • The serum concentrations of some biological parameters such as neutrophil gelatinase-associated lipocalin (NGAL), parathyroid hormone (PTH), albumin, and brain natriuretic peptide (BNP) vary with the level of renal function. What is New: • The addition of BNP and PTH to the combined quadratic formula improved its performance only slightly. NGAL and albumin failed to improve the prediction of GFR further.
Introduction
Estimation of glomerular filtration rate (eGFR) is important in clinical practice [1]. Several formulas have been developed for this purpose; however, all these formulas lack precision, and much effort is actually made in order to improve eGFR prediction. Serum creatinine (Scr) is the most commonly used endogenous marker to estimate GFR but with relative imprecision owing to variation of non-GFR determinants of Scr, e.g., muscle mass, high meat-containing diet, or tubular secretion. Moreover, an increase in Scr may not be observed until a substantial decrease of at least 40% in GFR has occurred [2]. Recently, serum cystatin C (Scyst C) has emerged as an alternative to or a complement of creatinine measurement in the evaluation of kidney function. Several studies have shown a better estimation of GFR using combined Scr and Scyst C measurements [3,4]. However, Scyst C can be influenced by uncontrolled thyroid disease or immunosuppressive therapies [5,6], also leading to a decrease in eGFR precision. Until now, measurement of inulin clearance represents the gold standard method to assess GFR. Other exogenous markers (i.e., Cr-51 EDTA, iohexol) have been used as well; however, all these methods can be cumbersome and are not always available in clinical practice [7].
To overcome the above-mentioned limitations, many serum biomarkers have been incorporated in mathematical models in order to improve bedside formula for GFR estimation [8,9]. The Kidney Disease Outcomes Quality Initiative (K/DOQI) guidelines recommend estimating GFR in adults and children using Scr and/or Scyst C based predictive equations [1]. The most widely used formula is the Schwartz formula, which was developed in 1976 and revised in 2009 [10][11][12]. We have also developed a Scr and a combined (Scr and Scyst C) quadratic formula in order to better assess GFR in Swiss children [13,14]. However, all these new formulas lack precision when compared to measured GFR. Much effort is invested in finding new biomarkers to improve the accuracy of bedside GFR estimation. This study aimed to evaluate if the addition of serum neutrophil gelatinase-associated lipocalin (NGAL), brain natriuretic peptide (BNP), parathyroid hormone (PTH), and albumin improves the performance of the Schwartz and the quadratic formulas, compared to urinary inulin clearances.
The choice of biomarkers to be included in this study was based on several considerations. NGAL has recently emerged as a predictive biomarker in chronic kidney disease (CKD) [15]. PTH is a recognized biomarker in CKD. Fu et al. [16] showed a significant negative correlation between PTH and eGFR. In addition, Okamoto K et al. [17] demonstrated that changes in PTH levels were associated with kidney function and renal outcome. Regarding albumin, patients with CKD often present with decreased appetite secondary to uremia and consequently may develop malnutrition [9,18]. Finally, BNP has been used in CKD progression and GFR estimation in adult patients [19,20].
To the best of our knowledge, there are no adult nor pediatric studies evaluating the impact of BNP, PTH, NGAL, and albumin in estimating GFR.
Patients
Children with chronic kidney diseases (CKD) who were referred to our clinic for GFR measurement between August 2012 and June 2013 were retrospectively included. Children were aged between 3 and 18 years. Causes of CKD included congenital and acquired single kidney, obstructive or reflux uropathy, polycystic kidney disease, and other various diagnoses (Bartter syndrome, patients with history of hemolytic uremic syndrome, post-chemotherapy). Patients who were unable to void spontaneously were excluded from the study.
Measurements and analytic methods
GFR was measured using the gold standard method-urinary inulin clearance (iGFR). Inulin was measured using an AutoAnalyzer 3 system, as previously reported by our research group [13,14]. At the time of measuring patient's GFR, additional data were collected: height (cm), weight (kg), Scyst C (mg/l), Scr (mg/dl), serum blood urea nitrogen (BUN) in mmol/l, serum NGAL (ng/ml), intact serum PTH (pg/ml), serum BNP (pg/ml), and serum albumin (mg/l). Scyst C was measured using the particle-enhanced nephelometric immunoassay (Siemens Healthcare Diagnostics). Scr was analyzed using the kinetic colorimetric compensated Jaffe method (Roche Diagnostics, cobas 8000) which was calibrated against the enzymatic method and standardized against the reference isotope dilution mass spectrometry method. Serum BUN was measured with the enzymatic assay (Roche Diagnostics, cobas 8000), and PTH was measured with enzyme immunoassay (Immulite 2000 XPi/Siemens), BNP with electro-chemiluminescence immunoassay (Roche Diagnostics, cobas 8000/e801), and serum albumin with bromocresol green (Roche Diagnostics, cobas 8000). The eGFR was calculated with the following equations: revised Schwartz, combined Schwartz, quadratic, and combined quadratic formulas (Table 1).
Statistical analysis
Continuous data are presented as median with interquartile range [IQR] and categorical as percentages. We used a simple linear regression to examine association of continuous variables (age, sex, height, serum cystatin C, creatinine, NGAL, BNP, PTH, and albumin) with inulin clearance. We first obtained the variance in inulin clearance explained by the basic formulas, i.e., the coefficient determination (r-squared). Then we added to the best fitted linear regression each of the additional variables (serum cystatin C, NGAL, BNP, PTH, and serum albumin), one at a time, obtained the adjusted r-square of the models, and tested the improvement in variance explained by means of the likelihood ratio test. Significance was set at p < 0.05.
We attempted to construct a parsimonious multivariate model predicting inulin clearance including all statistically significant predictors.
Results
A total of hundred and sixteen patients were enrolled in this study. Patients' median [IQR] age was 12.6 [8.4-15.5] years, and 52% were boys. Children's demographic characteristics are summarized in Table 2. Ninety-one percent of the patients had CKD stage I and II, and 9% had a CKD stage III to V. All patients' data were well documented and recorded, without any missing data.
Median [IQR] iGFR was 82.5 [71.0-98.8] ml/min/1.73 m 2 . The median values for eGFR using the revised and combined Schwartz formula, the quadratic, and the combined quadratic formula were 89.5, 91.3, 91.5, and 88.0 ml/min/1.73m 2 , respectively. Results confirmed that the best empirical model for inulin clearance (r 2 = 0.75) was the association of sex, inverse creatinine, inverse creatinine squared, height squared, and inverse cystatin C.
The combined quadratic formula performed best, with the highest r 2 value at 0.730 (Table 3) compared to other formulas (r 2 at 0.569, 0.676, and 0.708 for the Schwartz, combined Schwartz, and quadratic formulas, respectively).
The addition of cystatin C to the Schwartz and the quadratic formulas improved r 2 the most (from 0.569 and 0.708 to 0.700 and 0.734, respectively). Adding BNP and PTH also improved the variance of the combined quadratic formula slightly (from 0.708 to 0.738 for both, p = 0.042 and 0.046, respectively) ( Table 3). The addition of NGAL and serum albumin did not increase r 2 (Table 3).
After adding BNP and PTH to the combined Schwartz formula, the r 2 changed from 0.676 to 0.694 and 0.673, respectively. Regarding the addition of NGAL and serum albumin the combined Schwartz formula, we observed a change of r 2 from 0.676 to 0.673 and 0.682, respectively. In neither case were we able to specify a multivariate model that was significantly better than the basic formulas plus cystatin C.
Discussion
In this study, we found that the addition of serum NGAL, BNP, PTH, and albumin to the combined Schwartz and quadratic formulas for estimating GFR did not improve GFR prediction in our population. This study represents a new step in developing better GRF estimation equations with the addition of several renal biomarkers. We attempted to evaluate the usefulness of adding BNP, PTH, NGAL, and albumin in the estimation of GFR. BNP is a peptide hormone whose Ht, height expressed in cm; Scr, serum creatinine expressed in mg/dl; Scys, serum cystatin c expressed in mg/l; BUN, blood urea nitrogen expressed in mg/dl; age expressed in years [20]. We found a positive effect of adding BNP to the combined quadratic formula; however, the effect remains extremely small, albeit significant. Regarding PTH, it is well known that PTH increases with CKD progression. Several studies analyzed the correlation between PTH and GFR. This correlation may depend on the degree of CKD [21][22][23]. Similar to BNP, this study showed that adding PTH led to a slight increase in the performance of the combined Scr and Scyst C formulas. However, the clinical impact is limited, creating a more complicated formula. In a recent study, there was a significant correlation between serum NGAL and measured GFR in children with CKD. The authors concluded that serum NGAL may prove useful in the quantitation of CKD, especially at low levels of measured GFR, where, by correlation analysis, NGAL outperformed cystatin C [15]. In addition, Bolignano et al. [24] demonstrated that NGAL is an independent predictor of CKD progression. Despite the fact that NGAL has been shown to be a good diagnostic and prognostic biomarker of acute and chronic renal failure [15,25,26], the addition of NGAL did not lead to a better prediction of measured GFR. Finally, albumin has been previously used in adult GFR estimation formulas [9]. Levey A et al. showed that for every 10% of albumin change, there was a 3.1% change in GFR [9]. This study is the first one to include albumin in eGFR pediatric formulas. Similar to NGAL, albumin did not increase the performance of combined creatinine and cystatin C GFR estimation formulas. In the same vein, other researchers attempted to develop formulas incorporating new biomarkers such as β2-Microglobulin or β-Trace Proteins, with conflicting results [27,28]. Recently, Inker LA et al. nicely demonstrated that the incorporation of β2-Microglobulin and β-Trace Protein into adult prediction formulas improved their performance, without incorporating the race data [29]. Chen N et al. also showed the beneficial addition of β2-Microglobulin and β-Trace Protein in improving GFR estimation in a Chinese population, however without significant clinical impact [30]. The utility of these two biomarkers in predicting GFR should be investigated further.
Our study results confirm that the addition of Scyst C improves the performance of estimating GFR formulas, in particular the Schwartz formula. Recently, den Bakker et al. demonstrated in a large cohort of 408 patients that Scr and Scyst C were the most powerful predictors of GFR, even in the absence of height data [31]. Deng et al. confirmed a better accuracy and applicability of the multivariate eGFR equations including Scr and Scyst C compared to univariate formulas [32].
The strength of this study is the use of inulin clearance. All variables were measured simultaneously during inulin test, therefore reducing potential variation in their level due to progression of renal failure with time. In addition, to the best of our knowledge, this is the first study investigating the potential role of NGAL, BNP, PTH, and albumin in GFR estimation in a pediatric population.
However, the study results and conclusions are limited by the fact that we have included only Swiss children, and by a small number of patients with iGFR < 60 ml/min/1.73m 2 . These limit the generalizability of the results to non-Caucasian children and to those with moderate to severe CKD.
In conclusion, serum creatinine and cystatin C remain the best predictors for GFR estimation. The addition of other parameters (i.e., BNP and PTH) may be beneficial but should balance the complexity of the new equations and its associated cost.
Abbreviations BNP, Brain natriuretic peptide; BUN, Serum blood urea nitrogen; CKD, Chronic kidney disease; GFR, Glomerular filtration rate; iGFR, Urinary inulin clearance; K/DOQI, Kidney Disease Outcomes Quality Initiative; NGAL, Serum neutrophil gelatinase-associated lipocalin; PTH, Parathyroid hormone; Scr, Serum creatinine; Scyst C, Serum cystatin C Availability of data and material The datasets generated during and/or analyzed during the current study are available from the corresponding author. NGAL neutrophil gelatinase-associated lipocalin, BNP brain natriuretic peptide, PTH parathyroid hormone
Authors' contributions
Funding Open Access funding provided by Université de Lausanne. The study has been supported by the pediatric nephrology laboratory of Lausanne University Hospital.
Declarations
Ethics approval The study has been approved by the local Ethic Committee (Lausanne Ethic Research Committee Protocol number 345/12).
Consent to participate and for publication The study has been approved by the Research Ethic Committee of Lausanne.
Competing interests The authors declare no competing interests.
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/.
|
2021-03-11T14:43:46.768Z
|
2021-03-10T00:00:00.000
|
{
"year": 2021,
"sha1": "17cb83e5588319d7b97810458adfb96242179e23",
"oa_license": "CCBY",
"oa_url": "https://link.springer.com/content/pdf/10.1007/s00431-021-04019-w.pdf",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "17cb83e5588319d7b97810458adfb96242179e23",
"s2fieldsofstudy": [
"Medicine",
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
213906383
|
pes2o/s2orc
|
v3-fos-license
|
Environmental characteristics and management challenges of brackishwater fish ponds in Cirebon Regency, West Java Province, Indonesia: a fine-scale GIS Approach
This study presents a fine-scale Geographic Information System (GIS) method to explore existing environmental and farm management constraints at extensive brackish water aquaculture ponds located in Losari District, Cirebon Regency, and West Java Province. The sampling method followed a detailed survey method, employing high-resolution satellite imagery of worldview-2 and GIS. Detailed pond unit and layout canal maps were used to guide soil and water sampling, measurement of land elevation and tide observation. The result showed that the existing pond canals were not able to supply 4,800,000 m3 daily required quantity of culture water for the total 2,360 ha pond area due to high sedimentation at the mouth and along the canals. The result of spatial analysis of soil and water variables also indicated that, despite generally good soil and water quality, some existing environmental variables could potentially limit the productivity and sustainability of the existing brackishwater ponds. The soil variables such as phosphate (PO4-P) content (79.16 ± 28.34 mg / L) exceeded the optimum standard value of 35-46 mg/L, whilst total nitrogen (Total-N) content (0.07 ± 0.025 %) was lower than the optimum value of > 0.2%, for brackishwater aquaculture. Water quality variables comprising (NO2-N), total ammonia nitrogen (NH3-N) and nitrate (NO3-N) were also identified as limiting factors. Without a significant improvement in pond engineering and better understanding and management of existing environmental limiting factors, the increase in productivity and the sustainability of brackishwater fish ponds in the region seem almost impossible.
Introduction
The potential brackishwater pond areas in Indonesia are estimated at about 2.9 million ha with the total utilization of 715,846 ha (24.14%) in 2015. Of the total utilization area, 60% is under the category of traditional ponds with low productivity. Traditional pond relies mainly on natural food propagated in the pond with or without fertilization and from water supply from tidal sources. Many of these traditional ponds have been left unproductive or abandoned mainly due to diseases or environmental degradation. To maintain the fish or shrimp production, since 2012, the Ministry of Marine Affairs and Fisheries (MMAF) of Indonesia has implemented a national program called "pond revitalization program", focusing on planning and rehabilitation of pond infrastructure (e.g. dykes and canals) and general layout of existing ponds [1]. It should be noted that in every effort to increase land-based aquaculture production, it will involve an expansion of cultivated areas, increased density of aquaculture installations and increased use of feed, fertilizer and chemical inputs, including increased land and water use [2]. Inadequate aquaculture infrastructure such as pond canals will directly or indirectly lead to a decline in water quality which in turn triggers the emergence of fish or shrimp diseases [1]. Thus, land development, rehabilitation, and pond revitalization activities must pay attention to environmental characteristics and specific problems which might be the limiting factors for each location.
Most issues affecting the development and management of aquaculture, including brackishwater aquaculture are spatially referenced problems or at least have spatial components [3]. Geographic Information Systems (GIS) has various advantages compared to other conventional methods to handle processes and solve more complex spatially referenced problems in aquaculture. GIS, for examples, can provide a visual inventory of the physical, biological, and economic characteristics of the environment and also enable the generation of suitability maps for different uses [4,5]. [6] and [7] demonstrated the great advantage of GIS for brackish water aquaculture site selection by identifying land use/land cover patterns combined with 6 main group variables namely, engineering variables, water quality variables, soil quality variables, pond infrastructure facility, meteorological variables, and social restrictions. To apply spatial analysis in aquaculture [3] recommended three different spatial scales, comprising small farm/cluster scale; waterbody /watershed zone; global market-trade scale. A semi-detailed land suitability map can be assumed to be equivalent to the scale of watershed zones and is applicable for general aquaculture evaluation and planning purposes at provincial or regency level, whilst a detailed land suitability map of the small farm or cluster scale can be applied to help various stages in pond management, evaluation of environmental impacts and calculation of land carrying capacity [3,8,9,10].
The pond revitalization program requires a robust method as well as supporting detailed spatial dataset to evaluate the suitability of local farm management policies with consideration of land characteristics. Understanding engineering suitability of canal design and networks, pond units' layout, coastal hydrology, soil, and water quality and also aquaculture production history are of important information in the implementation of the pond revitalization program. Therefore, the farmscale approach is considered very relevant and can help the pond revitalization program succeed. This study presents a detailed scale approach of evaluating environmental characteristics and pond management challenges in Cirebon Regency which is one of the main shrimps producing area and targeted locations for pond revitalization program in Indonesia.
The study site
The study was conducted on existing brackishwater ponds in Losari District, Cirebon Regency, West Java Province of Indonesia, which lies between 6 o 44' 6" and 6 o 50' 7.7" south latitude, and longitudes 108 o 43' 34.8" and 108 o 50' 29.1" east longitude ( Figure 1). The topography in Losari District is generally low-lying coastal urban area and is facing directly the Java Sea. Land with elevation ≤ 25 m accounts for more than 65% of the district area [11]. Based on DKP-Cirebon report in 2012, brackishwater aquaculture ponds locates in seven coastal districts, of which the largest potential area is in Losari Subdistrict. However, despite Losari's largest potential brackish water pond area, the actual land usage for shrimp or fish farming is relatively low covering only about 1,360 of the total 2,500 ha potential areas. The existing brackishwater ponds have been utilized mainly for low-density milkfish (Chanos chanos) and tiger shrimp (Penaeus monodon) culture. Since 1993 recurrent crop losses due to white-spot disease and unfavorable environmental condition have limited production of P. monodon in traditional ponds throughout Indonesia. Because of this, many aquaculture ponds have fallen into disuse, or their use is limited to only milkfish production [12].
Soil and water sampling and laboratory analysis
A total of 63 for each soil and water samples was taken following a modified transect method, by which additional point samples in-between the transect lines were collected to improve distribution and density of the sampling points and to capture landscape feature of the study area. The soil samples were collected from mostly fallowed pond bottoms (depths 0.1 to 0.4 m and 0.4 to 0.6 m) using a Jarret bucket auger. These soil samples were transported to the Soil and Water Laboratory of the Research Institute for Coastal Aquaculture (RICA), in Maros District of South Sulawesi Province, for further analysis. Field and laboratory soil analysis was undertaken for 5 chemical and 1 physical variable. The soil physical variable was soil textural classes, whilst soil chemical analysis included pHF, pHFOX, total nitrogen (Total-N), Total and available phosphorus (P-PO4) and Total Organic Matter (TOM) (%). Field pH (pHF) and field pH after oxidation with 30% analytical grade hydrogen peroxide (pHFOX) were directly measured on a 'soil: water' paste [13] in the field using pH meter (Hanna-HI 8424). N TOT was determined according to Kejdahl method using Keltex apparatus [14], P-PO4 analysis followed Olsen dan Bray 1 method: Olsen is applied for soil with pH higher than 5.5, whilst Bray 1 if pH lower than 5.5 [15,16]. TOM was measured using Walkley-Black's wet oxidation method using 0.5 mm soil fraction [17,18]. The baseline water quality variables that were measured covered: temperature, salinity, pH, NO2, NO3, NH3, PO4, and Fe. This suite of variables is effective as a set of indicators for the changes in water quality. Salinity and pH were measured in situ using an YSI Pro Plus. Nitrite (NO2-N), Nitrate (NO3-N), Ammonia (NH3-N), Phosphate (PO4) and TOM were analyzed in the Water Laboratory of RICA following the procedure of [19,20]. Detailed mapping of pond bed elevation was carried out using Nikon ® Digital Theodolite NE-102 on 55 main control points and 212 survey points which were distributed throughout study areas. At the same time as pond bed elevation measurement, a 15-day tidal observation was conducted using an automatic tide gauge of Keller Water Level Logger and calculated using least square method to predict average high tidal surface [21]. Seven primary tidal constituents including the principal lunar semidiurnal components (M2), Principal solar semi-diurnal components (S2), lunar elliptic semi-diurnal constituents (N2), lunar declination semi-diurnal constituents (K2), Luni-solar diurnal component (K1), Principal lunar diurnal component (O1) and solar declination diurnal constituents (P1) were obtained from this analysis. The output of the tidal analysis of local mean sea level was used as a reference datum for elevation data to generate a detailed mapping of the pond bottom. Key soil and water variables considered as key limiting factors for the production of brackishwater ponds in the study area were selected and spatially analyzed using ordinary kriging (OK) method [22,23]. The output of kriging maps of key soil and water variables were used to explain spatial variability and context of the environmental and management challenges in the study area. All statistical analysis was performed by SPSS software version 17. The descriptive statistics analysis (e.g., mean and standard deviation) was first done to gain insight into the distribution values of soil data. Spatial analysis that includes vectorization, spatial interpolation, and combination of thematic maps was analyzed using ArcGIS 10.3, Map Info 9.0 and SURFER 8.0 software.
Layout, topography, and hydrology of brackishwater ponds area
The map of pond layout which was manually extracted and vectorized from the WorldView-2 imagery was presented in Figure 2. As shown in the detailed map, there was an estimate of 3000 existing pond units covering the total area of 2.360 ha and stretching on 20.7 km coastline of the survey area in Losari District. Fish or shrimp pond units generally were range in size from 0.2 to 3 ha (average 0.8 ha). The existing ponds were generally converted from mangrove and rice fields, which were located 1 to 10 meters from the primary or secondary canals. The analysis also indicated that pond units received seawater from 20 main canals during high tide, and used the same canals to discharge effluent at low tide. Canal length varied from 400 to 3000 m with an average width of 5 m. Water used for fish culture come from sea canals, brackish water rivers, or a mixture of the two which were then spread to pond units through secondary or tertiary canals. Mangroves (mainly Rizophora sp) can now only be observed in several spots along primary canals or edge of the pond units nearby coastline. Based on the analysis of local tide data, it was known that the tidal type for the study region was a mixed-type with semi-diurnal dominance (formzhal numbers/F = 0.95). With this mixed tidal cycle, water exchange in pond culture can only be effectively done twice daily, with the average 5 hours water intake period in every high tide. With amplitude values of major harmonic constituents (M2=3; S2=42, N2=3; K2=47; K1=39; O1=4 and P1=26), a relatively small (≤1 m) were very likely to occur in the study region even at average high tides.
The detailed topographic maps of sea bed and pond areas were shown in Figure 3. The general topography of the sea bed was gently sloping (slope <2%) within a distance of 300 m from the coastline, but subsequently changes to > 10% seaward. This indicated that a fairly strong wave often occurred in this region. Sediment from bank or pond wall erosion and sludge was removed during harvest and pond preparation, transported to the rivers or pond canals when high waves will be pushed and deposited along the beach and around the mouth of rivers or canals. The deposited sediments form arising coastal land overgrown with mangroves. Unfortunately, most of the newly growing land has been converted into fish ponds, by which creating more problems in management.
Water availability or good water supply is one of key environmental factors affecting the productivity of small-scale pond-based aquaculture. Water in the pond must be kept at certain level and the quality must be kept high to ensure optimal growth of cultured organisms. If it is assumed that the average depth of pond canals is 1 m, the total surface area of the (primary, secondary and tertiary) canal is 2,400,000 m 2 , and the tidal type of semi-diurnal (tidal range = 0.9 m); then daily water available is approximately 4,800,000 m 3 . The available volume of water is expected to supply 3000 pond units having a total area of 2360 ha (23,600,000 m2) with water requirements 8,850,000 m 3 (50% of the total needs of the pond units with an average water depth of 0.75 m). If existing canals can effectively function as expected, the available volume of water could technically meet the water requirements for fish pond culture. According to [24] water availability is considered sufficient if it can be obtained within a period of 5 high tides. The result of calculations seemed to exceed the real capacity of existing canals in providing water because with an average maximum flow rate of 0.5 m / sec, high tide water could only reach a maximum distance of ± 1 km (high tide time period = 5 hours). The shape of curved and shallow canals also inhibits the flow of water into fish ponds. Additionally, primary canals required times to distribute available water into secondary and tertiary canals. Some of the ponds in this study area also obtained water from rivers and freshwater canals of which the seawater supply was only possible with the help of pumps, especially those that are further away from the coastline. Under these circumstances, maintaining the required volume and optimum quality of water for pond aquaculture in this location seemed very difficult due to the limited number and improper design of primary canals. The accumulated organic matter as results of dead benthic algae, fish feces, and uneaten feed along with hazardous pesticide residues was usually washed and discharged by farmers through primary canals with the hopes that it will be wasted and neutralized when reaching the sea as receiving the body. Unfortunately, the pond effluents very often before reaching the sea were pushed back by high tide and incoming wave and deposited at the canals' mouth. This inefficient effluent treatment could create environmental conditions that are not suitable and even dangerous not only for cultured organisms but also other coastal habitats.
Soil and Water Quality
The result of the physicochemical analysis (pH, salinity, temperature, DO, nitrate, ammonia, nitrate, phosphate, and TOM) was summarized in Table 1. The result of in-situ measurements of water quality showed that salinity, temperature, pH and DO of pond water were still in the tolerable level and safe for brackishwater aquaculture. However, the result of the spatial analysis revealed temperature values ranged from 29.7 o C to 38.0 o C, of which some have exceeded the optimum value for shrimp culture, particularly for tiger shrimp (Penaeus monodon). The spatial distribution of pond water temperature values was described in Figure 4. Lying in the tropical region, Indonesia does not experience large seasonal temperature fluctuations, only daily temperature fluctuations that were observed to cause problems in pond culture. Variations in temperature and high values of water temperature measured in some spots were mainly due to differences in measurement time and the difference in pond water depth. According to [24], water temperature for the growth of tiger shrimp ranges from 26 o C to 32 o C, but the optimum temperature for increased productions is between 29 o C and 30 o C [24]. [25] reported the optimum temperature to support the growth of whiteleg shrimp (Litopenaeus vannamei) ranges from 25 o C to 35 o C. Given this, the kriging map does not only describe the spatial distribution of key variables values but also explains the reason for the spatial variability. 10 26 -60 * [24,6] The result of the laboratory analysis of pond water quality in Table 1 generally indicated that phosphate and BOT values of pond water were still within a suitable range for aquaculture. However, the concentrations of NO2-N, NH3-N, and NO3-N must be monitored because most of it exceeded the acceptable concentration for fish culture. NH3-N concentration obtained in this study ranged from 0.0013 -0.762 mg/L (0.32 ± 0.163), that were mostly higher than the standard seawater quality for the growth of marine biota, which was less than 0.10 mg / L [26]. Ammonia can be in the form of NH3 molecules or NH4 ions, of which NH3 was more toxic than NH4 [24]. NH3 can penetrate into cell membranes faster than NH4 [27]. The NH3 concentration of 0.05 to 0.20 mg/L could inhibit the growth of aquatic organisms in general. If NH3 concentration was more than 0.2 mg/L, it was toxic for some fish species [28]. [29] reported the acceptable concentration of ammonia for tiger shrimp culture was less than 0.1 mg / L. Fish cannot adapt with high concentrations of NH 3 , because it can affect the oxygen binding by blood and eventually leads to suffocation. The high concentration of ammonia and nitrite was also an indication of polluted environmental which was usually attributed to domestic waste, industrial waste and fertilizer run-off from agriculture activities. In addition, the high level of ammonia concentration in pond water was a good description of the anaerobic condition of pond bottom due to high decomposition of organic matter and problems of pond waters exchange. The nitrate concentration obtained in this study ranged from 0.001 mg/L to 0.640 mg/L, with the average concentration was higher than the acceptable seawater standard for marine organisms (0.008 mg/L) [26]. Table 2 summarizes the result of descriptive statistical analysis of soil variables. The result indicated that most soil variables under consideration were generally acceptable for aquaculture standard. The pHF values ranged from 6.16 to 7.98 and were still within the acceptable criteria and will not be a limiting factor for aquatic organisms. According to [6], pond soil with a pHF ranging from 6.5 to 8.5 was good because the limiting factor can be very easily managed. The difference between pHF and pHFOX (pHF-pHFOX) of soil samples which was a good indicator for potential acidity ranging from 0.0 to 1.23 (0.41 ± 0.35) was also still in the suitable category for aquaculture. The greater pH F -pH FOX value indicated the greater the potential for acidity in acid sulfate soils (ASSs). The soil quality variable observed to be a limiting factor for aquaculture was the concentration of phosphate. As can be seen in Table 2, phosphate values of soil samples obtained from the study area ranged from 31.99 -180.88 (79. 16 ± 28.34). Although the result, in general, showed a high level of soil fertility, in some locations the concentrations have exceeded the optimum standard for aquaculture activities. Spatial distribution of PO4 values as shown in Figure 4 indicated that the highest phosphate content values were concentrated at the location around the intensive ponds and in the ponds neighboring rice fields. The high phosphate content around intensive ponds could be attributed to effluents such as water, sediment and uneaten feed released to nearby environments. Excessive application of phosphate fertilizers on agriculture farms will also eventually be drained through canals that interconnected with aquaculture ponds. The results in Figure 4 further showed that the average total nitrogen (Total-N) concentration was lower than the optimum value required for aquaculture, with concentrations of 0.02 -0.18%. Nitrogen (N) plays important role in the dynamics of the aquaculture system; however, using only the total N level, it was still difficult to determine the role of various derivative N compounds at the study location. This was due to dual roles of N compounds as nutrients and as poisons [33]. As with the total N-content, TOM of the soil was also lower than the required for growth of aquatic organisms. TOM content of soils samples obtained from the study area ranged from 0.14% to 3.78%, which were lower than suitable values for aquaculture [34]. Low TOM content of the pond soil, efforts were needed to increase the level through the application of organic fertilizers in order to increase the biomass of natural feed. However, the high content of organic matter in soils can cause an increase in the population of bacteria, CO2, H2S, and CH4 that potentially inhibit the growth of vannamei shrimp. It also required a high oxygen demand and causes severe oxygen depletion, which in turn affected fish yields [35]. Since most fish and shrimp ponds in the study area were managed following traditional plus (improved extensive) and semi-intensive technologies, it was necessary to provide natural food to support the growth of cultured organisms. As with other traditional aquaculture practices in Indonesia, the management of brackishwater ponds in this location is carried out with rudimentary methods including sub-surface soil removal, use of chemicals (lime, fertilizers and pesticides), regular water exchanges to improve water quality and enhance nutrient recycling in pond, removal of accumulated organic materials, and repairing of cracked dykes. Bearing in mind that lime and fertilizer requirements, for example, might not be longer suitable with the current conditions of pond environmental quality. Based on field observation, it was very clear that most farmers did not fully take into account the aforementioned key soil and water variables during site election, pond preparation, and pond management. These pond management methods have been practiced from generation to generation. Therefore, the brackish water aquaculture management practices must be (re)evaluated regarding significant changes in the quality of the pond environment and specific characteristics of pond clusters or pond units.
Conclusion
The present study has shown that the fine-scale spatial analysis approach provided a very useful information that was applicable to the farm level of aquaculture management in general. The finescale spatial analysis approach accurately maps and gives insight into environmental and pond engineering constraints in the study site. Given its appropriate soil quality, water quality and climate conditions; there was a continuous shrimp or fish loss that causes the total production from brackishwater aquaculture below target. The present aquaculture practices in the study site have not fully adopted or complied with good aquaculture practices and farmers have not properly managed the existing limiting factors. The environmental limiting factors identified must be handled appropriately to ensure the increased production and sustainability of brackishwater aquaculture in the study area.
|
2019-12-05T09:26:37.404Z
|
2019-12-02T00:00:00.000
|
{
"year": 2019,
"sha1": "7d1efff1fa558a6a3f043dd6e46f31d81e64f294",
"oa_license": null,
"oa_url": "https://doi.org/10.1088/1755-1315/370/1/012067",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "2e49553d010761148701050f2262c317eb96971d",
"s2fieldsofstudy": [
"Environmental Science"
],
"extfieldsofstudy": [
"Environmental Science"
]
}
|
10479853
|
pes2o/s2orc
|
v3-fos-license
|
Paradox of schizophrenia genetics: is a paradigm shift occurring?
Background Genetic research of schizophrenia (SCZ) based on the nuclear genome model (NGM) has been one of the most active areas in psychiatry for the past two decades. Although this effort is ongoing, the current situation of the molecular genetics of SCZ seems disappointing or rather perplexing. Furthermore, a prominent discrepancy between persistence of the disease at a relatively high prevalence and a low reproductive fitness of patients creates a paradox. Heterozygote advantage works to sustain the frequency of a putative susceptibility gene in the mitochondrial genome model (MGM) but not in the NGM. Methods We deduced a criterion that every nuclear susceptibility gene for SCZ should fulfill for the persistence of the disease under general assumptions of the multifactorial threshold model. SCZ-associated variants listed in the top 45 in the SZGene Database (the version of the 23rd December, 2011) were selected, and the distribution of the genes that could meet or do not meet the criterion was surveyed. Results 19 SCZ-associated variants that do not meet the criterion are located outside the regions where the SCZ-associated variants that could meet the criterion are located. Since a SCZ-associated variant that does not meet the criterion cannot be a susceptibility gene, but instead must be a protective gene, it should be linked to a susceptibility gene in the NGM, which is contrary to these results. On the other hand, every protective gene on any chromosome can be associated with SCZ in the MGM. Based on the MGM we propose a new hypothesis that assumes brain-specific antioxidant defenses in which trans-synaptic activations of dopamine- and N-methyl-d-aspartate-receptors are involved. Most of the ten predictions of this hypothesis seem to accord with the major epidemiological facts and the results of association studies to date. Conclusion The central paradox of SCZ genetics and the results of association studies to date argue against the NGM, and in its place the MGM is emerging as a viable option to account for genomic and pathophysiological research findings involving SCZ.
Background Schizophrenia (SCZ) is a common and highly heritable form of psychosis associated with a remarkable biological disadvantage [1]. Accordingly, the central paradox of SCZ genetics is how susceptibility genes are preserved in the human gene-pool against a strong negative selection pressure [2][3][4][5]. Although recent large-sample epidemiological studies [6][7][8][9] have consistently shown that the reproductive fitness of the unaffected female siblings of patients with SCZ is slightly increased (1.02-1.08), it is not large enough to compensate for the gene loss due to the decreased reproductive fitness of patients (0.2-0.3 in males and 0.4-0.5 in females) and their unaffected male siblings (0.9-1.0) in the nuclear genome model (NGM). Furthermore, the latest meta-analysis shows that parents of patients with SCZ have fertility similar to the general population [10]. Therefore, heterozygote advantage does not seem to work in the NGM.
However, it works in the mitochondrial genome model (MGM) because mtDNA is transmitted to the next generation only through females. Indeed, we can see that this slightly elevated fitness of the unaffected female siblings, coupled with the less pronounced decreased fitness of female patients, is sufficient to compensate for the gene loss in the MGM; when we calculate -Δ, the cross-generational reduction of the frequency of females carrying putative pathogenic mtDNA in the general population, using data from the largest-sampled cohort study to date [8], we have ÀΔ < 5:06 Â 10 À3 [11]. This figure implies that the gene loss can be balanced by de novo mutation in the mtDNA which occurs at a rate of 8:8 Â 10 À4 ∼1:3 Â 10 À2 per locus per generation (4.3 × 10 -3 on average) [12].
Thus, in our previous paper [11], we carefully reexamined the necessary conditions for putative nuclear susceptibility genes for SCZ and deduced a criterion (the persistence criterion, or 'P-criterion') for nuclear susceptibility genes for SCZ based on the general assumptions of the multifactorial threshold model. Because SCZ must be sustained by mutation-selection balance without heterozygote advantage in the NGM and the mutation rates in the nuclear genome are much lower than in the mitochondrial genome [12][13][14], the criterion is strong. It demands 0 < d < ν 1Àsp ð Þ μ 1Àp ð Þsp (Appendix A), while the principle of association study demands d>0. Here d, μ, p and s denote the case-control difference of allele frequencies, the mutation rate at a putative risk locus, and the prevalence and the selection coefficient of the disease respectively. According to epidemiological data collected by Haukka et al. [8] p ¼ 1:29 Â 10 À2 and s ¼ 6:54 Â 10 À1 ð Þ , ν is very small for SCZ. Given the average mutation rate in the autosomes and the X chromosome μ ¼ 1:48 Â 10 À5 À Á , we have ν ¼ 1:76 Â 10 À3 , and given the highest mutation rate μ ¼ 1:48 Â 10 À4 ð Þ , we have ν ¼ 1:76 Â 10 À2 (Appendix B). Using this criterion, we can calculate a lower boundary for the sample size required in an association study of a given power (Appendix C). Thus, we can see that more than 29,000 case-control pairs are required in a genome-wide association study (GWAS) with a power of 0.8 to detect a susceptibility variant of the highest mutation rate and a population frequency between 0.005 and 0.995, and that more than 2.9 million case-control pairs are required in a GWAS with a power of 0.8 to detect a susceptibility variant of the average mutation rate and a population frequency between 0.0005 and 0.9995.
In the past two decades, mitochondrial dysfunction (MD) in SCZ has been suggested by several independent lines of evidence [15][16][17][18], that include mitochondrial hypoplasia, elevated oxidative stress (OS), disturbed oxidative phosphorylation (OXPHOS), and altered mitochondrialrelated gene expression in several cell lines. It is still unknown whether MD in the brains of patients with SCZ is a primary pathogenic change or a secondarily induced process, and the aforementioned inconclusive findings create more questions than answers about the origin of SCZ. We contend, however, that the MGM may very well have the answers.
In the present study, we apply the P-criterion to the results of association studies to date and survey the distribution of SCZ-associated genes that could meet or do not meet the P-criterion. We discuss why those SCZassociated genes that do not meet the criterion must be protective genes (PGs) for SCZ, and suggest that the MGM can replace the NGM as a basis for genomic research of SCZ. Based on the MGM we propose a new hypothesis that assumes brain-specific antioxidant defenses in which neurotransmissions are involved, and present ten predictions of the hypothesis. 45 in the SZGene Database [19] (the version of the 23 rd December, 2011) were selected. Based on the genotype distributions in the meta-analyses, allele frequencies and the case-control differences were calculated.
Results and discussion
Distribution of SCZ-associated variants that could meet the P-criterion Association studies to date have detected many SCZassociated genes on various chromosomes [19,20]. However, most of these do not meet the criterion; only 21 variants on18 loci in the top 45 genes in the SZGene Database could meet the criterion under the assumption that the mutation rates at those loci are near the upper limit on the autosomes and the X chromosome (Tables 1 and 2). In other words, association studies have detected only a few susceptibility genes in the NGM, if any, that produce small effects (in most cases OR < 1.4). Association studies to date (sample size <50,000 case-control pairs) may lack the power to detect pathogenic variants of less than the highest mutation rates. The question remains: how many samples would be required to detect them?
The P-criterion implies that an enormous sample size is required to identify a pathogenic variant of the average mutation rate. Indeed, more than 2.9 million case-control pairs should be recruited to a GWAS to identify a susceptibility variant of a population frequency between 0.0005 and 0.9995 with a power of 0.8 (Appendix C). If the mutation rate at a putative risk locus is relatively low μ ¼ 1:48 Â 10 À6 À Á , more than 290 million case-control pairs are required in a GWAS to detect a risk variant of a population frequency between 0.000005 and 0.999995 with a power of 0.8 (Appendix C). Since the total human population today is~7 billion [21] and the overall prevalence of SCZ is~0.7% [22], the total number of patients with SCZ in the world today is~49 million. Therefore it would take more than several hundred years to gather the required number of samples even if all of the affected individuals in the world were to be recruited. The journey to search for susceptibility genes for SCZ in the NGM seems far more difficult and would take longer than previously thought.
Distribution of SCZ-associated variants that do not meet the criterion
Now, let us consider the nature of those SCZ-associated genes that do not meet the criterion. The inequality d≥ v implies s M < s, where s M and s denote the selection coefficient in the affected subpopulation with an allele M and in the affected population as a whole respectively (Appendix A); otherwise, the frequency of M in the affected population must have been decreased to the same level in the unaffected population. Therefore, such genes, if sustained by mutation-selection balance, cannot be susceptibility genes but instead are PGs that decrease the risk and the severity of the disease.
Since a PG in the NGM cannot be associated with the disease unless it is linked to a susceptibility gene, PGs in the NGM should be located in the vicinity of susceptibility genes, something that is contrary to the results of association studies to date. For example, on chromosome 1, all of the SCZ-associated genes that could meet the criterion (RGS4, PLXNA2, DISC1) are located on 1q, while 4 PGs (MHTFR, GRIK3, PDE4B, GSTM1) are on 1p (Table 1). 15 PGs are located on 2q, 5q, 7q, 10q, 11p, 12p, 12q, 13p, 13q, 16p, 17p, and 19q, where no SCZ-associated variants that could meet the criterion are located ( Table 2). Therefore, the results of association studies to date do not support the NGM.
The only plausible interpretation that accords with the NGM may be that many nuclear susceptibility genes of less than the highest mutation rates have not been detected by association studies to date due to lack of power. If we assume this case, however, an enormous sample size (more than 2.9~290 million case-control pairs) would be required to identify them. Although the possibility of the NGM cannot be entirely denied, the proof of the NGM may be extremely difficult because it requires such an enormous sample size.
The MGM is compatible with the results of association studies to date A case can be made for the MGM however, because every PG on any chromosome can be associated with SCZ in the MGM. This is because mtDNA is transmitted only via females and there is no link between the nuclear genome and the mitochondrial genome; such that every nuclear genome that interacts with a pathogenic mitochondrial genome to alter the risk and the severity of the disease is subject to natural selections in the predisposed maternal lineage that succeeds the same pathogenic mitochondrial genome. Therefore, in the MGM every PG on any chromosome is subject to positive selection in the predisposed maternal lineage, and is thereby associated with SCZ.
Thus, the MGM is compatible with the results of association studies to date.
It should be noted that in the MGM the frequency of every facilitating gene (FG; a gene that increases the risk and the severity of SCZ in the presence of a pathogenic mitochondrial genome) on any chromosome may diminish in the predisposed matrilineal pedigrees by negative selection; thereby, negatively associating with SCZ.
Pathophysiological aspect of SCZ and the MGM
Mitochondria are involved in a variety of major cellular events such as OXPHOS, free radical production and Ca 2 + buffering, and play an active role in apoptosis. They possess two classes of antioxidant defenses (non-enzymatic or enzymatic) [23], and structurally and functionally intact mitochondria serve as a net sink rather than a net source of reactive oxygen species (ROS). ROSdefenses are severely undermined in structurally compromised mitochondria [23], and MD, presumably through imbalance of ROS production and removal [23], raises ROS emission [24,25], causing a strong intracellular OS. Mitochondrion has dual genetic bases (ncDNA and mtDNA), and there are 37 loci in the mtDNA (22 tRNAs, 2 rRNAs, and all of the 13 OXPHOS subunits: ND1, ND2, ND3, ND4, ND4L, ND5, ND6, CO1, CO2, CO3, CYB, ATPase6, and ATPase8). Therefore, abnormal mtDNA may cause disturbed OXPHOS and enhanced OS. In the past decade, MD in SCZ has been suggested by several independent lines of evidence (for review, see [15][16][17][18]); those include increased OS, mitochondrial hypoplasia, disturbed OXPHOS, and altered mitochondrialrelated gene expression in several cell lines.
The pioneering works in this field are noteworthy. As early as 1950, Hayashi [26], in a longitudinal study on glucose metabolites in blood sampled from the superior bulb of the internal jugular vein of patients with SCZ, observed a lower degree of carbonic dioxide production in the brain and a higher level of lactate and glutathione, the brain's dominant free radical scavenger, in patients in an acute exacerbation of the illness. Utena and Ezoe (1951) [27] reported decreased glucose consumption in vitro in cortical brain tissues sampled from patients with SCZ who underwent prefrontal leukotomy. Takahashi (1953) [28] confirmed this finding and recommended further investigations on OXPHOS in the brain tissue of schizophrenics. In line with those findings was the report by Stabenau et al. (1969) [29], who observed in a biochemical study of discordant monozygotic twin pairs that lactate production and the lactate-pyruvate ratios were higher in the affected twins than in the unaffected co-twins. More recently, Prabakaran et al. (2004) [30], in a large-scale functional genomics study, suggested a state of intermittent or chronic hypoxic stress and MD in the brains of patients with SCZ.
There has been consistent evidence for higher incidence of underdevelopment of muscular system in patients with SCZ [31]. Many authors have confirmed the Kretchmer's observation (1925) of a greater incidence of asthenic or leptosomatic constitution in schizophrenic or pre-schizophrenic individuals, which is characterized by narrowness of muscles, small shoulders, flat chest, and little muscular development [31]. These early observations have been supported by later studies of muscle biopsy and/or electromyography, which showed abnormalities of both muscle function (more hypotonus) and muscle morphology (atrophic muscle fibers scattered or in groups, regardless of the patients' motor activities) in SCZ [32][33][34][35][36].
There also have been many case reports that describe patients with mitochondrial disease and neuropsychiatric disorders including SCZ (for review, see [37,38]). Most of the patients have psychotic symptoms as well as fatigue and muscle weakness that preceded the diagnosis of mitochondrial disease. Furthermore, it has been shown that mtDNA, in interaction with ncDNA, modifies learning, exploration, and sensory development as well as the brain morphology in mice [39].
Those findings indicate that abnormal mtDNA could underlie the etiology of SCZ.
The DA hypothesis and NMDA-R hypofunction theory as well as the recent finding of enhanced carbonyl stress in SCZ may be incorporated into the MGM as will be discussed in the following sections.
Contradictory aspects of dopamine-SCZ connection and the MGM Substantial evidence has been accumulated that supports aberration in the dopamine (DA) system [40,41] and its interactions with other neurotransmitters [42] in SCZ. This concept (DA hypothesis of SCZ) mainly stems from indirect observations that there is a strong relationship between clinical potency of neuroleptics and their DAD2 receptor affinities in vitro and that DA agonists can induce psychotic symptoms with marked resemblance to SCZ. More recently, Laruelle et al. [43] reported that DA transmission was increased in patients experiencing an acute exacerbation of the illness, but not in patients in a state of remission.
Those findings strongly suggest that increased DA transmission provides a neuronal basis for positive or productive symptoms in SCZ. Since negative symptoms and cognitive impairments that can precede initial psychotic episode are difficult to explain by the increased DA transmission, a plausible interpretation may be that the increased DA transmission in the brain is a secondary process associated with SCZ. However, since excessive DA release seems to cause MD [16] and apoptosis [44], it could produce additional deterioration of the brain function and morphology after psychotic episodes.
On the other hand, Friedhoff and colleagues [45][46][47] reported some adaptive aspects of the DA system in SCZ, and proposed that the DA system may be a biological stress buffer. This concept is based on a consistent observation that patients responding favourably to neuroleptic treatment showed higher levels of plasma homovanillic acid (HVA), a metabolite of DA, in the pretreatment and/or the early treatment period, resulting in a marked decrease of plasma HVA level in a later steady state [46,[48][49][50]. This observation suggests that increased DA turn over in the pre-treatment or early treatment period may be relevant to a certain adaptive process in patients with SCZ. In line with those findings is the discovery of Cys311 of DRD2 polymorphism [51], which is listed in the top 45 in the SZGene. This firstly discovered SCZ-associated variant supresses the internalization of DRD2 and strengthens the DA transmission [52] while the cases carrying this variant show less negative symptom and less cognitive impairment [51].
Then, how can we reconcile those two contradictory aspects of the DA-SCZ connection?
At low doses catecholamines, such as DA, noradrenaline (NA) and adrenaline (A), are thought to protect nerve cells by virtue of their antioxidant activities; whereas, at high doses they are thought to be neurotoxic, acting as pro-oxidants [44]. The protective effect of DA, but not of NA and A, has been shown to be through the activation of DA D4 receptors [53]. It is also shown that DA D2 and/or D3 receptor activation protects neurons [54] and/or oligodendrocytes [55] against OS. Thus it is suggested that DA, through the activation of D2, D3 and/or D4 receptors, may protect cells from gradual changes in low levels of OS that occur in mild pathological insults and aging [53][54][55]. Therefore, it seems not implausible to suppose that increased DA transmission in the brain could occur as an adaptive process against OS, especially in the individuals with MD and resulting vulnerability to OS.
Indeed, it has been shown that in vitro and in vivo treatment with the mitochondrial complex I inhibitor rotenone enhances the sensitivity of striatal DA release to OS induced by H 2 O 2 [56]. Thus, if MD and resultant vulnerability to OS underlie the pathophysiology of SCZ, positive symptoms of the disease might be explained as 'the price that the brain pays for its homeostasis at OS'.
NMDA-R hypofunction theory of SCZ and the MGM
N-methyl-D-aspartate-receptor (NMDA-R) antagonists, such as phencyclidine and ketamine, produce SCZ-like symptoms including negative symptoms and cognitive impairments in healthy individuals [57][58][59], and exacerbate both positive and negative symptoms in patients with chronic SCZ [60][61][62]. In addition, NMDA-R antagonists induce disinhibition of neurotransmitter systems through specific effects on inhibitory circuits [63]. These indirect observations have led to the NMDA-R hypofunction theory of SCZ [42,64], which proposes that NMDA-R hypofunction provides a neuronal basis for the negative symptoms and cognitive impairments associated with SCZ and induce aberrations in the monoaminergic systems that could result in positive symptoms. Postmortem studies have revealed alterations in pre-and post-synaptic markers for glutamatergic neurons in several brain regions in patients with SCZ [64]. However, no conclusive evidence of NMDA-R hypofunction in SCZ has been obtained.
It has been shown that a rapid increase in ROS occurs in vitro [65] and in vivo [66] after exposure to NMDA-R antagonists. Exposure to ketamine suppresses the expression of GABAergic interneurons by inducing a persistent increase in brain superoxide through activation of NADPH-oxidase in the prefrontal cortex [63]; thereby, causing increased excitatory neurotransmission. Thus, the evidence for NMDA-R hypofunction theory may also support the hypothesis that OS plays a central role in the pathophysiology of the symptoms associated with SCZ.
It should be noted that while excessive activation of NMDA-R triggers neuronal degeneration through intracellular OS and resulting MD [67], NMDA-R activation at subtoxic doses has a neuroprotective property due to brain-derived neurotrophic factor (BDNF) release [68][69][70] and/or blockade of c-Jun N-terminal kinase (JNK) activation [71]. NMDA-R has regulatory redox sites through which the oxidation status regulates the physiological activity of the receptors [72][73][74][75]. More recently it has been shown that trans-synaptic stimulation of NMDA-R boosts the class II antioxidant defenses by modifying the thioredoxin-peroxiredoxin system [76]. Thus, it would be conceivable that the brain, the highest energy consuming organ vulnerable to OS, has a third class of antioxidant defenses ('class III defenses') in which neurotransmissions are involved, and where transsynaptic activation of NMDA-R as well as DA-R may occur, through unknown pre-synaptic redox regulation mechanisms, as an adaptive response to OS which could not be suppressed by the class I and II defenses. Failure to suppress OS by those three classes of antioxidant defenses may lead to persistent and strong OS and through suppressing the expression of inhibitory interneurons, may cause prolonged excessive glutamate and/ or DA release, leading to excitation toxicity and additional deterioration of the brain function ( Figure 1).
Enhanced carbonyl stress (CS) in SCZ and the MGM
Recently Arai et al. [77] reported that a subset (47%; N = 21) of patients with SCZ (N = 45) exhibit CS with high plasma pentosidine levels (>mean +2SDs of controls), without underlying diabetes mellitus or chronic kidney disease, the two major causes of elevated AGEs (advanced glycation end-products). In addition, marked decreases in serum pyridoxal levels were found in 50% (N = 11) of those patients with CS. Since pyridoxal is a reactive carbonyl compounds (RCOs) scavenger, decreased pyridoxal levels may be due to the consumption of pyridoxal during prolonged CS. However, the molecular mechanisms underlying the CS in those patients remain unclear.
They also identified 2 rare frameshift variants of loss-of function type in GLO1 that encodes glyoxalase 1 (GLO1), a detoxifying enzyme of RCOs. While marked CS characterized by elevated AGEs and decreased pyridoxal was associated with those variants in patients with SCZ, it was not in healthy controls. Then, how did these rare variants of GLO1cause CS only in patients with SCZ?
A number of enzymatic pathways including glyoxalase pathway contribute to the detoxification of RCOs, and redox coenzymes (reduced GSH and NAD(P)H) are particularly important elements for the activity of these pathways [78]. On the other hand, it has been well known that OS increases RCOs production [78]. Therefore, enhanced OS and secondarily induced alteration of redox coenzyme homeostasis may cause CS in the affected population (Figure 1). And while these missense variants in GLO1 may cause marked CS in patients with SCZ, they may not cause CS in healthy controls because of normal redox coenzyme activities and lack of enhanced OS.
In this case, the rare frameshift variants of GLO1 detected by Arai et al. must be FGs. Indeed, patients carrying those rare variants have been severely affected and treatment-resistant cases while the frequency of those variants was not higher in patients with SCZ (5/1761 = 0.28%) than in healthy controls (10/1921 = 0.52%) [77].
Predictions of the hypothesis
This hypothesis predicts: (1) a higher maternal transmission of SCZ (H1), (2) positive associations between PGs and SCZ as well as negative associations between FGs and SCZ (H1), and (3) various pathogenic alterations such as genomic instability, aberrations in neuromuscular development, and brain dysfunction, that can be caused by disturbed OXPHOS and enhanced OS in predisposed individuals (H1 + H2). In addition, the hypothesis predicts a possible therapeutic effect of ROS scavengers (H2).
Most of these predictions seem to be consistent with major epidemiological findings and the results of genetic studies to date.
The mode of transmission
The hypothesis (H1) predicts a higher maternal transmission. Although there has been no convincing evidence for an exclusively maternal transmission of SCZ, several reports suggest a higher maternal transmission of SCZ [79][80][81][82].
It also predicts that a certain proportion of patients are sporadic cases due to de novo mutation in the mtDNA. Given the average mutation rate in the mtDNA (4.3 × 10 -3 ) and the overall prevalence of SCZ (~0.7%), the expected proportion of de novo sporadic cases in the affected population is~0.6 (= 0.0043/0.007). This figure accords with the results of the epidemiological studies, and the fact that 60% of patients with SCZ have no affected first-and second-degree relatives [1].
Some researchers have hypothesised that SCZ is associated with de novo mutations arising in paternal germ cells [83][84][85][86][87]. This is based on the observation Figure 1 Proposed pathophysiology of schizophrenia. Mitochondrial dysfunction (MD), through imbalance of ROS production and removal, raises ROS emission, causing a strong intracellular OS. Disturbed OXPHOS and enhanced OS in predisposed individuals may cause various pathogenic alterations such as genomic instability, aberrations in neuromuscular development, brain dysfunction, and apoptosis. On the other hand, enhanced OS and secondarily induced alteration of redox coenzyme homeostasis may cause enhanced CS. It is assumed that the brain has the third class of antioxidant defenses in which neurotransmissions are involved, and that trans-synaptic activation of NMDA-R as well as DA-R may occur, through unknown pre-synaptic redox regulation mechanisms, as an adaptive response to OS which could not be suppressed by the class I (non-enzymatic) and II (enzymatic) defenses. Failure to suppress OS by those three classes of antioxidant defenses may lead to persistent strong OS and, through suppressing the expression of inhibitory interneurons, may cause prolonged excessive glutamate and/or DA release, leading to excitation toxicity (ET) and additional deterioration of the brain function.
MD OS CS
('paternal age effect') that the risk of SCZ in the offspring seems to increase as the paternal age advances from 20 years to over 50 years. However, the difference in the mean age of fathers of affected and unaffected individuals is not very large (< 1.7 years) [83,86]. Furthermore, there has always been increased risk of SCZ for the offspring of younger men (< 21 years) [83,86,87] and younger women (< 20 years) [86]. Therefore, major roles of paternally derived mutations in SCZ seem to remain unsubstantiated. Indeed, no available data can exclude the possibility that the 'paternal age effect' has a 'maternal origin'; while women in many countries today may be expected to bear children after the age of 20 years and not to marry much older or younger men unless the men have special socio-economic benefits, a certain proportion of women genetically predisposed to SCZ might behave differently.
It should be noted that in the famous twin study by Gottesman and Bertelsen [88] that included almost equal numbers of male and female monozygotic twins discordant for SCZ, most schizophrenic twins whose offspring were affected were females (12 out of 14), implying that the transmission was mainly via females P ¼ 14 C 12 Â ð 0:5 14 þ 14 C 13 Â 0:5 14 þ 14 C 14 Â 0:5 14 < 0:007Þ.
Every SCZ-associated gene is either a PG for SCZ or a PGlinked gene
In the MGM, G-G interactions between the nuclear genome and the pathogenic mitochondrial genome give rise to SCZ-associated PGs that decrease the risk and the severity of SCZ. Thus, the hypothesis (H1) predicts that every SCZ-associated gene is either a PG for SCZ or a PG-linked gene.
Cys 311 of DRD2 polymorphism and 'high risk' variants of NRG1, DTNBP1, and NPY may be such examples because schizophrenics carrying those variants exhibit less negative symptoms [89][90][91]. Unfortunately, most association studies to date lack sufficient clinical data about the onset and the clinical symptoms of subjects to determine the nature of the SCZ-associated variants detected.
An apparent signature of positive selection in SCZassociated genes
Since the positive selection of the SCZ-associated alleles (= protective alleles) mentioned above occurs only in the predisposed matrilineal pedigrees, a ubiquitous subpopulation in humans, frequencies of those alleles may not be as high as if the selection had occurred recently in the general population. Thus the hypothesis (H1) predicts that every SCZ-associated nuclear gene has an apparent signature as if it has been subject to positive selection in the recent evolutionary history of humans.
Two recently published reports [92,93] seem to concur with this prediction. On the other hand, the NGM predicts that every SCZ-associated nuclear gene has an apparent negative selection signature due to the strong negative selection pressure, which is contrary to these reports.
Sex differences and the protective effect of estrogen in SCZ
The hypothesis (H2) predicts that endogenous antioxidants exhibit a protective effect against SCZ, and may give a plausible explanation for sex differences in the disease.
A consistent and specific finding for SCZ is that the age at onset is significantly lower in males than in females [1,94]. SCZ starts earlier on average in males and reaches its peak between 15 and 25 years of age; whereas, in females it occurs almost between 20 and 30 years of age and shows a less steep curve after that age. It also appears that women are vulnerable to relapses or first episodes of SCZ in the perimenoposal period (the second peak of onset for females) [94], when estrogen production diminishes. A close association between premenstrual or menstruation phase and exacerbation of the illness in females has been well documented [94]. In addition, less negative symptoms, less brain morphological changes, and a better response to neuroleptics are relatively consistent findings in female patients with SCZ [95].
From these observations, one may infer that estrogen protects predisposed females [94]. However, the exact mechanism for this finding remains unclear [94,95]. The hypothesis (H2) may provide a plausible explanation, because estrogen has been shown to exhibit antioxidant activity due to its intrinsic antioxidant structure that lies in the phenolic moiety of the steroidal compound [96], to increase antioxidant enzyme activities [97,98], and to have a neuroprotective effect against OS [96,99]. Furthermore, mitochondrion has estrogen binding sites [100,101] and estrogen increases mitochondrial efficiency and reduces intracellular OS [102].
Prenatal risk factors for SCZ
The hypothesis (H2) predicts that early-life exposure to environments that induce OS can increase the risk of later development of SCZ in the predisposed population.
Increased obstetric complications in the birth of patients with SCZ
It has been suggested that MD may be involved in the etiology of preeclampsia [123][124][125]. Furthermore, a high incidence of preeclampsia, eclampsia, and stillborn infants has been observed in a family with a known mitochondrial disorder [126]. Folgero et al. [127] demonstrated two separate mtDNA point mutations in two families having a high incidence of preeclampsia and eclampsia.
Therefore, the hypothesis (H1) predicts that the risk of preeclampsia and stillbirth may be increased in the birth of patients with SCZ as well as in the pregnancies of women with SCZ.
There has been a body of evidence used to account for an increased risk of obstetric complications (OCs) in the birth of patients with SCZ [128,129]. A meta-analysis of population-based data [129] found significant estimates for three main categories of OCs: (1) complications of pregnancies, (2) abnormal fetal growth and development, and (3) complications in delivery. Among all, preeclampsia was the strongest individual risk factor detected in the largest single population-based cohort study to date [128]. Although obstetrical events in SCZ are often considered as having a direct causative effect, no available data have been able to refute the theory that these events are merely markers of some other causal process [130], such as MD implicated in this hypothesis (H1). An excess of stillbirths and neonatal deaths among women with SCZ has been reported by several investigators [131][132][133].
Low comorbidity between SCZ and rheumatoid arthritis (RA)
The hypothesis (H1) predicts that diseases predisposed by FGs, if present, would be negatively associated with SCZ. RA could be one such candidate.
OS plays a role in the pathogenesis of RA (for review, see [134]). In addition, it has been shown that chronic OS in the synovial T lymphocytes is not secondary to exposure to environmental ROS, but instead originates from intracellularly produced ROS [135]. Therefore, a presumed susceptibility gene for RA that may cause intracellular OS in several cell lines could be a FG for SCZ in the MGM, and is likely negatively associated with SCZ.
Thus the MGM may offer a new explanation for the low comorbidity and provide an additional prediction: most of patients with both of SCZ and RA would be females because the survival rate of males in early life stages must be remarkably reduced due to lack of protection by estrogen, and show more negative symptoms, poorer response to neuroleptic medication, and/ or more pronounced morphological changes in the brain.
Genomic or epigenomic instability
It has been shown that endogenous mitochondrial OS can induce many types of DNA damage including double strand breaks, end-to-end fusions, base and sugar modifications, DNA-protein cross-links, and gross chromosomal rearrangements [139,140]. Therefore, the hypothesis (H1) predicts that the enhanced OS may cause genomic or epigenomic instability during meiosis and/or early phase of ontogeny, producing increased rates of random point mutations and/or structural variants of the nuclear genome in the affected population. In addition, genomic instability may be more pronounced in male patients due to lack of antioxidant protection by estrogen.
There have been numerous reports of associations between SCZ and chromosomal abnormalities including fragile sites, reciprocal translocations, inversions, insertions, deletions, disomy and trisomy in many autosomes, and sex chromosome aneuploidies [141]. However, with the exception of 22q11 deletion, none of these have been consistently replicated.
A popular explanation in the NGM may be that most of these structural variants are coincidental findings of no clinical significance. Alternatively, those alterations may indicate genomic instability in SCZ. An increased risk of SCZ in individuals with 22q11 deletion [142,143] might be due to haplodeficiency of presumptive PGs of gain-offunction type and/or presumptive FGs of loss-of-function type aggregated on 22q11.
More recently, it has been reported that rare structural variants such as microdeletions or microduplications of sizes ranging from 100 kb to 15 MB throughout the genome are more frequent among individuals with SCZ than unaffected individuals [144]. While many of those structural variants duplicate or delete genes in neurodevelopmental pathways, one third of those do not disrupt genes, leaving their role in causation of the disease unwarranted. Another recent report [145] has shown that de novo copy number mutations are increased in sporadic SCZ. However, the cytobands of those copy number mutations are diverse among the affected individuals and their roles in the pathogenesis still remain unclear. Therefore, no available data can refute the possibility that those structural variants and copy number mutations are not the cause of SCZ but the results of the genomic instability in SCZ predicted by our hypothesis.
Indeed, direct measure of the de novo mutation rates shows an increased mutation rate in SCZ [146], and genomic or epigenomic instability has been suggested in SCZ [147]. Furthermore, it has been shown that blood cells from patients with SCZ present a higher rate of folate-sensitive fragile sites, and that male patients exhibit twice as many fragile sites as females while there are no age effects [148]. This sex difference may indicate that increased fragile sites expression (genomic instability) is the results of enhanced OS associated with SCZ.
Inconsistency in the results of the association studies on the nuclear genome
Results of association studies based on the NGM have been inconsistent, and SCZ-associated genes including copy number variations differ across populations or even across individuals of the same ethnicity [19,20,144,145].
The only plausible explanation for this inconsistency in the NGM is that there is an extreme heterogeneity in the genetic basis of SCZ.
On the other hand, these perplexing results of the association studies to date accord with a specific prediction of the hypothesis: in the MGM, a disease-associated allele is likely to be a PG, and may show a stronger association with milder cases and unaffected siblings but not with severe cases, while a negatively associated allele is likely to be a FG, and may show a positive association with severer early-onset cases. Therefore, an association detected in a study with mildly affected subjects may not be supported in studies with severely affected subjects, and an association detected in a study with severely affected subjects may not be supported in studies with mildly affected subjects. This is a specific prediction of the hypothesis (H1).
Therapeutic effects of ROS scavengers
In this hypothesis (H1 + H2), OS and secondarily induced CS cause pathogenic alterations in SCZ. Therefore, it predicts that ROS scavengers have therapeutic effects on SCZ.
Indeed, adjuvant therapy using neuroleptics and antioxidants such as omega-3 poly unsaturated fatty acids, ascorbic acid, α-tocopherol or N-acetyl cysteine have been shown to improve clinical outcomes of patients with SCZ (for review see [18]). However, the most promising therapeutic antioxidant for SCZ may be molecular hydrogen, which rapidly diffuses across membranes and selectively reacts with cytotoxic ROS, but not with other ROS that play physiological roles [149]. This prediction can be easily tested in the near future.
Possibilities and limitations of the hypothesis
This new hypothesis seems to explain various and specific aspects of SCZ and somewhat perplexing results of association studies to date. It meets the four criteria for evolutionary hypotheses of SCZ proposed by Brüne [4] because it: (1) provides a convincing mechanism for the preservation of genes in the human gene pool associated with SCZ, (2) explains potential sex differences, (3) accounts for the multifaceted symptomatology, and (4) is consistent with neuropsychological, developmental and evolutionary findings regarding the human brain.
However, there are several limitations to this hypothesis. First of all, it is only a theory that lacks direct and compelling evidence. Especially, the assumption of the class III antioxidant defenses in the brain should be tested in neurophysiological studies in the future. Second, the exact mechanism of discordance in monozygotic twins remains unclear. Although the hypothesis may explain the discordance due to different mutation loads in the first ovum cleavage, it still requires future investigation. Third, while there have been several studies that report new variants in the mitochondrial genome associated with SCZ [150][151][152][153][154][155][156], there have been no consistent findings on the pathogenic variants in the mitochondrial genome. However, it should be noted that any missense variant in the mitochondrial genome can be a pathogenic genome in the MGM because it may contribute to cause disturbed OXPHOS and enhanced OS. There could be heterogeneity in the genetic basis of SCZ in the MGM, as are the cases with known mitochondrial diseases [157].
Although a recent study failed to detect any pathogenic mitochondrial genome in the blood cells [158], it cannot necessarily deny the possibility of the MGM because of the heteroplasmy of the mitochondrial genome. It has been well known that pathogenic mitochondrial genome cannot necessarily be detected in the blood cells in known mitochondrial diseases [159]. Therefore, other cell lines such as muscles, neurons, or hair follicles should be examined to detect pathogenic variants in the mitochondrial genome in the future.
Conclusions
Genetic research of SCZ based on the NGM has been one of the most active areas in psychiatry for the past two decades. Although this effort is ongoing, results of association studies based on the NGM have been disappointing, or rather perplexing. No particular susceptibility gene that accounts for a large proportion of heritability has been identified by association studies. The results of association studies have been inconsistent, and SCZ-associated variants including copy number variations differ across populations or even across individuals of the same ethnicity.
The central paradox of SCZ genetics and the results of association studies to date argue against the NGM for SCZ today, and in its place the MGM is emerging as a viable option to account for genomic and pathophysiological research findings involving this enigmatic disorder.
Appendix A. Deduction of the persistence criterion
Deduction of the P-criterion is explained in detail elsewhere [11]. Therefore, we present here the essence of the method. At first we describe our three basic assumptions.
An ideal human population
We assume here a random-mating human population with a sufficiently large effective population size at equilibrium, where negative selection pressures on the susceptibility alleles for SCZ are predominant and the effect of genetic drift is negligibly small. The prevalence p of SCZ in this ideal human population is assumed to be stable across generations by mutation-selection balance. Therefore, the gene frequency in the general population (m G ) is given in terms of the gene frequencies in the affected population (m A ) and in the unaffected population (m U ): Mutation-selection balance in each risk locus We assume here that the total number of population frequencies of the pathogenic alleles at each risk locus is preserved by mutation-selection balance. Therefore, ÀΔm G , the cross-generational reduction in the frequency of a pathogenic allele should not be more than the rate of mutations that produce pathogenic variants at the locus. On the other hand, since mutations at the locus include mutations of two directions that produce pathogenic or non-pathogenic variants, the mutation rate at the locus should be greater than the rate of mutations that produce pathogenic variants at the locus. Thus we have: Multifactorial threshold model We assume the multifactorial threshold model, in which quantitative traits such as liability to the disease are determined by multiple genetic and non-genetic factors including a stochastic and/or an epigenetic effect. Under this assumption, the relative fitness as a quantitative trait in the affected population is determined by multiple factors and approximately follows a gamma distribution with a mean(1-s). (s is the selection coefficient of SCZ; the mean relative fitness in the normal population is defined as unity.) The distribution curve of the fitness in the affected subpopulation with a SCZ-associated allele M never shifts to the right unless M has a strong protective effect (i.e. the effect of elevating reproductive fitness by reducing severity of and liability to the disease). Therefore, we can assume that s M , the selection coefficient in the affected subpopulation with a SCZ-associated allele M, is not smaller than s s≤s M < 1 ð Þfor a susceptibility allele. The inequality s>s M implies that M is a protective gene that reduces the severity and the risk of the disease.
No special assumptions are required on the allelic structure in each locus, or penetrance of each susceptibility gene, and possible interactions among the loci. Now we proceed to deduce the persistence criteria. From the assumptions, m′ G , the population frequency of the SCZ-associated allele M in the next generation, is given by: Therefore, from (A1), the reduction of the population frequency of the SCZ-associated allele M per generation is From (A2) and (A3) we have Since s M 1Às M p is monotonically increasing for s M 0 < s M ð < 1Þ and s≤s M < 1 holds for susceptibility allele M, we have Thus we have the criterion for a susceptibility gene: where ν is defined as ν 1Àsp From the observation (A5), we can see that d≥v implies s>s M for any SCZ-associated variant M which is sustained by mutation-selection balance.
Appendix B. Mutation rates and parameterestimations for SCZ
Mutation rates on autosomes and the X chromosome almost always fall within the range between 10 -6 and 10 -4 per locus per generation (usually < 10 -5 ; one generation = 20 years) [13,14] and can be approximated by a linear function of the parental age at least under 30 years for maternal age and under 40 years for paternal age [160]. Large-sampled cohort studies in Israel, Sweden and Denmark show that the mean age of parents in the general population is~28 years for mothers and 31 years for fathers; the mean age of both parents is < 29.6 years [83,86]. Therefore we can assume: 10 À6 < μ < 29:6 20 Â 10 À4 ¼ 1:48 Â 10 À4 .
Appendix C. Estimation of the required sample size of a GWAS
Concerning the required sample size 2 N (N case-control pairs) and the power 1-β of an association study, we have the well-established formula [161]: , where x (population frequency of the allele), d (case-control difference of allele frequencies) and γ 2 are defined as: x 1 2 m A þ m U ð Þ, d m A À m U , and γ 2 m A 1 À m A ð Þþm U 1 À m U ð Þ¼ 2x 1 À x ð ÞÀ 1 2 d 2 .
For the average mutation rateμ ¼ 1:48 Â 10 À5 , we have ν ¼ 1:76 Â 10 À3 . Supposing 0:0005 < x < 0:9995 , we have 2x 1 À x ð Þ> 0:9995 Â 10 À3 . From the persistence criterion, we have: 1 2 d 2 < 1 2 ν 2 < 1:6 Â 10 À6 < 2x 1 À x ð ÞÂ0:002. Therefore, we have the following approximation with an error smaller than 0.2%: 10 6 for x=0.5. Therefore, more than 2.9 million case-control pairs are required in a genome-wide association study with a power of 0.8 to detect a susceptibility variant of the average mutation rate and a population frequency between 0.0005 and 0.9995. Similarly we can see that more than 29,000 case-control pairs are required in a genome-wide association study with a power of 0.8 to detect a susceptibility variant of the highest mutation rate μ ¼ 1:48 Â 10 À4 ð Þ and a population frequency between 0.005 and 0.995.
Finally, let us consider the case with a relatively low mutation rate μ ¼ 1:48 Â 10 À6 , which corresponds to ν ¼ 1:76 Â 10 À4 . In this case, more than 290 million case-control pairs are required in a GWAS with a power of 0.8 to detect a susceptibility variant of a population frequency between 0.000005 and 0.999995.
|
2017-07-09T09:19:20.336Z
|
2012-05-31T00:00:00.000
|
{
"year": 2012,
"sha1": "66ecbf81d48553cf105c0aea02c32e22cc265ab8",
"oa_license": "CCBY",
"oa_url": "https://behavioralandbrainfunctions.biomedcentral.com/track/pdf/10.1186/1744-9081-8-28",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "66ecbf81d48553cf105c0aea02c32e22cc265ab8",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Biology",
"Medicine"
]
}
|
14783202
|
pes2o/s2orc
|
v3-fos-license
|
Modulation of nociceptive threshold by combined hormonal contraceptives in women with oestrogen-withdrawal migraine attacks: a pilot study
Background Menstrually-related headache and headaches associated with oestrogen withdrawal are common conditions, whose pathophysiology has not been completely elucidated. In this study we evaluated the influence of combined hormonal contraceptives (CHC) on pain threshold in women presenting migraine attacks during hormone-free interval. Findings Eleven women with migraine attacks recurring exclusively during the oestrogen-withdrawal period were studied with the nociceptive flexion reflex, a neurophysiological assessment of the pain control systems, during the third week of active treatment and during the hormone-free interval. During the hormone-free interval, nociceptive withdrawal reflex threshold was significantly lower (12.8 ± 8.0 mA) as compared to the third week of hormonal treatment (15.6 ± 6.6 mA) (p = 0.02). No change was observed in the pain perceived and in the temporal summation. Conclusions Oestrogen withdrawal may mediate an increased sensitivity to somatosensory stimuli in women with migraine attacks recurring during the hormone-free interval.
Introduction
Migraine is a common and disabling condition mostly affecting women during fertile age; reproductive milestones, as well as hormonal fluctuations across menstrual cycle and exogenous hormonal interventions hugely influence migraine clinical course in women. In observational studies, up to 60 % of migrainous women reported an association between menstruation and migraine attacks, often described as more severe, longer lasting and more resistant to treatment than non-menstrual ones [1]. Since a long time, oestrogen fall during late luteal phase has been implicated as a major event triggering migraine attacks [2]; oestrogen withdrawal has also been reported to induce headache in non migrainous patients [3]; this clinical entity has been codified in the International Headache Disorders Classifications (ICHD) as oestrogen-withdrawal hedache [4].
Events following oestrogens drop and predisposing to headache remain to be fully elucidated. Among the others, modulation of pain control system by sex hormones has been suggested to play a role: in a previous study, we assessed pain threshold in healthy women during the follicular and luteal phases of spontaneous cycles by investigating the threshold for nociceptive withdrawal reflex triggered by electrical stimulation of the sural nerve [5]. We observed that the pain threshold was significantly lower during the luteal phase and this finding was associated with an increased perception of physical disturbances, as assessed through the menstrual distress questionnaire. This is likely to result from a complex interaction between neurotransmitters and sex steroids, whose fluctuations however cannot be easily predicted in spontaneously menstruating women. That being so, combined hormonal contraception (CHC) represents a good model to study the "menstrual window" of vulnerability, because during the hormone-free interval the fluctuations of exogenous hormones are sizeable and predictable. Even though modern hormonal contraception encompasses multiple choices, the most commonly used is "the pill" in which the hormone free interval lasts seven days (21/7 regimens) and CHCs are discontinued to obtain withdrawal menstrual bleeding following 21 days of use. The same is true for the patch and the ring, which contain as well a combination of estrogen and progestogen [6]. Because of the steep decline of ethinyl oestradiol levels, ranging usually from 15 to 30 μg, the so-called oestrogen withdrawal symptoms may occur and likely prevented by extended CHC regimens to avoid the hormone-free interval [7].
In order to further explore the relation between oestrogen-withdrawal and headache, in this study we evaluated changes in the nociceptive withdrawal reflex threshold in women using CHC and reporting migraine attacks consistently and exclusively in the hormone-free interval.
Methods
We enrolled in a pilot study 11 women who were using CHC with a 21/7 pattern and who reported migraine attacks that occurred exclusively during the hormone-free interval of CHC in the previous 3 months. Patients fulfilling the inclusion criteria were consecutively enrolled among those attending the outpatient clinic of the Headache Science Center of the "C. Mondino" National Neurological Institute. The CHC formulation was oral (20 or 30 mcg ethynil estradiol pill) in 5 patients, transdermal (20 mcg ethynil estradiol patch) in 2 patients and vaginal (15 mcg ethynil estradiol ring) in 4 patients. All the patients followed the same therapeutic cycle with 21 days of active treatment followed by a seven-day break (the hormone-free interval) during which withdrawal (scheduled) bleeding occurred.
Their mean age was 30.2 ± 8.4 years (range 19-42 years). An exploratory analysis was also performed according to the onset of headache with respect to CHC treatment: in 5 women (mean age 30.6 ± 8.2 years) a headache that fulfilled criteria for migraine without aura was present prior to the exposure to CHC, while in the other 6 women the onset of headache occurred in strict temporal association with CHC use. It is noteworthy that according to ICHD-III criteria beta version, patients in the first subset received a double diagnosis (1.1 migraine without aura + 8.3.3 Oestrogen withdrawal headache), while patients in the second subset were diagnosed as 8.3.3 Oestrogen withdrawal headache.
Each subject was assessed twice across the same CHC cycle: during the third week of active treatment (T0) and during the hormone free interval (T1); all women were headache-free at the time of testing. Nociceptive withdrawal reflex was evaluated at the lower limb level, through electrical stimuli delivered percutaneously at the sural nerve; the muscular response was recorded from the biceps femoris using an electromyografic technique. The reflex threshold following a single stimulus (RT-SS) was considered as the lowest current intensity needed to elicit a stable electromyografic response (mA) (see ref. [8]). The area under the curve was also calculated (μV*ms), as an estimate of the motor units recruited upon reflex trigger. Subjective pain intensity perceived by the patient during the delivery of stimuli with an intensity corresponding to RT-SS was rated by participants on an 11-point scale (Visual Analogue Scale, VAS), ranging from 0 = no pain to 10 = unbearable pain.
Temporal summation of the nociceptive flexion reflex (RT-TS), which represents a reliable objective measure of the functional activity of the nociceptive system in the spinal cord, was also evaluated using train of 5 stimuli at a frequency of 2 Hz. It was defined as the lower stimulation intensity (mA) generating a clear facilitation of the reflex response size (greater than 20 µV for 10 ms or more) in the fourth and fifth trace during the course of the five individual pulses train (for details see ref. [9]).
Findings
During the third week of CHC (T0), the mean RT-SS was 15.6 ± 6.6 mA, with an average area under the EMG track of 1287.9 ± 738.1 μV*mS, while during hormone free interval (T1), mean RT-SS was found to be significantly lower at 12.8 ± 8.0 mA (T1 vs T0, p = 0.02) (Fig. 1), with a tendency toward an increase in the reflex area, which however did not reach a statistically significant level (p = 0.158). No change was observed in terms of subjective pain perception at the reflex threshold as reported by the patient through VAS (RT-VAS at T0 = 5.41 ± 2.08; RT-VAS at T1 = 5.41 ± 2.02; p = 1.00). No differences were seen as regards temporal summation recorded at T0 and at T1.
As far as the exploratory analysis is concerned, the drop in reflex threshold at T1 with respect to T0 was similar in the 2 subgroups of women: migraine onset before CHC Δ RT-SS = −18.90 ± 12.66; migraine onset after CHC Δ RT-SS = −16.18 ± 29.31) (Fig. 2).
Discussion
Our findings indicate that the hormone-free interval of CHC is associated to a reduction in the RT-SS of the lower limb nociceptive reflex in women with attacks of migraine induced by oestrogen-withdrawal. No change in subjective pain intensity was perceived at RT-SS from T0 to T1, meaning that during hormone free interval, participants rated an electrical stimulus at lower current intensity as painful as an electrical stimulus at higher current intensity necessary to elicit the nociceptive withdrawal reflex during the third week of active treatment. Thus, the objective lowering of neurophysiologically recorded RT-SS during hormone-free intervals was indeed coupled to an increase in the subjective self-reported pain sensitivity.
Although not statistically significant, likely due to the paucity of the study sample, we found a trend towards an increase in the area under the curve at RT-SS during the withdrawal week. This observation suggests that during the hormone-free interval more motor units can be recruited at RT-SS intensity. If this result is confirmed in It is noteworthy that we observed the same pattern of neurophysiological and subjective responses in two different subgroups from a nosographic point of view: 1) migraineurs whose pre-existing headache was markedly modified by CHC as regards the temporal pattern of occurrence of attacks and 2) healthy women whose headache is secondary to the use of CHC. Though obtained in a small sample, this finding prompts the hypothesis that the observed neurophysiological similarities between these 2 subgroups do reflect a shared pathophysiological alteration of the nociceptive network mediated by the drop in oestrogen levels. In this regard, it is noteworthy that the working methodology in this study was to enroll patients who shared the same migrainous phenotype of attacks together with the same temporal pattern of their occurrence.
We are aware that our study suffers some limitations mainly represented by the small sample of patients and the lack of a control group. We tried indeed to limit the impact of these biases with a strict selection of patients and with the choice of a validated neurophysiological methodology to assess pain threshold.
Moreover, the main purpose was to validate and extend our previous observation of a reduction of the RT-SS in healthy women during the luteal phase [5]. Taken together, these data suggest that the decrease in oestrogen levels makes women more 'receptive' to pain as a default response. The predisposition to develop headache or migraine attacks could be related to events downstream of the oestrogen drop and that remain mostly elusive. Previous studies attempted to investigate the influence of hormonal contraceptives on pain threshold, with conflicting results [10,11], probably because of methodological differences. Teepker et al. [10] in particular evaluated the pain thresholds of different types of stimuli (heat, cold, pressure and electrical) and found no difference between patients with or without CHC. In agreement with our results, other authors described fluctuations electrical pain thresholds along the menstrual cycle [11]. A recent study that evaluated the effect of hormonal contraception on pain sensitivity showed that users of progestin only contraceptive have higher pressure pain thresholds than CHC users or nonusers, thus suggesting that also progestin do play a role, besides oestrogens, in the modulation of pain control systems [12].
The association between migraine and oestrogens has been widely studied and recently Chai et al. wrote a comprehensive review on this issue [13]. In particular, data on the effect of exogenous sex hormones showed that oestrogens are able to modify headache patterns in patients with migraine and that the avoidance of a marked drop of oestrogen levels in the blood may beneficial in preventing migraine occurring during estrogen withdrawal.
Different hypotheses have been proposed to explain the association between migraine attacks and hormonal fluctuation. According to the "mismatch theory" [14], a sudden oestrogen drop would result in a transient disequilibrium between genomic and non-genomic oestrogen-mediated actions on neurotransmission, vascular reactivity, neuroinflammation and other systems, overall promoting neuronal activation. Among the others, oestrogen modulation of serotonergic system seems to play a major role in influencing pain control system: a previous study conducted in women presenting status migrainosus during the hormone free interval of CHC, showed a blunted cortisol and prolactin secretion, following a serotonergic challenge (M-phenylchloropiperazine); moreover, the status migrainosus could be prevented by transdermal estradiol supplementation against placebo [15].
Conclusions
In this pilot study we described a reduction of neurophysiologically recorded pain threshold in women with migraine/headache attacks during the hormone-free interval of CHC. The drop in oestrogen levels is likely to play a role in modulating nociception during the hormone-free interval. Further investigations in the neurobiological events leading to these attacks are necessary for understanding the biological aspects of CHCinduced attacks. The observation of the same pattern in women with or without a positive history of migraine could be important from both a nosographic and a therapeutic point of view.
Abbreviation CHC, combined hormonal contraceptives; RT-SS, nociceptive withdrawal reflex threshold elicited with a single stimulus; RT-TS, nociceptive withdrawal reflex threshold elicited with temporal summation; VAS, visual analogue scale
|
2017-08-02T21:20:15.929Z
|
2016-08-04T00:00:00.000
|
{
"year": 2016,
"sha1": "2aea412eb53b07d945b974c261657ee319039c3e",
"oa_license": "CCBY",
"oa_url": "https://thejournalofheadacheandpain.biomedcentral.com/track/pdf/10.1186/s10194-016-0661-6",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "2aea412eb53b07d945b974c261657ee319039c3e",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
204249370
|
pes2o/s2orc
|
v3-fos-license
|
First record of Tarbinskiellus terrificus (Walker, 1869) (Grylloidea: Gryllidae) from Central India
The family Gryllidae is one of the family of super family Grylloidea. The family Gryllidae commonly known as crickets, body size small to large (1 or 2 to 50 mmAlexander, 1968) can be easily distinguished from other ensiferan families by following characters: antenaae filiform, usually long, ocelli presents or absent, auditory organs situated in the anterior tibiae, male elytra usually with stridulatory organs, jumping hind legs, tarsal formulls 3-3-3, and long tactual cerci bearing clumps of knobbed hairs, ovipositor usually long. The subfamily Gryllinae Laicharting, 1781 is one of the 13 subfamilies of the family Gryllidae Laicharting, 1781. About 7 tribe and 117 genera are globally known under the subfamily Gryllinae (Eades et al., 2015), of which 29 genera are presently known from India (Shishodia et al., 2010). Serville, 1839 erected genus Brachytrupes for the species Brachytrupes membranaceus membranaceus. After this Chopard, 1969 reported Brachytrypes genus for the species B. portentosus (Lichtenstein, 1796), Tamil Nadu: Coromondal coast (Type), from Bihar, Uttarakhand: Dehradun, West Bengal: Darjeeling, Siliguri; Assam: Kalligunga, Cachar, Sibsagar, Garo hills; Karnataka: Mysore; Tamil Nadu: Chennai; B. orientalis (Fabricius, 1775) from India: West Bengal: Kolkata; Assam: Sibsagar, Ledo; Orissa: Balighai; Tamil Nadu: Trichinopoly, Negapatam; Karnataka: Mysore; B. terrificus Walker, 1869 from India: Tamil Nadu: Chennai (Type specimen in British Museum). Later on Gorochov, 1983 erected Tarbinskiellus genus and placed all three species in the genus Tarbinskiellus Gorochov, 2001 studied Tarbinskiellus genus and describe one new species T. neotropicus sp. nov. from Brazil and also studied T. portentosus (Lichtenstein, 1796) from India: Assam and provide results T. portentosus is similar also to T. terrificus while T. portentosus species differs from strongly reduced ovipositor (a female type series of T. terrificus from BMNH is designated as lectotype to fix the application of the name and examined type specimen Brachytrupes terrificus and one of Walker series so named Brachytrupes terrificus Walker, syntype (det. B.C. Townsend). Till date genus Tarbinskiellus have 4 species e.g. T. neotropicus, T. orientalis, T. portentosus, T. terrificus are globally known Eades et al., 2015, out of which three species reported from India Shishodia, et al., 2010.
Introduction
from India: Tamil Nadu: Chennai (Type specimen in British Museum).
Later on Gorochov, 1983 erected Tarbinskiellus genus and placed all three species in the genus Tarbinskiellus Gorochov, 2001 studied Tarbinskiellus genus and describe one new species T. neotropicus sp. nov. from Brazil and also studied T. portentosus (Lichtenstein, 1796) from India: Assam and provide results T. portentosus is similar also to T. terrificus while T. portentosus species differs from strongly reduced ovipositor (a female type series of T. terrificus from BMNH is designated as lectotype to fix the application of the name and examined type specimen Brachytrupes terrificus and one of Walker series so named Brachytrupes terrificus Walker, syntype (det. B.C. Townsend). Till date genus Tarbinskiellus have 4 species e.g. T. neotropicus, T. orientalis, T. portentosus, T. terrificus are globally known Eades et al., 2015, out of which three species reported from India Shishodia, et al., 2010.
Study Area
The survey was carried out in Chhattisgarh state, covering two protected areas. Barnawapara wildlife sanctuary is situated in Raipur, covering an area of about 244.66 sq. km., lies between latitudes 21°18ʹ45˝to 21°30ʹN and longitudes 88°22ʹ30˝to 82°37ʹ30˝E. Udanti wildlife sanctuary, is situated in Gariyaband district of Chhattisgarh state, Sunil Kumar Gupta which lies between 20°0ʹN to 20°15ʹN, 82°30ʹE to 82°0ʹE with an area of about 247.80 sq. km.
The specimens were collected by light trap, using usually white sheet or mulmul cloth, hung between the two trees or pole, with the light placed using mercury bulb with Honda generator, or gas in front of it at night collection at Barnawapara camp near Barnawapara village and Udanti Wildlife Sanctuary near Ganga Jamuna village and killed by keeping in a killing bottle containing benzene vapour in killing jar. The specimens were preserved dry. The genitalia of male specimen was dissected out, in 10% Potassium hydroxide for 2 to 12 hours and then preserved in 90 % alcohol. Live image of specimen was capture by using Nikon D 300S with 105 macro lens. The specimen was studied under Leica stereozoom Microscope (Leica M205 A) and photographs were taken using the software Leica Application Suite (LAS V3.8) and deposited in National Zoological Collections of Zoological Survey of India, Kolkata.
Head: Head as wide as pronotum with dense bristles ( Fig 1C). Vertex flat, convex. Eyes large protruding.
Ocelli locate in triangle. Median ocelli more than twice as wide as lateral ocelli. Distance between lateralocelli slightly longer than the distance between one lateral and the median ocelli. Maxillary palpi large, joints 1 and 2 subequal; joint 3 the longest, joint 5 slightly short and widened from base. Pronotum: Pronotum 1.6 times wider than the length in dorsal view ( Figure 1D), transverse; anterior margin concave, posterior margin slightly straight and wider than anterior margin; anterior margin with dense bristles, lateral lobe of pronotum ventral margin straight. Mesosternum rectangular; metasternum broader than mesosternum, pentagonal form.
Legs: Fore legs: Clothed with bristles; femora 2.9 times as long as wide, cylinder, tibiae with a large elongated external tympanum and internal small, ovoid tympanum ( Figure 2A). 1 outer apical spur, short, 2 inner apical spurs, of which one long and one some short, wide at base, acute at apex. III tarsal segment slightly longer than first. Mid legs: Clothed with bristles but shorter than fore legs, cylinder; femora 4 times as long as wide, tibiae with 4 apical spurs, outer one spur 1.4 times longer than dorsal and inner spurs (Figure 2 C); dorsal and inner spurs slightly subequal. III tarsal segments 1.3 times longer than first. Hind legs: Femora 3.4 times as long as wide, and 2.4 times longer than the middle femora; basal part dilated; dorsal and ventral margin slightly hairy; dorsal and median external area glabrous; tibiae clothed with bristles but shorter than fore legs, with 5 outer subapical spurs and 4 inner subapical ( Figure 2D); except for the 4 apical spurs; outer IInd spurs longest, stout, tip acute and straight, Ist some curved and tip more acute, inner spurs subequal together; basitarsus 1.2 times longer than the combining length of the remaining two apical tarsomeres; furrowed dorsally; with outer margin 3 denticles and inner margin 5 denticles or spines; 2 apical spurs of basitarsus, inner apical basitarus spur 2.7 times longer than outer basitarsus spur, straight, stout and acut tip ( Figure 2F). Abdomen: Abdomen 1.4 times shorter than the hind femur. Subgenital plate with dense bristles, setose; basal part wide and apically triangulate; 1. 2 shorter than its width acute ( Figure 2B). Cerci male 1.4 times longer than the hind femur ( Figure 3D). Ovipositor short, straight ( Figure 2G). Male Genitalia: Ectoparamere large. Virga slightly large and stout. Ectoparameres process notch like, tip acute, large ( Figure 3C
Conclusion
The result of this study is based on the examination of 2 specimens belonging to genus Tarbinskiellus (Gryllidae: Gryllinae). Both the specimens were collected from Barnawapara Wildlife Sanctuary (district Raipur) and Udanti Wildlife Sanctuary (district Gariaband). This species has been not reported except Walker, 1869 from, Chennai, Tamil Nadu (type specimen) later Chopard 1969 recorded from Karnataka (Mysore: Bhadravati) India. Since then, there were no reports on the occurrence of this species in India except Tamil Nadu and Karnataka. This species can be distinguished from related species by more robust and stout body, size more large, contrasting colour of elytra, pronotum and legs feebly pubescent, pronotum slightly widening, and female with ovipositor reduced.
The specimens were collected during night in the months of April and June respectively. The field images along with the illustrations of male genitalia and stridulatory files are provided for the first time with the distribution map of the species.
|
2018-12-15T10:21:55.170Z
|
2017-08-01T00:00:00.000
|
{
"year": 2017,
"sha1": "994934061af3fe7da5536c8c781cb738674eba4c",
"oa_license": "CCBY",
"oa_url": "http://52.172.159.94/index.php/zsoi/article/download/117281/81605",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "994934061af3fe7da5536c8c781cb738674eba4c",
"s2fieldsofstudy": [
"Biology",
"Environmental Science"
],
"extfieldsofstudy": [
"Geography"
]
}
|
53039815
|
pes2o/s2orc
|
v3-fos-license
|
Some Challenges of Specifying Concurrent Program Components
The purpose of this paper is to address some of the challenges of formally specifying components of shared-memory concurrent programs. The focus is to provide an abstract specification of a component that is suitable for use both by clients of the component and as a starting point for refinement to an implementation of the component. We present some approaches to devising specifications, investigating different forms suitable for different contexts. We examine handling atomicity of access to data structures, blocking operations and progress properties, and transactional operations that may fail and need to be retried.
Introduction
The objective of this paper is to address challenges to do with specifying concurrent program components in order to promote discussion about alternative approaches. Our main foci are atomicity, blocking operations and transactional operations in the context of rely/guarantee specifications [13,14]. Our aim is to present the ideas rather than a fully formal development. Specifications play an important role in decoupling the use of a component from its detailed implementation. Often the role of specifications as a starting point for refinement to an implementation is emphasised but here we would like to balance that with their role of being used by other components. Hence we focus on a top-down approach to concurrent program specification, rather than a bottom-up approach.
Sequential programs. For sequential programs conventional, Floyd/Hoare-style specifications [3,9] in terms of preconditions and postconditions form the basis of component specifications, however, pre and post conditions alone are inadequate for specifying concurrent operations because they do not handle interference between the operations.
Shared variable concurrency. First, to state the obvious, variables that are local to a thread are not subject to interference and hence can be treated in a manner similar to their use in a sequential program. For variables shared between parallel threads, interference becomes an issue. An important consideration is whether access (e.g. read or write) to variables is atomic or not. At the lowest level, atomicity is determined by the machine hardware and properties like its atomic access "word" size. A further complication at the hardware level is that, due to caches and write buffers, the order of write/read accesses to shared memory may not respect the sequential order of instruction execution. But perhaps we get a bit ahead of ourselves if we worry about these issues when considering specifications.
A thread that needs to perform multiple atomic accesses can be subject to data races where variables are updated in parallel by concurrent threads and the thread may see inconsistent data. At a more abstract level, operations may be required to be atomic with respect to a shared data structure. The implementation of such operations may require locks to ensure sequentialisation of access to the data structure or it may use more sophisticated non-blocking algorithms that achieve the effect of operation atomicity by utilising hardware-level atomicity [22].
A concept commonly used to show an implementation is valid is linearizability [8], whereby parallel execution of a set of operations on a shared data structure is considered valid if it is equivalent to some linear (sequential) execution of the same operations (subject to certain requirements).
Verifying concurrent programs. Early rules for reasoning about parallel programs by Hoare [11] utilised preconditions and postconditions but had strict disjointness requirements on program variables occurring in parallel threads, which effectively ruled out interference between parallel threads. The approach of Owicki and Gries [19,20,21] treats parallel components like sequential programs with intermediate assertions between each atomic step but then requires an extensive interference-freedom proof. Concurrent separation logic [1,18] also leverages disjointness but does so in a more fine-grained and dynamic manner because it handles assertions over objects in the heap.
An early compositional approach to handle interleaved interference on shared variables was the rely/guarantee approach of Jones [12,13,14]. Jones extended pre/post specifications with a rely condition, a binary relation between program states expressing an assumption about the tolerable interference that any step of the environment of the thread can impose on the shared variables. To constrain the interference generated by a thread, Jones uses a guarantee condition, also a binary relation on states limiting the changes a program step of a thread can make to the shared variables. The guarantee is required to be reflexive (i.e. it contains the identity relation) so that the program may make stuttering steps that do not change the observable variables. The rely of each thread must be implied by the guarantees of all threads that run in parallel with it. The rely/guarantee approach does not dictate any particular granularity of atomicity, however, it does require all steps of a thread to satisfy its guarantee as long as all steps of the thread's environment satisfy its rely.
Blocking. Operations may block waiting for "communication" from another thread. For example, an operation wanting to read a message from a communication channel may need to wait for a message to be written to the channel by another thread. Specifications of such operations need to be able to express such waiting criteria. This affects the termination behaviour of the operation. For example, if a message is never written to a channel, a read from the channel will never terminate (block forever). We consider two approaches to specifying such operations.
• Using an explicit await construct that allows both nonterminating behaviour if the await blocks forever and terminating behaviour if it becomes unblocked.
• An implicit approach that specifies under what conditions an operation is guaranteed to terminate as well as its behaviour when it does terminate, e.g. a read on a message channel is guaranteed to terminate if the channel is non-empty.
These forms can give equivalent specifications, for which the explicit form may be more useful for refining the operation, while the implicit form makes it easier to reason about using the operation. Another source of blocking is at the implementation level where atomicity constraints on operations may lead to the use of locks that lead to an implementation blocking awaiting a lock.
Transactional operations. Utilising locks can generate bottlenecks because operations requiring the locks are sequentialised, and on multi-processor architectures they also generate more costly memory synchronisation primitives. One approach to avoiding (or minimising) locks is to implement operations that do most of their work locally and then perform a final atomic commit step that may fail if another operation has committed while the first operation was executing based on the old data [22]. Such implementations may suffer from starvation where they can repeatedly try and fail, potentially forever if there is continual interference from competing parallel operations. Note that in these situations, one of the competing threads may succeed but an individual thread may be pre-empted every time and never succeed. Fair scheduling is assumed, i.e. every thread is executed eventually, but that does not prevent an operation that a thread is attempting to execute from failing due to interference and needing to be retried. Specifying such operations has to allow for the case in which an operation is continually thwarted and may never terminate, while guaranteeing termination if the interference on the data structure eventually quiesces. Using c ω to represent the execution of a command c zero or more times, including possibly infinitely many times, such specifications have the general form where fail represents the operation failing due to interference (and not changing the state) and succeed represents the operation successfully completing (and updating the state once). The iteration fail ω may execute fail infinitely many times representing the continual thwarting by interference from parallel operations. Arguments about termination usually need to resort to either • arguments that interference eventually quiesces altogether, • timing arguments based on minimal separation between operations in any single thread leading to a situation in which interference will eventually quiesce for long enough for the operation to succeed, and • arguments based on probabilities of two (or more) operations overlapping and competing.
Note that probability bounds can be derived from timing bounds. The probability can be sensitive to load, i.e. the more threads competing, the lower the probability of success of any single operation. And probabilities can also be sensitive to the execution time of an operation: the longer it executes, the more likely it is to overlap with a competing operation. Of course, such timing and probability arguments depend on the context of the use of the data structure and can be tricky in practice. Implicit specifications can also be used for such operations by specifying the conditions under which they are guaranteed to terminate.
Overview. Sect. 2 addresses specifying atomic operations on a shared data structure (or resource). It examines the use of Hoare's with statement [10] and how it interacts with rely and guarantee conditions. Sect. 3 examines blocking operations giving both explicit waiting conditions and more implicit specifications using a temporal logic formula under which an operation terminates. Sect. 4 looks at transactional operations that may either succeed, or try and fail, possibly indefinitely. Both explicit waiting and implicit temporal logic specifications are considered.
Specifying atomicity
As an example, consider a message queue with operations to enqueue and dequeue messages. If there are separate concurrent threads enqueuing and dequeuing, each operation needs to (appear to be) be atomic, i.e. other operations cannot observe the state part way through the operation. If the queue were used in a purely sequential program, the write and read operations can be specified by Morgan-style specification commands [16,17] as follows.
The write operation takes a value v to append to the queue. Its postcondition is qu ′ = qu ⌢ [v], in which qu ′ stands for the final value of the queue, qu stands for the initial value, " ⌢ " is sequence concatenation, and [v] is the singleton sequence containing v. The write operation modifies only the queue and hence it has a frame of qu (before the colon). The read operation returns a value res that is the head of the queue and removes it from the queue in the process. It has a precondition that the queue is non-empty, which is written as an assertion command (in braces).
To extend the operation specification to handle concurrency, as presented in Fig. 1, the operations need to be augmented with rely and guarantee conditions and the atomicity of the operations needs to be handled. In Fig. 1 the relies and guarantees are represented as rely and guarantee commands [2,5,7]. The guarantee command (guar g) restricts every atomic program step of the thread to satisfy g. The rely command (rely r) represents an assumption that every environment step satisfies r; if an execution trace performs an environment step not satisfying r, any behaviour whatsoever is allowed from that point (i.e. it aborts in a manner similar to the precondition command {p} aborting if the initial state does not satisfy p). The rely and guarantee commands are combined with the remainder of the specification using the weak conjunction operator "⋓" [6,4]. The weak conjunction c 0 ⋓ c 1 performs steps allowed by both c 0 and c 1 unless either c 0 or c 1 aborts at some point, in which case their weak conjunction aborts from that point. Weak conjunction is a specification operator, rather than a programming operator.
For the message queue we assume there is a single writer thread performing write operations and a single reader thread performing read operations. A suitable rely condition for write is that elements are only ever removed from the front of the queue, i.e. the queue after any interference is a suffix of the queue before. The rely condition for read is that the interference only ever adds elements to the end of the queue and hence the queue before the interference is a prefix of the queue after. Note that the precondition of read is stable under its rely condition [24,25] The guarantees of each operation match the rely of the other operation. This version requires that there is just one reader thread and one writer thread because the rely condition of read assumes the queue can only be extended and hence it is not concurrently being dequeued by another thread, and the rely condition of write assumes the queue can only become a suffix of its previous state and hence it is not concurrently being enqueued by another thread.
For sequential programs and in the original rely/guarantee approach the postconditions of operations are considered end-to-end, i.e. they must hold between the states at the start and end of an operation invocation. Such a postcondition is problematic in the context of a parallel thread modifying the queue, for example, after an write operation is initiated but before it can complete (or lock the data structure), the reader thread may read a value. If the writer then completes the write without further interference, the end-to-end effect is that of both the read and the write, not just the write. An end-to-end postcondition is not suitable in this case. The alternative is a specification whereby the postcondition holds for some "atomic" step between the start and end of the operation and the operation makes no changes to the queue before or after that step, although other threads may. Atomicity can be handled by using resources (see Sect. 2.1). The environment may also interfere with operation preconditions. For example, if multiple readers were allowed on a queue, a precondition that the queue is non-empty may be invalidated by a parallel read operation that reads the last value from the queue thus making it empty.
Resources
Early work of Hoare [10] introduced the idea of a resource and a with statement that provides access to the resource. A resource represents a shared data structure that is only accessible to a thread within a command of the form, that ensures the resource d is not modified by the environment while the thread is executing c. A data structure d is declared as a resource by a declaration of the form resource d and, within the scope of the resource declaration, all uses of d must be within a with d do ... od statement. 1 It is required that the data structure of the resource is only accessed within with statements; this may be checked syntactically. The implementation is responsible for ensuring the data structure is not modified by the environment while executing within the with statement. The with statement allows any number of stuttering steps before the body and finite stuttering after the body of the with is executed. As entry to a with statement by one thread may block other threads wishing to gain access to the same resource, it is prudent to require that the bodies of with statements terminate; that is required within this paper. The data structure of the resource also often has a data-type invariant associated with it that is established by its initialisation and maintained by each operation (see the example in Section 3). An operation whose body consists of a single with statement executes the with statement atomically at some point of time between its invocation and completion. Hence multiple operations that overlap in time will have their bodies executed in some sequence (i.e. their execution is linearized [8]) and hence they will appear to behave equivalently to some sequential execution of the same operations, i.e. their execution is sequentially consistent.
Rely/guarantee laws for resource access
The concept of a resource may be combined with the rely/guarantee approach. When a thread enters the body of a with d do c od statement, the rely can be strengthened for the duration of c with d ′ = d and the guarantee weakened so that d only need satisfy the guarantee over the complete operation, not over every step. The initial step of the refinement of operations specified via a with statement needs to "move" the rely and guarantee conditions into the body of the with but in the process the rely and guarantee conditions are transformed.
For the write operation above (ignoring the guarantee for the moment) this law can be applied as follows.
(rely qu ′ suffixof qu) ⋓ with qu do qu: A guarantee condition surrounding a with may be weakened so that it only has to apply for the resource data structure over the body of the with command. It is assumed that the guarantee is of the form g d ∧ g x , where d is the only shared variable g d refers to, and g x does not refer to d. The weakened guarantee g x is retained to handle references to variables other than d within the guarantee. The specification g d requires g d to hold end-to-end over the body of the with. The lack of a frame allows any variables to be modified but when it is combined with c, any frame of c will apply to their weak conjunction.
For the write operation above (ignoring the rely for the moment) this law can be applied as follows.
(rely qu ′ suffixof qu) ⋓ (guar qu prefixof qu ′ ) ⋓ with qu do qu: qu ′ = qu ⌢ [v] od ⊑ using Law 2 (guar-with) and Law 1 (rely-with) with qu do (rely qu ′ = qu) ⋓ qu: Note that in the body of the with there is no explicit guarantee and the rely condition assumes that qu is not modified by the environment and hence the refinement of the body of the with is effectively a sequential refinement (as one would expect).
Blocking using an explicit await condition
Consider the example in Fig. 2 of a message queue with a bounded capacity of N messages. It has a write operation that waits until the queue is not full and appends a value to the tail of the queue, and a read
Blocking using temporal logic termination conditions
The specifications of read and write in Fig. 2, by including await statements, allow non-terminating behaviour in the cases where the await condition never becomes true, e.g. a read will wait forever if the queue remains empty because no writes are performed. However, read terminates if the queue is eventually non-empty and write terminates if the queue is eventually non-full. That leads to an alternative specification using temporal logic termination conditions in Fig. 3. To accommodate the termination conditions, a command of the form, is introduced, in which the command term only allows any terminating behaviour and for a temporal logic formula t, the command encode t allows just those traces that satisfy t. 2 If the temporal logic formula t holds, terminate t must terminate, but if t does not hold termination is not required, but is allowed. Neither of the specifications of write and read in Fig. 3, contain the explicit awaits used in More subtly, any finite prefix of a trace of a read operation for which qu = [ ] in every state cannot have satisfied the postcondition of read and hence cannot have terminated. However, it is still possible that another thread may execute a write at some later time establishing qu = [ ] and allowing the read to terminate. For the finite prefix of the trace, the behaviour of the read must correspond to the stuttering allowed by the with statement before it enters its body. Hence for every finite prefix of a trace of read in which qu = [ ] in every state, the operation performs only finite stuttering steps and, further, nontermination is allowed if (qu = [ ]). Hence the behaviours allowed by the specifications in Fig. 3 are actually equivalent to those of the specifications in Fig. 2.
For the queue example, if multiple readers and writers are allowed, the specifications need to be adapted. Firstly, the rely conditions of the write and read operations need to be removed because the rely of write is broken by a concurrent thread also performing writes, and similarly for the rely of read if there are concurrent readers. Secondly, the termination conditions need to be strengthened. There are two strengthenings that correspond to weakly and strongly fair interpretations of the await in specifications using await.
For weak fairness the termination condition of read is strengthened to ♦ (qu = [ ]), that is, eventually the queue remains non-empty, which implies that concurrent threads are not making the queue empty through read operations. For the equivalent specification using an await rather than a terminate, , that is, it is always the case that eventually the queue is non-empty, which implies that concurrent threads can repeatedly make the queue empty through read operations but if the queue becomes empty there must then be one of more write operations that make it non-empty. Because the termination condition is weaker, this version of read must terminate for a wider range of behaviours of its environment and hence requires a stronger implementation, e.g. one that strictly sequences access to the queue, perhaps using a ticket lock.
For the equivalent specification using an await rather than a terminate, the await can be given a strongly fair interpretation, i.e. if the guard of the await for read, qu = [ ], is always eventually enabled (but the queue may alternate between empty and non-empty), the await will succeed.
An advantage of the specifications using terminate clauses is that the condition under which an operation is guaranteed to terminate is made explicit, unlike specifications using await statements where the semantics of the await needs to be changed. This makes it clearer what a programmer using a component can rely on. It also avoids the complication of having different interpretations of await guards as used by Liang and Feng [15].
Transactional operations
Some implementations of operations are optimistic in that they complete most of the operation locally within a thread and then have a final commit phase that may fail if another operation has committed. Such operations consist of a repeated failure behaviour (that does not change the shared data structure) in the presence of interference followed by a successful commit phase. Of course, in the presence of repeated interference the successful commit may never occur. There are many examples of lock-free algorithms that exhibit the above behaviour [22] but to make the discussion concrete, we make use of Treiber's well-known concurrent stack data structure [23].
Specification using explicit failure
Treiber [23] provided a non-blocking lock-free implementation of a stack in which the push and pop operations may be concurrently executed by many threads. However, an attempt to push or pop a value In the worse case, if every time an operation is tried it fails due to interference, that invocation of operation may never terminate. Fig. 4 gives a specification of a stack with push and pop operations that may fail and need to retry, possibly indefinitely. If the stack is empty, pop returns the special value null, which may not be pushed onto the stack. 3 The push fail operation may be executed any number of times but each time it is executed the environment makes a step that changes the stack s. If from some point of time the environment never changes s, then the push fail becomes infeasible and the operation must perform push success which pushes the value on the stack and terminates; the command term only allows any terminating computation and hence when conjoined with a specification restricts it to just its terminating behaviours. The command ε s ′ = s corresponds to the environment performing a step that modifies s; it may also perform a finite number of stuttering program steps (i.e. steps that do not change observable variables). If the environment performs a step modifying s, ε s ′ = s terminates but if not, it becomes infeasible and termination of the iteration is forced, so that the push success alternative is taken. The definition of pop is similar.
Specification using temporal logic termination conditions
Note that if the stack s is never changed by the environment, the behaviour of the push (or pop) operation reduces to just its successful behaviour. More subtly, if the environment eventually stops changing s, then there can only be a finite number of failure iterations before the operation succeeds. This latter condition can be converted into a temporal logic termination condition ♦ ε (s ′ = s), i.e. eventually all environment steps do not change the value of the stack, which leads to the specification given in Fig. 5. An extended form of temporal logic is used here that distinguishes program and environment steps and allows one to specify a constraint on a step in the form of a relation, in this case s ′ = s. If parallel activity on the stack eventually quiesces, Treiber's push and pop operations are guaranteed to terminate, and hence the specifications with the quiescence termination conditions do not need to include the failure possibilities.
The negation of the termination condition is ♦ ε (s ′ = s), i.e. from every state there is eventually an environment step that modifies s. However -unlike for the blocking queue-that does not make the postcondition of either operation unsatisfiable and hence if the negation of the termination condition holds, each operation may either terminate satisfying its postcondition or never terminate. In a similar manner to the blocking queue, for a finite trace for which an operation has not yet satisfied its postcondition, it is still possible to extend the trace so that the postcondition is satisfied, and hence the only allowable behaviour of the operation for a finite trace that has not yet satisfied its postcondition is finite stuttering. Hence if the termination condition is not satisfied an operation may either terminate successfully satisfying its postcondition or fail to terminate but only ever perform stuttering steps, i.e. it never modifies s. Hence the specifications in Fig. 5 are equivalent to those in Fig. 4.
If the termination conditions on the stack operations are replaced by true, the operations must always terminate, even under interference from other threads performing push and pop operations. That gives strictly stronger specifications because their termination conditions are weaker. An implementation might then be required to make use of a lock that sequentialises access to the stack in the order in which the lock is requested (such as a ticket lock) in order to ensure termination.
Conclusions
To specify concurrent program components one needs to be able to address issues such as operation atomicity, operations blocking on conditions or locks, and transactional operations that may fail and need to be retried. Hoare's resource concept provides a notion of atomicity with respect to a resource. A contribution of this paper is to examine its interaction with rely and guarantee conditions in order to enable the initial refinement step of Hoare's with statements to code. Brookes [1] also makes use of the concept of a resource in concurrent separation logic. He generalises Hoare's concept to handle the heap as well as variables.
The specifications of operations using with statements do not dictate whether they are refined to implementations using locks or to non-blocking implementations or even a programming language that supports with statements. One issue not addressed here is that operations requiring multiple resources, e.g. an operation that needs to perform operations on two separate resources and needs to be considered atomic as a whole. In this case the with statement needs to allow for multiple resources, and if locking is used in the implementation of the operations, to avoid deadlock, the locking has to ensure that resources are locked in the same order. Note that with statements may be nested but they do not allow the complete flexibility of locks. For example, algorithms that lock one node in a list and then lock the next node before releasing the first node do not match a nesting structure.
Operations that block waiting for some condition have the potential for non-terminating behaviour. That has been addressed via two ways of specifying such operations: a form that makes use of an await construct that blocks until its guard holds, and an implicit form that includes a condition under which termination is required.
Non-blocking algorithms can provide more efficient solutions for managing shared data structures than using locks, but some algorithms have the issue that, under interference, they may fail and need to be retried. In the worst case an operation may be continually thwarted and never get a chance to complete and hence its specification needs to either allow for that possible behaviour or provide conditions under which it will terminate successfully, e.g. that the interference quiesces so that it can complete.
The approach taken in this paper is to indicate some directions for devising specifications for concurrent program components. In doing so we have shown that specifications with explicit await clauses can be expressed in a more abstract form with a temporal logic formula giving the condition under which termination is guaranteed. Such specifications have greater expressive power than those using explicit await constructs because temporal logic formulae allow termination conditions that cannot be expressed as await conditions, for example, if the blocking queue allows multiple readers and hence its await condition was no longer stable, an alternative termination condition of ♦(qu = [ ]) would require an implementation to guarantee the termination of each read operation under interference from other reads, provided a writer was also actively appending values to the queue. An implementation of read might, for example, use a lock that sequentialises access in the order in which the lock is requested (such as a ticket lock) in order to ensure termination.
Liang and Feng [15] have addressed handling progress conditions for blocking operations (which they refer to as partial methods). Their approach makes use of await statements but they give four different semantic interpretations to await statements depending on whether one requires the operations to be starvation free or deadlock free, and depending on whether the enabling conditions are treated as weakly or strongly fair. The weakly and strongly fair interpretations correspond to different termination conditions for operations. The explicit use of a temporal logic termination condition differentiates these two cases while avoiding the issues involved with giving different semantic interpretations to the await construct.
At this stage our treatment has not been fully formalised and further work is required to support refinement of such specifications to code.
|
2018-10-23T00:47:50.000Z
|
2018-10-23T00:00:00.000
|
{
"year": 2018,
"sha1": "faad1015e583769a1964bb796e1b3c70874c974d",
"oa_license": "CCBYNCND",
"oa_url": "https://arxiv.org/pdf/1810.09611",
"oa_status": "GOLD",
"pdf_src": "Arxiv",
"pdf_hash": "9152014c510f745913840b0fdc6805a70662ac86",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science"
]
}
|
248879963
|
pes2o/s2orc
|
v3-fos-license
|
Structural alterations of branched versus linear mixed-surfactant micellar systems with the addition of a complex perfume mixture and dipropylene glycol as cosolvent
Personal care products commonly contain perfume mixtures, consisting of numerous perfume raw materials (PRMs), and cosolvents. The lipophilicity and structure of an individual PRM is known to affect its localization within the surfactant self-assembly as well as the micellar geometry. However, because multiple PRMs are used in formulations, significant intermolecular interactions between the PRMs and between the PRMs and the surfactant tail may also influence the location of the PRMs and their effects on the self-assembly. Herein, two anionic/zwitterionic mixed-surfactant systems (sodium trideceth-2 sulfate (ST2S)/cocamidopropyl betaine (CAPB) and sodium laureth-3 sulfate/CAPB) were formulated with a cosolvent (dipropylene glycol (DPG)) and 12 PRMs of varying structures and lipophilicities. This 12 PRM accord is simpler than a fully formulated perfume but more complex than a single perfume molecule. The geometric variations in the self-assemblies were evaluated using small-angle neutron scattering, perfume head space concentrations were determined using gas chromatography-mass spectrometry, and perfume localization was identified using NMR spectroscopy. The addition of the perfume accord caused enlargement of the micelles in both surfactant systems, with a greater change observed for ST2S/CAPB formulations. Furthermore, the addition of DPG to ST2S/CAPB resulted in micelle shrinkage. The micelle geometries and PRM localization in the micelles were affected by the degree of branching in the surfactant tail.
Introduction
Odiferous molecules are ubiquitous in everyday life; they are used in foods to enhance avor; in house-cleaning products, cosmetics and perfumery, and personal care products to impart pleasant aromas; in cleansers when added as an essential oil; and in aromatherapy for benecial psychological effects. [1][2][3][4] Because fragrance is a primary factor in consumer purchasing choice 5,6 and requires the use of expensive ingredients, efficient ingredient deposition onto surfaces is crucial to minimize product loss during rinse-off.
Perfumes are commonly classied by their functional groups, such as aldehydes, alcohols, and phenols, and by their log P values, which denote the lipophilicity or hydrophilicity of a perfume raw material (PRM). 1,2 Their inherent high volatility, hydrophobicity, and susceptibility to oxidation upon storage render formulation rather difficult. Methods such as spray drying, extrusion, coacervation, and emulsication are used to appropriately encapsulate and release perfumes from colloidal domains. Gas chromatography-mass spectrometry (GC-MS) is traditionally used to assess the correlation between the log P values of perfumes and their release prole. However, ambiguity about the colloidal domains of formulations complicates such analyses. This complexity is compounded by the use of numerous PRMs to generate a fragrance within a formulation. As the fragrance is generated based on a complex PRM mixture, the relationship of the fragrance with the overall formulation is not straightforward. 3 Our previous work with a 3 PRM accord in a mixed-surfactant system with a cosolvent 7 revealed that surfactant, cosolvent, and oil combinations were not conducive to the formation of single-phase microemulsions. NMR analysis showed that the 3 PRMs were preferentially localized within the micelle core, regardless of their lipophilicities, indicating that perfume-perfume interactions strongly inuence the location of a PRM within a surfactant self-assembly. Additionally, smallangle neutron scattering (SANS) measurements showed that increasing the concentration of the 3 PRM accord caused continual growth of the self-assemblies, which maintained spherical or ellipsoidal geometries at high surfactant concentrations. This behavior contrasts with the planar, vesicular, or worm-like structures that are expected to form based on the lipophilicities and molecular structures of specic PRMs in the system. Hence, in the present work, we created a more complex perfume accord consisting of 12 PRMs with various structures and physical properties, which approaches the complexity of a full perfume formulation and has a good presence in the headspace. It is envisioned that this approach will advance the understanding of how an actual perfume mixture inuences the aggregation properties of rinse-off systems at concentrations that better reect those of personal care products, and whether the PRMs are again preferentially localized somewhere within the surfactant self-assembly.
Specically, we studied the location of the 12 PRMs and the colloidal domains in two mixed-surfactant systems: sodium trideceth-2 sulfate (ST2S)/cocamidopropyl betaine (CAPB) and sodium laureth-3 sulfate (SLE3S)/CAPB. These surfactant systems were chosen as representative of the mixed-surfactant systems commonly used in personal care products for mildness. Moreover, these systems allowed us to expand upon our previous work 7 by investigating the inuence of surfactant structure (branched (ST2S) versus linear (SLE3S)) on perfume localization. Furthermore, we determined the concentration of each PRM in the headspace of each mixed-surfactant system. In addition, we investigated the effects of perfume concentration and the addition of a cosolvent (dipropylene glycol; DPG) on the mixed-surfactant systems. The addition of a cosolvent is of particular interest, as it reduces the cohesive energy density and dielectric constant of the medium by affecting polar intermolecular forces including hydrogen bonding. 8,9 DPG is commonly used as a cosolvent in detergents, personal care products, and cosmetics owing to its low toxicity and high dielectric constant. 8 It is also commonly used in perfumery as a solvent and diluent because of its high boiling point, transparency, and lack of odor. Although there are reports pertaining to the effect of cosolvent, 8,10-13 perfume accords, 2,14 and individual perfumes, in surfactant systems, 15-17 information on the combined effects of a cosolvent and a perfume accord on a surfactant system is limited, 14 likely because of the complexity associated with the use of multiple additives.
We used SANS to elucidate the size and shape of the selfassemblies in solution. [18][19][20][21][22] In addition, we used NMR spectroscopy to investigate the location of the PRMs and GC-MS to study perfume release from the formulations. The use of a moderately complex formulation and perfume accord allowed for a more realistic investigation of how a fragrance composition consisting of multiple PRMs affects the structure of the self-assemblies.
Sample preparation
All raw materials were provided by Procter & Gamble. In total, eight samples were tested. The compositions of the SANS and NMR samples are detailed in Table 1. The GC-MS samples had the same compositions, except that only H 2 O was used instead of H 2 O/D 2 O mixtures. The rst ve samples contained ST2S, a branched-tail surfactant, and the remaining three samples contained SLE3S, a linear-tail surfactant. In addition to these primary surfactants, each system also included CAPB. The surfactant structures are shown in Scheme 1. Samples 1-3 and 6-8 had increasing amounts of the perfume accord, whereas samples 2, 4, and 5 had increasing amounts of DPG and a constant 0.5 wt% perfume accord. The perfume accord used in this study consisted of 12 PRMs, chosen to cover a wide range of log P values and molecular structures ( Table 2). The samples were placed in titanium cell holders with quartz windows, which were separated by a 1 mm sample gap. Different neutron wavelength (l) and sample-to-detector distance (SDD) congurations were used to achieve the desired q-range, where the scattering vector q ¼ (4p/l)sin(q/2) and q is the scattering angle. Samples 1-3 were irradiated with neutrons of wavelength l ¼ 6Å at SDDs of 1 m (high-q conguration) and 4 m (mid-q conguration), and with neutrons of wavelength l ¼ 8.09Å at an SDD of 15.3 m (low-q conguration) to cover a wide q-range of 0.0027-0.5568Å À1 . The data collection times were 5 min, 20 min, and 1 h for the high-, mid-, and low-q ranges, respectively. For samples 4-8, the high-q conguration was l ¼ 5Å with an SDD of 1.2 m, the mid-q conguration was l ¼ 5Å with an SDD of 4.6 m, and the low-q conguration was l ¼ 12Å and an SDD of 4.6 m, which covered a q-range of 0.00318-0.5666Å À1 . The data collection times were 10, 20, and 45 min for the high-, mid-, and low-q ranges, respectively. All sample data were collected at 25 C.
Using the data reduction macros developed for Igor Pro by the NCNR, 23 the data were reduced by correcting for background scattering, detector resolution and sensitivity, and beam transmission. Then, the data were radially averaged to obtain the absolute scattering intensity, I(q). The high-, mid-, and low-q data were combined to obtain the complete data set. Each data set was analyzed in SasView 5.0 ref. (24) using the smearedresolution ellipsoid form factor coupled with the Hayter-Penfold rescaled mean spherical approximation (MSA) structure factor model to account for electrostatic effects. The equations used for the ellipsoid form factor in SasView are given by Feigin and Svergun. 25 The equations describing the structure factor were developed by Hayter and Penfold 26 and by Hansen and Hayter. 27 A polydispersity term was included in the model for the equatorial radius (R e , perpendicular to the axis of rotation) using a Schulz distribution. Finally, the decoupling approximation, which was developed by Hayter and Penfold 28 and discussed further by Kotlarchyk and Chen 29 to correct for errors in the structure factor calculation caused by polydispersity and nonsphericity, was also used in modeling the data.
The magnitudes of the polar radius (R p , parallel to the axis of rotation) and R e in the ellipsoid model determine whether the ellipsoid is oblate or prolate. For R e > R p , the ellipsoid is oblate (disc-like); for R p > R e , the ellipsoid is prolate (melon-shaped); and for R p ¼ R e , the ellipsoid is spherical. The data were modeled using both the oblate and prolate ellipsoid cases, and which type of ellipsoid best described the data was determined based on the minimized residuals and the visual t quality. Particular attention was paid to minimizing the residuals in the high-q region, as this region is most sensitive to the form factor and least affected by the structure factor. If both cases t the data similarly, then the model with the smallest sqrt(c 2 /N) was chosen as the best t. The solvent scattering length density (SLD) was calculated and xed based on the weighted average of D 2 O, H 2 O, citric acid, and DPG in each sample. The micelle (i.e., particle) SLDs were also calculated and xed to the weighted average of the primary surfactant (i.e., ST2S or SLE3S), CAPB, and the perfume accord, if applicable. To simplify the SLD calculations, an average molecular formula and density were calculated for the perfume accord based on the weighted average of its components. The SLDs for the solvents and micelles as well as the parameters used to calculate them can be found in the ESI. †
GC-MS studies
GC-MS samples were prepared at room temperature two days before analysis to allow for equilibration. The components were added in the following order: ST2S or SLE3S, CAPB with 3 wt% citric acid, DPG, water, and the PRM accord. The sample was then vortexed until the components were thoroughly blended and set aside until testing. The GC-MS data were obtained using an Agilent Model 6890 gas chromatograph tted with a Gerstel MPS-2 autosampler, a 0.75 mm ID SPME injection port liner (Supelco, Bellefonte, PA, USA), and a J&W DB5-MS GC column with an ID of 30 m  0.25 mm and a lm thickness of 1.0 mm (Agilent Technologies, Inc., Wilmington, DE, USA). The detector was a Model 5973 mass selective detector (Agilent Technologies, Inc., Wilmington, DE, USA) with a source temperature of approximately 230 C and a MS Quad temperature of approximately 150 C. Ultra-pure helium was used as the carrier gas with a ow rate of 1 mL min À1 .
Each sample (3 g) was placed in a clean 20 mL headspace vial with a magnetic stir bar. Aer the vials were closed with PTFE septum caps, the samples were stirred and allowed to equilibrate at room temperature for at least 30 min. The sample vials were then placed in an autosampler tray. Each sample vial was automatically placed in the sampling chamber, where it equilibrated at 30 C for 1 min. The sampling time was 1 min., during which the sample was stirred at 500 rpm and taken up by a 50/ 30 mm, 24 ga, 1 cm long DVB/CAR/PDMS SPME ber assembly. The sample was injected at 270 C, and the GC-MS analysis was begun aer allowing the sample to desorb from the SPME assembly for 5 min. The temperature was initially held at 50 C for 30 s, and then increased to 275 C at a rate of 8 C min À1 and held for 2.5 min. The individual PRMs were identied using MS spectral libraries from John Wiley & Sons and the National Institute of Standards and Technology, which were purchased and licensed through Agilent. The chromatographic peaks of specic ions were integrated using the MassHunter soware (Agilent Technologies, Inc., Wilmington, DE, USA).
NMR studies 1 H-NMR spectroscopy was performed using a Bruker AV 400 MHz spectrometer (Billerica, MA, USA). The NMR samples were prepared 3 days before analysis using D 2 O as the solvent. The H 2 O peak at 4.69 ppm was used as a reference for data analysis.
Effect of 12 PRM accord addition
Mixed micelles composed of a primary surfactant (ST2S or SLE3S) and a secondary surfactant (CAPB) will form because the total surfactant concentration of 7.5 wt% is above the critical micelle concentration. The hydrophobic PRM accord is expected to localize within the micelles, while DPG is expected to primarily locate in the aqueous phase based on literature. 8 The SANS data modeling results are summarized in Table 3 and the individual tting parameters for each sample are available in the ESI. † The SANS tting results in Table 3 show how the micelle radii and volumes of the ST2S and SLE3S systems vary with the perfume content, and Fig. 1 shows an overlay of the SANS data and the model ts as functions of perfume concentration and mixed-surfactant system.
Effect of DPG addition
The self-assembly volume fractions determined by SANS were generally similar to the volume fractions calculated from the amount of each component expected to be in the dispersed phase (ST2S or SLE3S, CAPB, and perfume). However, once perfume was added to the ST2S/CAPB system, the volume fraction from SANS was somewhat higher than expected, as if the perfume added disproportionately more volume to the dispersed phase. This behavior suggests that the perfume formed a small droplet surrounded by surfactant monomers in the core of the ST2S/CAPB micelle instead of being solubilized by the surfactant tails, as presumed to occur in the SLE3S/CAPB system. As R p > R e for the branched-tail and linear-tail mixedsurfactant systems, prolate ellipsoids were formed for both systems. Interestingly, the size and geometry of the ST2S/CAPB micelles were more strongly affected by perfume addition than those of the SLE3S/CAPB micelles. Specically, R p increased for both systems as perfume was added, sharply increasing in the ST2S/CAPB system but gradually increasing in the SLE3S/CAPB system, whereas R e remained virtually constant in both systems. As a result, the volume change of the ST2S/CAPB micelles upon perfume addition was larger than that of the SLE3S/CAPB micelles.
As perfume accords are water-insoluble, surfactant micelles are required for solubilization in aqueous environments. The ST2S surfactant forms smaller micelles because the branchedtail structure causes the alkyl chains to pack differently in the core than the linear chains of SLE3S. 30 This packing is reected in the R p values, which show that the tails pack together to form elongated micelles in the ST2S/CAPB system. The tails may not be packed quite as tightly along the long axis of the micelle, which could allow the micelle to elongate further as perfume is taken up. Because of its longer linear-tail and extended head group, the SLE3S monomer is longer than the ST2S monomer, which explains why the SLE3S/CAPB micelles expand more gradually with perfume addition than the ST2S/CAPB micelles. When the 3 PRM accord and the present 12 PRM accord were added to a concentrated ST2S/CAPB system, self-assembly began with the formation of spherical structures that enlarged and eventually formed oblate ellipsoids in both cases. 7,31 Although the oil composition inuences the extent to which a microemulsion can be diluted and the range of compositions which form microemulsions, 31 the differences in size and geometry between the present work and our past work could be caused by a few key differences-namely, the total surfactant concentration (7.5 wt% versus 20-32.5 wt%, respectively), and the ratio of perfume : surfactant : DPG, or both. Table 4 shows the change in the ST2S/CAPB volume fraction, micelle radii and volume with the addition of DPG, and Fig. 2 shows the overlay of the scattering data and model ts. The volume fractions calculated from the amounts of ST2S, CAPB, and perfume in the system were constant because the surfactant and perfume concentrations did not change; however, the volume fractions determined from SANS decreased from 0.10 to 0.07 as DPG was added. With respect to the micelle radii, although the micelles were prolate ellipsoids at all DPG concentrations (R p > R e ), the micelles became less elongated with the addition of DPG owing to a decrease in R p . Specically, the decrease in micelle volume as a function of DPG concentration was due to shrinkage along one axis (R p ), resulting in less elongated micelles. Notably, this trend is contrary to that obtained with perfume addition. The change in R e with DPG addition was small, as also observed upon perfume addition. Table 3 Volume fractions (calculated from the amounts of surfactant and perfume in the system, and as determined by SANS), dimensions, and volume of branched-tail and linear-tail mixed surfactant self-assemblies as a function of perfume concentration DPG aids in solubilizing the perfume oil in the continuous phase, which explains the decreases in the volume fraction, micelle radii, and micelle volume, as less perfume needs to be solubilized inside the micelles. Recent work on the effects of DPG on the self-assembly of a cationic cetyltrimethylammonium tosylate (CTAT) system 11 and a sodium laureth-1 sulfate (SLE1S)/CAPB system 8 revealed that DPG reduces the micelle size. With DPG primarily located in the aqueous phase, the solvent dielectric constant is decreased, which increases electrostatic repulsions between the head groups, resulting in shrinkage of the micelle radius through an increase in the curvature of the micelle surface. 8 Apparently, the known effects of DPG still apply to the overall perfume accord, despite its complexity. As noted above, the geometry and sizes of the colloidal domain in this work differ from those in our previous work. 31 The 5 branched-tail mixed micelle systems studied in this work can be approximately located in the surfactant corner of the tertiary phase diagram at 90 wt% water (Fig. S1 † in the ESI †) that was developed in our previous work, 31 whereas the samples made as part of that work are located along the 1 : 1 surfactant : DPG line. This means that the 5 present samples contain more surfactant than the previous samples 31 (7.5 wt% vs. 3.7-5 wt%), which is why there are also higher volume fractions of micelles in the present systems. Additionally, the amount of DPG in the current samples is overall lower than it was in the previous samples 31 (0-3 wt% vs. 3.7-5 wt%), which further results in a surfactant : DPG ratio that is heavily skewed toward surfactant in this work. Here, the micelles not only shrank with DPG concentration, but also the aspect ratio decreased so that the micelles became less prolate; in comparison, the structures observed in our previous work 31 were oblate ellipsoids. This could suggest that the surfactant : DPG ratio, in addition to the overall DPG concentration, inuences the micelle geometry as well as its size.
To obtain insights into perfume localization, we performed NMR studies on the mixed-surfactant systems in presence and absence of the PRM accord. The NMR peak shis for the two surfactant systems with varying perfume concentrations were compared (Table 5 and Fig. S1 †). The changes associated with each signal were small because of the very low perfume content. Nevertheless, these changes provided useful information about the preferential distribution of perfumes in the micelles. 7 In the SLE3S/CAPB system, protons '1', '2', and '3' exhibited shielding effects upon perfume addition, indicating the localization of the perfume molecules near these surfactant protons. Similar effects were observed in the ST2S/CAPB system, but the shielding effect was more prominent for hydrophobic protons '1' and '2' than the more hydrophilic proton '3'. This observation implies that the PRMs (irrespective of their individual log P values) are preferentially located near the micelle core. This nding is very similar to our recent results for a 3 PRM system, 7 but the implication here is more signicant, as the increased complexity in the system with 12 PRMs did not alter the perfume localization tendency. Furthermore, this nding differs from the behavior reported for individual PRMs in earlier studies. 17,32,33 Apparently, the methyl group branches in the ST2S tail facilitate the retention of perfume molecules near the micelle core by creating an efficient hydrophobic pocket. Although Fieber et al. showed that the surfactant composition and molecular structure of cosurfactants inuences the partitioning of perfume molecules, 34 the collective ndings from our previous 3 PRM system and the present 12 PRM system clarify the importance of the tail nature and conformation of the primary surfactant in driving the localization of PRM accords within micelles. In addition, weak (because of low PRM concentrations) intermolecular interactions among different PRMs become operative in mixtures. It is worth noting that the log P values of individual PRMs become less relevant in dictating the localization of PRMs in the micelle. The lipophilicity or hydrophilicity of each PRM is indicated by the color bar, with blue indicating that the PRM is hydrophilic, orange means the PRM is lipophilic, and green means it is intermediate. Within each column, the PRMs are ordered from lowest log P value (blue) to highest log P value (orange). The legend is below the color bar. In (b), there is a negligible amount of perfume in the headspace of the SLE3S/CAPB sample containing 0 wt% perfume owing to carry-over from the previously measured sample, and the corresponding headspace composition was omitted from (d) to prevent confusion.
Perfume release from surfactant systems
To study how effectively each PRM was released from each mixed-surfactant system as a function of perfume accord concentration and DPG concentration, the headspace concentrations were measured using GC-MS ( Fig. 3 and 4). The PRM headspace concentrations increased less than two-fold when the perfume concentration was doubled in each mixedsurfactant system, i.e., sample 2 versus sample 3 ( Fig. 3a and b) and sample 7 versus sample 8 ( Fig. 3c and d). Thus, the chemical activity of the PRMs increased as more perfume was added, resulting in the headspace concentrations increasing in accordance with Raoult's law. The headspace compositions did not change signicantly with increasing perfume concentration for either surfactant system. The headspace compositions were dominated by the PRMs with intermediate log P values (2.0-3.5), followed by the highly hydrophilic PRMs and the highly lipophilic PRMs. The concentrations of the PRMs within each log P range varied greatly (e.g., dihydromyrcenol had the highest headspace concentration among the PRMs with intermediate log P values), suggesting that the molecular structure of the PRM must be considered in addition to its log P value when attempting to improve the headspace concentration. Fig. 4 shows the change in head space concentration and composition as the DPG concentration increased from 0 to 3 wt% (samples 2, 4, and 5). The total headspace concentration and the headspace concentrations of most PRMs increased upon DPG addition, despite the perfume concentration remaining constant. The headspace mainly consisted of PRMs with intermediate log P values (2.0-3.5), both in terms of the headspace concentration and the composition, followed by the highly hydrophilic PRMs and then the highly lipophilic PRMs. DPG addition caused the headspace concentration of each PRM to increase, but the relative ratio of the PRMs in the headspace remained constant, indicating that DPG addition did not preferentially improve the chemical activity of any particular PRM. The headspace concentration increasing without a signicant change in its composition as DPG was added may have been caused by the micelles shrinking with DPG addition. Although the perfume is preferentially located in the center of the micelle based on the above NMR results, a smaller micelle would still have less volume in which the perfume could be solubilized. Therefore, as the micelles became smaller, the total amount of perfume the micelles could solubilize would also decrease; thus, the activities of every PRM would uniformly increase in kind, leading to the headspace concentration increasing without affecting the headspace composition.
The headspace concentrations of the PRMs within each log P range varied greatly (e.g., dihydromyrcenol had the highest headspace concentration among the PRMs with intermediate log P values), as was the case when the perfume concentration was increased. Interestingly, the PRMs with the highest headspace concentrations in the intermediate and lipophilic log P groups (dihydromyrcenol and b-ionone, respectively) had the lowest molecular weights within these groups. For the hydrophilic PRMs, phenylethyl alcohol had the lowest molecular weight, but benzyl acetate had a signicantly higher headspace concentration. The only structural difference between these PRMs is an acetate group versus a -CH 2 OH group. Within each log P group, the differences in the relative concentrations of each PRM in the perfume accord alone cannot explain the signicant variations in the headspace concentrations. These observations conrm that the molecular structure of the PRM is also important to consider for improving perfume release from rinse-off systems. corresponding headspace compositions. The lipophilicity or hydrophilicity of each PRM is indicated by the color bar, with blue indicating that the PRM is hydrophilic, orange means the PRM is lipophilic, and green means it is intermediate. The legend is below the color bar. In each column, the PRMs are ordered from lowest log P value (blue) to highest log P value (orange).
Conclusions
Past work on perfumes in surfactant systems has primarily focused on the effects of a single perfume molecule on the geometry of the surfactant self-assembly 15,16 and where the perfume molecule localized within the assembly. 17,33 In contrast, our previous work on a 3 PRM mixture investigated the effectiveness of the ST2S/CAPB mixed-surfactant system in creating single-phase microemulsions with perfume, the inuence of perfume-perfume interactions on PRM localization within the self-assembly, and the effect of the perfume on the self-assembly. 7 The present work is more industry relevant and built upon our previous work by employing an even more complex 12 PRM accord, consisting of PRMs with various structures and c log P values; by comparing two different mixed-surfactant systems with branched or linear tails; and by considering the cosolvent (DPG) concentration as a variable. SANS modeling showed that the geometry of the mixed-surfactant self-assemblies remained the same despite the tail structure, with both systems forming prolate ellipsoids. Perfume addition caused the micelle volumes in both systems to increase, but the ST2S/CAPB micelles were more strongly affected. Upon DPG addition, the ST2S/CAPB micelles became smaller and less elongated while retaining a prolate ellipsoidal geometry, which is consistent with ndings of Jiang et al. 8 and Padasala et al. 11 The GC-MS headspace results showed that the headspace concentrations of all the PRMs increased as the perfume concentration increased, and also that the headspace concentrations increased slightly when DPG was added at a constant perfume concentration. The log P values and the PRM molecular structures appear to be important variables that inuence the headspace concentration of each PRM. NMR analysis showed that the perfume accord preferentially localized amongst the surfactant tails in the branched-tail surfactant system, whereas localization along the tail and ethoxy groups was observed in the linear-tail surfactant system. Thus, perfume molecule localization can be inuenced by more variables than those identied by Fan et al. 17 and Fischer et al. 33 Perfume-perfume interactions and the surfactant tail structure appear to be responsible for this behavior. These results show that the surfactant molecular structure and the changes to the system properties caused by cosolvent and perfume addition inuence the geometry of the dispersed phase structures. Additionally, the release of active ingredients, such as fragrance, can be improved or hindered by the molecular structures and degree of lipophilicity/lipophobicity of the components as well as the physical properties of the overall system.
Conflicts of interest
There are no conicts to declare.
|
2022-05-19T15:12:00.660Z
|
2022-05-17T00:00:00.000
|
{
"year": 2022,
"sha1": "dff0214f82300a7ba765c75e4ea1ea8253401e86",
"oa_license": "CCBY",
"oa_url": "https://pubs.rsc.org/en/content/articlepdf/2022/ra/d2ra00688j",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "fa3147bf96518d55de45e29585526fa36d42c8ee",
"s2fieldsofstudy": [
"Chemistry"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
41285336
|
pes2o/s2orc
|
v3-fos-license
|
Quantum trajectories for the realistic measurement of a solid-state charge qubit
We present a new model for the continuous measurement of a coupled quantum dot charge qubit. We model the effects of a realistic measurement, namely adding noise to, and filtering, the current through the detector. This is achieved by embedding the detector in an equivalent circuit for measurement. Our aim is to describe the evolution of the qubit state conditioned on the macroscopic output of the external circuit. We achieve this by generalizing a recently developed quantum trajectory theory for realistic photodetectors [P. Warszawski, H. M. Wiseman and H. Mabuchi, Phys. Rev. A_65_ 023802 (2002)] to treat solid-state detectors. This yields stochastic equations whose (numerical) solutions are the ``realistic quantum trajectories'' of the conditioned qubit state. We derive our general theory in the context of a low transparency quantum point contact. Areas of application for our theory and its relation to previous work are discussed.
I. INTRODUCTION
The field of research that surrounds the quest for a large-scale quantum computer is very exciting. At present, solid-state proposals 1,2,3,4,5 seem promising. The ability to read out the state of the quantum bits (qubits) of information is of obvious importance in any quantum computational scheme. In this paper we consider continuous measurement of the state of a pair of coupled quantum dots (CQDs) occupied by a single excess electron. This constitutes a charge qubit. It is worth mentioning that spin qubits 1,2,3,4,5 are considered more favorably for solid-state quantum computation due to their relatively long coherence times, 6 but read-out may have to be performed via charge qubits using spin-tocharge conversion. 1,7,8 The evolution of solid-state qubits subject to continuous measurement has received considerable theoretical consideration recently. 9,10,11,12,13,14,15,16,17,18,19 Single realizations of the continuous measurement of a solid-state CQD qubit, known as conditional (or selective) evolution, have been treated by a number of groups. 11,12,13,14,15,16,17,18,19 These works conditioned the qubit evolution on quantum processes (such as tunneling) at the scale of a mesoscopic detector. They did not consider conditioning on the macroscopic current that is realistically available to an observer. In particular, they ignored the noisy filtering characteristic of the external circuit, including an amplifier. It is worth noting that non-idealities have been considered in some of these works. Ref. 13 considered a detector with excessive backaction. Ref. 17 did this also, and also considered extra classical noise, phenomenologically. Ref. 16 considered "inefficient" measurements. None of these considered filtering.
In this paper we consider the evolution of a solid-state qubit conditioned on the output available to a realistic observer, which has been filtered and degraded (i.e. made more noisy) by an external circuit. That is, we are interested in the evolution of the system conditioned on information available to an observer, not on the microscopic events occurring within the detector to which a real observer has no direct access. Being able to determine the state of a quantum system conditioned on actual measurement results is essential for understanding and designing feedback control. 12,20,21,22,23,24,25,26,27 As well as being intrinsically interesting, this is also expected to be important in quantum computing, both for state preparation and quantum error correction. 28,29,30 A quantum trajectory 13,31,32,33 describes the Markovian stochastic evolution of an open quantum system conditioned on continuous monitoring of its output by a bare detector. A "bare" detector is one which does not include the noisy filtering characteristic of realistic measurements. In an experiment the output from this detector is filtered through various noisy electronic devices. Due to the finite bandwidth of all electronic devices, the evolution of the conditional state of the quantum system must be non-Markovian. A general method of describing this evolution was presented in recent papers 34,35 by two of us in the context of photodetection, where it was applied to an avalanche photo-diode and a photo-receiver. In the present paper the theory of Ref. 35 is applied to a solid-state detector -the low transparency quantum point contact 9,36 (QPC), or tunnel junction, which is an ideal detector. 11 In our approach an equivalent circuit is used to model the effects of a realistic measurement. Note that for clarity we will use the terminology detector for a bare detector and measurement device for a detector embedded within a measurement circuit.
The paper is organized as follows. We begin in the next section by describing our models for the qubit and the QPC (including the monitored qubit's conditional and average dynamics in the bare detector case). We then introduce and analyze our equivalent circuit for realistic measurement in Sec. III. The method of deriving realistic quantum trajectories is presented in Sec. IV, in the context of a QPC. We discuss our results in Sec. V and conclude in Sec. VI with a summary, comparison with previous work, and prospects for future work.
II. SYSTEM
In this section we describe the models for the qubit and the detector. Using a master equation formalism we present the conditional and ensemble average dynamics of the qubit state when measured by a low transparency QPC. The conditional qubit dynamics in the bare measurement case are represented by a stochastic master equation. We choose to present stochastic differential equations in the Itô formalism rather than the alternative Stratonovich formalism. 37 Extension of the theory to our more realistic measurement case occurs later. Figure 1 is a schematic representation of the CQD qubit and nearby low transparency QPC or tunnel junction. The CQDs (labeled 1 and 2) are occupied by a single excess electron, the location of which determines the logical state of the qubit. We assume that each quantum dot has only one single-electron energy level available for occupation by the qubit electron. These energy levels are denoted by E 1 and E 2 . Using the convention of = 1 (as we will for the entire paper), the total Hamiltonian for the qubit can be written aŝ where Ω 0 is the co-efficient of tunneling between the qubit dots andĉ 1 (ĉ 2 ) is the Fermi annihilation operator for the single available electron state within the qubit dot labeled 1 (2). The qubit electron tunnels between the two dots at the Rabi frequency Ω = Ω 2 0 + ε 2 , where ε ≡ E 1 − E 2 is the asymmetry in the CQD energy levels.
The state of a measured quantum system is affected by the detector in two ways. First, there is the measurement back-action caused by their mutual interaction. Second, if the output of the detector is observed, then the state of the system is conditioned by the stochastic outcomes. We describe the conditional dynamics of the measured qubit, including the measurement back-action, using a stochastic approach. In the case of measurement with a bare ideal detector, the state of the qubit is conditioned by electron tunneling events through the detector which constitute an idealized output current. For such an ideal detector the measurement back-action is quantumlimited, also called Heisenberg-limited. 38 A number of formalisms exist that describe the evolution of a measured quantum system conditioned on a particular measurement result from the detector. The conditional dynamics of continuously measured CQD systems have been treated by Bloch-type equations, 9,19 quantum trajectory theory 13,14,16 and a Bayesian formalism. 11,12,15,17,18 This Bayesian formalism has been shown to coincide with the quantum trajectory formalism with only notational differences (see the appendix of Ref. 14). All three formalisms coincide for the ensemble average dynamics of the measured CQD system. In the stochastic approach, the (Markovian) conditional dynamics of the measured qubit state is described with a stochastic master equation. This generates a "quantum trajectory", so called because it tracks the state of the quantum system in time. We also present the ensemble average master equation.
The equivalent circuit for the QPC coupled to the qubit is shown in Figure 1. We represent the tunnel junction by a capacitance C J , which contains the charge Q J . The stochastic electron tunneling events through the junction are represented by a current source. The location of the CQD electron changes the height of the potential barrier in the QPC and consequently the current through it, thus providing the means to measure the qubit state. For simplicity, we assume that electrons tunnel only from source to drain. This tunneling occurs at two different rates, namely r and r ′ , which correspond to the near (dot 1) and far (dot 2) CQD being occupied, respectively.
The ensemble average master equation for the qubit state, ρ, when measured by a low transparency QPC, or similar single tunnel-junction device, is 9,11,13,14 Heren =ĉ † 1ĉ1 is the occupation of the near dot. The Lindblad superoperator D represents the irreversible part of the qubit evolution -the decoherence. It is defined in terms of two other superoperators, J and A: where J (the 'jump' superoperator) and A (the anti-commutating superoperator) are defined by These superoperators, introduced in Ref. 32, are used commonly in quantum optics measurement theory. For simplicity we assume real tunneling amplitudes whereby which implies that X < 0. Complex tunneling amplitudes are allowed in the model of Ref. 14 and the generalization here would be straightforward. A realistic observer may not be able to tell when a tunneling event through the QPC occurs. However, we argue that in principle this information would be contained in the movement of the Fermi sea electrons in the leads attached to the QPC. Thus, we can legitimately represent the conditional evolution of ρ(t) (denoted with a superscript µ) in terms of these microscopic events. Using the method for quantum jumps introduced in Ref. 13, this conditional evolution is described by the following Itô stochastic master equation 14 where we have introduced the classical point process dN (t) that represents the number (either 0 or 1) of electron tunneling events through the QPC in an infinitesi- In Eq. (7) we have also introduced the convenient superoperator H and operatorR which are defined by It can be seen from the master equation (2) that the minimum tunneling rate through the QPC, r, occurs when qubit dot 1 is occupied (n = 1). This is due to maximum electrostatic repulsion between the qubit electron and electrons in the QPC vicinity. Accordingly, the maximum QPC tunneling rate r ′ occurs when n = 0. These tunneling rates could be functions of the voltage across the detector, which we consider as changing with time. However, this would necessarily mean that the measured qubit's evolution cannot be described by the quantum master equation formalism. To allow for this would be to go beyond what anyone has done in this area. Our equivalent circuit for realistic measurement of the CQDs is shown in Figure 2. We emphasize that this circuit models effects of realistic measurement (additional classical noise and filtering of the signal), not an actual experimental apparatus.
The circuit is biased by a non-ideal DC voltage consisting of a noiseless voltage ε and a noisy voltage source e i . This (white) noise source could be considered as the Johnson-Nyquist noise from the equivalent circuit resistance R i at some effective temperature T i . We emphasize again that this is a model only and need not correspond to a real temperature in an actual apparatus. The small current through the detector is amplified, then measured. In this process an observer will see white noise in addition to the current through the detector. This is modeled by adding a noisy output current e o /R o to the signal from the detector prior to measurement by a perfect ammeter, yielding the current I. The parasitic capacitance C across the detector is due to the large cross-sectional area of the leads relative to the detector junction.
Again, it is important to note that the circuit components are not necessarily representative of an actual experimental setup. For example, an amplifier does not consist of a noisy voltage and a resistor, rather the observed effect of amplification of the current through the detector can be modeled as the addition of an output noise e o /R o to the current through the detector. Although our description of the circuit is rather simple, we believe that it is a reasonable starting point that models some essential effects of a realistic measurement. Future improvements to this circuit model could include considering an actual circuit from an experiment.
We analyze the equivalent circuit with the low transparency QPC as detector and produce expressions for the measured current I(t) and the time evolution of the parasitic capacitor charge Q(t). The variable Q(t) is used to describe the state of the circuit part of the measurement device.
For the moment, ignore tunneling through the QPC.
Analysis of the measurement circuit using Kirchhoff's electrical circuit laws yields the following Itô differential equation for the increment in Q (the charge on the parasitic capacitor) due to the circuit components Similar analysis yields an expression for the measured current as a function of time: For the purposes of our work it is useful to express the (Johnson-Nyquist) noise sources e i and e o in terms of stochastic increments. In the steady state, Johnson-Nyquist voltage noise is white noise and has a flat spectrum where T is the temperature of the resistor R and k B is Boltzmann's constant. The current spectrum ('spectral density') definition 40 used here is where G(τ ) is the two-time autocorrelation function of the measured current. 54 Obviously a current flow is not an equilibrium situation, but for reasonable bias voltages the approximation of Eq. (13) remains valid. 41 For simplicity, we take the flat spectra of the input and output voltage noises to be 2D i R 2 i and 2D o R 2 o , respectively. This allows us to write Eqs. (11) and (12) in terms of the input and output Wiener processes, dW i (t) and dW o (t), as where These expressions correct the expressions in Refs. 34, 35 and 42 from 4k B T /R to 2k B T /R. The Wiener increment is related to Gaussian white noise ξ(t) by dW (t) = ξ(t)dt. 37 Now consider a single electron tunneling event through the QPC (dN (t) = 1). The charge on the parasitic capacitor will change by an amount edN (t), where e is the charge on an electron. This gives (17) where we have introduced the simplifying notations α = 1/R i C and β = ε/R i . The solution to this differential equation gives the value of Q(t) that may be substituted into Eq. (16) to give a lengthy expression for the measured current. 42
IV. DERIVATION OF REALISTIC QUANTUM TRAJECTORIES
The derivation of realistic quantum trajectories follows a number of well defined steps as presented for photodetectors in Ref. 35. We refer the reader to Ref. 35 for specific details of the derivation steps and only present the essential points and details that are unique to the solid-state situation. Note however that we use a somewhat simpler derivation, using the Zakai equation in Sec. IV B rather than the Kushner-Stratonovich equation of Ref. 35.
A. Stochastic differential Chapman-Kolmogorov equation
Eq. (15) describes the evolution of the circuit state for situations where Q(t) is known. A realistic observer will not have direct access to the precise value of Q(t) due to the randomness of the microscopic events occurring within the device. We therefore require an equation for the evolution of the probability distribution for Q, written P (q). Following the procedure outlined in Ref. 35, we obtain the stochastic differential Chapman-Kolmogorov (SDCK) equation for the evolution of P (q): where m = −αq + β. This equation gives the increment in the probability distribution for the charge on the parasitic capacitor conditioned by the unobserved microscopic events (µ) occurring within the measurement device.
B. Zakai equation
The state of the circuit part of the measurement device is now represented by the probability distribution P (q) that was introduced in the previous section. The state of this classical system changes upon measurement and so P (q) must be updated. The best estimate of the new probability distribution representing the conditioned state of the measurement device, given a measurement result I, is found using Bayesian analysis 43 to bẽ where Λ(I) = P (I|q = β/α). HereP (q|I) is read 'the probability of q given I'. The tilde denotes an unnormalized distribution and the value of q = β/α is chosen for convenience. The Zakai equation tells us how to update the probability distribution P (q) when the measurement result I is obtained. The quantity P (I|q) is the probability of obtaining the result I given that the state is q. We will use the simpler notation P q (I) ≡ P (I|q), where the subscript denotes the result upon which the conditioning is performed. Λ(I) can be thought of as the ostensible probability distribution, 33 as opposed to the actual probability distribution P (I) = dqP (I|q)P (q) = Λ(I) dqP (q|I) , (20) which replaces Λ(I) in the expression for the normalized distribution P (q|I). 35 From our expression for the measured current, Eq. (16), P q (I) is a Gaussian distribution with a variance of ν = (D i + D o )/dt and a mean of m = −αq + β. Thus, Eq. (19) gives the the Zakai equation (to order dt): where we have defined D Σ = D i + D o for convenience. Note that I has the ostensible distribution Λ(I) = exp −I 2 /2ν / √ 2πν.
C. Combining the stochastic increments
Our description of the stochastic conditional evolution of the measurement device is found by combining the increments dP µ (q) and dP I (q) given in the previous two subsections. The stochasticity of these two increments is related as the input noise dW i plays a role in both. For this reason we must combine them into one increment using rather than by simply adding them together. Remembering that we will eventually average over unobserved processes, the input noise needs to be separated into observed and unobserved parts. We express this as where Here a, b and c are as yet undetermined expressions and dW ′ is unobserved, normalized white noise that is unrelated to the known output I. When averages are taken, dW ′ will be averaged over and I kept. The observed output [Eq. (16)] can be expressed as Using Eqs. (24) and (25) gives the expression for dW ′ : Using this in Eq. (23) and equating the left and right hand sides allows a, b and c to be determined. Substitution of a, b and c back into Eq. (23) yields Using this result and the SDCK equation [Eq. (18)] in Eq. (22) gives This result represents the evolution of the circuit state conditioned on both the microscopic events occurring within the device and the observed current I.
D. Joint stochastic equation
The stochastic state of the joint classical-quantum system is found by forming the new conditional quantitỹ The evolution ofρ µ I (q) is described bỹ ρ(q) + dρ µ I (q) = P (q) + dP µ I (q) ρ(t) + dρ µ (t) . (30) The result of this process is the joint stochastic equation 35 Averaging over unobserved processes (dW ′ and dN ) is the next step in the derivation of realistic quantum trajectory equations and yields an expression for dρ I (q). This procedure removes the stochasticity associated with the unobserved processes within the detector and leaves the stochasticity associated with the measurement (I). The resulting equation is called a superoperator Zakai equation as we have obtained a quantum analogue of the Zakai equation in that from measurement we are conditioning the state of a super-system that contains a quantum system. It is important to realize that after averaging over unobserved processes the super-system statẽ ρ(q) + dρ I (q) will not factorize as in Eq. (30).
E. Normalization
Normalization of the superoperator Zakai equation is the final step in our derivation and yields the superoperator Kushner-Stratonovich (SKS) equation. The normalization is achieved as follows: After normalization the true expression for the observed current I should be substituted into the SKS equation. The true probability distribution for I can be found using Eq. (21) in Eq. (20) to yield where Q = qP (q)dq. Thus, the true expression for the observed current is where dW is the observed white noise (a Wiener increment). Here the average is Q = qTr [ρ I (q)] dq, since we are considering the output I for the combined classical-quantum super-system. Averaging over the unobserved noise dW ′ and tunneling process dN yields the superoperator Zakai equation, which upon normalization via Eq. (32) and substitution of Eq. (34) for I produces the SKS equation: This is the main result of our paper. The first line of Eq. (35) describes the evolution of the classical measurement device. The second line consists of two terms: the first term describes information gain about the measurement device (q) from its output -Eq. (34); the second term describes back-action on the classical device due to the observed noise. The third line describes the average evolution of the quantum system, including quantum back-action. The final line describes the effect of the quantum system on the measurement device. It is worth noting that the term involving J represents average evolution due to electron tunneling events through the QPC. It changes the most likely value for the charge of the parasitic capacitor from Q to Q − e when an electron tunnels through the QPC -effectively counting the average number of electrons passing through the QPC. The approach of Ref. 10 (also used in Ref. 39) involves a similar technique in which the exact number of electrons that have tunneled through the detector is tracked. This was also considered in the (earlier) derivation of the rate equations in Ref. 9.
The numerical solution of Eq. (35) would produce a trajectory for the state of the combined circuit-qubit system conditioned by a particular realization of the measured current I(t). The normalized conditioned qubit state is found from Thus, the realistic quantum trajectories (for the qubit state alone) are obtained by numerically solving the SKS equation and using Eq. (36). The results of this procedure will be presented in a future paper.
V. DISCUSSION
A simple consistency check for our SKS equation [Eq. (35)] is to integrate it over all q and recover the unconditional master equation. It is easy to confirm that this is indeed the case using the fact that well-behaved probability distributions (and their derivatives) vanish at ±∞.
A considerably more difficult task is to attempt recovery of the ideal conditional master equation (2) from the SKS equation (35). In theory, this should be possible in the limit of a measurement circuit with a small response time given by R i C (large bandwidth α = (R i C) −1 ) and low noises D i and D o . We now explore this question in detail.
The time taken to determine which CQD is occupied by the qubit electron is equal to the time required to ascertain the rate of tunneling through the detector. This task is made considerably more difficult by the white noise and finite bandwidth of the circuit containing the detector. Without the white noise the observer would see a spike in the current every time there was a tunneling event. Depending on the relative sizes of the noise, the tunneling rates and the circuit response time, the white noise will obscure the spikes in the current so that the observer must rely on the average current to distinguish between the two qubit states. The consequence of averaging out the white noise (by integrating the current I over some time τ ) is that if the qubit electron is tunneling on a time scale Ω −1 shorter than τ then the state of the quantum system cannot be followed.
We will now present an order of magnitude estimate of the effective bandwidth of the measurement device, which is defined as the frequency at which a signal (power) to noise (power) ratio of unity is obtained. Here we take the signal as being the current that flows through the QPC when it is in the more conducting state.
To find the noise and signal power we take the Fourier transform of Eqs. (16) and (17) in order to obtain the spectrum of the current I. 42 The signal power is and the noise power is Upon equating the signal and noise powers, and at this frequency setting ω = α eff , we have an effective bandwidth of where the dimensionless noise parameter N is defined according to The approximation of Eq. (39) holds in the limit where the noise power D o is small compared to the signal power r ′ e 2 (which is the regime that will lead to good measurements of the qubit state). For an observer to be able to follow the evolution of the qubit reasonably well we must have α eff > Ω. 35,44 VI. CONCLUSION
A. Summary
We have presented a new model for continuous measurement of a coupled quantum dot (CQD) charge qubit by a low transparency quantum point contact (QPC). We considered the evolution of this solid-state qubit conditioned on the output available to a realistic observer, which has been filtered and degraded (i.e. made more noisy) by an external circuit. This description is closer to the true conditioned evolution of the system, not a hypothetical evolution conditioned on the microscopic events occurring within the detector, to which a real observer has no direct access. Knowledge of the state of a quantum system conditioned on actual measurement results is essential for understanding and designing feedback control. 12,20,21,22,23,24,25,26,27 It is also expected to be important in quantum computing, both for state preparation and quantum error correction. 28,29,30 Our model for the conditional dynamics of the qubit due to measurement by a low transparency quantum point contact (QPC) was based on the quantum trajectory models of Refs. 13 and 14. We have presented a stochastic master equation that describes the time evolution of the measured qubit conditioned on a hypothetical detector output. Korotkov has derived equivalent conditional dynamics equations for the QPC (with notational differences) using a Bayesian formalism. 11,12,15,17,18 We generalized a realistic quantum trajectory theory 34,35 (recently developed for photodetectors) to treat solid-state detectors. The solutions of the resulting stochastic equations are the "realistic quantum trajectories" of the measured qubit state. These will be presented elsewhere.
B. Comparison to previous work
The conditional dynamics of continuously monitored CQD systems has been studied in Refs. 11,12,13,14,15,16,17,18,19. However, the model of realistic measurement that we have presented in this paper is new. Korotkov has recently presented a phenomenological theory 17 involving "non-ideal" detectors that is in the same spirit as ours (see Appendix A for a derivation of Korotkov's result using our stochastic master equation approach). However, he still assumed an infinite bandwidth detector and also assumed that the ideal detector could be described by diffusion rather than jumps. We believe that our approach offers a more satisfying description of this measurement process because the non-Markovian 55 effects of a realistic measurement circuit are included and the tunneling process through the QPC is described as a point process (jumps) as one would expect.
Finite detector temperature effects in the case of bare measurement were not considered in our model. The effects of a non-zero detector temperature T d have been considered previously 12,14,45 and result in an approximately linear (coth [eV d /2k B T d ]) increase in the ensemble decoherence rate and shot noise level with T d for eV d < 2k B T d , where V d is the detector bias voltage. With a detector temperature of the order of mK, 46 finite temperature effects could be expected to become important at bias voltages of V d < 0.3µV . These voltages are several orders of magnitude below a sample bias voltage for maximum response of a single electron transistor (SET) detecting the charge state of a quantum dot, 47,48 which suggests that our omission of finite detector temperature effects in the bare detector scenario is reasonable.
C. Future Work
There are many possibilities for future work in the theory of realistic quantum trajectories. Other detectors will be considered -for example, the single electron transistor 49,50 (SET). As the field of mesoscopic electronics is progressing at such a tremendous rate it is likely that the choice of detector will quickly become outdated. In fact, the SET has already been surpassed by the radiofrequency (RF) SET 51,52,53 as the measuring device of choice for the read out of the charge state of a mesoscopic qubit. This is one reason why we view the work in this paper as preliminary. The extension of realistic quantum trajectories to the RF-SET is a future aim.
Further work is also appropriate for the circuit model, which at present is considerably simplified, but is a good starting point which models some essential effects of a realistic measurement.
These and other possibilities for work on the theory of realistic quantum trajectories will be pursued in the future.
|
2017-10-24T04:56:39.086Z
|
2004-01-13T00:00:00.000
|
{
"year": 2004,
"sha1": "e490d768bb9fb65f7a8e694a7bb6c4219884e7ac",
"oa_license": null,
"oa_url": "https://espace.library.uq.edu.au/view/UQ:74829/UQ74829.pdf",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "ff61d862dbae4272cb4ebeea9d606f6f0c1ce3d1",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
}
|
7443029
|
pes2o/s2orc
|
v3-fos-license
|
Coronary Angiography Utilization and Costs for Coronary Artery Bypass Graft Surgery Patients in Turkey
Introduction To assess excess use of coronary angiography prior to coronary artery bypass graft surgery and its association with mortality, health care costs, and hospital quality in Turkey. Methods Using Turkish National Health Insurance Data (2009–2011) that included patients who underwent cardiac surgery, coronary angiography utilization was identified. Propensity score matching was used to compare survival rates and annual health care costs of patients in a coronary angiography excess-use group (>1 angiogram) and in a standard-therapy group (1 angiogram). The empirical Bayesian approach was used to combine mortality and hospital volume for quality index. The relationship between hospital quality and excess use of coronary angiography was assessed using Chi-squared tests. Results Out of 20,126 patients identified, 7.27% of patients underwent excessive coronary angiography procedures (excess-use group), with an average annual cost at 9.7% higher than those who had a single angiography (standard-therapy group; P < 0.01). Operational mortality associated with excessive use was significantly higher as well (7.4% versus 5.4%, P < 0.02). There exists variation in the use of coronary angiography across cities and hospitals. Patients who underwent cardiac surgery in high-quality hospitals were less likely to have excessive angiography use than those in low-quality hospitals (7.0% versus 9.5%, P < 0.01). Conclusion In Turkey, excess use of coronary angiography prior to coronary artery bypass graft surgery is associated with higher operational mortality, higher expenditures, and lower hospital quality.
INTRODUCTION
Coronary artery bypass graft (CABG) surgery is one of the most common surgical procedures performed to relieve angina symptoms in patients with coronary artery disease [1,2].
The surgery improves the quality of life and survival rates of patients with coronary artery disease [3][4][5].
Coronary angiography is considered the standard procedure for CABG assessment [6].
Every year, millions of patients get an angiogram, wherein a long, thin, flexible tube called a catheter is inserted in the arm or groin, and threaded to the heart to check for blocked arteries that may lead to a heart attack. Dye is injected through the tube to make blockages visible on X-rays [7]. Although angiographies are effective in select populations, the overall appropriateness of their use is questioned [8][9][10]. The test carries a small (\1%), but serious, risk of causing a stroke or heart attack, and also entails radiation exposure [11].
The troubling association between the increased number of CABG surgeries and high frequency of costly angiograms to check for heart disease has caused payers to request studies to outline the benefits, risks, and outcomes of these procedures. The cost of these procedures will be major contributors to increased health care spending [12]. Prior studies have indicated that excess use of and increased costs associated with these procedures are related to variation in the quality of care [13].
This study aimed to demonstrate the variation in the use of coronary angiography for patients who underwent CABG surgery in Turkey using nationwide, real-world data. The difference in mortality rates and annual health care costs associated with coronary angiography overuse was calculated. The correlation between the number of angiographies and hospital quality based on a previously validated claimsbased quality index was estimated. This article does not contain any studies with human or animal subjects performed by any of the authors.
Statistical Analysis
Paired t tests and Chi-square tests were used to compare the demographic, clinical, and provider characteristics between the two groups: standard-therapy group and excess-use group. To estimate risk-adjusted differences in mortality and annual health care costs with the overuse of angiography, the propensity score matching (PSM) method was used. This technique, employed by the predicted probability of group membership, isolates the observed bias from the estimation [16].
Following previously published guidelines for choosing the most appropriate matching technique for the present data, one-to-one matching with caliber was used [17]. Caliber was selected as 25% of the standard errors of the estimated propensity score. Using regression analysis, mortality rates and total health care costs for patients with standard therapy and excess use were also estimated. In particular, Cox regression was used for operational mortality estimation, and generalized linear models were used for total health care cost estimation for both groups.
Hospital quality was characterized by a previously validated composite score, which combines operational mortality with hospital volume information using Bayesian techniques [18]. Hospital procedure volume was assessed as the total number of procedures performed by index is presented in quartiles to rank low-to high-quality hospitals. Since this index combines information from multiple quality domains into a single measure, it has been referred to as the ''gold standard'' for hospital performance comparisons on US national standards of safety, quality, and efficiency that are most relevant to patients and payers [19][20][21].
All statistical analyses were conducted using SAS Ò V.9.3 (SAS Institute Inc., North Carolina, USA) and STATA Ò V11 (StataCorp LP, Texas, USA). In particular, SAS V.9.3 is used for data management, and STATA V11 is used for statistical analysis.
RESULTS
A total of 20,126 patients satisfied all inclusion criteria, of which, 88.12% (n = 17,734) underwent CABG surgery after one coronary angiography (standard-therapy group), and 7.27% (n = 1,464) had more than one coronary angiography (excess-use group) within 3 months prior to CABG surgery. The distribution of excess use is presented in Fig. 1.
Few CABG surgery patients (4.61%) did not utilize coronary angiography prior to surgery.
Less than 1% of these patients were admitted into the emergency care settings. The majority of surgeries were performed in high-volume hospitals (58.75%), and 11.16% of the surgeries occurred in teaching hospitals (Table 1).
Geographic variation in coronary
angiography use is presented in Fig. 2. To increase the reliability of the estimate, cities with\20 angiography procedures were not taken into consideration in the rates. The proportion of CABG surgery patients with excess use was the highest in Erzurum (13.36%), followed by Batman (12.28%), and Antalya (11.95%).
Thirty-six out of 81 cities did not have centers for coronary angiography, and 1,288 patients underwent CABG surgery without having a coronary angiography. A total of 33.17% Fig. 4). For the complete overview of the study population, patients who did not receive the angiography were also compared with the main study groups. After applying risk adjustment, standard-therapy patients use had lower mortality rates (3.96% versus 6.01%, P\0.010) and incurred lower annual health care costs (€7,850 versus €8,081, P\0.010) relative to patients who did not have angiography. Patients with excess use of angiography incurred higher annual health care costs (€8,695 versus €8,148, P\0.010) and had no significant operational mortality rate advantage (5.74% versus 6.09%, P = 0.675), relative to patients without angiography use.
In this study, angiography utilization in a disease-specific (CABG surgery) cohort of patients with uniform insurance coverage in Turkey was examined. Results determined that after adjusting for demographic and clinical factors, there was significant variation in coronary angiography use. Regional practice patterns emerged, showing that more procedures were performed in cities such as Erzurum, Batman and Antalya, relative to the rest of Turkey.
Current guidelines for patients undergoing assessment for obstructive coronary artery disease are designed to limit the number of invasive angiographies and thereby enhance the diagnostic yield of cardiac catheterization [26]. However, a substantial increase in these procedures has contributed to a rise in medical costs. Several factors are responsible for this increase, including wider availability of technology, increased demand by patients and physicians, and favorable reimbursements [27]. significantly [28]. Therefore, the rise in procedure utilization is expected. However, the association with excess use and hospital quality is of particular interest. Often, coronary costs for the excess-use group were 9.7% higher than for standard-therapy group.
Mortality rates were 1.37 times higher.
Hospital quality was assessed using a previously validated composite score, which has been applied by a large group of coalitions in the United States [19,21]. Traditional quality indicators, whether they are based on structure, process, or direct outcomes measures, have significant limitations. For example, hospital volume is important for high-risk procedures but does not reliably predict performance for individual hospitals. Process measures often fail to explain observed variations in hospital mortality rates. Direct outcomes measures, because of sample size limitations, are often imprecisely estimated and do not produce a reliable estimate of hospital performance. The composite measure used in this study has the potential to obviate these limitations, because it combines information from surgical volume and operational mortality into a single summary measure. This study presents a negative relationship between high-quality hospitals and excess use of coronary angiography.
Numerous studies have been done to evaluate factors that may influence the results of a variety of surgical procedures. Almost invariably, when assessed, surgeon experience is an important factor in determining the likelihood of successful outcomes from surgery. For example, in shoulder replacement procedures, a surgeon who performs 30 replacements over 6 years is considered ''high volume'' [29]. These doctors had lower complication rates and their patients had shorter hospital stays than ''low volume'' surgeons. In hip replacement surgery, surgeons who performed fewer than 30 hip replacements each year had more patients requiring revision surgery [30]. In another study, 5 years of scoliosis surgery experience did not improve outcomes [31]. Therefore, it is difficult to define a specific number of procedures that deem a surgeon proficient. In the present study, the authors empirically determined the effect of the volume of procedures by including them in the hospital quality index.
The present study has several strengths. To the authors' knowledge, this is the first study in Turkey that includes a large, diverse, and geographically varied sample of patients who received coronary angiography before CABG surgery. It is also the first study that shows the relationship between hospital quality and the excess use of coronary angiography. Excess costs associated with angiography overuse and operational mortality are also presented for the first time in Turkey.
Data were limited in their clinical information related to disease severity in coronary angiography patients. The Elixhauser Index and admission acuity were used in the attempt to control for differences. These measures are not representative of disease severity; therefore, the impact of severity on coronary angiography excess use may be underestimated. Although claims data are extremely valuable in revealing the real-world practice of procedures and determining [33].
Regression analysis uses the outcome as a lefthand-side variable, which is not supposed to be available during randomization. Second, confounding by treatment variables is often the main challenge to validity, and the propensity score focuses directly on treatment variables [34]. Third, a matched analysis can eliminate noncomparable exposed subjects [35]. Finally, PSM provides robust estimates when there are relatively few outcomes to compare with the number of potentially few outcomes [36][37][38]. Proponents of regression analysis, however, focus on statistical advances in regression literature [39,40] Table 4.
|
2018-04-03T03:39:14.034Z
|
2013-06-27T00:00:00.000
|
{
"year": 2013,
"sha1": "fd9a22165324729fbaa4c1433b1c46cad382017b",
"oa_license": "CCBYNC",
"oa_url": "https://link.springer.com/content/pdf/10.1007/s40119-013-0018-z.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "fd9a22165324729fbaa4c1433b1c46cad382017b",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
249661244
|
pes2o/s2orc
|
v3-fos-license
|
Recent progress in the CoCrNi alloy system
The exceptional mechanical properties, particularly at cryogenic temperatures, of the equiatomic CoCrNi alloy are documented in numerous published studies. Similar to the equiatomic CoCrFeMnNi (so called Cantor alloy), from which the ternary alloy was derived, the CoCrNi ternary possesses low stacking fault energy that promotes complex deformation modes, as well as the activation of deformation twinning at ambient temperatures and increased strain. In addition to outstanding deformation mechanisms, chemical short-range order and face-centered cubic (FCC)-hexagonal close packed (HCP) transitions have been verified in this alloy and prove to be key factors contributing to the alloy’s notable properties. The relationship between stacking fault energy and FCC → HCP phase transitions has been developed over the years through other low stacking fault materials, but the question that arises is: do well established physical metallurgical mechanisms require modification when applied to sys- tems such as CoCrNi given their compositional complexity? Local chemical order plays an important role in that it brings the deviation from the random solid solution behavior generally expected from complex concentrated alloys. In this review, the fundamental atomistic deformation mechanisms of the CoCrNi alloy will be reviewed with a focus on deformation substructures and chemical short-range ordering. Recent studies on microstructural engineering through thermo-mechanical processing and efforts to enhance the tensile properties of the CoCrNi derived systems with minor alloying additions are discussed. Finally, future directions of research, which involve applying current understanding of the underlying mechanisms towards alloy design strategies, are
Introduction
The development of high entropy alloys (HEAs) or so called complex concentrated alloys (CCAs) has become a predominate class within the field of materials science. The field's understanding of HEAs evolved significantly over the last two decades and has led to the development of alloys which might be future replacements for the existing conventional materials, such as high-strength steels and nickel-based superalloys [ 1 , 2 ]. One of these CCAs is the equiatomic CoCrNi, known for its excellent combination of tensile strength, ductility, and fracture resistance especially at cryogenic temperature, which will be discussed in detail in this review. This section of the review is dedicated to the evolution of CCAs and how the ternary CoCrNi, a subset of the equiatomic CoCrFeMnNi Cantor alloy, emerged as a prominent CCA system.
Conventional alloying strategies, where elements are added in minor amounts to a base metal, have been the common approach of developing metallic alloys since the Bronze Age. In 2004, two individual research groups [ 3 , 4 ] introduced the concept of concentrated multicom-ponent alloys or HEAs which deviates from the solvent-solute paradigm. It was found that the equiatomic composition of multi-component alloys stabilized solid solution phases over intermetallic phases and led to simpler microstructures than originally expected. Yeh et al. [2] attributed this behavior to the high configurational mixing entropy of equiatomic multicomponent alloys. While the motivation behind HEAs was to explore the central region of the compositional space, most of the early research was focused on alloys with equiatomic composition that yielded relatively simple microstructures with minimal intermetallic content. To avoid the barriers posed by specific definitions, terms such as CCAs and multi principal element alloys (MPEAs) were coined thus removing the requirement that the number of components need to be greater than or equal to 5, or that the value of configurational entropy be greater than 1.5R [3] . This shift is seen in recent publications where a vast number of compositions with no limits on the minimum number of elements are being explored, both single-phase and multiphase in equiatomic, as well non-equiatomic ratios with an understanding that configurational entropy alone cannot determine the stability of phases and that the role of enthalpy must be considered [4][5][6][7][8] . In this review, the term CCAs will be used to address this class of multi-component alloys. Our understanding is that increasing the number of elements does [11] and [15] , respectively. not necessarily result in improved properties. An example of this is the equiatomic CoCrFeMnNi Cantor alloy and its ternary derivates, particularly the equiatomic CoCrNi alloy. This alloy has become the focus of many investigations based on its unique mechanical properties which include tensile strength of ∼1 GPa at failure strain of ∼70%, and fracturetoughness above 200 MPa m 1∕2 . These properties synergistically increase even down to cryogenic temperatures which is attributed to a range of underlying mechanisms detailed in the next section.
CCAs with elemental 3d transition metals have been investigated by many researchers. Among them, the equiatomic CoCrFeMnNi was one of the earliest and most well-researched CCAs [9][10][11] . Since then, a significant amount of research has been conducted within this system exploring different alloying additions, non-equiatomic composition, mechanical behavior, corrosion behavior, and radiation damage tolerance [9] . The primary focus of studies on this alloy has been the mechanical properties, which show a room temperature yield strength of ∼400 MPa and tensile strength of ∼700 MPa and with an elongation of almost 56%. Gludovatz et al. highlighted the exceptional strength-ductility synergy in the equiatomic CoCrFeMnNi at cryogenic temperatures [11] . It was reported that as the temperature decreased from 293 K to 77 K, the yield strength and ultimate tensile strength increased by ∼85% and ∼70%, respectively with a simultaneous increase in the tensile ductility by ∼25%. The fracture toughness of the quinary alloy was also found to have a crack initiation toughness of 200 MPa·m 1∕2 or more which was retained even at cryogenic temperatures. These properties were attributed to the activation of deformation-induced nanotwins [12] as an additional deformation mechanism with the decreasing temperature giving rise to the dynamic Hall-Petch effect by introducing additional interfaces as well as delaying the onset of necking due to increased strain hardening.
In 2014, Wu et al. [ 13 , 14 ] investigated a total of ten binary, ternary, and quaternary subsets of the Cantor alloy with equiatomic compositions to understand their phase evolution and mechanical behavior. It was concluded that it is not just the number of elements but also the nature of elements that dictate the mechanical behavior of the CCAs, with Cr being a potent strengthener in the Cantor alloy family. Among all the subsets, this work highlighted the ternary CoCrNi alloy, with a singlephase FCC structure, having the highest strength and ductility in the temperature range of 77-673 K. This alloy was further studied by Gludovatz et al. [15] who observed activation of an additional deformation mode of nano twinning which led to a tensile strength of 1.3 GPa while displaying failure strains of 90% and fracture toughness of over 430 MPa m 1∕2 at 77 K. Fig. 1 compares the tensile properties of the equiatomic CoCrFeMnNi and CoCrNi at different temperatures. The most prominent difference between the equiatomic CoCrFeMnNi and ternary CoCrNi is the activation of twinning at room temperature as an additional mode of deformation in the latter. Activation of twinning increases as the temperature decreases or the strain increases.
Through the first section of this review, the current understanding of the atomistic mechanisms of the CoCrNi CCAs is discussed. Knowledge of stacking fault energy (SFE) and FCC →HCP phase transitions have been developed over the years through other low stacking fault materials however, the question that remains to be answered in the CoCrNi system is the role of compositional complexity. Accordingly, local chemical ordering and its effect on deformation mechanisms are discussed in detail. The next section of the review focuses on microstructural engineering to further enhance tensile properties of the CoCrNi alloy via thermo-mechanical processing. Lastly, derivative systems obtained by minor alloying additions to the base CoCrNi ternary are reviewed with an emphasis on their effect on mechanical properties.
Atomistic mechanisms driving mechanical behavior
The synergy between strength and ductility in the CoCrNi CCAs can be attributed to the underlying atomistic mechanisms. Besides that, the constituent elements also play an important in determining the mechanical properties. Since the individual elements In CoCrNi alloy have similar atomic radius, one would expect lattice distortion to be quite less. But counterintuitively, the solid solution strengthening effect in the ternary CoCrNi is found to be high [16] . It has been established that in case of CCAs, the lattice distortion and hence the solid solution strengthening effect is less related to the atomic radius and directly proportional to the atomic volume misfit and/or shear modulus [17] . The calculated atomic volumes for Co, Cr and Ni are respectively 11.12, 12.27 and 10.94 A 3 [17] . Hence, there is an appreciable volume misfit that leads to solid solution strengthening. Additionally, it has also been found that since Cr has a larger elastic modulus in FCC structures, Cr-containing alloys have higher shear modulus, which is the case with CoCrNi. That is why solid solution strengthening effect is high in CoCrNi. Another study by Oh et al. [18] highlights that a relatively large electronegativity difference between the individual atoms of Co, Cr and Ni also contribute to the increased solid solution strengthening effect in CoCrNi. All these effects together lead to the relatively high yield strength of the ternary CoCrNi alloy. While the effect of individual elements is significant, the mechanical behavior of these CCAs is heavily dependent on the atomistic mechanisms the are highlighted in detail in this section. We discuss the evolution of deformation substructures in the CoCrNi ternary alloy enabling it to overcome the strength-ductility tradeoff. The influence of SFE on the phase transition from FCC to HCP is then highlighted and the probability of a reverse HCP to FCC transition is reviewed. In the last part of this section, the presence of local chemical order within the CoCrNi alloy is discussed with an emphasis on how it can further affect SFE, deformation mechanisms, and eventually the mechanical properties.
Evolution of deformation substructures
The stacking fault energy (SFE) of a metal is a key factor in determining the dominant deformation mechanisms. In the case of face-centered cubic materials, when a perfect dislocation dissociates into two Shockley partials, the region between these two defects is a stacking fault (SF) and the energy associated with its formation is known as SFE. The width of a SF is inversely proportional to the SFE. The equiatomic CoCrNi alloy is a low SFE material with SFE between 18 and 45 mJ m − 2 , similar to MP35N/ MP159 alloys, high-Mn austenitic twinning-induced plasticity (TWIP) steels and the CoCrFeMnNi Cantor alloys. The low SFE causes the SFs to be wide between the Shockley partials which makes cross slip energetically difficult and leads to activation of twinning as a secondary deformation mechanism. It is expected that as the SFE decreases, the deformation mechanisms change from dislocation glide to dislocation glide + twinning and further to dislocation glide + FCC →HCP phase transformation, or the ɛ -martensitic transformation in the case of steels [ 19 , 20 ], It is interesting that twinning in CCAs leads to a simultaneous increase in both tensile strength and ductility [21][22][23][24] . Twinning leads to the formation of a new interface that acts as an obstacle to dislocation motion, while also contributing to work hardening by delaying necking instability. Fig. 2 (a) shows the activation of multiple twins in different directions and with different widths in CoCrNi during in situ straining experiments. Curtze and Kuokkala [20] suggested the formation of a 3D twin architecture inside individual grains (within the size range of ∼5-50 microns) of an equiatomic CoCrNi alloy that stabilized the twin boundary even when the dislocations glide through it as shown in Fig. 2 (b). Hence these twins contribute to plastic deformation while also strongly impeding dislocations therefore increasing the strength of the material.
The propensity for twinning increases with decreasing temperature and increasing strain. That is the reason why twinning activates as a secondary deformation mechanism in austenitic TWIP steels and in the Cantor alloys at lower temperatures (between 77 K to below RT) and at higher strain values [ 12 , 20 ]. On the other hand, the ternary CoCrNi shows twinning formation even at room temperature which further intensifies towards cryogenic temperatures. This behavior was attributed to the SFE in equiatomic CoCrNi being 25% lower than that of the equiatomic CrMnFeCoNi and reflects in the mechanical properties of CoCrNi showing higher strength, ductility, and fracture toughness than the CrMnFeCoNi both at room temperature and cryogenic temperature [ 15 , 23 ]. There is a stark difference in experimental and predicted values of the SFE in the CoCrNi ternary alloy. While experimentally the SFE is estimated to be in the range of 18 ± 4 mJ m − 2 [25] to 22 ± 4 mJ m − 2 [23] , the values calculated through first-principle calculations are generally negative: -24 mJ m − 2 [22] , or -41 mJ m − 2 [26] . Simulations and calculations are generally performed at 0 K where the entropic contributions can be neglected, however at a finite temperature, e.g., room temperature, the atomic vibrations are appreciable enough to be considered while computing SFE values. This can be one explanation for the difference in experimental and simulated values. Also, it is worth mentioning that experimental measurements cannot be negative as the SFs observed during imaging have to be at finite distance to each other which will always lead to positive SFE values. Despite the inherent limitations of both experimental and theoretical approaches, such variation highlights that the compositional complexity of CCAs needs to be considered, as the variation in local atomic configuration will affect the SFs and corresponding SFE.
At the most fundamental level, SFE depends on the electron orbital interactions and the free electron density. For a pure FCC metal, it was demonstrated through ab-initio calculations [27] that for a metal like Cu, the charge distribution is nearly spherical which makes redistribution of electrons easy after shearing the close-packed plane to form a SF, hence reducing the energy penalty and leading to a low SFE value. On the other hand, Al has an angular charge distribution that poses difficulty in electron redistribution, giving rise to a high SFE value. Also, the d-electron density plays a significant role during shearing caused by the formation of SFs and hence plays an important role in determining SFE [28] . Moving from pure metal to conventional alloys, the addition of another element to the base element will also have its effect on the charge density distribution and atomic bonding. If an alloying element causes more charge accumulation, the directionality of metallic bonding is decreased which leads to a decrease in SFE and vice-versa. It was observed in the case of Co-based alloys [29] that the addition of Cr, W, and Mo leads to weaker atomic bonding due to increased charge distribution in the interstitial regions which in turn decreases the generalized SFE of Co. Whereas, Ni, Mn, Al, and Fe have the opposite effect and increase [32] . the generalized SFE of Co. Hence, every individual element has a unique effect on the SFE of the alloy and the same holds true for the CoCrNi alloy. Besides the electronic configurations, volume and magnetic effects also contribute to the determination of SFE and the same element may also have a different impact on a different base alloy [30] . This review, however, is concerned with the effect of alloying elements in the base CoCrNi ternary which is discussed in detail in a later section.
FCC →HCP phase transformation
Glide of a single Shockley partial on a closed pack plane in FCC stacking sequence generates an intrinsic stacking fault (ISF) whereas when the partial dislocations glide on two consecutive closed packed planes, an extrinsic stacking fault (ESF) is generated [31] . Fig. 3 shows how ESF leads to the formation of deformation twins while ISF serves as an embryo for the HCP phase [32] . Therefore, an ISF the FCC sequence can be seen as a layer of HCP stacking which can act as a favorable site for the growth of the HCP phase. The difference in energy of formation of these FCC and HCP phases can be related to the SFE. A low SFE suggests that the energy of formation of both FCC and HCP phases will be quite close to each other and other factors such as temperature and strain will stabilize one phase over another. Deformation-induced phase transformation, analogous to transformation-induced plasticity (TRIP) effect in steels [33] , has been observed in other CCAs as well [34][35][36][37] .
Compared to CoCrFeMnNi, where FCC →HCP phase transition was induced at high pressure [38] , HCP phase formation seems to be more convenient in CoCrNi. Using spin polarized density functional theory (DFT) calculations [39] , it was predicted that at 0 K, the HCP phase has lower energy than the FCC phase in CoCrNi. It can be reasoned that the higher Co (which is an HCP phase) content in CoCrNi alloy compared to CoCrFeMnNi stabilizes the HCP phase. A similar observation was made by Zhao et al. [40] in CoCrNi x alloys where single-phase FCC structure formed for x = 1 to 0.8 but as the Ni content decreased from x = 0.7 to 0.5, eventually increasing the Co content, both FCC and HCP phases coexisted. There is a higher probability of HCP phase formation in equiatomic ternary CoCrNi, than there is in the quinary CoCrFeMnNi. Niu et al. [26] rationalized this behavior by taking the magnetic frustration into account which is strong in the case of CoCrFeMnNi resulting in comparable FCC and HCP energies and hence no driving force for HCP transformation. In the case of CoCrNi, however, even if one considers magnetic effects the HCP phase remains more stable than the FCC at lower temperatures.
For the equiatomic CoCrNi alloy, Miao et al. [39] reported a similar FCC to HCP phase transformation at higher strain levels both at room and cryogenic temperatures. The deformation substructure followed an expected pattern where at lower strain mainly planar dislocation slip existed and with increasing strain, nanotwins and HCP phase evolved and dominated the deformation. This work highlighted the formation of nanotwin-HCP lamella as shown in Fig. 4 , suggesting that the HCP phase forms due to the interaction of the SFs and nanotwins. The HCP phase grows and remains localized in this interaction zone throughout. Even though the volume fraction of the HCP phase increases with increasing strain, more so at the cryogenic temperature than at RT, it should be noted that the highest amount of HCP phase detected by this work is only ∼3% in a sample deformed to 53% true strain at cryogenic temperature. Another large-scale molecular dynamics simulation study by Yuan et al. [41] showed the HCP phase formed due to nucleation and propagation of multiple SFs at adjacent slip planes from the same grain boundary. The HCP phase in CoCrNi CCAs can be an additional strengthening mechanism in CoCrNi based alloys and can have a synergistic effect in improving the mechanical properties [42] .
Contrary to TRIP steels, where a large fraction of FCC austenite transforms to HCP ɛ -martensite [43] , the volume of transformed HCP phase in CoCrNi ternary alloy is quite low. This happens despite the sufficient driving force due to combined combination of low SFE, temperature and high strain. In their work, Niu et al. [26] suggest that the mixed dislocations trapped at the interface of the HCP region may facilitate the formation of new Shockley partials, causing the structure to heal back to its original FCC form. This explanation needs to be validated by further study. Previously, such reverse phase transformation from HCP to FCC has been suggested in another CCA system [36] Fe 50 Mn 30 Co 10 Cr 10, where the bidirectional phase transformation was attributed to local dissipative heating and stress-strain fields acting as the driving force for HCP to FCC transformation.
While there are no reports of any HCP to FCC transformation in bulk CoCrNi, magnetron sputtered thin films of CoCrNi undergo a reverse transformation from HCP to FCC [ 44 , 45 ] when subjected to in situ tensile testing. Chen et al. [44] attribute this bidirectional FCC →HCP →FCC transformation to the small SFE of CoCrNi alloy and hence the flexible stacking sequence which gives the unique capability of facile slip on the close-packed {111} FCC || (0 0 01) HCP planes in both FCC and HCP Fig. 4. Annular dark field (ADF) -scanning transmission electron microscope (STEM) image of CoCrNi specimen tested to 53% true strain at cryogenic temperature showing HCP band forming near the twin boundary. The inset in (b) is the center of symmetry (COS) map also highlighting HCP stacking [39] .
phases of the CoCrNi alloy. It is to be noted that the reversible phase transformation occurs mostly near the central region of the micropillars which undergoes the highest deformation through shear banding during in situ testing. In another similar work Chen et al. [45] showed the dependency of deformation on the size of the pillars where high shear stress in sub-micron pillar can be enough to facilitate a lattice rotation along the basal plane of the HCP phase causing an increase in resolved shear stress and the glide of partial on every other basal plane of HCP stacking promoting the phase transformation FCC inside the shear band. While the phase transformation in the CoCrNi thin film is insightful and leads to superior yield strength up to ∼4 GPa, parallels cannot be drawn between the thin film and bulk samples because of the inherent difference between them, namely: (1) the initial grain size in thin films is small and one of the reasons of high yield strength, (2) sputtered thin films have a strong texture which is generally not seen in bulk sample, (3) the area of the film considered in these works consisted of 60% HCP and 40% FCC phase to start with. The high density of SFs already present in the HCP phase also promotes the HCP to FCC transformation, while in the case of the bulk sample, the initial microstructure is essentially single-phase FCC. There is a need to explore these sequential FCC →HCP →FCC transformations with respect to CoCrNi related alloys in bulk samples. The limited knowledge from the thin film work can be applied and extrapolated to nanocrystalline bulk samples and the possibility, as well as conditions, of bidirectional transformations need to be investigated.
Chemical short-range order
Majority of early research assumed CCAs exhibited completely random solid solution phases where each element has an equal probability of occupying any given lattice sites [2] . Over the years, the importance of enthalpic interactions in determining phase stability at lower temperatures has been recognized. Experimental evidence demonstrates that previously considered single phase CCAs exhibit phase decomposition at intermediate temperatures, confirming that the single-phase state is a high temperature state of the alloy, metastable at room temperature through quenching from high temperature. It was also predicted that in CCAs, instead of a random distribution of elements, there are regions where an element has more preference to chemically bond with some elements than others. The deviation from the random solid solution distribution may lead to formation of nanoclusters, long-range order or short-range order. In general, nanoclustering refers to preferential segregation of like elements together locally in form of nanoprecipitates [46] . Long-range order refers to an arrangement where some elements occupy one sublattice while some occupy the other sublattice with a strict distinction between the two sublattices. Secondary precipitates with L1 2 , B2, etc. ordered lattice structure correspond to long-range order in CCAs. On the other hand, chemical short-range order (CSRO) occurs when an element prefers/avoids other elements as the nearest or second-nearest neighbor [ 47 , 48 ] CSRO is fundamentally a fluctuation in the local atomic arrangement that deviates from the random arrangement of the CCA matrix, and no distinct lattice structure is formed. CSRO can influence the SFE, deformation mechanisms and mechanical properties of CCAs which will be discussed in this section, with a focus on the CoCrNi ternary alloy.
One of the early efforts to detect CSRO in CoCrNi ternary alloy was made by F.X. Zhang et al. [49] using EXAFS measurements, where these authors observed that Ni-Cr and Co-Cr bonds had smaller Debye-Waller factors. This implied that these bonds are favorable and could lead to short-range ordering within the system. Another study by Zhang et al. [50] , using energy-filtered TEM, verified that CSRO was present in the CoCrNi alloy which was homogenized at 1200 ˚C and aged at 1000 ˚C, followed by furnace cooling. No such ordering was found in case of the alloy homogenized and quenched to room temperature. Similarly, CSRO was observed by Zhou et al. [48] in alloys homogenized at 1100 ˚C and recrystallization annealed at 600 and 1000 ˚C. Characterizing CSRO has been a challenging task for the CCA community because of the need to acquire detailed information at the atomic scale. Experimental efforts are underway to understand CSRO in the CoCrNi system; theoretical studies also predict the formation of CSRO in this system [51][52][53] . Walsh et al. [52] explained bonding preferences using the magnetic interaction within the CoCrNi system that dominates the CSRO. According to their spin-polarized DFT calculations: (1) Frustration of magnetic moment in case of Cr pair make them unfavorable; (2) Anti-alignment of Co and Cr spins is preferred which agrees with the tendency to form Co-Cr pairs; and (3) Replacing Co with Ni helped in reducing the same-spin Co-Cr, hence supporting the favorable Ni-Cr bonds earlier predicted. Ding et al. [53] showed similar ordering arrangement through Monte Carlo simulations using Δ k ij (i,j = Cr/Co/Ni) a parameter to quantify local chemical ordering around individual atoms in CoCrNi. Positive values of this parameter as shown in Fig. 5 The occurrence of CSRO in the CoCrNi alloy is particularly interesting and not intuitively predicted as all the three elements have similar atomic radii and show a small heat of mixing. In the same way, the equiatomic ternary CoCrNi is the stable single phase FCC material. Though no secondary phase has been detected experimentally when the alloy is conventionally cast, the possibility exists that the FCC phase could be a metastable phase. Other phases may appear at intermediate temperature during ideal equilibrium conditions as predicted by Ther-moCalc (shown Fig. 5 (b)). Hence, the arrangement of atoms leading to [54] ; (d) Probability density distributions of local SF energy of random solid solution (with no short-range order) and for samples annealed at 650, 950, and 1650 K, respectively (higher the temperature, lesser the ordering) [55] .
CSRO could be an artifact of the potential secondary phase formation. This hypothesis needs to be backed by both experimental and theoretical studies. Though the reason behind CSRO in CoCrNi needs more investigation, several recent studies do confirm its presence in the ternary alloy, and it is important to understand the factors that can tailor the CSRO. The local chemical environment in CoCrNi as well as other CCAs is not just determined by the thermodynamically lowest energy configuration but also by the heat treatment and kinetics of thermal processing. Generally, lower annealing temperature can decrease the entropic effects and make it more favorable for the atomic pairs with lower energy to bond and hence give rise to chemical ordering within the system [ 55 , 56 ]. Besides the heat treatment temperature, the cooling rate could also be a deciding factor. For instance, in the work done by Zhang et al. [50] , CSRO occurs in alloy that is aged, and furnace cooled. In contrast, in the study conducted by Yin et al. [57] , the alloys were not slowly cooled and no CSRO was found in the CoCrNi alloy. As important as the ageing temperature is, the slow rate of furnace cooling could allow the sample to acquire the energetically favorable CSRO arrangement. Hence, the cooling kinetics of the processing could play a significant role in promoting/suppressing CSRO. More experimental evidence can further establish the connection between heat treatment temperature, cooling rate and CSRO in the CoCrNi alloy.
Such a processing-structure relationship gives opportunities to not only tailor the CSRO but also effects the SFE and deformation substructures. This kind of CSRO, if in the vicinity of SFs, will influence the SFE locally. Ikeda et al. [54] highlighted how the intrinsic SFE can vary based on the local concentration of individual elements which is depicted in Fig. 5 (c). Such dependence of SFE on the local chemistry is one of the reasons for the vast difference even in the computed SFEs for the CoCrNi equiatomic ternary [ 25 , 26 , 28 , 58 ]. Through DFT-based MC simulations (at 0 K), Ding et al. [53] showed that there is a large distribution in the predicted SFE values which correspond to the heterogeneity in local chemical order within the sample under consideration. It was also observed that the SFE increased with increasing CSRO [50] which can be correlated to the increase in energy requirement to not just create a fault in the FCC structure but also disrupt the ordered state near it. This can further be related to the variation in computed and experimental SFEs where the calculated SFEs are generally negative while the experimental SFE was around ∼22 mJ m − 2 . The effect of temperature has already been mentioned but it is equally important to consider the incorporation of the possible short-range order while simulating or calculating the SFE. For example, Li et al. [55] showed that in their calculation the complex SFE in the case of CoCrNi as random solid solution is -24 mJ m − 2 but with taking local chemical ordering into account the CSFE varies from ∼1.4 to ∼57 mJ m − 2 as annealing temperature varies from 1650 to 650 K as shown in Fig. 5 (d).
The variation in CSRO, hence the variation in SFE, will influence a range of phenomena within the CoCrNi alloy including the deformation mechanisms, phase transformations, and defect kinetics. As mentioned earlier formation of short-range order implies extra energy expenditure for the motion of dislocations making them difficult to move. Moreover, deformation twinning and FCC →HCP phase transformation in CoCrNi is dependent on the generation of SF which can disrupt the ordering and hence cause an increase in energy. And this effect will vary throughout the sample owing to the heterogeneous nature of the ordering and SFE making it a complex competition between dislocation slip, twinning, and phase transformation. Besides the deformation mechanisms, the CSRO can also intensify the heterogeneity in inter-diffusion of vacancies and interstitials limiting their jump due to added energy barriers. CSRO will also make the grain boundary migration difficult due to the same reason. These effects will stabilize defects like vacancies, interstitials, and grain boundaries by limiting their motions, which has been exploited to develop resistance to radiation damage in the CoCrNi alloys [59] .
CSRO in the CoCrNi could also affect the mechanical properties of the system. It was found that the hardness of the aged alloy with short-range order was ∼0.3 GPa higher than that of the quenched alloy, while the tensile tests show an approximate 25% increase in yield strength for the alloy with SRO. Moreover, the initial work hardening rate in the aged sample is double than that of the water quenched sample. While these results indicate the potential strengthening effects of CSRO in this alloy system, a theoretical study performed by Yin et al. [ 57 ] emphasized that CSRO may not have much influence on the strength and hardness of the CoCrNi alloy, instead the strengthening effects are majorly from misfit volumes. It is difficult to conclude the effect CSRO has on mechanical properties of the CoCrNi alloy with limited experimental data. Further efforts are needed in the direction of characterizing the CSRO, verifying the predictions made by simulation and understanding the effect of CSRO on mechanical properties of CoCrNi alloy.
In summary, whereas the SFE may be at the center of all the fundamental mechanisms of the CoCrNi alloy and is conveniently generalized, it is important to understand that it can be influenced by other factors such as composition and local chemical ordering which will lead to local SFE variations within this material. Both experimental and theoretical calculations have their pros and cons, and hence should be analyzed accordingly. While the twinning and phase transformation in CoCrNi is similar to existing steels, the basic difference between CCAs and conventional alloy should be considered while drawing the comparison. Moreover, CSRO should be given special importance given its ability to change the basis of all the mechanisms. Careful control of the thermodynamics and kinetics of processing of these alloys can be useful in tuning the CSRO. Despite their complex nature, efforts need to be made towards the characterization of CSRO, both qualitatively and quantitatively, along with establishing the relationship between short-range order and mechanical properties.
Microstructural features and their role in modifying mechanisms
Microstructural features and atomistic mechanisms in the CoCrNi alloys are interdependent, allowing researchers to employ microstructural engineering to further enhance mechanical properties. As an example, the equiatomic CoCrNi alloy shows an excellent combination of ultimate tensile strength, ductility, and fracture toughness at micrometer level grain size, however, its yield strength at room temperature can yet be improved. In this section, we review and discuss the ways in which microstructure can be designed to affect deformation mechanisms as well as mechanical properties. Firstly, the techniques such as thermo-mechanical processing and torsional deformation processing are discussed to improve the mechanical behavior of CoCrNi alloys. More-over, surface modification processes and grain size effects are discussed in the latter part of the section.
Gradient microstructures
One of the ways to improve the yield strength of the ternary alloy, without any composition change, is to tailor heterogeneity in microstructure in such a way that it is a composite of harder regions that give the strength and softer regions that accommodate strain, hence mitigating the strength -ductility tradeoff. In the case of the CoCrNi alloy system, the gradient does not necessarily relate to grain size but also to deformation substructures: dislocations, SFs, twins, and the HCP phase. A good portion of research related to CoCrNi is focused on achieving heterogeneous microstructure through gradients in either recrystallized/partially recrystallized grains or gradients in strain/deformation. Both approaches are discussed below.
Gradient in recrystallization
One of the most common techniques to achieve a gradient microstructure is subjecting the alloy to cold rolling and annealing in a way that a partially recrystallized microstructure can be obtained [60][61][62][63][64][65][66] . Slone et al. [67] investigated the cold-rolled CoCrNi alloy after annealing at 873 K for 1 hand 4 h which exhibited partially recrystallized structure with 39 and 76% of the volume being recrystallized respectively and the non-recrystallized structure strained the twins and dislocations produced during cold rolling. The sample annealed at 873 K for 4 h showed an increased yield strength of around 797 MPa with a uniform elongation of 19% which could be attributed to the variation in local strain along the sample as shown in Fig. 6 where the recrystallized regions deform more than the non-recrystallized ones. The recrystallized grains, being softer, can accommodate more strain and are ductile, while the non-crystallized regions, having high dislocation densities and twins, contribute to the hardening and strengthening of the material. When combined, such a heterogeneous microstructure gives the optimum combination of both strength and ductility. Moreover, such heterogenous structures also perform well under impact loading [66] . Apart from cold rolling, cryo-rolling and subsequent annealing are also potential thermo-mechanical routes for CoCrNi as the cryogenic temperature will further intensify the deformation mechanisms in the system. Along this vein, Zheng et al. [68] observed that cryo-roll deformed CoCrNi consisted of dislocation structures, nano-scale deformation twin bundles, and nano-sized grains which partially recrystallized on annealing at 973 K for 10 min, leading to a yield strength of ∼1123 MPa and a uniform elongation of 31.2%.
Gradient in deformation structures
Another useful technique to achieve a heterogeneous microstructure for a single-phase alloy is through torsional deformation. When a CoCrNi alloy is subjected to high-pressure torsion (HPT) [69] , the surface of the rod that undergoes severe torsional deformation will give a high density of twin and dislocation density while the center of the rod which is deformed the least shows a low density of deformation structures. HPT in combination with annealing [70][71][72][73] can be used to tailor grain size as well as obtain heterogeneous deformation structure. Liu et al. [69] through experimental and theoretical study described the heterogeneity induced by high-pressure torsion in the microstructure and hence the properties which are shown in Fig. 7 . The optimum torsional angle for HPT that gives a good combination of strength and ductility needs to be determined. In this case for an angle of 20 , the yield tensile strength increased from 245 MPa (for initial undeformed sample) to 760 MPa while ∼31% uniform elongation is maintained. Another torsional deformation process like hot swaging has been utilized for the same purpose of creating a gradient microstructure [ 74 , 75 ].
Surface modification
Manipulating surface microstructures using severe plastic deformation (SPD) techniques such as shot peening and ultrasonic nanocrystal surface modification (UNSM) enhances, not only the mechanical properties, but also the fatigue and corrosion resistance. The SPD techniques generally form a gradient nanostructured surface to introduce residual compressive stresses that help in preventing the onset of failure [76] . The interior of the material remains relatively undeformed leading to gradual transition in grain size and strain from surface to the interior. The accumulation of geometrically necessary dislocations at the coarse-fine grain interface activates back-stress hardening [77] . These techniques have been utilized to improve strength of metallic alloys while maintaining their ductility [77][78][79][80][81] . For instance, UNSM processed Fe x (CoNi) 90-x Cr 10 medium-entropy alloy [81] exhibited yield strength twice as much as that of as-annealed specimens. The yield strength and tensile strength of UNSM treated Fe 57.5 (CoNi) 32.5 Cr 10 is 528 MPa and 676 MPa respectively as compared to 223 MPa and 513 MPa of the as-annealed material. Similar observations were made for CoCrFeMnNi alloy [79] where the UNSM treatment led to enhanced strength and the loss in ductility was directly proportional to the UNSM load. Additionally, the tendency of nano-sized deformation twinning in areas closer to the surface and this effect increases with the increase in the load applied during the UNSM treatment. The CoCrNi alloy sys-tems can show similar, potentially enhanced gradients of deformation structure and hence increased synergy between strength and ductility. Also, it is important to understand if such SPD techniques can lead to any possible phase transition within the system and how it affects the mechanical properties.
Modifying critical twinning stress by changing the grain size
Besides a favorable SFE, for twins to nucleate and grow, it is important that the stresses within the materials exceed the critical stress required to generate twinning. Since the nuclei for twins, i.e., SFs, already exist in low SFE materials, this stress mainly corresponds to the growth of twins. If the critical twinning stress is above the ultimate tensile strength, twinning will not be activated. Laplanche et al. [23] calculated the twinning stress for both equiatomic CoCrNi and CrMnFeCoNi to be 260 ± 30 MPa and 235 ± 10 MPa, respectively and emphasized that it is easier to reach the twinning stress in case of ternary alloys because of its higher yield strength and work hardening rate. It is to be noted that this calculation was done at a fixed grain size. As mentioned previously, the extent of twinning increases as the temperature decreases because the twinning stress does not vary much with temperature, whereas flow stress increases rapidly as the temperature decreases making it easier to reach the critical twinning limit at lower temperatures leading to the higher extent of twinning [20] . While twinning stress shows weak de- pendency on temperature, it can be considerably affected by grain size. For instance, in the case of TWIP steel with a composition 15Mn-2Al-2Si-0.7C (wt.%), it was found that the critical twinning stress increases with decreasing grain size [82] . This grain size-twinning stress relationship has two major implications. First, it suggests that thicker (or wider) twins will form in coarser grains while the fine grains will have thinner twins; therefore, coarse grains will be able to accommodate more plastic strain during cyclic loading. Second, in the case of broader grain size distribution, twins will be found in only some grains while the rest will contain planar slip [83] . Though no direct relationship has been established for critical twinning stress and grain size for CoCrNi alloy, there might be a possibility that similar grain size effects are also observed for CoCrNi and an optimum grain size may define critical twinning stress of CoCrNi in a way that the twins and dislocations can synergistically contribute to the mechanical performance of the CoCrNi alloy.
In summary, the presence of a gradient in the microstructure, whether through surface modification and/or in the form of grain size, the extent of recrystallization, or the density of deformation structures, can be introduced through a series of deformation processing and thermal treatments. If engineered properly, the strength-ductility tradeoff can be mitigated. Another important property to consider while subjecting these alloys to thermo-mechanical processing is their effect on the fracture toughness of the CoCrNi which is quite good in the fully recrystallized state. Further work is needed in this direction. It will be also beneficial to examine the effect of such heterogeneous microstructure on the cryogenic properties of this material.
Derivative systems
This section covers findings related to derivative systems which are defined as those obtained via some minor alloying addition to a base ternary system to enhance solid solution strengthening, and in some cases, precipitation strengthening as well. This section focuses on five important alloying additions made to the CoCrNi systems: Aluminum, Titanium, Molybdenum, Tungsten, and Silicon. The influence of these elements on the phase structure and mechanical properties of the derivative alloy is discussed for specific systems as well as in the context of the overall influence of alloying additions on atomistic mechanisms on CoCrNi.
Some of the relevant properties of the base elements and alloying additions are listed in Table 1 . If additional phase/s appear there could be a synergy between solid solution strengthening, precipitation strengthening, and grain boundary strengthening. In related studies, Hu et al. [84] pointed out the significance of self-diffusion activation energy (SDQ) of alloying elements which revolves around the idea that a higher drag will be imposed by the solute with a low diffusion rate which in turn will inhibit grain coarsening and vice versa. Alloying elements may also delay recrystallization leading to finer grain size. These effects are common for conventional alloys and other CCAs as well. In the case of CoCrNi, it is also important to account for the effect these alloying elements will have on the SFE, CSRO and hence deformation twinning and FCC →HCP phase transformation.
Aluminum
One of the important alloying additions that is added to the base ternary CoCrNi is Al. Al is a strong BCC phase former and due to the atomic size difference will distort the lattice. It has been shown that a single-phase FCC structure is stable with up to 7 at.% addition of Al [86] . Lattice distortion in the matrix is increased and consequently, the yield strength increases as the Al content increases from 0 to 7 at.%. Al also reduces the mass density of the alloy. Lee et al. [87] investigated the alloy with composition Al 7.5 CoCrNi which showed an additional B2 phase in the FCC phase. It was observed that the propensity for twinning was reduced in the sample. Though the effect of Al on SFE was not clear, Al led to the refinement of grains and was found to increase Table 1 Atomic radius [85] , crystal structure, and melting point [5] [88] studied Al x CoCrNi alloys and found that at low Al concentration (x < 12 at.%), the alloys had a singlephase FCC structure which evolved to a dual-phase FCC + BCC structure for 12 at.% < x < 22 at.% and to a double BCC phase structure for 22 at.% < x < 32 at.%. This observation highlights the role of Al as a BCC phase stabilizer. The yield strength increases, but at the expense of ductility, as the Al content increases which contributes to the formation of hard BCC phases. It should also be pointed out that for Al concentration greater more than 22 at.%, the alloy has a composition similar to a quaternary alloy than the ternary alloy and the results may not be directly compared to the ones with minor alloying additions. Nevertheless, the general trend observed with the Al addition involves, formation of the BCC phase above a certain Al concentration accompanied by an increase in yield strength and a decrease in elongation. While Agustianingrum et al. [86] did not observe any secondary phases below 7 at.% Al, work done by Sathiyamoorthi et al. [89] shows an FCC + B2 + phase for the Al 0.3 CoCrNi system. One of the basic differences in both cases lies between the processing and heat treatment of the alloy which has a significant impact on the initial microstructural conditions and phase stability. Processing details and the results obtained in the studies mentioned above are listed in Table 2 .
Titanium
Another widely used alloying addition is Ti. Specifically, when Al and Ti are added to the CrCoNi system, in the right proportion, the ´ phase forms which is already known to yield excellent properties in Ni-based superalloys [90] . The ´ phase has a L1 2 ordered FCC structure which is generally coherent with the disordered FCC matrix and contributes to the strength of the alloy without compromising ductility. With the same understanding, Al and Ti are generally added in mutual proportions to the base CoCrNi alloy giving rise to an FCC + L1 2 dualphase structure. Zhao et al. [91] studied the (CoCrNi) 94 Al 3 Ti 3 alloy and found two different precipitate distributions as shown in Fig. 8 ; one originated from the grain boundaries and precipitated in a discontinuous fashion, and the other in the interior of the grains where continuous precipitates with spherical morphologies were observed. Both types of precipitates were rich in Ni, Al, and Ti. The (CoCrNi) 94 Al 3 Ti 3 exhibited 750 MPa yield strength while maintaining good elongation of 45%, in contrast to a yield strength of 430 MPa in CoCrNi, a 44% increase. Another important observation made by this study is that the deformation twinning was less prevalent in the alloy with Al and Ti and that deformation was mainly dominated by SFs. While the effect of alloying on SFE was not calculated it was emphasized that the small width of the gamma phase, presence of very fine precipitates, and the anisotropy of the dislocation mobility might have impeded the nucleation of mechanical twins.
The same composition was studied by another group of researchers [92] who also observed L1 2 phase formation, though only continuous spherical precipitates and similar effects on mechanical properties were observed. Their study highlighted the slow coarsening rate of the precipitates when aged at 800 ˚C for various times. The precipitate particles maintained their spherical morphology and nanoscale size during all the aging time which was contributed to the small interfacial energy of these coherent precipitates and possible slow diffusion behavior in the CCAs in general. In a study of different alloy composition (CoCrNi) 94 Al 1.5 Ti 1.5 by Guo et al. [93] no precipitates or secondary phases were observed but twinning was found to be reduced when the alloy was subjected to two different degrees of torsion both at room temperature and 77 K. It was again correlated to the increase in critical stress for twinning and the bending effect of torsion on the twin boundary which makes the twin boundary deviate from the ideal angle eventually converting them to low angle boundaries. Slone et al. [94] also observed similar deactivation of twinning in CoCrNi based alloy with ∼4.1 at.% Al, 2.3 at.% Ti and 0.9 at.% Nb; through DFT calculations Slone showed that the Al-Al bonds are highly unfavorable in the HCP phase and this could be the reason shearing via partial dislocation is prevented as they lead to a formation of local HCP stacking, and instead the glide of perfect a/2 < 110 > dislocations is preferred. Since partial dislocation shearing, i.e., formation of SFs is a precursor to twin formation, twinning will be re- Fig. 8. Dark-field TEM micrographs showing continuous precipitation (on the left) and discontinuous precipitation (on the right) in the (CoCrNi) 94 Al 3 Ti 3 [91] . duced as well due to the addition of Al and Ti. The addition of these two elements introduces a trade-off between dislocations and twinning but the formation of coherent L1 2 precipitates has a positive effect on the strength with a slight loss in ductility.
Ti when added with C [95] was found to change the morphology of the as-cast specimen from columnar to equiaxed even in the regions closer to mold. It is believed that though Ti and C do not form a carbide, they may be rejected into the liquid ahead of the solid-liquid interface, causing constitutional supercooling and destabilizing the planar solidification front. The (CoCrNi) 99.2 Ti 0.4 C 0.4 system showed the finest grain size as well as mechanical properties that exceeded those of the base ternary CoCrNi.
Molybdenum
Mo, besides inducing precipitation and solid solution strengthening within the CoCrNi matrix [96][97][98][99][100] , has a very significant effect on the grain growth kinetics of the alloy it is added to. Mo has high self-diffusion activation energy which intensifies the solution drag effect and inhibits the grain growth which leads to an added grain boundary strengthening in the CoCrNi-Mo alloys [84] . Depending on the Mo concentration and processing conditions, Mo tends to form a sigma phase which is rich in both Mo and Cr. Chang et al. [96] showed that it was only after 5 at.% Mo that the sigma phase could be seen in the CoCrNiMo x alloys. Since the sigma phase is hard and brittle, it may contribute to strength but has a detrimental effect on the ductility of the alloy. That is why in this work best strength-ductility synergy was attained for the CoCrNiMo 3 alloy. Li et al. [98] observed a similar trend for CoCrNiMo x (x = 0, 0.1, and 0.2 in molar ratio) where they found grain size decreased and the sigma phase content increased with increasing Mo content. Through DFT calculations (at 0 K) it was found that SFE of CoCrNiMo 3 was similar to that of CoCrNi and both, twinning and dislocations, contribute to the deformation. In another study by Chang et al. [97] , it was observed that for an alloy with composition (CoCrNi) 93 Mo 7, there was a difference in how precipitates were distributed in as cast + cold-rolled + annealed samples, where one region is rich in precipitates while the other is depleted of it, as compared to the uniform grain boundary precipitation in case of homogenized + coldrolled + annealed samples. The as-cast sample showed better strength and ductility than the homogenized one because the presence of brittle precipitates at the grain boundaries (in the case of homogenized samples) lead to the weakening of the interface and hence an early onset of instability. Such a heterogeneous precipitation behavior highlights the importance of each processing step in microstructural evolution and mechanical properties. Table 3 lists the results from published studies on CoCrNi-Mo alloys.
Tungsten
Similar to Mo, W also has high self-diffusion activation energy which leads to slow grain growth kinetics which adds to strengthening mechanisms caused by large lattice and modulus mismatch of W with Cr, Co, and Ni. Guo et al. [93] observed that that as the W content increased, the recrystallization retarded, and grain size decreased which is intuitive. With additions of up to 3 at.% W, no secondary phases were observed. They calculated the grain growth activation energy which was 301.0 ± 7.1 kJ/mol, 314.3 ± 5.8 kJ/mol, and 393.3 ± 9.2 kJ/mol for 0 at.% W, 1 at.% W, and 3 at.% W respectively highlight the solute drag effect of W leading to slower grain growth. Though this work did not calculate the effect of W on SFE of CoCrNi, using the previous literature, it predicted that SFE will increase with W doping. This claim warrants further theoretical and experimental study. Wu et al. [101] also studied CoCrNi-W alloy with 3, 6, and 9 at.% W doping. Though they did not find any secondary phase with 3at.% W, the other two compositions had W-rich secondary phases which agrees with ThermoCalc predictions showing the evolution of the secondary phase at ∼4 at.% W. The mechanical testing showed an increase in both strength and ductility of CoCrNi-3 at.% W as the temperature changed from 293 to 77 K and it was also observed that deformation twinning was prevalent at 77 K while at 293 K, only dislocation glide and cross-slip dominated the deformation. Hence W leads to solid solution and grain boundary strengthening in the alloys. The mechanical behavior of W alloyed CoCrNi CCAs is listed in Table 4 . In a sense, the prediction of increased SFE with W doping [102] aligns with the dominating presence of cross-slip at room temperature in the work by Wu et al. [101] as the width of SFs decreases with increasing SFE making cross-slip easier but this hypothesis needs to be backed by relevant data to make it more reliable.
Silicon
Si is another alloying element that has maintained the strengthductility synergy when added to the CoCrNi ternary. In the case of TWIP steel, Si was found to lower critical twinning stress and SFE, hence promoting higher ductility and work hardening [103] . Chang et al. [104] investigated the CoCrNiSi x alloy for x = 0.1, 0.2 and 0.3. It should be noted that x here is the molar ratio and not the atomic percent. These authors observed that for homogenized, cold-rolled, and annealed samples, Si promotes recrystallization (that is contrary to other alloying elements retarding it) which eventually leads to an increase in grain size with increasing Si content ruling out the possibility of grain boundary strengthening. Furthermore, multiple twins were observed in CoCrNiSi 0.2 and CoCrNiSi 0.3 , with the twin volume fraction increasing with the increase in Si. Additionally, for CoCrNiSi 0.3 , an HCP phase was observed at a strain value of 65%. The enhanced twinning and FCC →HCP phase transformation indicate towards a low SFE and through thermodynamic calculations, the effective SFEs of CoCrNiSi x (x = 0, 0.1, 0.2) were calculated to be 17.4, 12.9, and 7.6 mJ m − 2 This calculation illustrates the positive effect of Si in lowering the SFE of CoCrNi alloys. In another study by Liu et al. [105] for the same alloy compositions showed the addition of Si to base ternary maintains the SFE. They also found two types of precipitate formation for the CoCrNiSi 0.3 alloy, which in this case was -type tetragonal-structure phase enriched with Cr and Si, and the other one was enriched with Cr with an unidentified crystal struc- Table 5 .
To sum up, the addition of an alloying element to exploit various strengthening mechanisms like solid solution strengthening, precipitation strengthening, GB strengthening, etc. definitely helps in enhancing the properties of the base CoCrNi ternary alloy. It will also be more productive if alloy design concepts can be applied while tuning the composition. While few efforts are being made to understand how SFE changes with the minor addition of other elements, another area that still needs to be explored is the effect of alloying on the short-range order of the CoCrNi alloys. Even the minor addition distorts the lattice as has been observed in many works, it can be expected that it will also disturb the neighboring arrangement within the system and hence short-range order. It will be insightful if the effect of alloying on the SFE and shortrange order be studied and correlated.
Comparison of CoCrNi CCAs with conventional FCC alloys
While the properties shown by these CCAs are remarkable, other 3d-transition metal alloys exhibit similar behavior at low temperatures. Before the discovery of CCAs, various alloys including cryogenic steels and Ni-Co based superalloys have been known for their cryogenic performances [107][108][109][110][111][112][113][114][115][116][117][118][119][120][121] . The compositions of some of these alloys are listed in Table 6 and Appendix A. Among these alloys, MP35N alloy exploits the FCC →HCP martensitic phase transition, which is similar to that in CoCrNi alloy, to display excellent mechanical properties. For instance, cold worked and unaged MP35N alloy has shown tensile strength of up to 2.41 GPa with a tensile elongation of 42% at 77 K [113] . MP35N also exhibits remarkable corrosion resistance and has applications as a conducting wire in biomedical devices [122] as well as reinforcement material for high field pulsed magnets [20] . MP35N was limited by its maximum service temperature making them unsuitable for temperatures higher than 573 K and hence a new alloy MP159 with more alloying addition to the base Co-Ni-Cr matrix was developed. The additions of Al, Ti, and Nb elements into the matrix can lead to the formation of Ni 3 (Al,Ti) ′ phase after suitable annealing treatment [123] . Both the MP alloys exhibit low stacking fault energy due to a high content of Co which facilitates the formation of deformation twinning as explained in Section 2.1 . This deformation twinning can positively contribute to strain hardening and lead to strengthening to these alloys. Another member of this family is the AEREX350 alloy, which besides showing a ′ phase in aged material may also show and ƞ phase with D024 structure may appear at grain boundaries on high temperature aging [124] . Both of the phases contribute to the strain-hardening of the alloy and it was also found that cold working prior to aging can be used to alter relative nucleation and growth kinetics of both the ′ and ƞ phase to ensure uniform distribution and enhance the mechanical properties [114] . It is clear that alloys showing excellent mechanical properties at temperatures as low as 77 K have been developed as early as 1980s and have found many suitable applications ranging aerospace industry to orthopedic implants. With such similar composition of the MP alloys and the equiatomic CoCrNi ternary, it is important to take cues from the work already done in these materials to advance research in the CCAs.
While the basic features of twins, slip and FCC to HCP phase transition remain the same, the most prominent difference between conventional FCC alloys and CCAs is the compositional complexity in the latter. This complex compositional landscape and a highly distorted lattice leads to non-uniform interactions between the atoms and deformation structures like twins, SFs and dislocations. Unlike conventional alloys, bonds will not be uniform in the case of the base CoCrNi equiatomic ternary and other CCAs, thus, some bonds will be harder or easier to break than others. Such variation will result in non-uniform dissociation of Shockley partials and hence the local variation in SFE values. Smith et al. [126] investigated the equiatomic CrMnFeCoNi alloy and found that there was a large variation in dissociation distances where the mean SF width was 4.82 nm with a standard deviation of 1.93 nm. Using computational methods, it was also shown that at several sites, the local SFE was negative while the average SFE was positive. Shih et al. [127] also pointed out the high solute-dislocation interactions in CCAs which should be taken into consideration while experimentally calculating SFE. Using the Ni-Co model, Shih et al. [127] demonstrated the presence of an additional resistance force due to the interaction of dislocations with the large solute concentration in the alloy, which balances the forces in such a way that the equilibrium separation distance between partial dislocation remains finite even at zero or negative SFE values. Even with limited work in this direction, it can be emphasized that the value of SFE should not be taken as a single global value for CoCrNi and other CCAs. The connection between compositional complexity and the local SFE needs to be further explored. Another reason for such a difference between calculated and experimental SFE values can be the presence of local chemical order in these alloys. It should also be noted that some of the fundamental mechanisms like chemical short-range ordering which are currently very prominent in the field of CCAs might not have been considered when studying the conventional alloys. The current and future study in the CoCrNi based alloys can be an advancement of what is already known as well as filling any gaps in the understanding of CoCrNi based alloys.
Conclusion and future opportunities
The equiatomic CoCrNi ternary and related alloys represent a model for a scientifically interesting and technologically promising CCA system; not surprisingly there have been numerous studies published on this family of systems. Similar and yet different from TWIP/TRIP steels, CoCrNi alloys have the potential to be used as a structural material for low-temperature applications. The compositional complexity, the range of deformation structures, phase transformation, and short-range order in the CoCrNi system opens a broad range of opportunities to tailor and engineer the processing-microstructure-properties relationship. Current research in this alloy system has been insightful and mainly focused on enhancing tensile properties, i.e., a good combination of yield strength, tensile strength, and ductility, through various thermo-mechanical processing and alloying additions. Moreover, research into deformation mechanisms suggest that indeed, the compositional complexity of these systems does influence the underlying deformation mechanisms. For example, the presence of an additional resistance force due to the interaction of dislocations with the large solute concentration in the alloy [ 127 ], which balances the forces in such a way that the equilibrium separation distance between partial dislocation remains finite even at zero or negative SFE values. Although results suggest that SFE plays a key role in the deformation of CoCrNi alloys, it is important to understand that it can be influenced by other factors such as composition and local chemical ordering which will lead to local SFE variations within this material.
As the research in the CoCrNi CCAs progresses, it is important to look into other important properties like fracture toughness and fatigue which are frequently overlooked in current research. Furthermore, investigating all these properties at the cryogenic temperature will be crucial in establish this alloy as potential candidate for cryogenic applications. With the increasing focus on CSRO, it will be important to establish an in-depth understanding of why ordering occurs and if the FCC phase is stable or metastable. It will also be beneficial to understand how the modification in processing (heat treatment temperature and cooling kinetics) affects the occurrence of CSRO and if such ordering has any effect on the mechanical properties at the bulk level, or that it is it just at the micro/nano-level. Reversible phase transformation from HCP to FCC needs to be further explored. The basic concepts inspired by the conventional alloys need to be modified to consider the compositional complexity in the CoCrNi system and the possibility of CSRO within the system. This will also help in bridging the gap between the experimental and computational efforts. Furthermore, alloy design strategies can be used to modify the derivative systems and their processing to optimize the material performance of CCAs.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have influenced this paper.
Acknowledgment
This research was primarily supported by the National Science Foundation Materials Research Science and Engineering Center program through the UC Irvine Center for Complex and Active Materials ( DMR-2011967 )
Appendix
Appendix
|
2022-06-15T15:03:48.938Z
|
2022-06-01T00:00:00.000
|
{
"year": 2022,
"sha1": "48fc8b536a1bb9699193eaf0412c24e76b570156",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.1016/j.mtla.2022.101476",
"oa_status": "HYBRID",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "fa8e6fb25f55fe99665a69667f3eaf02e566db00",
"s2fieldsofstudy": [
"Materials Science"
],
"extfieldsofstudy": []
}
|
9752013
|
pes2o/s2orc
|
v3-fos-license
|
Crataegus ×ninae-celottiae and C. ×cogswellii (Rosaceae, Maleae), two spontaneously formed intersectional nothospecies
Abstract Crataegus monogyna Jacq. is naturalized in North America, where it has hybridized with native diploid hawthorns at least twice. We provide names for the two nothospecies (as well as for the corresponding nothosections and nothoseries), referring to existing documentation in the literature for nothosp. nov. Crataegus ×ninae-celottiae K.I. Chr. & T.A. Dickinson (C. monogyna × C. punctata Jacq.). New data are provided to further document nothosp. nov. Crataegus ×cogswellii K.I. Chr. & T.A. Dickinson (C. monogyna × C. suksdorfii (Sarg.) Kruschke). In both cases, the striking differences in leaf shape between most New World hawthorns and Old World section Crataegus, and the intermediacy of the hybrids, account for the relative ease with which these hybrids can be recognized. Finally, new sequence data from ITS2 and chloroplast DNA barcoding loci confirm the genetic relationships between the two nothospecies and their respective parents.
Introduction
. Sites in Canada and the United States at which collections of native and naturalized diploid (unless indicated otherwise) Crataegus were made as vouchers for morphological, chemical, and molecular (boldface) observations ( Fig. 1-3; Tables 2-4). Sampled individuals are listed by their collector and collection number; principal collector is T. A. Dickinson (D) der to draw more unbiased samples, with the inevitable consequence that these samples refl ect the greater frequency of the introduced species and its hybrids. To mitigate this, we have included individuals of mostly diploid C. suksdorfi i from other sites in order to refl ect the variation found in this taxon. Note that we distinguish the taxon referred to here as C. suksdorfi i from the other western North American black-fruited hawthorn with 20 stamens per fl ower, C. gaylussacia A. Heller. Th is is because these two taxa are allopatric (Coughlan 2012 and unpubl. data), and diff er in morphology and cytotype. Crataegus gaylussacia has shorter petioles and thorns that are thicker at their base than is the case with diploid C. suksdorfi i (Dickinson unpubl. data). Molecular data are consistent with C. gaylussacia being an autotriploid derivative of diploid C. suksdorfi i (Zarrei et al. http://2012.botanyconference.org/engine/search/index.php?func=detail&aid=536 and unpubl. data; see also Lo et al. 2009). In contrast, the C. suksdorfi i complex has been shown to comprise, in addition to diploids, allotriploids and allotetraploids (Zarrei et al. http://2012.botanyconference.org/engine/search/index.php?func=detail&aid=536 and unpubl. data).
In order to increase our sample for molecular studies we have supplemented fi eld collections of leaf tissue and herbarium vouchers with tissue removed from existing specimens in the ROM Green Plant Herbarium. Historical records of the distribution of C. monogyna were collected from fi ve herbaria across Canada (TRT, MTMG, MT, QFA and UBC). Online databases of Canadian and U.S. herbaria used included ACAD, the Invader Database System of the University of Montana (which contains information for fi ve northwestern states: Idaho, Montana, Oregon, Washington, Wyoming), OSC, and WTU. Distribution maps were prepared from specimen locality data using SimpleMappr (Shorthouse 2010). Names of Crataegus sections and series used here follow those published by VASCAN (Brouillet et al. 2010), and are accepted names sensu FNA Ed. Comm. in prep.
Morphology. For this study we concentrated on capturing and analyzing leaf shape data, as described elsewhere (Dickinson et al. 2008). Many previous studies of hybridization involving C. monogyna (Bradshaw 1953;Byatt 1975;Love and Feigen 1978), and of leaf shape variation in Crataegus generally (e.g. El-Gazzar 1980;Phipps and O'Kennon 2007), have attempted to quantify leaf lobing by means of a ratio of two measurements, x and y , where x is the distance from the tip of a lobe (usually the most basal one) to the deepest point of the sinus between that lobe and the adjacent one above it, and y is a measure of leaf size, usually the parallel distance from the tip of the lobe to the midrib. Th is approach can be eff ective when comparisons involve only leaves that have some degree of lobing (e.g. studies of hybridization between C. monogyna and C. laevigata in Europe, or of the lobed leaves of many species belonging to North American C. sect. Coccineae , such as C. punctata ). However, when lobing is absent altogether the necessary landmarks (lobe tip, deepest point of the sinus) are absent, and the distance x is undefi ned or is set to zero (Love and Feigen 1978). In this case, a better approach is to carry out multivariate analyses of additional measurements of leaves and other organs (Wells and Phipps 1989), or to quantify variation in the leaf outline as a whole. Elliptic Fourier coeffi cients obtained from digitized leaf outlines captured using MorphoSys (Meacham and Duncan 1991), or the Fourier amplitudes derived from them, provide a useful method for doing just this (Dickinson et al. 2008;McLellan and Endler 1998;Rohlf and Archie 1984).
Leaf outline data were collected from two overlapping samples: (1) short shoot leaf spectra (Dickinson and Phipps 1984) collected from a random sample of individuals at the Cogswell-Foster Preserve (comprising one C. suksdorfi i , seven C. monogyna , and 12 putative hybrids), and (2) leaves on herbarium specimens from the Cogswell-Foster Preserve and other locations in the Pacifi c Northwest. In the latter the attempt was made to sample the leaf shape variation seen in C. suksdorfi i as widely as possible. In both cases, variation in the shape of the leaf blade (i.e. excluding the petiole) was summarized by means of 39 Fourier amplitudes, and displayed by means of principal components analysis.
For each leaf outline we also obtained the area ( A ) and perimeter ( P ), so as to calculate the inverse of the dissection index described by Kincaid and Schneider (1978), i.e. inv(D.I.) = 2( A π) ½ / P , a parameter that has an upper bound of one for a perfect circle regardless of size, and approaches zero as the length of the perimeter increases with increased lobing of the outline (Dickinson 2003;Dorken and Barrett 2004). In addition to outline data we made linear measurements with which to index overall leaf shape: X , leaf blade length above the widest point; Y , leaf width; and Z , leaf blade length below the widest point (Marshall 1978). On some of the fl owering specimens in our sample we collected additional data on stamen number, style length and style number (in fruiting specimens, equivalently, pyrene number), and stigma width, in order to compare these with data collected by others from the introduced species and C. punctata . After transformation to a common [0,1] range these data were also summarized using principal components analysis. Analyses of variance were carried out on selected measurements. All data analyses described above were carried out using the R environment for statistical computing (R Core Team 2013). Signifi cance of individual principal component axes was evaluated using the broken-stick criterion (Frontier 1976) with the help of R function evplot (Borcard et al. 2011).
Molecular methods. Four DNA barcodes ( rbcL , matK , trnH -psbA , and ITS2; CBOL Plant Working Group 2009; Chase et al. 2007;Hollingsworth et al. 2011) were generated directly from genomic DNA for a worldwide sample of Crataegus (Dickinson et al. http://2011.botanyconference.org/engine/search/720.html;Zarrei et al. unpubl. data). Th e plastid origin of the markers was used to establish the maternal parentage of the hybrids. DNA was extracted and amplifi ed from leaf tissue of individuals representing the two hybrids and their parent species (Table 2) using Canadian Centre for DNA Barcoding (CCDB) protocols (Ivanova et al. 2011;Kuzmina and Ivanova 2011a, b). Th is sample overlapped partially with the cloned ITS2 one (below), and provided an additional two C. suksdorfi i , 10 C. monogyna , and fi ve C. punctata individuals, as well as one more of each of the two hybrids (Table 2).
We also analyzed data from another project (Zarrei et al. http://2012.botanyconference.org/engine/search/index.php?func=detail&aid=536 and unpubl. data) in which ITS2 was cloned for a sample of individuals that included 14 C. suksdorfi i , four C. monogyna , three C. punctata and two each of the two hybrids (Table 3). Meth- Table 2. Results of Neighbor-joining clustering of sequence data for chloroplast DNA barcode loci. GenBank accession numbers indicate cluster affi liation (Cluster 1 or 2) for Crataegus species and their putative hybrids. Details of the BOLD data can be found at dx.doi.org/10.5883/DS-CRATMONO. See Table 1 for sites and collectors; eight-digit ROM Green Plant Herbarium (TRT) accession numbers identify vouchers.
Flow cytometry. Flow-cytometric methods for quantifying nuclear DNA in embryo and endosperm followed Talent and Dickinson (2007a). Embryo DNA amounts of 1.48-1.70 pg were taken to indicate diploids, and an endosperm to embryo ratio of approximately 1.5 was taken to indicate sexual reproduction with meiosis.
Results and discussion
Morphology. Despite diff erences in sample size, the Pacifi c Northwest hybrid, Crataegus × cogswellii , appears more variable than either of its putative parents, C. monogyna or C. suksdorfi i (Fig. 1). Th e hybrid is clearly intermediate with respect to both leaf lobing (the inverse Dissection Index; Fig. 1) and style number (STYLE; Fig. 1). Principal components analyses of leaf outlines from Pacifi c Northwest C. monogyna , C. suksdorfi i , and their putative hybrid, demonstrate variation in leaf shape both within and between these three entities ( Fig. 2A, B). Th e fi rst principal component refl ects the contrast between the unlobed leaves of C. suksdorfi i and the markedly lobed ones of C. monogyna , as well as the intermediacy of the hybrid ( Fig. 2A, B), much as illustrated earlier by Love and Feigen (1978;their Fig. 3), and by Wells and Phipps (1989) for the Ontario hybrid and its parents (their Fig. 4). Th e second principal component refl ects variation in the relative overall lengths and widths of the leaf outlines ( Fig. 2A).
DNA barcode loci . Analyses of both the directly sequenced and the cloned ITS2 ribotypes demonstrate the parentage of both putative hybrids ( Fig. 3; Table Taxa Voucher GenBank accession number ON45 Purich and Talent MP85 (TRT00002250) KC174190, KC174191, KC174192, KC174193, KC174194, KC174195 ON45 Purich and Talent MP86 (TRT00002251) 3); no signs of recombination were detected in the cloned ITS2 dataset. ITS2 sequences from the hybrids resemble either C. monogyna or one of the native North American species. Th e way in which both parental ribotypes are maintained in each of the hybrids examined here is probably due to how recently the hybrids have been formed: less than 200 years ago in the case of the Ontario hybrids (Douglas 1914;Kirk 1819;Provancher 1863), and less than 100 years ago in the case of the Oregon ones (the earliest specimen of C. monogyna was collected in 1914 in Douglas Co. Oregon;Phipps 1998). Th ese time periods are evidently too short for genome homogenization (concerted evolution) to have taken place, even in diploids reproducing sexually. Our small sample of seed from the hybrids (Table 4) parallels earlier invDI , inverse dissection index = 2( A π) 1/2 / P , where A is the leaf area and P is the leaf perimeter; STAM , number of stamens per fl ower; STYL , number of styles per fl ower. Both axes shown account for signifi cant portions of the total variance according to the broken-stick criterion (Frontier 1976). results (Talent and Dickinson 2007a) showing diploidy and sexual reproduction in both parental taxa.
Only two of the three chloroplast genome barcode loci showed suffi cient variation for individuals from Crataegus section Crataegus to be distinguished from ones belonging to either C. section Coccineae or C. section Douglasia (Table 2). Sequence data from both rbcL-a and the trnH-psbA spacer region showed the same two clusters, C . sections Coccineae and Douglasia (Cluster 1), and C . sect. Crataegus (Cluster 2; Table 2). Th e way in which the hybrids fell into one of these clusters or the other demonstrates that, with one exception, C. monogyna is the female parent of the Ontario hybrids with C. punctata studied here, while C. suksdorfi i is the female parent of the Pacifi c Northwest hybrids.
Th ese results corroborate earlier observations based on seed-set in artifi cial crosses between the parent species (Love and Feigen 1978;Wells and Phipps 1989). In reciprocal pollinations seed set was greatest (32-34%) when C. monogyna stigmas received pollen from C. punctata (Wells and Phipps 1989). Fruit set was most successful when C. monogyna pollen was applied to the stigmas of C. suksdorfi i fl owers (mean 42%, range 25-73%, compared to a 29% mean fruit set by C. suksdorfi i with open pollination; Love and Feigen 1978). However, all reciprocal crosses between C. monogyna , C. suksdorfi i , and their hybrid yielded seeds (R. M. Love, personal communication). (Table 1). In both A and B the two PCA axes shown are signifi cant according to the broken-stick criterion (Frontier 1976). In B arrowed point 1 represents the single individual of C. suksdorfi i for which individual leaves are represented in A , while arrowed point 2 represents the averaged data for the six leaves of C. monogyna shown in grey in A . Table 1 for details). Dashed lines indicate the sectional affi nity of the sequences B Th e corresponding Neighbor-Net network for the cloned ITS2 sequences has three branches representing: ( a ) ribotypes from individuals of C. monogyna , and from its hybrids with both C. suksdorfi i and C. punctata ; ( b ) ribotypes from individuals of C. suksdorfi i and C. × cogswellii ; and ( c ) ribotypes from individuals of C. punctata and C × ninae-celottiae (Table 3). Th e numbers shown are the % bootstrap support for each of the three branches.
Our use of data from DNA barcoding is not a test of the value of DNA barcoding in Crataegus , as this is discussed elsewhere (Dickinson et al. http://2011.botanyconference.org/engine/search/720.html; Zarrei et al. unpubl. data). Rather, we have taken advantage of our barcode sequence data from individuals unequivocally identifi able as C. monogyna , C. punctata , C. suksdorfi i and their hybrids in order to use sequence similarity to inform us about the hybridization process.
Hybridization. Since its introduction to North America during the late 18 th and the 19 th centuries (Kirk 1819;Provancher 1863;Douglas 1914), fi rst on the east coast and nov. and C. monogyna in Ontario. Filled square, holotype of Crataegus × ninae-celottiae ; Crosses, TRT specimens of C. × ninae-celottiae ; asterisks, C. × ninae-celottiae specimens cited by Wells and Phipps (1989); stars, specimens of C. monogyna in MT, MTMG, QFA, TRT, and UBC. Crataegus punctata occurs throughout the region depicted (Phipps and Muniyamma 1980; this paper also maps additional records for C. monogyna ). Table 4. Flow-cytometric results from seeds of the two described Crataegus nothospecies. Th e ratios shown for endosperm and embryo nuclear DNA contents are well within the ranges observed for sexually reproducing C. monogyna (Talent unpubl. data) and diploid C. suksdorfi i (Lo et al. 2013). (Phipps and Muniyamma 1980;Phipps 1998;Lin 2009). Nevertheless, except for isolated occurrences in northern Delaware and adjacent Pennsylvania, as well as in Kentucky, Utah, and the San Francisco Bay area in California, C. monogyna in North America is not found south of 40°N latitude. In Ontario, C. punctata appears to be the only native diploid with a similarly late fl owering period that is also frequently sympatric with C. monogyna ( Fig. 1 in Campbell et al. 1991;Fig. 4). Crataegus suksdorfi i is the only native hawthorn in the Pacifi c Northwest known to include diploid individuals, and these are restricted to Oregon and adjacent California and Washington, west of the Cascades (Fig. 5; Lo et al. 2013). Where they co-occur, diploid C. suksdorfi i and C. monogyna fl ower at the same time, the latter species much more abundantly than the former (Love and Feigen 1978). Crataegus monogyna may never have been commonly planted in boundary hedges in Canada as it was in Europe. Fences and hedges appear to have been only rarely constructed in 17th Century Canada by European settlers to confi ne ruminant animals (Greer 2012); the animals were instead fed indoors, but allowed to roam the arable land for a short season after harvest, confi ned by the wall of surrounding forest. To this day, the hawthorn commonly growing along Ontario fence lines consists of native species, perhaps naturally occurring there. In Ontario forests we often encounter remnants of zig-zag post-and-rail fences, and these had the advantage over a hedge that they could be rapidly constructed as needed to mark property boundaries or to keep animals out of particular areas. In the United States hedging had its advocates in the early nineteenth century, but one of these described the superiority of native species like C. crus-galli ("cockspur" or "Newcastle thorn") and C. marshallii ("parsley-leaved" or "Virginia thorn") over introduced C. monogyna (Kirk 1819; "to sow or plant without fencing, would (in this country) be a useless labour").
Taxon/TRT accession/site/collection
Flow cytometry of seeds from both hybrids was consistent with diploid embryos and triploid endosperm, except that the embryos from C. × cogswellii show slightly higher than diploid measurements, higher than the 1.39-1.66 pg measurements previously obtained from leaf data (Table 4; Talent and Dickinson 2005). Whether the seeds involved would have germinated is unknown, but in contrast to the large healthy looking seeds from C. × ninae-celottiae , those from C. × cogswellii had smaller embryos and were variously misshapen. We noted that some individual trees of C. × cogswellii have a high degree of parthenocarpy-completely seedless fruitand the seeds we collected may therefore have been supernumerary to any strongly viable seeds. We can only state that C. × cogswellii apparently carries out both meiosis and fertilization, as expected of other diploid Crataegus (Table 4; Talent and Dickinson 2007b).
In her examination of hybridization between C. punctata and C. monogyna in Ontario, Purich (2005) found that the styles of C. punctata are signifi cantly longer than those of C. monogyna (mean mono = 4.1 mm; mean punc = 7.3 mm; sample sizes 5/52 and 7/116, individuals/styles). Diff erences between the two species in pollen grain diameter, hence volume, are not signifi cant (Purich 2005). No such diff erence in style length is present when comparing C. monogyna and C. suksdorfi i . Th ese results suggest that in Ontario, at least, the longer styles of C. punctata could act as a barrier to the successful penetration of C. punctata ovules by pollen tubes from C. monogyna pollen grains (Table 2). With style lengths and pollen grain diameters in C. monogyna and C. suksdorfi i similar (Dickinson unpublished data), it may be that the more abundant fl ower production of C. monogyna (Love and Feigen 1978) contributes to its role as the predominant pollen parent of C. × cogswellii . Th e exception to the summary above (TRT203 in Table 2; C. punctata as the maternal parent) refl ects the way in which diff erences in style length likely act to infl uence the direction of hybridization in a probabilistic rather than an absolute way. (Fig. 6) Remarks. Shrub or tree up to ca. 6 m tall. Twigs of the current year densely to sparsely hairy or glabrous, hairs appressed to patent, straight or slightly curly; twigs of the previous year pale grey or ash-grey; aphyllous thorns 0.5-2 cm long, stout, straight; spinetipped, leaf-and dwarf-shoot-bearing branchlets lacking. Leaf blades ovate, obovate or elliptical, acute at apex, attenuate, cuneate or rounded at base, shallowly or deeply and regularly lobed, lobes with an acute apex, basal pair of veins convergent, straight or slightly divergent, intercalary veins running to the sinuses partly present, upper surface with ± deeply impressed veins at maturity, dull or lustrous bright or dark green, sparsely hairy and often becoming glabrous except along the veins, hairs appressed or semi-patent; lower surface dull, pale green, sparsely hairy throughout or only along the major veins and in the vein axils, hairs appressed or semi-patent; margin regularly crenate-serrate or serrate, teeth minutely glandular, glands less than 0.1 mm; petiole eglandular, narrowly winged in upper part. Subterminal leaf blade of fl owering shoots 30-55 mm long, 16-38 mm wide, shallowly and regularly lobed, lobes 2-5 pairs, basal pair extending 0.2-0.4 times the width of lamina to midrib, each lobe with 6-11 teeth, basal pair of sinuses in apical 1/4 to basal 1/3 of lamina; petiole 6-20 mm long; stipules caducous, membranous or herbaceous, 4-8 mm long, irregularly or regularly glandular-denticulate, with 20-30 teeth. Leaf blades of elongate shoots 35-45 mm long, 25-35 mm wide, shallowly or deeply and regularly lobed, lobes 3-5 pairs, basal pair extending 0.2-0.6 times the width of lamina to midrib, each lobe with 4-11 teeth, basal pair of sinuses in basal 1/2-1/3 of lamina; petiole 8-12 mm; stipules caducous, herbaceous, ca. 6 mm long, regularly glandular denticulate-serrate, with ca. 15 teeth. Infl orescence 3-4 cm long, lax, corymbose, 5-17-fl owered, densely to sparsely hairy, hairs appressed, semi-patent or patent, straight or slightly curly; pedicels 3-18 mm, densely to sparsely hairy, hairs appressed, semi-patent or patent, straight or slightly curly; bracts caducous, membranous or herbaceous, 3-4 mm long, 0.2-0.4 mm wide, linear-lanceolate, 10-15 times as long as wide, irregularly glandular-denticulate, with 5-7 teeth. Hypanthium 3-4 mm long, densely to sparsely hairy, hairs appressed, semipatent or patent, straight or slightly curly; sepals 2-4 mm long, 1.5-2 mm wide, triangular-lanceolate or triangular, 1-2.7 times as long as wide, entire or rarely irregularly and minutely glandular-serrate, teeth 0-2, apex acute or obtuse; petals 6-7 mm long and wide; stamens 18-20, anthers 1-1.2 mm long, pink or purple; styles 2-3; hypostyle pilose. Fruit 9-12 mm long, 8-12 mm in diameter, 1.0-1.1 times as long as wide, globose, broadly ellipsoidal or obovoid, ± lustrous, red or orange, punctate with small, pale brown lenticels, up to ca. 0.2 mm in diameter, sparsely hairy, crowned by the persistent, refl exed sepals; calyx tube indistinct, ca. 0.5 mm long, 3-4 mm wide; fl esh yellowish, hard and mealy; pyrenes 2-3, ventro-laterally smooth; hypostyle pilose. Phenology. Flowering in May-June. Fruiting in August-September. Reproductive biology. Sexual. 2 n = 2 x (2 n = 34? Muniyamma and Phipps 1979;Talent and Dickinson 2005); diploid embryos and triploid endosperm.
Etymology. Crataegus × ninae-celottiae honors Nina Celotti
Crataegus × ninae-celottiae diff ers from C. punctata in: aphyllous thorns shorter, 0.5-2 cm long (not 2-5 cm long); leaf blades regularly lobed almost to the base (not unlobed or shallowly lobed towards apex), intercalary veins running to the sinuses sometimes present; subterminal leaf blade of fl owering shoots usually smaller, up to ca. 55 mm long, and veins 2-5 pairs (not up to ca. 85 mm and veins 6-10 pairs); stipules often herbaceous and irregularly glandular-denticulate; sepals shorter, 2-4 mm long, and wider, 1-2.7 times as long as wide (not 3-7 mm long and 2-4.7 times as long as wide); styles and pyrenes 2-3 (not 3-5); fruit usually smaller, up to ca. 12 mm long and in diameter (not up to ca. 15 mm long and in diameter) and less distinctly punctate with smaller lenticels up to ca. 0.2 mm in diameter (not up to ca. 0.4 mm in diameter).
Phenology. Flowering in April-May. Fruiting in September. Some individuals strongly parthenocarpic.
Distribution. Northwestern U.S.A.; western Oregon ( Figure 5); potentially present in adjacent northwestern California and southwestern Washington where the parent species are sympatric.
Etymology. Crataegus ×cogswellii honours the Cogswell family, and Mr. and Mrs. Lee Foster, of Halsey, Oregon. In 1872 John Cogswell, Mrs. Foster's grandfather, purchased the land that the Fosters gave to the Oregon Nature Conservancy as the Cogswell-Foster Preserve (Lopez 1971), and at which C. × cogswellii has been most intensively studied (Love and Feigen 1978).
Similar taxa. Crataegus × cogswellii diff ers from C. monogyna in: leaf-and dwarfshoot-bearing branchlets usually lacking; stipules of leaves of fl owering shoots irregu- their specimens. Jen Byun and Kathleen Buck helped us collect data from herbarium material. Tara Winterhalt adjusted the contrast in Fig. 6 and Fig. 7. We are grateful to the herbaria mentioned in the fi gures for lending specimens or making specimen data available to us. Financial support to KIC from the Carlsberg Foundation (Grant 2008_01_0155) and to TAD from the Natural Sciences and Engineering Research Council of Canada (Discovery Grant A3430), the Royal Ontario Museum, the Department of Ecology & Evolutionary Biology (formerly the Botany Department) of the University of Toronto, and the Canada Foundation for Innovation and Ontario Research Fund (funding through Canadensys for equipment and personnel for specimen documentation) are gratefully acknowledged. DNA barcoding was funded by the Government of Canada through Genome Canada and the Ontario Genomics Institute (2008-OGI-ICI-03).
|
2017-06-03T07:06:16.610Z
|
2014-04-09T00:00:00.000
|
{
"year": 2014,
"sha1": "defa8f2e1eb26f70292672c59f8470e690861ca3",
"oa_license": "CCBY",
"oa_url": "https://phytokeys.pensoft.net/lib/ajax_srv/article_elements_srv.php?action=download_pdf&item_id=1524",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "866fef234ee2e22074187e78d83047aa198fb91f",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Biology",
"Medicine"
]
}
|
245668764
|
pes2o/s2orc
|
v3-fos-license
|
Multi-day dataset of forearm and wrist electromyogram for hand gesture recognition and biometrics
Surface electromyography (sEMG) signals have been used for advanced prosthetics control, hand-gesture recognition (HGR), and more recently as a novel biometric trait. For these sEMG-based applications, the translation from laboratory research setting to real-life scenarios suffers from two major limitations: (1) a small subject pool, and (2) single-session data recordings, both of which prevents acceptable generalization ability. In this longitudinal database, forearm and wrist sEMG data were collected from 43 participants over three different days with long separation (Days 1, 8, and 29) while they performed static hand/wrist gestures. The objective of this dataset is to provide a comprehensive dataset for the development of robust machine learning algorithms of sEMG, for both HGR and biometric applications. We demonstrated the high quality of the current dataset by comparing with the Ninapro dataset. And we presented its usability for both HGR and biometric applications. Among other applications, the dataset can also be used for developing electrode-shift invariant generalized models, which can further bolster the development of wristband and forearm-bracelet sensors.
Background & Summary
Recent advances in machine learning techniques have enabled applications of hand gesture recognition (HGR) using surface electromyographic (sEMG) signals. This has further boosted the development of advanced prosthesis control systems for rehabilitation of upper limb amputees 1 , and recent industrial applications have emerged using HGR for human-machine interactions in industrial 2 and consumer applications scenarios 3,4 . For these applications, the sEMG signals features are extracted and used as inputs for various machine learning techniques such as linear discriminant analysis (LDA) 5 , support vector machines (SVM) 6 etc. for detecting hand gestures. More recently, advanced techniques such as deep neural networks (DNN) have achieved highly accurate classification performance even with simple architectures 7,8 .
However, extensive investigation has demonstrated sEMG-based HGR has poor cross-user transference performance 9 , suggesting a calibration-free and one-size-fits-all model for all users is still elusive, which suggests that sEMG signals inherently contains individual differences, i.e., biometric information. This has provided motivation for investigating the potential of sEMG as a biometric trait. Combined with the HGR property, it enables the user to set user-defined gestures as a password for enhanced security, which is not possible with other bio-signals such as electroencephalogram (EEG) and electrocardiograph (ECG). Our recent studies have provided a framework for the fusion of these codes and to facilitate such a dual-mode (password and biometrics) authentication system 10 .
Although, high performance of both HGR and biometric models has been previously reported, there always exists a gap between the real-world conditions and the laboratory settings, under which most of the current HGR and biometric research have been conducted conditions. It has been established in the literature, that in a multi-session protocol spreading across days, non-stationary factors including electrode shifts, skin and physical conditions will seriously affect the performance of an sEMG processing system 11 . As there are numerous such factors, experimentally controlling each of them would increase the number of trial repetitions exponentially www.nature.com/scientificdata www.nature.com/scientificdata/ and hence can be arduous. Nevertheless, a multi-day dataset with a sufficiently large subject pool is warranted for validating the effectiveness of sEMG applications such as HGR and biometrics. Some open-access databases of multi-day sEMG recordings of forearm muscles are publicly available [12][13][14][15][16] . Two databases with large subject pool (>40) involved two days of data collection 14,15 . While one of them has as low as six channels 15 , some others utilized a high-density (HD) sEMG setup 12,13,16 . Only three studies involved more than two of data collection days, but the number of subjects was smaller (<11) 12,17,18 . Only one study with three days of data recording, had a sample size of 20 subject, with only signals from the forearm 19 . To explore the robustness and accuracy of HGR and biometrics, it is imperative to have a database with larger subject pool sizes, recorded across multiple days and comprising numerous gestures.
In the current study, we present an open-access named Gesture Recognition and Biometrics electroMyogram (GrabMyo) Dataset 20,21 . GrabMyo consists of 43 participants (subsequently termed as users), three sessions in three separate days (subsequently termed as sessions) of data collection, 16 hand and finger gestures each with seven repetitions (subsequently termed as trials). To our best knowledge, the presented dataset is the largest sEMG dataset in terms of the total number of recording sessions (43 users × 3 days = 129 recording sessions). A unique feature of GrabMyo is that the sEMG signals were recorded from both forearm and wrist positions. A graphical representation of electrode positions and the list of gestures investigated in the study is shown in Fig. 1a,b. The sampling frequency of the recorded signals was selected as 2048 Hz. To obtain generalizable data, special effort was taken such as 1) electrode positioning protocol for each session, 2) normal level force instruction, 3) rest duration for avoiding fatigue, 4) un-uniform interval between data collection sessions (Days 1, 8, and 29) and 5) data collection from healthy users (subjects with a single session of sickness have been eliminated from the study). These considerations are explained in greater detail in the Methods section. The dataset provides a valuable resource for sEMG-based HGR and biometrics research, particularly for improving algorithms' robustness in a multi-day scenario and cross-user generalization ability.
Methods
Subjects and ethical requirements. We recruited 43 healthy participants (23 M, 20 F) for the study spanning three different days: day 1, day 8 and day 29. The participants were students and staffs from the University of Waterloo. The average age was 26.35 ± 2.89, and the average forearm length (measured from the styloid process on the wrist to the olecranon on the elbow) was 25.15 ± 1.74 cm. More details about the dataset and the participant characteristics are reported in Table 1. Individuals with any existing muscle pain, skin allergies, and those who were unable to complete the three sessions due to any unprecedent circumstances were excluded from the study. For the third session (day 29), a range of 3 days was provided for some participants who couldn't make it to the specific appointment. Before the enrolment, the participants were provided an oral and written explanation of the procedures and signed an informed consent form. They were informed that they could withdraw from Acquisition setup. The experimental setup consisted of a PC and a monitor mounted on a desk, 0.75 m in front of a height-adjustable chair. The EMGUSB2 + (OT Bioelettronica, Italy), a commercial amplifier, was used for acquiring the sEMG signals. The gain of the device was set to 500, and the sampling rate was set to 2048 Hz. Pre-gelled skin-adhesive monopolar sEMG electrodes (AM-N00S/E, Ambu, Denmark) were used.
Prior to the experiment, the user's forearm length is measured as the distance between the olecranon process and the ulnar styloid process. The forearm circumference is measured at one-third of the forearm length from the olecranon process. The wrist circumference is measured at 2 cm away from the ulnar styloid process. Prior to electrode placement, the skin surface was shaved to remove hairs, cleaned with an alcohol swab, and abraded with a paper towel. For the forearm electrode placement, sixteen sEMG electrodes were placed in the form of two rings, each consisting of eight electrodes equally spaced around the forearm, forming eight bipolar pairs. The center-to-center distance between the two forearm-rings was maintained at 2 cm. For the wrist electrode setup, twelve monopolar sEMG electrodes of the same type as the forearm rings were placed in the form of two rings, each consisting of six electrodes equally spaced around the wrist and forming six bipolar pairs. The center-to-center distance between the two wrist-rings was maintained at 2 cms, the same as the forearm setup. Therefore, a total of 28 monopolar sEMG electrodes were used for each session, which forms four electrode rings: proximal and distal rings for the forearm and wrist. A detailed pictorial representation is provided in Fig. 1a. To maintain consistency of the positions of the electrodes across all participants, the first electrode in each ring (total rings = 4) was anatomically positioned on the center-line of the elbow crease as shown in Fig. 1a 22,23 . Although a standardized electrode placement protocol was followed for each session, no marks were left on the participant's skin. This intentionally induced uncertainty in the exact electrode positions corresponds to the real-life scenario where it is not realistic, if possible at all, to place an electrode band at the same position across multiple sessions.
Acquisition protocol. For each experimental session, following the acquisition setup, the participant is seated comfortably on the chair with both their upper limbs in a resting position. Visual instructions for performing the gestures were provided on the computer screen placed in front of the participants. The participants were instructed to perform the gestures at a normal force level, or similar to how they would normally do it during daily activities. For an estimate of "normal force", the participants were asked to perform multiple trial contractions at three self-defined force levels: soft, hard and medium contractions, where a medium level corresponds to the contraction of the normal force. The following 16 hand and wrist gestures were included in the current study (presented in Fig. 1b . The order of the 16 gestures was randomized and a resting (REST) trial was collected after all 16 gestures were performed once. A ten-seconds relaxing period was provided between each trial. One continuous data acquisition of 17 gestures (including the REST) is called one run. First, a trial run was performed to ensure that the participants understood the experimental protocol. Following the trial run, seven runs were recorded for each user, resulting in 119 contractions (17 × 7). Any accidental gesture or no-activity/delayed-activity was noted and the respective gesture's replacement contraction were performed after each run. The user could also request additional rest when he/she felt necessary. The entire session was repeated on day 8 (after 1 week) and day 29 (after 1 month).
Signal processing. The sEMG signals were bandpass filtered between 10 Hz and 500 Hz using a fourth-order Butterworth filter. A notch filter of 60 Hz was employed to remove the powerline noise that might have affected the signal recording. Table 1. Database Summary. The forearm length is measured from the olecranon process to the ulnar styloid. The forearm circumference is measured at a distance one-third from the elbow joint. The wrist circumference is measured 2 cms away from the ulnar styloid process.
Data Records
Data records 20,21 presented in this section and accompanying description files are available online in PhysioNet (https://physionet.org/about/database/) and IEEE Dataport (https://ieee-dataport.org/datasets). The database consists of 43 participants, three sessions in three separate days (subsequently termed as sessions) of data collection, and 17 gestures (including REST) each with seven repetitions. All the sEMG recordings are 5 seconds in duration, collected from 28 channels (16 forearm and 12 wrist), and sampled at 2048 Hz.
The components of the released repositories are described in detail (Table 2) and are organized as follows: waveform data and additional files.
Waveform data. For the PhysioNet Database, the sEMG recordings are provided in the waveform database (WFDB) format, which is considered the most widely used medium for storing physiological signals and waveform data 24 . There exist numerous open-source WFDB libraries for commonly used analysis tools using MATLAB and Python. The WFDB allows a structured way of storing the sEMG recording in the form of a tuple of two files: a dat-file containing the binary raw data and a corresponding header file with the same name and a hea-extension. The hea-extension contains all the signal-specific metadata such as the channel names, sampling frequency, and the scaling factor for converting the signal to physical units (in mVs).
For the IEEE database, the signal files are saved as .mat files. The mat file extension allows users a convenient approach for obtaining the sEMG recordings in the physical unit format (additional conversion is not necessary). Additionally, the mat files allow data to be organized as cells where multiple gestures and repetitions from a single session can be presented together.
Additional files. The additional files provide supplementary information such as the participant anthropometrics, electrode positioning guidelines, device configuration, the sequence of gestures, and their descriptions. MATLAB codes are also provided for reading the data files and subsequent feature extraction as explained in the Code Availability section. MotionSequence.txt provides gesture definitions and their sequence DeviceInfo.pdf describes the configuration of acquisition device.
technical Validation
A specific trial of an individual user case was chosen at random and the amplitude and frequency components were analyzed as shown in Fig. 2. It was observed that all 28 channels of forearm and wrist possessed equal amplitude and power spectrum distribution. In the following sections, signal processing techniques are discussed for analysis and technical validation of the sEMG signals. As well, the original references where the methods were described in depth are provided. For comparison, the validation results are provided along with those for the widely used Ninapro DB2: Exercise B comprising 40 users and 17 gestures.
Signal-to-noise ratio (SNR). The spectral properties of each sEMG signal recording were analysed and the SNR (in dB) was measured as the ratio between the power of the signal to the power of the noise 25,26 . As the types of the artifacts were unknown in our case, the power of the noise was estimated as the power of sEMG recordings during the rest trial. The average SNR of all the signals (14.565 ± 6.385 dB) was in range with the SNR values suggested for the wrist and forearm 25 and they are consistently higher than those from the Ninapro database (8.001 ± 4.051 dB) (see Fig. 3a).
correlation coefficient of Normality (ccN).
The CCN was measured to analyze the amplitude distribution. For a static contraction with moderate force, the sEMG can be modeled as a filtered, random, white Gaussian noise process 27 . It has been suggested that a test of normality can provide a measure of biosignal quality, where a signal amplitude with a non-normal distribution would be considered contaminated. A Gaussian distribution with equal mean and variance to that of the recording is generated 28 . The CCN is defined as the Pearson correlation coefficient between the histogram bin values of the sEMG recording and the normal density function value for the corresponding bins. A value close to 1 is considered a normal distribution. It was observed that the CCN of www.nature.com/scientificdata www.nature.com/scientificdata/ all the signals (0.975 ± 0.041) was close to 1 which was consistently higher than those from the Ninapro database (0.848 ± 0.075) (see Fig. 3b).
Standard HGR and biometric analysis.
Prior to HGR and Biometric evaluation, the signals were first processed and then features were extracted as described below. The forearm rings (eight channels) and the wrist rings (six channels), the monopolar sEMG signals were first re-referenced by a common average procedure. The processed signals were then segmented into 200 ms-width windows, with a 150 ms overlap. Each window was then processed using Hudgins's time-domain (TD) feature extraction 29 . Time-domain features (mean absolute value, zero crossing, slope sign changes, and waveform length) were extracted from filtered data 22 . Therefore, for a forearm setup, the feature vector consisted of 8 × 4 = 32 features, while the wrist setup consisted of 6 × 4 = 24 features.
For the biometric analysis, a matching score, commonly the Mahalanobis distance, is used to assess if it's a match (access granted) or no match (access denied) 10,22,23 . To maintain consistency in the HGR and biometric analyses, a Mahalanobis distance classifier was implemented for both analyses. For a given feature vector sample p (the input), its Mahalanobis distance S i.j , with the ith gesture and the jth user, was defined as where µ i,j is the centroid of the ith gesture class and the jth user and ∑ i, j is the covariance matrix for the specific gesture and user class. Both the parameters are calculated from the system training data and the sample p is from the system testing data. The leave-one-out (LOO) cross-validation scheme was used, where six trials were used for training and one trial for testing. Figure 3a,b, and Table 3 demonstrate the results of the technical validation as described in the following sections.
HGR evaluation. In this study, the HGR analysis was performed in a user-specific scheme. For a particular user and a particular gesture, the true class consisted of the feature vectors from the target gesture of the user and the false class consisted of the feature vectors from the remaining 15 gestures of that user (the rest gesture was excluded from the HGR analysis to maintain consistency with biometric analysis). Similarly, for the Ninapro database, the true class consisted of feature vectors from the target gesture of a specific user and the false class consisted of the feature vectors of the remaining 16 gestures. For the performance analysis a receiver operating characteristic (ROC), where the true positive rate (sensitivity) was plotted against the false positive rate (1specificity) by varying the threshold distance of correct gesture prediction. The true positive rate or sensitivity www.nature.com/scientificdata www.nature.com/scientificdata/ represents the probability of detecting a correct gesture, while the false positive rate is the probability of detecting an incorrect gesture. The area under the curve (AUC) is calculated from the ROC curve 30 . The ROC curve and the AUC values for all the users, days, and gestures are averaged and reported separately for the forearm and wrist electrode positions. Figure 4a shows the ROC plots for HGR analysis. It was observed that the AUC for the forearm was 0.948 (±0.018) and for the wrist was 0.941 (±0.021). Both the values were comparatively higher than the corresponding AUC value of 0.875 (±0.034) for the NinaPro data.
Biometric evaluation. In this study, the verification mode of biometrics was used to demonstrate the feasibility of multi-day sEMG-based biometrics. In this mode, the true user's identity and the corresponding gesture are known to the system. As such, for a specific user, the true class consisted of the feature vectors from the target gesture (e.g. the passcode) of that user and the false class consisted of the feature vectors of the remaining 42 users for that gesture. Similarly, for the Ninapro database, the true class consisted of feature vectors from the target gesture of a specific user and the false class consisted of the feature vectors from the remaining 39 users for that gesture. For the performance analysis a receiver operating characteristic (ROC), where the true positive rate (sensitivity) was plotted against the false positive rate (1 -specificity) by varying the threshold distance of correct biometric authentication. The true positive rate or sensitivity represents the probability of detecting a correct hand gesture, while the false positive rate is the probability of detecting an incorrect hand gesture. The equal error rate (EER), is obtained from the ROC curve, where the false positive rate is equal to the false negative rate (1-sensitivity) 30 . The ROC curve and the EER values for all the users, days, and gestures are averaged and reported separately for the forearm and wrist electrode positions. Figure 4b shows the ROC plots for Biometric analysis. It was observed that the EER for the forearm was 0.028 (±0.007) and for the wrist was 0.038 (±0.006). Both the values were comparatively lower than the corresponding EER value of 0.038 (±0.013) for the NinaPro data.
Usage Notes
As evident from the title, the GRABMyo Dataset has two major applications: HGR using machine learning approaches and sEMG-based biometrics. The true potential of GRABMyo are three-fold: 1) large subject-pool, 2) multi-day session, and 3) both wrist and forearm channels. Through the Signal Analysis section, we will explain processing sEMG recordings and the cross-validation approach for obtaining reliable benchmark results using machine learning algorithms. Following this, the Future Direction lists novel research strategies for HGR and Biometric analysis which could be bolstered using the large multiday dataset.
Signal analysis. The MATLAB code fileread.m allows the users to read data files along with their signal-specific metadata. After conversion to a numeric format, the data is structured as matrixes with timepoints as rows and monopolar sEMG channels as columns. The channel names provided can be used for separate forearm and wrist electrode rings and also for forming bipolar pairs between them 31 . The processed data is then windowed, and features can be extracted using multiple techniques 32 . A widely used frequency division technique 33 is provided as a sample feature extraction in feature_extraction.m. The gesture labels are provided by MotionSequence.txt which allows the extracted features to be matched to the specific gestures and their respective users. For HGR and biometric applications, the output classes are the gestures and users, respectively which can be used for developing machine learning and deep learning architectures. The data provides seven trial repetitions, hence seven-fold cross-validation (six-folds for training and one-fold for testing) can be used to simulate a near-practical scenario, where the testing data is recorded separately from the testing data. A more robust analysis i.e., the multi-day analysis with a 3-fold cross-validation can be employed, where subsets of data from two days are used for training and the data from the remaining day is used for testing. Future Directions. Some future research directions are: Improving biometric authentication. One unique advantage of the EMG biometric trait, in particular for authentication applications, is the combination of user-specific biometrics with user-defined gestures as passcodes, the latter of which is not possible with other bio-signals such as electroencephalogram (EEG) and electrocardiogram (ECG). A multi-code EMG-based biometric framework can be used to combine the gestures and improve authentication performance and security 10 .
Biometric identification. Another major biometric application is the identification mode where the system predicts the identity of the presenting user by finding the closest match. The identification is a more error-prone application as the system makes N comparisons, where N is the number of users. Therefore, the factors affecting system performance such as multiple days and sample size of the database need to be investigated for real-life applications. www.nature.com/scientificdata www.nature.com/scientificdata/ Subject independent gesture recognition. Extensive research on EMG has been performed on gesture recognition with application in rehabilitation using prosthetic and orthotic devices, home application for assisting daily activities, virtual environment control, and sign language recognition 3,4,34 . Recent studies have suggested deep learning techniques for cross-user calibration-free which trains generalized models using the population data, and hence reduces the training burden of the user 9,35,36 . The presented large-sample dataset can provide resources for such calibration-free models.
Electrode shift-invariant techniques. One of the significant factors affecting the cross-day sEMG performance is the shift in the electrode positions. It is impossible to fix the location of armband electrodes on the forearm and wrist for daily-wear use. These variations affect the performance of both the sEMG-based biometric and gesture recognition applications. Some techniques such as classification model adaptation 37,38 and feature space transformation using transfer learning 35,39,40 have been suggested to address the electrode shift variations. These techniques could be further investigated to potentially improve biometrics and HGR performance.
Code availability
The custom codes used for reading the signals of the database was created in MATLAB R2017b and is freely accessible at Physionet and IEEE Dataport 20,21 . To implement the codes, the users will need a MATLAB License • A readme file (readme.txt) with instructions about how to run the code in a 2017b or higher MATLAB version.
• A Matlab script (fileread.m)with a simple example about how to read WFDB files and convert them to. mat format.
• A text file(MotionSequence.txt)which provides the gesture sequence, and thus can be used to assign class labels to input data.
• A Matlab script(feature_extraction.m)allows a simple example to extract frequency features using featiDFTI.m and segmentEMG.m functions.
• A Matlab function(featiDFTI.m)for generating frequency division technqiue features from sEMG Data.
• A Matlab script(segmentEMG.m)for implementing windowing of sEMG Data.
|
2022-01-05T02:16:24.673Z
|
2022-01-04T00:00:00.000
|
{
"year": 2022,
"sha1": "4c60b4b881a731f350adc05ba50faae2541513b1",
"oa_license": "CCBY",
"oa_url": "https://www.nature.com/articles/s41597-022-01836-y.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "b1968bd898f079e81cbca90a09676778ab9c3772",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science",
"Engineering",
"Mathematics"
]
}
|
252917468
|
pes2o/s2orc
|
v3-fos-license
|
The abnormal level and prognostic potency of multiple inflammatory cytokines in PCI‐treated STEMI patients
Abstract Objective Inflammatory cytokines modulate atherogenesis and plaque rupture to involve in ST‐segment elevation myocardial infarction (STEMI) progression. The present study determined eight inflammatory cytokine levels in 212 percutaneous coronary intervention (PCI)‐treated STEMI patients, aiming to comprehensively investigate their potency in estimating major adverse cardiac event (MACE) risk. Methods Serum tumor necrosis factor (TNF)‐α, interleukin (IL)‐1β, IL‐6, IL‐8, IL‐10, IL‐17A, vascular cell adhesion molecule‐1 (VCAM‐1), and intercellular adhesion molecule‐1 (ICAM‐1) of 212 PCI‐treated STEMI patients and 30 angina pectoris patients were determined using enzyme‐linked immunosorbent assay. Results TNF‐α (52.5 (43.9–62.6) pg/ml versus 46.4 (39.0–59.1) pg/ml, p = 0.031), IL‐8 (61.6 (49.6–81.7) pg/ml versus 46.7 (32.5–63.1) pg/ml, p = 0.001), IL‐17A (57.4 (45.7–77.3) pg/ml versus 43.2 (34.2–64.6) pg/ml, p = 0.001), and VCAM‐1 (593.6 (503.4–811.4) ng/ml versus 493.8 (390.3–653.7) ng/ml, p = 0.004) levels were elevated in STEMI patients compared to angina pectoris patients, while IL‐1β (p = 0.069), IL‐6 (p = 0.110), IL‐10 (p = 0.052), and ICAM‐1 (p = 0.069) were of no difference. Moreover, both IL‐17A high (vs. low) (p = 0.026) and VCAM‐1 high (vs. low) (p = 0.012) were linked with increased cumulative MACE rate. The multivariable Cox's analysis exhibited that IL‐17A high (vs. low) (p = 0.034) and VCAM‐1 high (vs. low) (p = 0.014) were independently associated with increased cumulative MACE risk. Additionally, age, diabetes mellitus, C‐reactive protein, multivessel disease, stent length, and stent type were also independent factors for cumulative MACE risk. Conclusion IL‐17A and VCAM‐1 high level independently correlate with elevated MACE risk in STEMI patients, implying its potency in identifying patients with poor prognoses.
| INTRODUC TI ON
ST-segment elevation myocardial infarction (STEMI), defined as STsegment elevation in two anatomically contiguous leads, is the most urgent manifestation of coronary artery disease. 1,2 Over the past decade, the number of new STEMI cases is increasing in China annually, meanwhile, its mortality is constant. 3 Concerning the STEMI treatment, for patients with symptom onset within 12 h, percutaneous coronary intervention (PCI) is the preferred strategy to restore myocardial perfusion as soon as possible. [4][5][6] Unfortunately, the major adverse cardiac event (MACE) frequently occurs after PCI treatment, implying a poor prognosis in STEMI patients. 7,8 Hence, it is meaningful to explore reliable biomarkers for monitoring MACE risk in PCI-treated STEMI patients.
Inflammation is non-negligible in STEMI etiology, which is considered an essential process that participates in the initiation and progression of atherosclerotic plaque. [9][10][11] Consequently, some studies notice that the increased level of inflammatory cytokines possesses a certain value in predicting the MACE risk of STEMI patients, including interleukin (IL)-1, IL-6, IL-17, IL-37, etc. [12][13][14][15] For instance, one study suggests that the high level of IL-6 is correlated with elevated MACE risk in STEMI patients who receive primary PCI treatment. 13 Another study shows that the high circulating IL-37 level is a predictor of in-hospital MACE rate in STEMI patients treated with PCI. 15 However, the previous studies only focus on a single or a few inflammatory cytokines (no more than three), and comprehensive studies focusing on the prognostic value of plentiful inflammatory cytokines in PCI-treated STEMI patients remain rare.
Therefore, the present study determined eight inflammatory cytokines in 212 PCI-treated STEMI patients, aiming to comprehensively investigate their potency in estimating MACE risk and identify the prognostic indicators for PCI-treated STEMI patients.
| Data collection
Clinical characteristics of STEMI patients were obtained, which contained demographics, underlying diseases, and blood laboratory detections. Besides, disease characteristics and PCI operation information of STEMI patients were also collected, which included symptom-toballoon time, multivessel disease, culprit lesion, thrombus aspiration, number of implanted stents, type of stent, stent diameter, stent length, and infarct size (recorded as the infarct size at 3rd day after PCI). 1.6-50 ng/ml). The assays were performed in strict accordance with the kit protocols from manufacturers. In the analysis, the inflammatory cytokines were classified as high and low based on the median values.
| Follow-up and assessment
Standardized follow-up was carried out for the STEMI patients until tion. 16 Then, the cumulative MACE rate was calculated.
| Correlation of inflammatory cytokines with infarct size and lipid metabolism in STEMI patients
Subgroup analysis was conducted to investigate the association of inflammatory cytokines with infarct size, which showed that TNFα (Table S1).
Moreover, it was also observed that IL-17A was positively linked (Table S2).
Additionally, ROC curves were performed to determine the predictive ability of IL-17A and VCAM-1 for MACE occurrence risk, which showed that IL-17A could hardly predict MACE occurrence were linked with increased cumulative MACE risk ( Table 3).
| Independent factors of cumulative MACE risk in STEMI patients
To further eliminate the potential confounding factors, the The bold values represent the results with statistical significance, whose p value < 0.050.
were independently associated with increased cumulative MACE risk. Differently, the application of the everolimus-eluting stent (vs. sirolimus-eluting stent) (p = 0.034) was independently related to decreased MACE risk ( Table 4).
| DISCUSS ION
Among the complicated etiologies of STEMI, the most widely accepted one is the plaque rupture-caused coronary artery thrombosis. 2 When vascular walls are exposed to a proinflammation environment, the recruitment and accumulation of leukocytes are enhanced, which further promotes endothelial injury and the formation of thrombosis. 17 Subsequently, considering the essential role of inflammation in the pathogenesis of STEMI, the previous studies notice the abnormal level of several inflammatory cytokines in STEMI patients. [18][19][20][21] For instance, one previous study observes that serum The bold values represent the results with statistical significance, whose p value < 0.050.
Abbreviations: CI, confidence interval; CRP, C-reactive protein; cTnI, troponin I; HR, hazard ratio; IL, interleukin; MACE, major adverse cardiac events; STEMI, ST-segment elevation myocardial infarction; VCAM, vascular cell adhesion molecule. Fifthly, the comparison of inflammatory cytokine levels between PCI-treated and coronary artery bypass grafting (CABG)-treated STEMI patients was undetected in this study, which deserved further investigation.
In summary, IL-17A and VCAM-1 high level independently correlate with elevated MACE risk in STEMI patients, which might serve as useful indicators in assistance of STEMI diagnosis and identifying patients with poor prognosis.
DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.
CO N S E NT TO PA RTI CI PATE
Each patient or family member signed the informed consent.
|
2022-10-18T06:17:34.722Z
|
2022-10-17T00:00:00.000
|
{
"year": 2022,
"sha1": "e67a5f5ade9c7c5e75937e38cc41a30b1020e72a",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Wiley",
"pdf_hash": "ba6a1b0eab6dfe46dd70289ad516ed16cb7edb5c",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
228090737
|
pes2o/s2orc
|
v3-fos-license
|
Evaluation Model of Loop Stray Parameters for Energy Storage Converter of Hybrid Electric Locomotive
When the silicon carbide (SiC) power module is applied to the energy storage converter of a hybrid locomotive, under the action of di/dt and loop stray inductance, it is easy to produce excessively high voltage overshoot, which affects the battery life and stimulates high-frequency oscillations, causing power devices to withstand greater electrical stress. In order to optimize the system layout and improve system performance, it is extremely necessary to accurately extract and evaluate the loop inductance. This paper takes the typical high-frequency converter structure as the object, establishes an equivalent model of the circuit, and quantitatively analyzes the loop inductance from a mathematical point of view. For the circuit after the parallel of absorption capacitor, the small signal model is used to analyze and reveal the role and influence of the absorption capacitor. Finally, the calculation results of the model are compared through a double-pulse experiment. The results show that the error between the model and the experimental results is about 1%, and the effect of evaluating the stray parameters of the converter circuit is good, and it can provide a theoretical support for the selection and design of the absorption capacitor.
I. INTRODUCTION
Bidirectional DC/DC converters are widely used in energy storage converters of hybrid locomotive, connecting power batteries or super capacitors as auxiliary power [1], [2]. With the development of power electronic devices, reducing volume, reducing cost and improving reliability have become research hotspots. Silicon carbide (SiC) devices have the characteristics of high thermal conductivity, fast switching speed and high blocking voltage [3]. Power electronic devices based on SiC modules solve the contradiction between high converter operating voltage and low power module withstand voltage, and also greatly increase the switching frequency of the converter and reduce the volume of the output filter devic [4], [5], which has extremely broad potential in the application of energy storage converters for hybrid locomotives.
The associate editor coordinating the review of this manuscript and approving it for publication was Yijie Wang .
Under the high-frequency and high-voltage working environment, the action of SiC MOSFET switching momentary excessive di/dt and loop inductance will induce higher electrical stress and reduce battery life. In addition, high-order harmonics are generated, which affects the performance of the system, and may also cause problems such as switching losses [6], electromagnetic interference [7], and noise pollution [8], and even endanger the normal operation of the system in severe cases. In order to implement targeted suppression measures, it is extremely necessary to accurately extract and evaluate the stray inductance of the power loop.
Loop inductance mainly includes device parasitic inductance, stray inductance of connecting parts and bus capacitance parasitic inductance [9]- [11], etc. Existing literature researches on the extraction method of stray inductance are mostly for laminated busbars. Literature [12] sorts out the transient process of turn-on and turn-off, and finds the most favorable stage for the extraction of stray inductance, combining the method of integral operation to extract the stray VOLUME 8, 2020 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ inductance. Literature [13] studies the laminated busbar of the five-level converter module, and tests the proposed method by comparing simulation and experiment. Reference [14] improves the traditional integration method by considering stray resistance and measurement offset, and improves the calculation accuracy. Literature [15] proposes to use the oscillation frequency during the turn-on and turn-off process to extract the loop inductance by analyzing the switching transient process of SiC MOSFET, which improves the accuracy. Although many literatures have proposed the optimization of the extraction method of stray inductance, most of the analysis is to improve the accuracy of the experimental extraction method from the analysis of the transient process of the device switching. The mathematical quantitative analysis of the power loop stray parameter model is not performed. In order to enrich the theoretical basis and reduce the test cost, based on the previous research, this paper takes the bidirectional DC/DC converter based on SiC power module as the object, and establishes the stray parameter mathematical model of the power loop in sections. We quantitatively calculate the stray inductance of each part of the loop, and evaluate the proportion and influence of each section of stray parameters in the power loop. Finally, a double-pulse experiment is conducted to verify the accuracy of the theoretical analysis and the applicability of the evaluation model. Figure 1(a) is the main circuit diagram of the bidirectional DC/DC converter, including support capacitor C dc , support capacitor parasitic inductance L dc , stacked bus bar L bus , parallel SiC power module, filter inductor L b and filter capacitor C b , L d1 is SiC module parasitic inductance. This article takes the typical structure of the DC/DC converter shown in Figure 1(b) as the object, and builds the test platform shown in Figure 1(c) to extract the stray inductance of the converter power loop. Including DC voltage V DC , support capacitance C dc1 , C dc2 , support capacitance parasitic inductance L dc1 , L dc2 , laminated bus bar L bus , load inductance L load and DUT.
II. EXPERIMENTAL PLATFORM BASED ON SiC MOSFET
The device under test is selected from Cree's CAS300M17BM2 half-bridge power module. The support capacitor and the power module are connected by a busbar. The busbar adopts a laminated structure, and the current between the positive and negative plates is reversed, which makes the magnetic field cancel and the busbar inductance reduced. In order to further reduce the loop inductance, absorption capacitors can be connected in parallel at the positive and negative ports of the power device. The absorption capacitor, the support capacitor and the power device are connected to the bus bar through bolts. The DC input terminal is supported by two supporting capacitors in parallel; during the test, the two ends of the upper tube gate source are turned off by applying negative pressure.
The test method is the traditional double-pulse test method. Figure 2 shows the timing diagram of the double-pulse test. After the device under test is pressurized to stability, at time t 0 , the SiC MOSFET is turned on for the first time, and the drain current rises; At the first turn-off at time t 1 , the current change rate di/dt acts on the loop stray inductance and forms a voltage spike at both ends of the device after being superimposed with the bus voltage; At time t 2 , the device is turned on for the second time, and the reverse parallel diode of the upper tube reversely recovers, resulting in an overshoot of the drain current; at t 3 , the device is turned off, and a large voltage overshoot is formed again across the device under test. At this moment, the current change rate changes significantly, and the SiC junction capacitance has little effect, which is suitable for the extraction of stray parameters.
III. MODEL OF CONVERTER CIRCUIT CONSIDERING PARASITIC PARAMETERS
The converter circuit includes four parts as shown in Figure 3: support capacitor, laminated busbar, absorption capacitor and power module. The parasitic inductance of the power module can be obtained from the data sheet. This article mainly analyzes the other three parts.
A. SUPPORT CAPACITANCE PARASITIC INDUCTANCE
The wide-band characteristic of the support capacitor is characterized by an RLC circuit, where R dc1 , C dc1 , and L dc1 are the stray resistance, main capacitance, and parasitic inductance of the support capacitor, respectively. Single support capacitors are directly connected in series by RLC. In this paper, two support capacitors are connected in parallel as an example for analysis. The equivalent circuit is shown in Figure 4.
Before the test process, first charge the supporting capacitor. Therefore, the positive electrodes of C dc1 and C dc2 are equipotential points. Equations (1)∼(3) give the parameter values of each element in the process of equivalent transformation. (2)
B. STRAY INDUCTANCE OF LAMINATED BUSBAR
According to the circuit structure of FIG. 1, FIG. 5 draws a schematic diagram of each stray inductance branch in the bus bar. In the figure, L 1d , L 2s , L 3d , L 4s , L pd and L ns are the inductance from C dc1 to SiC MOSFET positive electrode, from C dc1 to SiC MOSFET negative electrode, from C dc2 to SiC MOSFET positive electrode, from C dc2 to SiC MOSFET negative electrode, from DC power supply to SiC MOSFET positive electrode, from DC power supply to SiC MOS-FET negative, respectively. There is direct mutual inductance in each branch inductance. Figure 6 shows the topology of the double-pulse test circuit. v 1d , v 3d , v s4 , v s2 are the terminal voltages of L 1d , L 3d , L 2s , L 4s , respectively. The volt-ampere characteristic formula of the stacked busbar with two supporting capacitors VOLUME 8, 2020 in parallel is shown in equation (4).
In equation (4), L bus is a 4 × 4 inductance matrix, and the inductance in FIG. 6(b) can be expressed using equation (5). Decouple the two parallel coupled inductors to obtain a structure in which L 12 -M ds and L 34 -M ds are connected in parallel and then in series with M ds . The calculation formula of the two parallel coupled inductors is shown in equation (6).
For a single support capacitor loop, the model calculation formula is: In the inverter commutation circuit, the branches L pd and L ns between the DC power supply and the busbar should also be considered. Figure 7 shows the distribution of the stray inductance of each branch of the inverter commutation 212592 VOLUME 8, 2020 circuit. The analysis method is the same as that of the test platform with double supporting capacitors in parallel. According to equation (8), the inductance of the power supply branch can be obtained: For the coupled inductance of three branches in parallel, there is independence during decoupling conversion, and only one or a few of them can be decoupled. The circuit after decoupling the star-connected circuit is shown in Figure 7(c).
Due to the parasitic inductance of the supporting capacitor, the DC side ends of the three equivalent branch inductances are no longer equipotential points. The two branches connecting the support capacitor and the positive pole of the power module should be connected in series with the support capacitor first, and then in parallel with the equivalent inductance of the DC voltage branch in the laminated busbar. At this time, the equivalent inductance of the inverter commutation loop can be obtained as shown in equation (9).
C. THE FUNCTION AND INFLUENCE OF ABSORPTION CAPACITOR
After absorbing capacitors in parallel, the overvoltage phenomenon caused by stray inductance can be suppressed, but at the same time, the circuit structure is changed, making the circuit into a higher-order system, and it is difficult to obtain an analytical expression using time-domain analysis. This paper analyzes the circuit using the small-signal model of SiC MOSFET in the complex frequency domain. SiC MOSFTE is approximately a constant current source with a certain increase. According to the principles of DC current source open circuit and voltage source short circuit, Figure 8 shows the small signal equivalent model of the test platform. The expression of loop inductance is: According to the equivalent circuit, the input admittance of the circuit is: When the imaginary part is 0, the circuit resonates. Therefore, the resonance condition of the circuit is: The resonance frequency is: In order to ensure that the circuit resonates, it should be guaranteed R dc < (L loop /C oss ) 1/2 . When R dc (L loop /C oss ) 1/2 , the resonance frequency of the circuit is approximately: At this time, the circuit is approximately GCL parallel, and its equivalent admittance is: The loop impedance is: The closer the absorption capacitor is to the SiC module, the better the absorption effect. Figure 9 is a test platform circuit and its equivalent model of parallel absorption capacitors.
After the parallel absorption capacitor, during the switching process, the support capacitor will first charge the absorption capacitor, and then the absorption capacitor will discharge to the power module, so Figure 9(a) can be equivalent to Figure 9(b). According to the foregoing, when the absorption capacitor is close to the port of the power module, selecting an absorption capacitor with a sufficient value can completely absorb the voltage spike introduced by the parasitic inductance of the support capacitor and the stray inductance of the busbar, but at the same time, it will introduce parasitic inductance of the absorption capacitor.
When the voltage spikes caused by L dc and L bus are completely absorbed, the resonance impedance of L dc , L bus VOLUME 8, 2020 and the absorption capacitor is much smaller than the L stray impedance at the resonance frequency.
Simultaneous (17) and (18): In addition, the absorption capacitor is connected in parallel at both ends of the device. After complete absorption, only the absorption capacitor parasitic inductance L snub and power device stray inductance L stray are in the loop. Therefore, the selection of the absorption capacitor is based on the two principles of capacitance and parasitic inductance. When L snub < L stray /10, the influence of L snub can be ignored, so the requirement for its parasitic inductance when selecting the absorption capacitor is L sn = nL snub , L sn is the parasitic inductance of a single absorption capacitor, n is the number of parallel absorption capacitors.
At this time, the loop stray inductance is:
IV. STRAY PARAMERER EXTRACTION BASED ON CALCULATION MODEL AND Q3D
The stray parameters include four parts: supporting capacitance parasitic inductance, stacked busbar stray inductance, absorption capacitance parasitic inductance and power module parasitic inductance. Two supporting capacitors are connected in parallel and three absorption capacitors are connected in parallel. The parameters in the analysis model are shown in Table 1. According to formula (1) ∼ formula (3), the supporting capacitance parasitic inductance L dc can be calculated as 20nH; The inductance of the laminated busbar is composed of mutual inductance and self-inductance. Figure 10 shows the busbar model built in Q3D software according to Figure (1 The principle of applying the excitation source is based on the actual test. For example, the supporting capacitor and power device are connected to the busbar with bolts, so the excitation source can be applied to the corresponding screw hole of the positive and negative busbars, and the simulation step and frequency are set according to the double pulse test time. After the inductance of each branch is obtained, calculation is performed according to equation (7), and the result is 25.1 nH. The parasitic inductance of the power device is obtained from the data sheet. According to equation (10), the loop inductance of the test platform can be obtained as 60.1nH.
V. EXPERIMENTAL EXTRACTION AND VERIFICATION OF STRAY PARAMETERS
Build a dual-pulse test platform based on the dual-pulse test schematic shown in Figure 1(b), and extract the stray inductance of each part of the power loop according to the waveform of the SiC MOSFET module off process.
A. STRAY PARAMETERS WITHOUT ABSORPTION CAPACITANCE
Select the positive-negative port, output-negative (down tube DS) and supporting capacitor port of the power module as the voltage test points to obtain the experiment waveform of voltage V DS across the lower tube of the SiC MOSFET, the power module terminal voltage V pn , the supporting capacitor terminal voltage V dc and the drain current I d , which is shown in Figure 11.
1) The stray inductance of the power loop is extracted according to the relationship between the volt-ampere characteristics of the inductor given by equation (22).
Among them, V Rdc is the voltage drop of the stray resistance.
Similarly, equation (24) gives the calculation formulas for the supporting capacitance parasitic inductance and the stray inductance of the laminated busbar.
B. STRAY PARAMETERS AFTER PARALLEL ABSORPTION CAPACITORS
After the parallel absorption capacitors at the positive and negative ports of the power module, the experimental waveforms of the voltage V DS across the lower tube of the SiC MOSFET, the voltage V pn at the power module terminal, and the drain current I d are shown in Figure 11. Table 3 statistics the inductance values of each part after theoretical analysis and experimental measurement after parallel absorption capacitors. After the parallel absorption capacitor, the absorption capacitor absorbs the overvoltage caused by the bus capacitor and the laminated bus bar, but at the same time introduces parasitic inductance. According to the experimental results, after a 1uF absorption capacitor is connected in parallel, the loop inductance is reduced by 21.95nH, and an absorption capacitance parasitic inductance of 13.6nH is introduced at the same time.
C. COMPARATIVE ANALYSIS OF EXPERIMENTAL RESULTS AND MODELS
When there is no parallel absorption capacitor, the measured loop inductance is basically consistent with the theoretical calculation. The parasitic inductance of the supporting capacitor and the parasitic inductance of the power device are VOLUME 8, 2020 slightly larger than the theoretical value. This is due to the fact that the voltage probe is clamped on the busbar and bolts are introduced during the experimental measurement, which causes the terminal voltage generated by the busbar inductance to be small and L dc and L d too large. The equivalent inductance of the bus capacitor and the stacked busbar series is 44.1nH. After the absorption capacitor absorbs it completely, the parasitic inductance of 13.6nH is introduced, which is close to the theoretical analysis. This error is caused by the position where the probe is clamped during the measurement. Because the actual structure uses bolt connection, the probe can only be clamped inside the busbar, causing the measured value of the voltage spike at both ends of the busbar inductance to be small, and the L dc and L d ends are measured The value is too large. This is an inherent measurement error. Combined with the ANSYS Q3D simulation results of the busbar, the analysis of the error is correct and reasonable, and the measured value of the total loop inductance is consistent with the theoretical analysis, which can prove the effectiveness of the method. The existence of this error can better prove the value of the calculation model proposed in this article, because it can compensate for the inherent error of the measurement. We have further analyzed the mechanism of this error in the updated article. Table 4 makes statistics on the proportion of inductance of each part in the power loop inductance before and after the parallel absorption capacitor. The parasitic inductance of the power module is based on the measured value when there is no parallel absorption capacitor. Before the parallel absorption capacitor, the parasitic inductance of the supporting capacitor accounted for the highest proportion of 38.3%, followed by the laminated busbar, accounting for 34.5%, and the parasitic inductance of the SiC module depends on the manufacturing process. When the busbar structure is fixed, the more supporting capacitors are connected in parallel, the smaller the loop equivalent inductance value. Therefore, we can try to use more supporting capacitors in parallel to reduce the loop inductance.
After the parallel absorption capacitor, the parasitic inductance introduced by the absorption capacitor accounted for the highest proportion, and the extra stray parameters due to insufficient contact of the absorption capacitor pin and contact accounted for 23.4%. When installing the absorption capacitor, multiple capacitors with small parasitic inductance should be connected in parallel as much as possible, and the capacitor can be fully contacted with the SiC module. In addition, attention should be paid to the parasitic inductance of the capacitor pin when selecting.
VI. CONCLUSION
In this paper, based on the SiC module-based hybrid locomotive energy storage converter converter circuit, a loop stray parameter evaluation model is established, and the stray inductance of each part of the power circuit is quantitatively calculated from a mathematical point of view. Firstly, the distribution of loop stray inductance is analyzed, and then the mathematical model of supporting capacitance equivalent circuit and laminated busbar considering the self-inductance and mutual inductance of the branch is established, and the effect of the absorption capacitance on the converter circuit is revealed through the small signal model. And influence. Finally, the results of model calculation and experiment extraction are compared to verify the accuracy and practicability of the model. The model can be used to calculate the power loop inductance of the converter in sections, clarify the distribution of stray parameters in the loop, and provide strong support for further optimization of the system structure layout. The model is still applicable to other similar topologies.
YUNXIN FAN (Member, IEEE) is currently the Director of the State Key Laboratory of Heavy Duty AC Drive Electric Locomotive Systems Integration. He is also a Professor Level Senior Engineer. He was selected into the second batch of national ten thousand talents program leading talents, in 2016. He has been committed to the research and development of rail transit traction power equipment technology, and has made outstanding contributions to the research and development of heavy haul traction equipment and the mastery of key technologies, the upgrading of passenger rail transit equipment and the research and development of high-end equipment. His current research interests include theory of rail transit electrical traction systems and optimization of energy matching strategy for hybrid locomotive. Integration. His current research interests include simulation theory of rail transit electrical traction systems, optimization algorithm of energy matching strategy for hybrid locomotive, and high-performance auxiliary system of urban rail vehicles.
XIAOYU REN received the B.Eng. degree in electrical engineering and automation from Xinjiang University, in 2018. He is currently pursuing the master's degree in power electronic and ac drives with Beijing Jiaotong University. His research interests include energy management strategies, heterogeneous multi-power drive systems (HMDS), and power electronic drive.
CHUNMEI XU received the Ph.D. degree in control theory and control engineering from Beihang University, Beijing, China, in 2004. She was an Associate Professor and a Graduate Supervisor of the School of Electrical Engineering, Beijing Jiaotong University. From August 2014 to August 2015, she was an Academic Visitor of the University of Wisconsin Madison. Her major research interests include control theory and its applications, including intelligent control, predictive control, servo system control, and control of traction drive system for urban mass transit.
|
2020-11-26T09:01:40.962Z
|
2020-01-01T00:00:00.000
|
{
"year": 2020,
"sha1": "0d9a32b6d901500eb1b0cadc1886b1b5b63fa1ab",
"oa_license": "CCBY",
"oa_url": "https://ieeexplore.ieee.org/ielx7/6287639/6514899/09265229.pdf",
"oa_status": "GOLD",
"pdf_src": "IEEE",
"pdf_hash": "09578a050f08c75ec103e2a6f3e5d44f88fb5dd4",
"s2fieldsofstudy": [
"Engineering"
],
"extfieldsofstudy": [
"Computer Science",
"Materials Science"
]
}
|
210295438
|
pes2o/s2orc
|
v3-fos-license
|
Geomorphological processes, forms and features in the surroundings of the Melka Kunture Palaeolithic site, Ethiopia
The landscape of the surroundings of the Melka Kunture prehistoric site, Upper Awash Basin, Ethiopia, were studied intensively in the last decades. Nonetheless, the area was mainly characterized under a stratigraphic/geological and archaeological point of view. However, a detailed geomorphological map is still lacking. Hence, in this study, we identify, map and visualize geomorphological forms and processes. The morphology of the forms, as well as the related processes, were remotely sensed with available high-resolution airborne and satellite sources and calibrated and validated through extensive fi eld work conducted in 2013 and 2014. Furthermore, we integrated multispectral satellite imagery to classify areas a ff ected by intensive erosion processes and/or anthropic activities. The Main Map at 1:15,000 scale reveals structural landforms as well as intensive water-related degradation processes in the Upper Awash Basin. Moreover, the map is available as an interactive WebGIS application providing further information and detail (www.roceeh.net/ethiopia_ geomorphological_map/).
Introduction
The prehistoric site of Melka Kunture, on the right banks of the upper Awash River, is a rich and complex archaeological and palaeontological area, extending over 90 km 2 , which was discovered in 1963 . A French mission directed by Jean Chavaillon (1965Chavaillon ( -1982Chavaillon ( /1993Chavaillon ( -1995 investigated at first systematically the site. Since 1999, an Italian mission directed by Marcello Piperno (1999-2010, and currently by Margherita Mussi of 'La Sapienza' University, Rome, conducts in-depth research at Melka Kunture. The lithostratigraphy of the area was investigated from the beginning of the excavations, revealing at least 70 archaeological horizons (Altamura et al., 2018). About 30 of them were extensively excavated or tested (Piperno, 2001). Since the upper Miocene, volcanic activity interrupted frequently the erosion and sedimentation stages (Chavaillon & Taieb, 1986;Taieb, 1974).
Volcanic products such as tuffs and lavas are generally interbedded with fluvial gravels, pebbles and sands (Raynal & Kieffer, 2004;Taieb, 1974). The tephra records allow for a stratigraphic correlation of archaeological horizons found in the gullies of the tributaries of the Upper Awash, or along the River itself (Geraads, Gallotti, Raynal, Bonnefille, & Mussi, in press;Mendez-Quintas et al., 2019;Morgan et al., 2012;Piperno, 2001). Endogenic activities related to the development of the Rift Valley such as volcanism and tectonics (Kusák, Kropáček, Vilímek, & Schillaci, 2016) are undoubtedly the dominant driving factors in landscape evolution. Hence, lava deposits such as basalts and pyroclastics like ignimbrites characterize the region.
However, even though the area was studied under various aspects, up till now a detailed geomorphological map of the region is not available. Thus, the main aim of this study is to generate an inventory of the geomorphological processes and related forms and features of the Melka Kunture area. Moreover, the study contributes to the understanding of the process dynamics and genesis of the landscape. In this context, we conducted several geomorphological surveys to reveal the components of the landscape, subject to alteration, erosion, transport and deposition processes induced by endogenic and exogenous driving forces as well as human activities. With the increasing amount of high-resolution satellite information, landforms and processes, especially of remote areas, can be assessed (Daba, Rieger, & Strauss, 2003;Tamene and Vlek 2008;Bouaziz, Matschullat, & Gloaguen, 2011). We mapped the dominant morphogenetic processes taking into account detailed geologic information (Poppe et al., 2013;Salvini, Riccucci, & Francioni, 2012;Sembroni, Molin, Dramis, & Abebe, 2017;St.-Onge, 1981;Verstappen & Van Zuidam, 1968) as well as additional geomorphic, pedologic and remote sensing information (Kropáček, Schillaci, Salvini, & Maerker, 2016;Maerker, Schillaci, & Kropáček, 2018). In order to comprehend and illustrate the complexity of the study area our integrated approach exploits (i) high-resolution topographic data for morphometric analyses, (ii) local information on geology and morpho-structural settings, (iii) multispectral information to characterize surface substrates and processes and (iv) analogue and digital visualization techniques to illustrate the final map content.
Study area
The study site is located 50 km southwest of Addis Ababa (Figure 1). It is close to the Butajira Road and to the Awash village, located on the Ethiopian plateau at about 2000 m a. s. l. (8°46 ′ 13 ′′ N; 38°3 3 ′ 30 ′′ E, 8°46 ′ 13 ′′ N; 38°38 ′ 3 ′′ E, 8°41 ′ 27 ′′ N; 38°3 3 ′ 30 ′′ E, 8°41 ′ 27 ′′ N; 38°38 ′ 3 ′′ E) (Figure 1). The Awash basin drains into the Rift Valley system, and represents one of the major tributaries covering large parts of the Oromia region. The tributaries of the Upper Awash River consist of ephemeral flows fed by a seasonal precipitation regime: Belg (June-October) and Keremt (February-April) (Billi & Dramis, 2003). The tributaries are often falling dry during the late dry season (December-February). The main archaeological excavation area lies on the right bank of the Awash River but also extends to the left bank, and covers an area including some of its small and/ or ephemeral tributaries such asAtebela and Tuka Meja (Figure 1). Endogenic processes such as the extensive presence of uplift, jointing and faulting phenomena are related to the development of the main Ethiopian Rift, which is linked to Miocene and Plio-Pleistocene volcanism. Hence, the relief of the study area is characterized by a semi-graben fault system (Gallotti, Raynal, Geraads, & Mussi, 2014) striking from ENE to WSW and induced by volcanic events. Tuffaceous beds in the study area are recognized as distal volcanic products. They are described in detail by Raynal and Kieffer (2004) and mapped by Salvini et al. (2012). Tuff formations are caused by Phreatomagmatic and Plinian-phreatic activities and build up the so-called non-welded ignimbrites (Salvini et al., 2012), which are extensively characterizing the southern parts of the archaeological area.
The archeological sites on the right bank of the Upper Awash River which cluster in the Gombore, Garba and Simbiro gullies Egels, 1971) have been extensively investigated. The archaeological sequences include volcanic and sedimentary deposits that are affected by water erosion resulting in gully systems with a typical length of 0.5-1 km. These gullies represent small torrential tributaries of the Upper Awash River.
The precipitation record of the Awash-Melka Kunture station (located ca. 1 km northeast of the Melka Kunture Museum), shows an average yearly rainfall of approximately 1000 mm y −1 with a clear pattern of high inter-annual variability. The precipitation regime is characterized by a short rainy season between March and April and a main wet season between July and September. The maximum precipitation of 73.4 mm/day was recorded on 4 September 1993, whilst the monthly maximum of 506 mm was registered in August 2011 (National Meteorology Agency, Addis Ababa). High intensity rainfall due to convection occurs especially during July, August and September.
During dry periods, a scarce vegetation cover characterizes the area. The main land use is represented by extensive agriculture with Lentil, Tef, Sorghum, Wheat as well as pasture (see Figure 2). Slopes around the villages are cultivated, but, the high density of sheep, goats and cattle lead to overgrazing, resulting in the degradation of vegetation cover, soil compaction, increased surface runoff and soil erosion. Close to the villages, Eucalyptus plantations are frequently present (Yitebitu, 2010). The vegetation formations of the region consist of secondary grassland and wooded grassland with various Euphorbia ssp. and Acacia ssp.. Lowlands are dominated by C4 grasses (White, 1983).
The typical soils of the area are shallow and with high skeleton content (Leptosols/Regosols according to WRB) (IUSS Working Group WRB 2014). In gently sloping and flat areas, brown or black clay rich soils dominate such as Vertic Umbrisols and Vertisols (IUSS Working Group WRB 2014) (see Figure 3).
Flood plains are characterized by clay rich Fluvisols (Gleysols). Foothills consist of colluvial sediments of different texture and depth. Vertisols ( Figure 3) and Fluvisols ( Figure 4) represent the main soil types of the area (Debele 1985).
Materials and methods
We performed a multiresolution interpretation of aerial photos and satellite images supported by an extensive validation in the field conducted in 2013 and 2014. We assessed typology and activity status of the landforms as well as of the geomorphological processes currently affecting the area. We mapped the dominant landforms taking into account detailed geologic information (Salvini et al., 2012;Verstappen & Van Zuidam, 1968)
Digital elevation model generation
We collected aerial photographs at 1:30,000 scale at the Ethiopian Mapping Agency EMA (Addis Ababa). In total, we purchased 34 images from 1972 scanned with 1024 dpi resolution and 1-meter ground resolution. The DEM was derived using an aerial photogrammetric technique called Structure from Motion (SfM) (see Supplementary materials section for details).
Based on the DEM we calculated several topographic indices following Wilson and Gallant (2000) and Hengl and Reuter (2009) that facilitate the geomorphological interpretation of forms and processes using the System for Automated Geoscientific Analyses (SAGA) (see Conrad 2006;Olaya and Conrad, 2008). The DEM was hydrologically corrected to eliminate sinks using the algorithm proposed by Planchon and Darboux (2001). The following topographic indices were derived: (i) Slope, (ii) plan, (iii) profile curvature, (iv) flow accumulation/catchment area using multiple flow algorithm (Tarboton, 1997), (v) transport capacity index (TCI), (vi) stream power index (SPI), and (vii) altitude above channel network (AACN) (see Vogel & Maerker, 2010). Slope gives general information on terrain steepness like escarpments, faults, or embankments. Curvatures indicate erosion transport and deposition landforms. TCI highlight areas potentially affected by sheet erosion whereas linear erosion features are revealed by the SPI. AACN shows surfaces that have a specific vertical distance to the river network such as fluvial terrace systems.
Mapping approach
The selection of mapping techniques and symbols followed an adapted approach combining the GIS related system proposed by Gustavsson, Kolstrup, and Seijmonsbergen (2006) for geology and substrates and the traditional Italian geomorphological classification system for gravity and hydro-erosive slope processes and forms. Moreover, we used part of the classical symbology for the landforms following Otto and Dickau (2004), and some symbols proposed by Gustavsson et al. (2006Gustavsson et al. ( , 2008. Specifically, the legend is composed by elements describing: (i) Topography, (ii) Hydrology, (iii) Structural elements and landforms, (iv) Lithology, (v) Gravitative landforms, (vi) Morphometry, (vii) Exogenic Geomorphological Processes and (viii) Erosional landforms.
The road infrastructure is based mainly on Google open street map (Google) whereas the toponyms were taken from the official topographical sheet Melka Kunture 1:25,000 provided by the Ethiopian Mapping Agency. We derived contour lines from the high-resolution DEM with 5 m resolution based on stereo aerial photographs from 1972 (see above section 3.1). Nonetheless, the DEM represents the situation in 1972 fieldwork conducted in 2013/2014 reveals that the topography is changing only very little especially in the areas where we have gully erosion (headcuts) or intensive bank erosion. We corrected the DEM with the respective GPS information accordingly. Furthermore, we adopted colours of the geomorphological features following Gustavsson (2006) and Gustavsson et al. (2006). Specific codes and hatchings were used to represent the present lithology based on the ArcMap USGS lithology library palette.
We assessed the spatial distribution of the erosion features by extensive field mapping (2013/2014). The mapping campaigns were prepared by a detailed stereo photo interpretation (Van Zuidam, 1985) morphometric analysis and expert knowledge based on the high-resolution DEM and the satellite imagery. Features such as surface soil erosion, linear erosion, and rock toppling have been successfully mapped. Additionally, a detailed terrain analysis aiming at the characterization of the landforms has been performed yielding a slope map that was subsequently used to delineate significant slope breaks. Areas affected by or with a high potential for sheet erosion are assessed using the TCI whereas linear erosion features are revealed by the SPI.
The breaks in slope have been represented with linear symbols (after Gustavsson et al., 2006). Significant changes in slopes such as fault systems, river bed incisions or meander banks were mapped overlaying the slope map to the aerial photographs and the Rapi-dEye imagery. Specific morphological features like gully erosion, have been illustrated with line type symbols, whilst badlands/heavily eroded areas and sheet erosion were mapped with polygonal features and illustrated in a purple scale. For the gully susceptibility, we used the Stream Power Index (SPI) following Finlayson and Montgomery (2003). Specifically, the legend is composed by elements describing: (i) Topography, (ii) Lithology, (iii) Hydrology, (iv) Structural elements and landforms, (v) Gravitative landforms, (vi) Morphometry, (vii) Exogenic Geomorphological Processes, and (viii) Erosional landforms.
Geomorphological processes assessment
We assessed the spatial distribution of the geomorphological features and processes combining fieldwork information, morphometric analysis based on the high-resolution DEM and a stereo aerial photo interpretation (API) approach following Van Zuidam (1985). We identified and mapped in detail the morphological forms and features according to their typical shape and according to their position in the landscape. Therefore, we selected the potential areas for specific processes using the topographic indices: TCI as proxy for sheet erosion, SPI for rill and gully erosion, AACN for river terraces identification. Flow accumulation/ Catchment area to specify the river network. Specific morphological features like gully erosion were illustrated with line type symbols whilst badlands/heavily eroded areas and sheet erosion were mapped with polygonal features. The indices also yield valuable information about the tectonic fault zones or uplifted sections since they are traced by abrupt changes in the landscape gradients. Especially, fault scarps in the basalts are the most evident sign of tectonic activity in the area. Furthermore, river systems and drainages are characterized by the river bankfull height and river floodplain area (Rosgen, 1996); the bankfull stage is represented by a blue hatched symbol while the floodplain is represented by green colour. Moreover, hydrological features, are represented in blue colours. Linear symbols were utilized to map active river banks. The bankfull levels were mapped according to Arnett and Williams (1978) using the AACN. The main drainages contribute significantly to the sediment transport. Based on the high-resolution DEM and extensive fieldwork we delineated drainage cross-sections and longitudinal profiles. Particularly their longitudinal profiles give important information on the erosional (steep parts or knickpoints) and depositional (flat sections) areas, as well as on the local erosion base levels.
Finally, the areas, identified by the abovementioned terrain indices, that show high susceptibilities for the respective forms and processes were checked with a detailed API and an extensive ground truth in 2013 and 2014.
Visualization
In order to cover the study area with appropriate detail, we present the mapped geomorphological forms and processes in two different formats: (i) an analogue classic map and (ii) a digital WebGIS-based visualization (see: www.roceeh.net/ethiopia_geomorphological_map/). The classic analogue map at 1:15,000 scale is projected in the UTM system. For orientation, an index grid was plotted on top as well as major UTM coordinates. The grid lines are numbered from A to E for the longitudinal axis and from 1 to 5 for the latitudinal axis. In this paper, we use this grid to localize the mapped features (e.g. A3, B4, etc.). The legend is aligned to the right of the map. Some of the mapped features are shown in the lower right of the map. In the upper right a general overview is given to locate the study area in central Ethiopia.
More details are found in the Supplemental materials section.
Results and discussion
The geomorphological map represents an inventory of landforms, and geomorphological processes. Endogenic processes result in structural features like faults, and volcanic surfaces. The latter are the major landforms dominating the area. Moreover, exogenic processes related to climatic conditions and specific land use pattern profoundly shape the landscape. At the mesoscale, landform units and processes of the surroundings of Melka Kunture are represented by the following forms and features related to (i) structural elements and forms, (ii) fluvial processes and forms; (iii) water erosion processes and forms; (iv) gravitative processes and forms; and (v) anthropogenic forms and features.
Structural elements and landforms
The major fault systems were identified using the terrain indices TCI and the visual interpretation of the shaded relief high-resolution DEM. The map shows parallel faults that can be attributed to a half-graben system striking in a WSW to ESE direction. Perpendicular to this direction secondary fault systems can be observed. The Melka Kunture archaeological area is located in a graben zone. Since the surrounding areas are uplifted the study region was subject to deposition processes documented by sedimentary units (Tuka Meja-, Kellaand Melka Formations) (see Figure 5).
Moreover, the river morphology reflects the main direction of the fault systems. Finally, the tectonic structures, in turn, trigger the lithological contacts between pyroclastic deposits and effusive rocks as well as the contacts to the sedimentary deposits. The so-called welded ignimbrites at the northern headwaters form major scarps. Basalts, welded ignimbrites and Kella Formation deposits made up of ignimbritic tuff and vesicular pumices. They occur at the south eastern border of the area, where tectonic contacts between these units can be observed.
Fluvial processes and related forms and features
The Upper Awash flows from the West to the East across the Oromia region, characterized by a low gradient meandering river morphology. It shows a high incision rate showing a moderate riffle/pool bed morphology. The classification proposed by Rosgen (1996) indicates a Stream Type C, consisting in a single thread channel with moderate to high depth and moderate to high sinuosity. At the confluence with the Atebela drainage, the banks are often less than two metres high and are detectable only by field observations. The tectonic activity played a strong role controlling river morphology and valley forms and features. Floodplains are well defined for low gradient river section whereas an increase of their slopes can be observed at the confluence with the Melka Dula tributary (see E3). At this section, the river cuts into the bedrock and becomes steeply entrenched corresponding to Stream Type A according to Rosgen (1996). Moreover, the section is characterized by high energy cascades. We define the Upper Awash tributaries in the study area generally as low gradient meandering with a low width/depth ratio and frequently eroded banks ( Figure 5). Their cross-sections are rectangular with almost vertical banks, of 2-4 m high, corresponding to Stream Type E following (Rosgen 1996). In the lower terrain positions the soils are well-drained, porous, and classified as Fluvisols showing recent signs of stratification and sedimentation (IUSS Working Group WRB 2014). The Fluvisols (Figure 4) receive fresh material due to regular flooding events.
Water erosion processes and related forms and features
Erosion processes affect mainly the soils and weak sedimentary deposits even on low sloping areas. Especially the Vertisols show high potential for soil erosion processes due to active layer clays and the resulting soil cracks forming during the dry season with up to 2 m depth. These cracks represent preferential pathways for water infiltrating during first precipitation events. Hence, forming pipes within the substrates and soils that grow and finally collapse representing initial stages of gully erosion ( Figure 6). Once the soil gets wet, the active layer clays swell and subsequently cracks close. Therefore, in wet conditions Vertisols may produce a remarkable amount of surface runoff following intensive precipitation events due to infiltration excess related to the small pore diameter of the clay dominated soil matrix. The surface runoff, in turn, detaches soil or substrate particles and transports them along the drainage network.
In the study area erosion processes can be observed particularly on the south-western slopes, adjacent to the Godeti gully (see map B5). Fine grained slope wash colluvial material is deposited at the base of the mayor incisions. High amounts of sediments are produced especially when single, intensive rainfall events affect freshly ploughed fields. Agricultural activity is mainly concentrating in the area north-west of the Wufi-Atebela archaeological sites.
Deep linear incision such as gully systems are found in the core archaeological area (e.g. Figure 6). We found continuous retrogressive dendritic gullies and discontinuous gullies related to local incisions. The length of these gullies varies between 10 and 1000 m. Gully systems tend to have a dendritic shape and a headwater catchment area of about 1000/10,000 m 2 . The headcuts of gullies are often characterized by pipe outlets. The depth of the gullies increases downstream, reaching a few metres close to the Awash River (e.g. Godedi area, B5). For and in-depth description of the gully erosion please see the Supplemental Materials section.
Gravitative processes and related forms and features
Rockfalls are the most frequent mass wasting process in the area, their shape ranges from 100 to 1000 m 2 . The main reason for these processes consists in the lithological discontinuities and the undercutting of the middle Pleistocene ignimbrites and tuffs belonging to the Kella formation (D4-D3). In fact, the escarpments of the Kella formation in the Tuka Meja drainage system are strongly eroded due to agriculture and cattle farming. Rock toppling also occur in the Kella formation and along the flood basalt escarpments (D4-B4) Because of the basalt pillars structure that has distinct and predefined fractures hence producing a sort of toppling.
Anthropogenic forms and features
Some parts of the study area are subject to the rapid growth of settlements. In particular, the Awash village expanded intensively in the last decades. This rapid development is connected to road constructions such as the new road connecting Butajira with Addis Ababa. The roads concentrate runoff and hence, road gullies developed. Moreover, road cuts led to accelerated erosion processes.
Human activities are mainly concentrated around the settlement areas and are characterized by intensive erosion processes, mainly due to over grazing and soil compaction. However, also the agricultural activity, especially after ploughing, increases the amount of sediments in the runoff after intensive rainfall events.
Conclusion
We utilized and integrated approach to identify and map the geomorphological processes of the Melka Kunture study area. Therefore, different methods like remote sensing, API, terrain analysis, as well as field measurements and observations were combined to generate a geomorphological map of the wider Melka Kunture area. We considered endogenic as well as exogenic geomorphological processes and related forms and features. The dominant landscape forming processes are related to the rift valley tectonics and the subsequent volcanic activities with their different depositional products. The latter changed the drainage pattern and hence fluvial erosion and deposition processes. Moreover, the stratigraphic sequence of the different volcanic and fluvial deposits leads to slope instabilities and mass movements like rockfalls and toppling (e.g. Figure 7).
Sheet and gully erosion processes contribute to the landscape evolution and are mainly triggered by anthropogenic activities such as agriculture and cattle farming. The latter lead to overgrazing and hence compaction of soils with increased surface runoff.
We provide the geomorphological map [Main Map] in the classic analogue 2D format and an appropriate scale of 1.15,000 as well as in an interactive WebGIS-based version using WMS layer and a Leaflet application offering additional information. The geomorphological map [Main Map] represents an overview of the different forms and features related to specific geomorphological processes. Hence, the map contributes to the understanding of landscape evolution in the area and provides basic information to assess geo-hazards affecting the Melka Kunture archaeological sites.
Disclosure statement
No potential conflict of interest was reported by the authors.
Funding
This study was carried out in the framework of the project entitled 'Integrated assessment of geomorphological processes dynamics on different spatiotemporal scales in the Ethiopian Highlands using remote sensing and advanced modelling approaches' financed by German Research Foundation (DFG) [grant number HO-1840/11-1212042855]. We would like to thank the ROCEEH project financed by Heidelberg Academy of Sciences and Humanities for travelling support. University of Tübingen provided lab and computing facilities.
Data availability statement
The digital data is hosted at the ROCCEH project facility at University of Tübingen and available upon request (www.roceeh.net). Derived data supporting the findings of this study are available from the corresponding author [M. M.] on request.
|
2019-10-10T09:16:31.088Z
|
2019-07-03T00:00:00.000
|
{
"year": 2019,
"sha1": "818c1ce26ec3c60304b70e46eac07fa9819f0c35",
"oa_license": "CCBY",
"oa_url": "https://www.tandfonline.com/doi/pdf/10.1080/17445647.2019.1669497?needAccess=true",
"oa_status": "GOLD",
"pdf_src": "Adhoc",
"pdf_hash": "f545f200dfb6f1b3fb101062e0d0eb188ee7434d",
"s2fieldsofstudy": [
"Geography",
"Geology"
],
"extfieldsofstudy": [
"Geography"
]
}
|
247147048
|
pes2o/s2orc
|
v3-fos-license
|
Amiodarone-Induced Multi-Systemic Toxicity Involving the Liver, Lungs, Thyroid, and Eyes: A Case Report
Objectives Amiodarone is widely used to treat arrhythmia. However, amiodarone is known for its severe toxicity to the liver, lungs, and thyroid. Amiodarone causes liver damage ranging from asymptomatic serum aminotransferase elevation to hepatic failure requiring liver transplantation. Although amiodarone toxicity has been reported, its simultaneous multi-organ toxicity is not well-known. Here, we introduce a novel case of multi-systemic amiodarone toxicity involving the liver, lungs, thyroid, and eyes. Case Presentation A 61-year-old woman visited the emergency room due to general weakness, nausea, visual disturbance, heat intolerance, and a non-productive cough. The patient had been using clopidogrel and amiodarone due to underlying atrial fibrillation. The total level of bilirubin was 0.71 mg/dL, aspartate aminotransferase was 358 U/L, alanine aminotransferase was 177 U/L, and prothrombin time was 27.1 s. Computed tomography showed diffuse increased liver intensity and scattered hyperattenuated nodular consolidations in both lungs. Transthoracic needle lung biopsy revealed fibrinoid interstitial inflammation with atypical change of type II pneumocytes and intra-alveolar foamy macrophages. In addition, the thyroid-stimulating hormone level was <0.008 μIU/mL, and free thyroxine was 4.67 ng/dL. The thyroid scan showed diffuse homogenous intake of technetium-99 m pertechnetate in both thyroid lobes. The ophthalmologic exam detected bilateral symmetrical corneal deposits in a vortex pattern. With these findings, we could diagnose amiodarone-induced hepatic, pulmonary, thyroid, and ophthalmologic toxicity. Liver function was restored after cessation of amiodarone, and thyroid function was normalized with methimazole administration. However, due to aggravated lung consolidations, systemic steroid treatment was administered, and improvement was seen 1 week after, at the follow-up exam. As her symptoms improved, she was discharged with a plan of steroid administration for 3 to 6 months. Conclusions This case implies the possibility of multi-systemic amiodarone toxicity. Thus, the toxicity of amiodarone to multiple organs must be monitored. Prompt cessation of the drug should be considered upon diagnosis.
INTRODUCTION
Amiodarone is an iodine-containing benzofuran derivative classified as a class III antiarrhythmic drug. It is clinically used for the treatment of tachyarrhythmias, including atrial fibrillation and reentrant tachyarrhythmias (1), which is one of the most prevalent types of arrhythmia (2). However, with the increasing use of amiodarone, there have been reports of side effects to various organs, including the lungs, liver, heart, thyroid gland, eyes, skin, and the nervous system (3).
Overall incidence of adverse effects can range from 30 to 90%, and serious side effects can take place in 10 to 26% of patients (1). Pulmonary toxicity is one of the fatal adverse effects of amiodarone, with mortality estimated between 1 and 33% and incidence of approximately 10% shown in previous studies (4). Patients with amiodarone-induced pulmonary toxicity usually present with dyspnea, non-productive cough, malaise, fever, and pleuritic chest pain (5). The most fatal manifestation of pulmonary toxicity is a rapidly progressing diffuse pneumonitis with acute respiratory distress requiring mechanical ventilation with mortality as high as 50 to 100% (6).
In addition, about 14-18% of patients taking amiodarone for a long period showed thyroid dysfunction, including hypothyroidism and thyrotoxicosis (15). However, a third of Korean patients developed thyroid dysfunction, and most cases involved hypothyroidism (16). Patients with amiodaroneinduced hypothyroidism usually present with fatigue, cold intolerance, and dry skin, similar to symptoms of classic hypothyroidism (17).
The most typical symptom of amiodarone-induced ocular toxicity is corneal microdeposits, 98% of which are found after 2 months of treatment (18). Asymptomatic corneal changes are observed in 50-60% of patients, and visual disturbance is rarely reported (19).
Although much is known about amiodarone toxicity, cases of simultaneous toxicity in various organs have rarely been reported. Here, we present a novel case of multi-systemic amiodarone toxicity, involving the liver, lungs, thyroid, and eyes.
CASE REPORT
A 61-year-old woman visited the emergency room due to general weakness, nausea, visual disturbance, heat intolerance, and a non-productive cough with dyspnea, which had persisted between 3 and 6 months. The patient had underlying ischemic heart disease and paroxysmal atrial fibrillation, for which she was being treated with clopidogrel, pitavastatin, valsartan, rivaroxaban, nicorandil, diltiazem, furosemide, and amiodarone. More specifically, amiodarone was prescribed for 34 months at an initial dose of 200 mg/day for 11 months, and then due to intermittent chest discomfort with palpitation, her cardiologist increased the maintenance dose to 400 mg/day for the next 23 months. The patient was not taking any other medications, including herbal agents, and did not have a history of alcohol consumption.
The laboratory workup showed that the level of the total bilirubin was 0.71 mg/dL (reference range, 0.22-1.3), the aspartate aminotransferase level was 358 U/L (reference range, 10-37), the alanine aminotransferase level was 177 U/L (reference range, , and the prothrombin time was 1.90 International Normalized Ratio (INR) (reference range, 0.8-1.2). Enhanced computed tomography (CT) showed diffusely increased liver intensity and scattered hyper attenuated nodular consolidations in the subpleural areas of both lungs (Figures 1A,B). To evaluate the possibility of malignancy, positron emission tomography-CT (PET-CT) was performed, and multiple hypermetabolic lesions (maximal standardized uptake values 7.3, Figure 1C) were examined. To confirm the diagnosis of the lung lesions, transthoracic needle biopsy was performed, and the pathologic examination showed fibrinoid interstitial inflammation with atypical change of type II pneumocytes and intra-alveolar foamy macrophages (Figure 2). Using both the CT and biopsy findings, we could diagnose the amiodarone-induced hepatic and pulmonary toxicity.
Further, to assess the possibility of amiodarone-induced thyroid dysfunction, we examined thyroid function tests. The thyroid-stimulating hormone level was less than 0.008 uIU/mL (reference range, 0.55-4.78), free thyroxine was 4.67 ng/dL (reference range, 0.89-1.76), triiodothyronine was 1.27 ng/mL (reference range, 0.6-1.81), thyroid binding inhibitor immunoglobulin (TBII) was 18.3 IU/L (reference range, 0-1.5), anti-microsome antibody was 139.0 U/mL (reference range, 0-60), and anti-thyroglobulin antibody was 125 U/L (reference range, 0-60). The PET-CT scan of the thyroid showed no metabolic lesion ( Figure 3A). The thyroid scan (technetium-99 m pertechnetate scintigraphy) showed diffuse homogenous intake of technetium-99 m pertechnetate in both thyroid lobes ( Figure 3B). Anti-microsomal antibody and anti-thyroglobulin antibody may be detected in Graves' disease (anti-microsomal antibody: 69.2%, anti-thyroglobulin antibody: 23-30%) (20); furthermore, with the results of TBII and the thyroid scan results, we diagnosed our patient as having type 1 amiodarone-induced thyrotoxicosis, a form of iodine-induced hyperthyroidism caused by excessive, uncontrolled biosynthesis of thyroid hormone by autonomously functioning thyroid tissue in response to iodine load, which typically develops in underlying latent Graves' disease (21). After consultation with the endocrinology department, methimazole (20 mg for 17 days followed by 5 mg/d) was used as treatment to manage excessive thyroid hormone synthesis.
The patient persistently complained of blurred vision, and the ophthalmologic exam detected bilateral symmetrical corneal deposits in a vortex pattern (Figure 3C), which was compatible with amiodarone-induced vortex keratopathy.
As amiodarone toxicity was considered as the possible cause of disease, amiodarone was discontinued from the first day of admission. As there was neither extensive lung involvement nor hypoxemia, steroid treatment was not immediately initiated. However, serial follow-up chest CT scans showed aggravation of previous lung lesions. After exclusion of lung malignancy via transthoracic needle biopsy, we decided on treatment with a systemic steroid regimen (methylprednisolone 40 mg) for pulmonary toxicity from day 29 of hospitalization. After 1 week of steroid medication, a follow-up CT showed an overall decreased extent of high attenuated subpleural masslike consolidations and small nodules in both upper lung fields and the right middle lobe, suggesting an improved state of R/O amiodarone-induced pulmonary toxicity. After 4 months of steroid medication, a follow-up CT showed resolution of previous lung consolidations (Figure 4A). The laboratory examination showed that both aspartate aminotransferase and alanine aminotransferase levels were within the normal range after cessation of amiodarone for 1 month without any treatment ( Figure 4B). Free thyroxine and thyroid-stimulating hormone were within the normal range after 6 weeks of methimazole medication ( Figure 4C). The patient recovered from her general weakness, cough, and hand tremor 1 month after amiodarone cessation and was discharged after most of the symptoms were resolved.
DISCUSSION
To the best of our knowledge, this is the first report demonstrating multi-systemic toxicity, involving four vital organs, induced by amiodarone. Amiodarone is a widely prescribed antiarrhythmic drug used to treat tachyarrhythmias, including atrial fibrillation and reentrant tachyarrhythmias of the accessory pathways. Amiodarone principally acts through multiple mechanisms, such as depressing the sinus and atrioventricular nodes and prolonging repolarization and refractoriness in the myocardium.
Amiodarone is a lipophilic drug, which accumulates mainly in adipose tissue and organs with high blood perfusion, such as the liver, lungs, and skin. Its long-term use leads to numerous adverse reactions, including photosensitivity, hypothyroidism, hyperthyroidism, hepatic dysfunction, bone marrow suppression, corneal microdeposits, and neuromotor defects (3). Due to its long biological half-life ranging up to 100 days, the drug activity may last more than 3 to 4 months after discontinuation (1).
Hepatotoxicity is a relatively uncommon adverse reaction to amiodarone. Amiodarone induces histological findings similar to alcohol-induced steatohepatitis, so differential diagnosis from alcoholic liver disease should be considered. Pathologic findings include macro-and microvesicular steatosis, Mallory bodies, polymorphonuclear leukocyte infiltration, ballooning degeneration of hepatocytes and phospholipidosis (22). The possible reason for hepatotoxicity is amiodarone, as its principal metabolite, desethylamidoarone induces production of reactive oxygen species and leads to hepatic triglyceride accumulation and microvesicular steatosis in hepatocytes (23). Pulmonary toxicity, an uncommon but serious adverse effect of amiodarone, reveals various patterns of lung involvement, including chronic interstitial pneumonia, bronchiolitis with or without organizing pneumonia, acute respiratory distress syndrome, diffuse alveolar hemorrhage, pleural effusion, and pulmonary nodules or masses. Especially, pulmonary mass and pulmonary nodules can mimic lung malignancy (24). The mechanism underlying the development of nodular pulmonary disease may be the inflammatory response to the accumulation of phospholipids in alveolar cells induced by the drug (25). Corticosteroid therapy should be considered in patients showing extensive lung involvement in imaging or those that develop hypoxemia, with a starting dose of 40 to 60 mg/day of prednisolone (26,27). Due to the structural similarity between amiodarone and thyroid hormones, amiodarone causes thyroid dysfunction due to excessive iodine overload or due to its direct cytotoxicity to the thyroid gland (28). Amiodarone can induce both hypothyroidism and hyperthyroidism. The incidence of amiodarone-induced thyrotoxicosis is 2 to 10%, and it more commonly develops in areas of the world where iodine deficiency is common (29). The type of thyroid dysfunction caused by amiodarone treatment may depend on the presence of underlying thyroid disease or dietary iodine content (30). Moreover, amiodarone leads to drug-induced lipidosis, which leads to typical side-effects in the eyes, such as vortex keratopathy (31). Although the most common findings are corneal and lens opacities, optic neuropathy has also been reported (32). There have been reports of amiodarone toxicity to one or two vital organs. Cho et al. (33) reported a case of a 65-year-old man with amiodarone-induced hepatitis and hypothyroidism. Turk et al. (34) also reported a case of a 54year-old female with amiodarone toxicity to the skin, thyroid, and eyes.
In this study, we observed multi-systemic amiodarone toxicity involving the liver, lungs, thyroid, and eyes. The main mechanism of amiodarone-induced toxicity is known as direct cytotoxicity and immunologic reaction. Patients who have received a daily dose of 400 mg or more for more than 2 months or a lower dose for more than 2 years are considered at high risk (26). Our patient can be considered as high-risk, as she was on 400 mg/day amiodarone for nearly 2 years. The long duration and high dosage of amiodarone were combined with treatment that may interact pharmacologically with amiodarone metabolism. First of all, she had been taking a statin concurrently with amiodarone for 3 years. Many studies have reported an interaction between statins and amiodarone as the potential cause of hepatotoxicity due to inhibition of the mitochondrial enzyme CYP3A4, which metabolizes statins, by amiodarone (35)(36)(37)(38). Additionally, the patient in this study took diltiazem, a recognized CYP3A4 inhibitor that may have jeopardized amiodarone metabolism. Furthermore, our patient showed significant improvement of lung lesions with the use of systemic steroid therapy, suggesting the usefulness of corticosteroids in extensive lung involvement, as reported in previous studies (26,27).
Our study has some limitations. First, although our report indicates multi-systemic amiodarone toxicity, the number of cases examined was very small. Second, the precise mechanism of amiodarone toxicity to multiple vital organs is not clearly verified. Lastly, the drug interactions associated with amiodarone metabolism, which may be the cause of amiodarone toxicity in our patient, had not been thoroughly examined. Further studies are warranted to elucidate the mechanism of multiorgan toxicity and classify high-risk patients to prevent amiodarone-induced toxicity.
In conclusion, this case indicates that multi-systemic amiodarone organ toxicity is a cause of concern in high-risk patients. Amiodarone toxicity should be considered in patients receiving amiodarone with any new symptoms, because this may involve various organs. A multisystem approach to amiodarone toxicity is important, because adverse events may occur in various organs simultaneously. Guidelines for monitoring adverse events in patients taking amiodarone for a long time must be established. Once diagnosis is determined, amiodarone must be discontinued promptly, and the use of systemic steroids needs to be assessed.
DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.
AUTHOR CONTRIBUTIONS
H-SY and JY conceived and designed the study, reviewed the literature, and contributed to manuscript drafting. SBC and SKC contributed to manuscript drafting. JY reviewed the case and edited the manuscript. Y-DC reviewed the pathologic findings. YK reviewed the ophthalmologic findings. WC and H-CK contributed to interpretation of data and reviewed of the case regarding amiodaroneinduced thyrotoxicosis. All authors approved the final version to be submitted and approved the publication of the manuscript.
|
2022-02-28T14:12:08.979Z
|
2022-02-28T00:00:00.000
|
{
"year": 2022,
"sha1": "a02309d6a5f151a19f917a3bdff32b74ab2452f7",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Frontier",
"pdf_hash": "a02309d6a5f151a19f917a3bdff32b74ab2452f7",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
257671266
|
pes2o/s2orc
|
v3-fos-license
|
PDF Malware Detection Based on Fuzzy Unordered Rule Induction Algorithm (FURIA)
: The number of cyber-attacks is increasing daily, and attackers are coming up with new ways to harm their target by disseminating viruses and other malware. With new inventions and technologies appearing daily, there is a chance that a system might be attacked and its weaknesses taken advantage of. Malware is distributed through Portable Document Format (PDF) files, among other methods. These files’ adaptability makes them a prime target for attackers who can quickly insert malware into PDF files. This study proposes a model based on the Fuzzy Unordered Rule Induction Algorithm (FURIA) to detect PDF malware. The proposed model outperforms currently used methods in terms of reducing error rates and increasing accuracy. Other models, such as Naïve Bayes (NB), Decision Tree (J48), Hoeffding Tree (HT), and Quadratic Discriminant Analysis (QDA), were compared to the proposed model. The accuracy achieved by the proposed model is 99.81%, with an error rate of 0.0022.
Introduction
Intelligent attacks utilizing documents with malicious codes have been increasing rapidly in recent years as file transfers expand.The majority of Internet users are aware of the risk posed by execution files that are attached to emails or web pages.However, because users are unaware of the documents, they serve as an effective means of spreading malware.Because Portable Document Format (PDF)s are more flexible than other document formats, they are one of the main attack vectors among the malware that has been identified.The majority of malicious PDF documents include JavaScript or binary scripts that exploit certain security flaws and carry out destructive deeds, as explained in [1].On the internet, there are countless billions of PDF files.Not all of them are as benign as one might expect.In actuality, PDF files may include a variety of objects, including binary or JavaScript codes.These items might occasionally be harmful.Malware software could try to infect a computer by finding a reading weakness [2].In 2017, Adobe Acrobat Reader was found to have sixty-eight vulnerabilities.There are about fifty of them that may be used to run arbitrary codes.Each reader has certain weaknesses, and a malicious PDF file may discover a method to exploit them [3].
Reason for the Selection of PDF Files
The PDF format is one of the most widely used file types for sharing digital documents between different platforms and applications.Refs.[4,5] contain a full description of the PDF standard.A few elements make PDFs one of the preferred file types for malware authors to disseminate dangerous content.(a) PDF is extensively utilized by people in both professional and social settings.Academic papers, technical reports, design documents, and electronic receipts are a few examples of typical instances; (b) PDF is independent of platforms and operating systems (OS).A standalone PDF reader or a modern web browser can be used to access a PDF file on a Windows PC, a Linux system, or a mobile device (with a PDF viewer plug-in); (c) it is a very versatile file format.In addition to text, PDF also allows other sorts of data, such as video files, interactive forms, links to other files, JavaScript, Flash, and unified resource locators (URLs).Additionally, different encoding and compression techniques can be utilized to reduce file size, conceal important material, or both; and (d) it is stealthy and sophisticated.In general, executable files are thought to be more dangerous than PDF files.Setting a policy to prohibit staff members from downloading executable files from the internet or including them in email attachments is a common security measure, but it is uncommon to do the same with PDF documents.The enormous ubiquity and adaptability of the PDF file format also provide attackers with several opportunities to spread malware through PDF documents.
PDF-Based Malware
Phishing and exploits are the two main types of PDF-based attacks.Phishing attempts frequently appear in emails.A typical instance is a PDF delivery or purchase confirmation receipt attached to an email that seems to be from a trustworthy online store or logistics company.Apart from the social engineering techniques used to persuade recipients to open phishing PDF attachments, the text content of such emails is largely meaningless.These PDF documents are typically one page long and include social engineering elements as well as a phishing URL that leads to a suspicious website where malicious downloads, personal data collection, and other activities can be carried out.In contrast to plain-textbased phishing efforts, PDF documents include binary or a blend of binary and ASCII languages, making them harder to detect.This is one of the factors contributing to the rise in the popularity of phishing attempts based on PDFs.Its motivation is identical to that of standard phishing scams.The information that attackers collect from victims may be used by them or may be sold on the illicit market for use in the so-called shadow economy [6].
The PDF file format is a popular option for use in offices because of its high efficiency, dependability, and interactivity.The development of non-executable file assault technologies and attack techniques such as the advanced persistent threat has seriously jeopardized PDF's security since malicious PDF files are the most researched infection routes in adversarial scenarios [7,8].With the development of machine learning (ML) technology [9] in recent years, researchers have developed a variety of ML-based techniques to recognize distinct attack types related to PDF files.The technology for detecting malicious PDF files may be divided into techniques based on static analysis, dynamic analysis, and techniques based on a combination of static and dynamic analyses.Modern research has demonstrated that PDF detectors based on ML may achieve excellent accuracy with a remarkably low false positive rate (FPR) [10].However, such a study focuses on the proposed Fuzzy Unordered Rule Induction Algorithm (FURIA) for malware detection in PDF files compared with Naïve Bayes (NB), Decision Tree (J48), Hoeffding Tree (HT), and Quadratic Discriminant Analysis (QDA).These models are compared based on some of the well-known assessment measures, including accuracy (ACC), F-measure (FM), recall, precision, Matthew's correlation coefficient (MCC), and mean absolute error (MAE).This study has two primary objectives:
•
To propose a malware detection model that will protect the systems from any harmful activity caused by PDF malware;
•
To compare the findings from the suggested and existing models in use to discover a better and more effective solution for PDF malware detection.
The main contributions of this study are summarized as follows: • We propose a FURIA-based model for the PDF malware detection; The rest of this paper is organized as follows: Section 2 summarizes the literature review.Sections 3 and 4 discuss the methodology and results analysis and discussion, respectively.Finally, Section 5 concludes this work.
Literature Review
Several studies have been conducted on PDF malware detection using various ML and deep learning (DL) models.The use of the Portable Document Format was explained by Reum et al. [11].They provided a comprehensive study of the JavaScript content and structure found in the XML-embedded PDF.After that, they developed a range of features, including configuration and metadata such as file size, keywords, versions, and content features, as well as encoding strategies such as keywords, names, and JavaScript-readable strings.Due to the complexity of its features and the robustness of machine learning algorithms to small modifications, adversarial examples are challenging to construct.To reduce the possibility of adversarial assaults, they also develop a recognition model utilizing black-box-style models with structure and content properties.Utilizing observable robustness features, Chen et al. [12] described how to train robust PDF malware classifiers.For instance, a classifier must always recognize PDF malware as dangerous, no matter how many pages from benign forms are placed into the document.They show how to rigorously assess a malware classifier's worst-case behavior concerning specific robustness characteristics.
ML techniques have been used to create classifiers for PDF malware in several projects.Wepawet [13] and PJScan [14] were two earlier efforts that concentrated on the harmful JavaScript that was included in PDF malware.These tools include a JavaScript code extractor and a classifier for malicious JavaScript that can be either dynamic or static.Recent PDF malware classifiers have concentrated on structural aspects of PDF files since not all PDF malware contains embedded JavaScript and because PDF malware developers have learned several ways to conceal JavaScript codes [15].We aim to develop cutting-edge structural feature-based classifiers in this effort.There have been studies that specifically looked at the JavaScript codes in PDFs.Features based on functions, constants, objects, methods, and keywords, as well as lexical characteristics of JavaScript scripts, were established by Khitan et al. [16].Zhang [5] also utilized elements from the PDF structure, entity characteristics, metadata information, and content statistics, along with JavaScript features, including the number of objects, number of pages, and stream filtering information.Based on the finding that malicious JavaScript functions differ from legitimate JavaScript functions, Liu et al. [17] presented a context-aware technique.This method opens the PDF file while monitoring suspicious behavior based on JavaScript statements by passing the original code as input to the "eval" function.
In [18], Smutz and Stavrou combined the PDF parser with a random forest classifier to identify fraudulent PDF files using information gleaned from document metadata and file structure.They looked into 202 features, including /Font and /JavaScript.According to Liu et al. [17], current protections against malicious PDFs are inadequate, prone to evasion, and too computationally costly to be utilized online.They recommended leveraging static and run-time features to identify JavaScript in context.A software engineering approach was used to provide a detection method based on behavioral differences in those systems since a PDF document acts identically on different platforms [19].A malicious document, on the other hand, will act differently depending on the platform.According to Li et al. [20], the weakness of all harmful detection methods that extract JavaScript is their dependency on 3rd-party extraction tools that precisely follow the Acrobat standard.A bigger training dataset does not always result in improved detection, according to Scofield et al. [21], and very little research has been conducted to establish the minimal size of a dataset required to achieve high detection accuracy.As a consequence, ref. [21] proposes a dynamic analysis-based detection technique.
Research Methodology
This study aims to develop a FURIA-based model for PDF malware detection.The overall research methodology is presented in Figure 1, which starts from data acquisition to comparison and performance analysis of each employed model.The dataset used in this study has been taken from the University of New Brunswick (UNB), Canadian Institute for Cybersecurity, http://205.174.165.80/CICDataset/CICEvasivePDFMal2022/Dataset/ (accessed on 5 February 2023).The dataset consists of 33 features, of which 32 are independent and 1 is dependent.The first 11 features are removed because they do not act in the analysis phase.These attributes are known as general features, including PDF size, metadata size, encryption, header, page number, text, image number, font objects, object number, number of embedded files, and the average size of all the embedded media.For extracting such features, we have used ClassifierAttributeEvaluator methods using a ZeroR classifier and the Ranker searching method.The selected features are ranked as follows: Selected Features: These features are titled as: colors, startxref, pageno, ObjStm, trailer, xref, endstream, stream, endobj, encrypt, JS, XFA, launch, EmbeddedFile, Javascript, RichMedia, JBIG2Decode, Acroform, OpenAction, AA, and obj, respectively.Table 1 presents the description of each selected feature.
the Acrobat standard.A bigger training dataset does not always result in improved detection, according to Scofield et al. [21], and very little research has been conducted to establish the minimal size of a dataset required to achieve high detection accuracy.As a consequence, ref. [21] proposes a dynamic analysis-based detection technique.
Research Methodology
This study aims to develop a FURIA-based model for PDF malware detection.The overall research methodology is presented in Figure 1, which starts from data acquisition to comparison and performance analysis of each employed model.The dataset used in this study has been taken from the University of New Brunswick (UNB), Canadian Institute for Cybersecurity, http://205.174.165.80/CICDataset/CICEvasivePDFMal2022/Dataset/(accessed on 5 February 2023).The dataset consists of 33 features, of which 32 are independent and 1 is dependent.The first 11 features are removed because they do not act in the analysis phase.These attributes are known as general features, including PDF size, metadata size, encryption, header, page number, text, image number, font objects, object number, number of embedded files, and the average size of all the embedded media.For extracting such features, we have used ClassifierAttributeEvaluator methods using a ZeroR classifier and the Ranker searching method.The selected features are ranked as follows: Selected Features: These features are titled as: colors, startxref, pageno, ObjStm, trailer, xref, endstream, stream, endobj, encrypt, JS, XFA, launch, EmbeddedFile, Javascript, RichMedia, JBIG2Decode, Acroform, OpenAction, AA, and obj, respectively.Table 1 presents the description of each selected feature.
JS
The proportion of Javascript-containing objects.
Javascript
This indicates the amount of items that include a Javascript code, the most often used feature, as is clear.AA specifies a particular response to an event.OpenAction Defines a specific action to be taken when the PDF file is opened.The bulk of common malicious PDF files have been found to use this functionality in conjunction with Javascript.Acroform Form fields in Acrobat forms, which are PDF files, offer scripting technology that may be abused by hackers.JBIG2Decode A popular filter for encoding harmful stuff is JBig2Decode.How many items have nested filters?Nested filters can make decoding more challenging and may be an indicator of evasion.
Richmeddia
The quantity of flash files and embedded media is indicated by the number of RichMedia keywords.Launch A command or program can be run by using the term launch.
EmbeddedFile
PDFs can attach or embed a variety of things inside themselves that may be exploited, such as additional PDF files, Word documents, pictures, etc. XFA Certain PDF 40 files contain XFAs, which are XML Form Architectures that offer scripting technologies that can be abused by attackers.
Color
In the PDF, many colors are utilized.
Class
Classify as malicious or benign.
For model training and testing, a standard method of K-fold validation [23,24] is used.Here, the value of K is selected as 10.The performance of each employed model is evaluated using some of the standard evaluation metrics, including mean absolute error (MAE), recall, precision, Matthew's correlation coefficient (MCC), FM, and classification ACC.These measures can be calculated as follows: Here, the important thing is to discuss the use of MAE in this study.MAE is typically used as an evaluation metric in regression problems, where the goal is to predict a continuous numerical output.However, in some cases, MAE can also be used in classification problems to evaluate the performance of the classification model.In classification, the output is a categorical variable, so using MAE as the primary evaluation metric might not be as informative as other classification-specific metrics such as accuracy, precision, recall, F1-score, or AUC-ROC.These metrics provide a more detailed understanding of how well the model performs in terms of correctly identifying positive and negative examples.However, in some cases, using MAE in classification can provide additional insights into the model's performance.In this scenario, MAE can be used to evaluate how far the predicted probabilities are from the true labels.
Fuzzy Unordered Rule Induction Algorithm (FURIA)
The FURIA is a new algorithm introduced by Huhn and Hullermeier that is responsible for generating fuzzy logic rules from a given database and classifying it using the obtained rules [22].Fuzzy logic algorithms are well-known for their properties, such as classification rules that are simply understood by the reader, the capacity to analyze linguistic input, and the ability to enable expert judgment.They can also be used as a tool for classification purposes [25,26].FURIA is the advanced version or derivative of the Repeated Incremental Pruning to Produce Error Reduction (RIPPER) algorithm.The RIPPER Algorithm is a classification algorithm based on rules.The training set is used to generate a set of rules.It is a famous rule induction algorithm.It generates fuzzy rules rather than traditional rules to replicate more flexible classification parameters.The fuzzy rules are constructed by substituting fuzzy intervals with a trapezoidal relevance function in association with the original RIPPER algorithm's advanced rule induction approach [27].
It has an appealing feature, which is the rule's extension.The generalization of the laws to include every option is the extension.It is a local technique that looks for information near the query.The simplest way to find the smallest generalization of a rule is to exclude those antecedents that are not met by the query [22].The pseudo-code for a single rule, r, is shown in Algorithm 1 below, and the flowchart of FURIA is presented in Figure 2.
Let A be the set of numeric antecedents of r While A = ∅ do a max ← null (a max denotes the antecedent with the highest purity ) pur max ← 0 (pur max is the highest purity value, so f ar) f or i ← 0 to size (A) do Compute the best fuzzification of
Results, Analysis, and Discussion
This section presents and discusses the study's findings.A new model, namely FURIA, is presented for PDF malware detection.FURIA and other benchmarked models are evaluated using a variety of criteria, which are ACC, FM, MCC, MAE, recall, and precision.Figure 3 illustrates the true positive rate (TPR) and false positive rate (FPR) analyses of each model compared with the proposed model.These analyses show the better performance of the FURIA, with the lowest FPR and better TPR.In both situations, it can be found that NB shows the worst outcomes.Figure 4 illustrates the outcome assessed via MAE.The MAE analysis shows the better performance of the proposed model with the lowest error rate, which is 0.0022, and the worst performance of NB with an error rate of 0.0147.All these values are achieved using a confusion matrix.Confusion matrix values achieved via each model are presented in Table 2.
Figure 5 presents the outcomes assessed via precision, recall, and FM.These outcomes also depict the better performance of FURIA, with a value of 0.998 for precision, recall, and FM, respectively.The HT and QDA have the same outcomes of 0.993 for recall, precision, and FM, respectively, while the NB shows the poorest performance with a value of 0.985 individually for recall, precision, and FM.
Results, Analysis, and Discussion
This section presents and discusses the study's findings.A new model, namely FURIA, is presented for PDF malware detection.FURIA and other benchmarked models are evaluated using a variety of criteria, which are ACC, FM, MCC, MAE, recall, and precision.Figure 3 illustrates the true positive rate (TPR) and false positive rate (FPR) analyses of each model compared with the proposed model.These analyses show the better performance of the FURIA, with the lowest FPR and better TPR.In both situations, it can be found that NB shows the worst outcomes.Figure 4 illustrates the outcome assessed via MAE.The MAE analysis shows the better performance of the proposed model with the lowest error rate, which is 0.0022, and the worst performance of NB with an error rate of 0.0147.All these values are achieved using a confusion matrix.Confusion matrix values achieved via each model are presented in Table 2 Figure 5 presents the outcomes assessed via precision, recall, and FM.These outcomes also depict the better performance of FURIA, with a value of 0.998 for precision, recall, and FM, respectively.The HT and QDA have the same outcomes of 0.993 for recall, precision, and FM, respectively, while the NB shows the poorest performance with a value of 0.985 individually for recall, precision, and FM.
x FOR PEER REVIEW 9 of 12 Figure 6 illustrates the evaluation of each model using R 2 , accuracy, and a logarithmic trendline.The logarithmic trendline, which is extremely useful when the rate of change in the data is rapidly increasing or falling and then leveling out, is the best-fit curved line.The positive and negative values can both appear on a logarithmic trendline [28] Figure 6 illustrates the evaluation of each model using R 2 , accuracy, and a logarithmic trendline.The logarithmic trendline, which is extremely useful when the rate of change in the data is rapidly increasing or falling and then leveling out, is the best-fit curved line.The positive and negative values can both appear on a logarithmic trendline [28].It may be obtained as follows: trendline.The logarithmic trendline, which is extremely useful when the rate of change in the data is rapidly increasing or falling and then leveling out, is the best-fit curved line.
The positive and negative values can both appear on a logarithmic trendline [28].It may be obtained as follows: Here, "ln" is the natural logarithmic function, and a and b are constants in the equation.The following generic equations, which differ only in the most recent input, can be used to retrieve the constants: R 2 , also known as the coefficient of determination, is a statistical measure that represents the proportion of variance in the dependent variable (or the outcome) that can be explained by the independent variable(s) (or the predictor(s)) in a regression model [28].It can be calculated as: R − squared = Explained Variation Total Variation (10) R 2 is consistently between 0% and 100%.The model does not take into consideration any fluctuation in the answer data around its mean, as shown by the 0%.100% means that the model fully accounts for all the variability in the response data surrounding its meaning.
The properties of PDF encouraged hackers to take advantage of several security flaws and circumvent security measures, making the PDF format one of the most effective attack vectors for harmful malware.Therefore, it is essential for information security to accurately recognize malicious PDF files.To this end, this study proposes a model based on FURIA for PDF malware detection.According to the analysis described in the preceding section, the proposed FURIA performs better than other employed models in terms of increasing accuracy and reducing the error rate.The accuracy percentage difference (PD) between FURIA and other applied algorithms is shown in Figure 7.This analysis shows that there is very little difference between the proposed model and the J48, which is only 0.12%, while the difference between the proposed model and the NB is greater than other employed models, which show the poorest performance of the NB as compared to the proposed model.The value of PD can be obtained using Equation (11), where x1 represents the value of FURIA and x2 represents the value of other employed algorithms.11) difference (PD) between FURIA and other applied algorithms is shown in Figure 7. T analysis shows that there is very little difference between the proposed model and the J which is only 0.12%, while the difference between the proposed model and the NB greater than other employed models, which show the poorest performance of the NB compared to the proposed model.The value of PD can be obtained using Equation ( The advantage of using FURIA is that it works well on datasets with imbalanced cla distributions.If a dataset has a large number of records and the majority of those recor fall into one class but the remaining records fall into other classes, the dataset is said have an unbalanced distribution of classes [22,28].We also have an unbalanc distribution of the data in the dataset, which is why the performance of the project model is better as compared with other employed models.The advantage of using FURIA is that it works well on datasets with imbalanced class distributions.If a dataset has a large number of records and the majority of those records fall into one class but the remaining records fall into other classes, the dataset is said to have an unbalanced distribution of classes [22,28].We also have an unbalanced distribution of the data in the dataset, which is why the performance of the projected model is better as compared with other employed models.
FURIA and NB
The data obtained from the UNB are used in all of the experiments.The rest of the algorithms in use are assessed using several common assessment metrics, such as MAE, precision, FM, recall, MCC, and accuracy, along with the proposed model.The models are trained and evaluated using the 10-fold cross-validation method.The threat now is that if the dataset is changed, the new results could outperform our analysis.The findings might potentially be affected by changing the criteria for data training and testing in place of the 10-fold cross-validation, for example, by using a percentage division.Another risk is that if a new algorithm is developed and it proves to be more effective than the one we now use, the results might be improved.
Conclusions and Future Direction
This study proposed a FURIA-based model for PDF malware detection.The proposed model is benchmarked with some of the well-known ML models, which are NB, J48, HT, and QDA.The performances of all these models are evaluated using some of the standard assessment measures that include MAE, ACC, FM, MCC, precision, and recall on the dataset taken from the UNB repository.The overall outcome presents a better performance of the proposed model, with an accuracy of 99.81% and a lowest error rate of 0.0022.FURIA outperforms other models; however, there are some limitations of FURIA.FURIA generates a large number of rules, which can make it difficult to understand and interpret the resulting model.This complexity can also lead to longer processing times and increased computational resources.Although fuzzy rules can be more interpretable than other machine learning models such as neural networks, they can still be difficult to interpret in complex data sets, which can limit their usefulness in some applications.
The FURIA-based model outperforms other well-known machine learning models for PDF malware detection; there are several potential future directions for research.There may be opportunities to refine the model further.For example, the model's parameters could be further optimized, or new features could be added to improve the model's performance.It may be beneficial to explore the potential benefits of combining the FURIA-based model with other ML models.Such hybrid models have the potential to improve the overall performance of the model.
Figure 3 .
Figure 3. TPR and FPR analysis of each model.
Figure 4 .
Figure 4. Proposed model comparison based on MAE.
Figure 3 .
Figure 3. TPR and FPR analysis of each model.
Figure 4 .
Figure 4. Proposed model comparison based on MAE.
Figure 4 .
Figure 4. Proposed model comparison based on MAE.
Figure 5 .
Figure 5. Models comparison based on precision, recall, and F-measure.
Figure 5 .
Figure 5. Models comparison based on precision, recall, and F-measure.
Figure 6 .
Figure 6.Accuracy analysis through each employed model.
Figure 6 .
Figure 6.Accuracy analysis through each employed model.Here, "ln" is the natural logarithmic function, and a and b are constants in the equation.The following generic equations, which differ only in the most recent input, can be used to retrieve the constants: a = INDEX(LINEST(y, LN(x)), 1) (8) b = INDEX(LINEST(y, LN(x)), 1, 2) (9)
Table 1 .
Selected features with their descriptions. .
Table 2 .
Confusion matrix values achieved via each employed model.
1 where x1 represents the value of FURIA and x2 represents the value of other employ algorithms.
Figure 7. Accuracy percentage difference between FURIA and other employed models.
|
2023-03-23T15:13:25.327Z
|
2023-03-21T00:00:00.000
|
{
"year": 2023,
"sha1": "1eac6768ba5297e39a32cfa69abba8a3e8f86223",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2076-3417/13/6/3980/pdf?version=1679390791",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "c3613e6f5f394220608378de0c038a2e716d62cc",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": []
}
|
119566961
|
pes2o/s2orc
|
v3-fos-license
|
An Algebraic Multigrid Method for Eigenvalue Problems
An algebraic multigrid method is proposed to solve eigenvalue problems based on the combination of the multilevel correction scheme and the algebraic multigrid method for linear equations. The algebraic multigrid method setup procedure is applied to construct the hierarchy and the intergrid transfer operators. In the algebraic multigrid scheme, a large scale eigenvalue problem can be solved by some algebraic multigrid smoothing steps in the hierarchy and some eigenvalue problems solving in a very small dimension. Some numerical experiments are presented to validate the efficiency of the proposed algorithm.
Introduction
Algebraic multigrid (AMG) method was introduced first in [2], where the main idea is to design a similar multigrid method for matrices. However, since there is no geometric background, the convergence has been proved only for some special matrices, such as symmetric positive definite M-matrices with weak diagonal dominance [18] and without the assumption of M-matrices in [11,16]. The essential difficulties for AMG method lie in the choice of coarse grid and intergrid transfer operators, which fully depend on our understanding of algebraic smooth error under certain smoothing processes. The classical coarsening strategy was introduced in [18], and others like aggregation and smooth aggregation in [17,20], compatible relaxation [4,14], based on element interpolation [5], energy-based strategy [3] and so on. The paper [8] presents some numerical experiments to study the robustness and scalability of the AMG method. Parallel and adaptive AMG methods have also been studied in [6,10]. Due to its simplicity, the AMG method has been applied to many problems, such as [1,9,15], etc.
In this paper, we are interested in the generalized eigenvalue problem: Find (λ, u) ∈ R × R n such that u T Mu = 1 and Au = λMu, (1.1) where A and M are real, symmetric N ×N matrix, and u is a vector in R n . The concerned generalized eigenvalue problem (1.1) always aries from the discretization of the elliptic partial differential equations involved in several scientific and theoretical fields such as material sciences, electromagnetics, quantum chemistry, acoustic, etc. These important applications usually require high resolution which means the discretization results in large scale algebraic eigenvalue problems. Then it is very useful to design efficient eigensolvers which need nearly optimal computational complexity. It is a natural idea to use the AMG method for eigenvalue problems. A very good review of the application of AMG method to eigenvalue problems is given in [12] and references cited therein. Roughly speaking, in the normal strategies, the AMG method is adopted as the smoother for linear equations in the inner iteration combined with some types of outer iterations for eigenvalue problems such as inverse power, shift-and-inverse, Rayleigh-quotient, locally optimal block preconditioned conjugate gradient and so on. Recently, a type of multilevel correction method is proposed to solve eigenvalue problems in [13,21,22]. In this multilevel correction scheme, the solution of eigenvalue problem on the final level mesh can be reduced to a series of solutions of standard boundary value problems on the multilevel meshes and a series of solutions of the eigenvalue problem on the coarsest mesh. Therefore, the computational work and required memory can arrive at the optimality. Similarly to the AMG method for boundary value problems, we can also design a type of AMG method for eigenvalue problems based on the multilevel correction method. The aim of this paper is to present an AMG method for eigenvalue problems (1.1).
The rest of this paper is organized as follows. In the next section, we introduce the classical AMG method, mainly the constructing of "coarse-grid". An AMG algorithm for solving the eigenvalue problem is presented and analyzed in Section 3. In Section 4, some numerical tests are presented to validate the efficiency of the proposed algorithm. Some concluding remarks are given in the last section.
Classical AMG
This section is devoted to introducing the classical AMG method which aims at solving the ill-conditioned linear system Au = f similar to geometric multigrid (GMG) method. Since there is no true geometric background, the main content is to determine the "coarse-grid" and intergrid transfer operators directly from the matrix A. By analogy, we define grid points, Ω, as the indices {1, 2, · · · , N} of u = (u 1 , u 2 , · · · , u N ) T , and choose a subset of Ω as the coarse grid points according to the undirected adjacency graph of the matrix A.
Denote C as the coarse grid points and F := Ω\C the fine points. For any vector in the coarse grid, v c , the interpolation (prolongation) operator to fine grid can be defined as follows: where C i is some small sets of interpolation points C i ⊂ C.
Following [18,19], we define the strong dependent set S i := {j |a ij | ≥ θ max ℓ =i |a iℓ |} and the strong influence set S T i := {j i ∈ S j } with 0 < θ < 1 (usually 0.25). Then the coarsening process goes as the follows: Denote A 1 = A, M 1 = M and the finest grid Ω 1 = Ω. Based on A 1 , the AMG setup procedure builds up the prolongation and restriction operators I k k+1 and I k+1 k = (I k k+1 ) T , respectively for k = 1, 2, · · · , n − 1. The coarse matrices are defined with the Galerkin projection as follows: We use d 1 , · · · , d n to denote the dimension in each level grid Ω 1 , · · · , Ω n .
AMG algorithm for eigenvalue problem
In this section, we introduce an AMG method for solving eigenvalue problems. Similarly to the geometric case in [21], assume we have obtained eigenpair approximations {λ k } q j=1 to our desired eigenpairs. Now we introduce an AMG correction step to improve their accuracy.
Perform m AMG iteration steps with the initial value u (j,ℓ) k to obtain a new eigenfunction approximation u leads to the right hand side term of the linear equation, u (j,ℓ) k denotes the initial guess and m is the number of AMG iteration times.
End Do
Summarize above two steps by defining Based on the above algorithm, we can construct an AMG method for eigenvalue problem which is a combination of the nested technique and the AMG correction step defined by Algorithm 3.1.
Solve this eigenvalue problem to get eigenpair approximations {λ (j) which are approximations to our desired eigenpairs.
Different from the GMG method [21], we do not have the exact prolongation and restriction operators. In the practical computation, we can choose the suitable iteration times p k to meet the accuracy requirement. Compared to other AMG methods, the proposed method here only need to do smoothing iterations for the standard elliptic type of linear equations and the AMG method can act as a blockbox. Furthermore, the required memory for the eigenpair solving is only about qN and we can also compute effectively eigenvalues in a given interval in the middle of the spectrum. Inspired by the analysis for the GMG method [21], it is known that the AMG method can have very good convergence rate if the coarse grids capture the low frequency information of the finest grid well.
Numerical examples
In this section, two numerical examples are presented to illustrate the efficiency of the AMG method proposed in this paper. where Ω = (0, 1) × (0, 1). The stiff and mass matrices A and M in problem (1.1) are obtained by discretizing problem (4.1) with the linear finite element method [7]. In order to show difference of the AMG method from the GMG method, we generate the mesh by Delaunay method which has no the hierarchy structure. In this example, we use two meshes: the coarse one with mesh size h = 0.01 and the finer one with mesh size h = 0.005. Algorithm 3.2 is applied to solve the algebraic eigenvalue problem (1.1) derived from the discretization of (4.1). In this subsection, we choose m = 2 and 2 conjugate gradient smoothing steps for the presmoothing and postsmoothing in each AMG iteration step in Algorithm 3.1. In the k-th level (k = 1, · · · , n 1 − 1) grid of the AMG scheme defined in Algorithm 3.2, we only do p k AMG correction steps defined in Algorithm 3.1. In order to measure the algebraic error of the AMG method, we also solve the eigenvalue problem by the direct method. Figure 1 gives the numerical results (algebraic errors) for the first 13 eigenvalues λ = [2,5,5,8,10,10,13,13,17,17,18,20,20] on the coarse mesh with the mesh size h = 0.01 and Figure 2 gives the corresponding numerical results on the finer mesh with mesh size h = 0.005, respectively. From Figures 1 and 2, we can find that the AMG scheme exhibits the uniform convergence rate which is the same as the AMG iteration for boundary value problems.
More general eigenvalue problem
Here we give numerical results of the AMG method for solving a more general eigenvalue problem on the unit square domain Ω = (0, 1) × (0, 1): Find (λ, u) such that −∆u − 1 |x−Z| u = λu, in Ω, u = 0, on ∂Ω, where Z = (0.5, 0.5). The stiff and mass matrices A and M in problem (1.1) are also obtained by discretizing problem (4.2) with the linear finite element method [7].
In this example, we also use two meshes (the coarse one with mesh size h = 0.01 and the finer one with mesh size h = 0.005) generated by Delaunay method to investigate the convergence behaviors.
Here, we also choose m = 2 and 2 conjugate gradient smoothing step in the presmoothing and postsmoothing procedure. Here we also compare the numerical results with the direct algorithm. Figure 3 gives the numerical results (algebraic error) for the first 13 eigenvalue approximations on the coarse mesh and
Concluding remarks
In this paper, we present a type of AMG method to solve algebraic eigenvalue problems arising from the discretization of partial differential equations. The AMG setup procedure is applied to construct the hierarchy and the intergrid transfer operators of the algebraic problems. Based on the combination of the multilevel correction method and the AMG method for linear equations, an AMG method for eigenvalue problems is proposed. This type of AMG method need almost the optimal computational work and the least memory. Finally, the efficiency of the proposed AMG method is exhibited by two numerical examples which show that the AMG method has uniform convergence rate. The choices of parameters m and p k , presmoothing and postsmoothing operators and AMG coarsening strategy should be considered and tested in future.
|
2019-04-21T04:11:48.951Z
|
2015-03-29T00:00:00.000
|
{
"year": 2015,
"sha1": "3ae702cf35adc22f65426163a1d60d592d982179",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "3ae702cf35adc22f65426163a1d60d592d982179",
"s2fieldsofstudy": [
"Mathematics"
],
"extfieldsofstudy": [
"Mathematics"
]
}
|
213687342
|
pes2o/s2orc
|
v3-fos-license
|
PREVALENCE OF TOOTH DISCOLORATION INDUCED BY ENDODONTIC MATERIALS
1. Consultant EndodontistMember of the scientific council, Saudi Board of Endodontics, Swedish board in Endodontics, Fellowship in Endodontic MicrosurgeryJeddah specialty dental Centre, Ministry of Health. 2. Medicinae Baccalaureus Baccalaureus Chirurgiae, bn Sina National College, KSA (Postal Address: Jjeddah, Al Thaghr neighbourhood., 22338. 3. General dentist, minestry of health, Riyadh, KSA. 4. BSC, Alfarabi Dental College , Jeddah , KSA 5. Medical Interns, Batterjee Medical College, KSA. ...................................................................................................................... Manuscript Info Abstract ......................... ........................................................................ Manuscript History Received: 15 November 2019 Final Accepted: 22 December 2019 Published: January 2020
ISSN: 2320-5407
Int. J. Adv. Res. 8(01), 01-08 2 The result was: The color change was significantly less with biodentine/saline than MTA/saline and MTA/blood (P < 0.05). Regardless of the material type and blood presence, discoloration increased after 3 months (P < 0.05). (Adl A, 2019) Which means that there is a clear influence and change in the color of the teeth of the MTA material this study is very good and compatible with our research since it has also been published recently although it is outside Saudi Arabia.
Another study: Tooth crown discoloration induced by endodontic sealers: a 3-year ex vivo evaluation. The result was: All sealers caused discoloration. At 1 month, ΔE values were significantly higher in groups 1 and 2 than in groups 3 and 4 (p < 0.05). At 1 year, ΔE values were significantly lower in group 3 than in the other groups (p < 0.05). At 3 years, ΔE values were significantly higher in group 1 than in the other groups (p < 0.05). (Ekici MA, 2019) Which means This study was clear for the time it took the teeth be affected and rehearsed after the use of endodontics materials. However, it is not focused on the type of materials, and it is also not established in Saud i Arabia.
The current research aims to determine Prevalence of tooth discoloration induced by endodontic materials.
Research Objectives:- 1. To determine a relationship between age and discoloration of teeth with endodontic materials 2. To determine a relationship between gender and discoloration of teeth with endodontic materials 3. To determine the most endodontic material used at the present time 4. To determine the most resulting material discoloration teeth 5. Determine the duration since the use of endodontic material until the color of the teeth is shown Methodology:-Study design: This is an analytical cross-sectional study.
Study Setting and period:
This is an analytical cross-sectional study conducted in kingdom of saudia Arabia (from the general population IP and OP), from may 2019 till December 2019.
Study population and sampling:
General population (IP and OP) Sampling Method:-Study Participants: Participants were conducted study be carrying the questionnaire during the period of data collection from 11/5/2019 till 12/11/2019. Inclusion criteria; who has already treated his teeth by endodontics materials. Exclusion criteria; other's Sampling size: Sample size was calculated using OpenEpi for sample size calculation for cross sectional studies, hypothesizing the true answers Accordingly, 481 participants were gathered, 35.14% of the respondents answered that the discoloration occurred in their teeth is differentiate from the natural color.
According to the prevalence study, 20.37% of the respondents used Gutta percha, and Half of them took more than 1 month. The results indicated that most respondents say either the color has not changed or that it has changed slightly.
3 Measurements:-Explanatory variables: 1. Sociodemographic characteristics: age, and sex. 2. Disease-related information: Have ever treated by endodontics, name of endodontics material, duration, differentiate of discoloration from the natural color, and description the colour change.
Outcome Measure:
The outcome measure is by counting the ratio of the number of patients have tooth discoloration this will be measured using: By determining the extent of the discoloration in addition to the endodontics that lead to suffering from tooth discoloration.
Prevalence study:
Will be carried to test the questionnaire if easily understood and the response of the participants. Data from the cross-sectional study will be used to calculate the sample size. This was performed using Chi-squared test and Mann-Whitney test. Multivariate analysis to investigate factors independently was performed using binary logistic regression. P value was set at a significance level of < 0.05.
Results:-
In this study, the aim was to determine the prevalence of tooth discoloration induced by endodontic materials. 481 patients' participants, were consecutively recruited from in patients' clinics, during a period from 11/5/2019 to 12/11/2019 Socio-demographic characteristics of the studied group , The results indicate that most of the respondents are between 16-30 years, The results indicate that most respondents are males. According to disease The results show that most of the respondents treated their teeth by filling the teeth (Endodontics), 20.37% of the respondents used Gutta percha , Half of the respondents took more than 1 month, 43.66% of the respondents answered thet the discoloration occurred in their teeth is not diffrentiate from the natural color.
Discussion:-
This study was based on a number of 481 participants, but to ensure the accuracy of the filling and the correctness of the information entered, a question was asked about have ever treated by endodontics, as the study is limited to discoloration induced by endodontic materials only, otherwise the endodontic substances cause a slight change in the color of the teeth in most of participants.
Conclusion:-
The responses of the participants in the questionnaire indicated that the endodontic substances cause a slight change in the color of the teeth, and there are a large number of participants who did not notice any change in the color of their teeth.
Recommendation:-
we recommend setting up health education programs about the dental health and the discoloration by endodontics materials this health problem must be presented broadly and beneficially and, in a way, that everyone understands, as most deal with the tooth discoloration by ignoring and not being important, work should be done on health conferences and medical discussions on that.
8 Ethical considerations:-Administrative approval will be sought from the unit of biomedical ethics research committee Ethical approval will be sought from the ethical committee of the faculty of medicine, king abdulaziz university. An informed consent will be sought from the participants.
|
2020-01-23T14:46:09.358Z
|
2020-01-31T00:00:00.000
|
{
"year": 2020,
"sha1": "69b73ceff50742a9e4fa09fe6158cde1a9e238bf",
"oa_license": "CCBY",
"oa_url": "http://www.journalijar.com/uploads/454_IJAR-30230.pdf",
"oa_status": "HYBRID",
"pdf_src": "Unpaywall",
"pdf_hash": "e101c175116a03037af2fe71c7701d9f772d0a70",
"s2fieldsofstudy": [
"Medicine",
"Materials Science"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
218505044
|
pes2o/s2orc
|
v3-fos-license
|
Grainyhead-Like Genes Family May Act as Novel Biomarkers in Colon Cancer
Objective The Grainyhead-like (GRHL) genes family were reported to participate in the development of a number of diseases. This study was designed to investigate the role of GRHL genes family in colon cancer (CC). Methods In this study, the transcriptional levels of GRHL genes family in patients with CC from GEPIA were explored. Meanwhile, the immunohistochemical data of the GRHL genes family were also obtained in the HPA database. Additionally, we re-identified the mRNA of these genes via real-time PCR. Furthermore, the association between the levels of GRHL genes and stage plot as well as survival condition including overall survival and disease-free survival of patients with CC was analyzed. Finally, by transfecting with specific-siRNA, clone formation assay was performed to observe the role of GRHL genes family in the proliferation of SW480 human colon cancer cells. Results We found that the mRNA and protein levels of GRHL1, GRHL2 and GRHL3 were significantly higher in CC tissues than in normal colon tissues. Additionally, GRHL1, GRHL2 and GRHL3 were significantly associated with the stages of CC. The Kaplan–Meier plotter showed that the low levels of GRHL1, GRHL2 and GRHL3 conferred a better overall survival of patients with CC while the high levels of GRHL1 and GRHL3 were associated with poor disease-free survival. Knockdown of GRHL1, GRHL2 and GRHL3 siHgnificantly inhibited the ability of colony formation of human colon cancer cells. Conclusion Our study demonstrated that GRHL genes are involved in the prognosis and survival in patients with CC, the inhibition of which may suppress the proliferation of colon cancer cells.
Introduction
Colon cancer (CC) is one of the most prevalent malignant tumors, with a high recurrence and the fourth most common cause of cancer-related deaths globally 1,2 . The prognosis of CC is predominantly determined by the clinicopathological features and tumor stages. 3 However, the heterogeneity of the CC makes it is not easy to predict patient prognosis according to these conventional strategies. 4 Additionally, the 5-year overall survival rate of CC remains still very low. 5 Hence, a set of sensitive prognostic markers and potential drug targets should be identified to improve the accuracy of prognosis and individualized treatments.
The Grainyhead-like (GRHL) family of transcription factors possess 3 members, namely GRHL1, GRHL2 and GRHL3, which were firstly discovered in the fruit fly Drosophila melanogaster. 6 The 3 transcription factors were found to adopt a DNA-binding immunoglobulin fold homologous to the key domain of p53 (a well-known tumor suppressor). 7 The expression of the 3 transcription factors is in a tissue and developmentally specific manner, which is primarily expressed in epithelial tissues of organs including olfactory and oral epithelium, urogenital tract, kidney, digestive tract, lung and myocardium. 8 At present, a great many studies unveiled the roles of GRHL genes in different types of cancer: gastric cancer, squamous cell carcinoma of the skin, breast cancer, prostate cancer, colorectal cancer, hepatocellular carcinoma, clear cell renal cell carcinoma, neuroblastoma, and cervical cancer. 6,9 The roles of GRHL genes in various types of cancer are complicated, and even in some cases appear to be contradictory: sometimes they promote cancer development, sometimes they may serve as tumor suppressors. For example, knockdown of GRHL2 in colorectal cancer cells could inhibit cell proliferation by targeting ZEB1. 10 Meanwhile, the protein level of GRHL2 was significantly higher in colorectal cancer tissues, which was positively correlated with tumor size and TNM stage. Furthermore, GRHL2 was also an independent prognostic index for recurrence-free survival and overall survival. 11 The dysregulated expression levels of GRHL genes and their relationship with clinicopathological features and prognosis have been partly reported in human CC. To the best of our knowledge, bioinformatics analysis has yet been applied to investigate the roles of GRHL genes in human CC. Based on the analyses of thousands of gene expression or variation in copy numbers published online, we analyzed and identified the expression and different GRHL transcription factors in patients with CC in detail to investigate their expression patterns, potential functions, and distinct prognostic values of transcription factors in CC.
Ethics Statement
Our study was approved by the Academic Committee of The First People's Hospital of Jiashan and conducted according to the principles expressed in the Declaration of Helsinki. All the datasets mentioned in this study were retrieved from the published literature, indicating that it was confirmed that all written informed consent was obtained. A total of 10 pairs of carcinoma tissues and adjacent tissues from patients diagnosed with CC were collected.
GEPIA (Gene Expression Profiling Interactive Analysis) Dataset
GEPIA (http://gepia.cancer-pku.cn/index.html) serves as a newly generated interactive web server designed by Zefang Tang, Chenwei Li, and Boxi Kang of Zhang Lab, Peking University, aiming to analyze the RNA sequencing expression data of 9736 tumors and 8587 normal samples from the GTEx projects the TCGA database in a standard processing manner. GEPIA provides customizable functions including tumor/normal differential expression analysis, profiling according to cancer types or pathological stages, patient survival analysis, similar gene detection, correlation analysis, and dimensionality reduction analysis. 12 In this study, we mainly employed the boxplot to detect the mRNA expression of GRHL genes in CC and normal colon tissues.
The Stage Plot and Survival Condition of Patients with Different Levels of GRHL Genes
Similarly, we used the GEPIA database to obtain stage plot, overall survival and disease-free survival information of GRHL genes. The log-rank P value and hazard ratio (HR) with 95% confidence intervals were shown on the plot. P < 0.05 was statistically significant.
Human Protein Atlas
The Human Protein Atlas (HPA, https://www.proteinatlas. org/) is a Swedish-based program initiated in 2003 with the aim to map all human proteins in cells, tissues, and organs using the integration of various omics technologies, including antibody-based imaging, mass spectrometrybased proteomics, transcriptomics, and systems biology. 13 By getting immunohistochemical data of patients with or without CC on the basis of HPA, we further verified the protein expression levels of GRHL genes.
Colony Formation Assay
The colony-forming ability of SW480 human CC cells transfected with siRNA or siNC was detected using a colony formation assay. In detail, the transfected cells (siGRHL1, siGRHL2, siGRHL3 and siNC) growing in log phase were trypsinized and seeded into six-well plates with a density of 2000 cells per well. The cells were kept in an incubator at 37°C for 7 days. Seven days later, the colonies were washed with phosphate-buffered saline (PBS), fixed with formalin (10%; Beyotime Institute of Biotechnology, Haimen, China) and stained with methyl violet. Finally, the methyl violet dye was washed off with PBS. The number of colonies was counted using a microscope (Olympus IX53; Olympus Corporation, Tokyo, Japan). Colony-inhibition rate=[1−(number of colonies in experimental groups/control group)]×100%; and colony-forming efficiency=(1−colonyinhibition rate) were calculated.
Real-Time PCR
Total RNA was extracted from cell cultures using Trizol (Invitrogen, Grand Island, NY, USA) based on the manufacturer's protocol. Subsequently, the cDNA was amplified by a reverse transcriptional kit (Promega, Madison, WI, USA). The real-time PCR was performed using cDNA as a template and Universal PCR Master Mix (Applied Biosystems, Carlsbad, CA, USA) by an Applied Biosystems 7900HT sequence detection system (Applied Biosystems, Foster City, CA, USA). The relative amount of mRNA was calculated and normalized using GAPDH as internal reference. The primers used in this study are presented in Table 1.
Western Blot
Lysates from CC cells were immunoblotted according to protocol. Briefly, protein extracts were resolved on SDS-PAGE and transferred to PVDF membrane. After blocking with 5% non-fat dry milk, the membrane was incubated overnight at 4°C with primary antibodies: anti-GRHL1, anti-GRHL2, anti-GRHL3 or anti-GAPDH. After incubation with primary antibodies, membranes were incubated with secondary antibodies for 50 min. To confirm equal protein loading, blots with antibody against GAPDH were used. Proteins were visualized with Amersham ECL™ Western blotting system.
Statistical Analysis
The obtained data were presented as the mean ± SD (standard deviation) and assessed by the two-tailed Student's t-test. A difference of P<0.05 was considered statistically significant.
Transcriptional Levels of GRHL Genes in Patients with CC
By analyzing 349 normal colon tissues and 275 CC tissues based on GEPIA online website, we found that the mRNA levels of GRHL1, GRHL2 and GRHL3 were significantly higher in carcinoma tissues than in normal colon tissues (P<0.05), indicating that the GRHL family of transcription factors may act as potential markers in the diagnosis of CC ( Figure 1A-C).
Immunohistochemical Data of GRHL Genes in Patients with CC
To further obtain the protein expression levels of GRHL genes, we obtained the immunohistochemical data of the 3 genes from HPA. Consistent with the above results based on GEPIA, the immunohistochemical data from HPA confirmed the protein levels of GRHL genes in CC (Figure 2A-C).
Correlation Between GRHL Genes Expression and Tumor Stage in Patients with CC
Subsequently, we analyzed the association between GRHL genes expression levels and tumor stage in patients with
Correlation Between GRHL Genes Expression and Survival Condition in Patients with CC
Meanwhile, we further analyzed the potential association between the expression levels of GRHL genes expression levels and the survival condition of patients with CC. The Kaplan-Meier showed that the levels of 3 transcription factors displayed significant correlation with the overall survival and disease-free survival of patients with CC. In detail, the low level of 3 genes may contribute to better overall survival of CC ( Figure 4A-C) while the high level of GRHL1 and GRHL3 may contribute to better diseasefree survival ( Figure 4D-F) (P<0.05).
Re-Identification of the Expression of GRHL Genes
To enhance the reliability of database, we detected the mRNA expression levels of GRHL genes in carcinoma tissues and adjacent tissues from patients with CC using real-time PCR, the results ( Figure 5) of which showed further validated the previous hypothesis from bioinformatics analysis.
Knockdown of GRHL Genes Reduces Proliferation of Human CC Cells
Finally, a colony formation assay and a CCK8 assay were performed to explore the effects of GRHL1, GRHL2 and GRHL3 on proliferation ability. Western blot showed that the 3 genes were successfully knocked down by siRNA Figure 6A. As shown in Figure 6B and C, the proliferation ability and the relative colony number were significantly declined after theses 3 genes were inhibited in CC cells (P<0.05). Also, we performed the same experiments to detect the proliferation ability in another cell line LoVo, the results of which further proved the anti-proliferation ability of GRHL genes (Supplementary materials).
Discussion
The dysregulation of GRHL transcription factors has been reported in various types of cancers. Although the role of GRHL transcription factors in the carcinogenesis and prognosis of some certain cancers has been partially unveiled, further bioinformatics analysis of CC on GRHL has yet to be performed till now. To our knowledge, our study is the first time to investigate the mRNA and protein expression as well as prognostic values of different members in GRHL genes family in CC. We sincerely hope that our study will contribute to available knowledge, improve treatment designs, and enhance the accuracy of prognosis for patients with CC. The GRHL1 transcription factor is tissue-specific and is predominantly expressed in the kidney. The GRHL1 gene is located at the chromosomal position 2p25 in humans. 14 In a mouse model, the loss of GRHL1 significantly DovePress affected the heart rate but had no effects on blood pressure. 15 GRHL1 was also reported to serve as a tumor suppressor by affecting mild chronic skin inflammation and aberrant terminal differentiation of keratinocytes in the squamous cell carcinoma of the skin. 16 Additionally, high levels of GRHL1 expression were also associated with a favorable prognosis for patients with neuroblastoma. In vitro experiments further revealed that GRHL1 could inhibit the development of neuroblastoma by regulating MYCN and HDAC3. 17 In our study, we found that the mRNA and protein expression of GRHL1 was significantly higher in patients with CC, the higher level of which may be associated with tumor stage and survival condition.
GRHL2 plays vital roles in embryonic neural tube closure, epidermal integrity, and wound healing processes. Mountainous evidence has disclosed that GRHL2 may be a novel proto-oncogene which could regulate epithelial plasticity by suppressing endothelial-mesenchymal transition in several types of tumor. 18 For example, GRHL2 expression level was positively associated with CD133 and E-cadherin in primary pancreatic ductal adenocarcinoma and was highly expressed in liver metastatic pancreatic ductal adenocarcinoma tissues. 19 Meanwhile, Yang et al 20 have identified 6 genes involving FN1, CDH2, CTNNB1, CITED2, as well as CTNNA3 as GRHL2-related genes together with GRHL2 in breast cancer metastasis, the expression levels of which were associated with clinical features. On the contrary, other studies also revealed that GRHL2 may serve as a tumor suppressor in cervical cancer, sarcoma, gastric cancer, and clear cell renal cell carcinoma. 21 Exogenous GRHL2 transduced into gastric cancer cells obviously suppressed the proliferation and enhanced apoptosis. At the same time, over-expression of GRHL2 significantly reduced the protein expression levels of c-Myc and Bcl-2. 22 Our study demonstrated that GRHL2 may serve as a potential proto-oncogene in CC and is involved in the proliferation of CC cells, at least proliferation ability. GRHL3 is very important for epidermal development and homeostasis in a wide range of species.
GRHL3 deficiency could induce epidermal keratinocyte hyperproliferation during embryogenesis by targeting PTEN directly in squamous cell carcinoma. 23 In colorectal cancer cells and tissues, the GRHL3 expression at both mRNA and protein levels was significantly increased. Knockdown of GRHL3 with siRNA obviously repressed colorectal cancer cell viability, proliferation, and migration, apart from promoting cell cycle arrest at G0/G1 phase and inducing cell apoptosis. 24 In accordance with this study, our study also proved the expression level of GRHL3 was correlated with prognosis and tumor stage of patients with CC. More importantly, GRHL3 also affected the proliferation of CC cells, which was evidenced by the altered colony formation ability of CC cells.
Conclusion
Taken together, in this study, we systemically analyzed the expression and prognostic value of GRHL genes family in CC and provided a thorough understanding of the heterogeneity and complexity of the molecular biological properties of CC. Our studies indicated that the increased expression of GRHL1, GRHL2 and GRHL3 in CC tissues might play a vital role in CC oncogenesis; therefore, they may be promising diagnostic biomarkers for CC. Additionally, the levels of GRHL1, GRHL2 and GRHL3 were significantly associated with survival and tumor stages of the patients with CC, suggesting that Pitx1 could be potential therapeutic targets for CC.
Disclosure
The authors report no conflicts of interest in this work.
Publish your work in this journal
OncoTargets and Therapy is an international, peer-reviewed, open access journal focusing on the pathological basis of all cancers, potential targets for therapy and treatment protocols employed to improve the management of cancer patients. The journal also focuses on the impact of management programs and new therapeutic agents and protocols on patient perspectives such as quality of life, adherence and satisfaction. The manuscript management system is completely online and includes a very quick and fair peer-review system, which is all easy to use. Visit http://www.dovepress.com/ testimonials.php to read real quotes from published authors.
|
2020-04-23T09:07:03.703Z
|
2020-04-01T00:00:00.000
|
{
"year": 2020,
"sha1": "20e5ff665bdbf0b3d0f715e2b5e214a93adc8fc6",
"oa_license": "CCBYNC",
"oa_url": "https://doi.org/10.2147/ott.s242763",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "b5d1eb65b96347468e3807f6b0f8ddcfb14add9d",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Biology",
"Medicine"
]
}
|
3290529
|
pes2o/s2orc
|
v3-fos-license
|
Integrated omics data of two annual ryegrass (Lolium multiflorum L.) genotypes reveals core metabolic processes under drought stress
Background Annual ryegrass (Lolium multiflorum L.) is a commercially important, widely distributed forage crop that is used in the production of hay and silage worldwide. Drought has been a severe environmental constraint in its production. Nevertheless, only a handful of studies have examined the impact of short-term drought stress on annual ryegrass. The aim of this study was to explore how stress-induced core metabolic processes enhance drought tolerance, or adaptation to drought, in annual ryegrass. Results We profiled the transcriptomes, proteomes, and metabolomes of two annual ryegrass genotypes: the drought-resistant genotype “Abundant 10” and drought-susceptible genotype “Adrenalin 11.” We identified differentially expressed metabolites and their corresponding proteins and transcripts that are involved in 23 core metabolic processes, in response to short-term drought stress. Protein–gene–metabolite correlation networks were built to reveal the relationships between the expression of transcripts, proteins, and metabolites in drought-resistant annual ryegrass. Furthermore, integrated metabolic pathways were used to observe changes in enzymes corresponding with levels of amino acids, lipids, carbohydrate conjugates, nucleosides, alkaloids and their derivatives, and pyridines and their derivatives. The resulting omics data underscored the significance of 23 core metabolic processes on the enhancement of drought tolerance or adaptation to drought in annual ryegrass. Conclusions The regulatory networks were inferred using MCoA and correlation analysis to reveal the relationships among the expression of transcripts, proteins, and metabolites that highlight the corresponding elements of these core metabolic pathways. Our results provide valuable insight into the molecular mechanisms of drought resistance, and represent a promising strategy toward the improvement of drought tolerance in annual ryegrass. Electronic supplementary material The online version of this article (10.1186/s12870-018-1239-z) contains supplementary material, which is available to authorized users.
Background
Drought is a severe environmental constraint to seed germination, plant growth, and productivity [1,2]. Plants employ physiological and molecular mechanisms of drought tolerance to cope with water shortages. Indeed, the drought-response mechanisms developed by plants at the cellular level are essential, as they allow tolerance in plants that facilitates cellular homeostasis [3]. Mounting evidence has confirmed that plant species that are more tolerant to drought stress maintain higher levels of unsaturated fatty acids, such as hexadecenoic acid, palmitic acid, pimelic acid, stearic acid, and linolenic acid, all of which intervene in cases of compromised membrane fluidity and cellular functions [4][5][6]. In addition, the synthesis of amino acids, such as valine, leucine, phenylalanine, proline, and histidine, can contribute to turgor maintenance through osmotic adjustment when plants, especially resistant cultivars, are subjected to gradually increasing drought stress [7].
Plants exhibiting high drought tolerance are promising candidates for studies in drought-related genes, proteins, and metabolites [8]. Transcriptomic, proteomic, and metabolomic profiling are highly useful approaches to dissecting the complex networks of regulatory mechanisms at multiple levels in plants [9,10]. In addition, multiple co-inertia analysis (MCoA) is a critical method for integrating data from several multi-omics datasets [11]. One of the advantages of MCoA is that it can be used to analyze a subset of variables (e.g., transcripts, proteins, and metabolites) that are present in two or more datasets. Several recent studies using comparative physiological, metabolomic, and transcriptomic analyses have provided deep insights into the mechanisms associated with improving abiotic stress resistance through the application of exogenous melatonin [12]. Similarly, combined analyses of the transcripts, proteome, and metabolites have been used to understand the manner in which core metabolic processes enhance carotenoid synthesis in transgenic maize [13]. "Omics" studies have also been performed, using large drought datasets for forage grasses with differing levels of sensitivity, both under water stress and non-stress conditions [14,15].
Annual ryegrass (Lolium multiflorum L.), a species closely related to perennial ryegrass (Lolium perenne), is a commercially important forage crop that is widely cultivated for the production of hay and silage worldwide [16], including southern China. In recent years, severe short-term (daily to monthly) droughts have occurred frequently over southern China, causing a severe adverse effect on grass productivity [17,18]. As such, the economic importance of annual ryegrass has encouraged many researchers to study the physiological and molecular bases of drought tolerance. Nevertheless, only a handful of studies have examined the impact of shortterm drought stress on annual ryegrass, using an omics approach [19].
In our recently published study, we developed two annual ryegrass varieties, Abundant 10 and Adrenalin 11, with differing degrees of drought tolerance [20]. In order to explore the molecular mechanism of drought tolerance in these two annual ryegrass genotypes, we identified differentially expressed metabolites and their corresponding proteins and transcripts that are involved in 23 core metabolic processes under short-term drought treatment. The associated regulatory networks were inferred using MCoA and correlation analysis, to reveal the relationships among the expression of transcripts, proteins, and metabolites that highlight the corresponding elements of these core metabolic pathways. This study provides valuable insight into the molecular mechanisms of drought resistance and represents a promising approach toward the improvement of drought tolerance in annual ryegrass.
Plant samples and drought treatments
Two L. multiflorum genotypes, drought-resistant "Abundant 10" and drought-susceptible "Adrenalin 11," were used in this study. Seedlings were transplanted 7 days after germination into Hoagland's nutrient solution. All seedlings were grown in temperature-controlled growth chambers with 16-h photoperiods (25°/18°C day/night temperature) and relative humidity of 60%. At day 20, after germination, the seedlings were divided into four groups to be used as the control (0 h) and various drought-treated samples. The drought-treated samples were placed on plastic trays and naturally air-dried for 1, 2, and 24 h, respectively (Additional file 1 Figure S1). Ten individual plants from each treatment group were considered a biological replicate. After treatment, three biological replicates were used for total RNA and protein extraction, and six biological replicates were then subjected to gas chromatography-mass spectrometry-(GC-MS)-based metabolite identification.
Relative water content (RWC) was calculated using the formula described by Barrs and Kozlowski [21]. The chlorophyll content was calculated according to the method described by Lakra et al. [22]. Relative electrical conductivity (REC) was calculated as the ratio of the initial electrical conductivity (EC) to the final EC [23]. The changes and activity of malondialdehyde (MDA) concentration, superoxide dismutase (SOD), catalase (CAT), and ascorbic acid peroxidase (APX) were each assayed with MDA, SOD, CAT, and APX assay kits, respectively (Comin Biotechnology Co. Ltd., Suzhou, China). A oneway ANOVA was performed using the SPSS Statistics 20.0 software (IBM Corp, Armonk, NY, USA), and means were compared using the least significant difference (LSD) test to determine significant differences (P < 0.05).
Transcriptome sequencing and analysis
Total RNA was isolated from samples, using the TRIzol reagent (Agilent Technologies, Santa Clara, USA). A mass of 5 μg per sample was collected for cDNA library construction, using the NEB Next® Ultra™ Directional RNA Library Prep Kit for Illumina® (NEB, USA). The cDNA Library was validated by two different methods to determine the average molecular length: 1) the Agilent 2100 Bioanalyzer (Agilent DNA 1000 Reagents; Agilent Technologies) and 2) real-time quantitative PCR (QPCR; TaqMan Probe; Thermo Fisher Scientific, Waltham, MA, USA). The qualified libraries (average length of fragment was between 250 and 350 bp) were amplified on the cBot System to generate the cluster on the flow cell using the TruSeq PE Cluster Kit V3-cBot-HS (Illumina, San Diego, CA, USA). The amplified flow cell was subjected to paired-end sequencing using the HiSeq 2000 sequencing system (Illumina, USA). As a reference genome had not been previously created, the clean reads were assembled as the reference genome via the Trinity software (https://github.com/trinityrnaseq/trinityrnaseq/ wiki) [24]. Read counts per gene were expressed as the expected number of fragments per kilobase of transcript sequence per million base pairs sequenced (FPKM). The P-values were adjusted using the Benjamini and Hochberg's approach to control the false discovery rate (FDR). Gene Ontology (GO) enrichment analysis of differentially expressed genes (DEGs) was implemented with the GOseq R software package, in which the gene length bias was corrected. The GO terms with DEGs (FDR ≤ 0.001 and a fold change ≥2) were used for functional enrichment analysis. Genes with an adjusted P-value below 0.05, as determined by the DESeq software, were assigned as differentially expressed, and employed in the GO and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses. The KEGG enrichment analysis of DEGs in the KEGG database (http://www.genome.jp/ kegg/) and the KOBAS (KEGG Orthology Based Annotation System) software [25,26] were used to test the statistically significant enrichment of DEGs in KEGG pathways.
Protein identification and data analysis
Proteins were extracted from samples at four time points throughout the drought stress treatments: 0, 1, 2, and 24 h. The extractions were performed with Lysis Buffer 3 containing 1 mM phenylmethylsulfonyl fluoride (PMSF) and 2 mM ethylenediaminetetraacetic acid (EDTA), and suspended at 200 W for 15 min. Proteins were isolated by centrifugation at 30000 g for 15 min at 4°C, after which 5× volume of chilled acetone and 10% (v/v) trichloroacetic acid were added at − 20°C.
After two rounds of centrifugation, the supernatant was carefully discarded and the precipitate was washed three times with cold acetone. The protein pellet was air-dried by lyophilization and dissolved in Lysis Buffer (7 M urea, 2 M thiourea, 4% NP40, 20 mM Tris-HCl, pH 8.0-8.5). The protein pellet was then suspended for 15 min and centrifuged at 4°C at 25000 g for 15 min, and the supernatant was collected.
To reduce the number of disulfide bonds in the proteins of the supernatant, 10 mM dithiothreitol (DTT) was added, and the mixture was left for 1 h at 56°C. Subsequently, 55 mM iodoacetamide (IAM) was added to block the cysteines, after which samples were kept in a dark room for 1 h. The supernatant of the proteins was kept at − 80°C. Protein samples of 100 μg each were added to 2.5 μg Trypsin (Promega, Madison, WI, USA) with a weight ratio of 40 protein: 1 trypsin, and kept at 37°C for 4 h. The peptides were vacuum-dried using Strata-X, and reconstituted in 0.5 M triethylammonium bicarbonate (TEAB) based on the manufacturer's protocol for the 8-plex iTRAQ reagent (Applied Biosystems, USA). This solution contained one unit of thawed and reconstituted iTRAQ reagent in 24 μL isopropanol. The peptides were labeled with isobaric tags, pooled, and then vacuum-dried.
Each fraction was re-suspended in buffer A (2% acetonitrile [ACN], 0.1% formic acid [FA]) and centrifuged at 20000 g for 10 min. The final concentration of peptides was on average approximately 0.5 μg/μL. The supernatant (10 μL) was evaluated using a LC-20 AD Nano-HPLC pump (Shimadzu, Kyoto, Japan) with an autosampler, and peptides were eluted onto an analytical C18 column (inner diameter 75 μm and column length 15 cm). The samples were loaded over 4 min, and the solvent gradient was run from 5% buffer B (96% ACN, 0.1% FA) for 0-8 min, with a linear gradient to 35% buffer B for 8-43 min, maintained at 60% buffer B for 43-48 min, and returned to 5% buffer B for 55-65 min.
Data acquisition was performed using a Triple TOF 5600 System (SCIEX, Framingham, MA, USA) fitted with a Nanospray III source (SCIEX) and a pulled quartz tip as the emitter (New Objectives, Woburn, MA, USA). Data were acquired using an ion spray with a 2.5 kV voltage. The curtain and nebulizer gases were set at 30 psi and 15 psi, respectively, and the interface heater temperature was 150°C. The mass spectrometer was operated with a resolving power (RP) of 30,000 FWHM (full width at half maxima) for time-of-flight mass spectrometry (TOF/MS) scans. Survey scans were obtained in 250 ms and up to 30 product ion scans were acquired (cut-off threshold was 120 counts/s). Raw data files were transformed into Mascot generic format (MGF) files using the Proteome Discoverer software.
Raw data files acquired from the Orbitrap analyzer were converted into MGF files using the Proteome Discoverer 1.2 software (Thermo Fisher). The Mascot 2.3.02 search engine (Matrix Science, London, UK) was used to identify and quantify proteins. An automatic decoy database search was performed in Mascot, by choosing the decoy checkbox in which a random sequence database was generated and tested for raw spectra. The real database was also tested for raw spectra. Only peptides with a 95% confidence interval from the Mascot probability analysis were counted as identified. The identification of each protein involved at least one unique peptide. The quantitative protein ratios were weighted and normalized using the median ratio determined by Mascot. We considered only data with values of P < 0.05 and fold changes > 1.2 as significant. Functional annotations of identified proteins were performed using the Blast2GO program against the non-redundant (NR) protein database. The KEGG and clusters of orthologous groups (COG) databases were applied to classify the identified proteins.
Metabolome analysis
Six biological replicates of each sample were analyzed for non-treated and drought-treated seedlings (drought for 24 h) in the two L. multiflorum genotypes. Approximately 50 mg of powdered samples were extracted in 1 mL of an 80% methanol, 20% distilled water solution for 30 min at 4°C. During the extraction process, the samples were centrifuged for 10 min at 12,000 g. The supernatant was removed and the pellet was subjected to further extraction in 60% methanol, and then in water at 4°C, as described above. Metabolite profiling was performed on an Agilent 1290 Infinity Liquid Chromatography System (Agilent Technologies) equipped with a 2.1 mm × 100 mm C18 reverse-phase column, with a 1.8-μm particle size (Waters Corp., Milford, MA, USA). The column was maintained at 40°C and the injected sample volume was 4 μL. Mass spectrometry (MS) experiments were performed on an Agilent 6530 Accurate-Mass Q-TOF/MS (Agilent Technologies) equipped with an electrospray ionization source. The binary gradient had a flow rate of 0.35 mL/min, and a gradient elution from 5% to 95% acetonitrile separated all compounds. The total run time was 34 min.
The electrospray source parameters were optimized as follows: positive mode-sampling cone 35 kV, capillary voltage 4 kV, extraction cone 3 V, source temperature 100°C, desolvation temperature 350°C, cone gas flow 50 L/h, and desolvation gas flow 600 L/h. Negative mode-sampling cone 50 kV, capillary voltage 3.5 kV, extraction cone 4 V, source temperature 100°C, desolvation temperature 300°C, cone gas flow 50 L/h, and desolvation gas flow 700 L/h. For accurate mass acquisition, a lock mass of leucine enkephalin (Lock mass) at a concentration of 0. After filtration, a peak table was created that included information on the retention time (RT), mass, and ion intensity of all identified components. Raw data were first preprocessed with a Mass Profiler (Agilent Technologies) and input to the Simca-P 13.0 program for multivariate analysis (i.e., principal component analyses [PCA] and partial least squares discriminant analysis [PLS-DA]). Data were expressed as the mean ± SD. Variable importance in the projection (VIP) values were obtained for variables in the orthogonal projections to latent structures discriminant analysis (OPLS-DA) model in order to select the differentially expressed metabolites. An independent t-test (P < 0.05) was used to determine whether the individual metabolites (candidate biomarkers) obtained from PLS-DA modeling, exhibited statistically significant differences between groups at the univariate analysis level. Only the metabolites with a VIP > 1 and a P-value < 0.05 were defined as differential metabolites. The MS analysis system was used to identify the characteristic metabolites corresponding to the featured peak in the Metlin database (http://metlin.scripps.edu) [27]. The KEGG database was also used to link differential metabolite levels to functional metabolic pathways in the drought-treated groups, compared to those in the control groups.
Integrative analysis using multiple co-inertia analysis and gene set enrichment analysis The MCoA was performed using the Bioconductor software package omicade4 [28]. For consistency, we selected 1-, 2-, and 24-h treatments as the drought treatment groups, and 0 h as the control group. The 0-h and 24-h metabolomic profiles were used as the control and drought treatment profiles, respectively. The mean FPKM in the transcriptomic data (three replicates), ratios in the proteomics data (two replicates), and peak intensity in the metabolic data (six replicates) were used in the MCoA. All values were log-transformed. The MCoA can deal with multiple datasets; therefore, we did not merge the positive and negative channels in the metabolite profiling, which resulted in four datasets for each of genotypes. The Gene Set Enrichment Analysis (GSEA) software (version 2.2.2) was used for gene set enrichment analysis [29]. The gene set information was collected from GO and KEGG pathway databases, as in the differential expression analysis. Gene sets containing more than 1000 or fewer than five candidates were excluded. A permutation text with 1000 permutations was used to evaluate the significance of the enrichment results. The default settings for other parameters were used.
Quantitative real-time-PCR analysis and western blot
Changes in the expression of 26 genes were verified by real-time PCR analysis, and detailed information on these genes is listed in Additional file 2 Table S1. Realtime RT-PCR was performed on the ABI7500 Real-Time PCR System (Applied Biosystems) using the SYBR® Premix EX-Taq™ II kit. Reverse transcription was performed using the Bio-Rad iScript cDNA Synthesis Kit. The PCR amplifications were conducted in a volume of 20 μL, containing 10 μL PCR-mix, 2.5 μL of genomic DNA, 5.5 μL ddH2O, and 1 μL of each primer. The thermocycler was set to touchdown mode according to the following program: 95°C for 30 s, 1 cycle; 95°C for 5 s, 40 cycles; and 60°C for 34 s, 1 cycle. A melting curve was generated by heating the sample to 95°C. Real-time PCR data was then analyzed by the comparative CT method.
An equal amount (10 μg) of each protein sample was loaded for sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (1.5 mm gel thickness). Migration of proteins in the gel was conducted at 150 V until the blue band from the sample buffer ran out of the gel. Protein-Marker IV was also loaded to determine the molecular weight of the proteins. Proteins were then transferred onto a polyvinylidene fluoride (PVDF) membrane (Millipore, USA). The following antibodies were used in the western blot analysis: actin (ACT), chitinase (CHN), glutamine synthetase (GS1), methionine synthase (MTR), UDP-glucose pyrophosphorylase (UGPase), glutamate dehydrogenase 1 (GDH1), serine hydroxyl methyltransferase (SHMT), and glutamine synthetase (GlnA) (Agrisera, Sweden). The PVDF membrane was probed with primary antibodies and developed using enhanced chemiluminescence detection (PerkinElmer, Waltham MA, USA). The blots were detected using the BeyoECL Plus (P0018) system. The images were obtained using the ChemiDoc TM MP imaging system, and the quantifications were conducted with the Image Lab TM V5.1 software.
Results
Drought stress induced growth and physiological changes in L. multiflorum To determine the morphological and physiological impact of drought stress on annual ryegrass, both drought-resistant and drought-susceptible genotypes were subjected to sustained drought for 5 weeks ( Fig. 1a-b). Differences between the genotypes were observed in seedling height (SH), chlorophyll (a + b) content, REC, RWC, root-to-shoot ratio (RSR), malondialdehyde (MDA) content, catalase (CAT), superoxide dismutase (SOD), and ascorbic acid peroxidase (APX) activities ( Fig. 1c-k). The impact of drought stress on RWC was apparently reduced among treated seedlings after the 14-day drought treatment, and there was a significant difference in response between the two annual ryegrass genotypes (Fig. 1c). The REC and chlorophyll (a + b) content of the susceptible plants exhibited a dramatic reduction, in comparison to those of the tolerant plants ( Fig. 1d-e). Prior to the application of drought stress, MDA content was invariant among the control samples of the two annual ryegrass genotypes. However, susceptible genotype revealed significantly higher MDA contents, relative to those of the tolerant genotype after 14 days of drought stress (Fig. 1f ). Higher levels of CAT, SOD, and APX activity were observed among the tolerant genotype exposed to long-term drought ( Fig. 1g-i), whereas no significant changes were noted in SH when subjected to drought stress (Fig. 1j). However, a highly significant increase in RSR was observed in the tolerant genotype in comparison to the susceptible genotype after annual ryegrass seedlings were treated with drought stress for 21 days (Fig. 1k).
Metabolite profiling of two L. multiflorum genotypes revealed changes in metabolites under drought stress The differentially expressed metabolites of the two L. multiflorum genotypes were identified following exposure to 24 h of drought stress, using LC-MS and the OPLS-DA model. Different types of compounds were detected in positive and negative modes including: lipids; amino acids; organic acids; carbohydrates and carbohydrate conjugates; nucleosides, nucleotides, and their analogs; indoles and their derivatives; alkaloids and their derivatives; amine compounds, pyridines and their derivatives, among others (Additional file 3: Figure S2). additional(DEMs) exhibited contrasting expression levels between the drought-tolerant and drought-sensitive genotypes under drought stress, particularly lipids, amino acids, organic acids, amine compounds, and pyridines and their derivatives. Significant correlations were noted among the levels of these compounds, according to Spearman two-tailed correlation analyses (P ≤ 0.05 and r 2 > 0.65 or P ≤ 0.05 and r 2 < − 0.65; Additional file 4: Figure S3). Lipids and amino acids more favorably ionized in positive ionization mode (Additional file 4: Figure S3A), whereas organic acids generated higher signal intensities in positive ionization mode (Additional file 4: Figure S3). We found that lipid levels exhibited the greatest number of significant correlations with other metabolites, followed by amino acids, and organic acids.
Comparative proteomics and transcriptomic profiling reveals differences in the expression of proteins and genes regulating core metabolism To investigate the differentially abundant proteins induced by drought treatment, we performed comparative proteomics analysis on the two annual ryegrass genotypes at four different time points 0, 1, 2, and 24 h. A total of 26,189 unique peptides matching 8224 proteins were identified by Mascot with a high level of confidence (all with a 1% FDR), of which 1395 were differentially abundant between the drought-susceptible and droughtresistant genotypes (Fig. 2a). To understand further the function and features of these proteins, differentially expressed proteins (DEPs) were identified using Blast2GO and the KEGG database. Most of these proteins were located in organelles, cells, membranes, or nuclei (Fig. 2b). The KEGG pathway enrichment analysis revealed that the abundance of DEPs involved in metabolic pathways that are related to the metabolites identified by LC-MS were dramatically affected by drought (Fig. 2c). This identified the drought-induced proteins associated with stress responses that might contribute to enhanced tolerance or adaptation to drought in annual ryegrass.
To gain insight into gene expression in annual ryegrass across all four time points of the drought treatment, we performed transcriptomic analysis to identify droughtmediated genes. After filtering out contaminated and low-quality Illumina HiSeq sequencing reads, approximately 137,708 unigenes were assembled. These annotated unigenes were used to search various functional databases, and unigenes that encoded transcription factors (TFs) were classified into different TF families (Additional file 5: Figure S4). Using a fold change > 2 and FDR < 0.05 as thresholds, transcripts from both resistant and susceptible genotypes were identified as DEGs. Many of the upregulated and downregulated DEGs that were highly expressed following drought treatment in the two L. multiflorum genotypes at different time points (Additional file 6: Figure S5), were enriched in various core metabolic pathways. Interestingly, a large number of DEGs that were involved in the pathway terms "sphingolipid metabolism," "pantothenate and CoA biosynthesis," "histidine metabolism," "glycerolipid metabolism," and "galactose metabolism," were observed in both L. multiflorum genotypes before drought treatment (Fig. 3a). In comparison to the drought-susceptible genotype, the expression of many DEGs were significantly increased in tolerant genotype after 1 h of drought, including those associated with "nitrogen metabolism," "cysteine and methionine metabolism," "citric acid cycle (TCA cycle)," "biotin metabolism," "arginine and proline metabolism," and "2-oxocarboxylic acid metabolism" (Fig. 3b). Similarly, more than 100 genes involved in "nitrogen metabolism," "cysteine and methionine metabolism," "biotin metabolism," and "biosynthesis of amino acids" were significantly enriched in tolerant genotype subjected to drought for 2 h (Fig. 3c). The number of DEGs involved in "tryptophan metabolism," "purine metabolism," "nitrogen metabolism," and "cutin, suberine, and wax biosynthesis" was increased in tolerant genotype, compared to sensitive genotype after 24 h of drought stress (Fig. 3d). These results indicate that the number of genes involved in these core metabolic processes showed significant changes.
To further confirm the RNA-Seq data, qRT-PCR was performed. The expression trends of 26 amino acid and lipid metabolism-related genes were consistent with RNA-Seq data in the two L. multiflorum genotypes Fig. 3 Differentially expressed genes involved in core metabolic pathways in two L.multiflorum genotypes at four time points (0, 1, 2, and 24 h) during drought treatment (a-d) (Additional file 7: Figure S6), and fold changes in gene expression were significantly correlated in both the resistant and susceptible genotypes (r 2 > 0.9).
Multiple co-inertia analysis to evaluate the integration of omics datasets
To demonstrate the relationships among transcripts, proteins, and metabolites, we collapsed our samples into control and drought-treated seedlings using the means of each omics profile. A strong correlation between the four datasets was observed (Fig. 4a). The top two MCoA components, responsible for more than 80% of the total variation (Fig. 4b), clearly separated the drought stress sensitive and resistant genotypes (Fig. 4a) The first and second components accounted for roughly comparable proportions of the variation, implying that both genetic background and drought treatment underlie considerable variation in the data.
As in the PCA, in which the contributions of variables (e.g., transcript, protein, and metabolite levels) to a component could have been evaluated by their loading vectors, variables with either positive or negative associations (i.e., with high absolute values) had significant influence on the co-inertia components. In contrast to the PCA, the variables comprising MCoA loading vectors were derived from all four datasets. In consideration of the biological meaning of the MCoA components, we subjected the loading vectors of the two components to GSEA. Pre-ranked analysis was performed, in which variables were ranked according to their loadings. The loading vectors for component 1 were significantly and positively associated with amino acid metabolism and lipid metabolism (P < 0.001; Fig. 4c-d). Taken together, the omics data sets subjected to MCoA indicated that amino acid and lipid metabolism play important roles in annual ryegrass drought response. Fig. 4 Integrative analysis of multiple omics data using MCoA and GSEA. The first two components defined by MCoA (a). The proportion of variation explained by each component (b). The GSEA of the first loading vector for the lipid metabolism (c) and amino acid metabolism gene sets (d). Each dot represents a plant genotype; the same genotype from different datasets are linked by segments. The length of segments connecting annual ryegrass is correlated with the similarity among datasets Targeted analysis revealed a significant association among transcripts, proteins, and metabolites in core metabolic processes Correlation analysis was performed among differentially expressed metabolites, proteins, and genes (Fig. 5). We first matched all transcripts with their detected proteins, and an intermediate Pearson correlation coefficient of r = 0.724 was observed. The DEGs with corresponding DEPs were classified into different groups exhibiting upregulation or downregulation. A total of 47 DEGs exhibited trends that matched those of their corresponding proteins in the two annual ryegrass genotypes subjected to drought stress. These down-regulated DEG and DEP pairs were involved in "phenylalanine, tyrosine, and tryptophan biosynthesis," "tropane, piperidine, and pyridine alkaloid biosynthesis," "phenylalanine metabolism," "phenylpropanoid biosynthesis," "biosynthesis of amino acids," "2-oxocarboxylic acid metabolism," and "glycerolipid metabolism." Similarly, 51 DEGs had contrasting expression changes with their corresponding proteins under drought stress, including genes involved in "valine, leucine, and isoleucine degradation" and "biosynthesis of amino acids." In addition, a large number of transcripts or proteins that had no significant regulation were also observed in core metabolic pathways.
Protein-gene-metabolite correlation networks were generated to examine responses of the 23 metabolic pathways in drought-resistant annual ryegrass (Table 2; Fig. 6). Integrated metabolic pathways were used to observe changes in enzymes, corresponding to levels of amino acids, lipids, carbohydrate conjugates, nucleosides, alkaloids and their derivatives, and pyridines and their derivatives. Several amino acids, such as L-valine, L-leucine, phenylalanine, L-histidine, and proline, were found in lower concentrations in the droughtsusceptible genotype than in the tolerant genotype under drought conditions. This may be a consequence of the downregulated key enzymes, including branched-chain amino acid aminotransferase (EC 2.6.1.42), tyrosine aminotransferase (EC 2.6.1.5), 5-methylthioribose kinase (EC 2.7.1.100), ATP phosphoribosyl transferase (EC Expression of L-tryptophan-pyruvate aminotransferase (EC 2.6.1.99) is responsible for the formation of tryptophan from indolepyruvate, which is then converted into
Discussion
Plants subjected to water stress undergo a range of physiological changes [30]. In our previous study, in comparison to susceptible genotype, a higher level of CAT activity was exhibited in resistant genotype against short-term drought; however, SOD and MDA activity showed no significant differences between the two L. multiflorum genotypes when treated with 1 h of drought [20]. In contrast, in the present study, a significant difference was noted between the two annual ryegrass genotypes in response to long-term drought (Fig. 1). As the number of days of drought increased, tolerant plants exhibited higher levels of CAT and SOD activity, higher RWC and chlorophyll content, and lower MDA content (Fig. 1). Research conducted by Mastalerczuk et al. [31] and Borawska-Jarmujłowicz et al. [32] on perennial ryegrass showed that significant reductions in chlorophyll contents were observed after subjecting seedlings to abiotic stresses. These findings are consistent with those of the present study. Thus, the activity of antioxidant enzymes and changes in chlorophyll content play important roles in increasing the tolerance of resistant genotype to drought stress.
Alterations in membrane lipid composition represent an important response to environmental stresses [33,34]. Drought stress reduces the accumulation of fatty acids, such as palmitic acid, hexadecenoic acid, picolinic acid, and stearidonic acid [4,35,36], while substantially increasing the levels of most fatty acids, with the exception of pimelic acid, capric acid, (2'E,4'Z,7'Z,8E)-colnelenic acid, alpha-linolenic acid, traumatic acid (TA) and acetyl-CoA carboxylase (ACCase) [37]. Tayeh et al. [38] showed that phospholipase A1 (PLA1) in plants plays important roles in the hydrolysis of phospholipids during cold stress. Andrade et al. [39] found that lipoxygenase (LOX) is involved in defense against drought stress in rice and sugarcane transgenic plants. Uma and Podile [40] reported that the transcripts encoding 9-LOX and a 9-LOX-derived compound, colnelenic acid, are significantly upregulated in resistant host tomato plants. Similarly, Cao et al. [41] reported that levels of alpha-linolenic acid and colnelenic acid were increased following treatment with exogenous methyl jasmonate (MeJA) . In the present study, phospholipase A1 enhanced the production of alpha-linolenic acid, which is converted to colnelenic acid; an effect that was directly induced by lipoxygenase in the tolerant annual ryegrass subjected to drought conditions (Fig. 6). These results showed that alpha-linolenic acid and colnelenic acid are involved in the regulation of membrane composition in drought-tolerant genotype.
In addition, TA is a plant wound hormone that can eliminate the negative effects of salt stress and oxidative stress in Chlorella vulgaris [42,43]. Consistent with our previous results, a high TA content was detected in the resistant plants, indicating that the metabolite contributes to enhanced stress tolerance in drought-treated annual ryegrass. Moreover, levels of acetyl-CoA carboxylase (ACCase) and omega-hydroxypalmitate Oferuloyl transferase in the annual ryegrass plants were enhanced by drought treatment. We found that reductions in the levels of palmitic acid and hexadecenoic acid were controlled by ACCase. The changes observed in ACCase and palmitic acid in lipid metabolism were consistent with the results of Kwan et al. [44].
Most free amino acids accumulate under drought stress [45], and this is thought to be an important adaptation [46,47]. In the present study, downregulated branched-chain aminotransferase (BCAT) directly reduced L-valine and L-leucine levels. Anthranilate synthase (AS) causes a leucine-to-proline change, which leads to a reduction in the proline content of annual ryegrass. In addition, the upregulation of histidinolphosphatase is involved in L-histidine degradation. Furthermore, tyrosine amino transferase and phenylpyruvate tautomerase might be jointly involved in the degradation of phenylalanine. This finding has not been observed in previous studies.
The response of 5-methylthioribose to drought stress in drought-tolerant annual ryegrass is induced by cysteine and methionine metabolism, which was controlled in the present study by the two downregulated enzymes, tyrosine aminotransferase and 5methylthioribose kinase. Moreover, this compound has not been associated with drought tolerance in previous studies. Anthranilate acts as a precursor for tryptophan biosynthesis and reduces indoleacetic acid (IAA) synthesis [48]. Anthranilate formation resulted in indirect induction of IAA production, a novel finding in annual ryegrass that was consistent with the results of Hartmann and Zimmer [49].
The behavior of 2′,3′-cyclic nucleotide variants has been characterized in animal tissues, but not in plant materials [50]. According to our research, only 2′,3′-cyclic GMP was detected in annual ryegrass, which is involved in guanine production. Glucosylglycerol, found in many cyanobacteria, and galactosylglycerols (floridoside and isofloridoside) accumulate in eukaryotic algae under salt stress conditions [51]. In the present study, galactosylglycerol production was involved in glycolysis and fructose/mannose metabolism in response to drought, indicating that the concentration of galactosylglycerol plays a role in drought tolerance of annual ryegrass. In another study, galactose levels in drought-resistant potato genotypes were observed to be much higher than those in corresponding controls [52], a finding that is consistent with our results.
Transcription factors involved in transcriptional and post-transcriptional mechanisms regulate the accumulation of various metabolites by activating the expression of biosynthetic enzymes. Based on our omics data, various TF families, including MYB, bZIP, and bHLH, might regulate core metabolic processes during drought stress. The R2R3-MYB transcription factor AtMYB41 is induced in response to desiccation, and activates cuticle biosynthesis In Arabidopsis thaliana under biotic stress conditions [53]. In addition, AtMYB96 positively regulates the expression of lipidtransfer protein 3 (LTP3) via direct binding to the LTP3 promoter, and thereby enhances plant drought stress tolerance [54].
In response to drought stress, many plant species accumulate high levels of compatible osmolytes, such as Pro (proline), Gly betaine, or sugar alcohols, which are thought to be associated with stress adaptation. In higher plants, L-Pro is synthesized from L-Glu via delta-1-pyrroline-5-carboxylate (P5C) by two enzymes, P5C synthetase (P5CS) and P5C reductase (P5CR) under drought stress. The L-Pro is metabolized to L-Glu via P5C by two enzymes, Pro dehydrogenase (ProDH) and P5C dehydrogenase, during recovery from stress [55,56].
Satoh and colleagues [57] first identified the group of S bZIP TFs, such as AtbZIP11, AtbZIP44, AtbZIP2, and AtbZIP53, which become involved in Pro metabolism by binding to the cis-acting element ACTCAT in the ProDH promoter. Later, Hanson and colleagues [58] reported that AtbZIP11 could affect amino acid metabolism by regulating the expression of ASPARA-GINE SYNTHETASE1 and ProDH2 during low energy stress. Post-transcriptional regulatory mechanisms are also very important for drought stress tolerance in plants. It has been reported that AtbZIP1 and Atb-ZIP53 can be specifically heterodimerized with group C bZIP TFs to bind directly to the promoters of AS-PARAGINE SYNTHETASE1 and ProDH2 [59]. In addition, the Arabidopsis type-B response regulator 18 (ARR18) physically interacts with AtbZIP63; an interaction that negatively interferes with the transcriptional activity of AtbZIP63 on the ProDH1 promoter under osmotic stress [60]. Recently, the key TF AtbHLH112 involved in Pro metabolism was identified by Liu et al. [61]. AtbHLH112 could activate the expression of P5CS and repress the expression of P5CDH and ProDH to increase Pro levels under abiotic stress tolerance. Taken together, the omics data suggest that post-transcriptional and transcriptional regulation of TFs that affect proteins and enzymes could mediate lipid and amino acid metabolism, thereby promoting adaptation to drought stress in annual ryegrass.
Over the last few decades, extensive efforts have been made to produce drought-tolerant ryegrass genotypes using molecular and biotechnological methods, such as the production of genetically modified or transgenic plants, with a focus on lipid and amino acid metabolism [62,63]. Transgenic expression of some drought stress-related genes involved in lipid and amino acid metabolism could provide a more rapid strategy to achieve improved drought stress tolerance in ryegrass.
Conclusions
In order to explore the molecular mechanism associated with drought tolerance in two annual ryegrass genotypes, we identified differentially expressed metabolites and their corresponding proteins and transcripts that are involved in 23 core metabolic processes, in response to short-term drought stress. The regulatory networks were inferred using MCoA and correlation analysis to reveal the relationships among the expression of transcripts, proteins, and metabolites that highlight the corresponding elements of these core metabolic pathways. This study provides useful insights into the molecular mechanisms of drought resistance and represents a promising approach toward enhancing drought tolerance in forage grass.
|
2018-01-31T18:41:03.397Z
|
2018-01-30T00:00:00.000
|
{
"year": 2018,
"sha1": "a9f1a9e02db751bb0594f4d5f896232e245a2b17",
"oa_license": "CCBY",
"oa_url": "https://bmcplantbiol.biomedcentral.com/track/pdf/10.1186/s12870-018-1239-z",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "a9f1a9e02db751bb0594f4d5f896232e245a2b17",
"s2fieldsofstudy": [
"Biology",
"Environmental Science"
],
"extfieldsofstudy": [
"Medicine",
"Biology"
]
}
|
258281557
|
pes2o/s2orc
|
v3-fos-license
|
Correlation between total phenolic and flavonoid contents and biological activities of 12 ethanolic extracts of Iranian propolis
Abstract Propolis is a resinous substance produced by honey bees that is very popular as a natural remedy in traditional medicine. The current research is the first study on the biological properties of ethanolic extracts of propolis (EEP) from several different regions (12) of Iran. Total phenolic and flavonoid contents (TPC and TFC) of Iranian EEPs were variable between 26.59–221.38 mg GAE/g EEP and 4.8–100.03 mg QE/g EEP. The DPPH scavenging assay showed all the studied EEP samples, except for the sample with the lowest TPC and TFC (P6), have suitable antioxidant activity. All the EEPs inhibited both cholinesterase enzymes (acetylcholinesterase: AChE, butyrylcholinesterase: BuChE) but most of them exhibited a distinct selectivity over BuChE. Evaluation of the antibacterial activity of the EEP samples using four pathogenic bacteria (B. cereus, S. aureus, A. baumannii, and P. aeruginosa) demonstrated that the antibacterial properties of propolis are more effective on the gram‐positive bacterium. Spearman correlation analysis showed a strong positive correlation between TPC and TFC of the Iranian EEPs and their antioxidant, anticholinesterase, and antibacterial activities. Considering that there is ample evidence of anticholinesterase activity of flavonoids and a significant correlation between the anticholinesterase activity of the studied Iranian EEPs and their total flavonoid content was observed, the interaction of 17 well‐known propolis flavonoids with AChE and BuChE was explored using molecular docking. The results indicated that all the flavonoids interact with the active site gorge of both enzymes with high affinity. Summing up, the obtained results suggest that Iranian propolis possesses great potential for further studies.
| INTRODUC TI ON
Propolis (bee glue) is a natural product created by honey bees (Apis mellifera L.) by mixing their saliva, beeswax, and exudates derived from different parts of the plants (Marcucci, 1995;Silva-Carvalho et al., 2015). This resinous material forms the inner lining of the beehives and is used as a protective barrier against foreign invaders, wind, and rain. The honey bees also use propolis to repair cracks and holes in the walls of the hive and seal openings and smooth the inner walls, maintain the hive's internal temperature and cover (embalm) the corpses of large invaders that are difficult to transport out of the hive. Propolis has a very pleasant aroma and due to its antimicrobial activity, prevents the growth of bacteria and fungi inside the hives (Burdock, 1998;Martinotti & Ranzato, 2015).
The exact composition, aroma, and color of raw propolis (green, red, brown, and yellow) depend on several factors such as botanical source, collection season, and geographical area (Wang et al., 2016).
There are also a small number of enzymes in propolis including glucose 6-phosphatase, succinic dehydrogenase, acid phosphatase, adenosine triphosphatase, and beta-amylase (Bhargava et al., 2021;Lotfy, 2006). Propolis has been used by humankind for different medicinal and nonmedicinal purposes since ancient times (Kuropatnicki et al., 2013). According to the published scientific data, various biological properties have been attributed to propolis, such as anticancer, antioxidant, anticholinesterase, antihypertensive, liver protection, wound-healing, oral health, anti-inflammatory, antiulcer, immunomodulatory, antimicrobial, antiviral, antifungal, and antiparasitic (Baltas et al., 2016;Bhargava et al., 2021;Dilokthornsakul et al., 2022;Kocot et al., 2018;Pasupuleti et al., 2017;Rezvannejad et al., 2017;Saeed et al., 2021;Suran et al., 2021;Viuda-Martos et al., 2008). The data provided by in vitro studies, animal models, and human clinical trials demonstrate that propolis can reduce the manifestations of neurological and brain disorders through its protective and therapeutic effects (Zulhendri, Chandrasekaran, et al., 2021;. The current research is the first comprehensive study on the biological properties of several propolis samples collected from different regions of Iran. This study covers the evaluation of the total phenolic and flavonoid contents (TPC and TFC) and the antioxidant, anticholinesterase, and antibacterial activities of 12 ethanolic extracts of Iranian propolis. Although the antibacterial and antioxidant activities of Iranian propolis have been previously studied, its anticholinesterase activity has not been studied until now. Meanwhile, previous studies have been conducted on one or a few limited Iranian propolis samples. In this research, also for the first time, the interaction of 17 well-known propolis flavonoids with the cholinesterase enzymes (AChE and BuChE) has been investigated using molecular docking.
| Propolis samples and preparation of the ethanolic extracts
Twelve Iranian propolis samples (P1-P12) used in this research are listed in Table 1, along with coordinates. Raw propolis samples were collected by experienced beekeepers from beehives located in various regions of Iran. The propolis samples were stored at −20°C until use. Preparation of the ethanolic extracts of propolis (EEP) was performed in the following steps: Raw propolis was pulverized by liquid nitrogen. Ten grams of powdered propolis with 100 mL of 80% ethanol was placed in a dark glass flask and stirred on a shaker at room temperature for 72 h and the mixture was then filtrated by centrifugation in two steps. The supernatant was concentrated using rotary evaporation at 40°C and then placed in a vacuum oven at 40°C for complete drying. The ethanolic extracts were kept in the dark containers at −20°C until further steps.
| Determination of total phenolic content
The total phenolic content (TPC) of the EEP samples was measured utilizing the colorimetric Folin-Ciocalteu method and gallic acid was used as the calibration standard (Singleton & Rossi, 1965). The results were expressed as milligrams of gallic acid equivalents (GAE) per gram of EEP (mg GAE/g EEP). For each sample, the experiment was performed in triplicate. For this purpose, 100 μL of 0.2 N Folin-Ciocalteu reagent, 590 μL of distilled water, and 10 μL of EEP in 80% ethanol (final concentration of 50 μg/mL) were placed in a cell for 1 min at room temperature in the dark. Three hundred microliters of sodium carbonate (7.5%) was then added to the cell and the mixture was incubated in the dark for 30 min at room temperature. The absorbance was read at 735 nm with a spectrophotometer (Cary 50, Australia). A solution containing 100 μL of 0.2 N Folin-Ciocalteu reagent, 600 μL of distilled water, and 300 μL of sodium carbonate (7.5%) was utilized as the blank sample.
| Determination of total flavonoid content
The aluminum chloride colorimetric method was used to measure the total flavonoid content (TFC) of EEP samples (Wang et al., 2016).
Five hundred microliters of EEP in 75% ethanol (final concentration of 50 μg/mL) was added to 500 μL of 2% aluminum chloride in a tube. The tube was then incubated in the dark for 15 min at room temperature. The absorbance was recorded at 435 nm with a spectrophotometer (Cary 50, Australia). A solution containing 500 μL of 2% aluminum chloride and 500 μL of distilled water was used as the blank. Three replications were performed for each sample.
Quercetin was applied as the calibration standard and the TFC of EEP samples was expressed in milligrams of quercetin equivalents (QE) per gram of EEP (mg QE/g EEP).
| DPPH-free radical scavenging assay
To evaluate the antioxidant activity of EEP samples, the DPPH (2, 2-diphenyl-1-picrylhydrazyl) free radical scavenging assay was used (Bondet et al., 1997). All the EEP samples were tested at six different concentrations. For each concentration, three replications were performed. Briefly, 750 μL of 0.4 mM DPPH solution was added to 250 μL of EEP dissolved in methanol. After 30 min of incubation in the dark at room temperature, the absorbance of the sample was recorded at 517 nm. A sample containing 750 μL of DPPH solution (0.4 mM) and 250 μL of methanol was considered as the control sample. Ascorbic acid was used as the positive control. For this aim, six different concentrations of ascorbic acid (0.5-12.5 μg/mL) were used and for each concentration, three replications were considered.
The following formula was used to calculate the percentage of DPPH inhibition free radical by each concentration of the EEP sample: A control , absorbance of the control sample; A sample , absorbance of the EEP sample.
The IC 50 value (the concentration of a sample that has the ability to inhibit DPPH radical by 50%) for each sample was obtained using a dose-response graph plotting the percentage of DPPH inhibition versus the concentration logarithm of the EEP sample.
| Measurement of anticholinesterase activity
To determine the half maximal inhibitory concentration (IC 50 ) of EEP samples for acetylcholinesterase (AChE) or butyrylcholinesterase (BuChE), the activity of the enzymes was measured in the absence and presence of six different concentrations of each EEP sample by the Ellman method (Ellman et al., 1961). Each concentration was analyzed in triplicate and a well-known inhibitor of the cholinesterase enzymes, neostigmine, was used as the positive control. All the enzyme assays were performed on a 96-well plate with a final volume of 200 μL.
The assay mixture consisted of 0.1 M potassium phosphate buffer (pH 8.0) and the enzyme (AChE or BuChE) and DTNB with the final concentrations of 0.1 unit/mL and 0.5 mM, respectively. The EEPs (in 70% ethanol) were incubated with the assay mixture for 10 min. Substrate (acetylthiocholine iodide or S-butyrylthiocholine iodide) was then added to the assay with the final concentration of 1 mM and the absorbance was read after 10 min at 405 nm by a microplate reader (Elx808 Biotek Instruments). A sample containing all assay components, except the enzyme was applied as the blank.
Finally, the percentage of enzyme activity inhibition at each concentration of the EEP sample was determined and the IC 50 value was calculated from the dose-response curve plotting the percentage of inhibition versus concentration logarithm of the EEP sample.
The results were reported as the mean ± standard deviation (SD).
| Evaluation of antibacterial activity
To evaluate the antibacterial activity of EEP samples, four pathogenic bacteria, Bacillus cereus, Staphylococcus aureus, Acinetobacter In the next step, the antibacterial activity of the EPP samples was investigated using the agar well diffusion method (Balouiri et al., 2016;Domingue et al., 1994).
| Statistical analysis
The data were calculated in the form of arithmetical mean values and standard deviations. The correlation between data was statistically analyzed using the one-way ANOVA method, Spearman correlation of SPSS software (version 13.0). The diameter of growth inhibition halo data was also statistically analyzed using the ANOVA method and SAS software (version 9.1) and means comparison has been performed by Tukey method and proc GLM.
y i denotes observations of studied parameters. x i denotes different concentrations of the EEP samples. ɛ i is the deviation vector.
| Molecular docking studies
To explore the possible binding sites of 17 well-known propolis flavonoids on the cholinesterase enzymes (AChE and BuChE) by molecular docking, AutoDock Vina 1.1.2 software (Trott & Olson, 2010) was used. For this aim, the crystal structure of human AChE (PDB entry 4M0E) (Cheung et al., 2013) and human BuChE (PDB entry 4TPK) (Brus et al., 2014) were used as the receptors. Before using the protein structures in the docking studies, the ligand and water molecules were removed and the missing residues were then added using MODELLER software (Webb & Sali, 2016). The protein structures were finally energy minimized by GROMACS (Berendsen et al., 1995). The TPC and TFC of the studied Iranian EEP samples are shown in Table 6, there is a perfect positive correlation between TPC and TFC (R 2 = 1), which means the EEP samples with higher TPCalso possess higher TFC and vice versa. There are frequent scientific reports that the vegetation of the propolis collection area has a very significant effect on its composition (Anjum et al., 2019;Huang et al., 2014;Wieczorek et al., 2022).
Oxidative stress is mainly created as a result of the imbalance be- Socha et al., 2015). There are also published documents demonstrating the positive correlation between the antioxidant capacity of the propolis extracts and their TPC (Ahn et al., 2007;Hamasaka et al., 2004;Kalogeropoulos et al., 2009;Moreira et al., 2008;Wang et al., 2016) or TFC (Isla et al., 2009).
| Anticholinesterase (anti-AChE and anti-BuChE) activity
Based on several published papers in recent years, propolis can be considered as a promising therapeutic natural substance to protect the brain and treat neurological injuries and disorders (Ayikobua et al., 2018;Bhargava et al., 2021;Zulhendri, Chandrasekaran, et al., 2021;.
The administration of cholinesterase enzymes (AChE and
BuChE) inhibitors is the main therapeutic strategy for the symptomatic treatment of mild to moderately severe forms of AD (Colovic et al., 2013;Sharma, 2019). The anticholinesterases are also used to manage other forms of neurological disorders, such as ataxia, (Lotfi et al., 2020;Mariki et al., 2021;Sharma, 2019). The identification of new potent anticholinesterase compounds through the study of natural resources has also attracted a lot of attention (Dey et al., 2017;Dos Santos et al., 2018;Houghton et al., 2006;Mukherjee et al., 2007).
The results corresponding to the evaluation of anti-AChE and anti-BuChE activity of the EEP samples are represented in Table 3. As shown in Table 3, the anti-BuChE activity of three EEP samples (P4, P5, and P11) is slightly higher than anti-AChE activity, but the rest of the samples show a distinct selectivity over BuChE. The data published on the anticholinesterase activity of Algerian propolis methanolic extracts also demonstrated the higher ability of the extracts to inhibit BuChE than AChE (Boulechfar et al., 2019(Boulechfar et al., , 2022.
Considering that in the late stages of AD, there is a decrease in AChE level and an increase in BuChE level in the brain, selective BuChE inhibitors are of great importance in the treatment of advanced AD .
As shown in Table 6
| Antibacterial activity
The In addition to the determination of MIC (minimum inhibitory concentration) and MBC (minimum bactericidal concentration) ( As mentioned earlier, the antibacterial activity of the EEP samples was also investigated using the agar well diffusion method. As shown in Table 5, seven different concentrations (7.81-500 μg/mL) of each EEP sample and one concentration (500 μg/mL) of ciprofloxacin (positive control) were studied in this test. An overview of the results indicated that as the concentration of EEP samples increases, the diameter of the growth inhibition halo also increases. The inhibition halo diameter of ciprofloxacin is 9 mm for B. cereus and S. aureus and 11 and 10 mm for A. baumannii and P. aeruginosa. The diameters of growth inhibition haloes corresponding to 500 μg/mL con-
The existence of a strong positive correlation between TPC
and TFC of the EEPs and the results obtained from agar well diffusion 1 (R 2 = .81), 2 (R 2 = .83), 3 (R 2 = .85), and 4 (R 2 = .82) ( Table 6) demonstrates that the EEP samples with higher TPC and TFC inhibit the growth of pathogenic bacteria with more power. These results are consistent with some previously published reports (da Silva et al., 2006;Górniak et al., 2019;Inui et al., 2014;Yuan et al., 2021).
| Molecular docking
As implied before, the propolis flavonoid composition, which is responsible for many of its biological and medicinal properties, is used as a criterion for the assessment of propolis quality (Huang et al., 2014;Wang et al., 2016). Scientific reports indicate the anticholinesterase activity of the flavonoid compounds (Dzoyem et al., 2017;Khan et al., 2009Khan et al., , 2018. Considering that the results obtained from the current research also represent a strong positive correlation between the anticholinesterase activity of EEP samples and their TFC, it was decided to investigate the interaction of 17 well-known propolis flavonoids (Kocot et al., 2018;Pasupuleti et al., 2017;Zhang et al., 2021) with the active site gorge of the cholinesterase enzymes (AChE and BuChE) by molecular docking studies. The structure of these flavonoid compounds is illustrated in Figure 1.
The AChE docking results demonstrated that all 17 flavonoid compounds have the ability to interact with the active site gorge of the enzyme and their best binding energy varies between −8.9 and −7 kcal/mol. The lowest docking energy is corresponding to rutin and the highest is related to fisetin (Table 7). Table 8 and Ala204) (Atanasova et al., 2015;Damuka et al., 2020;Kua et al., 2003;Wiesner et al., 2007;Zhang et al., 2002). (Brus et al., 2014;Macdonald et al., 2012).
The amino acid residues participating in the binding of BuChE to the flavonoids are represented in Table 9. As well seen, all the TA B L E 6 The Spearman correlations between studied parameters calculated by SPSS software. Figure 2c,d.
F I G U R E 1
The structure of 17 well-known propolis flavonoids.
According to the docking results, all the studied flavonoids bind to the gorge region of both enzymes with high affinity ( Table 7).
ACK N OWLED G M ENTS
The authors gratefully appreciate the funding support received for the project from the Research Council of the Graduate University of Advanced Technology, Kerman, Iran.
CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare no conflict of interest.
DATA AVA I L A B I L I T Y S TAT E M E N T
Data are available on request from the authors.
|
2023-04-23T15:12:30.844Z
|
2023-04-21T00:00:00.000
|
{
"year": 2023,
"sha1": "d6cf9a5f5229b71ed1837bea12da5ea5e3c12e31",
"oa_license": "CCBY",
"oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/fsn3.3356",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "35f36cf233f1884d652d677176ae055f9b41d483",
"s2fieldsofstudy": [
"Agricultural And Food Sciences"
],
"extfieldsofstudy": []
}
|
3816568
|
pes2o/s2orc
|
v3-fos-license
|
Using the Electronic Medical Record to Correlate Kawasaki Disease Phenotypes With Clinical Outcomes
Abstract Background We sought to systematically standardize the documentation of clinical and laboratory features in Kawasaki disease (KD) on the day of initial treatment and correlate the presentation with clinical outcomes. Methods Kawasaki disease features and classification were documented by the attending physician using a standardized documentation tool on the day of treatment for KD, including confidence in the KD diagnosis on a 4-point scale. Incomplete KD was further classified using American Heart Association (AHA) criteria (sufficient or insufficient) and baseline echocardiogram data. We prospectively recorded intravenous immunoglobulin (IVIG) resistance, coronary artery abnormalities (CAAs), periungual peeling, and retrospectively identified subsequent diagnoses of autoimmune/inflammatory disease. Results From November 2012 to October, 2015, 162 patients were treated for KD: 105 with complete KD (Group 1), 7 with incomplete KD based on CAAs on day of KD diagnosis (Group 2), 23 with incomplete KD meeting AHA criteria (Group 3), and 27 with incomplete KD and insufficient AHA criteria (Group 4). Group 4 patients had lower baseline median C-reactive protein levels (Group 4 median 4.65 mg/dL [interquartile range {IQR}, 2.3–13.6] vs Group 1 median 8.0 mg/dL [IQR, 4.5–17], Group 2 median 13.9 mg/dL [IQR, 1.4–18.2], Group 3 median 13.3 mg/dL [IQR, 4.9–20.2]), and no coronary abnormalities developed, although 11% had IVIG resistance. Group 4 had higher rates of subsequent autoimmune/inflammatory conditions diagnosed (11.1% in Group 4 vs <5% for all others, P = .02). Conclusions Standardized documentation and classification of KD features may be useful to correlate with clinical outcomes, including subsequent diagnosis of autoimmune/autoinflammatory disease. Among patients with incomplete KD who did not meet AHA criteria and had a normal baseline echocardiogram, the IVIG resistance rate may have been related to a lower likelihood of an accurate diagnosis of KD.
Kawasaki disease (KD) requires timely diagnosis and treatment to prevent development of coronary artery abnormalities (CAA) [1]. The clinical diagnosis of KD can be somewhat subjective, because clinical features can be transient and vary in intensity with some children lacking complete features (incomplete KD) [2]. Kawasaki disease requires multidisciplinary care, and those who may diagnose KD (ie, hospitalist, rheumatology, or infectious diseases [ID] specialist) and cardiologists involved in the management of patients. Care may sometimes be fragmented with multiple providers and transfers of care at different stages of illness and/or follow up, particularly for those patients with refractory and recurrent disease. This may lead to lack of communication of pertinent history. To address this need for continuity of care, in 2012, we created a multidisciplinary team for the care of all KD patients at our hospital, which comprised ID and cardiology clinicians with rheumatology care as needed. The attending ID physicians, who are typically the clinicians who first evaluate KD patients in our hospital, provided standardized clinical and laboratory data documentation of criteria for diagnosis on the day of treatment with intravenous immunoglobulin (IVIG). Data documentation was performed using an electronic data collection form specifically designed for this condition.
The objective of this study was to correlate documented clinical and laboratory findings and variables on the day of treatment with IVIG with subsequent clinical outcomes (IVIG resistance, CAA, periungual peeling, and subsequent diagnosis of autoimmune/autoinflammatory conditions including periodic fever syndromes) based on the documented type of KD presentation on day of treatment.
METHODS
The multidisciplinary KD clinic (staffed by ID and cardiology physicians) was created in 2012. The custom online form was created at the same time within the electronic medical record (EPIC software, Madison, WI) to detail the clinical and laboratory KD data, which were completed by the attending ID physician on the day of treatment with IVIG (Supplementary Figure 1) and on the day of patient discharge. Due to the subjective nature of some KD features, the level of confidence in the diagnosis of KD by the attending physician on the day of treatment was also recorded on a 4-point scale: 1 was the highest confidence in the KD diagnosis and 4 was the least confidence in the KD diagnosis ("possible KD and benefits of treatment outweigh risks of not treating").
Children diagnosed with KD were classified as complete or incomplete KD. For those treated for incomplete KD, cases were further classified as to whether sufficient or insufficient American Heart Association (AHA) clinical and laboratory criteria were present [2]. At minimum, all KD patients had an echocardiogram at baseline, 2 weeks, and 5-8 weeks after diagnosis. Coronary artery abnormalities were defined as those with a z score ≥2.0, based on the maximal internal diameters of the proximal right coronary artery or left anterior descending coronary artery. The z score of ≥2.0 was chosen because it is included as part of the echocardiographic considerations of treatment for incomplete KD in the published AHA guidelines when additional echocardiographic changes are present.
All inpatients diagnosed with acute KD were scheduled for follow up in the multidisciplinary KD clinic, staffed by both ID and cardiology clinicians, at a minimum of 2 weeks and 6-8 weeks post-IVIG treatment, and more frequently if required by their clinical case and/or coronary artery evolution. After the 6-to 8-week visit, patients assessed as uncomplicated were transitioned for follow up by cardiology care only. We used the data collection tool to prospectively record the presence of arthralgia/arthritis, presence of periungual peeling, and CAA during follow up at both the 2-and 6-to 8-week follow-up visits.
Intravenous immunoglobulin resistance was defined as persistent or recrudescent fever from 36 hours to 7 days after the completion of the first IVIG. Diagnosis of autoinflammatory/autoimmune syndromes, which included periodic fever syndromes, were recorded as of May 1, 2016 by retrospective review of the electronic medical record for all included patients, with Institutional Review Board approval. Autoinflammatory/ autoimmune conditions included juvenile idiopathic arthritis, Bechet's disease, systemic lupus erythematosus, polyarteritis nodosa, ankylosing spondylitis, dermatomyositis, and inflammatory bowel disease. Macrophage activation syndrome within 1 month after the KD diagnosis alone was not included because this is a known complication of KD. Periodic fever syndromes required either a clinical diagnosis by an ID or rheumatology attending physician of periodic fever, aphthous stomatitis, pharyngitis, and adenopathy (PFAPA) syndrome or another recurrent fever syndrome [3].
Kruskal-Wallis tests were used for comparisons, and a χ 2 test was used for proportions. Two-tailed P < .05 was considered significant. All analyses were performed using GraphPad Prism (GraphPad Software Inc., San Diego, CA).
Kawasaki Disease Patients Treated With Intravenous Immunoglobulin
From November 2012 to October 2015, our institution provided initial IVIG treatment to 162 patients with acute KD. The clinical classification tool was used in 143 (88.3%) cases. For those patients in whom the tool was not used, we retrospectively classified the patients into the appropriate KD type based on available data. Respiratory viral testing was performed in 156 (96.2%) of patients and 52 were positive (33.3%), 4 of whom had coinfections with more than 1 respiratory virus. The most common viruses identified were rhinovirus/enterovirus (n = 36), followed by human adenovirus (n = 12), parainfluenza (n = 2), human metapneumovirus (n = 2), influenza (n = 2), respiratory syncytial virus (n = 1), and coronavirus (n = 1).
Clinical and Laboratory Characteristics by Type of Kawasaki Disease
Clinical and laboratory characteristics by type of KD are presented in Table 1. Group 1 presented with complete KD (n = 105), Group 2 presented with incomplete KD and baseline echocardiographic abnormalities (n = 7), Group 3 presented with incomplete KD and sufficient AHA-defined criteria (n = 23), and Group 4 (n = 27) presented with incomplete KD and insufficient AHA criteria. There were 24 patients (14.8%) with IVIG resistance in the defined time period, and there were no significant differences in IVIG resistance among groups. The median C-reactive protein (CRP) was significantly lower in Group 4 than all others (median 4.65 vs 8.0-13.9 mg/dL in other groups) ( Table 1), and respiratory viral detection was most frequent in Group 4, but this was not statistically significantly different from the others. Subsequent autoimmune/autoinflammatory diagnoses were most common in Group 4 (11.1% vs 0%-4.1% in other Groups).
Timing of Coronary Abnormalities
Coronary artery abnormalities had been identified in 17 of 162 patients (10.4% of the cohort). Of these, CAA were present at baseline in 13 patients (76.5%). Four patients with normal baseline coronary arteries progressed to CAA (time to progression on echocardiogram was 7, 11, 13, and 15 days after baseline). All 4 patients had either IVIG resistance or required intensive care. Three patients required intensive care for fluid refractory hypotension. Additional autoimmune/autoinflammatory diagnoses documented in patients who were treated for acute KD are described below and listed by type of KD presentation.
Complete Kawasaki Disease
A 6-year-old African American boy was treated after 15 days of fever. He had a history of oropharyngeal changes, rash, and hand/feet and lymph node swelling. On physical exam, perianal peeling was noted with oropharyngeal erythema; he lacked scleral injection and hand and feet changes on that day. He was treated for likely complete KD in subacute phase of illness. The level of confidence in the KD diagnosis was rated as 2 on the 4-point scale. He had previously experienced intermittent episodes of unexplained fever. He never developed CAA but did have periungual peeling. Two years after the KD diagnosis, he was seen in rheumatology clinic with a history of intermittent fevers and arthralgia. Genetic testing was positive (heterozygous) for a variant in the MEFV gene, and he was diagnosed with familial Mediterranean fever with cessation of symptoms after starting colchicine treatment.
Incomplete Kawasaki Disease: Meeting American Heart Association Criteria
As previously published in more detail [4], a 5-year-old white male presented with 7 days of fever and a 1-month history of rash. The treating clinician categorized the lowest confidence in the KD diagnosis before treatment. He was subsequently diagnosed with systemic onset juvenile idiopathic arthritis with macrophage activation syndrome at 7 days after the KD diagnosis. He did not demonstrate any periungual peeling or CAA.
Incomplete Kawasaki Disease: Non-American Heart Association Criteria
A 10-year-old white boy was treated for KD with fevers for 9 days at the time of treatment. The physician categorized the lowest confidence in the KD diagnosis, documented that uveitis was found on ophthalmologic exam, and prompted treatment for peeling found on the pads of the fingers and palms (not periungal). The patient did not have subsequent CAA. Two years later, he developed bloody stools and mouth ulcers, and intestinal biopsies confirmed a diagnosis of Crohn's disease. A 10-month-old white female was treated on day 8 of fever for conjunctival injection only. The treating clinician classified the confidence in KD diagnosis as "probably" (confidence level 3). Treatment was provided when the nasopharyngeal swab for respiratory viruses was negative and because the CRP was elevated at 13.6 mg/dL. She did not demonstrate any periungual peeling and had no CAA. The following year, the child was diagnosed with PFAPA syndrome in an ID clinic after she had repeated, intermittent monthly episodes of fever lasting 2-5 days while remaining well in between the episodes. No genetic testing was performed.
A 6-year-old white male presented with 6 days of fever, strawberry tongue, maculopapular rash by history (although absent on the day of presentation), unilateral cervical lymphadenopathy, and generalized myalgias. The treating clinician categorized the lowest confidence in the KD diagnosis before treatment. During initial hospitalization, he developed severe arthralgias. He received 2 doses of IVIG and corticosteroids, and joint symptoms progressed. He was subsequently diagnosed with systemic onset idiopathic juvenile arthritis. He did not demonstrate any periungual peeling and had no CAA.
Classification of Kawasaki Disease (KD) Phenotype by Clinical Confidence in the KD Diagnosis
There were 142 patients in whom the treating clinician provided data as to the confidence in the KD diagnosis (Table 2). No cases of CAA were found among those with 2 lowest clinical confidence ratings for the KD diagnosis. The groups that had the lower confidence in the KD diagnosis had higher frequency of subsequent autoimmune/autoinflammatory diagnosis. Those with the lowest confidence in the KD diagnosis had the highest frequency of IVIG resistance.
DISCUSSION
Our study presents data linking the clinical features at the initial KD presentation to the clinical outcomes of periungual peeling, CAA, IVIG resistance, and subsequent diagnosis of autoimmune/autoinflammaotry diagnoses. We added a rating scale to the overall confidence in the KD diagnosis, which allowed us to gauge the overall clinical impression on the day of treatment for KD. There are potential benefits to standardizing documentation of the clinical presentation of KD on the day of treatment. First, the standardization of data collection allows for reassessment of the diagnostic certainty before proceeding with subsequent treatments for those who fail to respond to initial IVIG. This is important before considering additional doses of IVIG, because adverse effects can be dose dependent, such as hemolytic anemia [5][6][7]. Second, more precise associations of clinical outcomes with the type of KD, especially incomplete KD, can be made using this approach. Utilization of the AHA algorithm for incomplete KD has been shown to improve the diagnosis of KD in children who developed CAA [8], but the accuracy of the diagnosis for those with incomplete KD who do not meet these criteria is less clear. Among children with incomplete KD who did not meet the AHA criteria and whose initial echocardiograms were normal in this cohort, the IVIG resistance rate remained relatively high at 11%, but none developed CAA, and they had a trend towards lower rate of periungual peeling than other groups. Finally, utilization of this approach can allow for more precise epidemiologic data for the incidence and characteristics of IVIG resistance. Intravenous immunoglobulin resistance rates may vary [9], but these rates may be related to a lower likelihood of an accurate diagnosis of KD initially in some patients.
Patients treated for incomplete KD without baseline coronary abnormalities and who did not meet the AHA-defined criteria were more likely to have a subsequent autoimmune/ autoinflammatory diagnosis. We have previously reported the clinical presentation of systemic onset juvenile arthritis after the KD diagnosis and noted on literature review of prior reported cases that most were diagnosed as incomplete KD specifically lacking scleral injection [4]. Mevalonate kinase deficiency and PFAPA diagnoses causing recurrent fevers have been previously described after initial treatment for incomplete KD [10,11]. Whether children who were subsequently diagnosed with an autoimmune/autoinflammatory illness actually had KD initially or had an alternative diagnosis is unclear. It has been suggested that recurrent fever disorders and KD may share the same propensity for activation of innate immune responses causing elevations in cytokine profiles and that the diseases may be related in those predisposed [11]. In either case, clinicians should observe these children for development of other symptoms. Longer-term cardiology follow-up visits for KD patients may represent an ideal time to screen for autoimmune/autoinflammatory symptoms and referral when appropriate. The generally accepted practice in our institution is to order viral testing for patients being evaluated for KD. We recognize that among patients diagnosed with KD, respiratory viral detection is fairly common [12] and should not impact the decision to treat for KD, particularly for those with complete KD. However, viral characterization of adenovirus in children may be useful in distinguishing the 2 illnesses when coupled with the clinical presentation, especially for those being evaluated for incomplete KD [13,14]. In addition, viral testing is often obtained before an inpatient attending evaluation. We urge clinicians to use caution in interpretation of viral detection from the respiratory tract of patients being evaluated for KD; findings should be interpreted only in conjunction with the clinical and laboratory evaluation of KD.
Our findings are consistent with other recent US cohorts in that the majority of CAA in affected KD patients are present during the initial hospitalization, with prior reports ranging from 81% to 84% [15,16]. Before discharge, we identified children as high or low risk for development of CAA based on known risk factors [9,17]: ie, age (infants), those requiring intensive care, and those with nonresponse to initial IVIG treatment. For these high-risk patients and children with known CAA, follow up is 3-7 days after discharge. Among the 4 children with baseline normal echocardiograms who went on to develop CAA, all met at least 1 high-risk criterion.
CONCLUSIONS
Limitations of this study include that it was from a single center, the subsequent diagnoses were retrospectively identified, and the numbers of patients were small. Moreover, the tool in the electronic medical record required continued education of providers to increase compliance. Nonetheless, these data suggest that further study correlating the clinical KD phenotype on the day of presentation with clinical outcomes, including subsequent diagnoses of autoimmune/autoinflammatory disorders, should be considered. Prospective and standardized collection of KD criteria on the day of treatment together with clinical outcomes can provide useful information to further refine the KD diagnosis and the potential for risk of subsequent autoimmune/autoinflammatory conditions.
|
2018-04-03T05:32:33.890Z
|
2017-04-05T00:00:00.000
|
{
"year": 2017,
"sha1": "dd935e1392dc5e97fe50767c51e324e2578258bf",
"oa_license": null,
"oa_url": "https://academic.oup.com/jpids/article-pdf/7/2/119/24815916/pix016.pdf",
"oa_status": "BRONZE",
"pdf_src": "PubMedCentral",
"pdf_hash": "dd935e1392dc5e97fe50767c51e324e2578258bf",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
220885742
|
pes2o/s2orc
|
v3-fos-license
|
Niche and neutral processes leave distinct structural imprints on indirect interactions in mutualistic networks
Indirect interactions are central to ecological and evolutionary dynamics in pollination communities, yet we have little understanding about the processes determining patterns of indirect interactions, such as those between pollinators through shared flowering plants. Instead, research has concentrated on the processes responsible for direct interactions and whole-network structures. This is partly due to a lack of appropriate tools for characterising indirect interaction structures, because traditional network metrics discard much of this information. The recent development of tools for counting motifs (subnetworks depicting interactions between a small number of species) in bipartite networks enable detailed analysis of indirect interaction patterns. Here we generate plant-hummingbird pollination networks based on three major assembly processes – neutral effects (species interacting in proportion to abundance), morphological matching and phenological overlap – and evaluate the motifs associated with each one. We find that different processes produce networks with significantly different patterns of indirect interactions. Neutral effects tend to produce densely-connected motifs, with short indirect interaction chains, and motifs where many specialists interact indirectly through a single generalist. Conversely, niche-based processes (morphology and phenology) produced motifs with a core of interacting generalists, supported by peripheral specialists. These results have important implications for understanding the processes determining indirect interaction structures.
54
Species in a community are often influenced by other species they do not interact with directly 55 (Strauss, 1991;Wootton, 1994Wootton, , 2002. Such indirect interactions are a fundamental component 56 of communities, governing ecological and evolutionary processes as much as, or more than, For each of these sets of abundance, morphology and phenology data, we generated matrices Neutrality was simulated using an abundance matrix, A. Elements of A were the product of 117 each species' relative abundance. Thus, element aij represents the interaction probability 118 between plant species i and hummingbird species j and is equal to the product of the relative 119 abundances of i and j. This matrix therefore represents neutrality: the likelihood of species 120 interacting randomly in proportion to their abundance. 121 We create two morphological match matrices, corresponding to two different methods in the hummingbird bill length (Weinstein & Graham, 2017). This approach relaxes the assumption 132 that a hummingbird is equally likely to interact with all flowers that have a floral corolla equal 133 to or shorter than its bill, and makes morphological match a continuous, rather than binary, 134 quantity. If the difference between floral corolla depth and hummingbird bill length was 0, the 135 difference was set to the minimum non-zero difference between corolla depth and bill length 136 in the web to prevent errors when dividing by zero values.
137
Elements of the phenological overlap matrix, P, were calculated using matrix multiplication For each assembly process, and for each dataset, we generated 1000 binary interaction matrices 144 from the probability matrix using the 'mgen' function in the 'bipartite' R package (Dormann,145 Frund, Bluthgen, & Gruber, 2009). In total there were 96,000 binary matrices (1000 generated 146 matrices ´ four assembly processes ´ 24 sets of abundance, morphology and phenology data).
147
Generated matrices had the same connectance as their corresponding empirical matrices.
148
Characterising indirect interactions using motifs 149 We next characterised the different patterns of indirect interactions for each network and Macro-scale metrics, such as modularity, nestedness and connectance, summarise the structure of the whole 174 network. Micro-scale metrics, such as d´, degree or dependencies, characterise the structure of a single node.
175
Motifs sit between these two extremes, at the meso-scale, capturing local-scale patterns of indirect interactions 176 between species. The 'meso-scale' level shows the five types of motif that make up the macro-scale network.
177
Note that the network itself is a four-species motif and so, for this example, we only consider motifs with fewer 178 than four species (two-and three-species motifs). Importantly, motifs do not discard information about macro- pathways than the main set of core-peripheral motifs, and thus slightly stronger indirect effects.
230
As the generalists in these motifs might be able to buffer changes in each other's abundances, 231 it is likely that the generalists have a stronger effect on the specialist, than vice versa (Simmons,232 Cirtwill, et al., 2019). The specialist species' generalist partner has high levels of redundancy 233 in its interactions and thus may be a reliable partner for the specialist. However, asymmetric 234 complete motifs are likely less effective than core-peripheral motifs at curbing the spread of Statistical analysis 295 We used an ANOVA framework to assess statistical differences between the frequencies of 296 motifs in networks generated using different assembly processes. First, a MANOVA was used 297 with frequencies of all 16 motifs as dependent variables and assembly process as the 298 independent variable to determine whether there was an overall effect of assembly process on 299 motif frequency distribution. Then, to identify how assembly processes affect specific 300 dependent variables, we conducted univariate ANOVAs for each motif. For this, pairwise 301 comparisons between assembly processes were calculated using the 'multcomp' R package 302 (Hothorn, Bretz, & Westfall, 2008). Adjusted p-values were used to account for multiple 303 comparisons, using the 'single-step' method in 'multcomp'. with more occurrences of core-peripheral motifs (motifs 5, 10 and 14) (Figures 1b, 3 and 4). 311 Furthermore, some differences were observed between morphological matching and 312 phenological overlap matrices: phenological overlap matrices had significantly higher 313 frequencies of motif 9 (a core-peripheral motif) than morphological matching, but significantly 314 lower frequencies of motif 14 (another core-peripheral motif; Figures 1b, 3 and 4). asymmetric complete motifs 6, 11, 12 and 16) (Figures 3 and 4). Conversely, networks 344 produced assuming niche-based processes -those determined by morphology or phenology -345 contain more core-peripheral motifs, that comprise a core of interacting generalists, supported 346 by peripheral specialists (core-peripheral motifs 5, 10, 14) (Figures 3 and 4). This has important implications for whole-network dynamics, as it suggests that under neutral 379 processes, the average length of the interaction chain between any two species will be lower,
407
While there were few differences between different niche-based processes, networks based on 408 phenological overlap had significantly higher frequencies of motif 9 (a core-peripheral motif 409 with two generalists interacting) and significantly lower frequencies of motif 14 (a core-410 peripheral motif with three generalists interacting) than morphological matching models. This
|
2020-05-07T09:07:26.795Z
|
2020-05-02T00:00:00.000
|
{
"year": 2020,
"sha1": "717159c15ee93779274ee42cebd286b4a714f824",
"oa_license": "CCBY",
"oa_url": "https://www.biorxiv.org/content/biorxiv/early/2020/07/27/2020.04.30.070391.full.pdf",
"oa_status": "GREEN",
"pdf_src": "BioRxiv",
"pdf_hash": "7263d883a0de88e5b7bf96f4ea9d24884b6d5286",
"s2fieldsofstudy": [
"Environmental Science"
],
"extfieldsofstudy": [
"Biology"
]
}
|
267546730
|
pes2o/s2orc
|
v3-fos-license
|
A Dress Is Not a Yes: Towards an Indirect Mouse-Tracking Measure of Men’s Overreliance on Global Cues in the Context of Sexual Flirting
Assessing another person’s intention to flirt and, relatedly, their sexual interest is based on the interpretation and weighting of global (e.g., clothing style) and specific (e.g., facial expression) cues. Since cue incongruency increases the risk of erroneous judgments and thus can entail undesirable outcomes for both parties involved, detection of an individual propensity for overly relying on global (sexual) rather than specific (affective) cues is of social and clinical-forensic importance. Using a purpose-designed and pre-validated stimulus set, we developed a mouse-tracking task as an indirect behavioral measure for males’ overreliance on global cues (OGC) in the context of sexual flirting. In a convenience sample of heterosexual cisgender men (N = 79), experimentally induced sexual arousal was shown to increase the probability of OGC as a function of task difficulty (i.e., congruent or incongruent combinations of global and specific cues displayed by a potential female flirting partner). While error rate and reaction time proved to be indicators of OGC, the spatial measures maximum deviation and area under the curve provided less consistent results. In addition, error rate suggested sex drive and sexual objectification to act as moderators of the relationship between sexual arousal and OGC. Exploratory analysis further revealed a theoretically meaningful pattern of correlations between mouse-tracking measures and self-report measures of problematic (e.g., disinhibited, exploitative) sexuality. Implications of the results are discussed and a framework for differentiating potential causes of OGC (i.e., misperception, lack of self-control, and egocentric hedonism) is proposed. Supplementary Information The online version contains supplementary material available at 10.1007/s10508-023-02798-x.
Introduction
Sexual interest in a woman sparks a heterosexual man's desire to approach her (Welsch et al., 2020).Prior to a sexually motivated approach, it is advantageous for the man to assess whether the woman is more likely to reciprocate his sexual interest or not, i.e., whether she engages in flirting with him, as flirting is commonly understood to signal sexual availability (Buss, 1988). 1 In the initial phase of flirting, such assessment is usually based solely on visually discernible cues, since verbal exchange has not yet taken place.For the purpose of assessment, cues that are specifically directed at the male interpreter of these cues (e.g., facial expression, body language) appear to be particularly promising, in contrast to cues that-although perhaps perceived as sexually appealing-are not specifically directed at him (e.g., physical attractiveness, provocative clothing).
Overreliance on such nonspecific cues obviously increases the risk of misjudging a woman's (lack of) sexual interest toward the assessing male.As a result, sexually motivated approaches based on such nonspecific cues are more likely to have negative consequences for one or both parties involved, ranging from mild frustration to norm-violating (e.g., sexually offensive) behavior, depending on whether (and when) the man eventually recognizes the diverging interests and how he subsequently responds.Individual differences in the interpretation and weighting of cues in the context of flirting thus prove significant not only from a merely social but also from a clinical-forensic perspective, the latter especially in light of research showing that a subgroup of men particularly prone to perceive sexual intent in women is also more likely to engage in sexual coercion (Farris et al., 2008).
Global vs. Specific Flirting Cues
In the broad context of social encounters, nonverbal cues can be divided into cues that are (1) global, i.e., situationally stable (e.g., physical attractiveness, clothing style) and (2) specific, i.e., situationally variable (e.g., body language, facial expression).Global and specific cues differ in what information they carry for a particular (dyadic) social interaction, since specific cues by their nature pertain to that interaction only, whereas global cues do not, but at most refer to a general intention or mood.Failure to adequately account for this difference in informational value in the context of flirting increases the likelihood of overreliance on global cues (OGC), which means that inadequate (i.e., nonspecific) information is used to assess another person's current (lack of) sexual interest.For example, while in most public social contexts (e.g., when going for a night out) it would be appropriate to think of a woman's clothing as independent of her attitude toward a particular person, her facial expression should be understood as a cue directed specifically at the person with whom she is having eye contact at that very moment (Adams & Kleck, 2005).
Since facial expressions reliably distinguish between different situations and interaction partners, primary attention should be paid to them when assessing another person's intention during a social interaction.Especially if no verbal communication has taken place yet, a facial expression conveying positive affect can be understood as a possible invitation to further contact and approach (Miles, 2009).Using a mouse-tracking (MT) paradigm, Treat et al. (2020) showed that men at greater risk of exhibiting sexual aggression against women are more inclined to evaluate women's sexual interest based on inappropriate global rather than appropriate specific cues, i.e., based rather on normative attractiveness and provocative clothing than on the affect a woman displays.The correlations found by Treat et al. (2020) point to the also theoretically plausible link between individual propensity for OGC and an increased risk for problematic sexual behavior.
Sexual Arousal and Cue Relevance
Motivational (visceral) states affect decision-making and selection behavior in favor of need-congruent stimuli, i.e., stimuli that match the state of deprivation an individual is currently experiencing (e.g., food when feeling hungry; Seibt et al., 2006).As has been shown, this is also true for sexual arousal (Ariely & Loewenstein, 2006;Skakoon-Sparling et al., 2015), as sexual arousal increases the appeal of sexual stimuli by the prospect of sexual pleasure (Meston & Buss, 2007).As a result, sexually aroused individuals-especially men-are more prone to engage in sexually disinhibited behavior (Imhoff & Schmidt, 2014;Wiemer et al., 2023).This also applies to situations in which inhibition of sexual behavior would be socially appropriate (e.g., when sexual interest is not reciprocated), while acting out is sanctioned (Bouffard, 2011).In light of this, it seems obvious to investigate whether the "motivational myopia" (Ditto et al., 2006) created by sexual arousal "in the heat of the moment" (Ariely & Loewenstein, 2006) affects men's cuebased assessment of potential flirting partners, i.e., to what extent sexually aroused men tend to disregard a discrepancy between their own and a woman's sexual interest by focusing on global sexual rather than specific affective cues.Such an approach is further warranted by findings suggesting that a state of sexual arousal in men might prominently contribute to misunderstandings in sexual communication (Rerick et al., 2019).
Mouse-Tracking Selection Tasks
The premise underlying MT is that motor movements contain information about task-relevant cognitive processes occurring within the same time frame (Spivey & Dale, 2006).To access this information, MT continuously captures the mouse cursor movements toward a response option caused by participants' hand movements during a computer experiment (Freeman & Ambady, 2010a).Specifically, MT paradigms have been repeatedly used to measure the temporal unfolding of decision-based selection processes by asking participants either to choose between two competing stimuli (e.g., healthy vs. unhealthy food; Stillman et al., 2017) or to assign a single stimulus to one of two competing categories (e.g., feminine vs. masculine stereotypes; Freeman & Ambady, 2009).In both designs, participants are required to select one of two options and are thus subject to a (more or less pronounced) selection conflict.
Measures calculated by MT software (Freeman & Ambady, 2010b) are reaction time (RT), maximum deviation (MD), area under the curve (AUC), and error rate (ER).RT corresponds to the time participants need to move the mouse cursor from the starting field to the target.As has been frequently shown, response latency increases as a function of task difficulty (e.g., Hick, 1952).Consequently, a stronger conflict in the course of a selection task should be reflected in the time it takes participants to decide on one of the offered options.MD and AUC represent the spatial path of a mouse trajectory in terms of numerical parameters.While MD is defined as the point of an empirical trajectory farthest from an ideal straight line between the starting field and the target, AUC refers to the area between the ideal straight line and the empirical trajectory.High values for these two measures thus reflect an increased curvature of the mouse trajectory and point to a correspondingly high attraction of the nonselected option (thereby indicating a temptation or distraction of some kind emanating from this option).A normative error occurs when an option is not only partially headed for, but actually chosen, even though this choice contradicts a given instruction (e.g., when ice cream is chosen instead of broccoli, even though the instruction calls for making the choice in accordance with a healthy diet).Smith et al. (2018) used an MT paradigm to examine the real-time dynamics of men's decisions about women's sexual interest by presenting full-body images of women to be assigned to one of two categories ("sexually interested" or "rejecting").The depicted women varied in affect (sexual interest or rejection, i.e., a specific cue), clothing style (provocative or conservative, i.e., a global cue), and normative attractiveness (attractive or unattractive, i.e., another global cue).The authors' conjecture was that normatively incongruent cue combinations would lead to increased selection difficulty.While overall participants' assessments were normatively correct (i.e., consistent with the specific cue), for both global cues the expected congruency effect was found, as more errors occurred when specific and global cues were presented incongruently (e.g., when the woman showed rejection but at the same time wore provocative clothing).In addition to assessment accuracy, Smith et al. (2018) tracked mouse movements to examine how easy or difficult it was for participants to choose between the two categorical options.As expected, the mouse trajectories in incongruent trials were found to converge more closely to the normatively incorrect option than those in congruent trials, suggesting increased selection difficulty.Thus, a congruency effect was found not only for the selection outcome in terms of accuracy, but also for the selection process manifested in mouse trajectories.
Sex Drive and Sexual Objectification as Potential Moderators
Sex drive refers to the strength of dispositional sexual motivation (Baumeister et al., 2001).Individual differences in sex drive are evident in the frequency of sexual desire, the variety of (desired or performed) sexual acts, the number of (desired or real) sexual partners, the frequency of sexual fantasies and masturbation, and the willingness to make sacrifices in other areas of life for the sake of sex (Baumeister et al., 2001).Since Moholy et al. (2015) demonstrated that individuals with high sex drive are less successful in controlling their impulse-driven behaviors in a state of sexual arousal, it seems justified to speculate that such individuals may also be more prone to overly rely on global sexual cues, i.e., cues that are congruent with their sexual desire.
Sexual objectification is a form of perception or action in which a person is regarded or used as a mere object of sexual desire or as a sexual resource by reducing them to their body or to specific sexual functions of their body (Bartky, 1990).Sexual objectification also makes the sexually desired person appear more attainable by denying them relevant needs and interests of their own (e.g., Vaes et al., 2011), especially since these needs and interests may constitute an obstacle to the objectifier's intentions.Accordingly, if sexual objectification leads to the subjective perception that one's own sexual interest is not countered by an opposing interest, or that such an opposing interest is at least irrelevant, OGC should become more likely, as in this case, there is no motivation to pay attention to cues that might indicate such an opposing interest (i.e., specific affective cues).As has been demonstrated, sexual objectification can be provoked or enhanced by presenting images of women in sexualized clothing (e.g., Gurung & Chrouser, 2007).In the current study, a woman wearing a sexually appealing outfit was shown as one of two stimuli in each trial.This choice of visual stimuli is also in line with research findings that an objectifying gaze (the "male gaze"; Gervais et al., 2013) is the most common expression of sexual objectification in everyday life (Holland et al., 2016).
Current Study
Previous studies were successful in using MT to reveal individual difficulties in solving selection tasks (e.g., Lim et al., 2018;Smith et al., 2018).Thus, it seemed reasonable to assume that an MT selection task would also be a suitable method for measuring individual propensity for OGC in a state of sexual arousal, provided that (non)compliance with an instruction given to solve the MT task would be reflected in behavioral correlates of OGC (i.e., in the four MT measures RT, MD, AUC, and ER).Normative errors in particular as well as an altered RT, but also more curved mouse trajectories pointing to selection difficulties, could then be interpreted as indicators of an increased individual risk for real-life occurrence of OGC and its possible negative consequences.Such a conclusion would be even more warranted if (1) constructs of socially inappropriate or dysfunctional (e.g., disinhibited, exploitative) sexuality could be found that moderate the relationship between sexual arousal and OGC as indicated by MT and if (2) indicators of OGC were positively correlated with self-reports of problematic sexual attitudes, behaviors, experiences, and motivations (i.e., if they shared a nomological network with traits related to problematic sexuality). 2uilding on the findings mentioned above, we extended earlier approaches (e.g., Smith et al., 2018;Treat et al., 2020) by (1) designing two different trial types, with each trial type requiring a choice between two competing stimuli depicting the same woman displaying different cue combinations (see Stimuli and Trial Types section for details) and by (2) introducing repeated measurements (about 20 min apart).We expected the incongruent trials to cause more pronounced selection difficulties than the congruent trials, i.e., that a congruency effect would occur for both the selection outcome and the selection process.In accordance with their presumed different potential to elicit selection conflicts, we referred to the incongruent trial condition as the high conflict (HC) condition and the congruent trial condition as the low conflict (LC) condition.Additionally, sexual arousal was induced between T1 and T2, such that the baseline condition at T1 was completed in a "cold" visceral state, whereas the putative effect of "hot" states on decision-making (Ariely & Loewenstein, 2006;Skakoon-Sparling et al., 2015), i.e., on the selection of potential flirting partners, could be observed at T2 and relative to T1.While we expected selection in line with the given instruction to be generally more difficult in a state of sexual arousal at T2 than in the baseline condition at T1, we also expected this difficulty to increase further when, in a state of sexual arousal, the given instruction would conflict with a spontanous preference for need-congruent (global) sexual stimuli, i.e., when the incongruent cue combinations displayed in HC trials would interact with a state of sexual arousal.
As potential moderators of the hypothesized relationship between sexual arousal and OGC, we chose two constructs based on both theoretical considerations and previous research (i.e., sex drive and sexual objectification).Furthermore, to test our conjecture of a nomological network of OGC and aspects of problematic sexuality, we had several standard self-report questionnaires answered and correlated their scale scores with the MT measures, expecting a theoretically meaningful pattern of correlations.In addition to these established questionnaires, we had participants answer a self-constructed questionnaire on their propensity for socially problematic (e.g., manipulative) flirting behavior (see Measures Section for details on all questionnaires).We made this particular addition since we also chose a flirting framework for our selection task to draw on an everyday realworld experience often accompanied by a sexual objective (Henningsen, 2004;Moore, 2010).Moreover, flirting usually takes place dyadically, so prioritizing specific affective over global sexual cues should provide a favorable (i.e., socially appropriate as well as more promising) basis for selection.The instruction for the flirting task (i.e., "Click on the woman who is more likely to flirt with you right now") was designed to focus on the perceived sexual interest (present or absent) of the depicted women, whereas participants' own sexual interest was not addressed and should therefore be irrelevant to the decision.
Hypotheses
The above research findings and the considerations derived from them led us to the following hypotheses: 1. HC trials will differ from LC trials by (1) increased RT, (2) increased MD, (3) increased AUC, and (4) increased ER (i.e., main effect of trial type).2. Sexual arousal at T2 will result in (1) increased RT, (2) increased MD, (3) increased AUC, and (4) increased ER compared to T1 (i.e., main effect of time).
The influence of sexual arousal as formulated in
Hypothesis 2 will be more evident in HC trials than in LC trials (i.e., interaction of trial type and time).4. Higher sex drive will be associated with a stronger manifestation of the interactional relationships as formulated in Hypothesis 3. 5.A higher propensity for sexual objectification will be associated with a stronger manifestation of the interactional relationships as formulated in Hypothesis 3.
Participants
Initially, the convenience sample comprised 89 adult cisgender males recruited through flyers, posters, and email distribution lists within a medium-sized German college town, including recruitment on campus and at local soccer clubs.
Participants were required to self-identify as exclusively or predominantly heterosexual, thereby indicating a preference for women as potential sexual partners.Nine participants had to be excluded from further analyses after evaluating their sexual orientation and manipulation checks.Of these, three reported values > 2 on the Kinsey scale (Kinsey et al., 1948), so that for them the required sexual attraction of the female stimuli was not ensured.Seven participants stated that they had not been sexually aroused at all by the erotic audio story used to induce sexual arousal between T1 and T2.One of the nine excluded participants met both exclusion criteria.All remaining participants in the effective sample (N = 79) stated that they had answered the questionnaires honestly. 3he majority (70%) of the effective sample consisted of university students from various disciplines.Participants' age ranged from 18 to 50 years (M = 25.47,SD = 5.62).Most participants (68%) classified themselves as exclusively heterosexual (level 0 on the Kinsey scale), while the rest stated that occasionally they also had homosexual fantasies or contacts (24% level 1, 8% level 2).About half (51%) were in a committed relationship.The highest educational attainment reported by 44% was a university degree, whereas 51% reported the general higher education entrance qualification ("Abitur") and 5% the secondary school leaving certificate ("Mittlerer Schulabschluss").Participants either had normal vision or wore visual aids.All were right-handed or could easily operate a mouse with their right hand.As an incentive, participants could choose between € 5 (about US$ 5.5; chosen 39 times) and course credit (chosen 40 times).
Stimuli and Trial Types
Prior to the present study, 378 pictorial stimuli (i.e., fullbody photographs of adult women in both casual and rather revealing clothing with different facial expressions) were created and validated (for detailed information on the full stimulus set, see Landwehr & Mundloch, 2023).For the MT paradigm introduced here, a subset of these stimuli consisting of 192 images was used and divided into two parallel sets (each stimulus was presented only once).These parallel sets were matched on mean and distributed sexual attractiveness values (for more details on the stimuli used in this study see the Electronic Supplement section ES1).The pictorial stimuli had to meet certain criteria such as (1) high authenticity in terms of both individually chosen clothing (models brought their own clothes which they would actually wear in situations similar to the task's conditions, namely either when "visiting a good friend in casual clothing to hang out together" or "having a night out in sexy clothing with the general intention to flirt") and affective facial expressions matching two to-be-imagined social situations (in which models should either flirt with a person they are sexually interested in or unequivocally reject an obnoxious person as a flirting partner), (2) a clear contrasting effect per model between the two outfits (Casual vs. Sexy) as well as the two facial expressions (Flirting vs. Rejecting), and (3) standardization of various external factors, such as lighting and final image editing.Since Van den Stock et al. (2007) demonstrated that body language greatly influences the perception of facially communicated emotion, we (4) asked our models to maintain a standardized neutral posture regardless of their respective facial expression.In this way, we ensured that potential effects could be clearly attributed to variation in global (clothing) and specific (facial expression) cue combinations.
In contrast to Smith et al. (2018), we used an MT paradigm in which participants did not assign single stimuli to one of two categories but had to select one of two stimuli presented simultaneously, i.e., two images of the same woman.Importantly, this design allowed us to avoid a confounding effect of differences in participants' subjective perceptions of different women's attractiveness.Also, because the differences between the two images were entirely due to the clothing and facial expression cues displayed by the same woman, we expected the selection task to be generally quite easy to solve, emphasizing the diagnostic relevance of any instruction-inconsistent behaviors (i.e., indicators of OGC) that might manifest in the four MT measures.
In the HC condition, one of the images showed a woman in casual everyday clothing with a flirtatious facial expression (the Casual × Flirting combination), while the other image showed her in rather sexy clothing with a rejecting facial expression (the Sexy × Rejecting combination).This juxtaposition was based on the assumption that both a flirtatious facial expression and a sexually suggestive outfit are cues that men perceive as (sexually) appealing and thus facilitate male sexual selection (e.g., Grammer et al., 2004;Walsh & Hewitt, 1985).At the same time, it was reasonable to assume that both a rejecting facial expression and inconspicuous clothing would not be perceived as equally positive.Instead, these cues should be perceived as rather neutral (in the case of the casual outfit) or even as negative (in the case of the rejecting facial expression), which is why male sexual selection should not be facilitated by them and might even be inhibited (e.g., Goodboy & Brann, 2010;Moore, 1998).Since in the HC condition both images were incongruent, i.e., both contained a positive and a non-positive cue, a prompt to choose between them could be expected to evoke a selection conflict of some degree.
We intermixed the HC trials with trials of the LC condition.In the LC condition, one of the images showed a woman with both positive cues (the Sexy × Flirting combination), while the other image showed her with both non-positive cues (the Casual × Rejecting combination).When confronted with two such juxtaposed congruent images, the participants' choice was expected to fall clearly on the image showing the Sexy × Flirting combination, as this combination would correspond to both the more sexually appealing (clothing) option and the affective (facial expression) option normatively consistent with the given instruction to select "the woman who is more likely to flirt with you right now." Fig. 1 shows the HC and LC conditions along with their respective cue combinations.
Using Krippendorff's α coefficient (Hayes & Krippendorff, 2007), interrater reliability for the overall discriminability of the two clothing cues, Casual and Sexy, as well as the two facial expression cues, Flirting and Rejecting, was determined.While reliability was high for the clothing cues (r α = .82)and very high for the facial expression cues (r α = .93),most importantly, for all images-which were presented in the same pairwise combinations as they would later form the HC and LC trials of the experiment-the cues contained were assigned to the respective normatively correct category (i.e., Casual or Sexy, Flirting or Rejecting) by the majority of raters.Also, as expected, images of the Flirting condition were rated as more sexually attractive than images of the Rejecting condition (p < .001,d z = 1.32), while images of the Sexy condition were rated as more sexually attractive than images of the Casual condition (p < .001,d z = 0.94).Crucially, the expected group difference in perceived sexual attractiveness was particularly pronounced between the two LC cue combinations (1) Sexy × Flirting and (2) Casual × Rejecting (p < .001,d z = 2.04), while, as anticipated, no group difference emerged between the two HC cue combinations (1) Casual × Flirting and (2) Sexy × Rejecting (p = .197,d z = 0.12).These pretest results clearly corroborated the paradigm's fundamental premises in terms of either distinct or analogous attractiveness-related stimulus effects (see ES1 for detailed information).
Measures
We employed seven self-report questionnaires assessing problematic sexuality, presented sequentially in a fixed order to maximize between-subjects effects.To avoid participant fatigue, shortened versions were used if available.Acceptance of rape myths was assessed using the Acceptance of Modern Myths about Sexual Aggression Scale (AMMSA; Gerger et al., 2007; German short version Süssenbach & Bohner, 2011; nine five-point items from 1 = does not apply at all to 5 = totally applies).Rape myths are descriptive or prescriptive beliefs about rape (i.e., about its causes, context, consequences, perpetrators, victims, and their interactions) that serve to deny, trivialize, or justify men's sexual violence against women (Gerger et al., 2007).The Hypersexual Behavior Inventory-19 (HBI-19;Reid et al., 2011;German version Klein et al., 2013; eight five-point items from 1 = does not apply at all to 5 = totally applies) assesses hypersexual urges, thoughts, fantasies, and behavior associated with clinically relevant distress or functional impairment in important life domains.The Sexual Desire Inventory-2 Fig. 1 Trial types and cue combinations (SDI-2; Spector et al., 1996; German short version Kuhn et al., 2014; ten eight-point items, two items from 1 = never to 8 = more than once a day, five items from 1 = no desire to 8 = strong desire, two items from 1 = much less desire to 8 = much more desire, one item from 1 = unimportant to 8 = extremely important) measures sexual desire based on two subscales (Dyadic Sexual Desire and Solitary Sexual Desire) that can be combined into a total score.To assess individual differences in sexual inhibition and excitation, we used the Sexual Inhibition/Sexual Excitation Scales (SIS/ SES; Janssen et al., 2002;German short version Turner et al., 2014; SIS: eight five-point items from 1 = does not apply at all to 5 = totally applies; SES: six five-point items from 1 = does not apply at all to 5 = totally applies).The revised Sociosexual Orientation Inventory (SOI-R; Penke & Asendorpf, 2008; nine five-point items, three items from 1 = 0 to 5 = 8 or more, three items from 1 = I do not agree at all to 5 = I agree completely, three items from 1 = never to 5 = almost every day) measures individual preferences for impersonal sex and casual sexual encounters.Since SDI-2, SES, and SOI-R represent different (yet usually highly intercorrelated) facets of sex drive, an aggregated index of their z-standardized values was calculated to measure sex drive on a higher level of aggregation in the present study.The Sexual Objectification of Others Inventory (SOOI; Anslinger, 2019; ten five-point items from 1 = strongly disagree to 5 = strongly agree) was used to assess the propensity to perceive women primarily in terms of their sexual appeal and functionality based on two factors (Instrumental Objectification and Visual Objectification). 4articipants also answered the Socially Undesirable Sexual Selection Scale (SUSS; Schmidt, 2023; 13 items to be answered on a visual analogue scale from 0 = very unlikely to 100 = most likely).This unidimensional scale was developed to capture indications of manipulative flirting behavior.We assumed that willingness to use manipulative flirting tactics in particular would be associated with an increased likelihood of OGC, as both phenomena stem from an egocentric selection criterion, i.e., the sexual interest of the selecting male, as opposed to an assessment of the woman's interest (example item: "I am willing to deceive the woman so that she will have sex with me"; see ES2 for validity details).
Following T2, participants answered three manipulation check questions concerning (1) the perceived extent of sexual arousal evoked by the erotic audio story (i.e., Arousal Story ; "How sexually arousing did you find the story you heard?" from 1 = not arousing at all to 5 = very arousing), (2) the perceived extent of current sexual arousal (i.e., Arousal Now ; "How sexually aroused are you right now?" from 1 = not aroused at all to 5 = very aroused), and (3) their honesty in answering the questionnaires ("How honestly have you answered the questions during this study?"from 1 = not honest at all to 5 = completely honest).
For the MT task, RT, MD, AUC, and ER aggregated per trial type and measuring time were used as dependent variables.An initial preparation of these data was carried out with the software MouseTracker (Freeman & Ambady, 2010b).Each recorded trajectory was normalized and divided into 101 time steps.
Apparatus and Procedure
The experiment was conducted using a custom laptop with a 17-inch screen set to a resolution of 1366 × 768 pixels, standard headphones, and a standard optical mouse with a resolution of 800 DPI connected via USB cable.The software MouseTracker (Freeman & Ambady, 2010b) was used to record RT, MD, AUC, and ER.The software Inquisit 5 Lab (Millisecond, 2016) was used for presenting the questionnaires and recording participants' answers.
First, participants provided consent to take part in a study on flirting behavior.They were informed that they would be listening to an erotic audio story and answering various questions about their sexuality.They were also assured that all data would be collected anonymously, and that they could withdraw from the study at any time without any disadvantages.They then read the description of the MT task which explained that they should move the mouse as fast as possible while answering as correctly as possible according to the given instruction (i.e., "Click on the woman who is more likely to flirt with you right now").T1 started with four practice trials (unbeknownst to the participants and excluded from the analyses) after which 48 trials followed in a fixedrandomized order to maximize between-subjects effects.Of these trials, 24 belonged to the HC condition and 24 to the LC condition.Following T1, participants completed seven questionnaires (see Measures section for details).Next, they were asked to put on headphones and listen to a 332 s audio story of explicit erotic content used to induce sexual arousal (taken from Imhoff & Schmidt, 2014, Study 1).Subsequently, T2 followed with two practice trials and, again, 48 regular trials.Finally, participants answered three manipulation check items on state sexual arousal and honesty (see Measures section for details).The extent of sexual arousal evoked by the erotic audio story was assessed only after T2 and not immediately after the audio story (1) to prevent an interruption/a distraction between the induction of sexual arousal and the onset of T2 by a question whose answer requires cognitive processing and thus could diminish the extent of sexual arousal and (2) to avoid overly emphasizing the significance of sexual arousal for the subsequent task, so as to preempt altered (e.g., inhibited) response behavior.
At the beginning of each MT trial, participants had to click the start button, only after which the two images appeared in the upper left and right corners of the screen (see Fig. 2).Now, participants had to move the mouse within a 1000 ms response window, creating a dynamic starting condition which proved to be methodologically advantageous by enhancing the transfer of decisional processing into mouse movements (Scherbaum & Kieslich, 2017).Participants next had to choose one of the two presented images within a 5000 ms response window.Both response windows were chosen rather liberally (Hehman et al., 2014).If the start button or an image had not been clicked within the respective response window, a prompt in red letters appeared reminding participants to move the mouse faster.A shortened version of the instruction (i.e., "Who's flirting?")was displayed above the start button before and during each trial to remind participants of the task at hand.
Statistical Analyses
Trials of both conditions in which participants took too long to initially move the mouse (> 1000 ms) or to choose one of the two images (> 5000 ms) were discarded from the analyses (0.7% of all trials).Prior to conducting the 2 (Trial Type: HC vs. LC) × 2 (Time: T1 vs. T2) repeated measures ANOVAs, the data were screened for outliers.One data set proved to be a clear outlier according to the Tukey criterion (> 3 interquartile ranges) and was thus excluded from further analyses.The significance level for all calculations was set to 5%.Depending on the statistical method, R 2 , η p 2 , d z , and ß are reported as effect sizes.
For the moderation and correlation analyses conducted in this study, MT measures were first converted into trial condition-based difference scores for T1 and T2, respectively (e.g., ΔRT T1 = RT T1_HC − RT T1_LC ).While fluctuations in RT, erratic deviations in mouse movements, and unsystematic errors should occur in both trial conditions, systematic effects were expected to occur only due to the more demanding selection task in the HC condition.These difference scores therefore represent adjusted measures of selection difficulty solely attributable to the incongruent presentation of global and specific cues, and can thus be interpreted as indicative of individual differences in OGC.The presumed moderating effect of (1) sex drive and (2) sexual objectification on the relationship between sexual arousal and OGC was tested by moderated hierarchical regression analyses for each dependent variable with (1) sex drive (in form of the aggregated index) and (2) sexual objectification (in form of the SOOI score) as moderators.Two corresponding MT difference scores from T1 and T2 were used as predictor and criterion for each regression (e.g., ΔMD T1 as predictor and ΔMD T2 as criterion).
Mean Differences
First, in order to test the internal validity of the MT paradigm, a 2 (Trial Type: HC vs. LC) × 2 (Time: T1 vs. T2) repeated measures ANOVA was conducted for each of the MT measures.Crucially, Table 1 indicates a significant main effect of trial type on the four MT measures.As expected, all means were higher in the HC condition than in the LC condition.In the HC condition, participants thus took longer to select an option, were more prone to normative errors, and their mouse trajectories showed greater curvatures than in the LC condition, supporting Hypotheses 1a-d.
A significant main effect of time (Table 1) emerged for RT and ER, but not for MD and AUC, so that hypotheses 2b and 2c had to be rejected.A comparison of the means further revealed that RT unexpectedly decreased from T1 to T2, which is why Hypothesis 2a had to be rejected as well.However, in line with Hypothesis 2d, ER increased significantly from T1 to T2.
Interaction of trial type and time (Table 1) revealed no significant effect for the spatial MT measures MD and AUC, so Hypotheses 3b and 3c were rejected.Moreover, although an interaction was observed for RT, this effect was not in line with the direction of Hypothesis 3a, as instead of an increase, RT decreased from T1 to T2, even more so in the HC (p < .001,d z = − 0.42) than in the LC (p < .001,d z = − 0.32) condition.In accordance with Hypothesis 3d, however, an interaction was found for ER: While in the LC condition participants made an equal number of errors at T1 and T2 (p = .964,d z = 0.01), in the HC condition ER increased significantly from T1 to T2 (p < .001,d z = 0.46).Error frequencies as a function of trial type and time are shown in Fig. 3.
Sex Drive and Sexual Objectification as Moderators
Since the mean differences reported above provided preliminary support for the validity of the MT paradigm, as a next step, we conducted moderated hierarchical regression analyses to investigate two potential moderators of the relationship between sexual arousal and OGC.MT difference scores were entered in separate hierarchical regression analyses with (1) sex drive and (2) sexual objectification as moderators.As it turned out, the relationship between sexual arousal and OGC was significantly moderated by both sex drive and sexual objectification only when OGC was operationalized by ER, but not when operationalized by RT, MD, or AUC (see Table 2).Thus, while Hypotheses 4a-c had to be rejected, Hypothesis 4d was supported since in a state of sexual arousal, participants with higher sex drive made more errors than participants with lower sex drive.Likewise, hypotheses 5a-c had to be rejected, while Hypothesis 5d was supported: In a state of sexual arousal, participants with a higher propensity to sexually objectify women were more prone to errors indicative of OGC than participants with a lower propensity to do so.See Figs. 4 and 5 for graphs of the moderation effects.
Intercorrelations Among Mouse-Tracking Measures and Self-Report Measures
Table 3 shows significant correlations between MT measures at T1 and T2 within the same trial condition (e.g., between AUC HC_T1 and AUC HC_T2 ) that demonstrate the reliability (i.e., rank order stability) of the MT paradigm.Furthermore, Table 3 shows the intercorrelations of MT measures at T1 and T2, respectively.In both trial conditions, the high correlations between MD and AUC illustrate the close relationship between these spatial measures.Regardless of the trial condition, RT did not show any significant correlations with other measures at the same measuring time, at least indicating an independent explanatory power of RT.The intercorrelations of the scales (AMMSA, HBI-19, SDI-2, SIS/SES, SOI-R, SOOI, and SUSS) as well as their means, standard deviations, and internal consistencies (αs ≥ .72)can be found in Table 4.As expected, all scales were meaningfully intercorrelated, justifying post hoc the formation of an aggregated sex drive index consisting of SDI-2, SES, and SOI-R to be used as a single moderator.Regarding the two measures of state sexual arousal, descriptively both Arousal Story and Arousal Now correlated positively with all scales (rs ≥ .19;see Table 4), indicating a plausible link between dispositional sexual arousability and individual differences in problematic sexuality as measured by the questionnaires.As expected, the two self-report measures of state sexual arousal correlated substantially.Participant age was not related to any self-report measure, nor to the MT difference scores (ps ≥ .191).Likewise, participant relationship status was not related to the MT difference scores (ps ≥ .284).Descriptively, however, relationship status did exhibit a consistently negative correlation pattern with all assessed self-report measures (rs ≤ − .11;see Table 4), echoing prior findings suggesting that men's current relationship status contributes to predicting their overall perception of women's sexual exploitability (Lewis et al., 2012).
Nomological Network of Overreliance on Global Cues and Problematic Sexuality
In order to test whether there is a nomological network of OGC as assessed by MT measures and common aspects of problematic sexuality as assessed by self-report measures, correlations of these self-report measures with the MT difference scores were examined (Table 5).Since individual differences in problematic sexuality can be considered general correlates of individual differences in appropriate cue interpretation and weighting when assessing another person's sexual interest, a pattern of positive correlations between MT measures and self-report measures would also indicate convergent validity of the present paradigm.
Descriptively, all MT difference scores except for ΔMD T2 showed theoretically meaningful positive correlations with the self-report measures, with most effect sizes ranging between rs of .1 and .3.The most pronounced convergent validity pattern was found for ΔER (both at T1 and T2), whereas the highest correlations were found with the SOOI.Notably, at T1 a much stronger correlational pattern with 41% of the possible correlations being statistically significant was produced (as opposed to only 20% at T2). Mirroring the pattern of convergent validities, the internal consistencies of the MT difference scores (Table 5) were good only for ΔER, whereas the other MT difference scores turned out to be largely unreliable (a common finding for difference scores; Thomas & Zumbo, 2012).Self-reported state sexual arousal at the end of the study (Arousal Now ) was positively associated with almost every single MT difference score regardless of measuring time.
Effect of Incongruent Global and Specific Cues
Our first hypothesis stated that there would be differences in RT, MD, AUC, and ER between the two trial conditions HC and LC.In line with their designation, incongruent HC trials were expected to cause some degree of selection difficulty, whereas in congruent LC trials, selection was expected to be rather facilitated.Remarkably, all four MT measures actually differed between the two trial conditions in the predicted manner.Consistent with the results by Smith et al. (2018), a congruency effect was found for both the selection process and the selection outcome, as RT, MD, AUC, and ER were higher in the HC condition, i.e., participants took longer to select an option, mouse trajectories showed more curvature, and more errors were made than in the LC condition.From these findings, it can be tentatively concluded that in situations in which global sexual and specific affective cues co-occur incongruently, men are generally more prone to OGC, even though these two cue types should be regarded as of entirely different informational value for assessing a woman's current sexual interest in the context of a dyadic social encounter.It seems plausible that the difficulty in appropriate cue interpretation and weighting in the presence of incongruent global and specific cues represents a ubiquitous effect that occurs in various social settings.This difficulty is also likely to increase further when a global and a specific cue belong to the same category (e.g., affective expression).Saal et al. (1989) demonstrated that in business and academic settings, men tend to overly rely on the global cue of "professional friendliness" when displayed by women, misinterpreting it as (1) being specifically (i.e., selectively) directed at a particular male interaction partner and (2) driven by sexual interest, even if such an interpretation is not supported by context and conversational content (i.e., by specific verbal cues).Given these findings, the general connection between a heightened likelihood of OGC and an elevated risk of sexual harassment becomes notably evident.
Effect of Sexual Arousal
Our second hypothesis focused on the effect of sexual arousal on OGC.We presumed that in a state of sexual arousal, OGC should be more likely to occur as a result of motivational states affecting decision-making and selection behavior in favor of need-congruent stimuli (e.g., Seibt et al., 2006).As the results show, sexual arousal led to the predicted increase in ER, but not in RT, which instead decreased significantly from T1 to T2 (an unexpected finding reported similarly by Benbouriche et al., Obviously, mouse movements did not differ spatially between the two measurements, but were executed faster at T2, which at first glance suggests a practice effect and thus a performance improvement.This interpretation, however, is contradicted by the increase in ER, which indicates that the induced state of sexual arousal did, in fact, make it more difficult to choose the normatively correct option, thus also indicating an increase in OGC.For this reason, the decrease in RT might rather be seen as the result of a more disinhibited-and therefore more error-prone-decision behavior in a state of sexual arousal as demonstrated in previous studies (Imhoff & Schmidt, 2014;Wiemer et al., 2023).
Interaction of Cue Combination and Sexual Arousal
Hypothesis 3 stated an interaction of trial type and time such that the effect of sexual arousal should be more pronounced in HC trials than in LC trials.Analogous to the result pattern of Hypothesis 2, sexual arousal indeed led to a greater increase in ER in the HC condition than in the LC condition, whereas MD and AUC did not differ over time between the two trial types.Contrary to our expectation, RT decreased from T1 to T2 in both trial conditions, with this decrease being even higher in the HC condition.Based on the particularly expressive ER measure, it was thus possible to show that sexual arousal increases the difficulty of making normatively appropriate choices when confronted with incongruent global sexual and specific affective cues in the context of sexual flirting.From this, it can be concluded that apparently for some men, their own visceral/motivational state-and relatedly, their own sexual interest-rather than the information-based assessment of another person's sexual interest, is the determining factor for sexual selection, which translates into a higher risk for OGC and its detrimental consequences in real social situations.The significant increase in ER from T1 to T2 observed in the HC condition but not in the LC condition also emphasizes the diagnostic importance of normative errors for the MT paradigm.Contrary to our implicit expectation that differences both between trial types and over time would rather be revealed by the more subtle measures RT, MD, and AUC, the particularly informative ER measure turned out to be the clearest indicator in this respect.This seems encouraging from a methodological point of view, especially as this result suggests that the selection task described above is not susceptible to social desirability bias.While the large relative increase in errors in the sexual arousal condition compared to the baseline condition supports the basic theoretical assumptions of the paradigm and can also be seen as additional evidence for the known disinhibitory effect of sexual arousal (Imhoff & Schmidt, 2014;Wiemer et al., 2023), the low absolute numbers of errors are indicative for the selection task being both comprehensible and well manageable for nonclinical participants.At the same time, the distribution of individual errors in the HC condition at T2 (Fig. 3) points to a subgroup of participants in whom sexual arousal led to a relatively high number of errors (i.e., ≥ 6, indicating that errors occurred in at least 25% of the HC trials at T2). 1 3 Such a data-driven narrowing of the sample, combined with the pattern of positive correlations between ER and various self-report measures of problematic sexuality, speaks to the forensic-diagnostic potential of the present paradigm.
Impact of Sex Drive and Sexual Objectification
Hypotheses 4 and 5 stated that both dispositional sex drive and a propensity for sexual objectification would moderate the relationship between sexual arousal and OGC.According to our reasoning, higher sex drive should be associated with higher MT measures, as should a stronger propensity to perceive women primarily in terms of their "sexual usefulness" (Loughnan & Pacilli, 2014).
For both potential moderators, although the hypothesized effect could not be confirmed by RT, MD, and AUC, it could be confirmed by ER.Given research findings on the link between men's sexual objectification of women and an increased likelihood of physical and sexual aggression against women (e.g., Gervais et al., 2014;Vasquez et al., 2017), this result further emphasizes the forensic-diagnostic potential of the MT paradigm.The same applies to the revealed negative impact of high sex drive on appropriate cue differentiation in sexual partner selection, which is consistent with previous findings (Kafka, 2003;Moholy et al., 2015), especially as sex drive is a well-established risk factor for sexual offending (Seto, 2019).
Association Between Overreliance on Global Cues and Problematic Sexuality
In addition to our hypotheses, we explored our conjecture of a nomological network of OGC and various aspects of problematic sexuality.To this end, Pearson correlations of the MT difference scores with self-report measures assessing problematic sexual attitudes, behaviors, experiences, and motivations were conducted, the latter of which can be considered general correlates of difficulties in appropriate cue interpretation and weighting when assessing another person's sexual interest.On a descriptive level, all MT difference scores except for ΔMD T2 showed theoretically meaningful positive correlations with most of the self-report measures applied in this study (i.e., AMMSA, HBI-19, SDI-2, SIS/ SES, SOI-R, SOOI, and SUSS), thereby indicating convergent validity of the present paradigm.The strongest convergent validity pattern at both measuring times was found for ΔER, which included the two measures of state sexual arousal (Arousal Story , Arousal Now ), while also showing good internal consistencies.Descriptively, both measures of state sexual arousal correlated positively with all self-report measures and all MT measures, thus pointing to the importance of dispositional sexual arousability in the context of both problematic sexuality and OGC.Taken together, these correlational patterns confirm a nomological network of various aspects of problematic sexuality, higher sexual arousability, and OGC.From a methodological perspective, these distinct patterns can be considered a strong safeguard against artificial or coincidental single correlations driven by outliers or spurious effects.Surprisingly, all meaningful correlations between MT difference scores and self-report measures could already be found at T1, suggesting that a single measurement might be sufficient from a diagnostic point of view.Apart from that, this finding highlights the usefulness of difference scores as indicators of OGC as conceptualized in the current study.
Limitations
The findings presented in this study are inevitably subject to some limitations.First, the comparatively small sample size limits the interpretability of the results.At the same time, the within subjects-design used is superior to a between-subjects design in terms of power under the assumption of normally distributed data (Maxwell & Delaney, 2004), so preliminary conclusions can certainly be drawn from the results that warrant further research.Nevertheless, a future increase in sample size would be necessary to potentially support the findings of the present study with higher statistical power, which is especially true for the moderating effects.
Another potential drawback of our study could be seen in the lack of a (non-aroused) control group at T2, which would have served to control for possible learning effects or test motivation effects.However, we minimized the risk of overlooking such effects by the 2 × 2 design of our paradigm: While a learning effect is ruled out by the performance decline observed in the HC trials between T1 and T2 (as evidenced by the increase in ER), a fatigue or test motivation effect is contradicted by the fact that the same number of errors were made in the LC trials at T1 and T2, respectively, strongly indicating that there was no general response bias on the part of the participants.Thus, the design of our experiment rendered a control group largely unnecessary.
We used a convenience sample according to the WEIRD criteria established by Henrich et al. (2010), i.e., participants came from a population that can be characterized as western, educated, industrialized, rich, and democratic, with nearly 70% being university students.While such sample composition usually implies a lack of generalizability of results (Hanel & Vione, 2016), it can be considered methodologically promising in the case of the present study.Despite the non-representative composition of the sampleespecially with regard to criterion-related attitudes toward sexuality, sexual consent, and women in general, as well as executive functioning, educational level, personality traits, and socioeconomic status-the collected data already showed enough systematic variance to draw conclusions about the relationship between our experimentally manipulated factors and individual differences in both propensity for OGC and self-reports of problematic sexuality.It is thus to be expected that the present effects would be all the more evident in a larger and more mixed sample of men with more diverse attitudes, social backgrounds, and sexual behaviors and motivations, especially with a future focus on populations with increased risk for sexual (re)offending.
Trials in the HC condition consisted of two images of the same woman, each featuring a combination of incongruent cues (i.e., Casual × Flirting and Sexy × Rejecting).Following ratings for perceived sexual attractiveness per image, only those image pairs were used in the study for which the difference between the two attractiveness rating scores was close to zero in order to ensure that both images conveyed a similar level of sexual attraction (see ES1 for details).Instead, however, it would have been perfectly reasonable (and probably closer to reality) to maximize selection conflict by using those image pairs for the HC condition in which the sexual attraction of the Sexy × Rejecting image clearly exceeded that of the Casual × Flirting image, thereby potentially increasing baseline occurrence of OGC.Therefore, the more conservative stimulus selection performed for the current study might have led to smaller effect sizes for all MT measures.Especially concerning ER, a floor effect was observed, as most participants made few or no errors not only at T1, when it was expected, but also at T2.Since ER proved to be a particularly informative indicator of OGC in the present paradigm, it would be methodologically desirable to avoid or mitigate such a floor effect in the future.Another advantage of using image pairs with more varying attractiveness-related difference scores would be the opportunity to meaningfully correlate these differences scores with MT measures, further validating the conflict-inducing effect of the HC condition (in addition to the t-tests conducted for this purpose; see Stimuli and Trial Types section) at the level of single trials.
Strengths
Despite the above limitations, the present study is characterized by several strengths of design and data collection.First, the pictorial stimuli used were created specifically to meet predefined conceptual requirements.These requirements allowed for the two trial conditions HC and LC, thus ensuring the functionality of the selection task.Moreover, the stimuli can be expected to increase the ecological validity of the paradigm, as we afforded our models as much individual freedom as possible in interpreting-at the same time well-defined-categories of (1) clothing and (2) facial expression.Effect-related properties of the stimuli were validated prior to the study in a multi-step procedure (see ES1 for details).This extensive preliminary work is a clear advantage over comparable studies that rely on less specific/less standardized stimuli to answer their questions, which also means that confounding variables (e.g., body language, differences in attractiveness on multiple levels) cannot be controlled for to the same extent.Smith et al. (2018) examined the relationship between men's subjective ratings of women's sexual interest and their propensity to engage in sexually aggressive behavior.While for this purpose Smith et al. (2018) presented one stimulus per trial, which participants were asked to assign to one of two categories ("sexually interested" or "rejecting"), in the current study the choice was not between two semantic options, but between two simultaneously presented images.We see the advantage of such a two-image design (1) in that there are no detours due to image-to-text-abstraction, as participants' selection conflict arises directly from two competing visual stimuli depicting the same woman.Moreover, as a direct result of the two-image design, (2) the choice to be made is not based on absolute, i.e., simplified categories (flirting or rejecting), but on a relative instruction ("Click on the woman who is more likely to flirt with you right now") by which, in addition, (3) the concept of rejection is not explicitly introduced to the participants, allowing for more authentic decisions.
The selection task takes the form of a flirting scenario and is thus embedded in a social context characterized by its everyday relevance and comprehensibility.The task does not require any particular abstraction from the participants, and since the instructions used are brief and simple, they should prove easy to understand even in target populations with lower levels of education (e.g., forensic samples).Induction of sexual arousal was achieved noninvasively by means of an erotic audio story, i.e., without recourse to explicit visual material.Spontaneous positive feedback suggests that the study did not elicit substantial reactance among participants.The time required for both parts of the selection task and listening to the audio story in between is about 12 min, which is within reason for a diagnostic task.Finally, the in-person survey, as opposed to a more easily conducted online survey, ensured that the participants in this pilot study all used the same equipment, complied with the general instructions, and, in particular, actually listened to the erotic audio story, which was crucial for the induction of sexual arousal between the two measurements.
Research Desiderata
From the above, suggestions for improving the present paradigm can be derived for future research.Since one focus of this study was to maximize ecological validity, effect-related properties of the stimuli were kept within a range that corresponds to real social settings (e.g., authentic clothing).This closeness to everyday life is at the expense of effect maximization, which could be achieved by means of several-also combinable-approaches.The main consequence of such effect optimization should be an increase in selection conflict experienced by future participants, which in turn should lead to greater variability in MT measures.This could be done, for example, by enhancing clothing contrasts (e.g., by increasing skin display and clothing tightness in the Sexy clothing condition) or by adding information in the form of iconic or semantic cues (as successfully done by Lim et al., 2018) that might alter the perceived sexual attractiveness of the depicted women (e.g., cues with regard to purported [non] promiscuous behavior; see Stewart-Williams et al., 2016).
As a prerequisite for measuring propensity for OGC using the present paradigm, participants must be sexually aroused at T2.For this purpose, an erotic audio story was used in the current study.The manipulation check question about its effect showed that about 8% of the participants claimed to have found the story not to be sexually stimulating at all.Although this is already a fairly small proportion (which may even have been increased by social desirability bias), a more reliable method of inducing sexual arousal in male participants, e.g., by using audiovisual erotic stimuli, would certainly be beneficial.This is all the more true since participants were found to be less aroused in response to erotic audio-only stimuli than audiovisual stimuli once they became familiar with the procedure (Golde et al., 2000).However, to avoid compromising the applicability of the MT paradigm in practical contexts, a trade-off between reliability and invasiveness of the method used for inducing sexual arousal should be considered.
In a broader context, it is also conceivable to measure additional or alternative independent variables (i.e., aside from sexual arousal) using the MT paradigm, such as alcohol or heat.As previous research has shown, alcohol consumption in men leads to misjudgments regarding women's sexual interest (Benbouriche et al., 2018;Farris et al., 2010) and adds to the risk of sexual violence perpetration among men already predisposed to sexual aggression (Abbey, 2011).Likewise, it has been demonstrated that higher temperatures are associated with an increased number of sexual assaults (Anderson, 1989;Field, 1992).These effects might be detectable via the present paradigm by manipulating participants' alcohol consumption and/or the ambient temperature, with OGC as assessed by MT measures remaining the central outcome variable.The paradigm's within-subjects design would prove particularly advantageous here, as a comparison between T1 and T2 would allow for the capture of intraindividual changes occurring after exposure to the factor(s) of interest and, thus, to focus on when some men exhibit sexually problematic behavior, rather than on who displays such behavior (see Abbey, 2011).
Theoretical Implications: Different Causes for Overreliance on Global Cues
Male sexual misperception of women's sexual interest has been thoroughly studied under the premise that for consensual sexual acts to occur, another person's current (lack of) sexual interest must be correctly perceived (e.g., Perilloux et al., 2012;Treat et al., 2016Treat et al., , 2017)).According to our view, OGC in the context of sexual misperception can be explained at least partially by a failure to recognize the different informational value of global sexual and specific affective cues.But even if this different informational value in assessing another person's (non)sexual intentions is correctly discerned, OGC may still occur as a result of insufficient sexual self-control.Exerting self-control proves necessary when a subjectively experienced conflict arises in choosing between mutually exclusive options, where one option corresponds to a spontaneous (often pleasureoriented, yet in some respect detrimental) preference, while the other option is consistent with a more long-term (often normative) objective that is rationally prioritized (e.g., Gillebaart, 2018).These mutually exclusive options, however, may well have been adequately assessed in terms of their respective meanings and possible consequences.Thus, OGC in the context of sexual self-control occurs not because of an inadequate interpretation of global and specific cues, but because the attraction of the global sexual cue-which may be further enhanced by a state of sexual arousal-cannot be resisted despite an "honest attempt" to adhere to a social or personal norm or an instruction demanding the opposite.The absence of such an "honest attempt" constitutes the third theoretical reason for OGC, which we see in egocentric hedonism (i.e., a strong emphasis on one's own [sexual] satisfaction) on the part of the selecting male.In this case, it does not matter whether the woman's rejecting or consenting attitude is correctly perceived, since this aspect has no relevance for the man's selection, which is instead based mainly or exclusively on a stable hedonic preference for sexual cues that goes subjectively unchallenged by opposing interests and social norms.In the perceived absence of a competing (i.e., normative) option, no conflict occurs, so self-control is not a factor in this scenario.Obviously, such a sexual selection strategy greatly increases the likelihood of detrimental consequences of OGC, since no subsequent behavioral adjustment can be expected, neither from a corrected perception nor from a renewed effort of will.
On the surface, it may seem that a mere distinction into sexual selection based on either global or specific cues should prove sufficient in terms of experimental measurability, especially for eventual application in the forensic field, where (predicted) misconduct is judged in absolute terms, i.e., based on its adverse effect on others.In fact, however, it would undoubtedly be beneficial for both forensic and diagnostic/therapeutic purposes to distinguish between different mechanisms that underlie OGC as assessed by sexual selection tasks.While we assume that in our experiment participants' errors at T2 were likely not due to sexual misperception-as recognition and normatively correct interpretation of global and specific cues at T1 has been almost entirely successful-it remains unclear whether errors at T2 were rather due to a failure of sexual selfcontrol or whether breakthroughs of egocentric hedonism also occurred at this point.One way to distinguish between these latter two possible causes of OGC might be to analyze the mouse trajectories of error trials.We hypothesize that errors resulting from a lack of sexual self-control would be associated with rather curved trajectories (since there has usually been an unsuccessful attempt to resist the attraction of the sexual cue and to target the normatively correct option), whereas errors stemming from egocentric hedonism would be found at the end of rather straight trajectories (since a non-conflicting decision has been made based solely on personal preference for the sexual cue).A prerequisite for such an analysis would be a sufficiently high error baseline at T2, which could be achieved, among other things, by (1) enhancing the attraction of the sexual cue, (2) increasing participants' level of sexual arousal, and/or (3) reducing the response window for the forced selection.
Conclusion
The present study was able to show that an MT selection task is a promising approach for the (indirect) behavioral measurement of male OGC.Both manipulated factors-i.e., trial type and measuring time-proved to be of additional methodological value.In particular, the induction of sexual arousal prior to the second measuring time (1) highlighted the importance of ER as an indicator of OGC in the present paradigm and (2) allowed to demonstrate the moderating function of two constructs associated with socially problematic sexual behavior, namely dispositional sex drive and sexual objectification.
The distinct advantage of MT is seen in that recording the movement trajectories of a computer mouse makes it possible to trace the progression of selection processes up to their respective outcomes in real time (Freeman & Ambady, 2010a).MD and AUC, as numerical indicators of the spatial course of a trajectory, represent the procedure's genuine measures.While some researchers highlight the novel utility of MT and emphasize its potential (e.g., Kieslich & Henninger, 2017;Wulff et al., 2019), other voices are critical and express concerns about a possible redundancy of MT's genuine measures compared to more conventional ones, such as RT and ER (e.g., Bartels et al., 2019).
As for the results of the current study, MD and AUC were indeed found to be considerably less conclusive than RT and ER.However, not least due to the limited sample size, it currently remains unclear whether RT and ER really provide more reliable estimates or whether MD and AUC are mere epiphenomena of the measurement (Bartels et al., 2019), so that one could possibly forego the use of a computer mouse (as well as a comparatively intricate data analysis) in favor of a keyboard-based approach.Also, both spatial measures may still prove crucial for differentiating causal mechanisms underlying OGC and thus provide important diagnostic information, e.g., in the context of therapeutic interventions and risk assessments.Finally, it cannot be ruled out that a keyboard-based approach would not yield results of the same quality even if MD and AUC should indeed turn out to be informatively redundant, as the moving of a computer mouse as such (due to the thereby given possibility to change one's mind, and even multiple times) might be a necessary precondition for producing errors of diagnostic and forensic significance.
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/.
Fig. 2
Fig. 2 Structure and course of a mouse-tracking trial in the high conflict condition
Fig. 3
Fig. 3 Numbers of errors for both trial conditions.Notes.Left chart: low conflict trial condition; right chart: high conflict trial condition
Fig. 4
Fig. 4 Moderation of sex drive on the relationship between sexual arousal and overreliance on global cues as indicated by error rates
Fig. 5
Fig. 5 Moderation of sexual objectification on the relationship between sexual arousal and overreliance on global cues as indicated by error rates.Note.SO = sexual objectification significant changes were found for MD and AUC.
Table 1
Repeated measures ANOVAs for mouse-tracking measures N = 79.AUC = area under the curve; ER = error rate; HC = high conflict condition; LC = low conflict condition; MD = maximum deviation; RT = reaction time
Table 2
Moderated hierarchical regressions across mouse-tracking measures N = 79.AUC = area under the curve; ER = error rate; MD = maximum deviation; RT = reaction time; SDI = sex drive index; SO = sexual objectification
Table 5
Correlations between self-report measures and mousetracking difference scores as a function of time N = 79.AMMSA = Acceptance of Modern Myths about Sexual Aggression Scale; AUC = area under the curve; ER = error rate; HBI-19 = Hypersexual Behavior Inventory-19; MD = maximum deviation; RT = reaction time; SDI-2 = Sexual Desire Inventory-2; SD Index = sex drive index; SIS = Sexual Inhibition Scale; SES = Sexual Excitation Scale; SOI-R = Revised Sociosexual Orientation Inventory; SOOI = Sexual Objectification of Others Inventory; SUSS = Socially Undesirable Sexual Selection Scale.Δ based on the difference of high conflict trials minus low conflict trials.Statistically significant results are in bold
|
2024-02-09T06:17:35.055Z
|
2024-02-07T00:00:00.000
|
{
"year": 2024,
"sha1": "b9b8f0370b9869f9032e8bd769c13f591563e9a4",
"oa_license": "CCBY",
"oa_url": "https://link.springer.com/content/pdf/10.1007/s10508-023-02798-x.pdf",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "7ca2555f79304e6d2c5309ba97dc64a52dc1ca40",
"s2fieldsofstudy": [
"Psychology"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
257714795
|
pes2o/s2orc
|
v3-fos-license
|
Heart Failure, Female Sex, and Atrial Fibrillation Are the Main Drivers of Human Atrial Cardiomyopathy: Results From the CATCH ME Consortium
Background Atrial cardiomyopathy (atCM) is an emerging prognostic factor in cardiovascular disease. Fibrotic remodeling, cardiomyocyte hypertrophy, and capillary density are hallmarks of atCM. The contribution of etiological factors and atrial fibrillation (AF) to the development of differential atCM phenotypes has not been quantified. This study aimed to evaluate the association between histological features of atCM and the clinical phenotype. Methods and Results We examined left atrial (LA, n=95) and right atrial (RA, n=76) appendages from a European cohort of patients undergoing cardiac surgery. Quantification of histological atCM features was performed following wheat germ agglutinin/CD31/vimentin staining. The contributions of AF, heart failure, sex, and age to histological characteristics were determined with multiple linear regression models. Persistent AF was associated with increased endomysial fibrosis (LA: +1.13±0.47 μm, P=0.038; RA: +0.94±0.38 μm, P=0.041), whereas total extracellular matrix content was not. Men had larger cardiomyocytes (LA: +1.92±0.72 μm, P<0.001), while women had more endomysial fibrosis (LA: +0.99±0.56 μm, P=0.003). Patients with heart failure showed more endomysial fibrosis (LA: +1.85±0.48 μm, P<0.001) and extracellular matrix content (LA: +3.07±1.29%, P=0.016), and a higher capillary density (LA: +0.13±0.06, P=0.007) and size (LA: +0.46±0.22 μm, P=0.044). Fuzzy k‐means clustering of histological features identified 2 subtypes of atCM: 1 characterized by enhanced endomysial fibrosis (LA: +3.17 μm, P<0.001; RA: +2.86 μm, P<0.001), extracellular matrix content (LA: +3.53%, P<0.001; RA: +6.40%, P<0.001) and fibroblast density (LA: +4.38%, P<0.001), and 1 characterized by cardiomyocyte hypertrophy (LA: +1.16 μm, P=0.008; RA: +2.58 μm, P<0.001). Patients with fibrotic atCM were more frequently female (LA: odds ratio [OR], 1.33, P=0.002; RA: OR, 1.54, P=0.004), with persistent AF (LA: OR, 1.22, P=0.036) or heart failure (LA: OR, 1.62, P<0.001). Hypertrophic features were more common in men (LA: OR=1.33, P=0.002; RA: OR, 1.54, P=0.004). Conclusions Fibrotic atCM is associated with female sex, persistent AF, and heart failure, while hypertrophic features are more common in men.
in cardiovascular disease. Fibrotic remodeling, cardiomyocyte hypertrophy, and capillary 48 density are histological hallmarks of atCM. However, the contribution of various etiological 49 factors and atrial fibrillation (AF) to the development of differential atCM phenotypes has not 50 been robustly quantified. We aimed to evaluate the association between histological features of 51 atCM and the clinical phenotype. Many cardiovascular diseases substantially impact atrial function. Heart failure, hypertension, 79 supraventricular tachycardias, valvular diseases and thyrotoxicosis can, over time, deteriorate 80 the mechanical, electrical and endocrine function of the atria. In some patients, these 81 mechanisms cause a progressive atrial cardiomyopathy (atCM) whichin the consensus paper 82 of Goette et al. in 2016was defined as "any complex of structural, architectural, contractile 83 or electrophysiological changes affecting the atria with the potential to produce clinically 84 relevant manifestations" (1). Importantly, atCM has a substantial impact on cardiac 85 performance, and also on the occurrence of atrial fibrillation (AF) and stroke (2-4). Once AF 86 occurs, the arrhythmia itself accelerates the progression of atCM (1, 5). AtCM mechanisms 87 vary between individual patients and over time (6), likely contributing to the limited efficacy 88 of rhythm control therapies for AF (7,8). 89 One of the key features of atCM is enhanced fibrosis, as extensively reported in multiple 90 animal models (9-17) and various clinical settings, including left/right ventricular dysfunction 91 (18), AF (19,20) or heart failure (21,22). We recently reported that endomysial fibrosis, 92 defined as extracellular matrix deposition between myocytes within the muscle bundles, rather 93 than overall connective tissue content, is the main determinant of conduction disturbances in 94 human AF (23). Fibroblasts are not only responsible for extracellular matrix formation, but they 95 may also interact electrically with cardiomyocytes, thereby affecting conduction (24,25). 96 Increased cardiomyocyte size relates to increased wall thickness and chamber dilatation, both 97 of which can alter electrical propagation (26)(27)(28). Finally, capillary rarefaction may be a marker 98 of prolonged myocardial stress, as observed in patients with heart failure with preserved 99 ejection fraction (HFpEF), hypertensive heart disease and AF (29,30). 100 Although a first classification of atCM was proposed in the consensus paper from 2016 101 (31), a detailed investigation of the relationship between histological features of atCM and 102 clinical traits is still missing. Here, we sought to comprehensively appraise histological changes 103 in both right and left atrial samples in a large European cohort of patients undergoing cardiac 104 surgery for a variety of indications. We aimed to systematically quantify the type of atrial 105 fibrosis, fibroblast density, myocyte size, capillary density and capillary size using a previously 106 validated, semi-automated quantification method developed for high throughput histological 107 studies (32). Additionally, we aimed to cluster left and right atrial samples based on their 108 histological features and to identify the clinical traits associated with these clusters. LA, n=76 RA, 20 of which were paired) were of sufficient quality and included in the analyses.
117
The study was performed in accordance with relevant guidelines and was approved by the local 118 ethical committees of the sampling centers. Netherlands) was applied for 120 minutes. Prolong Gold Antifade mounting medium 131 (ThermoFisher (P10144), Netherlands) was added when coverslips were mounted. Sections 132 were imaged using a Leica microscope (DM4B) and camera (MC170HD). Only sections with 133 transversely cut cardiomyocytes were selected and photographed for further analysis. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint To study differences in atrial histology between LA and RA samples, all samples 157 (n=171) were combined for analysis using a multivariate model including the covariate 'tissue 158 location'. Additionally, the interaction between heart failure and tissue location (LA/RA) was 159 studied. Next, multivariate regression models were constructed to study the association between is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 24, 2023.
177
Tissue sample selection 178 A total of 260 samples from 227 patients were collected from the CATCH-ME biobank.
179
Dropout occurred due to low tissue quantity (inability to perform cryosectioning, n=66), poor Moreover, we found a significant negative interaction (-2.52±0.73, p<0.001) between atrial side 199 and heart failure status on endomysial fibrosis, indicating that heart failure is more strongly 200 associated with endomysial fibrosis in LA than RA samples. Given the substantial difference 201 in atrial histology between atria, further analyses were stratified by atrial side. Table 2). Rhythm history, heart failure, sex, and age were included as covariates 208 in the multivariate analysis as they were univariately associated with clinical traits. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 24, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 Additionally, 4 principal components containing univariately associated clinical traits were also 210 included. Overall, heart failure, rhythm history and female sex showed the strongest 211 associations with atrial histology, with stronger associations in the LA than RA (Figure 4).
212
Female sex and heart failure were associated with higher ECM content in the LA ( Table 3 summarizes the numeric details of this 219 analysis.
220
After exclusion of younger patients with end-stage heart failure receiving an HTX, age 221 was associated with increased total ECM ratio (LA, p=0.022) as well as with enhanced 222 endomysial fibrosis (LA: p=0.05 /RA: p=0.048). Capillary density was no longer increased in 223 heart failure patients after exclusion on HTX-patients (Supplemental Tables 3 and 4).
224
A moderate (Pearson) correlation was observed between endomysial fibrosis and total 225 ECM in the LA (r=0.68, p<0.001) and RA (r=0.60, p<0.001) samples ( Figure 5). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 24, 2023. ; https://doi.org/10.1101/2023.03.23.23287667 doi: medRxiv preprint than samples in the hypertrophic cluster. Fibroblast-specific vimentin signal and capillary size 243 did not differ in the RA clusters ( Figure 6).
244
Univariate analysis indicated that patients with fibrotic atCM in the LA more often had 245 a history of AF (p=0.041) or heart failure (p<0.001). Heart failure with moderately or severely 246 reduced left ventricular ejection fraction (HFrEF, p=0.003), in particular, was more common in 247 this cluster (Supplemental Table 5). Moreover, the primary indication for surgery differed 248 between patients showing fibrotic and hypertrophic atCM in the LA (p=0.007), with more 249 coronary artery bypass grafting (CABG) or aortic valve surgery in patients with hypertrophic 250 cardiomyopathy, but more HTX in patients with fibrotic cardiomyopathy (Supplementary Table 251 5). Patients with fibrotic remodeling in the RA were more often female (p=0.002) and shorter 252 (p=0.042). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The aim of this study was to better understand the relationship between the main AF-related is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 24, 2023. ; https://doi.org/10. 1101/2023 Prior research by our group showed that endomysial fibrosis, but not total ECM content, 296 correlates well with AF complexity, i.e. the degree of conduction block in fibrillating atria (23).
297
In the present study, the correlation between total ECM and endomysial fibrosis was significant 298 but only moderately strong. This suggests that, in order to assess the contribution of atrial 299 fibrosis to the electrophysiological substrate for AF, determining endomysial fibrosis may be 300 more relevant than total ECM content quantification. In addition, it is unknown whether LGE-301 MRI can accurately quantify endomysial fibrosis; it relies on regional differences in a shift in is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint Our data showed an independent association between AF and endomysial fibrosis in human LA 333 and RA, while AF was not associated with total ECM, altered myocyte size, capillary density is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint connective tissue content. In the present study, an association between age and endomysial 366 fibrosis was observed in heart failure patients who did not receive a transplant. Electrical Our, and the referenced experimental data, suggest that endomysial fibrosis might contribute to 372 the association between AF and age to some degree. In the present study, however, the relative 373 contribution of age was quantitatively limited. A 5-year age difference, causing a 1.5-fold 374 increase in AF prevalence in the general population (72), was associated with an 0.06 µm 375 increase in endomysial fibrosis in our study (comparable to 3.3% or 5.6% of the effect of heart 376 failure or persistent AF).
379
In AF patients, female sex is associated with a higher risk of stroke (73), higher recurrence rates AF compared to those without AF while such a difference was not found in males (79). Our 387 study is the first to demonstrate a clear association between female sex and both LA total ECM 388 and endomysial fibrosis. Interestingly, the effect size was comparable to that of persistent AF 389 or heart failure. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 24, 2023. ; surgery. This result can be seen as a first step towards the development of a classification of 397 atCM based on histological features and clinical traits associated with them. To a certain degree, 398 our results support aspects of the EHRA's classification proposed in the consensus paper in 399 2016 (1). Our data confirm the association between fibrotic alterations and AF or heart failure, 400 while other aspects identified by our study, such as sex-differences in endomysial fibrosis and 401 cardiomyocyte size, were not included in the EHRA classification (1).
402
Our study also stresses the need for high throughput, standardized, automated Data regarding pre-operative medication or nicotine and alcohol consumption were, 413 unfortunately, not complete and confounding of our data by these factors cannot be ruled out.
414
Also, we did not assess the role of fat tissue, inflammatory changes or endothelial remodeling. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 24, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 Funding 421 This study has been supported by grants of the Netherlands Heart Foundation (CVON2014-09, is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint Detection and quantification of left atrial structural remodeling with delayed-558 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 24, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 24, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 24, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 24, 2023. ; https://doi.org/10.1101/2023.03.23.23287667 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 24, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 24, 2023. ; https://doi.org/10.1101/2023.03.23.23287667 doi: medRxiv preprint
|
2023-03-25T01:25:49.749Z
|
2023-03-24T00:00:00.000
|
{
"year": 2023,
"sha1": "db5dc4bf7a5e38640f3e32b5d95683e19c7ae868",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "MedRxiv",
"pdf_hash": "db5dc4bf7a5e38640f3e32b5d95683e19c7ae868",
"s2fieldsofstudy": [
"Medicine",
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
73312682
|
pes2o/s2orc
|
v3-fos-license
|
COMMUNITY NURSES ’ PERCEPTIONS OF AND EXPOSURE TO CHILDREN WITH SEVERE DISABILITIES AND THEIR PRIMARY CAREGIVERS
In primary health care clinics nurses are faced with individuals of different ages with different problems, ranging from minor ailments to severe disabilities. Particularly vulnerable are those with severe disabilities, with specific reference to those with communication and intellectual impairments because of their inability to articulate their needs, feelings and rights. Community nurses are often the first contact that primary caregivers (parents) of children with severe disabilities (CSDs) have with health professionals and they often remain the only professionals who support and assist these caregivers. It is therefore clear that these nurses need to be equipped with the necessary knowledge and skills to assist caregivers in dealing with their CSDs. This is a descriptive study aimed at determining community nurses’ perceptions regarding disability and their exposure to CSDs. Quantitative (questionnaire) and qualitative (focus groups) methods were used to obtain data. Results indicated that nurses regarded their knowledge and skills in dealing with CSDs and their caregivers as inadequate, despite the fact that they were exposed to them and were expected to provide services to them. A need for knowledge and skill training in this regard was identified.
INTRODUCTION
Since the birth of South Africa's democracy in 1994, a new health policy, namely primary health care (PHC) was accepted in order to address equality in health services given the manpower and resource constraints of the developing world (Department of Health, 1999:4).
The Alma Ata declaration identified PHC as "… essential health care made universally accessible to individuals and families in the community by means acceptable to them, through their full participation and at a cost that the community and country can afford" (WHO/ UNICEF in Rifkin, 1986:240).PHC is thus not only seen as the first level of care, but rather as a reorientation of the health care system from its present concentration on late stage, high technology hospital services to community and preventive services (Willis, Biggins & Donovan, 1999:210).In order to meet the needs of as large a number of people as possible in a culturally appropriate, cost effective way, all possible resources must therefore be utilised (House, McAlister & Naidoo, 1990:18;Stanhope, 1995:50).Consequently professionals began to realise that primary caregiver involvement is central to service delivery, that a high degree of collaboration and joint decision-making between primary caregivers and professionals should be established so that a joint vision and joint decision-making can be achieved, and that the cross-training of professionals and paraprofessionals in providing services to everyone should be highlighted (Chapman & Ware, 1999:105;O'Toole, 1988:324).These three aspects form the basis of a transdisciplinary approach to service delivery -an approach that not only reduces the duplication of services, but one that also addresses the fragmentation that currently exists amongst differ-ent service providers.The transdisciplinary approach is therefore closely linked to PHC as it becomes meaningful within a more comprehensive model of service delivery.
Nurses are the frontline practitioners in the provision of PHC services.Community nurses do not merely perform nursing activities in the community setting.
Their task should rather be seen as the provision of ongoing, comprehensive and general practice that is not limited to a specific age or diagnostic group aimed at the promotion of health and the prevention of illness (Clark, 1996:18;Lerner & Ross, 1991:47).Community nursing therefore "requires planning and sharing with others in the community to promote health for the community, family and individual.Through the collaborative process, the special abilities of others are used to communicate, plan, solve problems and evaluate services" (Standards for Community Nursing in Wilkey & Gardner, 1999:307).The role of the nurse has thus shifted away from the role of servant to the medical profession towards the role of helper and partner of people and communities.
One particularly vulnerable group of individuals within communities is children with severe disabilities (CSDs).
It is evident that there is a high incidence of disability in South Africa.It is estimated that 5 -12% of South Africans are moderately to severely disabled (NPA, 1999:9).Of South Africa's total population of 40 million, 73% are women and children (Government Gazette no 17910, 1997:11) placing them at a higher risk of disability (Ramey & Ramey, 1992:337).More than 80% of black children with a disability live in extreme poverty and have poor access to appropriate health facilities or early childhood developmental opportunities (NPA, 1999:9).In the sphere of individuals with severe disabilities a particularly neglected group is that of individuals with communication and intellectual impairments due to their inability to articulate their needs, feelings and rights (Thorburn & Marfo, 1990:180).A study conducted in schools for children with cognitive impairment in and around Pretoria indicated that 39% of these children were using less than 15 intelligible words (Bornman & Alant, 1997:17).This percentage is significantly higher than that in comparable international studies such as the 2,4% reported in North Dakota (Burd, Hammes, Bornhoeft & Fisher, 1988:376) and 6% in rural areas in Washington State (Matas, Mathy-Laikko, Beukelman & Legresley, 1985:26).
There are different reasons for this, including the fact that persons with severe disabilities in South Africa are not exposed to Augmentative and Alternative Communication (AAC) strategies, as it is still a relatively new field of expertise in this country.Internationally it is also reported that programmes to specifically address the communication needs by equipping these children with the necessary skills to interact, are limited (Werner, 1987:3-10).
Apart from the high incidence of disability in developing countries, these individuals also share commonalties in terms of the services offered.It is well documented in the literature that people with severe disabilities receive less education and vocational training, and are often unemployed, resulting in poverty (Lundgren-Lindquist & Nordholm, 1993:83).The aura of charity still pervades many of the services offered and rarely are the persons with disabilities given the opportunity to help themselves (McConkey, 1996:28).
Due to long held low expectations of people with severe disabilities, they are often excluded from leadership positions or from any planning and decision-making in their communities, resulting in a lack of representation of their needs.
In addition, very few professionals in South Africa are trained in AAC implementation, and are thus unable to provide AAC intervention to people in need.This is due to the fact that communication intervention for children with severe disabilities (CSDs) has not been a priority, and that international sanctions limited South Africa's participation in the international AAC field until the early nineties.In addition, the majority of services provided by professionals are located in a few large cities, making it inaccessible to the rural population, leading to services being provided to less than three percent of disabled people in need (Thorburn & Marfo, 1990:22).Professionals also tend to work in isolation with minimal integration between services, which led many people to believe that service provision to disabled people was too complicated to be administered by persons other than professionals.It was thought that any other form of rehabilitation would be ineffective (Thorburn & Marfo, 1990:24).
When placing disability within the PHC paradigm, it is clear that the first contact that many CSDs and their primary caregivers have with professionals, is contact with community nurses.This is due to a number of factors, including the relative ease of access to primary health care clinics for South Africans, even those living in rural and remote areas; the fact that primary caregivers attend well-baby clinics and also visit these clinics for inoculations; and the fact that community nurses are often the only health professionals in the area.Furthermore, these nurses often remain the only professionals who provide support and assistance to caregivers of pre-school children.
Little information is, however, available as to community nurse's perceptions regarding CSDs; their exposure to CSDs and how they perceive the needs of CSDs and their primary caregivers.This information is central when planning a training programme for nurses to equip them with the necessary skills and expertise to assist primary caregivers in dealing with their CSDs.
The current research project therefore discusses the issue of nurse's perceptions in dealing with CSDs and their primary caregivers.In order to achieve this a descriptive needs analysis was conducted.
AIMS OF THE STUDY
The main aim of this study was to conduct a needs analysis in order to: ascertain the perceptions of community nurses regarding service delivery to CSDs; determine the exposure of community nurses to CSDs; establish how community nurses perceive the needs of CSDs and their primary caregivers.
RESEARCH DESIGN AND METHO-DOLOGY
A descriptive design was used employing both quantitative and qualitative measures, as it would be appropriate in meeting the aims of the study (Leedy, 2001:191).Quantitative data was obtained by using a short two-page questionnaire that was developed specifically for this purpose.Qualitative data, on the other hand, was obtained by conducting two focus groups with community nurses as they yield information from multiple sources, contain rich contextual data and are also excellent tools for capturing the nature of experiences, individual perspectives and opinions (Brotherson & Goldstein, 1992:336;Krogh & Lindsay, 1999:223;Krueger, 1988:47;Morse, 1996:467).The methodology will be described in terms of the participants, data collection and analysis as well as reliability and validity.
Participants
Community nurses employed by the Moretele Health District, a semi-rural and rural area situated in the North West Province (former Bophuthatswana), participated in this research.This Health District is viewed as a previously deprived area requiring ongoing Government support to uplift it and ensure equality of services.Despite this, the particular area is a good example of close collaboration between the Health District, which provides PHC and the Jubilee Hospital, which provides secondary health care.It is also relatively close to Pretoria where tertiary care, viz.traditional rehabilitation by therapists, is provided.This Health District is made up of a total of 25 clinics and 14 mobile points, which serve an estimated average of 4365 persons per day.
The clinics differ in size, ranging from some with only two nurses serving approximately 30 people daily to some clinics with eight nurses serving 300 people daily.
Participants were either working in clinics or at the various mobile points in the area at the time when the research was conducted and nurses who were not working in a nursing capacity (e.g.only administratively), who were positioned at the Jubilee hospital (providing secondary care), who were between jobs or retired, were excluded.The qualifications of the nurses were not included as a selection criterion, so that it covered a whole range, from nursing assistants to professional nurses.
Background information on the participants in terms of their age, gender, nursing status and experience, is provided in Table 1.
Data collection
In order to obtain the quantitative data, the researcher and fieldworker visited all 25 clinics and 14 mobile points in the particular health district in order to complete the questionnaire.Table 2 provides a summary of the most important areas covered in the questionnaire, with specific reference to the motivation for inclusion.The category, type and number of each question in each category are also discussed.
All the nurses who were present on the day that the researcher visited the particular site (ranging from nursing assistants to professional nurses) were included.
A total of 111 questionnaires were completed.This implies that 92% of the total population was surveyed (111 of 121 nurses employed in the Moretele Health District).
In order to obtain the qualitative data, 21 community nurses from this district who took part in an in-service training programme called "Nurses-in-Training" that meets monthly and receives talks on various nursingrelated topics, were involved.Due to the vast spread out nature of the district and the impracticality of closing clinics so that nurses could participate in the research, it was decided to include all the nurses who attended the first "Nurses-in-Training" programme for 2000 in the focus group.This resulted in two focus groups: one focus group with 10, and the other with 11 participants.Literature suggests groups of 8 -10 participants (Frey & Fontana, 1993:28).As a result of the commonalties that these nurses share and the fact that they mostly knew each other, rapport was quickly es Nurses generally tended to be older (85% in their thirties and forties).This finding is similar to a study conducted in the Eastern Cape that reported that 80% of their sample was aged 30 -40 years (Thipanyana & Mavundla, 1998b: 28).This implies that they are a mature workforce in this particular health district.This also impacts on the teaching principles that should be considered when planning a training programme, as issues related to adult learning styles will have to be adhered to.
Age in years
Percentage By far the majority of participants were female (98%).
This was expected as nursing has traditionally been considered a female occupation.
Percentage
More than half the total number of nurses (59%) had more than ten years experience, highlighting the fact that they tended to be a stable workforce who stayed employed for a long time and therefore had a vast amount of experience.This makes them excellent candidates for additional in-service training, as they will be able to utilise the new knowledge and skills for a long time.The movement between clinics is varied as 27% had spent less than one year in the present clinic but The degree to which nurses feel comfortable in handling the different disability types might be used as a predictor of how they view their skills and knowledge in this regard.It is assumed that if nurses have the knowledge and skills to deal with a child with a particular disability that they would then feel comfortable in handling this child.
However, a disadvantage of self-evaluation is that it is difficult to be objective and to accept one's own strengths and weaknesses (Bradshaw, 1989:108).tablished, so that the focus group was experienced as non-threatening.Each focus group was facilitated by a speech language therapist (SLT) with a primary interest in the severe disability field.It was aimed at determining nurses' perceptions and their exposure to CSDs as well as encouraging them to respond from their own experience.The researcher acted as one of these facilitators.English was used as the focus group medium as all nurses were conversant and comfortable with it.Both facilitators led the respective focus groups in a semi-structured discussion of three openended questions (Frey & Fontana, 1993:36;Krueger, 1988:60), namely: How often do you see the primary caregivers of CSDs?
Tell us about your experience.
What do you think are the biggest problems these primary caregivers have?
What can we do to help you to help the primary caregivers of CSDs better?
During the discussion the facilitators asked for clarification when some concepts were unclear or in cases where the data were open to misinterpretation (Bryman, 1994:5;Krefting, 1991:219).Nurses were also encouraged to take part actively, to share their wealth of experience, to understand that no comments would be regarded as "stupid" or "silly" and were assured that they would remain anonymous.Flexibility was allowed in terms of the sequence of questions so that facilitators were able to listen to the discussion, to observe and to respond to what they saw and heard.Structural coherence of the topic was thus maintained, increasing credibility (Krefting, 1991:217).In an attempt to enhance trustworthiness member checks were included, which entailed that the facilitators gave a short summary at the end of the discussion of each of the three questions.Participants were then asked whether they agreed, disagreed, or if any important issues raised had been overlooked (Hoffart, 1991:526).The two facilitators spent some time debriefing directly after the focus groups to discuss their interpretations in order to enhance trustworthiness (Peshkin, 1993:24).No areas that needed additional probing or clarification were noted.Debriefing is an important part of investigator triangulation and was included to heighten the credibility of the data obtained (Brotherson & Goldstein, 1992:337;Kimchi, Polivka & Stevenson, 1991:364-365).Verbatim transcriptions of these two focus groups were made.Focus group 1 lasted 58 minutes and Focus group 2 lasted 67 minutes.
Data analysis
For the analysis of the quantitative data descriptive statistics (frequency distribution counts and percentages) were used, implying that all variable values were listed and counted each time they occurred (Babbie & Mouton, 2001:460).In addition, contingency tables were created as they indicate the relationship between sets of nominal data (e.g.determining the relationship between exposure to the different disability types and how comfortable nurses felt in dealing with these children).
The Pearson Chi-square test was then used to determine the significance of this relationship.
Regarding the qualitative data, the researcher delineated themes from each of the two focus groups and came to some tentative conclusions.Transcriptions were then handed to the other focus group facilitator and she was also asked to delineate themes (peer debriefing).Care was taken to avoid "coding fetishism" where coding becomes a mechanical process of labelling involving obsessive and unnecessary details (Webb, 1999:325).Rather, coding was seen as part of the whole analytic process that also involves theoretical perspectives and interpretations.When two re-searchers separately code and then cross-check data, the issue of analytic stability comes to the foreground (Brotherson & Goldstein, 1992:338).In order to address this, both researchers received a set of coding guidelines pertaining to the content and the analytic procedure, e.g. the initial questions that were appropriate as well as the decision rules for determining the categories.Following the development of the themes the two facilitators met to review the theme analysis with a request for verification, correction, clarification and/or elaboration of tentative conclusions drawn from the data analysis.This process is essential in order to ensure the credibility of the data (Krueger, 1988:60).
Apart from obtaining only significant and consistent themes in the data it is also rich in providing illustrative examples.In order to not fragment, decontextualise and/or misinterpret data, it was crucial to go back to the complete text frequently (Webb, 1999:327).
Validity and reliability
Regarding the quantitative data (questionnaire) the following factors were taken into consideration: ended format, so that the participants were not guided in a specific direction, and also to allow them to answer in their own words.
• Likert scales did not include more than five options (so that the options did not become visually overwhelming and so that participants could remember the options (Babbie & Mouton, 2001:153).
• Items were mostly short e.g.Question 8 "How serious are the following problems in your opinion?" and care was taken to ensure that each question dealt with only one concept (Babbie & Mouton, 2001:234).
• Care was taken to avoid any biased or leading questions, jargon and other difficult terminology, and also to ensure the clarity of the questions, so that all the participants understood the same concepts with the vocabulary used (Bynner & Stribley, 1979:141;Mertens, 1998:115).
• An attempt was made to keep the questionnaire as short as possible to ensure that participants did not lose interest and/or become fatigued resulting in the omission of important information (Leedy, 2001:206).
• Meticulous care was taken with the visual appearance of the questionnaire.Questions were all numbered, organised in logical sequence and did not have too many items per page.
• Clear, brief instructions to guide the participants were included at the top of the questionnaire.They were encouraged to complete all questions, as incomplete responses would impact on the reliability of the data (Leedy, 2001:207).
In qualitative research, on the other hand, the terms reliability and validity cannot be used, as the nature of the data is different from that of quantitative data.Instead, a new set of criteria, developed for and fit for qualitative research should be used, one of which is the concept "trustworthiness" (Cutcliffe & McKenna, 1999:375;Kirk & Miller, 1986:23).In the present research attempts were made to increase the trustworthiness of the data through the use of specific techniques, which are displayed in Table 3 (Brotherson & Goldstein, 1992:336-338;Cutcliffe & McKenna, 1999:376-379;Kimchi, Polivka & Stevenson, 1991:364-366;Klopper, 1995:25-28; Krefting,
ETHICAL CONSIDERATIONS
Permission to perform the study was obtained from the relevant authorities.All participants from clinics in the given health district who were able to participate, were included.Data collection was conducted within the current structures for training, i.e. the existing "Nurses in Training" group to minimalise any disruption to the general treatment of patients.All participants were informed of the aims and nature of the research and gave consent that the information could be used.
To protect the identity of the participants no identifying information is provided.
RESULTS AND DISCUSSION
Results revealed the nurses' current exposure to CSDs, their rating of the seriousness of various health condi- In order to further explore the latter aspect, contingency tables were created to correlate the relationship between exposure to the different disability types and how comfortable nurses felt in dealing with these children.
The Pearson Chi-square test was administered to all the categories.Results are shown in Table 5. (Table 5 is on page 45.) From Table 5 it is clear that intellectual impairment and little or no functional speech (LNFS), which implies a severe communication disorder, proved to be statistically significant, and therefore further testing was done.
These results are shown in Table 6.(Table 6 is on page Peer group examinations and discussions Two independent researchers involved in PhD studies with experience in qualitative research but diverse backgrounds (nursing and speech-language pathology) were involved in order to assist with the exploration of concepts and the analysis of data.This provided the researcher with the means to discuss insights and problems, to "test" ideas regarding certain aspects and also to prevent personal bias.This bias can be attributed not only to the researcher's own culture, background, interest, etc. but could also arise from the prolonged engagement in the particular community.These peer discussions also led to deeper reflexive analysis of the data.The background information of the participants for the focus groups was obtained from a short questionnaire.This was then compared with the available demographic information for the whole group in order to ensure representativeness.It is important to determine whether the data obtained can be seen as typical or atypical of the phenomenon under investigation.This is briefly described in the results section.
Indirect method
Look at internal and external validity measurements The above-mentioned discussion on credibility and transferability is important for this section, as Krefting (1991:221) states that no validity is possible without reliability (and thus no credibility without dependability).A demonstration of the former is sufficient to establish the latter.
Triangulation
This method and its application in the current research has already been discussed in detail.
Stepwise replication
In qualitative research the researcher always aims to provide a "thick" or "rich" description of the context.Rigour was applied in describing the context, how data was obtained, analysed (i.e. the development of themes) and interpreted.The methodology should be explicit enough for an independent researcher to repeat.An independent knowledgeable researcher with experience in the field of qualitative research was asked to evaluate the degree in which the research process, including the raw data, data reduction and analysis products (condensed notes and qualitative summaries), data reconstruction (thematic categories), findings, interpretations and recommendations follow acceptable research practice.This researcher (the "auditor") stood completely neutral to the research.It should be emphasised that neutrality is not seen as a way to avoid "contamination".Rather the researcher's characteristics, attitudes and feelings are recognised as influencing the research and hence made explicit.
Percentage
Weighted frequencies indicate that there were no significant differences in the nurses' level of comfort in dealing with children with different types of disabilities, i.e. they did not feel more uncomfortable with a child with intellectual impairment (II) than with a child with CP.
D is a b ilit y t y p e
Percentage More than half the total number of nurses (59%) had more than ten years experience, highlighting the fact that they tend to be a stable workforce that stayed employed for a long time and because of this had a vast amount of experience.This makes them excellent candidates for additional in-service training, as they will be able to utilise the new knowledge and skills for a long time.The movement between clinics is varied as 27% had spent less than one year in the present clinic but 25% had spent 6 to 10 years and 20% had spent 4 to 5 years at the present clinic.Given the high rate of staff turnover during the past decade, the stability of the nursing workforce in the Moretele Health District is important.port, 1986:15).In addition, the role of community nurses in the rehabilitation of children with intellectual impairments has become more prominent in recent years (Alant, 1998:18).This table also indicates that despite their exposure to children with intellectual impairment all the nurses felt more comfortable than uncomfortable in handling these children.However, the nurses who had never seen children with intellectual impairment were notably the group who felt the most comfortable in handling them (93.7%).This might be due to the fact that they were unaware of the potential problems and issues of dealing with these children.
The second disability category that indicated significance on the Pearson Chi-square test was children with LNFS.Data for this group is shown in Table 7. in handling these children.However, once they were exposed to these children they became more uncomfortable in handling them, implying that even very infrequent The last section of the quantitative data relates to the type of service delivery and the needs of the CSDs as perceived by the nurses.Results are shown in Table 8.Table 8 follows on next page.
From Table 8 it is clear that the services to CSDs and their primary caregivers mostly entail direct referral, irrespective of the disability type.This is an important factor, as it was noted during the focus groups that the level of compliance with referral is low, as the primary caregivers of CSDs often do not have the necessary resources (e.g.financial constraints and transport difficulties).This correlates with observations from a study conducted in the Eastern Cape which also reported on the prominence of transport issues (Thipanyana & Mavundla, 1998a:23).The sustainability of this type of service delivery is therefore questionable indicating the need to train community health by providing appropriate knowledge and skills so that direct service delivery to this population becomes feasible.According to Musholt (1995:307) this type of training should be transdisciplinary in nature so that community nurses can start acting as generalists as they need to draw on many resources to provide quality care.This is confirmed in the nurses' evaluation of what CSDs and their primary caregivers need.Apart from schools (22%), the most pressing issues were trained professionals (14%) and good parental care (14%) that could be achieved through parent training programmes.
Regarding the qualitative data, a summary of the themes delineated from the two independent focus groups is provided in Appendix A. However, nurses acknowledged that in practice followups are not done and that most of the service delivery is based on referral, indicating a disagreement between policy and practice.They spoke spontaneously about the problems they encountered in the workplace, despite the fact that this information was not probed.This included aspects such as feelings of inadequacy due to the fact that they do not know how to handle primary caregivers of CSDs, depression, attitudinal barriers due to limited staff and time constraints and the fact that no follow-ups are done.Regarding their experiences it A wide range of needs was identified.Nurses perceived the need for a school to be the most pressing for CSDs, followed by the need for trained and skilled people to work with these children (14%) and more effective parental care that emphasises love and care through the training of the primary caregivers (14%).The latter two aspects have also been mentioned in the literature as pressing needs for CSDs (Alant, 1998:20).
Percentage
Table 8: Service delivery and needs of CSDs as perceived by nurses was noted that their exposure varied.Although exposure was limited in some instances, the impact was high as most nurses could recall details about the CSDs, even when they had seen them more than a year ago.In addition, it was interesting to note how many CSDs are identified informally (e.g.neighbours inform nurses).When asking nurses what they required to enhance their service delivery it became clear that knowledge (education), skills, a positive attitude and raising community awareness were high priorities.no similar kits exist that can address this need at a primary health care level.
LIMITATIONS OF THE RESEARCH
• English was used as the medium for data collection as all nurses reported that they felt proficient in using it.However, the richness of the data provided may have been enhanced had the study been conducted in the nurses' various mother tongues.
• Limited available research information regarding this particular topic exists for the South African context, and therefore the researcher had to rely on information obtained from other developing countries.
IMPLICATIONS FOR FURTHER RE-SEARCH
The results from both the quantitative and qualitative data have definite implications for service delivery to CSDs and their primary caregivers.Data clearly indicated that nurses felt inadequate about the way in which they handle CSDs and their primary caregivers and that they required more knowledge and skills that would in turn have a positive impact on their attitudes.It was also evident that nurses are well aware of CSDs and their primary caregivers' problems and that they are trusted by primary caregivers regarding discussions of their personal problems.This would make the nurses ideally suited to provide training to primary caregivers of CSDs so that they would be able to stimulate their own children.It was also clear that many primary caregivers in the Moretele Health District have poor parenting skills and that nurses would like to educate them.Furthermore, there are limited resources in the area, making appropriate referrals difficult.This might impact on the poor feedback that nurses receive (primary caregivers probably do not take their children to the referred places due to poverty, transport issues, etc.).This makes it clear that the training of nurses should focus on equipping nurses to train these primary caregivers so that their CSDs can remain in the community and that they do not have to move away.
25% had spent 6 to 10 years and 20% had spent 4 to 5 years in the present clinic.Given the high rate of staff turnover during the past decade, the stability of the nursing workforce in the Moretele Health District is important.
were used for identification purposes and to ensure that all clinics were visited.Gender, birth date, nursing status, years of experience as a nurse and experience in the present setting were used for descriptive purposes as all these factors could im pact on the type of responses provided.This inform ation is im portant for com piling a profile of the nurses as these aspects im pact on knowledge and skills as well as on the type of services they provide.regarding nurse's exposure to the following disability types: intellectual im pairm ent, cerebral palsy (CP), deafness, blindness, LNFS and an open category for "other".Exposure was m easured in term s of frequency, and included the following categories: never, every week, every 2 weeks, every m onth and less than once a m onth.Exposure is an im portant factor as this directly im pacts on perceptions and knowledge about the different disability types (i.e.intellectual im pairm ent, cerebral palsy (CP), deafness, blindness, LNFS and multiple disabilities) categories described in the W hite paper for the transform ation of the health system in South Africa (Government Gazette 17910, 1997:24), as priority areas for children, namely HIV/AIDS, Tuberculosis and chronic diseases, e.g.asthm a, were also included.Nurses were asked whether they regarded these conditions as very serious, serious, average, mild or as no problem in their particular work context.This is a crucial elem ent in determining perceptions, as nurses will have no desire to acquire skills and knowledge in a particular area if they do not regard this area as a serious problem .
to determine the type of service provided to children with disabilities, i.e. direct referral to hospital/therapists; direct referral to schools; give advice or counsel parents and refer; obtain case history, screen and refer or interact with the child and provide guidelines to handle the child without referral.This information is important, as it is known that direct referral is problematic in rural areas (due to limited referral options, poverty and transport problems, little or no feedback to the referring nurse, etc.included to obtain information about the priority areas of service delivery as seen by nurses.It was important to keep this as an open-ended question in order to not miss out any data that the researcher might have overlooked in compiling the questionnaire. been involved in the Moretele Health District for approximately eight years.Before conducting the research, one month was spent observing the clinics in an informal way.This prolonged engagement ensured trust between the researcher and the participants, providing her with the opportunity to become entrenched in the culture and work ethics of the nurses and aided in the removal of certain misconceptions (e.g. the number of CSDs seen by nurses).the given health care district was visited, not only a select few.Disability and the provision of services at PHC level were not only viewed from the nurses' perspective, but information was also obtained from the health care district managers and primary caregivers of children with severe disabilities.Triangulation A number of different triangulation types were used for the purpose of the research.They are as follows: the collection of data on individual levels (questionnaires) and on group level (focus groups).Data collected from one source was used to validate data from the other source.Method triangulation A between-methods triangulation was used as the research employed qualitative data (obtained from the focus groups) as well as quantitative data (obtained from questionnaires).Investigator triangulation Researchers from diverse backgrounds (Nurse, SLP, AAC specialist) with experience in qualitative research were included in the research.This enhanced the quality of data as each researcher brought her own expertise to the research.Analysis triangulation Both qualitative and quantitative approaches were applied in analysing the data.By comparing the results of these two techniques the researcher was able to note similar patterns and thus verify findings.Multiple triangulation Multiple triangulation occurs when more than one type of triangulation is used, intensifying the advantages of the different types of triangulation, further validating the findings and obtaining a more comprehensive and satisfactory sense of the phenomenon.The present research achieved this.
of the most important aspects of ensuring credibility.After themes were developed and the data analysed, results were presented to the participants to ensure the correctness of conclusions.It provided nurses with the opportunity to correct factual mistakes and to volunteer new information, and also provided the researcher with the opportunity to record the nurses' reactions after seeing the data in the new format.Cutcliffe and McKenna (1999:376) highlighted the importance of using the actual words of the participants as they recognise their own words and respond better, as it has specific meaning for them.Stakeholder reviewStakeholders were involved throughout the process to assist with reviewing the findings, clarifying points of interpretative or descriptive disagreement, prioritising unresolved issues and collecting information on those issues.not to work in isolation, and had frequent meetings with stakeholders, experts in the nursing and disability fields, government representatives as well as primary caregivers of CSDs throughout the whole research project.Examples include the assistance of nursing experts when the questionnaire was drawn up, discussions with disability and nursing experts in determining applicable questions for the were used to ensure rich descriptive data and to enhance the transferability of the data.Results from the two groups were compared and this was also compared with data obtained from the questionnaire.
from the present research was asked to establish how dependable data presented is, and to determine the accuracy of the data (including analysis and recommendations).
This concern is echoed by the World Health Organisation (WHO), who took the unprecedented step and declared TB a global emergency due to their grave concern about the modern TB epidemic(WHO, 2000:1).This document further states that more than 1.5 million TB cases per year occur in sub-Saharan Africa.This number is rapidly increasing as a result of the lethal HIV-TB combination, each speeding up the other's progress.TB is a leading cause of death among people who are HIV-positive and accounts for 15% of AIDS deaths worldwide (WHO, 2000:2).In Africa, HIV is the most important factor determining the increased incidence of TB the past ten years.Other factors that contribute to the worsening impact of TB include malnourishment, the breakdown of health services in many areas, poverty and the emergence of multi- Five major themes crystallised, namely the nurses' perception of problems experienced by primary caregivers of CSDs, how nurses perceive their role regarding disability and the type of services they currently provide, problems they experience, their own experiences regarding CSDs and, finally, what they need in order to enhance their service delivery.Regarding problems experienced by primary caregivers aspects such as disempowerment (including a lack of knowledge, poor parenting skills, poverty, limited resources, unemployment and illiteracy), external influences (religion, culture and tradition, family and community), emotional problems (ranging from denial to anger) and the impact (e.g. a loss of status in the community) were mentioned.These findings are in accordance with literature which reports that community nurses often lack the necessary information and training to meet the unique health care needs of people with disabilities (Musholt, 1995:303).Nurses perceived their current role mostly as obtaining a casehistory, conducting a physical examination, detecting the problem, conducting health promotion and coordinating community-based rehabilitation (CBR) activities, treatment of minor medical ailments, referral and follow-up.These are also the traditional roles ascribed to nurses in the literature (Thomas, 1999:734).
-
m ary of problem s experienced b y p rim ary careg ivers of children w ith disabilities Lack of knowledge § D isability § N orm al developm ent, in particular sexual developm ent w hich results in abuse and rape § M edical issues and problem s § Poor com pliance w ith appointm ents and m edication as w ell as im m unisations that are not up-to-date § "S hop around" resulting in intervention starting very late passive: "don't take action" -no follow-ups are done attitudes of nurses not alw ays conducive to the establishm ent of rapport due to tim e constraints and staff shortages § H ow to access a disability grant § H ow to access services Poor parenting skills § U naware of seriousness / not concerned § Im pact on siblings, e.g.siblings have behaviour problem s such as juvenile delinquenc y § Focus on care, not training and stim ulation § Lack initiative § O verprotect or abandon children § D eprivation § Young m others do not care for their typically developing children (com e to school dirty, hungry, sick, etc.).This phenom enon is even m ore prevalent in C SDs Poverty § T ransport struggle to use public transport to school and clinic som e public transport does not take wheelchairs § C linic attended infrequently (cannot afford) § Lack of own hom es and thus live w ith relatives § M alnourishm ent of pregnant m others and children lim ited and scattered: people unaware of them expensive long waiting lists § W heelchairs (acquiring and repairing) § U ntrained teachers § C hildren stay at hom e w ith no or little stim ulation § N o control of quality of services the community § Hide / dump the child § Rejection by family / in-laws § Family members blame each other § Blame partners (especially in the case of an unwanted pregnancy) § Sexual abuse and sex for money (CSDs) Theme 2: How do nurses perceive their role and what they currently do?although it is difficult to provide disability specific advice.ostly general advice, e.g.regarding nutrition rovide counselling (comfort) and family colds, cuts, fever, gastro enteritis, etc. Referral § Hospital (secondary / tertiary) § Social worker § Therapist § School § Medical specialist § Religious support § Genetic clinic § Hospitals refer back to clinics despite the fact that clinics are often unable to offer any services § Referral line problematic § No feedback after having attended § Often primary caregivers do not attend referred service due to transport problems Services provided Follow-up to monitor § Difficult, due to limited time and manpower Nutrition Hygiene transport problems Difficult, due to limited time and manpower Oftern primary caregivers do not attend referred service due toTheme 3: Problems experienced by nurses (This information was spontaneously provided and not probed for) epressed urses do not feel supported by the secondary and tertiary levels of care ncapable eel that they are not helping primary caregivers of CSDs optimally as therefore have to rush patients through clinics annot refer to school nurses as they too have limited resources and ptions for placing CSDs o follow-ups are done, despite the importance thereof o co-ordination of staff on setting (e.g.school nurses and nurses at the genetic clinic see SDs more often) § Some never see any CSDs § Comment: Although exposure might be limited, the impact is high as most nurses can recall details of clients even if they had not been seen for more than a year.§ Comment: Nurses are keen to know more about disability because they acknowledge the fact that although they do not frequently see CSDs at the clinics; they know that CSDs are out there in the community.Knowledge Disability types ntellectual impairment § Cerebral palsy (CP) § Sensory deficits (deaf or blind) § Epilepsy § Hydrocephalic § Micro-cephalic § Physical appearance, e.g.strange or small for age § Neglect, e.g.blind after running into a fence § Abuse and sexual abuse § At-risk factors, e.g.multiple pregnancies, low birth weight Skills Identification evelopmental milestones (children younger than 2 years) § Through schools (e.g.slow learner, repeated Grade 1 three times) § Neighbours inform nurse who then does home-visit § Detect at birth § If missed at birth, detect at baby clinic § Go to traditional healers before clinics / hospitals § Come to clinics when applying for a disability grant Some never see any CSDs Comment: Although exposure might be limited, the impact is high as most nurses can recall details of clients even if they had not been seen for more than a year.Own acceptance of CSDs before teaching primary caregivers to accept § Love CSDs: "Have a big heart" § Interest § Not irritable § Feel supported and "cared for" § Moral and spirit of nurses should be pepped up § Empathy (not sympathy) § Address and change parental attitudes § Focus on abilities and skills § Focus on a person, not on a disability § Peers -children do not have prejudices, adults instil them § Motivated to help CSDs and their primary caregivers Community awareness campaign § Drama / demonstration about severe disability § Remove stigmatisation through education § Empower community to supply own services Community awareness campaign Drama / demonstration about severe disability Remove stigmatisation through education Empower community to supply own services Motivated to help CSDs and their primary caregivers Morale and spirit of nurses should be pepped up
Table 7 : Frequency of exposure to children with LNFS vs. comfortableness in handling them
High stress levels § "Give up" -despondent § Fear: don't know what to do / community reaction § Shock § Shame § Expectations for "typically developing child" is shattered § "Breaks you" § Worry § Anger (with God and with self)
|
2018-12-21T11:03:48.582Z
|
2002-11-03T00:00:00.000
|
{
"year": 2002,
"sha1": "5a94c9b47ea39b1d4c27a55cbdfbe5f8b6c28a0a",
"oa_license": "CCBY",
"oa_url": "https://hsag.co.za/index.php/hsag/article/download/93/87",
"oa_status": "HYBRID",
"pdf_src": "Anansi",
"pdf_hash": "5a94c9b47ea39b1d4c27a55cbdfbe5f8b6c28a0a",
"s2fieldsofstudy": [
"Medicine",
"Education",
"Sociology"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
92613279
|
pes2o/s2orc
|
v3-fos-license
|
Reallocation to lateral and early-emerging axial roots allows maize (Zea mays L.) with reduced nodal root number to more efficiently forage for nitrate
Previous simulations indicated reduced nodal root number (NRN) was promising for maize (Zea mays L.) breeding, and were partially confirmed using variation in NRN among inbreds. However, the exact mechanism was unknown, therefore manipulative experiments were conducted in hydroponics and tall solid-media mesocosms with treatments involving no nodal root excision (0% NRE) or excising either 33% or 67% of the nodal roots (NR) as they emerged under high or low levels of nitrogen (N). Reduced NRN was hypothesized to increase elongation of all remaining root classes, increase N acquisition under low N, and increase shoot mass. In both experiments, plants with 67% NRE had 12% and 19% less root fraction of total biomass, 61% and 91% greater lateral-to-axial root length ratio regardless of N levels; and 61% and 182% greater biomass of embryonic roots under low N, compared to 0% NRE for hydroponics and mesocosms studies, respectively. In hydroponics, regardless of NRE level, specific root respiration under high N was 2.6 times of low N, and was greatest at depth. Under low N in mesocosms, plants with 67% NRE had 52% greater shoot biomass, 450% greater root length at depth, and 232% greater deep-injected 15N content in the shoot relative to 0% NRE, however biomass in hydroponics did not differ based on NRE. These results reveal the mechanism by which plants with fewer nodal roots increase N capture and shoot mass by reallocation of biomass to lateral, embryonic, and first whorl nodal roots that increases foraging efficiency in solid media. Summary Reallocating root biomass from nodal roots to lateral and early-emerging axial roots allows grasses to capture more nitrogen under limiting conditions, including by increasing foraging at depth.
INTRODUCTION
Roots are the major interface between plants and the soil, with a key function to extract nutrients and water that are required for plant productivity (Smith and De Smet, 2012;Meister et al., 2014). Plants have evolved the ability to proliferate roots in response to heterogeneity of resources, belowground competition, and the inevitability of root loss (Lynch, 2018). Plants respond to resource availability by adjusting their relative shoot and root mass allocation to maximize their relative growth rate (Ågren and Franklin, 2003), and generally allocate relatively more biomass to roots if the limiting factor for growth is a soil resource (Poorter et al., 2012). The fraction of newly fixed carbon from photosynthesis allocated to roots can exceed 50%, and the proportion to roots significantly increases under edaphic stress (Lambers et al., 1996;Rachmilevitch et al., 2015). Root phenes are the general elemental units of phenotype (Lynch and Brown, 2012) influencing resource acquisition, while phene states represent the particular value a phene has taken (York et al., 2013). Studies have shown that root phene states that reduce the metabolic cost of soil exploration permit greater root growth, which improves the capture of deep nitrate (Chimungu et al., 2014;Saengwilai et al., 2014), deep water (Gao and Lynch, 2016), and shallow phosphorus (Strock et al., 2018). 'Cheap root' phene states are those that reduce the metabolic burden of the root system with benefits for plant performance, including decreased root diameter, increased allocation to cheaper root classes such as lateral roots and root hairs, reduced root cortical cell area, and reduction in root secondary growth in dicots (Lynch, 2013;Meister et al., 2014;Galindo-Castaneda et al., 2018;Strock et al., 2018).
Maize plays an important role in the global food supply and industrial production, including starch, sweeteners, oil, beverages, glue, industrial alcohol, and fuel ethanol (Ranum et al., 2014). Breeding for optimized maize root system architecture is required to obtain maximum grain yield while reducing nutrient leaching and improving drought resistance in high-input agroecosystems by increasing soil resource acquisition efficiency (Mi et al., 2016;Lynch, 2018). The mature maize root system consists of the embryonic root (ER) and postembryonic root systems. The embryonic primary and seminal roots together with their laterals are important for maize seedling vigor during early development, but the postembryonic shoot-borne nodal roots (NR) and their laterals become the dominant root system by mass as the crop matures (Hochholdinger, 2009;Hochholdinger et al., 2018). Nodal roots that are formed belowground are called crown roots, whereas those that are formed aboveground are designated brace roots (Hochholdinger, 2009;Saengwilai et al., 2014). Evolution of root system architecture towards more nitrogen-efficient states corresponds to historical maize yield trends and increased tolerance of nitrogen (N) stress in the U.S. (Hammer et al., 2009;York et al., 2015).
The 'steep, deep, cheap' (SCD) ideotype of the maize root system has been proposed for optimal N and water foraging, and consists of root architectural, anatomical, and physiological phene states to increase rooting depth and N and water use efficiency in specific environments (Lynch, 2013;York and Lynch, 2015). Nodal root (NR) number is an aggregate phene consisting of the number of nodal whorls and the number of roots per whorl (Saengwilai et al., 2014;York and Lynch, 2015), and flowering time played a key role in shaping NR number via indirect selection during maize domestication (Zhang et al., 2018). The functional-structural plant model SimRoot predicted that reduced NR number could enhance maize growth under low levels of N by increasing growth of lateral roots and N acquisition (York et al., 2013;York, 2014). Field and greenhouse studies later confirmed maize genotypes with fewer nodal roots had greater rooting depth and ability to acquire N from deep soil of low N soils (Saengwilai et al., 2014), and improved water acquisition from drying soil (Gao and Lynch, 2016). However, these studies have relied on intraspecific diversity to create root phene variation, which could be confounded by the multitude of differences that potentially exist among genotypes that were not necessarily measured or reported (Lynch, 2011). In this study, we used root excision to manipulate NR number of maize inbred line B73 grown in growth chamber hydroponics and in greenhouse tall mesocosms to explore the effect of reduced NR number on N acquisition. The objective of this study was using maize plants with the same genetic and phenotypic background to test the hypotheses that reducing NR number would (1) increase elongation of embryonic roots (ER) and the remaining NR and their laterals due to reallocation of carbon, (2) increase shoot biomass in all conditions due to reallocating carbon to the shoot, and (3) increase N acquisition in low N environments.
Specific root classes are referred to as ER and NR, which are further grouped by whorl with appended numbers referring to order of emergence; therefore NR1 is the first whorl of NR that emerges.
67% NRE had no impact on SRR by root mass across all root classes under both N levels (Supplemental Under high N conditions, 67% NRE led to 55% reduction of root biomass (Supplemental Fig. S2A) and 47% increase of SRR by mass at the depth of 90-120 cm (Supplemental Fig. S2F). Under low N conditions, 67% NRE reduced root respiration by 40% in 60-90 cm and by 43% in 90-120 cm (Supplemental Fig. S2K), reduced SRR by length by 43% in 60-90 cm, reduced axial root length by 33% in 0-30 cm, by 51% in 30-60 cm, by 62% in 60-90 cm and by 53% in 90-120 cm (Supplemental Fig. S2P), and led to 48% increase of specific root length and 196% increase of lateral-to-axial root length ratio in 60-90 cm (Supplemental Fig. S2R). NR excision had no effect on root length and lateral root length across all depths under both N levels (Supplemental Fig. S2C, 2L, 2H and 2Q). Regardless of NR excision levels, SRR by mass increased with at the greatest depth under both N levels (Fig. 3).
The 33% NRE level decreased NR number by 34% (Fig. 4A), increased specific root length by 25% (Fig. 4K) and lateral-to-axial root length ratio by 65% (Fig. 4L), and had no impact on shoot biomass (Fig. 4B) and root mass fraction ( Fig. 4E) compared to 0% NRE across all N levels. Increasing NRE level to 67% decreased NR number by 60% (Fig. 4A) and root mass fraction by 19% (Fig. 4E), and increased shoot biomass by 35% (Fig. 4B), specific root length by 41% (Fig. 4K) and lateral-to-axial root length ratio by 91% (Fig. 4L), as compared with 0% NRE across all N levels. 67% NRE led to 5 th whorl of NR emerging 2 days earlier and 6 th whorl of NR emerging 8 days earlier compared to 0% NRE under high N level, respectively (Fig. 5). The interactions of N treatment and NRE treatment had significant effects on 15 N content in shoot, N content in shoot, and axial root length (Supplemental Table S1). The 33% and 67% NRE levels significantly increased N content in shoot by 48% and 53%, and 15 N content in shoot by 162% and 232% under low N, respectively, but these was not observed under high N ( Fig. 4F and 4G). Also, both excision levels significantly decreased axial root length under high N, but this was not observed under low N (Fig. 4I).
Removal of NR had no impact on root length and lateral root length across all root classes under both N levels (Supplemental Fig. S4A, S4F, S4C and S4H).
ER and the first four whorls of nodal roots under high N conditions, and ER and the first two whorls of nodal roots under low N conditions reached the depth of 120-150 cm at 42 DAT. In depth of 120-150 cm, 67% NRE increased root length by 630% in ER and by 690% in NR1 under low N conditions (Fig. 7D), and a trend of greater root length in ER induced by 67% NRE was also observed under high N conditions (Fig. 7C).
DISCUSSION
Reducing the construction and maintenance metabolic costs of roots has been proposed as a strategy for crop breeding to increase N and water use efficiency (Lynch, 2013(Lynch, , 2018. In this study, root system architecture (RSA) was manipulated in order to reduce the number of nodal roots (NR) to test hypotheses that this would influence RSA, increase shoot biomass in all conditions, and increase nitrogen (N) uptake in low N conditions. In hydroponics and mesocosms, 67% nodal root excision (NRE) led to increased embryonic root mass accumulation and greater overall lateral-to-axial root ratio, indicating an inherent reallocation of root system carbon to other root sinks. However, contrary to our hypothesis, 67% NRE had no effect on shoot biomass or shoot N content in either hydroponics N level, but increased shoot biomass by 35% across all N levels in mesocosms. Remarkably, under low N in mesocosms, 67% NRE allowed the proliferation of 450% more root length at the deepest depth, increased deep-injected 15 N content in the shoot by 232%, and increased shoot biomass by 52%, but no effect was observed on total root system biomass. These results indicate that maize has a remarkable ability to maintain the total root system sink strength for carbon that leads to inherent reallocation of mass and length to embryonic roots (ER), first whorl nodal roots, and lateral roots of all classes. NR excision had no effect on shoot biomass in hydroponics, possibly showing that the root system was a stronger sink than the shoot for the excess available carbon. The increased shoot biomass, N content, and 15 N content under low N in mesocosms indicates that the inherent modifications to root system architecture including greater root length of embryonic and all class lateral roots due to NRE leads to greater N uptake in solid media with positive effects on plant productivity. For the first time to our knowledge, a manipulative empirical study has confirmed that reduced nodal root number enhances N acquisition in maize under N replete conditions (York et al., 2013;Saengwilai et al., 2014) and has provided a mechanism through increased allocation to more efficient root classes.
Previous simulations using SimRoot indicated reduced NR number was a promising target for maize breeding (York et al., 2013;York, 2014), and these simulations were partially confirmed in greenhouse and field studies relying on genotypic variation within and among recombinant inbred line (RIL) populations (Saengwilai et al., 2014). Many NR-related quantitative trait loci (QTL) in maize have been reported for dry weight, length of NR (Burton et al., 2014), and approximately half of the total genetic variation in NR number was derived from QTLs for flowering time (Zhang et al., 2018). Though flowering time is related to NR number variation, the variation in NR number that is not due to flowering time will be more important for breeding. However, the NR-associated QTL on chromosomes have variation among maize populations, and no genes underlying the QTLs have been identified (Bray and Topp, 2018). Due to the fact maize RSA is controlled by many genes with generally small effects (Voss-Fels et al., 2018), genes like rt1 (Jenkins, 1930), rtcs (Hetz et al., 1996) and rtcl (Taramino et al., 2007) in mutants of maize not only regulate NR formation, but also affect the development of other root classes and plant vigor, which means controlling NR number genetically for physiological experiments is difficult. Using root excision to manipulate NR number of maize may produce severe changes in metabolic processes, hormone homeostasis, and gas fluxes (Bloom and Caldwell, 1988), possibly like the reduced assimilation rate observed in the hydroponics study. Early changes in hormone levels in the rooting zone induced by root excision (Druege et al., 2016) and possible hormonal regulation of specific transporters to facilitate nutrient acquisition (Ghanem et al., 2011), were hypothesized to partially lead to increased shoot production in all experiments. However, no effect on shoot mass was observed in hydroponics due to NRE, therefore the fundamental relation of reduced nodal rooting to shoot mass likely is not due to hormonal differences or increased carbon availability. In fact, nodal root excision eliminates the confounding effects of many phenotypic differences that exist among genotypes, and can increase understanding of adventitious root formation and function (such as nodal roots), as well as allocation of carbon to the most efficient root classes (Poorter et al., 2012;Steffens and Rasmussen, 2016). Nodal root number in maize has been now investigated using simulations (York, 2014), field phenotyping of populations (Trachsel et al., 2011), physiological studies of contrasting lines for nodal root number (Saengwilai et al., 2014), and now a manipulative physiological study, which makes this phene aggregate one of the most complete examples of using the functional phenomics pipeline as outlined in York (2019). Therefore, NR number manipulation in large mesocosms using NRE is a valuable tool for better understanding the complex relations among roots, shoots, and soil resource uptake.
Carbon partitioning is governed by the relative source and sink strengths of all organs (Poorter et al., 2012), and there are trade-offs in carbon allocation between leaf, stem and root functions (McCarthy and Enquist, 2007). In general, plants should allocate carbon to various organs such that the marginal potential gain of carbon is equal to the marginal carbon allocation (Bloom et al., 1985). Early 6 th whorl NR formation was observed at 67% NRE under high N in mesocosms, implying NR initiation is partially determined by carbon availability. The 67% NRE level significantly decreased root mass faction by 12% in hydroponics (Fig. 1E) and by 19% in mesocosms (Fig. 4E), which suggests that these architectural changes led to more efficient N uptake throughout the experiment, resulting in greater shoot growth in LN. In turn, greater shoot mass may allow greater carbon fixation and subsequent allocation to both roots and shoots, causing positive feedback for plant growth, as predicted by York (2014). In general, no differences in overall root respiration and root system mass were observed due to NR excision, indicating a remarkable capacity of maize to maintain total root system sink strength and metabolic burden. For plants supplied with low N in mesocosms, the removal of 67% of nodal roots increased shoot biomass by 52% as compared to no excision control (Supplemental Table S2), and led to significant increases of root length in the deepest soil layer and increase of lateral-to-axial root length ratio across all the rooting depths except 90-120 cm, which could enhance greater soil exploration of heterogeneous supplies of nitrate in soil (Hodge, 2004), and may explain the enhanced plant growth by reducing NR number of plants in mesocosms with LN.
Specific root length is a good indicator of the relative ability to explore soil by roots, largely driven by thinner roots, and is strongly associated with the resource absorption function (Freschet et al., 2015).
Removal 67% of NR significantly increased specific root length of NR5 under high N and NR3 and NR4 under low N in hydroponics and overall specific root length regardless of N levels in mesocosms, facilitated greater lateral-to-axial root length ratio of NR2, NR3 and NR4 in hydroponics and NR4 in mesocosms under low N, showing compensatory growth of lateral roots following root excision (Rubio and Lynch, 2007), which indicates enhanced capacity for resource acquisition (Lynch, 2013). Specific root respiration (SSR) by length and SRR by mass of the whole root system under high N were 2.0 and 2.6 times of those for roots under low N, which supports that root respiration is positively correlated with plant N concentration (Reich, 2014), and may be related to the changes in root morphological and biochemical phenes (Saglio and Pradet, 1980;Roumet et al., 2016). The reduction in SSR under low N could suggest it is adaptive in infertile conditions as a 'cheap root' phene state.
Surprisingly, the deepest roots had greater SRR (mass based) regardless of N levels, possibly due to more axial or lateral root tips of per unit of root mass in the deepest layer. Young roots or growing root tips are important in resource acquisition at a high cost of energy consumption, and respiration declines with the aging of the roots (Wells and Eissenstat, 2003). This study indicates reduced specific root respiration could be an adaptive response to infertile soil conditions, which could possibly be affected by a multitude of root anatomical phenes that were not measured here (Burton et al., 2013). Overall, the results indicate that compensatory growth by embryonic, first whole nodal, and lateral roots of all root classes allowed the plant to maintain total metabolic burden of the root system while increasing soil exploitation under low N, but that further improvements could possibly be made by targeting reduced specific root respiration while increasing allocation to root classes with greater specific root length (cheap root classes).
CONCLUSION
These results support the hypotheses that decreasing NR number increases the elongation of remaining of primary, seminal, nodal roots and their laterals, and increases N acquisition under low N condition in physical substrate. While compensatory growth of laterals was observed in the hydroponics study after nodal root excision, this did not lead to greater N uptake and shoot biomass, which indicates that excision of nodal roots doesn't directly affect shoot biomass production as might be expected due to hormonal changes and overall plant carbon budget. Rather, our results indicate that by reallocating root mass to less expensive embryonic, early emerging nodal, and lateral roots throughout the soil profile, but especially at depth, plants with fewer nodal roots were able to more effectively exploit soil resources which led to greater N uptake and shoot growth. This research also highlights the importance of the embryonic root system due to its persistence over the course of the experiment and having the greatest rooting at depth relative to other axial classes in maize near flowering. Therefore, the reduced nodal root number ideotype is confirmed as a potential target for breeding programs, and a mechanism through greater soil resource uptake by laterals and early-emerging axial roots was revealed.
Experimental Design
A hydroponics nutrient solution experiment was conducted in two growth chambers, arranged as a randomized complete block design replicated five times with a 2×2 factorial arrangement of treatments.
The factors were two levels of N supply (high-and low-N conditions), and two levels of NR removal.
Growth Conditions
Black plastic pails ( Madison, WI, USA. Seeds were surface-sterilized in 0.05% NaOCl for 15 min and rinsed three times using DI water, then pre-germinated in medium size (0.3-0.5 mm) premium sand placed in darkness at 28 ºC for 3 days, and transferred to growth chamber for 2 days. After that, uniformly germinated seedlings were washed out of the sand, wrapped around the junction between mesocotyl and coleoptiles with L800-D Identi-Plugs foam (Jaece Industries, NY), plugged in 5 cm Gro Pro net pot (Gro Pro, WA), and transplanted to a hole with diameter of 5 cm drilled into the lid with a hole saw. Each pail received one seedling, and the roots were directly submerged in nutrient solution. The nutrient solution for the high N treatment level was composed of (in µM) 190 KH2PO4, 2260 KNO3, 750 CaCl2, 380 MgSO4, 17.29 H3BO3, 2.63 ZnSO4·7H2O, 3.38 MnCl2 ·4H2O, 0.12 CuSO4·5H2O, 0.04 (NH4)6Mo7O24·4H2O, and 300 Fe(III)-EDTA (C10H12N2NaFeO8). For the low N treatment level, KNO3 was reduced to 280 µM, and the differences between potassium supplies was balanced with KCl. The nutrient solution was continuously aerated with an air pump attached to airstones, and solution pH was maintained between 5.9 and 6.1 by additions of KOH or HCl throughout the experiment.
The nodal root (NR) number manipulation treatment levels included: excising two-third of all emerged nodal roots (67% NRE), and no nodal root excision control (0% NRE). NR removal was started 2 days after transplanting, plants were observed every other day due to the continuous emergence of nodal root, and roots were excised as needed to maintain the 67% target. Foam was gently squeezed to make counting of nodal roots emergence for both root excision and no root excision control treatments, and nodal roots were then excised close to its corresponded internode using scalpel for root excision treatment only. The targeted 67% NRE target was based on the calculation made from NR numbers of plants with excision.
Sample collection and measurements
One day before harvest, leaf gas exchange of the second youngest fully expanded leaf was measured in the growth chamber with a LI-6800 Portable Photosynthesis System equipped with the Multiphase Flash Fluorometer (LI-COR Inc., Lincoln, NE, USA). Leaf chamber conditions were as follows: fan speed of 10000 rpm, flow rate of 600 µmol mol -1 , overpressure of 0.2 kPa, vapor pressure deficit at leaf of 1.2 kPa, leaf temperature of 25 ºC, saturating irradiance of 1200 µmol m -2 s -1 , and reference CO2 at 400 µmol mol -1 . CO2 exchange was logged manually using stability criteria of both ΔH2O and ΔCO2 stdev limit of 0.1 over a period of 10 s.
Plants were harvested 24 days after transplanting when high N plants were at the 9-leaf stage and low N plants were at the 7-leaf stage. Shoots were dried at 60 ºC for 3 days prior to dry weight determination. After a plant was removed from the nutrient solution, roots were immediately separated to the embryonic root system (including primary, seminal roots and their laterals) and different whorls of nodal roots, and each class was further divided into 30-cm sections along the axial roots of that class for an indication of rooting at various depths. All roots for a class and depth sample were blotted using tissue paper to remove excess water, then placed in a 19 ml custom chamber connected to the LI-8100 Automated Soil CO2 Flux System. The observation duration was 90 s, and the dead band was set at 20 s. All the respiration measurements were performed in an air-conditioned room with temperature maintained at 24 ºC. After root respiration measurements, roots from each section were scanned as described below.
Following scanning, the roots were dried at 60 ºC for 3 days and weighed.
Greenhouse Mesocosm Study
The mesocosm experiment was conducted in a greenhouse of Noble Research Institute from July 30 to September 10, 2017, and arranged as a randomized complete block design replicated four times with a 2×3 factorial arrangement of treatments. The factors were two levels of N supply (high-and low-N conditions), and three levels of NR excision.
Growth Conditions
The Kaeppler, University of Wisconsin, Madison, WI, USA. Seeds were surface-sterilized in 0.05% NaOCl for 15 min and rinsed three times using DI water, and were then pre-germinated in rolled germination paper (Anchor Paper, MN, USA) soaked in 0.5 mM CaSO4 and placed in darkness at 28 ºC in a germination chamber for 2.5 days. Two days before planting, each mesocosm received 6 L of nutrient solution. Each mesocosm received one seedling with the seed at a depth of 5 cm. Each plant was watered with 200 mL nutrient solution every other day. The nutrient solution for the high N treatment was composed of (in µM) 500 KH2PO4, 6000 KNO3, 2000CaCl2, 10006Mo7O24·4H2O, and 77 Fe(III)-EDTA (C10H12N2NaFeO8).
For the low N treatment, KNO3 was reduced to 600 µM, and the differences between potassium supply was balanced with KCl. KOH was used to adjust nutrient solutions to pH 6.0. Three NR number manipulation treatments consisted of removing approximately one-third of all emerged nodal roots (33%), removing two-third of all emerged nodal roots (67%), and no root excision control (0%). NR excision was started 6 days after transplanting and continued every other day till the end of the experiment due to the continuous emergence of NR.
The growth media permitted root excision without damaging remaining roots. The approach of NR excision was to brush solid media away from the base of the stem where nodal roots emerged, use a scalpel to excise the necessary roots near their base, cover the stem again with the media, and give 200 ml nutrient solution to each plant in order to allow the media to settle. The targeted NRE levels of 33% or 67% were based on the calculations made from NR numbers of the corresponding plants with no excision. The no root excision control (0% NRE) received the same procedure for counting NR number over time, but none were excised.
Sample collection and measurements
The ability of roots to acquire N from deep soil layers was quantified using deep injection of 15 NO3 -(98% atom) 1 day before harvest. One hole was drilled at 143 cm depth in each mesocosm, and 5 ml of were carefully washed away from roots using a low pressure water hose. After washing, roots were separated into embryonic roots (including primary, seminal roots, and their laterals) and different whorls of NR classes, and the roots of each class were divided into 30-cm sections along the axial roots for quantifying root distribution over depths for each class individually.
Root Scanning and Image Analysis
For both the hydroponics and mesocosm experiments, roots from each class and depth sample were spread in a 5-mm layer of water in transparent plexiglass trays and imaged with a flatbed scanner equipped with a transparency unit (Epson Expression 12000XL, Epson America) at a resolution of 600 dpi. Images were analyzed using WinRhizo software (WinRhizo Pro, Regent Instruments, Quebec, CA) individually so that diameter thresholds could be adjusted for each image to distinguish axial roots from lateral roots, which would not be possible using batch mode. Root mass fraction was calculated as root dry weight proportion of total plant dry weight. Specific root length was calculated by dividing root length by the corresponding dry weight. The oven-dried root mass and root length quantified using WinRhizo were used to calculate the specific root respiration by root dry mass (SRR by mass; nmol g -1 s -1 ) and the specific root respiration by root length (SRR by mass; nmol cm -1 s -1 ), respectively. Individual axial root length was calculated by dividing total axial root length of each root class by the corresponding number of axial roots.
Statistical Analysis
Statistical analyses were conducted using R version 3.5.1 (R Core Team, 2018) through RStudio version 1.1.45 (RStudio, 2016). Prior to data analysis, all the data were checked for normality with Shapiro-Wilk test and log-transformed when necessary. Regardless of data transformation, results presented in tables and figures report non-transformed values. Analysis of variance (ANOVA) was performed across nitrogen (N) levels and nodal root excision (NRE) levels for comparisons between N levels, NRE levels, and their interaction based on the phenes measured at the whole plant level (not separated by root class or depth). N levels, NRE levels and their interactions were analyzed as fixed effects in a mixed effects model with block as random effects. Root distribution data across root classes and rooting depths were analyzed with one-way ANOVA under each N level separately, and with pairwise comparisons of NRE levels under either same root classes or same rooting depths, because the focus was on NRE levels and otherwise the models were difficult to interpret and report. Tukey's test at α = 0.05 was used for multiple comparisons among different treatments. The R packages 'emmeans' (Lenth, 2018), 'lme4' (Bates et al., 2014), 'multcompView' (Graves et al., 2012), were used for statistical analyses and the R package 'ggplot2' (Wickham, 2016) for data visualization.
Supplemental Data
The following supplemental materials are available. Table S1. Summary of ANOVA (F values and significance lvels) of whole-plant shoot and root phenes as influenced by N treatment, NR excision treatment and their interactions for the two experiments. Table S2. Mean ±SE of shoot biomass as influenced by NR excision treatment under low N conditions in the greenhouse mesocosms experiment. Figure S1. Effect of NR excision treatment on root phenes of different root classes under high N (A, B, C, D, E, F, G, H) and low N (I, J, K, L, M, N, O, P) level in growth chamber hydroponics study. Each column represents the mean of n=5 for each treatment. Error bars represent standard errors, and columns with the same letter under same root class were not significantly different at p≤0.05 according to Tukey's test. ER: embryonic roots, NR1: 1 st whorl nodal roots, NR2: 2 nd whorl nodal roots, NR3: 3 rd whorl nodal roots, NR4: 4 th whorl nodal roots, NR5: 5 th whorl nodal roots, SRR: specific root respiration. Supplemental Figure S4. Effect of NR excision treatment on root phenes of different root classes under high N (A, B, C, D, E) and low N (F, G, H, I, J) level in in the greenhouse mesocosms. Each column represents the mean of n=4 for each treatment. Error bars represent standard errors, and columns with the same letter under same root class were not significantly different at p≤0.05 according to Tukey's test. ER: embryonic roots, NR1: 1 st whorl nodal roots, NR2: 2 nd whorl nodal roots, NR3: 3 rd whorl nodal roots, NR4: 4 th whorl nodal roots, NR5: 5 th whorl nodal roots, NR6: 6 th whorl nodal roots. Figure S5. Effect of NR excision treatment on root phenes of different rooting depths under high N (A, B, C, D, E) and low N (F, G, H, I, J) level in greenhouse mesocosms study. Each column represents the mean of n=4 for each treatment. Error bars represent standard errors, and columns with the same letter under same depth were not significantly different at p≤0.05 according to Tukey's test.
ACKNOWLEDGMENTS
We thank Bryce Walker, Anand Seethepalli, Na Ding for assistance sampling and root scanning, and Effect of NR excision treatment on root biomass of different root classes under high N (A) and low N (B) conditions in growth chamber hydroponics study. Each column represents the mean of n=5 for each treatment. Error bars represent standard errors, and columns with the same letter under same root class were not significantly different at p≤0.05 according to Tukey's test. ER: embryonic roots, NR1: 1 st whorl nodal roots, NR2: 2 nd whorl nodal roots, NR3: 3 rd whorl nodal roots, NR4: 4 th whorl nodal roots, NR5: 5 th whorl nodal roots. Figure 3. Effects of rooting depths on SRR by root mass under high N and low N conditions in growth chamber hydroponics study. Each column represents the mean of n=10 for each depth. Error bars represent standard errors, and columns with the same letter under same N level were not significantly different at p≤0.05 according to Tukey's test. SRR: Specific root respiration. of different root classes under high N and low N conditions in the greenhouse mesocosms. Each column represents the mean of n=4 for each treatment. Error bars represent standard errors, and columns with the same letter under same root class were not significantly different at p≤0.05 according to Tukey's test. ER: embryonic roots, NR1: 1 st whorl nodal roots, NR2: 2 nd whorl nodal roots, NR3: 3 rd whorl nodal roots, NR4: 4 th whorl nodal roots, NR5: 5 th whorl nodal roots, NR6: 6 th whorl nodal roots. Each column represents the mean of n=5 for each treatment. Error bars represent standard errors, and columns with the same letter under same root class were not significantly different at p≤0.05 according to Tukey's test. ER: embryonic roots, NR1: 1 st whorl nodal roots, NR2: 2 nd whorl nodal roots, NR3: 3 rd whorl nodal roots, NR4: 4 th whorl nodal roots, NR5: 5 th whorl nodal roots. of different root classes under high N and low N conditions in the greenhouse mesocosms. Each column represents the mean of n=4 for each treatment. Error bars represent standard errors, and columns with the same letter under same root class were not significantly different at p≤0.05 according to Tukey's test. ER: embryonic roots, NR1: 1 st whorl nodal roots, NR2: 2 nd whorl nodal roots, NR3: 3 rd whorl nodal roots, NR4: 4 th whorl nodal roots, NR5: 5 th whorl nodal roots, NR6: 6 th whorl nodal roots.
Figure 7.
Effect of NR excision treatment on root length distribution across rooting depths under high N level (A) and low N (B) conditions, and root length of different root classes in 120-150 cm (C, D) in the greenhouse mesocosms. Each column represents the mean of n=4 for each treatment. Error bars represent standard errors, and columns with the same letter under same depth or root class were not significantly different at p≤0.05 according to Tukey's test. Table S1. Summary of ANOVA (F values and significance levels) of whole-plant shoot and root phenes as influenced by N treatment, NR excision treatment and their interactions for the two experiments. Figure S1. Effect of NR excision treatment on root phenes of different root classes under high N (A, B, C, D, E, F, G, H) and low N (I, J, K, L, M, N, O, P) level in growth chamber hydroponics study. Each column represents the mean of n=5 for each treatment. Error bars represent standard errors, and columns with the same letter under same root class were not significantly different at p≤0.05 according to Tukey's test. ER: embryonic roots, NR1: 1 st whorl nodal roots, NR2: 2 nd whorl nodal roots, NR3: 3 rd whorl nodal roots, NR4: 4 th whorl nodal roots, NR5: 5 th whorl nodal roots, SRR: specific root respiration. Supplemental Figure S4. Effect of NR excision treatment on root phenes of different root classes under high N (A, B, C, D, E) and low N (F, G, H, I, J) level in in the greenhouse mesocosms. Each column represents the mean of n=4 for each treatment. Error bars represent standard errors, and columns with the same letter under same root class were not significantly different at p≤0.05 according to Tukey's test. ER: embryonic roots, NR1: 1 st whorl nodal roots, NR2: 2 nd whorl nodal roots, NR3: 3 rd whorl nodal roots, NR4: 4 th whorl nodal roots, NR5: 5 th whorl nodal roots, NR6: 6 th whorl nodal roots.
|
2019-04-03T13:09:44.175Z
|
2019-01-29T00:00:00.000
|
{
"year": 2019,
"sha1": "39793e13f3f37a1a6a2b971285e90046fd92f678",
"oa_license": "CCBYNC",
"oa_url": "https://doi.org/10.1093/jxb/erz258",
"oa_status": "GREEN",
"pdf_src": "BioRxiv",
"pdf_hash": "22db5d3af90cf52fe73d3ee409301b0c3fcb4120",
"s2fieldsofstudy": [
"Agricultural and Food Sciences"
],
"extfieldsofstudy": [
"Biology"
]
}
|
33612716
|
pes2o/s2orc
|
v3-fos-license
|
An accurate optical technique for measuring the nuclear polarisation of 3 He gas
. In the metastability exchange optical pumping cells of our on-site production unit and of our other experimental set-ups, we use a light absorption technique to measure the 3 He nuclear polarisation. It involves weak probe beams at 1083 nm, that are either perpendicular or parallel to the magnetic field and cell axis, with suitable light polarisations. When metastability exchange collisions control the populations of the sublevels in the 2 3 S state, absolute values of the 3 He ground state nuclear polarisation are directly inferred from the ratio of the absorption rates measured for these probe beams. Our report focuses on the transverse detection scheme for which this ratio, measured at low magnetic field for σ and π light polarisations, hardly depends on gas pressure or the presence of an intense pump beam. This technique has been systematically tested both in pure 3 He and isotopic mixtures and it is routinely used for accurate control of the optical pumping efficiency as well as for calibration of the NMR system.
Introduction
Metastability exchange optical pumping (MEOP) allows the polarisation of 3 He nuclear spins in pure helium gas and this technique is successfully used for production of highly polarised 3 He with large throughput. MEOP requires the electronic excitation of a small fraction of the He atoms to the metastable (2 3 S) state and involves two co-operative processes: strong selective excitation on the 2 3 S−2 3 P optical transition line at 1083 nm with transfer of angular momentum from polarised light to metastable He atoms; metastability exchange collisions with, for 3 He, transfer of nuclear polarisation from the metastable state to the ground state [1]. In pure 3 He gas, the driven radiative transitions enforce simultaneous orientation of the electronic and nuclear spins in the 2 3 S state, entangled by the hyperfine interaction. In isotopic gas mixtures, radiative excitation can selectively enforce orientation of 4 He electronic spin in the metastable state but 2 3 S polarisation builds up also for the 3 He isotope through metastability exchange between ground state 3 He and optically pumped 4 He atoms. For a detailed description of the MEOP process, we refer the reader to the companion contribution on MEOP basics in the same conference proceedings [1] as well as to comprehensive regular articles [2,3,4].
Radiofrequency (rf) discharges are frequently used for electronic excitation of a small (ppm) fraction of He atoms to the metastable state and MEOP performs better at low or moderate pressures (typically 0.5−5 mbar for operation at mT field strengths and 5−250 mbar for operation between 0.4 T and 4.7 T). Gas production units thus typically include a combination of several systems for high-purity gas handling, high-power optical pumping, polarisation preserving compression, and long-term or short-term storage for massive centralised production or flexible on-site production, respectively. Diagnostic systems are also implemented for monitoring of the operating conditions and optimisation of the performances of the gas production units. NMR or magnetometry can be used to measure polarisation in the storage cell where large magnetisation is accumulated. In the optical pumping cell, optical methods such as polarimetry of the visible fluorescence light emitted by the plasma [5,6,7] or 1083 nm light absorption measurements [2,4,8,9] are preferred because they provide much higher sensitivity and operate in the presence of the rf discharge.
All these techniques require calibration for accurate polarisation measurements, except the light absorption technique. The latter has actually been used for calibration purposes in the past decades [10,11]. In contrast with the visible light polarimetry technique, that becomes inefficient above approximately 10 mT due to strong hyperfine decoupling in the highly excited energy levels involved in the radiative cascade, the absorption technique potentially operates at all field strengths [4]. It has been successfully used for high field MEOP investigations [4,12,13]. At low magnetic fields, recent work [14] has motivated more quantitative investigations of its potential artefacts in the context of very high power optical pumping.
In this contribution, we describe an accurate way to measure the 3 He nuclear polarisation at low magnetic field with the absorption technique, in various MEOP conditions. We report on systematic dedicated tests performed in small sealed He cells filled with pure 3 He gas or isotopic gas mixtures, we address the practical aspects of the experimental implementation, and quantitatively discuss the most relevant identified sources of errors in polarisation measurements.
Polarisation measurements with the light absorption technique
The absorption technique is based on the very strong link between the ground state nuclear polarisation M and the distribution of populations between the three (for 4 He) and six (for 3 He) Zeeman sublevels in the metastable state that is established by the frequent exchange collisions between the 2 3 S atoms and the much more numerous ground state atoms. When metastability exchange dominates, a spin temperature distribution is enforced and the populations of the 2 3 S sublevels scale exponentially with their respective angular momentum with a Boltzmann parameter e β =(1 + M )/(1 − M ) [4]. The polarisation M can then be simply inferred from the ratio of populations in two 2 3 S sublevels with distinct angular momenta for either isotope, or from independent linear combinations of such populations.
At low magnetic field, the 2 3 S Zeeman sublevels are almost degenerate and two absorption measurements can be simultaneously performed with a monochromatic light source using different light polarisations. Three 2 3 S-2 3 P 0 lines are well resolved in a low-pressure gas: the C 8 line (2 3 S, F =1/2−2 3 P 0 ) and the C 9 line (2 3 S, F =3/2−2 3 P 0 ) in pure 3 He, and the 4 He D 0 line (2 3 S−2 3 P 0 ) in isotopic mixtures. Picking up one of these lines, one may choose among the following simple schemes: 1/ Transverse probe: a weak laser beam propagates perpendicular to the axis of the magnetic field B and transmitted powers are measured for the orthogonal linear polarisations π (parallel to B) and σ (perpendicular to B). M is inferred from the ratio r ⊥ =A π /A σ of the corresponding absorption signals (see Section 3.5) A π and A σ , respectively. At zero magnetic field, the ground state 3 He nuclear polarisation is given by: and B=1 mT. For higher magnetic fields, they can still be used provided that r ⊥ is replaced by the normalised ratio R ⊥ =r ⊥ (M )/r ⊥ (0) obtained by scaling r ⊥ to its value at null polarisation, r ⊥ (0). If the probe laser is not tuned to the centre of the null-field or null-polarisation absorption profile, the detuning introduces a field-dependent error on |M | of order -0.02/GHz/mT. Polarisation cannot be measured with probe light tuned to C 8 because this line addresses the same pair of 2 3 S sublevels (F =1/2, m F =±1/2) for π and σ polarisations, hence the ratio r 8 ⊥ does not depend on M . The C 9 line cannot be used for polarisation measurements in mixture cells because it overlaps with the strong unresolved D 1 and D 2 lines (2 3 S−2 3 P 1 and 2 3 S−2 3 P 2 , respectively) of the 4 He isotope. Therefore, absorption measurements are performed with the probe laser tuned to the C 9 line in pure 3 He gas and to the D 0 line in isotopic mixtures.
2/ Longitudinal probe: the weak laser beam propagates parallel to B, transmitted powers are measured for the two circular polarisations σ + and σ − , and M is inferred from the ratio r =A σ+ /A σ− of the absorption signals A σ+ and A σ− , respectively. At zero magnetic field, the ground state 3 He nuclear polarisation is given by: The polarisation components σ + and σ − individually address a single sublevel for these two single-component lines (m F =−1/2 and +1/2, respectively, for C 8 and m J =−1 and +1, respectively, for D 0 ) and the ratio r directly measures at B=0 the ratio of populations in the probed Zeeman sublevels. As is done for the transverse probe, r can be replaced by the normalised ratio R =r (M )/r (0) in formulas (3) and (4) to obtain accurate results in an applied field. Moreover, the use of reduced ratios also eliminates the error introduced by probe detuning for the longitudinal probe scheme.
No simple analytical formula exists for the C 9 line for which each polarisation component addresses simultaneously two Zeeman sublevels. Instead, the relation: has to be solved (e.g., finding the root of a cubic polynomial, numerically, or using polynomial fits...) to infer M from r 9 . Figure 1 shows the variations of r ⊥ and r with M . For the transverse probe scheme, C 9 and D 0 lines provide a comparable sensitivity for polarisation measurements, quite high above |M | ≥ 0.2. The sensitivity is reduced at small M , with a quadratic departure of r ⊥ from 1 that results from the fact that each light component simultaneously addresses sublevels with opposite angular momenta. In contrast, the longitudinal probe scheme has a linear response at small M for all lines, with a higher sensitivity for the C 9 and D 0 lines. At high M , it has a reduced sensitivity for C 9 and D 0 with r ∝ (1 − M ) 2 ; the C 8 line retains a linear sensitivity, yet provides a limited precision since both absorption signals vanish at M =1.
During optical pumping, a strong pump beam tends to enforce a distribution of populations in the 2 3 S states that departs from the spin-temperature distribution that would result from metastability exchange only. The above equations then provide apparent polarisations M app that depend on the choice of detection scheme and probe line and differ from the actual nuclear polarisation M . Examples of experimental deviations of M app from M are discussed in Section 5.
Probe laser
The probe laser consists of a linearly polarised 70 mW DFB laser diode (Toptica Photonics AG) housed within a temperature-regulated aluminium box for wavelength stability. Inside this box the divergent elliptical beam from the laser diode is collimated and then reshaped to circular with a pair of anamorphic prisms. The beam passes through a Faraday isolator before being coupled into a polarisation preserving monomode optical fibre as it exits the aluminium box. The laser can thus be situated remotely to the helium polarising equipment for both convenience and reduced impact on the magnetic field map. Typically 15 milliwatts are coupled into the fibre when the laser diode is running at maximum current. Laser stability and control are achieved with an in-house electronic controller. The aluminium box is heated to a fixed set point above room temperature with a resistive element for long term thermal stability and the laser frequency is adjusted using the on-chip thermoelectric cooler. Figure 2 shows a schematic diagram of the probe assembly and its caption includes all technical details. The laser diode light that exits the optical fibre tip is collimated and split in two adjacent beams with σ and π polarisations. This is in contrast with prior implementation [10], in which probe beams spatially overlapped. The beam separation preserves the accuracy of the polarisation measurements for all probe tunings, as discussed in Section 5.1. The beams propagate across the helium cell, entering and exiting through the curved glass walls. A cylindrical lens is used to correct for the resulting wavefront distortions before back-reflection through the lens and cell. One of the beams is then deflected by a polarising beam splitter (PBS) cube. Both beams are coupled by lenses into 1-mm core multimode plastic fibres that are connected to the photodiodes. The photodiode voltages are recorded on a PC using a USB data logger for post processing.
Pump laser
For optical pumping a 5 Watt broadband (1.7 GHz fwhm) 1083 nm fibre laser is used (Keopsys, model YFL-1083-50-COL). The collimated Gaussian output beam is circularly polarised and expanded to a diameter of 2 cm fwhm. It propagates along the direction of the magnetic field and passes twice through the cell. The transmitted light, deflected by the circular polariser, is monitored by a photodiode for checks of the laser tuning and measurements of the pump power absorption.
Helium test cells
The sealed helium cells used for these investigations are pyrex glass tubes with flat end windows (5 cm in diameter 5 cm long). They contain 3 He at various gas pressures, either pure or mixed with set ratios of 4 He. However some of the "pure 3 He" cells actually contain traces of 4 He plausibly left from the cleaning process prior to the filling of the cells. 1 A discharge is generated in the cell using two external wire electrodes connected to a tuned voltage transformer fed by a rf generator able to produce 5 W in the 1 to 3 MHz range. This rf excitation is amplitude modulated at around 70 Hz and lock-in detection is used to improve sensitivity of the absorption measurements.
Absorption signals
The absorption signals A π and A σ used for nuclear polarisation measurement are the amplitudes of the modulation depth (i.e., the ratios of the ac amplitudes to the dc components) of the corresponding probe powers exiting the cell. Each ratio is robust against fluctuations of the incident power I 0 and directly proportional to the absorbance −ln (I/I 0 ), with a coefficient that only depends on the plasma response to the rf excitation [4,15]. For improved accuracy, the PBS cube (transmission efficiency: 1-=0.96) is oriented so as to transmit the π component for which the absorbance is very small at high nuclear polarisation (A π −→0 for M −→1) and to deflect the σ component. In this configuration, the amplitudes S π and S σ of the modulation depth of the photodiode voltages are related to the absorption signals by S π =(1-)A π and S σ =A σ + A π . This is used to compute the ratio r ⊥ of absorption signals from processed photodiode voltage data:
NMR detection
For comparison with optical measurement results, NMR measurements have been performed at 28 kHz (the Larmor precession frequency for B=0.87 mT) using a homebuilt NMR detection system that consists of crossed induction and detection coils set around the test cell. Since the gas density is low, 90 • tipping pulses are routinely used to obtain good signal-to-noise ratios (typically up to 100), with tens of seconds of precession lifetimes T 2 . NMR signals are proportional to the fixed number N 3 of 3 He atoms contained in the glass cell and to the nuclear polarisation M . They have to be scaled for cell content and volume for direct comparison to the nuclear polarisations measured using the absorption technique. The gas pressures slightly differ from the filling pressures for the sealed cells that are several decades old and the actual pressures are measured by NMR. For each cell, the decay rates of free precession signals, 1/T 2 , have been measured for various applied magnetic field gradients to obtain the 3 He diffusion coefficient D [16]. D is inversely proportional to gas pressure P for pure 3 He gas (D=1.997×[T /300] 1.71 /P , with temperature T in K and P in atm) and depends on the partial pressures of 3 He and 4 He gas for isotopic mixtures (see reference [17], and corrected values of diffusion coefficients in [18]). Figure 3 shows probe recordings and processed data from a typical polarisation experiment. During the acquisition, the photodiode voltages and TTL reference signal are continuously recorded. Figure 3a shows a small section of the recorded transmitted probe powers in which the modulation of the absorption is clearly seen. Initially (part C) the gas is not polarised (M =0) and the modulation depth of the π and σ components are equal; later (part D) the gas is polarised (M =0.5) and both absorption rates and modulation depths are different for the two components. Figure 3b shows the time evolution of the average absorption signals and computed nuclear polarisation during the entire experiment. The plot is divided into six regions that are related to the six main steps of the acquisition protocol, as follows:
Polarisation measurements
(A) the laser beams are blocked and the rf discharge is off; offset voltages are recorded. (B) the probe block is removed and discharge is off; incident probe powers I 0 are measured. (C) the rf discharge is on; transmitted probe powers I are measured at null polarisation.
(D) the pump block is removed; optical pumping takes place and polarisation increases.
(E) the pump beam is blocked again when polarisation is close to steady-state; polarisation decreases due to plasma-induced decay for some arbitrary time period.
(F) the rf discharge is turned off; NMR signal is acquired to measure the magnetisation associated with the remaining polarisation M . Nuclear polarisation is fully destroyed in the cell by the 90 • rf pulse used for the NMR measurement. MEOP is performed to repolarise the gas and another measurement can be made for a different polarisation M after a shorter or longer decay period. Excellent linear correlation is obtained between the two independent sets of polarisation measurements (NMR and optical probe). In all the pure 3 He cells the absorption technique is checked to yield data that consistently agree with the NMR data scaled to actual gas pressures. In the mixture cells, the absorption technique using the D 0 line yields data that also fully agree with the NMR data scaled to 3 He partial gas pressures. This agreement confirms that the nuclear polarisation of ground state 3 He atoms imposes a spin temperature distribution in the 4 He metastable state as well.
Implementation and design
The absorption measurements are performed on the same infrared transition as the optical pumping. Contribution to the measured transmitted powers from the 1083 nm fluorescence light emitted by the gas subjected to intense pumping should thus be limited, as usual, by reducing the detected solid angle to a minimum. Alternatively, when fluorescence light and stray light from the intense pump laser cannot be avoided, a double modulation scheme (e.g., with a probe frequency swing in and out resonance or mechanical beam chopping) may help. The stray light contribution is a minor issue for the transverse probe scheme, when compared to the longitudinal one. However the absorbed probe power (hence the signal-to-noise ratio) is generally lower with transverse probe scheme, especially in the long cells used for gas production. In order to measure absolute ratios of populations, the two probe components should ideally overlap exactly. However, in this situation, polarisation components reaching the same 2 3 P Zeeman sublevel may interact. The scaled ratio of absorption signals R ⊥ = r ⊥ (M ) /r ⊥ (0) has been observed to exhibit small frequency-dependent variations at finite nuclear polarisations, both when the probe is tuned to the C 9 line in 3 He gas and when it is tuned to the D 0 line in isotopic gas mixtures. Similar observations have been made with the longitudinal probe scheme for R = r (M ) /r (0) in pure 3 He gas when absorption measurements are performed on the C 9 line, but no such artefact has been noticed when the C 8 line is used. Systematic investigations have been made for both probe schemes in a variety of experimental conditions and detailed results will be reported elsewhere. Using two separate beams for the absorption measurements, as described in Section 3.2 for the transverse probe scheme, has been checked to make the optical technique robust against probe detuning and accurate at all polarisations (for both probe schemes). These probe beams still remain very close so as to address nearly identical sets of atoms, and any small local differences (e.g., in the number density of metastable atoms) cancel out when reduced ratios R ⊥ or R are used.
Impact of the probe light intensity
The light intensity in the probe beam must be low enough so as not to induce significant distortion of the 2 3 S population distribution in the metastable state [4]. This distortion may be expected to scale as γ p /γ e in pure 3 He gas, where γ e is the metastability exchange rate and γ p is the relevant average optical pumping rate for the probed line components. γ e is proportional to gas pressure P (γ e =3.75×10 6 s −1 /mbar for 3 He [19]) and γ p is proportional to the light intensity and scales as 1/P , due to the impact of velocity-changing collisions on the fraction of the atomic Doppler profile actually addressed by the single-frequency probe light [3]. We have performed checks of this effect by measuring the ratios r 9 ⊥ and r 0 ⊥ for M =0 in two 3 He cells and in a mixture cell, respectively. They decreased linearly with decreasing probe intensity, with slopes consistent with expectations and obeying the 1/P 2 scaling for pure 3 He.
The light intensity should be reduced until it does not affect the measurements. For instance, for a probe intensity of 100 μW/cm 2 , M is overestimated at low pressure (P =0.43 mbar) by as much as 0.05 for M =0 and 0.01 for M =0.1. These errors are reduced to 0.023 (M =0) and 0.002 (M =0.1) for a probe intensity of 25 μW/cm 2 . It is possible to keep the intensity of the probe low by reducing the incident power and/or increasing the beam diameter. The second option is usually preferred because it preserves the signal-to-noise ratio. These limits on probe intensity are significantly less stringent if a broadband laser is used, such as the monitor output of the fibre laser used for optical pumping. The actual distortion of population distribution is then reduced by the ratio of linewidth (of order 100) at given laser intensity, hence intensities of order 1 mW/cm 2 may be used.
Impact of the probe light power
The probe-driven optical pumping cycles may also influence the time evolution of nuclear polarisation because they contribute to the overall angular momentum budget [1]. No angular momentum is associated to the π and σ components of the transverse probe but, since the probe light equally addresses sublevels with opposite angular momenta, it tends to more efficiently depopulate the overpopulated ones. As a result it drives the 2 3 S state towards a null average nuclear orientation and, through metastability exchange, contributes to the 3 He polarisation decay in the ground state. This may limit the MEOP performances in optical pumping experiments, in a way that is similar to what occurs for an imperfectly polarised pump beam. This may also influence the evolution of the polarisation in the absence of pump light and introduce a noticeable additional decay rate. The additional decay rate is expected to scale with the absorbed power (hence, with the 2 3 S number density) and to be inversely proportional to the number of ground state atoms (hence, to the 3 He gas partial pressure and to the cell volume). This has been experimental checked by recording polarisation decays in the plasma for various probe beam powers. Figure 4 shows that decay rates increase linearly with probe powers at fixed rf discharge excitation, as expected. Depending on the objective of the polarisation measurement, a trade-off between a high signal-to-noise ratio and a negligibly small action of the probe has to be found; probe powers in the range 0.1-1 mW are usually a safe choice. When long cells are used, the increase in their volume proportionally decreases the impact of probe power, but in the transverse scheme only.
Impact of the pump light
The deviation from spin temperature distribution of populations induced by the strong pump light, that provides the driving term for the build up of the ground state 3 He polarisation, results in modified absorption rates for the two probe beams [1]. The apparent polarisation values M app obtained using the relevant spin temperature formula (Eqs. (1) to (5), depending on the detection scheme) overestimate the actual polarisations M and the pump-induced differences M app -M can be experimentally measured when the pump is rapidly applied or blocked, at steady-state or at arbitrary M during polarisation build-up. Results of dedicated experiments performed to quantify this difference for both probe schemes in similar OP conditions are shown in Figs. 5 and 6. Data points in Fig. 5 have been individually obtained using the protocol described in Section 4 for various polarisation growth times. The uncertainties mainly result from polarisation changes during the time required to manually block the OP beam. The pump-induced differences decrease when gas pressure rises (from 0.43 to 5.4 mbar), due to the increased metastability exchange rate that competes with optical pumping for population distribution, and steadily decrease with high nuclear polarisations at all pressures. When the same line is used for OP and for detection, the pump and probe beams address the same 2 3 S Zeeman sublevels and the pump-induced differences are observed to be about twice as large as in Fig. 5 (data not shown).
Each of the four data sets in Fig. 6 has been obtained from a single recording, where short periods of polarisation build-up alternated with periods with a blocked pump. For each choice of probe line, the pump-induced differences systematically decrease again with pressure and polarisation, and they increase by a factor of five when the C 8 line is used instead of the C 9 line for the probe. When the C 9 line is used for pumping (data not shown), the pump-induced differences for each pressure are similar for the two probe lines and about twice as large as those represented by the solid symbols in Fig. 6.
Comparison of Fig. 5 to Fig. 6 shows that the transverse probe scheme is significantly less prone to the impact of pump light than the longitudinal probe scheme (5 to 10 times less, OP is made with 2 W laser output power tuned to the C 8 line (beam diameter: 1.4 cm fwhm) in 6 cm diameter 30 cm long cells filled with pure 3 He gas (squares: 0.63 mbar; circles: 1.19 mbar). Solid symbols: C 9 probe; open symbols: C 8 probe. The schematic of the longitudinal probe assembly used for the measurements is given in [1].
depending on the chosen OP transition). This is due to partial compensation of the effects on the transverse absorption signals of the pump-driven depopulation and overpopulation for pairs of 2 3 S sublevels with opposite angular momenta. For both detection schemes, pump-induced differences increase with laser power until optical saturation of the 2 3 S-2 3 P 0 occurs (typically above 1W/cm 2 ).
In previous work using a transverse probe scheme [10], accurate comparisons of the absorption and polarimetry techniques had been performed without pump light, during polarisation decay. The reported absence of noticeable difference between the polarisations measured with (M app ) and without (M ) pump beam explicitly refers to comparisons performed at steady-state polarisations. It can be explained by the limited signal-to-noise ratio (Fig. 4 of [10]) obtained with the single-pass probe path and by the high nuclear polarisations achieved with a broadband 3W laser (M ≥0.5).
For standard operating conditions, the nuclear polarisation exceeds 0.5 in the optical cells of our gas production unit and the steady-state pressure is on the order of 1.33 mbar. Overlooking the small difference between M app and M results in an absolute error smaller than 0.25% and relative errors systematically lie below 0.5% when the transverse probe scheme is used.
Conclusions
We have reported implementation and systematic tests of the transverse probe scheme for absolute polarisation measurements with the absorption technique. The consistency and high accuracy of the measurements have been demonstrated and assessed by correlation with NMR measurements as well as by comparison with numerical results obtained with a MEOP model. Potential systematic errors have been discussed and a number of simple experimental tests have been described to check that they actually remain well below the percent level. Most experiments have been performed using a single-frequency laser diode to probe 2 3 S-2 3 P 0 absorption. The technique can be applied also using a broadband light source. The overall decrease in probe light absorption has no influence on the computed M values, but data reduction becomes less straightforward as some contribution to light absorption comes from the wings of neighbouring resonance lines for very broad lasers. We have performed test measurements with the weak monitor beam provided by the supplementary output of the pump laser, but low accuracy has been obtained with this more noisy light source. Moreover, its use is restricted to situations where MEOP is performed using the C 9 or D 0 line.
The probe assembly that we have described has also been used for polarisation measurements in the long valved cell (50 cm length, 6 cm diameter) of our gas production unit to characterise the efficiency of MEOP in various operating conditions. For improved performance, a double cell with two such glass cylinders connected in series has been recently implemented in the unit. The same transverse probe assembly is currently used for simultaneous polarisation measurements in the two gas compartments. Successful measurements are performed with a configuration similar to that of Fig. 2 where the probe beam propagates successively through the two glass cells before back-reflection by the mirror. Using two different frequencies for the discharge modulation in each cell, their individual contributions to the absorption signal can be singled out by selective lock-in detection. In spite of the lower signal-to-noise ratios obtained with the transverse probe scheme, the simpler technical implementation makes this scheme quite advantageous over the longitudinal one for polarisation measurements in multiple cells with a single laser beam.
The transverse probe scheme provides a very robust way to perform absolute measurements of the 3 He nuclear polarisation in all situations where the gas pressure is not well known or variable in time. Gas production units operate either in batch mode (no gas flow) or in continuous mode (gas flow through the optical pumping cell where fresh gas gets polarised), in general with no easy and non-relaxing way to measure on-line the pressure inside the MEOP cell. The longitudinal probe scheme is more sensitive at low polarisations, but dedicated experiments must be performed to establish the pressure-dependent correspondence between M app and M . The use of this scheme may be restricted to batch mode operation or to reproducible stationary conditions of gas flow. The transverse probe scheme is therefore the best suited one for accurate on-line polarisation measurements.
|
2017-09-16T01:56:29.656Z
|
2011-06-01T00:00:00.000
|
{
"year": 2011,
"sha1": "d59f22f5e7e7c1ed3ca60134be6dce190cee8f52",
"oa_license": null,
"oa_url": "https://doi.org/10.1088/1742-6596/294/1/012008",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "6fac0eecb771827d68466d5b30ca606c18e5de26",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics",
"Chemistry"
]
}
|
261896913
|
pes2o/s2orc
|
v3-fos-license
|
The effect of abdominal bracing on respiration during a lifting task: a cross-sectional study
Background Abdominal bracing is a maneuver widely used by rehabilitation specialists and sports trainers to improve spinal stability. This study aimed to investigate how lifting tasks with and without abdominal bracing affect the respiratory function of the diaphragm. Methods M-mode ultrasonographic assessment of diaphragmatic motion combined with spirometry was performed on 31 healthy adults. Participants were asked to breathe continuously whilst lifting a load with spontaneous abdominal muscle contraction (natural loaded breathing) and abdominal bracing (AB loaded breathing). Results Pearson’s correlations revealed strong correlations between ultrasonography and spirometry measures (p < 0.001) for all types of breathing: tidal breathing (r = 0.709, r2 = 0.503), natural loaded breathing (r = 0.731, r2 = 0.534) and AB loaded breathing (r = 0.795, r2 = 0.632). Using paired-samples t-tests, the natural loaded breathing ultrasonography revealed more caudal diaphragm positions during inspiration (p < 0.001) but not during expiration (p = .101). Spirometry demonstrated lower lung volumes (L) at the end of inspiration and expiration (p < 0.001), with no changes in total lung volume (p = 0.06). The AB loaded breathing ultrasonography revealed more caudal diaphragm positions during inspiration (p = 0.002) but not during expiration (p = 0.05). Spirometry demonstrated lower lung volumes at the end of inspiration (p < 0.001), expiration (p = 0.002), and total lung volumes (p = 0.019). Conclusion This study demonstrated that abdominal bracing performed during a lifting task reduces lung volume despite an increase in diaphragmatic motion. Diaphragm excursions strongly correlate with lung volumes even under postural loading. Trial registration The study was prospectively registered on 8 April 2021 at ClinicalTrials.gov with identification number NCT04841109.
Background
The diaphragm is the first muscle activated during inspiration [1], and its function is closely related to changes in lung volumes.The diaphragm is responsible for 60-80% of the inspiratory work [2], and its contribution to lung volume changes is estimated between 75-90% [3][4][5].At higher lung volumes, the diaphragm's ability to generate force decreases as its muscle fibers shorten and change their orientation [6][7][8], resulting in its lower contribution to changes in tidal volumes and greater involvement of the extra-diaphragmatic inspiratory muscles (EIMs) [3][4][5].In addition to the respiratory function, the diaphragm is involved in the postural stabilization of the spine by increasing intra-abdominal pressure (IAP) [6][7][8].The diaphragm can perform these two actions simultaneously.When breathing in a posturally challenging situation, its EMG activity is higher compared to tidal breathing [6,7], and the diaphragm achieves a more caudal position with greater respiratory excursions [9].
Along with the diaphragm, the muscles of the abdominal wall are mostly responsible for regulating IAP [6,10], which is important not only for spinal stability but also for maintaining sufficient ventilation.If the tone of the abdominal muscles is impaired, the diaphragm descends, and tidal volume decreases substantially [11,12].A natural increase in the activity of all abdominal wall muscles is associated with lifting and lowering a load [13].Voluntary isometric contraction of the abdominal muscles before and during any loaded exercise is also known as abdominal bracing (AB) [14,15].Several studies [14][15][16][17] have consistently demonstrated that AB increases spinal stability, which is also an effective technique for stabilizing the spine when lifting weights [18].As it is believed that spinal instability may be one of the causes of low back pain [19], stabilization maneuvers are commonly used in rehabilitation and training programs [20,21].In comparison to other maneuvers such as abdominal hollowing, AB has demonstrated a superior improvement in spinal stability [16,22].
Lifting weights, as part of resistance training, is widely recommended in pulmonary rehabilitation for patients with respiratory disorders [23,24], including those with post-COVID-19 syndrome [25].One of the main symptoms of these conditions is dyspnea [26][27][28], which is associated with increased respiratory effort, decreased ventilation, or both [29].Therefore, it seems important to investigate how lung volumes change during postural loading with abdominal bracing in order to determine whether stabilization maneuvers may increase the risk of dyspnea in these patients.
During tidal breathing, the relation between diaphragmatic excursions and inspired volume has been shown to be linear [30].With increased ventilation, the respiratory activity of the diaphragm may conflict with its postural function, resulting in the suppression of the latter [31].However, it has not yet been reported whether sufficient ventilation can be maintained at higher postural demands.Therefore this study aimed to investigate how diaphragm movement correlates with lung volumes when lifting a load (natural loaded breathing) and when lifting a load with abdominal muscle preactivation through the AB maneuver (AB loaded breathing).We hypothesized that lung volumes would not correlate during both natural loaded breathing and AB loaded breathing with diaphragmatic excursions; as the postural demands on the diaphragm increase, the EIMs should compensate for its respiratory function.
Participants
In this study, 31 healthy adults (average age = 28.7 ± 5.8 years) from the general population, including both athletic and non-athletic individuals, were recruited via social media.The procedures were explained in detail, and signed informed consent was obtained from each participant.Table 1 displays the demographic characteristics of the study group.Similarly to other studies evaluating the function of the diaphragm [32][33][34], the following excluding criteria were applied: low back pain, previous abdominal or spine surgery, respiratory or musculoskeletal disorder (e.g., scoliosis, chest deformities, ankylosing spondylitis), any symptoms of any kind of disease, medical/surgical procedure or trauma within four weeks before initiation of the study, pregnancy and waist to height ratio (WHtR) greater than 0.59.All procedures were performed in accordance with the 1964 Declaration of Helsinki and were approved by the institutional ethical committee (Ref.No. EK-237/21).
Instrumentation
The methods used in this study are identical to those already published in our previous article [35].
M-mode ultrasonography
The ultrasonographic examination was performed with Toshiba (Canon Medical Systems Corporation, Otawara, Japan) Aplio i600 ultrasound system by a single experienced ultrasound operator.The movement of the diaphragm was assessed in the M-mode, which is used to track the motion of a given point in time.M-mode assessment of the diaphragm has been described in several studies [36][37][38][39].A low frequency convex 3.5 MHz transducer was placed in the subcostal region between the midclavicular and anterior axillary lines with the probe tilted cranially, medially, and dorsally to scan the posterior third of the right hemidiaphragm perpendicularly.The right subcostal approach provides a good visualization to assess the movement of the hemidiaphragm, as the liver serves as an acoustic window.In contrast, imaging of the left hemidiaphragm is more difficult because the air in the gastrointestinal tract can interfere with visualization, and therefore a smaller acoustic window through the spleen must be used.For deeper breathing, the left hemidiaphragm is not possible to visualize at all in a large proportion of individuals [38].
In the M-mode trace, the diaphragm is shown as an echogenic line.The position of the diaphragm was determined as the distance between the inspiratory peak and the expiratory peak of the curve from the probe (Fig. 1).This distance was measured vertically from the center of the echogenic line to the baseline.Excursions were then calculated as the difference between the end-expiratory and end-inspiratory positions.All ultrasonographic measurements were reported in millimeters.
Spirometry
Lung volumes were measured using a portable spirometer Jaeger MasterScope (VIASYS Healthcare, Hoechberg, Germany) with an original heated pneumotachograph.The initial calibration was performed with a one-liter pump in MasterScope software, followed by a second calibration using the original software called Breath-Recorder, described in previous studies [9,40,41].Raw flow-time data were stored directly on the spirometer's hard disk, while the flow signal was integrated to obtain a time-volume signal.All records were corrected for body temperature and ambient pressure saturated with water vapor (BTPS) to increase measurement accuracy.To analyze the recorded spirometry data was used original Grapher software [9,41].In this software, we can see any changes in time using a special cursor (Fig. 2).Therefore, any specific time-volume records of inspiratory/expiratory volume changes were accurately measured.
Procedures
Participants were asked not to eat for at least 1.5 h before the procedure.All examinations were performed in the same quiet room with a stable temperature by the same operators, who were blinded to the results of other assessments.Ultrasonographic and spirometric examinations were performed simultaneously with the subjects in a standing position.Subjects stood with their feet shoulder-width apart, elbows close to their body, and both hands placed on the handle of one kettlebell.They were instructed to lift the kettlebell only by bending their elbows to avoid tilting the trunk and loss of ultrasound imaging of the The procedure began with a deep inspiration followed by a sharp expiration, which was used as a time marker for both records.One recording lasted up to 20 s.Every procedure was repeated three times in all patients, and the average values of these measurements were then calculated.Diaphragm motion and lung volumes were measured in the following two scenarios: 1) Natural loaded breathing scenario (N-LBS): two tidal inspirations and expirations, then the kettlebell was lifted, followed by two inspirations and expirations while the kettlebell was held.2) AB loaded breathing scenario (AB-LBS): two tidal inspirations and expirations, then the participants were instructed to contract the abdominal muscles and lift the kettlebell, followed by two inspirations and expirations while holding the kettlebell and having the abdominal wall tensed.
Statistical analysis
Descriptive statistics were calculated for all variables.Data are mean ± standard deviation (SD) unless otherwise stated.Univariate outliers were assessed for each dependent variable by calculating z-scores using complete data for all scenarios (n = 31).Normality was assessed using ± 1.96 as the cutoff for the absolute z-score skew and kurtosis (respectively) for each variable [42].
Results evidenced three probable-outliers according to z-score values greater than 2.58, which occurred within variables that were not normally distributed.These outliers were handled by winsorization; where the outlier retained its rank value and was replaced with the next largest value [43].This process improved the normality of those variables to within the acceptable range, where no absolute z-scores were larger than ± 1.96 for skew or kurtosis after the correction of the outliers.The reliability of the ultrasonography and spirometry measures was calculated from averaging measurements of two tidal breaths recorded at different time points for each subject.Table 2 presents intraclass correlation coefficient estimates (ICC 2,k ), 95% confidence intervals, and standard error of measurement (SEM) calculated from the tidal inspiration and expiration values.The ICC's were calculated based on a mean-rating (k = 3), absoluteagreement, 2-way random-effects model.Reliability was interpreted as poor (< 0.5), moderate (0.5 -0.75), good (0.75 -0.9), and excellent (> 0.9) [44].Pearson's correlations were used to assess the relationship between diaphragm excursions and lung volumes for each scenario described, and paired-samples t-tests were used to determine changes in the diaphragm position and lung volumes for each scenario.Power analysis, using G*Power 3.1, indicated an 80% chance of detecting a strong correlation of 0.50 in 29 subjects (two-tailed), and 27 subjects were needed to achieve a medium effect size of 0.5 in paired t-tests (one-tailed), with statistical significance determined a priori at p < 0.05.When relevant, Bonferroni corrections were utilized when testing multiple hypotheses.The strength of the correlations was interpreted as weak (< 0.30), moderate (0.30-0.50), or strong (> 0.50), and effect sizes were interpreted as small (< 0.2), medium (0.5), or large (> 0.8) [45].All data analyses were conducted using the Statistical Package for the Social Sciences (SPSS version 28.0 for Mac; IMB Corp, Armonk, NY).
Hypothesis testing
Pearson's correlations demonstrated strong statistically significant positive relationships between the ultrasonography and spirometry measures for all types of breathing: tidal breathing: r(29) = 0.709, p < 0.001; natural loaded breathing: r(29) = 0.731, p < 0.001; and AB loaded breathing: r(29) = 0.795, p < 0.001 (see Table 3).Table 4 displays the means ± standard deviation (SD), mean differences, and outcomes of each scenario.These data are presented graphically in Figs. 3 and 4.During the natural loaded breathing scenario, ultrasonography demonstrated the inspiratory position of the diaphragm (mm) when holding the weight was significantly lower, i.e., more caudal, compared to tidal inspiration (t(30) = 4.83, p < 0.001), but not for the expiratory position (t(30) = 1.31, p = 0.101).Spirometry values demonstrated that the lung volume (L) at the end of inspiration when holding the weight was significantly lower than the end of tidal inspiration (t(30) = 4.53, p < 0.001), as well as the end of expiration when holding the weight (t(30) = 4.75, p < 0.001).Total lung volume, calculated as the difference between the average inspiratory and expiratory values, was lower when holding the weight but not enough to be statistically significant (t(30) = 1.65, p = 0.06).During the AB loaded breathing scenario, ultrasonography demonstrated the inspiratory position of the diaphragm (mm) when holding the weight with AB was significantly lower compared to tidal inspiration (t(30) = 3.08, p = 0.002), but not for the expiratory position (t(30) = 1.69, p = 0.05).Spirometry values demonstrated that the lung volume (L) at the end of inspiration when holding the weight with AB was significantly lower than the end of tidal inspiration (t(30) = 3.52., p < 0.001), as well as the end of expiration
Discussion
This study has demonstrated that total lung volumes may be reduced during lifting a load compared to tidal breathing, although diaphragm movement increases.We assume that this is due to an increase in the stiffness of the chest by the insertional action of the abdominal muscles on its lower part [46], since a significant decrease in total lung volume occurred only when the abdominal muscles were voluntarily contracted.Contrary to our hypothesis, these findings imply that there is no greater contribution of the EIMs to ventilation during lifting tasks, but the expiratory muscles are more involved instead.The data presented in Table 4 show that when lifting weight, inspiratory volume is being reduced, even though the diaphragm reaches a more inspiratory (caudal) position compared to tidal breathing.Conversely, the expiratory volume increases while lifting weight, even though the diaphragm reaches the same expiratory position as during tidal breathing.This further supports the theory of greater involvement of the expiratory muscles.Such results may seem opposing to the findings of Hagins & Lamberg [47][48][49], who reported that whole-body lifting tasks resulted in a significant increase in inspiratory volume.However, these studies did not compare tidal and loaded breathing but only investigated breathing behavior during lifting a load.Since inspiration was identified in studies by Hagins & Lamberg before lift-off, inspiratory volume was thus increased.
Considering that the diaphragm plays a crucial role both in breathing and spinal stabilization [6,8], its insufficient respiratory function can affect the postural function and vice versa.This is supported by several studies that showed the reduced respiratory or postural-respiratory movement of the diaphragm in patients with low back pain (LBP) [41,50,51], which may be associated with the decreased magnitude of the force the diaphragm is capable to exert [52].Some previous studies suggest that lifting weights could improve diaphragm strength.DePalo et al. [53] demonstrated that progressive, graduated training of biceps curls and sit-ups for 16 weeks led to diaphragmatic hypertrophy and an increase in maximal inspiratory pressure.In addition, subjects who trained with weights had greater diaphragm thickness and achieved greater maximal inspiratory pressures compared to those who were untrained [54].On the other hand, the controls in these two studies [53,54] were non-training, so it is unclear whether these effects were caused by a specific type of workout or exercise in general.Janssens et al. [55,56] also found that individuals with LBP were more prone to diaphragmatic fatigue compared to healthy controls.Even training the inspiratory muscles in LBP patients reduced pain intensity.This may help explain the association found between the presence of some respiratory disorders and LBP [57,58].
Patients with chronic respiratory diseases often have lower diaphragmatic excursions than healthy controls [59][60][61][62].Many of these patients suffer from dyspnea and associated exercise intolerance [63][64][65][66].From the patient's perspective, dyspnea is a major factor that impairs the quality of life in chronic respiratory disorders, such as chronic obstructive pulmonary disease (COPD) [67].An increase in dyspnea in COPD patients was found to be related to reduced exercise capacity and smaller diaphragmatic excursions [68].Although resistance training is generally beneficial and recommended for COPD patients [69,70], postural training of the diaphragm using lifting weights may not be advisable.For upper-body resistance training, including weight-lifting, studies have found no significant benefit in improving dyspnea and changes in maximal inspiratory pressure in patients with COPD [71,72].Moreover, the use of the upper limbs during different tasks can overload the EIMs and lead to dyspnea in these patients [73].Our results demonstrated slight decreases in lung volume when lifting a load, but not statistically significant.However, when using AB during resistance training, consideration should be given to the fact that reduced lung volumes may contribute to the development of dyspnea in patients with respiratory disease.
The current study also shows a strong positive correlation between diaphragmatic movement and lung volumes during individual postural-respiratory maneuvers, such as tidal breathing, natural loaded breathing, and AB loaded breathing.This suggests that even during loaded breathing, the diaphragm is responsible for the majority of the inspired volume.However, this correlation does not fully explain why the overall lung volume decreased when lifting a load.As discussed above, we presume that lung volumes may be reduced by the postural engagement of other trunk muscles.We also observed a significant reduction in lung volume occurring during lifting a load with the simultaneous AB.In a related article [35], we reported that adding the AB maneuver during lifting weight results in a twofold increase in abdominal wall tension (AWT) compared to the spontaneous contraction of the abdominal muscles.Since the magnitude of AWT and IAP are strongly correlated [74], resistance to caudal diaphragmatic movement increases during AB.This suggests that adding AB when lifting a load can improve not only spinal stability but also the strength of the expiratory muscles as well as the diaphragm.Further research is needed to determine the effect of posturalrespiratory training with or without AB on diaphragm thickness, movement, and strength in both healthy individuals and patients with LBP.
Our study has several limitations.Firstly, all subjects were healthy young volunteers, so the findings cannot be interpreted for a population of elderly individuals or those with respiratory disorders or musculoskeletal dysfunction.Secondly, it is unknown whether the same results would be confirmed for lifting heavier loads than 20% of body weight.In the present study, the weight could not be increased much more, as then the lifting would cause adverse body movement that would affect the ultrasonographic imaging of the diaphragm.Further research is needed to determine how heavier loads influence diaphragm movement as well as lung volumes.Lastly, we have not monitored the activity of the extradiaphragmatic respiratory muscles while performing postural-respiratory maneuvers; thus, we can only speculate on their involvement during loaded breathing.
Conclusion
Postural-respiratory contraction of the trunk muscles may reduce lung volume despite an increase in diaphragmatic motion during lifting a load.For patients at risk of dyspnea, it should be taken into account that lung volumes decreased significantly during loaded breathing with abdominal bracing.Still, strong correlations were found between lung volume and diaphragm movement for all types of breathing; suggesting major contributions of the diaphragm to respiration even during postural loading tasks.
Fig. 1 M
Fig. 1 M-mode ultrasonographic image of the diaphragm motion.The cursor is placed on the peak of the curve at the end-inspiratory (upper peaks B, D, F) and end-expiratory (lower peaks C, E, G) phase of breathing with a vertical line to the baseline
Fig. 2
Fig. 2 Spirometric time-volume curve displayed in Grapher software.The vertical axis represents the lung volume (in liters), and the horizontal axis represents the time (in seconds).Placing the cursor on the spirometric curve allows us to determine the exact value of the volume at a given time Note: mm millimeter, L = Liter a Subject performs two tidal breaths, then holds the weight and performs two additional breaths with spontaneous abdominal muscle activity b Subject voluntarily contracts abdominal muscles prior to holding weight, followed by two breaths Note: N-LBS=Natural loaded breathing scenario AB-LBS Abdominal bracing -loaded breathing scenario Effect size = calculated Cohen's d * Statistically significant difference observed (Bonferroni Correction P < 0.025)
Table 1
Descriptive statistics of participants (Mean ± Standard Deviation)
Table 2
Intraclass Correlation Coefficients of ultrasonography and spirometric values during tidal inspiration and expiration (ICC 2, k )
Table 4
Changes in Ultrasonography values (mm), Spirometric values (L), and Lung volume (L) during different scenarios of holding a load equivalent to 20% body weight (Mean [Standard Deviation])
|
2023-09-16T13:04:06.296Z
|
2023-09-15T00:00:00.000
|
{
"year": 2023,
"sha1": "5f54c68cd86a657d7f0bb082bd21a46f6eb9f5f3",
"oa_license": "CCBY",
"oa_url": "https://bmcsportsscimedrehabil.biomedcentral.com/counter/pdf/10.1186/s13102-023-00729-w",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "710b49696fe006d34bc3954dd33cb3ae93dbb382",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
3870209
|
pes2o/s2orc
|
v3-fos-license
|
Multicenter, phase II clinical trial of peptide vaccination with oral chemotherapy following curative resection for stage III colorectal cancer
The safety and immunological responsiveness of a peptide vaccine of ring finger protein 43 and 34-kDa translocase of the outer mitochondrial membrane combined with uracil-tegafur/leucovorin (UFT/LV) was previously demonstrated in metastatic colorectal cancer (CRC) in a phase I clinical trial. To clarify the survival benefit of a peptide vaccine combined with UFT/LV as adjuvant treatment, a phase II clinical trial was conducted involving patients with stage III CRC. All enrolled patients, whose human leukocyte antigen (HLA)-A status was double-blinded, were administered the same regime of a peptide vaccine and UFT/LV chemotherapy. The primary objective of the study was to compare relapse-free survival (RFS) in patients with HLA-A*2402 vs. those without HLA-A*2402. Secondary objectives included comparisons between the two groups regarding overall survival, safety, tolerability and peptide-specific activities of cytotoxic T lymphocytes (CTLs) as measured by the ELISPOT assay. Between December 2009 and December 2014, a total of 46 patients were enrolled to the present study. Three-year RFS was not significantly different between HLA-A*2402 matched and unmatched groups [67.8 vs. 73.6%, respectively; hazard ratio (HR)=1.254, 95% confidence interval (CI): 0.48–4.63; P=0.706]. Three-year RFS was significantly better in patients with positive CTL responses in the HLA-A*2402 matched group compared with those without (85.7 and 33.3%, respectively; HR=0.159, 95% CI: 0.023–0.697; P=0.011). In conclusion, vaccination-induced immune responses combined with UFT/LV were positively associated with survival benefit in patients with HLA-A*2402-positive stage III CRC. Further study is required to clarify whether vaccination-induced immune responses shortly following the initiation of therapy can predict the therapeutic effect and help develop a promising therapeutic strategy for patients with stage III CRC.
Introduction
Colorectal cancer (CRC) represents the most common cause of cancer-related mortality and the second most common malignancy in Japan (1). Both the mortality rate and the prevalence of CRC have increased in Japan during recent decades (2), and around 30% of CRC patients have stage III disease. Postoperative adjuvant chemotherapy is systemic chemotherapy that is performed after surgery to prevent recurrence and improve the prognosis of patients who have undergone R0 resection (3). In general, patients with stage III CRC for whom R0 resection has been performed are indicated for adjuvant chemotherapy because of the relatively high recurrence rate for patients with stage III CRC (30.8%) (4). 5-fluorouracil (FU)/leucovorin (LV) as adjuvant chemotherapy after surgery has prolonged survival for patients with advanced colon cancer compared with surgery alone (5). Moreover, the development of oral adjuvant chemotherapy such as uracil-tegafur (UFT)/LV (6), capecitabine (7), and S-1 (8) has further improved outcomes.
Chemotherapy of recurrent or metastatic CRC has improved substantially over the last 10 years through the development of new cytotoxic drugs including oxaliplatin (L-OHP) and irinotecan. Several studies in Western countries have demonstrated that the addition of L-OHP to FU/LV or capecitabine improved the adjuvant treatment of colon cancer (9)(10)(11). In Japan, however, some surgeons are skeptical of the use of L-OHP as adjuvant treatment for stage III colon cancer because the outcomes of Japanese randomized trials are better than those of Western countries. The ACTS-CC trial involving S-1 or UFT/LV reported a 3-year disease-free survival (DFS) rate using S-1 monotherapy for stage III colon cancer of 75.5%, while the JCOG0205 trial involving 5-FU/LV vs UFT/LV reported a 3-year DFS rate using UFT/LV of 72.5%; these results are comparable to those seen in trials in Europe and the US in which L-OHP was added to chemotherapy regimens (7,8). Another disadvantage of L-OHP-based therapy is its cumulative neurotoxicity (12).
We previously reported a phase I clinical trial of a peptide vaccine ring finger protein 43 (RNF43) and 34-kDa translocase of the outer mitochondrial membrane (TOMM34) combined with UFT/LV for metastatic CRC, and demonstrated the safety and immunological responsiveness of this combination therapy (13). To clarify the survival benefit of a peptide vaccine combined with UFT/LV as adjuvant treatment, we conducted a multicenter, phase II clinical trial of patients with stage III CRC.
Materials and methods
Patients and eligible criteria. Patients were eligible for enrollment if they were 20-80 years old with histologically confirmed stage III CRC, had adequate critical organ functions, and had an Eastern Cooperative Oncology Group performance status of 0 or 1. Patients were excluded if they were pregnant, breastfeeding, were trying to become pregnant, had an active infectious disease, had multiple cancers, or took steroids or immunosuppressive therapy. The study was carried out in accordance with the Declaration of Helsinki on experimentation with human subjects, and was approved by Institutional Ethical Review Boards of Kindai University (approval no. 20-110) and of Yamaguchi University School of Medicine (approval no. H22-175). It was registered at the UMIN Clinical Trials Registry as UMIN 000003552 (http://www.umin. ac.jp/ctr/index.htm). Written informed consent was obtained from all patients at the time of enrollment.
Peptides and drugs. HLA-A*2402-restricted RNF43 (NSQPVWLCL) and TOMM34 (KLRQEVKQNL) peptides were synthesized by American Peptide Company Inc. (Sunnyvale, CA, USA) according to a standard solid-phase synthesis method; preclinical trials previously confirmed that the peptides did not produce acute toxicity (13).
Montanide ISA-51 (also known as incomplete Freund's adjuvant) is a sterile vaccine adjuvant manufactured by SEPPIC Co. (Puteaux, France) in accordance with good manufacturing practice standards. Montanide is currently used as an adjuvant in vaccine therapies worldwide, and no serious adverse events caused by Montanide have been reported.
UFT is a relatively old oral fluoropyrimidine that was developed in Japan in the 1980s. It has many indications for metastatic and advanced solid cancers including those of the colon, lung, breast, and pancreas, and gastric cancer (14). In metastatic CRC, UFT/LV was demonstrated to have the same clinical efficacy as 5-FU/LV and comparable pharmacokinetics between Japanese and American patients (15)(16)(17). We previously demonstrated that the standard dose of UFT/LV did not impede the immune responses of patients with advanced CRC to peptides administered as cancer vaccination (18).
Study design. This phase II, non-randomized, single arm study in which the HLA-A status was double-blinded aimed to clarify the survival benefit of a peptide vaccine in combination with UFT/LV as adjuvant treatment for patients with stage III CRC. The therapy consisted of a cocktail of two epitope peptides with UFT/LV. Although the peptides used in this study were HLA-A*2402-restricted, all enrolled patients whose HLA-A status was double-blinded were administered the same regime of peptide cocktail and UFT/LV chemotherapy. The cocktail of two peptides (1 mg of each peptide) was mixed with Montanide ISA 51 and subcutaneously administered to patients once every 7 days five times. All patients also received daily oral doses of UFT (300 mg/m 2 /day) plus LV (UZEL: 75 mg/day) for 28 days. Each cycle of treatment was followed by 1 week of rest. Patients received six cycles of treatment unless their disease relapsed.
Study objectives. The primary objective was the comparison of RFS between patients with HLA-A*2402 vs. those without HLA-A*2402. Secondary objectives included comparisons between the two groups regarding overall survival (OS), safety, tolerability, and peptide-specific activities of cytotoxic T lymphocytes (CTLs). Adverse events resulting from the peptide vaccine were evaluated using the National Cancer Institute's Common Terminology Criteria for Adverse Events v.4.0 (19).
ELISPOT assay. Peptide-specific CTL responses were estimated by the in vitro ELISPOT assay as previously described (20). Briefly, frozen peripheral blood mononuclear cells (PBMCs) from each patient were thawed simultaneously. PBMCs (5x10 5 /ml) were then cultured with 10 microgram/ml of each peptide and 100 IU/ml of interleukin-2 (Novartis, Emeryville, CA, USA) at 37˚C for 2 weeks. Peptides were added to the cultures on day 0 and day 7. Following CD4+ cell depletion by the Dynal CD4 positive isolation kit (Invitrogen, Carlsbad, CA, USA), interferon (IFN)-γ ELISPOT assays were performed using the Human IFN-γ ELISpot PLUS kit (MabTech, Nacka Strand, Sweden), according to the manufacturer's instructions. The positivity of antigen-specific T cell response was quantitatively defined according to the evaluation tree algorithm described by Kono et al (21). Briefly, the peptide-specific T cell responses were classified into four grades (-, +, ++, and +++) depending on the peptide-specific spots at different responder/stimulator ratios. We judged to be positive case, when the algorithm indicated +, ++, or +++.
Statistical analyses. This study defined the HLA-A*2402 matched group as the study group and the unmatched group as the control group. We estimated that a minimum of 42 patients would be required for the HLA-A*2402 unmatched group and 64 for the HLA-A*2402 matched group, assuming a RFS of 75% in the HLA unmatched control group and that of 85% in the HLA matched study group with a two-sided α level of 0.05 and a beta level of 0.2. Considering the distribution (approximately 67%) of HLA-A*2402 in the Japanese population and some dropout cases, we decided to enroll a total of 110 patients.
Qualitative data were reported as the number of patients, and were compared using either Pearson's χ 2 test or Fisher's exact test, as appropriate. The Wilcoxon rank sum test was used to compare qualitative data. Survival curves were plotted using the Kaplan-Meier method and compared with the log-rank test.
Survival was measured from the first vaccination until recurrence, death, or the last follow-up. Tests were always two-sided and the level of statistical significance was set at P<0.05. Statistical analysis was performed using JMP 11 software (SAS Institute Inc., Cary, NC, USA).
Results
Demographics. The patient flow diagram is shown in Fig. 1. Between December 2009 and December 2014, a total of 46 patients were enrolled in the study. Although we planned to recruit 110 patients, the registration was prematurely closed because of slow patient recruitment. This might reflect the fact that adjuvant chemotherapy with oxaliplatin-based regimens were approved in August 2009, so patients may have wished to receive an oxaliplatin-based regimen rather than an uncertain peptide vaccine treatment. Moreover, approximately 40% of patients were predicted to have no benefit with the peptide vaccine treatment.
Twenty-eight patients had at least one allele of HLA-A*2402 and 16 had no HLA-A*2402 allele. Among the 46 patients, 44 received peptide vaccine therapy with UFT/LV (Fig. 1). Two patients were excluded because they withdrew consent. As shown in Table I, there was no significant difference between the HLA-A*2402 matched and unmatched groups regarding gender, age, location of the primary tumor, dose of vaccine peptides administered, and the number of positive lymph node metastases (i.e., ≤3 vs. >3) which were synonymous with stage IIIa vs. stage IIIb based on the Japanese Classification of Colorectal Cancer (22). Fewer patients with stage IIIb were in the HLA-A*2402 unmatched group than the HLA-A*2402 matched group, but this difference was not significant (P=0.059).
Subgroup analysis.
To minimize the potential bias regarding fewer stage IIIb CRC patients in the HLA-A*2402 unmatched group, subgroup analysis was conducted for each stage IIIa or stage IIIb group. No significant difference was observed within stage IIIa or stage IIIb regarding gender, age, location of the primary tumor, or the dose of vaccine peptides administered (Table IIA and B). There was also no significant difference between HLA-A*2402 matched and unmatched stage IIIa CRC groups regarding 3-year RFS (86.7 and 91.7%, respectively, HR=1.63, 95% CI: 0.156-35.0, P=0.687) (Fig. 3) HLA-A, human leukocyte antigen-A; n.s., not significant. (Fig. 4). There was no significant difference in OS between HLA-A*2402 matched and unmatched stage IIIb CRC groups.
Safety. The most frequent adverse event was vaccination-site reaction (n=39), and all of the events were grade 1 or 2. The vaccination therapy was well-tolerated with no treatment-associated adverse events ≥grade 3, except for two cases. One patient had grade 3 transaminase elevation, which recovered after cessation of the drug. However, after 28 vaccine peptide administrations, the recurrence of peritoneal dissemination was detected so the adjuvant therapy was discontinued (Table III). Another patient developed bowel perforation after 21 administrations of the vaccine peptide, so the adjuvant therapy was discontinued, though a causal relationship with the adverse event was not proven.
Immunological evaluation in the HLA-A*2402 matched group.
Peptide-specific CTL responses were estimated by the in vitro ELISPOT assay before the initiation of therapy and after two cycles of treatment. In the HLA-A*2402 matched group, positive or negative CTL responses specific for the RNF43 and/or TOMM34 peptides after two cycles of treatment were observed in 14 and nine patients, respectively; the CTL responses of five patients were not detected (Table IV). The 3-year RFS was significantly better in patients with positive CTL responses than in those without in the HLA-A*2402 matched group (85.7 vs. 33.3%, respectively, HR= 0.159, 95% CI: 0.023-0.697, P=0.011) (Fig. 5).
Discussion
This phase II clinical trial demonstrated that vaccination-induced immune responses combined with UFT/LV are positively associated with survival benefit in patients with HLA-A*2402-positive stage III CRC. Importantly, none of the patients in the stage IIIa group with positive CTL responses specific for RNF43 and/or TOMM34 peptides after two cycles of treatment relapsed at all, while two out of three patients with negative CTL responses did relapse (data not shown). In our earlier trial of the same treatment strategy, no remarkable clinical responses were observed in patients with metastatic CRC refractory to standard chemotherapy (13). However, in this trial, the clear advantage of positive CTL responses was observed as an adjuvant setting for CRC patients with lymph node metastasis after curative resection; thus, peptide vaccination is likely to be more effective for low-risk CRC patients than high-risk or metastatic CRC patients. In this trial, two patients developed grade 3 or higher adverse events, although these were not proven to have a causal relationship with the peptide vaccination combined with UFT/LV. Moreover, the vaccination therapy was otherwise well-tolerated with no other treatment-associated adverse events of grade 3 or higher. With the exception of the vaccication site reaction, the rates of other adverse events did not exceed those of UFT/LV chemotherapy, and the addition of the peptide vaccination did not increase the adverse events, which was the same as that reported in our previous trial (13).
Peptide-specific CTL responses were estimated by the in vitro ELISPOT assay before the initiation of therapy and after two cycles of treatment. In the HLA-A*2402 matched group, patients with positive CTL responses after two cycles of treatment showed a significantly better survival than those without. Peptide-specific CTL responses were also observed in HLA-A24 unmatched group, and patients with positive CTL responses in this group had a nonsignificant improvement in prognosis than those with negative CTL responses (data not shown). These peptides used in this study had been considered HLA-A*2402-restricted, however, the possible cross reactivity of the peptides to other serotypes should be taken into account. Since the affinity of the peptides-HLA-A24 should be higher than that of the peptides-other serotypes, the difference of the affinity might influence the clinical outcomes. Vaccination-induced immune responses are positively associated with survival benefit in HLA-A*2402-positive group, and vaccination-induced immune responses that occur shortly after the initiation of therapy could be used to predict the therapeutic effect.
This study has a number of limitations. First, its sample size was small. Therefore, to confirm the survival benefit of peptide vaccination with UFT/LV for patients with HLA-A*2402-positive stage III CRC, additional cases should be recruited to achieve adequate statistical power because some patients lacked positive CTL responses specific for RNF43 and/or TOMM34 peptides. Additionally, the background of the patients was biased. The HLA-A*2402-positive stage IIIb group only contained three patients, of whom most had rectal cancers. Thus, the outcomes of these patients were worse than expected. Second, although the HLA-A*2402 matched stage IIIb group had a nonsignificant trend toward better survival than the HLA-A*2402 unmatched stage IIIb group, its 3-year RFS was 46.1%, which was inferior to that seen in the JCOG0205 trial. While the stage IIIb group in Table IV. In vitro ELISPOT assay prior to the initiation of therapy and following 2 cycles of treatment in the human leukocyte antigen-A2402 matched group. this study was supposed to have a worse survival because of the high proportion of rectal cancer patients, we found no evidence to support the validity of an L-OHP-free regimen as adjuvant chemotherapy. Third, this study did not compare outcomes between patients receiving peptide vaccination combined with UFT/LV vs. those who did not. In assessing the value of the peptide vaccination, an appropriate control would be HLA-A*2402-positive CRC patients who received UFT/LV with no peptide vaccination.
In conclusion, vaccination-induced immune responses combined with UFT/LV are positively associated with survival benefit in patients with HLA-A*2402-positive stage III CRC. Further study is needed to clarify whether vaccination-induced immune responses that occur shortly after the initiation of therapy can be used to predict the therapeutic effect and help develop a therapeutic strategy.
|
2018-04-03T04:33:27.130Z
|
2018-01-29T00:00:00.000
|
{
"year": 2018,
"sha1": "649ba076933f0913373930d0a8b0d028605103a5",
"oa_license": "CCBYNCND",
"oa_url": "https://www.spandidos-publications.com/10.3892/ol.2018.7905/download",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "649ba076933f0913373930d0a8b0d028605103a5",
"s2fieldsofstudy": [
"Medicine",
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
257836016
|
pes2o/s2orc
|
v3-fos-license
|
Safety of nursing professionals and patient facing COVID-19 pandemic in critical care unit
Abstract Objective: to evaluate nursing professionals and patient safety culture during the professional performance in the care of suspected or infected patients with COVID-19. Method: a cross-sectional study carried out with 90 professionals from critical care units of two teaching hospitals. An instrument for sociodemographic characterization and health conditions was used, in addition to the constructs “Nursing professional and patient safety” and the Hospital Survey on Patient Safety Culture. Univariate analyzes were performed between the diagnosis of COVID-19 and the characteristics of Nursing professionals, applying Kendell’s correlation between the constructs. Results: the COVID-19 diagnosis presented a significant statistical difference between nursing professionals that worked for more than six years at the critical care unit (p=0.020) and the items of the construct “Nursing professional and patient safety” regarding the doubts about how to remove the personal protective equipment (p=0.013) and safety flow (p=0,021). The dimensions 2 (p=0.003), 3 (p=0.009), 4 (p=0.013), 6 (p<0.001), and 9 (p=0.024) of the Hospital Survey on Patient Safety Culture were associated with the accomplishment of training. Conclusion: a higher professional nursing experience time was associated with non-infection by COVID-19. The perception of the safety culture of the patient was related to the accomplishment of training.
Introduction
Experiencing a pandemic has several negative consequences, especially for nursing, a workforce that is at the forefront of hospital care. For these professionals, the development of new functions during the pandemic adds to the responsibilities of comprehensive care of managers and team leaders, in decision-making management, forecasting and provision of equipment and materials, implementing control and prevention strategies in health care, among others. These attributions were decisive for the nursing team, whose responsibility is to value their own safety and that of the patient as one of the fundamental pillars of quality in the provision of safe care (1)(2)(3) .
The pressure of the pandemic scenario, many times with the continuous improvement of good safety and communication practices (3)(4)(5)(6) .
During the COVID-19 pandemic, nursing was exposed to: work overload, high virus transmissibility, manipulation of specific protective equipment and high technology, which led health team members to physical and mental exhaustion in the workplace experienced worldwide (7) .
Thus, the attitudes of nursing professionals, their training, the availability and safe use of PPE, psychological followup, peer support and workload influence workers in this area regarding the assessment of the safety of patients who were hospitalized. during the COVID-19 pandemic (8) .
In this context, the present study carried out in two teaching hospitals in different Brazilian states, had as objective to evaluate nursing professionals and patient safety culture during the professional performance in the care of suspected or infected patients with COVID-19.
Method Study design
This is an exploratory cross-sectional study carried out in critical care units (prompt care, Intensive Care Unit (ICU)/semi-intensive and surgical center), guided by the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) (9) tool. (10) . Strongly agree, aiming to attribute a numerical value to the given answers by the participants according to the level of agreement with the statements.
Setting and sample size of the study
We elaborated the construct "Professional and patient safety" and we submitted to face and content validity in 2020. For that, a committee composed of three nurses specialized in nursing management, patient security and infection control and prevention related to the health care thematic evaluated the construct for validation. We invited the specialists to participate in the validation through an invitation letter sent via email with the ICF. They analyzed the instrument items regarding the content (relevance of the items in the instrument) and adequacy to measure with clarity what needs it proposes to measure (face validity). It was considered the level of agreement of 80% among the judges for each item to be evaluated (11) .
We carried out the validation phase of the construct at to the dimension X (we will exclude positive, neutral, and negative from missing data)] x 100 (10) .
We collected sociodemographic data and COVID-19 data from nursing professionals. The categorical variables
Data analysis
We carried out the analysis of the sociodemographic data and COVID-19 infection through descriptive statistics using the R 4.1.1 software. We carried out descriptive statistics for every variable aiming at the overall characterization of the sample. We described the qualitative variables through absolute and relative frequency. For the continuous and discrete numerical variables, we calculated the average, standard deviation, median, and variation. We applied the Cronbach alpha coefficient to the construct "Professional and patient safety" and to the HSOPSC instrument to verify the data reliability and intern consistency, establishing as evidence of satisfactory internal consistency values superior to 0,70.
The association of COVID-19 infection between variables: sociodemographic characteristics, professionals and the construct "Professional and patient safety" utilized Pearson's chi-squared test or Fisher's exact test. In the identification of the correlation between the HSOPSC dimensions and the "Professional and patient safety", we applied the Kendall correlation coefficient.
We considered α de 5% for all analysis.
Research demonstrates that higher exposure to labor risks is due to prolonged contact time with infected patients that need frequent interventions (13) .
The findings show that the working time in the unit was associated with COVID-19 infection and most professionals (89.77%) worked between 20 to 39 weeks hours only in one hospital (62.52%). The pandemic demanded a great volume of professionals as a working force, aiming to provide the challenges of the overload of highly complex care activities directed to critical patient care and in extreme vulnerability situations (14) .
Given the new challenges, a study carried out in
Pakistan with nursing professionals carried out educational interventions and learning through WhatsApp, and the results showed that the group participants reported significant improvement in the learning for "infections prevention and control", "knowledge COVID-19", as well as "leadership and communication" (16) . Moreover, in this same study, we verified that nursing professionals need longer periods of training and support for continuous learning and not only training restricted to hours or a few days (16) .
An investigation carried out with 2,707 professionals from 60 countries highlighted that the excessive training demanded in care during the pandemic influenced professional exhaustion, besides the exposure during patient care to COVID-19 and life prioritization; however, the use of PPE positively contributed to minimizing the burnout symptoms (17) . It is important to highlight that the institutional psychological support to the collaborators during the stress confrontation and emotional support during the pandemic were associated with better teamwork, a safe work environment, job satisfaction and job conditions and stress recognition (7) .
During the COVID-19 pandemic, the managers (3,(18)(19) . The management during the peak of service and the necessity of allocating resources to guarantee the hope and well-being of the health professional are essential attributes to balance the patients' demands in the ICU and the maintenance of the worker's health status (3,(18)(19) . (8,(20)(21)(22)(23)(24) .
The pandemic brought grave consequences that will
|
2023-03-31T06:17:33.905Z
|
2023-03-27T00:00:00.000
|
{
"year": 2023,
"sha1": "b22a66ca82fcf1e59720035a2be04bc7e6962e80",
"oa_license": "CCBY",
"oa_url": null,
"oa_status": null,
"pdf_src": "PubMedCentral",
"pdf_hash": "0a5fff240174f4926d528bf480a2621c77dfe123",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
21128836
|
pes2o/s2orc
|
v3-fos-license
|
Structural and functional determination of homologs of the Mycobacterium tuberculosis N-acetylglucosamine-6-phosphate deacetylase (NagA)
The Mycobacterium tuberculosis (Mtb) pathogen encodes a GlcNAc-6-phosphate deacetylase enzyme, NagA (Rv3332), that belongs to the amidohydrolase superfamily. NagA enzymes catalyze the deacetylation of GlcNAc-6-phosphate (GlcNAc6P) to glucosamine-6-phosphate (GlcN6P). NagA is a potential antitubercular drug target because it represents the key enzymatic step in the generation of essential amino-sugar precursors required for Mtb cell wall biosynthesis and also influences recycling of cell wall peptidoglycan fragments. Here, we report the structural and functional characterization of NagA from Mycobacterium smegmatis (MSNagA) and Mycobacterium marinum (MMNagA), close relatives of Mtb. Using a combination of X-ray crystallography, site-directed mutagenesis, and biochemical and biophysical assays, we show that these mycobacterial NagA enzymes are selective for GlcNAc6P. Site-directed mutagenesis studies revealed crucial roles of conserved residues in the active site that underpin stereoselective recognition, binding, and catalysis of substrates. Moreover, we report the crystal structure of MSNagA in both ligand-free form and in complex with the GlcNAc6P substrate at 2.6 and 2.0 Å resolutions, respectively. The GlcNAc6P complex structure disclosed the precise mode of GlcNAc6P binding and the structural framework of the active site, including two divalent metals located in the α/β binuclear site. Furthermore, we observed a cysteine residue located on a flexible loop region that occludes the active site. This cysteine is unique to mycobacteria and may represent a unique subsite for targeting mycobacterial NagA enzymes. Our results provide critical insights into the structural and mechanistic properties of mycobacterial NagA enzymes having an essential role in amino-sugar and nucleotide metabolism in mycobacteria.
Mycobacterium tuberculosis (Mtb) 2 is a major human pathogen and is the causative agent of tuberculosis (TB). In 2016, there were ϳ10.4 million people with new TB infections, and 1.7 million people died of the disease (1). The emergence of drug-resistant strains of Mtb has resulted in untreatable forms of TB, and there is therefore an urgent requirement for novel strategies to address this global health threat. The success of Mtb as a pathogen is due in part to the complex mycobacterial cell wall that comprises three distinct macromolecules: peptidoglycan (PG), arabinogalactan (AG), and mycolic acids that form the mAGP complex and is further interspersed with distinctive lipids and an outermost capsule of polysaccharides and proteins (2). The cell envelope has been implicated in key roles in the survival and virulence of Mtb and as such is an important target for current TB drugs and compounds that are in clinical development (3).
Hence, the Mtb GlmS and NagA enzymes have key roles in controlling the intracellular pools of GlcN6P, and the in vitro essentiality of both genes has been established in Mtb from transposon mutagenesis studies (18). In addition, metabolic profiling of guinea pigs infected with Mtb showed up-regulation of NagA in infected lung tissue (13,19).
NagA enzymes (EC 3.5.1.25) are widespread in nature and found in a range of eukaryotic and prokaryotic species (20) as metal-dependent enzymes with a physiological role in catalyzing the deacetylation of GlcNAc6P to yield GlcN6P and acetate ( Fig. 1) (21)(22)(23). A putative catalytic mechanism for NagA enzymes has been proposed in which nucleophilic attack occurs via a water/hydroxide ion. The mechanism proceeds via a strictly conserved active-site aspartic acid residue that has an essential role in catalysis and is understood to act initially as a base to activate the hydrolytic water molecule and then as an acid to protonate the amine leaving group (23)(24)(25). High-resolution X-ray structures have been determined of NagA from three Gram-negative bacteria species: Escherichia coli (PDB codes 1YMY, 1YRR, 2P50, and 2P53) (21,23) with and without a transition state inhibitor (N-methylhydroxyphosphinyl-Dglucosamine-6-phosphate), Thermotoga maritima (PDB code 1O12), and Vibrio cholerae (PDB 3EGJ and 3IV8); and the Gram-positive Bacillus subtilis organism (PDB code 2VHL) with the GlcN6P reaction product (24). All NagA structures characterized to date reveal a similar overall architecture and arrangement of two domains. Domain I comprises a (/␣) 8 barrel structural fold that forms the dimeric interface with domain I of the neighboring subunit. This dimeric interface enables the formation of two identical active sites that are involved in substrate and metal co-factor recognition. The smaller second domain of NagA enzymes comprises a -barrel with unknown biological function. There are intriguing differences in the active site metal-binding site. The NagA enzyme from E. coli contains a mononuclear metal-binding site (21)(22)(23), whereas the corresponding enzymes from T. maritima and from the Gram-positive B. subtilis both have binuclear metalbinding sites, though in the former case only one metal ion is necessary for catalysis (22,24). It is therefore possible that NagA enzymes have evolved to have a different role or catalytic function depending on the organism in which they are found.
Within Mtb, NagA (Rv3332) is found in a four-gene operon also encoding a putative aminotransferase (Rv3329); dacB1 (Rv3330), an enzyme involved in PG maturation; and SugI (Rv3331), a putative sugar-importer (17, 26) (Fig. 2). Despite its importance in Mtb, detailed studies of NagA from mycobacte- (GlcNAc6P deacetylase) is highlighted in red. Glc6P, glucose-6-phosphate; F6P, fructose-6-phosphate; GlcN6P, glucosamine-6-phosphate; GlcN1P, glucosamine-1-phosphate; GlcNAc1P, GlcNAc-1-phosphate; SugI, integral sugar transporter; Pgi, glucose-6-phosphate isomerase; GlmS, glutamine-fructose-6-phosphate aminotransferase; GlmM, phosphoglucosamine mutase; GlmU, bifunctional acetyltransferase/uridyltransferase. NagA from mycobacteria specific for GlcNAc6P ria have not yet been reported. This study focuses on the NagA homologs from Mycobacterium marinum and Mycobacterium smegmatis, close relatives to Mtb. Here, we describe the functional characterization and structural determination of mycobacterial NagA enzymes. We report the crystal structure of WT and mutant M. smegmatis NagA, both in ligand-free form and complexed to the GlcNAc6P substrate, revealing the molecular determinants and structural framework of the active site. Furthermore, we have carried out site-directed mutagenesis targeting putative catalytic residues allowing us to elaborate on the molecular recognition of NagA indicating the importance of conserved active-site amino acids in the catalytic function and substrate and stereochemical requirements of this important Mtb enzyme.
Production of MMNagA and MSNagA
To produce recombinant NagA protein, the nagA gene was amplified by PCR and cloned into either the pET28a plasmid for co-expression in E. coli containing the Mtb GroES chaperone or pYUB1062 (27) for expression in M. smegmatis, both bearing an N-terminal His 6 tag. It proved difficult to obtain the recombinant Mtb NagA protein, and therefore homologous enzymes from M. marinum and M. smegmatis were used in these studies, which have a high degree of sequence identity to the TBNagA enzyme with 71 and 52% sequence identity, respectively, at the amino acid level (28) (Fig. S1). Soluble, active MMNagA and MSNagA proteins were obtained and purified to apparent homogeneity using Co 2ϩ affinity, anion exchange, and size-exclusion chromatography ( Fig. S2 and S3). The identities of the NagA proteins were confirmed using in-gel trypsin digestion and analysis of the peptides by MS.
Substrate specificity of MMNagA and MSNagA
NagA enzymes from other species (Fig. S1), are known to catalyze the deacetylation of GlcNAc6P (20 -23), and we therefore speculated that mycobacterial NagA enzymes catalyze a similar reaction. The ability of the MSNagA and MMNagA proteins to catalyze the deacetylation of various N-acetyl amino sugars was tested using a fluorescence-based assay by monitoring the production of the fluorescent product formed from the reaction of fluorescamine and the free-amino group that is formed following deacetylation (29). The panel of carbohydrates tested is shown in Fig. 3. We first examined the NagA reaction with various phosphorylated N-acetylated amino sugars that included the predicted GlcNAc6P substrate, alongside the epimeric analogs GalNAc-6-phosphate (GalNAc6P) and ManNAc-6-phosphate (ManNAc6P) (Fig. 3). The measure-ments indicated that MMNagA and MSNagA are active deacetylase enzymes with markedly similar kinetic profiles (Table 1). In each case we were able to detect the formation of GlcN6P, GalN6P, and ManN6P in the reaction, with both MSNagA and MMNagA showing a clear preference for the GlcNAc6P substrate (Table 1). Although MSNagA and MMNagA show activity toward GalNAc6P and ManNAc6P, kinetic analysis revealed that the K m value was 40-fold higher for GalNAc6P and 6-fold higher for ManNAc6P than for GlcNAc6P (Table 1 and Fig. 4). Intriguingly, the apparent kinetic constants varied for the MSNagA enzyme depending on whether the protein had been expressed in an E. coli or M. smegmatis host expression system with a higher K m value of 5.2 mM and almost 5-fold reduction in the k cat found for MSNagA when expressed in E. coli (Table 1).
The importance of the phosphate group in substrate recognition and catalysis was examined through the reaction of NagA with GlcNAc-6-sulfate (GlcNAc6S) and GlcNAc. MMNagA and MSNagA were both able to catalyze the reaction with GlcNAc6S; however, in both instances the K m values were 7-fold higher (Table 1 and Fig. 4) than the phosphorylated analog. No catalytic activity toward GlcNAc was observed for either enzyme, indicating the importance of an anionic group at the 6-position and, in particular, a preference for the phosphate group for substrate recognition and catalysis. Next, we determined the importance of the N-acetyl group position at C2 through replacement with an N-glycolyl group, which can be tolerated by mammalian NagA homologs (30); however, only low levels of enzymatic activity for both MMNagA and MSNagA enzymes with this derivative with an extended linker at the C2 position were observed. Finally, we determined whether the mycobacterial NagA enzymes could act as a deacetylase for N-acetyl muramic acid 6-phosphate, which incorporates a C3 lactoyl group into GlcNAc6P, given that N-acetyl muramic acid is the main other sugar constituent of the peptidoglycan backbone. However, no detectable catalytic activity was observed in this instance indicating selectivity of mycobacterial NagA enzymes for the GlcNAc6P cell wall peptidoglycan fragment.
Metal dependence of MMNagA and MSNagA
NagA homologs from other organisms require a divalent metal ion for catalysis (20 -23). To further determine the metal content of MMNagA and MSNagA enzymes, protein samples were analyzed for the metals: manganese, iron, cobalt, nickel, copper, zinc, and cadmium using inductively coupled plasma MS (ICP-MS). From these studies, it was clear that when both MMNagA and MSNagA enzymes were expressed using an E. coli expression system in either LB medium or terrific broth, Figure 2. The NagA operon in M. tuberculosis. The operon organization was taken from Xbase. The accession numbers for these genes in M. tuberculosis H37Rv are as follows: Rv3329 (probable aminotransferase), Rv3330 (dacB1, probable penicillin-binding protein D-alanyl-D-alanine carboxypeptidase), Rv3331 (sugI, probable sugar-transport integral membrane protein), and Rv3332 (nagA, GlcNAc-6-phosphate deacetylase).
NagA from mycobacteria specific for GlcNAc6P
both zinc and iron were the predominant metals and were found in a ϳ1:2 zinc:iron ratio (Table S2). In comparison, when the MSNagA enzyme is expressed using a M. smegmatis expression system in LB medium, the protein contained predominantly more zinc in a 12:1 zinc:iron ratio. These differences in the selected metal ion that is incorporated into the protein during protein production may explain the increase in the K m and reduction in k cat for the MSNagA protein that is expressed in an E. coli host system ( Table 1). Expression of MSNagA in E. coli in LB medium supplemented with 1 mM ZnCl 2 did not alter the zinc:iron ratio, suggesting that the host expression system plays a significant role in metal co-factor selectivity.
CD spectroscopy revealed that removal of the metal ions, confirmed by ICP-MS, from either enzyme by extensive dialysis against chelex-treated buffer containing 10 mM 1,10-phenanthroline and 10 mM EDTA, resulted in unfolding and loss of catalytic activity (Fig. S4). Activity could not be restored by addition of ZnCl 2 to the metal-free enzymes, indicating that the metal ions are tightly coordinated in the active site and have an essential structural role in mycobacterial NagA enzymes.
Analysis of the pH dependence of MMNagA and MSNagA
To investigate the pH optimum for NagA catalysis with GlcNAc6P, the specific activities of the MMNagA and MSNagA enzymes were determined with GlcNAc6P across a range of pH values (Fig. S5). This shows the loss of catalytic activity at a pH below 5.5. The pH profile resembles that found for the E. coli NagA enzyme (23), where deprotonation of a catalytic acidic aspartic acid residue is necessary for activity.
Site-directed mutagenesis of MMNagA and MSNagA
To assess the significance of selected MMNagA and MSNagA residues in molecular recognition and function, we generated single point mutations in five amino acids in MMNagA, with one corresponding mutation in MSNagA, that were expected to form either the ligand-or metal-binding sites. In addition, we generated two point mutations in the binuclear metal binding site of the MMNagA enzyme. The HXH motif (His-56 -His-58), also found in B. subtilis and T. maritima NagA enzymes, was mutated to either a QXN motif to replicate the single metal-binding motif found in both the E. coli and V. cholerae NagA enzymes, or to AXA. For each mutant, we confirmed that the introduced mutation was not detrimental to the correct folding of the protein by CD spectroscopy (Fig. S4).
In each case the single amino acid mutation resulted in complete loss of activity of the MMNagA and MSNagA proteins (Table 1) under the conditions tested, supporting the notion that these residues have a key influence on substrate selectivity and/or catalysis. To ensure that the mutation did not result in reduction/loss of metal ion during expression and/or purification, zinc was added to the reaction mixture, but this did not restore activity. Increasing the concentration of GlcNAc6P substrate and protein along with the reaction time did result in the observation of low (Ͻ5%) residual activity for three MMNagA mutants: H139A, H249A, and R225A compared with WT. The QXN and AXA double mutants in the second metal-binding site motif resulted in protein that, although correctly folded, was inherently less stable. However, catalytic activity for the MMNagA QXN mutant with GlcNAc6P was observed, with a 7-fold increase in K m value compared with WT MMNagA (Table 1). In contrast, no activity was detected for the AXA MMNagA mutant with GlcNAc6P. Despite these mutations in the active site of MMNagA and MSNagA being detrimental for the catalytic activity of these enzymes, the mutants were still able to bind and recognize the GlcNAc6P substrate, as evidenced from microscale thermophoresis (MST) studies (Table S3).
Overall structure of MSNagA
To determine the molecular mechanism and structural basis of catalysis, we solved the crystal structure of MSNagA with and without the GlcNAc6P substrate present. The MSNagA protein in the absence of ligand crystallized in space group P1 with four molecules in the asymmetric unit. Phases for the structure of ligand-free MSNagA were determined by molecular replacement using the structure of E. coli NagA (PDB code 2P50) as a search model, and the structure was refined at a resolution of 2.6 Å. To investigate the interaction of the GlcNAc6P substrate with the protein, we mutated Asp-267, which has been shown to be important for catalysis in other NagA enzymes (21)(22)(23)(24), to alanine. This resulted in significant loss of enzymatic activity (Table 1) . The initial velocity data were plotted against the substrate concentration, and each assay was carried out in triplicate and expressed as a value Ϯ standard error of mean.
NagA from mycobacteria specific for GlcNAc6P
diffraction data and structure refinement are summarized in Table 2.
Overall, the structure of MSNagA is very similar to other NagA enzymes with NagA from E. coli, the closest structural match according to PDBeFOLD (31), aligning with a root mean square deviation of 1.4 Å for 334 aligned residues (target residues, 354; sequence identity, 35%; PDB code 3IV8). The structure can be divided into two domains. Domain I (residues 52-344) is a twisted (/␣) 8 -TIM barrel that is comprised of eight alternating -stands and ␣-helices and encloses the catalytic site of the enzyme and the metal-binding site. Domain II is a small -barrel comprising eight -strands contributed from the N and C termini of the MSNagA protein (amino acids 1-51 and 344 -382) (Fig. 5). In both of the MSNagA crystal structures, the protein forms a homodimer (Fig. 5) burying a total of 3,200 Å 2 of the surface area (32). By comparison to the determined crystal structures of NagA enzymes, this dimeric interaction is similar across bacterial NagA enzymes (21,22,24). The two active sites of the MSNagA homodimer are formed at the interface of domain I between the two monomeric subunits (Fig. 5) with hydrogen bonding between residues on ␣ 8 and ␣ 9 and salt bridges between Glu-230 and Arg-251. A long flexible loop formed from residues 203-223 of one subunit extends out to ␣ 9 of its neighboring subunit (residues 248 -258), providing additional stability at the dimer interface. In solution, MMNagA and MSNagA are found as dimers as shown by size exclusion chromatography (Figs. S2 and S3), and it is likely that this dimeric oligomerization state is also the biologically relevant unit for mycobacterial homologs, which has been found for other NagA enzymes (21,22,24,33).
The ligand-binding site of MSNagA
From our MST studies (Table S3) the D267A MSNagA mutant was still able to interact with the GlcNAc6P substrate, although no catalytic activity was observed ( Table 1). The cocrystal structure shows clear electron density for GlcNAc6P in the substrate-binding pocket of Chain A, enabling the substrate to be modeled in the active-site pocket at the dimeric interface ( Fig. 6 and Fig. S6). However, it may be that the site is not fully occupied, and no electron density for the GlcNAc6P substrate was observed in the second active site of chain B where instead a flexible loop region from Glu-122 to Pro-136 partially occludes the active site (Fig. S7). The reasons for these differences between the equivalent active sites in the co-crystallized complex are not clear; however, it is intriguing that a similar situation is also observed in the co-crystal structure of the enzyme from B. subtilis with the GlcN6P reaction product, where only partial occupancy (Ͻ0.3) was observed in the second binding pocket (24).
The GlcNAc6P substrate is situated in the cavity of the (␣/) barrel of domain I at the dimer interface, recruiting Arg-219 from the opposing monomer through an interaction with the phosphate moiety. An arginine at this position is specific to MSNagA and in both M. marinum and Mtb NagA homologs the equivalent residue is replaced by a histidine residue (Fig. S1). The side chain of the conserved arginine residue corresponding to Arg-220 in MSNagA points away from the GlcNAc6P ligand (Fig. S8).
The GlcNAc6P substrate is found in the ␣-configuration and is anchored in place through the formation of hydrogen bonds
NagA from mycobacteria specific for GlcNAc6P
with the side chains of His-244 that interacts with the anomeric GlcNAc6P hydroxyl group and with Asn-212 to the phosphate group. The remaining hydrogen bonds are through interactions with the main chain, with the equatorial C3-hydroxy group interacting with Ala-133 and the equatorial C4-hydroxy with Ile-301. These critical residues that interact with GlcNAc6P are conserved in homologous NagA enzymes with the exception of Ile-301 (MSNagA), which is instead replaced by a leucine residue; however, this modification is unlikely to affect its interaction to the substrate with the interaction occurring through its backbone carbonyl group. Additional interactions of the phosphate group are with the backbone amino group of Ala-213 and the guanidine group of Arg-219 from the opposing molecule (Fig. 6).
The carbonyl group of the GlcNAc6P substrate is positioned ϳ2.5 Å from the M1 metal ion, which is ideally primed to stabilize the transition state intermediate formed during catalysis (Fig. 6). It is of interest to note, by comparison of the apo-and ligand-bound structures, that there is little conformational change in the binding pocket upon substrate binding to MSNagA. However, there is a difference in two loop regions that impact on the accessibility of the active site. The surface-based loop that comprises the amino acids Met-274 -Ser-304 forms a dynamic "lid" that is likely to be functionally important in facilitating ligand binding through an opening/closing mechanism. An additional loop that differs between the two structures, comprising residues Glu-122-Pro-136, partially occludes the active site in chain B and prevents substrate binding (Fig. S7). Although domain II is not involved in substrate binding, we do observe a change in the position of the -barrel in chains A and B of the ligandbound structure.
Binding of metal ions to MSNagA
Although no divalent cations were added during the purification steps, well defined electron density corresponding to two showing GlcNAc6P with green carbon atoms, the metal ion as a pale-yellow sphere, and selected amino acid residues in stick representation (colored gray for chain A, and magenta for chain B). B, schematic diagram of the interactions of MSNagA with GlcNAc6P. The residues that interact with the M1 metal are shown in green, the residues that interact with the M2 metal are shown in red, the residues that interact with the GlcNAc6P substrate are shown in black, the residues that interact with the GlcNAc6P substrate from the opposing MSNagA monomer are shown in blue, and the Asp-267 residue that is mutated in the ligand-bound structure is shown in purple.
NagA from mycobacteria specific for GlcNAc6P
bound metal ions was clearly visible in the cavity of the (␣/) barrel buried at the bottom of the active-site cleft (Fig. 7). Each MSNagA monomer in both the ligand-free and ligand-bound models contains two metal ions positioned between 3.1 and 3.4 Å apart from each other. The proteins were shown to contain both zinc and iron in the ICP-MS analysis ( Table S2). Given that the same protein expressed in M. smegmatis contains predominantly zinc, it was decided to model the metal-binding sites as zinc rather than iron (see "Experimental procedures" ; Fig. S9). However, for the ligand-bound mutant structure, residual density in the M1 metal-ion binding site indicated a larger ion. We therefore modeled a cadmium ion at this position, which is likely to arise as a result of the addition of CdCl 2 in the crystallization conditions with the metal-binding site perhaps being destabilized by the D267A mutation.
All of the ligands that interact with the metal ions are contributed by conserved residues from one subunit of the dimer. In all chains of the apo-structure, the metal ion that is more buried (M1) makes four interactions coordinated by the side chains His-56, His-58, Glu-122, and Asp-267 (Fig. 7). The second metal ion (M2) is held in position by the side chains of residues His-188, Glu-122, and His-209 (Fig. 7). Coordination is completed by water molecules with one water molecule modeled as bridging both of the metals and another bound to M1. In the ligand-bound structure, the water molecule that bridges the two metal ions in the apo-structure is instead replaced with a chloride ion. The coordination of the metal ions in chain A to His-56, His-58, and Glu-122 is almost identical to that of the apo-structure with loss of coordination occurring to the Asp-267 residue, which has instead been mutated to an alanine. The M2 metal ion is similarly coordinated to Glu-122, His-188, and His-209, and it is striking that an additional interaction is formed with the carbonyl group of the GlcNAc6P substrate, replacing the water molecule in the apo-structure. The second chain of the co-crystal structure, where there is no GlcNAc6P substrate, can also be considered as another view of the apostructure. Here there is a translational movement of His-134, which is located within the Glu-122-Pro-136 loop that blocks the active site, such that the histidine side-chain is 2.6 Å from the cadmium ion (Fig. 7). In the ligand-bound structure, this histidine residue is within 3 Å of the carbonyl oxygen of the substrate. This histidine is important in catalysis because its mutation in MMNagA (His-139) results in loss of activity (Table 1).
Discussion
In prokaryotic species, it has been demonstrated that NagA is involved in the deacetylation of GlcNAc6P with an important role in carbohydrate metabolism and the recycling of murein fragments from the cell wall (11,12,34). A putative NagA enzyme, organized in an operon that differs from other bacterial organisms, has been identified in Mtb through genomic and proteomic studies and is predicted to be essential in the Mtb pathogen (17)(18)(19). Our biochemical and structural analyses of mycobacterial NagA enzymes are an important step to understanding amino-sugar assimilation, recycling, and metabolism in Mtb.
Our X-ray crystallographic structure determination reveals that MSNagA conforms to the urease structural superfamily, with the same overall fold, domain topology, and architecture as other NagA enzymes whose structures have been determined (21,22,24). MSNagA comprises two domains with formation of the active-site binding pocket occurring within domain I of the enzyme. Recruitment of a nonconserved Arg-219 from domain I from the adjacent monomer explains the dimeric nature of the enzyme. It is noteworthy that the mycobacterial NagA enzymes also form functional dimers. Mapping of Arg-219 to TBNagA reveals that this residue is instead replaced by a histidine residue, and it is probable that the invariant TBNagA arginine residue (Arg-220 MSNagA, corresponding to Arg-225 MMNagA and Arg-221 TBNagA) is utilized, along with the adjacent histidine residue that facilitates additional interactions with the 6-phosphate group of GlcNAc6P in a similar pincer manner to that of B. subtilis NagA (24). Modification of Arg-225 in MMNagA resulted in loss of catalytic function, indicating the important functional role of this conserved residue. Domain II appears to have no role in substrate binding, and the biological role of this small -barrel/sandwich domain remains unclear. Our attempts to investigate the role of the -barrel domain of MSNagA through the generation of a truncation mutant was hindered by the loss of production of soluble recombinant protein and suggests that the -barrel domain has a role in the stability of MSNagA. (chain A, blue). Silver spheres, Zn 2ϩ ; red spheres, H 2 O. GlcNAc6P with green carbon atoms and selected amino acid residues are shown in stick representation (colored by apo (magenta) or ligand-bound (blue)). B, superposition of the metal-binding site of chain A and chain B of the ligand-bound structure. Light blue, chain A; light brown, chain B; silver spheres, Zn 2ϩ ; orange sphere, Cd 2ϩ ; green sphere, Cl Ϫ . Selected amino acid residues are shown in stick representation (colored by chain A (blue) or ligand-bound (salmon)).
NagA from mycobacteria specific for GlcNAc6P
Importantly, in these studies we are able to unambiguously determine the position and orientation of the carbonyl group of GlcNAc6P, which is displaced in the NagA catalytic reaction. The binding mode of the biological GlcNAc6P substrate with MSNagA is very similar to that identified in the structure of B. subtilis NagA in complex with the GlcN6P product (24) and an E. coli NagA catalytically inactive mutant in complex with a N-methlhydroxyphosphinyl-Dglucosamine-6-phosphate transition state inhibitor (22). The MSNagA complex structure indicates that the active site does not undergo any major structural change on binding the GlcNAc6P substrate; however, two loop regions demonstrate a capacity for conformational flexibility and an importance in facilitation of substrate binding. A notable difference between mycobacterial NagA enzymes and NagA enzymes from other bacterial species is the presence of a cysteine at position 131 (MSNagA), which is replaced by a corresponding lysine residue in nonmycobacterial species and is located in the flexible loop that precludes the physiological substrate from binding. On the basis of the significant variation of a noncatalytic cysteine located within a region linked with substrate binding, there is the potential to exploit this reactive sulfhydryl moiety with the view to designing inhibitors specific for the TBNagA enzyme. A recent example of a cysteine-targeted therapeutic approach is in the development of irreversible peptidomimetic inhibitors to a noncatalytic cysteine located in the hepatitis C virus protease (35).
Our kinetic analysis reveals that MMNagA and MSNagA enzymes catalyze the deacetylation of the amino-sugar GlcNAc6P to form GlcN6P and acetate. Biochemical characterization has provided the first evidence of the function of mycobacterial NagA enzymes and demonstrated that MMNagA and MSNagA have comparable kinetic profiles to each other, and the kinetic constants reported here for MSNagA and MMNagA are comparable with the E. coli NagA enzyme in our fluorescent assay. The presence of a binuclear metal-binding site and two divalent metal ions located in each active site of MSNagA, which are both required for efficient catalysis and structural stability, is comparable with the Grampositive NagA homolog and substantiates evidence that NagA enzymes have diverged with an evolutionary adaption that enables the mechanism of activation and stabilization of the carbonyl intermediate to differ depending on the metal requirements of the specific organism.
From the extensive panel of carbohydrates tested in this study, there was a clear preference of MMNagA and MSNagA for GlcNAc6P over the epimeric amino-sugar analogs ManNAc6P and GalNAc6P, indicating that the mycobacterial NagA enzymes are sensitive to the stereochemical requirements of the substituent at the C2 and C4 of the substrate. Although there is a clear preference for an equatorial N-acetyl group at C2, an axial C2 N-acetyl group, in the case of ManNAc6P, can be moderately tolerated. In contrast, there is a highly stringent selection for the equatorial C4 hydroxyl group of GlcNAc6P, with a marked reduction in the catalytic efficiency k cat and an increase in the observed K m for GalNAc6P, where the C4 hydroxy group is positioned in an axial conformation. This was also observed for the E. coli NagA enzyme with the epimeric GalNAc6P substrate (23). Furthermore, the presence of a phosphate group at C6 is required for efficient catalysis to proceed. The biochemical results are consistent with the crystal structure of the complex and reveal that mycobacterial NagA enzymes are fine-tuned with a precise binding network that allows for the specific recognition and catalysis of the GlcNAc6P amino-sugar. Mammalian species have the ability to accommodate the larger N-glycolyl group of N-glycolylglucosamine-6-phosphate and deacetylate the acetyl and glycolyl GlcN6P analogs with equivalent efficiency (30). However, our studies demonstrate that mycobacterial NagA enzymes are distinct from these mammalian deacetylase enzymes and do not have this capacity to tolerate the introduction of a larger glycolyl group at the C2 position, indicating that mycobacterial NagA enzymes have evolved to have unique specificity and function.
The stringent recognition of the murein fragment GlcNAc6P over N-acetyl muramic acid 6-phosphate by mycobacterial NagA enzymes is interesting and may occur to ensure the integrity of Mtb peptidoglycan recycling and amino-sugar metabolism. Mtb nagA is encoded along with genes involved in peptidoglycan biosynthesis and carbohydrate uptake (Fig. 2) (17,26), indicating the Mtb NagA is primed to play a role in the recycling of GlcNAc6P derived from the breakdown of cell-wall peptidoglycan. NagA has been shown to be involved in amino-sugar acquisition and peptidoglycan recycling in a number of other organisms including E. coli (36) (11), B. subtilis (37), Staphylococcus aureus (38), and Streptomyces coelicolor (34,39). Hydrolysis of Mtb peptidoglycan is mediated through lytic transglycosylase resuscitation factors (40), and it is likely that Mtb has evolved functional pathways to enable the recovery of murein fragments within the nutrient poor macrophage environment. The Mtb solute-binding protein UspC of the UspABC transporter has recently been shown to be involved in the recognition of amino sugars with a potential role in recycling deacetylated PG fragments (41). Intriguingly, Mtb lacks the genetic machinery for phosphotransferase systems, and given the genomic organization of nagA, it is tantalizing to link the functional role of the Mtb SugI transporter to the import of phosphorylated carbohydrates, which subsequently acts in concert with the deactylase NagA enzyme to recycle and optimize the use of restricted carbohydrates during intracellular infection.
In conclusion, the NagA enzymes from M. smegmatis and M. marinum, with high sequence identity to the M. tuberculosis analog, are the first NagA enzymes to have been characterized both structurally and kinetically from mycobacteria. Given that there is almost identical sequence identity of these enzymes with the active-site of TBNagA, these mycobacterial homologs represent good model systems to understand the NagA enzyme from Mtb. Our data clearly indicate that mycobacterial NagA enzymes have a clear substrate preference for GlcNAc6P and other amino-sugar derivatives are poorly tolerated. Eukaryotic NagA enzymes convert GlcNAc6P directly to GlcNAc-1-phosphate, have low sequence identity (34%) to TBNagA, and differ in substrate selectivity. In this context, the conversion of GlcNAc6P to GlcN6P by mycobacterial NagA is a bacterial specific process that occurs at an essential metabolic NagA from mycobacteria specific for GlcNAc6P chokepoint in the GlcN6P degradation pathway. Given the importance of TBNagA enzymes in the synthesis of essential mycobacterial cell wall components, TBNagA may serve as a novel molecular drug target, and these studies provide a framework to exploit this enzyme in the search for new TB therapeutic agents.
Materials and reagents
All chemicals and reagents were purchased from Sigma-Aldrich, unless specified, with the exception of all of the carbohydrates used in this study, which were purchased from Carbosynth. PCR and restriction enzymes were obtained from New England Biolabs. Double-distilled water was used throughout.
Plasmid constructs, protein expression, and purification
Full-length GlcNAc-6-phosphate deacetylase (nagA) genes from M. tuberculosis, M. marinum, and M. smegmatis were amplified by PCR from the corresponding gDNA. The primer sequences are listed in Table S1. The PCR products were ligated into either NdeI and HindIII sites of the pET28a (ϩ) (Novagen) for expression in E. coli or pYUB1062 (27) for expression in M. smegmatis mc 2 4517, resulting in the constructs TBnagA_ pet28a, MMnagA_pet28a, MSnagA_pet28a, and MSnagA_ pYUB1062. Targeted single-site or double-site substitutions were introduced into MMnagA_pet28a and MSnagA_pet28a using the primers (Table S1), with Phusion Polymerase and the PCR cycle (98°C, 30 s; 25 cycles of 98°C, 10 s; 60°C, 30 s; 72°C, 4 min; followed by 5 min at 72°C), followed by digestion with 1 l of DpnI. Plasmid sequences were verified by sequencing (GATC) and used for protein expression.
Protein expression of NagA in E. coli
E. coli BL21(DE3) competent cells were co-transformed with the appropriate nagA_pet28a expression plasmid and the GroES 60.2 Mtb chaperone and grown at 27°C to an optical density at 600 nm (A 600 ) of 0.5 in terrific broth medium (Difco) supplemented with 50 g ml Ϫ1 kanamycin and 100 g ml Ϫ1 ampicillin. Protein production was induced with 0.5 mM isopropyl--thiogalactopyranoside, and the cultures were grown at 16°C overnight with shaking (180 rpm). The cells were harvested (4,000 ϫ g, 30 min, 4°C) and resuspended in lysis buffer (20 mM Tris, 300 mM NaCl, 10% glycerol, pH 8.0 (buffer A)) supplemented with 0.1% Triton X-100 and frozen at Ϫ80°C until further use.
Protein expression of NagA in M. smegmatis
M. smegmatis mc 2 4517 electrocompetent cells were transformed with the appropriate nagA_pYUB1062 construct and grown at 37°C to an A 600 value of 0.8 in LB medium supplemented with 0.05% Tween 80, 0.2% glycerol, 25 g ml Ϫ1 kanamycin, and 100 g ml Ϫ1 hygromycin. Protein production was induced with 0.2% acetamide, and the cultures were grown at 37°C for an additional 20 h with shaking (180 rpm). The cells were harvested and resuspended in lysis buffer (buffer A) supplemented with 0.1% Triton X-100 and frozen at Ϫ80°C until further use.
Protein purification
Complete protease inhibitor mixture (Pierce), 5 mM MgCl 2 , 2 mg of DNase, and 20 mg of lysozyme were added to the resuspended pellet, and the pellet was either sonicated (E. coli pellets) on ice (Sonicator Ultrasonic Liquid Processor XL; Misonix) or passed through a cell disruptor at 4°C (M. smegmatis pellets) (Constant Systems, 25 kpsi). Following centrifugation (27,000 g, 40 min, 4°C) the supernatant was filtered (0.45-m pore size) and loaded onto a pre-equilibrated HisPur Co 2ϩaffinity resin (Thermo Scientific, Pierce). The column was washed with buffer A (5 column volumes), and the recombinant NagA protein was eluted from the Co 2ϩ resin with increasing concentrations of imidazole. Fractions containing the protein were dialyzed at 4°C for 12 h against buffer A, and either a second HisPur Co 2ϩ -affinity resin purification step was undertaken or fractions containing NagA protein were dialyzed against 20 mM Tris-HCl, 100 mM NaCl, 10% glycerol pH 8.0 (buffer B) at 4°C for 12 h and applied to a HiTrap Q-column (1 ml; GE Healthcare Life Sciences) pre-equilibrated with buffer B and eluted with NaCl (0.1-1 M). Fractions containing NagA were pooled and purified further using size exclusion chromatography. Gel filtration experiments were carried out on a Superdex 200 16/60 column (GE Healthcare) using 20 mM Tris, 300 mM NaCl, 10% glycerol, pH 8.0. Fractions containing NagA were pooled, 0.03% DDM and 1 mM DTT were added, and the protein was concentrated to 5-10 mg/ml (Vivaspin 2; GE Healthcare) and stored at Ϫ80°C. The identity of the proteins was confirmed by tryptic digest and nanoLC-electrospray ionization-MS/MS (WPH Proteomics Facility, University of Warwick).
NagA assay
Concentrated MSNagA and MMNagA enzymes were dialyzed into 20 mM Bis-Tris, 300 mM NaCl, 10% glycerol, pH 7.0. The E. coli NagA enzyme was commercially available and obtained from NZYtech. The activity of the NagA enzyme was measured at 37°C in an end point assay by following the production of the fluorescent product formed with fluorescamine and primary amines at ex of 340 nm and em of 460 nm. Unless otherwise stated, the reaction was carried out in a 96-well microtiter plate in 20 mM Bis-Tris, 300 mM NaCl, 10% glycerol, pH 7.0, in a total reaction volume of 50 l. The reaction was initiated by the addition of substrate and terminated by the addition of 50 l of 0.4 M borate buffer pH 10, 40 l of 5 mM fluorescamine, and 50 l of dimethylformamide to label the free amines and the production of fluorescence was monitored at ex of 340 nm and em of 460 nm (Tecan Infinite M200). The production of free amine was quantified with a glucosamine standard. For characterization of the kinetics of NagA substrate initial utilization rates of NagA catalysis were obtained. Kinetic parameters were calculated and analyzed using nonlinear regression analysis (GraphPad Prism, v7). All measurements were performed in triplicate. The impact of pH on activity was determined at 37°C in the pH range 4.0 to 9.0 with GlcNAc6P (5 mM concentration). The following buffers were used: phosphate-citrate buffer, 300 mM NaCl, 10% glycerol (pH 4.0 -7.0), 20 mM Bis-Tris, 300 mM NaCl, 10% glycerol (pH 6.0 -7.0), and NagA from mycobacteria specific for GlcNAc6P 6.0 -9.0).
Affinity studies with MST
NagA protein was labeled using Monolith His-tag labeling kit RED-Tris-nickel-nitrilotriacetic acid in 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 10 mM MgCl 2 , 0.05% Tween 20, and a constant concentration of NagA (50 nM) was used. Carbohydrates were prepared in water in the concentration range 0 -0.5 M. The samples were loaded into the MonoLite NT.115 standard treated capillaries and incubated for 10 min before analysis using the Monolith NT.115 instrument (NanoTemper Technologies) at 21°C using 40% laser power and 20% LED power. The binding affinities were calculated with GraphPad Prism software. All experiments were carried out in triplicate.
Circular dichroism
Purified NagA proteins (0.3 mg/ml) were dialyzed in 20 mM Tris, 100 mM NaCl, 10% glycerol, pH 8.0 buffer, transferred into a 1-mm path-length quartz cuvette, and analyzed on Jasco J-810 DC spectrometer from 198 to 260 nm. The spectra were acquired in triplicate and averaged after subtraction of the buffer background.
Inductively coupled plasma MS (ICP-MS)
NagA proteins were diluted into 72% ultrapure HNO 3 and heated to 70°C for 30 min and then diluted with MilliQ deionized water to give a final concentration of 2% HNO 3 . The metal contents of the NagA samples were determined by ICP-MS (Agilent ICP-MS 7500cx) and by comparison to solutions of known metal concentrations using external calibration standards, which were prepared through a serial dilution of a single ppm stock mixture of zinc, cadmium, cobalt, copper, manganese, nickel, and iron in 2% nitric acid. External standards were prepared at 1, 2.5, 5, 10, 25, 50, 100, 250, 500, and 1,000 ppb. The masses of the isotopes detected were 55 Mn, 57 Fe, 59 Co, 63 Cu, 60 Ni, 66 Zn, and 111 Cd. 166 Er was used as an internal standard. Operating conditions were: plasma gas flow rate, 15.0 liters/ min; auxiliary gas flow rate, 0.15 liter/min; nebulizer flow rate, 0.808 rps; and RF power, 1,550 W.
Crystallization and structure determination
Crystals of NagA were grown initially by vapor diffusion in 96-well plates (Swiss-Ci), using a Mosquito liquid handling system (TTP LabTech) by mixing 1:2 volumes (150 nl) of concentrated NagA (10 mg/ml) with reservoir solution. NagA crystals grew within a week at 22°C in 0.12 M monosaccharide mix (Morpheus, Molecular Dimensions), 0.1 M imidazole/MES, pH 6.5, 20% (v/v) PEG 500 MME, 10% (w/v) PEG 20,000 with the addition of 10 mM CdCl 2 additive. For co-crystallization experiments, the D267A MSNagA was incubated with 5 mM GlcNAc6P and incubated at 4°C for 30 min before crystallization in the same reservoir conditions. Crystals were flash frozen in liquid nitrogen prior to data collection.
Structure determination
The X-ray diffraction data for both apo-and ligand-bound MSNagA crystals were collected at I04 and I24 beamlines of Diamond Light Source, respectively. Apo-MSNagA diffraction data were indexed, integrated, and scaled with XDS (42) through the XIA2 pipeline and the CCP4 suite of programs (43). Initial phases were obtained by molecular replacement using PHASER (44) with the NagA structure from E. coli (PDB code 2P50, chain A) as a search model. AutoBuild (45) was initially used for the model building followed by iterative cycles of alternating manual rebuilding in COOT (46), and reciprocal space crystallographic refinement with PHENIX-REFINE (47). The electron density in one of the flexible loop region (residues 275-300) was very weak, and the model building in this region was done manually in O (48). The ligand-bound MSNagA crystal data were also indexed, integrated, and scaled with XDS (42) through the autoPROC (49) pipeline. Initial phases were obtained by molecular replacement using PHASER (44) with the structure of the apo MSNagA (Chain A) as a search model, and the structure was refined as per the apo MSNagA. A feature enhanced 2F o Ϫ F c map (fem map) was calculated to enhance the fine details of the ligand density and the GlcNAc6P ligand was fitted into the electron density map. The Grade Web server (Global Phasing Ltd., http://grade.globalphasing.org/cgi-bin/ grade/server.cgi) 3 was used for the ligand restraint generation and optimization. The conformation of the GlcNAc6P ligand was validated with Privateer (50). The occupancy of the refined ligand was set at 0.75 because the associated density suggested that the site was not fully occupied. At this resolution it is difficult to distinguish zinc and iron ions in the electron density. Based on the ICP-MS results, we decided to model the metal ions as zinc. This resulted in B-factors of the metal ions similar to those of the surrounding residues. For one metal site in the D267A ligand-bound MSNagA structure, the residual density suggested a more electron-rich ion, and therefore cadmium was modeled at this position. In both the apo-and ligand-bound MSNagA structures, the interacting distances between the amino acids and the zinc ion at the active site were restrained to 2.1 Å, whereas no such restraints were applied to the cadmium ion interactions. Noncrystallographic symmetry restraints were initially used during refinement cycles; however, these were not used in the final refinements. Because of the higher resolution of ligand-bound MSNagA, it was used as a reference model for generating dihedral angles restraints during the refinement of apo-MSNagA.
The model of the apo-MSNagA structure determined comprises residues 1-377 in all chains (A-D), with an additional two residues defined (1-379) in chain B. MSNagA is a 385amino acid protein, and clear density for the final C-terminal residues was not observed. The first residue of the partially ordered N-terminal His 6 tag can also be observed in chains A-D. In the apo-MSNagA structure, there is one disordered region in chain D between residues 276 and 302. In chains A-C, weak electron density for this region was observed, indicating that this loop region of MSNagA is flexible. In comparison, the ligand-bound GlcNAc6P-MSNagA complex structure residues 1-378 (chain A) and residues 1-382 (chain B) are well defined by their electron density maps, enabling the positions of these amino acid residues to be determined with an additional three residues of the N-terminal His 6 -tag linker observed in both chains. Structure validations were done by MolProbity (51).
In the Ramachandran plot of the ligand-bound structure, there is only one serious outlier per chain. This is for His-209, one of the metal ligands. This residue forms part of a -turn and coordinates Zn1. Phe-211, which is on the margin of the allowed zone is also in this region. The density is unambiguous in this region, and the homologs in the PDB file have similar outliers. In the apo-structure, some of the residues in the flexible loops are on the margins of the allowed regions in the Ramachandran plot (see above). The figures were drawn using PyMOL (PyMOL Molecular Graphics System, version 2.0; Schrödinger), except those showing electron density, which were made using CCP4 mg (52).
Synthesis
NMR spectroscopy ( 1 H, 13 C) was conducted on either a Bruker DPX-300 or Bruker DPX-400 spectrometer, and all chemical shifts (␦) are given in ppm relative to the solvent reference. The data are recorded as follows: chemical shift (multiplicity (s for singlet, d for doublet, t for triplet, m for multiplet, and br for broad), coupling constant(s) J are quoted in Hz). Mass spectroscopy was obtained using a Bruker HCT Ultra machine. TLCs were performed on Merck silica gel 60 F-254 TLC sheets and were visualized by staining with 10% H 2 SO 4 in ethanol followed by heating. Flash chromatography was carried out using Sigma-Aldrich technical grade silica gel (pore size, 60 Å; particle size, 40 -63 m) as the stationary phase.
|
2018-05-09T00:44:54.447Z
|
2018-05-04T00:00:00.000
|
{
"year": 2018,
"sha1": "1b02c95e2c226a94ad1337785ba1378e18ae9521",
"oa_license": "CCBY",
"oa_url": "http://www.jbc.org/content/293/25/9770.full.pdf",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "9cda62043a4a8437e1f0718d072bbaf23a0247ab",
"s2fieldsofstudy": [
"Biology",
"Chemistry"
],
"extfieldsofstudy": [
"Medicine",
"Chemistry"
]
}
|
261695987
|
pes2o/s2orc
|
v3-fos-license
|
Urinary acrolein protein conjugates-to-creatinine ratio is positively associated with diabetic peripheral neuropathy in patients with type 2 diabetes mellitus
Acrolein, an unsaturated aldehyde, plays a pathological role in neurodegenerative diseases. However, less is known about its effects on peripheral neuropathy. The aim of this study was to investigate the association of acrolein and diabetic peripheral neuropathy in patients with type 2 diabetes. We recruited 148 ambulatory patients with type 2 diabetes. Each participant underwent an assessment of the Michigan Neuropathy Screening Instrument Physical Examination. Diabetic peripheral neuropathy was defined as Michigan Neuropathy Screening Instrument Physical Examination score ≥ 2.5. Serum levels and urinary levels of acrolein protein conjugates were measured. Urinary acrolein protein conjugates-to-creatinine ratios were determined. Patients with diabetic peripheral neuropathy had significantly higher urinary acrolein protein conjugates-to-creatinine ratios than those without diabetic peripheral neuropathy (7.91, 95% CI: 5.96–10.50 vs 5.31, 95% CI: 4.21–6.68, P = 0.029). Logarithmic transformation of urinary acrolein protein conjugates-to-creatinine ratios was positively associated with diabetic peripheral neuropathy in univariate logistic analysis, and the association remained significant in multivariate analysis (OR = 2.45, 95% CI: 1.12–5.34, P = 0.025). In conclusion, urinary acrolein protein conjugates-to-creatinine ratio may act as a new biomarker for diabetic peripheral neuropathy in type 2 diabetes. The involvement of acrolein in the pathogenesis of diabetic peripheral neuropathy warrants further investigation.
Introduction
Diabetic peripheral neuropathy (DPN) is the most common microvascular complication of type 2 diabetes mellitus (1).Most patients with DPN have large-fiber and small-fiber dysfunction (2).Peripheral neuropathy involving large fibers causes muscle weakness and sensory disturbances, and small fiber involvement is characterized by neuropathic pain, impaired temperature sensation, and autonomic dysfunction (2).While poor glycemic control and cardiovascular risk factors have been demonstrated to contribute to DPN, the pathogenetic mechanisms of DPN are multifactorial and complex (3).It has been proposed that exposure to environmental toxic agents may be associated with the development of DPN (4).Acrolein is a highly reactive α,β-unsaturated aldehyde to which humans are exposed in many situations (5,6,7).Acrolein is generated from environmental sources such as engine exhaust, cigarette smoke, industrial waste, or combustion of organic substances (6).It can be released from overheated vegetable and animal fats (6).Furthermore, acrolein can be endogenously produced through myeloperoxidase, lipid peroxidation, and amine oxidase-mediated metabolism of polyamines (7).Acrolein exerts diverse toxic effects through various mechanisms, including DNA and protein adduction, disruption of mitochondria and endoplasmic reticulum, cell membrane damage, oxidative stress, and inflammation and immune dysfunction (8).Due to an abundance of polyunsaturated fatty acid without adequate antioxidant enzymes, neurons are vulnerable to oxidative stress and acrolein toxicity (6).Acrolein impairs mitochondrial function and increases endoplasmic reticulum stress in neuronal tissue (9).Acrolein can damage axonal membrane and myelin structure, resulting in neural injury (9).The overall neurotoxic effects of acrolein induce neuroinflammation and neurodegeneration and may play a role in the pathophysiology of peripheral neuropathy (10,11).
Several studies have shown the association of acrolein with diabetes mellitus and its complications (11,12,13,14,15).Patients with diabetes mellitus had higher serum and urinary levels of acrolein adduct in comparison with healthy subjects (11,12,13).Evidence indicates that acrolein may contribute to diabetic retinopathy and diabetic nephropathy (14,15).Recently, Yao et al. found that treatment with hydralazine, an acrolein scavenger, attenuated neuroinflammation and neuropathic pain in a rat model of diabetes (11).The authors proposed that acrolein might be involved in the pathogenesis of DPN (11).
Treatment of hyperglycemia alone has limited impacts on established DPN in type 2 diabetes (16).However, a number of disease-modifying agents for DPN have shown disappointing results (16).There is an unmet need for relevant biomarkers of DPN to facilitate drug development.Examination of the association between acrolein and DPN may lead to potential biomarkers and targeted treatments in the future.Therefore, the purpose of the present study was to investigate the association of acrolein and DPN in patients with type 2 diabetes.We hypothesize that exposure to acrolein is associated with the development of DPN in patients with type 2 diabetes.
Patients
Patients from the outpatient clinics at the Taipei Veterans General Hospital were eligible if they were at least 50 years of age and had type 2 diabetes mellitus for at least 1 year.Patients were excluded if they had an estimated glomerular filtration rate (eGFR, as calculated by the Modification of Diet in Renal Disease Study equation) below 15/min/1.73m 2 of body surface area or had a history of macrocytic anemia, severely affected acute stroke (National Institutes of Health Stroke Scale > 15), New York Heart Association Functional class III to IV congestive heart failure, muscular diseases, or peripheral arterial occlusive disease.Patients with a recent diagnosis of foot ulcers or nondiabetic neuropathy in the last 6 months were excluded as well.The ethics committee of the Taipei Veterans General Hospital approved the study protocol, and all participants provided written informed consent at the baseline examination.
Measurements
After a 12-h overnight fast, participants received anthropometric assessment and blood pressure measurement at 08:00 h.An automated blood pressure recorder (HEM-7310, Omron Healthcare Inc., Kyoto, Japan) was used to determine the right arm blood pressure while each patient was seated.Height and weight were measured, and body mass index was calculated as weight in kilograms divided by the square of height in meters (kg/m 2 ).Waist circumference was measured to the nearest millimeter with anthropometric tape around the umbilicus.Fasting blood samples were drawn for the measurements of glucose, glycated hemoglobin (HbA 1c ), lipids, biochemistry, and acrolein protein conjugates (Acr-PC).Urinary samples were obtained for the measurements of creatinine and Acr-PC.Urinary acrolein protein conjugates-to-creatinine ratios (Acr-PC/creatinine) were determined.
The Michigan Neuropathy Screening Instrument Physical Examination (MNSI-PE) was used for screening of DPN.The participants received physical assessment of lower extremity based on MNSI-PE, including evaluation of feet appearance, ulceration, ankle reflexes, vibratory perception using a 128 Hz tuning fork, and pressure perception using a 10 g monofilament.Abnormal findings were recorded and scored according to the Data in relation to demographic features and physical activity were collected using an interviewadministered questionnaire.The participants were asked to recall the amount of time per day for each of five levels of activity in a representative day within the past two weeks.The five levels of activity were basal (sleeping or lying down), sedentary (sitting or standing), slight (e.g., casual walking), moderate (e.g., aerobic dancing), and heavy (e.g., running).A physical inactivity score was calculated as sedentary hours divided by waking hours according to the formula reported previously (17).
Biochemical parameters including serum glucose, lipids, creatinine, and alanine aminotransferase (ALT) were measured using commercial assay kits (Roche Diagnostics) in an automatic clinical chemistry analyzer (Roche-Hitachi 7180, Roche Diagnostics).The measurement of HbA 1c was performed using capillary electrophoresis.Serum and urinary levels of Acr-PC were determined using the acrolein-lysine/cysteine/ arginine adduct competitive enzyme immunoassay kit (Takara Bio.Inc.), and the details of detection methods were described elsewhere (18).
Statistical analysis
Data are expressed as arithmetic mean ± s.d. for normally distributed continuous variables, geometric mean (95% CI) for non-normally distributed continuous variables, and number (percentage) for noncontinuous variables.Logarithmic transformation was used to reduce the skewness of fasting plasma glucose, HbA 1c , ALT, triglyceride, serum and urinary Acr-PC levels, and urinary Acr-PC/creatinine levels prior to statistical analysis.Student's t-test and χ 2 test were performed to compare the DPN group and non-DPN group for the continuous and categorical variables, respectively.Pearson correlation analyses were applied to investigate the association between DPN and serum Acr-PC levels, urinary Acr-PC levels, and urinary Acr-PC/creatinine levels.Univariate and multivariate logistic regression analyses were performed to analyze the association of different variables with DPN.A P-value less than 0.05 was considered statistically significant.All analyses
Results
A total of 148 patients were included in the current study.The clinical characteristics of the participants with and without DPN are listed in Table 1.Patients with DPN are older and had higher body mass index and waist circumference.The two groups of participants had comparable rates of smoking status, physical inactivity score, blood pressure, fasting plasma glucose, HbA 1c , and lipid profiles.Patients with DPN had lower eGFR levels comparing to those without DPN.Serum levels and urinary levels of Acr-PC were not significantly different between the two groups.Urinary Acr-PC/creatinine levels were higher in the DPN group than in the non-DPN group (7.91 (5.96-10.50) in the DPN group, 5.31 (4.21-6.68) in the non-DPN group, P = 0.029; Fig. 1).
The results of the Pearson's correlation analyses are summarized in Supplementary Table 1 (see section on supplementary materials given at the end of this article).There was no significant correlation between DPN and logarithmic transformation of serum Acr-PC levels and urinary Acr-PC levels.A significant but weak correlation was detected between logarithmic transformation of urinary Acr-PC/creatinine levels and DPN (r = 0.18, P = 0.029).The results of the logistic regression analyses are shown in Table 2.In univariate analyses, logarithmic transformation of urinary Acr-PC/creatinine levels, age, body mass index, waist circumference, and eGFR were associated with DPN.The association between logarithmic transformation of urinary Acr-PC/creatinine levels and DPN remained significant in multivariate analysis (OR = 2.45, 95% CI = 1.12-5.34,P = 0.025).Collinearity statistics were carried out and evaluated based on the variance inflation factors.All variance inflation factors (range 1.01-3.74)less than ten indicated no multicollinearity problems among the independent variables.
Discussion
In this study, we found that urinary Acr-PC/creatinine levels were higher in patients with type 2 diabetes mellitus and DPN, compared with patients with type 2 diabetes mellitus without DPN.The association between urinary Acr-PC/creatinine levels and DPN remained significant after adjustments for clinical variables, including eGFR.These results suggest that urinary acrolein protein conjugates-to-creatinine ratio may be a potential surrogate biomarker of diabetic peripheral neuropathy in patients with type 2 diabetes mellitus.
Acrolein is one of the common environmental pollutants, and it is known to cause neurotoxicity and various neurological disorders (9).Oxidative stress is recognized as the main mechanism of acrolein-induced neurotoxicity (19).As a product and an initiator of lipid peroxidation, acrolein is considered as a perpetrator of oxidative stress (6).The in vitro and in vivo studies have identified that acrolein-induced oxidative damage can lead to mitochondrial and DNA dysfunction and exacerbate apoptosis (20).Acrolein also inhibits antioxidant enzymes and subsequently increases the rate of lipid peroxidation and oxidative stress (21).Acrolein is known to downregulate Nrf2, a key antioxidant regulator, and decreases capacity of antioxidant enzymes (22).Besides, acrolein exposure contributes to cellular deterioration by axonal membrane damage and myelin disruption (23).It represents another pathogenesis of functional deficits and cellular death triggered by acrolein (23).
Recent studies have shown an emerging role of acrolein-induced oxidative stress in the pathogenesis of diabetic complications (24,25).Acrolein protein adducts are proposed to be associated with carbonyl stress in diabetic glomerular lesions (24).Acrolein compromises antioxidant defense system and mediates Geometric mean and 95% CI of urinary acrolein protein conjugates-tocreatinine ratio in patients with and without diabetic peripheral neuropathy.Student's t-test on the logarithmic transformed data was used to compare the two groups.Urinary Acr-PC/creatinine, urinary acrolein protein conjugates-to-creatinine ratio; DPN, diabetic peripheral neuropathy.
T Wu et al. mitochondrial dysfunction in the diabetic retina (25).Acrolein is also involved in the neurodegenerative process in diabetic retinopathy (26).Besides, Yao et al. reported that treatment with acrolein scavenger hydralazine attenuated neuroinflammation and neuropathic pain in a rat model of diabetes (11).In our study, patients with type 2 diabetes mellitus and DPN had higher urinary Acr-PC/creatinine levels than non-DPN group.Urinary Acr-PC/creatinine levels is positively associated with occurrence of DPN after adjustments for clinical factors, including eGFR.Our results implicate that exposure to acrolein may increase the risk of peripheral neuropathy in patients with type 2 diabetes mellitus.
In the current study, we found a negative association between eGFR and DPN in patients with type 2 diabetes.In diabetic microvasculature, chronic intracellular hyperglycemia causes damage to vascular endothelium, leading to diabetic microvascular complications (27).Patients with DPN are more likely to have diabetic kidney disease (28).Meanwhile, patients with diabetic nephropathy tend to suffer from DPN and diabetic foot complications (28,29).In our study, the correlation among the diabetic microvascular complications was compatible with previous studies.However, the association between DPN and urinary Acr-PC/creatinine levels kept significant after adjustments for clinical variables, including eGFR.Our results suggested urinary Acr-PC/creatinine level as an independent biomarker for DPN in patients with type 2 diabetes.
In this study, logistic regression analyses showed no significant association between DPN and parameters of glycemic control (Table 2).Improved glycemic control reduces incidence of DPN dramatically in patients with type 1 diabetes, but the beneficial effect becomes modest in patients with type 2 diabetes (1).The associations between DPN and blood glucose control in type 2 diabetes from different studies have been inconsistent (30,31,32).The Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial showed decreased risk of DPN in intensive glycemic control group after 5 years of follow-up, but other large trials have shown no treatment effects (30,31,32).This discrepancy may be attributed to polypharmacy, concomitant comorbidities, hypoglycemia, and weight gain (1).A latent period of asymptomatic hyperglycemia before diagnosis of type 2 diabetes may also account for the inconsistent results (1).Further studies are needed to investigate the association between DPN and glycemic control.
In the present study, univariate analyses showed that occurrence of DPN is positively associated with age, body mass index, and waist circumference (Table 2).Older age is one of the well-documented risk factors for the development of diabetic peripheral neuropathy (33).The characteristics of patients with metabolic syndrome, including overweight and abdominal obesity, are associated with diabetic neuropathy in patients with type 2 diabetes as well (33).The pathophysiology of peripheral neuropathy associated with obesity includes altered lipid metabolism and inflammation, leading to microvascular and peripheral nerve injury (34).In our study, the results of the association between these well-known risk factors of peripheral neuropathy and DPN are compatible with previous studies.The association between DPN and urinary Acr-PC/creatinine levels kept significant after adjustments for the clinical risk factors.
There are some limitations in our study.First, the cross-sectional design of the study limits the ability to make casual relationships between variables, although it is beneficial in supporting the hypotheses.The lack of a control group prevents comparison with normal populations without diabetes as well.Second, the limited availability of analysis of acrolein renders its application in routine clinical practice.However, this noninvasive assessment of DPN can be easily applied in large epidemiological studies.This approach may also be used as an initial assessment of patients at high risk when the equipment is available.Last, serum and urinary levels of acrolein did not necessarily reflect the concentrations in target tissues.Further studies are required to explore acrolein-induced tissue injury in the peripheral nerve in experimental models of diabetes.
Conclusions
In conclusion, urinary Acr-PC-to-creatinine ratio is associated with diabetic peripheral neuropathy in patients with type 2 diabetes mellitus after adjustments for clinical variables including estimated glomerular filtration rate.Urinary Acr-PC-to-creatinine ratio may be a biomarker for identification of patients with type 2 diabetes mellitus at high risk for diabetic peripheral neuropathy.
T
Wu et al.
Figure 1
Figure 1Geometric mean and 95% CI of urinary acrolein protein conjugates-tocreatinine ratio in patients with and without diabetic peripheral neuropathy.Student's t-test on the logarithmic transformed data was used to compare the two groups.Urinary Acr-PC/creatinine, urinary acrolein protein conjugates-to-creatinine ratio; DPN, diabetic peripheral neuropathy.
T Wu et al.
Table 1
Clinical characteristics of study participants with and without diabetic peripheral neuropathy.a Student's t-test and χ 2 test were used to compare the two groups.aDiabetic peripheral neuropathy is defined as Michigan Neuropathy Screening Instrument Physical Examination score ≥ 2.5; b P < 0.05.Acr-PC, acrolein protein conjugates; ALT, alanine aminotransferase; BMI, body mass index; BW, body weight; DBP, diastolic blood pressure; DPN, diabetic peripheral neuropathy; eGFR, estimated glomerular filtration rate; HbA1c, glycated hemoglobin; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure; TG, triglycerides; WC, waist circumference.T Wu et al.
e230253 12:11 were
conducted using the Statistical Package for the Social Sciences Statistics Program (version 22.0, IBM Corporation, Armonk, NY, USA).
|
2023-09-13T06:17:07.569Z
|
2023-09-01T00:00:00.000
|
{
"year": 2023,
"sha1": "b36054e06293cecd806bc9cd9f6601b7550ab785",
"oa_license": "CCBY",
"oa_url": "https://ec.bioscientifica.com/downloadpdf/journals/ec/aop/ec-23-0253/ec-23-0253.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "3807e2d331c3e8ec927be9338e70a90cf08cb6d5",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
151694508
|
pes2o/s2orc
|
v3-fos-license
|
Some Difficulties in Understanding Negative Numbers Faced by Students : A Qualitative Study Applied at Secondary Schools in Indonesia
This study revealed how students’ understanding of negative numbers and identified their difficulties related with the concept of integer and its counting operation as part of identifying epistemological obstacles about negative numbers. Even though teachers have explained counting operation procedure of integer, but there was concept misunderstanding among students. The concept difference between what was comprehended by respondents in their learning process and knowledge science has resulted in wrong perceptions about the negative numbers. In this article, the authors explained how these misunderstanding in concepts occurred among students and how these ideas were expressed by the students in solving the problem which was related with counting operation of negative integer. This study was part of Didactical Design Research using qualitative approach in negative number learning by involving 96 students of 7th grade as participants in three different schools. This study showed some difficulties by the student to understanding negative numbers in terms of prerequisite knowledge, understanding concepts, procedures, principles and problem solving.
Introduction
In daily practice, students naturally experience the situation which is called as learning obstacles.These learning obstacles can be experienced by students with various ability levels and groups.If we examine, learning obstacles are not only experienced by students who has 'low' ability but also can be experienced by students who has 'good' ability.Even mathematicians have entire domains of mathematics with which they do not feel comfortable (Einsenberg, 2002).Students start their learning process in an environment that is unbalanced and full of difficulties and obstacles just like human society so the new knowledge comes from the skill to adapt to the new circumstances and stimuli and a new reaction to the environment is the proof that a learning process has taken place (Manno, 2006).This obstacles need to be identified to find out the way to overcome it and determine the strategies to reduce or minimize this obstacle.
Balacheff (Brown, 2008) stated that mathematical error in students occurs from the way student to adapt as their knowledge response in a milieu or it's surrounding.Brousseau (2002) refers to obstacle from theory of Bachelard (1938) and Piaget (1975) about "error", that error and failure have a role which is not simple.This type error is unexpected and uncertain, which is called obstacles.This error is part of knowledge acquisition.Learning obstacle, according to Brown (2008) is a process to know (obtain the knowledge).In other word, this obstacle can be avoided because this obstacle is important part in acquiring new knowledge.
Many factors can cause learning obstacles among the students.One of them is learning pattern which is delivered by teachers as a pedagogical aspect (Widdiharto, 2008).In learning activities, context usage is usually needed to use by teachers and it is a synergy in imparting a concept to students.However, context usage which is not balanced to understanding its mathematics content results in the context which is functioned as example of application only.Suryadi (2013) suggested that if context only used as application, then student only imitated procedure which was showed by the teachers without understanding concept and when it was confronted to different situations, the students were most likely unable to answer and solve them, then it resulted in what is called as learning obstacles.
In Indonesia, negative integer is encountered in mathematics curriculum at the beginning of secondary school, ie at the seventh grade (Kemdikbud, 2013).Some other countries, such as Singapore and Japan, also provide these materials in seventh grade.Researchers conducted a short interview to some 7th grade students and teachers of mathematics in secondary schools.According to them, even though it seems easier in grade seventh and teachers in secondary schools, this material is quite difficult to be understood, particularly in operation of number which involves negative number.Unlike positive number, negative number has not perception referential which is clear, and therefore, students should try harder to learn about negative number (Blair et al, 2012).Many of these are very good models of the positive and negative numbers but difficulties arise when we use them to illustrate the operations of addition, subtraction, multiplication and division (Galbraith, 1974).
When solving the problem, students should firstly understand the related concepts (Schmidt, 1997).Students who experience misunderstanding when receiving a learning concept for the first time, not only influence him/her when he/she learns that concept but will be result in next learning which is development of that concept.Basic concept which will be understood well by the students will give influence in understanding a context in higher level.Research about learning difficulties which is experienced by students will give underlying picture in developing effective intervention in teaching and learning activity, and therefore create hopes to improve mathematics teaching and learning in school and also for the development of knowledge itself (Li & Li, 2008).
Based on that view, this study was intended to find out students' understanding about the definition of negative numbers in daily life.This study is also more specifically to the students' difficulties in understanding the concepts of negative numbers and their counting operation.
The Concept of Negative Numbers
Negative numbers always become an interesting topic, particularly from mathematics didactic point of view (Thomaidis, 1993).The need for negative numbers arises both in the everyday world of measurement and in the mathematical world of number (Galbraith, 1974).The concept which is imparted in respondents thus far is that negative number has nature which is different to positive number which finally become obstacle in homogenization of (N, +) ∪(N, -).Two concepts which are different about positive and negative number will bring us to opposite pole and different domain.Thomaidis explained that the term 'positive label' and 'negative label' did not mean that positive and negative numbers but 'what is added' and 'what is subtracted'.Therefore, the natures of positive number operation actually are not prevailed in negative number.
Negative integer is part of integer, Z.In set N = {0, 1, 2, 3,…} from non-negative integer there is N = {1, 2, 3, 4….} which is positive number in which and for all a, b, c, ἐ Z, the operation nature in positive number are prevailed.For each a ε Z where -a ε Z it is prevailed that a + (-a) = 0 (invers of addition), then it can be written to become a -b for a + (-b) (Hungerford, 2000).The concepts of negative number like this can pose problem in respondent to understand it.Statement of a -b = a + (-b) is difficult to be understood by students if they not introduced first by addition invers.Through this invers concept, it can be determined that a -(-b) = a + b.For some secondary school mathematics textbooks, the operation nature of integer addition and subtraction is written directly in the form of formula.
In particular, Sfard (1991) analyzed mathematical concepts in two underlying aspects namely structural and operational which was refer to two things separated: "object" and "process".Sfard showed that there are two stages of children's learning development in order to understand the numbers.When children learn number concept, they start from counting first.This is "process" stage, which is natural and relatively easy for child.Anyway, a child should change counting process toward abstract number concept.This is called "object" stage.Sfard argued that child should "transfer" from process to object in order to understand the concept entirely.In case of negative number, according to Sfard, child ability to think about a process entirely can be assessed through respondent ability in joining basic process with counting operation namely its ability to do arithmetic manipulation of negative and positive number.
The negative sign can take on at least three meanings in mathematics: unary, binary, and symmetric functions (Bofferding, 2014;Vlassis, 2008).Bofferding (2014) uses the term "minus" generically to refer to the "-" symbol, "subtraction sign" to refer to the binary meaning of the minus sign, and "negative sign" to refer to the unary meaning of the minus sign.The three concept dimensions of negative number which was explained by Altiparmark and Ozdogan (2010), namely: 1) the negative case which is focused on meaning of number, direction and quantity of number (numerical system), 2) meaning of counting and number line which focused on counting operation which connected between what have been understood and the real world (we limited it only on counting operation of addition and subtraction) and the use of number-line as concrete illustration which described context of a situation, and 3) interpretation and explanation, that was how student understand the connection of negative number in a context, interpret, make generalization, determine calculation operation, and connect a context to another context.
Understanding and Epistemological Obstacles in Mathematics
In mathematics learning, understanding was a fairy common word spoken and often became the object of study in mathematics education.Understanding revealed to the meaning that the materials taught to the students not only as are citation, but also the students could understand the concept of the subject matter.Sierpinska (1990) regards that understanding as an act that involved in a process of interpretation, this interpretation being a developing dialectic between more elaborate guesses and validations of these guesses.The process of understanding starts with a guess which we further try to justify and validate.In the course of validation the guess may be improved, changed, or rejected.The new guess is then subjected to justification and validation.It continues as a spiral until it should be understood considered to have been appropriate.
In many cases the understanding and epistemological obstacles are two ways concerning about the same thing (Sierpinska, 1990), epistemological obstacles will appear if they are paired with understanding.Epistemological obstacles look backwards, attention focused on what is wrong in the ways of knowing.Understanding is the way to find and to know in the future perspective.Children will pass through both, if we look from the perspective of his past knowledge then we will see how these children overcome these obstacles.But if we look from the perspective of the future knowledge then we will look their understanding.Overcoming epistemological obstacles and understanding are two images that complement each other through epistemological analysis of mathematical concepts.
Concept of epistemological obstacles occurs in knowledge philosophy which is pioneered by Bachelard (Schneider, 2014).Brousseau then developed epistemological obstacles notion into didactic mathematics as one of learning obstacles.According to Brousseau (2002), based on its cause, learning obstacles are categorized into three types, namely: (1) onto genic obstacle; (2) didactical obstacle; and (3) epistemological obstacle.Ontogenic obstacle is obstacle which is related to child mental development stage in accord with its age and biological development.Some children sometimes lack of ability which is needed for cognitive aim related with age.If its obstacle derived from slow mental development (and not because pathology situation), then it will disappear together with its growth.Didactical obstacle is the obstacles which occur as result of learning choice which related with education system.This obstacle can be avoided through development of alternative learning approach (which is called by Brousseau as didactical engineering).The Introduction of integers by using a horizontal number line can cause conceptual onto genic and didactical obstacle (Suryadi, 2016).In understanding of number, onto genic obstacle, partly due to the inability of the children experienced the process of sense making, especially interpreted the 0 and negative integers.Forming process of the concept image of 0 and negative integers are difficult to represent in a concrete that make didactical obstacles, so it requires more thinking allows children.The last, epistemological obstacle is obstacle which can be avoided because of its important role in constructing its knowledge concept itself.According to Duroux (Brouseau, 2002, p. 99), epistemological obstacle is one's knowledge is limited to a certain context only.If the person is faced with a different context, the knowledge becomes unusable or she/he had difficulty to use it (Suryadi, 2016).Brown (2008) stated that "Epistemological obstacles can be construed as faulty ways of thinking but such a perspective ignores their importance, their developmental necessity, and their productivity in specific settings."Cornu (2002) argued that epistemological obstacles are not related with learning approach which is used by teacher, but as result of mathematical concept nature itself.The argument proposed by Cornu refers to Bachelard (1938) who explained that epistemological obstacles occured both in history of scientific thinking development or in education activity.According to him, epistemological obstacle is important thing to obtain knowledge and it is found in history of concept development.Therefore, educators (in this case, teachers) should learn deeply epistemological obstacle in history of mathematical development in order to understand student obstacle in learning the concept.Those two things will complement each other as revealed by Brousseau.According to him, the most important things is to learn deeply epistemological obstacle by identifying concept error which is experienced by students which then compared to obstacle in its mathematical history to determined its epistemology character (Brousseau, 2002).
In terms of epistemological obstacles and understanding of the concept, Schneider (2014) explains that epistemological obstacles have some relations with conception or concept error in the form of mental object, concept image, and student's point of view or thinking.It often happens that mathematical concept possessed by students is based on definition they know from their point of view.This may become one of student's obstacles in thinking mathematically to solve the problem (Vinner, 2014).This opinion is also revealed by Schmidt (1997) that difficulty often occured when students' ideas are different from the definitions accepted by experts.Initial concept which is understood by students sometimes not suited with concepts of scientist.Concept differences between what is understood by students in their learning process and knowledge science which cause concept misunderstanding or misconception (X.Li & Y. Li, 2008).From a child's point of view, this is normal and reasonable concept based on their experience in different context or in their activity in daily life thus it is not error from child's point of view (Fujii, 2014).
Related with concept of a material, there are some possibilities that cause learning obstacles, namely: (1) prerequisite knowledge; (2) concept understanding; (3) procedure; (4) principle; and (5) problem solving (Depdiknas, 2007).Procedure, principle and problem solving more related with structured knowledge which is student ability to apply concept (Widdiharto, 2008), included here the ability to translate the problem, identify problem solving scheme, determine strategy and solve the problem with appropriate algorithm.In relation with number, Bruno and Martinon (1999) use term 'numerical knowledge' which refer to concept, procedure, representation and algorithm.
Studies conducted on negative numbers showed the result that negative numbers have posed the problems for many high school students (Almeida & Bruno, 2014;Bofferding, 2014;Vlassis, 2008).The problems of negative numbers were caused by structure, position which was unknown, type of number or its context.Altiparmark and Ozdogan (2010) can reveal the difficulty in understanding concept of negative number that is understand numerical system, its direction and its amount, understand operation of arithmetic, and difficulty which is related with negative sign.Those three things are tightly related with abstract concept.This abstract concept of negative number is known as result in epistemological obstacle in development of mathematical knowledge (Larsen, 2012).
Objectives of Study
The underlying concept in mathematical material need to get more attention, because basic concept which can be understood well by students give influence in understanding learning material in higher level.In accord with what had been explained by Brousseau, researcher considered that it is important to identify student difficulty in understanding negative number.Through this identification, it is hoped that it can force teacher to create learning situation which is effective.The aims of this study are: 1) To study students' understanding about the definition of negative numbers in daily life.
2) To identify students' difficulties in understanding the concepts of negative numbers and their counting operation.
Participants
Totally 96 students (11-13 years old) in 7th grade in three different schools in Palembang, Indonesia, were involved as participants.Researchers choose this grade because negative number material as part of integer subject matter had been given in 7th grade.Respondents consisted of 58 students of two different Public Junior High Schools (32 in Group A and 26 in Group C) and 38 students of Private Junior High School (Group B).
Instrument
Researchers observed students' epistemological obstacles by investigated their difficulties in understanding negative numbers by administering test to 96 students of seventh grade which had learned about integer.25 problems (fill in blank and essay) were given to respondents related with integer and operation of integer.The questions were designed to identify the difficulties of respondents in understanding negative numbers which comprise aspects of prerequisite knowledge, concept of integer, counting operation procedure of integer, principle understanding, and problem solving which is adapted from diagnosis instrument of mathematical learning difficulty according to Depdiknas (2007) and Widdiharto (2008).Researchers also adapted three dimensions of negative number concept revealed by Altiparmark and Ozdogan (2010) as guidance in determining indicator of each difficulty aspect.
Instrument pass through validation qualitatively in material domain, construction and language by involving two mathematics education experts and three mathematics teachers.After revision, instrument was defined suitable to be used in accord with the aim of education.To obtain the accurate data of respondent understanding about negative number, researcher give problem which is related with procedure, principle, problem solving strategy in addition and subtraction operation which involve negative number.
Procedure
Test was given to all respondents as participants who agreed to take the test.Each respondent obtained item sheet and answer sheet and were not allowed to use calculator to ensure the objectivity and authenticity.Respondents are free to choose the number of item which will be solved first.Test implementation was also supervised by each math teacher in that class.Time provided is 90 minutes.
Data Analysis
Researchers gave score with 0 until 4 scales for each item.Therefore, the whole score was 100.This score was given based on the quality of respondent's answer to each item.This score was converted become 0 or 1 only with criteria 0 if score before are 0, 1, or 2 and 1 if score before are 3 or 4.This conversion considered that respondent who get score 0 -2 not fully understand the intention of item, false in concept, principle or procedure.
As for respondents who get score ≥ 3 are respondents who generally had answered correctly, even though there is miscalculation, but they had understand the intention of item, used concept, procedure and principle correctly.
Results
Analysis is done based on component of difficulty aspect to describe how many respondents who experience difficulty based on average score in each number of item.This analysis then continued by determining percentage of respondents who experience difficulty in each group.In the whole, researchers see the difficulty which almost distributed in all respondents in aspects of concept, procedure, principle and problem solving.This description showing information about difficulty experienced by respondents based on its source of problems in ies.ccsenet.
negative n Table 2. Pe
N =96
The numbe item
Standard deviation
In Table 2 and proced understood
Prereq
In average This was b integers are less than zero but argue that -5 is greater than -3 because -5 is further away from zero than -3 is, just as 5 is further away from zero than 3 is (Bofferding, 2014).For them, negative number is something abstract so when represent that number into concrete form they experience obstacles in translating it.
Understan counting o who have the propert
Furthermore, based on Table 4 also found that more than 50% of respondents unable to represent negative number in all contexts.The abstract nature of negative number become main epistemological obstacles which occur in students (Almeida & Bruno, 2014;Altiparmak & Ozdogan, 2010;Blair , 2012;Larsen, 2012).This can occur because thus far student is introduced to number concept as a real thing.Negative number which has different sign with positive number cannot be imagined in reality by students.The problem aroused in understanding the sign of number play a role in determining strategy which is used to solve the problem because the sign which opposite is more complex compared to the sign which is similar (Almeida & Bruno, 2014).
The procedure of counting operation has highest figure as one problem source of student's difficulties.Based on Table 5, showed that 95% respondents who are unable to finish calculation the operation of addition with negative number.The concept of counting operation occured because there was false perception held by students.Perhaps this false concept is result of perception they received from explanation which is given by teachers in the form operation of a + (-b) to become a -b.Respondents consider that positive number which side by side with negative will result in negative number.As for the form operation of a -(-b).The false perceptions become the main problem of arithmetic counting operation in students.Bofferding (2014) stated that when students interpreting movements as "more negative" or "less negative," they might ignore the direction of "positive" or "negative" and move more or less based on the positive number line, or they might pay attention to moving in the positive or negative direction without paying attention to whether they should move "more" or "less" in that direction.
On the average based on Table 7, it found that more than 85% respondents having obstacle on problem solving which one of difficulties related to negative numbers.The general problem encountered is that respondents are unable translate information which contain in this problem so respondents do not understand counting operation procedure is appropriate to describe problem condition.When they are able to understand the intention of problem, the main difficulty is determining strategy which is appropriate to solve the problem.The problem is that most students are unable to interpret a context into negative number by counting operation of subtraction.Glaeser (Heeffer, 2011) state that some epistemological obstacles experienced by students in learning negative number namely inability to manipulate the numbers of negative and difficulty to give the meaning on negative number.Students find it difficult in making meaning of negative number because the numbers of objects surround them is symbolized as positive number.It makes the students difficulty when being confronted with integer operation involving negative number.This is strengthened by result of this study that 60.4% of respondents experienced difficulty when they give context example which represent integer which involve negative number.
The strategy to solve the problem become not suitable which result in respondent solve the problem by false procedure.From a study delivered by Hughes (1986), some students who are able to translate the problem into mathematical model can be false in determining the solution of that problem.Student's knowledge which is very limited about counting operation, skill to apply basic operation process into new problem become the causing factor why students are unable to solve the problem (Hughes, 1986, p.8).Some studies which had been conducted in Indonesia showed that student understanding toward number still low particularly in doing counting process (Purnomo et al., 2014).
Children in aged 11 years old experience transition from concrete thinking into abstract.As revealed by Piaget, children begin to be able to do symbolic manipulation and enhancement of abstract thinking, and have intellectual potential to do formal reasoning (Hill, 2011).From this study, it can be see that students in grade 7th are transition from elementary school into secondary school, when initially they think informal with concrete operation switch to formal operation which is more abstract.It is not surprising that understanding the concept of abstract negative number become obstacle for student.A process bridging those two matters will be very helpful for students to minimize the difficulties occurred.
Negative numbers are an abstract concept for which students need phenomenological guidance in order to avoid epistemological obstacles (Larsen, 2012).Heeffer (2011) states that it is important for mathematics teachers to understand the history of the epistemological obstacle of negative numbers in the history of development of mathematical concepts to explore their students' difficulties when learning concepts of negative numbers.Identifying students' difficulties in understanding the negative numbers allows teachers to design a studying of the negative numbers as a follow-up.In planning to teach a mathematical concept it is of the utmost importance to determine the possible obstacles, particularly the endemic epistemological obstacles (Cornu, 2002).Therefore, an instructional design activity through the right didactical situation expected to minimize the students' difficulties on negative integers.
Table 1 .
The possibility aspect of students' difficulty problem source
|
2018-06-17T01:51:05.533Z
|
2016-12-26T00:00:00.000
|
{
"year": 2016,
"sha1": "6ab53901a70a6a6eda867d77ea7095a80cb9f805",
"oa_license": "CCBY",
"oa_url": "https://www.ccsenet.org/journal/index.php/ies/article/download/60549/35256",
"oa_status": "HYBRID",
"pdf_src": "Anansi",
"pdf_hash": "6ab53901a70a6a6eda867d77ea7095a80cb9f805",
"s2fieldsofstudy": [
"Education",
"Mathematics"
],
"extfieldsofstudy": [
"Psychology"
]
}
|
2268501
|
pes2o/s2orc
|
v3-fos-license
|
NOTES OF A RECENT VISIT TO SEVERAL PROVINCIAL ASYLA FOR THE INSANE IN FRANCE
Having taken considerable interest in promoting the study of mental diseases, and endeavoured to direct professional attention to that important subject, particularly in reference to the advantages of rendering the public institutions of London?appropriated for the reception and care of persons labouring under insanity, more available than heretofore to medical students, anxious to investigate the nature and study the treatment of that interesting class of maladies, 1 was induced, during past years, to pay various visits to the Parisian lunatic hospitals, in order to see the system pursued in regard to this point, at the great establishments of the French metropolis. Although well satisfied with the information then acquired, yet feeling desirous of farther examining the provincial French asyla for the insane, more at leisure, than previous tours enabled me to make, I accordingly visited several of these establishments during last autumn; and being much interested with many things which came under observation , I am therefore led to believe the short notes cursorily made during my peregrina
Besides, as the Minister can revoke the appointments of director or physicians, upon the report of the Prefet; as he settles the amount of the salaries of these officers; and farther, as the Prefets are the servants of the Minister, by whom they are appointed, and at whose pleasure they retain their offices, the Minister of the Interior thus becomes the sole patron and dispenser of all the important appointments now attached to the public insane asyla in France; much in the same way as the Minister of Justice has the patronage of those in the law nnd legal tribunals.Although the physicians must reside, according to this ordonnance, within the asylum, some may, by favour, nevertheless, obtain a special permission from the Minister to live elsewhere; but in that case the party ought to visit the lunatics confldcd to his care, at least once every day, and should lie be prevented doiug so, this duty must be performed by another physician.
The above are some of the general regulations respecting public insane asyla in France; but when any person is desirous of obtaining a licence to open a private establishment, the applicant presents a petition to the Prefet of the Department in which the proposed asylum will be situated, to whoso satisfaction the petitioner must prove that he is twenty-one years of age, and in the enjoyment of all his civil rights; that his conduct and morals have been good during the three previous years, as shown by a certificate from the Mayor of the Commune in which he has resided; and lastly, that he is a Doctor of Medicine.However, where the individual does not possess the latter qualification, he may produce an obligation from some physician who engnges, with the Prefet's approval, to undertake the medical duties of, and to reside in, the asylum; and as the Prefet can, at any time, revoke this appointment, it is not likely the treat- ment of the patients will be much neglected.Further details respecting the constitu- tion and government of the public and private insane establishments of France, might he given; but enough having already been said regarding the general administration of these institutions, I will only now add, that besides the official persons previously mentioned, there are also two Inspectors-General of all the lunatic institutions of the Republic, whose special duties, amongst others, are to visit and report upon these establishments to the Minister of the Interior; and as these responsible offices are now ably filled by M. Ferrus, formerly physician to Bicetre, and M. Parchappe, lately physician to the asylum of Saint Yon, both very well known to the medical profession by their works on insanity; it is superfluous to speak of either of these gentlemen's qualifications for such important appointments.
Anterior to the first French revolution, the lunatic hospitals of that country, like other parts of Europe, then stood much in need of amelioration.The path first pointed out by the philanthropic Tenon, having been, however, zealously pursued by Pinel, that physician soon directed his benevolent mind to the treatment of insanity, and to improve the internal economy of those institutions, whereby great changes were effected, and the wretched condition of many insane patients greatly ameliorated.
Subsequently, Esquirol, and other eminent persons, made the study of mental diseases, as also the moral and medical treatment of the insane, the subject of their particular inquiries, so that increased attention being bestowed upon these important questions, great improvements were made in the science and treatment of mania, not only in France, but throughout the entire civilized world.
Considering it would be superfluous, as also uninteresting, to extend the notes I propose transcribing to an unreasonable length, my remarks will therefore be limited to eight provincial French asyia recently visited; since, to attempt more, would render the present r.eport unnecessarily voluminous.For similar reasons, I have also thought it advisable to refrain at present from alluding, except cursorily, to the public lunatic institutions of Paris, already so well known?namely,Bicetre for males, and the Sal- petriere for females; which are the only public hospitals for insane pauper patients in the French metropolis.There is, however, the hospital at Charenton, for those of either sex who can pay a stipulated pension towards their treatment and maintenance.Of these institutious, the Salpetriere hospital is, perhaps, the largest establishment of the kind in Europe, being not only a workhouse for infirm women, but likewise an asylum for pauper insane females belonging to Paris and the environs.At present, the entire population of the Salpetriere is about 5350, of whom nearly 1C00 are insane and epileptic patients.
To give some idea of the immense extent of this charitable institution, it may be stated, that upwards of 3000 pounds' weight of animal food are daily used in cooking, and in the kitchen, one of the largest in the world, the number of dinners daily prepared for the various inmates is often 5200; the materials of which are excellent.Bicetre, like the Salpetriere, is both a poor-house and an asylum for insane males belonging to the department of the Seine; but, although the entire population is under that of the Salpetriere, it is always very considerable, the number of lunatics being seldom below 1200, including paralytic cases, epileptics, and idiots; the improvement of whose lamentable condition, through the skill and treatment of M.
Voisin, has of late justly attracted much attention.Again, the asylum at Charenton, likewise a public establishment for lunatics, differs from Bicetre and the Salpetriere in two important particulars?viz., both sexes are admitted as patients, and they pay for their treatment and maintenance.This institution is administered under the authority of the Minister of the Interior; and the [scale of payments for board embraces three classes.The first pay 1300 francs, the second 1000 francs, and the lowest 720 francs per annum; but the Minister may authorise the admission of gratuitous patients.
Here, the total number of inmates varies from 450 to 500, of whom some belong to the highest orders in society.
Confining my notes, therefore, to asyla situated in the west and central departments of France, and for the reasons already stated, the first public institution I shall now bring under notice is that of Bon Sauveuk, at Caen.This hospice, besides being an asylum for lunatics of both sexes, is a religious establishment, containing a large population, which is thus enumerated in the official table given me by the authorities.1st.The choir and lay-sisters consist of 237 MOTES OF A RECENT VISIT TO individuals.2nd.There are five priests, composed of the superior and four chaplains.3rd.20 free boarders, all resident; 20 being ladies, and 0 gentlemen.4th.The deaf and dumb, of which there are 155, compi'ising 05 males, and 00 females.5th.
The resident domestics amount to 128, of whom 08 are males, and 00 females.And lastly, 002 lunatics; of whom 302 were men, and 390 women, at the period of my visit; thus making an aggregate population of 1243 persons, all living within the precincts of this institution.Besides the above numbers, there are also 2 physicians, and 80 non-resident work people; so that 1325 individuals now belong to this immense esta- blishment.
This very important, and truly, one of the most remarkable institutions of France, is situated in the city of Caen, or rather, in one of its fauxbourgs, and not far from the old abbey of St. Etienne, which contains the tomb of William the Conqueror.It has a fine, if not an extensive prospect of the neighbouring country; still the situation is low, and I should consider somewhat damp, especially as the meadows immediately adjoining are flooded every winter.The governing body of the establishment is composed of "les Sccurs Keligieuses," and the priests; the superior of whom is an autocrat, possesses supreme and irresponsible power in the management of the institu- tion ; whilst the order account to no one respecting their large revenue, expenses, or administration.This may in part be accounted for by the circumstance that its wealth and prosperity are mainly owing to the exertions of the clergy, and to former religious benefactors; particularly to the late Abbe Jamet, who endowed Bon Sauveur with all his property.This venerable ecclesiastic died in 1845, at the advanced age of eighty-three, having devoted his long life to promote the welfare of this his favourite establishment.
Besides the present hospice, the Order possesses two succursal institutions, one about thirty miles from Caen, the other near Albi, in the south of France; which last, I understood, is quite as important as that now described.Although the property of an ancient religions order, the surveillance of all lunatics resident within its walls rests, as in other French asyla, in the central government, who appoint the physicians charged with the treatment of every insane patient; still, the medical attendants' power is here not so defined as prevails in several institutions I could name, but of which more hereafter.
At the period of my visit to Bon Sauveur, the total number of lunatic inmates in the various divisions amounted to 002, as stated in a previous paragraph; of whom 302 ?were male, and 300 female patients; but of these, 212 women and 141 men pay for their board, lodging, and treatment, sums varying from 400 to 4000 francs annually; the remaining 330 inmates being all indigent persons.In reference to receiving private patients at this institution, it may be perhaps interesting to readers to mention, that the celebrated Beau Brummell died within the walls of Bon Sauveur; the apartment lie occupied in this madhouse being shown me, where that quondam companion of George the Fourth was supported, during the last years of his chequered life, by the liberality of Mr. Armstrong, the late British consul at Caen.Notwithstanding the numerous lunatic patieuts treated at this asylum, the medical staff consists of only two visiting physicians, Dr. Vastel taking charge of the male lunatics, and Dr. Fancon Duquesnay of the female department; the arduous duties of which they each perforin regularly and zealously.But as neither of those experienced gentlemen reside at the asylum, both being practitioners of repute in Caen, and there are no internes, it hence follows, when any emergency occurs requiring medical aid, the attending medical officer must be sent for to his private residence.This is a very great defect, aud I said as much to the authorities; but the holy sisterhood did not think it advisable to have pupils living within the walls of their convent, or even to allow a resident physician, although such is the strict law of the country, and as now carried out in most of the insane asyla throughout France.
During the year 1840, the movements of insane patients at Bon Sauveur was as follows:?
Died
Males, 31 Females, 21 Total, 52.Amongst the above deaths it should, however, be stated, that 18 men and 11 women died last year by cholera, whereby the mortality of the lunatics was considerably augmented beyond the ordinary average; and as similar effects from the recent epidemic will be noticed in the other lunatic institutions referred to in the present communication, this peculiar feature of the year 1819 must not l)e overlooked, when drawing conclusions respecting the ratio of deaths met with amongst the insane, in the various establishments I visited.
Not having been originally constructed for the reception of lunatics, the buildings are defective in many respects, especially the cells for the agitated patients, which have stone walls, iron bars, and unglazed windows ; the beds in such localities being frequently wooden cages, into which the afflicted inmate is even locked at night, besides being tied to the bed, or confined by a strait waistcoat.The superior class of patients liave, however, often excellent accommodation, some in detached houses, with gardens, particularly on the female side; whilst the gardens generally are beautiful, and kept iu excellent order.
The sisterhood, and others attached to Bon Sauveur, have only occupied the present residence since 1805, when they first received insane meu into their establishment; although previously, and even so early as 1728, they took charge of female lunatics, but more from charity than gain.However, consideriug the large number of paying inmates, amounting at present to 353 individuals, the sum now received by the executive must be considerable; and a director of an insane establishment in another part of France, who seemed as well acquainted with Bon Sauveur as he is also with the cost of keeping such institutions, told me distinctly the profit derived from the lunatics at this asylum, could not amount to less than 80 or 100,000 francs annually.
As in most French public asyla for the insane I have ever visited, the female lunatics in this hospice appeared much more agitated and noisy than the male patients; at the same time, mechanical restraint seemed oftener employed in the former than the latter sex.Thus, amongst the 390 insane females in the various wards on the day of my visit, 12 were in strait-waistcoats, some being also tied to seats or chairs, and one to a tree in the garden; besides which, three or four were very furious, and shut up in solitary cells, two of whom I noticed looking into the court-yard through a small opening made in the lower part of their cell door, the same as we usually find in English dog-kennels.This is no exaggeration, as I saw them with my own eyes, and also heard the poor sufferers howling within, and that even so recently as the month of August 1850!On the other hand, amongst the 302 male lunatics, only 3 were in strait-waistcoats, whilst three others had their arms restrained by leather straps.Besides these six patients, one man was also shut up in his iron-barred cell, who certainly made a very great noise, and thrust his clenched hands through the unglazed window as if to strike.This individual, the attendants said, was so dangerous, that they did not think it safe to approach within arms'-length of such an excited maniac.
Another male patient, who had recently committed murder, was likewise confined to his caged bed in a solitary cell; but this man was quiet, although very filthy.Refer- ring again to the female wards, I would repeat, the inmates were exceedingly noisy; and when passing through a court-yard, in which the most agitated female patients were confined, being accompanied only by the attendant sisters, I must acknowledge our position then seemed somewhat dangerous; and although I have often perambulated similar departments iu France, as well as in other countries, our sojourn here was far from agreeable ; and my kind conductors thought so likewise, whilst they told me that one of their sisterhood had been almost strangled by a furious maniac then at large in the court-yard we had just visited.
So little importance seems attached at this institution to employing tlie strait- waistcoat in refractory cases, that I was informed, if any sudden fit ofplirenzy seized a patient, the sisters in the ward, at their own discretion, would at once put the party in a camisole, or shut up the lunatic in one of the caged beds, to which they were likewise even bound by ligatures; and this'was done, they said, as well for the patient's own safety as that of others, and even of the attendants'; the physician being, of course, duly informed of such proceedings at his subsequent visit to the institution.
Notwithstanding the remarks now made respecting the employment of mechanical restraint at Bon Sauvejir, which I have now detailed from personal observation and inquiry on the spot, tie benevolent sisters perform their painful duty amongst the many afflicted fellow-creatures by whom they are surrounded, most zealously and courageously, according to their own views respecting the nature and treatment of insane persons ; but the system and machinery now in operation in such an establishment is bad; and 1 unhesitatingly assert, however well disposed or charitable the holy sisterhood are undoubtedly, there ought to be resident medical attendants, with internes, having the sole and responsible control over every patient; the entire management being also remodelled and made conformable to the reeent laws enacted by tlie legislature, according to which every public lunatic establishment in France ought to be now governed.
Employment to a considerable extent is carried out, both amongst the male and female patients; the former being frequently occupied in cultivating the extensive gardens of the hospice, in the large drying-house, or in various trades and handicrafts ; as also at a farm belonging to the establishment in the vicinity, where a portion of the insane patients are engaged in agricultural operations; and although axes in cutting ?wood, or other dangerous weapons, are often put into the hands of the lunatics, no harm or accident, I was informed, lias ever resulted from such employments.Amongst the female patients, knitting, sewing, household work, and various other occupations, for which women are by nature adapted, were likewise zealously encouraged and put in practice; so much so, that it gives me infinite satisfaction to finish this brief notice of the Bon Sauveur Asylum, by saying, however much I differ from the authorities of this institution, respecting their frequent recourse to mechanical restraint, the endea- vours constantly made to employ the insane patients are highly creditable and satis- factory.
When alluding to the holy sisterhood of this large institution, and the important position they still hold, it may be considered interesting to state, as an illustration of the persons often composing this religious body, that many belong to the upper ranks of society, of which the following is an instance.During my perambulations through the extensive apartments of the establishment, 1 was introduced to a noble " Comtesse" connected with a distinguished family, formerly attached to the court of one of the recent kings in France, when royalty was still in the ascendancy.This lady had also moved in the first circles of London, -where she likely joined, as also at Paris, in the gaieties of fashionable life.However, here she now was, clad in the plain flowing black dress of her order, with her head enveloped in an ample snow-white hood; the only ornament she wore being a silver crucifix on the breast.Although no friend to monastic institutions, I could not but admire the devotion of this worthy individual, ?who had dedicated the remainder of her future existence and energies to attendance upon the sick, and to works of charity.That she had, nevertheless, not forgotten the ?world, her previous acquaintance, or the places she had formerly visited, was evident from our conversation; and amongst other subjects we talked of, and to which I was able to reply to her inquiries, one had reference to a distinguished physician, a friend of my own, now resident in London, of whom she spoke in complimentary language, having been attended by that gentleman when indisposed in the English metropolis.But similar examples to the above are even now not uncommon in France ; and there is no class of persons who are so zealously disposed to dedicate their time and attentions to alleviate the miseries and bodily sufferings of their fellow-creatures, as many of these sisters of charity, who are often attached to the hospitals of this country.
This lunatic institution is situated near the ancient capital of Brittany, upon a rising ground, about half a league from the city of Rennes.It possesses a good situation, apparently healthy, from its natural advantages; has a fine prospect of the neighbouring country, and is not overlooked by contiguous buildings.The house is, however, old, particularly one part; and this portion of the building, not having been con- structed for an insane asylum, is not at all adapted for that purpose; indeed, one of the courts in the more modern quarter has been closed, because the female patients, who formerly there took out-of-door exercise, being overlooked by the male patients occupying the gallery above, the two sexes could thus converse together, which proved most improper and injurious.Even the best part of this structure looks more like a prison than a lunatic institution, having strong iron bars in the windows; and alto- gether it is not a proper receptacle for the insane, which seems also to be the opinion some of the managing authorities; whilst discussions had, I understood, actually taken place respecting the suppression of this establishment, and the removal of the patients elsewhere.From this and other causes, very little money was last year expended in repairs or improvements ; and as differences of opinion prevailed, according to report, betwixt the local and central government in reference to its management, these circum- stances only render the efforts made by the officials in the establishment, towards improving the well-being of the patients, more meritorious.; At present no resident physician is attached to St. Meen's, although there was one formerly.Since his removal by the Minister of the Interior to another asylum, the institution is only visited every other day by Dr. Pechot an eminent physician, in general practice at Rennes, appointed temporarily for that purpose, from his high quali- fications and reputation.Two internes, however, reside in the house, but as both can never be absent at the same time, medical assistance is always at hand; and if any case of emergency occurs, the visiting physician is immediately sent for.Should he be otherwise engaged, or out of the way at the time, then one of the professors of the medical school is called in to see the case, and give directions.Farther, in order that no delay may ever occur under such circumstances, a horse and carriage is always kept ready to bring the physycian or deputy to St. Meen's; whilst on the days Dr. Pechot does not visit the asylum, an interne goes round the wards, sees all the patients, and prescribes whatever he thinks necessary.
The gardens of the asylum are extensive, and seem productive; and recently, there was also a farm attached to the institution, where the male lunatics were often employed ; but this has been let to a tenant, whereby the agricultural occupations by patients has been interrupted, which is much to be regretted; however, hopes are entertained that the farm will be again resumed.
During the year 1849, the following was the movement of the insane patients at St. Meen's Asylum:?Admitted Males, 47 Females, 44 Total, 91.
Of the 49 deaths now reported, it is necessary to mentiou, that 21 were caused by cholera, 10 of whom were men, and 11 women.Again, with reference to the total number of lunatics in the institution on the day I visited the wards, they consisted of 143 men and 1G8 women; making an aggregate population of 311 lunatics then under treatment.
This Asylum seemed more tranquil than others I had visited in France during previous years; especially on the female side of the house, the appearances in this respect being somewhat similar to an English institution for lunatics, and certainly very different from the female wards of Bon Sauveur.This feature of St. Meen's, which I could not help observing, may perhaps be owing to the more staid Breton character, compared with other French provinces; but whatever may be the cause, it was cer- tainly not less remarkable, whilst, on the male side, there also reigned considerable tranquillity.I saw no patients confined in their cells, and only two females were in strait- waistcoats ; whilst three male patients were in camisoles, one of whom, however, I was informed, had been so kept for upwards of five years consecutively, excepting when locked up in his cell during the night, or occasionally left free for an hour or two in the day-time, to ease his limbs; but then he was always put into a cell for safety.No straps are used in this Asylum, and the strait waistcoat is the only instrumental restraint permitted, which measure must never be employed by any attendant without the specific order of the physician, or, in his absence, of the interne in charge of the patients; the great desire of all being to use restraint as seldom as possible; and doubtless this contributed to the marked quietude of the wards above observed.
Many of the patients were epileptics; several were affected with general paralysis, and some with erotomania; and what is somewhat singular, especially in France three, if not four, could not be prevailed upon to speak at any time, and one of these' a man, had remained silent for years.Two female idiots also attracted my special attention, on account of the great mutual attachment these unfortunate sufferers exhibited.They were constantly together, slept in one bed, sat on the same bench, walked always with each other, ate at the same time, and they even laughed, or attempted to speak simultaneously; in short, the two were inseparable; and my informant added, he had long noticed this peculiarity.
The bodily health of the patients appeared in a satisfactory condition; they are well fed, comfortably clothed, and although the house was old, and inappropriate for lunatics, its management seemed, on the whole, creditable to the authorities.Every effort is made to employ the patients, both male and female; the men in the gardens, cutting firewood, weaving, shoemaking, carpenters' work, also in the kitchen, out-offices of the institution, and so forth.The women seemed especially busy in various occu- pations, which was greatly promoted by the zeal and perseverance of the chief female attendant?somewhatsimilar to the matron in an English asylum.This lady attached great importance to employing lunatics, as it materially promoted their recovery; and in consequence, nearly all the female patients in the wards seemed employed, either in spinning, knitting stockings, making clothes, mending, sewing, and similar occupations.Others were working in the kitchen, assisting in the hath-liouse, washing clothes, and cleaning apartments; in short, almost every female seemed more or less engaged, and as a proof that the labour of the patients was profitably promoted, the whole clothing in the establishment is made on the premises; whilst the inmates are encouraged to work, by giving them a certain proportion of the money they thus gain, as a recom- pence; whereby they may in the meantime obtain a few little comforts, and also possess some money, however small, when they leave the institution.
As in most other public asyla for the insane, private patients, who pay for their treat- ment and maintenance, are likewise admitted at St. Meen's; but owing to the uncer- tainty respecting the future fate of the establishment, and other causes, the number of that class of lunatics at present is inconsiderable ; there being only 23 male patients and 19 females in the house, who pay from 500 to 1000 francs annually; and con- sidering all things, the charges seemed fair and moderate.
Although a lunatic asylum, strictly speaking, and not a workhouse, like some other establishments, there are still a few patients in St. Meen's not insane, being only afflicted with cutaneous complaints.This exception is in compliance with an ancient regulation still in force; but the number of individuals of this description was very small when I visited the institution, and they are not enumerated in the statistical details given of the resident population; the 311 patients reported, being all more or less afflicted with mental diseases; many of whose maladies, as in most public French lunatic asyla, were of long continuance, and therefore incurable.
Nantes Asylum.This celebrated establishment, like that of Bon Sauveur, forms only one division of the " Hospice General" of the ancient and populous city of Nantes.It is situated near the banks of one of the numerous branches of the Loire, in the fauxbourg of St. Jacques, on the south side of the river, having the water on one side, and aflat, open, yet richly cultivated country on the other.The gardens are extensive, very fruitful; and although the eye may be pleased with the external aspect of the buildings and locality, the situation is not well chosen, nor seems salubrious; and as the lunatic department forms only one division of an extensive charitable institution, having a large population, some of whom now labour under various forms of disease, whilst many others are paupers or infirm old persons; but especially as the entire establishment has only one large kitchen to supply the whole inmates, I consider the lunatic asylum, now under review, objectionably placed, notwithstanding the admirable arrangements of the building appropriated to the insane, which is one of the best throughout France; and may well be taken as amodel for similar structures by public bodies, alone anxious to promote the comfort and proper classification of the insane, without looking too much to architectural orna- ments, or ad captandum appearances.
Tlie total population of the Hospice General of Nantes, at the period of my visit, com- prised 1190 individuals, who were thus distributed:?1st, 190infirm old men, and 205 women of the same description, making a total of 455 inmates of what we call in Eng- land a workhouse.2nd, 25 deaf and dumb persons.3rd, 14 male and female boarders.4th, 101 orphans or foundlings?composed of 82 boys and 79 girls; and lastly, 391 lunatics; to which must be added 150 individuals who compose the staff of officers, internes, sisters, nurses, teachers, and attendants of every description; thus making the aDoreSate number of ] 190 persons, all resident in the institution.
Formerly, only one physician attended to the medical treatment of this large popula- tion ; now, two practitioners of repute, resident in Nantes, and in general practice, have the medical charge of all the inmates excepting those who are insane; which important duty is confided to Dr. Bouchet, so well known in France as elsewhere, for his scientific attainments, and experience in mental diseases.The physician resides at the institution, and gives his whole time and attention to the insane patients of the asylum, being debarred from pursuing private practice.He has two internes, who reside in the house constantly; and if Dr. Bouchet considers it necessary, at any time, to have a consulta- tion of physicians, or surgeons, especially in physical disease, or if an epidemic prevails amongst the inmates of the lunatic wards, he may call in any of the civil practitioners resident in Nantes he thinks advisable, who are paid by the administration of the hospice, for services so rendered, as if in private practice.This is an admirable arrangement, and worthy of imitation, particularly in cases where surgical operations are required.
During the year 1849, the following official report shows the movement of the patients in the lunatic department.
Amongst the 100 deaths above enumerated, cholera carried off 37 male, and 24 female lunatics; thus making til deaths by tliat epidemic malady; whereby the mortality of the insane inmates was considerably augmented above the ordinary average.At the period of my visit, the total population of the division for lunatics amounted to 391 persons, as already stated, and was thus divided?Insane men, 181; women, 210: of whom 119 males, and 157 females were classed as indigent patients; whilst those who paid for their board, lodging, and treatment, consisted of 02 gentlemen, and -53 ladies; being altogether 115 persons, or nearly one-third of the entire number.
The pensioners are divided into three classes, who pay from 750 to 1700 francs annually; but one female patient, belonging to the upper class of society, has a detached house and garden to herself, with servants, for which she pays 4000 francs per annum.
Having been constructed expressly for the reception of lunatics, the interior accom- modation of the asylum is of a superior description; the dormitories are well ventilated; the court-yards judiciously arranged, and entirely separate; the one not overlooking the other, and having easy communication with the main galleries; whilst the attendants, from their own sleeping apartments, can easily inspect the various patients in each dor- mitory.Some of the private single sleeping rooms are perhaps rather small in dimension ; nevertheless, the interior arrangements are better than in many other establishments, and the classification of the inmates appeared judicious.There are no iron bars in the windows; the safety of the inmates being assured by wire screens, the same as may be seen in ordinary houses, to prevent the glass from being broken.In the out-galleries overlooking the gardens, and in which the patients walk during bad weather, there are, however, iron bars between the stone pillars, to prevent accidents, which is a proper precaution.
The general aspect of the dormitories and court yards, when the lunatics were assem- bled, was quiet and orderly; although the female patients seemed rather noisy, as in most French asyla; but they were much less agitated than at Bon Sauveur.The number of patients under restraint, on the day of my visit, was twelve females?all in the strait-waistcoat, with nine males, two of whom had only their arms tied by straps, the hands being free, and the remaining seven in camisoles.In addition, one of the male patients so restrained, had also his legs tied together by hobbles, whereby he could ?only move very slowly through the court-yard.This patient was reported to be very dangerous, both to himself and others; and that cause was assigned as the reason for such treatment.No lunatic was otherwise prevented from locomotion, or taking exer- cise ; and every cell was empty of tenants.In reference to the employment of the strait-waistcoat, it maybe further mentioned, that all the male patients then under me- chanical restraint were of the indigent class; but amongst the females thus treated, three were not of that description.
As elsewhere, many of the inmates were incurable patients, their disease being of long standing.One feature rather peculiar to this institution is, however, well worth recording?namely, the very small number affected with general paralysis, otherwise so ?common in France, especially in Paris, and upon which serious form of malady tbe French physicians have recently done much to illustrate; in this asylum the disease seemed of rare occurrence, only three cases of the kind being found amongst the whole 391 insane patients at present under treatment, two being males, and only one female.
Amongst the insane inmates, occupations of various kinds are carried on to a great -extent in this asylum, many being employed in the gardens, or in different trades and handicrafts, unnecessary now to particularise; employing the mind diseased, if possible, through bodily labour, being considered one of the chief means of amelioration.Several were working in the smithy; some as carpenters, or shoemakers; others were digging in the gardens, or carrying away loads and assisting in forwarding the improvements still in progress.Amongst the females, many were engaged in sewing, knitting, making clothes, cooking, and in other household duties.Besides the above occupations, a small NO.XII.
N N farm, of about thirty acres, being also attached to the hospital, the lunatics have thus a locality where they may be occupied in agricultural employments.But if I understood Dr. Bouchet's opinions correctly, that distinguished psychologist, although a decided advocate of out as well as in-door employments for lunatics, prefers horticultural to agricultural occupations; and in such a distinction I have heard other physicians also acquiesce.
In support not only of the advantages, but of the safety of employing lunatic patients in many ordinary occupations, although dangerous instruments may be even placed in their hands, it is most satisfactory to state, on the authority of such a witness as Dr. Bouchet, that during the last eighteen years he has been a zealous promoter of tlie labour system; no accident has resulted from so employing insane patients, excepting in a single instance a few years ago, when one of the lunatic labourers struck another inmate with a deal board he was then carrying towards another part of the premises.
But on this point it may be reasonably asked; might not similar consequences follow, under the same circumstances, even amongst sane persons ?The question is conclu- sive; whilst the idea of danger appears small, wherever the attendants adopt due precautions.
Before taking leave of the asylum at Nantes, it ought to be stated, that the system pursued in visiting the various patients, the minuteness and regularity adopted in all the details, in reference to their condition, treatment, and employment, struck me as being particularly worthy of imitation; whilst the registry of cases, the amount and quality of occupation performed by individual patients were particularly interesting ; in short, I have no hesitation in saying, this institution deserves much of the praise it has already received from other observers, whether native or foreign.
St. Gemmes Asylum, near Angeks.This provincial establishment for the reception of insane patients belonging to the department of the Maine et Loire, is situated in the village of St. Gemmes, on the neck of land which is bounded on the one side by the Maine, and on the other by the river Loire.It is built upon a rock close to the latter stream, about three miles distant from the ancient city of Angers, the former capital of Anjou?so celebrated in history, and formerly designated " La Ville Noire."Unlike the institution at Nantes, this asylum is appropriated exclusively to lunatics, and was originally an old chateau, first constructed in 1701 by a rich farmer-general of the public revenue.Many new buildings have, however, been recently added; and extensive alterations, with various improvements, are likewise now in progress; so that when all the contemplated structures shall be finished, it will exhibit quite a different aspect from the present, and better answer the purposes proposed.Until completed, strangers must therefore reserve their opinion; but judging from what I saw, and the plans of future operations obligingly exhibited for my inspection, the asylum of St. Gemmes-sur-Loire will doubtless become a first-rate institution.
The official machinery of this insane establishment is different from the institutions referred to previously, the resident physician being likewise superintending director, thus combining all the attributes of supreme medical and local administrative functions.There are no internes in the establishment, but the head of the asylum has a secretary and other subordinates to execute his orders and to manage details.Here, as in other asyla, there are also sisters of charity, who attend on the patients; but as all power centres in one individual, the discipline pursued was better than I have always noticed elsewhere.
During the year 1849, the movement amongst the insane patients was reported to be as follows:?Admitted Males, G9 Females, 57 Total, 12G.
Amongst the above seventy deaths, thirty-six males and twenty-six females, making a total of sixty-two individuals, were carried off'by cholera; hence only eight patients of the entire number died from other causes.It is also worthy of notice, that nearly all the deaths by cholera occurred from the 28th of August to the 16th of last September, being the identical period when the recent epidemic prevailed most fatally in the Britiph metropolis, the first seven days of the latter month having been designated by myself, in another publication, the " black week of 1849;" since 2298 human beings died in London from cholera alone during that short period of time, besides 885 by other causes.
On tbe day of my visit to St. Gemmes, the asylum contained 1C1 male patients and 179 females, making a total of 340 lunatics.Many of the inmates were incurable, their disease being of long continuance.Epileptics were numerous ; general paralysis was common; a good many were idiots; several were cretins; and what is of rare occurrence in French asyla, two or three pellagreuse insane patients were pointed out for special observation.The women generally seemed more agitated than tbe male lunatics, the same as elsewhere observed, although not quite so noisy as I have some- times witnessed.The men were more quiet, and certainly less talkative than the other sex, which peculiarity accords with general observation throughout France.
A very large proportion of the inmates were indigent persons, about thirty only, out of the total of 840 lunatics, being pensionary or paying patients, for whom tbe sums received varied from 500 to 1700 francs per annum.But from the information given me, the wish of the authorities, and especially of the medical director, seemed to be to render this institution a receptacle for the treatment of poor maniacs, rather than a place for receiving rich patients.The motive is highly creditable, and deserves recording, which I now do with much satisfaction.
No iron bars are seen in the dormitories, or any other apartment, excepting the eells for agitated patients, the windows throughout the establishment being protected bywire screens.The dormitories are well aired and spacious, the new portions being superior to the old; but as the latter will ultimately be replaced by improved struc- tures, the defects now apparent will then disappear.The court-yards, although spacious, are defective of ornamental plots or flowers, as in most other asyla, and a cabinet d'aisance being placed in the centre of most, that object does not improve their appearance.Cleanliness and strict discipline pervade the whole establishment, and 1 could not avoid remarking the benefits accruing to the inmates by the orderly and systematic management pursued, which was creditable to the medical director, whose authority over every detail, whether lay or professional, was absolute; so different from the concurrent jurisdictions I have occasionally elsewhere observed.
During the peregrinations I made through the court-yards and dormitories, the number of patients then under restraint was as follows:?Male patients, twelve, all in camisoles, and two, if not three, also tied to their beds, having been very violent during the preceding night.The female patients in restraint were more numerous than the male lunatics, fifteen persons of tbe softer sex being also in strait-waistcoats, four of whom were likewise confined in bed, to keep them from injuring others or themselves.None, however, were shut up in any of the cells; and all the twentyseven I saw under mechanical restraint were indigent patients.
Mental disease is rather more common in this district of France than in some neighbouring departments; intemperance and religion, but especially hereditary tendency, being a frequent cause of insanity amongst the native population.So great an influ- ence, indeed, has hereditary tendency in the production of mental maladies, that in at least three-fifths of the cases met with, I was informed, such a predisposition may be traced.This makes a higher proportion than other physicians have usually reported; but it only renders the above fact more interesting.As a collateral illustration of the transmission to offspring of peculiar features in physical, as it is in mental organiza- tion, through many generations, I would mention tbe marked resemblance which several of the insane patients bore to the countenance and bodily frame of the ancient Romans; and one maniac in particular was pointed out to me, whose features, shape of head, and general contour of person, seemed exactly those of an old Roman.Indeed, if his bust bad been taken in marble, it might almost pass for that of some antiquated senator or emperor.This will, however, appear less singular, when it is remembered, that the neighbouring district constituted formerly a very important Roman station, and was long occupied by that warlike people, who built a bridge over the Loire, near St. Gemmes, at a place still called " Pont de Ce," or the bridge of Caesar.Besides which, tbe remains of an amphitheatre, and other interesting antiquities, are still visible in other localities near Angers.
Although the gardens of the asylum are frequently flooded in spring and autumn by inundations of the Loire, during summer, there is often a great want of water for necessary purposes.This deficiency of so essential an element, especially in an asylum for the insane, tbe director proposes to remedy by the erection of a steam engine, or a machine driven by horse power, to pump water from the adjacent river.This will prove a great adjunct to the establishment, and one which is very much required for the health and comfort of every resident.
N N 2 ON TiEDIUM VIT.E.
Occupations of various kinds are carried on at tlie asylum by the inmates of both sexes, and much of the labour in the gardens is performed by patients.The alterations and improvements also in progress are advanced by their means; and in the adjoining farm ?having upwards of forty acres, a similar system is pursued, the pigs, cows, and so forth, being often tended by insane patients, wherever practicable or judicious.A considerable part of the household work is also done by the female patients.Others are likewise occupied in knitting, sewing, mending, and making clothes; which employments are both profitable to the institution, as in part, likewise, to the individuals themselves.In fact, here, as in all French lunatic asyla, employing the inmates is considered essential, and a great adjunct in promoting recovery.
Having stated in a previous paragraph, that the offices of physician and director of the asylum are combined at St. Gemmes, before concluding this brief sketch of the above institution, it now only remains for me to say, Dr. Levincent fills these two important and responsible appointments, as he has, I believe, done for several years past, with much credit to himself and benefit to the institution.The task is, however, herculean, as he not ouly treats medically the various insane patients under his charge, registers their history and cases, but likewise superintends the feeding, clothing, and domestic arrangements of this extensive establishment, to say nothing of the financial and other departments under his immediate surveillance.Besides such laborious occu- pations, the numerous improvements and alterations now in progress likewise occupy much of his time and attention; but as Dr. Levincent is known to be a physician of energy and talent, he can, therefore, undertake official duties which to men otherwise constituted, would seem insuperable.
(To be concluded in the January number.)
|
2018-05-31T08:20:17.042Z
|
2016-01-01T00:00:00.000
|
{
"year": 2016,
"sha1": "b3d6e15d49d958955a3aff6da33265555c05cf3d",
"oa_license": "CC0",
"oa_url": null,
"oa_status": null,
"pdf_src": "ScienceParseMerged",
"pdf_hash": "b3d6e15d49d958955a3aff6da33265555c05cf3d",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"History",
"Medicine"
]
}
|
240505289
|
pes2o/s2orc
|
v3-fos-license
|
Abortion in Captive Gray Brocket Deer (Mazama gouazoubira) Associated with Colloid Goiter, Hemonchosis and Necrotizing Rumenitis
Background: The gray brocket deer (Mazama gouazoubira) is a specie that shows great adaptability in different habitats and it is the most abundant deer specie in South America. The present work describes for the first time a case of abortion followed by death associated with colloid goiter, massive hemonchosis and necrotizing rumenitis in a captive female gray brocket deer. Case: A 4-year-old female gray brocket deer (M. gouazoubira) raised in captivity had a history of abortion during the last third of gestation. The animal was kept in an enclosure together with 3 other gray brockets deers, being 1 male of the same age and 2 juvenile brocket deer of approximately 1 and 2 years old. The animals were fed with concentrated used as cattle feed and dewormed annually with 1% Ivermectin. The animals’ enclosure had vegetation cover formed by grasses and soil. The animals appeared healthy with no behavioral changes. The day after the stillbirth, the mother was found dead in the enclosure and sent to the animal pathology sector of the University of Vila Velha (UVV), Brazil. Necropsy revealed that thyroid lobules were highly increased in volume and histopathological findings were compatible with colloid goiter. A large number of nematodes were found in the abomasal content, totalizing 11,626 helminths, which were morphologically characterized as Haemonchus contortus. Grossly, the serous and ruminal mucosa exhibited an extensively reddish focal area with irregular contour, surface ulceration and a firm consistency. Microscopically, a severe necrotizing rumenitis was diagnosed. The liver showed pale multifocal areas on the subcapsular surface, friable to the touch which deepened when cut. Histopathological analysis revealed an accentuated multifocal panlobular coagulative necrosis, characterizing an acute liver necrosis. Discussion: Iodine is a mineral of great importance for thyroid hormones synthesis and your requirements are higher during pregnancy and lactation. Diets deficient in iodine causes a reduction in the basal activity of the hormones triiodothyronine (T3) and thyroxine (T4) and over-stimulation of the thyroid by thyroid stimulating hormone (TSH), resulting in goiter. In the present case, it is possible that the shortage of iodine in diet caused a goiter and, as a consequence, triggered the abortion. Haemonchus contortus is a pathogenic nematode of small ruminants, leading to decreased productivity and death in some cases as a result of anemia and hypoxia. The contact between domestic and wild animals, resulting in the emergence of infectious diseases and the spread of pathogens among species. In the present case, manual counting accounted for 11,626 H. contortus larvae, characterizing a massive infection and justifying the condition of severe anemia. The high parasitic load shown in this case points out this parasite’s importance related to this species in captivity. In general, inflammatory lesions in the rumen are results of excessive intake of fermentable carbohydrates, which leads to a considerable decrease in ruminal pH and leads to a high proliferation of lactic acid bacteria. This lesion has been previously reported in cervids. This case of comorbidities demonstrates that failures in nutritional and health handling, may cause simultaneous multiple diseases leading to death. Preventive measures for helminth parasite control and a proper feeding management with an adequate diet must be provided in order to preserve the species in captivity.
INTRODUCTION
Cervids are ruminant ungulates of the order Artiodactyla, distributed worldwide in multiple countries and biomes [2]. Feeding is one of the most important aspects in the management of these animals in captivity, and basically consists in concentrate, such as cattle food, and roughage [2]. Diets containing low levels of iodine highly predisposes to development of goiter, clinically characterized by hypothyroidism and abortion in some cases [8]. The gray brocket ( Mazama gouzoubira) is the most abundant deer species in South America, and as far as we could notice, there is no previous report of nutritional goiter followed by abortion in this species [6]. The natural ability to adapt to different environmental conditions allows contact between the gray brocket deer and some domestic ruminants, which may result in the spread of pathogens among species, such as gastrointestinal parasites [3,9].
Haemonchus contortus is a gastrointestinal nematode frequently diagnosed in small domestic ruminants [12]. In the gray brocket, this parasite has been described in studies concerning the description of the helminth fauna [3], and indicators of parasitic infection [4]. Although haemonchosis has been additionally reported in free-living gray brocket deer [3,4], its importance in captivity deer has not yet been documented.
Ruminal acidosis due to carbohydrate overload is a major problem in livestock [13], but its occurrence is still poorly described in wild ruminants [6]. Therefore, the aim of this study was to report a case of abortion associated with colloid goiter, massive hemonchosis, and necrotizing rumenitis in a captive gray brocket deer.
CASE
A 4-year-old captive female gray brocket deer (Mazama gouazoubira) presented abortion during the last third of gestation. The animal was being kept in captive, with no previous history of any clinical comorbidities. A day after abortion, the mother was found dead in the enclosure and was referred to necropsy at University of Vila Velha, Brazil. The female was being kept in an enclosure with 3 other gray brocket deer, 1 male of same age, and 2 younger brocket deer (1 and 2 years old, respectively).
The animals were been fed every 24 h with concentrate used in cattle feeding, and were annually dewormed with Ivermectin 1 . Vegetal cover inside the enclosure was composed mostly of Brachiaria spp. On post mortem examination blood was hydremic and mucous membranes were intensely pale. The thyroid lobules were bilaterally and diffusely enlarged, and firmly consistent ( Figure 1A). Within the uterine body there was a moderate amount of cloudy and mucous content. In the gross evaluation of the forestomaches, a focally extensive reddish area of ulceration with irregular and firm borders was observed in the ruminal mucosa ( Figure 1B). In the abomasal content there was a large number of white and red striped and cylindricalshaped helminths, with average corporal length of 30 mm ( Figure 1C). The large intestine loops were diffusely filled with a dark red fluid (digested blood). A few areas of serosal hemorrhages were additionally seen in the centripetal gyri of the colon. In the subcapsular hepatic surface, there were multifocal pale, deep, and friable areas ( Figure 1D). During necropsy, the abomasal content was sent for laboratory analysis in order to perform morphological characterization of the helminths and the parasitic load. Based on morphological features, parasites were characterized as Haemonchus contortus [12]. A total of 11,626 helminths were accounted for analyzed abomasal content.
Samples of thyroid, forestomaches, abomasum, small and large intestines, liver, uterus, trachea, lung, heart, lymph nodes, spleen and kidney were fixed at 10% buffered formalin, and submitted for routine histological processing. Microscopically, thyroid follicles were lined by a single layer of flattened cells, and markedly dilated and filled by a homogeneous, dense and eosinophilic material (colloid) [ Figure 2], which characterizes colloid goiter. In the ruminal mucosa, there was an intense necrotizing rumenitis.
Histopathological analysis of the liver revealed a marked multifocal panlobular coagulative necrosis, with severe and diffuse vacuolar glycogenic degeneration. Such findings were characteristic of acute liver necrosis.
Additional findings included acute moderate catarrhal metritis and mild neutrophilic interstitial pneumonia.
DISCUSSION
Iodine is an essential mineral for human and animal health, since it is essential during the synthesis of thyroid hormones. Deficiency of iodine in the diet causes a decrease in the basal activity of the hormones triiodothyronine (T3) and thyroxine (T4), and over-stimulation of the thyroid by the thyroid stimulating hormone (TSH), resulting in goiter [8]. Iodine requirements are higher during pregnancy and lactation. In humans, inadequate mineral intake during pregnancy may lead to goiter, neonatal hypothyroidism, intellectual disabilities, and abortion [7]. In the present case, no clinical manifestations of hypothyroidism were previously noted. So, it was not affordable to perform serological dosages of any thyroid hormones by the time of the postmortem exam.
A retrospective study carried out in the state of Mato Grosso, Brazil, analyzed some epidemiological and anatomopathological findings of three outbreaks of goiter in cattle, which lived in properties with history of abortion and birth of weak and underdeveloped animals, highlighting the importance of the disease in livestock [5]. Thus, it is possible that the shortage of iodine in diet has triggered goiter, which resulted in abortion in the present case. According with a previously study [6], the most frequent causes of death in gray brockets were traumatic damage (36.6%), respiratory (15.2%), and capture myopathy, with no report of goiter in this study.
During the last few years, the constant loss of habitats has approached domestic and wild animals, resulting in the emergence of several infectious diseases [11] by the spreading of pathogens among species [3,4,9]. A study performed in the Brazilian Pantanal with 10 gray brockets suggested that the helminth fauna of these animals may have been derived from domestic cattle introduced in that area [3]. It has been additionally described that H. contortus is frequently associated with infections in this cervid species by comparing both the number of infected animals and the parasite load, suggesting that the host may be fairly adapted. Due to its wide distribution in South America, cervids are often seen around areas occupied by agriculture and domestic animals [3], which may act as important sources of infection for cervids.
Parasitism by H. contortus represents an important issue in small ruminants breeding, leading to production drops and death in some cases [10,12]. Anemia and hypoxia occur as a result of the bloodsucking habits of the parasite, and are the main reasons for death. The parasitic load is an important factor in the severity of haemonchosis, once a previous study reported that each adult helminth ingests about 0.05 mL of blood per day [1]. In the present report, manual counting totalized 11,626 larvae, characterizing a massive infection of H. contortus, and justifying the condition of severe anemia. A few strategies have been developed in order to reduce the animal losses related to the parasitism. An experimental study described a reduction of H. contortus larvae (L3) in sheep feces through the use of conidia of the nematophagous fungi Duddingtonia flagrans and Monacrosporium thaumasium [10], which can be an alternative control method for hemoncosis in captive wild ruminants. The high parasitic counting described here enhances the importance of this host-parasite relationship in captivity situations.
Inflammatory lesions in the rumen are frequently reported in most domestic ruminants, and in free-living gray brockets [6]. Basically, these chemical injuries occur as result of excessive intake of highly fermentable carbohydrates, which leads to a sudden decrease in ruminal pH, leading to a high proliferation of lactic acid producing bacteria [13]. The ruminal lesions found in the present case were consistent to those reported in cases of rumenitis in the literature [13]. The feeding management of the reported deer was based on measures of concentrate for cattle, which was probably the cause of the ruminal lesion. Although it was not possible to identify the bacterial agent involved in the present case, the bacterial agents frequently involved in cases of ruminal acidosis followed by necrotizing rumenitis used to be from the own ruminal microbiota, such as Fusobacterium necrophorum and Clostridium spp. [13]. Moreover, liver damage may have resulted from septic thromboembolism, secondary to the primary ruminal lesion.
In conclusion, it is reported a case of abortion associated with colloid goiter, massive haemonchosis and necrotizing rumenitis in a captive deer (M. gouazoubira). This is the first report of colloid goiter in this species. Moreover, successive management failures may cause concomitant diseases and lead to death in a short term. Preventive actions to control parasites and provide adequate diets during pregnancy and postpartum must be adopted to preserve the species in captive.
|
2021-11-04T00:08:05.517Z
|
2021-01-18T00:00:00.000
|
{
"year": 2021,
"sha1": "5613fd56665b747d46d371fe0dc922921118709e",
"oa_license": "CCBY",
"oa_url": "https://seer.ufrgs.br/ActaScientiaeVeterinariae/article/download/114003/pdf",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "91d148156be72bf6766d6a2f11b42fd48eca0d9d",
"s2fieldsofstudy": [
"Environmental Science",
"Medicine",
"Biology"
],
"extfieldsofstudy": [
"Biology"
]
}
|
270487450
|
pes2o/s2orc
|
v3-fos-license
|
Global research landscape and emerging trends in Graves’ disease: A bibliometric analysis
Background: Graves’ disease is a prevalent thyroid disorder and is the primary cause of hyperthyroidism. Significant progress has been made in understanding the epidemiology, pathogenesis, diagnosis, treatment, and prognosis of this disease. However, bibliometric analyses on Graves’ disease are lacking. We aimed to comprehensively summarize the research, progression, and focal points of Graves’ disease through data mining and integrated analysis of the existing literature. Methods: We retrieved relevant literature on Graves’ disease from 2003 to 2023 from the Web of Science database. We performed bibliometric analysis using CiteSpace and the R package Bibliometrix. Results: We identified 10,901 publications from 132 countries, with a steady rise in the number of publications over the past 5 years. The US leads in publication volume, with the University of California System being the primary contributing institution. The journal Thyroid had the highest publication output, while the Journal of Clinical Endocrinology and Metabolism was the most frequently cited. These publications involved 2305 authors, with Antonelli Alessandro and Smith Terry being the most prolific. The most frequently cited articles were the “2016 American Thyroid Association guidelines for diagnosis and management of hyperthyroidism and other causes of thyrotoxicosis” and the “Thyroid Association/European Group on Graves’ orbitopathy guidelines for the management of Graves’ orbitopathy.” Analysis of the bursts of cited references, keywords, and their clustering revealed that research on Graves’ disease predominantly centers on clinical management, thyroid-stimulating hormone receptors, thyroid hormones, autoimmunity and inflammation, Graves’ ophthalmopathy, thyroid nodules, and thyroid cancer. Conclusion: This is the first comprehensive bibliometric study to summarize progress and trends in Graves’ disease research. These results highlight recent research hotspots and promising directions, thereby providing a valuable reference for other scholars.
Introduction
[3][4] It affects approximately 3% of women and 0.5% of men globally. [5]][9] Currently, the research on the pathogenesis of GD is not very clear.Complex interactions between genetic susceptibility and environmental factors, such as viruses, radiation, drugs, and iodine, are considered major contributors to GD. [10,11] These risk factors may lead to an imbalance of T lymphocyte subsets, resulting in increased differentiation of pro-inflammatory cells and decreased differentiation of anti-inflammatory cells.The immune imbalance in turn induces thyroid autoimmunity, resulting in thyroid cells presenting autoantigens directly to autoreactive CD4 + T helper cells.These cells are then activated, proliferate, and trigger the production of plasma cells that produce TSHR antibodies, leading to the occurrence of GD. [2,3] The primary treatment for this condition comprises antithyroid medications.However, with the identification of risk factors and a clearer understanding of their pathogenesis, more promising therapeutic approaches (e.g., immunomodulation) are undergoing extensive clinical research. [12]Over the past 2 decades, significant progress has been made in understanding the occurrence, development, and treatment of GD.The outcomes of prior studies form a foundation for future extensive in-depth research, aiding in the identification of hotspots and exploring new perspectives.
Bibliometric analysis is an emerging research method that involves the comprehensive analysis of various aspects of specific field-related literature, including publication quantity, journals, author information, citation frequency, and keywords.This approach aims to summarize past academic achievements and infer directions for future research and trends in a particular field. [13]CiteSpace and the R package Bibliometrix are commonly used tools in bibliometric analysis.Besides extracting the aforementioned information, they facilitate visualization of the results, [14,15] intuitively reveal the scientific development of cutting-edge knowledge, and help researchers grasp the current research status and trends in the field.Several bibliometric analyses have been conducted on Graves' Ophthalmopathy, autoimmune thyroiditis, Hashimoto' s thyroiditis, and thyroid cancer [16][17][18][19][20][21] ; however, to date, no bibliometric analyses exist on GD.Therefore, we conducted a bibliometric and visual analysis of GD research published in the Web of Science Core Collection (WoSCC) database from 2003 to 2023.We aimed to summarize research achievements, unearth research hotspots, help researchers quickly and comprehensively understand the current research status, and provide directions for future studies.
Search strategy
The WoS database, recognized as an authoritative source, served as the literature retrieval platform.We conducted a literature search on August 1, 2023, and downloaded all relevant documents.The search formula employed was TS = (Graves' disease), covering the period from January 1, 2003 to July 31, 2023.The document types included articles and reviews, and the language was restricted to English, with no additional search constraints (see Figure 1).
Data analysis and visualization
In this study, the advanced version of CiteSpace and the R package Bibliometrix served as analysis tools to extract key information and integrate relevant data.We employed them to visualize analyses related to countries and institutions, journal and co-cited journal analyses, author and co-cited author analyses, co-cited literature analyses, and co-occurring keyword analyses.Additionally, we generated overlay maps for journals and employed citation burst analysis for the literature and keywords.
Quantitative analysis of publications
The results of the literature analysis spanning the past 20 years revealed 12,798 studies related to GD, including 11,372 articles and reviews.Among these, 10,901 were English-language articles (see Fig. 1).The analysis of the annual publication volume (see Fig.As depicted in Figure 2, Phase 1 contained relatively fewer studies on GD.In Phase 2, there was a slight increase in research quantity compared with Phase 1, although the annual publication rate remained relatively low.Notably, Phase 3 demonstrated a substantial increase in the number of publications compared with the preceding phases, indicating a generally rising trend.Importantly, the decrease in the annual publication count for 2023 was attributed to the inclusion of articles published until July 31st of that year.
Country analysis
The analyzed articles originated from 132 countries and regions, with the top 3 contributors being the US (n = 2377), China (n = 1315), and Italy (n = 916), accounting for 35.75% of all publications.The top 10 countries collectively contributed 64.69% of all publications (see Table 1).
To visualize collaboration patterns between countries, we constructed a cooperative network graph based on the publication volume and interactions of each country.Noteworthy collaborations include those between the US and China, the US and the UK, the UK and Germany, and Germany and Italy (see Fig. 3).
Institutional analysis
The corpus of literature for this study originated from 708 institutions, with the top 3 institutions being the University of California System (n = 310), the University of Pisa (n = 283), and the N8 Research Partnership (n = 221).Among the top ten institutions, 5 were from the US, 2 from the UK, one from France, one from Italy, and one from China (see Table 1).
The collaborative networks between institutions revealed significant interactions.Notably, a close collaboration is observed among institutions within the University of California System, including the University of California, Los Angeles (UCLA); the University of California Los Angeles Medical Center; and the David Geffen School of Medicine at UCLA.Additionally, active cooperative relationships exist between the N8 Research Partnership and Harvard University as well as between the University of Pisa and the University System of Ohio (Fig. 4).
Analysis of journals and co-cited journals
The studied literature was distributed across 2402 journals.Table 2 presents the top 10 journals by publication volume, each featuring over 100 articles.The top 3 journals, Thyroid, the Journal of Clinical Endocrinology and Metabolism, and Clinical Endocrinology, have each published over 200 articles, totaling 1070 documents, which constitute 9.81% of the entire corpus.Among all journals, Thyroid boasts the highest output and impact factor (n = 469, 4.30%, IF = 6.6), followed by the Journal of Clinical Endocrinology and Metabolism (n = 375, 3.44%, IF = 5.8).The influence of journals is proportional to the number of co-citations.Thirty-two journals have been cited over 1000 times.Specifically, the Journal of Clinical Endocrinology and Metabolism has the highest number of citations (n = 6752), followed by Thyroid (n = 6218) and Clinical Endocrinology (n = 4935) (Table 3).
Furthermore, co-citation analysis revealed robust citation relationships.The Journal of Clinical Endocrinology and Metabolism exhibited active citation ties with Thyroid, New England Journal of Medicine, and Lancet (see Fig. 5).The co-citation overlay map (see Fig. 6) elucidates the citation relationships, with the left side representing citing journals, the right side representing cited journals, and the lines indicating citation connections.The 4 thickest lines denote the main citation paths, suggesting that papers published in journals related to molecular/biology/immunology and medicine/ medical/clinical fields were most frequently cited by articles in journals related to molecular/biology/genetics and health/ nursing/medicine.Among the 436 co-cited authors, the top 10 have garnered over 700 citations each, and the top 5 authors have received over 1000 citations (see Table 4).Bartalena (n = 1791) emerges as the most frequently cited author, followed by Bahn (n = 1434) and Weetman (n = 1391).Notably, Smith Terry, Kahaly George, and Tomer Yaron are the only 3 scholars ranking in the top 10 for both publication volume and citation count.
Analysis of co-cited references
Over the past 2 decades, there have been 21,968 co-cited references in GD research.The top 10 co-cited references, each cited at least 100 times, are listed in Table 5. Notably, 4 of these articles originated from the New England Journal of Medicine and had the highest IF.Among all co-cited references, the paper by Ross et al, titled "2016 American Thyroid Association Guidelines for diagnosis and management of hyperthyroidism and other causes of thyrotoxicosis," published in Thyroid, stands out as the most-cited (n = 408).Following closely is the article "The 2016 European Thyroid Association/European Group on Graves' orbitopathy guidelines for the management of Graves' orbitopathy" published in the European Thyroid
References with citation bursts
"Bursting references" denote literature that experiences a high-frequency of citations within a specific field during a particular period.Figure 9 shows the top 20 bursting references.The blue line segments represent time intervals, whereas the red line segments indicate periods of frequent citations.Figure 9 indicates that bursting references emerged each year.The burst strength of these 20 references ranged from 39.01 to 158.75, with durations of endurance spanning 2 to 5 years.
Among these references, the top 3 in burst strength are the "2016 American Thyroid Association guidelines for diagnosis and management of hyperthyroidism and other causes of thyrotoxicosis" (strength = 158.75), the "2016 European Thyroid Association/European Group on Graves' orbitopathy guidelines for the management of Graves' orbitopathy" (strength = 115.85),and "Graves' disease" (strength = 113.87).Notably, these references not only exhibited the highest burst strength but also endured for a period of 4 or more years.
Keyword analysis
A network graph of keywords revealed research hotspots within a specific field.Table 6 presents the top 40 high-frequency keywords used in GD research."Graves' disease," "autoimmune thyroid disease," and "Graves ophthalmopathy" claim the top 3 spots, each appearing over 2000 times, while "hyperthyroidism" and "management" exceed 1000 occurrences.The top 15 keywords surpassed 700 occurrences, and the top 40 keywords appeared at least 190 times, representing the primary research direction for GD.
General information
Medical journals have played a pivotal role in the historical dissemination of medical information worldwide, spanning over 3 centuries from the inception of the first traditional journals to today' s open-access publications. [32]The proliferation of journals and the increasing volume of published articles reflect the continuous progress and evolution of medical research.However, the advent of the information explosion has also posed challenges in obtaining a comprehensive understanding of knowledge in specific fields, during a given period, and across particular topics. [33]he emergence of bibliometric theories and tools offers effective solutions to these challenges. [34]As a research methodology, bibliometrics enables the systematic and quantitative analysis of scientific literature, providing valuable insights into the landscape of scholarly communication.By leveraging bibliometric approaches, researchers can navigate the vast sea of published literature, identify trends, and extract meaningful patterns in scientific research.The history of GD dates back to 1835 when Irish physician Robert James Graves first reported this condition, marking nearly 2 centuries of research. [35,36]In this study, we conducted a systematic search of the WoS database spanning the past 2 decades (2003-2023) to compile a comprehensive overview of GD research.After applying meticulous exclusion criteria, we included 10,901 English-language papers from 708 institutions across 132 countries, involving 2305 scholars.These papers were published in 2402 different journals and referenced in 21,968 articles.The analysis revealed a noticeable increase in publications since 2017, suggesting a growing focus on GD research in recent years.Collaboration appears to be a key trend, with multi-institutional and international cooperation emerging as a direction for future research.Although current collaborations primarily involve a few countries, institutions, and domestic authors, fostering broader partnerships is imperative for long-term development. [37]In examining publication trends and citation patterns, we observed that most GD research was published in endocrinology-related journals.However, noteworthy contributions from high-impact multidisciplinary journals underscore the significance and broad relevance of GD research.Dualmap overlay analysis further implies that research on GD straddles both the basic and clinical domains, reflecting a phase of concurrent development in these areas.Nevertheless, there is a substantial difference in the progress achieved by basic and clinical research, with clinical studies, especially guidelines and management-related publications, dominating the landscape.[24][25][26][27][28][29][30] This study, conducted by Ueda et al, identified a close association of the T-cell regulatory gene CTLA-4 with GD pathogenesis through comparative gene analysis in different participants and splice studies in mice published in Nature. [31]A noteworthy observation emerged from the analysis of the authors and co-cited authors.The top 10 authors, all with 40 or more publications, included Smith Terry, who contributed 2 highly co-cited articles published in the New England Journal of Medicine.A PubMed search indicated that many of these authors had consistently published studies over the past 5 years of the study period.Hence, these prolific researchers warrant attention as their work may represent potential paths for future exploration, providing valuable insights and inspiration.
Clinical management of
sweating, palpitations, tremors, behavioral changes, a lack of concentration, emotional instability, accelerated growth and bone maturation, thyroid enlargement, pretibial myxedema, and infiltrative exophthalmos (Graves' ophthalmopathy). [3]econd, biochemical blood tests should indicate suppressed thyroid-stimulating hormone (TSH), elevated free triiodothyronine, and/or elevated free thyroxine concentrations (an elevated elevated free triiodothyronine level is a more sensitive marker of GD than an elevated and/or elevated free thyroxine level). [38]Third, immunological blood tests should reveal elevated titers of TSH receptor antibodies (TRAb).The diagnosis of GD is confirmed when all 3 criteria are met. [22]or patients with moderately elevated TRAb, a diagnosis can be established if radioactive iodine uptake testing indicates diffuse uptake and an increased uptake rate. [24]The diagnostic process for GD in children mirrors that in adults.However, owing to its rarity and often subtle symptoms, this condition is frequently misattributed to primary psychological conditions or gastrointestinal and cardiovascular disorders, leading to potential harm to developing children and adolescents. [39]otably, pregnancy itself may cause palpitations and sweating, and increased estrogen secretion can elevate total serum thyroxine levels. [40]Since pregnant women cannot undergo radioactive testing, diagnosing GD during pregnancy is extremely challenging when the patient tests negative for TSH receptor antibodies. [41]2.1.2.Treatment.Although treatment approaches for GD vary among countries and regions, it is generally agreed that patients should undergo one of the following treatment methods: antithyroid drug therapy, radioactive iodine therapy, or thyroidectomy.[22] America, and Japan, antithyroid drug therapy is the primary treatment option [22,25] ; although radioactive iodine therapy is favored by physicians in the US, the use of antithyroid drugs is gradually increasing, indicating a trend toward becoming a primary treatment approach.[24,42] Although antithyroid drugs do not cure GD, a sufficient and prolonged course of treatment is highly effective in controlling hyperthyroidism.[22] Additionally, antithyroid drug therapy is the preferred option for GD in children, adolescents, and pregnant women.[43] The choice of antithyroid drugs includes propylthiouracil during the first trimester of pregnancy, thyroid storms, and adverse reactions to other antithyroid drugs.Methimazole is recommended for all other patients, including children and adolescents.[22,43] The typical treatment duration with methimazole is 12 to 18 months, with discontinuation when the TRAb and TSH levels normalize.However, if hyperthyroidism persists after completion of the treatment regimen, options include continued low-dose methimazole, radioactive iodine therapy, or surgical intervention.[22] 4.2.1.2.2. Raditive iodine therapy.Radioactive iodine is generally considered the optimal treatment for patients with mild thyroid enlargement or those for whom antithyroid drug therapy is ineffective or contraindicated.[22] Typically, a single dose of adequate radioactive iodine is administered to induce thyroid hypofunction in patients with GD. In hildren, a reduced dose of radioactive iodine is used to achieve a fairly rapid reduction in thyroid function and minimize the risk of recurrence.[22,44] Thyroid function should be monitored continuously for 6 months after radioactive iodine therapy until the patient' s thyroid function stabilizes through thyroid hormone replacement therapy.[22] Radioactive iodine is not recommended for patients with Graves' ophthalmopathy owing to the inherent risk of exacerbating the condition.Smoking individuals are also discouraged from radioactive iodine therapy, as they are more prone to Graves' ophthalmopathy.[45,46] Furthermore, studies suggest a long-term increase in cancer risk in patients receiving radioactive iodine; hence, this viewpoint remains controversial.[47,48] 4.2.1.2.3. Surgica treatment: thyroidectomy.Thyroidectomy is the preferred treatment for symptomatic compression or large goiters (≥ 80 grams) in patients with moderate to severe active Graves' ophthalmopathy.[22] Better results and lower postoperative morbidity rates were associated with surgeons who had more experience and higher surgical volumes. [49]Although the mortality rate of thyroidectomy is extremely low (<0.1%),complications such as postoperative hypocalcemia warrant attention. [50]For the clinical management of GD, addressing the challenge of a rational, early, and accurate diagnosis in children and pregnant women is imperative and represents a potential avenue for future research.Additionally, further research should focus on the dosage, treatment duration, and cessation criteria for antithyroid drug therapy; the potential long-term risks of radioactive iodine therapy for cancer incidence; and predictive and preventive measures for postoperative hypocalcemia following surgical treatment.4.2.1.2.4.Thyroid-stimulating hormone receptor and thyroid hormones.As mentioned earlier, GD is characterized by the abnormal increase in TRAb, leading to enhanced synthesis and secretion of thyroid hormones. [51] The types of TRAb have been identified: thyroid-stimulating antibodies, thyroidblocking antibodies, and neutral antibodies.[7,8] GD is closely associated with thyroid-stimulating antibodies.[7] Duan et al [52] examined the molecular structures of TSHs, TSH receptors, human-activated antibodies, and inhibitory antibodies.They identified the key amino acid sites crucial for hormone receptor-specific recognition.By comparing the states of activated antibodies, inhibitory antibodies, and the binding of TSHs to TSH receptors, activated antibodies were found to induce an upright active conformation of the extracellular domain (ECD) of TSH receptor (TSHR), causing receptorsimilar conformational changes and subsequent activation.[52] By contrast, inhibitory antibodies competitively bind to hormone receptors and inhibit receptor activation without inducing receptor-similar conformational changes.[52] This molecular-level study provides the first explanation of the pathogenic mechanism of GD and lays the foundation for the development of novel therapeutic drugs.Further research by Faust shed light on the aforementioned pathogenic mechanism, proposed a model for the physiological and pathological activation of TSHR, and suggested the potential extension of this mechanism to other G protein-coupled receptors with large ECDs. [53]an et al [54] designed a novel therapy using chimeric antigen receptor T cells.These cells demonstrate the ability to recognize and effectively eliminate B lymphocytes producing TRAb both in vitro and in vivo, offering a promising and innovative immunotherapy for GD.[54] Figure 10.Top 50 keywords with strong citation bursts.
On the clinical front, the critical importance of TSH receptor antibodies diagnosis and treatment, there is a major research focus on developing more sensitive detection methods and effectively reducing TSH receptor levels.Currently, 3 common methods are used to detect TSH receptor antibodies: competitive immunoassays, bioassays, and enzyme-linked immunosorbent assays (ELISA). [55]Competitive immunoassays do not distinguish between stimulating and blocking active antibodies but allow for quantification of the binding capacity of TSH receptor antibodies. [56,57]Bioassays measure the stimulatory or inhibitory effects of TRAb by detecting cAMP production and intracellular TSHR signals. [57,58]ELISA, which is more commonly used in research than in clinical settings, is based on the inhibition of human monoclonal TRAb (M22) binding. [59]everal studies have compared instruments from different manufacturers for TSH receptor antibody detection, aiming to identify faster, more sensitive, accurate, and cost-effective detection methods. [56,60,61]2.1.3.Autoimmunity and inflammation.GD, an organspecific autoimmune disorder, has been studied continuously, revealing novel findings related to its autoimmune response and inflammatory processes.[1,2] Janyga et al [62] identified a significant elevation in IL-23 and IL-31 levels in GD patients by measuring 15 cytokines associated with Th17 and Treg lymphocytes in the serum of different populations.IL-23, a key cytokine controlling peripheral tissue inflammation, is positively correlated with the severity of autoimmune disease symptoms, along with IL-31 and IL-33 levels.[62,63] When IL-23 is activated, certain lymphocytes, including T cells, secrete cytokines such as IL-17, IL-22, TNF-α, and IFN-γ, enhancing the immune response.[64] IL-23 induces the formation of pathogenic Th17 cells and inhibits the reactivity of Treg cells by suppressing the responsiveness of IL-33 and reducing the differentiation of Treg cells.[65,66] This indicates a crucial role for IL-23 and IL-31 in GD pathogenesis.
Another study suggests that deficiencies in regulatory T cells (Treg) and regulatory B cells, imbalances between Treg and Th17 lymphocytes, and abnormal production of anti-inflammatory cytokines significantly impact the progression of autoimmune thyroid diseases, including GD. [67] This also implies that deficiencies in minerals such as selenium, zinc, iron, copper, calcium, and magnesium may promote oxidative stress, affect immune function, and further contribute to GD. Supplementation with these minerals has been proposed as an immunomodulatory approach for the treatment of GD. [67] Further research indicates that an imbalance between Treg and Th17 cells may lead to the development of GD.Various factors associated with the differentiation of Treg and Th17 cells have been identified in various pathological states.[70] Although earlier studies have attributed the occurrence of GD to an imbalance between helper T (Th) 1 and 2 cells, recent research on the balance between Th17 and Treg cells has introduced a new perspective. [71]Investigating the differentiation of various T-cell subsets in GD (especially Th17 and Treg cells) to identify potential therapeutic targets is a promising current research focus.4.2.1.4.Graves' ophthalmopathy.Graves' ophthalmopathy is an autoimmune disease affecting the tissues behind the eyes.It predominantly occurs in patients with hyperthyroidism, with approximately 30% of cases being associated with GD.However, it is occasionally observed in patients with normal thyroid function or even hypothyroidism due to chronic thyroiditis. [27,29]The exact mechanism underlying this condition is not fully understood; however, TSH receptor antibodies are widely recognized as significant contributors. [72]esearch suggests that fibroblasts initiate 2 key pathways under the stimulation of TSH receptor antibodies and T-cell activation. [73]First, they respond to stimuli by producing glycosaminoglycans (GAGs), of which hyaluronic acid is the most prominent component.The accumulation of these hydrophilic molecules leads to the swelling of the extraocular muscles.Second, these cells further increase the volume of the extraocular muscles by promoting fat generation.Eyeball protrusion and other clinical manifestations of Graves' orbitopathy result from these processes.In addition, it has been hypothesized that the IGF-1 receptor present in fibroblasts synergistically interacts with TSH receptor antibodies to promote the production of GAGs.Although this reaction provides some insight into the pathogenic mechanism of Graves' ophthalmopathy, it is neither detailed nor comprehensive.
Currently, research on Graves' ophthalmopathy primarily focuses on its clinical management.The European Group on Graves Orbitopathy released clinical practice guidelines in 2016 and 2021, ranking among the top 10 most-cited studies in the past 2 decades.The latest guidelines provide evidence-based treatment recommendations for Graves' ophthalmopathy at different levels of severity and incorporate clinical management during viral pandemics. [29]Recent research on Graves' ophthalmopathy has emphasized clinical management and the inadequacies of basic research.The cyclic process of identifying clinical issues, addressing them through scientific research, and applying the results to clinical practice constitutes a beneficial feedback loop between clinical and basic research.This approach holds promise as a potential research method for Graves' ophthalmopathy.
Thyroid nodules and thyroid cancer in Graves'
disease.In recent years, an increasing number of institutions have published case reports on the association between GD, thyroid nodules, and thyroid cancer, indicating a growing scholarly interest in their correlation.A study involving 5025 patients with GD and 20,100 frequency-matched patients without GD demonstrated a higher risk of cancer among patients with GD.In particular, the risks of developing breast and thyroid cancers within 3 and 6 years were elevated, prompting the proposal of preventive strategies for thyroid and breast cancers. [74]Another study suggested that the occurrence of thyroid nodules and the risk of thyroid cancer in patients with GD might increase with age. [75]These conclusions are consistent with those of other studies. [76,77]However, the mechanisms underlying the development of thyroid nodules and cancer in patients with GD remain unclear.Some studies have proposed a crucial role of thyroid-stimulating immunoglobulin in this process.However, there is an ongoing debate regarding whether TRAb levels positively or negatively correlate with the incidence of thyroid cancer in patients with GD. [75,77,78] Future research should provide evidence to substantiate this correlation and unveil the underlying mechanisms to achieve early prediction, diagnosis, and treatment.4.2.1.6.Advantages and shortcomings.This study has several unique advantages over traditional reviews.First, we conducted a systematic analysis of GD using a bibliometric methodology, providing comprehensive guidance for researchers in this field.Second, our use of bibliometric tools such as CiteSpace and the R package Bibliometrix ensures a more objective data analysis.Finally, our results showcase trends and hotspots in GD research.However, this study has some limitations.First, we sourced our research data solely from the WoSCC database, which may have led to a reduction in the number of literature items that met the criteria.Second, the cutoff date for the publications included in our analysis was July 31, 2023; thus, our study may not entirely reflect the actual situation in 2023.Finally, we only included studies published in English, which might imply the omission of literature in other languages.
Conclusion
Our study represents the inaugural application of CiteSpace and the R package Bibliometrix, which enabled us to conduct a bibliometric analysis of the progress and trends in GD research.We comprehensively summarized the global research outcomes of GD from 2003 to 2023.The findings indicate a gradual increase in publications over the past 5 years, highlighting the growing trend of collaboration among various countries, institutions, and scholars.Current research focuses primarily on clinical management, TSH receptors and thyroid hormones, autoimmunity and inflammation, Graves' ophthalmopathy, thyroid nodules, and thyroid cancer.We hope that this study serves as a valuable resource for fellow scholars, offering insights into the current state of GD research and aiding in the identification of prospective research directions.
Figure 1 .
Figure 1.Flowchart of the publication screening process.
Figure 2 .
Figure 2. Annual output of publications.
Figure 7
shows the collaborative relationships among several authors, such as Antonelli Alessandro with Smith Terry and Kahaly George with Douglas Raymond.Interestingly, Antonelli Alessandro collaborated closely with the top 5 authors.
Figure 3 .
Figure 3. Geographic distribution (A) and visualization (B) of countries where GD research has been conducted.
Figure 4 .
Figure 4. Visualization of institutions on GD research.
Figure 5 .
Figure 5. Visualization of co-cited journals on GD research.
Figure 6 .
Figure 6.Dual-map overlay of journals on GD research.
Figure 7 .
Figure 7. Visualization of authors (A) and co-cited authors (B) who have performed GD research.
Figure 8 .
Figure 8. Visualization of co-cited references on GD research.
Figure 11 .
Figure 11.Visualization of the analysis of keywords (A) and keyword clusters (B) on GD research.
Table 1
Top 10 countries and institutions in GD research.
Table 2
Top 10 journals in GD research.
Table 3
Top 10 co-cited journals in GD research.
Table 4
Top 10 authors and co-cited authors in GD research.
Table 5
Top 10 co-cited references in GD research.
Table 6
Top 10 keywords in GD research.
|
2024-06-15T05:16:17.930Z
|
2024-06-14T00:00:00.000
|
{
"year": 2024,
"sha1": "a761bc1556a79a467d002ca9765c2b20357360d0",
"oa_license": "CCBY",
"oa_url": null,
"oa_status": null,
"pdf_src": "PubMedCentral",
"pdf_hash": "a761bc1556a79a467d002ca9765c2b20357360d0",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
18921730
|
pes2o/s2orc
|
v3-fos-license
|
IDS in Telecommunication Network Using PCA
Data Security has become a very serious part of any organizational information system. Internet threats have become more intelligent so it can deceive the basic security solutions such as firewalls and antivirus scanners. To enhance the overall security of the network an additional security layer such as intrusion detection system (IDS) has to be added. The anomaly detection IDS is a type of IDS that can differentiate between normal and abnormal in the data monitored. This paper proposes two types of IDS, one of them can be used as a network intrusion detection system (NIDS) with overall success (0.9161) and high detection rate (0.9288) and the other type can also be used as a host intrusion detection system (HIDS) with overall success (0.8493) and very high detection rate (0.9628) using NSL-KDD data set.
1.INTRODUCTION
In the age of information technology revolution the telecommunications networks have been developed from circuit switched network to packet switched network, after that it has Mutations enormous towards all-IP based networks. These developments make the communication of applications and services such as data and voice are being transferred on top of the IP-protocol [1].
The development of data transmission speeds in both uplink and downlink has increased considerably from the second generation (2G) of radio access networks to the third generation (3G) of radio access networks and the development of devices that subscribers of telecommunications networks make the boundary between computers and mobile phones has become unspecified.
With the smart phones, the subscriber can do almost everything and can dispense on the basic personal computers. This means that the full data on the Internet is now in the hands of each smart phone owners. Technologies in communications networks have become more progress and it has raised new unwanted possibilities. Risks and threats that were applicable only in the fixed networks are now feasible in the radio access networks. The security systems have to become more intelligent because of threats are becoming more advanced.
The basic security measurements such as firewalls and antivirus scanners cannot keep pace with the overgrowing number of intelligent attacks from the Internet. A solution to enhance the overall security of the networks is to add an additional security layer to increase the security layers by using intrusion detection systems (IDS). Intrusion Detection System (IDS) designed to complement other security measures based on attack prevention [2]. Amparo Alonso-Betanzos et al. [3] say 'The aim of the IDS is to inform the system administrator of any suspicious activities and to recommend specific actions to prevent or stop the intrusion'.
There are two types of intrusion detection, one of them is signature-based and the other is anomaly-based intrusion detection. The signature-based or misuse detection method use patterns of well-known attacks to identify intrusions [4].
The anomaly-based intrusion detection uses network traffic which has been monitored and compared versus any deviation from the established normal usage patterns to determine whether the current state of the network is anomalous. An anomalous traffic can considered as intrusion attempt.
Misuse detection uses well-defined patterns known as signatures of the attacks. Anomaly-based detection builds a normal profile and anomalous traffic detected when the deviation from the normal model reaches a preset threshold level [5].
The anomaly-based intrusion detection depends on features selection. Well selection of features will maintain accuracy of the detection while speeding up its calculations. Therefore, any reduction in the number of features used for the detection will improve the overall performance of the IDS. If there are no useless features, focus on the most important ones expected to improve the execution speed of IDS.
This increase in the detection speed will not affect accuracy of the detection in a significant way. Incorrect selection of the features may reduce the speed of the operation and reduce detection accuracy [6]. This aim of this paper is to improve the intrusion detection system by using Principal Component Analysis as a dimension reduction technique. The Paper Compares between two different features selections, i.e.6 features and 10 features. One of this features selections can be used in Network Intrusion Detection System (NIDS) and the other can be used in Host Intrusion Detection System (HIDS). Chakraborty [7] has reported that the existence of irrelevant and redundant features generally affects the performance of machine learning part of the work. Chakraborty proved that good selection of the feature set results in better classification performance.
2.RELATIVE WORK
A. H. Sung et al. [8] have demonstrated that the elimination of these unimportant and irrelevant features did not reduce the performance of the IDS.
Chebrolu et al. [9] reported that an important advantage of combining redundant and complementary classifiers is to increase accuracy and better overall generalization. Chebrolu et al. [9] have also identified important input features in building IDS that are computationally efficient and effective. This work shows the performance of three feature selection algorithms: (1) Bayesian networks, (2) Classification and Regression Trees and (3) an ensemble of Bayesian networks and Classification and Regression Trees. Sung and Mukkamala [8], have explored SVM and Neural Networks that can categorize features with respect to their importance. Use SVM and Neural Networks to detect specific kinds of attacks such as probing, DoS, Remote to Local, and User to Root. Prove that the elimination of less importance and irrelevant features has no effect on reducing the performance of the IDS.
Chebrolu et al. [9] suggested CART-BN approach, where CART has a better performance for Normal, Probe and U2R and the ensemble approach worked has a better performance for R2L and DoS. Meanwhile, A. Abraham et al. [10] proved that ensemble of Decision Tree was suitable for Normal, LGP for Probe, DoS and R2L and Fuzzy classifier was good for R2L attacks.
A. Abraham et al. [11] prove the ability of their suggested on Ensemble structure in modelling lightweight distributed IDS.
Manasi Gyanchandani et al. [12] improved the performance of C4.5 classifier over NSL-KDD dataset using different classifier combinations techniques such as bagging, boosting and stacking. Gholam Reza Zargar et al. [2] show that dimension reduction and identification of effective network features for category-based selection can reduce the processing time in an intrusion detection system while maintaining the detection accuracy within an acceptable range.
Distance
Many multivariate techniques applied to the anomaly detection problem are based upon the concept of distances. The most familiar distance metric is the Euclidean or straight-line distance. In most cases, it is used as a measure of similarity in the nearest neighbour method. Let x = (x 1 , x 2 , x 3 , …, x p ) ´ and y = (y 1 , y 2 , y 3 , …, y p ) ´ be two p-dimensional observations, the Euclidean distance between x and y is Since each feature contributes equally to the calculation of the Euclidean distance, this distance is undesirable when different features measured on different scales or the features have very different variability. The effect of the features that have high variability or large scales of measurement would control others that have less variability or smaller scales. As an alternative, a measure of variability can be incorporated into the distance metric directly. One of these metrics is the well-known Mahalanobis distance Where S is the sample covariance matrix.
Principal Component Analysis (PCA)
Naturally in intrusion detection problems Data found in high dimensions. To easily explore the data and further analysis, the dimensionality of the data must be reduced. The PCA is often used for this purpose. PCA is a predominant linear dimensionality reduction technique, and it has been widely applied to datasets in many different scientific domains [13].
PCA is concerned with explaining the variance covariance structure of a set of variables through a few new variables, which are linear combinations of the original variables. Principal components are particular linear combinations of the p random variables {x 1 , x 2 , x 3 , …, x p } with three important properties. The first one is the principal components are uncorrelated. The second one is the first principal component has the highest variance and the second principal component has the second highest variance and so on. The third one is the total variation in all the principal components combined equal to the total variation in the original variables {X 1 , X 2 , X 3, …, X p }.
The new variables with such properties are easily obtained from eigenanalysis of the covariance matrix or the correlation matrix of {X 1 , X 2 , X 3 , …, X p } [14]. Let the original data X be a n x p data matrix of n observations on each of p variables (X 1 , X 2 , …, X p ) and let R be a p x p sample correlation matrix of X 1 , X 2 , …, X p . If (λ 1 , e 1 ), (λ 2 , e 2 ), (λ 3 , e 3 ), … (λ p , e p ) are the p eigenvalue and eigenvector Pairs of the matrix R, λ ≥ λ ≥λ ≥ … ≥ λ ≥ 0, then ith sample principal component of an observation vector x= (x 1 , x 2 , x 3 , …, x p ) ʹ is y i = e′ i z y i = e i1 z 1 + e i2 z 2 + e i3 z 3 +...+ e ip z p , i =1,2,3,.., p Where e = (e , e , e ,..., e )′ is the ith eigenvector.
And Z = (z , z 2 , z 3 , …, z p ) is the vector of standardized observations defined as Where x is the sample mean of the variable x . The ith principal component has sample variance λ and the sample covariance or correlation of any pair of principal components is equal to zero. The PCA produces a set of independent variables so the total variance of a sample is the sum of all the variances accounted for by the principal components. The correlation between any two variables is Where σ is the standard deviation of x which is a sample of data. The principal components of the sample correlation matrix have the same properties as principal components from a sample covariance matrix. As all principal components are uncorrelated, the total variance in all of the principal components is The principal components produced by the covariance matrix are different from the principal components produced by the correlation matrix. Eigenvalues have larger weights because of some values are much larger than others. Since The NSL-KDD data set has many items with varying scales and ranges so the correlation matrix will use.
Applying PCA to Outlier Detection
PCA applied as an outlier detection method. In applying PCA, there are two main issues, (1) how to interpret the set of principal components and (2) how to calculate the notion of distance. First, each eigenvalue of a principal component corresponds to the relative amount of variation it encompasses. The larger the eigenvalue is the more significant its corresponding projected eigenvector should be. Therefore, the most significant principal components sorted before the least significant principal components. If a new data item projected along the upper set of the significant principal components, it is likely that the data item can be classified without projecting along all of the principal components. Second, the data sample can represent by the axes of eigenvectors of the principal components. Those axes considering a normal when the data sample is the training set of normal network connections. If any points lie outside these axes by far distance then the data connection would exhibit abnormal data connection. Outliers measured using the Mahalanobis distance are presumably network connections that are anomalous, any network connection with a distance greater than the threshold value (t) is considered an outlier. In this work, any outlier represents an attack. Consider the sample principal components of an y , y , … , y observation x where y i = e′ i z , i =1,2,... , p z = x − x , k=1, 2, 3, ..., p The sum of scores that are squares of the partial principal component is equal to the principal component score Equating to the Mahanobolis distance of the observation X from the mean of the normal sample data set [15].
Anomaly detections Needs an offline training or learning phase whether those methods are outlier detection, statistical models, or association rule mining. PCA has two clearly separate phases (the offline and online detection phases). These two separate phases are an advantage for hardware implementation. Another advantage of PCA is reduction of features. As we will show in our experiment, PCA effectively reduces the number of processed features from 41 to 10 or 6 features.
The outline steps involved in PCA are shown in (figure 1). Training data take as input and a mean vector of each sample calculate in the offline phase. Ideally, these data sets are a snapshot of activity connections in a real network environment. In addition, these data sets should contain only normal connections. Second, correlation matrixes calculate from the training data.
A correlation matrix normalizes all of the data by calculating the standard deviation. Next, eigenanalysis performed on the correlation matrix to create independent orthonormal eigenvalue and eigenvector pairs. The set of principal components can use in online analysis because of these pairs. Finally, the sets of principal components sort by eigenvalue in descending order. The eigenvalue is a relative measure of the variance of its corresponding eigenvectors.
Using dimensionality-reducing method such as PCA to extract the most significant principal components, so only a subset of the most important principal components needs to classify any new data. In addition to using the most significant principal components (q) to find intrusions, we have found that it is helpful to look for intrusions along a number of least-significant components (r) as well.
then (q) is a subset of the highest values and is a subset of the smallest components. The MajC threshold is referred (t ) while the MinC threshold is referred to (t ). An observation (x) is an attack if
The online portion takes major principal components and minor principal components and maps online data into the eigenspace of those principal components
Data Set Description
Mostly all the experiments on intrusion detection are done on KDDCUP '99 dataset, which is a subset of the 1998 DARPA Intrusion Detection Evaluation data set and is processed extracting 41 features from the raw data of DARPA 98 data set. Defined higher level features that help in differentiating between "good" normal connections from "bad" attacks connections [16]. KDDCUP 99 data set can be used in host-based systems, network-based systems, signature systems and anomaly detection systems.
A connection is a sequence of Transmission Control Protocol (TCP) packets starting and ending with the time between which data come from a source IP address to a target IP address under some protocol. Each connection described as a normal or as an attack with defined the attack type. Each connection record consists of about 100 bytes [17].
KDD train and test set contains a huge number of records and huge number of redundant records. Almost about 78% and 75% of the records duplicated in the train and test set respectively. The classification will be wrong because of these redundant records and thus these records prevent classifying the other records that is not redundant. To solve this problem, a new dataset was developed NSL-KDD [18]. One copy of each repeated record was not removed in the KDD train and test set.
Performance Measures
Metrics, which are mainly used to evaluate the performance of classifiers are presented in [19], [20] and are given here for ready reference.
Experiment steps and results
In our experiments we use KDDTrain_20Percent [21] in both the training and testing stages. First, we select 6 features from the basic features of TCP connections which used with NIDS because these features do not need any host logs. Second, we add 4 features from traffic features, which based on time window and this collection (10 features), used in HIDS is shown in (Table 1).
Sev-count
Number of connections to the same service as the current connection in the past two seconds from the same source
Continuous (3)
Dst-host-count Number of connections to the same host as the current connection in the past two seconds Continuous (3) Dst-host-srv-count Number of connections to the same service as the current connection in the past two seconds to the same host
Continuous (3)
We used a Matlab program to design our IDS. Based on [22], we suggest using (q) major components that can explain about 50 -70 percents of the total variation in the standardized features. When the original features are uncorrelated, each principal component from the correlation matrix has an eigenvalue equal to 1. So the minor components are those components whose variances or eigenvalues are less than 0.20, which would indicate some relationships among the features (r).
First step we selected 6 features and suggested using q = 3, r =0. Second step we added 4 features and suggested using q= 3, r =2. In a multiclass prediction, the result on a test set is often displayed as a two dimensional confusion matrix with a row and a column for each class. Each matrix element shows the number of test examples for which the actual class is the row and the predicted class is the column. Good results correspond to large numbers down the main diagonal and small, ideally zero, off-diagonal elements. The confusion Matrix is showed on the ( Table 2). The Performance Measures are shown in (Table 3) and (Table 4). Step (1) Step ( Both recall and precision have good value in these two steps but one of steps can be used as NIDS another can be used as HIDS which has a better detection rate.
CONCLUSION AND FUTURE WORK
Future network intrusion detection system generation will most likely employ both signature detection and anomaly detection modules. Anomaly detection methods process a large amount of data in order to recognize anomalous behaviour or new attacks. This paper used PCA as an effective way of outlier analysis. PCA is particularly useful because of its ability to reduce data dimensionality into a smaller set of independent variables from which new data can be classified.
This paper has two steps in its experiment. The first step takes six features from the basic features of TCP connections that can used in NIDS and this step has an overall success rate (0.9161) with high detection rate (0.9288). The second step takes ten features {six features from the basic features of TCP connections plus four features from traffic features} which can be used in HIDS and this step has an overall success rate (0.8493) with very high detection rate for Anomaly class (0.9628).
Plan for the future work is to use these two steps to make an integrated intrusion detection system by using relationship between these two steps.
|
2013-08-13T01:10:06.000Z
|
2013-08-01T00:00:00.000
|
{
"year": 2013,
"sha1": "2da32f1ce35caebe880e0be4a643a3e58d2e03cc",
"oa_license": null,
"oa_url": "https://doi.org/10.5121/ijcnc.2013.5412",
"oa_status": "BRONZE",
"pdf_src": "Arxiv",
"pdf_hash": "2da32f1ce35caebe880e0be4a643a3e58d2e03cc",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science"
]
}
|
222093905
|
pes2o/s2orc
|
v3-fos-license
|
LncRNA MALAT1 Affects Mycoplasma pneumoniae Pneumonia via NF-κB Regulation
Our aim was to determine whether the long non-coding RNA (lncRNA) metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) is involved in Mycoplasma pneumoniae pneumonia (MPP), and its possible mechanism of action. MALAT1 expression in the bronchoalveolar lavage fluid of 50 hospitalized children with MPP was compared to its expression in 30 children with intrabronchial foreign bodies. MALAT1 expression was higher in children with MPP, accompanied by increased inflammatory mediators interleukin 8 (IL-8) and tumor necrosis factor alpha (TNF-α), compared to the controls. In human airway epithelial cells infected with wild-type Mycoplasma pneumoniae (strain M129), MALAT1, IL-8, and TNF-α expression significantly increased, and increased expression of IL-8 and TNF-α could be suppressed by MALAT1 knockdown. Luciferase reporter gene assay and western blot showed that knockdown of MALAT1 reduced nuclear factor-κB (NF-κB) activation. In vivo, RNAi packaged with adenovirus (Adv) was nasally transfected into BALB/c mice to silence MALAT1, and an MP-infected mouse pneumonia model was prepared. The results demonstrated that the degree of pulmonary inflammatory injury, vascular permeability, secretion of inflammatory factors, and expression of phosphorylated p65 (pp65) in MP-infected mice were partly reversed after MALAT1 knockdown compared to those in the controls. In conclusion, MALAT1 is involved in the regulation of airway and pulmonary inflammation caused by MP infection via NF-κB regulation.
INTRODUCTION
Mycoplasma pneumoniae (MP) is one of the main causes of community-acquired pneumonia in children (Jain et al., 2015). Pneumonia caused by MP often leads to serious complications that affect the quality of life in children (Zhang et al., 2016). In patients with Mycoplasma pneumoniae pneumonia (MPP), excessive inflammatory response is a leading cause of pulmonary inflammation and is predictive of poor prognosis. However, the mechanism that leads to the uncontrolled inflammatory response in MPP is still unclear, and further investigation is necessary.
Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), one of the first lncRNAs associated with human disease to be identified, is of interest because it is widely evolutionarily conserved. MALAT1 was originally identified as a prognostic marker for poor clinical prognosis in patients with early non-small cell lung cancer (NSCLC) (Ji et al., 2003). After that discovery, the function of MALAT1 was studied in various cancers for its role in cancer development and progression (Sun and Ma, 2019). Recent findings suggest that MALAT1 also plays an important regulatory role in inflammatory diseases. For example, MALAT1 was reported to be involved in the regulation of lipopolysaccharide (LPS)induced acute kidney injury (Ding et al., 2018) and lung injury (Liang et al., 2019). MALAT1 is also an inflammatory regulator in human systemic lupus erythematosus (Yang H. et al., 2017). Downregulation of MALAT1 alleviates saturated fatty acid-induced myocardial inflammatory injury (Jia et al., 2019), while silencing of MALAT1 ameliorated inflammatory injury after lung transplant ischemia-reperfusion (Wei et al., 2019). Therefore, we aimed to determine whether MALAT1 also plays a regulatory role in the inflammatory response in MPP.
Previous reports have shown that MALAT1 regulates the nuclear factor-κB (NF-κB) signaling pathway to modulate the inflammatory response (Lei et al., 2019). MALAT1 affects NF-κB via direct binding or indirect modulation, thus altering the transcription and activation of its downstream inflammatory factors and resulting in varied inflammatory responses (Dai et al., 2018;Ding et al., 2018). Cytokine production in airway epithelial cells can lead to inflammation associated with respiratory diseases. Interleukin-8 (IL-8), a potent chemoattractant and activator of neutrophils, is associated with the initiation and maintenance of inflammation (Pease and Sabroe, 2002). In humans, the airway epithelium is the main source of IL-8, and the pathogenesis of pneumonia is positively correlated with IL-8 expression (Martin et al., 1997). Tumor necrosis factor alpha (TNF-α) is an important bioactive inflammatory mediator that plays a pivotal role in inflammation and immune regulation. High levels of TNF-α can damage capillary endothelial cells, promote microthrombosis, and lead to ischemic necrosis; thus, TNF-α is related to the severity of pneumonia (Salvatore et al., 2007). Therefore, we conducted experiments in vitro and in vivo to investigate the role of MALAT1 in MPP.
Study Population
Fifty children who were diagnosed with MPP and underwent bronchoscopy while hospitalized at Children's Hospital of Nanjing Medical University were included in the study. Thirty children with intrabronchial foreign bodies (FB) were included as the control group. MP infection was diagnosed by polymerase chain reaction (PCR) testing of MP in nasopharyngeal secretions and/or serological testing of MP immunoglobulin M (IgM) by enzyme-linked immunosorbent assay (ELISA). Patients with chronic disease, heart disease, immune deficiency disease, or immunosuppressive drugs were excluded. All enrolled children tested negative for respiratory syncytial virus, influenza virus, parainfluenza virus, pneumovirus, adenovirus, chlamydia trachomatis, and bacterial culture. The study protocol was approved by the Ethics Committee of the Children's Hospital of Nanjing Medical University (approval number: 201801126-1), and the informed written consent of at least one parent/guardian of each patient was obtained.
Bronchoalveolar Lavage (BAL)
Guidelines for bronchoscopy and alveolar lavage were reviewed previously Gao et al., 2018). Children with MPP underwent bronchoscopy and bronchoalveolar lavage fluid (BALF) collection within 1 week after admission. Children with FB were immediately treated with bronchoscopy to remove the foreign bodies, and BALF specimens were collected during reexamination. Irrigation was performed with sterile normal saline (0.3-0.5 mL/kg). Within 1 h after BALF collection, the supernatant was separated, divided, and stored at −20 • C. RNA was isolated with TRIzol reagent (Invitrogen, Carlsbad, United States) and stored at −80 • C.
Mycoplasma Culture, Preparation, and Infection
Mycoplasma pneumoniae (strain M129) was provided by Professor Chen Z. M. (Children's Hospital, Zhejiang province), as described in our previous study (Shi et al., 2019). The strain was cultured in a Mycoplasma broth, which consists of Mycoplasma broth base (Oxoid, Basingstoke, United Kingdom), 0.5% glucose, 0.002% phenol red, and Mycoplasma selective supplement G (Oxoid). MP was quantified by counting the number of colonyforming units (CFU) in Mycoplasma agar plates.
siRNA and Cell Transfection
SiRNA against MALAT1 and scrambled siRNA were designed and synthesized by RiboBio biotech (RiboBio, Guangzhou, China). Vector transfection was performed using the ribo FECT TM CP transfection kit (RiboBio, Guangzhou, China) in accordance with the manufacturer's protocol. Sequences of custom siRNA are listed in Table 1.
Laboratory Animals
Forty male specific-pathogen-free (SPF) BALB/c mice aged 6-8 weeks (body weight: 20-25 g) were purchased from the Laboratory Animals Center of Nanjing Medical University. The mice were housed in conditions with temperatures between 21 and 25 • C and relative humidity between 55 and 65%, with free access to food and water. This study was approved by and
Adenovirus (Adv)-RNAi Transfection and Subsequent Establishment of Mouse Model of MP Infection
Fluorescent (GFP)-labeled adenovirus packaged with shorthairpin RNA (shRNA) against MALAT1 (Adv-MALAT1-RNAi) and scrambled shRNA (Adv-shNC) were synthesized by GeneChem (Shanghai, China). Mice were randomly divided into four groups (n = 10 per group): shNC + PBS; sh-MALAT1 + PBS; shNC + MP; and shMALAT1 + MP. To administer the virus, 4% chloral hydrate (0.1 mL/10 g body weight) was given by intraperitoneal to lightly anesthetize the mice, and 30 µL of virus containing 5 × 10 8 PFU of Adv-shMALAT1-RNAi was inhaled into the lungs through the nose of each mouse. Two doses of this virus were administered to each mouse in 24 h intervals. Mice in the shNC group were treated with control recombinant adenovirus containing scrambled shRNA. After 3 days, mice in the MP group were intranasally administered 10 8 CFU of M129 in a 50 µL solution, and mice in the PBS group were administered 50 µL of PBS. During administration, their heads were kept tilted at an angle of 30 • -45 • to ensure fluid flow through the lower respiratory tract. After 4 days, the mice were euthanized, and test samples were collected. Sequences for shRNA are listed in Table 2.
Biochemical Analysis of Oxidative Stress in Lung Homogenate Samples
After the mice were euthanized, left lung tissue samples were collected, and 10% homogenate was prepared by addition of PBS and centrifugation at 12,000 × g at 4 • C for 15 min. The supernatant was stored at −80 • C for analysis. The concentrations of superoxide dismutase (SOD) and malondialdehyde (MDA) were determined by visible spectrophotometry (Jiancheng Bioengineering Institute, Nanjing, China).
Collection of Blood and BALF Samples
After euthanasia, 1 mL of blood from each mouse was collected from the eye socket in 1.5 mL Eppendorf tubes. The trachea of each mouse was carefully separated, and 1 mL of normal saline was used to wash and collect the BALF samples three times. Both the blood and BALF samples were centrifuged at 3,000 × g at 4 • C for 5 min, and the supernatant was stored at −80 • C for analysis.
Hematoxylin-Eosin (HE) Staining
After mice were euthanized using a chloral hydrate injection, specimens of the lower lobe of the right lung were fixed in 4% formaldehyde solution, followed by dehydration, paraffin embedding, sectioning, and HE staining for histopathological analysis.
Quantitative Real-Time PCR (qRT-PCR)
Total RNA was extracted from cells or mouse lung tissue using TRIzol reagent according to manufacturer's instructions (Invitrogen, New York, United States
Western Blot
Nuclear protein was extracted from cells and mouse lung tissue using a Nuclear Cytoplasm Extraction Kit (KeyGen Biotech, Mouse IL-8-R 5 -CTACTACACAGGGATCAGGGC-3 Mouse TNF-α-F 5 -GTGCCAGCCGATGGGTTGTAC-3 Mouse TNF-α-R 5 -TGACGGCAGAGAGGAGGTTGAC-3 Mouse GAPDH-F 5 -GGTTGTCTCCTGCGACTTCA-3 Mouse GAPDH-R 5 -TGGTCCAGGGTTTCTTACTCC-3 Frontiers in Cell and Developmental Biology | www.frontiersin.org Nanjing, China). Aliquots of 30 µg of protein lysate were separated using SDS polyacrylamide gel electrophoresis and transferred to PVDF membranes (Millipore, Bedford, MA). After blocking with 5% skim milk at room temperature for 1 h the membrane was incubated overnight with primary antibody at 4 • C, and then incubated with secondary antibody at room temperature for 1 h. Finally, ECL reagent (Beyotime, Nanjing, China) was used for chemiluminescent detection. Primary antibodies used for immunoblot analysis included phosphorylated p65 (Cell Signaling Technology, Danvers, United States) and histone H3 (Proteintech, Rosemont, United States).
Luciferase Reporter Assays
NF-κB transcriptional activity reporter plasmid pNFκBluc (Beyotime Institute of Biotechnology) and Renilla luciferase expression vector pRL-TK (Promega, Madison, United States) were co-transfected into normal or MALAT1depleted A549 and BEAS-2B cells using Lipofectamine 2000 (Invitrogen, Carlsbad, United States). After 24 h, cells were treated with MP and harvested 24 h later. Luciferase activity was measured using the dual-luciferase reporter assay system (Promega, Madison, United States) in accordance with the manufacturer's protocol. Relative firefly luciferase activity was normalized to Renilla luciferase activity.
Statistical Analysis
Each experiment was performed in triplicate. All statistical analyses were conducted by SPSS software, version 19.0 (IBM, Armonk, United States). One-way ANOVA, Student's t-test, or Mann-Whitney test were used for statistical comparisons as indicated. P < 0.05 was defined as statistically significant.
Analysis of MALAT1 Expression in BALF From Children
First, qRT-PCR was used to detect MALAT1 expression in BALF from children. Expression of MALAT1 in BALF from children with MPP was significantly increased compared to the control group ( Figure 1A, P < 0.01). In addition, ELISA results suggested that the level of inflammatory factors IL-8 and TNF-α in BALF of the MP group was significantly higher than that of the control group (Figures 1B,C, P < 0.001). Further, correlation analysis showed that expression of MALAT1 was positively correlated with expression of IL-8 and TNF-α (Figures 1D,E, P < 0.001).
Knockdown of MALAT1 Inhibited the Increased Secretion of Inflammatory Cytokines by Airway Epithelial Cells After MP Infection
To investigate the role of MALAT1 in MP infection-induced inflammation, siRNA was used to knockdown MALAT1 in A549
Knockdown of MALAT1 Inhibited NF-κB Activation Induced by MP Infection in Airway Epithelial Cells
When NF-κB is activated, the subunit of NF-κB, p65, is phosphorylated and translocated to the nucleus. Therefore, we detected pp65 in the nucleus of A549 and BEAS-2B cells by western blot, and the results showed that MALAT1 knockdown reduced pp65 expression in the nucleus of MP infected cells (Figures 3A,B, P < 0.05). This is in accordance with the results of the dual luciferase reporter assay (Figures 3C,D, P < 0.01). These results suggest that MALAT1 knockdown inhibited NF-κB activation induced by MP infection in airway epithelial cells.
MALAT1 Knockdown Reduced Histopathological Damage Caused by MP Infection in BALB/c Mice
Lung tissue samples were collected from each group of mice, and MALAT1 expression was detected using qRT-PCR. The expression of MALAT1 was upregulated in the shNC + MP group and downregulated in the shMALAT1 + PBS group, compared to the shNC + PBS group. In addition, the expression of MALAT1 was decreased in the shMALAT1 + MP group compared to the shNC + MP group (Figure 4A, P < 0.001). As shown in Figure 4B, histopathological changes were evaluated by HE staining of the lung tissue samples. The shNC + MP group showed significant histological changes, including cell infiltration, alveolar wall thickening, significantly increased secretion, and even structural collapse. These changes were reversed upon MALAT1 knockdown in the infected group.
MALAT1 Knockdown Reduced Lung Injury Caused by MP Infection in BALB/c Mice
After MP infection, the shNC + MP group exhibited increases in protein concentration in lung homogenate and ratio of protein concentration in lung homogenate and serum. In terms of these two indexes, the shNC + MP group exhibited a greater increase than the shMALAT1 + MP group (Figures 5A,B, P < 0.05). The level of SOD activity indirectly reflects the body's ability to remove oxygen free radicals, and the level of MDA indirectly reflects the severity of free radical attack on cells. The results showed that compared with the shNC + MP group, the SOD activity of mice in the shMALAT1 + MP group significantly increased, while the MDA concentration decreased (Figures 5C,D, P < 0.05).
MALAT1 Knockdown Reduced the Expression of Inflammatory Cytokines in Mice With MP Infection
The mRNA expression of IL-8 and TNF-α in mouse lung tissue was detected using qRT-PCR, and the protein expression of IL-8 and TNF-α in mouse lung homogenate and BALF was detected using ELISA. The results showed that both the mRNA and protein expression of IL-8 and TNF-α in the shNC + MP group significantly increased compared to the shNC + PBS group, and knockdown of MALAT1 reversed the increase (Figure 6, all P < 0.05).
NF-κB Is Involved in MALAT1-Mediated Regulation of the Inflammatory Response Induced by MP Infection
Using western blot, we detected phosphorylated p65 expression in the lung tissue of each group. As shown in Figure 7, expression of phosphorylated p65 in the shNC + MP group was significantly increased compared to the shNC + PBS group; however, the above change was reversed after MALAT1 knockdown. These results suggested that the NF-κB signaling pathway is involved in MALAT1-mediated regulation of the inflammatory response induced by MP infection.
DISCUSSION
MP infection may cause severe local inflammation, leading to cytotoxicity and tissue damage. For this reason, patients with MPP are prone to atelectasis and necrotizing lesions and often require fiberoptic bronchoscopy. According to previous studies, the local inflammatory response is initiated by the release of proinflammatory cytokines from epithelial cells (Yang et al., 2002). Among the induced cytokines, IL-8 and TNF-α play pivotal roles in the inflammatory response of MPP. Our data showed that the levels of IL-8 and TNF-α were significantly elevated in BALF from children with MPP, supernatants of MP-infected epithelial cells, and lung tissues and BALF of mice infected with MP compared to the non-infected groups. These results confirmed the role of IL-8 and TNF-α in airway inflammation caused by MPP. MALAT1 has been reported to be involved in multiple diseases. Recently, researchers have focused on its role in inflammatory diseases. Based on the findings in this study that MALAT1 was highly expressed and expression was positively correlated with inflammatory factors IL-8 and TNF−α in BALF from children with MPP, we performed in vitro and in vivo experiments to determine the role of this lncRNA in MPinduced inflammation. Our in vitro results showed that MALAT1 knockdown inhibited the secretion of inflammatory factors from epithelial cells after MP stimulation. In vivo, after MP infection, the expression of IL-8 and TNF-α in lung homogenate and alveolar lavage fluid increased. SOD activity in lung homogenate decreased, while MDA concentration increased, indicating the existence of anti-oxidation imbalance. The ratio of lung homogenate protein concentration to serum total protein concentration was significantly increased, indicating the presence of pulmonary permeability damage, which is consistent with the previous study (Shi et al., 2019). However, MALAT1 knockdown in BALB/c mice alleviated lung inflammation and lung injury induced by MP infection. These results demonstrated for the first time the regulatory role of MALAT1 in MPinduced inflammation.
Previous studies have shown that the regulatory role of MALAT1 in inflammation is complex. For instance, a recent study reported that MALAT1 knockdown markedly reduced lung injury induced by sepsis (Lin et al., 2019). Furthermore, knockdown of MALAT1 was reported to suppress inflammatory response in LPS-induced acute lung injury and kidney injury (Dai et al., 2018;Ding et al., 2018;Cui et al., 2019;Liang et al., 2019). Knockdown of Malat1 in mice alleviated inflammatory injury after cerebral ischemia, and overexpression of Malat1 aggravated ischemic brain inflammation (Cao et al., 2019). In addition, MALAT1 has been reported to promote inflammation in septic heart damage (Chen et al., 2018), systemic lupus erythematosus (Yang H. et al., 2017), and hyperglycemia-induced inflammatory process (Puthanveetil et al., 2015). The above findings together with the results from the current study suggest that MALAT1 plays a pro-inflammatory role in many diseases. However, other studies demonstrated that lncRNA MALAT1 performs a protective function against inflammation in several diseases (Zhao et al., 2016;Cremer et al., 2019;Zhu and Men, 2019). These inconsistencies may be due to the complex mechanisms of MALAT1 regulation in different diseases.
With respect to the mechanism of MALAT1, our study indicated that the pro-inflammatory function of MALAT1 at least in part depends on NF-κB. NF-κB is involved in the inflammatory response of MPP. The lipoprotein components of MP recognize toll-like receptors and activate NF-κB (Shimizu et al., 2007(Shimizu et al., , 2008. In our present study, we detected NF-κB activation in MP-infected epithelial cells and mouse lung tissue and confirmed that MP infection activates NF-κB. According to previous studies, MALAT1 can regulate inflammation by modulating the activation of NF-κB. MALAT1 can bind directly to the subunit of NF-κB (Zhao et al., 2016;Zhu and Men, 2019) or act as competing endogenous RNA (ceRNA) and compete with miRNA functions toward NF-κB (Ding et al., 2018;Cao et al., 2019;Jia et al., 2019;Li et al., 2019;Liang et al., 2019). Our data suggested that MALAT1 knockdown inhibits phosphorylation of p65 caused by MP infection in airway epithelial cells and lung tissue of mice. Therefore, the regulatory role of MALAT1 in inflammation of MPP is associated with NF-κB activation. Although the precise mechanism remains unclear, acting as a certain ceRNA might be the way by which MALAT1 facilitate p65 phosphorylation in MP infection. Further study is needed to reveal the exact regulatory mechanism.
CONCLUSION
In conclusion, our present study demonstrates that lncRNA MALAT1 plays a key regulatory role in MPinduced inflammation. The regulatory function of MALAT1 likely proceeds through NF-κB signaling. By performing in vivo experiments, we found that downregulation of MALAT1 reduces pulmonary inflammation caused by MP infection. Our results suggest that MALAT1 may be a new therapeutic target for MPP.
DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article.
ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Ethics Committee of the Children's Hospital of Nanjing Medical University. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin. The animal study was reviewed and approved by the Institutional Animal Care and Use Committee, Nanjing Medical University.
AUTHOR CONTRIBUTIONS
HG performed the experiments and statistical analysis, contributed to the interpretation of data, made the figures and tables, and drafted the manuscript. YiZ participated in study design, collected and interpreted the clinical information, and determined the clinical status for each children involved in the study. YaZ, TH, and SZ participated in animal experiments and data interpretation. DZ and FL designed the study, analyzed the data, and revised the manuscript. All authors contributed to the article and approved the submitted version.
|
2020-10-02T13:05:42.506Z
|
2020-10-02T00:00:00.000
|
{
"year": 2020,
"sha1": "de784477e667d72bd835754417907736f35044c4",
"oa_license": "CCBY",
"oa_url": "https://www.frontiersin.org/articles/10.3389/fcell.2020.563693/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "de784477e667d72bd835754417907736f35044c4",
"s2fieldsofstudy": [
"Medicine",
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
248726927
|
pes2o/s2orc
|
v3-fos-license
|
A SYSTEMATIC REVIEW OF INTERVENTIONS USED TO REDUCE UNIVERSITY STUDENTS’ SPEAKING ANXIETY
ABSTRACT
INTRODUCTION
Mastering spoken English has become a crucial requirement in educational settings, in the job market, and in career development.The speaking skill is defined as "a productive aural/oral process that consists of using grammatical rules, cohesive devices, lexical items, phonological rules for expressing one's thoughts and feelings in speech" (Hammad & Ghali, 2015).Many learners consider conversing in English as a challenging and daunting task (MacIntyre, 2017).This is associated with several factors that may hinder and impede S/FL learning, including anxiety and low motivation.Anxiety has been acknowledged as the most widely researched emotion in S/FL learning.Compelling evidence demonstrates that speaking is the most anxiety-evoking aspect in S/FL learning (Gosiewska-Turek, 2018;Huashan, 2019;Kasap, 2019;Lu, Lee, & Lin, 2019;Milan, 2019;Saputra, 2018;Wilang & Vo, 2018).Traditionally, SA refers to the fear of oral language usage (Balemir, 2009 cited in, Miskam and Saidalvi classification involves using direct strategies (memory, cognitive, and compensation strategies) and indirect strategies (metacognitive, affective, and social strategies).For example, Akkakoson's (2016) study with 88 Thai students identified a wide range of strategies applied to cope with SA, of which social strategies were the most frequently used, followed by metacognitive, compensatory, cognitive, and memory-related strategies, respectively.
More recently, examining the effectiveness of various educational interventions on SA has sparked increased scholarly attention.In plain words, many studies have designed and applied various SAI and evaluated their effectiveness on S/FL SA (Altunkaya, 2018;Dincer, OzÇelİk, Zülfünaz, & BahÇecİk, 2020;El Shazly, 2021;Pontillas, 2020;Zacarin, Borloti, & Haydu, 2019).Two methods have been observed about the SAI mechanism in dealing with SA, including intervening directly or indirectly.For instance, facing learners' SA directly through the employment of virtual reality-oriented interventions has been found to lower undergraduate learners' SA (Stupar-Rutenfrans, Ketelaars, & van Gisbergen, 2017).Improving and substituting negative emotions with more positive ones can help learners to boost positive emotions while lowering SA.For example, the emotional freedom technique (EFT), as employed in Dincer et al. (2020) study, can help decrease learners' SA.
Another method is intervening indirectly to mitigate SA, as seen in Pontillas (2020) study, which employed a feedback-oriented intervention that developed learners' speaking proficiency in addition to alleviating their SA.
Also, developing learners' speaking skills through face-to-face or technological devices assistance by using various pedagogical means to combat SA indirectly has been noticed.For example, whether output-based or input-based, the application of strategic planning contributed to reducing Korean learners' SA (Lee & Kim, 2018).Although many educational interventions have been developed and employed and their impact on SA reduction were evaluated, their various features make it challenging to synthesize the results and implement them in practice.
Fortunately, one systematic review has reviewed foreign language anxiety reduction intervention regardless of learners' educational level (school or university) or learning skill evoked by anxiety (speaking, writing, reading, and listening).While this has offered useful indications concerning foreign language anxiety interventions (FLAI); however, its broad focus can make it complex and challenging to digest, plus it mainly included experimental studies, which can eliminate valuable information regarding language anxiety reduction interventions.In addition, Toyama and Yamazaki (2021) limited their methodological focus to include mainly experimental or quasiexperimental studies.This led to neglecting studies that adhered to other methodological designs that might arguably be of great significance in reducing foreign language anxiety.Thus, the current study offers a rigorous and comprehensive analysis of interventions used to lower the undergraduate learners' SA regardless of their methodological designs.This will provide educators, scholars, instructors, and learners with a vast array of valuable evidence to intervene and minimize the harmful effects of SA and optimize the learning experience.
METHODOLOGY
A systematic review is a literature review technique that includes "clearly formulated questions" which "uses systematic and explicit methods to identify, select and critically appraise relevant research, and to collect and analyze data from the studies that are included in the review" (Cochrane Collaboration, 2003, as cited in Siddaway, Wood, andHedges (2019)).The method adopted in this review adheres to the guidelines offered by the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) (Moher, Liberati, Tetzlaff, Altman, & Group, 2009) framework to locate and identify relevant papers for the current review.In addition, for synthesis purposes, this review utilized a narrative approach which "can be used to synthesize both quantitative and qualitative studies".Aligned with the purpose of the study, a narrative approach was adopted to allow text-based synthesis instead of a statistical summary to transparently and systematically review the collected studies that utilized different methodologies (qualitative, quantitative, and mixed methods).Further, when a review includes experimental and quasi-experimental studies that are not sufficiently similar, a narrative approach can be employed to facilitate meta-analysis (Mays, Pope, & Popay, 2005).Moreover, in developing the current systematic review process with the narrative approach, researchers primarily depended on four studies on systematic reviews (Jiménez, Ramos, González-Moraleda, & Resurreccion, 2020;Pluye & Hong, 2014;Siddaway et al., 2019;Toyama & Yamazaki, 2021) of which valuable insights were gained.
Literature Search
Three online databases, including Scopus, Web of Science (WOS), and Science Direct, were systematically searched to identify relevant studies to F/SL SAI.Key search terms were formed by employing single words and nested clauses combined with Boolean operators (AND, OR).These included ("speaking anxiety") AND ("treatment OR therapy OR intervention OR reduction OR reducing OR decrease OR decreasing OR alleviate OR lowering"), ("communication apprehension") AND ("treatment OR therapy OR intervention OR reduction OR reducing OR decrease OR decreasing OR alleviate or lowering"), ("fear of speaking") AND ("treatment OR therapy OR intervention OR reduction OR reducing OR decrease OR decreasing OR alleviate OR lowering").Various forms of the same term (e.g., decrease, decreasing) were employed to capture more papers.This review was embarked on in mid-June 2021.The first author performed the search, and the second author supervised the systematic review process.The search yielded a total of 2280 studies (Scopus = 1669, Web of Science = 143, Science Direct = 331), of which 541 were excluded due to duplication, as exhibited in Figure 1.After analyzing the remaining 1739 potential related studies' titles and abstracts, further 1683 were excluded.Consequently, 60 studies were eligible for full-text screening and analysis against the inclusion and exclusion criteria as stated in Table 1: To this point, because of not fully addressing the inclusion criteria, 44 studies were eliminated from the review.
These studies were excluded because their participants were not undergraduates (n = 11), not all undergraduates (n = 6), not clear if they are undergraduates (n = 1), pilot study (n= 1), do not primarily focus on SA (n = 4), merely explain SAI (n = 4), interventions not used to decrease SA (n = 1), addressed other types of anxiety (n = 14), or full text not available (n = 2).Accordingly, 16 papers were included in this review and progressed to the data extraction process.
It is worth mentioning that reference harvesting for 16 studies was carried out to identify other relevant papers; however, no additional related studies were found.
Data Extraction
Each of the 16 studies identified was read at least twice and then data was documented and imported into a
SYNTHESIS OF RESULTS: STRENGTH AND WEAKNESSES
This review aimed to identify and systematically review published SAI studies between January 2015 and mid-June 2021 in an undergraduate context around the globe.Our search identified 16 studies that involved using interventions and investigating their effects on learners' SA.A descriptive analysis was conducted before analyzing key features of SAI to summarize trends of SAI interventions in the literature.
Although we limited our search to start with articles published in 2015, the earliest SAI study identified in this review was published in June 2016, and the final one in January 2021.An increase in SAI publications over that period was noted, except for the first half of 2021.Our quality assessment identified a total of 16 studies which comprised one qualitative, 11 either experimental or quasi-experimental, and four mixed methods.Studies were conducted in nine countries: Turkey (n = 5), the Netherlands (n = 1), Korea (n = 2) Taiwan (n= 1), Malaysia (n = 2), the USA (n = 2), Brazil (n = 1), the Philippine (n= 1), Egypt (n = 1); further, over half of the studies (n = 10) were performed in east or west Asia.
Regardless of their methodological designs, all studies aimed to investigate the effect of their interventions on undergraduate students' SA.Regarding intervention duration, the longest lasted for three semesters (Lee & Kim, 2018), while a one-session intervention was the shortest (Dincer et al., 2020;Wang et al., 2020).Less than half of the interventions took between one to ten weeks to be conducted (n = 7), 37% were performed over a period of one semester/course to three semesters (n = 6), whereas three studies either did not specify the duration of its intervention (Hashim et al., 2019;Nazligul et al., 2019) or merely reported the number of sessions (12 sessions) conducted (Zacarin et al., 2019).
All of the SAI studies included in this review had a sample size of less than 100, except for two studies: LeFebvre et al. ( 2020) and Lee and Kim (2018), which involved 559 and 168 participants respectively.Using software (G*Power) to calculate sample size as recommended by Dincer et al. (2020) and Kim et al. (2018), while sample representativeness was discussed in accordance with El Shazly's (2021) study.In addition, 37% of the studies reported their participants' numbers without mentioning their gender (n = 6).About 70% of the studies included participants aged 18 to 40, exempting five studies that did not mention their participants' age range.
Ignoring reporting participants' age range and gender can be, to some extent, tolerated.However, insufficient explanation concerning sampling techniques used to recruit the participants and their representativeness of the target population can seriously impact the findings' generalizability and transformability to similar populations or contexts.
Even though only 16 studies were included in this review, it can be noticed that a large number of scales have been reported above.This is because several studies used more than one scale in their data collection.Besides using well-known SA measurements, mixed-method studies (n = 4) developed various other tools.This comprised a Client Satisfaction Questionnaire (CSQ), recording sheet, interview script (Zacarin et al., 2019), observation, impromptu speech rubric, and reflective journal (Pontillas, 2020), and four items used experience evaluation and three open-ended questions (Wang et al., 2020).Meanwhile, Hashim et al.'s (2019) qualitative study used students' opinions and reflections to explore the effectiveness of their intervention on learners' SA. 2010) assertiveness (Kim et al., 2018) (30 item schedule designed by Rathus (1973) Rathus (1973), and satisfaction with simulation practical education (Kim et al., 2018) (Satisfaction with Simulation Experience Scale (SSES) devised by Levett-Jones et al. (2011).Moreover, one study (Kim et al., 2018) integrated two instruments to measure nursing clinical self-efficiency, which are the adjusted form of Nursing Clinical Self-Efficiency Scale (MNCSES) (the original scale is constructed by Rathus (1973)), and self-efficiency instrument established by Parker (1993).
Having employed various tools to examine the effectiveness of the SAI, all studies reported using at least one validation method, except for two studies that did not clearly present their validation techniques (Han & Keskin, 2016;Hashim et al., 2019).Considering analytic methods of the SAI studies, descriptive statistics (percentages, mean, standard deviation, and median) was the most popular, with more than two-thirds of the studies reported using at least one or a mixture of the descriptive-analytic tools (n = 13).Following this, almost all studies (n = 14) indicated performing a statistical test/s or/and qualitative analysis, while two studies did not clearly articulate their analytical techniques (Chen, 2018;Hashim et al., 2019).About 43% of studies performed paired t-test (n = 7), followed with Wilcoxon signed-rank test (n = 4), thematic analysis (n = 3) independent t-test (n = 3), Friedman test (n = 2), Mann-Whitney U test (n = 2), ANOVA (n = 2), MANCOVA (n = 1), Multiple Comparison Test (n = 1), Pearson's chi-square (n = 1), Kruskal-Wallis H test (n = 1), and Scheffe test (n=1).
QUALITY APPRAISAL
A quality assessment for each of the 16 papers was conducted to ensure their methodological rigor.Examining the research design of the included papers revealed that 11n of them were either quasi-experimental or experimental, four employed mixed-methods design, and only one adopted a qualitative research approach.Quasiexperimental and experimental studies were assessed by drawing on nine questions offered by Joanna Briggs Institute (JBI), as shown in Table 3 and Table 4.As depicted in Table 5, the Mixed Methods Appraisal Tool (MMAT) (Hong et al., 2018) was employed to evaluate the quality of mixed-method papers.Meanwhile, the risk of bias for the qualitative paper was assessed by employing the Critical Appraisal Skills Program (CASP) checklist, as shown in Table 6.
To be included in this review, papers should at least fulfill four criteria of their assessment tool.The full criteria of each assessment tool are mentioned in the Tables (3,4,5).This means that to be considered in this review, each study must score 4 in its quality appraisal.When calculating the sum score of each evaluation, all 16 papers scored four and above, so no study was excluded from this review.Conducting the quality appraisal of the included papers revealed that six papers were classified as of high quality (Dincer et al., 2020;Kilic et al., 2018;Lee andKim, 2018, Nazligul et al., 2019,;Wang et al., 2020, Zacarin et al., 2019), and ten papers were deemed to be of moderate quality (Altunkaya, 2018;Chen, 2018;El Shazly, 2021;Han & Keskin, 2016;Hashim et al., 2019;Kim et al., 2018;LeFebvre et al., 2020;Pontillas, 2020;Ramamuruthy, 2019;Stupar-Rutenfrans et al., 2017.
TYPES OF INTERVENTIONS USED TO REDUCE SA
In the current review, SAI were conceived to include all scholar performed attempts that target assuaging SA within the context of undergraduate F/SL Learners.Our SAI classification was informed by their mechanism and, in particular, how they operated to address the SA issue.According to this conceptualization, SAI can be categorized into two main and broad categories: direct and indirect.This distinction is supposed to be of usefulness as it isolates anxiety stimuli and ties it with the most appropriate intervention.Although the proposed categories share the same aim, i.e., alleviating SA, they adopt different manners to achieve this goal.Figure 1 management.These Inventions can be further subdivided according to the issue(s) they are designed to deal with into habituation-based and management-oriented subcategories.These subcategories will be analyzed in the following paragraphs.
Habituation-Oriented on one hand, the basic premise of habituation-based interventions, as the name implies, is that individuals, over time, become habituated to SA evoking stimuli and recognize that the anticipated catastrophic situation is unlikely to occur (Milosevic & McCabe, 2015).As a result, they get multiple opportunities to re-evaluate the irrationality of their beliefs and manage their fears appropriately.The current review found three studies (Nazligul et al., 2019;Stupar-Rutenfrans et al., 2017;Zacarin et al., 2019) that employed virtual reality exposure.
This exposure can be classified as a habituation-based SA intervention.Gradual exposure technique regarding the number of audiences (empty class, small audience, and large audience) and/or the size of the avenue (small followed by medium and then large) was employed by these studies to elicit participants' SA and familiarize them with virtual reality avenue.EFT consists of preparation (identifying reasons and measuring levels of anxiety), followed by repeating some sentences (e.g., I forgive myself) while a therapist was tapping on participants' meridian system.At the feel yourself stage, participants would take a breath and recite inside some sentences such as 'I love myself'.Although Dincer et al. (2020) study included a control group and a comparison group, the intervention was applied in a single session without follow-up.
Drawing on cognitive reconstructing principles, Wang et al. (2020) designed a stimulated tutoring session that offered training and practice on overcoming SA through a conversational agent.The conversational agent was presumed to serve as a psychological counselor by which its users divulged private information.Although Wang et al.'s (2020) paired t-test demonstrated no significant difference in participants' pre-and post-speech anxiety scores, the qualitative data revealed that the conversational agent contributed to mitigating pre-speech anxiety.In a similar vein, a psychoeducational group training led to a significant reduction in SA and a higher life satisfaction score in the experimental group compared with the control group (Kilic et al., 2018).More interestingly, the training's positive impact was still evident at the time of the follow-up one year later.
Indirect SAI
In Further, simulation education combined with problem-based learning decreased nursing students' communication apprehension in group discussion and boosted their clinical self-efficacy, yet there was no significant difference in their assertiveness (Kim et al., 2018).However, what specifically served to reduce nursing's communication apprehension in group discussion through the stated intervention remains unclear.It is essential to interpret the study's results with some caution due to several concerns, including the small sample size consisting of mainly male participants, the lack of a control group and follow-up, and using the same speaking task to assess participants' speaking ability in pre-and post-intervention.The employment of a task-based approach positively impacted participants' post-test grades, but its effectiveness in reducing learners' SA remains unclear (Ramamuruthy, 2019).However, developing speaking skills might not result in SA reduction.For instance, an activity-based oral presentation course increased psychological counseling department students' SA rather than decreasing it (Altunkaya, 2018).Using a mobile application (e.g., Flip Grid) to record participants' impromptu speech videos, Hashim et al. (2019) found that adopting speaking routines in a classroom environment helped their participants develop their speaking skills and self-confidence while reducing SA.However, the effectiveness of Hashim et al. (2019) intervention was based on reflections and opinions rather than statistical analysis.
In a similar vein, the employment of the WhatsApp mobile application also resulted in SA decline to some extent (Han & Keskin, 2016).Notwithstanding that the male participants experienced higher SA levels than their female counterparts between pre-and post-intervention, females lowered their SA towards the end of the said intervention more than male participants.Chen (2018) provided learners with a holographic learning support system that helped them practice speaking in different scenarios and ultimately declined their SA.Practicing speaking with the assistance of technology may not result in SA decrement.For instance, artificial intelligence (chatbots and Mondly) slightly intensified undergraduate learners' SA, yet it contributed to facilitating their language learning and resulted in some ensuing gains; however, the study lacked a control group (El Shazly, 2021).
Feedback-Oriented.The current review also found one study utilizing teacher feedback to improve participants' speaking abilities and reduce SA.Notwithstanding that there was no control group, Pontillas (2020) found that Popsispeak (Outcome Based Teaching and Learning Strategy) lowered participants' SA and enhanced their speaking abilities.Interestingly, the teacher-researcher provided feedback in the form of motivational letters to each student after his/her speech delivery, which included suggestions for improving participants' speaking skills.
DISCUSSION
The current study has examined a range of SAI employed in undergraduate context around the globe between January 2015 and Mid-June 2021 in terms of their association with SA decrement to summarize and clarify the current state of the field and offer directions for future research.A total of sixteen studies that have met the inclusion criteria and have had a quality rating of at least moderate were included in this review.Notwithstanding that this is a small sample, it demonstrates a range of SAI used in undergraduate education in various countries with varying durations and results.Generally, the review offers a promising indication regarding the efficacy of various interventions in the mitigation of SA.Some of the most pertinent issues from this review will be discussed next.
In line with previous research, virtual reality exposure interventions decreased S/FL learners' SA (Nazligul et al., 2019;Stupar-Rutenfrans et al., 2017;Zacarin et al., 2019).In a virtual environment, usually with a trained therapist conducting the treatment, individuals confront computerized simulation of anxiety-evoking situations wherein the environment can be manipulated and customized according to the individuals' fears (Chesham, Malouff, & Schutte, 2018).Virtual reality exposure habituates individuals to anxiety-producing situations by gradually confronting each situation, starting with the least then the most feared until anxiety decline (Carl et al., 2019).In the studies, Nazligul et al. (2019) and Zacarin et al. (2019), computerized virtual realities were used solely in the research site and under the guidance and supervision of an experimenter or a therapist.In contrast, participants used (Stupar-Rutenfrans et al., 2017) mobile video virtual reality exposure in their home environments, making it perhaps more flexible, cost-effective, and feasible concerning time and human resources.
Giving that each study's sample size was relatively small (Nazligul et al., 2019;Stupar-Rutenfrans et al., 2017;Zacarin et al., 2019), the generalizability of the findings to a broader or a similar population might be questionable.
Thus, further work is required with a larger population.Additionally, due to the Covid-19 pandemic, online teaching and learning have become the accepted norm instead of face-to-face instructions across different countries around the world.Thus, the virtual world has become the actual one in many aspects of our lives, and education is no exception.Indeed, learners now have access to various programs, such as Zoom and WebEx, that may substitute virtual reality exposure programs in the sense that in many educational institutions, these programs form the virtual environment where learning and teaching take place and, therefore, provide learners with more authentic SA experiences.Additionally, these programs may also be combined self-perceptions by determining irrational and negative self-statements and then substituting them with more rational and positive ones (Fremouw & Zitter, 1978).The employment of the conversational agent did not demonstrate a significant statistical SA reduction.However, the qualitative analysis demonstrated that the stated intervention contributed to alleviating pre-speech anxiety.Nevertheless, it is essential to note that evaluating the efficacy of the intervention from one laboratory session does not permit long-term evaluation.Therefore, longitudinal research in a classroom environment is needed to assess whether the said intervention can have a positive long-term effect on individuals' SA.In addition, some participants complained about not receiving personalized feedback from the conversational agent about their performance, so future designs can consider incorporating this technique into their conversational agents' designs.
In contrast to Wang et al. (2020)) cognitive restructuring, Kilic et al. (2018) psychoeducational training sessions were carried out by a human trainer rather than a computerized conversational agent, wherein SA was reduced significantly.Cognitive-behavioral techniques (CBT) and subjective well-being-increasing activities (SWIA) were both utilized in the development of Kilic et al. (2018) intervention which aimed to improve learners' self-awareness and helped them to discover the most appropriate SA reduction methods on their own.During the intervention sessions using group dynamics, participants discussed causes and levels of anxiety arousal and used effective activities to address suitable solutions.Despite the small sample size, this SA reduction intervention seemed feasible, cost-effective, and realistic.It can be implemented in educational settings without much effort from instructors since their roles were limited to planning, moderating, creating a positive environment, and facilitating the whole process.Further, our literature review identified some interventions designed to improve and foster speaking skills using various pedagogical and technological training approaches.For example, when reporting the effectiveness of the task-based approach apropos decreasing SA, Ramamuruthy (2019) stated that the intervention positively affected participants' grades as learners with high anxiety levels achieved low grades compared with less anxious ones.
These study results need to be interpreted cautiously due to several concerns, including recruiting a small sample wherein the vast majority were male participants.There was no control group nor continuation besides using the same set of pre and post-intervention questions to assess participants' speaking abilities.Furthermore, despite different terminology applied, two speaking skills training with technology interventions (Han & Keskin, 2016;Hashim et al., 2019) were found to share the core idea of Lee and Kim (2018) output-based planning strategy, which provided learners with some preparation time before the spoken task, yet they recorded their participants' speech using mobile applications.Moreover, the efficacy of engaging in a communication course (LeFebvre et al., 2020) or using a holographic learning support system (Chen, 2018) Principally, EFT regulates energy flow by stimulating the meridian points and relaxes an individual while concentrating on emotional reactions causing anxiety.To put it differently, EFT combines "elements of exposure, cognitive therapy and somatic stimulation" (Stapleton, 2020).By contrast, as a form of meditation, breathing therapy regulates breathing by focusing on the mind and respiratory organs (Kuppusamy, Kamaldeen, Pitani, Amaldas, & Shanmugam, 2018).
Even though Dincer et al. (2020) (Pontillas, 2020).This method seems quite promising as it provides individualized feedback combined with motivation that considers learners' differences and learning needs besides maintaining their privacy as it is done in a written form.Additional research is required to supply more objective evidence concerning what works best, in which contexts, and why.
Generally, the included studies in this review mostly lack clarity of reporting criteria or factors that were leveraged to justify the appropriateness of their selected interventions prior to embarking on field research.In other words, robust reasons have not been provided in association with intervention suitability for a specific group of learners before embarking on scholarly research.Besides, a general trend with the reviewed studies is associated with unclear negotiation and construction of SA interventions with the educational setting that could compromise their transferability to other contexts and their real-life feasibility.Given this situation, several factors may improve the effectiveness of SA interventions and, thus, save time when searching for anxiety-combatting antidotes that need to be considered.
Based on the review findings, it is highly advisable to consider several issues when selecting and applying various SA reduction interventions.Despite not providing straightforward solutions for SA, some elements offered may aid educators, instructors, and learners in selecting the most appropriate SA reduction interventions to their context.Likewise, a doctor cannot prescribe medicine for patients in the medical sector unless they are diagnosed.
The same is perhaps true in the educational field.This implies that knowing what causes learners' SA is probably of high importance.For example, if learners are found to experience SA due to anxiety triggered from facing a crowd, it would be feasible to introduce them to the habituation interventions.In contrast, if SA occurred due to arousal of negative emotions or negative self-perceptions, it would be conceivable to substitute these emotions and perceptions with more positive ones.English speaking deficiency is another potential element that might significantly contribute to increasing learners' SA.Thus, it is highly recommended to examine learners' S/FL speaking competency before assigning SA reduction intervention.
LIMITATIONS
It is undeniable that every study has its limitations, and this review is no exception.The present review is
Figure 1 .
Figure 1.PRISMA flow chart of the SAI included in the review.
Besides assessing the interventions' efficacy in reducing undergraduate learners' SA, some studies measured other variables, such as emotion regulation (Stupar-Rutenfrans et al., 2017) (Emotional Regulation Questionnaire (ERQ developed by Gross and John (2003)), communication strategies (Lee & Kim, 2018) (Oral Communication Strategy Inventory (OCSI) created by Nakatani (2006), satisfaction with life (Kilic et al., 2018) (Satisfaction with Life Scale (SWLS) invented by Diener, Emmons, Larsen, and Griffin ( shows the classifications of SAI.The classifications on the left are based on our systematic review results with their examples provided on the right.Each main category (direct and indirect) is divided into more subcategories: Habituationbased, self-management, face-to-face speaking skill training with technology, and feedback-oriented.The following sections pull together all SAI reviewed in this paper and summarize the main results of each within their respective category.6.1.Direct SAI Generally, direct interventions address mitigating SA without the involvement of any other intermediary(ies).This implies that the employment of these interventions is directed to treat SA through following their diversified instructions and techniques.In other words, direct SAI constitutes various ranges of interventions designed to mitigate SA by leveraging their different techniques.In this sense, the direct interventions explicitly target SA the present review, indirect SAI refers to those interventions that reduce SA by addressing individuals' speaking skills rather than dealing directly with SA.In other words, indirect interventions used various techniques to develop and foster learners' speaking skills which are presumed to reduce SA.Although indirect SAI are not strictly targeting anxiety management, they can probably facilitate SA assuaging.These can be further subdivided into face-to-face speaking skill training, speaking skill training with technology and feedback oriented.Face-to-Face Speaking Skill Training.Some interventions reviewed in this study opted to improve learners' speaking ability through various pedagogical means, which are assumed to alleviate learners' SA eventually.LeFebvre et al. (2020) examined the changes in SA that participants experienced while engaging in an introductory communication course.Although LeFebvre et al. (2020) found that participants' SA decreased over time, there is a lack of understanding regarding what precisely reduced participants' SA.Applying strategic planning in which output-based planning group was compared to input-based planning group, Lee and Kim (2018) reported a significant decrease in SA over time regardless of planning strategy.
with other types of interventions designed to reduce SA.Accordingly, future research should examine the efficacy of these programs when combined with other interventions to reduce SA.Some interventions were designed to develop individuals' ability to manage SA.Wang et al. (2020) designed a conversational agent to assuage SA by drawing on cognitive reconstructing principles.Cognitive reconstructing is a part of the cognitive-behavioral interventions family, which is assumed to assist people to correct their negative limited to SA interventions published journal articles and conference proceedings between January 2015 and mid-June 2021 and applied in undergraduate educational contexts.Therefore, the omission of potentially relevant research in SA presented in other publication formats such as book chapters and unpublished doctoral thesis from the present review is considered one of the limitations.There have been no replications of most intervention procedures analyzed in this study.Thereby, further research needs to be conducted to ensure the efficacy of these SAI with diverse populations.Moreover, determining some methods' categories was not always straightforward owing to the complexity of SAI nature.Another limitation is associated with analyzing studies published only in English; therefore, no claims are made regarding having a fully comprehensive dataset.Despite these limitations, this systematic review is robust enough to represent accurate and meaningful trends concerning various SA reduction interventions applied in undergraduate education.9.CONCLUSIONThis article provides a systematic review of various SAI employed in undergraduate education and their influence on SA.It shows that SAI are studied by employing various methodological designs, namely experimental quasi-experimental, mixed-methods and qualitative designs.These research designs can provide useful information and insights which can pave the way for developing more sophisticated SA management interventions in future research.The review demonstrated that SAI could be categorized according to their mechanism in dealing with SA to direct and indirect interventions.It is highly hoped that the information presented in this systematic review can assist anxious S/FL learners in identifying the most appropriate SA reduction interventions.This review is believed to lighten the way for and inspire other scholars, educators, and instructors to develop SA management approaches that could create a more relaxed learning environment that suit their learners' individual needs.Finally, based on this study's results, we recommend examining the cause of SA besides the learners' speaking English proficiency levels before selecting or applying any SA reduction approach since these factors can significantly impact the success or failure of SAI.
Table 1 .
Inclusion and exclusion criteria for the review.
: FLCAS1, a modified version of Foreign Language Classroom Anxiety Scale; SD, Standard Deviation; PRCA-24, Personal Report of Communication Apprehension; SAS, State Anxiety Scale; ERQ, Emotional Regulation Questionnaire; SASPST, Speech Anxiety Scale for Pre-Service Teachers; ANOVA; Analysis of Variance; OCSI, Oral Communication Strategy Inventory; FLCAS, Foreign Language Classroom Anxiety Scale; MANCOVA, Multivariant Analysis of Covariance; PANAS1, a modified version of Positive and Negative Affect Schedule; STAI, State-Trait Anxiety Inventory; SWLS1, a modified version of Satisfaction with Life Scale; MWU, Mann-Whitney U test; Wilcoxon SRT; Wilcoxon signed-rank test; NCSES1, modified version of Nursing Clinical Self-Efficiency Scale; SSES, Satisfaction with Simulation Experience Scale; LSAS, Leibowitz Social Anxiety Scale; SUDS, Subjective Units of Distress Scale; IAS, Interaction Anxiousness Scale; SAAS, Social Appearance Anxiety Scale; BFNES, Brief Fear of Negative Evaluation Scale; AA, Audience Anxiousness; CAI, Communication Anxiety Inventory; PRPSA, Personal Report of Public Speaking Anxiety; SSPS, Self-Statement during Public Speaking Scale; CSQ, Client Satisfaction Questionnaire; MCT, Multiple Comparison Test; SUDS* Subjective Unit of Disturbance Scale; Kiesler's scale1, a modified version of Kiesler's scale, PRCA1 a modified version of Personal Report of Communication Apprehension; n.a.not applicable or insufficient description; α, Cronbach's alpha; B, Before intervention application; A, After intervention application.
table with predefined headings Table2.These headings cohered into various characteristics of the papers which included publishing information (researcher, year of publication and location of study), Intervention (Intervention type, objectives of research, duration) research methodology (research approach, data collection instruments, analytic method, and validity, reliability, trustworthiness, and credibility), and major research outcomes.In the Note
Table 1 .
Quality Appraisal for the included experimental studies.
Table 2 .
Quality Appraisal for the included experimental studies.
Table 3 .
Quality Appraisal for Included Qualitative study.
Table 6 .
Quality appraisal for mixed-method studies.
Dincer et al. (2020))al.(2017)utanexperimental study to test the effectiveness of computerized 3 dimensioned virtual reality exposure compared to psychoeducation training concerning reducing SA.As evidenced by Wilcoxon singed rank test.Both interventions led to a significant decline in SA; however, the effectiveness could not be determined in the long-term due to a lack of follow-up.The findings ofStupar-Rutenfrans et al. (2017), a longitudinal experimental study, revealed that virtual reality exposure was more effective in decreasing SA for participants with high SA levels than their counterparts with moderate levels of SA.Zacarin et al. (2019)found that virtual reality exposure was accompanied by a breathing exercise.Although they reported that the said intervention contributed to lowering the participants' SA with the conduction of two follow-ups, we need to interpret this result with extra caution due to the relatively small sample size (n = 6) besides the fact that all participants were females.Management-Oriented.On the other hand, management-oriented interventions are those which are directed at developing individuals' ability to manage SA by adopting diversified techniques.Scholars worked on improving and understanding learners' negative emotions associated with SA.Comparing the efficacy of breathing therapy, and Emotional Freedom Technique (EFT),Dincer et al. (2020)concluded that EFT was more effective than Breathing therapy concerning SA reduction, as demonstrated by Cohen's effect size.Breathing therapy was performed through the following three stages: relaxation stage, breathing stage, then feel yourself stage, whereas studies was that they recruited relatively small samples.However, the amount of information and descriptions provided by each author researcher of the three studies concerning virtual interventions designs and implementations provided sufficient details that might allow replication with large samples.SA Speaking Skill-Training with Technology.It is probably undeniable that technology has impacted how individuals do things and how it has become an indispensable part of life, wherein the educational environment is no exception.Speaking skill training with technology interventions may have the potential to achieve both SA reduction and S/FL development.They are designed to foster learners' speaking abilities through interacting with virtual characters or with each other through using a range of technological means to carry out that interaction.
Dincer et al. (2020)nd studies comparing the effectiveness of two interventions concerning SA reduction.When comparing EFT and breathing therapy,Dincer et al. (2020)found the EFT was more effective in reducing SA, as evident by Cohen's effect size.
to reduce SA remains unclear.In general, a serious concern with speaking skill training interventions is perhaps attributed to the lack of understanding and ambiguity surrounding what precisely lowered participants' SA and, thus, merits further research to determine what exactly contributed to decreasing learners' SA.
Zacarin et al. (2019)roup and a comparison group, the results still require further investigation because the interventions were compared based on a single session, and no follow-up was conducted.Therefore, to determine the long-term effect, future research can benefit from conducting a longitudinal study.Moreover,Lee and Kim (2018)study compared two strategic planning methods (input-based and outputbased planning) and found them effective in alleviating SA over time.Both methods helped learners better organize their thoughts and prepare more than directly engaging in the spoken activity.These strategic planning methods can be applied in a classroom setting without much effort from the instructors (i.e., practicing or training prior to implementation).On the contrary, the current review has identified two studies that used a combination of two interventions.Combining simulation education with problem-based learning reduced nursing students' apprehension regarding communication during group discussions, but not their public SA(Kim et al., 2018).According toZacarin et al. (2019), a combination of virtual reality exposure and behavioral therapy contributed to decreasing psychological students' SA and elevating their coping behaviors.It would be more beneficial for future research to enhance mitigation of SA by combining multiple interventions; however, these combinations should be well justified and scientifically based.
(Altunkaya, 2018)ative evidence regarding the effectiveness of SAI was identified.Engaging in an activity-based oral presentation course(Altunkaya, 2018)and using artificial intelligence (chatbots and Mondly) (El Shazly, 2021) increased learners' SA.Further research is necessary to examine these interventions' efficacy in other contexts using appropriate procedures and analytic methods with adequate sample size and sufficient duration.In addition, the present review has highlighted the necessity for additional research regarding feedback-oriented intervention as only one study employed teacher feedback to reduce learners' SA
|
2022-05-13T15:08:07.137Z
|
2022-05-11T00:00:00.000
|
{
"year": 2022,
"sha1": "325435677f0b0aeaf99cae4899a07850f7b19d60",
"oa_license": null,
"oa_url": "https://archive.conscientiabeam.com/index.php/61/article/download/2994/6455",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "702adb9cbd4c769ffa1a6721ab47933797f2244f",
"s2fieldsofstudy": [
"Education",
"Psychology"
],
"extfieldsofstudy": []
}
|
157513316
|
pes2o/s2orc
|
v3-fos-license
|
STRATEGIC ROLE OF WOMEN IN CHRYSANTHEMUM-BASED ENGINEERING DESIGN ON DISASTER AFFECTED AREAS Muafi
Strategic role of women makes a positive contribution in chrysanthemum-based engineering design on disaster affected areas. The Merapi eruption in 2010 in Special Region of Yogyakarta (DIY), Indonesia, especially in the village of Hargobinangun, brought tremendous impact in the field of environmental damage, social economy, and agriculture. Post-Merapi eruption, chrysanthemum flower cultivation in the Hargobinangun village became stagnant. Some farmers have started to despair and lost determination. But thanks to the encouragement and the strategic role of women as husbands’ companions, farmers began to rise up to build sustainable and competitive green areas. This is a primary research and the data obtained was from questionnaires and interviews with some respondents. This research also used a secondary data obtained from publications of related institutions. The statistical technique used is Partial Least Square. This research concludes that the strategic role of women (as strategic partners, fighters, administrative experts, agents of change) was able to provide a significant contribution to labor productivity, family welfare, and green environmental performance.
In order to support the strategy of the Medium Term Development Plan or Rencana Pembangunan Jangka Menengah (RPJMN) 2010-2014 Sleman and the Strategic Plan of Indonesian Ministry of Agriculture 2010-2014, the need to increase the competitiveness and performance of agricultural areas, especially green areas, is necessary.Agricultural development in green areas is expected to have a strategic role in improving a sustainable development, creating harmony and balance in the process of economic empowerment, and transferring environmentally friendly technologies.
The strategic role of agriculture in supporting food sovereignty is outlined in the Strategic Plan of the Ministry of Agriculture that targets the development of horticulture, one of which is chrysanthemum as one of non-food national commodities and as an import substitution.Achievements of this target is certainly not easy, given that the agricultural development is still faced with some fundamental problems.Some of these problems include an increase in environmental degradation and global climate change, limited availability of infrastructure, national seed and breeding systems that are not optimal, farmers have limited access to capital, and weak institutional capacity.This is presumably due to the unavailability of overall agricultural development engineering design that enables the creation of inter-regional cooperation so that agricultural development can be carried out effectively and efficient (RI Regulation No. 50/Permentan/CT.140/8/2012).
The Merapi eruption in 2010 brought tremendous impact in the field of environmental damage, social economy, and agriculture.The affected area of chrysanthemum plant development was in Hargobinangun village, Pakem, Sleman, Special Region of Yogyakarta (DIY), Indonesia.Whereas since 2005, the region has been appointed as the center of chrysanthemum plant in DIY Province considering the altitude of the area is 700-900 meters above sea level and could be eligible for the growth of chrysanthemum plants.Activities concerning chrysanthemum plants, such as flower cultivation, have been done by more than 100 local farmers from 13 farmer groups and could manage an area of 20,000 m2 with a production capacity of 15,000 flowers per week (Bappeda, 2003).After the eruption of Merapi in November 5th,2010, chrysanthemum flower cultivation in the Hargobinangun village became stagnant.Most farmers did not know what to do because crops were destroyed.That was because the area is very is close to Mount Merapi and is just 4-10 km from the peak of Merapi.Back in 2008, farmers of Hargobinangun have actually had a successful chrysanthemum business in which at the time obtained steady assistance from several University in Yogyakarta.
Therefore, at this time, both men and women farmers need to be synergized to emerge from the downturn as a result of the Merapi eruption.Men should not ignore the role of women because they can serve as agents of change.Women have an important involvement in family decision making, and some even become the backbone of the family.Well-educated women will create a generation of good and prosperous country (Khomeini, 2004;Nizar, 2014).Some research shows that women in Indonesia have a strategic role in building green areas and competitiveness (Nizar, 2014;Sujarwati, 2013;Sajogyo, 1985).
Strategic role of Human Resources Management (HRM)
One business issue that has been going on for the past few years is the ongoing number of important developments in literature on issues related to human resources management strategic human resource management in particular.Some academics and practitioners focus their attention on strategic human resource management approach.This means that approach in the functions of human resource management are not limited to the functions of personnel alone, but also extended to the involvement of strategic management process as a whole.Armstrong (2008) added that the major business issue affecting HR strategy includes: (1) the company's purpose, (2) an increase in competitive advantage, (3) the need to develop a more positive and performance-oriented culture, and (4) culture management.Human resource strategy is expected to help make the company's business strategy to work well.The business strategy must take opportunities and organizational barriers into account to achieve superior performance and competitive advantage.HR strategy is a series used by human resource to help the company achieve its strategic goals (Dessler, 2013).Mathis and Jackson (2008) suggested that the role of HR as a strategic partner has three views, namely; (1) HR activities are seen to be not strategic at all.This is because human activities are just mere operational.
(2) The strategic role of HR is to adjust individual HR practices (recruitment, rewards, etc.) to fit in with the company's strategy and competitive strategy.According to this view, the top management forms a particular strategy for the company, in which they ask and require HR to create human resources programs to successfully implement the company's strategy.(3) HR management is an equal partner in the process of strategic planning.According to this view, the role of HR management is not just adjusting their activities to the demands of business strategy; they also do not only run daily operational tasks.HR management should be equal partners, both in the formulation and implementation of competitive strategy and includes the entire company organization.
Ulrich (1997) added that the role of HR in building a competitive organization is represented in two points; focus and activities.Focus ranges from long term/strategic to short term/operational.HR professionals must learn to be both strategic and operational, focusing on the long and short term.Activities ranges from managing processes (HR tools and systems) to managing people.Furthermore, if this is associated with the strategic role of women farmers in Hargobinangun, it means that the organization managed by them should be focused on a few HR strategies and organizational practices in terms of a business strategy.Working professional women farmers should be strategic partners that can help to ensure the business strategy to run successfully.Translating business strategies in human resources practices can be done in three ways: (1) the business can adapt to changes since the time of conception to execute a strategy is short, (2) the business can be better if it meets customer demands because the strategy of customer service is translated into specific policies and practices, and (3) the business can achieve financial performance through a number of its effectiveness in executing the strategy (Ulrich, 1997).Women farmers that join Gapoktan (farming group) ASTHA BUNDA should, together with the management of SMEs, create a business strategy as a framework that directs the specific activities designed for HR, such as the process of recruitment and selections of employees, appropriate staffing, training and development, and performance assessment.This is expected to be able to generate competence and employee behavior, which in turn will definitely help the business and implement their business strategy according to the organization's objective (Dessler, 2013).The approach of strategic human resource should be able to ensure that the human capital SMEs (knowledge, skills, and abilities) can contribute to achieve the business organization's objectives (Huselid., Jackson, and Schuler, 1997).
Strategic Role of Women in Economic Development and Green Environment
Developments in society always include women's role in supporting the process of change.
Women can play a role in starting, moving, and disseminating the change process.Woman as an agent of change is one who plays as a catalyst and estimates responsibilities for managing the change activity (Robbins, 2001).Being an agent of change means having to build organizational capacity to capture and capitalize on change (Nurhayati, 2001;Ulrich, 1997).In organizations, experience shows that the success of change starts with changing individuals and then changing the organization, which is often referred to as an"individual out"approach (Black & Gregersen, 2003).Thus, the success of women in their role as agents of change needs to be focused on individuals prior to the conduct of re-drawing a mental map of every members involved in the organization (Zhou et al., 2004).If the individual has changed and is ready for a change, then the next role of agents of change is to motivate their members to behave accordingly and in harmony with the enthusiasm of the organization.One of the key indicators of using a role as an agent of change is that one is able to build a good relationship with others (Soetjipto, 2009).Change must begin by preparing all human resources to accept the change because human is always the subject and the object of change due to human nature, as well as having the nature to resist change.Changes in human resources must be initiated by the getting rid of the old behavior patterns that tend to maintain the status quo to be changed to be willing to accept a new mindset that evolve dynamically (Victor, 2002).Zaltman and Duncan (1977) suggested that a person, including women, who acts as agents of change must have these competence: (1) technical qualifications, (2) administrative capabilities, and (3) and interpersonal relationships.In this case, technical qualifications means having technical abilities in determining business locations, scale of operations, and production processes.Administrative capabilities means having administrative skills associated with business group activities.Interpersonal relationships means that women should have the abilities in dealing with other people or have a good empathy on community members and the surrounding environment.Besides, Zaltman and Duncan (1977) added that the effectiveness of the strategic role of women is needed to stimulate and accelerate the process of collaborative problem solving.Havelock and Havelock (1973) explained that in fact if someone wants to act as an agent of change, there are four ways that can be done, which includes: 1.As a Catalyst, who encourages people to change; 2. As a Solution Giver, who provides ideas about how change should occur; 3.As a Process Helper, who encourages the process of change; 4. As a Resource Linker, who helps others to obtain relevant resources and bridging people together to solve problems.
The purpose of Millennium Development Goals2015 in including women in environmental management so that women understand the importance of environments, thus women will maintain, preserve the environments well, and also keep the environments clean (Sudarwanto, 2010).This research is focused on the strategic role of women in economic development for families, especially in the development of chrysanthemum plants that is expected to have a broad impact on the competitiveness and performance of particular areas.Chrysanthemum as one of the commodities that can be used as ornamental plants, potted plants, and flowers is included as a major product in the production of ornamental plants nationwide.Currently, chrysanthemum is included as the most popular flower in Indonesia because of its advantages, namely a flower that is rich in color, relatively more durable, and also affordable.This is shown by an increase in the amount of production and the addition of chrysanthemum harvest areas each year in order to meet consumer demand.Increased production of chrysanthemum flowers indicates that consumer demand for these commodities is also increasing.Along with the increasing demand of chrysanthemum flowers, agribusiness opportunities need to be developed (Butaflika, 2008).
So far, we know that plants can act as a filter, a source of aesthetic and stress relievers (Soekartawi, 1996), provide coolness and comfort, creates a greener environment, lower the air temperature and clean the air at the same time (Sudarmono, 1997;Endah, 2001),a source of education, economic and social matters, as well as medicinal plants (Rukmana, 1997).Therefore, women as agents of change should be able to act as a catalyst, provides solutions, helps the change process, and links resources in matters relating to both the technical and non-technical aspects of business, and also production, institutional, and social aspects related to the chrysanthemum-based areas affected by the Merapi disaster.
Besides being agents of change, women can have strategic roles for instance: (1) as a strategic partners, in which women encourage the organization to implement a strategy, (2) administrative experts, in which women must improve the efficiency of the traditional functions and the overall organization, (3) fighters, in which women must continuously increase their performance so as to improve organizational commitment (Nurhayati, 2001;Ulrich, 1997).Several studies prove that there is a positive relationship between the strategic role of HR on productivity (Becker & Huselid, 1998;Huselid, 1995).Some research findings also give the same conclusion that the strategic role of HR is indispensable both in business and social organizations to increase organizational performance (Alleyne, et al., 2005;Bou & Beltran, 2005;Edelman, et al., 2005;Carmeli, 2004;Gilley, et al., 2004;Muafi, et al., 2016;Muafi, 2016).Schuler and Jackson (1987) added that if an organization is to manage its human resources, every individual there will be involved and that it will need quite a long time.This means that in solving the problems that exist now, the organization must keep having a long-term vision and continue to improve their ways of working so that the desired results can be obtained quickly.Walters (1985;Werther & Davis, 1996) added that human resource management is a life of organizations that focuses on the management and utilization of people effectively to increase their productive contributions to the organization through strategic ways, is ethical, and also is accountable.Ulrich (1997) stated that the role of HR strategy focuses on HR strategies and practices in business strategies.HR professionals work to become strategic partners and help to ensure business strategies.In essence, the functions of HR can no longer be administrative, but must simultaneously be strategically oriented.Ulrich's opinion (1997) is also reinforced by Husehild et al. (1997) which is splits the functions of HR into technical and strategic HR.Technical HR emphasizes on settings functions and administrative processes, while strategic HR emphasizes more on empowerment functions and social interactions.Findings of Husehild et al. (1997) stated that there is a significant inter-relationship between strategic HR with employee productivity, cash flow, and market value.
H1: Strategic role of women (strategic partners, administrative experts, fighters, and agents of change) has a significant positive effect on labor productivity.H2: Strategic role of women (strategic partners, administrative experts, fighters, and agents of change) has a significant positive effect on family welfare.H3: Strategic role of women (strategic partners, administrative experts, fighters, and agents of change) has a significant positive effect on performance improvement of green environments.
RESEARCH METHODS
This research was conducted on SMEs of chrysanthemum farmers in Hargobinangun village, District of Kaliurang, Yogyakarta, Indonesia.These chrysanthemum farmers have a Farmers Group Association (Gapoktan) called ASTHA BUNDA.They are traumatized and are almost desperate since the Merapi eruption in 2010, which has brought tremendous impact on the aspects of environments, socioeconomic, and agriculture.
Chrysanthemum farmers that only relied on their businesses by planting chrysanthemums then have lost motivation and passion to develop chrysanthemum because all the land has been covered by volcanic ashes.But due to the role of women, the husbands have a strong will and motivation to bounce back.The population in this research is every SME chrysanthemum farmers that join ASTHA BUNDA farming group, represented by a women chrysanthemum farmeran SME ownerwith a total population of 122 SMEs.The targeted number of respondents is 100 SMEs.This already meets the requirements for the survey research, with a Partial Least Square (PLS) statistical technique (Hair et al., 1995).The sampling technique used was purposive sampling.This is because the criterion of respondents has been chosen in advance; women chrysanthemum farmers that join Gapoktan ASTHA BUNDA.Respondents who returned the questionnaire were 89 (response rate 73%).
This research consists of primary and secondary data, obtained through questionnaires and some publications of relevant agencies.Publications are in the form of data regarding the number of employees, the types of chrysanthemums, varieties preferred by consumers, production cost, total income, and others issued by government agencies, universities, and from Gapoktan ASTHA BUNDA.This secondary data was very helpful to support the primary data used by researchers.The type of questionnaires given to the respondents were closed and asked the women's perceptions who manage SMEs of chrysanthemum in Hargobinangun, Kaliurang, DIY.The scaling technique used for the woman role variable (strategic partners/MS, fighters/PJ, administrative experts/AA, and agents of change/AP), productivity (prod), family welfare (welfare), and green environmental performance (Envir) was a Likert scale of 1 to 7. The number of items is referred to and modified from Ulrich's research (1997;Muafi et al., 2014); with 4 items of women as strategic partners, 4 items of women as administrative experts, 5 items of women as fighters, and 4 items of women as agents of change.Additionally, 4 items of labor productivity, 4 items of family welfare, and 4 items of green environmental performance, modified from Klein (2008;Chenoweth, 2011;Karnani, 2007).Data was analyzed using PLS (Partial Least Square) because it serves as a powerful technique for analyzing latent variables in structural equation models with a variety of indicators (Sirohi, et al., 1998;Roostika, 2011).
Description
Most respondents in this research have more than 5 years of business experience amounting up to 58 SMEs (84%), have a workforce of less than 5 employees (14%) and plant varieties of flowers such as Reegan, Puma, and Town Talk.Varieties in consumer preferences are the ones in yellow, white, and red with all the variance (60%).In order to produce 5,000 stalks of chrysanthemum flowers, it requiresa 120 m2 building area of plastic house, with a cost of Rp.6,000,000.00.The cost of production for 4 months requires Rp.2,021,388.If the selling price per stalk is Rp.1,000.00, the net income earned would make total of Rp.2,978,612.00.Meanwhile if farmers plant corn on the same area, they could only earn Rp.300,000.00.Chrysanthemum needs in DIY reach up to 5,000 belts/week (250,000 belts / year) with a price range of Rp.10,000-15,000/belt.This level Table 2 shows that most of the variables have a greater value than the correlation of other variables.There is only one variable that has a smaller value than the correlation of other variables.It is clear that the items in this study had a good discriminant validity that can be used as an instrument of data measurement.
Reliability Test
The reliability test results can be seen from cronbach's alpha and composite reliability, listed in Table 3.Based on Table 3, it can be seen that cronbach's alpha and composite reliability have a value greater than 0.7.This indicates that the questionnaire item in this research is reliable.
Determination Coefficient (R 2 ) of Endogenous Variable
Determination coefficient (R 2 ) is a measure that states the variance proportion size of a modifier with its predictor.In SEM, the value of determination coefficient (R 2 ) in an endogenous latent variable indicates the amount of variance, which can be explained by an exogenous latent variable.The endogenous variables in the inner model of structural comparison shows that productivity (prod), family welfare (welfare), and green environmental performance (Envir) are determined by women role (WR).Based on Table 4, the total value of determination coefficient (R2) can be calculated as follows: Q2 predictive = 1 -(1 -0,400)(1 -0,528)(1 -0,421) = 1 -(0,600 x 0.472 x 0.579) = 0,836 The total determination coefficient (R2) in this research is 0,836.This means that exogenous latent variables can be used to predict a model with an 83.6% value, while the remaining 16.4% is caused by other variables outside the model.
The evaluation of PLS model can also be done with Q2 predictive relevance, or as often called predictive sample reuse.This technique represents the synthesis of cross-validation and fitting function with using predictions of the observed variables and estimations of construct parameter, as listed in Table 5.Based on Table 5, it can be seen that the Q2 of all variables has a value greater than 0. This indicates that the model has a good predictive relevance.
Hypothesis Testing
Analysis results can be seen in Figure 1, as follows: The test result of inner model, as listed in Table 6, shows that out of the three correlations, they are all significant.Which means; (1) woman role has a significant positive effect on productivity, (2) woman role has a significant positive effect on welfare, and (3) woman role has a significant positive effect on green environmental performance.
DISCUSSION
It is recognized that the role of women as workers in the family is very important especially for husbands with a relatively small income.This research concludes that the strategic role of women has a significant positive effect on labor productivity, family welfare, and performance of green environments.This also supports the results of previous studies of Alleyne, et al. (2005;Bou & Beltran, 2005;Edelman, et al., 2005;Carmeli, 2004;Gilley., et al., 2004;Husehild et al., 1997) which confirmed that there is a significant inter-relationship between the role of strategic HR with employee productivity, cash flow, and market value.The first hypothesis is accepted.This indicates that the strategic role of women (as strategic partners, fighters, administrative experts, and agents of change) in improving labor productivity.The strategic role of women farmers should be stressed that the people involved in the organization is an important resource, as well as future investment.In order for HR to play a strategic role, the organization should focus on long-term human resource problems and implications.The strategic role in improving productivity can be done by not only addressing short-term needs but it also should be involved in solving problems related to programs effectiveness and strategic planning of the organization.In this role, they are required to solve the problems faced by the organization, such as productivity issues, work life quality, and an increase in competition, especially the invasion of flowers outside chrysanthemums and chrysanthemum supplies from outside Hargobinangun.The recent increased price of chrysanthemum production inputs can cause a decrease in farmers' income, so it is feared that this could lead to a slack of ornamental plants agribusiness.In addition, the application of SOP (Standard Operating Procedure) is one of the effective ways to get flower quality and high competitiveness.
By referring to the SOP, farmer manufacturers can meet the quality standards appropriate to the preference of international consumers as this will open up export opportunities to earn maximum foreign exchange (Direktorat Budaya Tanaman Hias, 2007).Flower qualities are factors that greatly affect the selling price of chrysanthemum cut flowers.In order for chrysanthemum cut flowers produced by women farmers in Hargobinangun to have high qualities, then an increase in production should be accompanied by improved farming technologies to improve the production quality to finally expect an increase in the selling price.This reparation of cultivation technique was done by applying a cultivation technology for chrysanthemums, both in specific locations and in an integrated manner.
Associated with labor productivity, the efforts continued to be made by women are: (1) restore and develop the system of managing chrysanthemum plants in an innovative way, (2) develop and cultivate chrysanthemum seed crops, (3) share their knowledge on the implementation of chrysanthemum plant technology to the ASTHA BUNDA Gapoktan group, (4) cooperate with banking companies and economic enterprises in the aspect of capital to develop their businesses, (5) arrange the overhauls of Gapoktan ASTHA BUNDA into a credible governance, transparent, and professional.Women who work will be able to increase family income which will then improve the quality of nutrition needed and family health.For women in Indonesia, it is not surprising if women have a dual role in which they have to work and be a housewife at the same time (Mudzhakhar, 2001).This research needs to consider the model of GAD (Gender Analysis Pathway) which offers a social construction of gender and equal roles between men and women, including when women are positioned equally with men as agents of change in terms of social change.Women can be treated as subjects of change and not merely empowered or simply viewed as objects of change (Abdullah, 2006).The findings of further research conclude that women have a strategic role (as strategic partners, fighters, administrative experts, and agents of change) in improving family welfare (the second hypothesis is accepted).As suggested by BKKBN (1995), family welfare is a family established by a legal marriage; is able to meet worthy spiritual and material needs of life; is devoted to God Almighty; has a harmonious relationship, and has a balanced life between family members, as well as between families within the community/society, and the environment.This means that a prosperous family is not only about the welfare aspects, but also includes the aspect of a meaningful and peaceful life.Women can complement the tasks of men and are an integral part of the role and responsibilities of the Father, family, community, and government.A family function that is well run and effective will clarify and strengthen the direction and purpose of the formation of a prosperous family with a good quality.Gender equality will occur if the wife and husband can fully put their human rights into realization and are equally potential to contribute ideas and energy to their family and community development.The third hypothesis was also proved acceptable.Women have a significant strategic role (as strategic partners, fighters, administrative experts, and agents of change) in improving green environmental performance.The attitude and behavior of women farmers should be oriented towards pro environmental (Bissing-Olson et al., 2012;Muafi et al., 2016;Muafi, 2016).Likewise, they must change their mindset, lifestyle, and even business practices by considering the environmental aspects around them.If the area that was built by women farmers in Hargobinangun can be made "green," then the public, both from inside and outside the area, has the potential to accelerate sustainable resources for economic development (McCauley and Stephens, 2012;Muafi et al., 2016).The impact will be able to grow and develop a sustainable cluster (Allen and Ptiowsky, 2008).Nevertheless, it needs the support of various stakeholders, namely the government, organizations, the surrounding communities, and media (Hendriques and Sadorsky, 1999).All of these parties are expected, if they have a high commitment in implementing environmental management towards green area developments, to have an impact on improving the competitiveness of the region (Muafi, et al., 2016;Muafi, 2016).
In the scope of chrysanthemum SME organizations, Aragon-Correa (1998) explained that organizations should adopt a consistent and proactive environmental management.This is in line with the research results of Aragon-Corea (1998;Muafi et al., 2016;Muafi, 2016), which found that there is a relationship between the dimensions of natural environment approach and proactive business strategies in order for SMEs to produce high competitiveness.
Tandon (2012) stated that in green economy, there are three things that need to be considered by women, namely environments, socioeconomic development, and gender equality.As an empowered individual, women should be productive, responsible, and have a high involvement in environmental health management.According to Sudarwanto (2010), the great potentials women have can be developed in fields of maintenance, preservation of environments, and prevention of environmental pollution.This is because in addition to the great number of women around, it has also been proven that women have been able to overcome problems surrounding the environments.Up until now, women are not included in environmental management, whether in the aspects of access, participation, control, or benefits.Women are also less educated in terms of the knowledge about environmental management given.Women are only referred to as objects of users in household consumption, without being given any knowledge of the harms from the materials given to them, their families and the environments.
Research Implications
Women can have a strategic role in the improvement of labor productivity, family welfare, and green environments.Increased labor productivity and family welfare can be done through: (1) improving their competence (technical and administrative qualifications as well as family relationships), professionalism, work ethic and work motivation, entrepreneurship skills, and the ability to lead organizations, (2) women could serve as strategic partners, fighters, administrative experts, and agents of change, and (3) women can also develop in fields of maintenance, preservation of environments, and prevention of environmental pollution.(4) increasing access to capital/credit, market information, and networking markets, both in collaboration with banks, economic enterprises, and state-owned or private enterprises.One thing to keep in mind is that running a business cannot be separated from the support of the community, so that the business itself should have a positive impact on the population and the surrounding environments (Smith, 2007).The findings of this research provide important implications that the strategic role of women is very important in chrysanthemum-based engineering design on disaster affected areas.
Table 3 :
Values of cronbach's alpha and composite reliability
Table 4 :
R 2 Values of Endogenous Latents on Inner Model
Table 5 :
Q 2 predictive relevance Result
Table 6 :
Inner Model Test Result
|
2018-12-04T12:47:23.625Z
|
2017-11-21T00:00:00.000
|
{
"year": 2017,
"sha1": "695759c87286e34e887b76aa114d9363df4c33ad",
"oa_license": "CCBYNCSA",
"oa_url": "http://publisher.unimas.my/ojs/index.php/IJBS/article/download/533/477",
"oa_status": "GOLD",
"pdf_src": "ScienceParseMerged",
"pdf_hash": "0c3efc3909a740777a1df55e187a248ff80a8f36",
"s2fieldsofstudy": [
"Economics"
],
"extfieldsofstudy": [
"Economics"
]
}
|
270035960
|
pes2o/s2orc
|
v3-fos-license
|
Persistent Challenges: A Comprehensive Review of Persistent Postural-Perceptual Dizziness, Controversies, and Clinical Complexities
Persistent postural-perceptual dizziness (PPPD) is a chronic and disabling disorder characterized by persistent dizziness, unsteadiness, and imbalance. It often arises without an identifiable cause and is exacerbated by upright posture, active or passive movement, and exposure to moving or complex visual stimuli. This complex pathophysiology and the psychological dimensions of its symptomatology pose a significant challenge to clinicians. PPPD presents diagnostic challenges and a lack of standardized treatment options, underscoring the need for multidisciplinary approaches encompassing pharmacotherapy, vestibular rehabilitation, and psychological interventions for effective management. Bridging the gaps in understanding PPPD requires collaborative efforts across disciplines, emphasizing integrated research approaches and patient support networks to enhance care and improve outcomes. This review explores the challenges, controversies, and clinical complexities of PPPD, highlighting the importance of a patient-centered approach.
Introduction And Background
Persistent postural-perceptual dizziness (PPPD) is a functional, multifaceted medical condition characterized by a prolonged, often daily, sense of dizziness, unsteadiness, or imbalance [1,2].Most of the epidemiological data on PPPD are derived from studies on phobic postural vertigo, visual vertigo, and chronic subjective dizziness [1,2].The estimated prevalence of PPPD among patients with vestibular symptoms is 15-20%, and the estimated incidence following acute or episodic vestibular disorders is about 25% of patients [3].The average age of patients presenting for evaluation is in the mid-40s, and female patients are predominant [3].Only a minority of patients experience spontaneous resolution of symptoms [3].This condition goes beyond typical vertigo and is not exclusively triggered by specific head movements [1,2].The exploration of PPPD involves an extensive review that delves into various facets, encompassing diagnostic criteria, clinical presentations, and underlying pathophysiological mechanisms [1,2].Despite the existence of the Bárány Society criteria, a major factor adding to the complexity of PPPD is the lack of universally accepted diagnostic and treatment guidelines [3].This uncertainty has sparked debates within the medical community [3].Some doctors believe that stricter criteria are necessary to ensure accurate diagnosis, while others think a broader approach is better to capture the wide range of PPPD symptoms [3].This ongoing discussion makes it difficult to create a standardized way to diagnose PPPD, which can be confusing for those new to the topic [3].Moreover, understanding the pathophysiological basis of PPPD remains elusive, adding another layer of complexity [4].The interplay between vestibular, proprioceptive, and visual sensory systems and potential central nervous system contributions raises intricate questions [4].
Managing PPPD involves various challenges, often requiring a holistic approach that includes pharmacotherapy, vestibular rehabilitation, and psychological interventions [4][5][6].Additionally, integrating mental health aspects is essential, given the strong link between PPPD and conditions like anxiety and depression, which adds another layer of complexity to patient care [5,6].This review aims to provide a comprehensive overview of PPPD, incorporating the most recent literature to offer insights into the current state of knowledge on the topic [3][4][5][6].
PPPD is a multifaceted disorder characterized by persistent dizziness and a sensation of unsteadiness, often exacerbated by upright posture or movement [1,3].This condition can be debilitating, affecting daily activities and overall quality of life [1,3].The criteria for diagnosing PPPD, as established by the Bárány Society and the International Classification of Vestibular Disorders, include a history of persistent nonvertiginous dizziness lasting three months or more, exacerbated by upright posture or movement and accompanied by a heightened visual dependency or postural instability [3].Patients may also exhibit symptoms such as generalized anxiety, avoidance of situations that provoke dizziness, and difficulty concentrating [3].The diagnosis of PPPD is clinical, based on thorough history-taking and examination to exclude other vestibular or medical conditions [3,7].Understanding these criteria is essential for healthcare professionals to accurately diagnose and manage PPPD, ultimately improving patient outcomes and quality of life [3,7].
In the late 19th century, German physicians described syndromes of dizziness and discomfort in motionrich environments, associated with autonomic arousal, anxiety, and avoidance behaviors [3].These included Platzschwindel (vertigo in a plaza or square), focusing on neuro-ophthalmologic processes, and Platzangst (fear in a plaza or square), emphasizing psychological origins [3].Another syndrome, Die Agoraphobie (fear of the marketplace), linked postural control, locomotion, and spatial orientation [3].These syndromes were debated as neurologic or psychiatric, but eventually faded from use, and agoraphobia became a psychiatric disorder without its original space and motion context [3].Around the same time, space-motion discomfort (SMD) was described as a condition marked by uneasiness about spatial orientation and heightened sensitivity to motion stimuli, particularly in visually rich environments [3].Visual vertigo (VV) was identified as a phenomenon in patients recovering from vestibular losses, manifesting as unsteadiness or dizziness triggered by complex visual stimuli, later termed visually induced dizziness (VID) [3].Additionally, chronic subjective dizziness (CSD) was defined as persistent non-vertiginous dizziness or unsteadiness, increased sensitivity to motion, and difficulty maintaining visual focus [3].These syndromes provided insights into the complex interactions between vestibular function, visual stimuli, and psychological factors in the experience of dizziness [3].
The Bárány Society criteria for PPPD are listed in Table 1.All five criteria must be fulfilled to make the diagnosis [3].The disorder is precipitated by conditions that cause vertigo, unsteadiness, dizziness, or problems with balance including acute, episodic, or chronic vestibular syndromes, other neurologic or medical illnesses, or psychological distress.(1) When the precipitant is an acute or episodic condition, symptoms settle into the pattern of criterion A as the precipitant resolves, but they may occur intermittently at first and then consolidate into a persistent course.(2) When the precipitant is a chronic syndrome, symptoms may develop slowly at first and worsen gradually.
Symptoms cause significant distress or functional impairment.
Symptoms are not better accounted for by another disease or disorder.
Pathophysiology and clinical features
The pathophysiology of PPPD is not fully understood, but it is believed to involve complex interactions between the sensory systems and central nervous system [1,4].It encompasses dysfunction in the integration of visual, vestibular (related to balance and spatial orientation), and proprioceptive (awareness of body position) information [1,4].
Sensory Mismatch
The concept of sensory mismatch is a crucial aspect in understanding the pathophysiology of PPPD because it is often associated with the information received from these sensory systems [4].This condition arises when there is a conflict between visual, vestibular, and proprioceptive inputs, leading to persistent dizziness and unsteadiness [4,7].The brain struggles to integrate these conflicting signals, which can result in a constant feeling of motion or imbalance, even in the absence of actual movement [4,5,7].Conflicting information between visual input and signals from the vestibular system can occur, especially during head movements or changes in posture [5].Patients often experience increased symptoms in environments with complex visual stimuli, such as crowded places or while using screens [4,5].This sensory overload can exacerbate the condition, making everyday activities challenging [4,5].Normally, the brain can adapt to short-term sensory conflicts, and the mismatch is resolved through central processing and integration [4,5].In PPPD, there may be a breakdown in this adaptive process, leading to a chronic and maladaptive response to sensory discrepancies [8].
Central Sensitization
It is a process in which the central nervous system becomes hypersensitive to stimuli, resulting in an exaggerated response to sensory input [5,9].This phenomenon is associated with various chronic pain and neurological conditions, and it plays a significant role in the pathophysiology of PPPD [9].Emotional factors, such as anxiety or stress, can contribute to this central sensitization [9].In PPPD, the interplay between central sensitization and psychological factors can create a cycle where heightened emotional responses contribute to increased symptom severity and vice versa [9].
In a study comparing PPPD patients and healthy controls, vestibular stimulation was administered using a motorized rotary chair [10].The findings indicated that PPPD patients have a reduced vestibulo-perceptual threshold, leading to increased motion sensitivity and a concomitant increase in vegetative responses, both of which are related to the duration of the disease [10].This threshold may be influenced by psychological factors, such as anxiety [10].Thus, the vestibular system appears to be susceptible to modulation, which is attributed to cortical-subcortical hyperexcitability, reduced sensory feedback, and a deficit in habituation, reflecting chronic maladaptation [10].A decreased motion perception threshold, combined with abnormal vestibular responsiveness in PPPD patients, leads to motion intolerance and the induction of dizziness when exposed to movement [10].
Therapeutic approaches often include interventions aimed at desensitizing the central nervous system, reducing hypersensitivity, and promoting adaptive changes in neural processing [11].Understanding the complex interplay between central sensitization and sensory processing is essential for developing effective treatment strategies for PPPD [12].A comprehensive and multidisciplinary approach that addresses both the physical and psychological aspects of the condition is often recommended for managing PPPD [12].
Adaptation Mechanisms
Under normal circumstances, the brain is adept at adapting to changes in sensory input to maintain balance.For example, when a person moves their head, the visual and vestibular systems work together to provide a stable perception of the environment [13].Maladaptive changes in how the brain processes sensory information may occur, leading to a persistent perception of dizziness even when there is no ongoing pathology [14].The brain's attempts to adapt to perceived threats or imbalances might contribute to the chronic nature of PPPD [14].
Neurotransmitter Involvement
The role of neurotransmitters is an area of ongoing research, and while the exact mechanisms are not fully understood, there is evidence suggesting that neurotransmitter dysfunction may contribute to the development and persistence of symptoms in PPPD, for example, those involved in mood regulation and sensory processing may contribute to the development or perpetuation of symptoms [14].
It is crucial to acknowledge that research on PPPD is continuously evolving and the precise mechanisms involved are likely to become clearer with the advancement of scientific understanding [1,4].The complex nature of PPPD, which involves both physiological and psychological components, underscores the necessity for a comprehensive approach in its diagnosis and management [1,4].
Individuals with PPPD experience a continuous and chronic sensation of dizziness or unsteadiness, often occurring daily [2,4].This sensation is not solely triggered by specific head movements; it could result from an acute vestibular disorder, a stressful life event, or a neurologic injury and could also be worsened by an upright posture, passive movement, and exposure to complex visual stimuli [2,14].PPPD may be precipitated by vestibular disorders, medical illnesses, or psychological distress that cause vertigo, dizziness, or unsteadiness [3].Also, this condition is frequently associated with psychological factors such as anxiety and depression [15].Emotional responses can influence the severity and persistence of symptoms, and these conditions often coexist [15].This association has a significant impact on an individual's quality of life, leading to functional impairment, reduced daily activities, and increased healthcare utilization [16].These key factors support that the treatment of PPPD often requires a holistic approach, addressing both physiological and psychological aspects [16], which is crucial for healthcare professionals to navigate the complexities and develop effective strategies for diagnosis and management [8,14].
Controversies
Major controversies surrounding PPPD primarily revolve around the duration of symptoms, overlap with other conditions, and treatment approaches.
Duration of Symptoms
Defining the duration necessary for a diagnosis of PPPD remains a point of contention [1][2][3].Some researchers and clinicians might have differing opinions on the appropriate time frame for categorizing symptoms as persistent.They argue for a specific time frame, while others emphasize the importance of recognizing the chronic nature of the condition, irrespective of a predefined duration [1][2][3].
Overlap with Other Conditions
There is an ongoing discussion about the potential overlap between PPPD and other vestibular disorders, migraines, and psychological conditions [17].Distinguishing PPPD from these related conditions poses a challenge in both research and clinical settings [17].Studies have provided evidence of an association between PPPD and migraine headaches [18].Sarna et al. demonstrated that most patients with PPPD meet many of the criteria for migraine headaches [18].They found a 53% prevalence of migraine headaches among individuals with PPPD, compared to the 8-13% prevalence of migraine headaches in the general population [18].This significant difference suggests a strong association between PPPD and migraine headaches [18].Migraine and PPPD share several characteristics, such as being attributed to cortical disturbances, manifesting hypersensitivity to various sensory stimuli (including visual motion, motion, light, sounds, and smells), and exhibiting a female predominance [18].Therefore, identifying the coexistence of both disorders is crucial, as treating the migraine may reduce the severity of symptoms in PPPD [18].
Anxiety disorders, such as generalized anxiety disorder and panic disorder, often coexist with PPPD [19].There are ongoing discussions regarding whether these factors are causative or secondary to persistent dizziness [19].The relationship is bidirectional, with anxiety potentially worsening dizziness symptoms and chronic dizziness contributing to heightened anxiety [20].Anxiety disorders can manifest with persistent postural dizziness in various ways [20].One manifestation involves triggering episodes of dizziness and unsteadiness crises [20].This is due to the heightened emotional state associated with these disorders, which amplifies the symptoms [20].Additionally, living with persistent dizziness and a sense of imbalance can be distressing [20].It may lead individuals to develop anxiety as a reaction to their ongoing symptoms and the impact these symptoms have on their daily lives [20].Also, both PPPD and anxiety can involve the dysregulation of the autonomic nervous system and other neurological processes [4].Shared neurobiological factors may contribute to the co-occurrence of these conditions [4].
Treatment Approaches
There isn't necessarily controversy surrounding the treatment of PPPD; there may be ongoing discussions and debates about the most effective strategies [21,22].Since vestibular rehabilitation, which involves exercises to improve balance and reduce dizziness, is a main component of PPPD treatment, controversies may involve the optimal duration, intensity, and specific exercises used in vestibular rehabilitation for PPPD [21,22].Pharmacotherapy agents, such as selective serotonin reuptake inhibitors (SSRIs) or vestibular suppressants, may be considered in the management of PPPD [23]; some dissension involves the potential benefits, risks, and appropriate duration of medication use [23].Addressing these controversies is crucial for refining diagnostic criteria, improving treatment outcomes, and advancing our understanding of PPPD [23].Ongoing research aims to unravel these complexities, providing a clearer framework for clinicians and researchers working with individuals affected by this challenging condition [21][22][23].
Challenges with Medications
Medications, such as vestibular suppressants or anxiolytics, may provide only partial relief, and their efficacy varies among individuals with PPPD [21][22][23].Achieving consistent and significant improvement can be challenging.Also, many medications prescribed for PPPD come with potential side effects, ranging from drowsiness to cognitive impairment [23].Balancing the benefits and drawbacks of medications poses a challenge, particularly when side effects impact daily functioning [23].
Individual responses to medications exhibit a high degree of variability, making it imperative to recognize that the efficacy of a particular treatment for one individual may not translate to effectiveness for another [24].The identification of the optimal medication and dosage often necessitates a meticulous trial-and-error approach [24].This method, while essential for achieving therapeutic success, can contribute to the lengthening of the overall treatment duration [24].
Pharmacotherapeutic strategies targeting anxiety or depression, frequently concomitant with PPPD, are often successful in addressing psychological manifestations but may not directly ameliorate the underlying vestibular-related issues [1,23].The synergistic implementation of psychological and pharmacological interventions presents intricate challenges in achieving optimal therapeutic outcomes.
Challenges with Vestibular Rehabilitation
Vestibular rehabilitation emerges as a pivotal component in the multifaceted management of PPPD [25].This therapeutic modality assumes a central role in addressing the intricate interplay of vestibular dysfunctions contributing to the chronic dizziness characteristic of PPPD [25].The efficacy of vestibular rehabilitation hinges on unwavering commitment and disciplined, consistent practice [26,27].Adherence to prescribed exercises poses a notable challenge, particularly during the initial phases when individuals may contend with discomfort or perceive a gradual pace of improvement [26].Clinicians navigating the landscape of PPPD therapeutics must underscore the imperative of sustained engagement in vestibular rehabilitation to optimize its impact on vestibular function and mitigate the enduring symptomatology associated with this complex condition [27].
Responses to vestibular rehabilitation vary widely among individuals.Some may experience significant improvement, while others may find only modest relief [27].Determining the most effective exercises for each person remains a complex aspect of treatment [27].Individuals with PPPD may have heightened sensitivity to certain movements, making it challenging to engage in exercises that induce dizziness or discomfort.Tailoring rehabilitation programs to accommodate individual tolerances is crucial but can be intricate [27].
PPPD often involves both physiological and psychological components.Coordinating vestibular rehabilitation with other therapeutic modalities, such as cognitive-behavioral therapy (CBT), poses challenges in ensuring a comprehensive and integrated approach [27,28].In cases where PPPD coexists with other vestibular disorders or medical conditions, tailoring vestibular rehabilitation becomes more complex [27,28].Addressing multiple contributing factors requires a nuanced and individualized treatment plan [27,28].Navigating these challenges underscores the importance of a personalized and multidisciplinary approach to PPPD treatment, where healthcare professionals collaborate to address the diverse aspects of this complex condition.Integrating medications, CBT, and vestibular rehabilitation, with a focus on individual needs and tolerances, is crucial for optimizing outcomes [27,28].A summary of PPPD challenges is seen in Figure 1. between the physical symptoms and psychological aspects of the disorder [29].While CBT is recognized for its effectiveness in managing anxiety and depression, which are commonly associated with PPPD, translating these benefits to the physical symptoms of dizziness and imbalance can be complex [29].Patients may experience difficulty in engaging with CBT due to the pervasive nature of their physical symptoms, which can affect concentration and the ability to participate in therapy sessions [25,29].Additionally, the somatic focus of PPPD symptoms might require therapists to adapt traditional CBT techniques to address the unique needs of this population, integrating specific strategies aimed at reducing symptom-related anxiety and improving coping mechanisms [29].The lack of widespread understanding and recognition of PPPD in the broader healthcare community can also limit access to appropriately trained therapists, further complicating the treatment landscape [29].Thus, while CBT holds potential as a component of a comprehensive treatment plan for PPPD, overcoming these challenges is essential to maximize its efficacy and improve patient outcomes [29].
Functional Impairment
The chronic nature of PPPD can have a profound impact on an individual's daily life, restricting their capacity to engage in routine activities [16].Functional impairment may extend to various aspects, affecting work, social interactions, and recreational pursuits.Individuals with PPPD commonly endure persistent, non-vertiginous dizziness exacerbated by upright posture and head movements [6].This ongoing sensation of unsteadiness, imbalance, and fear of falling can significantly hinder a person's ability to carry out daily activities that demand balance and coordination [30,31].The consequences of these symptoms highlight the substantial challenges faced by individuals living with PPPD in their day-to-day lives [30,31].
Reduced Quality of Life
The combined impact of the mentioned factors significantly diminishes the overall quality of life for individuals dealing with PPPD [30].Beyond affecting specific activities, functional impairment extends into various aspects of daily life [30].The persistent symptoms of PPPD cast a pervasive influence, leaving an impact on emotional well-being, imposing physical limitations, and intricately influencing social relationships [30].Managing chronic, non-vertiginous dizziness intensified by upright posture and head movements can create a sustained sense of unease, affecting mental health [1,30,31].Additionally, the intricate interplay of PPPD symptoms with social relationships compounds the challenges, potentially fostering a sense of isolation and hindering the ability to fully participate in interpersonal connections [1,30,31].Collectively, these diverse aspects contribute to a nuanced and comprehensive understanding of the reduced overall sense of well-being experienced by individuals with PPPD in daily life [6,30].Understanding the multifaceted impact of PPPD on quality of life is essential for healthcare professionals to develop comprehensive treatment plans that address not only the physical symptoms but also the emotional and social aspects of the condition [6,30].Holistic care that considers the individual's overall well-being is key to improving their quality of life [6,30].
Treatment Response Variability
Responses to treatment, whether pharmacological, rehabilitative, or psychological, can vary widely among individuals.Predicting and ensuring a positive response to interventions remains a challenge in managing PPPD effectively [31].Recognizing these challenges is crucial for healthcare professionals to develop tailored approaches that address the diverse clinical manifestations and limitations associated with PPPD [31].A comprehensive understanding of the disorder's impact is essential for providing patient-centered care and improving outcomes [31].
Balancing Work, Health, and Finances
Those affected by PPPD face a broad question: to find the stable work performance paralleled by potential complications of concentration and a necessity to make many breaks [14].PPPD and its accompanying struggle will on most occasions significantly increase the utilization of healthcare services as this is done to diagnose as well as curb the pains, necessitating increased consultations and interventions [1,6].As a result, the overall combination of medical costs, possible risks at work, and the necessity for alternative therapies results in much economic strain not only on an individual but also on his or her family members [30].
Integrated Research Approaches
Promoting interdisciplinary research that integrates neurology, otolaryngology, psychiatry, and rehabilitation sciences is imperative for advancing our understanding of PPPD [32].The collaboration of experts from these diverse disciplines contributes nuanced perspectives to elucidate the multifaceted nature of PPPD comprehensively [32].An interdisciplinary approach facilitates a holistic exploration of PPPD, considering not only its physiological underpinnings but also its psychological and rehabilitative aspects [32].This collaborative synergy of expertise is crucial for deciphering the complexities of PPPD and holds the potential to generate innovative, patient-centered interventions [32].Encouraging interdisciplinary research is instrumental in fostering transformative breakthroughs that not only enhance our understanding of PPPD but also elevate the standard of care provided to individuals contending with this challenging condition [32].
Longitudinal Studies
Advocating for ongoing, in-depth studies that follow individuals with this condition over the long term is crucial for truly understanding how this condition unfolds and evolves [33].Longitudinal studies, characterized by their extended temporal scope, afford a distinctive opportunity to capture the dynamic temporal features of PPPD, elucidating its manifestation patterns, fluctuations, and potential resolutions or transformations [33].Through methodical data collection over an extended temporal horizon, researchers can discern nuanced patterns, identify triggering factors, and delineate diverse trajectories of PPPD [33].Committing resources to longitudinal research not only contributes to the scientific elucidation of PPPD but also pledges advancements in patient care by offering evidence-based insights that can guide bespoke interventions and management [33].
Treatment Optimization Trials
Conducting randomized controlled trials to optimize and compare various treatment modalities, including medications, vestibular rehabilitation, and psychological interventions, can help establish evidence-based guidelines for effective PPPD management [23].
Telemedicine Solutions, Educational Initiatives, and Patient Support Networks
Exploring and advancing telemedicine solutions for PPPD facilitates remote monitoring, consultations, and follow-ups, thereby addressing challenges related to geographical and mobility constraints [34].This initiative not only enhances patient access to specialized care but also fosters a more patient-centric and inclusive healthcare environment [33].Concurrently, launching educational campaigns is pivotal in raising awareness among healthcare professionals, patients, and the general public, leading to early recognition, reduced stigma, and increased support for individuals grappling with PPPD [34].Moreover, fostering international collaboration in PPPD research is imperative, as it facilitates the exchange of diverse perspectives, methodologies, and findings, ultimately accelerating progress in comprehending and managing this intricate disorder on a global scale [34].The establishment and promotion of support networks for individuals with PPPD, including online communities and educational resources, contribute significantly to coping strategies and shared experiences, fostering a sense of community among those affected [34].This integrated approach reflects a commitment to advancing PPPD care through innovation, education, collaboration, and support [34].
Increased Recognition and Research Focus
In recent years, there has been a notable increase in the recognition and research focus on PPPD as a distinct clinical entity [17].This surge in attention reflects a growing awareness of the significant impact PPPD has on patients' lives and the need for more targeted treatments [17].Researchers have directed their efforts toward understanding various aspects of PPPD, including its prevalence, clinical characteristics, and underlying mechanisms [3,17].
Studies investigating the prevalence of PPPD have revealed its relatively high occurrence, highlighting the importance of addressing this condition in clinical practice [17].Additionally, research has sought to identify the clinical features that distinguish PPPD from other vestibular and psychological disorders, aiding in more accurate diagnosis and treatment planning [3,17].Moreover, efforts to uncover the underlying mechanisms of PPPD have shed light on the complex interplay between sensory inputs, central processing, and psychological factors [3,17].This deeper understanding is crucial for developing effective therapeutic strategies that target the root causes of PPPD [3,17].
Neuroimaging Studies
Advances in neuroimaging techniques have revolutionized researchers to explore structural and functional brain changes associated with PPPD [32].Functional magnetic resonance imaging (fMRI) has been instrumental in uncovering structural changes in the brains of individuals with PPPD [33].Researchers can identify alterations in specific regions, such as the vestibular system, cerebellum, and cortical areas, which play crucial roles in spatial orientation and postural control [34].The findings from these studies suggest that brain regions in individuals with PPPD may not be as active or as well-connected as in healthy individuals [33].This diminished activity and connectivity could lead to poorly integrated mechanisms for posture and gaze control [33,35].Investigators using fMRI in patients with PPPD have observed decreased functional connectivity in the temporal lobe but enhanced connectivity in the occipital lobe [35].These findings indicate that the interaction between the visual cortex and the vestibular cortex is abnormal, with this interaction being predominantly driven by visual information [36].This abnormality explains the impaired ability of individuals with PPPD to adjust posture and movement using both vestibular and visual information [35,36].It also accounts for their visual dependence and the dizziness experienced following exposure to complex visual stimuli [33,35].Also, there have been described some changes in brain function registered by magnetoencephalography (MEG) in a study where MEG recordings from patients with PPPD and control subjects were analyzed [36].The frequency-dependent alterations in neuromagnetic activity were observed in patients with PPPD compared with healthy controls [36].
PPPD represents a challenging frontier in the field of vestibular medicine, characterized by its multifaceted symptomatology, complex pathophysiology, and significant impact on patients' quality of life [1].Despite advancements in understanding its clinical presentation and associations with other conditions such as migraines and anxiety disorders, PPPD remains an enigmatic condition [3,5].Emerging research, particularly in neuroimaging, offers promising insights into PPPD's neurobiological basis, potentially paving the way for more targeted therapies [34].However, the variability in treatment responses and the ongoing challenges over treatment and rehabilitation highlight the need for continued investigation and collaboration across specialties [8].Future research should focus on longitudinal studies to understand PPPD's progression and on optimizing treatment strategies through randomized controlled trials.The management of PPPD, therefore, necessitates a patient-centered approach, recognizing the diversity of symptoms and their impact on individuals' lives.
Conclusions
PPPD is a complex condition that blurs the lines between physiological and psychological health, presenting significant challenges in diagnosis and treatment.The current body of research underscores the necessity of a multidisciplinary approach to manage PPPD effectively, integrating pharmacotherapy, vestibular rehabilitation, and psychological support tailored to individual patient needs.
One or more symptoms of dizziness, unsteadiness, or non-spinning vertigo are present on most days for three months or more.(1) Symptoms last for prolonged (hours-long) periods of time, but may wax and wane in severity.(2) Symptoms need not be present continuously throughout the entire day.Persistent symptoms occur without specific provocation, but are exacerbated by three factors: (1) upright posture, (2) active or passive motion without regard to direction or position, and (3) exposure to moving visual stimuli or complex visual patterns.
|
2024-05-26T15:27:30.591Z
|
2024-05-01T00:00:00.000
|
{
"year": 2024,
"sha1": "32f4feb29a65e8834bfeb742e492f38b185bdd62",
"oa_license": "CCBY",
"oa_url": "https://assets.cureus.com/uploads/review_article/pdf/253786/20240523-6008-l7fol3.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "e12aa708698325addb3cc7acc6e0caf312437bcc",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": []
}
|
211101956
|
pes2o/s2orc
|
v3-fos-license
|
Surface properties of SnO2 nanolayers prepared by spin-coating and thermal oxidation
In this work, comparative studies of the surface morphology and surface chemistry of SnO2 nanolayers prepared by spin coating with subsequent thermal oxidation (SCTO) in the temperature range of 400–700 °C using scanning electron microscopy (SEM), atomic force microscopy (AFM) and x-ray photoelectron spectroscopy (XPS) methods, are presented. The SEM images show that SCTO SnO2 nanolayers contain partly connected irregular structures strongly dependent on the final oxidation temperature, with interconnected single grains of longitudinal shape and size, resulting in a flatter surface morphology with respect to the commonly used three-dimensional (3D) SnO2 thin films. In turn, AFM studies additionally confirm that SCTO SnO2 nanolayers after post-oxidation annealing at higher temperatures contain isolated grains of average lateral dimensions in the range of 20–50 nm having a rather flat surface morphology of average surface roughness defined by the root mean square factor at the level of ∼2 nm. From the XPS experimental research it can be concluded that, for our SCTO SnO2 samples, a slight surface nonstoichiometry defined by the relative [O]/[Sn] concentration at the level of 1.8–1.9 is observed, also depending on the final post-oxidation temperature, being an evident contradiction to recently published literature using x-ray diffraction data. Moreover, XPS experiments show that there is also a permanent small amount of carbon contamination present at the surface of internal grains of our SCTO SnO2 nanolayers, creating an undesired potential barrier for interactions with gaseous species when they are used as the active materials for gas sensing devices.
Introduction
Tin dioxide (SnO 2 ) is a wide band gap (3.6 eV) n-type semiconductor with a rutile structure [1] that has fascinating physicochemical properties including, among others, a high electrical conductivity (∼10 2 W −1 ×cm −1 ) with the natural Nanotechnology Nanotechnology 31 (2020) 315714 (7pp) https://doi.org/10.1088/1361-6528/ab7586 Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. tendency to variability after exposure to various gases [2,3]. This is why SnO 2 , mainly in the form of thick and thin films, has been applied in resistivity-type gas sensor devices [4,5] for environmental and medical applications [6].
In the last decade, research has been focusing on low dimensional SnO 2 nanostructures, including among others nanowires and nanobelts, due to their enlarged surface-tovolume ratio as well as enhanced chemical stability and electrical performance [7,8]. However, thin film technology is a core high-yield fabrication method for real-world sensors because of its main advantages such as low power consumption. There are various physical and chemical techniques that have been implemented recently for the preparation of SnO 2 thin films, which were comprehensively reviewed in [5,6]. Depending on the deposition method, post annealing or manipulation of the substrate temperature and gas pressure can be used to control the size of the crystallites and to intentionally obtain the desired and optimized morphology. This is extremely important as the shape and size of SnO 2 nanostructures have a significant influence on their gas sensing properties [9,10].
Within the physical deposition methods, a great tendency has appeared in the last several years towards developing lower temperature and inexpensive methods for deposition of SnO 2 nanostructured thin films. Apart from the well-known sol-gel (SG) [11] and spray pirolysis (SP) [12] methods, one of the most promising low temperature technologies for preparation of SnO 2 nanolayers is a method of spin coating deposition of specific precursors on Si substrate proposed by Cukrov et al [13] and then further developed by Bazargan et al [14], Khuspe et al [15] and Uysala et al [16].
Bazargan et al [14] recently observed that using the SnCl 4 solution and performing spin-coating deposition on etched glass substrates combined with post-deposition annealing (oxidation) in an oxygen atmosphere in the temperature range of 350-500°C resulted in the formation of uniform, flat granular SnO 2 thin films containing monodisperse crystallites with sizes in the range of 7-10 nm and having a low root mean square (RMS) surface roughness (1.6-2.2 nm). Moreover, it was observed that after postannealing in an oxygen atmosphere at temperatures above 500°C the surface roughness increased since the RMS factor was evidently higher (∼6), whereas the dimensions of the more isolated crystallites increased up to ∼25 nm.
In turn, Khuspe et al [15] combined the sol-gel technique for the preparation of homogeneous solution containing SnO 2 powder with its subsequent spin coating deposition on glass substrate for the preparation of SnO 2 nanostructured thin films also containing tetragonal nanocrystallites with dimensions in the range of 5-10 nm for NO 2 gas sensing application.
A similar sol-gel spin coating deposition procedure combined with additional post-annealing in air in the temperature range 450-650°C was applied by Uysala et al [16] for the preparation of SnO 2 nanostructured thin films for potential photovoltaic applications.
However, it can be noticed that the issue of the local surface chemistry of SnO 2 nanolayers (namely the surface nonstoichiometry together with undesired carbon C contamination commonly adsorbed at the surface of various SnO 2 forms from the air atmosphere), has been rather neglected in the literature undertaking the subject of spin coated SnO 2 and listed above. This is of great importance for SnO 2 gas sensor performance, mainly for the gas sensitivity as well gas sensor aging effects, as observed in our recent studies of L-CVD SnO 2 nanolayers [17] and PVD SnO 2 nanowires [18].
Driven by these facts, in our last studies we have focused on the surface impact on the properties of SnO 2 nanolayers deposited by spin-coating and subsequent thermal oxidation (SCTO) based on the systematic comparative scanning electron microscopy (SEM) and atomic force microscopy (AFM) studies of their surface morphology, including grain dimension and roughness, combined with the x-ray photoelectron spectroscopy (XPS) studies of their surface chemistry (purity and stoichiometry) in view of their potential application in novel types of conductometric gas sensor devices.
Experimental
The SnO 2 nanolayers used in our studies have been deposited by the spin-coating method using SnCl 4 ·5H 2 O in isopropanol at the rotation speed of 1800 rpm on Si(111) substrate recently cleaned (etched) in HCl solution in order to remove the natural oxide and then covered with an approximately 8 nm Au film to improve the stability, adhesion to substrate and deposited thin film homogeneity. At the second step, after a short period of drying in dry air at 100°C, an additional thermal oxidation of the above mentioned deposited layer was performed for 1 h at various temperatures in the range of 400-700°C in a dry air atmosphere inside a reaction chamber of a typical diffusion furnace. The thickness of our SCTO SnO 2 was estimated to be about ∼200 nm, and does not significantly evolve during above mentioned post-oxidation procedure. Other experimental details can be found in [19].
The local surface morphology of our SCTO SnO 2 samples was controlled using the SEM method at Brescia University (Italy), with a Zeiss LEO 1530 Model SEM microscope. More experimental details regarding this method can be found in [18]. For a more quantitative analysis of the surface morphology of our SCTO SnO 2 samples including the shape of individual grains, the AFM method was additionally applied using the XE-70 Park model working in a non-contact mode. In turn, the surface chemistry of SCTO SnO 2 thin films was controlled by the XPS method using the SPECS XPS spectrometer equipped with x-ray Al Kα source at photon energy 1486,6 eV (XR-50 model, and a concentric hemispherical analyzer (CHA PHOIBOS 100 model). Other experimental details can be found elsewhere in [17,18].
Results and discussion
SEM and AFM investigations of SCTO SnO 2 thin films showed that their extremely complex morphological landscape is strongly dependent on the temperature of thermal post-oxidation during the second step of the preparation procedure.
The respective SEM images of our SCTO SnO 2 nanolayers post-oxidized (annealed) at various temperatures in the range of 400-700°C are shown in figure 1.
From the respective SEM images one can observe, that for the lower post-oxidation temperatures (below 500°C) SCTO SnO 2 nanolayers exhibit evidently continuous surface morphology, as observed recently by Bazargan et al [14], without visible single crystalline forms.
What is the most important, as our SEM studies confirmed, SCTO SnO 2 nanolayers after post-oxidation at temperatures higher than 500°C exhibit a well separated grain structure. In our previous paper [19], undertaking XRD phase analysis of SCTO SnO 2 nanolayers, it was shown that the samples demonstrated an irregular crystalline structure on the (110) facets and the Debye-Scherrer formula showed that the average sizes of the interconnected individual grain-type SnO 2 crystallites in the SCTO SnO 2 nanolayers after post-oxidation at temperatures of 500 and 700°C were 5.1 and 6.7 nm, respectively.
In addition to the above, for the SCTO SnO 2 nanolayers after post-oxidation at the highest temperature of 700°C, one can observe from the respective SEM image that the grain's shape becomes more longitudinal as the average width and length of a single grain are at the level of ∼20 nm and 50 nm, respectively. In relation to the above, it appears that our results of SEM characterization slightly differ from the SEM experiments of Bazargan et al [14], who observed almost isolated crystalline grains of the average lateral dimension ∼25 nm. However, there are several reasons for these differences beginning with the temperature range of the postoxidation procedure, the various lateral resolutions of respective SEM images, different surface preparation procedures of the Si substrate, various oxidizing atmospheres of the post-oxidation procedure and finally the respective flow rates.
As was mentioned above, for additional verification of the shape of individual grains of our SCTO SnO 2 nanolayers, AFM comparative studies were additionally performed. Figure 2 shows the 3D AFM image of the internal, local structure of a SCTO SnO 2 nanolayer after post-oxidation at the highest temperature of 700°C, together with the corresponding AFM profile.
As can be seen, the AFM experiments confirm that the nanograins of the formed SCTO SnO 2 nanolayers are more longitudinal with respect to the experiments of Bazargan et al [14] with an average maximum height below 10 nm, and average lateral dimension in the range of 20-50 nm.
However, at the same time, one can conclude that AFM imaging in the case of the presented SnO 2 nanolayers is moderately corrugated, which can be related to the influence of the tip convolution to a large extent. Moreover, the discussed SCTO SnO 2 nanolayers are rather flat as in our case the RMS factor, being the measure of the average surface roughness is at the level of ∼2 nm being evidently smaller (∼3 times) than the one presented by Bazargan et al [14]. The above conclusion that our SCTO SnO 2 nanolayers exhibit a very flat surface morphology in comparison to the commonly used SnO 2 thin films, as reviewed by Eranna [7], is of great importance from the point of view of possible applications in photovoltaics as transparent conductive electrodes. Furthermore, it should be emphasized that our SCTO SnO 2 nanolayers after post-oxidation at temperatures higher than 500°C contain irregular but isolated grains of average lateral dimension not far from the Debye length (∼several nm), for which the highest sensitivity of metal oxide gas sensor materials are commonly observed [7,8]. From this point of view the SCTO SnO 2 nanolayers would be highly promising candidates for potential gas sensing applications.
In parallel to the surface morphology, in this subsection the main results of the XPS studies of SCTO SnO 2 nanolayers after deposition are presented and analyzed.
For all the XPS survey spectra of SCTO SnO 2 nanolayers for different oxidation temperatures in the rangeof 500-700°C, which looked very similar, the contribution of the main O1s, single peaks, and the double Sn3d 3/2 and Sn3d 5/2 peaks corresponding to the two main elements were observed.
Moreover, for all XPS survey spectra, an evident contribution of single XPS C1s peak was also observed, that confirmed the existence of carbon C surface contamination on the surface of our SCTO SnO 2 nanolayers.
For a more precise analysis of the surface chemistry of our SCTO SnO 2 samples, including their stoichiometry, the core level XPS O1s-Sn3d spectral windows were used, shown in figure 3. Taking into account the area under the principal components of XPS O1s and Sn3d 5/2 peaks and using the analytical procedure based on the atomic sensitivity factor (ASF) [20], the relative [O/Sn] atomic concentration was determined. A similar procedure based on XPS survey spectra in the binding energy range of 600-0 eV (not presented here) and the area under the respective principal components of C1s and Sn3d 5/2 was used in the determination of relative [C/Sn] atomic concentration. The obtained results for all our SCTO SnO 2 samples are summarized in table 1.
As shown in table 1, the XPS experiments confirm that the surface of all our SCTO SnO 2 nanolayers is slightly nonstoichiometric, with an evident domination of tin dioxide which is crucial, and we have to underline at this point, is that the above conclusion cannot be understood as contradictory to information obtained from our previous XRD experiments [19]. The XRD results presented in [19] clearly indicated only the presence of SnO 2 formed in the SCTO SnO 2 nanolayers under study. However, in contrast to the XRD method, one has to bear in mind that XPS is primarily a surface sensitive technique, especially when using the x-ray photon energy (1486.6 eV), able to recognize the surface chemical species in the very first few layers in depth only up to 10 nm. This is why the XPS method was used in our studies for the determination of effective surface nonstoichiometry of SCTO SnO 2 nanolayers, which is extremely important regarding the value of the Debye length for SnO 2 at the level of several nm [5,7].
In order to prove the above statement, the decomposition of XPS O1s and Sn3d 5/2 peaks for the sample with the highest [O]/[Sn] relative concentration was performed, as can be seen in figure 4.
Concerning the XPS O1s line ( figure 4), a simple visual shape confirms that it is wide, asymmetrical and exhibits an evident shoulder at the higher binding energy side of the spectrum. After decomposition (deconvolution) using the fitting procedure with the Gaussian distributions, it becomes evident that it contains two components separated by 1.4 eV corresponding to O atoms (ions) in bonding with various surface atoms (ions). The main parameters used in the applied fitting procedure, as well as the obtained best fitting parameters are summarized in table 2.
The XPS O1s line component at lower binding energy (531.0 eV) corresponds to O 2 ions in the Sn-O band (named lattice oxygen), whereas the second one at higher binding energy (532.4 eV) can be attributed to hydroxyl groups (H-O band) adsorbed at the surface. Their relative area (intensity) is at the level of ∼1.8. This is important, similar XPS O1s line components were recently observed by Mazloom et al [21] for Co-doped SnO 2 thin films, also prepared using the sol-gel spin coating technique.
Concerning the XPS Sn3d 5/2 line ( figure 4), a simple visual shape analysis also confirms that it is wide, slightly asymmetrical and exhibits a small shoulder at the lower binding energy side of the spectrum. After decomposition (deconvolution) using fitting with Gaussian distributions, it is evident that it contains two components separated by 2.3 eV corresponding to Sn atoms (ions) bonding with various surface atoms (ions). The main parameters used in the fitting procedure, as well as the obtained fitting parameters are also summarized in table 2.
The main component of XPS Sn3d 5/2 line at higher binding energy of 471.2 eV corresponds to Sn 4+ ions in Sn-O band (lattice oxygen), whereas the second one (very small) at lower binding energy (∼485.0 eV) can be attributed to the existence of Sn o bondings related to the small amount of metallic Sn. Their relative area (intensity) is in good agreement with the information on relative concentration [O]/[Sn] ratios obtained from the O1s-Sn3d spectral windows that the SCTO SnO 2 nanolayers under the last analysis are only slightly nonstoichiometric, with an evident domination of tin dioxide SnO 2 . Similar XPS O1s line components were also observed in our recent studies of RGVO SnO 2 nanolayers [22].
As was mentioned earlier, the main difference in the surface chemistry of our SCTO SnO 2 nanolayers, observed in the XPS survey spectra, is the amount of C contaminations.
In general, they come from the different C gaseous species present in the natural air atmosphere, which immediately adsorb at the surface of all semiconductor materials (time ∼ms) and are controlled by the mass spectrometry For the samples after post-oxidation at the lowest temperature of 500°C it is at the level of 2.65, slightly lower than for the freshly deposited L-CVD SnO 2 thin films after air exposure observed in our recent studies [23][24][25]. In turn, for the SCTO SnO 2 nanolayers after post-oxidation at the highest temperature of 700°C it is almost three times lower. This is probably related to the fact that they exhibit more tight (packaged) surface morphology corresponding to the greater dimensions of the interconnected individual nanograins and then the respective smaller channels between them are usually 'open' for potential undesired diffusion of C contaminations (mainly as CO 2 ) from the residual gas atmosphere. Crucially, the average distances between the nanograins in our SCTO SnO 2 nanolayers are evidently smaller with respect to the most 'open' surface observed for the L-CVD SnO 2 thin films, as well as the SnO 2 thin films obtained by the rheotaxial growth and thermal oxidation (RGTO) in our recent studies [26]. Importantly, those C contaminations at the surface of our SCTO SnO 2 nanolayers are in the form of C-OH bonding, as it was recognized after decomposition of XPS C1s peak shown in figure 5.
The same shape of this XPS C1s peak was observed for all the samples under our studies.
Already a simple visual shape analysis also confirms that it is wide and symmetrical.
After decomposition (deconvolution) of the XPS C1s peak using fitting with Gaussian distributions, it is evident that it contains only one component at the binding energy 285.5 eV, that can be attributed to the C-OH surface bonding commonly observed at various semiconductor surfaces including oxides [1,20,27]. Unfortunately, the undesired fast C adsorption on the surface of SnO 2 nanolayers (grains) is extremely critical for their gas sensor application because it strongly affects the response time of gas sensor devices. This is because every active (toxic) gas has to flow towards the gas sensitive active centers, for instance SnO 2 nanolayers, through the C contamination (up to ∼3 atomic layers in average), which generates an undesired and uncontrolled barrier for potential toxic gas adsorption at the internal surface of the sensor material.
Conclusions
In this paper the results of comparative studies of the surface morphology and surface chemistry of SCTO SnO 2 nanolayers using a combination of SEM, AFM and XPS methods are presented. SEM and AFM studies show that the specific surface morphology of SCTO SnO 2 nanolayers exhibit partly connected irregular structures with interconnected single grains of more longitudinal shape and size, resulting in flatter morphology corresponding to an average roughness (RMS) below 2 nm, as derived from the AFM studies, being very promising for their potential photovoltaic applications as transparent conductive electrodes. Moreover, SCTO SnO 2 thin films contain almost isolated longitudinal grains of average width and length at the level of ∼20 nm and 50 nm, respectively, not so far from the Debye length (∼several nm). From this point of view, they can be considered as a promising novel form of metal oxide material for potential applications in novel types of conductometric gas sensors.
In turn, XPS studies confirm that for SCTO SnO 2 samples, a slight surface nonstoichiometry at the level of 1.8-1.9 is observed, together with C contamination, probably at the surface of internal grains. This undesired effect cannot be ignored because it generates an uncontrolled barrier for the potential adsorption of interacting gases at the internal surface of the sensor material. This is why, this undesired and uncontrolled C contamination at the internal surface of SCTO SnO 2 thin films appears to be the most important limitation for application in novel sensor devices. In relation to this, an understanding of the adsorption/desorption behavior of C contamination at the surface of SCTO SnO 2 nanolayers is crucial for the interpretation of the gas sensing mechanism. Such studies, also in comparison to the various 1D forms of SnO 2 , including nanowires and nanobelts, are currently in progress in our labs.
|
2020-02-14T14:05:59.487Z
|
2020-02-12T00:00:00.000
|
{
"year": 2020,
"sha1": "09c0563ce81e4ec6a622f603957cb273ac1c1234",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.1088/1361-6528/ab7586",
"oa_status": "HYBRID",
"pdf_src": "IOP",
"pdf_hash": "920d13ab0d951ef1cb293c2dcbeaa0445e5c368b",
"s2fieldsofstudy": [
"Materials Science"
],
"extfieldsofstudy": [
"Physics",
"Materials Science",
"Medicine"
]
}
|
6642724
|
pes2o/s2orc
|
v3-fos-license
|
The lipoprotein lipase gene in combined hyperlipidemia: evidence of a protective allele depletion
Background Lipoprotein Lipase (LPL), a key enzyme in lipid metabolism, catalyzes the hydrolysis of triglycerides (TG) from TG-rich lipoproteins, and serves a bridging function that enhances the cellular uptake of lipoproteins. Abnormalities in LPL function are associated with pathophysiological conditions, including familial combined hyperlipidemia (FCH). Whereas two LPL susceptibility alleles were found to co-segregate in a few FCH kindred, a role for common, protective alleles remains unexplored. The LPL Ser447Stop (S447X) allele is associated with anti-atherogenic lipid profiles and a modest reduction in risk for coronary disease. We hypothesize that significant depletion of the 447X allele exists in combined hyperlipidemia cases versus controls. A case-control design was employed. The polymorphism was assessed by restriction assay in 212 cases and 161 controls. Genotypic, allelic, and phenotypic associations were examined. Results We found evidence of significant allelic (447Xcontrol: 0.130 vs. 447Xcase: 0.031, χ2 = 29.085; 1df; p < 0.001) and genotypic association (SS: 0.745 vs. 0.939, and SX+XX: 0.255 vs. 0.061) in controls and cases, respectively (χ2 = 26.09; 1df; p < 0.001). In cases, depletion of the 447X allele is associated with a significant elevation in very-low-density lipoprotein cholesterol (VLDL-C, p = 0.045). Consonant with previous studies of this polymorphism, regression models predict that carriers of the 447X allele displayed significantly lower TG, low-density lipoprotein cholesterol (LDL-C) and TG/high-density lipoprotein cholesterol (HDL-C) ratio. Conclusion These findings suggest a role for the S447X polymorphism in combined hyperlipidemia and demonstrate the importance of evaluating both susceptibility and protective genetic risk factors.
Background
Abnormalities of lipoprotein metabolism play an important role in the development of atherosclerosis [1,2], and was one of the earliest established biochemical links to cardiovascular disease [3]. It is well established that elevated plasma triglycerides (TG) are an independent risk factor for coronary heart disease (CHD), and not simply a marker of an inverse relationship with high density lipoprotein-cholesterol (HDL-C), also an independent risk factor for disease [4]. The relative risk of CHD associated with a 1-mmol/L increase in TG was 1.32 (95% CI: 1.26-1.39) in men and 1.76 (95% CI: 1.5-2.07) in women [4]. Thus, identifying factors that determine plasma TG is desirable.
The contribution of lipoprotein lipase (LPL) in the development of dyslipidemia and atherosclerosis is increasingly recognized [5][6][7]. Numerous studies have shown that LPL is a key enzyme that plays a central role in lipoprotein metabolism and transport [8] and substantial evidence suggests that LPL has an important influence on TG levels [9]. In addition, LPL possesses a non-enzymatic bridging function, assisting in cellular lipoprotein uptake [10]. The finding of triglyceride-rich lipoproteins (TRL) in human atheromata [11,12] has provided substantial pathophysiologic evidence for a direct role of LPL in atherogenesis. Importantly, hereditary forms of dyslipidemia have been in part attributed to ablative mutations in LPL [13].
Though over 200 coding mutations within the LPL gene (MIM 609708) have been described, the role for common functional polymorphism of LPL, particularly in familial combined hyperlipidemia (FCH [MIM 144250]), remains poorly explored [14,15]. The most common amino acid change in LPL is Ser447X (rs328; formally p.Ser474X, c.1421C>G; but throughout the present report we have kept to the traditional nomenclature for this polymorphism), resulting in an LPL protein truncated by two amino acids [16]. It appears likely that the functional properties of the 447X truncated LPL protein is an enhanced bridging function, leading to increased clearance of TRL from the circulation [5]. Interestingly, recent cohort studies suggest that LPL 447X carriers appear to have a more favorable lipid profile and this allele appears to be a negative risk factor for coronary artery disease (CAD) [17][18][19]. An interesting recent report of neonatal somatic gene transfer with an adenoviral vector containing the S447X variant of LPL described rescue of 95% of LPL-deficient mice from lethality [20]. This study is further evidence of the gain of function of the S447X variant, as pointed out by Rader [21], because considerably fewer animals were rescued when the wild-type LPL vector had been used previously [22].
FCH is the most prevalent genetic lipid disorder observed in patients with CAD and their relatives, with a frequency of 1-2% in all populations examined and a prevalence of 10%-20% in patients with premature CAD [23][24][25][26][27]. Hypertriglyceridemia is a major component of the phenotype of FCH, the most common of the dyslipidemias [25][26][27]. The common metabolic defect in FCH appears to be hepatic overproduction of apolipoprotein B-containing TRL [28,29] and a preponderance of small dense low-density lipoprotein (LDL) particles [30,31]. Previously, we and others have proposed a genetic model for FCH that postulates the existence of both a common dominant major gene(s) that is further influenced by a number of modifier genes [23]. The identification of modifier genes could assist in finding the as yet undiscovered primary genetic determinant(s) by reducing the genetic heterogeneity of FCH [23].
Several reports have noted the presence of multiple lipoprotein phenotypes in obligate heterozygotes for LPL mutations that are reminiscent of FCH (i.e., hypertriglyceridemia, hypercholesterolemia, or both) [32]. As much as one-third of FCH patients have levels of post-heparin LPL activity and mass below the 10 th percentile for the general population [32,33]. Given the multifactorial nature of FCH, with other genetic or environmental factors (e.g., obesity) causing an increase in the hepatic production of lipoproteins, the catabolic capacity of LPL in individuals who are genetically predisposed to a low basal level of LPL activity may be overwhelmed, exacerbating the primary dyslipidemia and thus accelerating atherosclerosis [23]. Importantly, studies that have examined the role of genetic variation of the LPL gene in FCH have focused primarily on rare variations, mutations or intragenic markers [23,[34][35][36][37][38].
The possibility that a disorder could be explained, at least in part, by either enrichment of pro-atherogenic genetic variations or depletion of anti-atherogenic ones led to our interest in examining the later in combined hyperlipidemia. The occurrence of impaired LPL activity in individuals with FCH suggested that examination of the antiatherogenic LPL 447X allele could introduce a novel paradigm in the study of this disease. The aim of this study was to investigate the role of the LPL S447X polymorphism in a sample of subjects with combined hyperlipidemia and compare them with healthy controls. We provide evidence of a genetic association with LPL S447X with FCH and with lipoprotein composition in affected individuals.
Characteristics of the study groups
Based Baseline on the hypothesis that a functional variation at the LPL gene locus would differ in frequency between dyslipidemic individuals and healthy controls, we screened patients and control subjects for the S447X polymorphism. The clinical characteristics of the two groups are described in Table 1. Individuals with combined hyperlipidemia displayed significantly elevated VLDL-TG (t = -15.310, 286.34 df, p < 0.001), LDL-TG (Z = -13.656, p < 0.001) and HDL-TG (t = -9.524, 383.36 df, p < 0.001) and decreased HDL-C (t = 10.010, 425.09 df, p < 0.001) compared to controls.
Frequencies of the SNPs
Allelic and genotypic frequencies for S447X in cases and controls are listed in Table 2. The S447X genotype distribution did not deviate from Hardy-Weinberg expectations (χ 2 = 1.585; 2df; p = 0.453). It is noteworthy that the allelic and genotypic frequencies of the polymorphism in the control group were similar to previous reports [39,40].
Genetic association of lipid parameters with S447X
To examine the effects of the polymorphism on lipoprotein metabolism, fasting lipoprotein concentrations provided a metabolic 'snap shot' for comparisons between carriers of the 447X allele and the 447S homozygotes. No significant changes in mean measures of plasma lipoprotein compartments were observed in the control group (Table 3). With the case group the S447 homozygotes displayed elevated VLDL-C (SS: SX+XX, Z = -2.003, p = 0.045) ( Table 4). Suggestive evidence of an association Table 4).
Linear regression analysis of the effects of S447X on lipid measures
Regression models for lipid measurements included a data-driven selection among the polymorphism category SS versus SX+XX. Potential predictors included age, gender, and clinical category (case or control). Models in which a polymorphism effect did not achieve p < 0.10 are not reported. Among the lipid measures, a logarithmic transform was found to be generally appropriate for use in the models. No significant models were indicated for TC, VLDL-C, or VLDL-, LDL-, and HDL-triglyceride. For transformed lipid measures, estimated adjusted means on the original scale are presented (Table 5).
For total triglycerides (n = 372), the best model selected SX+XX vs. SS (p = 0.001). Selected effects include gender (p = 0.001), clinical category (p < 0.001) and genotype (p = 0.062). The adjusted R-square for the model was 0.697 ( Table 5). The group of S447 homozygotes displayed modestly elevated plasma triglycerides compared with carriers of the 447X allele (+19.7 mg/dL). For LDL-C (n = 345), the best model selected was also SX+XX vs. SS (p = 0.007). Selected effects included clinical category (p < 0.001) and genotype (p = 0.087). The adjusted R-square for the model was 0.309 ( Table 5). The group of S447 homozygotes displayed modestly elevated LDL-C as compared with carriers of the 447X allele (+11.1 mg/dL). For the ratio of total triglycerides to HDL-C (n = 370), the best model selected was again SX+XX vs. SS (p = 0.031). Selected effects include gender (p = 0.001), clinical category (p < 0.001) and genotype (p = 0.044); the adjusted R-square for the model was 0.651 (Table 5). The group of S447 homozygotes displayed modestly elevated ratio as compared with carriers of the 447X allele (+0.639).
Though BMI measurement was available on all subjects, a measure of BMI at the same time point as the baseline lipoprotein measurement was not available for all subjects. Regression models including BMI resulted in unacceptably small counts in cells, and thus analysis of the subset including BMI was not reported.
Discussion
The common S447X truncation polymorphism of LPL is associated with a cardio-protective lipid profile and a modest reduction in risk for CAD [17]. Data from this study support the hypothesis that there is significant depletion in allelic and genotypic frequencies for this protective polymorphism in individuals with combined hyperlipidemia compared to healthy controls. Specifically, we found that 447X exists in 13% of healthy controls but only in 3% of cases (p < 0.001). The frequency of 447X in the control group is similar to that in the Framingham Offspring Study that examined 1114 men and 1144 women and found the frequency of the S447X variant was 16 and 17.5% in men and women, respectively [19]. As in a previous report [19], data from this study showed no significant gender differences with respect to the LPL allelic or genotypic frequencies, though male cases versus controls displayed a slightly greater depletion of the 447X allele than did female cases versus controls.
Several population studies have reported cardio-protective alterations in lipoprotein profiles in subjects who carry the 447X allele (e.g., lower TRL, lower VLDL-C, higher apo AI levels, higher HDL-C, protection against CHD) [5,7,[41][42][43][44]. The linear regression analysis in this study showed that carriers of the 447X allele displayed modestly elevated plasma triglycerides (+19.7 mg/dL), LDL-cholesterol (+11.1 mg/dL), and ratio of TG over HDL-C (+0.639) as compared with carriers of the 447S allele. These results are congruent with population studies of cardiovascular disease that have reported similar effects on TG [4,45], increased HDL-cholesterol, and 0.8-fold reduced risk of ischemic heart disease in 447X carriers.
Although associations with favorable changes in both VLDL-C [42] and HDL-C have been reported, no association between LDL-C and this polymorphism has been demonstrated previously. We believe that it is likely that the case selection criterion we employed (elevated LDL-C) could potentially account for the association between the 447X allele and lower LDL-C. Although another polymor- phism within LPL has been associated with alterations in the TG/HDL-C ratio [46], no study has previously explored the impact of this polymorphism on the ratio. However, one may infer that similar findings were demonstrated in the longitudinal analysis of the Bogalusa Heart Study (i.e., increased frequency of 447X in subjects in the bottom quartile for plasma TG and HDL-C) [41].
Though several studies have examined the co-segregation of genetic variations within the LPL gene region with the occurrence of FCH [23], Potential study limitations include a possibility of selection bias of combined hyperlipidemia subjects without the use of familial primary dyslipidemia criterion. A goal of our group is to identify an endophenotype that would allow the identification of individuals with FCH without the resource intensive assessment of relatives. However, it is important to note that our group has successfully validated several genetic associations reported for FCH [48,49]. Another potential limitation involved the availability of BMI, an index known to impact lipoprotein metabolism and lipid homeostasis [46,50], in the entire study sample for inclusion in regression analyses.
Summary
As one would expect in a multifactorial model of a common disease, contributive alleles would be expected to add an incremental risk of disease (i.e., CAD). Common deficiencies in LPL may be the underlying causes of significant increases in CAD risk [1,2]. Understanding the role that functional gene polymorphisms play in risk and determining the levels of intermediate phenotypes (e.g., TG) is essential to our understanding of the important metabolic pathways in the diseased and disease-free state. Given the prevalence of the LPL S447X polymorphism in the population, greater knowledge of the underlying consequences of this variation may be of considerable importance in understanding genetic predisposition to atherosclerosis and heart disease [51]. S447X is a common, functional variant and is associated with a beneficial lipid profile, and is significantly depleted in patients with combined hyperlipidemia. Importantly, understanding the contribution not only of susceptibility alleles in dyslipidemia, but also of depletion of protective alleles, may contribute materially to the identification of the genetic determinants of heart disease.
Study design
This study was a retrospective analysis of the prevalence of a common LPL polymorphism in non-Hispanic Caucasian (European-descent) subjects with combined hyperlipidemia and a control group.
Genotypic and phenotypic studies
Genomic DNA was prepared from whole blood and was drawn after a 10-hour fast [52]. VLDL were prepared by ultracentrifugation [53]. HDL-C was measured after precipitation of apo-B-containing lipoproteins with dextran sulfate and magnesium [54]. Cholesterol and TG levels were measured in plasma and in lipoprotein fractions by either automated fluorescence method or automated chemical analysis [55]. LDL-C was calculated as TC minus HDL-C plus VLDL-C. Standards were provided by the Centers for Disease Control (Atlanta, Georgia, USA). Baseline lipoprotein measurements were obtained when patients had received no lipid lowering medication for at least 1 month. The presence or absence of the S447X polymorphism was determined as described previously [17].
Statistical methods
The SPSS for Windows (v11.0.1, 2001) system for statistical analysis was used. Allele and genotype frequencies were determined by the gene-counting method. Tests for Hardy-Weinberg equilibrium in controls, and allelic or genotypic association in cases versus controls, were evaluated by χ 2 test. χ 2 tests of allele frequency were adjusted using Yates' Continuity Correction. Power transformations of potential predictor variables were examined where appropriate. Two-group comparisons of means of transformed or normally distributed variables used the independent samples t-test. Two-group comparisons of means of untransformed, non-normally distributed variables used the Wilcoxon two-sample test. The procedure, general linear model (GLM), was used for linear regression models. Power transformations of potential predictor variables were examined to maximize the explanatory power of the overall model (by maximizing the F statistic). Interactions between covariates and genotypes were evaluated. Selected interaction effects and covariateadjusted means of the transformed responses for levels of categorical factors were tested using procedure GLM. Interaction effects with p < 0.10 were retained. For multiple comparisons between factor levels, Bonferroni-corrected p-values are reported.
|
2017-06-23T11:34:20.483Z
|
2006-07-05T00:00:00.000
|
{
"year": 2006,
"sha1": "73571efbcc9eae2fb1a50e2e5bda9a6e8b47c948",
"oa_license": "CCBY",
"oa_url": "https://lipidworld.biomedcentral.com/track/pdf/10.1186/1476-511X-5-19",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "845e13b80ffce89162c8a85a6e5201ab10fae000",
"s2fieldsofstudy": [
"Medicine",
"Biology"
],
"extfieldsofstudy": [
"Medicine",
"Biology"
]
}
|
219546331
|
pes2o/s2orc
|
v3-fos-license
|
Modification of Membrane Permeability by Animal Viruses
Animal viruses permeabilize cells at two well-defined moments during infection: (1) early, when the virus gains access to the cytoplasm, and (2) during the expression of the virus genome. The molecular mechanisms underlying both events are clearly different; early membrane permeability is induced by isolated virus particles, whereas late membrane leakiness is produced by newly synthesized virus protein(s) that possess activities resembling ionophores or membrane-active toxins. Detailed knowledge of the mechanisms, by which animal viruses permeabilize cells, adds to our understanding of the steps involved in virus replication. Studies on early membrane permeabilization give clues about the processes underlying entry of animal viruses into cells; understanding gained on the modification by viral proteins of membrane permeability during virus replication indicates that membrane leakiness is required for efficient virus release from infected cells or virus budding, in the case of enveloped viruses. In addition, the activity of these membrane-active virus proteins may be related to virus interference with host cell metabolism and with the cytopathic effect that develops after virus infection.
I. INTRODUCTION
Animal viruses permeabilize cells at two well-defined moments during infection: (1) early, when the virus gains access to the cytoplasm, and (2) during the expression of the virus genome. The molecular mechanisms underlying both events are clearly different; early membrane permeability is induced by isolated virus particles, whereas late membrane leakiness is produced by newly synthesized virus protein(s1 that possess activities resembling ionophores or membrane-active toxins. Detailed knowledge of the mechanisms by which animal viruses permeabilize cells adds to our understanding of the steps involved in virus replication. Studies on early membrane permeabilization give clues about processes underlying entry of animal viruses into cells; understanding gained on the modification by viral proteins of membrane permeability during virus replication indicates that membrane leakiness is required for efficient virus release from infected cells (or virus budding, in the case of enveloped viruses). In addition, the activity of these membrane-active virus proteins may be related to virus interference with host cell metabolism and with the cytopathic effect that develops after virus infection.
Apart from the interest in studying membrane permeabilization from a basic viewpoint, these studies can also provide novel, useful techniques. For instance, the finding that animal viruses increased membrane permeability to a number of compounds during the late phase of infection provided a new approach to block translation specifically in the infected cells. This selective inhibition of animal virusinfected cells was achieved by means of hydrophilic inhibitors of protein synthesis that do not permeate the membranes of normal cells, but readily penetrate into virus-infected cells (Carrasco, 1978). The use of nonpermeant translation inhibitors now constitutes a simple assay for changes in the permeability of the plasma membrane (virus induced or not) in both eukaryotic and prokaryotic cells (Carrasco and Vazquez, 1983;Lama and Carrasco, 1992a). Moreover, the initial finding that animal virus particles induce early membrane permeabilization and promote co-entry of macromolecules such as protein toxins into cells (Fernandez-Puentes and Carrasco, 1980) has been used to introduce plasmids, without size limit, to virtually transform 100% of cells grown in tissue culture (Cotten et al., 1992;Wagner et al., 1992b3. Several years have elapsed since our last review on the permeabilization of cells by animal viruses . Important insights into two main aspects of this area of research have been achieved since then. First, the mechanisms by which early and late membrane permeabilization occurs are now better understood. Second, we are now witnessing the identification of the viral proteins known as viroporins , which are involved in modifying membrane permeability. Studies on viroporins at the molecular level will soon provide details on their mode of action and the exact function they play in the virus replication cycle.
A. Mechanisms of Virus Entry
Two different modes of virus entry into cells have been described: (1) direct penetration through the plasma membrane and (2) entry through the endocytic pathway (Hoekstra and Kok, 1989; Marsh and Helenius, 1989;. Either of these pathways implies that the virus particle, the nucleocapsid, or at least the viral genome has to cross a membrane. This is particularly true in the case of viral particles devoid of a lipid envelope, such as poliovirus (Fig. 1). Following attachment of poliovirus to its receptor on the cell surface, the virus permeabilization to a-sarcin molecules. The virus particle first attaches to a cell surface receptor, followed by endocytosis of virus and toxin molecules. Binding of the virion to the receptor changes its conformation and promotes interaction of viral proteins with the membrane. This interaction promotes virus uncoating and translocation of the genome to the cytoplasm. In addition, the interaction of virus proteins with the membrane leads to the formation of transient pores that dissipate the proton gradient created by the vacuolar proton ATPase, promoting the passage of toxin molecules through the endosomal membrane. particle undergoes profound structural rearrangements (Gomez Yafal et al., 1993), with the result that VP4, which is found in the interior of the particle, interacts with the cellular membrane (De Sena and Mandel, 19771, and the amino terminus of VP1, which is also located within the virion, becomes exposed at the surface, ready to interact with the membrane (Fricks and Hogle, 1990). The interaction of these virion proteins with the membrane may open a pore through which the viral genome is extruded into the cell (Fricks and Hogle, 1990). This mechanical model of genome translocation across membranes does not take into account the requirement for an energized membrane during virus infection. Thus, the interaction of poliovirus particles with receptors leads to conformational changes in virions (Mason et al., 1993;Gomez Yafal et al., 1993;Haywood, 1994), but uncoating and RNA release require a metabolically active cell.
Viruses that possess a lipid membrane surrounding the nucleocapsid could in principle avoid passing through cellular membranes by fusion of their envelope with the plasma or endosomal membranes (Hoekstra and Kok, 1989; Marsh and Helenius, 1989). Some viruses, such as Sendai virus, are able to fuse their lipid envelope directly with the plasma membrane, whereas other viruses, such as Semliki Forest virus, fuse their envelope with the endosomal membrane (Lamb, 1993). In all circumstances, fusion is promoted by specialized viral glycoproteins endowed with this activity, and is triggered by conformational changes in the fusion glycoprotein (Stegmann et al., 1989). TWO major factors can contribute to these changes in glycoprotein conformation: (1) binding of the virus particle to the receptor and/or (2) low pH (Lamb, 1993;Haywood, 1994). The fact that a low-pH step is required during the entry of some viruses into cells has been taken as evidence that acidic pH is required specifically to induce conformational changes in the viral fusion glycoprotein (Stegmann et al., 1989). This model assumes that fusion necessarily leads to the delivery of the virus genome to the cytoplasma. No energy requirement has been implicated, provided that fusion takes place (Marsh and Helenius, 1989). However, this model does not explain how virus particles are able to cointernalize other macromolecules into cells (Fernandez-Puentes and Carrasco, 1980;Carrasco et al., 1989).
. Translocation of Toxins through Membranes
Before analyzing the co-entry of toxin moieties with virus particles, it is useful to summarize the process of toxin entry into cells. Several excellent review articles on the entry of toxins into cells have appeared (Olsnes et al., 1988;Wilson and Collier, 1992;FitzGerald and Pastan, 1993;Read and Stein, 19931, and the reader is referred to these articles for additional information. Diphtheria toxin is one of the intracellularly acting toxins whose action is best understood in molecular terms (Choe et al., 1992;Read and Stein, 1993). Its mode of entry into cells is briefly summarized here. Trypsinization of diphtheria toxin produces two fragments, A and B; the B polypeptide attaches the toxin to specific receptors present on the cell surface, followed by internalization of the toxin into endosomes (Wilson and Collier, 1992;Naglich et al., 1992). The acidic environment of endosomes triggers interaction of a hydrophobic protein domain with the membrane, leading to the formation of an ion channel (Zalman and Wisnieski, 1984;Beaumelle et al., 1992). This event is followed by the translocation of the accompanying fragment A to the cytoplasm (Olsnes et al., 1988;Ariansen et al., 1993), where it blocks translation (Wilson and Collier, 1992). It has been proposed that the transmembrane channel formed by fragment B aids fragment A in passing through the membrane, gaining access to the cytoplasm (Kagan et al., 1981). However, the pore opened by fragment B is about 25 A in diameter, which is incompatible with the A subunit traversing the B channel (Zalman and Wisnieski, 1984). The transient pores with cation-selective permeability resemble those formed during the entry of animal viruses (see below). Both fragments A and B interact with the membrane. The insertion of diphtheria toxin in the membrane does not suffice for the translocation of fragment A to the cytoplasm. This passage requires energy, provided by the protonmotive force, to push the protein moiety through the membrane (Beaumelle et al., 19921, as occurs for translocation of other proteins across mitochondria1 membranes (Pfanner and Neupert, 1990;Martin et al., 1991). Energy in the form of ATP may be required to translocate other proteins, such as ricin A chain (Beaumelle et al., 1993). Therefore, depending on the protein that is translocated across a particular membrane, energy in the form of a pH gradient (Olsnes et al., 1988;Beaumelle et al., 1992), a membrane electrical potential (Martin et al., 19911, or ATP (Beaumelle et al., 1993;Subramani, 1993) is necessary.
Early Membrane Permeabilization by Virus Particles
Several years ago, we found that viral particles efficiently permeabilized cells, allowing entry of protein toxins that are otherwise unable to cross the membrane because there are no receptors for them (Fernandez-Puentes and Carrasco, 1980). The proteins are efficiently delivered to the cytoplasm shortly after addition of virus to the medium, and almost 100% of the cells become permeabilized within a few minutes (Fernandez-Puentes and Carrasco, 1980;Carrasco, 1981;Otero and Carrasco, 1987). The majority of animal viruses tested were able to induce this phenomenon, including Semliki Forest virus (SFV), vesicular stomatitis virus (VSV), vaccinia virus, adenovirus, and poliovirus (Fernandez-Puentes and Carrasco, 1980;Carrasco, 1981;Carrasco and Esteban, 1982;FitzGerald et al., 1983;Otero and Carrasco, 1987;Lee et al., 1990), suggesting that the viral particle contains a component that not only promotes the entry of the viral nucleocapsid into the cell, but also translocates other macromolecules that are not physically bound to the particles across the cellular membrane to the cytoplasm (Fig. 1).
We know that not only protein toxins pass across membranes in the presence of animal virus particles, but that other proteins, such as luciferase or horseradish peroxidase (Otero and Carrasco, 19871, and even polysaccharides (Gonzalez and Carrasco, 19871, are also delivered into cells by these particles. It has been shown that infectious virions are not needed for early membrane permeabilization to occur, because this phenomenon takes place with UV-inactivated (but not heatinactivated) virions (Carrasco, 1981;Otero, 1986).
Poliovirus is one of the viruses that most efficiently promotes the entry of protein toxins, for instance a-sarcin, into HeLa cells (Fernandez-Puentes and Carrasco, 1980;Carrasco, 1981;Lee et al., 1990;Almela et al., 1991). This permeabilization is specific for the interaction of the viral particles with receptors, because it occurs only in cells that contain the poliovirus receptor, and not in cells that cannot be infected by poliovirus, such as L cells, or BHK cells (Otero, 1987). In addition, a virion component and not a contaminant of the viral preparation promotes permeabilization, because early permeabilization is blocked by antibodies directed against poliovirus structural proteins (Otero, 1987).
A number of compounds interact with poliovirus particles and inhibit the uncoating process without interfering with attachment or the internalization process (Rossmann, 1989;Almela et al., 1991). These compounds, in addition to blocking poliovirus uncoating, are also effective inhibitors of the early membrane permeabilization to a-sarcin (Almela et al., 1991). Hence, uncoating of the poliovirus particle to deliver the viral genome to the cytoplasm is required to permeabilize the endosomal membrane to the accompanying proteins (Almela et al., 1991).
More recently, adenovirus particle-induced cell permeabilization has been used to introduce DNA into cells by means of a receptor-mediated gene delivery system. Plasmids can bind to cells when they are complexed with transferrin-polylysine molecules Zenke et al., 1990). Entry of these complexes is enhanced when adenovirus particles are present in the culture medium. Under these circumstances more than 90% of the cells express the transfected gene (Cotten et al., 1992(Cotten et al., , 1993Curiel, 1993;Wagner et al., 1992a,b). Chicken adenovirus particles, which are replication defective in mammalian cells, have been successfully used to enhance receptor-mediated gene delivery into cells (Cotten et al., 1993).
The system developed tends to emulate virus particles in such a way that nucleic acids (plasmid DNA) are coated with proteins (transferrin) that interact with them; these proteins contain a moiety that binds the nucleoprotein complex to receptors. Although this complex is able to bind to the cell surface, it still lacks the permeabilizing capacity of virus particles that is provided by addition of inactivated virion particles. Even the conjugation of influenza fusogenic peptides, derived from the hemagglutinin molecule, with the DNA complexes enhances gene delivery (Wagner et al., 1992a), suggesting that a synthetic virus like system for efficiently transforming cells is feasible.
Inhibitors of Vacuolar Proton-ATPase
The use of several macrolide antibiotics that powerfully block proton-ATPase pumps in mammalian cells promises to be crucial in elucidating the mechanisms by which animal virus particles or genomes pass into the cell interior (Bowman et al., 1988). In addition, the forces that govern early membrane permeabilization by viruses is now better understood thanks to these antibiotics. Bafilomycin A (BFLA) and concanamycins are selective inhibitors of endosomal proton-ATPase (Drose et al., 1993). BFLA and concanamycin A strongly inhibited the entry of several animal viruses that possess a lipid envelope, such as SFV, VSV, and influenza virus Carrasco, 1993, 1994;Guinea and Carrasco, 1994b,c). On the other hand, the infectivity of poliovirus was not blocked by BFLA, indicating that poliovirus uncoats and releases its genome by a mechanism that is independent of an acidic environment . We are currently investigating the requirement of membrane potential for poliovirus entry into cells. Because entry of SFV and poliovirus was affected differentially by BFLA, they constituted good models to test the involvement of the proton-ATPase activity in the cointernalization process. Curiously enough, BFLA inhibited the co-entry of a-sarcin promoted by both SFV and poliovirus particles . This result clearly indicates that the exit of a-sarcin from endosomes requires a proton gradient and is coupled to the action of the proton-ATPase. These findings prompted us to suggest that the low-pH requirement previously reported for the entry of some animal viruses could be due to a requirement for a pH gradient, rather than to low pH per se . The pH gradient would provide the energy required in this process, as occurs during the entry of toxins into cells (see Section II,B,l). If so, a virion component would be able to couple the proton gradient to the release of macromolecules from the endosome. Therefore, the mechanism of the co-entry of protein toxins is more specific than previously thought and does not involve the physical rupture of endosomes by animal virus particles; rather the toxin would be translocated across the intact endosomal membrane by an activity present in virion particles .
C . A Proton Motive Model to Account for Virus Entry and Early Membrane Permeabilization
Current models that explain the entry of viruses into cells do not account for all the existing experimental evidence (see Section 11,A). The requirement of a metabolically active cell for delivery of virus genome into the cytoplasm is not rationalized. In addition, these models do not explain the mechanism by which early permeabilization, observed with virus particles, to other macromolecules takes place. We have proposed that the proton-motive force generated in endosomes by the activity of the vacuolar proton-ATPase pump can be used to promote uncoating and to drive the viral genome through the lipid barrier of the membrane . Therefore, the energy accumulated in endosomes is required for the viral nucleocapsid to enter cells. In this model the same route and energy can be used by macromolecules to pass across the membrane and would thus account for the co-entry process. A brief account of the different steps of virus entry according to the protonmotive model, would be as follows: 1. Viral proteins can alter their conformation on receptor binding and insert into the cellular membrane (Flynn et al., 1990;Meyer et al., 1992). Acidic pH could also influence these conformational changes (Marsh and Helenius, 1989), but this may not be strictly required for fusion (Haywood and Boyer, 1985;Edwards and Brown, 1991;Haywood, 1994). In fact, binding of virus to its receptor is not strictly required for foot-and-mouth virus infection (Mason et al., 1993 (Spruce et al., 1991;Schlegel et al., 1991;Tosteson et al., 1993). In fact, some virus particles contain "textbook examples" of ion channels (Kalko et al., 1992). 3. The viral proteins inserted in the membrane may couple the energy liberated by the movement of protons (or ions) to the cytoplasm down a concentration gradient, to the translocation of the viral genome (or protein toxins) in the same direction (see Fig. 1).
This model predicts that there are viral proteins that open pores in membranes and, together with receptors, are able to use energy to translocate substrates in a nonfavorable thermodynamic direction. In contrast to an uncoupler, which simply dissipates the energy stored in ionic gradients, this "virus-transducing complex" is able to couple the energy to genome translocation. The pore size would permit the passage of protons and other ions, but prevent diffusion of macromolecules.
Additional support for this concept is provided by the fact that virions modify membrane potential during entry (Fuchs and Kohn, 1983;Rosenthal and Shapiro, 1983;Seth et al., 1985;Bashford et al., 1985), probably as a consequence of the capacity of virion proteins to form ion channels Spruce et al., 1991;Tosteson et al., 1993). In addition, SFV does not enter cells when the membrane potential is abolished by modifying the concentration of monovalent cations, even under acidic conditions (Helenius et al., 1985). As indicated above, viruses are able to monitor and infect only cells that are metabolically active (Fuchs and Kohn, 1983). Even though viruses attach to dead cells (and even to isolated membranes, or truncated receptors) (Lentz, 1990;Haywood, 19941, they effectively fuse and enter only cells that possess an energized membrane (Fuchs and Kohn, 1983).
In addition to virus entry, the early permeabilization phenomenon observed with virus particles can also be easily rationalized by this proton-motive model. Thus, viral proteins involved in the translocation of the viral particle across the membrane would also translocate other macromolecules, depending on an existing pH gradient or membrane potential. Therefore, this model accounts for the observation that the toxin moieties involved in translocation through membranes are exchangable with virus particles (Fernandez-Puentes and Carrasco, 1980).
The mechanical and the proton-motive models can be easily differentiated by experimental tests. The classical low-pH model predicts that a low pH suffices for virus fusion and entry in a mechanistic way, i.e., low pH changes the conformation of a protein that is inserted into the cellular membrane and fusion ensues. The proton-motive model predicts that a pH gradient is required, thus low pH is not sufficient (Guinea and Carrasco, 1994b,c;Perez and Carrasco, 1994). Moreover, in principle, the pH gradient may not be necessary for entry of those virus species that could use either the membrane potential or the pH gradient. Future studies in this exciting field of virus entry directed to elucidating the exact molecular basis of early membrane permeabilization should decide which of these models is closer to reality.
A. Mechanism of Late Membrane Permeabilization
The molecular mechanisms by which animal viruses modify membrane permeability late during infection will be considered in this section. Because the work done in this field has been previously reviewed in depth (Carrasco et al., 19891, the late modifications of the membrane induced by animal virus infection will be briefly summarized and work done in the last few years will be highlighted. One of the most salient characteristics of the late membrane modifications is that they require viral gene expression , suggesting that one or several virus gene products are responsible for these changes. Figure 2 shows the pathway for synthesis and membrane insertion of virus proteins with membrane-active capacity and the postinfection alterations that influence membrane function. Two fundamental aspects are of interest in considering the late membrane leakiness induced by animal virus infection: (1) the nature of the cell membrane modifications at the molecular level and (2) the identity of the viral products responsible for these modifications. Considering the large number of cytolytic animal viruses identified, we have only a limited understanding of these two aspects in a few viral systems.
Each cytolytic animal virus modifies the permeability of the plasma membrane with different kinetics, most probably reflecting not only the different timing of virus gene expression but also the intrinsic activity of each particular viral protein involved in cell lysis. Moreover, the particular cell line considered and the culture conditions also influence these alterations in membrane permeability. The specific phenomenology of plasma membrane permeabilization is thus a characteristic not only of the animal virus considered but also of the cell it infects. Nevertheless, some general features describe the changes that accompany the disruption of the cell membrane.
Monovalent Cations and Membrane Potential
The most common change that occurs in the plasma membrane during infection of a susceptible cell by a cytolytic virus is enhanced permeability to monovalent cations. This effect is obviously accom-panied by a drastic drop in membrane potential (see Fuchs and Kohn, 1983;Carrasco et al., 1989). Sodium ions that are pumped outside the cell by Na+ ,K+-ATPase readily enter into infected cells, whereas potassium ions that accumulate inside the cell leak out. In the case of picornaviruses or togaviruses these modifications in ion distribution are initially observed 2 to 3 hr after infection (Carrasco and Smith, 1976;Nair et al., 1979;Nair, 1981Nair, , 1984Garry et al., 1979;Egberts et al., 1977;Muiioz et al., 1985a;Lacal and Carrasco, 1982;Ulug et al., 1984;Lopez-Rivas et al., 19871, so that the majority of viral macromolecular synthesis takes place in a cytoplasm in which the concentrations of monovalent ions are continuously changing (Munoz et al., 1985b;Carrasco and Castrillo, 1987;Ulug et al., 1987). Hence, the synthesis of poliovirus or Semliki Forest virus proteins is carried out in a cytoplasm that contains a higher concentration of sodium and a lower concentration of potassium ions compared to uninfected cells. Consequently, viral mRNAs that are translated late during virus infection have adopted special structures that cause them to be optimally translated under these altered conditions (Saborio et al., 1974;Nuss et al., 1975;Carrasco and Smith, 1976;Carrasco et al., 1979). Hence, the resistance of some viral mRNAs to translational inhibition by high concentrations of sodium ions in the external medium is taken as a reflection of the alterations that take place in ion concentration in the cytoplasm of virus-infected cells. This resistance is a widespread phenomenon observed for the translation of mRNAs from cytolytic animal viruses (Nuss et al., 1975;Saborio et al., 1974;Cherney and Wilhelm, 1979;Garry et al., 1979;Alonso and Carrasco, 1982a;Carrasco et al., 1979;Nair, 1981Nair, , 1984Mizzen et al., 1987;Castrillo et al., 1987). On the other hand, these ionic concentrations are inhibitory for the translation of most cellular mRNAs, although the synthesis of a few cellular proteins, such as the heat-shock proteins, is resistant to such modifications (Munoz et al., 1984;Macejak and Sarnow, 1991).
In summary, progressive membrane damage induces a collapse of ionic gradients and disrupts membrane potential in virus-infected cells (Fig. 2). Cell morphology a t this time of infection usually appears normal under the phase-contrast microscope; cell rounding and shrinkage take place later, a t a time when the membrane is highly damaged (Lopez-Rivas et al., 1987). The phenomenology of these changes is similar to that observed with membrane-active toxins (Bernheimer and Rudy, 1986;Dempsey, 1990;Bevins and Zasloff, 1990;Guihard et al., 1993), or ionophores (Alonso and Carrasco, 1982b;Heitz et al., 1989;Katsu et al., 1989).
Divalent Cations and pH
The concentrations of protons and calcium ions also change a t about the same time that monovalent ion concentrations are affected. Despite the importance of protons and divalent cations as regulators of a plethora of cellular functions (Carafoli, 1987;Whitaker, 1990;Madshus, 1988), few studies have been performed on this subject in animal virus-infected cells.
Intracellular alkalinization of about 0.3 pH units has been reported in poliovirus-infected cells (Holsey et al., 1990;Holsey and Nair, 1993). Virus gene expression is necessary for this modification to occur, but only the translation of the input viral RNA is required. These results contrast with the finding that Sindbis virus infection causes an acidification of about 0.5 pH units in BHK cells (Moore et al., 1988). A similar decrease in the cytoplasmic pH is observed in MDCK cells infected by influenza virus (Ciampor et al., 199213). In addition, variations in the pH of cellular compartments in influenza virus-infected cells are thought t o be mediated by the M2 protein (Ciampor et al., 1992a). Differences in the methodology used to estimate the cytoplasmic pH may account for the contradictory results obtained in poliovirus-and togavirus-infected cells. It is also possible, although unlikely, that different viruses have opposite effects on cytoplasmic pH.
Initial attempts to measure variations in the concentration of calcium ions in virus-infected cells showed that calcium uptake by mitochondria increased during the initial hours of Semliki Forest virus infection, perhaps reflecting a rise in cytosolic calcium (Peterhans et al., 1979). The increased uptake declined at late stages of infection, suggesting mitochondria1 injury by virus infection (Peterhans et al., 1979). As with poliovirus-infected cells (Lopez-Rivas et al., 1987), no significant differences were observed in total cellular calcium as determined by radioactive calcium (Peterhans et al., 1979). Using this technique, Nokata et al. (1987) noted a greater influx of calcium into cytomegalovirus-infected human cells during the initial hours of infection, leading to the suggestion that a viral gene product altered membrane permeability to this cation (Nokta et al., 1987). The use of fluorescent probes to estimate free cytosolic calcium provides a more accurate method than measurements based on radioactive calcium (Cobbold and Rink, 1987). The use of both radioactive calcium and fluorescent probes in rotavirus-infected cells clearly indicates that permeability to calcium is drastically altered and the concentration of this cation increases by severalfold in the infected cells (Michaelangeli et al., 1991). These modifications are clearly apparent from about the fourth hour postinfection and are dependent on virus gene expression (Michelangeli et al., 1991). The expression of individual rotavirus proteins in insect (Spodoptera frugiperda) cells points to NSP4 expression as the nonstructural glycoprotein responsible for elevating calcium levels during rotavirus infection (Tian et al., 1994). Influenza virus infection of neutrophils causes a rise in intracellular calcium very early during infection (Hartshorn et al., 1988). This effect has been related to a functional deactivation of neutrophils by influenza virus infection.
Curiously, to my knowledge, no studies measuring free cytoplasmic calcium in picornavirus-infected cells have been published. Attempts to estimate the concentrations of calcium in poliovirus-infected HeLa cells by means of radioactive calcium revealed almost no difference from concentrations in uninfected cells (Lopez-Rivas et al., 1987). However, these studies did not discriminate between free or bound calcium (Lopez-Rivas et al., 1987). More recent results using fluorescent probes indicate that calcium levels in the cytoplasm of poliovirus-infected cells rise by 5to 10-fold from the third hour postinfection (A. Irurzun, J. Arroyo, and L. Carrasco, unpublished observations).
Enhanced Permeability to Other Compounds
In addition to ions, other low-molecular-weight compounds readily diffuse through the membranes of virus-infected cells . Nucleotides and sugars leak from cells and a number of hydrophilic antibiotics that do not permeate the plasma membrane of normal cells selectively enter virus-infected cells (Carrasco, 1978;Contreras and Carrasco, 1979;Carrasco and Vazquez, 1983). The use of these nonpermeating antibiotics provides a sensitive assay for changes in membrane permeability induced by viruses or other compounds Carrasco, 1981, 1982b;Lama and Carrasco, 1992a), and also gives an accurate indication as to the timing of changes in permeability.
It could be speculated that changes in membrane permeability occur only in a fraction of "dead" cells, whereas metabolically active cells maintain intact membranes while synthesizing viral macromolecules. This is not so, because the normally nonpermeating translation inhibitors totally block viral protein synthesis, indicating that cells actively synthesizing viral proteins lose the permeability barrier to these compounds (Carrasco, 1978;Contreras and Carrasco, 1979). Still, it is possible that the viral proteins made in leaky cells are useless for the virus, i.e., these proteins will not participate in the formation of virions and are made after the virus progenies have escaped from the cells. This is unlikely, because the inhibition of protein synthesis by hygromycin B, a nonpermeating inhibitor, reduces progeny virus formation by several orders of magnitude (Benedetto et d., 1980;Lacal and Carrasco, 1983).
The conclusion reached from these studies is that membrane permeability nonspecifically increases at the beginning of the late phase of virus infection. Ions and other low-molecular-weight hydrophilic molecules readily diffuse through the plasma membrane in both directions. Phenomenologically, it seems that pores appear in a gradual fashion after virus infection, as previously predicted (Carrasco, 1977(Carrasco, , 1987.
B. Modification of Lipase Activity
Apart from the characteristic permeability properties of the membrane with respect to different solutes, a n important question relates to the physical integrity of the membrane phospholipids. The most surprising aspect of this area of research is the paucity of data regarding the molecular basis of membrane lysis by cytolytic viruses. The identification of virus gene products that enhance membrane permeability will promote more interest in analyses of phospholipid integrity and lipase activation in animal virus-infected cells.
HeLa cells infected with poliovirus undergo a marked stimulation of phospholipase C, whereas the activity of other phospholipases does not increase (Guinea et al., 1989;Irurzun et al., 199313). As a result of this stimulation, choline and phosphorylcholine appear in the culture medium from the third hour postinfection, and the levels of inositol triphosphate UP,) in the cytoplasm of poliovirus-infected cells is several times higher than in uninfected cells (Guinea et aZ., 1989). Thus, it is not surprising to find elevated concentrations of divalent cations in the cytoplasm of these cells (see Section 111,A,2). In conclusion, phospholipase C is activated in poliovirus-infected cells, leading to the release of choline to the medium and the formation of high amounts of IP, (Fig. 2). Perhaps this inositol metabolite, in turn, triggers the release of membrane-bound calcium into cytosolic free calcium (Berridge and Irvine, 1989;Whitaker, 1990). The consequences that these changes may have for cellular metabolism and for viral functions, such as assembly, have not been determined.
Not all viruses induce the activation of phospholipase C alone. In cells infected with other animal viruses there is an increase in choline and arachidonic acid release to the medium, suggesting that in addition to phospholipase C, phospholipase A, is also activated in these systems . Arachidonate metabolism is modified by a number of animal viruses, including orthopoxviruses (Palumbo et al., 1993), herpesviruses ( AbuBakar et al., 19901, Semliki Forest virus and vesicular stomatitis virus . During the infection of BSC-1 cells by cowpox virus or vaccinia virus there is an enhanced arachidonate product formation from the cyclooxygenase and lipoxygenase pathways. Possibly a virus gene product is involved in mediating the alterations of arachidonate metabolism (Palumbo et al., 1993). These alterations seem to be necessary for orthopoxvirus replication, because inhibitors of lipoxygenase pathway block virus replication (Palumbo and Buller, 1991). During the infection of HeLa cells with SFV or BHK cells with vesicular stomatitis virus (VSV), there is a significant increase in phospholipase C and A, activity. Entry of hygromycin B was prevented by zinc ions or chloroquine, indicating that the increase in membrane permeability in SFV-infected cells may be mediated in part by lipase activation (Perez et aZ., 1993). Respiratory syncytial virus infection of differentiated U937 cells induces a sustained stimulation of l-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine, a potent mediator of inflammation, suggesting that this compound plays an important part in RSV-induced inflammatory response in lungs (Villani et al., 1991). Even the addition of purified gp120 from HIV-1 to human monocytes results in the activation of the cyclooxygenase and lipoxygenase pathways (Wahl et al., 1989). These results lead to the suggestion that the induction of arachidonic acid metabolites is required for the replication of some animal viruses (Palumbo and Buller, 1991) and may also be important for virus pathogenesis.
Apart from the stimulation of lipase activity as a result of virus gene expression, animal viruses can also modify lipase activity during entry into cells, as observed with influenza virus (Harshorn et al., 1988) and cytomegalovirus ( AbuBakar et al., 1990).
The actual picture that emerges from these results is that the activation of these phospholipases depends very much on the virus and the cell type used in these studies. Two consequences may be predicted if these two phospholipases are activated during the course of virus infection: (1) the formation of prostaglandins and (2) the formation of diacylglycerol and lysophosphatidylcholine. These molecules are effectors of many cellular and viral functions (Exton, 1990;Billah, 1993). In addition, the presence of these phospholipid moieties in the plasma membrane destabilizes the lipid bilayer and enhances membrane permeability (Felix, 1982). It is not surprising to find that cellular membranes containing embedded pore-forming proteins display increased phospholipase activities. Many membrane-active compounds have a similar effect (Katsu et al., 1989;Young, 1992). It seems plausible to suggest that poreforming proteins disturb membrane integrity, leading to membrane permeabilization that in turn increases phospholipase activity, which further destabilizes the membrane. The membrane might be permeabilized even in the absence of phospholipase activation, as occurs with phage systems (Hardaway et al., 1975;Young, 1992). Future studies will define more precisely the role that each of these changes plays in the increased membrane permeabilization induced by animal virus infection.
Finally, infection of chicken embryo fibroblasts by b u s sarcoma virus, a nonlytic transforming virus, stimulates the release of arachidonic metabolites (Barker et al., 1989), an effect involved in cell transformation by tyrosine kinase-containing retroviruses (Barker et al., 1989).
C. Vesicular Traffic in Cells Infected with Animal Viruses
In addition to modifications in the structural and functional properties of the plasma membrane, animal viruses can also modify the vesicular system. However, these modifications may not be so relevant to the enhanced permeabilization effect but may instead be connected with other viral functions such as genome replication Carrasco, 1990, 1991;Irurzun et al., 1992Irurzun et al., , 1993a or virus assembly (Sodeik et al., 1993;Schmelz et al., 1994). Therefore, modification of the vesicular system by animal viruses will not be reviewed here. I will only point out that some animal viruses, such as poliovirus, not only use the vesicular system to transport and sort their proteins, but also to replicate their genomes. The replication of poliovirus genomes takes place in close association with newly made membrane vesicles (Caliguiri andTamm, 1969, 1970;Semler et al., 1988). During infection with a number of animal viruses, membrane traffic is profoundly modified in such a way that vesicles of various sizes appear in the cytoplasm. These distinctive structures were termed "viroplasm" in the case of poliovirus (Dales et al., 19651, "cytopathic vacuoles" in togaviruses (Acheson and Tamm, 1967;Grimley et al., 1968;Froshauer and Kartenbeck, 1988), or "factories" in vaccinia virus-infected cells (Cairns, 1960;Dales and Siminovitch, 1961; Moss, 1990). The macrolide antibiotic brefeldin A is active against animal viruses that contain glycoproteins and also shows antiviral activity against poliovirus (Maynell et al., 1992;Irurzun et al., 1992;Tucker et al., 1993;Molla et al., 19931, a virus lacking a lipid envelope and that does not encode glycoproteins (Urzainqui and Carrasco, 1988). The results obtained with brefeldin A indicate that the poliovirus replication complexes must pass through the Golgi apparatus in order to be functional (Irurzun et al., 19921, a process that may be mediated by protein 2C (Rodriguez and . Apart from poliovirus, the synthesis of viral RNA by other animal viruses is also depressed by inhibition of lipid synthesis or membrane traffic Irurzun et al., 1993a).
IV. VIROPORINS: VIRAL PROTEINS THAT MODIFY MEMBRANE PERMEABILITY
The great majority of animal viruses are cytolytic. During their growth cycle they lyse cells by mechanisms that are as yet poorly understood. Animal virus replication is not necessarily bound to cytopathogenicity. Thus, mutant animal viruses with reduced cytopathic properties are able, in many instances, to replicate and express their genomes to normal levels (Kowalski et al., 1991;Schlesinger et al., 1993), suggesting that at least in some viruses a single gene product is the major cause of the cytopathic effect. The concept advanced that the cytopathic effect is produced by the expression of one or a (limited) number of virus genes (Carrasco, 1987) is gaining support, as opposed to the idea that the bulk of virus gene expression, or the formation and accumulation of viral particles, is cytopathic. The virus products involved in cytopathicity most probably have the plasma membrane as the primary target of their toxic effects. The suggestion that animal viruses encode proteins "that form small pores in the lipid bilayer," acting as ionophores (Carrasco, 19771, is now leading to the recognition of a new family of virus proteins that enhance membrane permeability: the viroporins. With increasing knowledge of viral genomes and the application of cloning techniques, a number of proteins involved in modifying membrane permeability have been identified in recent years, and we are beginning to understand the function they play in cytopathogenicity and virus growth. In many respects cell lysis induced by animal viruses is similar to the lysis induced by membrane-active compounds or phages (see Section VI). The knowledge accumulated from these systems may benefit our understanding of the molecular-level processes involved in cell lysis during animal virus replication. Characterization of animal virus products involved in cell lysis is a major goal toward understanding their mode of action. Progress in elucidating the putative viral proteins involved in modification of membrane permeability has been achieved through their cloning and individual expression in bacteria and eukaryotic cells.
A . Picornavirus 3A Protein
Members of the Picornaviridae family have a positive, singlestranded RNA molecule as genome (Semler et al., 1988;Harber and Wimmer, 1993). Translation of this RNA yields a polyprotein that is proteolytically cleaved to produce the mature picornavirus proteins (Fig. 3). Growth of poliovirus in susceptible cells leads to drastic changes in membrane permeability, as described in Section II1,A. Picornavirus-infected cells perhaps represent the animal virus system in which modifications of membrane permeability have been most thoroughly analyzed. Despite a wealth of information, however, little is known about the picornavirus genes involved in these changes.
A s a first step toward the identification of poliovirus proteins involved in increased cellular permeability, most of the poliovirus nonstructural proteins were individually cloned in E . coli (Lama et al., 19921, and their expression was regulated by the use of a powerful inducible system (Studier et al., 1990;Dubendorff and Studier, 1991). These studies led to the identification of two toxic proteins with lytic capacity for bacteria: poliovirus proteins 2B and 3A (Lama and Carrasco, 1992a,b). Induction of elevated synthesis of either of these polypeptides permeabilized prokaryotic cells, whereas none of the other poliovirus nonstructural products produced this effect (Lama and Carrasco, 1992a). Polypeptide 3A, and its precursor 3AB, are known to interact with cellular membranes in infected human cells (Takegami et al., 1983a,b). Computer analyses of these proteins indicated that the hydrophobicity of 3A extended over a region of 20-22 amino acids that formed an amphipathic helix. Moreover, 3A is a basic protein that shows some homology with other lytic proteins . In contrast, protein 2B lacks an amphipathic helical domain. Because membrane-active proteins act on all types of cells, irrespective of whether they are eukaryotic or prokaryotic Bernheimer and Rudy, 1986;Dempsey, 1990;Kokryakov et al., 1993), their expression and analysis of their mode of action in E . coli cells provide a powerful model system. The inducible system developed by Studier et al. (1990;Dudendorff and Studier, 1991) has already proved to be suitable for the characterization of these toxic and highly lytic virus proteins, because their expression is tightly regulated (Lama and Carrasco, 1992b). In addition, these viral proteins modified membrane permeability in a manner similar to that observed using poliovirus in mammalian cells (Lama and Carrasco, 1992a). In this respect, changes in permeability to different small molecules (lactose analogs, lysozyme, nucleosides, nonpermeating inhibitors) induced by poliovirus 3A occurred in both directions (influx and efflux), a t a time when other larger molecules were effectively excluded by the plasma membrane. This suggests the formation of hydrophilic pores of a defined size (Lama and Carrasco, 1992a;Lama, 1994).
The prokaryotic system used in these studies was also useful for characterizing 3A proteins with mutations affecting their permeabilizing properties. Because synthesis of poliovirus polypeptide 3A in bacterial cells elicits an increase in membrane permeability, the host cells can be morphologically differentiated from E . coli clones expressing wild-type 3A in the presence of a chromogenic substrate (Lama, 1994). Characterization of the mutants indicated that lysis was a highly specific property of the protein, because changes in a single amino acid abolished the permeabilizing activity and hence the lytic properties of 3A. Even though some 3A mutants still encode a 3A basic protein that is synthesized in cells, usually to higher levels than wild-type 3A, they do not increase membrane permeability (Lama, 1994). Some of the mutations were located in the hydrophobic domain of protein 3A, giving rise to partially soluble viral polypeptides, whereas other mutations were located in two different domains of 3A, outside of the hy-drophobic region (see Fig. 3). An important concept that arises when considering the action of the different 3A mutants is that the ability of a protein to modify membrane permeability should be envisaged in a very specific way. As with enzymes, wherein a single amino acid change may lead to the loss of enzymatic activity, protein 3A loses its pore-forming potential as a result of a small change, for example, the modification of a single amino acid.
Although poliovirus protein 3A is lytic for E . coli cells in the presence of phage T7 lysozyme, and this lytic capacity is specific and does not simply arise from accumulation of 3A in cells, it has not been proved that this protein is responsible for the cytolytic effects of poliovirus. The findings described thus far favor the view that 3A is the poliovirus protein involved in cell lysis, but to reach this conclusion two fundamental predictions still need to be proved. One is that poliovirus protein 3A increases membrane permeability when it is individually expressed in uninfected mammalian cells. Second, poliovirus strains containing mutations in the 3A gene should be unable to induce membrane alterations. A definitive answer to the identity of the poliovirus protein involved in permeability changes must await these results.
B . Several HIV Proteins Modify Membrane Permeability
Human immunodeficiency virus (HIV) is a retrovirus that causes AIDS (Gougeon et al., 1993;Levy, 1993b). Great efforts have been devoted to understanding AIDS and HIV biology, giving rise in a rather short period of time to a detailed picture of the mechanism by which HIV infection modifies membrane permeability. As with other lentiviruses, HIV infection of susceptible cells leads to a number of cytopathic effects, including cell death (Levy, 1993b). Progression to disease in HIV-infected patients is accompanied by loss of CD4+ T lymphocytes and has been correlated with the emergence of more cytopathic virus variants (Levy, 1993a,b). These HIV variants replicate more rapidly, are more cytopathic for infected cells, and may contribute to lymphocyte depletion and the severity of HIV-induced pathogenesis (Gougeon et al., 1993).
The cytopathic effects of HIV infection are rather complex, involving syncytia formation, cell ballooning, and individual cell lysis . The severity of these effects depends on the HIV variant and on the cell type analyzed. As usually occurs with cytolytic animal viruses, the cytopathic effects are the outcome of interaction of a given virus with a given host cell; other conditions also modulate and influence the cytopathic effect (Bablanian, 1975;Schrom and Babla-nian, 1981;Fraenkel-Conrat and Wagner, 1984). Due to this complexity it is not surprising to find that different viral molecules, or different regions of a given protein, participate to different extents in the cytopathic effect. The proteins of fusogenic viruses, such as HIV, are involved in the formation of multinucleated cells (syncytia) and must be distinguished from the proteins that modify individual cell morphology. Both events might be classified as cytopathic effects, but the proteins contributing to them might be different. This distinction is not usually made in studies devoted to analysis of virus cytopathogenicity, sometimes creating confusion. Syncytia formation by HIV is related to the fusion properties of glycoproteins, whereas cell killing may be effected by a different gene product, although the molecular events underlying both phenomena may have in common the disturbance of membrane function Cloyd and Lynn, 1991). In conclusion, changes in membrane permeability by HIV, and perhaps by other enveloped animal viruses, may be produced not only by a typical viroporin, but also by several glycoprotein domains.
The major HIV-induced cell injury has been correlated with perturbation of cell membrane permeability (Lynn et al., 1988;Garry et al., 1988;Cloyd and Lynn, 1991), but not with membrane fusion (Somasundaran and Robinson, 1987). Alterations of membrane permeability by HIV infection are observed early during infection, induced by virus particles Fermin and Garry, 19921, and late in the infectious cycle, as a result of gene expression. These late alterations include nonspecific modification of the transport of low-molecularweight compounds and monovalent and divalent cations (Lynn et al., 1988;Cloyd and Lynn, 1991). Sucrose enters HIV-infected cells, whereas inulin is excluded, suggesting that the pores formed by the expression of HIV genes have a specific diameter that allows passage of hydrophilic molecules up to a given size (Cloyd and Lynn, 1991;Miller et al., 1993). Lipid synthesis is also affected in HIV-infected cells, as in other animal virus-infected cells . In HIV infection, synthesis of phospholipids decreases, whereas synthesis of neutral lipids is enhanced (Lynn et al., 1988).
. Portions of the HIV gp4Increase Membrane Permeability
Initial studies with HIV-1 mutants suggested that cytopathicity of T lymphocytes was affected by the carboxy region of the transmembrane (TM) glycoprotein (gp41) (Fisher et al., 1986;Sodroski et al., 1986). Cytopathogenicity induced by other retroviruses, such as feline leukemia virus or simian immunodeficiency virus (SIV), is also affected by the TM glycoprotein (Bosch et al., 1989;Thomas and Overbaugh, 19931, indicating that the interaction of a n HIV protein with the membrane is the basis of the cytopathicity. Several regions in HIV gp41 have been implicated in membrane interactions (see Fig. 4). One region is located at the amino terminus of the protein and is involved in cell fusion. The amino terminus of orthomyxovirus and paramyxovirus TM glycoproteins contains a region of hydrophobic amino acids involved in membrane fusion; this region has been termed "fusion peptide." Similarities between the amino terminus of HIV gp41 and these fusion peptides have been noted (Gallaher, 1987). Point mutations in the amino-terminal region reduced syncytium formation, without affecting glycoprotein processing (Freed et Kowalski et al., 1991). Mutations of individual amino acids located in the amino-terminal hydrophobic stretch of gp41 as well as polar amino acids that flank this region profoundly block syncytium formation (Freed et al., 1990). A similar result is observed with two mutants that contain several amino acid insertions in this region (KOwalski et al., 1991). Curiously, the replication ability and virion production by these mutants do not differ from that of wild-type HIV-1 (HXBC2) in Jurkat lymphocytes, indicating that the levels of HIV gene expression (Somasundaran and Robinson, 19881, or virus particles, are not sufficient to induce cytopathic effects (Kowalski et al., 1991). Single-cell lysis of Jurkat cells was also affected by modifying the hydrophobic amino terminus of gp41 after 5-10 days of infection, whereas maximal expression of p24 peaked a t day 5 for wild-type and mutant viruses (Kowalski et al., 1991).
al.,
Membrane fusion and syncytia formation are also effected by the membrane-spanning domain (Owens et al., 1994) and the cytoplasmic moiety (carboxy terminus) of HIV-1 gp41 (Lee et al., 1989). Deletion of the carboxy-terminal 6-14 amino acids of HIV-1 gp41 alters the infectivity and cytopathogenity, although, surprisingly, a larger deletion of 76 amino acids did not abolish viral replication (Lee et al., 1989). The conclusion reached was that conformational determinants, rather than protein length, are important for gp41 function. Deletion of 146 amino acids of the cytoplasmic domain of SIVmac (plus several point mutations throughout the protein) produces a nonpathogenic variant, SIVmac l A l l , in monkeys. Curiously, the expression of the recombinant TM from SIVmac l A l l produces much larger and more numerous syncytia in rhesus macrophages compared to the parental TM (Ritter et al., 1993). Clearly, cell killing is unrelated to fusion ability or pathogenicity, in both SIV and HIV (Somasundaran and Robinson, 1987;Chakrabarti et al., 1989;Ritter et al., 1993).
There are two other regions in the gp41 glycprotein of HIV that display high hydrophobic moment, suggesting that they may also associate with membranes (Venable et al., 1989). These two regions span approximately amino acids 772-790 and 828-848 (Fig. 4). Theoretically, these regions could form highly amphipathic a-helices that may reside in the membrane as aggregates, forming hydrophilic pores (Venable et al., 1989). The hydrophobic faces of these peptides would be in contact with membrane lipids; the hydrophilic faces would form the pore in contact with water (Venable et al., 1989;Eisenberg and Wesson, 1990). Indeed, gp41 tends to aggregate preferentially to form tetramers, although trimers are also detected (Pinter et al., 1989). Similarities between these regions and those in other lytic peptides, such as melittin, have also been noted (Eisenberg and Wesson, 1990).
Apart from structural considerations of gp41, two lines of evidence suggest that gp41 alters membrane permeability. Peptide homologs of the region spanning amino acids 828-855 are very toxic for prokaryotic and eukaryotic cells, eliciting membrane permeability changes and cell lysis (Miller et al., , 1993. It was suggested that the lytic peptides form pores of defined size in the plasma membrane (Miller et al., 1993).
In addition to the regions of gp41 discussed above, the transmembrane domain of this protein could, in principle, also participate in modifying membrane permeability. This region shows similarities to other viroporins, including the presence of an adjacent region containing basic amino acids. Future studies on the membrane-active properties of gp41 could be directed to the expression of this protein in model systems to analyze the capacity of the whole protein to increase membrane permeability and the capacity of truncated versions of gp41 to affect the membrane. These analyses would map with more precision each portion of gp41 that plays a role in enhancing membrane permeability.
HIV V p u Protein
The vpu gene of HIV-1 is located 3' to the first exon of tat and overlaps with the 5' and of enu (Strebel et al., 1988;Cohen et al., 1988). HIV-2 and SIV do not contain an analogous protein. Vpu is a typical viroporin molecule; it is a n integral membrane protein containing 81 amino acids and is phosphorylated on a serine residue (Strebel et al., 1989). It contains a hydrophobic stretch of 28 amino acids a t the amino terminus ( Fig. 4) embedded in the membrane, with the hydrophilic carboxy terminus oriented toward the cytoplasm (Maldarelli et al., 1993). Vpu is produced late in replication and it has not been found in virions; the protein is localized in a perinuclear region, most likely the Golgi complex (Klimkait et al., 1990).
Genetic analysis with Vpu mutants suggests that this protein participates in late steps of infection related to virus release from infected cells (Terwilliger et al., 1989;Klimkait et al., 1990;Yao et al., 1992;Gottlinger et al., 1993). The absence of Vpu results in an accumulation of cell-associated viral proteins. T cells infected with Vpu-positive virus release three-to fivefold more virions than do cells with Vpunegative virus. Virions from Vpu-negative provirus accumulate at the cell surface and in endosomes (Klimkait et al., 1990). The mutant virion particles released are not homogeneous particles, displaying various sizes and shapes, and some of them contain several nucleoids (Klimkait et al., 1990). Vpu complements Vpu-negative mutants in trans, but it has no effect on production of SIV that is devoid of a Vpu counterpart (Terwilliger et al., 1989). In contrast, more recent evidence indicates that the Vpu protein of HIV-1 not only facilitates the release of virus particles from the surface of infected cells, but also enhances capsid release from other retroviruses, such as HIV-2, visna virus, and Moloney murine leukemia virus (Gottlinger et al., 1993).
In addition to enhancing release of virus particles, Vpu also affects cell killing. Less cytopathic Vpu mutants have been isolated and show a decrease in syncytium formation and cell killing in CD4+ human T cells (Terwilliger et al., Kishi et al., 19921, although it is unclear if these mutant viruses contain additional mutations in other genes that are responsible for this phenotype. Apart from these effects on virus release and cell killing, Vpu induces the degradation of CD4 molecules in the infected cells, suggesting multiple roles for Vpu (Willey et al., 1992).
Similarities of Vpu with other amphipathic viral proteins, including influenza M2, have been noted (Maldarelli et al., 1993). Apart from these structural similarities, I would draw attention to the functional resemblance of Vpu to togavirus 6K and influenza M2, because these three proteins are involved in efficient release of virus particles from cells (see below). To my knowledge, there have been no studies on the modification of membrane permeability by the expression of individual Vpu molecules. Such studies would indicate the extent to which Vpu qualifies as a real viroporin molecule.
C . Influenza M2 Protein Forms Ionic Pores
Influenza virus is a negative-sense stranded RNA virus that contains a segmented genome. Segment 7 encodes a t least two proteins (M1 and M2) synthesized from two different mRNAs derived by transcription of segment 7 and differential splicing (Lamb et al., 1981). The structure of M2 resembles that of poliovirus 3A, togavirus 6K and HIV-1 Vpu (see above). M2 is also an integral membrane protein associated with both the Golgi complex and the cell surface Zebedee et al., 1985). In polarized epithelial cells (MDCK), M2, like hemagglutinin HA, is preferentially expressed a t the apical surface in close proximity to budding virions (Hughey et al., 1992). Although early studies indicated that M2 was absent from virions, as with Vpu and togavirus 6K, it was later suggested that M2 is also present in virions (Zebedee and Lamb, 1988;Hughey et al., 1992). M2 contains 97 amino acid residues, with a hydrophobic membranespanning region of about 20 amino acids that may form an amphipathic helix. M2, as well as HIV-1 Vpu, can form homotetramers that, in the case of M2, are disulfide linked (Sugrue and Hay, 1991). In addition to tetramers, higher molecular-weight structures containing 10-12 molecules of M2 are also detected (Sugrue and Hay, 1991). The amino terminus of the M2 protein is located in the extracellular space, whereas the carboxy terminus is cytoplasmic .
M2 is able to modify membrane permeability when expressed in Xenopus oocytes (Pinto et al., 1992) or E . coli cells (Guinea and Carrasco, 1994a). Patch-clamp studies of Xenopus oocytes expressing the natural influenza M2 protein show an increased ionic flux (sodium and potassium) through the plasma membrane (Pinto et al., 1992;. This increased membrane permeability to monovalent cations probably extends to protons and other low-molecular-weight molecules, although direct evidence for higher permeation to other solutes, including protons, has not been presented. Enhanced membrane permeability induced by M2 increased at low pH (6-6.5) in Xenopus oocytes (Pinto et al., 1992). Mutants of M2 with an altered transmembrane region were unable to permeabilize amphibian cells (Pinto et al., 1992). The use of E . coli cells expressing high levels of M2 protein in an inducible manner, as discussed in Section IV,A, indicated that this protein was lytic for prokaryotic cells that express lysozyme, and modified membrane permeability to a number of hydrophilic molecules (Guinea and Carrasco, 1994a). The nonspecific increase in permeability in prokaryotic cells indicates that M2, like other viroporins, forms hydrophilic pores in the membrane. These pores allow the passages of ions and other low-molecular-weight molecules. This membrane disruption was similar to the effects observed in infected cells (Lopez Vancell et al., 1984;Carrasco and Lacal, 1983). Therefore, M2 is a strong candidate to account for the changes in membrane permeability in influenza virus-infected cells. MDCK cells infected with influenza virus show enhanced passive permeability to sodium and potassium ions 5 h r postinfection (Lopez Vancell et al., 1984). Permeability toward to nonpermeating translation inhibitors was observed before the onset of synthesis of influenza virus proteins . A more detailed kinetic analysis of permeability changes in influenza virus-infected cells is required to assess more precisely the nature and the timing of membrane disruption that occur in these cells. Finally, influenza M2 mutants and the individual expression of M2 in mammalian cells can now be exploited to explore the capacity of M2 to promote permeability changes in mammalian cells.
In view of these findings, it is tempting to suggest that M2 is a viroporin that nonspecifically increases membrane permeability and participates in the influenza virus replication cycle as other viroporins do, i.e., by facilitating virus release from the infected cells. An alternative view suggests that M2 has proton translocation activity; as such, M2 would participate at two different steps during the influenza growth cycle, early during infection when the virus enters cells, by promoting the entry of protons into the viral nucleocapsid to destabilize or collapse this structure, and later in the infectious cycle during the synthesis of viral glycoproteins by modulating the pH in the Golgi apparatus (Sugrue et al., 1990a;. In this model, M2 would raise the pH in the Golgi complex, such that the integrity of HAUHA2 during transport through the vesicular system to the plasma membrane would be maintained Takeuchi and Lamb, 1994). As yet, there is no evidence that M2 has proton translocation activity. Rather, this protein nonspecifically increases membrane permeability. Clearly, experiments employing M2-deficient mutants are necessary to assess the exact role played by M2 during the influenza virus life cycle.
Further clues to the action of M2 come from studies with the antiviral compound amantadine. This agent blocks the growth of a number of animal viruses, including some strains of influenza virus (Tominack and Hayden, 1987;. The exact mode of action of amantadine in influenza virus replication remains unclear. It blocks two steps of infection. Early inhibition by amantadine is related to inhibition of virus uncoating . This effect is rather nonspecific because high concentrations of the drug are required and replication of other animal viruses is inhibited by these amantadine concentrations (Nicholson, 1984;Rosenthal et al., 1982Rosenthal et al., , 1988Schulman, 1982;Wittels and Spear, 1991). The molecular basis of this early inhibition by amantadine relies on the accumulation of the drug in endosomes, an event that leads to increased endosomal pH. It was suggested that amantadine binds to the pore region of the channel formed by M2 (Sugrue and Hay, 1991). The M2 protein present in virions would permit the flow of ions, including protons, from the endosome into the virus particle interior to disrupt macromolecular interactions, freeing the nucleocapsid from the M1 protein. This activity of M2 would be blocked by amantadine. More recent experiments suggest that amantadine acts as an allosteric blocker that binds to another portion of M2 not directly involved in the pore region .
At low concentrations amantadine blocks a very late step of influenza virus maturation; viruses are able to bud from the cell surface, but they are unable to pinch off from the cell (Ruigrok et al., 1991). This situation closely resembles HIV-1 viruses defective in Vpu (Klimkait et al., 1990) or togaviruses that lack the 6K protein (Liljestrom et al., 1991). Mutants of influenza virus resistant t o low concentrations of amantadine have been isolated and characterized. They map in protein M2 and, more precisely, in amino acid residues 27-35 (Hay et al., 1986). Inhibition of M2 function by amantadine causes a reduction in the pH of vesicles of the Golgi complex triggering the conformational change in HA (Ciampor et al., 1992a). Interestingly, this effect of amantadine is reversed by monensin (Sugrue et al., 1990a); this is not surprising in view of the idea that viroporins have ionophore-like activity (Carrasco, 1987). Thus, i t was suggested that amantadine brings about a premature conformational change in HA that occurs in the trans-Golgi complex during the transport of HA to the cell surface (Sugrue et al., 1990a). This change in HA would block virus release.
In support of the view that M2 is involved in cytopathicity is the recent finding that expression of M2 increased the cytopathic effect in insect Sf9 (Spodoptera frugiperda) cells infected with recombinant baculovirus. Addition of amantadine to these cells increased the yield of M2 by 10-fold (Black et al., 1993). These results are similar to our findings that expression of mutated poliovirus 3A protein that is unable to permeabilize bacterial cells is enhanced by more than 20-fold (J. Lama and L. Carrasco, unpublished results).
D . Togavirus 6K Protein
Togaviruses are plus-stranded RNA-containing viruses that possess a lipid envelope (Schlesinger and Schlesinger, 1993). During the late phase of Semliki Forest virus or Sindbis virus infection there is a drastic shut-off of host protein synthesis, and only the translation of a viral subgenomic mRNA takes place (Ulug et al., 1987), producing a polyprotein precursor that contains several proteins in the order C-p62-6K-El (Garoff et al., 1980). This precursor is proteolytically cleaved to generate the mature viral proteins: C, the capsid protein, is autocatalytically cleaved and interacts with the genomic RNA to form the nucleocapsid, whereas the rest of the proteins are processed by the cellular signal peptidase (Liljestrom and Garoff, 1991). The p62 precursor is hydrolyzed to generate E3 and E2 that, together with E l , comprise the three glycoproteins present in the lipid bilayer of virion particles (Schlesinger and Schlesinger, 1993;Barth et al., 1992). Togaviruses induce drastic changes in membrane permeability during infection (Carrasco, 1978;Garry et al., 1979;Ulug et al., 1984;Murioz et al., 1985b;Garry et al., 1986;Ulug et al., 1987). These modifications lead to an imbalance of ions in the cytoplasm, thus blocking host protein synthesis Ulug et al., 1987). The translation of Semliki Forest virus subgenomic 26s mRNA is not inhibited by increasing the concentrations of monovalent ions either in uitro or in the infected cells Alonso and Carrasco, 1982a). We have suggested that some of the changes in membrane permeability observed during togavirus infection could be explained by the membrane-active characteristics of the 6K protein (Sanz et al., 1994). This protein is a heavily acylated polypeptide of 60 amino acids that interacts with membranes Lusa et al., 1991). Although the 6K protein is initially associated with p62 and E l and travels through the ER to the plasma membrane as a complex with these proteins, it is not encapsidated in virions (Lusa et al., 1991). Thus, most 6K is excluded during budding of new virions, although some 6K can still be detected in these particles .
Once again the 6K protein shows structural similarities with other viroporins . The togavirus 6K protein is a hydrophobic membrane protein that increases membrane permeability and induces cell lysis in the bacterial model system described in Section IV,A (Sanz et al., 1994). Inducible expression of the Semliki Forest virus protein 6K is very toxic for E . coli, causing increased membrane permeability and cell lysis. Synthesis of 6K led to a sudden release of choline from preloaded cells and to the entry of the hydrophilic antibiotic hygromycin B (Sanz et al., 1994).
The exact role played by the 6K protein in the Semliki Forest virus replication cycle remains largely unknown. Complete removal of the 6K sequence from the genome had no effect on virus RNA replication, gene expression, or glycoprotein transport, but there was a reduction in virus release (Liljestrom et al., 1991). Mutants that contain an underacylated 6K protein also have defects in budding of new virions (GaedigK-Nitschko et al., 1990). Recently, a Sindbis virus 6K mutant containing an in-frame insertion of 15 amino acids has been generated (Schlesinger et al., 1993). This mutant was viable and the virions formed had no detectable 6K, but they were as infectious as wild-type Sindbis virus (Schlesinger et al., 1993). Synthesis of viral proteins occurred in the 6K-deficient mutants, but there was some interference with the proteolytic processing at the junction between p62 and 6K. Notably, this Sindbis virus mutant was defective in its ability to block host translation (Schlesinger et al., 19931, suggesting that synthesis of the 6K plays a part in the shut-off of host protein synthesis by togaviruses. These results agree well with the recent finding that hy-potonic medium reverses the shut-off of cellular translation induced by Sindbis virus (Garry, 1994).
No studies have analyzed the permeabilization of cells during infection with the mutant viruses defective in 6K. In agreement with the idea that the 6K protein modifies membrane permeability, thus facilitating virus release from cells, is the finding that the alterations in ion transport in Sindbis virus-infected cells modulate the release of progeny viruses (Ulug et al., 1989). As may occur with HIV, it is possible that alteration of membrane function by togaviruses is accomplished by several viral proteins, including the 6K protein and some glycoprotein moieties.
E . Other Putative Viroporins
Analyses of Kyte and Doolittle (1982) hydrophobicity plots and the actual hydrophobic sequence of poliovirus 3A indicate that a portion of this protein has an index above 2.5 and a hydrophobic stretch of about 20 amino acids . Close to this sequence there are basic amino acids, whereas acidic residues are not present. In addition to the viral proteins discussed above, i.e., poliovirus 3A, HIV-1 Vpu, influenza M2, and togavirus 6K, several known virus proteins also have such characteristic sequence elements (see Figs. 5 and 6). Some paramyxoviruses, such as respiratory syncytial virus, contain a small hydrophobic (SH) polypeptide that is an integral membrane protein (Olmsted and Collins, 1989). The RSV SH protein is composed of 64 amino acids and contains a very hydrophobic stretch of amino acids in the center of the protein, followed by a basic sequence, typical of membrane-active proteins. The SH protein is expressed at the cell surface (Olmsted and Collins, 19891, but the function of SH in the virus replicative cycle remains unknown. Several species of SH have been distinguished, including one that is glycosylated (Olmsted and Collins, 1989). The protein forms aggregates of various monomers, including species of about 8-10 monomers (Collins and Mottet, 1993). Infectious bronchitis virus, a coronavirus, encodes a small protein of 110 amino acids with a hydrophobic region of about 20 residues (Smith et al., 1990). This protein is membrane associated and is located in the Golgi apparatus and at the cell surface (Smith et al., 1990).
Among DNA-containing viruses several viroporin candidates can be described, based only on sequence data. However, I shall mention only those putative viroporins for which there is at least some experimental evidence for membrane association. Human papillomavirus expresses an E5 protein that is associated with the membrane and interacts with l r r l r l l r r l -2 -1 " 1 -1 -2 u SA-Poll0 1
molecules.
Hydrophobicity plots (Kyte and Doolittle, 1982) of putative membrane-active the pore-forming protein of the vacuolar proton ATPase (Goldstein et Conrad et al., 1993). The E5 proteins from several human papillomaviruses and bovine papilloma virus 1 also contain a very hydrophobic region of amino acids that most probably span the lipid bilayer. Adenoviruses encode a 11.6K protein that is glycosylated and localizes in the Golgi apparatus and the nuclear membrane Tollefson et aZ., 1992). This protein is of typical size and contains the hydrophobic region and basic amino acids of viroporins (Figs. 5 and 6). As yet, there is no evidence that these proteins are involved in altering membrane function.
A . Viroporin Structure
With the increasing availability of viral genome sequences, the putative viroporins encoded by animal viruses may eventually be predicted. Moreover, the ease of genetic manipulations that allow individual cloning and expression of these proteins will promote more interest in elucidating the mechanisms of action of this group of viral proteins. In a few years, further work in this field should provide us with a more complete picture of viroporin activity in the infected cells and could provide additional targets of antiviral intervention.
Although we are only beginning to glimpse the existence of this new family of animal virus proteins that are involved in cell lysis, some generalities of their structure and function can be summarized. Structurally, viroporins are in general short proteins containing around 50-120 amino acid residues. Analysis of the amino acid content of several viroporins shows that they contain a higher than average content of leucine plus isoleucine residues and a lower content of glycine. All of virporins possess a hydrophobic stretch of about 20 amino acids, which in some instances may form an amphipathic helix. Thus, they are integral membrane proteins with at least one membrane-spanning domain. Viroporins tend to form oligomers, most frequently tetramers, thus leaving a hydrophilic pore through which ions and low-molecularweight hydrophilic compounds pass in a nonspecific manner. In addition to tetramers, other oligomeric structures containing multiple viroporin molecules may be formed. In addition to the hydrophobic region, viroporins usually contain basic amino acids, as seen with membrane-active toxins (see Section VI). These regions of basic amino acids may participate in membrane permeabilization by destabilizing the lipid bilayer. In summary, the structure of viroporins allows formation of oligomeric structures spanning the membrane to form hydrophilic pores.
B . Viroporin Function
The main function of viroporins is to help in the release of progeny virus from the infected cells. Spread of virus to other surrounding cells or tissues is facilitated by cell lysis. Apart from this function, it seems logical to think that the expression of proteins that lead to cell lysis would also have detrimental effects for the infected cells, even before lysis occurs. These effects could be manifested as modifications in normal cell morphology and metabolism, including the synthesis of cellular macromolecules.
It may still be argued by some virologists that changes in membrane permeability are the result of the exit of whole virus particles from cells. I would stress that the opposite view is more plausible, i.e., that viruses exist the cell because the membrane is being lysed by specific membrane-active proteins encoded by the viral genome. Therefore, a new concept arises-namely, that animal viruses contain in their genomes one or more genes specifically devoted to lysing the plasma membrane. This may be particularly important for viruses that possess naked virion particles. In order to leave the cells, such viruses, which are devoid of a lipid membrane, need to lyse the cell membrane. In the case of enveloped virion particles, this envelope is acquired in the last step of assembly, by budding from the plasma membrane, which is modified by a specific viral product that produces pores. The increased membrane permeability due to pore formation would be required for efficient budding of enveloped viruses. In the absence of viroporins animal viruses are still made, but the number of virus progeny is greatly diminished. Little is known about the forces that induce a virus particle to bud and be released from the cell. Membrane budding could also be regulated by alterations that virus-infected cells suffer in calcium homeostasis. Apart from calcium, the levels of other cations and pH also change drastically during infection. Several years ago, it was indicated that the ionic milieu in the cytoplasm of a virusinfected cell during virus assembly is quite different from that in the early stages of infection, when the virus decapsidates (Carrasco, 1977). Virus assembly is known to be affected by cations and pH in uitro, so why not also in the infected cell?
Therefore, viroporin action a t the molecular level may be directed to deenergizing the membrane and disrupting its physical integrity. Formation of pores leads to the dissipation of the membrane potential and the ionic gradients that would be gradually destroyed as infection progresses. Not only the ionic gradients maintained by the plasma membrane, but also those present in intracellular compartments, including the proton gradient, would disappear. The physical integrity of the membrane could be disrupted initially by the formation of hydrophilic pores and by the insertion in the membrane of viroporin regions containing basic amino acids that destabilize the lipid bilayer in a detergent fashion. The focal distribution of viroporins in the bilayer would promote preferential virus release through these regions. Subsequently, stimulation of phospholipases may complete the physical disappearance of the lipid bilayer in certain regions of the plasma membrane, leading to cell lysis.
Finally, viroporin activity may also have consequences for cellular metabolism and morphology, as occurs with membrane-active toxins that affect an array of cellular functions by simply modifying membrane function. Viroporins could affect cellular metabolism, including macromolecular synthesis. It was initially proposed that the modifications in membrane permeability to monovalent ions in virus-infected cells is responsible for the virus-induced shut-off of host translation (Carrasco and Smith, 1976;Carrasco, 1977). Overwhelming evidence indicates that this may be the case of EMC virus-infected cells. Togaviruses block host translation by a similar mechanism Ulug et al., 1984Ulug et al., , 1987Munoz et al., 1985a). The translation of cellular mRNAs will be rapidly inhibited by the action of the lytic protein, whereas viral RNA translation tolerates these adverse conditions and is still translated. Perhaps shut-off of host translation is a by-product of a viral process, such as cell lysis, a requirement to liberate the virus particles made during infection.
All experimental evidence on the potential activity of viroporins has been obtained with whole, wild-type viruses; few virus mutants have been assessed for viroporin function. As yet, only the togaviruses defective in 6K and HIV mutated in Vpu have been examined. Furthermore, effects of the individual expression of viroporins in mammalian cells are still an unexplored field. At present we have evidence that the individual expression of poliovirus 3A, SFV 6K, and influenza M2 in E . coli and amphibian cells modifies membrane permeability in a nonspecific way. The inducible prokaryotic system described in Section IV,A promises to be of great value not only for identifying other putative viroporins, but also for obtaining mutants to explore the portions of the molecule most relevant to viroporin activity. The identification of pore formation by M2 expressed in amphibian cells constitutes another promising model system to characterize the animal virus proteins that modify membrane permeability.
C . Pore-Forming versus Ion-Channel-Forming Proteins
Resolution of three-dimensional structures of proteins, and more particularly of integral membrane proteins, has revealed the architec-ture of some proteins that form pores and ionic channels in membranes (Eisenman and Dani, 1987;Aschroft and Roper, 1993;Cowan, 1993;Catterall, 1993;Imoto, 1993;Hoshi and Zagotta, 1993;Massotte et al., 1993). The most common feature of proteins that form ion channels is the presence of domains containing a-helices that span the membrane (Eisenman and Dani, 1987;Eisenberg, 1984;Lear et al., 1988;Imoto, 1993). The association of several amphipathic helices builds up a pore in such a way that the hydrophobic region of the a-helix is in contact with the membrane lipids, whereas the hydrophilic residues form part of the hydrophilic pore lumen (Oblatt-Montal et al., 1993;DeGrado et al., 1989;Segrest et al., 1990). Visualization of botulinum neurotoxin-induced channels in vesicles has been achieved (Schmid et al., 1993). The channel is formed by interaction of four neurotoxin molecules. Pores can also be formed by amphipathic p-sheets, which have a tendency to associate and can traverse the membrane if they contain a t least 10 amino acids. The best-known members of this group of membrane proteins are the bacterial porins (Cowan, 1993;Schulz, 1993), which are proteins located in the outer membrane of gram-negative bacteria and which allow the passive passage of solutes (Cowan, 1993). The antiparallel p-barrel structure of porins allows their insertion into the membrane, leaving a pore that can be gated by other regions of the molecule (Cowan, 1993; Schulz, 1993;Weiss and Schulz, 1993). Lymphocyte perforins may also form similar structures responsible for cell lysis (Lichtenheld et al., 1988;Shinkai et al., 1988;Lowrey et al., 1989;Ojcius et al., 1991). Selective ion channels present a funnel-shaped architecture, with at least one narrow region within the channel pore, where ion selection takes place (Eisenman and Dani, 1987;Imoto, 1993). In contrast to this selectivity of ion channels, proteins involved in cell lysis allow the nonspecific passage of ions and of low-molecular-weight hydrophilic compounds, with size being the only constraint on permeability.
With respect to viroporins, it is of interest to point out the identification of a rather small cellular protein containing only 130 amino acid residues; this protein is involved in voltage-gated potassium channels (Takumi et al., 1988;Hausdorff et al., 1991;Murai et al., 1989). It differs from other known ion channel proteins; it is still unknown if this protein induces membrane depolarization by forming aggregates in membranes, but it constitutes a good model system to analyze fundamental aspects of cellular proteins endowed with ion channel activity.
In conclusion, ion channels have been defined as aqueous pores through which selected ions can move from one side of the membrane to the other. These channels are usually selective for a given ion and control the flux of ions by gating this pore in response to various factors. Ion channels are integral components of the membrane involved in the normal functioning of cells. Their channel-gating characteristics ensure that the ion channel remains open or closed, depending on the cellular functional state. Thus, ion channels are nonlytic for cells.
Membrane pores are discussed here in the context of their relationship to cell lysis. They are formed by alien proteins, whose ultimate function is to disorganize the membrane and kill the cell. Pores formed by these lytic proteins do not have a determined pore size, which may vary according to the number of proteins that interact to form a given pore. In this sense the formation of pores is plastic. The initial pore is formed by oligomerization of a determined number of pore-forming proteins. However, as infection progresses and the number of viroporins in the plasma membrane increases, other proteins can be recruited to widen the pore size. Pores are not selective with regard to the molecules that can pass through them. Ions, in the first instance, because of their smaller size, and low-molecular-weight compounds can pass through the pores. As the pore size increases, however, macromolecules, including enzymes, may leak through them. This is exactly the phenomenology observed in animal virus-infected cells: initially the permeability of ions and of low-molecular-weight compounds increases in both directions, and finally enzymes leak out of the infected cells.
VI. OTHER LYTIC PROTEINS OF NATURAL ORIGIN
It is beyond the scope of this work to review in depth the different proteins that possess lytic activity. However, detailed studies accomplished with lytic agents can greatly benefit our understanding of the action of viroporins a t the molecular level.
A . Lytic Proteins from Bacteriophages
Identification of phage proteins involved in the lysis of bacteria is far more advanced than is the case for analogous proteins from animal viruses (Young, 1992). Extensive studies have been carried out on the molecular basis of phage-induced lysis in bacteria. In all cases analyzed, the bacteriophage encodes a specific protein responsible for lysing the bacterial membrane (Young, 1992). These lytic proteins should not be confused with lysozyme, an enzyme also encoded by some phages (and by cells) that is able to digest peptidoglycans, and hence the bacterial cell wall. Some virologists still may have the idea that lysis of bacteria by phages is due to lysozyme activity. In fact, bacterial lysis by phage infection results from permeabilization of the lipid bilayer by specific lytic proteins, thus allowing lysozyme to act on the external cell wall (Garret and Young, 1992).
These proteins are integral membrane proteins and lyse bacteria by making holes; for this reason they were christened "holins" (Young, 1992). There is a strong structural (and perhaps functional) similarity between holins and viroporins (see Figs. 5 and 6). At present there is evidence that about seven proteins may act as holins. These proteins are S and S 2 1 from A phage and the lambdoid phage 21, respectively; protein t from phage T4; gp17.5 from T7 phage; LydA encoded by phage P1; the 429-encoded p14; and p10 from phage $6 (Young, 1992). As with viroporins, these proteins are unrelated in terms of primary structure, but they have in common one or two putative membranespanning domains of hydrophobic amino acids. It has been speculated that the two transmembrane domains can interact with each other to form a hole in the membrane. Perhaps, as in the case of viroporins, holins can form multimeric aggregates to generate pores in the membrane, even if they possess a single transmembrane domain. Bacteriophage lysis can be prevented without inhibiting the accumulation of intracellular phage particles (Josslin, 1971). Phospholipase activity occurs as a consequence of hole formation by phage T4, suggesting a role for lipases in lysis (Bennett et al., 1971;. However, phage release still occurs even if phospholipase activation is inhibited (Hardaway et al., 1975).
For historical and technical reasons, phage lysis has been more frequently studied in temperate phages, particularly in lambdoid phages. The proteins encoded by gene S of phage A and phage 21 contain two membrane-spanning domains separated by a p-turn region with a charge-rich carboxy terminus (Young, 1992). The protein encoded by gene S (107 codons) is a basic 8.5-kDa protein (Altman et al., 1983).
According to a proposed model, 12-15 S monomers could participate in the formation of a hole with a 5-nm diameter, allowing the passage of small proteins. Theoretical calculations indicate that the number of holes per cell may be as low as 200, making their detection by electron microscopy difficult (Young, 1992). In the absence of S gene function, the product of gene R, a transglycosidase that degrades the peptidoglycan of E . coli, accumulates intracellularly, without lysis or apparent cytotoxicity (Reader and Siminovitch, 1971). Expression of the S gene in the absence of activity of gene R (S+ Rphage) does not lead to cellular lysis by phage, but permeabilization of the membrane still occurs, leading to inhibition of host macromolecular synthesis and cell death (Reader and Siminovitch, 1971;Garret and Young, 1992). Ex-pression of the S lysis gene in Saccharomyces cerevisiae led to its accumulation in the plasma membrane and cytotoxicity (Garrett et al., 1990), further indicating that activity of holins or viroporins can be observed in both prokaryotic and eukaryotic cells (see Section IV,A).
The nature of the lesion produced by holins in the cytoplasmic membrane has been analyzed in some detail. As in animal virus-infected cells, a nonspecific increase in membrane permeability is observed in both directions across the membrane. Even low-molecular-weight proteins can cross the damaged membrane; lysozyme leaks out, whereas RNase I can enter the cytoplasm (Young, 1992). This situation resembles infection of mammalian cells by picornaviruses, wherein different molecules cross the membrane in both directions . Even the entry of protein toxins can be detected to a small extent, suggesting that the pores formed are heterogeneous in size or that small proteins can, to a limited extent, cross damaged membranes by unknown mechanisms. The efficiency of membrane penetration by a given compound would depend primarily on the size and the chemical characteristics of the molecule, whereas detection of its passage depends on the sensitivity of the assay employed. Thus, labeled ions or sugars can easily be shown to cross a leaky membrane, whereas no protein traffic is detected with labeled protein, unless enzymatic action is assayed.
Some genetically simpler phages, such as 4x174 or MS2, lyse cells without the participation of a phage-encoded lysozyme activity. In this case, too, the action of a single species of holin protein is required. The general features of these proteins conform to the features integral membrane proteins, with a highly hydrophobic stretch of amino acids (see Figs. 5 and 6). Digestion of the bacterial cell wall is most probably accomplished by cellular enzymes.
Bacteriophage 4x174 contains gene E , which is involved in cell lysis (Barrel1 et al., 1976). As with viroporins from animal viruses, the 91amino acid protein contains a transmembrane domain and several regions of basic amino acids. E protein is also able to form oligomers (Blasi et al., 1989). Inducible expression of E from a plasmid leads to a sudden collapse of Na+ and K4+ gradients with concomitant cell lysis (Witte et al., 1987;Witte and Lubitz, 1989). The formation of transmembrane tunnels has been observed by electron microscopy (Witte et al., 1990). Curiously, there is only one E hole per cell in a specific location of the cell envelope. It remains to be demonstrated that a similar structure appears in 4Xl74-infected cells.
RNA phages such as MS2 or f2 contain an L gene that encodes a membrane-bound protein (Model et al., Beremand and Blumenthal, 1979). L protein is involved in cell lysis, allowing the release of viral particles. The structure of L is again typical of a virus-encoded membrane-active protein. It contains 75 amino acids with a hydrophobic stretch of about 27 residues and a region of basic amino acids at the amino terminus. Even after deletion of a 40-amino acid N-terminal region, the L protein is competent for lysis (Berkhout et al., 1985). This led to the synthesis of a number of peptides comprising the hydrophobic region of L; these peptides are lytic for bacteria, dissipate the ionic gradients, and increase membrane permeability in model membrane vesicles (Goessens et al., 1988). L protein localizes in special cellular structures, the membrane adhesion sites, to form defined pores (Walderich and Holtje, 1989). Defects in lytic genes from bacteriophages lead to a n impairment in release of virus particles, whereas the synthesis of viral macromolecules and virion particles is normal, or even increases.
B . Nonviral Lytic Proteins from Natural Sources
To kill cells, nature has evolved compounds targeted to destruction of the plasma membrane. Apart from ionophores, a number of protein toxins act directly on membranes. This is the case, for example, for melittin (Bernheimer and Rudy, 1986;Dempsey, 19901, bacterial hemolysins (Felix, 1982;Suttorp et al., 19931, magainin (Bevins and Zasloff, 1990), cecropins van Hofsten et al., 1985), attacins (Hultmark et al., 19831, perforins (Podack, 19921, complement (Podack, 1986;Muller-Eberhard, 1988), defensins (Ganz et al., 1990;Cociancich et al., 1993), protegrins (Kokryakov et al., 19931, and Entarnoeba histolytica pore-forming peptide (Leippe et al., 19921. Understanding the mode of action of the increasing number of proteins discovered to lyse cells is simplified by analysis of two paradigmatic peptides: melittin, which modifies membrane permeability by disrupting the lipid bilayer of membranes, and alamethicin, a peptide able to form aggregates that span the membrane-forming aqueous pores. It is likely that viroporins, or some moieties of viral glycoproteins, could injure the plasma membrane by similar mechanisms. Moreover, some viroporins may contain different domains, one involved in pore formation and another in membrane disruption. The reader is referred to reviews on permeabilizing agents for additional details on these compounds (Felix, 1982;Bernheimer and Rudy, 1986;Kaiser and Kezdy, 1987;Vaara, 1992).
Melittin is the main toxic component of honey bee venom (Dempsey, 1990). It is composed of 26 amino acids, of which residues 1-19 are very hydrophobic and czn form a n amphipathic a-helix, whereas the domain containing amino acids 20-26 is very basic. Both domains are required for efficient lysis (Dempsey, 1990). Several agents, including divalent cations, have been reported to protect against the lytic effects of melittin and other toxins (Bashford et al., 1989). However, this finding has not been reproduced by others (Portlock et al., 1990). Melittin has been used in numerous studies as a model to investigate the interaction of peptides and membranes in order to elucidate the molecular basis of lytic activity. The action of melittin on membranes is complex and depends on a number of parameters, including the composition of the membrane, the membrane potential, and the melittin concentration. On interaction of melittin with membranes, there is an increase in ion conductance, membrane depolarization, membrane fusion, phospholipase activity, etc. (Dempsey, 1990). The initial effects of melittin on membranes may include the formation of pores that are permeable to ions but still impermeable to proteins (Dempsey, 1990). These pores probably form discrete channels before the more generalized effect on membrane lipid disorganization takes place. This leads to the breakdown of membranes into micelles, similar to the action of some detergents.
Alamethicin, like gramicidin A, is a linear peptide antibiotic that inserts into memb 'anes and produces voltage-gated ion channels. The major component of alamethicin is composed of a peptide containing 20 amino acids; it forms an amphipathic a-helix, leading to aggregates that form pores in the membrane (Fox and Richards, 1982;Kaiser and Kezdy, 1987). Recently, high-resolution analysis of dimeric gramicidin A inserted into lipid bilayers has revealed a structure that is consistent with channel formation by this peptide (Ketchem et al., 1993).
Melittin and gramicidin S, a cyclic peptide antibiotic, stimulate the release of phospholipids from erythrocyte membranes at concentrations that modify membrane permeability, indicating that membrane moieties are released with consequent enhancement in permeability (Katsu et al., 1989). The detailed molecular actions of melittin and alamethicin that are responsible for lysis may be different (Portlock et al., 1990).
VII. FUTURE PROSPECTS
The modifications in membrane permeability produced during virus infection are now clearly defined. Early membrane permeabilization is caused by the viral particles during the process of virus entry and uncoating. Insights into the molecular mechanisms of this process suggest that animal viruses require a proton-motive force to enter cells and to permeabilize them. Future work in this field would benefit from the use of selective inhibitors of vacuolar proton ATPase and from experiments aimed at elucidating the exact role that the pH gradient and/or the membrane potential play in the process of virus entry.
Late membrane leakiness requires virus gene expression and is manifested as a general increase in permeability to ions and lowmolecular-weight compounds, suggesting that physical pores in the membrane are generated during virus infection. These pores would arise by oligomerization of particular viral proteins designed to lyse the infected cells. Efforts should be made to understand the structural characteristics of these aqueous pores.
Knowledge gained in the past few years on animal virus proteins that enhance membrane permeability is delineating a new family of proteins, the viroporins, which are necessary for efficient exit of virus particles from infected cells. The general structural characteristics of these viroporins indicate that they are relatively small (40-120 residues) basic proteins that usually contain a hydrophobic region of about 20 amino acids. It can be anticipated that the future characterization of new viroporins will give a better idea of how general this structure is among virus proteins that are endowed with lytic activity.
Apart from the function of viroporins in the virus growth cycle, the way is now open to analyze the effects that these proteins have on different cellular functions, such as inhibition of host macromolecular synthesis and the cytopathic effect. At least two avenues of research could be followed. One is the cloning and expression of individual viroporins to assay their effects on the host cell. The second approach is to generate viruses with mutant viroporin function. Characterization of these viruses and the actions of the variant viroporins in cellular metabolism would provide more definitive answers to the questions concerning membrane modification and the inhibition of host macromolecular synthesis by virus infection.
|
2019-03-22T16:13:30.994Z
|
1995-01-01T00:00:00.000
|
{
"year": 1995,
"sha1": "f4d026ddab3ec8f7ae3e460a6515ee758e4b9d34",
"oa_license": null,
"oa_url": "https://doi.org/10.1016/s0065-3527(08)60058-5",
"oa_status": "BRONZE",
"pdf_src": "PubMedCentral",
"pdf_hash": "9a196f0dd2b2b9ca9f489a9b3185c6e6d9ee48ea",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Biology"
]
}
|
18565297
|
pes2o/s2orc
|
v3-fos-license
|
Development of a physiologically based pharmacokinetic model of trichloroethylene and its metabolites for use in risk assessment.
A physiologically based pharmacokinetic (PBPK) model was developed that provides a comprehensive description of the kinetics of trichloroethylene (TCE) and its metabolites, trichloroethanol (TCOH), trichloroacetic acid (TCA), and dichloroacetic acid (DCA), in the mouse, rat, and human for both oral and inhalation exposure. The model includes descriptions of the three principal target tissues for cancer identified in animal bioassays: liver, lung, and kidney. Cancer dose metrics provided in the model include the area under the concentration curve (AUC) for TCA and DCA in the plasma, the peak concentration and AUC for chloral in the tracheobronchial region of the lung, and the production of a thioacetylating intermediate from dichlorovinylcysteine in the kidney. Additional dose metrics provided for noncancer risk assessment include the peak concentrations and AUCs for TCE and TCOH in the blood, as well as the total metabolism of TCE divided by the body weight. Sensitivity and uncertainty analyses were performed on the model to evaluate its suitability for use in a pharmacokinetic risk assessment for TCE. Model predictions of TCE, TCA, DCA, and TCOH concentrations in rodents and humans are in good agreement with a variety of experimental data, suggesting that the model should provide a useful basis for evaluating cross-species differences in pharmacokinetics for these chemicals. In the case of the lung and kidney target tissues, however, only limited data are available for establishing cross-species pharmacokinetics. As a result, PBPK model calculations of target tissue dose for lung and kidney should be used with caution.
Introduction
Physiologically based pharmacokinetic (PBPK) modeling is widely held to be a useful methodology for improving the accuracy of chemical risk assessment (1)(2)(3)(4)(5)(6). The goal of PBPK modeling is to simulate the uptake, distribution, metabolism, and elimination of a chemical in an organism, using as realistic a description of the relevant physiology and biochemistry as is necessary and feasible (7)(8)(9)(10). For its use in risk assessment, PBPK modeling attempts to describe the relationship between external measures of exposure (e.g., amount administered or concentration in air) and internal measures of biologically effective dose (e.g., amount metabolized or concentration of an active metabolite in the tissue displaying the toxic response) in both the experimental animal and the human (11,12).
Simple pharmacokinetic approaches have occasionally been used by regulatory agencies in risk assessment; for example, the most recent U.S. Environmental Protection Agency (U.S. EPA) cancer risk estimates for trichloroethylene (TCE) were derived using estimates of metabolized dose (13,14). The recent U.S. EPA guidelines for the application of inhalation dosimetry in the derivation of inhalation reference concentrations (15) also make use of pharmacokinetic principles. However, the only case to date where a regulatory agency has used a full PBPK approach in a published risk assessment was in the U.S. EPA's latest revision of its inhalation risk assessment for methylene chloride (16). The decision to use the PBPK approach in this case was made only after a period of considerable controversy, including a workshop sponsored by the National Academy of Sciences at which the usefulness of PBPK modeling for chemical risk assessment was discussed. The scientific consensus following the workshop was that "relevant PBPK data can be used to reduce uncertainty in extrapolation and risk assessment" (1). In 1989, after a detailed multiagency evaluation of the available PBPK information and a review by the U.S. EPA Scientific Advisory Board, the U.S. EPA revised the inhalation unit risk and riskspecific air concentrations for methylene chloride in its Integrated Risk Information System (IRIS) (17) database, citing the PBPK model of Andersen et al. (18). The resulting risk estimates were lower than those obtained by the default approach by nearly a factor of 10. Application of the PBPK model for methylene chloride in a cancer risk assessment for occupational exposure has also been described (19)(20)(21), and a modified version of the model was used by the Occupational Safety and Health Administration (OSHA) in their rulemaking for methylene chloride (22).
Risk assessments using PBPK models have also been proposed for many other chemicals, including not only TCE (23)(24)(25)(26), but also perchloroethylene (PERC) (24,27,28), ethylene dichloride (29), vinyl chloride (26,(30)(31)(32), dioxane (33,34), chloroform (35), benzene (36), and ethyl acrylate (37). However, apart from the case with methylene chloride described above, there still have been no risk assessments published by a regulatory agency in which a PBPK model was used for estimating target tissue dose. Part of the reason for the slow progress of incorporating PBPK modeling in cancer risk assessment is the concern of regulatory agency risk assessors about uncertainties in its implementation. The potential impact of uncertainty in pharmacokinetic risk assessment has been a subject of some controversy (19)(20)(21)(38)(39)(40)(41)(42)(43). The purpose of the study reported here was to develop a PBPK model for TCE that included as complete a description as possible of all of the metabolites and target tissues that are relevant to the toxicity and carcinogenicity of TCE and to evaluate the suitability of the resulting model to provide dosimetry for each of the target tissues in support of a comprehensive pharmacokinetic risk assessment for TCE. For completeness, aspects of both the cancer and noncancer risk assessment contexts pertinent to the development and evaluation of the PBPK model are discussed in this article. However, a companion article in this same issue provides a detailed description of the application of the PBPK model in a noncancer risk assessment for TCE (44). Therefore, the discussions in this article will focus primarily on the cancer end points.
Toxicity and Carcinogenicity ofTCE TCE produces noncancer toxicity in a variety of tissues; principal noncancer end points include neurological, hepatic, renal, immunological, and developmental effects (45)(46)(47). The American Conference of Government Industrial Hygienists (ACGIH) threshold limit value (TLV) for TCE is currently 50 ppm as an 8-hr time-weighted average (TWA), based on central nervous system (CNS) effects (headaches, fatigue, irritability), with a short-term exposure limit of 200 ppm to protect against its anesthetic effects (48). In 1989, as part of a major rulemaking that promulgated standards for more than 200 chemicals, OSHA adopted a permissible exposure level (PEL) for TCE of 50 ppm as an 8-hr TWA based on the ACGIH TLV (49). However, since the entire 1989 rulemaking has now been overturned by the courts, the PEL for TCE has returned to its previous value of a 100-ppm 8-hr TWA, also based on CNS effects. TCE was widely used in industry for many years because of its relatively low toxicity, its excellent solvent properties, and its nonflammability. In recent years, however, use of TCE has been greatly reduced due to concerns regarding its carcinogenicity (46,47). Nevertheless, the question of the human carcinogenicity of TCE remains controversial. Although a large number of studies have demonstrated tumors in animals following exposure to TCE, the relevance of these animal results to the question of the human carcinogenicity of TCE has frequently been challenged (50)(51)(52). The ACGIH has classified TCE into Group A5, not suspected as a human carcinogen, based on a well-conducted, negative epidemiological study performed in an aircraft maintenance facility at Hill Air Force Base (Ogden, UT) (53,54). The International Agency for Research on Cancer (IARC), on the other hand, classifies TCE into Group 2A, probably carcinogenic to humans, based on their assessment of sufficient data in animals and limited data in humans (55). The human evidence considered significant by IARC was the consistency of an association of TCE exposure with slightly increased incidences of liver/biliary tract tumors and non-Hodgkins lymphoma in studies of three cohorts in the United States (53), Sweden (56), and Finland (57), despite the fact that all three studies were characterized as negative by the original investigators because the increases were not statistically significant.
Requirements for a PBPK Model to Support TCE Risk Assessments Quantitative cancer risk estimates for TCE are currently based on animal bioassays, specifically liver and lung tumors in mice. In 1983, the U.S. EPA calculated unit risks for TCE of 4.1 x 106 (pg/m3)-1 for inhalation and 0.54 x 10-(pg/L)-1 for drinking water using data on the incidence of liver tumors in male B6C3F, mice given TCE in an oil vehicle by gavage (58,59); the linearized multistage model (60) was used with a calculation of absorbed TCE dose scaled by body surface area (BSA) to obtain these estimates (61). In 1985, lower unit risks of 1.3 x 106 (pglm3)-1 for inhalation and 0.32 x 10-6 (jg/L)1' for drinking water were recalculated on the basis of the same oral bioassays, using the results of pharmacokinetic studies (50,62,63) to calculate total metabolized dose in both animals and humans, rather than absorbed dose. The BSA adjustment was still applied to obtain the human equivalent dose (13). In 1987, the U.S. EPA calculated a new inhalation unit risk of 1.7 x 10-(ig/m3)-1 based on the incidence of mouse liver and lung tumors in inhalation bioassays (64)(65)(66), again using a calculation of metabolized dose and the BSA adjustment (16). Statistically significant increases in the incidence of renal tubular cell adenoma and carcinoma have also been observed in male Fischer 344 (F344) rats exposed to TCE by gavage (67). Although not yet used in a quantitative risk assessment for TCE, the incidences of these kidney tumors in rats have raised concern, since they represent a rare tumor that has also been associated with human occupational exposure to TCE (68).
For each of these three target tissuesliver, lung, and kidney-there is evidence that the carcinogenicity of TCE may be associated with one or more of its metabolites: trichloroacetic acid (TCA) and dichloroacetic acid (DCA) in the liver (69,70), chloral (CHL) in the lung (71), and 1,2-dichlorovinylcysteine (DCVC) in the kidney (72). Thus, to be useful in a comprehensive cancer risk assessment for TCE, a PBPK model should include at least three target tissues-liver, lung, and kidney-along with a description of the kinetics of the metabolites imputed to play a role in the carcinogenic activity.
Several target tissues have also been identified for the noncancer toxicity of TCE, including the liver (73), kidney (65,66), CNS (74), immune system (75,76), and developing fetus (77). As in the case of the carcinogenicity of TCE, several of these noncancer end points appear to be associated with exposure to the metabolites of TCE rather than to the parent chemical itself (73,78). In addition to the metabolites mentioned above with regard to the carcinogenicity of TCE, it was felt (43) that a PBPK model developed to support a noncancer risk assessment for TCE should also include a description of the kinetics of trichloroethanol (TCOH), a major metabolite of TCE that has been suggested to be responsible for the observed neurological effects of chloral hydrate (79).
Previous PBPK Modeling ofTCE
A number of PBPK models have been developed for TCE. However, most are only parent chemical models; that is, they provide a pharmacokinetic description ofTCE itself, but do not include an explicit description of the pharmacokinetics of any of the metabolites (23,24,(80)(81)(82)(83)(84)(85)(86). These models have been used successfully for predicting TCE concentrations in the blood and tissues (86), for calculating the respiratory input from inhalation exposures (81,82,85), and for investigating the impact of variations in the physiological or biochemical parameters on the kinetics ofTCE (80,83,84). Parent chemical models have also been employed to calculate total metabolized dose in support of a cancer risk assessment for TCE (23,24). However, these parent chemical models cannot be used for predicting tissue exposure to specific metabolites.
Models of both TCE and its metabolites have also been developed. In a series of publications, Sato and co-workers have described the use of a simple PBPK model to study the kinetics of TCE and its metabolites in humans (87), to evaluate the impact of changes in physiological factors (88) and environmental factors (89) on the kinetics of TCE in the human, and to predict the effects of interactions with ethanol consumption on TCE kinetics (90). However, the structure of these models would not support the animalto-human extrapolation of pharmacokinetic dose metrics needed for risk assessment. Fisher and co-workers developed a PBPK model for TCE and its principal metabolite, TCA, in the pregnant (91) and lactating (92) rat, as well as in the mouse (93). These rodent models, together with a similar model of TCE and TCA in the human (94), served as the basis for a PBPK-based risk assessment for TCE liver carcinogenicity (25) based on either total metabolism or AUC for TCA. These models provided the first successful cross-species pharmacokinetic description for a metabolite of TCE. The model development performed in the current study builds on the work of Fisher and Allen (25) by adding limited descriptions of additional metabolites (TCOH, DCA, CHL, DCVC) and target tissues (lung and kidney).
Metabolism ofTCE
The following discussion summarizes the experimental evidence for the nature of the pharmacokinetics and metabolism of TCE, which formed the basis for the decisions that were made regarding the structure and parameterization of the PBPK model. TCE is a volatile, lipophilic chemical that distributes readily throughout all tissues, including the brain, but partitions preferentially into fat tissue. In contrast, its major metabolite, TCA, is a water-soluble chemical that preferentially distributes into the plasma and richly perfused organs and is found only in relatively lower concentrations in the musde and fat. The properties of TCOH are somewhat intermediate between the other two compounds (95). Clearance of TCE occurs both by exhalation and by metabolism. A schematic of the metabolic pathways for TCE is shown in Figure 1. The major oxidative pathway, which takes place primarily (but not exclusively) in the liver, is shown to the right of the diagram; the minor glutathione-dependent pathway, which involves several locations induding the liver and kidney, is shown to the left.
Oxidative metabolism. The primary route of metabolism for TCE, shown on the right side of the diagram in Figure 1, is oxidation via the microsomal mixed-function oxidase (MFO) system, now referred to as cytochrome P450, or CYP (96)(97)(98)(99)(100)(101)(102). A minor pathway for TCE metabolism involving conjugation with glutathione (GSH) by glutathione transferase (GST) has also been observed (103). This pathway, which is shown on the left side ofthe diagram, will be described in the section on conjunctive metabolism.
The principal oxidative metabolite formed in vitro is CHL (96,97,100), which is subsequently reduced to TCOH in the cytosol or oxidized to TCA in either the cytosol or mitochondria (96). CHL is not stable in vivo, and circulating concentrations are relatively low compared to its breakdown products, TCA and TCOH (50). Within a few hours of the administration of 50 mglkg chloral hydrate to a child, the rapid initial dearance of CHL was essentially complete. Subsequent blood concentrations parallel the time course for TCOH but are approximately an order of magnitude lower, suggesting a continuing production of CHL from TCOH (104).
The principal circulating metabolite of TCE in the blood is TCA, which accumulates in the body due to protein binding (105) and slow excretion (106), whereas TCOH is readily excreted (107,108). TCA appears to be derived both directly from CHL and indirectly from TCOH (107)(108)(109). TCE is much more extensively metabolized in the mouse than in the rat, whether TCE is administered orally (50) or by inhalation (91).
Based on both in vitro and in vivo studies, the metabolism ofTCE has been suggested to consist of oxidation of TCE to CHL by the MFO system, followed by either oxidation of CHL to TCA by an aldehyde oxidase (also known as chloral dehydrogenase) or reduction to TCOH by alcohol dehydrogenase (ADH) with subsequent glucuronidation. Oxidation of TCOH to TCA by the MFO system has also been proposed (110 description, oral co-administration of ethanol inhibited the metabolism of inhaled TCE to TCOH by more than 50% and the production of TCA was essentially abolished while ethanol was present (110). A similar study in rats demonstrated qualitatively similar, but quantitatively much less remarkable, effects of ethanol co-administration on the kinetics of orally administered TCE (111).
Inhalation exposures of human volunteers to TCE concentrations from 27 to 201 ppm showed no evidence of metabolic saturation or of a change in the proportion of TCA to TCOH (112). Saturation of TCE metabolism has been observed in mice, rats, and dogs (66,113). The relative proportion of the major metabolites does not appear to be a strong function of dose. However, repeated dosing does appear to increase the production of TCA at the expense of TCOH (114), and the relative production of CO2 increases with increasing dose in mice (115).
Human in vivo studies with TCE (116) have identified the major urinary metabolites to be TCOH (50% of the administered TCE dose), primarily as the glucuronide, and TCA (19%). Monochloroacetic acid (MCA) was also identified as a minor metabolite (4%) in these studies. Another minor metabolite, N-(hydroxyacetyl)-aminoethanol, has also been identified in human (and rodent) urine following TCE exposure, and TCE-derived oxalic acid has been detected in the rodent (117). A study of TCE metabolism in nonhuman primates (118) found that TCA was partially excreted as the glucuronide, particularly at longer times after dosing. The authors suggest that since the detection of TCA glucuronide had not been reported previously, TCA excretion may have been under-reported in earlier studies (such as the human study cited above). The glucuronidation of TCA is supported by the observation that TCA is excreted in the bile of rats and mice (114). Urinary excretion represents the major route of elimination of the metabolites; fecal excretion, in the form of TCOH glucuronide, accounts for less than 5% of the total (118). The low fecal excretion is apparently associated to some extent with enterohepatic recirculation of TCOH (i.e., biliary excretion of the glucuronide, followed by hydrolysis and reabsorption of TCOH), which has also been suggested to occur in rats (119). DCA has been identified as a minor urinary metabolite of TCE (on the order of 1%) in both rats and mice (114,115,117,120), whereas MCA appears to be present at less than 0.1% (114). A recent in vitro study with mouse and rat liver tissues concluded that unlike most other chlorinated compounds, which are metabolized primarily by the microsomal enzymes of the MFO system, DCA degradation appears to occur primarily in the cytosol in a process that requires Environmental Health Perspectives * Vol 108, Supplement 2 * May 2000 k \\// *1 GSH (121). Although DCA has not been detected as a urinary metabolite ofTCE in the human, its kinetics have been studied (122)(123)(124). The peak concentration and AUC of DCA in the plasma after intravenous administration is linear up to approximately 20 mg/kg, but above 20-30 mg/kg some individuals display evidence of saturation of metabolism (121). The apparent volume of distribution and half-life for DCA are 0.3 L/kg (range: 0.09-0.60) and 1.05 hr (range: 0.25-1.87), respectively (124). Significantly, the clearance in humans appears to be much more rapid than would be expected from allometric scaling of animal data. In a comparative study (122), the half-lives in rats, dogs, and humans were 2.1-4.4 hr, 17.1-24.6 hr, and 0.33-0.6 hr, respectively. The extremely high rate of metabolism of DCA in humans is probably responsible for the failure of investigators to detect it as a urinary metabolite ofTCE.
Studies have shown that DCA is produced from TCA (67). In addition, DCA is produced in a roughly linear fashion from PERC (28,125) at levels consistent with production from TCA, the principal metabolite of PERC. An analysis of data on the dose-response and elimination kinetics of DCA formed from TCE led to the conclusion that another source of DCA was required in addition to TCA, but that the data were inconsistent with the second source being the initial oxidation step (126). Instead, the production of DCA from TCOH was hypothesized. A metabolic study of TCA and DCA in rats and mice (67) found that DCA was more rapidly metabolized than TCA. More than 50% of the administered TCA from an oral dose was excreted unchanged in the urine as compared to only 2% of administered DCA. Plasma concentration-time curves for TCA were similar in mice and rats, whereas those for DCA were greater in rats than in mice, both when DCA was administered and when it was derived from TCA.
There is evidence that repeated exposure to high concentrations of DCA inhibits its own metabolism. In studies with human volunteers (124,127), the excretion half-life for DCA increased 2-to 6-fold following repeated intravenous doses of DCA on the order of 10-50 mg/kg. Inhibition was only slowly reversible, taking from 1 week to greater than 3 months. In the rat, but not the mouse, exposure to 2 g/L DCA in drinking water for 14 days curtailed metabolic production of CO2 from DCA by more than an order of magnitude and increased the excretion half-life by roughly 4-fold (128).
It has recently been reported that under some conditions DCA can be artificially produced from TCA in blood samples as an artifact of the analysis. If acidification of the sample is performed rapidly after collection of the sample before the hemoglobin iron has been oxidized by exposure to air, DCA can be generated from TCA present in the sample (129). This artifact may have compromised some of the data reported on DCA plasma concentrations following administration of TCA or TCE (66,67). Metabolism in the lung. As with most chemicals, the preponderance of the metabolic dearance of TCE appears to take place in the liver. It has been demonstrated, however, in studies with an isolated, ventilated perfused lung preparation (130), that the male F344 rat lung also possesses a limited oxidative metabolic capability for TCE. Although the affinity (K,, for the lung metabolism observed in that study was similar to the affinity observed in the liver, the capacity (Vmax) of the lung metabolism was less than 1% of the capacity of the liver. These results suggest that lung metabolism is not an important contributor to total in vivo metabolism in the rat and that the rat lung does not possess a significant first-pass (presystemic) clearance capability for inhaled TCE. However, these results do not rule out the possibility that metabolism in the lung could produce sufficient local exposure to metabolites to produce toxicity and/or carcinogenicity.
Conjugative metabolism. A small proportion of TCE appears to be metabolized by enzymatic conjugation with GSH, principally by GST in the liver, followed by further metabolism in the kidney to the cysteine conjugate DCVC (131). The GST metabolic pathway is shown on the left side of Figure 1. Delivery of DCVC to the kidney may also be mediated by enterohepatic recirculation, in which GSH conjugate excreted in the bile is converted by gut bacteria to the cysteine conjugate, which is then reabsorbed (132). The GSH conjugate has been identified both in vitro, with rat liver microsomes, and (at 5 nmol/L) in the bile of rats given 2.2 g/kg TCE in corn oil (103). The cysteine conjugate also has been identified in the urine of animals dosed with TCE (72).
The bioactivation of DCVC to a reactive and mutagenic thioacylating intermediate is performed by cysteine conjugate I-lyase in the kidney (133). Although similar P-lyase activity has been measured in the kidney and liver, the two enzymes are distinct (134). Detoxification and dearance of DCVC takes place by urinary excretion of the N-acetyl derivative (103,135). In a study with PERC (136), it was determined that the excretion of the N-acetyl derivative was dose related (a higher fraction of N-acetyl derivative was excreted at doses where the oxidative pathway was saturated) and was significantly greater in the rat than in the mouse. However, measurements of acid-labile protein adducts associated with DCVC suggest that the activation of DCVC in the kidney may be as much as 12-fold greater in mice than in rats and that the kidney tissue exposure to DCVC-derived reactive species from oral dosing with TCE may be twice as great in the mouse as in the rat (137,138).
The activity of P-lyase has been measured in the liver and kidney of both humans and rats. One research group has reported a specific activity in human kidney on the order of 2.0-3.6 nmol/min/mg cytosol (139,140), compared to 6.45-7.6 nmol/min/mg cytosol in the rat (134). Another research group, however, has reported a maximum velocity (Vmax) of only 0.8 nmol/min/mg cytosol in the human, with an affinity (Ki) of 0.29 mM, compared to Vmax = 7.5 nmol/min/mg cytosol and Km = 1.6 mM in rat kidney cytosol in the same study (141). Data for PERC on the relative activity of liver cytosolic GST and kidney cytosolic cysteine conjugate ,B-lyase suggest that the human activity of both enzymes is roughly 10-fold lower than that in the rat (136). On the other hand, the specific activity of N-acetyltransferase in kidney cytosol appears to be very similar across species: 0.41 nmol/min/mg cytosol in the human, compared to 0.35-0.61 in the rat and 0.94 in the mouse (142).
The fact that N-acetyl-DCVC has been identified in the urine of humans exposed to TCE both occupationally (142) and in controlled exposures (143) indicates that exposure of the kidney to DCVC does occur in the human. In the occupational study (142), the concentrations of N-acetyl-DCVC in the workers' urine was about one-third of that measured in rats dosed orally with 50 mg/kg TCE. The ratio of N-acetyl-DCVC to TCA in the workers' urine ranged from 0.03 to 0.3, while in rats it ranged from 0.025 to 0.045, and in mice from 0.014 to 0.065. However, more recent data (143) obtained in controlled studies with both rats and human subjects suggest that relative urinary excretion of GSTpathway metabolites from TCE is actually somewhat lower in humans than in rats. Table 1: CA, concentration in arterial blood; CF, concentration in the fat; CG, concentration in the gut; Cl, concentration in inhaled air; CL, concentration in the liver; CR, concentration in the richly perfused tissues; CS, concentration in the slowly perfused tissues; CTB, concentration in the tracheobronchial tissue; CV, concentration in venous blood; CVF, concentration in the venous blood leaving the fat compartment; CVG, concentration in the venous blood leaving the gut compartment; CVL, concentration in the venous blood leaving the liver; CVR, concentration in the venous blood leaving the richly perfused tissues; CVS, concentration in the venous blood leaving the slowly perfused tissues; CVTB, concentration in the venous blood leaving the tracheobronchial tissue; CX, concentration in exhaled air; KF, rate of production of DCVC in the kidney; PDOSE, administered oral dose of TCE; QCL cardiac output; QF, blood flow to the fat; QG, blood flow to the gut; QL, blood flow to the liver; QP, alveolar ventilation; QR, blood flow to the richly perfused tissues; QS, blood flow to the slowly perfused tissues; QTB, blood flow to the liver; VR, volume of richly perfused tissue; VS, volume of slowly perfused tissue; VTB, volume of the tracheobronchial tissue.
(kidney, brain, alveolar region of lungs, etc.) and slowly perfused (muscle, skin, etc.) compartments. The model includes both inhalation and oral routes of exposure. Oral gavage is modeled using a two-compartment description of the gastrointestinal tract (144), rather than the single-compartment description used by Fisher and Allen (25), in order to better simulate the time course for the uptake of TCE from corn oil gavage. Allometric scaling is used throughout the model (volumes scaled by body weight, flows and metabolic capacities scaled by body weight to the threequarters power, rate constants scaled by body weight to the negative one-quarter power) to simplify intraspecies and interspecies extrapolation. Parent chemical dose metrics provided in the model include the concentration of TCE in blood and tissues, as well as the AUC for TCE in the blood.
Lung submodel. The model includes three target tissue submodels in which metabolism takes place: lung, kidney, and liver ( Figure 3). Michaelis-Menten kinetics are assumed for all metabolic processes. The tracheobronchial region of the lungs, which receives its own arterial blood supply, is described separately to support the modeling of in situ metabolism in this region by the Clara cells. This approach for describing metabolism in the cells lining the airways of the lung was felt to be more biologically accurate than the sequential gas exchange and lung tissue compartments used in the methylene chloride model (18). However, as long as metabolism in the lung is unimportant for presystemic elimination, as is the case for TCE and methylene chloride, the two descriptions should yield similar results. The dose metrics provided for the lung are the instantaneous concentration and AUC for CHL in the tracheobronchial region, which is assumed to be produced by saturable production and clearance of CHL in Clara cells. No systemic circulation of CHL is considered in the model.
Oxidative metabolism. Apart from the limited metabolism occurring in the lung, the model assumes that all oxidative metabolism takes place in the liver. The dose metric provided to describe metabolism is the total amount of TCE metabolized divided by the body weight. The model does not actually calculate the formation and metabolism of CHL in the liver, but instead assumes that TCA and TCOH are formed in a fixed yield from the oxidative metabolism of TCE. In the model, TCOH can subsequently be oxidized to TCA, conjugated with glucuronic acid, or reduced to DCA. DCA can also be produced from the reduction of TCA. Biliary excretion of TCOH glucuronide and enterohepatic recirculation of free TCOH is described, with only the glucuronide being excreted in the urine. Dose metrics for use with the liver target tissue include the concentrations and AUC for DCA and TCA in the plasma, as well as a potency-weighted sum of the AUCs for DCA and TCA, which will be described in the discussion on dose metric uncertainty. The concentration and AUC for TCOH in the blood are also provided as a noncancer dose metric.
Conjugative metabolism. The model also includes a linear metabolic pathway representing conjugation of TCE by GST. The model implicitly assumes that all GSH conjugation of TCE in the liver leads eventually to the appearance of DCVC in the kidney. Clearance of DCVC by N-acetyltransferase into the urine is also modeled. The dose metric provided in the model for the kidney (KTOX) is the total production of a thioacetylating intermediate from DCVC divided by the volume of the kidney.
PBPK Model Parameters
The parameters for the model are listed in Table 1, with the parameters for the parent chemical portion of the model listed first, followed by the parameters for each of the metabolites in turn.
Parameters for the parent chemical. The physiological parameters, with two exceptions, Environmental Health Perspectives * Vol 108, Supplement 2 * May 2000 * urine were based on the recommendations of the International Life Sciences Institute Risk Science Institute Working Group on Physiological Parameters (145). The exceptions were the cardiac output (QCC) in the mouse, based on the recommendations of Arms and Travis (146), and the alveolar ventilation (QPC) in the human, obtained from Astrand and Rodahl (147). In the model, the tissue volumes and blood flows for the gut, liver, and tracheobronchial region are subtracted from the values shown for all rapidly perfused tissues to obtain the parameters for the rapidly perfused tissue compartment shown in Figure 2; those for the fat are subtracted from the values shown for all slowly perfused tissues to obtain the parameters for the slowly perfused tissue compartment. The kidney volume shown in the table is used only in calculations for the kidney dose surrogate; as shown in Figure 2, the kidney is not described separately in the parent chemical model. The partition coefficients for TCE were obtained from the work of Fisher and Allen (25,94) and Fisher et al. (93); the partition coefficients for the gut and tracheobronchial tissues were assumed to be the same as those reported for the richly perfused tissues. The oral uptake parameters were estimated from data on the appearance of TCE and its metabolites in the blood following corn oil gavage in mice and rats (50). For some parameters, identified in Table 1, values chosen for calculating risk assessment dose metrics were different from those chosen to reproduce pharmacokinetic data. For example, human dose metrics were calculated using a value for QPC of 24, which corresponds to the U.S. EPA's standard assumption of a total ventilation rate of20 m3/day; the corresponding value for QCC of 16.5 was estimated from Astrand and Rodahl (147). Similarly, animals used in pharmacokinetic studies tend to have lower average body weights than animals used in cancer bioassays, so body weights appropriate to each case were used in the model. Parameters for oxidative metabolism. Initial values for the metabolic parameters for TCE were obtained from the work of Fisher and Allen (25,94) and Fisher et al. (93); however, the metabolic and clearance parameters for TCA, TCOH, and DCA were derived primarily on the basis of fitting the pharmacokinetic data depicted in Figures 4-17. When possible, parameters were also estimated from independent studies; for example, data from rodents and humans dosed with TCA were used to estimate the volumes of distribution and urinary excretion of TCA (93,109). Since the model contains a large number of metabolic and clearance parameters, many of which are highly correlated, the parameter values estimated by this process (i.e., the kinetic parameters for TCA, TCOH, and DCA) cannot be considered to be unequivocally identified. However, an additional biological constraint was applied during the modeling by demanding that all parameters be essentially constant across exposure scenarios within a given species and (to the extent justified by the experimental data) across species. This constraint greatly reduces the likelihood that alternative parameterizations could demonstrate equivalent success in reproducing the entire body of data. Another constraint on the parameterization not obvious from the figures is that of the total TCOH extractable from the blood, roughly 80% is present as free TCOH in the human (110), whereas roughly 70-85% is present as glucuronide in the rodent (113,126). In the figures in this article, the model concentrations shown represent either free TCOH (in rodents) or total (in humans), corresponding to the experimental data provided.
It is informative to note the departures from simple allometric expectations that were required on the basis of the experimental data across species. As with most other xenobiotics, the mouse shows a relatively greater and more variable capacity (VMC) for oxidative metabolism ofTCE than the rat and human. Also in keeping with evidence from other P450 substrates, the affinity for oxidative metabolism of TCE in the human is roughly an order of magnitude less (i.e., the value of KM is larger) than in the rodents. A striking difference between humans and rodents, which was clearly demanded by the experimental data, was that the oxidation of TCOH to TCA appears to be a relatively high-affinity (small value of KMO), low-capacity (small value of VMOC) process in the rodent but low affinity, high capacity in the human. It may be that this disparity reflects the involvement of different enzymes (e.g., MFO in the rodent vs ADH in the human). The result of this species difference is that although the model uses a similar.value across species for PO (based on the initial split of TCA and TCOH from CHL), the apparent ratio of TCA to TCOH predicted (and observed) over the entire time frame of an exposure to TCE is much higher in the human than in the rodent. The capacity (VMG) for glucuronidation of TCOH in the human, on the other hand, is much lower than in the rodent, as reflected in the greatly different ratios of free TCOH to glucuronide in the blood, mentioned above. The prolonged time courses of TCOH in the human provide clear evidence of biliary excretion (KEHBC) and enterohepatic recirculation (KEHRC). Evidence for enterohepatic recirculation was equivocal in the rodents, however, with recirculation being required to reproduce some data, but being contradicted by other data in the same species.
The least well-characterized portion of the oxidative metabolism pathway is the description of the kinetics of DCA. The only species in which DCA has been reproducibly detected in the blood following TCE exposure is the mouse, and these data were used to obtain values for the production (VMRC and KMR) and clearance (VMDC and KMDC) in the mouse. (The artifactual production of DCA from TCA in blood samples, noted earlier in this report, may have compromised some of the data on DCA plasma concentrations used to parameterize the production and clearance of DCA in the mouse.) Assuming that the affinities (KMR and KMD) are constant across species, the capacities in the other species were estimated (for VMDC) from the reported half-lives of DCA across species (122), or (for VMRC in the human) from data on peak DCA concentrations in human subjects exposed to TCE by inhalation (148). Since the clearance of DCA in the rat is actually slower than in the mouse (70), the capacity for production of DCA (VMRC) in the rat was set to the lower human value rather than that of the mouse to be consistent with the failure of investigators to observe DCA in the plasma of the rat following administration of TCE (69). The renal clearance of DCA (KUDC) was assumed to be the same as that observed for TCA (KUTC) in the same species. As mentioned earlier, the most striking departure from allometric expectations for the kinetics of DCA is the extremely high clearance in the human compared to the other species.
Parameters for lung metabolism. The parameters in the PBPK model for predicting the lung dose metric are the capacity and affinity for the production of CHL (VMTBC and KMTB) and the capacity and affinity for its clearance (VMCTBC and KMCTB). In the model, the production of CHL in the tracheobronchial region was assumed to be associated with the P450 activity in that tissue. This is the assumption that was made in the pharmacokinetic risk assessment for methylene chloride (18). The approach used in that risk assessment was also used to obtain the parameters in this case: the affinity in the lung was assumed to be the same as in the liver for the same species, and the relative capacity of the lung compared to the liver was determined on the basis of P450 activity measured with standard substrates (18). Based on these data, P450 activity falls off much more rapidly with body weight than would be expected from allometric considerations. No data were available on the clearance of CHL in the lung across species; therefore, it was assumed to be a low-affinity, high-capacity enzyme system such as ADH. The parameters in the PBPK model were chosen such that concentrations of CHL in the lung of the mouse predicted by Environmental Health Perspectives * Vol 108, Supplement 2 * May 2000 the model were consistent with those observed in recent studies (149). It was further assumed that the clearance of CHL in the lung scales across species according to allometric expectations (i.e., by body weight to the 3/4 power). This assumption leads to much lower CHL concentrations in the lungs of rats and humans compared to mice for the same TCE exposure conditions. An alternative assumption would have been that the activity of the enzyme responsible for the clearance of CHL scales in the same way as P450. This assumption would lead to similar concentrations of CHL in the lungs of mice, rats, and humans for the same TCE exposure conditions. Parameters for conjugative metabolism. The parameters in the PBPK model for predicting the kidney dose metric are the production of DCVC by the GST pathway (KFC), its activation by 3-lyase (KBLC), and its clearance by N-acetyltransferase (KNATC). First-order rate constants are used because the production of metabolites by the GST pathway is quite low, and saturation of enzyme capacity is unlikely. As discussed earlier, the capacity and affinity of i-lyase in the kidney have been measured in both rats and humans (141). These data were used to estimate the apparent first-order rate constants (KBLC) used in the model. No data were available on the activity of I-lyase in the mouse, so the relationships between P-lyase metabolic parameters in mice and rats reported for trichlorovinylcysteine derived from PERC (136) were assumed to apply-for DCVC as well. For N-acetyltransferase, only specific activity data across species are available (142). These data were converted to the corresponding rate constants (KNATC) by assuming the affinity of N-acetyltransferase for DCVC is the same as that measured for ,B-lyase in the same species. This assumption is supported by the similarity of the affinities of N-acetyltransferase and I-lyase for DCVC in the rat: 3.3 mM and 1.6 mM, respectively (141,150).
Finally, measurements of oxidative and conjugative metabolites in the urine following TCE exposure (143) were used to obtain estimates of the GST pathway rate constant (KFC). The oxidative pathway was represented by total excretion of TCA plus TCOH, while the conjugative pathway was represented by excretion of 1,2-DCVC. Data from the same study on excretion of 2,2-DCVC were not used. Unlike 1,2-DCVC, there was no evidence of a dose response for 2,2-DCVC as a function of TCE exposure in humans or rodents; similar amounts of 2,2-DCVC were excreted for TCE exposures ranging from 40 to 160 ppm. Ignoring 2,2-DCVC is unlikely to significantly affect the risk assessment for TCE, since 1,2-DCVC is clearly the more toxic and mutagenic of the two isomers (151).
The results of this analysis are shown in Table 2. In performing this analysis, all of the parameters in the model were set at the default values except VMC, KM, and KFC (in particular, KBLC and KNATC were set to the values calculated as described above). The values of VMC, KM, and KFC in the model were then varied to bring the model into agreement with the data for both the oxidative (MFO) and conjugative (GST) pathways. It can be seen that the model could be made to agree quite well with the urinary data when allometric scaling was assumed for conjugative metabolism (i.e., using the same value of the scaled parameter KFC in rat and human). Although it was necessary to adjust KM to obtain the best agreement with the MFO pathway data, allometric scaling of conjugative metabolism also gave the best agreement with the GST pathway data when the default values for KM were used. This result is consistent with the observed allometric scaling of the GST pathway for methylene chloride (152).
Additional data on urinary metabolite concentrations following oral gavage of rats with 50 mg/kg TCE (142), although not suitable for comparing with the model, were consistent with the inhalation data, suggesting that there is not an effect due to route of exposure. Therefore, the value estimated from inhalation was used to obtain the kidney dose metrics for the rat for both inhalation and oral exposures. Figure 4 compares the predictions of the model with experimental data on the concentrations of TCE in the blood and TCA in the plasma of male and female B6C3F1 mice exposed to 110 and 368 ppm, respectively, of TCE by inhalation for 4 hr (93). The model overpredicts the blood concentrations of TCE during the exposure by about 50%, but provides a good description of the time course for TCA in the plasma. Figure 5 shows the ability of the model to simultaneously reproduce experimental data on the blood concentrations of TCE, TCA, and TCOH in mice exposed to 1,000 mg/kg TCE by corn oil gavage (50). The model is also able to simulate ( Figure 6) the time course ofTCE, TCA, TCOH, and DCA in the B6C3F, mouse following an oral dose of 499 mg/kg in an aqueous vehicle (126). The DCA data at 6 and 9 hr in Figure 6 are suspect due to problems with the analytical method (129). of the time course for TCA in the plasma of the female rats, and for both TCE and TCA in the male rats. Figure 9 shows the ability of the model to simultaneously reproduce experimental data on the blood concentrations of TCE, TCA, and TCOH in rats exposed to 1,000 mg/kg TCE by corn oil gavage (49). As shown in Figures 10-13, the model is also able to simuiate the time courses of TCE, TCA, and TCOH in the F344 rat following oral doses of 100, 197, 591, and 3,000 mg/kg in an aqueous vehicle (69,113). Tins (hours) Figure 8. Comparison of predicted and experimental concentrations of TCE in blood and TCA in plasma in F344 rats exposed to TCE by inhalation. The figures show (A) TCE blood concentrations in male rats exposed for 4 hr to 529 ppm TCE vapors and TCA plasma concentrations in male rats exposed for 4 hr to 505 ppm TCE vapors; (B) TCE blood and TCA plasma concentrations in female rats exposed for 4 hr to 600 ppm TCE vapors. Kinetic data are taken from Fisher et al. (93).
Finally, Figures 14-17 demonstrate the ability of the model to describe the human kinetics of TCE and its metabolites, TCA and TCOH. Figure 14 compares the model predictions with experimental data collected on two different occasions for TCE in the blood, as well as for TCA and TCOH in the plasma and urine, following a 6-hr exposure of human subjects to 100 ppm TCE (109,110). The model provides a reasonable simulation of the time course of TCE in the blood during the exposure and for several hours afterward but underpredicts the longterm concentration of TCE in the blood. Model predictions for TCA in plasma and urine are close to the experimental data throughout the experiment. The model overpredicts the early concentrations ofTCOH in the plasma while underpredicting the later concentrations; however, the model predictions of TCOH glucuronide in the urine are very close to the data throughout the experimental period. Figure 15 demonstrates the ability of the model to reproduce data on multiple exposures (in this case, 4-hr exposures of human subjects to 70 ppm TCE repeated daily for 5 days) (154). The model underpredicts the peak concentration of TCE in the blood in this experiment by a factor of 2 but does reproduce the progressive failure of TCE concentrations to return to zero at the end of the day after repeated exposure. The model also provides a reasonable simulation of the accumulation and excretion of TCA in the plasma and urine, as well as TCOH glucuronide in the urine. Figures 16 and 17 show similar results for TCE in exhaled air as well as TCA and TCOH in the urine of human subjects exposed 7 hr per day for 5 days to 200 ppm TCE (155), and for TCA and TCOH in the plasma and urine of human subjects exposed 5 days for 6 hr to 50 ppm TCE (110).
It can be readily seen from Figures 4-17 that it was not possible to obtain complete agreement between the model and each of the studies investigated using a single set of parameters in each species. This failure undoubtedly results from a combination of variation across individuals and animal strains, experimental error, and model error. Nevertheless, given the general agreement of the model with a variety of data on TCE, TCA, and TCOH concentration time courses in both rodents and humans, there can be relatively high confidence in dose metrics based on the Time (hours) Figure 9. Mean observed and predicted blood concentrations of TCE (A), TCA (B), and free TCOH (C) following corn oil gavage with 1,000 mg/kg TCE in rats. Kinetic data are taken from Prout et al. (50). predictions of the PBPK model for these chemicals. Similarly, model predictions for the total amount of TCE metabolized per kilogram body weight were generally within a factor of 2 of data on inhalation and oral exposures of mice, rats, and humans (50,62,63,114,117). Unfortunately, as mentioned earlier, there is a lack of similar data to provide confidence in the model predictions for DCVC in the kidney, CHL in the lung, and DCA in the human.
Sensitivity and Uncertainty Analysis
In order to evaluate the level of confidence that could be given to the calculations performed with the PBPK model, a series of quantitative and qualitative analyses were performed to characterize the uncertainty in the model structure, parameterization, and dose metric selection. Since the intended use of the PBPK model is to calculate target tissue dose metrics, any evaluation of the model should focus on this aspect of the model capabilities. Therefore, both the sensitivity analysis and the uncertainty analysis were conducted with respect to model predictions of the target 2 Table 1. Sensitivity coefficients of less than 0.1 in absolute value were omitted from the table for clarity, and coefficients above 0.5 are oudined. It can be seen that of the 29 parameters in the TCE/TCA portion of the model, 12 have essentially no impact on risk predictions based on the two dose metrics, and only 5 have a significant impact: the alveolar ventilation (QPC), the capacity for metabolism of TCE (VM), the fraction of TCA produced from the metabolism of TCE (PO), the volume of distribution (VDTCAC) and rate of excretion (KUTC) of TCA. None of the parameters are associated with sensitivities greater than 1.0, indicating that there is no amplification of error from the inputs of the model to the outputs. This is, of course, a desirable trait in a model to be used for risk assessment.
Sensitivities for the other metabolites in the model are not shown, but the results were similar to those for TCA; that is, none of the parameters were associated with sensitivities greater than 1 in absolute value, and (except for QPC) only the parameters directly related to the production and dearance of a metabolite were associated with significant sensitivities (close to 1) for dose metrics based on that metabolite.
PBPK Model Parameter Uncertainty
There are a number of ways of characterizing the uncertainty associated with use of a PBPK model. The best approach depends on the level of uncertainty. In the case of the TCE model, the level of uncertainty varies considerably from one portion of the model to another. Some parameters in the model, such as those for TCE, TCA, and TCOH, are relatively well established by data, and the uncertainties can be addressed fairly quantitatively. Under these conditions the preferred method for characterizing the overall model uncertainty is to perform Monte Carlo analysis, as discussed below. On the other hand, other parameters in the model, such as those associated with the production and clearance of DCVC in the kidney and CHL in the lung, are based on inadequate Table 3 shows the normalized analytical sensitivities for the TCE and TCA parameters in the PBPK model described above. The normalized analytical sensitivity coefficient represents the fractional change in output associated with a fractional change in the input parameter. For example, if a 1% change in the input parameter results in a 2% change in the output, the sensitivity coefficient would be 2.0. In Table 3, the outputs evaluated are the dose metrics for the total amount of TCE metabolized per kilogram body weight and the AUC for TCA. The parameters in the and often conflicting data, and the uncertainties cannot be adequately quantified to support such a rigorous analysis. For these aspects of the model, an appropriate method for characterizing uncertainty is to simply calculate the range of dose metrics that could reasonably be expected given the existing data. Liver dose metric: Monte Carlo analysis. The sensitivity analysis described above does not consider the potential interactions between parameters; the parameters are tested individually. Also, sensitivity analysis does not reflect the uncertainty associated with each parameter. For example, the fact that the output is highly sensitive to a particular parameter is not important if the parameter is known exactly. To estimate the combined impact of uncertainty regarding the values of the various parameters, a Monte Carlo analysis was performed on an early version of the PBPK model for a characteristic dose metric, the average daily AUC for TCA in the plasma. The version of the model tested in this analysis was essentially identical to that of Fisher and Allen (25) and did not include the description of TCOH and DCA kinetics provided in the current version of the model. In support of the Monte Carlo analysis, the distributions of possible values for each of the input parameters were estimated, as shown in Table 4. The Monte Carlo software (PBPK_SIM, K.S. Crump Group, Ruston, LA) randomly selects a set of parameter values from the distributions for the bioassay animal and runs the PBPK model to obtain dose metric values for each of the bioassay dose groups. It then selects a set of parameter values from the distributions for the human and runs the PBPK model to obtain a dose metric value for a specified human exposure scenario. This process is repeated the specified number of times (400 in this case) until the distributions of dose metrics have been obtained. Table 4 lists the means and coefficients of variation (CV) used in a Monte Carlo uncertainty analysis of the AUC-TCA dose metric. Truncated (greater than zero) normal distributions were used for all parameters except the kinetic parameters, which were assumed to be log-normally distributed. The CVs for the physiological parameters were estimated from data on the variability of published values (146,156). In order to maintain mass balance in the PBPK model, after sampled parameter values for cardiac output and fractional tissue blood flows were used to calculate blood flows to each of the tissues, total cardiac output was recalculated as the sum of the individual tissue blood flows. The CVs for the partition coefficients were based on repeated determinations for two other chemicals, PERC (28) and chloropentafluorobenzene (6). The CVs for the metabolic and kinetic constants were estimated from a comparison of reported values in the literature and by exercising the model against various data sets to determine the identifiability of the parameters estimated from pharmacokinetic data. It should be understood, however, that the The results of the Monte Carlo analysis are shown in Table 5 based on the average daily AUC the plasma, the 5th and 95th po the dose metric distributions a within a factor of 2 to 3 of the n of the 90% confidence intervals metrics range over somewhat ] order of magnitude. These result to those reported for a PBPK mo ylene chloride (21) and are probal tative of the uncertainty associate dose metrics in the validated poi model such as AUC-TCOH, AU the mouse), and total metabolism Lung dose metric. The great4 uncertainty regarding the calcul lung dose metric is the lack of metabolic clearance of CHL i Table 6 shows the predicted lung for the principal inhalation bioas ing a dose response for lung tum4 for the highest oral exposure of th the highest rat exposures, and human exposure scenarios. TI exposures have been ordered accc 165,66). 0Doses employed in oral bioassay (59). (25) bolically active Clara cells in the mouse than in the human (71). Although no data could be found on the cross-species activity of ADH in for TCA in the lung, ADH activity in the rat airway ercentiles of appears to be restricted to the Clara cell (157), Ire generally suggesting that the enzymes responsible for the nean, and all clearance of CHL may also fall off more for the dose rapidly than allometric expectations.
less than an
The alternative dose metric calculations, ts are similar shown in parentheses in Table 6, were del of meth-obtained under the assumption that the dearbly represenance of CHL scales in the same way as lung d with other P450. Using this assumption yields dose metrtions of the rics roughly 10-fold higher in the rat and JC-DCA (in 700-fold higher in the human. This signifi-ofTCE. cant uncertainty regarding the relative expoest source of sure to CHL across species could be resolved lation of the by the collection of data on CHL dearance in data on the the lung across species, similar to the data in the lung. that have been reported on P450. Of course, dose metrics as discussed in the section on the lung parassays provid-meters, there are also other sources of paraors (64)(65)(66), meter uncertainty associated with the lung be mouse, for dose metric. Additional experimental data are for several needed before a quantitative estimate of the he different overall uncertainty associated with this metric )rding to the could be confidently attempted.
Kidney dose metric. The overriding source of uncertainty regarding calculation of the kidney dose metric is the inadequate and often conflicting data in the literature for the conjugative pathway. Specific data gaps include the affinity of kidney N-acetyltransferase for DCVC in the human and the activities of the GST pathway in the rat and human. Table 7 shows the predicted kidney dose metrics for the principal bioassays providing a dose response for kidney tumors (65)(66)(67), for the highest oral exposure of the mouse, and for several human exposure scenarios. The different exposures have been ordered according to the predicted value of the LADD based on the production of the toxic thiol per gram of kidney tissue (KTOX). The human dose metrics shown in Table 7 are those calculated assuming allometric scaling of the GSH pathway rate constant. There is, of course, additional parameter uncertainty associated with the limited data on the other enzymes involved in the production, intoxication, and clearance of DCVC. As with the lung target tissue, it will not be possible to provide an accurate assessment of the overall uncertainty in the kidney dose metric until reproducible data are collected.
Dose Metric Sdection Uncertainty
The pharmacokinetic dose metrics most commonly applied to characterize the exposure of a tissue to a chemical are the peak concentration and the AUC, and these are the principal types of dose metrics provided in the PBPK model. However, there are other possible forms for dose metrics that might be useful for describing nonlinear processes. For example, time above a critical concentration (TACC) has been suggested as an appropriate dose metric for the effects of methotrexate, whose toxicity demonstrates a strong dependence on dose rate (158). The following discussion provides a rationale for Environmental Health Perspectives * Vol 108, Supplement 2 * May 2000 Significantly increased tumors in at least one study. *Increased tumors in at least one study (not statistically significant). 'Lifetime average daily area under the chloral concentration curve in the tracheebronchial region (mg/L-hr). bMaximum concentration achieved in the tracheobronchial region (mg/L). clnhalation, 7 hr/day, 5 the pharmacokinetic dose metrics provided in the PBPK model and considers other possible dose metrics that could be selected.
Liver dose metric. If, as was once thought, reactive species produced during the metabolism of TCE were responsible for its liver carcinogenicity, an appropriate dose metric would be total daily metabolism divided by the volume of the liver (13.18). However, current information suggests that two stable metabolites, TCA and DCA, are primarily responsible for the liver tumor incidence observed in mice dosed with TCE (69,70). The commonly accepted form of the dose metric for the chronic interaction of a stable metabolite with a tissue is the AUC in the tissue. This mathematical form implicitly assumes that the cumulative effect of the metabolite on the tissue is linear over both concentration and time. In this case, the most appropriate dose metric would reflect liver tissue exposure (AUC) to both TCA and DCA (69,70). If it is assumed that both DCA and TCA contribute to the carcinogenicity of TCE in the liver, the proportion of the observed tumor risk to assign to each metabolite could be based on their relative potencies when dosed directly. However, as mentioned earlier, data on the AUCs for DCA resulting from exposures to TCA [e.g., (70)] may have been compromised by a sampling artifact that could lead to overestimates of DCA concentrations in the presence of TCA (129), making it impossible to estimate the individual potencies of TCA and DCA. As a simplifying assumption, all of the tumorigenicity of TCE can simply be ascribed to TCA, as was assumed by Fisher and Allen (25). Since DCA has been detected in the mouse to a much greater extent than in the human following TCE exposure, the use of the AUC for TCA alone as the dose metric is almost certainly safe-sided (in the direction of overestimating the human risk estimate) compared to including potency-weighted AUCs for both TCA and DCA.
Strictly speaking, the AUC for TCA should actually be calculated for the concentration in the liver. However, the use of the AUC in the plasma provides a surrogate for the liver AUC that can be validated more readily against experimental data. Since risk estimates are based on the ratio of animal and human dose metrics, this effectively amounts to an assumption that the ratios of the plasma concentrations of the acids to their concentrations in the liver are constant across species. In fact, data on the binding of TCA in the plasma of rats and humans (113) suggest that TCA in plasma is bound to a much greater extent in the human (-80%) than in the rodent (-50%). Based on these data, it can be estimated that the liver-to-plasma TCA concentration ratio in the human is about 40% of the ratio in the rodent. This estimate is also consistent with the ratio of reported relative volumes of distribution of TCA in the two species, which are on the order of 10 and 25% of body weight in the human and rodent, respectively. Thus, using the AUC of TCA in the plasma as the dose metric provides a conservative estimate of the crossspecies relationship for the AUCs in the liver, tending to overestimate liver exposure to TCA in the human by about 2-fold. Table 8 shows the predicted liver dose metrics for the principal animal bioassays providing a dose response for liver tumors (59,(65)(66)(67)159), for the highest rat exposures in these same studies, and for several human exposure scenarios. The different exposures have been ordered according to the predicted value of the LADD based on the AUC for TCA. Bioassay exposures associated with LADDs for AUC-TCA of greater than 1,150 were uniformly positive, whereas bioassay exposures with LADDs less than 700 were uniformly negative. The highest exposures of rats produced AUC-LADDs considerably less than those producing tumors in the mouse, consistent with the negative results in the rat bioassays.
The most striking feature of the results for this target tissue compared to the lung and kidney is that two of the three highest dose metrics were obtained for the human occupational exposure scenarios. The relatively high dose metrics for AUC-TCA in the human reflect the much slower clearance of TCA compared to the rodent. It is interesting to note that the rank ordering in Table 8 would be different if it were based on AUC-DCA. In that case, all of the human dose metrics would be uniformly below the positive animal bioassay dose metrics, reflecting the rapid clearance of DCA in the human.
Although AUC is a standard metric for tissue exposure, other forms of the dose metrics for DCA and TCA might be more appropriate for their modes of action. If it is possible that the tumorigenic effects of these chemicals are related to some aspect of their interaction with a receptor, peak concentrations (CMAX), or TACC, might actually be more appropriate than AUCs. Another nonlinear dose metric recently discussed for receptor-mediated effects is based on average receptor occupancy (160). Unfortunately, the more an attempt is made to include pharmacodynamic events in a dose metric, the more difficult it becomes to collect the data necessary for its use. In the case of TCE, there are currently no experimental data available to evaluate the use of such alternative pharmacodynamic dose metric approaches. Of the possible dose metrics, only AUC, CMAX, and TACC can be estimated from the data currently available. All three of these metrics are available for TCA in the PBPK model. Lung dose metric. As described earlier, tumors have been observed in the lungs of mice exposed to TCE by inhalation. The mechanism in this case appears to be entirely different from that in the liver. In a welldesigned experimental effort (71), investigators at ICI combined in vivo and in vitro experiments to elucidate the mechanism of TCE carcinogenicity in the mouse lung.
In the in vivo studies, female mice and rats were exposed to TCE at a range of inhaled concentrations at and below the concentrations at which lung tumors have been observed in mice, and the effects of TCE in the lung were determined. A specific lesion, characterized by vacuolization of lung Clara cells, was observed in mice, but not rats. There was evidence of a threshold for the Clara cell effects at about 20 ppm. The majority of Clara cells were unaffected at 20 ppm and all enzyme markers were normal, whereas at 200 ppm most of the Clara cells showed marked vacuolization accompanied by marked loss of CYP450 activity. Mice exposed to 100 ppm CHL by inhalation displayed Clara cell lesions similar to those observed with 1,000 ppm TCE. In contrast to these results, only mild effects were observed with TCOH inhaled at 100 ppm, and none were observed with 500 mg/kg TCA given intraperitoneally. (The effects had been observed with intraperitoneally administered TCE at 2,000 mg/kg.) These results suggested that CHL was responsible for the toxicity.
In the in vitro studies conducted by Odum et al. (71), mouse lung Clara cells metabolized TCE to CHL, TCOH, and TCA, with CHL being the major metabolite. Significantly, no TCOH glucuronide was detected. In comparison with mouse Clara cells, mouse hepatocytes produced primarily TCOH and its glucuronide. In both cell preparations, a steady-state concentration of CHL was achieved. Separate in vitro studies demonstrated that mouse Clara cells possess a relatively low activity for the glucuronidation of TCOH compared either to the glucuronidation of other substrates in the lung or to the glucuronidation of TCOH in the liver. It has also been determined that ADH, the enzyme that converts CHL to TCOH, has low activity in the mouse lung (161), consistent with the relatively low production observed in the Clara cells. On the basis of this evidence, the investigators concluded that the observed acute toxicity in the lung was a result of accumulation of CHL in Clara cells, resulting from a limitation in the formation of TCOH and its glucuronide. The specificity of this lesion for the Clara cells can be rationalized in terms of their relatively high CYP450 activity, coupled with a limited ADH and glucuronosyltransferase activities.
The implications of these results for the lung tumorigenicity of TCE are 2-fold. First, the accumulation of CHL, if it does occur in vivo, has clear carcinogenic implications, since CHL was genotoxic in a number of studies (71). Second, the recurrent toxicity observed with intermittent exposure could produce increased cell proliferation, exacerbating the genotoxic effect. In terms of the requirements for a lung dose metric in the PBPK model of TCE, it would appear that the model should include, at minimum, a description of the in situ metabolism of TCE in the Clara cell, to the extent of providing dose measures based on achieved CHL concentrations. Although a number of significant qualitative and quantitative uncertainties remain concerning the carcinogenicity observed in the lung, the use of the PBPK model could provide insights on the quantitative consistency of various mechanistic hypotheses with experimental data. There do not appear to be sufficient data at this point to support a quantitative description of the species-dependent pharmacokinetic dose response for the lung carcinogenicity.
The lung dose metric calculations shown in Table 6 can be used to evaluate the consistency of the CHL dose metrics with the bioassay results. As mentioned previously, the entries in the table are arranged in decreasing order of LADD for AUC-CHL. Bioassay exposures associated with LADDs for AUC-CHL greater than 1.5 and CMAX for CHL greater than 0.3 tended to be positive, whereas bioassay exposures with AUC-CHL LADDs less than 1.5 and CAMX for CHL less than 0.3 were uniformly negative. The daily peak concentration dose metric (CMAX for CHL) appears to be more consistent with the negative response of the rats. Neither metric explains the fact that the oral mouse bioassays were negative for lung tumors, suggesting the possibility of a portal-of-entry effect.
Kidney dose metric. A variety of mechanisms have been identified for the kidney effects of halogenated hydrocarbons (162). The fact that tumors are observed only in the rat might suggest that they are associated with the male rat nephropathy described for many hydrocarbons, in which the accumulation of a male-rat-specific a20-globulin in proximal tubular cells leads to hyaline droplet accumulation, necrosis, increased cell proliferation, and cancer (163). However, a study designed specifically to evaluate this possibility showed evidence of the hyaline droplet accumulation and increased cell replication with PERC but not with TCE (163). It was also felt that the oxidative metabolism of TCE was unlikely to explain the kidney carcinogenicity in rats, since the rate of metabolism in the liver greatly exceeds that in the kidney, and no liver tumors are seen in the rat (72). An alternative mechanism was proposed, in which direct conjugation of TCE with GSH in the liver was followed by further metabolism in the kidney to a cysteine conjugate that could then be cleaved to a reactive intermediate in the kidney tubular cells (131). The cysteine conjugate formed from TCE, DCVC, has been shown to be highly nephrotoxic (78) and mutagenic in the Ames test (115).
As with the lung carcinogenicity of TCE, more than one mechanism may play a role in the kidney tumors. The tumors produced in the kidney by TCE are very rare tumors in control animals and do not appear to be associated with the exacerbation of spontaneous processes, suggesting that a genotoxic mechanism may be responsible. On the other hand, in the only bioassay that reported a significant increase in kidney tumors from TCE (67), cytotoxicity was observed in the kidney at both the low and high doses, whereas tumors were observed only at the high dose. Kidney cytotoxicity was also reported in association with a nonstatistically significant incidence of kidney tumors in another study (65,66). However, dosing of mice with DCVC in drinking water for 46 weeks produced clear evidence of toxicity at 87 weeks but no evidence of tumors (164). The significance of this result is enhanced by the observation that activation of DCVC in the kidney appears to be much greater in the mouse than in the rat, and the mouse also appears to be more Environmental Health Perspectives * Vol 108, Supplement 2 * May 2000 responsive to the induction of cell proliferation by DCVC than the rat (138). Moreover, measurements of acid-labile protein adducts in the kidney associated with DCVC suggest that the production of DCVC-derived reactive species in the kidney resulting from an oral dose of 1,000 mg/kg TCE may actually be greater in mice than in rats (137,138), and mice but not rats showed increased cell proliferation in the kidney in response to treatment with TCE at 1,000 mg/kg. Other studies in the rat also fail to support the suggestion that significant hyperplasia is produced in the kidney from exposure of rats to TCE (163). Thus, whether a genotoxic or cytotoxic mechanism involving DCVC is proposed, it is difficult to explain either the negative results of the DCVC bioassay in the mouse or the greater sensitivity of the rat compared to the mouse with regard to kidney tumors from TCE. Nevertheless, a mode of action for TCE in the kidney involving mutagenicity and cytotoxicity from DCVC is the most supportable choice at present, especially since no suggestion for an alternative source of the observed tumorigenicity has been provided in any of the studies just described. The kidney dose metric calculations shown in Table 7 can be used to evaluate the consistency of the KTOX dose metric with the bioassay results. Bioassay exposures associated with KTOX LADDs greater than 15 tended to be positive, whereas bioassay exposures with KTOX LADDs less than 15 were negative. The mouse dose metric is well below that of the rat, consistent with bioassay results. The lower dose metric values in the mouse result from the higher ratio of clearance (KNATC) to intoxification (KBLC) in the mouse as compared to the rat.
Noncancer dose metrics. The issues associated with the selection of dose metrics for the noncancer toxicity of TCE are discussed in a companion article in this same issue (44). Therefore, only a short summary of the rationale for the selection of noncancer dose metrics is included here. The relationship of various noncancer dose metrics across species is shown in Table 9. In this table, the values of the dose metrics are shown for equal administered dose. Clearly, the human equivalent dose or concentration for a given animal study depends on both the route of exposure and the dose metric chosen, which in turn depend on the mode of action assumed for the specific end point being considered.
Neurological effects. The neurological effects of short-term exposure to solvents such as TCE are rapidly reversible, suggesting that they result from a physicochemical effect of the parent chemical on the proper function of lipophilic cellular membranes. Appropriate dose metrics for these effects would be the peak concentration or AUC for the parent chemical in the brain. Since tissue-blood partition coefficients are relatively uniform across species, the peak concentration in the blood can be used as a surrogate. Alternatively, in the case of TCE, an appropriate dose metric might be the peak concentration for TCOH, which has been suggested to be responsible for the observed neurological effects of chloral hydrate (79). Hepatotoxicity. Pharmacokinetic studies (73) have demonstrated that the relationship between the acute hepatotoxicity of TCE and the total production of urinary metabolites is linear, and it has been suggested that this result is consistent with the hypothesis that the toxicity is produced by reactive intermediates (13). Based on this assumption, the most reasonable dose metric for the hepatic toxicity of TCE would be the total amount of metabolism divided by the volume of the liver (18). On the other hand, a comparison of the toxic potency for TCE and PERC (73) suggests that TCA, rather than total metabolism, is responsible for the liver effects of these chemicals (44).
Nephrotoxicity. As already discussed, the toxicity observed in the kidney appears to be due to metabolism of DCVC. Therefore, the cancer dose metric (KTOX) provides a useful metric for the kidney toxicity as well.
Immunological and developmental effects. Significant uncertainty exists regarding the appropriate dose metric for immunological and developmental effects. Possible metrics include the peak concentrations and AUCs for TCE and TCA. In the case of fetal effects, the dose metric would most properly be calculated using a PBPK model with a description of the fetus. However, dose metrics based on maternal blood should provide a reasonable surrogate for the effects of TCE, since TCE and its metabolites appear to move readily across the placenta (91). The discussion in this article has been restricted primarily to pharmacokinetic issues. However, the line between pharmacokinetics and pharmacodynamics is ill defined, and a pharmacokinetic model can often be extended somewhat into the pharmacodynamic realm. The following discussion touches on some of the areas where there is potential to develop improved dose metrics for the PBPK model and that include some level of pharmacodynamics.
Liver. Evidence regarding the mode of action of TCA and DCA could potentially be used to develop a dose metric more closely associated with tumorigenicity. In principle, if information on the differential response across species to mitogenic effects from TCA and DCA were obtained, it could be incorporated into a pharmacodynamic tissue dose metric. Possible pharmacodynamic metrics in the case of the liver carcinogenicity of TCE might be the expression of TGF-3 or a measurable suppression of cell proliferation in hepatocytes as a marker of an early response to a presumed mitogenic signal. The use of a similar approach, based on the observed dose dependence of hormonal response, has been proposed for an analogous carcinogenic mechanism associated with follicular cell carcinoma in the thyroid (165).
Lung and kidney. If the kidney and lung carcinogenicity of TCE is considered to result primarily from enhanced cell proliferation secondary to recurrent toxicity, possible dose metrics would include measures of cytotoxicity, cell death, or cell division, as has been proposed for the liver carcinogenicity of chloroform (35,166). In the case of chloroform, fairly complicated metrics involving the instantaneous rates of metabolism and distributions of cellular sensitivity have been suggested (5,35). To apply these approaches to (167)(168)(169)(170).
Of particular note, the dose response for cytotoxicity in these subchronic studies with chloroform is markedly different from that observed in acute and in vitro studies (35). It appears that caution must be used when the dose response for a surrogate measure of tissue response is derived from in vitro or short-term in vivo experiments. If exposure to a genotoxic metabolite is also considered to be quantitatively important in the lung and kidney target tissues, an even more complicated dose metric would be required. The use of a dose metric based on the product of DNA-protein crosslinks and cell-labeling index has been suggested for evaluating the incidence of nasal tumors from formaldehyde exposure, assuming a carcinogenic mode of action involving both genotoxicity and cytotoxicity (171). Again, extensive studies paralleling CIIT's efforts with formaldehyde would be required to apply this approach to TCE.
Target rIsue Correspondence
Another aspect of uncertainty relevant to the incorporation ofpharmacokinetic modeling in risk assessment is the question of how to deal with a lack of correspondence of target tissues across species. For example, none of the tumors observed in TCE bioassays were reproduced in both mice and rats. This behavior can be contrasted with that of a trans-species carcinogen such as vinyl chloride, which produces tumors in the same target tissue (the liver) in all species tested, as well as in humans, with similar potency (26). The lack of site correspondence for TCE in different animal species clearly has important implications for the expectation of site correspondence between animals and humans (which is implicit in the pharmacokinetic approach for risk assessment). Nevertheless, the assumption necessarily underlying the application of pharmacokinetic modeling in a risk assessment for TCE would be that the human target tissues of concern would be the same tissues identified in the animal studies. As pointed out earlier, there is at least some suggestion from epidemiological studies that another target tissue for TCE in the human could be the lymphatic system. If sufficient evidence of a link between TCE exposure and lymphoma were obtained from epidemiological studies, but without sufficient dose-response information to support a potency estimate, it is unclear how pharmacokinetic modeling could be used to provide an animal-based estimate unless a statistically significant dose response for lymphoma could be demonstrated in the rodent. Even then, data would also be required on the metabolism and mode of action of TCE in this target tissue.
Human Variability
Standard cancer risk assessment practice estimates the risk for an average individual, whereas noncancer risk assessment typically applies an uncertainty factor of 10 to account for human variability and the possibility of sensitive subpopulations. Human variability plays an important role in determining the actual risk to an individual compared to the average risk to a population. Part of this variability is pharmacokinetic and is subject to quantification. For example, the variation in the human dose metrics for TCA presented in Table 5 primarily reflects the impact of variability, as opposed to uncertainty, in human pharmacokinetics on target tissue dose (in this case, for tissue exposure to TCA resulting from environmental exposure to TCE). Pharmacokinetic factors affecting the response of an individual to the toxicity and carcinogenicity of a chemical such as TCE include size, weight, condition, fat content, and level of physical activity. These factors modify the uptake, distribution, and elimination of TCE associated with a given exposure (172). For example, an individual with a large proportion of fat will absorb more of a chemical such as TCE and retain it longer than a lean individual. This longer storage increases the opportunity for metabolism to the active species. Studies on normal human volunteers have shown significant variation in individual pharmacokinetic behavior, and it is clear that this variability in pharmacokinetic factors is an important component of the overall interindividual variability of susceptibility to the toxic effects of chemicals (173).
By far the most important variability impacting target tissue dose is in metabolism. Four different isozymes of CYP450 have been found to play a role in the oxidative metabolism of TCE in rodents: 1A1/2, 2B1/2, 2C1116, and 2E1. Of these, only 2C11/6 is not found in humans. CYP 2E1 appears to have the highest affinity for TCE, although the other isozymes can become important at higher concentrations (100). Sex, pregnancy, and age-related differences in metabolism can result from normal variations in CYP 2E1 content (101); increased metabolism can result from the inducibility of CYP 1A1/2 (e.g., by aromatics), 2B1 (e.g., by phenobarbital), or 2E1 (e.g., by ethanol) (102,115). Studies of human populations have shown that the activity of the CYP enzymes can vary by more than a factor of 10 between individuals (152,174,175), and that there is a genetic difference (polymorphism) between individuals with high activity and low activity that is associated with a different susceptibility to cancer (176). Genetic polymorphisms of the CYP enzymes across racial and ethnic groups have been observed (177), as have quantitative differences in metabolic capacity (178).
Sex differences in the excretion of TCE metabolites have also been noted in the human (179), with females excreting a larger proportion of TCA and a smaller proportion of TCOH than males. The difference between females and males in the ratio of TCA to TCOH excreted is greatest initially (as much as a factor of 5.5 during the first 24 hr after exposure), suggesting that the difference derives from a relatively greater rate of the production of TCA from CHL rather than from TCOH. The production of TCA in humans appears to be highly variable and generally somewhat higher than in other animals. For example, in one study the production of TCA from chloral hydrate in different individuals varied from 5 to 47% (107).
There are still other factors such as disease and hormonal status that could also affect the individual risk from exposure to a TCE, either because of an impact on pharmacokinetics or metabolism, or due to other interactions. Estrogens, for example, have been associated with both increased risk (for breast cancer) and decreased risk (for colon cancer) and are also metabolized by the CYP system (175). Therefore, the possibility of interaction with TCE exposure includes metabolic inhibition or induction as well as tumor promotion or repression.
Pharmacokinetic and metabolic differences alone cannot explain the overall interindividual variation in susceptibility observed in exposed populations (173,180). Clearly there are other, less well understood interindividual differences, both acquired (due to environmental exposures or disease states) and inherited (due to genetic differences) that are also important determinants of the individual risk for development of toxicity from exposure to a chemical. However, to the extent that we can quantitatively describe and evaluate pharmacokinetic and metabolic variation, it will become increasingly possible to estimate the range of risks in an exposed population and to identify the factors that put individuals at the greatest risk.
Conclusions
The PBPK model described in this article provides reasonably accurate and precise estimates of dose metrics based on TCE and its major metabolites, TCA, TCOH, and DGA, in both experimental animals and humans. Tissue dose metrics calculated with the model should therefore be useful in risk assessments for end points where the mode of action involves tissue exposure to these chemicals. Other target tissue dose metrics that can be calculated with the model, including CHL in the lung and DCVC in the kidney, are highly uncertain due to a lack of adequate pharmacokinetic data across species. Additional studies could greatly reduce the uncertainty associated with these dose metrics and make their use in risk assessments more viable. However, it must be understood that pharmacokinetics is only one dimension of the process of estimating human risk from animal studies; the other potentially more important dimension is pharmacodynamics. Species differences in pharmacodynamics may lead to wide differences in susceptibility to tumors or to other toxic outcomes at the same target tissue doses.
|
2014-10-01T00:00:00.000Z
|
2000-05-01T00:00:00.000
|
{
"year": 2000,
"sha1": "69f90c843e55561d4c23612163a3ef42c2e9c0d5",
"oa_license": "pd",
"oa_url": "https://doi.org/10.1289/ehp.00108s2283",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "181b9237b4ae5bc95dc5da18be525ab560002e35",
"s2fieldsofstudy": [
"Chemistry",
"Environmental Science",
"Medicine"
],
"extfieldsofstudy": [
"Medicine",
"Chemistry"
]
}
|
121571714
|
pes2o/s2orc
|
v3-fos-license
|
Operation of the Radio Occultation Mission in KOMPSAT-5
Received Jul 16, 2010 Revised Sep 06, 2010 Accepted Sep 11, 2010 Corresponding Author E-mail: cmsoo@kasi.re.kr Tel: +82-42-865-3244 Fax: +82-42-861-5610 This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. Operation of the Radio Occultation Mission in KOMPSAT-5 Mansoo Choi, Woo-Kyoung Lee, Sungki Cho, and Jong-Uk Park Space Geodesy Research Group, Korea Astronomy and Space Science Institute, Daejeon 305-348, Korea University of Science and Technology, Daejeon 305-333, Korea
INTRODUCTION
Korea multi-purpose satellite-5 (KOMPSAT-5) should have a very high level of orbit determination precision by the precise observation requirements of the synthetic aperture radar (SAR) which is the main payload.For this, Korea Astronomy and Space Science Institute (KASI) is to install the atmosphere occultation precision orbit determination (AOPOD), as a sub-payload, consisting of integrated global positioning system occultation receiver (IGOR), space-borne dual frequency GPS receiver, and a laser retro reflector array (LRRA) which is for the satellite laser ranging (SLR) used in the orbit determination verification, and carrying out the research on the installation of the AOPOD system, performance verification and the operation strategy.
We performed the establishment of the ground test bed using IGOR engineering model (EM) and carried out electrical test bed (ETB) test.Based on these, we conducted various test including flight model (FM) performance test, system integration and EMI/EMC (Montenbruck et al. 2005).We are also performing researches, based on the results, on the operation strategies for the individual scenarios for the IGOR operation.In this article, we introduce the AOPOD system that is supposed to be installed to the KOMPSAT-5 as the sub-payload and present the results of the study about the operation strategies for the effective GPS radio occultation data gathering which is the satellite mission firstly performed in Korea.GPS radio occultation data can be utilized in various scientific applications including monitoring of atmosphere and ionosphere of the earth as well as weather forecast (Wickert et al. 2001).
In Section 2 of this article, we describe the composition and functions of the space-borne dual frequency GPS receiver as well as the operation strategies depending on the operation phase and mode of the KOMPSAT-5.Section 3 presents the study of the GPS radio occultation operation strategies.Finally, the conclusions and future works are described in Section 4.
Composition of the space-borne dual frequency GPS receiver (IGOR)
The AOPOD system of KOMPSAT-5 is composed of the IGOR, the space-borne dual frequency GPS receiver (Fig. 1a), and the LRRA for SLR (Fig. 1b).The space-borne dual frequency GPS receiver, produced by Broad Reach Engineering (BRE), USA, is based on the BlackJack receiver developed by NASA/JPL of which performance and stability was already verified as it was installed to the successful low earth orbit (LEO) satellite programs including JASON-1, SAC-C, CHAMP, GRACE, COSMIC and TerraSar-X (Williams et al. 2002).The space-borne dual frequency GPS receiver employs two occultation antennas (Fig. 2a) and two POD antennas (Fig. 2b).
The laser reflector in Fig. 1b, constructed by connecting the prism type reflectors in a certain form, is the 4-corner cube laser reflector produced by GeoForschun-gsZentrum (GFZ) Potsdam, Germany.
The basis specifications of the IGOR receiver are shown in Table 1.As shown in Fig. 3, IGOR consists following 5 parts: the RF receiving part to receive 4 RF signals, the digital part where the GPS signals are digitalized, the data storage device part consisting of a 128 M byte solid state recorder (SSR), the 1553B interface part for the MIL-STD-1553B communication with the spacecraft, and the ground test interface part for the ground test and RS-422 communication.In particular, all the parts except the RF receiving part are duplexed to secure the receiver stability (KASI 2009b).In the case of KOMPSAT-5, a total of 4 filter/preamp assemblies are installed in between the RF input port and the antenna in order to reduce the signal interference caused by the SAR signal.Fig. 1a shows one filter/preamp assembly installed to the top of the receiver of the IGOR.Two POD antennas and two occultation antennas are connected to the four RF input ports of the IGOR receiver (Lee et al. 2007).
IGOR data gathering and function test
The GPS signals received by the GPS antenna are stored in the SSR, the data storage device part, through the RF receiving part and the digital part.Since the capacity of the SSR data storage device is limited, the data is overwritten from the initial storage space of the SSR if the capacity of the received GPS data exceeds the SSR capacity.The current storage address of the raw data can be known through the telemetry.The GPS raw data stored in this way is transferred to the mass memory of KOMPSAT-5 and then transferred to and stored in the data storage server of the ground data center.The stored GPS raw data is transferred to the data processing server which processes the data to the receiver independent exchange (RINEX) format, the standard GPS data format, and store it.The GPS data processed in the ground data center can be used in various studies of the atmosphere and ionosphere including determination of satellite orbit and estimation of the temperature and pressure, vapor distribution and electron density in the atmosphere.The function of the IGOR is to provide the POD data for the determination of the precise satellite orbit and the GPS radio occultation data for scientific application research.KASI and Korea Aerospace Research Institute (KARI) carried out the tests to verify the performance and functions of the IGOR using GSS7700, the GPS signal simulator produced by Spirent, the 1553B control PC which is the data input/output interface, and the IGOR EM.Currently, KASI is carrying out the test of FM.The key test items are listed in Table 2 and the ETB for the IGOR EM test is shown in Fig. 4 (Choi et al. 2008).
GPS radio occultation
Radio occultation refers to the phenomenon where the celestial body such as a planet or star of which apparent size is small is occulted by the celestial body of which apparent size is large such as the moon or Jupiter.The signal delay phenomenon caused by the refraction of the radio wave by the atmosphere of a planet that occurs when the radio wave passes through the planet's atmosphere is also referred to as radio occultation.The GPS signal that is transferred from the GPS satellites at the altitude of approximately 20,200 km to the LEO satellites at the altitude of 250-1,500 km may also undergo refraction as it pass through the earth's atmosphere depending on the geometric arrangement of the two satellites.This is the GPS radio occultation that is found in the LEO satellites and Fig. 5 shows the basic principle of it (Wickert et al. 2001).The degree of refraction by the atmosphere is measured as the degree of bending of the signal pathway, the bending angle, which is denoted as α in Fig. 5.The angle α can be expressed, as in Eq. ( 1), using the vertical component of the signal direction, a (impact parameter), the distance from the center of the earth to the intersection point of the GPS signal direction and the bent curve, r , the tangent radius, r t , and the atmospheric refraction index, n.
Eq. ( 1) can be arranged to Eq. ( 2), as the equation with respect to atmospheric refraction index, by the Abelian transformation (Fjeldbo et al. 1971, Hajj et al. 2002).
Using the atmospheric refraction index derived from the bending angle (α) through the signal delay considering Doppler's effect, the temperature, pressure and the amount of vapor can be calculated.These coefficients are necessary in constructing global atmospheric model and usefully applied in weather forecast and climate monitoring.In addition to the atmospheric refraction index, the total electron density (TEC) of the ionosphere can be calculated from the signal delay to apply it to the ionosphere modeling and the space environment forecast.Lee et al. (2007) estimated the number of GPS radio occultation occurrences obtainable in one day by KOMP-SAT-5 using the GPS radio occultation simulation software, end-to-end GNSS occultation performance simulator 4 (EGOPS 4).Considering the geometric arrangement of the GPS satellite and KOMPSAT-5 and the signal bending effect using the exponential atmospheric model, the number of GPS radio occultation occurrence obtainable from the occultation antennas installed to the satellite afterward and forward was 500, approximately.However, the actual number of GPS radio occultation occurrence may be less than 400 because, in the actual operation in the orbit, the field of view of the antennas is limited by the main body of the satellite and the solar cell plate, and the data-obtaining rate of the afterward antenna is lower than that of the forward antenna.
Operation of KOMPSAT-5 IGOR
The system operation phase of KOMPSAT-5 is constituted as shown in Fig. 6 (KARI 2008).The pre-launch phase is the ground test stage before the satellite launching and the test stage of the current KOMPSAT-5.The launch and early operation phase is the stage after the initial satellite launching but just before the satellite begins the normal mission in the normal orbit.Finally, the mission operation phase is the stage where KOMP-SAT-5 carries out its mission in the normal orbit.In this study, the IGOR operation scenario was constituted with the two phases and eight modes shown in Table 3 based on the KOMPSAT-5 operation phases in Fig. 6.The POD only phase is the stage that is operated in the case where the basic performance status of the IGOR receiver and the basic constitution status are verified and only POD data is collected.The POD+OCC phase is the operation stage to obtain POD and occultation data together and Fig. 6.Korea multi-purpose satellite-5 system operation phases.it generally follows POD only phase.In addition, the operation modes of the individual phases were constituted depending on the performed functions in relation to the duplexing of the internal system of the IGOR.
During the initial operation, the IGOR works in the POD primary mode of the POD only phase in the digital B board.When the POD primary is working abnormally, the mode is converted to the POD redundant mode of the digital A board.Abnormal IGOR operation include the cases when an abnormal electric potential is applied to the IGOR receiver, when the GPS signal is not received continually for a certain duration after the initial booting of the IGOR receiver, and when the GPS signal is not received continually after the GPS signal that has been received normally by IGOR is cut.The conversion from the primary mode to the redundant mode is not automatically performed but by the remote command transmission, because an abnormal IGOR power supply to the IGOR can cause malfunction of the primary mode and the automatic conversion to the redundant mode can cause malfunction of the redundant mode as well, and thus affect the entire IGOR receiver.The POD primary mode and the redundant mode of the IGOR POD cannot be operated at the same time, but only either of them can.In addition, the occultation cannot be performed independently, but always together with the POD primary or redundant mode performance.Table 4 shows the five operation command categories that were constituted to operate the phases and modes of the IGOR and the individual operation command categories are composed of the remote command sets.
The IGOR operation is conducted in the following four modes in order: booting, fast search, slow search and normal operation.The fast search is the stage to find the basic GPS satellite signals for the IGOR operation and the slow search is the mode where the GPS signals are received using all the channels including the GPS signals received in the fast search mode and the Doppler values are calculated.Following the initial power supply to IGOR, all the steps are automatically conducted in order from the booting to the normal operation.The time flow status according to the general IGOR operation is shown in Table 5.The receiver operation status according to the IGOR operation shown in Table 5 can be changed depending on the KOMPSAT-5 environment and the arrangement of the GPS satellites (KASI 2009c).
As shown in Table 6, IGOR should set the initial parameters to perform the basic functions of the GPS receiver.The IGOR carries out the functions as it is under operation by applying the values in Table 6, and the parameters can be varied depending on the operation scenario if necessary.After the initial booting of the IGOR, one POD antenna is activated and the other antennas are in the inactivated state as shown in Table 6.Hence, for the operation, the ground station of the KOMPSAT-5 should transmit remote commands related to the antennas to convert the antenna state to the one that is appropriate to the operation scenario.
IGOR occultation parameter setting
The operation of the GPS radio occultation of the IGOR receiver is for the effective gathering of various scientific http://janss.krdata to be used in the monitoring of the atmosphere and ionosphere of the earth as well as weather forecast.It consists of control of the receiver constitution status, remote commanding, monitoring of the receiver and data status, trouble-shooting and alteration of the occultation parameters.In Table 3, the POD+OCC phase is divided into six modes depending on the GPS radio occultation functions of the IGOR and the KOMPSAT-5 operation scenarios.Additionally, for the effective occultation in each mode, four sets of occultation parameters were generated: setting atmospheric occultation (OCCAFT), rising atmospheric occultation (OCCFWD), setting ionospheric occultation (IONOAFT) and rising ionospheric occultation (IONOFWD).
As shown in Table 7, the parameter sets for the individual occultation functions were divided into nine data fields and the values for each data field was set by considering the KOMPSAT-5 orbit and the occultation events (KASI 2009a).
RegionBottom, the key occultation parameter, means the altitude where the data gathering is started or finished by the occultation antenna referring the limb of the earth.RegionTop means the altitude where the data gathering is started or finished by the occultation antenna referring the limb of the earth or the angle of the satellite from the local horizontal.
The parameters in Table 7 can be modified or constituted into new parameters sets depending on the KOMP-SAT-5 and the space environment (KASI 2009b).In the IGOR FM test using the occultation parameters in Table 7, the generation of the GPS radio occultation data was verified depending on the antenna setting and the pa- rameter set constitution, and the information needed to make the occultation plans for the emergent situations during the KOMPSAT-5 operation in the future was obtained.
CONCLUSIONS AND FUTURE WORKS
KOMPSAT-5 is a LEO earth-observation satellite of which main mission is to obtain high-precision images by the earth observation and it will be launched in 2010.According to the requirements of such a high-precision observation, the AOPOD system consisting of the spaceborne dual frequency GPS receiver and a LRRA for SLR is supposed to be installed to it for the first time in Korea.Particularly, the GPS radio occultation data that will be obtained by the AOPOD system will provide the vertical distribution of the physical values of the atmosphere and the ionosphere in a global scale.
In this article, the AOPOD system, the sub-payload of the KOMPSAT-5, and its functions were introduced and the operation strategy for the GPS radio occultation was described.Two phases and eight modes were established for the GPS radio occultation operation based on the KOMPSAT-5 operation phases and the various functions of IGOR.In addition, five IGOR operation categories were established and applied for the effective data gathering and each of the categories were constituted by the combination of various remote commands to perform the basic IGOR functions as well as the POD and occultation missions.
The KOMPSAT-5 occultation parameters referring to the KOMPSAT-5 orbit operation scenarios and the IGOR characteristics and they will be used in the AOPOD system operation when KOMPSAT-5 is operated in the future.KASI has developed KASI radio occultation processing system (KROPS), a GPS radio occultation data processing system, and it is under test, currently.The vertical distribution of the physical quantities in the earth's atmosphere will be obtained using the KOMPSAT-5 GPS radio occultation operation strategies suggested in this article.It will contribute to the study of the atmosphere and ionosphere as it is connected to the results that can be obtained from other space programs related to GPS radio occultation.
Table 2 .
IGOR test items and descriptions.
Table 5 .
IGOR start-up and boot timeline.
Table 6 .
IGOR default setting values.
|
2018-12-21T23:26:00.505Z
|
2010-12-15T00:00:00.000
|
{
"year": 2010,
"sha1": "da07348c42e5b55a382057501e64bafcf4279109",
"oa_license": "CCBYNC",
"oa_url": "http://koreascience.or.kr/article/JAKO201004140972388.pdf",
"oa_status": "GOLD",
"pdf_src": "ScienceParseMerged",
"pdf_hash": "da07348c42e5b55a382057501e64bafcf4279109",
"s2fieldsofstudy": [
"Engineering",
"Physics",
"Environmental Science"
],
"extfieldsofstudy": [
"Computer Science"
]
}
|
261356083
|
pes2o/s2orc
|
v3-fos-license
|
TITLE AND ABSTRACT 1 a ) TITLE : Identi � cation as a randomized trial in the title 1 a )
The CONSORT-EHEALTH checklist is intended for authors of randomized trials evaluating webbased and Internet-based applications/interventions, including mobile interventions, electronic games (incl multiplayer games), social media, certain telehealth applications, and other interactive and/or networked electronic applications. Some of the items (e.g. all subitems under item 5 description of the intervention) may also be applicable for other study designs.
TITLE AND ABSTRACT 1a) TITLE: Identi cation as a randomized trial in the title 1a) Does your paper address CONSORT item 1a? * I.e does the title contain the phrase "Randomized Controlled Trial"? (if not, explain the reason under "other") yes Other: 1a-i) Identify the mode of delivery in the title Identify the mode of delivery. Preferably use "web-based" and/or "mobile" and/or "electronic game" in the title. Avoid ambiguous terms like "online", "virtual", "interactive". Use "Internet-based" only if Intervention includes non-web-based Internet components (e.g. email), use "computer-based" or "electronic" only if o ine products are used. Use "virtual" only in the context of "virtual reality" (3-D worlds). Use "online" only in the context of "online support groups". Complement or substitute product names with broader terms for the class of products (such as "mobile" or "smart phone" instead of "iphone"), especially if the application runs on different platforms. 1 2 3 4 5 subitem not at all important essential Does your paper address subitem 1a-i? * Copy and paste relevant sections from manuscript title (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study A "web-based" transdiagnostic randomized control trial for affective and mood disorders
1a-ii) Non-web-based components or important co-interventions in title
Mention non-web-based components or important co-interventions in title, if any (e.g., "with telephone support").
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 1a-ii?
Copy and paste relevant sections from manuscript title (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 4/38 The item is not applicable for the current study because we used only an web-based intervention with no additional components.
1a-iii) Primary condition or target group in the title Mention primary condition or target group in the title, if any (e.g., "for children with Type I Diabetes") Example: A Web-based and Mobile Intervention with Telephone Support for Children with Type I Diabetes: Randomized Controlled Trial 1 2 3 4 5 subitem not at all important essential Does your paper address subitem 1a-iii? * Copy and paste relevant sections from manuscript title (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study A web-based transdiagnostic randomized control trial for "affective and mood disorders" 1b) ABSTRACT: Structured summary of trial design, methods, results, and conclusions NPT extension: Description of experimental treatment, comparator, care providers, centers, and blinding status.
1b-i) Key features/functionalities/components of the intervention and comparator in the METHODS section of the ABSTRACT
Mention key features/functionalities/components of the intervention and comparator in the abstract. If possible, also mention theories and principles used for designing the site. Keep in mind the needs of systematic reviewers and indexers by including important synonyms. (Note: Only report in the abstract what the main paper is reporting. If this information is missing from the main body of text, consider adding it) 1 2 3 4 5 subitem not at all important essential Does your paper address subitem 1b-i? * Copy and paste relevant sections from the manuscript abstract (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 5/38 "Included participants (N=105) were randomized to either the immediate treatment or the wait-list control group (WLCG). The transdiagnostic treatment was based on Barlow's Unified Protocol that addresses the underlying mechanisms of anxiety and depression." 1b-ii) Level of human involvement in the METHODS section of the ABSTRACT Clarify the level of human involvement in the abstract, e.g., use phrases like "fully automated" vs. "therapist/nurse/care provider/physician-assisted" (mention number and expertise of providers involved, if any). (Note: Only report in the abstract what the main paper is reporting. If this information is missing from the main body of text, consider adding it) 1 2 3 4 5 subitem not at all important essential Does your paper address subitem 1b-ii?
Copy and paste relevant sections from the manuscript abstract (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "During the next 10 weeks, participants in the active treatment group were guided through the program by eight graduate students in clinical psychology. The therapist assistance consisted of asynchronous written communication on a secured web-platform." 1b-iii) Open vs. closed, web-based (self-assessment) vs. face-to-face assessments in the METHODS section of the ABSTRACT Mention how participants were recruited (online vs. o ine), e.g., from an open access website or from a clinic or a closed online user group (closed usergroup trial), and clarify if this was a purely web-based trial, or there were face-to-face components (as part of the intervention or for assessment). Clearly say if outcomes were self-assessed through questionnaires (as common in web-based trials). Note: In traditional o ine trials, an open trial (open-label trial) is a type of clinical trial in which both the researchers and participants know which treatment is being administered. To avoid confusion, use "blinded" or "unblinded" to indicated the level of blinding instead of "open", as "open" in web-based trials usually refers to "open access" (i.e. participants can self-enrol). (Note: Only report in the abstract what the main paper is reporting. If this information is missing from the main body of text, consider adding it) 1 2 3 4 5 subitem not at all important essential Does your paper address subitem 1b-iii?
Copy and paste relevant sections from the manuscript abstract (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 6/38 "Volunteer participants register for the study and completed a series of online self-report screening measures. Participants who fulfilled the basic inclusion criteria on self-report measures were further contacted for a telephone interview based on the Structural Clinical Interview for DSM-IV (SCID-I)." "After the treatment all participants were invited to complete a set of self-report measures ... and SCID-I interview ... Six-month after the intervention the treatment group was invited to participate at the follow-up assessment..." 1b-iv) RESULTS section in abstract must contain use data Report number of participants enrolled/assessed in each group, the use/uptake of the intervention (e.g., attrition/adherence metrics, use over time, number of logins etc.), in addition to primary/secondary outcomes. (Note: Only report in the abstract what the main paper is reporting. If this information is missing from the main body of text, consider adding it) 1 2 3 4 5 subitem not at all important essential Does your paper address subitem 1b-iv?
Copy and paste relevant sections from the manuscript abstract (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "Included participants were randomized to either the immediate treatment (N=69) or the wait-list control group (WLCG, N=36) using a 2:1 ratio." "At the end of the program participants from the active treatment group completed on average 19 homework assignments ...and for the self-report measures we collected from 78.3% at post-intervention and from 52.2% at six-month follow-up assessment. Post-intervention SCID-I interviews were collected from 78.3% participants from the active treatment group and 80.6% from the control group.
1b-v) CONCLUSIONS/DISCUSSION in abstract for negative trials Conclusions/Discussions in abstract for negative trials: Discuss the primary outcome -if the trial is negative (primary outcome not changed), and the intervention was not used, discuss whether negative results are attributable to lack of uptake and discuss reasons. (Note: Only report in the abstract what the main paper is reporting. If this information is missing from the main body of text, consider adding it) 1 2 3 4 5 subitem not at all important essential Does your paper address subitem 1b-v?
Copy and paste relevant sections from the manuscript abstract (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "Insignificant between-group differences for the Y-BOCS and PDSS could be explained by the small number of participants with the associated primary diagnostic (e.g., only three participants with obsessive-compulsive disorder), by the choice of outcome measure (PDSS was not rated among the evidence-based measures) and by the fact that these disorders could be more difficult to treat." Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "Our hypothesis was that participants receiving the treatment would display significantly lower levels of depression and/or anxiety symptoms at the end of the treatment compared to those in the wait-list control group (WLCG) and that these improvements would be maintained 6-months following treatment." METHODS 3a) Description of trial design (such as parallel, factorial) including allocation ratio Does your paper address CONSORT subitem 3a? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study ".. we used a (phase II) simple randomized control trial (RCT) design in which participants were assigned to either an immediate treatment or waitlist control group (WLCG). Randomization followed a 2:1 ratio, such that two thirds of the participants were assigned to the immediate treatment group in order to maximize engagement and retain most participants for the post-intervention assessment. The random allocation sequence was generated by one of the authors (A.R.) who was not involved in selection process. " 3b) Important changes to methods after trial commencement (such as eligibility criteria), with reasons Does your paper address CONSORT subitem 3b? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study There were no changes in methods or eligibility criteria during the trial.
3b-i) Bug xes, Downtimes, Content Changes Bug xes, Downtimes, Content Changes: ehealth systems are often dynamic systems. A description of changes to methods therefore also includes important changes made on the intervention or comparator during the trial (e.g., major bug xes or changes in the functionality or content) (5-iii) and other "unexpected events" that may have in uenced study design such as staff changes, system failures/downtimes, etc. [2].
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 3b-i?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The platform was previously tested during an open trial designed for healthy participants, and the only improvement consisted of the auto-save option for homework assignments."
4a) Eligibility criteria for participants
Does your paper address CONSORT subitem 4a? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The online recruitment process was designed to be broadly inclusive, with few exclusion criteria. Individuals were eligible for the study if they: (a) were fluent in Romanian, (b) were at least 18 years of age ... Participants were excluded from the study if they: (a) reported significant suicidal ideation... Individuals taking psychotropic medications at the time of enrollment were required to be stable on the same dose for at least four weeks prior to enrolling in the study..."
4a-i) Computer / Internet literacy
Computer / Internet literacy is often an implicit "de facto" eligibility criterion -this should be explicitly clari ed. 4a-ii) Open vs. closed, web-based vs. face-to-face assessments: Open vs. closed, web-based vs. face-to-face assessments: Mention how participants were recruited (online vs. o ine), e.g., from an open access website or from a clinic, and clarify if this was a purely web-based trial, or there were face-to-face components (as part of the intervention or for assessment), i.e., to what degree got the study team to know the participant. In online-only trials, clarify if participants were quasianonymous and whether having multiple identities was possible or whether technical or logistical measures (e.g., cookies, email con rmation, phone calls) were used to detect/prevent these.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 4a-ii? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The study was advertised in the local and national media following a press conference organized by the West University of Timisoara in April 2016. Interested participants could freely register or enroll for the study online. After electronically signing the Informed consent (i.e., compulsory check box), participants were instructed to complete a series of online self-report measures... Participants who ... scored in the range of the cut-off score(s) (mild to moderate clinical symptoms) were invited to take a phone interview based on the SCID-I." 4a-iii) Information giving during recruitment Information given during recruitment. Specify how participants were briefed for recruitment and in the informed consent procedures (e.g., publish the informed consent documentation as appendix, see also item X26), as this information may have an effect on user self-selection, user expectation and may also bias results.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 4a-iii?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "After electronically signing the Informed consent (i.e., compulsory check box), participants were instructed to complete a series of self-report measures." "If participants met the diagnostic criteria for at least one affective and/or anxiety disorders, they were invited to take part in the study. At the end of the recruitment process, all registered participants received a general feedback based on their screening results..." https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 11/38 4b) Settings and locations where the data were collected Does your paper address CONSORT subitem 4b? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study Data were collected in Romania. "The study was advertised in the local and national media following a press conference organized by the West University of Timisoara in April 2016. Interested participants could freely register or enroll for the study online (Romanian site https://www.e-cbt.ro/program/psitad/ archived at http://www.webcitation.org/6sZPMh6Jh).
4b-i) Report if outcomes were (self-)assessed through online questionnaires
Clearly report if outcomes were (self-)assessed through online questionnaires (as common in web-based trials) or otherwise.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 4b-i? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "participants were instructed to complete a series of online self-report measures as part of the screening process" "Participants who completed the online screening and scored in the range of the cut-off score(s) (mild to moderate clinical symptoms) were contacted and invited to take a phone interview based on the Structural Clinical Interview for DSM-IV (SCID-I)."
4b-ii) Report how institutional a liations are displayed
Report how institutional a liations are displayed to potential participants [on ehealth media], as a liations with prestigious hospitals or universities may affect volunteer rates, use, and reactions with regards to an intervention.(Not a required item -describe only if this may bias results) 1 2 3 4 5 subitem not at all important essential Does your paper address subitem 4b-ii?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 12/38 "The study was advertised in the local and national media following a press conference organized by the West University of Timisoara in April 2016. Interested participants could freely register for the study online" 5) The interventions for each group with su cient details to allow replication, including how and when they were actually administered 5-i) Mention names, credential, a liations of the developers, sponsors, and owners Mention names, credential, a liations of the developers, sponsors, and owners [6] (if authors/evaluators are owners or developer of the software, this needs to be declared in a "Con ict of interest" section or mentioned elsewhere in the manuscript).
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 5-i?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The web-platform was design as an infrastructure to facilitate written asynchronous communication....The web-platform was developed by a Romanian IT team coordinated by the first author. "
5-ii) Describe the history/development process
Describe the history/development process of the application and previous formative evaluations (e.g., focus groups, usability testing), as these will have an impact on adoption/use rates and help with interpreting results.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 5-ii?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 13/38 "Our web-platform consists of two distinct but interconnected modules designed for online assessment and online psychotherapy... Access to the platform is controlled by ID and Password and all sensitive content is encrypted and stored on a secure server. The platform was previously tested during an open trial designed for healthy participants, and the only improvement consisted of the auto-save option for homework assignments."
5-iii) Revisions and updating
Revisions and updating. Clearly mention the date and/or version number of the application/intervention (and comparator, if applicable) evaluated, or describe whether the intervention underwent major changes during the evaluation process, or whether the development and/or content was "frozen" during the trial.
Describe dynamic components such as news feeds or changing content which may have an impact on the replicability of the intervention (for unexpected events see item 3b).
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 5-iii?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "We retained only nine treatment sessions for our guided intervention, as our previous experience suggests this represents an adequate length for treating participants over the internet. The transdiagnostic program was designed as a stand-alone intervention, and the program content was unchanged during the trial."
5-iv) Quality assurance methods
Provide information on quality assurance methods to ensure accuracy and quality of information provided [1], if applicable.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 5-iv?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The Romania version of the program was based on Barlow and colleague's Unified Protocol for Transdiagnostic Treatment of Emotional Disorders (UP) [16]. UP treatment modules were adapted for the online environment, but the treatment structure was conceptually similar." "All participants were guided through the program in the same order (session 1 through 9), and access to the next session was granted if participants partially completed their homework assignments of the previous session." https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 14/38 5-v) Ensure replicability by publishing the source code, and/or providing screenshots/screencapture video, and/or providing owcharts of the algorithms used Ensure replicability by publishing the source code, and/or providing screenshots/screen-capture video, and/or providing owcharts of the algorithms used. Replicability (i.e., other researchers should in principle be able to replicate the study) is a hallmark of scienti c reporting.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 5-v?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study Our web platform represent just the infrastructure where the content of the intervention (any workbook or treatment manual) can be presented and homework assignments can be filled-in in an email-like format.
5-vi) Digital preservation
Digital preservation: Provide the URL of the application, but as the intervention is likely to change or disappear over the course of the years; also make sure the intervention is archived (Internet Archive, webcitation.org, and/or publishing the source code or screenshots/videos alongside the article). As pages behind login screens cannot be archived, consider creating demo pages which are accessible without login.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 5-vi?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "Interested participants could freely register or enroll for the study online (Romanian site https://www.e-cbt.ro/program/psitad/ archived at http://www.webcitation.org/6sZPMh6Jh). "
5-vii) Access
Access: Describe how participants accessed the application, in what setting/context, if they had to pay (or were paid) or not, whether they had to be a member of speci c group. If known, describe how participants obtained "access to the platform and Internet" [1]. To ensure access for editors/reviewers/readers, consider to provide a "backdoor" login account or demo mode for reviewers/readers to explore the application (also important for archiving purposes, see vi).
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 5-vii? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "Active treatment participants could access the nine sessions using their own device (computer, tablet etc.) at a time and place of their convenience." "All participants were assisted free of charge for a ten-week interval: nine weeks for each session and one extra week for their eventual delays." A backdoor login account (in Romanian) could be found at https://ecbt.ro/program/psitad/login.php user: 1529tyre, pass: t1rxj3fv subitem not at all important essential Does your paper address subitem 5-viii? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "At the beginning of the treatment participants were encouraged to complete one session per week ... The Romania version of the program was based on Barlow and colleague's Unified Protocol for Transdiagnostic Treatment of Emotional Disorders "The program first sought to increase participants' motivation for the transdiagnostic treatment ...(session 1).... Finally, participants were asked to review the strategies learned ... (session 9). "Eight graduate students in clinical psychology assisted and guided participants throughout the treatment ... "
5-ix) Describe use parameters
Describe use parameters (e.g., intended "doses" and optimal timing for use). Clarify what instructions or recommendations were given to the user, e.g., regarding timing, frequency, heaviness of use, if any, or was the intervention used ad libitum. not in the ms, or brie y explain why the item is not applicable/relevant for your study "At the beginning of the treatment participants were encouraged to complete one session per week and the associated homework assignments." "During the next 10 weeks, participants in the active treatment group were guided through the program by eight graduate students in clinical psychology. The guidance consisted of asynchronous written communication on a secure web-platform."
5-x) Clarify the level of human involvement
Clarify the level of human involvement (care providers or health professionals, also technical assistance) in the e-intervention or as co-intervention (detail number and expertise of professionals involved, if any, as well as "type of assistance offered, the timing and frequency of the support, how it is initiated, and the medium by which the assistance is delivered". It may be necessary to distinguish between the level of human involvement required for the trial, and the level of human involvement required for a routine application outside of a RCT setting (discuss under item 21 -generalizability).
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 5-x?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "Supervised by an experienced psychotherapist, the graduate students provided online personalized feedback for participants' homework assignments and answered their questions within a 24h interval. In terms of the frequency of message exchange, if participants did not initiate a written message with the graduate student assigned to them (which was seldom the case), they received a weekly feedback for their homework assignments. In case of inactivity, a participant received up to three written reminders on the platform, at a rate of one message per week."
5-xi) Report any prompts/reminders used
Report any prompts/reminders used: Clarify if there were prompts (letters, emails, phone calls, SMS) to use the application, what triggered them, frequency etc. It may be necessary to distinguish between the level of prompts/reminders required for the trial, and the level of prompts/reminders for a routine application outside of a RCT setting (discuss under item 21 -generalizability). "In terms of the frequency of message exchange, if participants did not initiate a written message with the graduate student assigned to them (which was seldom the case), they received a weekly feedback for their homework assignments. In case of inactivity, a participant received up to three written reminders on the platform, at a rate of one message per week." 5-xii) Describe any co-interventions (incl. training/support) Describe any co-interventions (incl. training/support): Clearly state any interventions that are provided in addition to the targeted eHealth intervention, as ehealth intervention may not be designed as stand-alone intervention. This includes training sessions and support [1]. It may be necessary to distinguish between the level of training required for the trial, and the level of training for a routine application outside of a RCT setting (discuss under item 21 -generalizability.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 5-xii? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The transdiagnostic program was designed as a stand-alone intervention, and the program content was unchanged during the trial." 6a) Completely de ned pre-speci ed primary and secondary outcome measures, including how and when they were assessed Does your paper address CONSORT subitem 6a? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study All self-report measures were completed online by participants.
Previous study show that no differences emerged between the online and the paper-and-pencil versions of the same measure.
6a-ii) Describe whether and how "use" (including intensity of use/dosage) was de ned/measured/monitored
Describe whether and how "use" (including intensity of use/dosage) was de ned/measured/monitored (logins, log le analysis, etc.). Use/adoption metrics are important process outcomes that should be reported in any ehealth trial.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 6a-ii?
Copy and paste relevant sections from manuscript text "Treatment adherence was estimated using participants online behavior: a) how often they accessed the online treatment (number of logins) and b) how often they were actively engaged with the content of the treatment (number of completed homework assignments -possible range: 0 -41). During the 10-week treatment period, the average number of platform accesses was 46.76 (SD = 29.86) per participant ... At the end of the treatment participants estimated to have spent an average 4 hours/week in treatment-related activities (SD = 3.53; Median = 2.5 hours/week)." 6a-iii) Describe whether, how, and when qualitative feedback from participants was obtained Describe whether, how, and when qualitative feedback from participants was obtained (e.g., through emails, feedback forms, interviews, focus groups).
2 3 4 5
subitem not at all important essential Does your paper address subitem 6a-iii?
Copy and paste relevant sections from manuscript text "After the intervention, most treated participants declared to be satisfied ... with the program...They also declared that the information offered within the program was highly qualitative ... Finally, using a 10-point scale, participants declared that the transdiagnostic program appeared logical (M = 8.57, SD = 1.99), and that they are confident to recommend it to someone facing similar difficulties (M = 8.38, SD = 2.18)." 6b) Any changes to trial outcomes after the trial commenced, with reasons Does your paper address CONSORT subitem 6b? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study No changes on trial outcomes during the trial 7a) How sample size was determined NPT: When applicable, details of whether and how the clustering by care provides or centers was addressed 7a-i) Describe whether and how expected attrition was taken into account when calculating the sample size Describe whether and how expected attrition was taken into account when calculating the sample size.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 7a-i?
Copy and paste relevant sections from manuscript title (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The RCT was powered to detect an effect size of d=.60 at post-treatment (alpha = .01) at a power of 95% (1-β). A total of 105 participants were needed for the study, when considering a potential dropout risk of 20%." https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 20/38 7b) When applicable, explanation of any interim analyses and stopping guidelines Does your paper address CONSORT subitem 7b? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study No interim analyses was performed.
No stopping guidelines were used.
8a) Method used to generate the random allocation sequence NPT: When applicable, how care providers were allocated to each trial group Does your paper address CONSORT subitem 8a? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study One of the authors (A.R.), who was not involved in selection process, generated the pseudorandom allocation sequence by using an available online tool (https://www.randomizer.org).
8b) Type of randomisation; details of any restriction (such as blocking and block size) Does your paper address CONSORT subitem 8b? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "... we used a (phase II) simple randomized control trial (RCT) design in which participants were assigned to either an immediate treatment or waitlist control group (WLCG). Randomization followed a 2:1 ratio ..." No other restriction on randomization was used.
https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 21/38 9) Mechanism used to implement the random allocation sequence (such as sequentially numbered containers), describing any steps taken to conceal the sequence until interventions were assigned Does your paper address CONSORT subitem 9? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The randomization was conducted with all included participants one day before starting the intervention program." No additional steps to conceal the sequence were taken.
10) Who generated the random allocation sequence, who enrolled participants, and who assigned participants to interventions Does your paper address CONSORT subitem 10? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "One of the authors (A.R.), who was not involved in selection process, generated the pseudorandom allocation sequence by using an available online tool (https://www.randomizer.org)." "Included participants were allocated to the either the active treatment group or the WLCG by the graduate students." "Interested participants could freely register or enroll for the study online " 11a) If done, who was blinded after assignment to interventions (for example, participants, care providers, those assessing outcomes) and how NPT: Whether or not administering co-interventions were blinded to group assignment 11a-i) Specify who was blinded, and who wasn't Specify who was blinded, and who wasn't. Usually, in web-based trials it is not possible to blind the participants [1, 3] (this should be clearly acknowledged), but it may be possible to blind outcome assessors, those doing data analysis or those administering co-interventions (if any). Does your paper address subitem 11a-i? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "the post-intervention SCID-I interviews were conducted by a different team of graduate students blinded to participants' group and diagnostic status." 11a-ii) Discuss e.g., whether participants knew which intervention was the "intervention of interest" and which one was the "comparator" Informed consent procedures (4a-ii) can create biases and certain expectations -discuss e.g., whether participants knew which intervention was the "intervention of interest" and which one was the "comparator".
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 11a-ii?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study Since we use only one active group (with the transdiagnostic intervention program) -we did not encounter this problem.
11b) If relevant, description of the similarity of interventions
(this item is usually not relevant for ehealth trials as it refers to similarity of a placebo or sham intervention to a active medication/intervention) Does your paper address CONSORT subitem 11b? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study 12a) Statistical methods used to compare groups for primary and secondary outcomes NPT: When applicable, details of whether and how the clustering by care providers or centers was addressed Does your paper address CONSORT subitem 12a? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The post-intervention and follow-up data were analyzed based on the intention-to-treat (ITT) framework by the means of linear mixed models. This approach uses all available observations (maximum likelihood estimation) and allows for correlation between longitudinal data. Time (baseline vs. post-treatment) was set as the within-group factor and trial condition (transdiagnostic vs. wait-list) was used as the between-groups factor...." 12a-i) Imputation techniques to deal with attrition / missing values Imputation techniques to deal with attrition / missing values: Not all participants will use the intervention/comparator as intended and attrition is typically high in ehealth trials. Specify how participants who did not use the application or dropped out from the trial were treated in the statistical analysis (a complete case analysis is strongly discouraged, and simple imputation techniques such as LOCF may also be problematic [4]).
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 12a-i? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The post-intervention and follow-up data were analyzed based on the intention-to-treat (ITT) framework by the means of linear mixed models. This approach uses all available observations (maximum likelihood estimation) and allows for correlation between longitudinal data." "...We computed Hedges's g effects size estimates for both betweengroups and with-group comparisons."
12b) Methods for additional analyses, such as subgroup analyses and adjusted analyses
Does your paper address CONSORT subitem 12b? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 24/38 No subgroup analyses was used. For the adjusted model "the fix effects of gender, previous psychotherapy experience in the past 4 years (dummy coded: yes / no), and treatment credibility were added to each model. ... The fix effect of total number of homework assignments completed by participants (treatment adherence) was also added to each of these models. "
X26) REB/IRB Approval and Ethical Considerations
[recommended as subheading under "Methods"] (not a CONSORT item) X26-i) Comment on ethics committee approval 1 2 3 4 5 subitem not at all important essential Does your paper address subitem X26-i?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The study was approved by the Ethical Commission of West University of Timisoara, Romania (4509/26.02.2016) and was registered on ClinicalTrials.gov as NCT02739607. Written informed consent was obtained from all participants by surface mail."
x26-ii) Outline informed consent procedures
Outline informed consent procedures e.g., if consent was obtained o ine or online (how? Checkbox, etc.?), and what information was provided (see 4a-ii). See [6] for some items to be included in informed consent documents.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem X26-ii?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 25/38 "After electronically signing the Informed consent (i.e., compulsory check box), participants were instructed to complete a series of online self-report measures as part of the screening..."
X26-iii) Safety and security procedures
Safety and security procedures, incl. privacy considerations, and any steps taken to reduce the likelihood or detection of harm (e.g., education and training, availability of a hotline) 1 2 3 4 5 subitem not at all important essential Does your paper address subitem X26-iii?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "Access to the platform is controlled by ID and Password and all sensitive content is encrypted and stored on a secure server."
RESULTS
13a) For each group, the numbers of participants who were randomly assigned, received intended treatment, and were analysed for the primary outcome NPT: The number of care providers or centers performing the intervention in each group and the number of patients treated by each care provider in each center Does your paper address CONSORT subitem 13a? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The included participants were randomized into the treatment (N=69) and the control group (N=36). After the intervention, we conducted the SCID-I interview with 54/69 participants (78.3%) from the active treatment group and with 29/36 participants (80.6%) from the WLCG. A similar percentage of post-intervention self-report measures were collected from both groups (see Figure 1)... Six-month follow-up assessment questionnaire were collected from 36/69 participants (52.2%) in the active treatment group. " https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 26/38 13b) For each group, losses and exclusions after randomisation, together with reasons Does your paper address CONSORT subitem 13b? (NOTE: Preferably, this is shown in a CONSORT ow diagram) * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "Participants who changed psychotropic medication or started another treatment were excluded from these analyses." "During the nine-week intervention program 5/69 participants (7.2%) from the treatment group and 3/36 (8.3%) from the control group started/changed their psychotropic medication or started another treatment."
13b-i) Attrition diagram
Strongly recommended: An attrition diagram (e.g., proportion of participants still logging in or using the intervention/comparator in each group plotted over time, similar to a survival curve) or other gures or tables demonstrating usage/dose/engagement.
2 3 4 5
subitem not at all important essential Does your paper address subitem 13b-i?
Copy and paste relevant sections from the manuscript or cite the gure number if applicable (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "Participants adherence during the program is detailed in the attrition diagram (see Figure 2)."
14a) Dates de ning the periods of recruitment and followup
Does your paper address CONSORT subitem 14a? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 27/38 "The study was advertised in the local and national media ... in April 2016." "After the ten-week program, all included participants were invited to complete the post-intervention online assessment, and were contacted by phone to complete a diagnostic interview based on SCID-I." "a six-month follow-up assessment was taken for the immediate-treatment group only" 14a-i) Indicate if critical "secular events" fell into the study period Indicate if critical "secular events" fell into the study period, e.g., signi cant changes in Internet resources available or "changes in computer hardware or Internet delivery resources" 1 2 3 4 5 subitem not at all important essential Does your paper address subitem 14a-i?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study No critical secular events occur during the trial.
14b) Why the trial ended or was stopped (early) Does your paper address CONSORT subitem 14b? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study The trial was not stopped / ended prematurely. Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "Details regarding participants' demographic characteristics are presented in Table 1.... Principal and comorbid clinical diagnoses are also presented in Table 1. Overall, disorders ranged between 1 and 5, with an average of 1.67 disorders per participant (SD = .93)."
15-i) Report demographics associated with digital divide issues
In ehealth trials it is particularly important to report demographics associated with digital divide issues, such as age, education, gender, social-economic status, computer/Internet/ehealth literacy of the participants, if known.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 15-i? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "Participant's overall mean age was 34.27, ... A lower proportion of males ended up in the WLCG compared to the active treatment group (P = .049)... Even though not statistically significant (P = .070), there was a lower proportion of participants who benefited from psychotherapy in the previous four years in the active treatment group than in the WLCG. In all other respects, the treatment and control groups were similar in terms of demographic characteristics, including their time spend online." 16) For each group, number of participants (denominator) included in each analysis and whether the analysis was by original assigned groups 16-i) Report multiple "denominators" and provide de nitions Report multiple "denominators" and provide de nitions: Report N's (and effect sizes) "across a range of study participation [and use] thresholds" [1], e.g., N exposed, N consented, N used more than x times, N used more than y weeks, N participants "used" the intervention/comparator at speci c pre-de ned time points of interest (in absolute and relative numbers per group). Always clearly de ne "use" of the intervention.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 16-i? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study Since we used linear mixed models as the main statistical methods, all available observations (maximum likelihood estimation) were utilized.
16-ii) Primary analysis should be intent-to-treat
Primary analysis should be intent-to-treat, secondary analyses could include comparing only "users", with the appropriate caveats that this is no longer a randomized sample (see 18-i).
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 16-ii?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study Same as above.
17a) For each primary and secondary outcome, results for each group, and the estimated effect size and its precision (such as 95% con dence interval) Does your paper address CONSORT subitem 17a? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study " Table 2 includes the observed means and estimated marginal means for the primary outcomes measures. Linear mixed model results for the effectiveness of the intervention at post-treatment and Hedges's g effect sizes are displayed in Table 3." " Table 4 includes the descriptive statistics for the secondary outcomes and Table 5 displays the group by time interaction results of the linear mixed models and the Hedges's g effect size estimates." 17a-i) Presentation of process outcomes such as metrics of use and intensity of use In addition to primary/secondary (clinical) outcomes, the presentation of process outcomes such as metrics of use and intensity of use (dose, exposure) and their operational de nitions is critical. This does not only refer to metrics of attrition (13-b) (often a binary variable), but also to more continuous exposure metrics such as "average session length". These must be accompanied by a technical description how a metric like a "session" is de ned (e.g., timeout after idle time) [1] (report under item 6a). subitem not at all important essential Does your paper address subitem 17a-i?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "... and on average, participants completed 2.18 weekly assignments. At the end of the treatment participants estimated to have spent an average 4 hours/week in treatment-related activities (SD = 3.53; Median = 2.5 hours/week)." 17b) For binary outcomes, presentation of both absolute and relative effect sizes is recommended Does your paper address CONSORT subitem 17b? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study Effect size was estimated only for continuous outcome measures.
18) Results of any other analyses performed, including subgroup analyses and adjusted analyses, distinguishing pre-speci ed from exploratory Does your paper address CONSORT subitem 18? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study All analyses were conducted for the crude and adjusted model. An example of this is the following comment "SPIN scores were significantly reduced in the crude model, but the adjusted one drifted towards a trend (P = .057)."
18-i) Subgroup analysis of comparing only users
A subgroup analysis of comparing only users is not uncommon in ehealth trials, but if done, it must be stressed that this is a self-selected sample and no longer an unbiased sample from a randomized trial (see 16-iii).
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 18-i?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study No subgroup analyses was conducted.
19) All important harms or unintended effects in each group
(for speci c guidance see CONSORT for harms) Does your paper address CONSORT subitem 19? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study No harm or unintended effects were detected during the trial.
19-i) Include privacy breaches, technical problems
Include privacy breaches, technical problems. This does not only include physical "harm" to participants, but also incidents such as perceived or real privacy breaches [1], technical problems, and other unexpected/unintended incidents. "Unintended effects" also includes unintended positive effects [2]. Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study Qualitative feedback from participants was not included in the paper. We could include one/two such quotes if needed.
22)
Interpretation consistent with results, balancing bene ts and harms, and considering other relevant evidence NPT: In addition, take into account the choice of the comparator, lack of or partial blinding, and unequal expertise of care providers or centers in each group 22-i) Restate study questions and summarize the answers suggested by the data, starting with primary outcomes and process outcomes (use) Restate study questions and summarize the answers suggested by the data, starting with primary outcomes and process outcomes (use).
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 22-i? * "Finally, our results should be considered in light of several limitations. ... we excluded participants with very complex symptoms (N=11) and the mean number of baseline diagnostics per participant were somehow smaller in our sample compared to other trials... Finally, roughly half of participants in the active treatment condition failed to complete the follow up questionnaires at six-month follow-up." 21) Generalisability (external validity, applicability) of the trial ndings NPT: External validity of the trial ndings according to the intervention, comparators, patients, and care providers or centers involved in the trial
21-i) Generalizability to other populations
Generalizability to other populations: In particular, discuss generalizability to a general Internet population, outside of a RCT setting, and general patient population, including applicability of the study results for other organizations 1 2 3 4 5 subitem not at all important essential Does your paper address subitem 21-i?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The treatment approach appeared logical to most participants and they seemed willing to recommend the transdiagnostic program to other people with similar problems. This means that the program appear useful to participants and it could eventually be implemented to similar samples of internet users, outside the present trial."
21-ii) Discuss if there were elements in the RCT that would be different in a routine application setting
Discuss if there were elements in the RCT that would be different in a routine application setting (e.g., prompts/reminders, more human involvement, training sessions or other co-interventions) and what impact the omission of these elements could have on use, adoption, or outcomes if the intervention is applied outside of a RCT setting.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem 21-ii?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study https://docs.google.com/forms/d/e/1FAIpQLSfZBSUp1bwOc_OimqcS64RdfIAFvmrTSkZQL2-3O8O9hrL5Sw/formResponse?hl=en_US 35/38 "One potentially positive feature that could be added in clinical practice is a short (bimonthly) telephone contact, that could contribute to solving simple problems and stimulate participants involvement."
OTHER INFORMATION
23) Registration number and name of trial registry Does your paper address CONSORT subitem 23? * Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study ClinicalTrials.gov (NCT02739607) 24) Where the full trial protocol can be accessed, if available Does your paper address CONSORT subitem 24? * Cite a Multimedia Appendix, other reference, or copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study X27) Con icts of Interest (not a CONSORT item) X27-i) State the relation of the study team towards the system being evaluated In addition to the usual declaration of interests ( nancial or otherwise), also state the relation of the study team towards the system being evaluated, i.e., state if the authors/evaluators are distinct from or identical with the developers/sponsors of the intervention.
1 2 3 4 5 subitem not at all important essential Does your paper address subitem X27-i?
Copy and paste relevant sections from the manuscript (include quotes in quotation marks "like this" to indicate direct quotes from your manuscript), or elaborate on this item by providing additional information not in the ms, or brie y explain why the item is not applicable/relevant for your study "The Romanian team developed the e-cbt.ro web-platform that is exclusively used for research purposes. All authors declare to have no conflict of interest (financial or otherwise) regarding this study."
About the CONSORT EHEALTH checklist
As a result of using this checklist, did you make changes in your manuscript? * yes, major changes yes, minor changes no What were the most important changes you made as a result of using this checklist?
Adding procedural details regarding the study. How much time did you spend on going through the checklist INCLUDING making changes in your manuscript * About 8-9 hours.
As a result of using this checklist, do you think your manuscript has improved? * yes no Other: Would you like to become involved in the CONSORT EHEALTH group? This would involve for example becoming involved in participating in a workshop and writing an "Explanation and Elaboration" document yes no Other:
Any other comments or questions on CONSORT EHEALTH
No additional comments.
STOP -Save this form as PDF before you click submit
To generate a record that you lled in this form, we recommend to generate a PDF of this page (on a Mac, simply select "print" and then select "print as PDF") before you submit it.
When you submit your (revised) paper to JMIR, please upload the PDF as supplementary le.
Don't worry if some text in the textboxes is cut off, as we still have the complete information in our database. Thank you!
|
2018-05-30T19:18:28.679Z
|
2016-01-01T00:00:00.000
|
{
"year": 2017,
"sha1": "00e927c174a3732c40d8578df1b14db76c390cdd",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.2196/mhealth.6967",
"oa_status": "GOLD",
"pdf_src": "MergedPDFExtraction",
"pdf_hash": "00e927c174a3732c40d8578df1b14db76c390cdd",
"s2fieldsofstudy": [
"Computer Science",
"Psychology"
],
"extfieldsofstudy": []
}
|
158055382
|
pes2o/s2orc
|
v3-fos-license
|
SPECIFIC OPERATIONS FOR ACCOUNTS ’ HOMOGENIZATION WITH A VIEW TO CONSOLIDATION
The preparation of annual consolidated financial statements by group companies represents a relatively new problematic issue for regulating bodies, the world of the academia, as well as for accounting practitioners. The current national and international accounting standards undergo a continuous change and updating to the economic and judicial reality of the business environment which, in turn, goes through a permanent transformation because of the globalization of national economies. These novelties raise the necessity of debates organized by accounting specialists with the aim of finding solutions that respect legal regulations related to the preparation of financial statements for consolidation. The present article approaches accounting aspects related to the operation of adjusting individual financial statements that are the object of consolidation.
Accounting standards regarding the preparation of annual consolidated financial statements 1 stipulate a series of rules concerning the homogenization of the financial statements of the entities that are to be consolidated, as follows: a) when an entity included in the consolidation perimeter/scope (group member) employs, for similar transactions and events under similar circumstances, other accounting policies than those adopted in the case of annual consolidated financial statements, in order to ensure conformity to the group accounting policies, before being included in the consolidated financial statements, the individual financial statements of the respective entity must be adjusted in the corresponding manner; b) the entity that prepares the annual consolidated financial statements must apply the same evaluation methods as in the case of its own annual financial statements, which means that the assets and the liabilities of the entities included in the consolidation perimeter/scope will be evaluated by means of the same methods when they are included in the annual consolidated financial statements, methods which the 1 Accounting Standards regarding annual individual financial statements and annual consolidated financial statements, approved by OMPF no.802 dated December 29 2014, published in the Official Gazette of Romania, no.963 dated December 30 2014, part 8.6, pos.517-522.⃰ Consolidation can also be carried out on the basis of cash flows (see Marian Săcărin, "Contabilitate aprofundată"par.Consolidation Modes, p. 228).
entity that prepares consolidated financial statements also uses for drawing up its own annual financial statements, respectively its annual consolidated financial statements; c) în the case of assets which were adjusted to fair market value, only for fiscal purposes, the annual consolidated financial statements will include them only after eliminating the respective adjustments.We can thus conclude that the annual individual financial statements of the entities included in the consolidation perimeter/scope, i.e. subsidiaries, associated entities or entities over which common control is exercised and which use evaluation methods or accounting policies that differ from those applied when drawing up annual consolidated financial statements, must be restated (adjusted) obligatorily and waivers from this requirement are accepted only in exceptional cases.Possible waivers or exceptions are presented and justified in explanatory notes for annual consolidated financial statements.
In the case of consolidation on the basis of accounts aggregation ⃰ , adjustments are made both for the differences corresponding to the current financial year and for the differences corresponding to the previous financial years, with the mentioning that the adjustments corresponding to the previous financial years affect equities (reported result, reserves, etc.) and the adjustments corresponding to the current financial year affect its financial result.The operations regarding the adjustments that occur when including the financial statements in the consolidated accounts of the entities that are part of the consolidation scope involve accounting reasoning that practitioners should perform with significant judgment and professional competence.We will further present the accounting treatment corresponding to the various types of adjustments (restatements) occurring as a result of different amortization or depreciation policies.
A.
Company "B" purchased on the 25 th of December 2010 a vehicle that cost 160,000 lei.According to the company's depreciation policy, the firm will depreciate the respective fixed asset in a straight line manner, over a period of five years.Starting with the 1 st of January 2013, company "B" becomes a subsidiary of company "A", with the latter having the intention to include the former in the scope of consolidation.Company "A" (the parent company), according to its own policies regarding the depreciation of tangible fixed assets, has its own rule concerning the same category of fixed assets, respectively a working period of 8 years and a straight line depreciation type.Given the different depreciation policies used by the parent company and its subsidiary for similar tangible fixed assets, at the time of the consolidation, company "A" will start adjusting (restating) the annual consolidated financial statements prepared by company "B" (subsidiary) in compliance with the Accounting standards regarding consolidated financial statements in effect, which were previously mentioned.The differences between the depreciation calculated according to the accounting policies of the parent company and the depreciation calculated by the subsidiary in line with its own accounting policies used for preparing annual consolidated financial statements, are presented as follows: N o.⃰ Note: The differences of depreciation (annual and cumulated) are calculated between the (annual and cumulated) depreciation determined in compliance with the accounting policies of the parent company and the depreciation calculated by the consolidated company (company "B", the subsidiary).
Financ
The accounting treatment regarding the restatement of differences between the depreciation calculated in compliance with the policies of the parent company and the depreciation calculated by its subsidiary, at the time of the consolidation, is the following: 1.In the first year of consolidation, i.e. in 2013, the difference of depreciation corresponding to the current year (-12,000 lei) will be recorded as depreciation reduction, on the basis of the current result by means of depreciation expenses, and the preceding differences of depreciation (corresponding to 2011 and 2012) will influence equities (retained earnings/reserves), as follows: 4. In the financial year 2016, in the accounting records of the subsidiary, the vehicle is entirely depreciated.In the consolidated accounts there should be recorded, for the current period, a supplement for expenses with depreciation corresponding to the annual depreciation quota established in compliance with the policies of the parent company (+20,000 lei) and there should be adjusted the depreciation calculated in excess of what is due in previous financial years for the sum of 60,000 lei, as follows: B. At the end of the financial year N, company A, a subsidiary of company B, purchases a tangible fixed asset which is worth 150,000 lei.In the individual accounts of the subsidiary, it is depreciated in a straight line manner, over a period of ten years, whereas the accounting policies of company B also stipulate a straight line depreciation but over a period of six years.At the end of the financial year N+3, the fixed asset is sold at a price of 125,000 lei.Differences of depreciation, resulting after the subsidiary company applies different accounting policies for similar transactions and events under similar circumstances, corresponding to the three financial years, are as follows: Company B will adjust the annual financial statements of subsidiary company A when preparing the annual consolidated financial statements, as shown below:
Adjustment of the financial statements of the subsidiary company: a) In the financial year N+1
In this financial year, the depreciation of 15,000 lei recorded by the subsidiary shall be increased with the sum of 10,000 lei.After the increase of the Depreciation, the net value of the tangible fixed asset in the consolidated financial statements is of 125,000 lei, whereas in the individual financial statements of the subsidiary it has a value of 135,000 lei.
The accounting recording corresponding to the adjustment is as follows: 6811 "Operating expenses related to the depreciation of fixed assets" = 281 "Depreciation of tangible fixed assets" 10,000 In the financial years N+2 and N+3, besides supplementing the expenses with the depreciation corresponding to the current financial year, the differences of depreciation corresponding to the previous financial years will also be adjusted, as follows: As a result of the fixed asset's sale by the subsidiary, at the end of the financial year N+2, its financial result shall be restated as follows: a) The financial result corresponding to the fixed asset's sale operation, as it is recorded in the books and financial statements of the subsidiary is shown below: Revenues from sale of fixed assets (7583) 125,000 Accounting value of fixed assets (212) 150,000 Accumulated depreciation at the time of sale 45,000 Net book value (unamortized value 6583) 105,000 Result obtained after sale of fixed asset 20,000 b) The financial result, in line with the accounting depreciation policy of the company that prepares consolidated financial statements (parent company B), obtained from sale of the fixed asset by the subsidiary: Revenues from sale of fixed assets (7583) 125,000 Accounting value of fixed assets (212) 150,000 Accumulated depreciation at the time of sale 75,000 Net book value (unamortized value 6583) 75,000 Result obtained after sale of fixed asset 50,000 The financial result reported in the subsidiary's financial statements (which are the object of the consolidation) shall have to be increased with the difference of 30.000 lei by deducing the expenses with the disposed assets, because the subsidiary recorded an unamortized value corresponding to the sold tangible asset of 105,000 lei as compared to 75,000 lei, which should have been the value recorded in agreement with the depreciation policy applied by the parent company.Accounting entries of the financial result adjustment in the financial statements of the subsidiary are as follows: C. Subsidiary company "A" recorded as intangible fixed assets, at the beginning of the financial year N, set-up costs/expenses of 60,000 lei.These expenses have to be amortized in a straight-line rate, over three years.However, in the consolidated accounts, the set-up costs are recorded as expenses of the financial year when they were spent.The entity that performs the consolidation shall have to adjust the financial statements of the subsidiary A, as follows: a) In the financial year N: Because in the case of the company that prepares the annual consolidated financial statements the set-up costs/expenses are not capitalized, the consequences of their capitalization by the subsidiary will be eliminated: * Set-up expenses shall be recorded here, in agreement with their economic nature (charges, taxes corresponding to the services of registering the company at the Registry of Commerce, advertising etc.) b) In the financial year N+1, according to the accounting rationale, adjustments will be made both for the differences corresponding to the current year and for the amortization difference corresponding to the previous year.
Conclusions
As it can be concluded from the contents of the approached theme, accounting standards specific to the consolidation of financial statements by group companies represent for specialists, both theoreticians and practitioners in the field of accounting, an issue undergoing full dynamic and which requires permanent debate with a view to finding the most adequate implementation solutions.The adjustment of the individual financial statements of companies included in the consolidation scope (subsidiaries, associated entities or entities controlled in common) and which "employ different accounting policies for similar transactions and events under similar circumstances" represents a compulsory requirement that needs to be put into practice by accountants.In this context, regulating or normative institutions from the field of accounting, professional associations and the academia should offer models and solutions for solving practical situations that require conformity to the group's accounting policies.and liquidation of companies as well as of the withdrawal or exclusion of certain associates from companies, with subsequent amendments and additions.
The same restatement entries shall be recorded in the subsequent financial years(2017 and 2018), until the vehicle shall be entirely depreciated in agreement with the accounting policies of the parent company:-in 2017 6811 "Operating expenses related to the depreciation of fixed assets" the financial year N+2 -(the last year of amortization) the accounting rationale is similar to the one corresponding to financial year N+1, as shown below:2801"Amortization of set-up costs"
2. In the financial year 2014:
In the financial years 2014 and 2015 the accounting rationale is the same, hence:
|
2019-05-20T13:03:45.776Z
|
2017-12-30T00:00:00.000
|
{
"year": 2017,
"sha1": "03f68cab59ca2a7bf3f1a65a0b36e43373875518",
"oa_license": "CCBY",
"oa_url": "http://sceco.ub.ro/index.php/SCECO/article/download/391/367",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "03f68cab59ca2a7bf3f1a65a0b36e43373875518",
"s2fieldsofstudy": [
"Business",
"Law"
],
"extfieldsofstudy": [
"Economics"
]
}
|
221821595
|
pes2o/s2orc
|
v3-fos-license
|
Post 9/11 American Footprints in Pakistani Media: A Critique of Semiotic Discourses of Pakistani Newspapers
ARTICLE DETAILS ABSTRACT History: Accepted 23 March 2020 Available Online 31 March 2020 This paper highlights the power of image in shaping perception of the people regarding post 9/11 American representation in Pakistani print media discourses. The study deconstructs the semiotic discourse(s) of Pakistani English newspaper Dawn (daily) from September 2018 to February 2019 to argue that linguistic and semiotic devices and techniques work discursively to shape the readers’ perception regarding American foot-prints in Pakistani print media. It employs Multimodal Critical Discourse analysis approach by drawing upon Machin (2007), Van Leeuwen framework for recontextualization (2008) and Fairclough’s (2003) for visual and linguistic analyses to lay bare embedded ideologies propagated through word-picture conjunction. The levels of analysis include participants, settings, poses, objects, metaphor, inclusion, exclusion and discourse. Moreover, the researchers have validated the findings of their semiotic analysis by conducting two focus group discussions among the students of linguistics and other disciplines. The findings reveal that print media semiotic discourses provide an appropriate use of language in graphic form. The findings reveal that no use of language is ideology free and words and pictures work in conjunction to propagate desired ideology to the target readership. Additionally, the study notices the visible change that has taken place regarding American representation from superordinate to back foot and ready-to-hold dialogue through semiotic discourses of mentioned newspaper.
Introduction
Semiotics, having non verbal power, acts as meta-language (Jacobson, 1975). People usually regard their sense of sight more reliable than their sense of hearing. Hence, a signifier, in an iconic mode, provides a deeper understanding than the words can. Eco (1976) asserts that a picture speaks more than thousand words. Through semiotics, social and psychological human nature can be understood. A sign or gesture conveys meanings more emphatically. Sign is a trace, clue or mark which stands for both universality and individuality, says Barthes (1974). Saussure (1916) considers semiotics a science that studies life of signs in the society. Pierce (1931) concurs and states that semiotics is a system of principles to study behavior based on signs. Krestiva (1969) opines that every speech act conveys a message. It may be done through language of words, posture, gesture, clothing or any other means in a social context. Pierce and Saussure both presented models to explain semiotics. Although their models are different from each other in some respects, yet both lay stress on the semiotics as a more powerful means for imparting a specific ideology in the most effective manner. In Sussreran model, a sign is made up of signifier and signified whereas Pierce claims that an object produces 3 signs, i.e. icon, index and symbols. He called signifier as representamen, signified as interpretant whereas the third component of his model is object.
Both these models are based on semiotics which is a means of reflecting social reality that is culturally oriented and reflected through using certain devices in order to impart specific ideology and to shape individual's worldview. Moreover, semiology has rocketed in the present era because every individual can extract meaning from semiotic discourses according to his/her mindset. One thing significant about semiotics is that it is culturally oriented and carries a plethora of interpretations. Barthes (1974) argues that our senses are culturally trained to extract meanings from semiotics. Additionally, it is also an observed phenomenon that human beings not only convey verbally but also employ different techniques of communication other than verbal language. Still another significant feature of semiotic discourses is that with the help of word picture conjunction, and exploiting the technique of denaturalization of language, many explicit and implicit ideologies are propagated to the target audience.
Semiotics: An influential genre of Pakistani news papers
Newspapers are considered one of the best sites for ideological investment. Print media is considered one of the best ways to influence and shape the mind of masses. Hence, it plays a significant role in shaping the mindset of the target readership accordingly. Still imagery means a lot where the propagation of implicit ideologies is concerned. Political cartoons are regarded as communicative weapons. Boorstin (1963) asserts that the meaning does not come out obviously; it often exists lengthily and waits out to be disclosed. He argues that political cartoons have become more interesting than original and in fact have become the original. In today's world order, visuals are dominant images (Kress 1966). Political cartoons are a benighted ideology of realism. The way a cartoon is photographed, and where it is photographed, is a significant feature to be kept in mind while decoding semiotic discourse.
Print media discourses provide an insight into social, political and cultural reality of a society. They examine the role of language and visual language. They are an instrument of hegemony of political authority. Moreover, political cartoons have a two-fold appeal, i.e. reproduction of ideologies and voice of mainstream media. Semiotic discourse is a fast emerging domain of media discourses. It is one of the most influential ways of representation. Even the most ideologically packed incident can easily be unpacked through semiotic discourses, which in turn are decoded by different mindsets differently. The process of unpacking the political cartoons needs social, political and cultural knowledge because they are culturally oriented. Print media discourses serve as an active tool to propagate the ideology of the controlling group to the target audience. Like many forms of discourses, print media discourses (visual and linguistic) implicitly/explicitly represent the world view of the controlling ideological group. It is observed that same incident most often finds different representation(s) in accordance with the desires of controlling ideological group. Different representation of the same event is very common phenomenon in Pakistani print media. Both linguistic and visual techniques are employed to propagate desired ideology and worldview to the people to win general consent and to make them behave in the manner as desired by controlling ideological group. Therefore, print media linguistic & semiotic discourses carry power within & behind them. Zubair and Sajid (2012) assert that in Pakistani newspapers, visual structures have strong ideological and semantic dimensions with deep influence over the masses instead of being more formal pictures. Semiotics is context based and embodies various cultural connotations besides representing language in denaturalized way. Research Question is: How is America represented discursively through semiotic discourses of Daily Dawn?
Methodology Employed
Post 9/11 semiological representation of America has been through many phases. Most often, the concept of super-ordination and sub-ordination has been plaid up. In the recent past, the ideology of doing more has been asserted on Pakistan. This is how print media semiotic discourses have been representing Pak-US relationship and American foot prints in Pakistani print media in a dominating way. The present research critically decodes the semiotics carrying Pak-US relationship collecting data from Pakistani English Daily 'Dawn' to highlight whether any change has occurred in Pak-US relationship through semiotic discourses or not.
The data for the present research has been collected from Daily Dawn. Out of 25 political cartoons regarding American representation in Pakistani mentioned newspaper, one has been purposively selected to decode at linguistic and semiotic levels. In order to analyse the selected image, the researchers have employed Fairclough (2003) model of CDA, Kress and Van Leeuwen's (2006) for visual analysis have been applied to analyse the data at linguistic and semiotic levels. The researchers find this model apt for the analysis as it investigates social practices, events, ideologies and texts being shaped and reshaped by power relations within a social context. Furthermore, discourse analysis is a form of speech act analysis to find the social meanings using different techniques employed.
Amended Model of Fairclough
Implicature representation assumption back/foregrounding Inclusion / exclusion Person as a State Metaphor Discourse as a social practice Use of Pronoun
Amended Model of Kress and Leeuwen
The rationale behind devising an amended model is that data for the present research comprises Linguistic and visual modes. In order to make semiotic analysis more valid, the researchers have also conducted Focus Group discussions on the selected images, concerning post 9/11 American foot print in Pakistani print media semiotic discourses. This has been done to counter the arguments that are usually leveled against semiotic analysis that most often it matches semioticians' specific way of decoding the images which, in certain cases, may not match with general perception on such texts. Two focus group discussions were held. The first group consisted of the students of M.Phil. linguistics studying at NUML Multan Campus. The other group included the participants from other disciplines like Education, Management Sciences, and Islamic Studies. Both the groups were of mixed gender and their age ranged from 25 to 50 years. Each Focus group discussion lasted for one hour and has been visually recorded. The levels which have been focused to analyze Focus Group discussion data are words, internal consistency, frequency, extensiveness and finding the big idea by employing Kruger's (2000) model for focus group data analysis. Gaze at the viewer 2.
Absence of the gaze at the viewer 3.
Linguistic and Semiotic Analysis
Wodak (2001) states that discourses are historically rooted and most often they cannot be analyzed comprehensively without having the knowledge of context. Similarly, the semiotic discourse under analysis has different historical associations. In the past the Taliban's were under the thumb of America and hence, were provided and exploited as desired by the controlling master of that time (America). Philipson (1997) argues that power exists at various modalities and discursive modes of power are more influential as compared to coercive modes. Talibanized ideology slowly and gradually continued prevailing among them and among other groups in Afghanistan. Therefore, they have not only gained strength in size as is obvious from semiotic under analysis but in number as well because at present more than 54 percent area of Afghanistan is under the control of Talibans. This is how discursivity worked to win general consent in terms of number and power. Since the conquest of Russia these trained fighters were not handled judicially, In this way they continued gaining strength in terms of number, and material goods, they continued propagating their ideology among the masses and hence at present have gained so much discursive and coercive powers as has been illustrated in the semiotic under analysis to challenge America and Afghan Govt. Fairclough (2003) opines that media discourses reflect social reality. Similarly, the image under analysis illustrates that at present Talibans being so powerful are in a position to decide whether to hold peace talk with the other allies (Afghan Govt. & America) or not. The semiotic under analysis illustrates that the way Talibans' representative is heading forward to the other allies (Afghan Govt. & America) is very threatening. He is strongly built and well armed. He is heading forward towards the allies waiting for him to come and hold table talk with them as the language in graphic form (political cartoon) represents.
Moreover, another important thing to note is that both the allies as have been represented in the image are smaller in size as compared to the Taliban's representative marching towards them frowningly. This implies that the Talibans' representative is least interested in holding table talk rather in confrontation. As far as visual discourses are concerned different techniques are used in order to propagate message comprehensively. The semiotic under analysis carries the linguistic text written in capital letters bold font. This typographic technique is used when something very important is meant to be highlighted. The linguistic text written in capital letters and within the box in the form of speech bubbles connotes the only and ever-growing worry in the minds of American and Afghan presidents is ever increasing size of the Talibans by all means. The cartoonist of the controlling ideological group wants to convey a message in accordance with the Western perception and representation of stereotypical Muslim who is bearded and caped. One thing is also important to note that danger or terrorism has been associated with he, manhood. It is also a western perception especially about the bearded Muslims that they are no less than a threat terror to the west. If we look carefully at the end of the linguistic message the cartoon carries 'He is growing bigger and bigger….' there are four dots which stylistically are named as 'ellipses'. The use of ellipses (four dots) implies that the rest of the things have intentionally been left unsaid, presupposing that all the things are clear and well known to the participants. Eco (2000) opines that word picture conjunction is a significant technique to enhance the effect of the message being propagated. Similarly, the ongoing action in the picture has been linguistically supported by present continuous "He IS GROWING….". Visually, this has been corelated by the representative of Talibans on his toes. Barthes (1974) argues that semiotic discourses are culturally oriented and carry plethora of information in them. This also holds valid because if the semiotic under analysis is decoded from western perspective it implies that Muslim militants are growing with every passing day and now, they have crossed the limits. All the efforts made by the West (America) are proving futile. The militant Talibans are least interested in holding talks, rather these attempts of restoring peace are unpleasant to them as is obvious from the face of the Muslim militant shown in the picture. This is how the West supports its stance that the East (Muslims) are the embodiment of evil and danger to the civilized west.
On one hand there is a concept of growing. The use of word growing implies one thing more important is that growing is used for living beings. Living beings have the capability of growing. It is pertinent to mention here that the Talibanized ideology has been growing since its inception. Ideology is regarded as a living being as it can drive the living beings. The astonished faces of American and Afghan presidents illustrate that the Talibans have gained strength beyond their expectations. Hence, both are terribly afraid. The nearer the Talibans' representative is coming, the more afraid they are getting. Another important point to notice as highlighted by the semiotic under analysis is that the Afghan president seems more frightened as compared to the American president. His foot prints are invisible the way he is sitting on the chair.
Still another, feature of the use of bigger is the use of comparative degree which when correlated with the size of the person is bigger as compared to the other two persons as mentioned in the image. One of them is the representative of America and the other is the representative of Afghan presupposedly. If the sizes of the two persons are combined, they cannot be compared even with the single size of the bearded person who is alleged as a terrorist. This implies that despite the combined efforts of American and Afghan governments are unable to crush the Talibanized ideology. Still another visual technique that has been employed here is that the person is approaching towards the chair vacant for him from the back side of the chair. Moreover, empty chair at the left side of the table also implies that the other party that is involved in the talk is not responding to the arguments or discussion. Moreover, one important thing is the size of the chair, the person is approaching to sit on seems too small as compared to the size of the person the chair is vacant for. It also implies that the Taliban representative is from the way he is being entertained in the upcoming session of dialogue among the three parties. It seems that he is not comfortable with the terms and conditions on which the other parties are ready to hold discussion. One the implications might be as connoted by the small size of the empty chair that the demands of the Taliban representative are as huge as his size is and he will be uncomfortable even if he sits at the dialogue table because of the too small size of the chair. This situation implies that although the three parties have been holding talks for sometimes yet not in accordance with the demands of the Talibans. Therefore, the representative of Taliban is angry and perhaps he has made up his mind that things are not going to be settled through talks rather with gun as is illustrated by the style with which he is approaching towards them.
Still another important feature to be noticed is the dress code, the person having the gun in his hand is wearing a shalwar-qameez, turban with bearded get up which is most of the time is associated with Muslim clergy man. According to western stereotypical perception about rigid and hardliner Muslims, they are extremist and are least inclined in embracing any type of change in the form of dialogues initiated by the peace-loving west. Moreover, the west perceives that such type of Muslims are so rigid and close minded as is highlighted by tightly sealed turban. The west perceives that the hardliner Muslims get the heat of extremism from the fundamentals of Islam and finally explode them in the form of suicide attackers. They are so much under the influence of violent ideology that they are least inclined to listen to others. They picture under analysis illustrates that the earholes of the Muslim / Taliban representative are almost closed, therefore, he is not ready to listen to others because he is totally being driven by his violent ideology. Rather he is fully inclined in doing whatever he has been fed-up with. The concept of persons as state metaphors has also been highlighted here. The three persons mentioned in the pollical cartoon respectively stand for their countries or school of thought. The American president as shown in the picture in sitting with flat footing which implies that up to a great extent, he is on backfoot. In the past it was observed that America has continuously been following the ideology of "Doing More" from the Taliban and Pakistani government, but the present image illustrates the change in political scenario that has taken place has forced America to be on back foot. Because of so many social political, geographical and economic factors, America has realized that holding talk may prove more beneficial than deciding things with coercive force.
Still another significant feature of the image under analysis is the boggled face of Afghan President which implies that he is absolutely worried about his future because after the withdrawal of American forces (if possible) and handing over rule to Taliban, he would have no ground to stand on as is obvious from the picture that his footprints are not visible. Kress (2010) argues that Cartoons are regarded as metaphorical representations of speech bubble; the two small speech bubbles on the hat of American President imply the thoughts in his mind that how to deal Afghan government and Taliban in such a troubling situation. Whereas, the three small speech bubbles on the head of Afghan President imply that he is more worried about his survival. If the things are settled between Taliban and American what would be his or present government's status. Still another important thing that has been shown in this image is that American president is wearing a flag of his country on his hat which implies that America has an identity of her own ; on the other hand Afghani president is bare headed which means he does not have any peculiar identity. At the same time his survival is at the risk. One of the implicatures may be that because of the destruction caused by terrorism, Afghanistan and the Afghan governments have totally lost their control and they are totally under the thumb of America and perhaps this might be the cause of their losing identity. Hence the overall image that appears after going through the linguistic and semiotic discourse is the political scenario regarding the America, Afghan government and the Talibans is very much vague till now. On one hand America with her allies is doing her best to restore peace and to abolish terrorism yet on the other hand despite doing their best the terrorism or the Muslim fanaticism is growing bigger and bigger as has been shown at linguistic and semiotic level. So the underlying message that the image suggests is that America is the representative of the civilized world and peace whereas, Muslim militants, especially the Talibans have been represented as anti to all the peace restoring forces. This is a message to the rest of the world America-Talibans dialogue are not going to get the desired results very soon because both are sticking to their deep-rooted ideologies. In this way Fairclough's (2003) claim seems justified that print media discourses are considered as one of the best sites for the ideological investment. So, the ideology that has been invested through the linguistic and semiotic discourses is that despite doing best America and the other allies, so far have not been able to come to restore the peace and so far have not been able to come to good terms to other parties because peace cannot be had with the gun rather it can be had with the dialogues. So, here we see two sides of the image, American and Afghani presidents are waiting for the peace talk whereas the Muslim militant / Taliban have been shown least inclined in holding dialogue with them rather they seem more willing to confront, it seems perhaps they have decided that things are not going to be settled with carrot rather with stick as is connoted by the get-up of the Taliban representative because he is marching towards America and Afghan presidents with rolled selves and finger at the trigger of the gun.
Focus Group Findings
As mentioned earlier in the methodology section the researchers have validated the findings of semiotic analysis through the remarks of focus group participants' remarks. The remarks of some of the participants have been included to support the stance. One of the focus group Participants while expressing her ideas on the semiotic under analysis said: "Post 9/11 media representation of America implies that perhaps America has failed in reaping her interests in Afghanistan. The picture represents a changing scenario showing, America unable to handle Taliban." Another participant remarks as: "The picture is meta-narrative. The chairs on which American & Afghan presidents are sitting are unarmed whereas the man coming to hold dialogue the two presidents is fully armed. Here the concept of unarmed chair and armed man has been juxtaposed. Interestingly, the unarmed chair is Vacant for the armed man".
One of female participants remarked as under: "The political cartoon embodies different counter narratives. The three men in the picture are representatives of three different ideological groups." Fairclough (2003) opines that language connotes and denotes. The counter narratives have been portrayed. According to the participants that apparently , the representative of the Talibans has been shown unwilling to hold dialogue but inwardly the American president is also unwilling to hold talk. This is a mere a show of. The Talibans are the product of America. This is how the Linguistic & Semiotic discourses are connoting and denoting things simultaneously. Hence, visuals carry bundle of interpretations.
Another participant remarked; "The way guy (representative of Taliban) is approaching towards them is astonishing to both American and Afghan president because perhaps his style is totally unexpected to them and hence, they are astonished." The overall picture that emerges after analyzing the remarks of the participants is that the previously trained or in Asian context the Badmash or the Terrorists/Talibans, who were beneficial for America, now have gone beyond her control. It means that Taliban and American are in direct contact or in direct clash with each other. Afghani President is just like a Harbinger between them. Irony of fate is that America is going to settle the issue with NATO forces, still trying to control the fate of the country. America is showing to the world that she is most concerned about the terrorist situation in Afghanistan. The participants concluded their remarks that America has been trapped in her own evil design. Now the talibanised ideology will continue increasing and there will be increase in their number with every passing day as is obvious from the picture. In this way the researchers validated their findings of semiotic analysis through focus group discussions.
Conclusion
The present research concludes that an obvious change is there regarding American representation through Pakistani Print Media semiotic discourses i-e a shift from super-ordinate position to back foot (ready to hold talk) . In the past, it was observed that America was represented as dominant and Pakistan and Talibans as dominated, through print media discourses but the findings of the present research reveal that power paradigm shift is obvious. Now print media representations of America through semiotic discourses is different in a sense that America has been represented as ready to hold talk instead of imposing her authority on Pakistan/Afghanistan. Still another significant finding of the research is that semiotic discourses are significant genre of print media and word-picture conjunction works a lot in shaping the mindset of the target audience. The way she (America) has been represented through the semiotic discourses of Pakistani print media is that America seems moving at back foot, ready for negotiation. The percentage of such data showing America on back foot and instead of asserting her super-ordination is ready to hold talk to resolve political issues. Hence, the present research concludes that print media semiotic discourses on one hand shape the mindset of its readers and on the other hand represent political changes taking place in the global scenario. The impact of political cartoons on the people in the back drop of Pak-Us relation is of vital significance.
|
2020-08-06T09:05:05.199Z
|
2020-03-31T00:00:00.000
|
{
"year": 2020,
"sha1": "b5baa7011c296e7221e1e021b548fa5b72e3ec5b",
"oa_license": "CCBYNC",
"oa_url": "http://reads.spcrd.org/index.php/reads/article/download/190/182",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "43b262e88cdba1c323142f124a04ff1cfe959a75",
"s2fieldsofstudy": [
"Linguistics"
],
"extfieldsofstudy": [
"Sociology"
]
}
|
237145193
|
pes2o/s2orc
|
v3-fos-license
|
Detection of Cryptosporidium spp. and Giardia duodenalis in small wild mammals in northeastern Brazil
This study investigated the occurrence of Giardia duodenalis and Cryptosporidium spp. in rodents and marsupials from the Atlantic Forest in southern Bahia, northeastern Brazil. Two hundred and four fecal samples were collected from different forest areas in the municipalities of Ilhéus, Una, Belmonte, and Mascote. Identifications were performed using PCR and nested PCR followed by sequencing of the gdh and tpi genes for G. duodenalis, and the gp60 and Hsp-70 genes for Cryptosporidium. The total frequency of positive PCR samples for both G. duodenalis and Cryptosporidium spp. was 5.4% (11/204). Giardia duodenalis occurred in 2.94% (4/136) of rodents and 2.94% (2/68) of marsupials. The prevalence of Cryptosporidium in rodents and marsupials was 1.47% (2/136) and 4.41% (3/68), respectively. In the areas sampled, the frequency of parasitism was 50% (7/14), while the Mascote region alone had no parasitized animals. The G. duodenalis subgenotype AI was identified in the rodent species Hylaeamys laticeps, Oecomys catherinae, Oligoryzomys nigripes and Akodon cursor, and in the marsupials Gracilinanus agilis and Monodelphis americana. In the rodents Rhipidomys mastacalis, H. laticeps and in the marsupial Marmosa murina the protozoa Cryptosporidium fayeri, Cryptosporidium parvum and Cryptosporidium ubiquitum with subtypes IIa and IVg by the gp60 gene were found. In conclusion, this study provides the genetic characterization of Giardia and Cryptosporidium species and genotypes in rodents and marsupials. And, these findings reinforce that the rodent and marsupial species mentioned above play a role as new hosts for Giardia and Cryptosporidium.
Introduction Small mammals such as rodents (Rodentia, Cricetidae) and marsupials (Mammalia, Didelphimorphia) transmit pathogens to humans and domestic animals; however, the consequent risk to public health is poorly understood [1,2]. Environmental disruption due to human activity influences the occurrence and spread of zoonotic and parasitic diseases (e.g., giardiasis and cryptosporidiosis) in these animals, affecting the wildlife species balance [3].
Giardia Kunstler, 1882 and Cryptosporidium Tizzer, 1907 are protozoa known worldwide for causing severe gastroenteric disease in humans, as well as domestic and wild animals [2,4,5]. These protozoa cause infections from cysts or oocysts found in environmental and water contaminations [4,6].
The role of wild animals in human giardiasis and cryptosporidiosis epidemiology is uncertain. However, molecular studies have allowed the identification of several species of Giardia and Cryptosporidium in wild animals [6,[7][8][9].
Molecular techniques have successfully determined and supported the understanding of epidemiological processes [9] by using several genes to identify distinct species of Giardia and Cryptosporidium. Additionally, they reveal genotypes and subgenotypes, of which some are specific to humans and others to animals [6].
Molecular studies to detect Giardia and Cryptosporidium in wildlife reported the presence of these protozoa in different species of small mammals. However, in northeastern Brazil, no studies have employed molecular genotyping to identify G. duodenalis and Cryptosporidium spp. Thus, the objective of this study was to identify, through a molecular technique at the level of genotypes and subgenotypes, G. duodenalis and Cryptosporidium spp. in fecal samples of rodents and marsupials captured in agroforestry areas (Cabruca) and the Atlantic Forest in southern Bahia, northeastern Brazil.
Collection area
Within the study area, 14 forest areas, distributed in four municipalities in the southern region of the State of Bahia, were sampled. These included three cocoa agroforestry areas located in the rural area of Ilhéus (areas 1-3), and 11 forest areas located in the municipalities of Una, Mascote and Belmonte (areas 4-14) (Fig 1). The study region is characterized by a hot and humid tropical climate, with an average relative humidity of 89-90% and an average temperature of 24-25˚C, predominantly covered by tropical forest vegetation and an agroforestry system, which preserves native forest [19]. In the region, it rains 150 days a year on average, with precipitation reaching 2,000 mm/year. The dry seasons are not well defined; occasionally, one to three months receive less than 100 mm of rain [20]. Elevation of the sampled areas ranged from 42-100 m above sea level and were georeferenced with a Global Positioning System (GPS).
Capturing animals and obtaining biological material
The capture period ranged from June 2015 to December 2016. The animals were captured using Sherman (23 × 8 × 9 cm), Tomahawk (50 × 17 × 17 cm), and pitfall traps. Each area was divided into three plots, with for a total of 24 traps per plot and 72 traps per area. After identification of the species, fecal samples were collected with subsequent release of the animals at the places of origin (Table 1). Fecal samples were stored in 1.5 mL microtubes, kept refrigerated and delivered to Laboratory of Veterinary Parasitology of the State University of Santa Cruz (LAPVET-UESC), weighed, and standardized between 180 and 200 mg.
DNA extraction and molecular characterization
The fecal samples were washed with sterile PBS (pH 7.2) and subjected to genomic DNA extraction using the QIAamp DNA Stool Mini kit 1 (Qiagen), according to manufacturer's instructions. After adding the lysis buffer, the samples were subjected to five cycles of heating (96˚C) and freezing (-196˚C), with 3 minutes of heating and 5 minutes of freezing, then homogenized in a vortex for 5 minutes with 0.2 g of glass beads (0.5 mm), following the kit's guidelines thereafter. The amount of extracted genomic DNA was established using a Nano-Drop 2000 (Thermo Scientific, USA), stored in boxes, and placed in a freezer at -20˚C.
To detect the presence of G. duodenalis and Cryptosporidium spp., each isolated DNA sample was subjected to nested PCR. For the amplification of Giardia fragments, gdh [16] and tpi coding genes [17] were used. Cryptosporidium fragments were amplified using gp60 [12] and Hsp-70 [11] genes ( Table 1).
The tests were carried out in a Proflex PCR system thermocycler (Applied Biosystems) using the Platinum Taq DNA polymerase kit (Invitrogen) for the mix. Positive fecal samples from Giardia cysts and isolates from the Veterinary Parasitology Laboratory at UESC were used as positive controls. Cryptosporidium (isolates 13F and 13C) from the Laboratory of Clinical Analysis (LAC) of the State University of Feira de Santana, Bahia [21] and ultrapure water were used as negative controls. The PCR products were subjected to 1% agarose gel electrophoresis, developed with SYBR 1 Safe, purified using the PureLink PCR Purification kit (Invitrogen), and sent for sequencing.
Statistical analysis
To verify the association between the positivity of the samples with the catch area (agroforestry and forest areas), statistical analysis was performed using Fisher's exact test with 95% confidence intervals using the Epi Info ™ 7.2.0.1 software.
The analysis of the tpi and gdh gene sequences demonstrated 100% genetic similarity with the G. duodenalis species of the subgenotype AI ( Table 2). The genetic analysis of Cryptosporidium identified C. parvum, C. ubiquitum, and C. fayeri, and subtypes that belong to the IIa and IVg allelic families. No subtype found for C. ubiquitum (Table 3).
Discussion
The present study investigated, for the first time, the presence of the protozoa Giardia and Cryptosporidium in rodents and marsupials captured in the northeast region of Brazil. The southern region of Bahia includes an extensive area of the Atlantic Forest with a richness of fauna and flora species, being an important area for the conservation of global biodiversity [20]. In addition to having areas of cocoa agroforestry, providing shade for planting and preserving native forests [22].
PLOS ONE
Detection of Cryptosporidium spp. and Giardia duodenalis in small wild mammals in northeastern Brazil Giardia duodenalis infection has been described in wild animals, such as rodents and marsupials, with a prevalence ranging from 2% to 12% [3,[23][24][25][26][27]. This defines a low prevalence in forest areas, compared to that in urban areas with rodents having a higher prevalence ranging from 24.4% to 64.3% [2,23,28]. In the present study, the frequency of positive animals was 5.4%, and such low positivity may be related to the sampling site, which has rich and abundant flora, low anthropization, and the presence of some arboreal animal species, such as G. The gdh and tpi genes demonstrated good sensitivity, allowing the generated sequences to identify the G. duodenalis species and the subgenotype AI in the six isolates. Because it has conserved regions, characterization of these genes can identify all genotypes and subgenotypes of G. duodenalis [36][37][38].
The Cryptosporidium frequency was 1.47% and 4.41% in rodents and marsupials, respectively, similar to that described by Santos [24]. The literature describes this protozoan infecting a variety of small mammal species [3,24,[39][40][41][42][43][44]. Studies in urban areas also show a greater degree of parasitism of this protozoan in synanthropic rodents [2,28,41,42]. The presence of this protozoan may be associated with anthropic action and the presence of domestic animals provides an interaction between humans and wild fauna, favoring its dissemination [45].
Cryptosporidium ubiquitum was found in Hylaeamys laticeps, the first finding in wild rodents captured in Brazil. This species has low specificity and is commonly reported in animals, including rodents, marsupials, and other host species [35,41,43,53,54]. Cases in humans
PLOS ONE
Detection of Cryptosporidium spp. and Giardia duodenalis in small wild mammals in northeastern Brazil have shown that [55,56] the most common route of C. ubiquitum transmission is through water [56]. The two genes assessed, gp60 and Hsp-70, have satisfactory sensitivity and can be used in studies to identify Cryptosporidium and verify its genetic diversity [45,53,57,58]. Using more than one gene provides a more detailed understanding of the protozoan's genetic variability and abiotic factors in the study population [59].
In this study, the occurrence of protozoa in small mammals was similar in the Atlantic Forest (Una and Belmonte) and agroforestry (Ilhéus) environments. The difference in the number of positive animals between capture areas was not statistically significant, demonstrating that agroforestry areas maintain low contamination due to the continued diversity of fauna and flora, despite greater anthropic action and transit of domestic animals that threaten the diversity of wild animals [60].
The close human relationship with wildlife as a result of disorderly urban occupation, illegal trade in wild animals, or the maintenance of these animals as pets, are some of the factors that enhance the transmission of zoonotic diseases between species, thus threatening both conservation of biodiversity, and public health [61,62]. Thus, surveillance and monitoring of wildlife pathogens is necessary for the detection, mitigation and prevention of diseases with zoonotic potential.
Conclusion
Results herein obtained pioneer Giardia and Cryptosporidium identification in rodents and marsupials from southern Bahia, northeastern Brazil, showing the present technique as sensitive enough to identify the subgenotypes of Giardia and Cryptosporidium through the gdh and tpi, and Hsp-70 and gp60 genes, respectively.
|
2021-08-18T05:38:37.419Z
|
2021-08-16T00:00:00.000
|
{
"year": 2021,
"sha1": "fe22ee36de4136c657a92df044ec673233b3d027",
"oa_license": "CCBY",
"oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0256199&type=printable",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "fe22ee36de4136c657a92df044ec673233b3d027",
"s2fieldsofstudy": [
"Environmental Science",
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
265240214
|
pes2o/s2orc
|
v3-fos-license
|
The impacts of rent burden and eviction on mortality in the United States, 2000–2019
Investments in stable, affordable housing may be an important tool for improving population health, especially in the context of rising costs and evictions for American renters. Still, a lack of longitudinal data linking these exposures to health outcomes has limited prior research. In this study, we use linked administrative data to estimate the associations of rent burden and eviction with all-cause mortality. We constructed a novel dataset linking renters in the long-form 2000 Census (N = 6,587,000) to mortality follow-up through 2019 from the Census Numident file. To measure exposure to eviction, we further linked this dataset to 38 million eviction records between 2000 and 2016 using names and addresses. For a subsample of renters, we also measured within-individual changes in rent burden between 2000 and 2008–2012 by linking to the American Community Survey. We estimated the associations of rent burden and eviction with mortality using Cox proportional-hazards models and discrete-time hazard models adjusted for individual, household, neighborhood, and state characteristics, examining varying associations by cohort, race, gender, and eviction risk. Higher baseline rent burden, increases in rent burden during midlife, and evictions were all associated with increased mortality. Compared to a baseline rent burden of 30%, a burden of 70% was associated with 12% (95% confidence interval = 11–13%) higher mortality. A 20-point increase in rent burden between 2000 and 2008–2012 was associated with 16% (12–19%) higher mortality through 2019. An eviction filing without judgment was associated with a 19% (15–23%) increase in mortality and an eviction judgment was associated with a 40% (36–43%) increase. Associations were larger for those at lower predicted risk of eviction. These findings reveal how rising costs and evictions are shaping mortality for American renters. Policies designed to increase the supply of affordable housing and prevent eviction may lead to widespread improvements in population health.
The study of housing as a social determinant of health has a long history (Leifheit et al., 2022;Engels, 2009;Du Bois, 1899;Hernández et al., 2019) with much of this research focusing on the quality and safety of housing (e.g., lead paint) and neighborhood effects (e.g., segregation) (Williams and Collins, 2001;Jacobs, 2011).However, there has been less attention paid to the economic and legal aspects of housing and their relationship to health, especially for renters (Sewell, 2016(Sewell, , 2021;;Reynolds, 2021).Despite stagnant wages, rental prices have more than doubled over the last two decades (Myers and Park, 2019).In 2020, half of poor renter households were severely rent burdened, with over 50% of income dedicated to rent-only to see rents increase at the fastest rate ever recorded in the following year (Bhattarai et al., 2022).Each year, 2.7 million households-roughly 7% of all renter households-face the threat of eviction (Gromis et al., 2022).
Homeowners benefit from predictable, fixed-rate mortgage payments, while renters face substantial volatility in housing costs due to the ability of most American landlords to raise rents at their discretion.In some markets, rent hikes in recent years have been dramatic, with average rents increasing by 26% in Gaston County, North Carolina, 28% in Maricopa County, Arizona, and 39% in Collier County, Florida, between 2019 and 2022 (Bhattarai et al., 2022).As rents rise, health-related spending is often crowded out.Poor households with children who are moderately rent-burdened (i.e., 30-50% of income dedicated to rent) spend 57% less on healthcare and 17% less on food compared to unburdened households (Airgood-Obrycki et al., 2022).
Households prioritize rent over other essential spending in part because landlords have the power to evict tenants.Eviction can be a disruptive event that leads to disenrollment from social safety net programs such as Medicaid, job loss, and a host of other negative consequences (Schwartz et al., 2022;Desmond and Kimbro, 2015;Collinson et al., 2022).Eviction can also compromise a person's physical and mental health by exposing them to prolonged periods of intense housing precarity, including homelessness, and acute stress (Desmond and Kimbro, 2015).In addition, eviction can increase exposure to infectious disease, as seen during the COVID-19 pandemic (Benfer et al., 2021;Leifheit et al., 2021;Sandoval-Olascoaga et al., 2021).The consequences of eviction-exposure to residential segregation, concentrated poverty, and labor precarity among them-have been established as "fundamental causes" of population health, influencing the distribution and life course trajectories of chronic disease, adverse health events, and death (Williams and Collins, 2001;Phelan and Link, 2015;Link and Phelan, 1995).
In this study, our objective was to estimate how rent burden (the percentage of a household income that is spent on rent) and eviction events (eviction filing without judgment, eviction judgment) are related to all-cause mortality in the United States from 2000 to 2019, adjusting for an extensive set of individual, household, neighborhood, and state characteristics.We constructed a novel longitudinal dataset linking individual renters from the long-form 2000 Census to 38 million eviction court records from 2000 to 2016, the American Community Survey (ACS) from 2008 to 2012, and administrative death records from the Census Numident file from 2000 to 2019.We analyzed the associations of rent burden and eviction events with mortality risk using Cox proportional-hazards models and discrete-time hazard models, examining how these associations varied by age, race, ethnicity, gender, and estimated eviction risk.Overall, this study provides evidence that housing cost burden and evictions are an important dimension for understanding mortality among American renters.
The political economy of housing as a fundamental cause of health
We consider the political economy of housing, particularly the economic and legal relations associated with renting (e.g., rental costs and evictions), within the fundamental cause theory of population health.The deployment of flexible resources-money, knowledge, power, prestige, social networks-is central to fundamental cause theory, as the specific mechanisms through which these resources are used in relation to health can be easily substituted over time and space (Link and Phelan, 1995).Those with flexible resources can more easily adapt to new settings of disease burden; for example, more easily accessing new health-enhancing technologies and behaviors (e.g., vaccines, working from home) for an emergent disease risk (e.g., . There has been some emphasis on housing within the fundamental cause framework, such that housing inequality puts certain individuals at higher "risk of risks"-for example, residential segregation patterns exposure to low-quality housing stock, physical hazards, and proximity to healthcare services (Williams et al., 2019).But in frameworks that treat socioeconomic position as a fundamental cause of health outcomes, characteristics such as "renter" or "owner" are largely treated as "upstream" features of the individual, in the sense that individuals derive different flexible resources from these statuses (money, social networks) and translate these to different health-relevant resources (healthcare, nutrition).
There has been less focus on the relational systems that maintain these connections and give these socioeconomic positions meaning, (Reynolds, 2021;Emirbayer, 1997;Muntaner, 2013;Williams, 2019) which is particularly important for considering the influence of renting conditions on health.Reynolds (2021) draws insights from the political economy of health tradition to describe a framework of health power resources.This framework focuses on how broader forms of power consolidation and power relations between specific actors organize both the distribution of socioeconomic and flexible resources via stratification (e.g., who is able to access favorable mortgage terms and high value homeownership vs. who is left to seek housing in racialized rental submarkets) and the translation of those resources to health via the commodification of health-relevant resources (e.g., largely privatized healthcare), material deprivation resulting from discrimination (e.g., segregation of rental costs and eviction risk in rental submarkets), and reduced efficacy resulting from devitalization (e.g., propensity to engage in risky health behaviors).In Fig. S1, we map these concepts onto the power relations underpinning housing stratification and mortality.Power consolidation in the housing market and policy stratifies ownership and translates the status of renting to increased mortality risk through many different mechanisms.In this paper, we focus on two important economic and legal mechanisms: rent burden and eviction.
Research broadly focused on the unequal effects of disruptive events (e.g., loss of employment, divorce) suggests that the consequences of eviction for mortality might vary across dimensions such as race, gender, and the likelihood of receiving an eviction judgment (Aquino et al., 2022).In this study, we test two diverging explanations as to why we might find varying impacts of eviction given the fact that the distribution of renting conditions and eviction is very racially stratified in the United States (Hepburn et al., 2020;Krysan and Crowder, 2017).Black renters live in worse renting conditions on average than white renters, contributing to differential baseline mortality between these populations (Desmond and Wilmers, 2019;Massey and Denton, 1993).On one hand, this implies we might find larger impacts of eviction on mortality for those at high risk of eviction (e.g., Black renters) due to existing resource disparities and cumulative disadvantage that make it more difficult to navigate the event of an eviction.On the other hand, we might expect to find larger impacts for those at relatively low risk of eviction ("negative selection") where an eviction may be a more unexpected and non-normative shock to baseline mortality rates that are relatively lower, such as among white renters (Aquino et al., 2022).
Research objectives
We have two objectives in our analysis of how the political economy of housing shapes mortality risk for American renters.First, we hypothesize the increased rents and eviction events are associated with increased mortality.Second, we hypothesize that the association between eviction events and mortality varies by cohort, race, gender, and eviction risk.Testing these hypotheses requires constructing new linked data, especially because administrative data from eviction court systems contain very limited information about each eviction case, typically only the names and address of defendants.Data linkage is therefore necessary to observe characteristics of those filed against and track outcomes.
Linking decennial census data to administrative death records
To construct our baseline sample, we used the long-form 2000 Decennial Census, which was completed by roughly 16% of the U.S. population.The Census allowed us to measure a set of baseline individual, household, neighborhood, and state characteristics that may confound the relation between our primary exposure measures-level of rent burden in 2000, change in rent burden from 2000 to 2008-2012, eviction filing without judgment, eviction filing with judgment-and our primary outcome, all-cause mortality.
To measure all-cause mortality, we obtained individual death records from the Census Numident file, which contains all interactions related to social security numbers (SSNs) that individuals have had with the Social Security Administration since 1972.This includes information on applications for SSNs, requested changes to SSN information, and date of death.The Numident file is the system of record for death information within the Social Security Administration.It also provides the foundation for the Personal Identification Validation System, which the Census Bureau uses to assign a unique identifier called a Protected Identification Key (PIK) to all individuals based on social security number, name, date of birth, address, and sex as available (Finlay and Genadek, 2021;Wagner and Layne, 2014).We used the Census Numident all-cause mortality data for this research, as opposed to cause-specific data from the Centers for Disease Control and Prevention, because it has been shown to be of high quality and is also a part of the Census Bureau's data linkage infrastructure, which allows us to link death information from this file to other administrative records and survey data.PIKs enable us to link records at the person-level over time and across Census survey and administrative sources-as well as to any other external data source that can be reliably assigned PIKs.We merged dates of death from the Numident file to virtually all renters in the 2000 Census using PIKs, allowing us to measure all-cause mortality from 2000 to 2019.
Exposure to baseline rent burden and rent burden changes in midlife
We measured our first exposure-baseline rent burden-among all renters in the 2000 Census.To measure our second exposure-withinindividual rent burden change-we created a subsample by merging renters from the 2000 Census to the 2008-2012 American Community Survey (ACS) by PIK.The ACS is only distributed to roughly 3.5 million households each year, so we pooled ACS data across these five adjacent years to avoid issues arising from small sample sizes.We used tenure status observed in the ACS to identify individuals still renting roughly a decade after completing the 2000 Census.
To analyze the association between levels of rent burden and mortality, we focused on the 2.1 million renters who were middle-aged (ages 40-65) in 2000.Similarly, we focus on the 93,000 persistent renters who were middle-aged in 2008-2012 for exposure to rent burden changes since 2000 (i.e., rent burden changes experienced from ages 30-55 to .This allowed us to analyze these associations during a period of the life course prior to retirement ages (when there is higher risk of mortality selection) and after young working ages (when income and tenure are more volatile).
Exposure to eviction filing and judgment
To estimate the association between eviction events and mortality, we used a national database of eviction records compiled by the Eviction Lab at Princeton University (Desmond et al., 2018a, b) (Fig. S2, Fig. S3).We submitted eviction records from 2000 to 2016 (58 million records) to the Census Bureau's Personal Identification Validation System, which assigned PIKs to individual defendants using a probabilistic linkage based on first name, last name, and address reported in eviction filings; doing so resulted in 38 million matches based on a 65% match rate (Census Disclosure Review Board Approval Number: CBDRB-FY23-CES004-013; sample sizes are rounded according to Census disclosure policy).We merged matched defendants to the 2000 Census by PIK and subset to those renting and age 22 or older at baseline (6.6 million renters, including 187,000 renters who received at least one eviction filing but never a judgment and 327,000 renters who received an eviction judgment between 2000 and 2016).The Supplementary Information includes more details on the construction of all samples (Fig. S4, Section S1).
Analytic strategy
To assess the association between baseline rent burden in 2000 and mortality from 2000 to 2019, we estimated hazard ratios using Cox proportional-hazards models adjusting for a set of individual, household, neighborhood, and state characteristics measured in 2000.These potential confounding factors included baseline age, race, ethnicity, gender, educational attainment, household income, nativity, number of children, household size, marital status, living in the same place five years ago, veteran status, disability status, being unemployed, number of bedrooms, residential building size, tract-level median household income, tract-level poverty rate, and state of residence.We used similar Cox models to assess the association between within-individual changes in rent burden (2000-2008-2012) and mortality (2008-2012 to 2019).For these models, we also adjusted for income trajectories from 2000 to 2008-2012 to isolate mortality risk associated with rent burden fluctuations due to changes in rent rather than changes in income (e.g., changes in wages or household formation).In sensitivity analyses, we estimated associations based on subsamples where we removed individuals who moved over the period and individuals who experienced large decreases in household income over the period.To assess the associations between time-varying eviction events (filing without judgment, filing with judgment) and mortality between 2000 and 2016, we estimated hazard ratios using discrete-time hazard models with the complementary log-log link adjusted for the same set of confounders above.We also estimated hazard ratios based on models stratified by cohort, race-ethnicity-gender, and eviction propensity.We used eviction propensity, which combined information across all our covariates on the likelihood of experiencing an eviction between 2000 and 2016, to examine possible "negative selection" into eviction.Negative selection refers to the fact that the population targeted for eviction was likely already disadvantaged in many other ways and may have some of the highest background mortality rates (Aquino et al., 2022;Brand and Xie, 2010).As such, the estimated hazard ratios between eviction events and mortality may be relatively lower for tenants at the highest risk of eviction exposure, owing to a combination of ceiling effects and the fact that eviction may be a more unexpected and non-normative shock for the population at lowest risk of exposure (i.e., where background mortality rates are relatively lower).(Aquino et al., 2022).
As a result of internal imputation procedures conducted by the Census Bureau there are no missing covariate data (U.S. Census Bureau, 2003).The proportional-hazard assumptions were tested and verified by the inclusion of an interaction term with time in the model.See the Supplementary Information for more details on the specification of mortality models (Section S2.1-2.3) and limitations (Section S2.4).
Descriptive statistics
Characteristics of renters in the 2000 Census (n = 6,587,000) are presented by eviction status in Table 1.Between 2000 and 2016, we observed 180,000 of these renters receive an eviction filing that did not result in a judgment and 327,000 receive an eviction judgment.We recorded 1,474,400 deaths among renters never filed against, 27,000 deaths among renters filed against without a judgment, and 45,000 deaths among renters evicted.
Many Americans who were persistently renting during midlife (2000 to 2008-2012) experienced decreases in rent burdens (Fig. S5).Still, a substantial proportion of both poor and non-poor renters experienced increases in their rent burden over the same period.This was especially true for poor renters, where 47% experienced an increase in rent burden of 5 percentage points or more by 2008-2012.For many poor renters, these increases were much larger.For example, among poor renters with a rent burden of 30-49% in 2000, 45% experienced an increase of 20 percentage points or more by 2008-2012 (Fig. S5).
Baseline levels of rent burden and mortality risk
Higher rent burdens in 2000 were associated with increased mortality through 2019, with a steep gradient in the adjusted hazard ratio between 20% and 70% rent burden (Fig. 1 reports estimates based on a penalized cubic spline for continuous rent burden; Table S1 and Fig. S6 include estimates based on a linear term).Compared to a rent burden of 30%, a rent burden of 50% was associated with 9% higher mortality from 2000 to 2019 (hazard ratio [HR] = 1.09, 95% confidence interval [CI] = 1.09, 1.10) (p < 0.001), and a rent burden of 70% was associated with 12% higher mortality (HR = 1.12,CI = 1.11, 1.13) (p < 0.001).Analyses stratified by race-ethnicity-gender indicated that the association between high rent burdens and increased mortality was pronounced for men, especially for Hispanic men and non-Hispanic Black men, where a rent burden of 70% compared to 30% was associated with a mortality increase of 16% (HR = 1.16,CI = 1.11, 1.22) (p < 0.001) and 14% (HR = 1.14, CI = 1.11, 1.17) (p < 0.001), respectively (Fig. S7).
Changes in rent burden and mortality risk
For middle-aged renters in the 2008-2012 ACS who had been renting a decade prior in the 2000 Census, a 10-point increase in rent burden over that period was associated with 8% higher mortality (HR = 1.08,CI = 1.05, 1.10) (p < 0.001) compared to no increase in rent burden while a 20-point increase was associated with 16% higher mortality (HR = 1.16,CI = 1.12, 1.19) (p < 0.001) (Fig. 2, Table S1).In sensitivity analyses, we estimated associations based on the subsample of renters that we observed living in the same Census block in both 2000 and 2008-2012 (33,000 renters) to exclude rent increases due to moves.We use Census block to determine non-movers, which means we may be including some individuals who did move within blocks over the period, but we suspect such moves are rare.We find that estimated hazard ratios are similar for this subsample (Fig. S8).
Eviction events and mortality risk
Across all renters, the threat of eviction-a court filing without judgment-was associated with a 19% increase in mortality (HR = 1.19,CI = 1.15, 1.23) (p < 0.001) and an eviction-receiving a judgment-was associated with a 40% increase (HR = 1.40,CI = 1.36, 1.43) (p < 0.001) (Fig. 3, Table S1).In analyses stratified by cohort, we find that despite much higher background mortality rates at older ages, hazard ratios are similar across ages 30-34, 50-54, and 70-74 in 2000.In analyses stratified by race-ethnicity-gender, eviction was associated with higher mortality across all groups and associations tended to be larger for groups where eviction was less common (e.g., white renters).
Analyses stratified by estimated eviction risk indicated negative selection into eviction (Fig. 3, Table S1): the association between eviction and mortality was significantly larger for those estimated to be in the lowest eviction risk quartile (HR = 1.89,CI = 1.66, 2.14) (p < 0.001).Still, we found that an eviction judgment was associated with a 27% increase in mortality even among those estimated to be at highest risk (HR = 1.27,CI = 1.23, 1.31).See Section S3 for additional sensitivity analyses.
Discussion
In a study matching millions of housing court records to decennial census, survey, and administrative data, we found that rent burden and eviction events were significantly associated with higher mortality risk.Notably, we found not only that rent burdens in excess of 30%-the common threshold of "rent-burdened"-are associated with increased mortality, but also that rent burdens below 30% are similarly protective.By using longitudinal data to observe within-individual rent burden increases over a decade, we found that the association between rent burden and mortality was significant even when analyzing only those who rented in the same place over time, suggesting that the association was not driven by selection into higher rents via moves.These rent increases may be considered a plausibly exogenous shock to renters since they were unlikely to depend on the characteristics or behaviors of the renters themselves.Rather, it is likely a large proportion of renters obtain housing in submarkets where rents continue to increase throughout midlife, far outpacing wage growth (Desmond, 2018).This may force renters to prioritize housing costs over other health-related spending such as preventative healthcare (Airgood-Obrycki et al., 2022;Desmond and Kimbro, 2015;Benfer et al., 2021).
A report from the Department of Housing and Urban Development in 2012 investigated households with severe rent burdens and described a "puzzling" relationship between rent burden and household mobility.The report found that households with severe rent burdens were no more likely to move in the following years than those with lower burdens (Eggers and Moumen, 2012).However, this lack of responsive moves may not be so puzzling after all.Due to substantial compression of rental prices (Desmond and Wilmers, 2019;Boeing and Waddell, 2016) and housing policies that have not kept pace with growing need, the supply of affordable housing has declined (National Low Income Housing Coalition, 2021).Consequently, low-income tenants are forced to adapt in place to rent hikes when resources are constrained, which may ultimately increase their mortality risk.
Eviction filings and judgments were associated with 19% and 40% increases in mortality risk, respectively.Most evictions occurred within highly marginalized renter populations where mortality rates were already very high.However, even when we estimated associations within the most disadvantaged renters-those estimated to be at highest risk of experiencing an eviction-we found that eviction was still associated with a 27% increase in mortality.The well-documented negative consequences of eviction, such as homelessness and "downward moves" that involve families relocating to neighborhoods with higher crime and poverty rates, can have an impact on one's health.Future research and data linkage will be needed to document which factors (e.g., homelessness, healthcare access, etc.) may be the most influential mechanisms connecting eviction to increased mortality risk.This study found high rent burdens and evictions are hazardous to health.Therefore, public policies designed to increase the supply of affordable housing-such as rental vouchers and small-dollar mortgages-and to prevent eviction-such as diversion programs and legal aid-may lead to improvements for public health.These policies may also help reduce racial disparities in health and mortality, given that Black and Hispanic families disproportionately rent their homes due to the nation's legacy of redlining and mortgage discrimination (Taylor, 2019;Rothstein, 2017).
Fig. 1 .
Fig. 1.Adjusted hazard ratio for all-cause mortality from 2000-2019 associated with changes in baseline rent burden in 2000.Estimates of the hazard ratio are reported based on a penalized cubic spline for continuous rent burden centered on 30% (the dashed lines at rent burden of 30% and a hazard ratio of 1.0 indicates the reference); Fig. S3 includes estimates of the hazard ratio based on a linear term.Census Disclosure Review Board Approval Number: CBDRB-FY23-CES004-016.Sources: 2000 Census linked to 2021 Numident file.
Fig. 3 .
Fig. 3. Adjusted hazard ratios for all-cause mortality from 2000-2016 associated with eviction events.Eviction events include an eviction filing that did not result in a judgment and an eviction judgment.The dashed line at a hazard ratio of 1.0 indicates the reference group of renters never filed against.Census Disclosure Review Board Approval Number: CBDRB-FY23-CES004-013.Sources: 2000 Census linked to 2021 Numident file and 2000-2016 eviction records.
Table 1 Characteristics of the study population at baseline in 2000, stratified by eviction filing status over the period 2000-2016.
Sample sizes and death counts are rounded according to Census disclosure policy.Census Disclosure Review Board Approval Number: CBDRB-FY23-CES004-016.
N.Graetz et al.
|
2023-11-17T16:11:35.226Z
|
2023-11-01T00:00:00.000
|
{
"year": 2024,
"sha1": "201ec4a62fc0462ad2413e12f219e7e2eab78b79",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.1016/j.socscimed.2023.116398",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "d21ca1f2cb080faed190d0c5964d32c8e7d83a8a",
"s2fieldsofstudy": [
"Economics"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
239997741
|
pes2o/s2orc
|
v3-fos-license
|
Diagnosis of Secondary Sclerosing Cholangitis by Port Site Metastasis
Port site metastasis is an uncommon but challenging pathological entity whereby metastatic cancer is discovered at the operative port site after surgery. Secondary sclerosing cholangitis is a multifocal stricture disease of the biliary system as the result of extra-biliary pathology; rarely, it is due to an infiltrative disorder such as neoplasia. This is the first reported case of secondary sclerosing cholangitis that was diagnosed with metastatic cancer following the discovery of port site metastasis after laparoscopic cholecystectomy.
Introduction
Port site metastasis (PSM) is a well-documented postoperative phenomenon in which cancer is discovered at port incision sites after laparoscopic surgery. Secondary sclerosing cholangitis (SSC) is characterized by multifocal biliary strictures as the result of a known nonbiliary pathology. We present herein a unique case of a young patient that presented with multifocal biliary strictures following laparoscopic cholecystectomy (LCCX). The diagnosis of SSC due to metastatic malignancy was ultimately made only after discovering PSM.
Case Presentation
A 28-year-old Hispanic female was referred back to our institution from an outside facility because of right upper quadrant abdominal pain, jaundice, and fever. She had undergone a robotic-assisted LCCX 11 months prior for symptomatic cholelithiasis. Preoperative bile duct imaging and serum liver chemistries were normal. Gallbladder pathology reported gallstones, chronic cholecystitis, and no evidence of neoplasia.
She was first referred to our institution about six months after LCCX with abdominal pain, jaundice, and pruritus. Initial endoscopic retrograde cholangiopancreatography (ERCP) revealed multifocal extrahepatic and intrahepatic hilar strictures with focal saccular dilation of the common hepatic duct ( Figure 1). Cholangioscopy was required to achieve guidewire access of the left and right hepatic ducts via very tight focal strictures and bilateral 8.5F stents were placed. The strictures appeared extrinsic without evidence of mass lesions; cholangioscopy-directed intraductal biopsies revealed necrosis with atypical epithelial cells. Serologic workup including IgG-4 was within normal limits. Hepatobiliary surgery was consulted. An abdominal CT-angiogram was negative for visceral vascular pathology (i.e. postoperative vascular injury). Given her age and the multi-focal nature of the disease, the patient was started on an empiric trial with prednisone for possible autoimmune sclerosing cholangitis. Repeat ERCP one month later demonstrated near resolution of strictures and no residual waist on fully inflated balloons, so a stent-free trial was planned. However, the patient returned within one week with worsening abdominal pain and recurrent jaundice that required repeat ERCP with stent placements.
She continued to experience recurrent symptoms and presented again 11 months post-LCCX. An abdominal CT scan revealed a new mass in the porta hepatis, omental caking, bilateral adnexal enlargement, and a large pelvic mass (Figure 2 a). In addition, masses were now noted bilaterally on the anterior abdominal wall at the laparoscopic port sites suspicious for PSM (Figure 2 b). Upon repeat review of previous abdominal imaging, subtle nodularity of the port sites on prior CT scans was noted. Axial images from three ( Figure 3) and eight ( Figure 4) months post-LCCX all showed findings consistent with PSM. Serum cancer antigen 19-9 and cancer antigen 125 levels were elevated (8080 U/mL and 202 U/mL, respectively). Diagnostic laparoscopy demonstrated ascites and peritoneal carcinomatosis. Liver wedge, omental biopsies, and ascites cytology revealed adenocarcinoma. Histologic staining could not differentiate between a hepatobiliary or mucinous ovarian neoplasm as the primary tumor site. She was referred to oncology and underwent chemotherapy. Unfortunately, the patient died 15 months post-LCCX.
Discussion
Most cases of SSC are due to autoimmune, toxic, ischemic, or infectious etiologies; rarely, infiltrative disorders such as cancer cause SSC [1]. Biliary strictures related to SSC mimic primary sclerosing cholangitis clinically but the clinical outcomes are worse [2]. Numerous underlying etiologies of SSC have been described, with the most common being immune-mediated IgG4-related disease.
Diffuse multifocal biliary strictures from SSC due to malignancy is a very uncommon phenomenon. Case reports have documented SSC secondary to lymphoma as well as metastatic prostate, ovarian, and gallbladder cancers [3][4][5][6]. Lew et al. documented complete resolution of SSC findings via ERCP after successful treatment of metastatic ovarian cancer [7].
The majority of PSM occurs in patients with cancer either known at the time of surgery, discovered during surgery, or identified postoperatively following pathologic review [8]. Most commonly, PSM after laparoscopy (1-4%) is in patients with metastasis from known colorectal or gynecologic primaries [9]. Among patients with known gynecologic cancers, the overall incidence of PSM is about 2%; PSM risk is higher (6%) in patients with ovarian cancer [10]. There are also reports of PSM after laparoscopic diagnosis of ovarian tumors thought to be of low malignant potential [11]. Gynecologic cancers, more specifically ovarian cancers, have been associated with short intervals from surgery to PSM [8]. However, there is also a report of PSM from ovarian cancer two years after a normal diagnostic laparoscopy was performed for pelvic pain [12].
Following LCCX, PSM is most often related to the delayed spread of gallbladder cancer that is only discovered postoperatively and less commonly due to colon, pancreatic, or ovarian cancer [13]. One case of PSM after LCCX due to cholangiocarcinoma was reported where the patient presented with a single focus of distal bile duct obstruction [14].
Occult ovarian cancer-associated PSM presentation post-LCCX can be obscure. Carlson [15]. Ovarian cancer was diagnosed six months later and treated with a complete response, but the patient presented with PSM 16 months after her ovarian cancer diagnosis. Another report described PSM four months after LCCX [16]. A microscopic focus of cancer was detected in the gallbladder specimen so it was initially believed to be unsuspected gallbladder cancer. However, ovarian cancer was diagnosed during the investigation of the PSM.
While the source of primary cancer was not clearly defined in our case, hepatobiliary or ovarian sources were considered most likely. We are aware of only a few other reported cases of PSM in which the primary malignancy was never discovered [17][18][19]. For example, a 50-year-old female presented with PSM 18 months after an elective LCCX for symptomatic gallstones [17]. Tumor immunohistochemistry studies were consistent with an ovarian primary, but no evidence of ovarian tumor was found during an exploratory laparotomy. Polychronidis et al. reported a 75-year-old male with PSM (well-differentiated extra-hepatic mucinous adenocarcinoma) 11 months after LCCX without a known primary malignancy; magnetic resonance cholangiopancreatography (MRCP) and ERCP were used to search for extrahepatic malignancy, but results were normal [18]. A mucin-secreting papillary adenocarcinoma PSM was diagnosed in a 45-yearold female 28 months after LCCX [19]. Pathology of the gallbladder specimen did not reveal any malignancy and the patient's operative report did not mention any omental nodularity, unexplained ascites, or obvious primary masses to suggest metastatic disease at the time of her surgery.
In this case, there was a subtle progression in the nodular appearance of our patient's port sites on crosssectional imaging over time that was not noted until the patient developed large hepatic and ovarian masses. Unfortunately, PSM portends a poor prognosis so earlier PSM detection would not likely have provided any chance for surgical cure. The two-year survival rate for patients diagnosed with PSM secondary to gallbladder cancer is <20% [13]. It is likely that PSM secondary to gynecologic malignancies is even more aggressive as they often present with shorter intervals (e.g., eight days) between laparoscopy and PSM [20].
Conclusions
Unique from the previously reported cases of PSM, our patient presented with multifocal biliary strictures and symptomatic jaundice. In this case, operative bile-duct injury after a complicated LCCX and autoimmune SSC was considered the most likely diagnosis. Retrospective review of imaging for PSM in cases of unexplained bile duct strictures may prove helpful for early diagnosis of SSC. Unfortunately, PSM portends a poor prognosis so earlier PSM detection likely would not have provided any chance for surgical cure. In summary, PSM due to an unknown primary is exceedingly rare in itself. This is the first reported case, to our knowledge, of PSM with an unknown primary that presented with SSC after an LCCX.
Additional Information Disclosures
Human subjects: Consent was obtained or waived by all participants in this study. N/A issued approval N/A. Informed patient consent was obtained from the patient's next of kin for publication of the case details.
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.
|
2021-10-19T16:06:17.239Z
|
2021-09-01T00:00:00.000
|
{
"year": 2021,
"sha1": "0f7a9c046b126dce924432069dde335024ff143b",
"oa_license": "CCBY",
"oa_url": "https://www.cureus.com/articles/71481-diagnosis-of-secondary-sclerosing-cholangitis-by-port-site-metastasis.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "c208de4c50b5c243bfb023df5b2a4a29dac91a0e",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
249101687
|
pes2o/s2orc
|
v3-fos-license
|
A systematic review on cough sound analysis for Covid-19 diagnosis and screening: is my cough sound COVID-19?
For COVID-19, the need for robust, inexpensive, and accessible screening becomes critical. Even though symptoms present differently, cough is still taken as one of the primary symptoms in severe and non-severe infections alike. For mass screening in resource-constrained regions, artificial intelligence (AI)-guided tools have progressively contributed to detect/screen COVID-19 infections using cough sounds. Therefore, in this article, we review state-of-the-art works in both years 2020 and 2021 by considering AI-guided tools to analyze cough sound for COVID-19 screening primarily based on machine learning algorithms. In our study, we used PubMed central repository and Web of Science with key words: (Cough OR Cough Sounds OR Speech) AND (Machine learning OR Deep learning OR Artificial intelligence) AND (COVID-19 OR Coronavirus). For better meta-analysis, we screened for appropriate dataset (size and source), algorithmic factors (both shallow learning and deep learning models) and corresponding performance scores. Further, in order not to miss up-to-date experimental research-based articles, we also included articles outside of PubMed and Web of Science, but pre-print articles were strictly avoided as they are not peer-reviewed.
INTRODUCTION
COVID-19 has ravaged the world since the World Health Organization (WHO) declared it a global health emergency on January 30, 2020. There are 446,511,318 total confirmed cases and 6,004,421 deaths worldwide as of March 8, 2022 (see Fig. 1) dashboard. https://covid19.who.int/). Despite all our best efforts, the human cost in terms of person-hours spent and lives lost to this virus is extensive. Although there are multiple ways to fight the virus, the greatest hope for ending the pandemic is vaccines, of which 10,704,043,684 vaccine doses (as of March 6, 2022) have been distributed globally (Ahmed et al., 2020). However, if we rely on vaccines alone, there will be many more deaths before humanity can put this pandemic behind it. Therefore, how else can we fight this virus other than the suggested preventative measures, like washing our hands and wearing masks? Other than diagnosis, screening tests will be one of the most integral parts of containing this virus completely, especially in resource-constrained regions across the world, where vaccines are unavailable or untrusted because of general vaccine skepticism within a community.
Cough is used to diagnose and predict many diseases to help clinicians better serve patients; some of these diseases include but are not limited to pneumonia and asthma. Multiple studies have recently used artificial intelligence (AI)-guided tools to help classify different coughs and help diagnose or predict certain diseases. For example, conducted a study to help classify the severity of chronic obstructive pulmonary disease by measuring the airflow and volume of patients breathing with spirometry tests. Nakamori et al. (2020) showed that using a simplified cough test, they can predict the risk of patients with acute stroke getting pneumonia. Another study conducted by used AI-guided tools to detect whether a specific cough was 'dry' or 'wet,' which would help establish the presence of sputum. In Rudraraju et al. (2020), a study also found that AI-guided tools can classify obstructive vs restrictive lung problems. Lastly, Han et al. (2020) used cough and other types of valuable data to estimate the severity level of illness. These previous five examples show that different cough metrics can get leveraged in multiple ways, with and without AI-guided tools, to help diagnose and predict diseases in humans. However, because cough has helped diagnose many other diseases, AI-guided tools that leverage cough sound samples must also take into consideration the works involved with diagnosing other diseases to more accurately screen COVID-19 (Imran et al., 2020). Furthermore, mel-frequency cepstral coefficients (MFCCs) processing helps COVID-19 screening because authors found high similarity in MFCCs between different COVID-19 cough and breathing sounds . Although this is only one cough feature that most robust AI-powered algorithms use, it is a good case-in-point for the research that we will explore in this article. Based on the features like MFCCs extracted from the audio recording, AI-guided tools can help detect a cough and use this process to screen COVID-19 patients (Chen et al., 2021;Miranda, Diacon & Niesler, 2019;Chuma & Iano, 2021). More often, accuracy for basic cough detectors is generally greater than 95%. While reviewing state-of-the-art works, we consider whether AI-guided tools are reproducible (with external validation) for COVID-19 diagnosis-'is my cough COVID-19?' (Topol, 2020).
A brief clinical analysis of the symptom cough in COVID-19 patients reveals many factors to consider related to AI-guided tools. The first factor to consider is that there is no international definition of the symptom of cough (Morice, 2007). This fact shows that diagnosis of the symptom cough is subjective, and clinician reporting of cough can be inaccurate, and it is even more inaccurate in the general population (Donnelly & Everard, 2019; Morice, 2007). The following factors should also be get considered: sociodemographics, the severity of the illness, the patient's age, and temporal considerations. Multiple socio-demographic studies on COVID-19 show that cough is ubiquitous in terms of the presentation rate of the symptom between different populations such as African-Americans and Latinos (Gayam et al., 2020;Weng et al., 2020). Furthermore, these studies show that there are differences in the presentation of other symptoms regarding race. Next, multiple studies regarding the severity level of the illness show little difference between mild, moderate, severe, and critical infections (Lapostolle et al., 2020;Wang et al., 2020;Vaughan et al., 2021;Matangila et al., 2020;Yang et al., 2020). However, there are a few studies, out of many, that show a difference (Fontana, 2007). Moreover, patients in the mild to moderate categories generalize well to the population at large because most COVID-19 infections will fall into this category (Yang et al., 2020). Age category is also an essential factor to consider and studies show there is little to no difference between age groups for the presentation rate of the cough symptom (Brendish et al., 2020;Cattelan et al., 2020;Guan et al., 2020;Morice, 2007). Lastly, considering the temporal aspects of cough in COVID-19 infections, we see a critical aspect that shows that cough is a reliable metric. That aspect is, the symptom of cough lasts longer and at a higher rate than all other symptoms (Faezipour & Abuzneid, 2020). This study analyzed a relatively small number of patients; however, if this study can be confirmed, it shows that cough as a screening metric may help catch an infection in the later stages. Overall, based on these studies, the cough presentation rate is high relative to all symptoms and can be considered the second most common symptom after fever.
This paper systematically reviews AI-guided tools that are used to analyze cough sound for COVID-19 screening. To avoid possible confusion between diagnostic tests and screening tests, a diagnostic test aims to establish the presence/absence of the disease, while a screening test is to detect potential disease indicators. In short, we systematically review screening tests with the use of AI-guided tools. AI-guided tools rely on fully observed clinically diagnosed data (cough sounds, in our case) and are considered one of the fastest and most accurate tools in classifying/screening COVID-19 (Imran et al., 2020;Ahmed et al., 2020). Other data, such as respiratory and breathing sounds, can also be considered for screening , but our study is limited to analyzing cough sounds because it is one of the most widespread symptoms that can easily get screened with low-cost equipment for many people in a short period of time.
The remainder of the paper is organized as follows. "Survey Methodology" presents our review methodology (inclusion criteria of the published research articles). It mainly includes AI-guided tools for COVID-19 screening using cough sounds in both crowdsourced and laboratory confirmed data ("AI-guided tools for COVID-19 screening using cough sounds"). In addition, in "Research articles outside of PubMed and Web of Science", we also review important findings outside PubMed and Web of Science. These sections get followed by discussion in "Discussion". "Conclusion" concludes our study.
SURVEY METHODOLOGY
An essential aspect of any review article is how the information was collected (McDonagh et al., 2013). For our systematic review, we follow a workflow representing different phases of systematic review, where it primarily includes identification, screening, eligibility and included criteria as shown in Fig. 2. In our study, we used PubMed central repository and Web of Science, and selected key words are (Cough OR Cough Sounds OR Speech) AND (Machine learning OR Deep learning OR Artificial intelligence) AND (COVID-19 OR Coronavirus). In our screening, duplicate items were removed, and experiment-based papers are included. For better meta-analysis, we screened for appropriate dataset (size and source), algorithmic factors (both, shallow learning and deep learning models), and corresponding performance scores. For meta-analysis, we included experimental-based research articles. Pre-print articles (e.g., ArXiv, medRxiv and TechRxiv) were strictly avoided as they are not peer-reviewed.
AI-guided tools for COVID-19 screening using cough sounds AI has contributed a lot in healthcare and integrating speech/audio processing tools is no exception (Santosh, 2019;Mukherjee, Salam & Santosh, 2021). We review state-of-the-art works on COVID-19 screening through cough sounds. Not only that, but we also address other data types such as sneezing, respiratory, speech, throat clearing, wheezing, and breathing. For a better understanding, Fig. 3 provides a workflow of how AI-guided tools are commonly employed. It takes cough sound data as an input, extracts features and differentiates COVID-19 positive human subjects from non-COVID ones.
It is possible to develop AI-guided audio/speech processing tools/techniques that extract and leverage acoustic biomarker features to pre-screen COVID-19 recordings from cough. Before developing detailed information, let us discuss a few cases where use of biomarkers and smartphone-based tools/techniques can be seen. Experts proposed a speech and signal processing approach to analyze COVID-19 (in both cases: asymptomatic and symptomatic). A dataset with a complexity of neuromotor coordination across speech subsystems that involve respiration, phonation, and articulation, encouraged by the distinct nature of COVID-19's lower vs upper respiratory tract inflammation, helps detect COVID-19 in asymptomatic and symptomatic patients (Quatieri, Talkar & Palmer, 2020). These biomarker features can get leveraged with AI-guided tools to have a significant effect to increase forced cough COVID-19 detection accuracy (Laguarta, Hueto & Subirana, 2020). Similarly, smartphone-based self-testing of COVID-19 using breathing sounds and their implications to find breathing complications by comparing specific acoustic signal patterns has also been gotten observed in this review. In Faezipour & Abuzneid (2020), the authors suggest their opinion for using advanced signal processing in tandem with new deep machine learning and pattern recognition techniques on smartphone technology. In Tables 1 and 2, we organize the results to contrast studies with and without laboratory confirmation of their datasets for better understanding. Not limited to cough sounds, other sources of data, such as speech, breathing and respiratory sounds, are considered in some studies (see Fig. 3).
Studies on non laboratory confirmed datasets
In Table 1, we summarize work in accordance with how features were extracted (shallow learning or deep learning). In other words, we consider handcrafted features (with typical machine learning classifiers) and deep learning algorithms to detect cough sounds in COVID-19 patients. Regardless of how samples were collected, we start with popular features, such as MFCCs, typical machine learning classifiers, and end with deep features. Table 1 follows the order of the description (provided below). • Handcrafted features (shallow learning): MFCCs contributed a lot in audio/speech processing. In Dash et al. (2021), the authors used the MFCCs of speech signals. From the Coswara dataset, designated database-1, the authors used 570 participants, and each participant contributed nine audio files to various categories of samples (3,470 are clean, 1,055 are noisy, and the rest are highly degraded sound samples). They did not, however, mention if a laboratory has verified the data. Next, the authors used a crowdsourced database designated database-2, with 6,631 users; 235 users were declared COVID-19 positive. The authors suitably optimized the frequency range and the conversion scale of the recordings. The authors used adaptive synthetic sampling approach for imbalanced learning to create synthetic data. In their experiments, they reported accuracy of 0.86 (0.74) for database-2 (database-1) cough sounds.
In a similar fashion, Mouawad, Dubnov & Dubnov (2021) analyzed symbolic recurrence quantification measures derived from MFCC features to detect COVID-19 cases using cough sounds and speech. They used recurrence dynamics and variable Markov model on sustained vowel 'ah' recordings and showed that their model is robust for detecting the disease in sustained vowel utterances. The dataset was composed of 1,927 cough sound samples with only 32 sick patients and 1,488 speech records with only 20 sick patients. The authors ensured that the classification model was not biased towards the majority class and experimented with a wide range of data sampling techniques such as oversampling the minority class or under-sampling the majority class. They reported the following test results: an accuracy of 0.91 and an AUC of 0.84.
In addition to handcrafted features, other studies highlight the use of typical machine learning classifiers, such as Neural Networks (NNs), Support Vector Machine (SVM), Random Forest (RF), and Logistic Regression (LR). In Shimon et al. (2021), the authors used multiple different shallow learning techniques like SVM and RF. Their dataset was from non-publicly available crowdsourced data that took multiple types of recordings from people over multiple days. The type of recording used in this article was both cough sounds and the vowel 'a'. In total, 1,296 cough recordings were used, and 428 "a" recordings were used. The experimenters used multiple different techniques to extract the audio feature, including the openSMILE toolkit and Librosa. After the classifiers classified each audio sample, the experimenters used simple majority voting to classify the patient as COVID-19 infected or non-COVID-19 infected. SVM showed the best result of the different classifiers with an accuracy of 0.78 and 0.74, an AUC of 0.64 and 0.60, a sensitivity of 0.95 and 0.90, and a specificity of 0.36 and 0.35 for 'a' and cough, respectively.
In Despotovic et al. (2021), the authors used wavelet scattering features and deep audio embeddings with multiple different shallow learning techniques. Their dataset used crowdsourced data with 1,103 participants where each produced two to five samples each. From these samples, the authors created a balanced dataset with a total of 496 cough samples. They used five-fold leave one out cross-validation for their experimental set-up. Boosting with the wavelet scattering features seemed to have the best overall performance with a 0.88 accuracy, a 0.87 sensitivity, and a 0.89 specificity. In Melek (2021), the authors used multiple different shallow learning architectures to classify cough as COVID-19 Positive or Negative. Among the different architectures are Polynomial-SVM, Linear-LDA, and Euclidean-kNN. They used a composition of two different datasets, 121 from Virufy and 59 from NoCoCoDa. Once combined, the dataset consisted of 107 positive samples and 73 negative samples. Unfortunately, the paper did not specify how they split their data, which limits the reproducibility of this model and limits interpretation of the results. Euclidean-kNN achieved their best results on a single cough with a 0.98 accuracy, 1.0 Specificity, 0.97 Sensitivity, and 0.99 AUC.
• Deep features (deep learning): The authors also introduced a Deep Neural Network (DNN) to analyze cough sounds. In Nessiem et al. (2021), the authors explore the usage of deep learning models as a ubiquitous, low-cost, pre-testing method for detecting COVID-19 from audio recordings of breathing or coughing taken with mobile devices or via the web. First, they collected 1,427 audio files from a crowdsourced database. Then, they adapted an ensemble of three Convolutional Neural Networks that utilize breathing and coughing raw audio, spectrograms, and Mel-spectrograms to analyze if a speaker is infected with COVID-19 or not. In their experiments, they reported an unweighted average recall/sensitivity of 0.75, AUC of 0.81, and accuracy of 0.73 by ensembling NNs.
In Gokcen et al. (2021), the authors used a "Deep neural network" for the classification of COVID-19 infected individuals. However, the paper did not give an excellent description of their neural network, and from what was gathered, they used some dense net with neurons. However, it is clear that the paper had a sample size of 822 coughs from the MIT open dataset. This seemingly small neural network produced the following results: an accuracy of 0.79 and a recall of 0.75.
In Feng et al. (2021), the authors used an rNN to classify COVID-19 infected individuals from non-infected individuals. They used two different datasets, Coswara for the one they split into validation and training sets, the next is a small Virufy dataset for its final test set. Coswara consisted of 1,433 participants, which they used a 0.80, 0.20 split. The small Virufy test set consisted of 16 recordings from seven patients. Their model had an encouraging result for the training and validation set; however, when they tried it on the Virufy dataset, they could only get around a 0.81 accuracy with an AUC of 0.79.
In Coppock et al. (2021), the authors used a deep neural network on a crowdsourced dataset. This study used a "CIdeR" Convolutional Neural Network (CNN) which is based on the ResNet architecture to help classify in four tasks: (a) COVID positive and stratum healthy-no-symptoms (62 vs 245 subjects); (b) COVID-positive with COVID cough and stratum healthy-with-cough (23 vs 30 subjects); (c) COVID-positive with COVID cough and stratum asthma-with-cough (23 vs 19 subjects); and (d) COVID positive and COVID negative (62 vs 293 subjects). In total, this dataset consisted of 517 samples from 355 participants, and each sample was chopped up into segments, then used a voting mechanism with a mean average across the segments to break ties. Their respective AUCs are: (a) 0.83, (b) 0.57, (c) 0.91, and (d) 0.85.
In Lella & Pja (2021), the authors extended the work of Brown et al. (2020) (using the exact same dataset) using a Deep CNN with multi-feature channels and data augmentation. In contrast to previous work, their results were improved. It was, however, not mentioned whether the improvement was due to data augmentation, the use of the DCNN classifier, or the increase in the dataset's size. These authors are also ambiguous in declaring whether their data is entirely non-laboratory confirmed, as they mention that users have been admitted 'into the clinic.' To classify between COVID-19 positive and negative cases, an accuracy of 0.95 was reported. No other metrics such as sensitivity and specificity were used. In Pahar et al. (2021), they used multiple different machine learning techniques on a crowdsourced publicly available dataset called the 'Coswara' dataset. It was composed of 92 (1,079) COVID-19 positive samples (negative samples). The authors generated synthetic COVID-19 samples to build dataset balanced. Of many classifiers (e.g., LR, LSTM, CNN and ResNet50), ResNet50 achieved the best-reported accuracy (0.95). However, throughout the entire reported results, sensitivity remained in the lower nineties (percentage-wise). Later, the researchers also tested their model on a clinically validated dataset called Sarcos after training on the Coswara dataset. Their best performing model was a combination of LTSM + SFS which had an accuracy of 0.9291, AUC of 0.938, sensitivity of 0.91, and a specificity of 0.96.
In addition, the authors have found that biomarkers are important features. In Laguarta, Hueto & Subirana (2020), their tool helped extract and leverage acoustic biomarker features to help pre-screen COVID-19 recordings. The dataset was composed of 475 laboratory-confirmed COVID-19 infected patients, 962 doctor assessments, and 1,223 personal assessments. This study used under-sampling to pull coughs from a group of 'hundreds of thousands' of non-infected COVID-19 people to balance the dataset with 224 official tests, 523 personal assessments, and 1,913 personal assessments. However, this study does not thoroughly explain what a 'personal assessment' is. However, this study does explain that a diagnostic test must have been complete within seven days of of collection of the sample, with symptom onset no later than 20 days. In total, 4,256 cough samples were used for training and 1,064 for validation. The CNN incorporates multiple biomarker feature models. These biomarkers included muscular degradation, vocal cords, sentiment, and lung and respiratory tract. They achieved a forced cough COVID-19 screening accuracy of 0.97, an AUC of 0.97, a sensitivity of 0.98, and a specificity of 0.94. Furthermore, asymptomatic patients achieved a sensitivity of 1.0 with 0.83 specificity. This work admits that both the inclusion of doctor assessments and personal assessments hurts the generalizability of their model, and are conducting future tests. Since the results are high, their methods of biomarker use should be further explored. In Jayachitra et al. (2021), the authors used a deep neural network architecture. The paper proposed a RNN that would fuse each mode of input's predictions and add the symptoms to the mix of fusions. The dataset used a mixture of audio and radiological pictures to get a yes or no for COVID-19 infection. After gathering many datasets, their cough sample size consisted of 153 positive samples and 348 negative samples. Again, artificial data was used in their sample size. Their testing procedure did not include a separate test set, so there is a potential bias toward the dataset with the .80 .20 split. However, they report that their multi-modal dataset scored 1.0 across the board on all metrics with just cough at 0.97 accuracy, 0.97 precision, and 0.94 recall. As the model is still progress, it may not be generalized well to the population.
Studies on laboratory confirmed datasets
In Table 2, we observe the use of CNN and/or DNN as well as ensemble learning to detect COVID-19 in the cough sounds of human subjects. Based on the authors' reports, we categorize their studies with laboratory confirmed datasets. In what follows, we discuss CNN-based works as well as a new paradigm of physiological impact.
CNN has been popularly used to develop AI-guided tools across many different computer vision tasks, and COVID-19 screening using cough sounds is no exception. In Imran et al. (2020), the app called 'AI4COVID-19' recorded three seconds of sound and provided results in less than 2 min. This app used a CNN to identify cough sounds. In their first test, they employed Deep Transfer Learning-based Multi-Class classifier (DTL-MC) CNN with multiple output classifiers for COVID-19 and three other diseases. The second used a Classical Machine Learning-based Multi-Class classifier (CML-MC) to judge whether the first one suffered from over-fitting issue. The last test employed a Deep Transfer Learning-based Binary Class classifier (DTL-BC) (similar to the first one), but it was limited to a binary output (yes/no) for a possible COVID-19 infection. For training, they used 1,838 cough sounds and 3,597 non-cough environmental sounds, and for testing, they used 96 bronchitis, 130 pertusses, 70 COVID-19, and 247 normal cough samples. The authors did not mention whether they were laboratory-confirmed, but they referred to the people that provided the samples as "patients," so we assume these samples were collected in a hospital/clinical environment, and thus are laboratory confirmed. The DTL-BC classifier in this study provided the best overall accuracy at 0.93 in differentiating COVID-19 from non-COVID-19 cough sounds and had a sensitivity of 0.96, and specificity of 0.91.
Another form of NN called Recurrent Neural Network (RNN) was used in the literature. In Pinkas et al. (2020), the authors collected self-recordings from phones (vocal utterances, speech and cough sound), and used RNN to produce specialized sub-models for the SARS-CoV-2 classification. From 29 laboratory-confirmed COVID-19 patients and 59 negative control subjects, 235 samples were used for training the model and 57 samples for testing. An ensemble stacking fused the predictions of the sub-models and pre-training, bootstrapping and regularization techniques were used to prevent over-fitting. They reported an accuracy of 0.79 (with a corresponding sensitivity of 0.79) based on leave-oneout validation protocol.
Cough sounds may not be sufficient to completely analyze COVID-19 positive cases, and physiological impact is another important data type that could help better analyze. In Lonini et al. (2021), the authors introduced a novel paradigm based on recording the physiological responses elicited by a short sequence of 2-min activities (physical activity, cardio-respiratory function and cough sounds). While validating the data, they employed a novel body-conforming soft wearable sensor placed on the suprasternal notch to capture physical activity data and cardio-respiratory function. Combining these features on snapshots from 19 COVID-19 positive and 14 healthy cases provided an AUC of 0.94 as compared to 0.64 (with only forced cough sounds).
Research articles outside of PubMed and Web of Science
In order not to miss up-to-date experimental-based research works, we used the exact same keywords in search engines: ACM, IEEE, Springer, and Elsevier. Using the exact same inclusion criteria, these articles are summarized in Table 3, and their respective results on laboratory confirmed data. Han et al. (2021) proposed a voice-based framework to automatically detect COVID-19 positive cases and evaluated the performance on a subset of data crowdsourced from the 'COVID-19 Sound App'. The authors used InterSpeech 09 Computational Paralinguistics Challenge (COMPARE) set, openSMILE toolkit, MFCCs features, and used SVM with linear kernel as the classifier. While using the app, users got asked to record information in the app by submitting their breathing, coughing, and voice samples along with reported symptoms, if any, and provide some basic demographic and medical information. On 828 samples (326 COVID-19 positives and 502 COVID-19 negatives), the authors reported an AUC of 0.79 with a sensitivity of 0.68 and a specificity of 0.74.
In Brown et al. (2020), the authors used data analysis over a large-scale crowdsourced dataset, however some of their data seems to be laboratory confirmed as they asked their patients whether or not they were in a hospital. Furthermore, the dataset consists of 154 cough and breathing sounds from self-reported COVID-19 infected users, of which 54 report a dry cough. Also, the control groups of the experiment consisted of 298 non-COVID-19 users, 32 non-COVID-19 users with a cough, and 20 non-COVID-19 users with asthma and cough. In all cases, they used an 80/20 split for training and analysis and under-sampled the majority for training. They tested classifiers, such as LR, Gradient Table 3 Cough sounds (including other data type) for COVID-19 screening performance in terms of Accuracy (ACC), Area Under the Curve (AUC), Sensitivity (SEN) and Specificity (SPEC).
Performance
Authors ( Boosting Trees, SVM (with a radial basis function kernel). They reported binary classification tasks for: (a) differentiating COVID-19 users from non-COVID-19 users; (b) differentiating COVID-19 users with a cough from non-COVID-19 users with a cough; and (c) differentiating COVID-19 users with cough from non-COVID-19 user with asthma and cough. In their results, the authors found an AUC of 0.80 across all tasks and a sensitivity of 0.69. Similarly, in Vrindavanam et al. (2021), the authors presented an approach to classify audio samples between COVID-19 patient and a healthy person by taking LR, SVM and RF classifiers into account. Their dataset was composed of 150 cough audio samples, of which 54 were COVID-19 positive, and the study did not state whether they were laboratory confirmed. In their test, SVM performed better than LR and RF on all performance metrics and reported an accuracy of 0.84 (with AUC of 0.88). As the dataset size was small, the results could possibly be biased, which was not mentioned in their article.
In Anupam et al. (2021), the authors analyzed COVID-19 cough sounds to detect COVID-19. The dataset is composed of 640 cough samples (source: Coswara database): 160 infected and 480 healthy cases. Interestingly, other papers written around this paper had almost double the number of cough samples. However, the paper did not specify what technique they used for eliminating assumingly half of the dataset, along with any other pre-processing techniques they may have used. For classification, the authors used shallow-based machine learning classifiers such as LR, KNN, SVM, and decision tree algorithms. The SVM classifier performed the best of all: AUC of 0.98, accuracy of 0.99, sensitivity of 0.97, precision of 0.99, and F1 score of 0.98.
In Mohammed et al. (2021), the authors developed a robust classifier for a COVID-19 pre-screening model from crowdsourced cough sound data. While detecting COVID-19 from sound datasets, the authors faced two main challenges. The first challenge being a variable number of coughs in each recording, and the second is the low number of COVID-19 positive cases compared to healthy coughs in the data. In total, they were able to obtain 8,886 cough samples which they then used under-sampling the majority to create a balanced dataset of 1,276 cough samples. After obtaining their balanced dataset, they used a VGG16 to extract the audio features such as Mel-spectrograms, MFCCs, spectrograms, and even used the raw audio data to obtain 25,088 feature vectors per audio input. Afterward, they used two separate training pipelines with ensemble learning, one with shallow-based learning such as LR, SVM, and K-Nearest Neighbor, and the second with ensemble learning of three different CNNs, one CNN built from scratch along with two pre-trained VGG models. Their method illustrated a respectable performance using an ensemble model on the testing dataset with AUC of 0.77, precision of 0.80, recall of 0.71, F1 score of 0.75, Kappa of 0.53, and an accuracy of 0.77.
In Bansal, Pahwa & Kannan (2020), the authors proposed a CNN-based audio classifier using an open cough dataset. They used a manually labeled dataset that was composed of two categories: COVID-19 and non-COVID. They proposed two different approaches: one is based on MFCC features and another used spectrogram images for CNN network. First, they took a dataset of 911 cough sounds, where 871 are from YouTube videos and 40 are from audio files. Afterward, they pared that down to 500 audio samples after labeling. The authors found that the MFCC approach produced 0.71 test accuracy, 0.81 sensitivity, 0.61 precision, and 0.69 F1 score. These results were better than the spectrogram-based approach. It would be interesting to see the results of this paper's methods along with mel-spectrograms as well. Grant, McLane & West (2021) developed a method that can be applied to analyze sounds to detect COVID-19 on a crowd-sourced data with sound recordings and these were self-identified. They took a total of 1,040 (78 COVID-19 positive) cough samples and 1,199 (81 COVID-19 positive) speech and breathing samples. MFCCs and relative spectra perceptual linear prediction features were evaluated independently with two different classifiers: DNN and RF. The following AUCs were reported: 0.6836 from cough sounds (DNN classifier), 0.79 from speech sound (RF classifier), and 0.76 from breathing sound (DNN classifier). Khriji et al. (2021) proposed a deep Long Term Short (LSTM) technique to detect COVID-19 infections from cough, breath and sneeze signals via smartphones or wearable sensors. The dataset was composed of audio signals like cough, sneeze and breath. Furthermore, this dataset was partitioned further into three subsets, including a training set (sick (1,435) + not sick (2,283)), a validation set (sick (468) + not sick (753)), and a test set (sick (642) + not sick (1,012)). The authors, however, did not mention whether dataset was laboratory confirmed. The authors reported an accuracy of 0.80 and a sensitivity of 0.78.
In Pal & Sankarasubbu (2021), the authors evaluated their model using a medical dataset containing symptoms and demographic data of 30,000 audio segments. They extracted 328 cough sounds from 150 patients with four cough classes (COVID-19, Asthma, Bronchitis and Healthy). They used a CNN to classify cough sounds, and the study showed that their model captured many robust features of cough sounds to distinguish between COVID-19 coughs and several types of non-COVID-19 coughs. The authors reported an accuracy of 0.95, a sensitivity of 0.90, a specificity of 0.97, F1 score of 0.90, and precision of 0.91.
In Wei et al. (2020), the authors analyzed a real-time robot-based tool to evaluate risk level due to COVID-19 infection. They used real-time speech analysis, temperature, keyword, cough, and other functions to convert live audio into structured data. The authors collected a dataset of 1,283 speech recordings via human-robot conversations from 184 people for the test evaluation. Of all, 392 segments from 64 people were laboratory-confirmed COVID-19 infected. The remainder of the samples were healthy individuals with a history of smoking, acute bronchitis, chronic pharyngitis, children with pertussis and healthy people with no smoking history. For cough detection, using a CNN, they reported an accuracy of 0.76, while their sensitivity was high (0.99).
Similarly, Hassan, Shahin & Alsabek (2020) studied early screening and diagnoses of COVID-19 patients by using RNN and leveraged its significant architecture to discover the acoustic features of cough, breathing and voice of the patients. In their study, 60 healthy and 20 COVID-19 infected patients were asked to record three separate samples: cough, breath and voice sounds. The COVID-19 infected patient samples got collected from hospitals in the UAE. With data split, 70/30 (train/test), they achieved the following accuracies: 0.97 for cough sounds, 0.98 for breathing sounds, and 0.88 for voice sounds.
DNN is no exception in cough sound analysis in detecting COVID-19. In Andreu-Perez et al. (2021), the authors used empirical mode decomposition with the tensor of speech features and a Deep Artificial Neural Network (ANN) to detect cough of COVID-19 patients. Of all, 8,380 samples of cough sounds were collected via a web app called 'Cough Detect' to record the coughs anonymously; 2,339 of the samples are from patients with confirmed qRT-PCR laboratory tests for the COVID-19 infection. The authors reported promising results: AUC of 0.99, sensitivity of 0.96, and specificity of 0.96.
DISCUSSION
In this section, we summarize our observations based on the current state-of-the-art works by considering cough sound as primary data. Our discussion is based on cough sound analysis via AI-guided tools for COVID-19 screening (ref. "AI-guided tools for COVID-19 screening using cough sounds").
(a) Cough symptoms present themselves at a high rate relative to other symptoms in the vast majority of studies (ref. "Introduction"). It stands to reason that this will also translate well into the population at large, and the majority of the general population would present a cough if infected by COVID-19. In addition to other clinical tests, cough tests could potentially help build clinical decisions for COVID-19 positive. In other words, cough is not the only reliable metric, it is however comparable to others, such as fever. AI-guided tools (ref. "AI-guided tools for COVID-19 screening using cough sounds") can analyze cough sounds and help detect COVID-19 at a very high rate regardless of cough diagnosis. This fact, along with affordability and accessibility, could make AI-guided cough screening a first-line defense against COVID-19 and similar infectious outbreaks, especially in resource-constrained regions.
(b) Following the state-of-the-art methods for screening COVID-19 infections (ref. "AI-guided tools for COVID-19 screening using cough sounds"), we observed a clear distinction between two types of datasets: with and without laboratory confirmation (see Tables 1 and 2).
On laboratory confirmed datasets, AI-guided tools performed better as compared to non-laboratory confirmed data in the vast majority of reviewed articles. Also, the laboratory confirmed data could potentially provide convincing results, as they were annotated by experts. Machine learning algorithms require enough training data (with all possible positive cases). Moreover, AI-guided tools trained with fairly large amounts of data can be accurately used to screen COVID-19 human subjects from a forced cough regardless of the dataset, even if performance degrades slightly on non-laboratory confirmed data. In all cases, importantly, the authors did not mention whether their AIguided tools are externally validated and reproducible (Topol, 2020). Moreover, most studies with high metrics in the non-laboratory confirmed sections have serious deficiencies in the reporting of their methods, or their model is admittedly biased.
(c) Integrating other data types such as sneezing, respiration, speech, throat clearing, wheezing, and breathing can help build a better decision-making process (see Tables 1 and 2).
(d) After reviewing the use of shallow learning and deep learning (ref. "AI-guided tools for COVID-19 screening using cough sounds"), we observed that the use of deep features could open the genericity of the model rather than relying on prior knowledge.
(e) In designing AI-guided tools, we observed the use of additional features, such as biomarkers and clinical data (Laguarta, Hueto & Subirana, 2020). These features get ensembled into the CNN that utilizes the standard features extracted from audio recordings. Such a feature integration could potentially drive future works. Introducing new paradigms (e.g., biomarkers) could help better analyze COVID-19.
CONCLUSION
In this article, we have systematically reviewed state-of-the-art works in both years 2020 and 2021 by taking into account AI-guided tools to analyze cough sounds for COVID-19 screening. Clinically, we have found that cough (via the use of cough presentation rate in ref. "Introduction") is considered as one of the primary symptoms in severe and nonsevere infections alike. COVID-19 screening (ref. "AI-guided tools for COVID-19 screening using cough sounds") using cough sounds is found to be potentially a cheap, effective, and available alternative to help people decide to quarantine or get tested. In other words, as cough is ubiquitously presented among varying populations, it is possible to develop AI-guided tools with high accuracy to do mass screening using cough sounds. This method of screening deserves to get more thoroughly investigated and developed into production via a mobile app.
Integrating other data types, such as sneezing, respiration, speech, throat clearing, wheezing, breathing, biomarkers and clinical data, can help build a better decision-making process; we will extend our work by implementing ensemble DNNs within a multimodal learning mechanism.
ADDITIONAL INFORMATION AND DECLARATIONS Funding
This work was supported by the Dean's opportunity fund, College of A & S, The University of South Dakota. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Grant Disclosures
The following grant information was disclosed by the authors: The University of South Dakota.
|
2022-04-29T15:49:42.889Z
|
2022-04-25T00:00:00.000
|
{
"year": 2022,
"sha1": "4a8bf67d9a59570dca08361d41a350c6ed77ed45",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.7717/peerj-cs.958",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "8a261c92c2c13ae784caa25840ff194d5e6c6375",
"s2fieldsofstudy": [
"Medicine",
"Computer Science"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
18056081
|
pes2o/s2orc
|
v3-fos-license
|
All flat manifolds are cusps of hyperbolic orbifolds
We show that all closed flat n-manifolds are diffeomorphic to a cusp cross-section in a finite volume hyperbolic (n+1)-orbifold.
Introduction
By a flat n-manifold (resp. flat n-orbifold) we mean a manifold (resp. orbifold) E n /Γ where E n is Euclidean n-space and Γ a discrete, cocompact torsion-free subgroup of Isom(E n ) (resp. Γ has elements of finite order). In [7], Hamrick and Royster resolved a longstanding conjecture by showing that every flat n-manifold bounds an (n + 1)-dimensional manifold, in the sense that each diffeomorphism class has a representative that bounds.
Flat manifolds and orbifolds are connected with hyperbolic orbifolds via the structure of the cusp ends of finite volume hyperbolic orbifolds-if M n+1 is a non-compact finite volume hyperbolic orbifold, then by a standard analysis of the thin parts, a cusp cross-section is a flat n-orbifold (see below and [14]). In the early eighties, motivated by this and work of Gromov [6], Farrell and Zdravkovska [5] posed the following geometric version: Question Does the diffeomorphism class of every flat n-manifold have a representative W which arises as the cusp cross-section of a finite volume 1-cusped hyperbolic (n + 1)-manifold?
We note that it makes sense to ask only for bounding up to diffeomorphism type, since it is a well-known consequence of Mostow Rigidity that there are algebraic restrictions on the isometry type of the flat n-manifolds that can arise as cusp cross sections of finite volume hyperbolic (n + 1)-manifolds.
In this generality, the question has a positive answer in dimension 2, that is, for the torus and Klein bottle, but as is shown in [10] is false in dimension 3 (and indeed in all dimensions 4k − 1 ≥ 3). On the other hand, since finite volume, non-compact hyperbolic manifolds exist in all dimensions (see §6 for example), this together with the fact that maximal abelian subgroups are separable (see below and [9]) can be used to deduce the n-torus is a cusp cross section of a hyperbolic manifold with possibly many cusps.
In this paper we shall show: Theorem 1.1 For every n ≥ 2, the diffeomorphism class of every flat nmanifold has a representative W which arises as some cusp cross-section of a finite volume cusped hyperbolic (n + 1)-orbifold.
As already remarked, much more is known in dimension 2, and we only include this since the argument also works there. In dimension 3, [11] proves a little more, namely that every flat 3-manifold is a cusp cross-section of a hyperbolic 4-manifold, but even there, the number of cusps is not known to be one. Also in dimension 3, in [13], it is shown by an ad hoc argument that of the 10 diffeomorphism types of flat 3-manifolds, 7 have representatives that arise as a cusp cross-section in some 5 cusped hyperbolic manifold related to gluings of the ideal 24-cell in H 4 .
The proof of Theorem 1.1 uses arithmetic methods, together with some extra control one can arrange to make a separability argument. The main ingredient in the proof is the following (see §2 for definitions and notation): Theorem 1.2 Let Γ be the fundamental group of a flat n-manifold. Then there is a quadratic form q n+2 defined over Q, of signature (n + 1, 1) for which Γ embeds as a subgroup of O 0 (q n+2 ; Z).
The group O 0 (q n+2 ; Z) is arithmetic, and hence of finite co-volume acting on H n+1 . The proof of Theorem 1.1 is then completed by using a subgroup separability argument to pass to a subgroup of finite index in O 0 (q n+2 ; Z) for which the group Γ is a maximal peripheral subgroup.
As with [11], we cannot guarantee 1-cusped examples, and at present, we have been unable to pass to manifolds, even assuming separability of geometrically finite subgroups.
It appears that the question in its full generality is much harder to approach, for example, it appears be unknown, whether, for n ≥ 4, there even exist 1-cusped hyperbolic n-manifolds of finite volume. As far as the authors are aware, this is known only in dimensions 2 and 3.
Quadratic forms
We need to recall some standard facts about quadratic forms and orthogonal groups of such forms; [8] is a standard reference.
If f is a quadratic form in n + 1 variables with coefficients in K and associated symmetric matrix F , let the Special Orthogonal group of f . These are algebraic groups defined over K . If L is a subring of C we denote the set of L-points of these groups by O(f ; L) (resp. SO(f ; L)).
Definition. Two n-dimensional quadratic forms f and q defined over a field K (with associated symmetric matrices F and Q) are equivalent over K if there exists P ∈ GL(n, K) with P t F P = Q.
If K ⊂ R is a number field, and R K its ring of integers, then SO(f ; R K ) is an arithmetic subgroup of SO(f ; R), [3] or [2]. In particular SO(f ; R K ) has finite co-volume acting on an associated symmetric space. The following is well-known and proved in [1] for example. Lemma 2.1 If f and q are equivalent over K then: • SO(f ; R) is conjugate to SO(q; R) and SO(f ; K) is conjugate to SO(q; K).
• SO(f ; R K ) is conjugate to a subgroup of SO(q; K) commensurable with SO(q; R K ).
Crystallographic groups and hyperbolic orbifolds
We record some facts about crystallographic and Bieberbach groups that we will need, see [4] for a comprehensive discussion of these groups, and [14] Theorem 4.2.2 for the theorem stated below.
An n-dimensional crystallographic group is a cocompact discrete group of isometries of E n . When Γ is torsion-free it is called a Bieberbach group. By Bieberbach's Third Theorem (see [4]), the number of n-dimensional crystallographic groups up to affine equivalence is finite. What we require is summarized in the following from [14] page 222: Theorem 2.2 An n-dimensional crystallographic group Γ contains a normal subgroup of finite index consisting of translations, that is free abelian of rank n. The maximal such subgroup is characterized as the unique maximal abelian subgroup of finite index in Γ.
Theorem 2.2 implies that associated to an n-dimensional crystallographic group Γ is a finite group θ(Γ), the holonomy group, and an extension: When Γ is a Bieberbach group we get a free action on E n , by rigid motions, that is for all for some t γ ∈ E n (see [4] for details).
2.2.1
We refer the reader to [12] or [14] for further details on what is contained in the next two subsections.
Equip R n+1 with the quadratic form f n = −1, 1 . . . , 1, 1 of signature (n, 1). The connected component of the identity in O(f n ;R) will be denoted O 0 (f n ;R). This group preserves the upper sheet of the hyperboloid f n (x) = −1 but contains reflections so reverses orientation. We identify O 0 (f n ; R) with Isom(H n ).
Passing to the connected component of the identity in SO(f n ; R), denoted SO 0 (f n ; R) (which has index 4 in O(f n ; R)), gives a group which may be identified with Isom + (H n ); it preserves the upper sheet of the hyperboloid f n (x) = −1 and the orientation. Given a (discrete) subgroup ∆ of O(n, 1; R), ∆ ∩ SO 0 (n, 1; R) has index ≤ 4 in Γ.
2.2.2
An element of O 0 (f n ; R) is parabolic (resp. elliptic) if it has a unique fixed point which lies on S n−1 ∞ (resp. has a fixed point in H n ). Given a non-compact hyperbolic n-orbifold Q = H n /∆ of finite volume, and C a cusp cross-section of Q, then there is a subgroup ∆ C < ∆ consisting of parabolic and elliptic elements having an invariant horosphere H such that H/∆ C = C . The group ∆ C is a crystallographic group. This group is called a maximal peripheral subgroup of ∆.
In terms of the model above, an element is parabolic if and only if it is not elliptic and leaves invariant a unique light-like vector v . Furthermore, in the context of ∆ and ∆ C above, all elements of ∆ C will preserve this unique lightlike vector. We summarize what we need.
2.2.3
We record the following for convenience concerning arithmetic subgroups of Isom(H n ). For more details, see [3], [2] and [16].
Let f be a diagonal quadratic form with rational coefficients and signature (n, 1). Thus there is a P ∈ GL(n + 1, R) such that P t F P = F n , and so the group P O 0 (f ; Z)P −1 defines a discrete arithmetic subgroup of Isom(H n ). The theory of arithmetic groups then gives, Theorem 2.4 In the notation above, the groups P O 0 (f ; Z)P −1 are finite covolume subgroups of Isom(H n ).
In what follows, we will suppress the conjugation by P . A group O 0 (f ; Z) (and hence the conjugate in Isom(H n )) is cocompact if and only if the form f does not represent 0 non-trivially with values in Q, see [3]. Whenever n ≥ 4, the arithmetic groups constructed above are non-cocompact, since it is well known every indefinite quadratic form over Q in at least 5 variables represents 0 nontrivially, see [8]. In fact up to commensurability, all non-cocompact arithmetic subgroups of O 0 (f n ; R) arise from this construction (see [16]).
Some technical lemmas
In this section we gather together a collection of well-known results on separability properties of groups that will be used to pass from Theorem 1.2 to Theorem 1.1.
Recall that a subgroup H of a group G is separable in G if, given any g ∈ G\H , there exists a subgroup K < G of finite index with H < K and g / ∈ K . G is called subgroup separable (or LERF) if all finitely generated subgroups of G are separable in G. The profinite topology on a group G is defined by proclaiming all finite index subgroups of G to be a basis of open neighbourhoods of the identity. Since open subgroups are closed in the profinite topology, the following reformulates separability: Lemma 2.5 Let G be a group and H < G is a subgroup. G is H -subgroup separable if and only if H is closed in the profinite topology on G.
Lemma 2.6 Let G be a group and H < K < G. Assume that H is separable in G and that [K : H] < ∞. Then K is separable in G.
Proof By Lemma 2.5, H is closed in the profinite topology on G. Standard properties of topological groups imply that any coset gH of H in G is therefore a closed subset. Since [K : H] < ∞, K is a finite union of closed sets, hence closed, and therefore separable in G.
The following is also well-known (see [9]): Lemma 2.7 Let G be a residually finite group, and A a maximal abelian subgroup. Then A is separable in G.
Proof of Theorems 1.1 and 1.2
This section is devoted to proving theorems 1.1 and 1.2. We prove the latter first, which we restate for the reader's convenience: Theorem 1.2 Let Γ be the fundamental group of a flat n-manifold. Then there is a quadratic form q n+2 defined over Q, of signature (n + 1, 1) for which Γ embeds as a subgroup of O 0 (q n+2 ; Z).
Before embarking on the proof we remark that the first part of the proof can be replaced by the argument in the proof of Bieberbach's third theorem giving an integral representation into GL(n + 1, Z) of a Bieberbach group. However, we will use some additional features of the construction given below in completing the proof of Theorem 1.1.
Proof of Theorem 1.2 Suppose that Γ is the fundamental group of a flat n-manifold, so as discussed in §2.2, we have a free action of Γ on E n by rigid motions. Thus, if g ∈ Γ, then g acts as v → θ(g)v + t g . and the assignment g → θ(g) is a homomorphism of Γ to its holonomy group θ(Γ), with kernel the maximal translation subgroup of Γ.
Suppose that µ 1 , ...., µ n generate the maximal normal free abelian Z n in G, where µ i acts as translation by m i , where we declare that this is the vector m i = (0, ..., 1, ..., 0), one in i − th place. The group Γ acts by conjugacy on the subgroup µ 1 , ...., µ n and a calculation reveals that gµ i g −1 is the translation given by The normality of the translation subgroup shows that for some collection of integers {a i,j (g)}. Equating these two statements gives a finite integral representation of G given by We now construct an integral linear representation of Γ, as follows. Choose a presentation for the group Γ using generators g 1 , ....., g p and with relations w t (g 1 , ..., g p ) = I , and add all relators which say w j (g 1 , ..., g p ) = µ j for each 1 ≤ j ≤ n. (These ensure that the chosen m i 's don't change.) Each g i acts as so that expanding the equations coming from the relators, we get a collection of equations for the t i 's with rational coefficients which have some solution (for example, that coming from the identity representation that we are given for Γ as a Bieberbach group).
It follows that there are rational solutions to these equations and we claim that any such solution gives a faithful and rational representation of G. Pick any rational solution and regard this as a representation ρ : Γ → ρ(Γ). The conditions imposed by the second batch of equations guarantee that the restriction of ρ to the translation subgroup of Γ is actually the identity homomorphism. Since the translation subgroup is isomorphic to Z n and this is Hopfian, it follows that ker(ρ) avoids the translation subgroup of finite index and hence ker(ρ) is trivial, since Γ is torsion free. This shows that ρ is an isomorphism as required.
It follows that there is a rational solution for the t i 's in terms of the m i 's. Convert the affine representation on E n to a rational linear representation on E n+1 by This is a faithful rational linear representation of Γ, which comes from coning the given action of Γ by rigid motions in the hyperplane e n+1 = 1 to the origin in E n+1 . By conjugating the representation, we may rescale the vector e n+1 and thus arrange that the representation Φ * is actually by integral matrices.
It is slightly more convenient at this stage to define a new faithful integral representation by setting where A T denotes transpose. We can then extend this representation to E n+2 by mapping g toΦ where the column vector v g is to be determined. Now let , be any θ T invariant positive definite inner product on the Z-module µ 1 , ...., µ n ; such an inner product exists by taking a random inner product and forming the θ -average. Let D be the symmetric rational matrix associated to this form in the {m i } basis. Extend this form to E n+2 by summing on a subspace H 2 , which in the language of quadratic forms is a hyperbolic plane. More precisely, we let H 2 denote the 2-dimensional form 2XY , with associated symmetric matrix 0 1 1 0 (see [8] Chapter 1). The form D ⊕ H 2 now has signature (n + 1, 1).
Denoting vectors lying in E n by w and the last two dimensions by v 1 and v 2 , it is a simple matter of linear algebra to show that v g = (W g , τ g ) ∈ E n ⊕ v 1 may be chosen so that eachΦ(g) is an isometry of the form D ⊕ H 2 .
The linear algebra suggests that the matricesΦ(g) may be nonintegral in the last column, since the initial solution vectors v g need only be rational. However, conjugating by a matrix of the form we may find a new collection of matrices which are the same save in the last column, and which has v g replaced by K.v g . In particular, for suitable K we may arrange that the conjugated representation is integral.
After this conjugation, the new representation now leaves invariant a different form, but this new form is rationally equivalent to D ⊕ H 2 ; in particular, it continues to have signature (n + 1, 1).
Claim With this choice, we get a faithful integral representation of the group Γ.
Proof We need only show that the relations in Γ hold. Faithfulness will follow, because if a product of these matrices is the identity, then it must at least be the identity in the n + 1 representation which is already a faithful representation of Γ.
We prove the claim by showing that any isometry, γ say, which is the identity on the upper left n + 1 × n + 1 block is in fact the identity.
This completes the proof of theorem 1.2.
Remark Note that the construction exhibits Γ explicitly as a subgroup of the stabiliser of the lightlike vector v 1 .
Proof of Theorem 1.1 To complete the proof of Theorem 1.1 we proceed as follows. Let q n+2 be the form constructed above and consider O 0 (q n+2 ; Z). As noted in the Remark above, the construction yields Γ as a subgroup of the stabiliser in O 0 (q n+2 ; Z) of the light-like vector v 1 , however the proof provides no control over whether Γ is actually equal to Stab(v 1 ). To achieve this, we show that the subgroup Γ is separable in O 0 (q n+2 ; Z) and then the theorem follows by a standard separability argument.
To this end, let C be the maximal peripheral subgroup of O 0 (q n+2 ; Z) fixing v 1 , so that Γ < C < O 0 (q k+2 ; Z), and C is a crystallographic group. We recall that by Theorem 2.2, C contains a translational subgroup T * which is free abelian of rank n, and is the maximal abelian subgroup of C .
We begin by observing that our construction of the group Γ began with a translational subgroup which contained translation by 1 in each of the coordinate directions, so that after the two dilation conjugacies which convert rational to integral, we see that for some integer p 1, the maximal translational subgroup T * of Stab(v 1 ) ≤ O 0 (q n+2 ; Z) contains the group T p consisting of translations by p in each of the coordinate directions of the first n + 1 coordinates.
We claim that T p is separable in O 0 (q n+2 ; Z). The reason is this: Firstly, any element of O 0 (q n+2 ; Z) which lies outside T * can be separated from T p since in fact it can be separated from T * by Lemma 2.7.
Secondly, we claim that any element of T * − T p may be separated from T p . This involves a few cases, which we now sketch.
In the first place, we observe that all the elements of γ ∈ T * have the first n + 1 entries of its last row being zero, since γ(w) Moreover, if we look at the upper left n × n block of any element of T p , this is constructed to be the identity matrix and if the element γ we wish to separate does not have this property then we may separate by choosing a random prime q not dividing some such entry and use the restriction of the homomorphism SL(n + 2, Z) → SL(n + 2, Z/qZ) to the integral subgroup O 0 (q n+2 ; Z).
It follows that it remains to separate an element γ ∈ T * − T p which is the identity matrix in the first n + 1 columns save for the first n entries in the n + 1-st row. It is these entries which contribute to the translational nature of the elements of T p . However, recall that we have proved that any isometry of , which is the identity on the upper left n + 1 × n + 1 block must in fact be the identity matrix. It follows that in the matrix γ these entries cannot all be divisible by p (else the upper left n + 1 × n + 1 block is identical with that for some matrix of T p and we deduce that γ ∈ T p ) so that we may use the reduction map SL(n + 2, Z) → SL(n + 2, Z/pZ) to separate γ in this case.
We note that this argument can actually be used to show a little more, namely that for any integer r, the subgroups T rp of T p are separable in O 0 (q n+2 ; Z).
The separability of Γ may now be deduced. For if we let T Γ be the maximal abelian subgroup of Γ, then T Γ ∩ T p is a subgroup of finite index in T p and it follows that there is an integer r for which T rp ≤ T Γ ∩ T p ≤ Γ. The separability of Γ follows from Lemma 2.6. This completes the proof of theorem 1.1.
Remark The number of cusps for the groups O 0 (q n+2 ; Z) can be greater than one, even for simple examples. For example, the groups O 0 (f n ; Z) have 1 cusp for 2 ≤ n ≤ 8, but O 0 (f 9 ; Z) has 2 cusps. This can be seen from [15] and [16] which describes these unit groups as groups generated by reflections in certain ideal simplices in H n . The number of cusps being easy to read off from the Coxeter diagrams.
Example
We finish off by giving an example of the construction as an aid to the proof of Theorem 1.2. We shall take the Hantsche-Wendt manifold, which arises as the 3-fold cyclic branched cover of the figure-eight knot. Its fundamental group is therefore a Bieberbach group in dimension 3. Representing matrices are provided on pp. 6-7 of [4] which in our notation are: preserving the rationally equivalent form x 2 + y 2 + z 2 + 4wt.
Note that the above form is equivalent over Q to f 4 = x 2 + y 2 + z 2 + w 2 − t 2 .
In [13] the authors obtain the Hantsche-Wendt manifold as a cusp cross-section of a hyperbolic 4-manifold arising from a torsion-free subgroup in O 0 (f 4 ; Z).
|
2014-10-01T00:00:00.000Z
|
2002-04-18T00:00:00.000
|
{
"year": 2002,
"sha1": "3da76f15043555b40670bda70dee8a90ce29b4e1",
"oa_license": null,
"oa_url": "http://msp.org/agt/2002/2-1/agt-v2-n1-p13-p.pdf",
"oa_status": "BRONZE",
"pdf_src": "Arxiv",
"pdf_hash": "3da76f15043555b40670bda70dee8a90ce29b4e1",
"s2fieldsofstudy": [
"Mathematics"
],
"extfieldsofstudy": [
"Mathematics"
]
}
|
244732730
|
pes2o/s2orc
|
v3-fos-license
|
Analysis of the energy balance and CO2 flow under the influence of the seasonality of climatic elements in a mangrove ecosystem in Eastern Amazon
An unprecedented study was carried out in a mangrove ecosystem in the northeastern coast of the Brazilian Amazon to understand the behavior of climatic elements in a year with the occurrence of El Niño (2015), associated with the seasonal function source/sink of CO2 by the ecosystem. Global radiation (Rg), net radiation (Rn), temperature, relative humidity, precipitation, horizontal wind speed and direction, as well as turbulent flows of sensible heat (H), latent heat (LE), and carbon (f_CO2) were recorded using eddy covariance, a system for studying turbulent flows of heat and gases in the atmosphere. We observed a drastic reduction in rainfall volumes, which accounts for 63.7% of the expected total according to the region’s climatology. Regarding f_ CO2, the highest values of photosynthesis, autotrophic, and heterotrophic respiration of the ecosystem occurred in the wet season due to precipitation, ideal photosynthetically active radiation, lower soil salinity, and higher NDVI of the ecosystem. In the 2nd semester of the year, we observed that the decrease in cloudiness, causing a higher radiation supply in the forest canopy, accompanied by a reduction in precipitation and an increase in the value of H and soil salinity, favored the increase of foliar abscission by the dominant genus Rhizophora and Avicennia, thus influencing the reduction of magnitudes of carbon source/sink functions in the ecosystem during this season, even on high tide days.
Introduction
Mangroves are coastal ecological systems typical of estuarine transition areas between terrestrial and marine environments, subject to tidal regimes. They represent a considerable flow of mass and energy, where the balance of this energy and the carbon involved in the biosphere-atmosphere exchange is transformed into organic matter, from simple assimilates in the form of structural and energetic molecules, contributing to the ecological dynamics of different biotic populations, represented by the flora and fauna of this ecosystem (Nagelkerken et al. 2008;Odum et al. 1982;Schedlbauer et al. 2010).
Mangrove forests account for around 8% of the entire coastline of earth's tropical and subtropical zones, with roughly 137,760 km 2 . Brazil is the third biggest country in mangrove extension, with an area of 9,623.83 km 2 , extending from the State of Amapá to its Southern limit in Santa Catarina. We observe in these ecosystems a significant variation of biological communities, where it operates at certain times of the year as a nursery for a considered variety of invertebrate and vertebrate species (Giri et al. 2011;Lee et al. 2014;Schaeffer-Novelli et al. 2000;Spalding et al., 1997).
Studies on the hourly variability of the main meteorological variables in a mangrove ecosystem in the northeastern state of Pará in the years from 2001 to 2003 verified that the highest average air temperatures occurred in December, while the lowest occurred in April, with the annual average air temperature around 27 °C; relative humidity had an annual average value of 83%. The energy balance showed a significant seasonality demonstrating the local cloudiness's effect in these energy flows (Silva Junior et al. 2006).
Micrometeorological measurements at mangroves in the Northeastern coast of Pará from November 2002 to August 2003 showed that the seasonal and hourly variations of E and LE fluxes and the evaluation of the energy partition and the Rn presented high values during the dry season. Bowen's ratio showed a generally low value in the wet season, indicating a higher proportion of energy was used as latent heat (Pereira and Rodrigues 2013).
Studies on net ecosystem exchange of CO 2 (NEE), annual net ecosystem production (NEP), and the meteorological and environmental conditions that favor such phenomena in a mangrove forest in Florida have shown that maximum daytime NEE estimated from photosynthetic activity varied from −20 to −25 μ mol (photons) m −2 s −1 between March and May. The ecosystem's respiration was highly variable 2.8 ± 2.4 μ mol (CO 2 ) m −2 s −1 , reaching its highest values during the summer wet season. During the winter dry season, CO 2 assimilation in the forest increased due to increased diffuse solar radiation in response to higher radiative transfer in the forest canopy. During 2004, the forest behaved as an atmospheric CO 2 sink, with net annual ecosystem production around 1.170 g Cm −2 . This exceptionally high NEP was attributed to year-round productivity and low ecosystem respiration reaching peak values of only 3 g C cm −2 d −1 (Barr et al. 2010).
Based on the theoretical foundations presented for this ecosystem, it was decided to test the following hypothesis: The ENSO's influence as a negative anomaly in precipitation over the study area and the consequent modulation of this phenomenon in the differentiated partitioning in the energy balance in the region.
Due to the increase in soil salinity in the 2nd semester of the year and the reduction in rainfall, a fact probably intensified by ENSO, the hypothesis of the reduction of carbon absorption by the forest canopy was also tested due to the occurrence of more significant abscission leaf.
The main goals of this study were to calculate the monthly rainfall accumulation for the year 2015, comparing it with the climatology of the region, analyzing the rainfall anomaly influenced by the South Oscillation Index (SOI), study through the diurnal cycle, considering the Local Time (LT), of Rg, Rn, as well as the energy balance, through the diurnal cycle of sensible and latent heat fluxes, for the mangrove ecosystem in a year with the occurrence of ENOS. From the observed data, the air temperature and relative humidity were plotted hourly, and the diurnal cycle of the CO 2 flux (f_ CO 2 ) was calculated. The behavior of the turbulent fluxes under the effect of the tidal cycles was carefully observed.
Geological and floristic features of the mangrove in the northeastern coast of the state of Pará
The northeastern region of the state of Pará has, on its coast, a geological formation dating from the Holocene period, where the rise of salinity of the tidal waters in the fluvial sector can be attributed to the sea level rise in the Atlantic Ocean. It is believed that the region may have been subjected to a complex interaction of several factors, mainly consisting of changes in sea level, subsidence rates, and climate changes, the latter being the one with the most significant potential to have affected the discharge of the Amazon river (Behling et al. 2001;Cohen et al. 2012;Pujos et al. 1996;Souza Filho 2005).
Mangroves on the northeastern coast of the state of Pará, known as the Macrotidal Amazonian Mangrove Coast, are characterized by a low relief, varying from 0 to 80 m, presenting a vast coastal plain, up to 70 km wide, and an adjacent continental shelf, approximately 200 km wide, being extremely irregular, indented, and cut by several estuaries (Souza Filho 2005).
The floristic features of this ecosystem consist of woody plant species, typical of this biome, with the occurrence of halophytic angiosperms belonging especially to the species Conocarpus erectus L., Laguncularia racemosa L., Rhizophora mangle L., Rhizophora racemosa G.F.W. Meyer, Avicennia germinans, and A. schaueriana (Ferreira 1989;Menezes et al. 2008).
Meteorological observations have shown that in this region, there is the formation of squall lines, sea breeze circulation, and the mean flow is from the Northeast, with an average wind speed velocity of 7 m s −1 from 6 to 12 am, LT Silva Dias and Nobre 1995;Germano et al. 2017;Moraes et al. 2005;Silva junior et al. 2006).
Studies on the salinity of the mangrove soil in the municipality of Bragança, Pará have shown that this environmental factor accompanies the rain cycle, when in the wet season, there is a decrease in soil salinity, with a gradual increase in the extent to which there is a reduction of precipitation in the second semester (Koch and Wolff 2002;Mehling, 2006).
Characterization of the experimental site
The micrometeorological data were collected at the experimental site of Cuiarana, a village near the urban center of the city of Salinópolis, Northeastern Pará (Fig. 1A), under coordinates 00 ° 39′50″S, 47 ° 17′10″O. This site presents an area in a state of ecological succession, with typical mangrove tree species, surrounded by mature mangrove (Fig. 1B, D), where the dominance of the genus Avicennia (85.56%) and Rhizophora (11.47%) occurs in a soil characterized as hypersaline, with trees with average heights of 9.5 ± 4.5 (m), and the species Rhizophora mangle L., Avicennia germinans (L.) Stearn., and Laguncularia racemosa (L.) Gaertn., with the following "relative densities" respectively: 30.22%, 58.27%, and 11.51% (De Carvalho and Jardim 2017).
Instrumentation of the micrometeorological tower
A micrometeorological tower was installed (1C) equipped with eddy covariance (EC), a high-frequency system for studying turbulent flows and low-frequency sensors for the collection of meteorological data.
The mangrove tower's high and low-frequency sensors were installed in an aluminum structure, which is 10 m high from the ground and 2 m above the forest canopy, providing the "footprint" an equivalent of 1 km 2 in diameter. The high and low-frequency data integrate two distinct sets of variables recorded in two different data loggers, model CR1000, equipped with a memory card adapter.
The following sensors collected the precipitation data (mm), Rg (W m −2 ), Rn (W m −2 ), air temperature (°C), and relative humidity (%): rain gauge (TE 525mm), Piranometer-CS 300/Campbell, net radiometer, NR Lite 2, K&Z, the sensor for temperature and relative humidity CS 215/ Campbell. The photosynthetically active radiation (PAR) μ mol (photons) m −2 s −1 was calculated empirically from the Rg, according to Aguiar et al. (2012b, a). For the calculation of the radiation budget, we considered Eq. (1) whose terms are expressed by Rn is the net radiation, H stands for sensible heat, LE is the latent heat, and G represents the flow of energy propagating in the substrate.
For the energy balance in this mangrove ecosystem, the estimation of soil heat flux (G) in Cuiarana was carried out by applying a regression equation between the balance of radiation and the heat flux in the soil measured by Barr et al. (2012), which presented a coefficient of determination with R 2 = 0.63, when they performed similar studies, in the mangrove swamp in Southern Florida, between 2004 and 2009. The reason for applying this methodology was the excellent correlation between the net radiation within wet (0.83) and dry (0.94) seasons among the Everglades mangroves on the Florida coast and Cuiarana on the northeastern coast of Pará state.
In order to obtain the turbulent data, the EC technique was used for the calculation of f_ CO 2 , LE, and H. This system consists of an open-path infrared gas analyzer coupled with a 3D sonic anemometer model CSAT-3A, which measures the three wind direction components in degrees and velocity in ms −1 , both from Campbell Scientific, Logan, UT, USA. The storage of this high-frequency data (10 Hz) was programmed to generate average values every 30 min, thus totaling 48 continuous data points for every 24 h, considering the LT (Moncrieff et al. 1997).
The EC technique consists of computing the covariance between the fluctuations of the vertical component of the wind velocity w′, simultaneously with the scalar of interest's conservative amount. Therefore, to calculate the carbon flux, we have to compute the covariance of w′ times the covariance of the concentration of carbon (1) R n = H + LE + G Fig. 1 Location of the study area on the northeastern coast of the state of Pará, with emphasis on the municipality of Salinópolis, where the Cuiarana experimental site is located dioxide (Eq. 2); for latent heat flux, we have covariance between w′ and specific humidity (Eq. 3); for the H, we have to obtain the covariance of w′ with air temperature (Eq. 4).
where ρ is the absolute density of air; w' is the fluctuation of the vertical component of the wind speed; C′ is the standard deviation from the mean CO 2 concentration; λ is the latent heat of water vaporization; q′ is the standard deviation from the mean of the specific humidity; C p is the specific heat capacity of the air at constant pressure; T′ is the standard deviation from the mean air temperature.
Data collection was done monthly, converted into the LoggerNet 4.3 program, and then processed in Alteddy 3.9 software. To calculate the flows of H, LE, and f_ CO 2 , the software was configured to apply a bidimensional rotation to the coordinate system, so the horizontal wind components were aligned with the main flux, and the mean vertical velocity was forced to zero. The Webb correction was used to correct the effects of the humidity of the study area on the temperature measured by the sonic anemometer, in order to adjust the effects of the air density on the measurements of the open path gas analyzer (Webb et al. 1980;Kaimal and Finnigan 1994).
Method applied to analyze errors in the generated data
The dataset of turbulent fluxes for the year 2015 presented periods with operational failures. These failures were mostly related to wet days, high air humidity, and salinity of the environment, with approximately 93 days without registration and months that were compromised, such as January in which there was no generation of data. For the remaining data processed by Alteddy, quality filters were applied to the output fluxes registers, choosing quality flags from 1 to 3 that point to a better information pattern. Reinforcing this methodological standard, it was also considered the dispose of the fluxes data interval in the hours at which there were night data with u* ≤ 0.20 m s −1 . Therefore, it was decided not to apply any filling methodology since the filtering of this data was equivalent to 17% of the data generated. After applying these criteria, we obtained 105 and 167 days of reliable data for the wet and dry seasons, respectively. (2)
Precipitation analysis
For comparing purposes of the precipitation variability, an analysis of monthly accumulated data for 2015 was carried out. These data were compared with the 33-year climatology (1978 to 2010) of the Brazilian National Water Agency (ANA) meteorological stations in Salinas' municipality. Considering the precipitation of 2015, we carried out a study for the anomalies associated with the analysis of the Southern Oscillation Index (SOI), obtained from the NCEP/ NOAA website, in order to understand the effects of the sea surface temperature (SST) of the Pacific Ocean, under the precipitation in the Amazon.
Energy analysis
For the data of H, LE, and f_ CO 2 , retrieved from the EC system, as well as Rg and Rn, we calculated the diurnal cycle to understand how these variables behave hourly in a year (2015) of El Niño with a magnitude of 1.25. The PAR was used to understand the hour and spectral range that allows the efficiency of photosynthesis.
Rg and Rn's data and H and LE obtained by the EC technique were transformed into hourly means to generate 24 data points, so there would be a representation for every hour of the day for 2015.
Soil salinity data and leaf area of the mangrove canopy
Two analyses were carried out to test the hypothesis of the seasonal behavior in carbon fluxes from the foliar abscission. The first one being on the salinity of the water of the soil pores in an area of 1 km 2 on the periphery of the EC tower with a handheld refractometer ATAGO, where six samples were collected in April 2015 with an average result equal to 32.3 ppt and another six samples collected in November 2015, with an average result of 70 ppt.
The second analysis was through remote sensing using the Normalized Difference Vegetation Index (NDVI) from Landsat 8 channels 3 and 4 (red and infrared respectively) for the months that showed the best condition analysis. Therefore, the images were selected for February and September of 2015 to justify foliar abscission by chlorophyll's radiation's reflectance. In February, we verified an NDVI of 0.75, and for September, the NDVI was 0.69 ± 0.04.
Tidal cycle
Another data analyzed was the local tidal cycle, classified as semidiurnal and weakly asymmetrical, with a flood period of up to 6:40, with records of maximum heights of 5.5 m during the spring equinoxes. Currents have a predominant northwest direction during flood tides and southeast direction at reflux tides. The highest current velocities were recorded during the flood tide, with a maximum of 0.5 m/s in March and September, while for the backflow tides (March and June), the maximum velocities reach a value of 0.4 m/s (Pinto et al. 2011).
The tide cycle information for the study year was retrieved from the naval anchorage in Salinópolis, semidiurnal. Only the highest tides' information was considered since they presented soil flood possibilities for the Cuiarana mangrove.
Computation of seasonal energy flow and balance under tidal cycle effect
To analyze the measurements of the turbulent energy fluxes (H, LE, and G), under the influence of high and dry tides during wet and dry seasons, the months of April and October were selected to represent both seasons, respectively. The four highest irradiance times (11:00 am, 12:00, 1:00 pm, and 2:00 pm) were selected, constituting the four data points analyzed in the days of the months studied under the tidal cycle's influence.
In April, the Julian days from 99 to 102 represented the flows during the high tide, and the days from 105 to 108 represented the days for the low tide. For October, the Julian days from 277 to 280 represented the days analyzed for the high tide, while from 297 to 300, represented the days of low tide. The selected data were averaged for the selected days' interval, generating a single value for each variable for April and October under high/low tide influence.
Results and discussions
An analysis of meteorological data for 2015: air temperature, relative humidity, and precipitation, with its respective anomaly The daily analysis of the air temperature in the mangrove ecosystem showed that March had inverse values for the relative humidity throughout the year, which we can highlight the biggest temperature occurring on December 11 with 31.2 °C at midnight, while the lowest temperature occurred on March 7 with 23.9 °C at 5:30 am. The diurnal cycle of the air temperature for 2015 shows that between 10 am and 4 pm, this variable reaches and maintains values between 27 and 28 °C for almost 6h.
The relative humidity maintained its highest values during the wet season with gradual decay during the wet season transition. In the diurnal cycle, the highest air humidity values were 77.8% and occurred at 9:00 am; the lowest values were recorded at 5:00 pm, with 64.7%.
The accumulated monthly precipitation data for 2015 showed a total of 1.73 mm. When compared to the climatological normal of 33 years (1978 to 2010) for the municipality of Salinópolis, whose accumulated annual value was 2.72 mm, we verified that the precipitation in 2015 accounted for 63.7% of the total expected, with a deficit of 988 mm. Regarding the rainfall volume of 2014, which had an annual accumulation of 2.60 mm, we verified that in 2015 it rained only 66.6% of the previous year's volume, representing a deficit of 868.6 mm compared with 2014.
Data from the NCEP/NOAA (2017) showed the Brazilian Amazon under the influence of the El Niño Southern Oscillation (ENSO) in 2015, whose annual average magnitude of the El Niño oceanic index was 1.25 with greater amplitude in the fourth quarter of 2014 (October, November, and December) (2.3), thus influencing a considerable reduction in precipitation seen in the region in 2015 (Fig. 2C).
During the wet season (from January to May 2015), the total precipitation accumulated was 1.58 mm, representing 69% of the total expected compared to this season (2.27 mm). The early months of the wet season (January to March) presented negative precipitation anomalies (696.1 mm). It must be observed, with particular attention, that these months are known for high precipitation volumes.
According to climatic studies by Pereira and Rodrigues (2013) and Silva Junior et al. (2006), in the mangrove ecosystem of Bragança (State of Pará), the months of February, March, and April are the rainiest, with March being the one with the highest accumulated precipitation, with values above 700 mm for the years from 2001 to 2003.
During the dry season of 2015 (August to December), the observed rainfall was quite anomalous compared to the season's total expected. The total precipitation accumulated for the season was 12.3 mm, whereas the climatology expected 159.8 mm. Therefore, in these months, it rained only 7.7% of the total expected for the season.
Silva Junior et al. (2006) reported a similar precipitation behavior to what we have seen in Cuiarana, with a reduction in precipitation in the dry season, noting that from August to November, rainfall accounted for 0.2% of the total annual volume for the years from 2001 to 2003. Zeng (1998), when conducting studies on the seasonal cycle and interannual variability of the Amazonian hydrological cycle, found that there is a good positive correlation between the SOI and precipitation anomalies occurring in the Amazon, with a temporal lag of 3 to 4 months related to events occurring in the Pacific equatorial. Therefore, it should be noted that the SOI of −1.5 in November 2014 was the largest negative anomaly of this year. It may have influenced the anomalous amounts of precipitation seen in February and March 2015 (Fig. 2). Marengo (1992) and Marengo and Nobre (2009) observed that the subtropical jet in the upper troposphere during the austral winter is more intense and closer to the equator than in the summer, being associated with the decrease of the convection in the Amazon Rainforest. Both authors verified that during the austral winter, the circulation in the low troposphere is characterized by the northernmost position of the Equatorial Trough in the region, decreasing the trade winds' intensity and the moisture advection from the Atlantic Ocean.
These observations reinforce the discussion of Nobre et al. (2009). They state that local tropical convection constitutes one of the main processes for the formation of precipitation in the entire Amazon basin, being modulated by large-scale circulations, such as the Hadley cell, the positioning of the Intertropical Convergence Zone (ITCZ), and the zonal circulation of the Walker cell.
De Souza et al. (2004), Marengo and Hastenrath (1993), and Ronchail et al. (2002) found that the weakening of convection in the Amazon Basin during El Niño events is explained by a change in the induced subsidence by the Walker circulation. These authors show that the zonal displacement of the Walker cell, as a consequence of the ENSO, favors the reduction of the upward movement during El Niño events over the north and northeast sectors of South America, which triggers the reduction of the precipitation in the Amazon.
We suspect that the occurrence of SOI for July (−1.9), August (−2.4), and September (−2.7) of 2015 influenced the low rainfall in September (0.0 mm), October (0.0 mm), and November (0.1 mm) of this year, with a great reduction in rainfall for the region when the climatology for this quarter expected a volume of 31.7 mm.
Analysis of the breeze circulation and the mean flow
The observations for wind direction and speed at the Cuiarana ( Fig. 2A, B) experimental site agree with Germano et al. (2017) studies on the breeze circulation in Eastern Amazon. The diurnal cycle of the wind direction shows a maximum of northeast and east, with an average wind speed of 2.21 ms −1 , for the season between 10:00 am and 5:00 pm, during the wet season and 3.0 ms −1 in the dry season, characterizing the occurrence of the sea breeze (SB). The SB occurs in the same direction as the trade winds (east), being more frequent from 9:00 am to 6:00 pm LT.
We observed that the LB occurred in the direction of southeast/south ( Fig. 2A), during the wet season, with a time of occurrence between 00:00, and 6:00 am LT. However, we should notice that the LB circulation is less intense than SB since it opposes the mean flow direction. The LB was weaker for the dry season than the wet season; the average wind speed for this circulation was 1.0 m s −1 .
We also observed that the horizontal wind speed increases during the day (Fig. 2B), with the occurrence of thermal turbulence associated to the SB, and also because this flow is located in the same quadrant of the trade winds, thus increasing the wind speed due to the superposition of the SB with the mean flow. We verified that increasing the horizontal wind speed from 1.0 ms −1 (Fig. 2B) promotes a decrease in CO 2 concentration of 500 parts per million/volume (ppmv) to 350 ppmv; these results concur with the findings by Silva Junior et al. (2004) for diurnal CO 2 concentration in a pasture area in the Amazon forest.
Studies of global, liquid radiation, and energy fluxes for the wet and dry seasons
The general observations of the radiation balance data on the mangrove ecosystem showed that the highest magnitudes for Rg, Rn, and H occurred in the dry season, while the values for the LE were higher in the wet season.
Rg's data show that at 1:00 pm, we observed the highest magnitudes of the net shortwave radiation, with 673.8 W m −2 ± 24.15 W m −2 in the wet season and 792.2 ± 10.7 W m −2 in the dry season (Fig. 3A, B). For the Rn, we observed at 1:00 pm, for both wet and dry seasons, the highest magnitudes, with 469 ± 18.2 W m −2 and 572.2 ± 7.9 W m −2 , respectively.
When we have the maximum shortwave and diffuse radiation (2:00 pm) from the dry to the wet season, the radiative difference is 148.4 W m −2 . In comparison, the net radiation at 1:00 pm is 103.9 W m −2 , confirming that the seasonality in the region is very pronounced, characterized by the decrease of cloudiness, due to the displacement of the ITCZ, in the months of the dry season, allowing a more significant entrance of radiation to the surface, favoring greater energy availability for the physical processes in the lower troposphere. Fernandes et al. (2018), studying the climatic configuration through the radiation balance and transmissivity indexes (Kt) of shortwave radiation in the mangrove ecosystem in Cuiarana, verified that for February, March, and April, the Kt assumes the lowest values of the year due to the higher reflectance of the radiation in this season, with a consequent decrease of the energy availability in the forest canopy.
The analysis of energy partition between latent heat and sensible heat for the two seasons shows that the LE predominates in energy processes much longer throughout the day (at night and dawn) and that the magnitude of H is more significant than LE in the wet season for 5 h, between 11:00 am and 4:00 pm, where the maximum values for H and LE occurred at 2:00 pm with 187.7 W m −2 ± 3.8 and 171.2 W m −2 ± 4, respectively. However, this observed feature in the wet season is marked by a very close partition between H and LE during the day, ratifying the water availability in the system in the form of moisture, rainfall, and tidal pulses, making possible the efficiency of the physiological phenomena of the mangrove, mainly in the gas exchanges between water vapor and atmospheric CO 2 , to increase biomass through photosynthesis (Fig. 3C).
From June to December, between 9:00 am and 5:00 pm, the energy partition reveals that the magnitude of H in relation to the LE prevails for 8 h in this season, with a maximum value at 1:00 pm with 241 W m −2 ± 4.1. This environmental behavior of the ecosystem is expected due to the decrease in cloud cover, with a consequent reduction in rainfall volumes, as well as the decrease of air humidity in the second half of the year, where the highest LE value was also observed at 1:00 pm with 139.5 W m −2 ± 3.4 (Fig. 3B). Pereira and Rodrigues (2013), analyzing the energy partition in the mangrove ecosystem on the coast of Pará, found that in the wet season, the maximum value of H occurred in January at 1:00 pm with 274.7 W m −2 . On the other hand, the maximum LE observed occurred in May at 11:00 am with 374.7 W m −2 . During the dry season, H's maximum value occurred in November at 11:30 am with 300.2 W m −2 ; the LE had its maximum value in June at noon with 267.5 W m −2 (Fig. 3B).
We should notice that the data recorded by these authors with their maximum values higher than those observed in Cuiarana and differences in the hours occurred due to the location of the meteorological tower. The meteorological tower in Bragança was installed in a deforested area, thus justifying a higher reflectance of the soil with few vegetal cover.
Studies of Barr et al. (2014) about seasonal mangrove evapotranspiration revealed that diurnal and seasonal controls of water vapor fluxes evidenced that the energy partition between H and LE was highly variable. The forest behaved as a semi-arid ecosystem during the dry season, with most energy partitions converted into H, and minimum LE values with 5 MJ d −1 . In contrast, during the wet season, the forest presented LE fluxes with 18 MJ d −1 .
The dry season, high salinity levels were observed, influencing the reduction of evapotranspiration and reducing the stomatal canopy conductance. These authors verified that the canopy's daily conductance to the water vapor decreased with increased salinity from the multiple linear regression analysis.
Study of incident radiation (K in ), fluxes of H, LE, and G and energy balance
Data of the incident radiation (K in ) represented by the Rn/K in relation for the days and hours studied showed that both K in and the net radiation in the ecosystem presented the seasonal and intraseasonal patterns for eastern Amazonia. This energy availability governs the physical and biological processes in the mangrove and the behavior of the other elements of the energy balance (Table 1).
The energy partition analysis in the Cuiarana mangrove showed that the energy balance closure through the ratio (H + LE)/(Rn−G) had a deficit of ~ 30% in the wet ~ 40% in the dry season. We should notice a predominance of LE concerning H during the months of the wet season. The main reasons for this are the precipitation of the first half of the year, the high humidity of the air, and tidal pulses, thus allowing a more significant amount of water vapor to the soil-biosphere-atmosphere system.
The evaporative fraction (LE/Rn) presented better results in the wet season months, with values similar to (H/Rn). (Table 1).
We noticed the dry season data, the seasonality's characteristic effects, with a predominance of H compared to LE, with an average value of 239 W m −2 for H between the months in the second half of the year. The effect of seasonality on energy partition is due to the region's hydrological and energetic cycle's different behavior, reflecting the decrease in LE/Rn values to ~ 20% and H/Rn elevation ~ 40% (Table 1).
The soil heat flux (G) accompanies the seasonality of energy availability in the ecosystem, with mean values of 11.1 W m −2 between March and June and 19.5 W m −2 between August and November. The lowest/highest energy availability in the wet/dry seasons reflects the region's largescale weather systems (Table 1).
The energy balance closure presented showed a slope of 0.68 and a determination coefficient of 0.91 for the wet season. For the dry season, the line's slope was 0.63, and the determination coefficient showed values of 0.92. The slope found is consistent with the values found in studies on mangrove ecosystems that used the EC technique, citing 0.82 in Barr et al. (2013).
Seasonal study of the effect of the tidal cycle on the mangrove energy balance
The energy partition during the wet and dry seasons, under the influence of the tidal cycle, showed that there was a predominance of H concerning the LE during the occurrence of low tides in both seasons (H= 203.3 ± 3.1 W m −2 ; LE=193.3 W m −2 /0.25m and H= 241.5 ± 2.7 W m −2 ; LE= 119.4 W m −2 /0.3m) and predominance of LE concerning H only during high tides (LE= 202.1±2.2 W/m −2 ; H= 172.1 W m −2 / 4.3 m) from the wet season. In the dry season, the LE was lower than H (LE= 125±1.7 W m −2 ; H= 232 W m −2 /4.1 m) ( Table 2).
We verified that the tidal cycle in the region influenced the energy partition, with the region's seasonal characteristics. In the first half of the year, we noticed that the high cloud cover in the local atmosphere, coinciding with the high tides, favored the predominance of LE, while in the second half of the year, with lower rainfall availability and higher incident radiation supply, there was a consequent predominance of H concerning other energy flows, both in the high and low tides.
Regarding the soil heat flux (G), we observed that during the wet season's low tides, the highest magnitudes of energy occurred with 19.7 W m −2 , equivalent to 4.7% of the Rn. During the dry season's low tides, the average G fluxes were around 9 W m −2 , which corresponds to 3% of the Rn.
The analysis of the energy balance closure for the high/low tide hours in both seasons showed that the better energy balance closure occurred during the high tide of April, with 0.74, evidencing that the energy transfer associated with tidal activity improved the energy balance of the flooded surface during the observation season, even though the wind speed was 2.7 m s −1 . Analyses by Barr et al. (2013), which quantified the energy transport, during the high tides (and low tides) in a mangrove in Southern Florida, by summing the enthalpy change (ΔH tot = total enthalpy variation (ΔH tot = ΔH stor + ΔH adv )) of heat stored (or released) in the water column (ΔH stor ) and the heat exchange advected by the tidewater (ΔH adv ), verified that when ΔH tot was included in the energy balance, at the times of the day with the greatest irradiance, the energy balance closure improved from 0.73 to 0.75. Therefore, these authors verified that the tidal waters attribute an energetic computation improving the microclimatic conditions within the forest canopy through these observations.
Regarding the source/sink function of CO 2 , under tidal effects, we noticed that when there was a supply of water vapor in the low troposphere, during the high tides, especially during the wet season, carbon assimilation was accentuated (12.3 μ mol (CO 2 ) m −2 s −1 ) by the stomatal exchange in the forest canopy. During the low tide in April, the average carbon assimilation was 16% smaller (10.3 μ mol (CO 2 ) m −2 s −1 ). In October, carbon assimilation was 9.3 μ mol (CO 2 ) m −2 s −1 on the days of high tide, while during the low tides, the forest absorbed 9.0 μ mol (CO 2 ) m −2 s −1 , corroborating the fundamental role of water in the stomatal exchange and improvement of carbon assimilation, mainly within the local seasonal characteristics.
Study of the diurnal carbon cycle
The CO 2 diurnal cycle during 2015 under the effects of ENSO showed that carbon sequestration in both the wet and dry season begins around 7:00 am when the solar zenith angle initiates the process of photosynthesis in the forst. A reduction of the mangrove's photosynthetic physiological activity occurs around 5:30 pm in the wet season and 6:00 pm in the dry season (4A).
From 6:00 pm, the ecosystem is performing only autotrophic and heterotrophic breathing. We should notice that along the year in different hours when PAR occurred between 1.9 μ mol (photons) m −2 s −1 and 3.1 μ mol (photons) m −2 s −1 , there was a predominance of photosynthesis concerning the respiration of the ecosystem (Fig. 4C).
We observed that photosynthesis was predominant in the wet season compared to the dry season. The precipitation in the first half of 2015 was associated with lowering horizontal wind speed, and a decrease in soil salinity, which favored the maximum photosynthetic activity, between noon with −12.9 μ mol (CO 2 ) m −2 s −1 and 1:00 pm −13.5 μ mol (CO 2 ) m −2 s −1 (Fig. 4C), where the PAR showed maximums of 2.727 μ mol (photons) m −2 s −1 in the wet season and 3.406 μ mol (photons) m −2 s −1 in the dry season. These results agree with observations made by Barr et al. (2014) and Leopold et al. (2016), associated with the mangrove phenology in delaying foliar abscission, as described by Nascimento et al. (2006). Barr et al. (2010), studying the net CO 2 exchange between the mangrove ecosystem and the coastal atmosphere of FL, USA using the EC technique, found that maximum daytime CO 2 absorption ranged from −20 to −25 μ mol (CO 2 ) m −2 Fig. 4 Diurnal cycle of the carbon flux for the months in the wet and dry season in 2015 at the Cuiarana experimental site, Pará s −1 between March and May (Spring). During the winter, the mangrove's CO 2 assimilation increased by the same proportion as the diffuse solar radiation due to the radiative transfer in the forest canopy.
The analysis of the data of the respiration of the ecosystem showed that the magnitudes of this physiological phenomenon were higher in the wet season, with peaks of 10 μ mol (CO 2 ) m −2 s −1 at 4:00 am, while in the dry season, we registered 2.8 (CO 2 ) m −2 s −1 at 5:00 am. The physiology of respiration was expected to have higher magnitudes in the second half of the year, due to a decrease in cloudiness and consequent increase in radiation availability, as observed by Barr et al. (2010).
However, we do not observe this in the ecosystem on the coast of Pará, by the seasonality of the region marked by the reduction of precipitation and elevation of salinity of the water of the mangrove soil from 32.3 ppt in the wet season and 70 ppt in the dry season, allowing the species Avicennia germinans and Rhizophora mangle to show more significant adaptation and dominance in this hyper hyaline ecosystem, as shown by studies of Menezes et al. (2008), Laurance et al., (2004), Do Nascimento et al. (2006, and ratified by Fernandes (2016).
Studies by Barr et al. (2014) on soil water salinity's seasonality to support seasonal mangrove evapotranspiration in Southern California found that the salinity varied the dry season rapidly at 35 ppt, to values between 26 and 17 ppt at the beginning of the wet season. The increase of precipitation in June and July and the increase of river discharges contributed significantly to the salt dilution of the soil's micro and macropores.
Observations by Leopold et al. (2016), in the "Coeur de Voh" mangrove in New Caledonia, also verified that the salinity of the water presented a high seasonal variation. The salinity of water from the soil interstice was much lower during the wet season than the dry season. These authors verified that the decrease of salinity was related to the system's freshwater intake during the wet season. It was also observed that the high volumes of precipitation increased the discharge of the river and consequent contribution to the tides flood with high freshwater content, which may have diluted the concentration of the saline solution from the soil interstices.
Seasonal remote sensing data for the Cuiarana mangrove for 2015 showed a variation of the vegetation cover through the NDVI analysis of the forest canopy between the wet and dry season values of 0.75 ± 0.05 and 0.69 ± 0.04, respectively. We noticed a variation in the forest canopy from the chlorophyll infrared reflectance, suggesting a decrease in the leaf area index. Therefore, with the beginning of the dry season, the genus Avicennia and Rhizophora assume, as an environmental adaptation strategy, the foliar abscission, reducing water vapor by leaf area, and acquiring the lowest values source/ sink function of carbon in this season. Cunha et al. (2006); Fernandes et al. (2007);Menezes et al. (2008) attribute to this phenological behavior of Avicennia and Rhizophora the striking feature of salinity seasonality and tidal levels. However, observations by Cunha et al. (2006), studying the frequency and structure of three mangrove species in Southeastern Brazil, and the observations through the EC technique for Cuiarana, showed that other meteorological elements with a synoptic scale influence the adaptive behavior of these genus in the mangrove, such as the meridional shift of ITCZ to its northernmost position, reducing the cloudiness in the region, with consequent reduction of precipitation, thus influencing the energy partition and LE reduction for the second half of the year.
In this way, with less water in the system, to make possible the stomatal changes in the most demanding season, with high air temperature, reduction of the relative humidity, and higher incident radiation, the species, in order to avoid water stress, promote foliar abscission in this season, corroborating the observations of Fernandes et al. (2007), in the Bragança mangrove that described the seasonality of foliar abscission of the genus Avicennia and Rhizophora within a same seasonal pattern that has been observed in Cuiarana through the CO 2 flux.
Conclusions
For 2015, under ENSO in the mangrove ecosystem of the experimental site of Cuiarana, on the northeastern coast of the state of Pará, we verified that the precipitation had volumes below the climatological normal, evidencing the strong influence of the sea surface temperature anomalies in the Pacific Ocean over the Amazon. Therefore, the energy partition was affected by this precipitation anomaly, inducing a predominance of LE at various times of the day, influenced by the small precipitation volumes. In the dry season, the H, temperature, and relative humidity followed the climatological normal expected for the region.
Regarding the carbon sink/source function performed by the mangrove forest, where there is a great abundance of the species Avicennia germinans and Rhizophora mangle, we verified that the largest magnitudes of photosynthesis and respiration occurred significantly during the wet season, since a series of physical factors such as lower soil salinity, high precipitation, cloudiness attenuating the radiation, and biotic factors such as transpiration through stomatal activity, favored the physiology of the ecosystem. However, there is an inversion of these physical patterns in the dry season, accompanied by foliar abscission and decrease in the mangrove tree species' physiological functions.
The tidal cycle effect was analyzed both in the partition of energy and in the source/sink function of carbon. We observed a favorable condition for carbon absorption on high tide days, both during the wet and dry seasons. The tidal cycle also affected the energy balance closure; we verified a rise in energetic computation during high tides hours, improving the energy balance closure.
|
2021-12-01T15:23:32.120Z
|
2021-12-01T00:00:00.000
|
{
"year": 2021,
"sha1": "a641e9c46da2fffddce6c4743547578ff9707e0a",
"oa_license": "CCBY",
"oa_url": "https://link.springer.com/content/pdf/10.1007/s00484-021-02224-8.pdf",
"oa_status": "HYBRID",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "a641e9c46da2fffddce6c4743547578ff9707e0a",
"s2fieldsofstudy": [
"Environmental Science"
],
"extfieldsofstudy": [
"Medicine"
]
}
|
246551870
|
pes2o/s2orc
|
v3-fos-license
|
A nonlinear radio-photon conversion device
: The article analyzes existing materials and structures with quadratic-nonlinear optical properties that can be used to generate a difference frequency in the terahertz and sub-terahertz frequency ranges. The principle of constructing a nonlinear optical-radio converter, based on an optical focon (a focusing cone), is proposed. Based on the assumption that this focon can be imple-mented from the metal-organic framework (MOF), we propose a technique for modeling its parameters. The mathematical model of the process of propagation and nonlinear interaction of waves inside the focon is based on a simplification of the nonlinear wave equation. Within the framework of the developed model, the following parameters are approximately determined: the 3D gradient of the linear refractive index and the function determining the geometric profile of the focon, which provide a few-mode-based generation of the difference frequency.
Introduction
One of the main strategic tasks for the development and implementation of new generation communication networks is the development of new information technologies that allow increasing the bit-rate of data transmission (up to 1 Tbit/s) and the spectral efficiency (60 bit/s/Hz) of the used radio frequency channels, as well as the use of ultrawideband communication channels (up to 100 GHz), high energy efficiency with ultrashort network propagation delays up to 1 µs. The solution of the set tasks can be achieved through the research and development of data processing technology based on new principles of access to the physical layer of the network, new modulation schemes and channel coding, new waveforms, improving the functioning of diversity reception technology, multiple access and security both in wireless and optical communication channels. Thus, according to the various 6G white papers, which regulate possible ways to implement broadband access in 6G wireless networks, it is assumed that sub-terahertz radio frequencies will be useful for the Internet of Things (IoT) and the Industrial Internet (IIoT). In these tasks, it is required to ensure the operation of non-extended radio channels, but with a significant information capacity not only between subscribers, a subscriber and a device, but also between smart devices.
Such tasks require both the use of new modulation formats and new physical layer technologies, which ensure, in particular, the "seamlessness" (absence of intermediate operations) of optical-radio conversion which is especially relevant in sub THz, THz systems where it is very difficult to generate high-frequency radio signals by electronic devices, and it also seems inefficient to decode a multichannel signal for subsequent conversion of its components [1][2][3][4]. Therefore, to ensure such seamlessness, in particular, for opticalradio conversion in two-frequency (or wavelengths λ: λ1, λ2) radio over fiber (RoF) networks transmitting vortex signals (carrying orbital angular momentum, OAM), a nonlinear converter was developed [1]. But according to the results of the simulation of the converter's parameters [1] and experimental results (see Section 2), the efficacy of this converter is very low. According to our study, this is due to: a) a very small coefficient of quadratic nonlinearity (2) of the ferroelectric crystal used in [1], and b) due to radiation losses in the specified crystal because of divergence of the input light beam, which is associated with the absence of any wave-guiding structures in the scheme proposed in [1]. The negative impact of these factors can be significantly reduced if the material and structure for the nonlinear element are properly selected. Thus, the purpose of this article is to determine the structure and the material for this type of frequency conversion.
Analysis of optical quadratic-nonlinear materials and structures potentially suitable for the design of a nonlinear converter
It was shown in [1] that during the nonlinear interaction of two input light beams in a quadratic medium, for example, in a lithium niobate (LiNbO3) crystal, an electromagnetic beam with a frequency equal to a difference between the frequencies of the input signals is generated (difference frequency generation, DFG). In this case, several physical properties of the input beams are also transferred to the output beam (conserved), namely: for OAM carrying input beams with topological charges ℓ1 and ℓ2, respectively, at the output we obtain OAM signal with ℓ3 which is equal to the difference between the input OAMs: ℓ3 = |ℓ1 -ℓ2| [1]. This fact arouses interest in the further development of the seamless principle of optical-radio conversion by a nonlinear element, presented in [1].
However, numerical estimates in [1] showed that the generation efficiency η of a vortex (OAM) radio beam for a nonlinear medium with a length of several millimeters is extremely low (about 10 -16 ). This efficiency significantly depends on the frequency ratio between the generated (difference) frequency 3 and input frequency 1, i.e.: where 3 = |1 -2|. Some causes of this were emphasized in the introduction: a small value of (2) and the absence of a guiding structure. If we look to the results of other authors obtained in a similar problem, we can see that this scheme is actively studied and successfully tested experimentally. Thus, in [5], the conversion of an OAM signal from the optical to the terahertz spectral region was experimentally studied using a quadratic-nonlinear element. The relations for the output signal parameters, in particular, the OAM topological charge (ℓTHz) and the conversion efficiency η DFG , are also obtained there, see (2) and (3). 21 12 21 λλ where OPA corresponds to optical signal. In this case, the conversion efficiency, according to the authors of [5], significantly depends on the beam interaction region, but the dependence on the input and output frequencies ratio, see (1), is not taken into account. The authors of [6,7] also assume that a quadratic ferroelectric crystal is quite applicable for the process of parametric frequency generation: terahertz difference frequency, half-frequency, as well as double frequency and sum-frequency. In this case, the efficiency of the output beam depends on the parameters of the system and is proportional to the propagation length z: ( ) 42 2 2 2 2 2 2 1 2 1 ωπ β sin 2 4 2 The authors of [8] also present the results of experimental approbation of the considered principle of frequency transformation. According to the studies carried out, it was found in [8] that a quadratic-nonlinear crystal can indeed be used for the considered task, however, the conversion efficiency is also low -it did not exceed 0.1% of the input radiation intensity.
Having analyzed the published works in the area under consideration [5][6][7][8], our scientific team has also performed an experimental study of this principle of generation using a periodically poled lithium niobate crystal -PPLN [3,4] and based on the model presented in [1], fig. 1. In this setup we used a coherent laser which signal is split into two paths and fed to two branches of free-space setup ( fig. 1, b); one of the free-space branches has a diffractive optical element (DOE), forming an OAM signal; both optical signals a coaxially focused on the heated PPLN and the radio signal from its output is detected in the radio domain. It was possible to establish that the quadratic-nonlinear element used actually provides the generation of a difference frequency even when the phase matching of the waves is performed only approximately -by PPLN structure. However, the intensity of the output radio signal remains at the noise level, which significantly limits the practical usability of the considered principle. We should note that in all analyzed papers on this topic, the authors make the same conclusion: the efficiency of the difference frequency generation using this type of crystal is unacceptably low. Our theoretical approach is described in details in [1]; in any case, the results of our experiment are not the subject of this article. a b c Figure 1. Photos of the fastening system (a-c) of a ferroelectric lithium niobate crystal on an experimental setup using a specialized heating interface and free-space setup (b), designed to increase the efficiency of the crystal (the heating driver is not shown; heater (a) is open); c) measurement setup including both optical and radio spectrum analyzers and a light source (laser).
At the same time, the physical reasons for this low efficiency, generally speaking, have not been studied in detail, but only assumptions have been put forward, which, moreover, differ from work to work. The general assumptions of the authors who published the results of independent studies are as follows: firstly, existing ferroelectric crystals have a small value of (2) , secondly, optical beams arriving at the input end of the crystal have a divergence, inverse proportional to the radiation wavelength. This leads to significant losses, moreover, one can list additional factors such as 1) the absence of a resonant propagation condition for the generated radio frequency; 2) violation of the few-mode propagation regime for the input optical radiation inside the crystal (this is in addition to the divergence). If the high-order modes are excited, the mode purity is violated and therefore efficacy decreases.
Indeed, ferroelectric crystals, such as LiNbO3, LiTaO3, KNbO3, KTaO3, PbTiO3, and the like, have such advantages as [9]: 1) low dependence of light transmission on frequency in the visible and near-IR radiation; 2) low cost and high availability. But at the same time, they have the following disadvantages: 1) the value of the nonlinear susceptibility (2) , defining the efficiency of nonlinear processes, is in the range (0.5…45)10 -12 m/V, i.e. is very small in comparison with other quadratic non-linear materials that exist today; 2) these materials can't be made in the form of a waveguide or fiber-optic structure, which leads to beam divergence, violation of single-mode regime and resonant propagation modes.
The small value of (2) in ferroelectric crystals is conditioned by the physical nature of nonlinearity -it is associated with electronic polarization, and, accordingly, with the so-called electronic nonlinearity [10,11]. This mechanism is caused by the distortion of the outer electron cloud of atoms, ions, molecules compared to their unperturbed state, and has a very fast response time (<10 -15 s), but low (2) . Although in [12], it is shown that the listed mechanism of nonlinearity can be enhanced by spatial redistribution of electrons due to a spatially modulated light flux, which is the characteristic of spatially resonant structures, such as photonic crystal fibers. However, the materials listed above have low values of (2) , which means that they will not allow one to obtain an efficient frequency transformation.
Regarding the materials with higher nonlinearity coefficients, one can mention organic materials [13], in which the value of (2) can be in the range 10 -6 ...10 -7 m/V, which is relatively large. However, the mechanism for achieving such values of (2) refers to molecular reorientation with a very big establishment time about ~10 -10 up to 10 -3 s, i.e. the greater the value of (2) can be reached, the longer one has to wait until the material enters the non-linear regime, i.e. the longer the molecule, the greater the nonlinearity achieved, but also the longer this molecule turns. Therefore, the authors of studies in this area focus not only on the values of the coefficients obtained but also on the dynamics of the observed processes. In addition, this class of materials exhibits a significant dependence of light transmission on its physical state and condition, for example, on humidity. There is also a significant oscillating dependence on frequency in the visible and near-IR radiation.
Nevertheless, in [13] it is said that these materials can be used to generate, for example, a difference frequency in the case when this is not associated with the signal modulation but is necessary only to obtain the indicated radio emission (in other words, when one wants to generate radio carrier frequency). The authors of [13] also show that (2) depends significantly on the ratio of the molecular axis and the electric intensity vector of the input light field, and can be found through relation (5).
From the point of view of the problem under consideration, in addition to the long time of the nonlinear regime establishment, a significant drawback of these materials is also the impossibility of their implementation in the form of extended waveguides or fiber optic structures.
An interesting subclass of the considered category of materials is the so-called metalorganic frameworks (MOF) [14][15][16][17]. In MOFs not only the organic filler is responsible for the macroscopic properties (see [13]), but also the metal structure, which, besides the structure strength (which was laid in the original idea of developing the MOF), can also form a field potential when irradiated with an electromagnetic field of the corresponding frequency range [14,18]. Strictly speaking, as we know, there are no results of experimental studies on the effect of the metal frame that makes up the MOF on the possible generation of plasmon field effects, which affects the macroscopic optical properties of the material. It should be noted that the nonlinear properties of MOF structures are studied quite intensively in both theoretical [19] and experimental [20] ways. For example, in [21], research is primarily focused on the design of a molecule that provides the required nonlinear properties.
According to publications [15,17], MOFs can be used to form an optical quadratic nonlinearity in problems of generating, for example, the second harmonic: The main properties of MOFs from the point of view of the problem under consideration include: 1) low dependence of light transmission on frequency in the visible and near IR region, 2) the (2) is in the range 10 -3 ...10 -9 m/V, which is by several orders of magnitude higher than in ferroelectric crystals, 3) non-linear regime establishment time is ~10 -14 up to 10 -3 s (the spread is very significant because the time depends significantly on the influence of the metal frame), 4) the properties of MOF can be controlled by an external field/potential, 5) MOFs cannot be made in the form of extended guides/fiber optic structures.
In the framework of the considered problem, MOF under certain conditions can be used as the basis of a quadratic-nonlinear element. Namely, in the case of a decrease in the nonlinear regime establishing time (to values of 10 -14 -10 -13 s) due to the use, for example, of plasmon effects.
Another option for constructing a quadratic-nonlinear element with properties suitable for the problem being solved is the use of a photonic-crystal fiber (PCF). It should be mentioned that PCFs have the widest range of macroscopic optical properties: in some cases, their nonlinearity is reduced to zero due to the propagation of the light field in areas filled with air [12]; in other cases, they obtain a significant macroscopic nonlinearity [12,[22][23][24][25][26][27] achieved, for example, by channeling the light field in regions of ultra-small diameter, which provides a significant increase in the intensity and a corresponding increase in the effective nonlinear response. As an example, this option is used for spectrum broadening by the interaction of nonlinearity on short pulses (with durations ~10 -12 s).
The results concerning the quadratic nonlinearity in PCF have been published widely [23][24][25][26][27]. So, in [25] it is said that (2) can be in the range 10 -3 ...10 -10 m/V, which many times exceeds the corresponding values for ferroelectric crystals. There are additional mechanisms for obtaining nonlinearity [23,24] associated with injected carriers, which are put into operation with a nonlinear shift of the photonic bandgap [23]. At the same time, this mechanism is inherently akin to electronic nonlinearity, which means that it is characterized by a short establishment time. Therefore, in quadratic-nonlinear PCFs constructed in the above manner, one can expect significant (2) along with short entry times into the specified regime. As we know, there is no published data on the dynamic parameters of the nonlinear regime of PCF. It is also obvious that PCFs are waveguiding structures, which is undoubted of interest for the problem being solved.
It is mentioned in [26,27] that quadratically nonlinear PCFs can be successfully used to generate the second harmonic and difference frequency [28], including the THz range [29]. But if a high nonlinearity is achieved by spreading the input light beam over highintensity narrow light "filaments", then this will actually lead to multipole [30] generation of the radio frequency wave and, quite possibly, to distortion of the information signal carried by it. Therefore, PCFs can potentially be used for the problem of square-nonlinear generation of the RF difference frequency, but the considerations presented above will require optimization of the PCF structure to reduce the distortion of the generated signal, as shown in the above-mentioned papers.
Thus, according to the above analysis of materials and structures with quadraticnonlinear optical properties, it can be concluded that MOFs and/or PCFs can be used for the problem being solved. In the first case, i.e. for MOF, one can either use the parameters of some known structure (material) to develop a non-linear element for DFG or study the effect of a metal frame and thereby optimize MOF parameters. It is quite possible that in the latter case, in order to form the desired optical properties (e.g., to control the time of nonlinear regime establishment) it may be necessary to use additional radiation (λpump), fig. 2, which shifts the operating point of the non-linear regime and thereby catalyzes the process. In the second case, i.e. for PCFs, it is necessary to take into account the properties of the photonic-crystal structure that determine the mechanism of nonlinearity and, accordingly, the mechanism for generating a radio wave, and generally speaking, to optimize the PCF's nonlinearity for the problem under consideration.
Proposed conversion device and formulation of the problem of modeling its parameters
As can be seen from the analysis above, the simplest solution is the use of MOF, as presented in [17]. In this regard, to minimize losses and increase the efficiency of a radiophotonic conversion device designed to generate a radio frequency, it is necessary to make a nonlinear element in the form of a guiding structure that retains resonant properties for both input optical and output radio radiation. This can be done by changing the diameter of said guide structure D(z) along the propagation axis z as in horn antenna or focon. In this case, for the input optical radiation (λ1, λ2), it is desirable to maintain a (quasi)singlemode propagation over the entire length L of the nonlinear element. Or, in any case, to optimize the parameters of the nonlinear element so that the single-mode propagation mode for the input radiation is maintained as long as possible. This can be achieved by changing the radial profile of the refractive index of the material along the length of the nonlinear element, i.e. providing the corresponding refractive index profile n(z,r). Note that by "(quasi)single-mode" we mean the OAM mode (within the scalar approximation it corresponds to a set of linear polarized (LP) modes containing angular exponent exp(iℓφ), and each of these modes is a superposition of degenerated transverse modes [31]), so generally speaking, we are talking about few-mode regime.
Thus, the proposed radio-photonic conversion device should have the following structure, fig. 2. Optical radiation is supplied to its input at two close wavelengths, the difference of which corresponds to the wavelength of the generated radio wave. Both input optical wavelengths may be fed through a single input optical fiber. One should use a few-mode fiber (FMF) if the optical wavelengths lie in C-band. In the case when it is necessary to use additional (corrective) radiation (λpump), the input optical channel can be implemented, for example, by a two-core concentric optical fiber, fig. 2. The output part of this radio photonic device must have a substantially larger diameter than the input part, which is necessary to ensure effective resonant properties for the generated radio wave. Since the relationships that interconnect the geometric parameters of the guiding structure with the mode wavelength are linear, we can estimate: where is the wavelength of the generated radio wave; λ is one of the input optical wavelengths. Therefore, we can consider the nonlinear element as a focon, whose geometric profile D(z) can either be a continuously increasing function or have a minimum (as shown in fig. 2) in the case of a significant inlet diameter. Such a focon can be made by MOF. The design of the focon parameters is reduced to finding D(z) and n(z,r) with known λ1, λ2, (2) , Din, Dout, and focon length L provided that the amplitude of the output radio wave A is above a given threshold A * : A A * . An additional physical condition that ensures A A * , as mentioned above, is the requirement to maintain a quasi-single-mode regime for input radiation at length H, which in the limit H → L. Note that if the focon is made of the PCF, the parameters of the PCF structure should also be included in initial design parameters since the PCF's nonlinearity cannot be replaced by the averaged (2) .
Figure 2.
An illustration of the structure of the proposed radio-photonic conversion device that provides the generation of differential radio signal when optical radiation is supplied to the input at two close wavelengths.
Based on the above, the problem is posed of modeling the magnitude of the electric field strength for the input (with indices 1 and 2) and output (3) waves to determine the spatial distribution of the refractive index of the focon and its geometric profile. We introduce the following parameters: -for the focon profile one can write: + = = (2) can be considered homogeneous over the focon volume, since it changes slightly in the presence of a linear refractive index gradient. This assumption is due to the fact that the effect of the gradient of the linear refractive index on the phase of the signal during the interference process is much more significant than the effect of a change on the nonlinear process (see below). In other words, when the linear part of the refractive index changes, its nonlinear part undergoes proportional changes, which is, however, very insignificant. We also neglect the dispersion of (2) for the considered waves according to [36].
-refractive index outside the focon = cl ( , ) n r z n for 2 1 () r f z , where we assume that ncl = 1.46 and index "cl" stands for "cladding".
Modeling the parameters of the proposed device
The propagation of an electromagnetic wave in a nonlinear medium is described by the following equation: where ( , , , ) E x y z t is the vector of the electric field strength, c is the speed of light in vacuum, and ε is the permittivity of the medium, and ( , , , ) NL P x y z t is the nonlinear polarization vector.
By analogy with [1] and using the technique described in [36], to obtain a solution (8), one should, firstly, use the separation of variables, and secondly, apply the Fourier transform of the given functions in the time domain. The latter allows in (8) to get away from time derivatives by performing the following replacement: d dt j → . So, the Fourier image of the solution (8) will be written in the form: where we omit the symbol ~ in the further notation for simplification. Now we change the Cartesian system to cylindrical coordinates (we consider the focon to be a body of rotation); ω is the radiation frequency, ω βε c = , (, ω) Az is a function describing the longitudinal (along the focon, fig. 2) dependence of E, and ( , ) Fr is the transverse component of the in E. In addition, in (8) and (9) the indices 1, 2 and 3 related to the input (1 and 2) waves and the output (3) wave, respectively, since the above expressions for all three waves appear to be identical.
Next, we find the derivative of (9) with respect to coordinates, and substitute the resulting relation into (8), taking into account the properties of the Fourier transform. Following [35], we neglect the second derivatives with respect to coordinates to obtain the final equation, since we assume that functions (, ω) Az and ( , ) Fr change much slower than β . iz e Now, if we divide β into components depending on the basic physical processes under consideration, we can obtain two equations of the form: where Φ β (ω) determines the additional phase incursion of the wave under the influence of the determining physical process. So, for example, if a light wave propagates in an extended optical path, then the determining physical process will be chromatic dispersion. In this article, we assume that there are no dispersion distortions during the propagation of waves inside the focon since we consider two close wavelengths. On the contrary, significant influence have quadratic-nonlinear change in the refractive index, characterized by (2) , when the waves propagate along the focon, and the interference of waves during transverse distribution, which ensures the addition of re-reflecting waves into a single mode. But the function F in (9) also has a z-axis dependence, moreover, as shown in [32][33][34], to obtain a proper solution one should divide the entire focon into transverse layers of width Δz where Δz →0, fig. 3. Within each layer, the "transverse" process can be considered independent of z, as it noted in the second equation in (10) and justified in [32][33][34]. Here index q is the counter of such layers. The dependence of F on z is ensured by the fact that the output of each q th layer is the input of each (q+1) th layer. In addition to previous assumptions, we also neglect the Rayleigh losses inside the focon since it has quite short length. The parameter β0 in (10) determines the phase incursion of a wave in a medium with an initially unperturbed refractive index (in the absence of nonlinear processes and any refractive index gradient), i.e. 00 ω β n c = . We make the following assumption: there is no re-reflection (and hence interference) between the layers in the longitudinal direction despite the difference in the refractive index between the layers. This is because the focon has a smooth gradient and does not consist of layers with a hop in the refractive index at the boundary. Moreover, when Δz→0, the increment of the refractive index between the layers is Δn→0.
Further, if we take into account that in a quadratic-nonlinear medium , and make mathematical transformations similar to the technique [36] for the function A, and represent the function F in the form of counterpropagating waves F (up) and F (down) relative to the focon axis, then for each of the waves (1, 2 and 3) the equations (10) will take a form that is suitable for numerical solution: (12) In equations (11) and (12) is the wavenumber of the q th layer, fig. 3, which has multipath interference properties. Each layer can be considered as a multibeam interferometer -a Fabry-Perot interferometer (FPI). But due to the fact that the medium is nonlinear, () q depends on how many re-reflections have already occurred in this q th layer; here l is the re-reflection counter, [1, ] l lK , Kl in the ideal FPI is equal to , but in this case it is determined by the number of effective rereflections and is calculated through the reflection coefficient of the focon profile. For each of the waves, we will assume that: where () β q MD is the initial (unperturbed) wave number of the q th FPI with the averaged refractive index over the radius. This leads to the fact that the calculation of ( ) , lq F is performed for a certain FPI refractive index averaged over the radial gradient, and thus allows one to obtain only an estimations, but, on the other hand, ensures avoidance of the recurrence. Moreover, we apply this approach at the first iteration of the calculation; when the main parameters have already been estimated, in particular, an approximate value ( , ) n r z has been obtained, then for the calculation of () β q MD in (13) it is possible to use the value ( , ) n r z found at the previous iteration. The parameter B in (13) determines the nonlinear contribution from the transmitted radiation, and is written by analogy with [36,37], ( . To calculate the coupling coefficient of interfering waves i , i = 1, 2, 3, similarly to [38], we will use the relation: according to which each of the waves interferes only with itself. In (14) the index i of the refractive index relates to the corresponding frequency (wavelength). The number of effective re-reflections in the FPI can be determined as [39,40]: is the reflection coefficient of the focon profile at arbitrary point z (due to reflection from the interface between media with different refractive indices): where the terms for each of the waves can be found using the properties of the FPI for re-reflecting (superposing) waves [38]: The parameter will be defined by analogy with (13): , q r f z q n r z q , and the initial amplitude value will be set for the first iteration as: ( ) ( ) ( ) = 0, 0, 1 4 q q q F F P , assuming that the refraction index is isotropic within the layer and therefore the power is divided equally, i.e. a half for each wave. In addition, for the components of the terms F under consideration, which characterize re-reflections from the focon profile in the transverse direction, the following one can use the following relations: The transformation (8)-(11) is generally known, so we considered it unnecessary to present a detailed derivation of equations (11) in the article. Let us pay attention to the fact that the left parts of each equation for the functions A in (11) do not depend on the angle φ, so the solutions of system (11) exist under the condition: = 0 . This suggests that in a quadratic nonlinear medium the input signals carrying OAMs ℓ1 and ℓ2 at frequencies ω1 and ω2 generate an output signal with OAM ℓ3=ℓ1-ℓ2 and frequency ω3=ω1-ω2, which ensures the seamlessness of optical-radio engineering conversion. Note that for the existence of OAM modes, as is well known [31], a few-mode regime (excitation of several transverse modes) is required. That's why we search ( , ) n r z and 1 () fz -to provide proper few-mode regime. Therefore, we reckon that the wave field (for each of the waves under consideration), determined by (12), should obey the following relation: where Jℓ is the Bessel function of the ℓ-kind, uℓm is the first ℓm th maximum of the Jℓ, ( ) q P is the power parameter, and * is the generalized mode propagation coefficient. The specified "quasi-single-mode" addition will ensure the absence of dark areas, which means the large volume of the interacting light, and with this the highest efficiency of the device (especially in case of low-order OAMs, e.g., ℓ1=1, ℓ2=0, etc.). Relation (18) is valid for E as a function of time, to which inverse Fourier transform (12) should be performed. Note that the stated condition can be simplified. Namely: the power parameter in (18) can be considered only as a multiplier i.e. as weight coefficient. In addition, the considered amplitude distribution has no dependence on z (except the phase exponent), but only on r,φ, so condition (18) applies only to function F. Therefore, we consider that relation (18), which is valid for ( ) q F as a function of time, complements relations (16). As a result of the numerical solution of equations (11), taking into account (13)- (18), iterative selection of the functions ( , ) n r z and 1 () fz was carried out, fig. 4. In calculations we assumed χ (2) = 10 -6 m/V. As we can see in fig. 4, the proposed converter structure can be easily coupled with few-mode fiber due to a comparative input diameter of a few microns.
Conclusions
In this article we provide an analysis of currently known materials and structures with quadratic-nonlinear optical properties. These materials and structures can be used in the process of non-linear parametric generation of the difference frequency in a twofrequency optical-radio engineering conversion.
|
2022-02-05T16:41:15.918Z
|
2022-02-01T00:00:00.000
|
{
"year": 2022,
"sha1": "0b1413f52da1282968651e7fdc5a3c2bc05ecb7a",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2304-6732/9/6/417/pdf?version=1655368456",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "b15a980d7ca2850a7566f6f8063a7069b51203fc",
"s2fieldsofstudy": [
"Physics",
"Materials Science"
],
"extfieldsofstudy": []
}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.